Citation
Chemical vapor deposition of thin films for diffusion barrier applications :

Material Information

Title:
Chemical vapor deposition of thin films for diffusion barrier applications :
Creator:
Bchir, Omar James
Publication Date:
Language:
English
Physical Description:
ix, 419 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Annealing ( jstor )
Atoms ( jstor )
Conceptual lattices ( jstor )
Electrical resistivity ( jstor )
Gibbs free energy ( jstor )
Grain boundaries ( jstor )
Lattice parameters ( jstor )
Polycrystals ( jstor )
Solvents ( jstor )
Vapor phases ( jstor )
Chemical Engineering thesis, Ph. D ( lcsh )
Dissertations, Academic -- Chemical Engineering -- UF ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2004.
Bibliography:
Includes bibliographical references.
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Omar James Bchir.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
003100349 ( ALEPH )
AA00004688_00001 ( sobekcm )
767969628 ( OCLC )

Downloads

This item has the following downloads:


Full Text










CHEMICAL VAPOR DEPOSITION OF THIN FILMS FOR DIFFUSION BARRIER
APPLICATIONS

















By

OMAR JAMES BCHIR


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2004




























I would like to dedicate this dissertation to my nephew, Matthew Edward Brown, and to
the memory of my grandparents.















ACKNOWLEDGMENTS

On a professional note, I would like to thank my supervisory committee

(especially my advisor, Dr. Tim Anderson, and Dr. McElwee-White) for their constant

guidance and assistance throughout my Ph.D. work. I would also like to thank the past

and present members of Dr. Anderson's, Dr. McElwee-White's and Dr. Norton's

research groups, with whom I have collaborated during my time at UF, and Dr. Jianyun

Shen, with whom I collaborated on thermodynamic modeling. I would also like to thank

Dr. Holloway for use of his 4-point probe device, and the staff at the Major Analytical

Instrumentation Center (MAIC), including Eric Lambers, Wayne Acree and Brad

Willenburg, for valuable assistance and training with various characterization techniques.

On a personal note, I would like to thank my parents (Sara and Hachemi Bchir)

for their encouragement, love and support throughout my life, and for their strong

emphasis on the importance of education. I would also like to thank my sister, Annissa

Brown, my brother-in-law Eddie, and my nephew Matthew, for their encouragement and

support. Thanks also go to my girlfriend Luciana Manfrim for her love and support.

Lastly, thanks go to all of the friends that I have made through the years while growing

up in Florida, attending Georgia Tech, working at Fluor Daniel and attending UF.














TABLE OF CONTENTS

Page
A CKN OW LED GM ENT S......................................................................................... iii

ABSTRACT .............................................. .................... .................................... viii

1 INTRODUCTION
1.1 Background ..................................................... .......................... 1
1.2 Comparison of Diffusivities for Al and Cu in Si ............................ 4
1.3 Diffusion Barrier Requirements and Properties............................ 6
1.3.1 Film Structure.......................... .................................... 9
1.3.2 Electrical Properties ......................................... .......... ... 11
1.3.3 Conform ality ................................................. .............. 14
1.3.4 A dhesion............................................ ............................ 16
1.4 Dielectric Material Considerations............................. ............ ... 17
1.5 Diffusion Barrier Failure Mechanisms.......................... .......... 17
1.6 Diffusion Barrier Deposition Techniques ....................................... 21
1.6.1 Physical Vapor Deposition (PVD) ................................... 21
1.6.2 Chemical Vapor Deposition............................................. 24
1.6.3 Atomic Layer Deposition (ALD) ......................................... 37
1.7 Copper Deposition Methods...................................... ............ .... 44
1.8 Statement of Problem ................................................ ........ ..... 49
1.9 H ypothesis...................................................... ........................... 51

2 REVIEW OF THE LITERATURE
2.1 Cu-Si Interconnects without a Barrier............................................ 53
2.2 Refractory Metal-Based Barriers .............................................. 54
2.2.1 Unary Refractory Metal Barriers...................................... 54
2.2.2 Binary Refractory Metal-Based Barriers.......................... 56
2.2.3 Ternary Refractory Metal-Based Barriers............................ 63
2.3 Justification for WNx (and WNxCy) as the Barrier Material............. 71
2.4 WNx Film Properties ............................................................... 72
2.5 WNxCy Film Properties ............................................... ........... 75
2.6 Amorphous WNx Film Deposition .................................................. 77
2.7 Demonstrated Uses of WN ........................................... ........... .. 78
2.8 WNx Deposition Techniques....................................... ............ .. 85
2.8.1 Annealing W in NH3 .............................................. ........... .. 85
2.8.2 Plasma Nitridation / Ion Implantation of W...................... 86
2.8.3 Pulsed Laser Deposition................................... ........... 86
2.8.4 Reactively Sputtered WNx Deposition.............................. 87









2.8.5 LPCVD WNx Deposition ............................................... 87
2.8.6 PECVD WNx Deposition ............................................... 91
2.8.7 MOCVD WNx Deposition ............................................... 93
2.8.8 ALD WNx Deposition .................................... ............. 97
2.8.9 ALD WNxCy Deposition ................................... .......... 99
2.9 Conclusions ..................................................................................... 100

3 EQUILIBRIUM MODELS FOR WNx-BASED BARRIERS
3.1 Motivation and Method..................................................................... 101
3.2 Constraints..... ....................................... 111
3.3 Degrees of Freedom ........................................................................ 111
3.4 Computational Methods .................................................................. 112
3.5 Phase Equilibrium in the W-N System .......................................... 114
3.6 Previous Studies of the W-N Phase Diagram................................. 125
3.7 W-N Optimization Results and Discussion.................................... 126
3.8 Stable Solid Phases in the W-C System......................................... 139
3.9 Previous Study of the W-C Phase Model....................................... 142
3.10 Equilibrium in the W-C-C1-H-N System....................................... 145
3.11 Optimization of the FCC P-WNxCy Gibbs Energy ........................ 153
3.12 Degrees of Freedom (DOF) Analysis.............................................. 168
3.12.1 Homogeneous Gas Phase Speciation .................................. 171
3.12.2 Heterogeneous Gas Phase Speciation ................................. 177
3.12.3 Heterogeneous Solid Phase Equilibrium............................. 182
3.13 Predicted W-C-N Ternary Phase Diagram.................................... 186
3.14 Predicted P-WCo.5 P-WNo.5 Pseudobinary Equilibrium............. 190

4 EXPERIMENTAL APPROACH FOR CVD
4.1 Substrate Preparation......................................................................... 194
4.2 Solvent Tests ................................................................................... 195
4.3 Description of CVD System Components ...................................... 196
4.4 Start-Up........ ....................................... 201
4.5 Copper D eposition............................................................................. 201
4.6 Analysis Techniques ....................................................................... 202
4.7 Precursor Screening Procedure ....................................................... 208

5 EVALUATION OF C14(CH3CN)WN-i-Pr AS A SUITABLE
WNx PRECURSOR
5.1 Synthesis of Isopropyl [C14(CH3CN)WN-i-Pr] Precursor............... 210
5.2 Solvent Selection............................................................................... 211
5.3 Precursor Mass Spectral Pre-Screen .............................................. 212
5.4 Film Structure ................................................................................... 217
5.4.1 X RD Results........... .............................................................. 217
5.4.2 TEM .. ................................................................................... 223
5.5 Film Com position.............................................................................. 225
5.5.1 A ES ..................................................................................... 225
5.5.2 AES Depth Profiling ........................................................... 229









5.5.3 XPS...................................................................................... 233
5.5.4 SIM S Depth Profiling............................................................ 253
5.6 Film Growth Rate (XSEM )............................................................. 255
5.7 Film Incubation Time and M orphology............................................ 259
5.8 Film Electrical Properties.................................................................. 264
5.9 Effect of NH3 and N2 Addition to Film Growth ............................. 267
5.9.1 Film Structure........................................................................ 267
5.9.2 Film Composition...... ... ............................................................ 270
5.9.3 Growth Rate ........................................................................ 277
5.9.4 Film Resistivity ................................................................... 279
5.9.5 Sheet Resistance.................................................................. 280
5.10 Conclusions for Use of NH3 and N2 with i-Pr................................ 281
5.11 Effect of Solvent Change on Carbon Content................................... 282
5.11.1 Film Structure........................................................................ 284
5.11.2 Film Composition...................................................................... 287
5.11.3 Carbon Deposition Rate ...................................................... 291
5.11.4 Conclusions On Effect of Solvent Change.......................... 297
5.12 Conformality Tests............................................................................ 298
5.13 Adhesion Tests .................................................................................. 298
5.14 Conclusions on Use of C14(CH3CN)WN-i-Pr to Deposit WNx....... 299

6 EVALUATION OF C14(PhCN)W(NPh) AS A SUITABLE WNx PRECURSOR
6.1 Film Growth Studies ....................................................................... 303
6.2 Synthesis of Phenyl [C14(PhCN)W(NPh)] Precursor ........................ 303
6.3 Solvent Selection............................................................................... 303
6.4 Precursor M ass Spectral Pre-Screen .............................................. 304
6.5 Film Structure ................................................................................... 308
6.5.1 XRD .. ................................................................................... 308
6.5.2 Lattice Parameter................................................................... 313
6.5.3 Polycrystal Grain Size......................................................... 315
6.6 Film Composition.............................................................................. 316
6.6.1 AES ..................................................................................... 316
6.6.2 Film Growth Rate (XSEM )................................................. 320
6.7 Film Electrical Properties.................................................................. 322
6.7.1 Film Resistivity ................................................................... 322
6.7.2 Film Sheet Resistance ......................................................... 323
6.8 Conclusions on Use of C14(PhCN)W(NPh) to Deposit WNx............ 324

7 EVALUATION OF C14(CH3CN)WN-C3H5 AS A SUITABLE
WNx PRECURSOR
7.1 Film Growth Studies ....................................................................... 331
7.2 Synthesis of Allyl [C14(CH3CN)WN-C3Hs] Precursor..................... 332
7.3 Solvent Selection............................................................................... 333
7.4 Precursor M ass Spectral Pre-Screen .............................................. 334
7.5 Film Structure ................................................................................... 336
7.5.1 XRD .. ................................................................................... 336









7.5.2 Lattice Parameter................................................................... 339
7.5.3 Polycrystal Grain Size......................................................... 341
7.6 Film Com position.............................................................................. 342
7.7 Film Growth Rate (XSEM)............................................................. 344
7.8 Film Electrical Properties................................................................ 345
7.8.1 Film Resistivity ................................................................... 345
7.8.2 Film Sheet Resistance ......................................................... 346
7.9 Conclusions on Use of C14(CH3CN)WN-C3H5 to Deposit WNx...... 347

8 COPPER TESTS ON BARRIER FILMS FROM C14(CH3CN)WN-i-Pr
8.1 Copper Deposition Results............................................................... 357
8.2 Electrical Measurements ................................................................. 357
8.3 Scotch Tape Tests............... ................................................................ 358
8.4 Barrier Integrity Tests ..................................................................... 358
8.4.1 X RD .. ................................................................................... 359
8.4.2 Electrical Measurements ..................................................... 364
8.4.3 Depth Profiling Analysis..................................................... 365
8.5 Conclusions Regarding Cu Testing of Deposited Barrier Layers..... 366

9 RECOMMENDATIONS FOR FUTURE WORK
9.1 Control Film Composition .............................................................. 368
9.2 Decrease Deposition Temperature .................................................. 371
9.3 Decrease Barrier Thickness............................................................... 371
9.4 Further Conformality Testing............................................................ 372
9.5 Further Adhesion Tests ................................................................... 372
9.6 Further Barrier Integrity Analyses .................................................. 373
9.7 Deposition on Alternate Substrate Materials .................................. 375
9.8 Testing of Films for X-ray Absorption Mask Applications.......... 376
9.9 Deposition of Alternate Barrier Materials....................................... 376
9.10 System Modifications........................................................................ 376

REFERENCES................................................................................ 378

APPENDIX

A CVD SYSTEM PROCESS FLOW DIAGRAMS ........................................ 409

B OTHER PRECURSORS TESTED............................................................. 415


BIOGRAPHICAL SKETCH ...................................................................................419




F-


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

CHEMICAL VAPOR DEPOSITION OF THIN FILMS FOR DIFFUSION BARRIER
APPLICATIONS

By

Omar James Bchir

August 2004

Chairman: Timothy J. Anderson
Major Department: Chemical Engineering

The semiconductor industry is transitioning from aluminum to copper as the

interconnect material for integrated circuits. Diffusion barriers are essential for

preventing copper migration into silicon in the resulting integrated circuits. Metalorganic

chemical vapor deposition (MOCVD) is a useful technique for conformal deposition of

thin barrier films. By manipulating the molecular structure of the MOCVD precursor(s),

it is possible to control the structure and properties of the deposited film.

The focus of this work has been to examine novel precursors for use in MOCVD

of thin film diffusion barriers. To date, the suitability of a variety of novel precursors for

deposition of tungsten nitride (WNx) for diffusion barrier applications has been tested.

The WNx precursors were of the form C14(CH3CN)WNR, where R represents the

isopropyl, phenyl, and allyl group, among others. Mass spectrometry fragmentation

patterns for each precursor were studied to pre-screen candidate precursors. Film

properties were examined by several different characterization techniques, including








XRD, AES, XPS, SEM, and 4-Point Probe. Data from these techniques were then

correlated to pre-screen fragmentation patterns to determine the impact of the imido (R)

group on film deposition. Apparent activation energies for film growth from the allyl,

isopropyl and phenyl precursors, for example, were 0.15, 0.84 and 1.41 eV, respectively,

and were directly related to the strength of the N-R bond in the precursors.

In addition to experimental testing of new precursors for WNx deposition,

thermodynamic phase equilibria for the W-N-H-C-C1 deposition system was assessed.

Modifications were made to the previously reported W-N binary model to include new

experimental data from the literature on the desired FCC WNx phase as well as the SHP

WN phase. Once the binary W-N diagram was reassessed, this model was merged with

the existing W-H-C-Cl database to create an initial model for solid-gas equilibrium in

our system at our experimental conditions. Using a combination of XRD, AES, and XPS

data from our films, the W-N-H-C-C1 model was modified to include

carbon-nitrogen-vacancy interactions in the face-centered cubic (FCC) WNxCy solid

phase, so that the new model reflected our experimental results. Finally, the

low-temperature ternary W-C-N phase diagram was predicted.















CHAPTER 1
INTRODUCTION

1.1 Background

The semiconductor industry continues to shrink the size of transistors on

integrated circuits (ICs) to increase the number of transistors per chip, which translates to

increased IC device speed. To shrink these transistors, the sizes of the device features

(trenches, vias, etc.) and the metal interconnects ("wires" that connect the individual

transistors on the chip) must be decreased. Recent accomplishments have reduced

feature sizes to 100 nm, yielding more than 400 million transistors per chip. The

International Technology Roadmap for Semiconductors [ITR02] predicts IC feature sizes

to shrink continuously over the next several years, in the following steps: 100 nm (in

2003), 90 nm (2004), 80 nm (2005), 70 nm (2006), 65 nm (2007), 45 nm (2010), 32 nm

(2013), and 22 nm (2016).

Aluminum (Al) and Al-based alloys have historically been used as the metal

interconnect materials on integrated circuits. Al is reaching its usability limit as industry

continues to shrink IC features. For a given current flow, decreasing the interconnect

cross-section leads to an increase in current flow density. At high current flow density,

Al suffers from electromigration, where electrons flowing through the metal interconnect

impart enough momentum to carry the metal atoms with them. Once current flow density

rises above the critical value (106 A/cm2 for Al), electromigration leads to voids

(openings) and hillocks (pileups) in the interconnect wiring (Figure 1-1). Formation of









voids in the Al interconnect leads to an open circuit, which causes device failure. Al

doped with Cu (termed Al(Cu)) was used to increase the electromigration resistance, but

this scheme is also reaching its usability limit. A different interconnect metal is required

to meet present and future current flow density demands.


electrons- (a)

hill k
void


electrons (b) .



Figure 1-1. Before (a) and after (b) electromigration induced hillock and void formation
in an Al interconnect.

Industry has already transitioned to copper as the interconnect material for the

intermediate and upper wiring levels on IC devices, due to copper's higher

electromigration resistance and 40% lower bulk resistivity (1.67 p&2-cm for Cu vs. 2.65

tQ-cm for Al) [Cha89, Kol91]. Copper has ten orders of magnitude higher

electromigration resistance than Al, which will increase device lifetime [Mur95], and its

lower resistivity will decrease the resistance-capacitance (RC) time constant, which is a

large factor in chip performance for ICs with feature sizes below 250 nm [Hu98]. The

following relation defines the resistance-capacitance time delay:


RC-= PI (1-1)
tM* t*

where p is resistivity (S2-cm), I is the length (cm) of the interconnect line, e is the electric

permittivity (C/V*cm) of the insulating film, tM is the thickness of the metal interconnect

(cm) and tiLD is the thickness of the neighboring insulator (cm). A decrease in the RC









time constant leads to higher device switching speed [Pau87], and the switch from Al to

Cu may lead to a 15% increase in overall microprocessor speed [IBM97]. While bulk Cu

has a resistivity of 1.67 i.Q-cm, the resistivity of IC Cu lines with thickness above 50 nm

ranges from 1.7 to 2.0 .Q-cm [Kal00]. In comparison, Cu-doped Al lines AI(Cu) have

resistivities ranging from 3.0 to 3.5 pQ-cm. In addition to resistivity and

electromigration improvements, Cu is said to provide higher IC production yield than

Al-based devices with similar design [Sin02a]. Other advantages of switching to Cu

interconnects include a decrease in the number of interconnect levels, a roughly 30%

decrease in power consumption for operation at a given frequency, and a cost savings of

roughly 30% per interconnect level due to integration of dual damascene processing

[Nov00]. Silver (Ag), which has an even lower bulk resistivity (1.59 uQ-cm) than Cu,

has been considered for interconnect metallization, but it suffers from poor

electromigration resistance [Mur95], making its use as the metallization material

unlikely.

The improved electrical properties of Cu are highly desirable, but with these

benefits come several drawbacks. These include copper's tendency to corrode under

standard fabrication conditions, copper's lack of a stable self-passivating oxide (like

A1203 on Al), the inability to chemically etch Cu, poor adhesion to low dielectric constant

(low-k) materials, and rapid diffusion/reaction with Si and Si02 [Hu98, MerOl]. Copper

diffusion into neighboring layers can cause an increase in contact resistance, a change in

the barrier height, leaky p-n junctions, embrittlement of the contact layer, deep level

traps and destruction of electrical connections to the chip [Kal00, Wit80]. Copper


M









contamination levels in Si at and above 1013 cm-3, for example, are believed to reduce

device yield [IstOO].

1.2 Comparison of Diffusivities for Al and Cu in Si

At high temperatures, both Al and Cu diffuse into Si, and when put into direct

contact with Si, they will intermix rapidly. While Si is non-reactive with Al [RamOO], it

has a small solubility range in Al (0.25 to 1.5% by weight) [Jae88]. At high temperatures

(>10000C), Al diffuses substitutionally into Si, where Al replaces a Si atom in the Si

lattice.

Copper, though, has much higher mobility in Si, and is very reactive with Si and

SiO2 substrates [Bau97], forming copper silicide compounds, such as Cu3Si and CusSi

[Bau97, Hym98, Kwo95, RamOO, Rei92], which cause strong deterioration of contact

systems [WanOlb]. Copper's high mobility in the Si lattice extends over a wide

temperature range. Cu moves into the Si lattice by substitutional diffusion at high

temperature (>8000C), but also by interstitial diffusion at low temperature (<7000C),

where Cu atoms diffuse between the Si atoms in the lattice (through the interstices). Both

interstitial and substitutional diffusion of Cu into the Si lattice occur simultaneously, with

interstitial diffusion dominating at low temperature and substitutional diffusion

dominating at high temperature.

While the diffusivity of Al in Si at 5000C is less than Ix10-20 cm2/s [IstOO], the

diffusivity for Cu in Si depends both on type and concentration of dopants in the Si

substrate. The nature of p-type silicon is such that acceptor atoms (such as boron, B)

substituted on Si lattice sites have a negative charge (B-). Since Cu diffuses through Si

as a positive ion (Cu') [Bai96, IstOO], the negatively charged B- ions tend to "trap" some









of the diffusing Cu ions, giving Cu an apparent diffusivity that is lower than that in

intrinsic (undoped) Si. In p-type Si with a boron concentration of 5.0 x 1020/cm3, for

example, the diffusivity of Cu is 2.0 x 106 cm2/sec at 5000C, while that in intrinsic Si is

2.0 x 10-5 cm2/sec [Ist00]. At room temperature, the diffusivity of Cu in p-type and

intrinsic Si, measured by the transient ion drift method [Hei97], ranged from 3.3 x 10-13

cm2/s in p-type Si up to 2.7 x 10-7 cm2/s in intrinsic Si [Ist98]. This indicates that after

30 minutes at room temperature, Cu can diffuse a distance of 0.5 pm in p-type Si and

400 glm in intrinsic Si.

Since the diffusivity of Al in Si is so small at the processing temperature, the main

problem with Al arises when it penetrates the barrier material and accumulates at the

barrier-Si interface. This accumulation can lead to formation of Al spikes at the

interface, which short out shallow junctions (Figure 1-2a). Copper's high diffusivity in

Si, however, means that Cu will not build up at the barrier-Si interface until the Si has

reached its saturation value of Cu. Hence, Cu can diffuse into the bulk of the Si wafer

and move laterally to contaminate a large area on the wafer [IstOO], as depicted in Figure

1-2b.

While Cu has high diffusivity in Si at room temperature, its solubility is very low

(less than 1 Cu atom per cm3) [IstOO], so formation of a Cu-Si solid solution is unlikely.

The diffused Cu will therefore form complexes, precipitates and agglomerates with Si.

Four typical reactions can occur for Cu in Si. These include formation of point

defects/complexes, copper silicide precipitates, addition to existing defects, and

outdiffusion to the Si surface. More information on Cu interdiffusion and reaction

mechanisms in Si is available elsewhere [IstOO]. In addition to Si, Cu also diffuses









rapidly through Si02 under thermal and electrical stress, leading to high leakage currents

and dielectric breakdown [Jai99]. Diffusion of Cu atoms into SiO2 typically occurs due

to application of an external electric field, which causes Cu to be ionized (to Cu') and

accelerated into the SiO2 [Kri01]. Copper's high diffusivity in Si and SiO2 makes the

need for a robust thin film diffusion barrier to separate Cu from neighboring layers on the

IC very critical. Figure 1-2c depicts a stable device structure resulting from use of a

diffusion barrier.


ILD Al ILD D LD LD Al or Cu ILD
Si a Si b Si c

Metal Diffusion Diffusion
Spikes Region Barrier


Figure 1-2. Metal to silicon contact layers: (a) Al on Si without a diffusion barrier
suffers from junction spiking. (b) Cu on Si without a barrier suffers from
massive diffusion (c) Metal to silicon contact with a diffusion barrier
remains intact.

1.3 Diffusion Barrier Requirements and Properties

A diffusion barrier is a material used to separate two layers that, if put in direct

contact, would interdiffuse and/or react with one another. Diffusion barriers are typically

classified into one of three types: passive, sacrificial and stuffed. Passive (ideal) barriers

are inert with respect to the layers that they separate, and have low solubility for the

neighboring semiconductor and metal. Sacrificial barriers react with one or both of the

neighboring layers, and are eventually consumed. The rate of reaction between the

barrier and neighboring layers must be slow enough that the lifetime of the barrier is

longer than that of the device. If the barrier has a shorter lifetime than the device, the

sacrificial barrier will be totally consumed and the device will fail prematurely. Stuffed









barriers are polycrystalline films that have their grain boundaries stuffed with a material

that blocks diffusion [Wol86]. By stuffing these boundaries, the dominant

low-temperature route for diffusion through the barrier is minimized. Stuffing occurs

due to both physical and chemical effects. The stuffing "agent" should physically reduce

the amount of pore space in the films and should chemically repel the Cu from diffusing

through the grain boundary [Par95a]. As an alternative to grain boundary stuffing, high

temperature annealing has been used to "cure" deposited barrier films, so that

micro-defects such as grain boundaries are removed. Although this technique does

decrease the number of grain boundaries in the film, the high temperatures also increase

diffusivity of Cu in the barrier film, and can cause Cu penetration [Bai96]. The "curing"

properties of a high temperature thin-film anneal are therefore tempered by the increased

mobility of Cu in the barrier film at higher temperatures.

In addition to stuffing the grain boundaries, barrier effectiveness can be further

enhanced by minimizing the number of grain boundaries in the barrier film. Deposition

at low temperatures provides this benefit and several more. First, low temperature

deposition minimizes the number of grain boundaries by favoring amorphous film

growth. Second, in nitride films, for example, excess nitrogen is retained at lower

deposition temperature, and is therefore available to stuff any grain boundaries that may

form during thermal cycling. Third, the likelihood of damaging temperature sensitive

components (such as low-k dielectric polymers) on the device during barrier deposition

is minimized, provided that a temperature ceiling of 4000C is maintained [Eis00a, Ele02,

HauOO, Kim03b, Les02, SunOlb]. Deposition above 4500C can increase the compressive

stress in Cu films, which is relieved by formation of Cu hillocks [Jai99]. Fourth, low









deposition temperatures require a smaller thermal budget. The advantages provided by

low temperature deposition can greatly reduce the probability of Cu penetration through

the barrier and decrease power consumption. In practice, low deposition temperatures are

desirable, but depositing the barrier films at the lowest possible temperature may not be

the optimal approach. Depositing the barrier film at the highest allowable processing

temperature, assuming that the film structure remains amorphous at this temperature,

would prevent significant changes to barrier film structure during subsequent thermal

cycling.

During IC processing, the diffusion barrier material is typically deposited between

a metal layer (usually Al or Cu) and a semiconductor layer (usually Si or GaAs), a metal

and a dielectric (e.g., SiO2 or low-k), or two metal layers [Lin98a]. The requirements for

an ideal diffusion barrier are listed below. Film structure, electrical properties,

conformality and adhesion property requirements will be discussed further in the

following sections.

Barrier must prevent diffusion between and be non-reactive with
neighboring device layers.

Barrier should have amorphous film structure to eliminate grain
boundaries, which are facile paths for Cu diffusion through the barrier.

Barrier should be deposited at low temperature to prevent damage to
temperature sensitive components on the IC device and to enable
amorphous film growth.

Barrier should have good conformality and uniformity across the wafer, to
ensure good barrier coverage over small device features and uniform
barrier deposition on all devices across the wafer.

Barrier should have low bulk electrical resistivity and low contact
resistivity to Cu, to minimize resistance to current flow.

Barrier should have good thermal conductivity to minimize heating of the
barrier layer.









Barrier should have minimal contamination levels; halide impurities can
cause corrosion of the Cu layer, while oxygen and free carbon impurities
can increase film's electrical resistivity.

Barrier should have minimum thickness to maximize copper's
cross-section in the interconnect and to foster low contact resistance in the
metallization stack.

Barrier should promote adhesion between device layers, as poor adhesion
leads to electromigration of Cu at the de-adhered interface.

Barrier should have high thermal and structural stability, to prevent failure
of the barrier during exposure to thermal and mechanical stresses at
processing conditions.

Barrier should enable direct Cu electroplating to eliminate the need for a
Cu seed layer, and promote nucleation of the Cu (111) orientation, which
has the greatest resistance to electromigration.

Barrier should be compatible with chemical mechanical planarization
(CMP) processes, and act as a good CMP stop layer to eliminate need for
a separate stop layer deposition step.

1.3.1 Film Structure

The ideal diffusion barrier structure is a defect free single crystal film [Nic78].

This structure does not contain grain boundaries, which are interface regions

("micro-defects") between crystal grains in a polycrystalline material. Grain boundaries

are a "short circuit" path for rapid intermixing of the two neighboring layers, and are the

dominant cause for barrier failure at low temperature [Bai96]. In a single-crystal barrier

film, all diffusion occurs through bulk defects (including vacancies and dislocations),

which is inherently much slower than diffusion through grain boundaries. Deposition of

single crystal barrier films is impractical, however, due to low growth rates, lattice

mismatch with the underlying substrate, and deposition temperature limitations [Kal00].

The next best solution is to deposit an amorphous, dense, smooth, defect free film [Bai96,

Kwo95]. Amorphous films have short-range order (< 5 A), but no long-range order (>








20 A) [E1190]; therefore no polycrystalline grain boundaries exist to enable rapid

diffusion. In addition to having an amorphous microstructure, the density of the barrier

material should be as close to the bulk value as possible, to eliminate voids as possible

diffusion paths. Polycrystalline films, especially those with a columnar microstructure or

those with equiaxial grains of size similar to film thickness, are the poorest performers in

diffusion barrier applications. These films contain grain boundaries that can extend from

one side of the barrier to the other, meaning an easy diffusion path for Cu through the

barrier. Nano-crystalline films, which are polycrystalline films with grain size below

-50 A [Kal00], are more effective than polycrystalline films with larger grains, but are

not as effective as amorphous diffusion barriers. Examples of single-crystal,

polycrystalline and amorphous films are shown in Figure 1-3.


Metal t, ..... Metal M Meta Meta




Substrate a Substrate / I Substrate c Substrate d

Polycrystals Grain Polycrystals Grain
Boundary Boundary
Boundary

Figure 1-3. Diagram of (a) Single crystal barrier. (b) Polycrystalline barrier with
equiaxial grains. (c) Polycrystalline barrier with columnar grains. (d)
Amorphous barrier without grain boundaries.

Several different methods are used to deposit amorphous films [Siv95]. The first

is to deposit films at low temperature, which enhances sticking probability and

suppresses surface diffusion, forcing films toward a disordered state. The second is to

deposit films at high nucleation (arrival) rates, so that many small nuclei form on the

substrate surface, preventing coalescence of small nuclei into larger polycrystals. The









third is to add contaminants (such as N, C, etc.) to impede surface diffusion of the metal

species, imparting disorder and decreasing the ability of nuclei to coalesce and grow. In

addition to these deposition methods, careful selection of materials can assist formation

of an amorphous film. A material's composition may be chosen, for example, so that

several stable phases exist together at equilibrium. In a binary system, a film

composition should be chosen in a two-phase equilibrium region, while in a ternary

system, a three-phase equilibrium region should be selected [RamOO]. The competing

stable phases neighboring these regions should be line compounds (have minimal solid

solubility regions), have very different compositions, and have complex crystal structures

[RamOO]. Compared to a broad solid solution phase, line compounds are inflexible to

stoichiometry deviations, which can disrupt polycrystal formation and trigger amorphous

growth. In a binary system, then, choosing a film composition in a two-phase

equilibrium region bounded by two line compounds should increase the likelihood of

amorphous film deposition.

1.3.2 Electrical Properties

To obtain any benefit from the lower bulk resistivity of Cu, the barrier (liner)

thickness should be less than 10% of the overall linewidth [Hu98]. A thicker barrier

means decreased cross-sectional area of Cu available to conduct current, which causes an

increase in the effective resistivity of the Cu line. For a Ti/TiN/AI(Cu)/Ti/TiN sandwich

structure, the effective resistivity was estimated to be 4.1 pQ-cm, due to the high

resistivities of the Ti and TiN layers, as well as that for TiA13, which forms during

annealing at the Ti-AI(Cu) interface [Hu98]. For Cu lines with widths ranging from 0.1









to 1 p.m, keeping barrier thickness < 10 % of the overall linewidth results in a constant

effective resistivity < 2.3 pLQ-cm for the Cu line [Hu98].

The barrier material will typically be used in two types of scenarios on the

integrated circuit, which are depicted in Figure 1-4. In the first case, the barrier runs in

parallel with the Cu line, as with a trench structure, where the barrier separates the Cu

trench from neighboring dielectric layers. In this case, the thickness of the barrier must

be minimized, in order to have maximum cross-sectional area for the Cu conductor line.

The resistivity of the barrier in this scenario is not critical, however, as electrical current

will be flowing through the Cu and not directly through the barrier [Tra03]. In the

second case, the barrier will be separating one metal layer from another, as in a via

structure, hence the current will pass directly through this intermetal barrier layer. The

resistance associated with current flowing from one Cu line through a via (and barrier) to

another Cu line is called the via resistance. The barrier film's thickness and bulk

resistivity must be minimized to limit the barrier's impact on via resistance. The

suggested ideal film resistivity is < 500 p.-cm [Ele03], and barrier thickness should be

<30 nm [Bai96]. Specific via resistance (Q-pLm2 or 2-cm2), defined as the electrical

resistance through the via multiplied by the via's cross-sectional area, is a common

measure of barrier impact on current flow through a via. The current, acceptable specific

via resistance value for the 100 nm device node is 0.1 9-2-m2 (1 x 10~9 Q-cm2), and this

is projected to decrease to <0.01 Q-pLm2 (1 x 10-10 o2-cm2) for the 22 nm device node in

2016 [ITRO2, Nic95, Tra03].

As an example, the change in specific via resistance with ITRS feature diameter

for a fixed 15:1 via aspect ratio was calculated. The via was treated as a cylinder, with









Cu in the center and the barrier layer lining the inside edges of the cylinder. Current is

assumed to travel in the Cu cross section of the via, which is the path of least resistance,

and then passes through the barrier layer at the via's bottom, as depicted in Figure 1-5.


Trench
Barrier

- Via Barrier


Intermetal Barrier


Figure 1-4. Schematic of trench and via applications for barrier materials.


-- Cu


- Barrier


Figure 1-5. a) Cut away view of Cu trenches with via contact, b) Expanded, 3D view of
Cu via with barrier layer on sides and bottom.

The thickness of the barrier layer corresponded to ITRS projections for each device node,

as follows: 12 nm (in 2003), 10 nm (2004), 9 nm (2005), 8 nm (2006), 7 nm (2007), 5

nm (2010), 3.5 nm (2013), and 2.5 nm (2016). The results for varying barrier

resistivities, along with projected requirements from the ITRS Roadmap, are shown in

Figure 1-6. The diagram indicates that by the 65 nm node (2007), the barrier layer's









resistivity must be < 200 p.2-cm to meet ITRS projections. Likewise, this value must be

S25 pQ-cm by the 32 nm node (2013).

1.2e-9
^-*- 10 pL-cm
1.0e-9- -0- 25 I-cm
"-0- 200 Q-cm
a 8.0e-10 -- 200 .pQ-cm
-0- 300 .Q-cm iS^ /
,- ITRS Projections
6.0e-1 -

> 4.0e-10

i 2.0e-I&

0.0
22 32 45 6570 80 90 100
Feature Diameter (nm)

Figure 1-6. Specific via resistance as a function of feature diameter, shown for various
barrier layer resistivities. Calculation is for a via with a 15:1 aspect ratio.

1.3.3 Conformality

Conformality (or step coverage) in device microstructures, especially in trenches

and vias, is of utmost importance in barrier design. Highly conformal films (approaching

100%) have nearly uniform thickness at all points on a substrate surface (on both the

sidewalls and bottom). Film conformality is determined by measuring the film's

thickness on the wafer surface (ts), on feature sidewalls (tw) and at feature bottom (tb)

(Figure 1-7). Conformality on the sidewall and bottom can be calculated by Equations

1-2 and 1-3, respectively. The feature's aspect ratio, which is the ratio of the feature's

height to its width, can be calculated by Equation 1-4.

Sidewall conformality (%)= 100(1-2)
Sidewall conformality (%) = ('- 100
t s











Bottom conformality (%) = tb 1 100 (13)
t{ )


Aspect Ratio = (1-4)



te -- Barrier
film


i .b


Substrate


Figure 1-7. Sketch of film dimensions for calculation of conformality.

Barrier Film







Substrate Substrate Substrate


Figure 1-8. Diagram of (a) Ideal barrier conformality. (b) Good barrier conformality. (c)
Poor barrier conformality.

Barriers with poor conformality have uneven thickness, with the film being

thinner in certain spots than others. These thin spots are "weak links" in the barrier,

which are more susceptible to diffusion than the thicker parts of the layer. Figure 1-8a

shows an ideal barrier with 100% conformality. Figure 1-8b shows a practical barrier

with rounded edges, while Figure 1-8c shows a poor barrier, with a large overhang near








the trench opening and sparse coverage at the trench bottom. Cu can easily penetrate to

the underlying Si substrate through the thin barrier at the bottom of the trench.

While device dimensions continue to shrink, the aspect ratio continues to

increase, and the ability to conformally cover the bottom and sidewalls of these features

becomes even more challenging. Deposition methods and chemistries must be developed

to accomplish this task.

1.3.4 Adhesion

The adhesion of the barrier to both the Cu and the neighboring dielectric or

semiconductor layers, along with the crystallographic orientation of the deposited Cu

(which is affected by adhesion issues), all impact copper's electromigration behavior

[Jai99]. Strong Cu adhesion to the barrier layer contributes to good electromigration

resistance [Pet03], whereas poor adhesion can cause self-diffusion of Cu along the

barrier interface. Self-diffusion of Cu along the interface is even more likely than

diffusion along grain boundaries, because grain boundary self-diffusion has a higher

activation energy than interface self-diffusion [Llo95]. Copper has an activation energy

for bulk self-diffusion of 2.19 eV [Mur95], while the values for grain boundary and

interface self-diffusion are 1.2 and 0.8 eV, respectively [Llo95]. Cu (111) is the

preferred orientation for the metal, as this plane has minimum surface energy, which

results in formation of low-angle grain boundaries [Llo95]. These low-angle grain

boundaries minimize flux divergence of electrons flowing through the metal, and

therefore minimize electromigration [Jai99]. Aside from the barrier-Cu interface,

adhesion at the barrier-dielectric interface can be adversely affected by out-gassing of









moisture or other organic materials from the dielectric surface, which can lead to

anomalous Cu removal behavior.

1.4 Dielectric Material Considerations

In addition to the changeover to Cu, parallel research is pursuing new, low-k

materials, which will also decrease the RC time constant. Historically, SiO2, with a k

value ranging from 3.9 to 4.1, was the dielectric material of choice [Hu98]. Transitioning

from Al/SiO2 devices to Cu/low-k based devices (with k-2) may lead to more than a 400

% reduction in RC delay [Kal00]. New films (such as polyimide, SiLKT, SiOC, etc.),

with dielectric constants below 3.9, are being investigated for use as low-k materials, but

they cannot tolerate high temperatures during processing [Hu98, Pet03]. The processing

temperature limit for these materials, along with the need for good adhesion to Cu, are

factors that must be considered when selecting a barrier material to separate them.

Another factor influencing barrier material selection is its compatability with

high-k materials, such as those in MOSFET gates or in DRAM capacitors. Barrier

materials have been used as top electrodes on high-k gates and DRAM capacitors, and

their favorable performance may dictate further use in the future.

1.5 Diffusion Barrier Failure Mechanisms

Thin films typically have higher diffusivity than thicker bulk films, due to a large

number of short circuit paths (grain boundaries, dislocations, etc.) distributed over a very

small volume [Bal75]. Thinning the barrier film increases the likelihood that these short

circuit paths will extend from one side of the film to the other, making it more susceptible

to diffusion. Thin film diffusion barriers typically fail in one of two ways. The first is

metallurgical failure, where the Cu content of the barrier increases to several atomic








percent, changing the barrier layer's composition. The second is electrical failure, where

little intermixing with the barrier occurs, but the Cu penetrates the barrier and is present

in the semiconductor/dielectric layer in sufficient quantity to alter the characteristics of

the device [IstOO]. Most barrier testing techniques, such as depth profiling or SECCO

etch tests [Iva99], are suited for detection of metallurgical failure. Use of XRD to detect

CuSix compounds, for example, is a fairly insensitive detection method, because the bulk

concentration of Cu in Si must already be above saturation before the silicide compounds

will form.

Measurement of electrical failure is a more sensitive way to determine if Cu has

penetrated the barrier and intermixed with the underlying substrate. To detect electrical

failure, the electrical properties of devices such as p-n junctions, Schottky diodes or

metal-oxide-semiconductor (MOS) capacitors, built under the Cu-exposed barrier or on

the wafer surface after the barrier has been removed, are measured. Leakage currents

are measured on p-n junctions and MOS devices and compared to Cu-free ones to

determine if Cu has penetrated the barrier. Changes in the capacitance and current-

voltage (I-V) characteristics of Schottky diodes are measured to determine if Cu has

penetrated into the active areas on devices.

Ideally, the performance of different diffusion barrier materials would be tested

by comparing the diffusivity of Cu in the different barriers. To determine diffusivity, the

migration of atoms, or diffusive flux (J, atoms/cm2-sec), through the barrier material

should be estimated. The flux can be described by Fick's law (Equation 1-5):

J=- C (1-5)
J = -D (1-5)
(dx








where D is the diffusivity (cm2/sec), C is the atomic concentration (atoms/cm3) and x is

the diffusion distance (cm). By measuring flux (J) through the barrier for a given

concentration gradient (dC/dx) across the barrier, the diffusivity can be determined.

Once this is known, the next level of analysis is to determine whether diffusion through

the bulk lattice or the grain boundary dominates. To distinguish between diffusion

through the bulk lattice and the grain boundary, two different expressions for diffusivity

have been given [Ohr92]. Equation 1-6 gives the expression for lattice diffusivity (DL):


DL = DoL exp(ERT (1-6)

where EL is energy per mole for atomic diffusion through the lattice, and DoL is the value

for lattice diffusivity at standard conditions. Equation 1-7 gives the expression for grain

boundary diffusivity (DB):


D, = DoB exp T) (1-7)

where EB is the energy per mole for atomic diffusion through the grain boundary, and DoB

is the value for grain boundary diffusivity at standard conditions. At low temperature, the

diffusion mechanism through a polycrystalline solid is typically controlled by grain

boundaries and other defects in the film [Nic78]. Since vacancy-assisted bulk diffusion

is negligible below about three-tenths of the solid's melting temperature (0.3Tm),

diffusion through the grain boundary is the path of least resistance at low temperature.

This is reflected by a lower activation energy for grain boundary diffusion relative to bulk

diffusion [Cha82, Nic78], with EB = 0.5EL for FCC metal thin films [Bal75]. As

annealing temperature is increased, a changeover from grain boundary controlled

diffusion to bulk diffusion occurs. The temperature at which this occurs, called the








Tammann temperature, is typically 0.5-0.66Tm [Nic78]. At higher temperature, the

values for DL and DB begin to approach each other, and flux through the bulk lattice

becomes the dominant route for diffusion, due to the much larger cross sectional area of

the bulk lattice relative to that of the grain boundaries.

Experimentally, determination of diffusivity from Cu flux measurements is

impractical. Exact measurement of the Cu flux through the barrier is difficult to

determine, as failure determination techniques usually focus on qualitative barrier failure.

Moreover, issues such as defect density can have a large impact on thin film diffusivity,

and can cloud efforts to distinguish between bulk lattice and grain boundary diffusion.

Diffusivity in thin barrier films may be estimated with Equation 1-8, which describes

diffusion from a limited source at a crystal surface [Wol86]:


Ck = exp(- d Dt (1-8)


where Ck is Cu concentration (cm-3) at distance d (in cm) from the surface, D is the

diffusivity (cm2/s), t is diffusion time (sec), and Qo (cm-3) is the concentration of Cu at

the surface. Since exact Cu concentrations at the two interfaces can be difficult to

measure, an apparent diffusivity (Dapp) is defined by Equation 1-9 [IstOO]:

d2
Dapp (1-9)
4t

where d is the film thickness (in cm) and t is the time (s) for the Cu to appear at the

barrier/substrate interface (i.e., the time it takes the barrier to fail). Comparison of

apparent diffusivities for different barrier films may be done to determine which barrier is

best.









1.6 Diffusion Barrier Deposition Techniques

Diffusion barrier films have been deposited by a variety of methods, including

physical vapor deposition (PVD) and chemical vapor deposition (CVD) techniques. The

required deposition temperature, substrate characteristics and feature size generally

dictate which deposition method is appropriate for a given application. A discussion of

some common PVD and CVD techniques to deposit metal nitrides and their inherent

advantages and/or disadvantages follows.

1.6.1 Physical Vapor Deposition (PVD)

Typical physical vapor deposition (PVD) methods to deposit barrier films include

electron beam evaporation and sputtering. In electron beam evaporation, an electron

beam is used to heat a small portion of the solid metal above its melting point. While at

high vacuum (10-3 to 10-4 Torr), the molten metal then evaporates atoms into the gas

phase. These atoms diffuse through the chamber's atmosphere and deposit on the

susbtrate. One or more reactive gases, such as NH3 or CH4, for example, may be

introduced into the chamber during evaporation to deposit multi-component films.

The sputtering process involves the use of a plasma gas (typically argon) to

physically "knock" atoms off of a metal target and onto a substrate. The sputter chamber

operates at low vacuum (-100-120 mTorr), and cations from the plasma impact the metal

target cathode to liberate metal atoms, which then deposit on the substrate. In the

reactive sputtering process, a reactive gas such as N2, for example, is introduced into the

sputter chamber along with the plasma gas. Once the atoms liberated from the metal

target cathode reach the substrate, they can interact with the nitrogen radicals formed by









the plasma and deposit either a metal nitride compound or a metal-nitrogen solid

solution.

Films produced by PVD processes suffer from high stress, a large number of

vacancy defects, excess free volume and poor conformality, leaving spots in the barrier

open to diffusion by neighboring species at via and trench sides/bottoms [Aff85, Bau98,

Bos91, Kel99, Lee93, Lu98b]. The poor conformality inherent to all PVD processes is

due to the directionality imparted to the atoms/clusters traveling toward the substrate.

This directionality causes step shadowing, where parts of the substrate surface are not

"seen" by the incoming sputter atoms. Step shadowing results in little or no film

coverage on certain sections of the substrate. These exposed substrate areas are then

vulnerable to reaction with subsequently deposited Cu atoms. Figure 1-9 depicts step

shadowing.

Directional Particles Directional Particles Directional Particles
0 4 Q 00 00 0 0 1
o .\ ** e




Substrate Substrate SuSubstrate Substrate Substrate
/Barrier
Film Shadowed Shadowed
/ wall t wall
Shadowed Barrier Barrier
walls Film Film
a b c


Figure 1-9. Step shadowing effects in small device features due to directional particles
arriving at the substrate. a) At an angle to the surface, b) Perpendicular to
surface, without resputter. c) Perpendicular, with resputter.

Three variations on sputtering have been developed to extend its applicability to

smaller device feature nodes. The first, collimated sputtering, involves the placement of









a plate (which contains small holes) between the sputter target and the substrate [Sin02b].

The plate collimates the sputtered atoms/clusters, so that only those traveling parallel to

each other in a direction perpendicular to the plate reach the substrate. Once at the

substrate, the atoms can deposit or can resputter the metal film at the bottom of a feature

(Figure 1-9c), moving it up onto the feature walls to ensure sidewall conformality. The

second variation, long-throw sputtering, improves the performance of the sputter

deposition system by increasing the distance between the target and substrate, so that

only sputtered atoms/clusters travelling in parallel (and perpendicular to the substrate)

reach the substrate [Sin02b]. Ionized-PVD (I-PVD) is the third variation used to extend

the applicability of sputtering processes beyond the 100 nm device node. I-PVD

involves use of a second rf coil to ionize neutral metal atoms that have been liberated

from the target by sputtering. These ionized atoms are then accelerated toward the

substrate, which is biased to attract the newly ionized metal atoms [RosOl]. Impinging

metal atoms resputter the metal that has already deposited on the substrate, and these

resputtered atoms can move from the bottom of a feature to the sidewalls and coat them.

This leads to improved step coverage in small features compared to other PVD methods.

Conformality improvements within the feature from the I-PVD resputter effect are

tempered by resputter thinning bevelingg) of the feature's top edges (Figure 1-9c). This

beveled edge is a weak point in the barrier, and is susceptible to attack by Cu. In

addition, although resputter associated with I-PVD improves sidewall coverage, this

coverage is still tenuous for films in high aspect ratio features [Vij99]. These extended

sputtering techniques, being directional in nature, also have a limited lifespan [Sin02b],

and may reach their usability limit at the 45 nm device node in 2010. Among the barrier









deposition techniques, only a chemical vapor deposition (CVD) based process offers the

conformality required for future device generations, with 100% conformality being the

goal for all barriers [Fal98, Fle98, Lev98, LiuOO].

1.6.2 Chemical Vapor Deposition

Chemical vapor deposition (CVD) is a widely used film growth technique. It

involves the use of one or more gas phase reactants to deposit a solid film on a substrate.

The reactant flux at the substrate surface is non-directional in nature, eliminating the

possibility of step shadowing in small device features. The substrate promotes reaction

between components from the gas phase, and successive reactions lead to film growth.

CVD processes typically grow highly conformal films at high growth rates, making them

superior to PVD techniques for deposition on aggressive device features. Unlike PVD

methods, CVD is capable of area selective growth, which can eliminate a patterning step

during device fabrication [Gat96]. CVD techniques are also characterized by high

throughput, minimal downtime and easy source changeout, which makes them very

attractive for industrial applications. The steps of the CVD process are as follows [Siv95,

Vos78]:

Reactants in the gas phase diffuse through the boundary layer (defined
below) to the substrate surface

Reactants adsorb onto the substrate surface

Adsorbed species move around surface and settle into available surface
sites

Reactants adsorbed to available surface sites undergo final reactions,
which are often catalyzed by the substrate surface, and reaction products
are incorporated into the growing film

By-products from the reaction desorb from the substrate surface









Desorbed by-products diffuse through the boundary layer and away from
the substrate surface

Conversion of gas phase precursors into deposited film can involve both gas

phase and surface reactions. Gas phase reactions can involve formation of intermediates

and by-products as well as parasitic reactions, such as gas phase nucleation. Surface

reactions can involve adsorption/desorption, film formation, and side reactions, which

can cause contaminant incorporation into the film. The total pressure in the reactor

determines the degree of coupling between gas and surface chemistry [Gat96]. At a total

reactor pressure of 1 Torr or greater, homogenous and heterogeneous processes are

tightly coupled, while at pressures < 10-4 Torr (where Xmfp = 50 cm), the probability of

gas phase collisions becomes negligible and the process becomes strictly surface

controlled [Gat96]. This surface control at low pressure enables finer control of film

properties, such as smoothness and thickness. At higher reactor pressures (>1 Torr), one

or more of the gas phase precursors may begin to decompose and/or react to form

intermediate gas phase compounds [Gat96]. These precursors (and/or intermediates) can

then travel to the substrate surface.

When the precursor (or intermediate) molecules first reach the substrate surface,

they can weakly bind (physisorb) to the surface by van der Waals forces or they can

immediately react and deposit chemisorbb) [Mas96]. Physisorbed molecules have some

degree of mobility, and can diffuse along the substrate surface until they either chemisorb

or desorb back into the gas phase. High surface mobility, coupled with the

non-directional nature of the reactant flux, are the key CVD features that enable highly

conformal film deposition on substrates with varying topography. The probability that an

impinging precursor chemisorbs at the initial point of impact on the substrate surface is








known as the reactive sticking coefficient. A precursor with high reactive sticking

coefficient (approaching unity) and low surface mobility yields a rough film texture,

while a precursor with a low reactive sticking coefficient (10-3 or less) and high mobility

yields a very smooth film [Gat96]. Hence, to deposit smooth films by CVD, a low

reactive sticking coefficient coupled with high surface mobility is desirable.

Surface mobility and reactive sticking coefficient are highly dependent on

substrate temperature, which must be controlled closely to ensure deposition of films

with appropriate structure and properties. Temperature must be high enough to activate

the reactions leading to CVD, but cannot be increased indiscriminately. At lower

temperature, the sticking coefficient approaches unity and surface diffusivity is minimal

[Gat96]. As temperatures increase, the sticking coefficient decreases, while the surface

diffusivity and deposition rate typically increase. Higher deposition rates are likely on

the flat surface of the substrate, but aspect ratio dependent effects (ARDE) can inhibit

deposition in small features. ARDE occur when by-products emanating from the

sidewalls and bottom of a small feature inhibit in-diffusion of fresh precursor to the

feature. While increasing temperature increases reaction rate on the flat surface, ARDE

can lead to a starved condition on the interior walls of the feature. Hence, although

increased deposition temperature improves surface diffusivity and decreases sticking

coefficient, it can negatively impact film conformality in small diameter, high aspect ratio

features (Figure 1-8c).

In addition to optimizing surface mobility and reactive sticking coefficient, the

operating regime must be considered when selecting deposition temperature. At lower

temperature, the deposition rate is exponentially dependent on substrate temperature,








which is indicative of the kinetically controlled regime. Increasing temperature in this

regime causes a large increase in film deposition rate, and the slow step in the deposition

process is the reaction on the surface. At higher temperature, the deposition rate is

relatively insensitive to deposition temperature, which is indicative of the gas-phase

mass transfer controlled regime. In this region, the temperature is high enough that

essentially all reactants reaching the substrates surface immediately react, and the

rate-determining step in the deposition process is transport of reactants to the surface.

While the temperature dependence of the deposition rate in this regime is mild (varying

from T1 5-2.), the lack of kinetic (surface) control in this regime exacerbates conformality

problems. Moreover, higher deposition temperature can alter film structure, causing a

shift from amorphous to polycrystalline films, which introduces grain boundaries that can

kill barrier performance.

In small features that are to be conformally coated, the surface diffusion length of

the physisorbed precursor should be the same order of magnitude as the feature diameter

[Coo89]. In larger features, surface diffusion alone is inadequate for conformal coverage,

so it must work in tandem with reflection, whereby molecules are reflected back and

forth from the sides of the feature until they find a place to chemisorb. Decreasing the

precursor's sticking coefficient increases the chance of reflection in a feature. For good

coverage, the dimension of the feature should be smaller than the mean free path in the

gas phase, so higher pressure (which has a shorter mfp) can be used with smaller device

features [Tsa86]. Pressures at and below 500 Torr would be adequate for deposition in a

100 nm feature.








Molecules designed for use as CVD precursors should have high volatility,

thermal liability, and have easily (and cleanly) removable supporting ligands [Kod94].

The CVD chemistry should be selected to avoid formation of non-volatile reaction

by-products, which can cause particle deposition on the substrate surface and high point

defect density [Jai99]. In addition, selection of precursors with low reactive sticking

coefficient and high mobility surface species will enhance conformality of the deposited

film.

1.6.2.1 Precursor Delivery

During CVD, one or more gas-phase chemical precursors must be delivered to

the reactor. The precursors themselves may exist in the gas phase, or they made be put

into the gas phase by a number of techniques. Liquid source precursors are typically put

into the gas phase by a liquid bubbler system, where a canister containing the liquid

precursor sits in a temperature controlled bath (which controls vapor pressure). Carrier

gas flows into the bottom of the canister, and bubbles up through the liquid. As the gas

bubbles up through the liquid precursor, it becomes saturated with precursor molecules.

This carrier gas, saturated with precursor, then flows into the reactor.

There are two typical methods to convey low volatility solid precursors to the

reactor: solid source and nebulizer delivery. In a solid source delivery system, the

precursor is held in a metal or glass tube. The tube and precursor are heated by thermal or

photon energy, and solid precursor near the top of the tube sublimes and is conveyed to

the reactor by a carrier gas. As a general rule, a solid precursor must have a vapor

pressure higher than 10-3 Torr at its melting point to be effectively sublimed and









conveyed by a solid source delivery system [Ohr92]. If the vapor pressure is below this

value at the melting point, minimal precursor delivery and film deposition will occur.

The second technique, nebulizer delivery, can be used to overcome this vapor

pressure limitation. This technique requires dissolution of the solid precursor in a liquid

solvent. The solvated precursor is pumped into the nebulizer, which contains a

piezoelectric quartz plate. This plate vibrates due to application of a high frequency

electrical current, and vaporizes the liquid droplets that contact it. Once the droplets are

vaporized into a mist, carrier gas flows through the nebulizer and conveys the mist into

the reactor. Figure 1-10 shows a schematic of a nebulizer.

Solvated Precursor /
Carrier Gas to CVD
Reactor


00
0 O
O0 Precursor
"Mist"


Vibrating
Quartz Plate Dissolved
S "' 4 i \ Precursor from
Plastic Syringe Pump
Tubing
Carrier Gas to
Nebulizer




Cable to Power Supply

Figure 1-10. Schematic of a nebulizer delivery system.

1.6.2.2 Variants of CVD

Several variants of CVD are typically used to deposit thin films. The first is

low-pressure chemical vapor deposition (LPCVD), which typically relies on metal halide









chemistry (e.g., TiCl4, WF6) to deposit films. This technique usually requires high

deposition temperature (>4500C) and suffers from halide incorporation into the barrier

films. The presence of halides in the barrier can lead to Cu corrosion, which decreases

Cu-barrier adhesion and reduces electromigration resistance [Huo02].

The same halide chemistries used for LPCVD are typically used for

plasma-enhanced chemical vapor deposition (PECVD), in which a plasma assists

fragmentation of the reactants, thereby lowering the deposition temperature. After

fragmentation by the plasma, the precursor fragments travel to the heated substrate, react

on its surface and deposit a film. Reduced conformality for PECVD films in high aspect

ratio trench features is due to the directional nature of the plasma [Tsa96], as shown in

Figure 1-9. Despite the low resistivity and deposition temperatures that PECVD offers,

its inability to deposit highly conformal films in high aspect ratio features makes its use

in future barrier deposition processes unlikely.

In metalorganic chemical vapor deposition (MOCVD), films are deposited by

reaction of one or more carbon-containing vapor phase precursor compounds. MOCVD

precursors typically have some or all of the halide ligands common to LPCVD and

PECVD replaced with carbon-containing ligands. These precursors afford little or no

halide incorporation into the deposited films, and MOCVD has been demonstrated to

deposit a variety of films at low temperature. By varying the structure of the precursor

molecule, the precursor can be optimized to dissociate in a "clean" fashion (i.e., with

minimal oxygen and carbon contamination) and at relatively low deposition temperature.

The presence of impurities (O, N, C) is reported to improve stability in contact structures,

because the impurities tend to stuff grain boundaries and inhibit diffusion [So88].








Oxygen's "stuffing" effectiveness, however, is lower for Cu than Al, because the

reactivity of Cu with O is less than that for Al with O [Kim99c]. Carbon bound to the

metal has also been reported to improve thermal stability of the barrier and to foster

growth of smaller grains [Wan01b], but free carbon in the films can scatter electrons,

increasing film resistivity. Among the challenges for MOCVD of refractory nitride

materials are controlling carbon and oxygen contamination, lowering the deposition

temperature, and minimizing film resistivity.

1.6.2.3 CVD Reactors

CVD reactors fall into two broad categories: hot wall and cold wall reactors. Hot

wall reactors have heated walls, with typically laminar flow profiles. These are preferred

for exothermic reactions, since the hot wall temperatures discourage or prevent unwanted

deposition on the reactor walls. In contrast, cold wall reactors have steep temperature

gradients surrounding the susceptor, which often leads to convection pattern formation in

the reactor. These are preferred for endothermic reactions, since the reaction will occur

most readily on the hottest surfaces in the system [Vos78].

CVD reactors are typically run in two modes: differential or starved. In a

differential reactor (similar to a CSTR), the ratio of reactant out to reactant in is

approximately one, so that the composition in the reactor is essentially constant. In a

starved (feed rate limited) reactor, the ratio of reactant out to reactant is much less than

one, meaning that there are large concentration gradients in the reactor caused by fast

reactions [Kod94]. ARDE can cause both differential and starved conditions to exist

simultaneously on a substrate with aggressive feature topography. Flat surfaces on the

substrate receive adequate precursor flux, and hence operate in a differential condition,








while ARDE in small features prevent good precursor flux, starving the sidewalls and

bottom. Increasing the deposition temperature increases the concentration disparity

between the inside of the feature and the flat surface of the substrate, further degrading

the conformality over these features.

1.6.2.4 CVD Transport Issues

Heat and mass transport in a cold-wall CVD reactor can be complex, due to large

temperature gradients. Depending on pressure in the reactor, mass transport may occur

by fluid flow and/or diffusion [Tim01]. At higher reactor pressures (near atmospheric),

both fluid flow and diffusion occur simultaneously. In fluid flow transport, gas

molecules follow streamlines through the reactor, while for diffusion transport,

concentration gradients drive transport across streamlines. Heat transport occurs by a

combination of convection, conduction and radiation in the reactor. Near the deposition

surface, the temperature, velocity and concentration vary significantly [Kod94].

Use of an impinging jet to feed reactants onto a small susceptor results in

stagnation point flow. The stagnation point is at the center (origin) of the surface, where

flow velocities are zero (where y=0 in Figure 1-11). With adequate pressure, a shear

layer of uniform thickness develops near the surface of the susceptor [Whi91]. The shear

layer, also called the momentum boundary layer, is the relatively stagnant region between

the surface of the substrate and the region where the fluid velocity (u) reaches 99 % of

the free stream velocity (U., cm/sec). Reactants and by-products travel to and from the

substrate surface, respectively, by diffusing through this layer. This boundary layer (SM)

displaces the outer inviscid flow away from the substrate surface, as depicted in Figure 1-

11. Thickening of the low-velocity shear (boundary) layer due to viscous diffusion is








balanced by thinning of this layer due to acceleration of the outer inviscid stream, which

leads to uniform boundary layer thickness.

The momentum boundary layer thickness (8M, cm) in stagnation point flow is

approximated by Equation 1-10:


6M = 2.4 VD5 (1-10)


where v is the kinematic viscosity (cm2/sec) and Ds is susceptor diameter (cm) [Whi91].

More information on the dynamics of stagnation point flow may be found elsewhere

[Whi91].












--- -- -- -- -- -- ---- -------- -- -





Figure 1-11. Stagnation-point flow [Whi91].

In addition to the momentum boundary layer, concentration and thermal boundary

layers, with thickness denoted 8c and &r, respectively, are also defined. The

concentration boundary layer extends from the substrate surface to a point where

precursor concentration is 99% of the bulk concentration, and likewise, the thermal

boundary layer extends to a point where the temperature is 99% of the bulk temperature.









The thickness of the boundary layers depends on several process variables,

including gas velocity, temperature, and pressure. Gas phase species outside of these

boundary layers (where y > SM, 8c, or &r) move by both convection and diffusion.

Within the boundary layer (where y < 6M, 6c, or &r), velocity, concentration, and

temperature are non-uniform. As species leave the gas and deposit onto the substrate,

their gas phase concentration drops near the substrate surface. To compensate for this

drop in concentration near the surface, a net diffusive flux occurs bringing depositing

species from regions of higher concentration in the boundary layer to the near-surface

region. As the surface reaction generates reaction by-products, the concentration of

by-products near the substrate surface is higher than it is at locations farther from the

surface. This higher concentration of by-products near the substrate surface leads to a

net diffusive flux of by-product away from the substrate. Likewise, temperature varies

from the substrate into the gas phase. The temperature at the substrate is high, and this

drops steadily through the thermal boundary layer, until it reaches the bulk value.

Typical CVD temperature and concentration profiles are depicted in Figure 1-12.


TB PiB




-t----- ------X -----------_____ ;_________

T, P.
Substrate


Figure 1-12. Variation of temperature and precursor partial pressure near the substrate in
a CVD reactor. Note that while &r and 6c are shown as being equal, this is
not always true.








Transport to the substrate during CVD is assumed to occur by diffusion of

reactants through the concentration boundary layer at the surface. This boundary layer

arises due to consumption of the precursor at the surface, which causes a concentration

gradient to form between the surface and the bulk. Flux of reactant through this layer (J,

kgmol/m2-sec) is given by Equation 1-11 [Kel91]:

SDo (Ts TbXPb P)
cToRln(TT (1-11)


where Do is the diffusion coefficient (m2/sec) at temperature To (K), 8c is the thickness of

the concentration boundary layer (m), Pb and Ps are the bulk and surface partial pressure

of the reactant (Pa), Tb and Ts are the bulk and surface temperature (K) and R is the gas

constant (m3-Pa/kgmol-K). Once the reactant species diffuse across the boundary layer,

they may or may not be incorporated into the growing film surface. The mass transfer

flux (Jtr, kgmol/m2-sec), which is the amount of precursor incorporated at the film

surface, is defined in Equation 1-12:

k,(P -Peq)
Jti = (1-12)
RTs

where kd is the mass transfer coefficient (m/sec) and Peq is the equilibrium precursor

partial pressure (Pa) at the surface. When the deposition process reaches steady state, the

diffusion and mass transfer fluxes are equal, and a dimensionless parameter called the

CVD number, NCVD, can be defined [Kel91]:


(Pb -P) ka6cTln(T T
CVD (Pe-P ) DTs(Ts -Tb) (1-13)









The denominator on the left-hand side of the equation represents the degree of precursor

supersaturation at the substrate surface. When the amount of supersaturation is small,

Ps=Peq and NCVD >>1, hence all reactant reaching the substrate reacts immediately. This

is known as the mass transfer controlled regime. Conversely, if the supersaturation is

large, Ps approaches Pb and NCVD <<1, and the species reaching the substrate react very

slowly. This is known as the kinetically (or surface reaction) controlled regime. When

NCVD >>1, high points on a rough surface have a higher value of Ps, and the growth rate

at a high point is increased relative to a low point. This causes rough points on the

surface to be amplified by film growth when operating in the mass transfer controlled

regime. In contrast, when NCVD <<1, Ps approaches Pb everywhere along the surface, so

that the growth rate is similar at all points along the surface. This leads to a smoothing of

rough spots during film growth when operating in the kinetically controlled regime.

The type of carrier gas used in the reactor can have a significant impact on the

films grown during CVD. For fixed bulk and surface temperatures, gases with a lower

thermal conductivity will have a sharper temperature profile, meaning less upstream

heating of the precursor species as it approaches the substrate. Using a carrier gas with

higher thermal conductivity results in a smoother temperature profile, enabling more

upstream heating of the precursor. Increased upstream heating promotes pyrolysis, or

gas-phase thermal breakdown of the precursor into the intermediate or final reactant

species. Increasing the time available for precursors to undergo pyrolysis increases the

likelihood that the precursors have decomposed into the intermediate components (if

necessary) for final surface reaction. If the precursors undergo sudden pyrolysis very








near the substrate surface, they will not decompose as thoroughly, leaving large

molecular fragments that can be incorporated into and contaminate the growing film.

Ideally, transport of reactants to and products from the substrate surface should

occur in such a way to deposit a "clean" film containing the desired components. In

reality, this is unlikely, with atoms from reaction by-products and precursor ligand

fragments typically depositing to some degree in the films. Control of contamination by

ligand decomposition is extremely important to get films with desirable properties. An

example of single-source CVD, which has some contamination from precursor ligands,

is depicted in Figure 1-13.


S Atom desired in film
--- --- Atom not desired in film



Pyrolysis

S---- 0




S/ Surface
x easel a e
Substrate


Figure 1-13. Diagram of the CVD process for a single-source precursor.

1.6.3 Atomic Layer Deposition (ALD)

Standard CVD techniques introduce all required precursor gases to the reactor

simultaneously, making deposition rate and precise thickness control difficult during film









deposition. Although the ITRS roadmap suggests that CVD deposition of barriers will be

important in the near term, atomic layer deposition (ALD) methods will emerge as the

dominant solutions because of their superior conformality and improved thickness

control. As barrier thickness drops below 100 A, standard CVD techniques will approach

their usability limit due to difficulties controlling the deposition rate. Atomic layer

deposition (aka ALD, ALE or ALCVD) will be necessary to deposit extremely thin films

on the IC devices [Dan02], and its use in IC production at the 65 nm node is projected

[Bey02, Cha04]. The highly conformal, ultra-thin barriers afforded by ALD will be

essential in the future to minimize the barrier's impact on the resistance per unit length in

Cu interconnects [Kap02].

As a variant of CVD, ALD is well suited for deposition of ultra-thin, highly

conformal films over small device features. ALD enables monolayer addition with

precise thickness and composition control, irrespective of the underlying substrate's

topography [Dan02]. ALD has been used to deposit metal, semiconductor, dielectric and

seed layers [Dan02, Kla00a-b, Les02]. In particular, ALD has been used to deposit

high-k gate dielectrics at the front end and diffusion barriers at the back end of the

process. ALD is a "digital" process, involving the stepwise use of two or more gas phase

precursors, each of which is self-limiting on the substrate [RosOl]. A single precursor

gas is present in the reactor at any given time, so that a uniform layer of the precursor

may chemisorb to the substrate surface. Once this chemisorbed layer forms, the chamber

is evacuated or swept with inert gas [Sun92]. By minimizing chamber volume, rapid

deposition and quick purge/vacuum steps are promoted. This is important to minimize

the incorporation of background impurities into the film and also to prevent the process









from shifting to CVD mode, which can occur if additional reactant from the previous

pulse is in the reactor's atmosphere during the subsequent pulse. A second precursor gas

is then introduced to react with the chemisorbed layer from the previous step, forming a

new monolayer of material. The second precursor gas is also self-limiting, so that

reaction stops after the monolayer of chemisorbed material from the previous step has

been consumed. Film thickness is more difficult to control for growth with CVD, which

has been reported to close extremely narrow vias rather than depositing conformal films

in them [HauOO].

ALD was first reported in 1977 by Tuomo Suntola to deposit ZnS films for

electroluminescent displays [Goo86]. Since this first report, the utility of ALD to deposit

a variety of materials has been demonstrated. Materials such as metals (e.g., W, Cu, Ni,

Co, Ti, Ta, Ru, Pt, Al), metal oxides (e.g., A1203, ZrO2, HfO2) and metal nitrides (e.g.,

TiN, TaN, WNx) have been deposited using this technique [Les02]. In contrast to the

thin, flat films typically desired by ALD, the process has also been demonstrated to coat

porous, high surface area substrates for catalysis [Hau93].

ALD exploits the difference in energy between chemisorption and physisorption

to deposit film layers in a self-limiting fashion [Gat96, Goo86]. Temperature is selected

so that Echemisorption > kT > Ephysisorption. In other words, the temperature is high enough to

overcome physisorption forces to enable desorption of any physisorbed species, but it is

also low enough to prevent removal of chemisorbed layers. For an adequate pulse time,

complete surface reaction occurs and the film composition is determined by

thermodynamics (stable thermodynamic phase forms for the selected conditions).









The precursors must be volatile, thermally stable, self-limiting and highly

reactive (fast, complete reactions) with each other and with the substrate. They must also

have sufficient purity, have unreactive by-products, and not undergo self-decomposition

or cause etching of the film or substrate. The precursors should have a vapor pressure of

at least 0.1 Torr for delivery during ALD. Desirable ALD reactions should have large

negative AG values to ensure rapid surface reaction after the precursor pulse into the

reactor [Les02]. High reactivity enables rapid saturation of the film surface, which in

turn enables a good deposition rate. The need for high precursor reactivity is contrary to

CVD precursors, which require smaller negative AG values to prevent gas phase

nucleation and particle formation. The precursor dose must be high enough to saturate

the substrate surface, which must have reactive adsorption sites [HauOO]. In addition, the

reaction temperature must be chosen to enable reaction chemisorptionn) between the

precursor and reactive sites, while also avoiding gas phase decomposition or

condensation of the precursors [HauOO].

The advantages of ALD relative to CVD are excellent conformality (-100%),

inherent elimination of pinholes, good thickness uniformity, and elimination of potential

gas phase reaction/nucleation [Goo86, Rit03, Sun92, Sun93]. The disadvantages of ALD

include slow growth rate (0.06-0.6 pm/hr) [Goo86, Ros01], the possibility for substantial

incubation time to deposit the first monolayer of film, contaminant incorporation and the

potential for surface reconstructions to adversely affect deposition rate [RosOl, Sun93].

In the most aggressive applications, a diffusion barrier will need to be deposited on four

different surface materials simultaneously: two different insulator materials, an etch stop

layer (such as Si3N4) and Cu [Ele02]. Varying incubation times on the different materials








can cause significant deposition difficulty, so choice of precursor and deposition

conditions is essential to deal with situations like these. Precursor testing should

therefore be done on all possible underlying substrates, to ensure that a given material

can be reliably deposited with a given precursor. Another disadvantage is decreased film

purity relative to CVD grown films. Background contaminants have considerable time to

incorporate into the film during the pulse and purge steps, which can degrade film

structure and electrical properties. The purity of purge gases and control of out-gassing

from reactor walls and seals are critical, because inadequate control can lead to

considerable impurity incorporation causing modification of film structure and properties.

Typically, a faster growth rate means less sensitivity to contaminant incorporation

[Han03]. ALD must compete with the purity of PVD films, whose growth rates are 2 to

3 orders of magnitude higher in an environment that is 3 to 6 orders of magnitude cleaner

[Han03].

Several adjustable parameters are available for the ALD process. The first

parameter is the type of ALD reactor system used for deposition. The two main

categories for ALD systems are the open type (used with molecular beam epitaxy

(MBE)) and the closed type (used with LPCVD) [Sun93]. In the open type system,

pressure is fixed, as with an MBE system, which typically operates at ultra-high vacuum

(UHV). Precursors are introduced from two or more sources (e.g., Knudsen cell), which

are turned on and off to generate the pulsing action. Desorbed surface species collect on

the cold walls in the reactor system or by the vacuum pump. In the closed type system,

the atmosphere is turned over between reactant exposures to generate the pulsing action;

this is done either by evacuating the system (pressure modulation) or by introducing an









inert purge gas. In addition to the type of system used, the chamber volume may be

adjusted to accommodate more substrates (i.e., to increase throughput) or to decrease

purge/vacuum times.

The second adjustable parameter is the type of precursors used for the deposition

process. The precursors must be self-limiting, very reactive with each other and with

substrate, and generate the desired film structure/stoichiometry. Many precursors have

been tested in the literature for various materials, with halide chemistry often being used

due to easy removal of the halide ligands (with H2, NH3, etc.). In addition to precursor

selection, pre-treatment of one or more of the precursors before introduction to the

reactor is another adjustable option. For example, Ta films were deposited using

plasma-enhanced ALD with sequential pulses of TaCl5 and H2 [Kim02b]. The H2 was

cracked to atomic H in an rf plasma system before being introduced to the reactor.

The third parameter is substrate temperature, which should be selected to enable

chemisorption of first layer and desorption of outer, physisorbed layers [Gat96]. A fourth

parameter is the reactant pulse time. A longer pulse time enables more complete surface

reaction, and closer approach to thermodynamic equilibrium at the film surface. The last

parameter, which is unique to closed type systems, is reactor pressure. ALD process

pressure in closed type systems typically ranges from 1 to 10 Torr [Les02], which can be

modulated to evacuate the chamber if an inert gas purge is not desired.

The process is depicted in Figure 1-14. The result of running in ALD mode is a

linear growth rate with the number of cycles. The number of monolayers grown per

cycle (normally <1), multiplied by the number of cycles, gives the film thickness, where

one cycle includes one pulse of each of the precursor gases. Several growth cycles are








typically necessary to deposit a monolayer of material, because steric crowding on the

film surface prevents 100% surface coverage for each pulse. In practice, monolayer by

monolayer growth is unlikely, and there may be two or more monolayer levels growing

together during deposition. Film growth is dependent on several factors, including

reactivity of the precursors, the number of reactive sites available on the substrate/film

surface, and the size of the precursor molecule (larger molecules typically yield lower

growth rates due to steric hindrance during the reaction cycle) [HauOO]. ALD cycle time

typically ranges from 0.5 to 5 seconds.


A-X A-X X X
A-X A-X X AX A







B--Y X-Y X-Y
B-Y
BB X-Y
B-Y X-Y
S X B B
SI
B B
X X A A
A A I
A




Figure 1-14. Diagram of the ALD process. (a) Introduction of first precursor. (b)
Absorption of first precursor. (c) Introduction of second precursor. (d)
Complete monolayer deposition of A-B.

The growth rate for an ALD process should be linear with the number of growth

cycles. If this is not the case, the process may not be self-limiting. To check this,

reactant pulse time can be changed to determine if the experiment is self-limiting, or if









longer pulse times result in thicker films. ALD growth reportedly occurs by formation of

islands during nucleation, hence the barrier should have thickness at least equal to that

required to close the surface of the substrate [Bey02]. This closure thickness depends on

the deposition conditions and the barrier materials being used. Electrical measurements

(such as 4-point probe) can be used to determine if the deposited film is continuous.

Films may have high resistivity until enough cycles have passed that the film is

continuous; the film's resistivity will then drop substantially. This is one way to

determine if the first monolayer has been completely filled in, if depositing a conductive

layer on Si, for example. Extrapolating from a plot of resistivity vs. the number of

deposition cycles, one can estimate the deposition cycle after which a continuous film is

finally deposited. Ion scattering spectroscopy has also been used to determine when the

surface has closed.

The appeal of ALD to industry lies in its ability to manage line resistance and to

provide better stress migration performance than PVD films [Pet03]. A thin ALD barrier

film at the via bottom is instrumental in lowering via resistance. In addition, Cu seed

layers must be ultra-thin, continuous, and have excellent conformality to ensure

consistent Cu deposition during ECD, so ALD will be essential for seed deposition. ALD

is also excellent for varying the composition of the deposited film with thickness

(composition grading), which can be important to ensure a film's compatibility with

underlying layers.

1.7 Copper Deposition Methods

Electrochemical deposition (ECD), also known as electrodeposition or

electroplating, is the technique currently used in industry to deposit void-free bulk Cu









layers on ICs with aggressive topographies [ITR02]. This technique has the advantages

of low deposition temperature, high deposition rate, and low manufacturing cost [Chi98].

The ECD process begins with the deposition of a Cu "seed layer" on the barrier surface,

because the barrier layer's resistivity is typically too high to enable uniform

electrodeposition. Seed layer deposition is currently done by PVD, which is projected to

be used down to the 45 nm device node [Pet03]. Once the seed layer is deposited, the

sample can be immersed in the electrochemical bath and the ECD process can begin.

During Cu ECD, an electrolyte containing Cu cations and sulfate anions (Cu2+ and SO42,

respectively) is put into contact with the desired deposition surface. Electrons are

introduced to the deposition surface (called the cathode) and reduce the Cu2+ ions in the

ECD solution, causing these ions to plate out as metallic Cu. A Cu seed layer deposited

by PVD (prior to the ECD process) is typically used as the cathode, on which Cu2+ ions

plate out to form Cu(0). The plating reaction is:

Cu2+ + 2e- Cu(0) (1-14)

To complete the electrical circuit, an anode is placed in the electrolyte solution.

The anode introduces current to the solution, which replaces positive charges lost by the

Cu plating reaction, enabling the electrolyte to remain electrically neutral [ReiOO]. Since

Cu2+ ions are present in the plating solutions for ECD, backside protection of the Si wafer

from the bath is essential during deposition. Failure to do this can result in massive Cu

in-diffusion through the backside of the wafer, which will occur rapidly based on the

diffusivity of Cu in Si (as discussed above). At the industrial level, the ECD plating tools

have an o-ring around the edge of the wafer, which allows the device side of the wafer to

contact the plating solution while preventing the solution from touching the backside. At








the research lab level, waxes or glues may be used to protect the backside of the

substrate.

The use of special additives enables the ECD technique to fill small features from

the bottom up with Cu (superfilling). To ensure deposition of a microvoid-free, uniform

thickness Cu layer by ECD, the plating current (and therefore the seed layer thickness)

must be essentially constant from edge to center across the wafer [Sin02a]. Hence, high

uniformity of the Cu seed layer is essential. The seed layer is usually deposited on the

barrier surface by sputtering, however, which is characterized by poor step coverage in

small features [And99, Wan03]. A Cu seed layer with poor step coverage leads to poor

Cu coverage during ECD. To overcome the conformality challenges associated with

sputtering, several different approaches are being pursued.

The first approach is to deposit Cu seed layers by CVD [NorOl]. Generally, Cu

CVD precursors are classified into two categories: Cu(I) and Cu(II) compounds. Cu(I)

compounds, such as Cu(hfac)(TMVS), where hfac is hexafluoroacetylacetonate and

TMVS is trimethylvinylsilane, are typically more volatile, and can be used without a

carrier gas or a reducing agent [NorOl]. These compounds tend to be liquids and are

capable of deposition below 2000C, but are therefore more reactive and less thermally

stable, making control of the deposition rate more difficult. Copper(H) compounds, such

as Cu(hfac)2, are more thermally stable, but typically require a reducing agent (such as

H2) to remove the ligands during deposition [Kod94]. In addition, these compounds tend

to be solids, requiring a deposition temperature higher than 2500C [NorOl]. The major

issue for Cu CVD from both Cu(I) and Cu(II) precursors is incorporation of halides and

carbon from the precursor ligands into the Cu film, which can have detrimental effects on








film properties such as adhesion and resistivity. More detail on Cu CVD precursors and

deposition is available elsewhere [Kod94].

Deposition of a Cu seed layer by CVD followed by PVD and reflow of Cu has

also been reported [Fri99a, Jai99]. Cu reflow typically involves the deposition of Cu by

PVD methods, especially sputtering, followed by subsequent annealing of the device.

The annealing process enhances surface diffusion of Cu, which enables Cu transport into

and filling of small features [Fri99a]. The driving force for this diffusion is the chemical

potential gradient associated with differences in surface curvature along the surface of the

deposited Cu. To minimize the total free energy on the Cu surface, Cu atoms diffuse

along the surface from convex regions to concave ones [Fri99a]. The net result is

thinning of convex overhang regions at the tops of features and filling of concave parts of

the features, such as via/trench sidewalls and bottoms. This diffusion-mediated process

is good for movement of Cu to fill submicron feature sizes. Anneal time can be modified

by changing the anneal temperature, but the anneal temperature cannot be higher than the

temperature stability limits for the device, and must be low enough to prevent bulk Cu

diffusion, which can lead to delamination and a change in Cu texture. Typically, Cu is

annealed for 13-14 minutes at 450C during the reflow process [Fri99a].

Another approach is to deposit Cu seed layers by ALD. Several different

precursors have been tested for ALD Cu seed deposition, including Cu(hfac)(TMVS),

Cu(hfac)2, Cu(thd)2, [Cu(C3H7)NC(CH3)N(C3H7)]2, and CuCl, where thd is tetramethyl

heptanedionate [Kod94, Lim03, Nor01]. The first four compounds are bulky,

metalorganic sources, which have a very low growth rate per cycle due to steric

hindrance. In addition, the first two compounds contain F, making the possibility of F








incorporation into the deposited Cu film an issue. The last compound, CuCl, is a solid

with very low vapor pressure, hence transport to the reactor is very challenging. [Huo02,

Jup97, NorO1].

Electroless Cu plating, which deposits Cu without use of a sputtered Cu seed

layer, may also be used [Wan03]. This technique involves deposition of Cu from an

ionic solution, where the deposition surface catalyzes a redox reaction, without any

external electrodes [Sha95]. A specific component from the solution (e.g., glyoxylic

acid) serves as a reducing agent, and is oxidized on the catalytic surface, releasing one or

more electrons [Wan03]. These electrons reduce Cu' ions from the solution, causing the

Cu to plate out as a film on the deposition surface. While this technique has been used to

deposit Cu seed layers with good conformality and low resistivity (1.7 jiQ-cm), and has

also been used to do Cu filling [Lee98a], control of the deposition rate is not as precise as

sputtering. More detail on this deposition technique is available elsewhere [Sha95,

Sha97].

The last approach is seedless ECD [ITR02, Sin02a], where Cu is directly

electroplated onto the underlying barrier material using an external electrode, but without

any seed layer. This technique is the most technologically and economically promising,

because it eliminates the seed layer deposition step from the current IC metallization

process and does not require new equipment expenditures. Selection of a diffusion

barrier material that enables seedless Cu electrodeposition could result in significant cost

savings to the metallization process.








1.8 Statement of Problem

The three main challenges of shifting to Cu interconnects on IC devices are: a) Cu

deposition technique b) patterning method for the Cu layer, and c) which barrier material

and deposition method will prevent Cu-Si interdiffusion while fostering adhesion to

neighboring layers [And99].

The first challenge has been addressed by use of ECD. Since Cu is resistant to

chemical etching, the application of standard reactive ion etching (RIE) techniques used

to pattern Al is not feasible for Cu. The second challenge, patterning of Cu

interconnects, has therefore been addressed by the dual-damascene deposition process.

In this process, Cu is deposited across a pre-patterned wafer by one of the

aforementioned deposition techniques. Chemical mechanical polishing (CMP) uses a

grinding pad and slurry to remove excess Cu, leaving only the desired spots (e.g.,

trenches and vias) filled with metal. Since the dual-damascene process deposits Cu in

both trenches and vias, the need for tungsten (W) plugs at the upper levels of the IC,

which are used to fill vias on Al based devices, is eliminated. Removing this W plug

lowers the electrical resistance of the device, both because Cu has a lower resistivity than

W and because the Cu-W interface is eliminated.

The third challenge, barrier material and deposition technique, is an ongoing issue

as device features continue to shrink. Ti/TiN barriers, deposited by sputtering, are still

used at the contact level with W plugs, where TiN protects the contact from F in the WF6

precursor and improves W adhesion [ITR02, Kal00]. Ti reacts with Cu to form cuprides,

making Ti thermodynamically unstable as a diffusion barrier for Cu [RamOO]. Moreover,

if TiN films are Ti rich, the excess Ti can react with Cu and lead to failure of the TiN








barrier. In addition, TiN barriers with thickness below 20 nm reportedly fail to prevent

Cu diffusion due to grain boundary diffusion [Kal00].

A Ta/TaN dual layer barrier structure is currently used as the Cu diffusion barrier

at the intermediate and upper wiring levels on IC devices. Cu has good adhesion to Ta,

which encourages deposition of low-resistivity Cu (111) on its surface [Pet03], while

TaN has good adhesion to the dielectric layer [ITR02, Sin02a]. The Ta/TaN dual layer

barriers are deposited by modified sputtering methods [ITRO2, Sin02a], however, which

have a finite lifetime due to aforementioned step shadowing issues. Barrier deposition

takes place on the industrial scale in a cluster tool, which contains several chambers

connected by a common vacuum transfer chamber. The barrier is deposited in one

chamber, and the wafer is then moved in-vacuo to a separate Cu seed layer deposition

chamber. Once the Cu seed layer is deposited, the wafer is pulled out of the cluster tool

and put into a separate system for Cu ECD.

The "holy grail" of diffusion barrier research is a robust material / deposition

method couple that meets all of the above listed barrier requirements and is extendable to

future device generations. In addition, a material/deposition pair which would enable Cu

metallization to extend from the upper levels of the device down to the contact level

would eliminate the need for Al metallization at the local level and the use of W contact

plugs, yielding greater device speed. The interconnect structure's features (trenches and

vias) become increasingly more aggressive (narrower diameter and higher aspect ratio) as

the wiring levels get closer to the contact level. Research should therefore focus on

meeting coverage requirements at device level feature dimensions to enable future

extension of Cu throughout the IC.









1.9 Hypothesis

It is evident that a variant of CVD will be required to meet future demands for

conformal diffusion barrier deposition on aggressive IC device topographies. New

precursors will be required to meet the increasingly stringent demands on film properties.

In an effort to avoid halide incorporation, synthesis and testing of novel WNx and WNxCy

MOCVD precursors were pursued in this work, for reasons that will become clear in

Chapter 2.

Depositing WNx and WNxCy films from novel precursors by MOCVD should

minimize halide content, have low temperature deposition, enable good adhesion to

neighboring layers, and enable a single-step CMP process. In addition, these materials

appear to be useful for direct, seedless electroplating, and have been reported to resist Cu

diffusion. WNx is also reported to deposit in amorphous form more easily than TaNx

[RamOO], and the addition of C to form WNxCy should enhance the ability to deposit

amorphous films even further.

Eventually, a shift from CVD to ALD will be necessary to meet conformality

requirements in small device features. Hence, our strategy is to initially pursue WNx and

WNxCy films by MOCVD and to eventually transition to precursor development and

testing for ALD.

The film growth and procedure and characterization techniques used to analyze

the films will be discussed in Chapter 4. Data, analysis and conclusions for each

precursor will be presented in a separate chapter, while the thermodynamic analysis for

the W-N-C-H-Cl system is presented in Chapter 3. Cu diffusion results for barriers






52

deposited with an isopropylimido-based W complex will be discussed in Chapter 8, and

future work will be discussed in Chapter 9.















CHAPTER 2
REVIEW OF THE LITERATURE

From a thermodynamic standpoint, a barrier material must be chosen so that the

chemical potential (the thermodynamic description of reactivity) at each interface ensures

little or no reaction. A major thermochemical difference between Al and Cu is their

reactivity with silicon. Al can exist in equilibrium with Si, while Cu cannot, readily

reacting to form silicides [RamOO]. This non-equilibrium suggests that the

thermodynamic driving force for diffusion in the Cu-Si system is larger than that in the

Al-Si one. An ideal barrier separating Cu from Si must therefore be stable to and

non-reactive with both Cu and Si. Moreover, the barrier should be stable with other

potential materials in the IC, such as low-k and high-k dielectrics.

Many materials have been examined to determine their potential as diffusion barriers

for Cu metallization schemes. A comprehensive list of various materials tested as

diffusion barriers can be found in previous review articles [Jai99, Kal00, Kim03b].

2.1 Cu-Si Interconnects without a Barrier

Ideally, the easiest route to IC device fabrication would be to deposit Cu directly

onto any neighboring layers. Hymes et al. [1998] formed copper silicide compounds by

sputtering Cu onto thin films of Si, with the goal of using these compounds as passivation

layers to prevent further Cu-Si interdiffusion and reaction. At room temperature, Cu3Si

and CusSi formed, while all Cu3Si decomposed to CusSi after annealing at only 3000C.

The CusSi compound disappeared after annealing at 5000C, leaving only pure Cu and








dissolving the Si into the Cu layer. The inability to form a thermally stable Cu-Si

passivation layer highlights the instability of the Cu-Si interface and underscores the

need for a diffusion barrier to separate these materials.

2.2 Refractory Metal-Based Barriers

The term refractory refers to a material with a melting point above 1800C

[Pie96]. Due to their low self-diffusion coefficients, refractory metal-based materials

have been frequently studied for use as diffusion barriers [Liu00], with current refractory

metals of interest for diffusion barrier applications including Ti, Ta, W and Ru.

2.2.1 Unary Refractory Metal Barriers

Unary refractory metals have been explored as potential barrier materials, due to

their low self-diffusivity and low electrical resistivity. Some drawbacks associated with

these materials include their tendency to crystallize (enabling grain boundary diffusion)

and to react with Si at temperatures greater than 4500C [JohOOa]. The applicability of

several unary refractory metals as barriers will be discussed in the following sections.

2.2.1.1 Ti Barriers

While Group V and VI transition metals (e.g., Ta and W) are stable and

non-reactive with Cu, Group IV transition metals such as Ti are unstable, forming

several cuprides, including Cu4Ti, Cu4Ti3, CuTi and CuTi2 [RamOO]. For this reason, Ti

is not a viable choice for a Cu diffusion barrier, as its reaction with Cu would degrade the

barrier structure and lead to rapid device failure.

2.2.1.2 Ta Barriers

Tantalum (Ta) does not react with Cu [Wan94], and has demonstrated good

oxidation resistance and enhanced Cu (111) texturing during Cu ECD [Chi02]. Ta








generally fails to suppress Cu diffusion, however, because grain boundaries readily form

during deposition and subsequent thermal cycling of the Ta layer. In addition, Ta reacts

with Si to form silicides, making the Ta-Si interface unstable, and disqualifying Ta metal

as a potential single barrier layer material to separate Cu and Si.

2.2.1.3 W Barriers

Tungsten (W) is another transition metal that does not react with Cu [RamOO,

Wan94]. It does, however, react with Si [RamOO]. W reportedly failed as a Cu barrier

due both to a silicidation reaction above 6700C [Tak97b] and crystallization, which

causes grain boundary formation [Cha97c]. Sputtered W barriers have also been reported

with columnar grains, which extended across the entire barrier thickness and caused

barrier failure above 7000C [Mer97]. While PECVD W has been reported with very low

resistivity (10 p-cm) [Cha97b], films deposited at 3500C had an open grain boundary

structure, which led to a low activation energy (0.46 eV) for Cu diffusion [Gup95].

Metallic W has also been deposited by ALD, using WF6 and Si2H6 precursors over a

temperature range of 30 to 3500C [Ela01]. Below 1000C, some Si surface species were

left on the surface and remained unconsumed by the WF6 precursor exposure.

2.2.1.4 Ru Barriers

Ru, like Ta and W, is non-reactive with Cu [Chy03], and has been tested as a Cu

diffusion barrier material. The tested Ru barriers were deposited by PVD methods,

however [Jos03], which have a limited lifespan, as mentioned in Chapter 1. In addition,

Ru silicides are reported to form after annealing at 5000C [Jel03]. A sputtered Ru layer

sandwiched between two TiN layers reportedly did not enhance the barrier's ability to

prevent Cu diffusion [Kim03c]. Ru deposited by CVD has been studied for use as the








metal electrode for DRAM capacitors [Aoy99, LeeOOa, Mat02], and Ru has also been

deposited by ALD [Aal03]; reports of CVD and ALD Ru as Cu diffusion barriers,

however, have not been given.

2.2.2 Binary Refractory Metal-Based Barriers

To overcome the limitations associated with unary refractory metal barriers,

binary compounds containing Ti, Ta and W have been widely studied. The addition of a

second, nonmetal element to the refractory metal tends to increase the likelihood of

depositing amorphous films, because the nonmetal element disrupts the crystallization

process. The non-metal also generally increases film resistivity, however. The

properties and applicability of several refractory metal carbides and nitrides will be

discussed in the following sections.

2.2.2.1 Refractory Metal Nitrides (M-N)

Adding nitrogen to refractory metals gives rise to refractory metal nitrides, where

the term nitride refers to compounds formed between nitrogen and other elements with

equal or lower electronegativity [Pie96]. Three characteristics play a role in the

formation of metal nitrides: the size ratio of nitrogen to the other element, the

electronegativity difference between nitrogen and the metal, and the electronic bonding

characteristics between nitrogen and the metal [Pie96]. Below a N/metal radius ratio of

0.59, interstitial nitrides form, while above 0.59, covalent nitrides (such as Si3N4)

typically form [Pie96]. The early transition metals (Groups IV, V and VI) have a large

enough host lattice to enable formation of interstitial nitrides, where the nitrogen atoms

sit on interstitial sites in the metal lattice. The atomic radius of N is 0.74 A, while that for

W, for example, is 1.394 A, yielding a N/W radius ratio of 0.53. The nitrogen and metal








atoms have a large electronegativity difference, so the N atoms "nest" in the interstitial

sites of the metal lattice [Pie96]. Interstitial nitrides have bonding with a combination

of metallic, covalent and ionic character, high hardness and strength, and high thermal

and electrical conductivity. These nitrides tolerate nonmetal vacancies, making them

susceptible to the presence of interstitial impurities such as oxygen. Most early transition

metals have a BCC structure, which cannot accommodate significant nitrogen levels at

the interstices. To form an interstitial nitride, the metals switch to a close-packed

structure, such as FCC or HCP, to ensure adequately sized interstitial sites that can

accommodate N. Close-packed FCC transition metals have both tetrahedral and

octahedral interstitial sites, but since the tetrahedral sites are too small to accommodate

the N atoms, N only occupies the octahedral sites [Pie96].

Nitrogen addition suppresses crystallization in the films, and helps to repel the

advance of Cu through the grain boundaries, due to a repulsive interaction between Cu

and N [EksOl, Tak97a]. Excess nitrogen in these films migrates from the bulk

polycrystals to the grain boundaries, where repulsive interactions between Cu and

nitrogen "stuff" the boundary and halt Cu diffusion [EksOl, Sha89]. Formation of copper

nitride (Cu3N) by reactive sputtering has been reported, but this compound is unstable,

decomposing to Cu and N under vacuum at 1000C [Liu98, Mar95].

Refractory metal nitrides tend to have high hardness, good chemical stability and

high conductivity [Lev98, Nag93]. Three nitrides of current interest are FCC titanium

nitride (TiN), FCC tantalum nitride (TaN), and FCC tungsten nitride (WNx). While FCC

TiN and TaN have melting points of 2950 and 3093C, respectively [Pie96], FCC WNx

does not melt, but instead decomposes to BCC W and N2 gas under vacuum at elevated








temperature (850C) [Suh01]. Since this decomposition temperature is substantially

higher than the typically processing window for ICs, FCC WNx is a viable diffusion

barrier candidate material. The barrier properties of TiN, TaN and WNx are discussed in

more detail below.

2.2.2.1.1 TiN

Titanium nitride (TiN) was widely used as a diffusion barrier for the Al-Si

system, and its reported resistivity ranges from 20 to 2000 pL2-cm [Par96]. PVD has

dominated TiN barrier deposition, resulting in non-uniform coverage of high aspect ratio

structures and a columnar grain structure [Gal99, Kal00]. TiN performs poorly with Cu

metallization due to this columnar microstructure [Nic95], which enables rapid Cu

diffusion through the barrier [Gal99]. Plasma treatment and exposure to SiH4, however,

have been reported to improve TiN's resistance to Cu diffusion [Jos02]. Another

difficulty involves the reactivity of Ti rich TiN films with Si and Cu. For TiN films with

a Ti/N ratio > 1 (i.e., film is unsaturated with nitrogen), the excess Ti will react with Si or

Cu, leaving the barrier permeable to Cu and the device vulnerable to rapid failure

[Kim92, Lee94a, RamOO].

To avoid the problems associated with PVD TiN, several different CVD

techniques have been tested. Lu et al. [2000] used LPCVD of TiC14 + NH3 at high

deposition temperature (6300C), but the resulting films had a columnar grain structure

and suffered from halide incorporation. MOCVD TiN has shown only 40-50%

conformality in small contact structures [Fal98], has high impurity incorporation, and

requires an additional plasma-processing step to stabilize the films in low resistivity

form [Gal99]. MOCVD TiN films deposited below 4500C were very porous, with low








density and high resistivity, and had poor resistance to Cu diffusion [Par95b]. Capping

the film with an ultra-thin Si3N4 layer has also enhanced performance of porous TiN

films deposited by MOCVD of tetrakisdimethylamido tantalum (TDMAT). This thin

capping layer improves barrier capability without significantly increasing the film

resistivity, but adds another step to the barrier deposition process [Lu98a].

While oxygen stuffing of the TiN barrier (due to air exposure) helps it to resist Al

in-diffusion (due to aluminum oxide formation), it does not prevent Cu in-diffusion, due

to the lack of a stable, passivating Cu oxide layer [Par95a]. In addition, TiN was not

suitable for Cu electroless plating, as its redox potential was higher than that of the Cu

electrode, preventing the initial displacement reaction from occurring at the TiN surface

[Wan03].

2.2.2.1.2 TaN

TaN is a proven Cu diffusion barrier, and is reportedly stable against Cu-Si

interdiffusion up to 7200C [Wan94]. Copper with a PVD Ta/TaN dual layer barrier is

currently used as the interconnect scheme for intermediate and upper level wiring on IC

devices. The dual layer barrier is required to overcome adhesion difficulties: Cu adheres

well to Ta, but not to TaN, while SiO2 and most dielectrics adhere well to TaN, but not

Ta. [ITR02, Sin02a]. In addition, the TaN layer promotes deposition of the low

resistivity (20-30 pQ-cm) a-Ta phase, whereas the high resistivity (180 p.2-cm) 3-Ta

phase forms when Ta deposits directly on SiO2 [Tra03]. But, since PVD TaN has much

higher resistivity (-200 to 250 iQ-cm [Sun98a, Tra03]) than Ta, the TaN layer thickness

should be held to a minimum. A 7 A thick TaN film was found to foster a-Ta deposition

and to prevent Cu interdiffusion [Tra03]. The lifespan of this bilayer deposition








technique is limited, however, due to the inability of PVD processes to deposit conformal

films at ever shrinking device dimensions.

LPCVD routes to TaN deposition have been examined as well, as this deposition

method provides superior conformality to sputtering techniques, which are projected to

reach their usability limit at the 65 nm node in 2007 [Han03]. CVD of TaN was reported

by reacting halide precursors, including TaBrs, TaF5 and TaC15, with NH3 [Kal99, Hil00].

Resistivities ranged from 395 to 5000 p.Q-cm depending on the precursor used, with the

lowest value corresponding to TaCl5 [Kal99, Che99b, Hil00]. Halide incorporation,

however, ranged from 0.5 to 4.5 at. %, with a value of 1.0 at. % for the TaC15 precursor

[Kal99, Hil00, Che99b], While plasma assisted CVD reduced the resistivity of films

deposited by TaBrs + NH3 to 150 pQ-cm, an increase in Br content up to 3 at.% occurred

[Che99a].

Attempts to grow TaN by MOCVD resulted in films with high resistivity and

some carbon contamination. When pentakis(dimethylamido)tantalum (PDMATa) was

used for MOCVD of TaN, the resulting films contained the insulating Ta3N5 phase

[Fix93]. Pentakis(diethylamido)tantalum (PDEATa) was also tested, but the films also

had very high resistivity (7000-60000 iQ-cm) [Cho98, Cho99]. MOCVD using

tert-butylimido-tris(diethylamido)tantalum (TDEATa) has also been tested, resulting in

good step coverage (-100%) but high deposition temperature (up to 6500C) and

resistivity (900-2000 pgQ-cm) [Tsa95]. Low film density and grain boundary formation

were the typical causes of barrier failure in MOCVD TaN [Kal00].

One study of an ALD TaN/PVD Ta barrier has also been given [Ho04]. The

precursors used in this study were not disclosed, but the barrier scheme was shown to








work with a 65 nm Cu back-end-of-line (BEOL) interconnect structure. The ALD TaN

barrier had good yield and reliability, and a 16% reduction in RC time delay over a PVD

TaN/PVD Ta barrier scheme was demonstrated.

2.2.2.1.3 WNx

As mentioned above, N has been added to W to deposit WNx films. These films

show promise as thermodynamically stable, stuffed barriers for separation of Cu and Si

and have been demonstrated as excellent glue layers for W/Si and W/SiO2 interfaces

[Gal97, Kim92]. Many deposition techniques, including ion implantation, sputtering,

LPCVD, PECVD, MOCVD and ALD have been used to deposit WNx films. A detailed

discussion of WNx films deposited by these techniques will be deferred until Section 2.7

below, as we have chosen to focus on this material for our work.

2.2.2.2 Refractory Metal Carbides (M-C)

Unlike the nitrides, refractory metal carbides from all three Groups (IV, V and

VI) tend to be hard, wear-resistant materials with high melting points and good chemical

resistance [Pie96]. The Group IV, V and VI refractory metals form interstitial carbides,

with crystal structures similar to those for the nitrides. The electrical resistivity of

carbides is typically lower than that for nitrides with equivalent crystal structure, due to

weaker bonding between the metal and C relative to bonding between metal and N

[Nak87]. While these properties make the carbides interesting diffusion barrier

candidates, C, unlike N, does not exhibit a repulsive interaction with Cu, making the

carbides less resistant to Cu diffusion. The barrier properties of TiC, TaC and WCx are

discussed below.








2.2.2.2.1 TiC

One report of TiC (deposited by sputtering) as a Cu diffusion barrier has been

given [WanOla]. While the TiC resisted metallurgical failure at temperatures up to

6500C, the same films suffered electrical failure after annealing at temperatures just

above 5000C.

2.2.2.2.2 TaC

Sputter deposited TaC films have also been tested as Cu diffusion barriers

[Ima97, Lau02, Mor98]. While these films had low resistivity (210 p.Q-cm for 100 nm

thick film), carbon-rich TaC films contained low-density carbon regions surrounding

TaC grains [Ima97]. These low-density regions are facile paths for Cu penetration

through the barrier, with 7 nm TaC films failing to prevent Cu diffusion above 550C

[Lau02]. This indicates that C, unlike N, lacks the ability to chemically repel Cu

diffusion through the film, and is reflected in TaC's lower activation energy for Cu

diffusion as compared to W2N or TaN [Mor98].

2.2.2.2.3 WCx

The bulk resistivity of I-WCx is slightly lower than that for P-WNx, with

reported values of ~ 41 jQ-cm and -50 dQ-cm, respectively [Lee93, Nic78]. MOCVD

WCx from W(CO)6 + C2H4 has been tested as a Cu diffusion barrier [Sun01b, Vel00].

Deposition temperatures ranged from 250 to 5100C, with 50 nm thick films deposited at

2900C having resistivity of 250 pQ-cm. A 70 nm thick WCx film annealed for 8 hours at

4000C resisted Cu diffusion. Sputtered WCx with resistivities ranging from 200 to 1000

p.2-cm were also tested as Cu diffusion barriers, with the incorporation of C resulting in

smaller grain sizes relative to pure W films [Vel00, Wan01b]. The onset of silicide








formation above 7000C, however, indicated the instability of the WCx/Si interface and

also compromised barrier integrity against Cu diffusion. Kim [2003a] reported

deposition of tungsten carbide (WCx) films by plasma assisted ALD (PAALD) using the

bis(tert-butylimido)bis(dimethylamido)tungsten [(tBuN)2(Me2N)2]W precursor with H2

and N2 carrier gas. The films were deposited on SiO2/Si substrates at 2500C, and had

growth rates ranging from 0.4 to 0.7 A/cycle, 100 % conformality in 0.15 pm features

with a 15:1 aspect ratio, and resistivities ranging from 295 to 22000 pQ-cm. The films

contained 63 at.% W, 28 to 42 at.% C, 2 to 7 at.% N, and 1 to 6 at.% O. Cu barrier

integrity tests were not reported, however.

2.2.3 Ternary Refractory Metal-Based Barriers

The addition of a third element to a binary refractory metal nitride matrix tends to

further disrupt the microstructure, increasing the likelihood of nanocrystalline or

amorphous phase formation [IstOO, Jai99, RamOO]. C, Si and B are the most examined

third elements added to metal nitride films.

2.2.3.1 Refractory Metal Carbonitrides (M-C-N)

The addition of C to refractory nitrides has two general purposes: to increase the

likelihood of amorphous film deposition, and to decrease film resistivity relative to the

binary nitride. The C and N intermix on the interstitial sublattice of the host metal, and

lattice structures are similar to those for the nitrides and carbides. The advantages of

adding C to the N on the non-metal sublattice are tempered by the decreased ability of C

(relative to N) to chemically repel Cu diffusion. The composition and microstructure of

the films must therefore be carefully controlled to deposit low resistivity films with high

stability against Cu diffusion. Free carbon (C not bound to W), in particular, should be








minimized, as it has minimal impact on Cu diffusion and detrimentally affects film

resistivity.

2.2.3.1.1 TiCN

One report of a TiCN Cu barrier, deposited by MOCVD of

tetrakis(dimethylamino)titanium (TDMAT), has been given [Eiz94a]. While the films

resisted Cu in-diffusion after annealing at 6000C, they had high resistivities, ranging

from 3000 to 20000 Q2-cm for 200 A films.

2.2.3.1.2 TaCxNy

TaCxNy films deposited by sputtering of a TaC target in an Ar/N2 atmosphere

were also tested as Cu diffusion barriers [SunOla]. TaCxNy was found to have superior

thermal stability to the respective binary phases, and its resistance to Cu diffusion was

greater than TaC due to stuffing of the grain boundaries with N. The films had relatively

low resistivity (-300 uQ-cm), and prevented Cu diffusion after a 30 min anneal at

6000C. Jun [1996] used pentakis(diethylamido)tantalum to deposit TaCxNy films by

MOCVD, but resistivity was high (2 6000 .Q-cm) and the barrier failed after annealing

for 1 hour at 5000C. MOCVD of TaCxNy films was also done using a mixture of

pentakis(dimethylamido)tantalum and pentakis(diethylamido)tantalum [HosOO]. These

films prevented Cu diffusion after a 30 min anneal at 5000C, but had high resistivity (>

4000 aQ-cm).

2.2.3.13 WNxCy

The bulk resistivity of 3-WCx is somewhat lower than P-WNx, and as expected,

the resistivity of P-WNxCy is reportedly lower than that for P-WNx. Carbon and

nitrogen in these films can deposit at interstitial lattice sites (i.e., bind to W) or as free C








or N. Given a choice, free N is more desirable, because although it increases film

resistivity (as does free C), it also repels Cu diffusion. Free C also increases resistivity,

but is less effective at preventing Cu diffusion. The difficulty, then, is to minimize or

eliminate deposition of free C, while encouraging some N to deposit interstitially and

allowing some to deposit as free N. This should give minimum film resistivity and

maximum Cu resistance.

WNxCy films were deposited by non-reactive sputtering of W-C and W-N

targets, with the resulting films being amorphous, but maintaining some short-range

order [Vie02]. These films were not tested for Cu barrier reliability, however.

MOCVD was used to deposit WNxCy as a barrier for Cu metallization in a patent

application [FukOO]. Amorphous WNxCy films were deposited by reacting a gas

containing W, such as WF6, W[N(CH3)]6, or W[N(C2Hs)]6, with a hydrocarbon gas, such

as CH4, and a nitrogen supply source, such as a nitrogen plasma, at 360C. X-ray

diffraction indicated a peak between 36 and 38 200 and a second position between 42 and

44 20, both of which are indicative of FCC WNxCy deposition. While resistivities down

to 275 pQ2-cm were reported, details of Cu barrier testing were not included.

Several studies have reported ALD deposition of WNxCy. The deposition used

sequential reactions of WF6, NH3 and Et3B [Ele03, Kim03d, Li02, Li03, Pet02, Smi02].

The lowest reported resistivity was 210 piQ-cm for a film deposited at 3500C [Ele03,

Li02]. One group reported that a 120 A thick WNxCy film was stable for a 30 min anneal

up to 7000C [Kim03d]. These ternary films had high density and excellent adhesion to

copper, but contained some amount of F (0.5-1 at.%) [Ele03, Li03].








2.2.3.2 Other Ternary Refractory Metal Compounds

Several ternary silicide systems, including TiSixNy [Bai96, ChiOl, EisOOb, Jos02],

TaSixNy [Lee99, Som97], and WSixNy [Bla97, Hir1O], have been examined for diffusion

barrier applications. While the addition of Si promotes amorphous film growth and

improves Cu adhesion [HarO1], it also increases film resistivity and can decrease failure

temperature [Kal00]. The ratio of metal/Si must be greater than 1.67 to ensure stability

with Cu [Nic95]. The failure mechanism in these films tends to be grain boundary

diffusion after barrier crystallization [Nic95], and the stability of these films is

questionable due to the potential for Si out-diffusion into and reaction with the

neighboring Cu layer. These barriers are also reported to have poor adhesion to low-k

dielectrics [Pet03], and the majority of research on ternary silicide barriers has relied on

PVD methods, which will have limited use in the future.

2.2.3.2.1 TiSiN

TiSiN films were deposited by reactively sputtering a Ta-Si target in an Ar/N2

gas mixture [Iij95, Rei94]. Reid et al. [1994] deposited amorphous films with a

stoichiometry of Ti0.34Si0.23N0.43 and resistivity of 680 pQ-cm. These films resisted Cu

diffusion after a 30 min anneal at 6500C. lijima et al. [1995] deposited amorphous films

with a stoichiometry of Tio.31Si0.19N0.50 and resistivity of 500 jQ-cm. These films

resisted Cu diffusion after a 30 min anneal at 6000C.

LPCVD of TiCl4+SiH4+NH3+H2/Ar at 5000C was also reported [Bla97].

As-deposited films were microcrystalline, with a composition of Tio.47Sio.o7No.46.

Chlorine and oxygen contamination levels ranged from 3 to 5 at.%. This film prevented

Cu diffusion after a 6000C anneal for 1 minute.


_~__ *








MOCVD of TiSiN films has been reported with multiple precursor schemes,

including TDMAT + SiH4 + NH3 [ChiOl, Jos02] and TDEAT +SiH4 + NH3 [Bai96]. A

50 A thick TiSiN film had good wettability to Cu and low resistivity (350 LQ-cm)

[Chi01]. The Si was suggested to improve Cu adhesion when thin oxidized layers are

present on the barrier, and also to reduce Cu agglomeration [Pet03]. TiSiN films

deposited by a combination of MOCVD and plasma assisted CVD, however, allowed Cu

to diffuse through after annealing at 5000C [Jos02]. Amorphous TiSiN films deposited at

4000C contained ~3 at.% C and 0, and had high resistivity (>2000 jIQ-cm) [Bai96].

Metalorganic ALD of TiSiN films was reported using TDMAT, NH3 and SiH4 at

1800C [MinOO]. The film contained small crystallites in an amorphous phase, and film

stoichiometry was Ti0.32Si0.18N0.50, but barrier film resistivity and Cu testing were not

reported.

2.2.3.2.2 TaSiN

Several reports of TaSiN thin films as Cu diffusion barriers have been given.

TaSiN films were deposited by reactive sputtering of a Ta-Si target in N2 [HarOl] and

Ar/N2 [Kim97a, Kol91, Lee99] gases. This ternary layer had better barrier performance

and better adhesion to Cu than a Ta/TaN dual layer barrier, but details of the barrier's

performance and electrical properties were not given. Lee et al. [1999] deposited films

with Tao.43Sio.o4No.53 stoichiometry, with a failure temperature of 8250C (decomposing to

Cu3Si +TaNx + TiSi2 above this temperature) but the film had high resistivity (1419

pQ-cm). Kolawa et al. [1991] deposited 100 nm thick amorphous films with a

stoichiometry of Ta0.36Si0.14N0.50. The films had high thermal stability, preventing

interdiffusion between neighboring Cu and TiSi2 layers up to 9000C, had O content








below 3 at.%, and had resistivity of 625 ji2-cm. Kim et al. [1997a] varied the N content

of the TaSiN films to determine the impact on barrier performance. Films with N content

greater than 40 at. % resisted Cu diffusion after an 8000C anneal, while those with lower

N levels failed after a 7000C anneal. Information on film resistivity was not provided,

however.

LPCVD of TaCl5+SiH4+NH3+H2/Ar at 500C was also used to deposit TaSiN

[Bla97]. As-deposited films were microcrystalline, with a stoichiometry of

Tao.35Si.11N0.54, and contained Cl and O levels ranging from 3 to 5 at.%. The films failed

to prevent Cu diffusion after a 1 minute, 6000C anneal, however.

2.2.3.2.3 WSixNy

Several reports of WSixNy thin films as Cu diffusion barriers have been given.

Reactive sputtering of a W-Si target in N2/Ar gas was used to deposit amorphous

WSio.6N films [Min96, Shi97], which crystallized after annealing at 8500C [Shi97]. Film

resistivity was 430-450 2Q-cm, and the WSio.6N film blocked Cu diffusion after a 30

min anneal at temperatures up to 6000C.

WSixNy films were also deposited by N2 plasma nitridation of sputtered WSix thin

films [Hir98, Hir99, HirOl]. As-deposited films were amorphous, and their effectiveness

decreased with increasing film crystallinity. A WSiN(6 nm)/WSix(14 nm) bilayer

resisted Cu diffusion after annealing for 1 hour at 400C [Hir99]. Electrical properties of

the barrier film were not reported, however. WSixNy thin films were inadvertently

formed by annealing a W/WNx/poly-Si structure at 800C [Nak97]. The resulting

W/WSixNy/poly-Si structure was stable up to 950C, above which silicidation of the W








layer occurred. The ternary film was not tested as a Cu diffusion barrier, however, and

bulk resistivity of this film was not reported.

WSixNy films were also deposited by LPCVD of WC16+SiH4+NH3+H2/Ar at

500C [Bla97]. As deposited films had a stoichiometry of Wo.54Si0.12No.34, contamination

levels of Cl and O ranging from 3 to 5 at.%, and high resistivity (1000 SgQ-cm). These

layers prevented Cu diffusion after a 6000C anneal for 1 minute, and crystallized at

9000C.

Amorphous WSixNy thin films were also deposited by PECVD of WF6+ N2 + H2

+ SiH4 at 380 to 4000C [Eck03]. While film resistivity was low (308 pQ-cm), layer

uniformity and thermal stability was poor, as films crystallized after a 1 hour anneal at

6000C. Amorphous WSixNy thin films were also deposited by PECVD of

WF6+Si2H6+NH3 at 3500C [Gok99]. Films with compositions of WL.21Si0.14N and

Wo.96Sio.24N were deposited, which crystallized after annealing at 800C for 30 min.

Film resistivities below 200 p2-cm were reported after annealing at 4500C, but film

performance with Cu was not reported.

2.2.3.2.4 WBxNy

Several reports of WBxNy thin films as Cu diffusion barriers have been given.

WBxNy films with a variety of stoichiometries were deposited by reactive sputtering of a

W-B target in N2/Ar gas [Rei95]. A W0.64B0.20No.16 film showed the best resistance to

Cu diffusion, preventing it after a 30 min anneal at 8000C. This stoichiometry also had a

low resistivity of 220 gi2-cm. WBxNy films were also deposited by reactive sputtering

of W and W2B5 targets in N2/Ar gas, with substrate temperature ranging from 25 to

5000C [LeeOlb, ParOO]. Film stoichiometries ranged from Wo.9oBo.o5No.o5 to









W0.57Bo.o5No.38, with resistivities as low as 140 pC2-cm obtained for the stoichiometry

with the lowest N content. As-deposited films were amorphous, crystallizing after a

7000C anneal [ParOO], and resisted Cu diffusion after a 30 min anneal at 8000C [LeeOlb].

WBxNy films deposited by reactive sputtering of a W2B5 target in N2/Ar gas at 350C were

used to form Schottky contacts to GaAs [Kim98b].

WBxNy films were formed by ion implantation of BF2+ into WNx thin films,

which were deposited by PECVD of WF6 + NH3 + H2 on Si (100) at 3000C [Kim97b,

Kim99a]. The ternary film was amorphous, and resisted N out-diffusion after annealing

at 8000C. As-deposited ternary bulk film resistivity was 200 p.Q-cm, which was slightly

higher than the value for the binary WNx film, and the ternary film resisted Cu diffusion

after annealing at temperatures up to 750C.

Deposition of WBxNy thin films by PECVD of WF6 + NH3 + BloH14 + H2 on Si

(100) at 3500C has also been reported [Kim97b, Kim98a, Kim02a]. As-deposited film

stoichiometries ranged from Wo.90B0.05No.05 to Wo.38B0.42N0.20, depending on the B1oH4/

NH3 ratio, with the optimal film stoichiometry being W0.46B0.25No.29. Resistivity ranged

from 100 to 844 jLQ-cm, with a value -700 pLQ-cm for a 2000 A thick film with optimal

stoichiometry. Films with optimal stoichiometry remained amorphous even after

annealing at 8000C, and resisted Cu-Si interdiffusion up to this temperature. Tungsten

rich as-deposited films were polycrystalline, and B atoms reportedly out-diffused even

faster during annealing that did the N atoms. Both the Wo.9oBo.o5No.o5 and Wo.80Bo.15No.05

stoichiometries failed to resist Cu diffusion after annealing above 6000C.









2.3 Justification for WNx (and WNxCy) as the Barrier Material

The microelectronics industry already uses Ti, Ta and W in one form or another

(such as TiN barriers, Ta/TaN barriers and Ta205 capacitors, and W plugs) for memory

and processor devices [Jai99]. TiN is used as a barrier for Al based metallization, but

TiN is unsuitable for Cu metallization, as mentioned earlier. Cu diffusion barrier

research has focused on TaN and WNx, with WNx appearing to have technological

advantages over TaN as a Cu diffusion barrier. WNx is known to be an effective

diffusion barrier against copper penetration at temperatures up to 750 *C [Pok91], and

adhesion of CVD Cu to WNx is stronger than that to TaN [Iva99]. In addition, WNx

outperforms TaN as a liner material for seedless electrochemical deposition (ECD), as

ECD Cu shows stronger adhesion to WNx film layers [ITR02, ShaOl, Sin02a]. A suitable

liner material would enable seedless ECD of Cu. Elimination of the Cu seed layer,

typically deposited by sputtering, would increase throughput by removing a process step,

as well as remove the complications associated with sputtering in ever-shrinking device

features. The combination of these steps would result in cost savings per wafer, higher

wafer throughput, and higher quality devices for a WNx based system as compared to its

TaN counterpart [Gal99].

In addition, to enable electroless plating of Cu onto the barrier materials, TaN

barrier layers require a pre-deposition HF etch, while WNx barriers do not. Tungsten

oxide, present on the WNx surface due to air exposure, is readily dissolved in the

electroless plating solution, hence the pre-deposition etch step is unnecessary [Wan03].

There are also processing advantages for WNx. During chemical mechanical

polishing (CMP), TaN is removed at a rate 18 times slower than Cu, while WNx is









removed 1.5 times faster [Gal99]. One report gives WNx removal rates of -100 A/min

for CMP pressures below 0.5 psi, while comparable removal rates for TaN require

polishing pressures close to 2 psi [Tak02]. The slow removal rate of TaN means that the

pad must be in contact with the Cu/TaN surface for an extended period of time to

planarize the wafer. This extended contact time leads to dishing of the copper, where

excessive copper is removed from the device. Dishing of both the Cu and dielectric

layers impacts line resistance and planarity of the device [Jai99]. A time-consuming

two-step CMP process, which includes a slurry and pad changeout, is required to

minimize copper dishing during planarization of TaN.

From the above discussion, it is evident that WNx films are the most promising

candidates for diffusion barrier materials. The resistivities for these films are reasonably

low, and WNx has shown good adhesion to Cu and other potential neighboring materials.

CVD and ALD routes to low-resistivity WNx have been established, although the

precursors must be optimized to minimize halide contamination and to control carbon

levels. Recent studies have examined WNxCy for diffusion barrier applications, due to its

ability to deposit in amorphous form, its lower resistivity relative to WNx, and its

excellent adhesion to Cu. A review of the properties of WNx, along with a brief

discussion of WNxCy films, follows.

2.4 WNx Film Properties

The four commonly seen phases for tungsten nitride are BCC WNx, FCC 0-WNx,

HCP WNx, and SHP-WN [Gui93, Nak87, Wri89]. The structure and properties of these

phases are discussed in more detail in Chapter 3. FCC I-WNx is the desired phase for

Cu diffusion barriers because it has the lowest bulk resistivity (-50 pIQ-cm) of the








various tungsten nitride phases [Lee93]. The structure of FCC WNx is NaCI type, with

W atoms at FCC sites and N atoms at octahedral interstitial sites. The FCC P-WNx

phase, with x = 0.5 (also called P-WNo.5 or 1-W2N) has lattice constant a=4.126 A,

hence it has a lattice mismatch of -24% with Si (a=5.431 A) and -27 % with GaAs

(a=5.653 A) [JCP88]. The P-WNo.5 stoichiometry is described as a defective NaCl-type

structure, with half of the octahedral interstitial sites filled with N and half of them vacant

[HonOO]. Bonding in metal nitrides is complex, with both metallic (valence electrons

delocalized; non-directional bonding) and covalent (valence electrons shared; highly

directional bonding) characteristics, due to the combination of metal-metal and

metal-nonmetal interactions [Tot71]. Hones et al. [2000] described metal-nonmetal

bonding in NaCI type structures as having 7n-like and a-like bands due to overlap of the

d and 2p orbitals from the metal and nonmetal, respectively. These bands (called pdn

bands) have ionic character, with bonds having more ionic character when this band is

populated and more covalent-metallic character when the band is depopulated [HonOO].

Shen et al. [2000a], using transmission electron diffraction (TED), found W-N and

W-W nearest neighbor distances of 2.08 and 2.92 A, respectively, in sputter deposited

P-W2N films. The P-W2N lattice parameter (4.126 A) represents the distance between

centers of W atoms at the covers of an FCC unit cell, while the 2.92 A W-W correlation

represents the distance between centers of the corer and face centered W atoms in the

FCC unit cell (Figure 2-1). The theoretical density of FCC P-WNo.5, based on 4 tungsten

and 2 nitrogen atoms per unit cell, is 18.0 g/cm3, while experimentally determined

densities range from 8.0-17.9 g/cm3 [Bos91, Gal97, Hec02, Mar93, Sam80, Sot03].














4.126 A N

2.92 A




Figure 2-1. Distances between W atoms in the FCC face of the 3-W2N phase.

Shen and Mai [2000b] have reported that binding energy shifts (per XPS) for W

and N in polycrystalline P-WNo.5 films indicate an ionic bonding character in the films

(consistent with W4-NI-). The apparent binding energy difference between the metal

and N atom, from XPS, gives an idea of the degree of charge transfer. A large binding

energy difference infers greater bond ionicity [Pri95]. Moreover, adding more nitrogen

to the films has been suggested to cause a decrease in the number of free electrons

provided by W in the solid film [Sot03], which is consistent with increasing film

resistivity with nitrogen content. At lower N levels, the metal's band structure is

retained, so that metal nitrides show metallic properties [Muk93]. Matsuhashi and

Nishikawa [1994] reported FCC P-WNx films to have a work function before annealing

of 5.0 eV, where work function is a measure of the energy required to remove an electron

from the material to a state of rest outside the material (i.e., at the vacuum level).

Transmission electron diffraction (TED) intensities were also used to determine the

concentration of W-N neighboring pairs in the P-WNx films. Results showed that the

concentration of W-N neighboring pairs did not increase when bulk N content was

increased above x-0.5 [ShenOOb]. This indicates that excess N in the films migrated to








the polycrystalline grain boundaries rather than filling the remaining vacant interstitial

sites in the FCC lattice. Shen et al. [2000c] reported no SHP 6-WN formation, even for

WNx with x=1.22, which shows that excess N migrated to the grain boundaries in the

polycrystalline FCC films.

Other reported properties for the binary nitride include a compressive stress of 4.3

+ 0.5 GPa, nanohardness of 30 GPa, and a Poisson's ratio assumed to be 0.25 for films

with the WNo.6 stoichiometry [HonOO]. In addition, depending on the W/N ratio, thick

WNx films were reported with refractive indices and absorption coefficients ranging from

3.20 to 4.00 and 1.20 to 3.95, respectively, as determined by ellipsometry [Boh90].

Oxygen impurities tend to form solid solutions with metal carbides and nitrides

[Tot71], and oxygen is reported to oxidize the surfaces of polycrystals [Par96].

Impurities such as O, N and C are reported to enhance the stability of diffusion barrier

films [Cha97c, Cha99, Lee94a]. Various stoichiometries of tungsten oxynitride have

been reported. The JCPDS powder diffraction standard indicates two phases of W(N,O)x,

one with x=0.62 and a=4.138 A, and the other with x= 0.57 and a=4.126 A [JCP88]. A

WN1.3400.42 phase with a=4.153 A has also been reported [Sel95].

2.5 WNxCy Film Properties

Little information is available in the literature on the structure and properties of

WNxCy films. In going from nitrides to carbides, electron density increases around the

metal atomic sphere and decreases around the non-metal atomic sphere, consistent with

more ionic bond character for the nitrides and more covalent bond character for the

carbides [Gub94]. This is also consistent with the higher resistivity of 3-WNx relative to

P-WCx [Kim03a, Nak87]. Adding carbon to the nitride phase therefore leads to a








decrease in film resistivity. The FCC WNxCy phase, with N and C intermixing on the

interstitial sublattice of FCC W, has been predicted by thermodynamic analysis [Fri99b,

Hua97, Jon93], and has been studied for diffusion barrier applications. A detailed

experimental analysis of the composition, structure and electrical properties of these films

has not been reported, however.

WNxCy film deposition was reported using WF6+NH3+B(C2H5)3 at 3500C [Ele03,

Li03]. The film composition was Wo.ssNo.15sC.30, and films had a nanocrystalline P-WCx

or P -WNx cubic structure in an amorphous matrix along with low resistivity (210 to 400

pQ-cm). The films had F, O and B impurities below 0.5 at.%, H content < 4 at.%,

density of 14 g/cm3 and resistivity of 300 to 400 JiQ-cm [Li03]. Another report using

the same chemistry gave a film composition of Wo.57No.13C0.30 and resistivity ranging

from 600 to 900 pQ-cm, but deposition temperature was not given [Smi02].

Kim et al. [2003d] also reported deposition of WNxCy films using WF6+NH3+

B(C2H5)3 at 3500C. Film composition from RBS was 48 at.% W, 32 at.% C and 20 at.%

N, and film density was 15.37 g/cm3. The film had low resistivity (350 p-Q-cm), and had

an electron diffraction pattern that closely matched those for P-WNo.5 and P -WCo.6.

One HR-TEM lattice fringe spacing was 2.39 A, which was between the interplanar

spacings for (111) P-WNo.5 (2.38 A) and P -WCo.6 (2.43 A). The other lattice fringe

spacing was 2.08 A, which was between the interplanar spacings for (200) P-WNo.5 (2.06

A) and P -WC0.6 (2.11 A). These values support the presence of a ternary WNxCy solid

solution. Moreover, a 12 nm WNxCy film prevented Cu diffusion during annealing up to

7000C.








Vlakhov et al. [1995] deposited W, WC and WNxCy films to study their electrical

resistivities, and stated that bulk tungsten nitrides, carbides, and carbonitrides were good

superconductors [Vla95]. The W films were deposited from both W(CO)6 and WCl6, the

WC films were formed by annealing the W films in a carbon-containing atmosphere, and

the WNxCy films were deposited by co-reacting W(CO)6 +NH3 + CH3COCH3 [Vla95].

The room temperature (300 K) resistivities of W, WC and WNxCy films with thickness of

0.4 gLm were compared. The W and WC films had similar resistivities, ranging from 109

to 191 tQZ-cm, while the WNxCy films had a much higher resistivity of 651 pQ-cm.

2.6 Amorphous WNx Film Deposition

Incorporation of smaller solute atoms into a metal matrix will cause a crystalline

solid solution to become unstable above a certain solute concentration, fostering

amorphous film growth [E1190]. To enable amorphous film deposition in binary systems,

the atomic radii of the two elements must differ by more than 10 % [E1190]. The atomic

radius of N is 0.74 A, while that for W is 1.394 A, hence N is 47% smaller than W, and

so WNx should be able to deposit in amorphous form. Moreover, nitrogen can act as a

roadblock to trap diffusing W on the film surface, or can serve as a nucleation site for

lattice defects. The trapped W and permanent defects caused by N incorporation prevent

crystallization, and amorphous growth occurs [SheOOb].

Several reports of amorphous WNx deposition by a variety of deposition

techniques have been given. These include reactive sputtering, LPCVD, PECVD,

MOCVD, annealing W in NH3, PLD and ion implantation. FCC f3-WNx films with x <

0.5 were typically amorphous, while those with x = 0.5 contained polycrystalline








P-WNo.5, and those with x > 0.5 contained P-WNo.5 and additional N at the grain

boundaries [She00e,f].

Depending on their composition, amorphous WNx films have been reported to

crystallize at temperatures ranging from 480-6200C [SuhOl]. Additional contamination

from C will likely increase the crystallization temperature of WNxCy films.

2.7 Demonstrated Uses of WNx

WNx as a Diffusion Barrier

Kilbane and Habig [1975] first reported growth of WNx thin films by reactively

sputtering a W target in a mixture of Ar/N2 gas. Reichelt and Bergmann [1975] also

studied deposition of WNx thin films by RF sputtering. Ten years later, two reports were

published describing the use of sputter deposited WNx films as diffusion barriers [Aff85,

Kat85]. Since then, many studies of WNx as a diffusion barrier separating Cu and Si

have been reported. References for some of these reports are listed in several review

articles [Jai99, Kal00, Kim03b].

The use of WNx as a diffusion barrier on low-k materials has also been examined.

The interaction between WNx and the low-k material hydrogen silsesquioxane (HSQ)

has been studied [ZenOOa]. Retention of H in the HSQ film is very important, as the

dielectric constant (k) of the film is a strong function of H content (as Si-H). A W2N

barrier film deposited by PECVD of WF6+N2+H2 prevented the release of H (as H2) from

HSQ more effectively than its TaN counterpart deposited by PVD [ZenOOa]. Adhesion

strength between WNx and two other low-k dielectrics, an aromatic hydrocarbon

(SiLKT, from Dow Chemical) and a polyarylene ether-based polymer (Flare 2.0T,

from Allied Signal), has also been studied [LanOO]. TiN, TaN and WNx all had similar








adhesion strengths to these low-k materials, suggesting that a similar bonding

mechanism exists between the barriers and these low-k films [LanOO]. These barriers are

expected to have stronger adhesion than Cu to the polymer layers, because Cu, unlike the

refractory metals, has fewer unfilled d-orbitals available for bonding to the surface

[LanOO].

WNx films have also been tested to separate Cu from the low-k material fluorine

doped silicon oxide (SiOF) [Lee98c]. WNx had poor adhesion to SiOF layers which had

not been pre-treated with 02 plasma at 3000C, and failed to prevent Cu-SiOF

intermixing. When the SiOF layer was pre-treated, however, WNx had good adhesion to

and prevented Cu diffusion into SiOF for a 30 second anneal at 9000C. Surface oxidation

and densification of the SiOF, caused by the plasma pre-treatment, are believed to

enhance barrier capability on pre-treated SiOF [Lee98c].

WNx films have also been tested as diffusion barriers between the Ni and W

layers in an InxGal-xAs/Ni/W contact structure for use with GaAs [Uch97]. These

reactively sputtered WNx films prevented out-diffusion of In from the InxGalxAs layer,

and also lowered the contact resistance from (4 to 1 -cm).

WNx formed by nitridation of an amorphous CVD-W layer using N2 plasma was

used as a barrier in the multilayer structure Cu/amorphous-WNx/amorphous-W/p+n-Si

diode [Cha99]. The structure maintained its integrity after annealing at 725C. The same

group determined the upper limit for a Cu/amorphous-WNx/amorphous-W/Si multilayer

structure to be 7500C [Cha97a, Cha97c].








WNx as Gate Electrode in MOSFET Devices

Historically, metal oxide semiconductor field-effect transistor (MOSFET)

devices have used SiO2 as the gate dielectric. Shrinking device sizes are forcing a

decrease in SiO2 layer thickness. Devices formed with SiO2 layers less than 3 nm thick

suffer from high leakage current due to direct tunneling of electrons through the oxide,

hence an alternative material with higher permittivity is required to prevent tunneling

[LeeO1c]. Ta20O is a potential alternative due to its high dielectric constant (20-25)

relative to SiO2 (3.9) [May90, Par98a]. New contact materials, including WNx, are being

investigated for devices with Ta205 gates. WNx/Ta205 structures had lower leakage

current than TiN/Ta205 structures after annealing at 9000C [LeeOlc]. In addition, the

WNx/Ta205 structure had no interfacial reaction, while the TiN/Ta20s interface degraded

after annealing. The superior performance of WNx is reportedly due to diffusion of N

from WNx into the Ta2O5 layer, which suppresses Ta and O out-diffusion [LeeOlc].

Diffusion of N into Ta205 did not occur for the TiN/Ta20O structure.

WNx has also been used as a diffusion barrier and W source in poly-Si gate stack

structures [Gal00, Yan02]. Without a barrier, W reacts readily with Si, forming a highly

resistive WSix phase [Kas94]. Kasai et al. [1994] formed a W/WNx/Poly-Si gate

structure with a 50 A thick layer of WNx. The sheet resistance of W in this gate structure

was low (1.6 Q/ ) compared to one without the barrier layer (18 Q/ ), which suffered

from silicide formation. First, PECVD WNx was deposited on poly-Si and annealed to

form a W-Si-N layer. Then, the structure was annealed at temperatures up to 10000C in

either an Ar or N2 atmosphere to convert some of the remaining WNx layer to W. The

newly formed layer of W can then be used as the contact metal for the gate. While








as-deposited WNx films had sheet resistance of 19.2 Q/ the W layer produced by

annealing had sheet resistances as low as 1.28 Q/ [Gal00]. Sputtered WNx was tested in

a similar manner; formation of WSi2 was suppressed due to formation of a Si3N4 layer at

the WNx-Si interface after a rapid thermal anneal (RTA) at 10000C [Yan02]. Kang et al.

[2001,2002] used a sputtered, 100 A thick WNx film as a barrier layer in a W/WNx/poly

Sil-xGex gate structure for use in a CMOS-FET. Takagi et al. [1996] used a W/WNx

polysilicon gate electrode to produce a CMOS device. The WNx layer suppressed silicide

formation after a 30 second anneal at 950C, and the specific contact resistance for a

W/WNx/poly-Si structure was 10-7 Q-cm2 [Tak96]. In addition, the sheet resistance of

WNx deposited on poly-Si was reported to be an order of magnitude lower than that for

WSix [Cho02].

WNx as a Gate Electrode in MESFET Devices

Rectifying (Schottky) contacts for GaAs self-aligned gate field-effect transistors

(SAGFETs) must have good thermal stability, good adhesion to GaAs, and high Schottky

barrier height [Lee95, Pac91]. WNx Schottky contacts formed by PECVD of

WF6+NH3+H2 on GaAs maintained their interface integrity after a 30 second, 1000C

RTA [Lee95]. Yu et al. [1988] sputter deposited thin WNx films for use as a Schottky

contact to GaAs, while Boher et al. [1990] deposited thick WNx films. Kim [1994] used

PECVD of WF6 + NH3 + H2 at 3500C to deposit WNx Schottky contacts to GaAs. WNx

nucleated easily on GaAs and had higher thermal stability than W films, maintaining the

W2N stoichiometry and blocking As out-diffusion during a RTA at 10000C for 30

seconds [Kim94]. Grain boundaries in the PECVD WNx films were stuffed by excess N,

preventing Ga and As out-diffusion, and this Schottky contact structure was stable up to








850C. Paccagnella et al. [1991] examined the effect of GaAs pre-treatment on the

performance of WNx/GaAs Schottky diodes. Schottky barrier height was highest for

WNx/GaAs diodes annealed at 800C, where GaAs was pretreated by H2 plasma [Pac91].

Sputter deposited WNx has also been used as a gate metal in GaAs SAGFETs [Nag94,

Gei86, Uch86]. In particular, WNo.o4 was found to produce Schottky barriers with

excellent stability after annealing for 20 min at 8100C [Gei86]. WNx films have also

been deposited onto 4H-SiC to form Schottky contacts with high temperature stability

[Pec97, Kak99]. The contacts, deposited by room temperature magnetron sputtering of a

W target in Ar/N2 gas, remained rectifying up to 1200"C despite formation of the W5Si3

and W2C phases [Pec97].

WNx as an Ohmic Contact in HBT Devices

Thermally stable, low resistance ohmic contacts are critical to produce high

speed, high frequency devices such as heterojunction bipolar transistors (HBTs). Park et

al. [1998b] deposited the Au/Pt/Ti/WNx metal contact structure onto n-InGaAs, which is

a cap layer used for AlGaAs/GaAs HBTs. The minimum contact resistivity for the

Au/Pt/Ti/WNx/n-InGaAs structure was 9.5 x 10-8 L-cm2, which was obtained after

annealing at 4000C [Par98b]. In addition, the morphology of the contact remained

smooth over a wide annealing range.

WNx as a Top Electrode in DRAM Devices

WNx films have also been investigated as top electrodes for Ta205 dynamic

random access memory (DRAM) capacitors [ChoOl, Mat94]. Cho et al. [2001] deposited

both TiN and WNx by reactive sputtering in Ar/N2 ambient to test their stability as a

diffusion barrier at the interface in a W/Ta205 gate structure. The structures with WNx








barriers had higher thermal stability, as they were more resistant to Ta and O

out-diffusion than TiN. The increased stability was presumably due to migration of N

from the WNx film to the WNx/Ta205 interface, which prevented Ta and O from diffusing

out of the Ta20s. Suzuki et al. [1998] used WF6 with two different N sources, NH3 and

NF3, to deposit WNx thin films by PECVD. When tested as top electrodes in DRAM

capacitors, these films had leakage currents that were an order of magnitude lower than

PVD TiN top electrodes. Matsuhashi and Nishikawa [1994] tested sputtered WNx, TiN

and TaN as top electrodes in Ta205 DRAM capacitors. After annealing for 30 min at

8000C, the WNx/Ta205 structure exhibited lower leakage current than structures with TiN

or TaN electrodes [Mat94]. Kim et al. [2001] used a W/WNx/Poly-Si gate stack to

produce polysilicon based DRAM devices.

WNx as an X-ray Absorber Mask

The use of amorphous WNx films in absorber masks for X-ray lithography has

also been demonstrated [Lee97, Lee98b]. WNx has good stress controllability, strongly

attenuates x-rays, and has a coefficient of thermal expansion (CTE) very close to

common x-ray membrane materials such as SiC, BN, SiNx, and CN, making it a

desirable x-ray mask material [Lee98b]. WNx masks were grown by reactive sputtering

of a W target in an Ar/N2 atmosphere on room temperature indium tin oxide (ITO) coated

Si substrates. Film microstructure depended on the nitrogen content of the films, with 20

at. % N being the upper limit for the amorphous phase. As N levels increased above this,

film microstructure shifted to polycrystalline P-WNo.s.








WN, as a Liner for Cu Deposition

Sputtered WNx films were also shown to be good liner materials for seedless

electrochemical deposition, due to strong Cu-WNx adhesion and high Cu nucleation

density on the WNx surface [Sha01]. ECD Cu was deposited on air-exposed WN, films,

and the composition of the ECD bath was controlled to strip off surface oxides [ShaOl].

Use of TiN as a seed layer for electroless Cu deposition resulted in only sporadic Cu

coverage on the barrier layer [Mur95], while electroless Cu deposition occurred readily

on WNx films [Wan03]. For the same bath temperature (700C), electroless Cu deposition

occurred more rapidly on WNx than on TaN [Wan03]. Moreover, Cu deposited by CVD

had low via resistance on WNx, but high via resistance with Ta based barriers [Jai99].

Other WNx Uses

P-WNo.5 has found use in bulk form (i.e., powder) as a catalyst for quinoline

hydrodenitrogenation [Abe93], n-heptane isomerization [Sel95], deamination of

2-octylamine and alcohol dehydration [Lee92, Luc96]. Reactively sputtered amorphous

WNx films have also been deposited and used as a means to form equiaxed, low

resistivity W films in W/poly-Si gate structures [Lee03]. Amorphous WNx films

annealed at 1273 K released nitrogen to form BCC W with a resistivity of 12 pQ-cm,

which was similar to the value for pure, sputtered W. The presence of a small amount of

residual N in the films at 1273 K suppressed silicide (WSix) formation, enabling the W

film to retain its low resistivity.

WN, Etching

Several etching studies have been done on WNx films. Lee [98b] first reported

use of SF6 + Ar + N2 in an inductively coupled plasma (ICP) etching system to








anisotropically etch WNx thin films. Etch rates ranged from 4000 to 14000 A/min.

Vijayendran et al. [1999] used an NF3/N2 reactive ion etch (RIE) to remove WNx films.

Reyes-Betanzo et al. [2002] used SF6 + Ar to anisotropically etch W and WNx films.

2.8 WNx Deposition Techniques

A variety of techniques to deposit WNx have been tested in order to optimize key

film properties, which include film resisitivity and deposition temperature. Ideal film

resistivity was reported to be 500 p.-cm, [Ele03], while a temperature ceiling of 400C

has been given for IC production by numerous reports [EisOOb, Ele02, HauOO, Kim03b,

Les02, Sun01b]. Depositing the barrier at the absolute minimum temperature, however,

may not be the best solution, as other processes occurring during IC processing will

typically take the device temperature up to ~4000C. If the barrier is deposited at much

lower temperature than the processing temperature ceiling, significant shift in barrier

structure can occur during subsequent high temperature processing, leading to Cu

diffusion and potential barrier failure.

The reported deposition techniques include reactive sputtering, annealing of W in

NH3, plasma nitridation of W, LPCVD, PECVD, MOCVD, and ALD. Reports of each of

these techniques to deposit WNx are discussed below.

2.8.1 Annealing W in NH3

Deneuville et al. [1989] reported formation of WNx by annealing sputtered W

films on Si in an NH3 atmosphere from 500 to 11000C. The N/W ratio increased with

temperature, going from 0.37 to 1.85 across the aforementioned temperature range.

While several polycrystalline structures were postulated based on film composition, XRD

results were not given to support them. High temperature and pressure, along with long








anneal times, are required to form WNx in this manner, making this an undesirable

deposition technique for IC manufacturing.

2.8.2 Plasma Nitridation / Ion Implantation of W

W films, deposited by CVD of WF6 + SiH4 + H2 were nitrided to form thin WN,

films using an in-situ N2 plasma treatment [Yeh96]. This nitride layer prevented WAll2

formation after annealing at 550C, where WA112 formation can degrade contact

structures using W plugs between Al and Si. WNx films were also formed by N+ ion

implantation into W metal substrates [Zha99]. Barrier testing with Cu was not done for

either of these films, however. N+ ion implantation was also done on W films deposited

by PECVD of WF6 + H2 to form amorphous WNx films [Kwo95]. These films resisted

Cu penetration after a 30 min anneal at 8000C, while polycrystalline PECVD WNx films

failed. The nitridation technique is directional in nature, however, meaning that good

conformality will be difficult in highly aggressive future device topographies. Chang

[1997a-c, 1999] reported N+ ion implantation into W films deposited by CVD of WF6 +

SiH4. The films had resistivity of 198 pQ-cm and resisted Cu diffusion for annealing

temperatures below 7000C, but had poor adhesion to the underlying Si substrate.

2.8.3 Pulsed Laser Deposition

Soto et al. [2003] first reported pulsed laser deposition of tungsten nitride films

from a W target in the presence of N2. Films were deposited on n-type Si (111) and on

Coming glass slide substrates, at N2 pressures ranging from lxl0-8 to lxl0-1 Torr. Films

contained W, N and some O, believed to originate from W target contamination, and had

higher density than films produced by DC magnetron sputtering. This technique, like








plasma nitridation, is also directional, making its ability to deposit highly conformal films

in future device generations questionable.

2.8.4 Reactively Sputtered WN, Deposition

Typically, sputtered WNx films are formed by sputtering a W target in a N2/Ar

atmosphere. A variant of this, known as nitrogen ion beam sputter deposition (IBSD),

has also been reported [Eiz94b, Gal93], where a nitrogen ion beam, rather than an Ar+

ion beam, is used to liberate W atoms from the W target. Nitrogen ions or radicals

backscattered from the target are then incorporated with W at the substrate to deposit

WNx. Both variants enable amorphous film deposition due to the decreased mobility of

W in the presence of N on the substrate surface. N can inhibit W mobility on the

substrate surface by blocking W diffusion or by trapping W. Decreased W diffusivity

leads to defect formation, failure of crystal growth and amorphization of the growing film

[SheOOb]. Sputtering can be done at low deposition temperature (at or near ambient),

which protects temperature sensitive components from thermal damage during the barrier

deposition process.

Many reports of sputtered WNx deposition have been given in the literature. The

key properties of these films are summarized in Table 2-1. While sputter deposited WNx

films are generally contaminant free, the major drawback of sputtering, as mentioned in

Chapter 1, is poor conformality in small, high aspect ratio device features.

2.8.5 LPCVD WNx Deposition

LPCVD of WNx has been researched as a possible alternative to sputtering for

smaller device features with high aspect ratios. Many reports of LPCVD WNx films have












0% e 00000
0%~0
88 a m

M~ dl~ =0 ? 3s
Q,- 00
kNn 06db~




00
o $~





cs~cq














'00
-B










bo






Wz







0 E U N






+++
cuoo
IS 8 S






























in 0
E- U

1^^ ? ^ ^

g^ ^5 ^

^
1^ ^
|^^
o
0

U5
._ in
.9 *-- Y o





.2 ,'-
1 t^AS (
(S~ A b ^





1-a a i .9 a
Ev l ,.P ,Z


been given in the literature. The key properties of these films are summarized in Table 2-








As evident from Table 2-2, LPCVD reactions are based on halide chemistry, and

include the reduction of WF6, WO3, or WC16 by NH3 or NF3 in a H2 atmosphere [Chi93,

Gon02, Mar93, Nag93, Nak87, Sak96, Suz98, Vol85]. Nagai and Kishida [1993]

reported that use of WCl6 as the tungsten precursor for reaction with NH3 and H2 was

more thermodynamically favorable than using WF6, as the Gibbs free energy change was

larger for the chloride precursor reaction. Marcus et al. [1993] have reported high

LPCVD deposition rates, ranging from 1800-4500 A/min.

One report (not included in the table) of the "indirect" deposition of J3-WNx has

been given, where this phase formed after annealing a metastable W3Ns phase [Zha97].

Reacting WNC13 with ZnN2 produced the W3N5 phase at 4000C, and subsequent

annealing at 6000C resulted in formation of the f3-WNx phase.

The halide precursors used for LPCVD can create difficulties during barrier

deposition and subsequent device processing. WF6 is reported to consume Si during the

reaction, forming SiF4, which leaves vacancies on the Si substrate [Kim91, Kim92,

Lai98a]. In addition, there are concerns that residual F and Cl in the barrier contribute to

corrosion of metal interconnects [Fal98, Kel99, Raa93]. Contamination levels of

0.1-0.9% F have been reported for LPCVD films [Mar93]. Moreover, reactive

by-products (e.g., hydrogen halides) resulting from LPCVD are a material handling

concern [Cur92, Kim91].

Adduct formation is another difficulty associated with LPCVD. Adducts are

compounds produced by chemical addition of two or more reactants. These adducts can

settle onto the substrate and barrier film during deposition, resulting in dislocations

and/or pinholes in the film that can lead to barrier failure. Gas phase adducts such as






90


WF6-4NH3, NH4F2, and NH4F have been reported during LPCVD reactions [Nak87,

Suz98, Tsa96].

I .


o E 2





a




0 0
a o < a


00

o












0 + + +







uder, rather than film, produced in this report
0 0






bulk -W2N powder, rather than film, produced in this report
E ^ ^^ <>
5~ ^' 2 2X 2 '^
c~Y 1/^ 9
cx s 3(
au_________Q___


(j te lulk B-2







2.8.6 PECVD WN, Deposition

The same halide chemistries used for LPCVD are also used for PECVD, whereby

the plasma assists fragmentation of the reactants, lowering reaction temperatures

considerably. After fragmentation by the plasma, the precursor fragments travel to the

heated substrate, react on its surface and deposit a film. Many reports of PECVD WNx

films have been given in the literature. The key properties of these films are summarized

in Table 2-3.

Films deposited by PECVD show excellent adhesion, with resistivities as low as

70 ~2-cm. Growth rates are moderate, ranging from 300-400 /min [Gal97], and

conformality up to 70 % has been reported for trench holes with diameter > 0.30 lm

[Kim93].

PECVD films suffer from reduced conformality in small diameter, high aspect

ratio trench features, however, due to the directional nature of the plasma [JohOOa,

Tsa96]. Despite the low resistivity and deposition temperatures that PECVD offers, its

inability to deposit highly conformal films in high aspect ratio features makes its use in

future barrier deposition questionable. Moreover, PECVD films, like LPCVD films,

suffer from halide impurities. Concerns about halides contaminating the barrier films

also exist for PECVD, since the same reactants are often used for both LPCVD and

PECVD. Contamination levels up to 4 at.% F, which are higher than those for LPCVD,

have been reported for PECVD films [Lee93]. Another difficulty associated with

PECVD is gas phase adduct formation; adducts such as (WF6)x(NH3)x and NH4F have

been reported during WNx PECVD [Lu98b, Suz98]. These adducts can contaminate

films grown by PECVD and deposit particles on the film surface, both of which can




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EZW2F09SE_86YZIN INGEST_TIME 2011-11-02T17:08:08Z PACKAGE AA00004688_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

CHEMICAL VAPOR DEPOSITION OF THIN FILMS FOR DIFFUSION BARRIER APPLICATIONS By OMAR JAMES BCHIR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

I would like to dedicate this dissertation to my nephew, Matthew Edward Brown, and to the memory of my grandparents.

PAGE 3

ACKNOWLEDGMENTS On a professional note, I would like to thank my supervisory committee (especially my advisor, Dr. Tim Anderson, and Dr. McElwee-White) for their constant guidance and assistance throughout my Ph.D. work. I would also like to thank the past and present members of Dr. Anderson's, Dr. McElwee-White's and Dr. Norton's research groups, with whom I have collaborated during my time at UF, and Dr. Jianyun Shen, with whom I collaborated on thermodynamic modeling. I would also like to thank Dr. Holloway for use of his 4-point probe device, and the staff at the Major Analytical Instrumentation Center (MAIC), including Eric Lambers, Wayne Acree and .Br.ad. Willenburg, for valuable assistance and training with various characterization techniques. On a personal note, I would like to thank my parents (Sara and Hachemi Bchir) for their encouragement, love and support throughout my life, and for their strong emphasis on the importance of education. I would also like to thank my sister, Annissa Brown, my brother-in-law Eddie, and my nephew Matthew, for their encouragement and support Thanks also go to my girlfriend Luciana Manfrim for her love and support. Lastly, thanks go to all of the friends that I have made through the years while growing up in Florida, attending Georgia Tech, working at Fluor Daniel and attending UF. lll

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGMENTS .... .... .................... .................................. ........ .. .. .... ... ..... iii ABSTRACT ....... ......... ............ ......... ..... ..... .......... ................... ............. ... ....... . ... viii 1 INTRODUCTION 1.1 Background ............. ........ ..... .................................................. . ........ 1 1.2 Comparison of Diffusivities for Al and Cu in Si ............... .............. 4 1.3 Diffusion Barrier Requirements and Properties ............ ...... ............. 6 1.3.1 Film Structure ........ .... ..................... . ........... ............. ... ... ..... 9 1.3 2 Electrical Properties ........... ....... ......... .. ..... ............ ...... ... .... 11 1.3 3 Conformality ............... ... ....... ..... .... .............. ..... .......... ........ 14 1.3.4 Adhesion ................................ ........... ......... . ........ .... ... . .... . 16 1.4 Dielectric Material Considerations ............... ...... ......... ..................... 17 1.5 Diffusion Barrier Failure Mechanisms .... ........................ ................ 17 1.6 Diffusion Barrier Deposition Techniques .............. .... ... ........... ..... 21 1.6.1 Physical Vapor Deposition (PVD) ............... ... ................. ... 21 1.6.2 Chemical Vapor Deposition ............. .... .. ... .... .. ...... ...... ..... 24 1.6.3 Atomic Layer Deposition (ALD) ................ ......................... 37 1.7 Coppe r Deposition Methods ..... ..... ............. ........ .. ................. . ........ 44 1.8 Statement of Problem ........ ................................ ..... ... .................... 49 1.9 Hypothesis .......... .. .... .. ....... ....... ... . ... .... . .. .. . ... ... ........... ... .............. 51 2 REVIEW OF THE LITERATURE 2.1 Cu-Si Interconnects without a Barrier. ...................... ..... .................. 53 2.2 Refractory Metal-Based Barriers ... ... .... ... ......... ... .... ........... ...... .. 54 2.2.1 Unary Refractory Metal Barriers .................. ..... ...... ...... ...... 54 2 .2 .2 Binary R e fractory Metal-Based Barrier s .... ..... .......... ... .... ... 56 2 2.3 Ternary Refractory Metal-Based Barriers ..... .. ..... .... .. ......... 63 2.3 Justification for WNx (and WNx C y ) as the Barrier Material ...... ..... 71 2.4 WNx Film Properties . ....... .. ... .... .................. ..... .... ... .......... . ... ....... 72 2.5 WNx C y F ilm Properties . ..... ......................... ............. ..... ... .......... ... 75 2 6 Amorphous WNx Film Deposition ........ .... ..... . .... . ... ..... ..... ..... ...... 77 2 7 Demonstrated Uses of WNx ........... . ........ ........................ ................. 78 2 8 WNx Deposition Techniques ........... ... ..... .... .......... .... .................... 85 2.8.1 Ann ea ling W in NH3 85 2.8.2 Plasma Nitridation / Ion Implant a tion of W ... ..... ... ....... ...... 86 2.8.3 Pulsed Laser Deposition ......... ......... .... ... ............... ........ . .. 86 2.8.4 Reactively Sputtered WNx Depos i tion ...... .......... ... ....... .. .. 87 iv

PAGE 5

2.8.5 LPCVD WNx Deposition ..................... ............. ...... ....... . . 87 2 8.6 PECVD WNx Deposition ......... . ....................... .......... .... .... 91 2.8 7 MOCVD WNx Deposition ................ ...................... ........... 93 2.8.8 ALD WN x Deposition ... ............................. .......................... 97 2 8.9 ALD WNxCy Deposition .... .... ................... ........... ....... ....... 99 2.9 Conclusions ...... ..... . ..... ..... ...... ... .............. ......... .... ........ ..... ..... 100 3 EQUILIBRIUM MODELS FOR WNx-BASED BARRIERS 3.1 Motivation and Method .......................... .......................................... 101 3.2 Constraints ............... . . ... .... ......... .... ... . ...... ....... .... . ... . ........... . 111 3.3 Degrees of Freedom ...... ....... ..... .............. . .... . ...... . .... ...... ......... ... 111 3.4 Computational Methods .............. .... .............................. ................. 112 3.5 Phase Equilibrium in the W-N System . .... . ..... ....... .... ...... . ... ..... 114 3.6 Previous Studies of the W-N Phase Diagram ........ ... ... .......... ... ... 125 3.7 W-N Optimization Results and Discussion ....... .... .... .... ..... ..... ..... 126 3.8 Stable Solid Phases in the W-C System ...... .... .... .... .... ............... 139 3.9 Previous Study of the W-C Phase Model.. ......... .... ............ .......... 142 3.10 Equilibrium in the W-C-Cl-H-N System .... .... . ... ..... ......... . ...... . 145 3 .11 Optimization of the FCC P-WNx C y Gibbs Energy . ....... . . . ... ........ 153 3.12 Degrees of Freedom (DOF) Analysis ... .... .... ... .... .... ...... .... ... ..... ... 168 3 12.1 Homogeneous Gas Phase Speciat i on ...... ......... .... ......... .. ... 171 3.12.2 Heterogeneous Gas Phase Speciation ... ............. .............. ... 177 3 12.3 Heterogeneous Solid Phase Equilibrium ...... ....... ........ . ...... 182 3.13 Predicted W-C-N Ternary Phase Diagram .... ........ ............... ..... . 186 3.14 Predicted P-WCo. s P-WNo.s Pseudobinary Equilibrium ... ........... 190 4 EXPERIMENT AL APPROACH FOR CVD 4.1 Substrate Preparation .................. ..... .... .... .................. . ............. .... ... 194 4.2 Solvent Tests ..... .................... .......... ......... ................... ... ........... 195 4.3 Description o f CVD System Components .... .... . .... . ....................... 196 4.4 Start-Up ....... .................................. ............................ ............ ....... 201 4 5 Copper Deposition . ... . ..................... .......... ... . .... ............ . ...... .. .... . 201 4.6 Analysis Techniques ... ..... ... ..... ........ . .... .... ... .... ......... ......... .... 202 4.7 Precursor Screening Procedure ... . .... .... ......... ... ......... ............ ..... . 208 5 EVALUATION OF Cl4(CH3CN)WN-i-Pr AS A SUIT ABLE WNx PRECURSOR 5.1 Synthesis of Isopropyl [Cl4(CH3CN)WN-i-Pr] Precursor ... ...... . ... 210 5.2 Solvent Select i on ....... ...... ..... .......... .... . ..... .......................... ......... ... 211 5.3 Pre cursor Mass Spectral Pre-Screen ....... ......... ....... .... ... ...... ........ . 212 5 4 Film Structure ...... ..... ..... ......... ......... ... . ........ .... . ...... ... ... ............ 2 17 5 4.1 XRD Results ... . . . ............... ......... . ....... . . ............ ...... ...... . 217 5 4.2 TEM ............. ... ... ......... . . .... ............ .............. .......... ........ 223 5 5 Film Composit i on .... ............. ..... .......................... .... .... ......... .......... 2 25 5.5.1 AES ...... ........ ..... . .... ..... ... ........ .... .... ....... ........ ............ . . 2 25 5 5 2 AES Depth Profiling ............. ....... . . ... ..... ............ ... .. .... ..... 229 V

PAGE 6

5 5 3 XPS ... ................ ............. . .... ..... . ..... .......... .. ... .... .... . ...... ... 233 5.5.4 SIMS Depth Profiling ................... ............... ........ ...... ......... 253 5 6 Film Growth Rate (XSEM) .............. .................................... ..... ..... 255 5.7 Film Incubation Time and Morphology . .............. ........................... 259 5.8 Film Electrical Properties .... ....... ... ..................... ... .......... ..... ........ 264 5.9 Effect of NH3 and N 2 Addition to Film Growth . . ...... ...... .... .......... 267 5.9.1 Film Structure ...... ................ .... .... ....... ...... ........ ......... ...... 267 5 9.2 Film Composition ... .... ... ... ... . . ... .... . ............ ................ ..... 270 5.9 3 Growth Rate .................................................................. ....... 277 5.9.4 Film Resistivity ..... ..... ..... ... . ... ............. ..... . .... . ........ . ... . 279 5.9.5 Sheet Resistance ............... .... ........... ....... ..... ........... ........... 280 5 10 Conclusions for Use of NH3 and N2 with i-Pr. .................... ........... 281 5.11 Effect of Solvent Change on Carbon Content.. ........ .... .................. . 282 5.11.1 Film Structure ...... ... .......... . .... .... .... ....................... ........... 284 5.11 2 Film Composition ............... . ...... ........ ...... ... ... ...... . ... . . . . 287 5.11.3 Carbon Deposition Rate . . . ... ....... ............. ............. . ... ... 291 5.11.4 Conclusions On Effect of Solvent Change . . ... ... ... .............. 297 5 12 Conformality Tests ....... ..... .... ............... . .... . ................. .................. 298 5.13 Adhesion Tests ...... .... .... ......... . ........ ....... ...... .. ... ... .... .... ........... . 298 5.14 Conclusions on Use of Cl4(CH3CN)WN-i-Pr to Deposit WN x 299 6 EVALUATION OF Cl4(PhCN)W(NPh) AS A SUITABLE WNx PRECURSOR 6 1 Film Growth Studies ... . .... .... . ... . ... . .... ... .......... .... ............ ...... .... 303 6.2 Synthesis of Phenyl [Cl4(PhCN)W(NPh)] Precu r sor ............ ..... .... 303 6.3 Solvent Selection ............................. ................. .... ................. ..... .... 303 6.4 Precursor Mass Spectral Pre-Screen .......... .................................... 304 6 5 Film Structure ... . .... ...... ... . ..... .... ...... ... .................. .... ............... ... 308 6.5.1 XRD ............. ..... ............. .... .... .... ..... ......... ... . .............. ... . 308 6 5.2 Lattice Parameter ..... ...... .... .... ...... ... ........ ................ ........... 313 6.5.3 Polyc r ystal Grain Size ................ ......... ........... ....... ......... ..... 315 6.6 Film Composition ......... . . . . .... .... ...... .............. .... . . ..... ... ............. 316 6.6 1 AES ... . .... ..... .... ..... .............. .... .... . .... ........... . ................ 316 6.6 2 Film Growth Rate (XSEM) ... .... .................. .... ........ .... ..... 320 6.7 Film Electrical Properties ... ... .... . ... ... ...... .... .............. ......... ..... ..... 322 6.7.1 Film Resistivity ................... ......... ................................ . ..... 322 6 7.2 Film Sheet Resistance .............. .... . ..... .... .... . ....... ............. 323 6.8 Conclusions on Use of Cl4(PhCN)W(NPh) to Deposit WNx 324 7 EVALUATION OF Cl4(CH 3 CN)WN C 3 H5 AS A SUIT ABLE WN x PRECURSOR 7 1 Film Growth Studies ..... ...... . ....... .... .... ... ........ .... ......... ... ... . . ..... 331 7.2 Synthesis of Allyl [Cl4(CH 3CN)WN-C3 H5 ] Precursor. ...... . .... . ..... 332 7.3 Solvent Selection ................. . ............. ............... ................. . ... . ..... 333 7.4 Pre cursor Mass Spectral Pre-Screen .... ... . .... ... .... .... .... . . .... ........ 334 7.5 Film Struc t ure ... ....... ......... . . ........ ... . ..... . ......... ........ .... ..... ... ... .. 336 7.5 1 XRD ... .......... .................. ........ ...... ...... ........ . ... .... ... . ... . 336 VI

PAGE 7

7.5.2 Lattice Parameter ................................................................... 339 7.5.3 Polycrystal Grain Size ............................................... ............ 341 7 .6 Film Composition ........... .............................................................. ... 342 7. 7 Film Growth Rate (XSEM) ............................................................... 344 7.8 Film Electrical Properties .................................................................. 345 7.8.1 Film Resistivity ..................................................................... 345 7 .8.2 Film Sheet Resistance ........................................ .... ............... 346 7.9 Conclusions on Use of Cl4(CH3CN)WN-C3Hs to Deposit WNx 347 8 COPPER TESTS ON BARRIER FILMS FROM Cl4(CH 3CN)WN-i-Pr 8.1 Copper Deposition Results ................................................................ 357 8.2 Electrical Measurements . ................................................................. 357 8.3 Scotch Tape Tests ................................................................... ..... ..... 358 8.4 Barrier Integrity Tests ....................................................................... 358 8.4.1 XRD ................................... .......... ....................................... 359 8.4 .2 Electrical Measurements .................. .. ................................ ... 364 8.4.3 Depth Profiling Analysis ................. ..................................... 365 8.5 Conclusions Regarding Cu Testing of Deposited Barrier Layers ..... 366 9 RECOMMENDATIONS FOR FUTURE WORK 9.1 Control Film Composition ................................................................ 368 9.2 Decrease Deposition Temperature .................................................... 371 9.3 Decrease Barrier Thickness ............................................................... 371 9.4 Further Conformality Testing ............................................................ 372 9.5 Further Adhesion Tests ......... ............................................................ 372 9.6 Further Barrier Integrity Analyses .............. ....................... ... ........... 373 9.7 Deposition on Alternate Substrate Materials .................................... 375 9.8 Testing of Films for X-ray Absorption Mask Applications ... ......... 376 9.9 Deposition of Alternate Barrier Materials ......................................... 376 9.10 System Modifications ............................................................ ........... 376 REFERENCES .............................................................. ................ .. 378 APPENDIX A CVD SYSTEM PROCESS FLOW DIAGRAMS .... .............. ... .................. 409 B OTHER PRECURSORS TESTED ...................................................... ........ 415 BIOGRAPHICAL SKETCH .................. .............................. ........ ... ... ................... 419 vii

PAGE 8

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CHEMICAL VAPOR DEPOSITION OF THIN FILMS FOR DIFFUSION BARRIER APPLICATIONS By Omar James Bchir August2004 Chairman: Timothy J. Anderson Major Department: Chemical Engineering The semiconductor industry is transitioning from aluminum to copper as the interconnect material for integrated circuits. Diffusion barriers are essential for preventing copper migration into silicon in the resulting integrated circuits. Metalorganic chemical vapor deposition (MOCVD) is a useful technique for conformal deposition of thin barrier films. By manipulating the molecular structure of the MOCVD precursor(s) it is possible to control the structure and properties of the deposited film The focus of this work has been to examine novel precursors for use in MOCVD of thin film diffusion barriers. To date, the suitability of a variety of novel precursors for deposition of tungsten nitride (WNx) for diffusion barrier applications has been tested. The WNx precursors were of the form Cl4(CH3CN)WNR where R represents the isopropyl, phenyl and allyl group among others. Mass spectrometry fragmentation patterns for each precursor were studied to pre-screen candidate precursors. Film properties were examined by several different characterization techniques including Vlll

PAGE 9

XRD AES XPS, SEM and 4-Point Probe. Data from these techniques were then correlated to pre-screen fragmentation patterns to determine the impact of the imido (R) group on film deposition. Apparent activation energies for film growth from the al l yl isopropyl and phenyl precursors, for example, were 0.15 0 84 and 1.41 eV, respectively and were directly related to the strength of the N-R bond in the precursors. In addition to experimental testing of new precursors for WN x deposition, thermodynamic phase equilibria for the W-N-H-C-Cl deposition system was assessed Modifications were made to the previously reported W-N binary model to include new experimental data from the literature on the desired FCC WN x phase as well as the SHP WN phase. Once the binary W-N diagram was reassessed, this model was merged with the existing W H C Cl database to create an initial model for solid gas equilibrium in our system at our experimental conditions. Using a combination of XRD AES and XPS data from our films the W N H C Cl model was modified to include carbon-nitrogen-vacancy interactions in the face centered cubic (FCC ) WN x C y solid phase, so that the new model reflected our experimental results. F inally the low temperature ternary W-C-N phase diagram was predicted. l X

PAGE 10

CHAPTER 1 INTRODUCTION 1.1 Background The semiconductor industry continues to shrink the size of transistors on integrated circuits (!Cs) to increase the number of transistors per chip, which translates to increased IC device speed. To shrink these transistors, the sizes of the device features (trenches, vias, etc.) and the metal interconnects ("wires" that connect the individual transistors on the chip) must be decreased. Recent accomplishments have reduced feature sizes to 100 nm, yielding more than 400 million transistors per chip. The International Technology Roadmap for Semiconductors [ITR02] predicts IC feature sizes to shrink continuously over the next several years, in the following steps: 100 nm (in 2003), 90 nm (2004), 80 nm (2005), 70 nm (2006), 65 nm (2007), 45 nm (2010), 32 nm (2013), and 22 nm (2016). Aluminum (Al) and Al-based alloys have historically been used as the metal interconnect materials on integrated circuits. Al is reaching its usability limit as industry continues to shrink IC features. For a given current flow, decreasing the interconnect cross-section leads to an increase in current flow density. At high current flow density, Al suffers from electromigration, where electrons flowing through the metal interconnect impart enough momentum to carry the metal atoms with them. Once current flow density rises above the critical value (106 A/cm2 for Al), electromigration leads to voids (openings) and hillocks (pileups) in the interconnect wiring (Figure 1-1). Formation of 1

PAGE 11

2 voids in the Al interconnect leads to an open circuit, which causes device failure. Al doped with Cu (termed Al(Cu)) was used to increase the electromigration resistance, but this scheme is also reaching its usability limit. A different interconnect metal is required to meet present and future current flow density demands. electrons -----i+ electrons --Figure 1-1. Before (a) and after (b) electromigration induced hillock and void formation in an Al interconnect. Industry has already transitioned to copper as the interconnect material for the intermediate and upper wiring levels on IC devices, due to copper's higher electromigration resistance and 40% lower bulk resistivity (1.67 .Q-cm for Cu vs. 2.65 .Q-cm for Al) [Cha89, Kol91] Copper has ten orders of magnitude higher electromigration resistance than Al, which will increase device lifetime [Mur95], and its lower resistivity will decrease the resistance-capacitance (RC) time constant, which is a large factor in chip performance for ICs with feature sizes below 250 nm [Hu98]. The following relation defines the resistance-capacitance time delay: (1-1) where pis resistivity (.Q-cm), I is the length (cm) of the interconnect line Eis the electric permittivity (CN*cm) of the insulating film, tM is the thickness of the metal interconnect (cm) and tILo is the thickness of the neighboring insulator (cm). A decrease in the RC

PAGE 12

3 time constant leads to higher device switching speed [Pau87], and the switch from Al to Cu may lead to a 15% increase in overall microprocessor speed [IBM97]. While bulk Cu has a resistivity of 1.67 .Q-cm, the resistivity of IC Cu lines with thickness above 50 run ranges from 1.7 to 2.0 .Q-cm [KalOO]. In comparison, Cu-doped Al lines Al(Cu) have resistivities ranging from 3 0 to 3.5 .Q-cm In addition to resistivity and electromigration improvements, Cu is said to provide higher IC production yield than Al-based devices with similar design [Sin02a]. Other advantages of switching to Cu interconnects include a decrease in the number of interconnect levels, a roughly 30% decrease in power consumption for operation at a given frequency, and a cost savings of roughly 30% per interconnect level due to integration of dual damascene processing [NovOO]. Silver (Ag), which has an even lower bulk resistivity (1.59 .Q-cm) than Cu, has been considered for interconnect metallization, but it suffers from poor electromigration resistance [Mur95], making its use as the metallization material unlikely. The improved electrical properties of Cu are highly desirab l e, but with these benefits come several drawbacks. These include copper's tendency to corrode under standard fabrication conditions copper's lack of a stable self-passivating oxide (like Ah03 on Al), the inability to chemically etch Cu poor adhesion to low dielectric constant (low-k) materials and rap i d diffusion/reaction with Si and Si02 [Hu98, MerOl]. Copper diffusion into neighboring layers can cause an increase in contact resistance, a change in the barrier height, leaky p-n junctions, embrittlement of the contact layer, deep level traps and destruction of electrical connections to the chip [KalOO, Wit80]. Copper

PAGE 13

4 contamination levels in Si at and above 1013 cm-3, for example, are believed to reduce device yield [lst00]. 1.2 Comparison of Diffusivities for Al and Cu in Si At high temperatures, both Al and Cu diffuse into Si, and when put into direct contact with Si, they will intermix rapidly. While Si is non-reactive with Al [RamOO], it has a small solubility range in Al (0.25 to 1.5% by weight) [Jae88]. At high temperatures (> 1000C), Al diffuses substitutionally into Si, where Al replaces a Si atom in the Si lattice. Copper, though, has much higher mobility in Si, and is very reactive with Si and SiO2 substrates [Bau97], forming copper silicide compounds, such as Cu3Si and Cu5Si [Bau97, Hym98, Kwo95, RamOO, Rei92], which cause strong deterioration of contact systems [Wan0lb]. Copper's high mobility in the Si lattice extends over a wide temperature range. Cu moves into the Si lattice by substitutional diffusion at high temperature (>800C), but also by interstitial diffusion at low temperature ( <700C), where Cu atoms diffuse between the Si atoms in the lattice (through the interstices). Both interstitial and substitutional diffusion of Cu into the Si lattice occur simultaneously, with interstitial diffusion dominating at low temperature and substitutional diffusion dominating at high temperature. While the diffusivity of Al in Si at 500C is less than 1x10-20 cm2/s [lstOO], the diffusivity for Cu in Si depends both on type and concentration of dopants in the Si substrate. The nature of p-type silicon is such that acceptor atoms (such as boron, B) substituted on Si lattice sites have a negative charge (B-). Since Cu diffuses through Si as a positive ion (Cu+) [Bai96 Ist00], the negatively charged Bions tend to "trap" some

PAGE 14

5 of the diffusing Cu ions, giving Cu an apparent diffusivity that is lower than that in intrinsic (undoped) Si. In p-type Si with a boron concentration of 5.0 x 1020/cm3 for example, the diffusivity of Cu is 2.0 x 10-6 cm2/sec at 500C, while that in intrinsic Si is 2.0 x 10-5 cm2/sec [lst00]. At room temperature, the diffusivity of Cu in p-type and intrinsic Si, measured by the transient ion drift method [Hei97], ranged from 3.3 x 10-13 cm2/s in p-type Si up to 2.7 x 10-7 cm2/s in intrinsic Si [lst98]. This indicates that after 30 minutes at room temperature, Cu can diffuse a distance of 0.5 m in p-type Si and 400 min intrinsic Si. Since the diffusivity of Al in Si is so small at the processing temperature, the main problem with Al arises when it penetrates the barrier material and accumulates at the barrier-Si interface. This accumulation can lead to formation of Al spikes at the interface, which short out shallow junctions (Figure 1-2a). Copper's high diffusivity in Si however, means that Cu will not build up at the barrier-Si interface until the Si has reached its saturation value of Cu. Hence, Cu can diffuse into the bulk of the Si wafer and move laterally to contaminate a large area on the wafer [lstOO], as depicted in Figure 1-2b. While Cu has high diffusivity in Si at room temperature its solubility is very low (less than 1 Cu atom per cm3 ) [lst00], so formation of a Cu-Si solid solution is unlikely. The diffused Cu will therefore form complexes, precipitates and agglomerates with Si. Four typical reactions can occur for Cu in Si. These include formation of point defects/complexes, copper silicide precipitates, addition to existing defects, and outdiffusion to the Si surface. More information on Cu interdiffusion and reaction mechanisms in Si is available elsewhere [lst00]. In addition to Si, Cu also diffuses

PAGE 15

6 rapidly through SiO 2 under thermal and electrical stress, leading to high leakage currents and dielectric breakdown [Jai99]. Diffusion of Cu atoms into SiO2 typically occurs due to application of an external electric field, which causes Cu to be ionized (to Cu+) and accelerated into the SiO 2 [Kri0l]. Copper's high diffusivity in Si and SiO2 makes the need for a robust thin film diffusion barrier to separate Cu from neighboring layers on the IC very critical. Figure l-2c depicts a stable device structure resulting from use of a diffusion barrier. Diffusion Region I s~D lc,Ul<,j1 ILDJ Diffusion Barrier Figure 1-2. Metal to silicon contact layers: (a) Al on Si without a diffusion barrier suffers from junction spiking. (b) Cu on Si without a barrier suffers from massive diffusion (c) Metal to silicon contact with a diffusion barrier remains intact. 1.3 Diffusion Barrier Requirements and Properties A diffusion barrier is a material used to separate two layers that, if put in direct contact, would interdiffuse and/or react with one another. Diffusion barriers are typically classified into one of three types: passive, sacrificial and stuffed. Passive (ideal) barriers are inert with respect to the layers that they separate, and have low solubility for the neighboring semiconductor and metal. Sacrificial barriers react with one or both of the neighboring layers, and are eventually consumed. The rate of reaction between the barrier and neighboring layers must be slow enough that the lifetime of the barrier is longer than that of the device. If the barrier has a shorter lifetime than the device, the sacrificial barrier will be totally consumed and the device will fail prematurely. Stuffed

PAGE 16

7 barriers are polycrystalline films that have their grain boundaries stuffed with a material that blocks diffusion [Wol86]. By stuffing these boundaries, the dominant low-temperature route for diffusion through the barrier is minimized. Stuffing occurs due to both physical and chemical effects. The stuffing "agent" should physically reduce the amount of pore space in the films and should chemically repel the Cu from diffusing through the grain boundary [Par95a]. As an alternative to grain boundary stuffing, high temperature annealing has been used to "cure" deposited barrier films, so that micro-defects such as grain boundaries are removed. Although this technique does decrease the number of grain boundaries in the film, the high temperatures also increase diffusivity of Cu in the barrier film, and can cause Cu penetration [Bai96]. The "curing" properties of a high temperature thin-film anneal are therefore tempered by the increased mobility of Cu in the barrier film at higher temperatures. In addition to stuffing the grain boundaries, barrier effectiveness can be further enhanced by minimizing the number of grain boundaries in the barrier film. Deposition at low temperatures provides this benefit and several more. First, low temperature deposition minimizes the number of grain boundaries by favoring amorphous film growth. Second in nitride films, for example, excess nitrogen is retained at lower deposition temperature and is therefore available to stuff any grain boundaries that may form during thermal cycling. Third, the likelihood of damaging temperature sensitive components (such as low k dielectric polymers) on the device during barrier deposition is minimized, provided that a temperature ceiling of 400C is maintained [Eis00a, Ele02, Hau00, Kim03b, Les02, Sun0lb]. Deposition above 450C can increase the compressive stress in Cu films which is relieved by formation of Cu hillocks [Jai99]. Fourth, low

PAGE 17

8 deposition temperatures require a smaller thermal budget. The advantages provided by low temperature deposition can greatly reduce the probability of Cu penetration through the barrier and decrease power consumption In practice, low deposition temperatures are desirable, but depositing the barrier films at the lowest possible temperature may not be the optimal approach. Depositing the barrier film at the highest allowable processing temperature, assuming that the film structure remains amorphous at this temperature, would prevent significant changes to barrier film structure during subsequent thermal cycling. During IC processing, the diffusion barrier material is typically deposited between a metal layer (usually Al or Cu) and a semiconductor layer (usually Si or GaAs) a metal and a dielectric (e.g., Si02 or low-k), or two metal layers [Lin98a]. The requirements for an ideal diffusion barrier are listed below. Film structure, electrical properties, conformality and adhesion property requirements will be discussed further in the following sections. Barrier must prevent diffusion between and be non-reactive with neighboring device layers. Barrier should have amorphous film structure to eliminate grain boundaries, which are facile paths for Cu diffusion through the barrier. Barrier should be deposited at low temperature to prevent damage to temperature sensitive components on the IC device and to enable amorphous film growth. Barrier should have good conformality and uniformity across the wafer to ensure good barrier coverage over small device features and uniform barrier deposition on all devices across the wafer. Barrier should have low bulk electrical resistivity and low contact resistivity to Cu to minimize resistance to current flow. Barrier should have good thermal conductivity to minimize heating of the barrier layer.

PAGE 18

9 Barrier should have minimal contamination levels ; hal i de impurities can cause corrosion of the Cu layer while oxygen and free carbon impurities can increase film's electrical resistivity. Barrier should have minimum thickness to maxnmze copper's cross-section in the interconnect and to foster low contact resistance in the metallization stack. Barrier should promote adhesion between device layers as poor adhesion leads to electromigration of Cu at the de-adhered interface. Barrier should have high thermal and structural stability to prevent failure of the barrier during exposure to thermal and mechanical stresses at processing conditions. Barrier should enable direct Cu electroplating to eliminate the need for a Cu seed layer and promote nucleation of the Cu (111) orientation, which has the greatest resistance to electromigration. Barrier should be compatible with chemical mechanical planarization (CMP) processes and act as a good CMP stop layer to eliminate need for a separate stop layer deposition step 1.3.1 Film Structure The ideal diffusion barrier structure is a defect free single crystal film [Nic78]. This structure does not contain grain boundaries, which are interface regions ("micro-defects") between crystal grains in a polycrystalline material. Grain boundaries are a "short circuit" path for rapid intermixing of the two neighboring layers, and are the dominant cause for barrier failure at low temperature [Bai96]. In a single-crystal barrier film, all diffusion occurs through bulk defects (including vacanc i es and dislocations) which is inherently much slower than diffusion through grain boundaries. Deposition of single crystal barrier films is impractical, however, due to low growth rates, latt i ce mismatch with the underlying substrate and deposition temperature limitations [KalOO]. The next best solution is to depos i t an amorphous dense smooth defect free film [ Bai96, Kwo95]. Amorphous films have short-range order(~ 5 A), but no long-range order(~

PAGE 19

10 20 A) [Ell90]; therefore no polycrystalline gram boundaries exist to enable rapid diffusion. In addition to having an amorphous microstructure, the density of the barrier material should be as close to the bulk value as possible, to eliminate voids as possible diffusion paths. Polycrystalline films, especially those with a columnar microstructure or those with equiaxial grains of size similar to film thickness, are the poorest performers in diffusion barrier applications These films contain grain boundaries that can extend from one side of the barrier to the other, meaning an easy diffusion path for Cu through the barrier. Nano-crystalline films, which are polycrystalline films with grain size below -50 A [KalOO], are more effective than polycrystalline films with larger grains, but are not as effective as amorphous diffusion barriers. Examples of single-crystal, polycrystalline and amorphous films are shown in Figure 1-3. a Polycrystals Grain Boundary Polycrystals Grain Boundary d Figure 1-3. Diagram of (a) Single crystal barrier. (b) Polycrystalline barrier with equiaxial grains. (c) Polycrystalline barrier with columnar grains. (d) Amorphous barrier without grain boundaries. Several different methods are used to deposit amorphous films [Siv95]. The first 1s to deposit films at low temperature which enhances sticking probability and suppresses surface diffusion, forcing films toward a disordered state The second is to deposit films at high nucleation (arrival) rates, so that many small nuclei form on the substrate surface preventing coalescence of small nuclei into larger polycrystals. The

PAGE 20

11 third is to add contaminants (such as N, C, etc.) to impede surface diffusion of the metal species, imparting disorder and decreasing the ability of nuclei to coalesce and grow. In addition to these deposition methods, careful selection of materials can assist formation of an amorphous film. A material's composition may be chosen, for example, so that several stable phases exist together at equilibrium. In a binary system, a film composition should be chosen in a two-phase equilibrium region, while in a ternary system, a three-phase equilibrium region should be selected [Ram00]. The competing stable phases neighboring these regions should be line compounds (have minimal solid solubility regions), have very different compositions, and have complex crystal structures [Ram00]. Compared to a broad solid solution phase, line compounds are inflexible to stoichiometry deviations, which can disrupt polycrystal formation and trigger amorphous growth. In a binary system, then, choosing a film composition in a two-phase equilibrium region bounded by two line compounds should increase the likelihood of amorphous film deposition. 1.3.2 Electrical Properties To obtain any benefit from the lower bulk resistivity of Cu, the barrier (liner) thickness should be less than 10% of the overall linewidth [Hu98]. A thicker barrier means decreased cross-sectional area of Cu available to conduct current, which causes an increase in the effective resistivity of the Cu line. For a Ti/fiN/Al(Cu)/fi/fiN sandwich structure, the effective resistivity was estimated to be 4.1 .Q-cm, due to the high resistivities of the Ti and TiN layers, as well as that for TiAh, which forms during annealing at the Ti-Al(Cu) interface [Hu98]. For Cu lines with widths ranging from 0.1

PAGE 21

12 to 1 m, keeping barrier thickness < 10 % of the overall linewidth results in a constant effective resistivity:::; 2.3 .Q-cm for the Cu line [Hu98]. The barrier material will typically be used in two types of scenarios on the integrated circuit, which are depicted in Figure 1-4. In the first case, the barrier runs in parallel with the Cu line, as with a trench structure, where the barrier separates the Cu trench from neighboring dielectric layers. In this case, the thickness of the barrier must be minimized, in order to have maximum cross-sectional area for the Cu conductor line. The resistivity of the barrier in this scenario is not critical, however, as electrical current will be flowing through the Cu and not directly through the barrier [Tra03]. In the second case, the barrier will be separating one metal layer from another, as in a via structure, hence the current will pass directly through this intermetal barrier layer. The resistance associated with current flowing from one Cu line through a via (and barrier) to another Cu line is called the via resistance. The barrier film's thickness and bulk resistivity must be minimized to limit the barrier's impact on via resistance. The suggested ideal film resistivity is :::; 500 .Q-cm [Ele03], and barrier thickness should be <30 nm [Bai96]. Specific via resistance (.Q-m2 or .Q-cm2), defined as the electrical resistance through the via multiplied by the via s cross-sectional area, is a common measure of barrier impact on current flow through a via The current, acceptab l e specific via resistance value for the 100 nm device node is 0.1 .Qm2 (1 x 10-9 .Q-cm2), and this is projected to decrease to <0.01 .Q-m2 (1 x 10-10 .Q-cm2 ) for the 22 nm device node in 2016 [ITR02, Nic95 Tra03]. As an example the change in specific via resist a nce with ITRS feature diameter for a fixed 15:1 via aspect ratio was calculated. The via was treated as a cylinder wit h

PAGE 22

13 Cu in the center and the barrier layer lining the inside edges of the cylinder. Current is assumed to travel in the Cu cross section of the via, which is the path of least resistance, and then passes through the barrier layer at the via's bottom, as depicted in Figure 1-5. Interrnetal Barrier Trench Barrier Via Barrier Figure 1-4. Schematic of trench and via applications for barrier materials. a Cu Barrier Figure 1-5. a) Cut away view of Cu trenches with via contact. b) Expanded, 3D view of Cu via with barrier layer on sides and bottom. The thickness of the barrier layer corresponded to ITRS projections for each device node, as follows: 12 nm (in 2003), 10 nm (2004), 9 nm (2005), 8 nm (2006), 7 nm (2007) 5 nm (2010), 3.5 nm (2013), and 2.5 nm (2016) The results for varying barrier resistivities, along with projected requirements from the ITRS Roadmap are shown in Figure 1-6. The diagram indicates that by the 65 nm node (2007) the barrier layer's

PAGE 23

14 resistivity must be~ 200 .Q-cm to meet ITRS projections. Likewise, this value must be 25 .Q-cm by the 32 nm node (2013). l.2e-9 -.-----------------------, ,....._ "' 8 l.0e-9 I a .._, 8 8.0e-1 Cl) -~ 6 0e-1 0:: ro > 4.0e-1 C) lC u 8_ 2.0e-1 CZ) 0.0 --10 -cm --CJ25 -cm -6--50-cm -<>100 -cm -0-200 -cm --CJ300 -cm --.-ITRS Projections 22 32 45 6570 80 90 100 Feature Diameter (mn) Figure 1-6. Specific via resistance as a function of feature diameter, shown for various barrier layer resistivities. Calculation is for a via with a 15: 1 aspect ratio. 1.3.3 Conformality Conformality (or step coverage) in device microstructures, especially in trenches and vias, is of utmost importance in barrier design. Highly conformal films (approaching 100%) have nearly uniform thickness at all points on a substrate surface (on both the sidewalls and bottom). Film conformality is determined by measuring the film's thickness on the wafer surface (ts), on feature sidewalls (tw) and at feature bottom (tb) (Figure 1-7). Conformality on the sidewall and bottom can be calculated by Equations 1-2 and 1-3, respectively. The feature's aspect ratio, which is the ratio of the feature's height to its width, can be calculated by Equation 1-4. Sidewallconf ormality (%) ( :: } 100 (1-2)

PAGE 24

15 Bottom conformality (%)=( :: )-100 (1-3) Aspect Ratio = ( J (1-4) Substrate Figure 1-7. Sketch of film dimensions for calculation of conformality. Barrier Film Substrate Substrate Substrate a b C Figure 1-8 Diagram of (a) Ideal barrier conformality. (b) Good barrier conformal i ty. (c) Poor barrier conformality. Barriers with poor conformality have uneven thickness, with the film being thinner in certain spots than others. These thin spots are "weak links" in the barrier, which are more susceptible to diffusion than the thicker parts of the layer. Figure l-8a shows an ideal barrier with 100% conformality Figure l-8b shows a practical barrier with rounded edges, while Figure 1-8c shows a poor barrier with a large overhang near

PAGE 25

16 the trench opening and sparse coverage at the trench bottom. Cu can easily penetrate to the underlying Si substrate through the thin barrier at the bottom of the trench. While device dimensions continue to shrink, the aspect ratio continues to increase, and the ability to conformally cover the bottom and sidewalls of these features becomes even more challenging. Deposition methods and chemistries must be developed to accomplish this task. 1.3.4 Adhesion The adhesion of the barrier to both the Cu and the neighboring dielectric or semiconductor layers, along with the crystallographic orientation of the deposited Cu (which is affected by adhesion issues), all impact copper's electromigration behavior [Jai99]. Strong Cu adhesion to the barrier layer contributes to good electromigration resistance [Pet03], whereas poor adhesion can cause self-diffusion of Cu along the barrier interface. Self-diffusion of Cu along the interface is even more likely than diffusion along grain boundaries, because grain boundary self-diffusion has a higher activation energy than interface self-diffusion [Llo95]. Copper has an activation energy for bulk self-diffusion of 2.19 eV [Mur95], while the values for grain boundary and interface self-diffusion are 1.2 and 0.8 eV, respectively [Llo95]. Cu (111) is the preferred orientation for the metal, as this plane has minimum surface energy, which results in formation of low-angle grain boundaries [Llo95]. These low-angle grain boundaries minimize flux divergence of electrons flowing through the metal, and therefore minimize electromigration [Jai99]. Aside from the barrier-Cu interface, adhesion at the barrier-dielectric interface can be adversely affected by out-gassing of

PAGE 26

17 moisture or other organic materials from the dielectric surface, which can lead to anomalous Cu removal behavior. 1.4 Dielectric Material Considerations In addition to the changeover to Cu, parallel research is pursuing new, low-k materials, which will also decrease the RC time constant. Historical l y Si02 with a k value ranging from 3 9 to 4 1, was the dielectric material of choice [Hu98]. Transitioning from Al/Si02 devices to Cu/low-k based devices (with k-2) may lead to more than a 400 % reduction in RC delay [KalOO]. New films (such as polyimide, SiLK, SiOC, etc.), with dielectric constants below 3.9, are being investigated for use as low-k mat erials, but they cannot tolerate high temperatures during processing [Hu98, Pet03]. The processing temperature limit for these materials, along with the need for good adhesion to Cu, are factors that must be considered when selecting a barrier material to separate them. Another factor influencing barrier material selection is its compatability with high-k materials, such as those in MOSFET gates o r in DRAM capaci tors. Barrie r materials have been used as top electrodes on high-k gates and DRAM capacitors and their favorable performance may dictate further use in the future. 1.5 Diffusion Barrier Failure Mechanisms Thin films typically have higher diffusivity than thicker bulk films due to a large number of short circuit paths (grain boundaries dislocations, etc.) distributed over a very small volume [Bal75]. Thinning the barrier film increases the likelihood that these short circuit paths will extend from one side of the film to the other, making it more susceptible to diffusion. Thin film diffusion barriers typically fail in one of two ways. The first is metallurgical failure where the Cu content o f the barrier increases t o seve r al atomic

PAGE 27

18 percent, changing the barrier layer's composition. The second is electrical failure, where little intermixing with the barrier occurs, but the Cu penetrates the barrier and is present in the semiconductor/dielectric layer in sufficient quantity to alter the characteristics of the device [IstOO]. Most barrier testing techniques, such as depth profiling or SECCO etch tests [Iva99], are suited for detection o f metallurgical failure Use of XRD to detect CuSix compounds, for example, is a fairly insensitive detection method, because the bulk concentration of Cu in Si must already be above saturation before the silicide compounds will form. Measurement of electrical failure is a more sensitive way to determine if Cu has penetrated the barrier and intermixed with the underlying substrate. To detect electrical failure, the electrical properties of devices such as p-n junctions, Schottky diodes or metal-oxide-semiconductor (MOS) capacitors, built under the Cu-exposed barrier or on the wafer surface after the barrier has been removed, are measured. Leakage currents are measured on p-n junctions and MOS devices and compared to Cu-free ones to determine if Cu has penetrated the barrier. Changes in the capacitance and current voltage (l-V) characteristics of Schottky diodes are measured to determine if Cu has penetrated into the active areas on devices. Ideally, the performance of different d i ffusion barrier materia l s would be tested by comparing the diffusivity of Cu in the different barriers. To determine diffusivity the migration of atoms, or diffusive flux (J, atoms/cmLsec), through the barrier material should be estimated. The flux can be described by Fick s law (Equ a tion 1-5): J =-D( :~) (1-5)

PAGE 28

19 where D is the diffusivity (cm2/sec), C is the atomic concentration (atoms/cm3 ) and x is the diffusion distance (cm). By measuring flux (J) through the barrier for a given concentration gradient (dC/dx) across the barrier, the diffusivity can be determined. Once this is known the next level of analysis is to determine whether diffusion through the bulk lattice or the grain boundary dominates. To distinguish between di ff us i on through the bulk lattice and the grain boundary two different expressions for diffusivity have been given [Ohr92]. Equation 1-6 gives the expression for lattice diffusivity (DL): (-E) DL = DoL exp RT (1-6) where EL is energy per mole for atomic diffusion through the lattice, and DoL is the value for lattice diffusivity at standard conditions. Equation 1-7 gives the expression for grain boundary diffusivity (De): ( -E1/R) DB= DoB exp RT (1-7) where Ee is the energy per mole for atomic diffusion through the grain boundary, and D08 is the value for grain boundary diffusivity at standard conditions. At low temperature, the diffusion mechanism through a polycrystalline solid is typically controlled by grai n boundaries and other defects in the film [Nic78]. Since vacancy-assisted bulk diffusion is negligible below about three-tenths of the solid's melting temperature (0.3T m ) diffusion through the grain boundary is the path of least resistance at low temperature This is reflected by a lower activation energy for grain boundary diffusion relat i ve to bulk diffusion [Cha82, Nic78], with Ee :::: 0.5EL for FCC metal thin films [Ba l 75]. As annealing temperature is increased a changeover from grain boundary controlled diffusion to bulk diffusion occurs. The temperature at which this occurs, called the

PAGE 29

20 Tammann temperature, is typically 0.5-0.66Tm [Nic78]. At higher temperature, the values for DL and D8 begin to approach each other, and flux through the bulk lattice becomes the dominant route for diffusion, due to the much larger cross sectional area of the bulk lattice relative to that of the grain boundaries. Experimentally, determination of diffusivity from Cu flux measurements is impractical. Exact measurement of the Cu flux through the barrier is d i fficult to determine, as failure determination techniques usually focus on qualitative barrier failure. Moreover issues such as defect density can have a large impact on th i n film diffusivity and can cloud efforts to distinguish between bulk lattice and grain boundary diffusion. Diffusivity in thin barrier films may be estimated with Equation 1-8, which describes diffusion from a limited source at a crystal surface [W 0186]: C Qo ( -d ) k = ~exp 4Dt v1tDt (1-8) where Ck is Cu concentration (cm-3 ) at distance d (in cm) from t he surface D is the diffusivity (cm2/s), t is diffusion time (sec), and Q0 (cm -3 ) is the concentration of Cu at the surface. Since exact Cu concentrations at the two interfaces can be difficult to measure, an apparent diffusivity (Dapp) is defined by Equation 1-9 [lst00]: d 2 D =a p p 4t (1-9) where d is the film thickness (in cm) and t is the time (s) for the Cu to appear at the barrier/substrate interface (i e ., the time it takes the barrier to fail). Comparison o f apparent diffusivities for different barrier films may be done to determine which barrier i s best.

PAGE 30

21 1.6 Diffusion Barrier Deposition Techniques Diffusion barrier films have been deposited by a variety of methods including physical vapor deposition (PVD) and chemical vapor deposition (CVD) techniques The required deposition temperature, substrate characteristics and feature size general l y dictate which deposition method is appropriate for a given application. A discussion of some common PVD and CVD techniques to deposit metal nitrides and the i r inherent advantages and/or disadvantages follows. 1.6.1 Physical Vapor Deposition (PVD) Typical physical vapor deposition (PVD) me t hods to deposit barrier films include electron beam evaporation and sputtering In electron beam evaporat i on, an electron beam is used to heat a small portion of the solid metal above its melting point. While at high vacuum (10-3 to 10-4 Torr) the molten metal then evaporates a t oms into the gas phase. These atoms diffuse through the chamber's atmosphere and deposit on the susbtrate. One or more reactive gases, such as NH3 or C~, for example, may be introduced into the chamber during evaporation to deposit multi-component fil ms. The sputtering process involves the use of a plasma gas (typically argon) t o physically "knock" atoms off of a metal target and onto a substrate. The sputte r chamber operates at low vacuum (-100-120 mTorr), and cations from the plasma impact the metal target cathode to liberate metal atoms which then deposit on the substrate. In the reactive sputtering process, a reactive gas such as N2 for example, is introduced i n t o the sputter chamber along with the plasma gas. Once the atoms liberated from the me t al target cathode reach the substrate they can interact with the nitrogen r a dicals formed by

PAGE 31

22 the plasma and deposit either a metal nitride compound or a metal-nitrogen solid solution. Films produced by PVD processes suffer from high stress, a large number of vacancy defects, excess free volume and poor conformality, leaving spots in the barrier open to diffusion by neighboring species at via and trench sides/bottoms [Aff85, Bau98, Bos91, Kel99, Lee93, Lu98b]. The poor conformality inherent to all PVD processes is due to the directionality imparted to the atoms/clusters traveling toward the substrate. This directionality causes step shadowing, where parts of the substrate surface are not "seen" by the incoming sputter atoms. Step shadowing results in little or no film coverage on certain sections of the substrate These exposed substrate areas are then vulnerable to reaction with subsequently deposited Cu atoms. Figure 1-9 depicts step shadowing. Directional Particles Directional Particles Direc ti onal Part i cles \\\ r r t tr Substrate Substrate Substrate Substrate Substrat e ,. Substrate I Barrier Film / wall i. wall i Shadowed Barner Barrier walls F i lm Film a b C Figure 1-9. Step shadowing effects in small device features due to di r ectiona l particles arriving at the substrate. a) At an angle to the surface. b) Perpendicular to surface, without resputter. c) Perpendicular, with resputter Three variations on sputtering have been developed to extend i t s applicability to smaller device feature nodes The first collimated sputtering involves the placement of

PAGE 32

23 a plate (which contains small holes) between the sputter target and the substrate [Sin02b]. The plate collimates the sputtered atoms/clusters, so that only those traveling parallel to each other in a direction perpendicular to the plate reach the substrate. Once at the substrate, the atoms can deposit or can resputter the metal film at the bottom of a feature (Figure 1-9c), moving it up onto the feature walls to ensure sidewall conformality. The second variation, long-throw sputtering, improves the performance of the sputter deposition system by increasing the distance between the target and substrate, so that only sputtered atoms/clusters travelling in parallel (and perpendicular to the substrate) reach the substrate [Sin02b]. Ionized-PVD (1-PVD) is the third variation used to extend the applicability of sputtering processes beyond the 100 nm device node. 1-PVD involves use of a second rf coil to ionize neutral metal atoms that have been liberated from the target by sputtering. These ionized atoms are then accelerated toward the substrate, which is biased to attract the newly ionized metal atoms [Ros0l]. Impinging metal atoms resputter the metal that has already deposited on the substrate, and these resputtered atoms can move from the bottom of a feature to the sidewalls and coat them. This leads to improved step coverage in small features compared to other PVD methods. Conformality improvements within the feature from the 1-PVD resputter effect are tempered by resputter thinning (beveling) of the feature's top edges (Figure 1-9c). This beveled edge is a weak point in the barrier, and is susceptible to attack by Cu. In addition, although resputter associated with 1-PVD improves sidewall coverage, this coverage is still tenuous for films in high aspect ratio features [Vij99]. These extended sputtering techniques being directional in nature, also have a limited lifespan [Sin02b], and may reach their usability limit at the 45 nm device node in 2010. Among the barrier

PAGE 33

24 deposition techniques, only a chemical vapor deposition (CVD) based process offers the conformality required for future device generations, with 100% conformality being the goal for all barriers [Fal98, Fle98, Lev98, Liu00]. 1.6.2 Chemical Vapor Deposition Chemical vapor deposition (CVD) is a widely used film growth technique. It involves the use of one or more gas phase reactants to deposit a solid film on a substrate. The reactant flux at the substrate surface is non-directional i n nature, eliminating the possibility of step shadowing in small device features. The substrate promotes react i on between components from the gas phase, and successive reactions lead to film growth. CVD processes typically grow highly conformal films a t high growth rates making them superior to PVD techniques for deposition on aggressive device features Unlike PVD methods, CVD is capable of area selective growth, which can eliminate a patterning step during device fabrication [Gat96]. CVD techniques are also characterized by high throughput, minimal downtime and easy source changeout, which makes them v ery attractive for industrial applications. The steps of the CVD process are as follows [Siv95 Vos78] : Reactants in the gas phase diffuse through the boundary laye r (defined below) to the substrate surface Reactants adsorb onto the substrate surface Adsorbed species move around surf ace and settle into available surface sites Reactants adsorbed to available surface sites undergo fi nal reactions which are often catalyzed by the substrate surfa c e and rea c tion produc t s are incorporated into the grow i ng film By-products from the reaction desorb from the substrate surfac e

PAGE 34

25 Desorbed by-products diffuse through the boundary layer and away from the substrate surface Conversion of gas phase precursors into deposited film can involve both gas phase and surface reactions. Gas phase reactions can involve formation of intermediates and by-products as well as parasitic reactions, such as gas phase nucleation. Surface reactions can involve adsorption/desorption, film formation, and side reactions, which can cause contaminant incorporation into the film. The total pressure in the reactor determines the degree of coupling between gas and surface chemistry [Gat96]. At a total reactor pressure of 1 Torr or greater, homogenous and heterogeneous processes are tightly coupled, while at pressures < 10-4 Torr (where Amtp 50 cm), the probability of gas phase collisions becomes negligible and the process becomes strictly surface controlled [Gat96]. This surface control at low pressure enables finer control of film properties, such as smoothness and thickness. At higher reactor pressures (> 1 Torr), one or more of the gas phase precursors may begin to decompose and/or react to form intermediate gas phase compounds [Gat96]. These precursors (and/or intermediates) can then travel to the substrate surface. When the precursor ( or intermediate) molecules first reach the substrate surface, they can weakly bind (physisorb) to the surface by van der Waals forces or they can immediately react and deposit (chemisorb) [Mas96]. Physisorbed molecules have some degree of mobility, and can diffuse along the substrate surface until they either chemisorb or desorb back into the gas phase. High surface mobility, coupled with the non-directional nature of the reactant flux, are the key CVD features that enable highly conformal film deposition on substrates with varying topography. The probability that an impinging precursor chemisorbs at the initial point of impact on the substrate surface is

PAGE 35

26 known as the reactive sticking coefficient. A precursor with high reactive sticking coefficient (approaching unity) and low surface mobility yields a rough film texture, while a precursor with a low reactive sticking coefficient (10-3 or less) and high mobility yields a very smooth film [Gat96]. Hence, to deposit smooth films by CVD, a low reactive sticking coefficient coupled with high surface mobility is desirable. Surface mobility and reactive sticking coefficient are highly dependent on substrate temperature, which must be controlled closely to ensure deposition of films with appropriate structure and properties. Temperature must be high enough to activate the reactions leading to CVD, but cannot be increased indiscriminately. At lower temperature, the sticking coefficient approaches unity and surface diffusivity is minimal [Gat96]. As temperatures increase, the sticking coefficient decreases, while the surface diffusivity and deposition rate typically increase. Higher deposition rates are likely on the flat surface of the substrate, but aspect ratio dependent effects (ARDE) can inhibit deposition in small features. ARDE occur when by-products emanating from the sidewalls and bottom of a small feature inhibit in-diffusion of fresh precursor to the feature. While increasing temperature increases reaction rate on the flat surface, ARDE can lead to a starved condition on the interior walls of the feature. Hence, although increased deposition temperature improves surface diffusivity and decreases sticking coefficient, it can negatively impact film conformality in small diameter, high aspect ratio features (Figure 1-8c). In addition to optimizing surface mobility and reactive sticking coefficient, the operating regime must be considered when selecting deposition temperature. At lower temperature, the deposition rate is exponentially dependent on substrate temperature,

PAGE 36

27 which is indicative of the kinetically controlled regime. Increasing temperature in this regime causes a large increase in film deposition rate, and the slow step in the deposition process is the reaction on the surface. At higher temperature, the deposition rate is relatively insensitive to deposition temperature, which is indicative of the gas-phase mass transfer controlled regime. In this region, the temperature is high enough that essentially all reactants reaching the substrates surface immediately react, and the rate-determining step in the deposition process is transport of reactants to the surface. While the temperature dependence of the deposition rate in this regime is mild (varying from T1.5 2 0), the lack of kinetic (surface) control in this regime exacerbates conformality problems Moreover, higher deposition temperature can alter film structure, causing a shift from amorphous to polycrystalline films, which introduces grain boundaries that can kill barrier performance. In small features that are to be conformally coated, the surface diffusion length of the physisorbed precursor should be the same order of magnitude as the feature diameter [Coo89]. In larger features, surface diffusion alone is inadequate for conformal coverage, so it must work in tandem with reflection, whereby molecules are reflected back and forth from the sides of the feature until they find a place to chemisorb. Decreasing the precursor's sticking coefficient increases the chance of reflection in a feature. For good coverage, the dimension of the feature should be smaller than the mean free path in the gas phase, so higher pressure (which has a shorter mfp) can be used with smal ler device features [Tsa86]. Pressures at and below 500 Torr would be adequate for deposition in a 100 nm feature.

PAGE 37

28 Molecules designed for use as CVD precursors should have high volatil ity, thermal !ability, and have easily (and cleanly) removable supporting ligands [Kod94]. The CVD chemistry should be selected to avoid formation of non-volatile react i on by-products, which can cause particle deposition on the substrate surface and high point defect density [Jai99]. In addition, selection of precursors with low reacti v e st i cking coefficient and high mobility surface species will enhance conformality of the deposited film. 1.6.2.1 Precursor Delivery During CVD, one or more gas-phase chemical precursors must be delivered to the reactor. The precursors themselves may exist in the gas phase, or the y made be put into the gas phase by a number of techniques. Liquid source precursors are typically put into the gas phase by a liquid bubbler system, where a canister contain i ng the liquid precursor sits in a temperature controlled bath (which controls vapor pressure). Carrier gas flows into the bottom of the can i ste r and bubbles up through the li quid As the gas bubbles up through the liquid precursor, it becomes saturated with precursor molecules. This carrier gas saturated with precursor, then flows into the reactor. There are two typical methods to convey low volatility solid precursors to the reactor: solid source and nebulizer delivery. In a solid source delivery system, the precursor is held in a metal or glass tube The tube and precursor are heated by thermal or photon energy, and solid precursor near the top of the tube sublimes and i s conveyed to the reactor by a carrier gas. As a general rule, a solid precursor must ha v e a vapor pressure higher than 10 -3 Torr at its melting point to be effectively sublimed and

PAGE 38

29 conveyed by a solid source delivery system [Ohr92]. If the vapor pressure is below this value at the melting point, minimal precursor delivery and film deposition will occur. The second technique, nebulizer deli very, can be used to overcome this vapor pressure limitation. This technique requires dissolution of the solid precursor in a liquid solvent. The solvated precursor is pumped into the nebulizer, which contains a piezoelectric quartz plate. This plate vibrates due to application of a high frequency electrical current, and vaporizes the liquid droplets that contact it. Once the drop l ets are vaporized into a mist, carrier gas flows through the nebulizer and conveys the mist into the reactor. Figure 1-10 shows a schematic of a nebulizer. Solvated Precursor/ Carrier Gas to CVD Reactor Vibrating Quartz Plate Cable to Power Supply Figure 1-10. Schematic of a nebulizer delivery system. 1.6.2.2 Variants of CVD Precursor Mist" Dissol v ed Precursor from Syringe Pump Carrier Gas to Nebulizer Several variants of CVD are typically used to deposit thin films. The first is low-pressure chemical vapor deposition (LPCVD), which typically relies on metal halide

PAGE 39

30 chemistry (e.g., TiC14 WF6 ) to deposit films. This technique usually requires high deposition temperature (>450C) and suffers from halide incorporation into the barrier films. The presence of halides in the barrier can lead to Cu corrosion, which decreases Cu-barrier adhesion and reduces electromigration resistance [Huo02]. The same halide chemistries used for LPCVD are typically used for plasma-enhanced chemical vapor deposition (PECVD), in which a plasma assists fragmentation of the reactants, thereby lowering the deposition temperature. After fragmentation by the plasma, the precursor fragments travel to the heated substrate, react on its surface and deposit a film. Reduced conformality for PECVD films in high aspect ratio trench features is due to the directional nature of the plasma [Tsa96], as shown in Figure 1-9. Despite the low resistivity and deposition temperatures that PECVD offers, its inability to deposit highly conformal films in high aspect ratio features makes its use in future barrier deposition processes unlikely. In metalorganic chemical vapor deposition (MOCVD), films are deposited by reaction of one or more carbon-containing vapor phase precursor compounds. MOCVD precursors typically have some or all of the halide ligands common to LPCVD and PECVD replaced with carbon-containing ligands. These precursors afford little or no halide incorporation into the deposited films, and MOCVD has been demonstrated to deposit a variety of films at low temperature. By varying the structure of the precursor molecule, the precursor can be optimized to dissociate in a "clean" fashion (i.e., with minimal oxygen and carbon contamination) and at relatively low deposition temperature. The presence of impurities (0, N, C) is reported to improve stability in contact structures, because the impurities tend to stuff grain boundaries and inhibit diffusion [So88].

PAGE 40

31 Oxygen's "stuffing" effectiveness, however, is lower for Cu than Al, because the reactivity of Cu with O is less than that for Al with O [Kim99c]. Carbon bound to the metal has also been reported to improve thermal stability of the barrier and to foster growth of smaller grains [WanOlb], but free carbon in the films can scatter electrons, increasing film resistivity. Among the challenges for MOCVD of refractory nitride materials are controlling carbon and oxygen contamination, lowering the deposition temperature, and minimizing film resistivity. 1.6.2.3 CVD Reactors CVD reactors fall into two broad categories: hot wall and cold wall reactors. Hot wall reactors have heated walls, with typically laminar flow profiles. These are preferred for exothermic reactions, since the hot wall temperatures discourage or prevent unwanted deposition on the reactor walls. In contrast, cold wall reactors have steep temperature gradients surrounding the susceptor, which often leads to convection pattern formation in the reactor. These are preferred for endothermic reactions, since the reaction will occur most readily on the hottest surfaces in the system [Vos78]. CVD reactors are typically run in two modes: differential or starved. In a differential reactor (similar to a CSTR), the ratio of reactant out to reactant m 1s approximately one, so that the composition in the reactor is essentially constant. In a starved (feed rate limited) reactor, the ratio of reactant out to reactant is much less than one, meaning that there are large concentration gradients in the reactor caused by fast reactions [Kod94]. ARDE can cause both differential and starved conditions to exist simultaneously on a substrate with aggressive feature topography. Flat surfaces on the substrate receive adequate precursor flux, and hence operate in a differential condition,

PAGE 41

32 while ARDE in small features prevent good precursor flux, starving the sidewalls and bottom. Increasing the deposition temperature increases the concentration disparity between the inside of the feature and the flat surf ace of the substrate, further degrading the conformality over these features. 1.6.2.4 CVD Transport Issues Heat and mass transport in a cold-wall CVD reactor can be complex, due to large temperature gradients. Depending on pressure in the reactor, mass transport may occur by fluid flow and/or diffusion [TimOl]. At higher reactor pressures (near atmospheric), both fluid flow and diffusion occur simultaneously. In fluid flow transport, gas molecules follow streamlines through the reactor, while for diffusion transport, concentration gradients drive transport across streamlines. Heat transport occurs by a combination of convection, conduction and radiation in the reactor. Near the deposition surface, the temperature, velocity and concentration vary significantly [Kod94]. Use of an impinging jet to feed reactants onto a small susceptor results m stagnation point flow. The stagnation point is at the center (origin) of the surface, where flow velocities are zero (where y=O in Figure 1-11). With adequate pressure, a shear layer of uniform thickness develops near the surface of the susceptor [Whi91]. The shear layer, also called the momentum boundary layer, is the relatively stagnant region between the surface of the substrate and the region where the fluid velocity (u) reaches 99 % of the free stream velocity (Uxi, cm/sec). Reactants and by-products travel to and from the substrate surface, respectively, by diffusing through this layer. This boundary layer (OM) displaces the outer inviscid flow away from the substrate surface, as depicted in Figure 111. Thickening of the low-velocity shear (boundary) layer due to viscous diffusion is

PAGE 42

33 balanced by thinning of this layer due to acceleration of the outer inviscid stream, which leads to uniform boundary layer thickness. The momentum boundary layer thickness (OM, cm) in stagnation point flow is approximated by Equation 1-10: (1-10) where vis the kinematic viscosity (cm2/sec) and Ds is susceptor diameter (cm) [Whi91]. More information on the dynamics of stagnation point flow may be found elsewhere [Whi91]. Figure 1-11. Stagnation-point flow [Whi91]. In addition to the momentum boundary layer, concentration and thermal boundary layers, with thickness denoted &: and &r, respectively, are also defined. The concentration boundary layer extends from the substrate surface to a point where precursor concentration is 99% of the bulk concentration, and likewise, the thermal boundary layer extends to a point where the temperature is 99% of the bulk temperature.

PAGE 43

34 The thickness of the boundary layers depends on several process variables, including gas velocity, temperature, and pressure. Gas phase species outside of these boundary layers (where y > OM, Oc, or o,,) move by both convection and diffusion. Within the boundary layer (where y < OM, Oc, or 0,,), velocity, concentration, and temperature are non-uniform. As species leave the gas and deposit onto the substrate, their gas phase concentration drops near the substrate surface. To compensate for this drop in concentration near the surface, a net diffusive flux occurs bringing depositing species from regions of higher concentration in the boundary layer to the near-surface region. As the surface reaction generates reaction by-products, the concentration of by-products near the substrate surface i s h i gher than it is at locations farther from the surface. This higher concentration of by-products near the substrate surface leads to a net diffusive flux of by-product away from the substrate Likewise, temperature varies from the substrate into the gas phase. The temperature at the substrate is high, and this drops steadily through the thermal boundary layer, until it reaches the bulk value. Typical CVD temperature and concentration profiles are depicted in Figure 1-12. pi,B Figure 1-12. Variation of temperature and precursor partial pressure near the substrate in a CVD reactor. Note that while &rand Be are shown as being equal t his is not always true.

PAGE 44

35 Transport to the substrate during CVD is assumed to occur by diffusion of reactants through the concentration boundary layer at the surface. This boundary layer arises due to consumption of the precursor at the surface, which causes a concentration gradient to form between the surface and the bulk. Flux of reactant through this layer (J, kgmol/mLsec) is given by Equation 1-11 [Kel91]: (1-11) where D0 is the diffusion coefficient (m2/sec) at temperature T0 (K), cSc is the thickness of the concentration boundary layer (m), Pb and Ps are the bulk and surface partial pressure of the reactant (Pa), Tb and Ts are the bulk and surface temperature (K) and R is the gas constant (m3-Pa/kgmol-K). Once the reactant species diffuse across the boundary layer, they may or may not be incorporated into the growing film surface. The mass transfer flux (J1r, kgmol/m2-sec), which is the amount of precursor incorporated at the film surface, is defined in Equation 1-12: kd(Ps -Peg) J =-----tr RT s (1-12) where kct is the mass transfer coefficient (m/sec) and P eq is the equilibrium precursor partial pressure (Pa) at the surface. When the deposition process reaches steady state, the diffusion and mass transfer fluxes are equal, and a dimensionless parameter called the CVD number, Ncvo, can be defined [Kel91]: kd8c T;ln(;{b) DoTJTs -Tb) (1-13)

PAGE 45

36 The denominator on the left-hand side of the equation represents the degree of precursor supersaturation at the substrate surface. When the amount of supersaturation is small, P5=Peq and Ncvo >>1, hence all reactant reaching the substrate reacts immediately. This is known as the mass transfer controlled regime. Conversely, if the supersaturation is large, Ps approaches Pb and Ncvo <<1, and the species reaching the substrate react very slowly. This is known as the kinetically (or surface reaction) controlled regime. When Ncvo >>1, high points on a rough surface have a higher value of Ps, and the growth rate at a high point is increased relative to a low point. This causes rough points on the surface to be amplified by film growth when operating in the mass transfer controlled regime. In contrast, when Ncvo <<1, Ps approaches Pb everywhere along the surface, so that the growth rate is similar at all points along the surface. This leads to a smoothing of rough spots during film growth when operating in the kinetically controlled regime. The type of carrier gas used in the reactor can have a significant impact on the films grown during CVD. For fixed bulk and surface temperatures, gases with a lower thermal conductivity will have a sharper temperature profile, meaning less upstream heating of the precursor species as it approaches the substrate. Using a carrier gas with higher thermal conductivity results in a smoother temperature profile, enabling more upstream heating of the precursor. Increased upstream heating promotes pyrolysis, or gas-phase thermal breakdown of the precursor into the intermediate or final reactant species. Increasing the time available for precursors to undergo pyrolysis increases the likelihood that the precursors have decomposed into the intermediate components (if necessary) for final surface reaction. If the precursors undergo sudden pyrolysis very

PAGE 46

37 near the substrate surface, they will not decompose as thoroughly, leaving large molecular fragments that can be incorporated into and contaminate the growing film. Ideally, transport of reactants to and products from the substrate surface should occur in such a way to deposit a "clean" film containing the desired components. In reality, this is unlikely, with atoms from reaction by-products and precursor ligand fragments typically depositing to some degree in the films. Control of contamination by ligand decomposition is extremely important to get films with desirable properties. An example of single-source CVD, which has some contamination from precursor ligands, is depicted in Figure 1-13. e Atom desired in film I) Atom not desired in film 'i -----~ -------. --,----o e '.1 1 ......... ... I Substrate Figure 1-13. Diagram of the CVD process for a single-source precursor. 1.6.3 Atomic Layer Deposition (ALD) Standard CVD techniques introduce all required precursor gases to the reactor s i multaneously making deposition rate and precise thickness control difficult during film

PAGE 47

38 deposition. Although the ITRS roadmap suggests that CVD deposition of barriers will be important in the near term, atomic layer deposition (ALD) methods will emerge as the dominant solutions because of their superior conformality and improved thickness control. As barrier thickness drops below 100 A, standard CVD techniques will approach their usability limit due to difficulties controlling the deposition rate. Atomic layer deposition (aka ALD, ALE or ALCVD) will be necessary to deposit extremely thin films on the IC devices [Dan02], and its use in IC production at the 65 nm node is projected [Bey02, Cha04]. The highly conformal, ultra-thin barriers afforded by ALD will be essential in the future to minimize the barrier's impact on the resistance per unit length in Cu interconnects [Kap02]. As a variant of CVD, ALD is well suited for deposition of ultra-thin, highly conformal films over small device features. ALD enables monolayer addition with precise thickness and composition control, irrespective of the underlying substrate's topography [Dan02]. ALD has been used to deposit metal, semiconductor, dielectric and seed layers [Dan02, KlaOOa-b, Les02]. In particular, ALD has been used to deposit high-k gate dielectrics at the front end and diffusion barriers at the back end of the process. ALD is a "digital" process, involving the stepwise use of two or more gas phase precursors, each of which is self-limiting on the substrate [RosOl]. A single precursor gas is present in the reactor at any given time, so that a uniform layer of the precursor may chemisorb to the substrate surf ace. Once this chemisorbed layer forms, the chamber is evacuated or swept with inert gas [Sun92] By minimizing chamber volume, rapid deposition and quick purge/vacuum steps are promoted This is important to minimize the incorporation of background impurities into the film and also to prevent the process

PAGE 48

39 from shifting to CVD mode, which can occur if adpitional reactant from the previous pulse is in the reactor's atmosphere during the subsequent pulse. A second precursor gas is then introduced to react with the chemisorbed layer from the previous step, forming a new monolayer of material. The second precursor gas is also self-limiting, so that reaction stops after the monolayer of chemisorbed material from the previous step has been consumed. Film thickness is more difficult to control for growth with CVD, which has been reported to close extremely narrow vias rather than depositing conformal films in them [Hau00]. ALD was first reported in 1977 by Tuomo Suntola to deposit ZnS films for electroluminescent displays [Goo86]. Since this first report, the utility of ALD to deposit a variety of materials has been demonstrated. Materials such as metals (e.g., W, Cu, Ni, Co, Ti, Ta, Ru, Pt, Al), metal oxides (e.g., Al203 Zr02 Hf02 ) and metal nitrides (e.g., TiN, TaN, WNx) have been deposited using this technique [Les02] In contrast to the thin, flat films typically desired by ALD, the process has also been demonstrated to coat porous, high surface area substrates for catalysis [Hau93]. ALD exploits the difference in energy between chemisorption and physisorption to deposit film layers in a self-limiting fashion [Gat96, Goo86]. Temperature is selected so that Echemisorptioo > kT > Ephysisorption In other words, the temperature is high enough to overcome physisorption forces to enable desorption of any physisorbed species, but it is also low enough to prevent removal of chemisorbed layers. For an adequate pulse time, complete surface reaction occurs and the film composition is determined by thermodynamics (stable thermodynamic phase forms for the selected conditions).

PAGE 49

40 The precursors must be volatile, thermally stable, self-limiting and highly reactive (fast, complete reactions) with each other and with the substrate. They must also have sufficient purity, have unreactive by-products, and not undergo self-decomposition or cause etching of the film or substrate. The precursors should have a vapor pressure of at least 0.1 Torr for delivery during ALD. Desirable ALD reactions should have large negative ~G values to ensure rapid surface reaction after the precursor pulse into the reactor [Les02]. High reactivity enables rapid saturation of the film surface, which in turn enables a good deposition rate. The need for high precursor reactivity is contrary to CVD precursors, which require smaller negative ~G values to prevent gas phase nucleation and particle formation. The precursor dose must be high enough to saturate the substrate surface, which must have reactive adsorption sites [HauOO]. In addition, the reaction temperature must be chosen to enable reaction ( chemisorption) between the precursor and reactive sites, while also avoiding gas phase decomposition or condensation of the precursors [HauOO]. The advantages of ALD relative to CVD are excellent conformality (-100% ), inherent elimination of pinholes, good thickness uniformity, and elimination of potential gas phase reaction/nucleation [Goo86, Rit03, Sun92, Sun93]. The disadvantages of ALD include slow growth rate (0.06-0.6 m/hr) [Goo86, RosOl], the possibility for substantial incubation time to deposit the first monolayer of film, contaminant incorporation and the potential for surface reconstructions to adversely affect deposition rate [RosOI, Sun93]. In the most aggressive applications, a diffusion barrier will need to be deposited on four different surface materials simultaneously: two different insulator materials, an etch stop layer (such as Si]N4 ) and Cu [Ele02]. Varying incubation times on the different materials

PAGE 50

41 can cause significant deposition difficulty, so choice of precursor and deposition conditions is essential to deal with situations like these. Precursor testing should therefore be done on all possible underlying substrates, to ensure that a given material can be reliably deposited with a given precursor. Another disadvantage is decreased film purity relative to CVD grown films. Background contaminants have considerable time to incorporate into the film during the pulse and purge steps, which can degrade film structure and electrical properties. The purity of purge gases and control of out-gassing from reactor walls and seals are critical, because inadequate control can lead to considerable impurity incorporation causing modification of film structure and properties. Typically, a faster growth rate means less sensitivity to contaminant incorporation [Han03]. ALO must compete with the purity of PVD films, whose growth rates are 2 to 3 orders of magnitude higher in an environment that is 3 to 6 orders of magnitude cleaner [Han03]. Several adjustable parameters are available for the ALO process. The first parameter is the type of ALO reactor system used for deposition. The two main categories for ALO systems are the open type (used with molecular beam epitaxy (MBE)) and the closed type (used with LPCVD) [Sun93]. In the open type system, pressure is fixed, as with an MBE system, which typically operates at ultra-high vacuum (UHV). Precursors are introduced from two or more sources (e.g., Knudsen cell), which are turned on and off to generate the pulsing action. Desorbed surf ace species collect on the cold walls in the reactor system or by the vacuum pump. In the closed type system, the atmosphere is turned over between reactant exposures to generate the pulsing action; this is done either by evacuating the system (pressure modulation) or by introducing an

PAGE 51

42 inert purge gas. In addition to the type of system used, the chamber volume may be adjusted to accommodate more substrates (i.e., to increase throughput) or to decrease purge/vacuum times. The second adjustable parameter is the type of precursors used for the deposition process. The precursors must be self-limiting, very reactive with each other and with substrate, and generate the desired film structure/stoichiometry. Many precursors have been tested in the literature for various materials, with halide chemistry often being used due to easy removal of the halide ligands (with H 2 NH3 etc.). In addition to precursor selection, pre-treatment of one or more of the precursors before introduction to the reactor is another adjustable option. For example, Ta films were deposited using plasma-enhanced ALD with sequential pulses of TaC15 and H 2 [K.im02b]. The H2 was cracked to atomic Hin an rf plasma system before being introduced to the reactor. The third parameter is substrate temperature, which should be selected to enable chemisorption of first layer and desorption of outer, physisorbed layers [Gat96]. A fourth parameter is the reactant pulse time. A longer pulse time enables more complete surface reaction, and closer approach to thermodynamic equilibrium at the film surface. The last parameter, which is unique to closed type systems, is reactor pressure. ALD process pressure in closed type systems typically ranges from 1 to 10 Torr [Les02], which can be modulated to evacuate the chamber if an inert gas purge is not desired. The process is depicted in Figure 1-14. The result of running in ALD mode is a linear growth rate with the number of cycles. The number of monolayers grown per cycle (normally <1), multiplied by the number of cycles, gives the film thickness, where one cycle includes one pulse of each of the precursor gases. Several growth cycles are

PAGE 52

43 typically necessary to deposit a monolayer of material, because steric crowding on the film surface prevents 100% surf ace coverage for each pulse. In practice, monolayer by monolayer growth is unlikely, and there may be two or more monolayer levels grow i ng together during deposition. Film growth is dependent on several factors, including reactivity of the precursors, the number of reactive sites available on the substrate/film surface, and the size of the precursor molecule (larger molecules typical l y yield lower growth rates due to steric hindrance during the reaction cycle) [Hau00]. ALD cycle time typically ranges from 0.5 to 5 seconds. A-X A-x c< ;;;::, a B-Y B-Y A-X A-X I B-Y B-Y X X I I A X A X X I I A X A I I 7 A I x-y x-y x-y x-y B B I I A B A Figure 1-14. Diagram of the ALD process. (a) Introduction of first precursor (b) Absorption of first precursor. (c) Introduction of second precursor. (d) Complete monolayer deposition of A-B. The growth rate for an ALD process should be linear with the number of growth cycles. If this is not the case the process may not be self-limiting. To check this reactant pulse time can be changed to determine if the experiment is self-limiting or if

PAGE 53

44 longer pulse times result in thicker films. ALD growth reportedly occurs by formation of islands during nucleation, hence the barrier should have thickness at least equal to that required to close the surface of the substrate [Bey02]. This closure thickness depends on the deposition conditions and the barrier materials being used. Electrical measurements (such as 4-point probe) can be used to determine if the deposited film is continuous. Films may have high resistivity until enough cycles have passed that the film is continuous; the film s resistivity will then drop substantially. This is one way to determine if the first monolayer has been completely filled in, if depositing a conductive layer on Si, for example. Extrapolating from a plot of resistivity vs. the number of deposition cycles, one can estimate the deposition cycle after which a continuous film is finally deposited. Ion scattering spectroscopy has also been used to determine when the surf ace has closed. The appeal of ALD to industry lies in its ability to manage line resistance and to provide better stress migration performance than PVD films [Pet03]. A thin ALD barrier film at the via bottom is instrumental in lowering via resistance In addition, Cu seed layers must be ultra thin, continuous and have excellent conformality to ensure consistent Cu deposition during ECD, so ALD will be essential for seed deposition. ALD is also excellent for varying the composition of the depos i ted film with thickness (composition grading) which can be important to ensure a film's compatibility with underlying layers. 1. 7 Copper Deposition Methods Electrochemical deposition (ECD), also known as electrodeposition or electroplating is the technique currently used in industry to deposit void-free bulk Cu

PAGE 54

45 layers on ICs with aggressive topographies [ITR02]. This technique has the advantages of low deposition temperature, high deposition rate, and low manufacturing cost [Chi98]. The ECO process begins with the deposition of a Cu "seed layer" on the barrier surface, because the barrier layer's resistivity is typically too high to enable uniform electrodeposition. Seed layer deposition is currently done by PVD, which is projected to be used down to the 45 nm device node [Pet03]. Once the seed layer is deposited, the sample can be immersed in the electrochemical bath and the ECO process can begin. During Cu ECO, an electrolyte containing Cu cations and sulfate anions (Cu2+ and so/-, respectively) is put into contact with the desired deposition surface. Electrons are introduced to the deposition surface (called the cathode) and reduce the Cu2+ ions in the ECO solution, causing these ions to plate out as metallic Cu. A Cu seed layer deposited by PVD (prior to the ECO process) is typically used as the cathode, on which Cu2+ ions plate out to form Cu(0). The plating reaction is: Cu2+ + 2e Cu(0) (1-14) To complete the electrical circuit, an anode is placed in the electrolyte solution. The anode introduces current to the solution, which replaces positive charges lost by the Cu plating reaction, enabling the electrolyte to remain electrically neutral [ReiOO]. Since Cu2+ ions are present in the plating solutions for ECO, backside protection of the Si wafer from the bath is essential during deposition. Failure to do this can result in massive Cu in-diffusion through the backside of the wafer, which will occur rapidly based on the diffusivity of Cu in Si (as discussed above). At the industrial level, the ECO plating tools have an o-ring around the edge of the wafer, which allows the device side of the wafer to contact the plating solution while preventing the solution from touching the backside. At

PAGE 55

46 the research lab level, waxes or glues may be used to protect the backside of the substrate. The use of special additives enables the ECD technique to fill small features from the bottom up with Cu (superfilling). To ensure deposition of a microvoid-free, uniform thickness Cu layer by ECD, the plating current (and therefore the seed layer thickness) must be essentially constant from edge to center across the wafer [Sin02a]. Hence, high uniformity of the Cu seed layer is essential. The seed layer is usually deposited on the barrier surface by sputtering, however, which is characterized by poor step coverage in small features [And99, Wan03]. A Cu seed layer with poor step coverage leads to poor Cu coverage during ECD. To overcome the conformality challenges associated with sputtering, several different approaches are being pursued. The first approach is to deposit Cu seed layers by CVD [NorOl]. Generally, Cu CVD precursors are classified into two categories: Cu(I) and Cu(II) compounds. Cu(I) compounds, such as Cu(hfac)(TMVS), where hfac is hexafluoroacetylacetonate and TMVS is trimethylvinylsilane, are typically more volatile, and can be used without a carrier gas or a reducing agent [NorOl]. These compounds tend to be liquids and are capable of deposition below 200C, but are therefore more reactive and less thermally stable, making control of the deposition rate more difficult. Copper(II) compounds, such as Cu(hfac)i, are more thermally stable, but typically require a reducing agent (such as H2 ) to remove the ligands during deposition [Kod94]. In addition, these compounds tend to be solids, requiring a deposition temperature higher than 250C [NorOl]. The major issue for Cu CVD from both Cu(I) and Cu(II) precursors is incorporation of halides and carbon from the precursor ligands into the Cu film, which can have detrimental effects on

PAGE 56

47 film properties such as adhesion and resistivity. More detail on Cu CVD precursors and deposition is available elsewhere [Kod94]. Deposition of a Cu seed layer by CVD followed by PVD and reflow of Cu has also been reported [Fri99a, Jai99]. Cu reflow typically involves the deposition of Cu by PVD methods, especially sputtering, followed by subsequent annealing of the device. The annealing process enhances surf ace diffusion of Cu, which enables Cu transport into and filling of small features [Fri99a]. The driving force for this diffusion is the chemical potential gradient associated with differences in surface curvature along the surface of the deposited Cu. To minimize the total free energy on the Cu surface, Cu atoms diffuse along the surface from convex regions to concave ones [Fri99a]. The net result is thinning of convex overhang regions at the tops of features and filling of concave parts of the features, such as via/trench sidewalls and bottoms. This diffusion-mediated process is good for movement of Cu to fill submicron feature sizes. Anneal time can be modified by changing the anneal temperature, but the anneal temperature cannot be higher than the temperature stability limits for the device, and must be low enough to prevent bulk Cu diffusion, which can lead to delamination and a change in Cu texture. Typically, Cu is annealed for 13-14 minutes at 450C during the reflow process [Fri99a]. Another approach is to deposit Cu seed layers by ALD. Several different precursors have been tested for ALD Cu seed deposition, including Cu(hfac)(TMVS), Cu(hfach, Cu(thdh, [Cu(C 3 H7)NC(CH 3 )N(C3H7)h, and CuCl, where thd is tetramethyl heptanedionate [Kod94, Lim03, NorOl]. The first four compounds are bulky, metalorganic sources, which have a very low growth rate per cycle due to steric hindrance. In addition, the first two compounds contain F, making the possibility of F

PAGE 57

48 incorporation into the deposited Cu film an issue. The last compound, CuCl, is a solid with very low vapor pressure, hence transport to the reactor is very challenging. [Huo02, Jup97, NorOl]. Electroless Cu plating, which deposits Cu without use of a sputtered Cu seed layer, may also be used [Wan03]. This technique involves deposition of Cu from an ionic solution, where the deposition surface catalyzes a redox reaction, without any external electrodes [Sha95]. A specific component from the solution (e.g., glyoxylic acid) serves as a reducing agent, and is oxidized on the catalytic surface, releasing one or more electrons [Wan03]. These electrons reduce Cu+ ions from the solution, causing the Cu to plate out as a film on the deposition surface While this technique has been used to deposit Cu seed layers with good conformality and low resistivity (1.7 .Q-cm), and has also been used to do Cu filling [Lee98a], control of the deposition rate is not as precise as sputtering. More detail on this deposition technique is available elsewhere [Sha95, Sha97]. The last approach is seedless ECD [ITR02, Sin02a], where Cu is directly electroplated onto the underlying barrier material using an external electrode, but without any seed layer. This technique is the most technologically and economically promising, because it eliminates the seed layer deposition step from the current IC metallization process and does not require new equipment expenditures. Selection of a diffusion barrier material that enables seedless Cu electrodeposition could result in significant cost savings to the metallization process.

PAGE 58

49 1.8 Statement of Problem The three main challenges of shifting to Cu interconnects on IC devices are: a) Cu deposition technique b) patterning method for the Cu layer, and c) which barrier material and deposition method will prevent Cu-Si interdiffusion while fostering adhesion to neighboring layers [And99]. The first challenge has been addressed by use of ECD. Since Cu is resistant to chemical etching, the application of standard reactive ion etching (RIB) techniques used to pattern Al is not feasible for Cu. The second challenge, patterning of Cu interconnects, has therefore been addressed by the dual-damascene deposition process. In this process, Cu is deposited across a pre-patterned wafer by one of the aforementioned deposition techniques. Chemical mechanical polishing (CMP) uses a grinding pad and slurry to remove excess Cu, leaving only the desired spots (e g., trenches and vias) filled with metal. Since the dual-damascene process deposits Cu in both trenches and vias, the need for tungsten (W) plugs at the upper levels of the IC, which are used to fill vias on Al based devices, is eliminated. Removing this W plug lowers the electrical resistance of the device, both because Cu has a lower resistivity than W and because the Cu-W interface is eliminated. The third challenge, barrier material and deposition technique, is an ongoing issue as device features continue to shrink. TiffiN barriers, deposited by sputtering, are still used at the contact level with W plugs, where TiN protects the contact from F in the WF6 precursor and improves W adhesion [ITR02, Kal00]. Ti reacts with Cu to form cuprides, making Ti thermodynamically unstable as a diffusion barrier for Cu [RamOO]. Moreover, if TiN films are Ti rich, the excess Ti can react with Cu and lead to failure of the TiN

PAGE 59

50 barrier. In addition, TiN barriers with thickness below 20 nm reportedly fail to prevent Cu diffusion due to grain boundary diffusion [Kal00]. A Ta/faN dual layer barrier structure is currently used as the Cu diffusion barrier at the intermediate and upper wiring levels on IC devices. Cu has good adhesion to Ta, which encourages deposition of low-resistivity Cu (111) on its surface [Pet03], while TaN has good adhesion to the dielectric layer [ITR02, Sin02a] The Ta/faN dual layer barriers are deposited by modified sputtering methods [ITR02, Sin02a], however, which have a finite lifetime due to aforementioned step shadowing issues. Barrier deposition takes place on the industrial scale in a cluster tool, which contains several chambers connected by a common vacuum transfer chamber. The barrier is deposited in one chamber, and the wafer is then moved in-vacuo to a separate Cu seed layer deposition chamber. Once the Cu seed layer is deposited, the wafer is pulled out of the cluster tool and put into a separate system for Cu ECD. The "holy grail" of diffusion barrier research is a robust material / deposition method couple that meets all of the above listed barrier requirements and is extendable to future device generations. In addition, a material/deposition pair which would enable Cu metallization to extend from the upper levels of the device down to the contact level would eliminate the need for Al metallization at the local level and the use of W contact plugs, yielding greater device speed. The interconnect structure's features (trenches and vias) become increasingly more aggressive (narrower diameter and higher aspect ratio) as the wiring levels get closer to the contact level. Research should therefore focus on meeting coverage requirements at device level feature dimensions to enable future extension of Cu throughout the IC.

PAGE 60

51 1.9 Hypothesis It is evident that a variant of CVD will be required to meet future demands for conformal diffusion barrier deposition on aggressive IC device topographies. New precursors will be required to meet the increasingly stringent demands on film properties. In an effort to avoid halide incorporation, synthesis and testing of novel WNx and WNxCy MOCVD precursors were pursued in this work, for reasons that will become clear in Chapter 2. Depositing WNx and WNxCy films from novel precursors by MOCVD should minimize halide content, have low temperature deposition, enable good adhesion to neighboring layers, and enable a single-step CMP process. In addition, these materials appear to be useful for direct, seedless electroplating, and have been reported to resist Cu diffusion. WNx is also reported to deposit in amorphous form more easily than TaNx [RamOO], and the addition of C to form WNxCy should enhance the ability to deposit amorphous films even further. Eventually, a shift from CVD to ALD will be necessary to meet conformality requirements in small device features. Hence, our strategy is to initially pursue WNx and WNxCy films by MOCVD and to eventually transition to precursor development and testing for ALD. The film growth and procedure and characterization techniques used to analyze the films will be discussed in Chapter 4. Data, analysis and conclusions for each precursor will be presented in a separate chapter, while the thermodynamic analysis for the W-N-C-H-Cl system is presented in Chapter 3. Cu diffusion results for barriers

PAGE 61

52 deposited with an isopropylimido-based W complex will be discussed in Chapter 8, and future work will be discussed in Chapter 9.

PAGE 62

CHAPTER2 REVIEW OF THE LITERATURE From a thermodynamic standpoint, a barrier material must be chosen so that the chemical potential (the thermodynamic description of reactivity) at each interface ensures little or no reaction. A major thermochemical difference between Al and Cu is their reactivity with silicon. Al can exist in equilibrium with Si, while Cu cannot, readily reacting to form silicides [Ram00]. This non-equilibrium suggests that the thermodynamic driving force for diffusion in the Cu-Si system is larger than that in the Al-Si one. An ideal barrier separating Cu from Si must therefore be stable to and non-reactive with both Cu and Si. Moreover, the barrier should be stable with other potential materials in the IC, such as low-k and high-k dielectrics. Many materials have been examined to determine their potential as diffusion barriers for Cu metallization schemes. A comprehensive list of various materials tested as diffusion barriers can be found in previous review articles [Jai99, KalOO, Kim03b]. 2.1 Cu-Si Interconnects without a Barrier Ideally, the easiest route to IC device fabrication would be to deposit Cu directly onto any neighboring layers. Hymes et al. [1998] formed copper silicide compounds by sputtering Cu onto thin films of Si, with the goal of using these compounds as passivation layers to prevent further Cu-Si interdiffusion and reaction. At room temperature, Cu3Si and CusSi formed, while all Cu3Si decomposed to Cu5Si after annealing at only 300C. The Cu5Si compound disappeared after annealing at 500C, leaving only pure Cu and 53

PAGE 63

54 dissolving the Si into the Cu layer. The inability to form a thermally stable Cu-Si passivation layer highlights the instability of the Cu-Si interface and underscores the need for a diffusion barrier to separate these materials 2.2 Refractory Metal-Based Barriers The term refractory refers to a material with a melting point above 1800C [Pie96]. Due to their low self-diffusion coefficients, refractory meta l-based materials have been frequently studied for use as diffusion barriers [Liu00], with current refractory metals of interest for diffusion barrier applications including Ti Ta, W and Ru. 2.2.1 Unary Refractory Metal Barriers Unary refractory metals have been explored as potential barrier materials, due to their low self-diffusivity and low electrical resistivity. Some drawbacks associated with these materials include their tendency to crystallize (enabling grain boundary diffus i on) and to react with Si at temperatures greater than 450C [JohOOa]. The applicability of several unary refractory metals as barriers will be discussed in the following sections. 2.2.1.1 Ti Barriers While Group V and VI transition metals (e.g., Ta and W) are s t able and non-reactive with Cu, Group IV transition metals such as Ti are unstable forming several cuprides, including Cu 4 Ti, Cu4Th, CuTi and CuTii [RamOO]. For this reason, Ti is not a viable choice for a Cu diffusion barrier as its reaction with Cu would degrade the barrier structure and lead to rapid device failure. 2 2.1.2 Ta Barriers Tantalum (Ta) does not react with Cu [Wan94], and has demonstrated good oxidation resistance and enhanced Cu (111) texturing during Cu ECD [Chi02]. Ta

PAGE 64

55 generally fails to suppress Cu diffusion, however, because grain boundaries readily form during deposition and subsequent thermal cycling of the Ta layer. In addition, Ta reacts with Si to form silicides, making the Ta-Si interface unstable, and disqualifying Ta metal as a potential single barrier layer material to separate Cu and Si. 2.2.1.3 W Barriers Tungsten (W) is another transition metal that does not react with Cu [Ram00, Wan94]. It does, however, react with Si [Ram00]. W reportedly failed as a Cu barrier due both to a silicidation reaction above 670C [Tak:97b] and crystallization, which causes grain boundary formation [Cha97c]. Sputtered W barriers have also been reported with columnar grains, which extended across the entire barrier thickness and caused barrier failure above 700C [Mer97]. While PECVD W has been reported with very low resistivity (10 .Q-cm) [Cha97b], films deposited at 350C had an open grain boundary structure, which led to a low activation energy (0.46 eV) for Cu diffusion [Gup95]. Metallic W has also been deposited by ALO, using WF6 and Sizllt, precursors over a temperature range of 30 to 350C [Ela0l]. Below 100C, some Si surface species were left on the surface and remained unconsumed by the WF6 precursor exposure. 2.2.1.4 Ru Barriers Ru, like Ta and W, is non-reactive with Cu [Chy03], and has been tested as a Cu diffusion barrier material. The tested Ru barriers were deposited by PVD methods, however [Jos03], which have a limited lifespan, as mentioned in Chapter 1. In addition, Ru silicides are reported to form after annealing at 500C [Jel03]. A sputtered Ru layer sandwiched between two TiN layers reportedly did not enhance the barrier's ability to prevent Cu diffusion [Kim03c]. Ru deposited by CVD has been studied for use as the

PAGE 65

56 metal electrode for DRAM capacitors [Aoy99, LeeOOa, Mat02], and Ru has also been deposited by ALD [Aal03]; reports of CVD and ALD Ru as Cu diffusion barriers, however, have not been given. 2.2.2 Binary Refractory Metal-Based Barriers To overcome the limitations associated with unary refractory meta l barriers, binary compounds containing Ti, Ta and W have been widely studied. The addition of a second, nonmetal element to the refractory metal tends to increase the likelihood of depositing amorphous films, because the nonmetal element disrupts the crystallization process. The non-metal also generally increases film resistivity however. The properties and applicability of several refractory metal carbides and nitrides will be discussed in the following sections. 2.2.2.1 Refractory Metal Nitrides (M-N) Adding nitrogen to refractory metals gives rise to refractory me tal nitrides where the term nitride refers to compounds formed between nitrogen and other elements with equal or lower electronegativity [Pie96]. Three characteristics play a role in the formation of metal nitrides: the size ratio of nitrogen to the other element the electronegativity difference between nitrogen and the meta l and the electronic bonding characteristics between nitrogen and the metal [Pie96] Below a N/metal radius r atio of 0,59 interstitial nitrides form while above 0.59, covalent nitrides (such as Si]N4 ) typically form [Pie96]. The early transition metals (Groups IV, V and VI) have a large enough host lattice to enable formation of inte r stitial nitrides whe r e the n i trogen atoms sit on interstitial sites in the metal lattice. The atomic radius of N is 0.74 A while that for W, for example, is 1.394 A yielding a N/W radius ratio of 0 .53. The n i trogen and metal

PAGE 66

57 atoms have a large electronegativity difference, so the N atoms "nest" in the interstitial sites of the metal lattice [Pie96]. Interstitial nitrides have bonding with a combination of metallic, covalent and ionic character, high hardness and strength, and high thermal and electrical conductivity. These nitrides tolerate nonmetal vacancies, making them susceptible to the presence of interstitial impurities such as oxygen. Most early transition metals have a BCC structure, which cannot accommodate significant nitrogen levels at the interstices. To form an interstitial nitride, the metals switch to a close-packed structure, such as FCC or HCP, to ensure adequately sized interstitial sites that can accommodate N. Close-packed FCC transition metals have both tetrahedral and octahedral interstitial sites, but since the tetrahedral sites are too small to accommodate the N atoms, N only occupies the octahedral sites [Pie96]. Nitrogen addition suppresses crystallization in the films, and helps to repel the advance of Cu through the grain boundaries, due to a repulsive interaction between Cu and N [EksOl, Tak97a]. Excess nitrogen in these films migrates from the bulk polycrystals to the grain boundaries, where repulsive interactions between Cu and nitrogen "stuff' the boundary and halt Cu diffusion [EksOl, Sha89]. Formation of copper nitride (Cu3N) by reactive sputtering has been reported, but this compound is unstable, decomposing to Cu and N under vacuum at 100C [Liu98, Mar95]. Refractory metal nitrides tend to have high hardness, good chemical stability and high conductivity [Lev98, Nag93]. Three nitrides of current interest are FCC titanium nitride (TiN), FCC tantalum nitride (TaN), and FCC tungsten nitride (WNx)While FCC TiN and TaN have melting points of 2950 and 3093C, respectively [Pie96], FCC WNx does not melt, but instead decomposes to BCC Wand N2 gas under vacuum at elevated

PAGE 67

58 temperature (850C) [Suh0l]. Since this decomposition temperature is substantially higher than the typically processing window for ICs, FCC WNx is a viable diffusion barrier candidate material. The barrier properties of TiN, TaN and WNx are discussed in more detail below. 2.2.2.1.1 TiN Titanium nitride (TiN) was widely used as a diffusion barrier for the Al-Si system, and its reported resistivity ranges from 20 to 2000 .Q-cm [Pai96]. PVD has dominated TiN barrier deposition, resulting in non-uniform coverage of high aspect ratio structures and a columnar grain structure [Gal99, KalOO]. TiN performs poorly with Cu metallization due to this columnar microstructure [Nic95], which enables rapid Cu diffusion through the barrier [Gal99]. Plasma treatment and exposure to Sil-Li, however, have been reported to improve TiN's resistance to Cu diffusion [Jos02]. Another difficulty involves the reactivity of Ti rich TiN films with Si and Cu. For TiN films with a Ti/N ratio> 1 (i.e., film is unsaturated with nitrogen), the excess Ti will react with S i or Cu, leaving the barrier permeable to Cu and the device vulnerable to rapid failure [Kim92, Lee94a, Ram00]. To avoid the problems associated with PVD TiN, several different CVD techniques have been tested. Liu et al. [2000] used LPCVD of TiC14 + NH3 at high deposition temperature (630C), but the resulting films had a columnar grain structure and suffered from halide incorporation. MOCVD TiN has shown only 40-50% conformality in small contact structures [Fal98], has high impurity incorporation, and requires an additional plasma-processing step to stabilize the films in low resistivity form [Gal99]. MOCVD TiN films deposited below 450C were very porous, with low

PAGE 68

59 density and high resistivity, and had poor resistance to Cu diffusion [Par95b]. Capping the film with an ultra-thin Si3N4 layer has also enhanced performance of porous TiN films deposited by MOCVD of tetrakisdimethylamido tantalum (TDMAT). This thin capping layer improves barrier capability without significantly increasing the film resistivity, but adds another step to the barrier deposition process [Lu98a]. While oxygen stuffing of the TiN barrier (due to air exposure) helps it to resist Al in-diffusion (due to aluminum oxide formation), it does not prevent Cu in-diffusion, due to the lack of a stable, passivating Cu oxide layer [Par95a]. In addit ion, TiN was not suitable for Cu electroless plating, as its redox potential was higher than that of the Cu electrode, preventing the initial displacement reaction from occurring at the TiN surface [Wan03]. 2.2.2.1.2 TaN TaN is a proven Cu diffusion barrier, and is reportedly stable against Cu-Si interdiffusion up to 720C [Wan94]. Copper with a PVD Ta/TaN dual layer barrier is currently used as the interconnect scheme for intermediate and upper level wiring on IC devices. The dual layer barrier is required to overcome adhesion difficulties: Cu adheres well to Ta, but not to TaN, while SiO2 and most dielectrics adhere well to TaN, but not Ta. [ITR02, Sin02a]. In addition, the TaN layer promotes deposition of the low resistivity (20-30 .Q-cm) a-Ta phase, whereas the high resistivity (180 .Q-cm) ~-Ta phase forms when Ta deposits directly on SiO2 [Tra03]. But, since PVD TaN has much higher resistivity (-200 to 250 .Q-cm [Sun98a, Tra03]) than Ta, the TaN layer thickness should be held to a minimum. A 7 A thick TaN film was found to foster a-Ta deposition and to prevent Cu interdiffusion [Tra03]. The lifespan of this bilayer deposition

PAGE 69

60 technique is limited, however, due to the inability of PVD processes to deposit conformal films at ever shrinking device dimensions. LPCVD routes to TaN deposition have been examined as well, as this deposition method provides superior conformality to sputtering techniques, which are projected to reach their usability limit at the 65 nm node in 2007 [Han03]. CVD of TaN was reported by reacting halide precursors, including TaBr5 TaF5 and TaC15 with NH3 [Kal99, HilOO]. Resistivities ranged from 395 to 5000 .Q-cm depending on the precursor used, with the lowest value corresponding to TaC15 [Kal99, Che99b, HilOO]. Halide incorporation, however, ranged from 0.5 to 4.5 at. %, with a value of 1.0 at. % for the TaC15 precursor [Kal99 HilOO, Che99b], While plasma assisted CVD reduced the resistivity of films deposited by TaBrs + NH3 to 150 .Q-cm, an increase in Br content up to 3 at. % occurred [Che99a]. Attempts to grow TaN by MOCVD resulted in films with high resistivity and some carbon contamination. When pentakis(dimethylamido)tantalum (PDMATa) was used for MOCVD of TaN, the resulting films contained the insulating Ta3N5 phase [Fix93] Pentakis(diethylamido)tantalum (PDEATa) was also tested, but the films also had very high resistivity (7000-60000 .Q-cm) [Cho98, Cho99]. MOCVD using tert-butylimido-tris(diethylamido)tantalum (TDEATa) has also been tested, resulting i n good step coverage (-100%) but high deposition temperature (up to 650C) and resistivity (900-2000 .Q-cm) [Tsa95]. Low film density and grain boundary formation were the typical causes of barrier failure in MOCVD TaN [Kal00]. One study of an ALD TaN/PVD Ta barrier has also been given [Ho04]. The precursors used in this study were not disclosed, but the barrier scheme was shown to

PAGE 70

61 work with a 65 run Cu back-end-of-line (BEOL) interconnect structure. The ALD TaN barrier had good yield and reliability, and a 16% reduction in RC time delay over a PVD TaN/PVD Ta barrier scheme was demonstrated. 2.2.2.1.3 WNx As mentioned above, N has been added to W to deposit WNx films. These films show promise as thermodynamically stable, stuffed barriers for separation of Cu and Si and have been demonstrated as excellent glue layers for W/Si and W/SiO2 interfaces [Gal97, Kim92]. Many deposition techniques, including ion implantation, sputtering, LPCVD, PECVD, MOCVD and ALD have been used to deposit WNx films. A detailed discussion of WNx films deposited by these techniques will be deferred until Section 2.7 below, as we have chosen to focus on this material for our work. 2.2.2.2 Refractory Metal Carbides (M-C) Unlike the nitrides, refractory metal carbides from all three Groups (IV, V and VI) tend to be hard, wear-resistant materials with high melting points and good chemical resistance [Pie96]. The Group IV, V and VI refractory metals form interstitial carbides, with crystal structures similar to those for the nitrides. The electrical resistivity of carbides is typically lower than that for nitrides with equivalent crystal structure, due to weaker bonding between the metal and C relative to bonding between metal and N [Nak87]. While these properties make the carbides interesting diffusion barrier candidates, C, unlike N, does not exhibit a repulsive interaction with Cu, making the carbides less resistant to Cu diffusion. The barrier properties of TiC, TaC and WCx are discussed below.

PAGE 71

62 2.2.2.2.1 TiC One report of TiC (deposited by sputtering) as a Cu diffusion barrier has been given [Wan0la]. While the TiC resisted metallurgical failure at temperatures up to 650C, the same films suffered electrical failure after annealing at temperatures just above 500C. 2.2.2.2.2 TaC Sputter deposited TaC films have also been tested as Cu diffusion barriers [Ima97, Lau02, Mor98]. While these films had low resistivity (210 .Q-cm for 100 nm thick film), carbon-rich TaC films contained low-density carbon regions surrounding TaC grains [Ima97]. These low-density regions are facile paths for Cu penetration through the barrier, with 7 nm TaC films failing to prevent Cu diffusion above 550C [Lau02]. This indicates that C, unlike N, lacks the ability to chemically repel Cu diffusion through the film, and is reflected in TaC's lower activation energy for Cu diffusion as compared to W2N or TaN [Mor98]. 2.2.2.2.3 WCx The bulk resistivity of ~-WCx is slightly lower than that for ~-WNx, with reported values of 41 .Q-cm and -50 .Q-cm, respectively [Lee93, Nic78] MOCVD WCx from W(CO)6 + C2Hi has been tested as a Cu diffusion barrier [Sun0lb, VelO0]. Deposition temperatures ranged from 250 to 510C, with 50 nm thick films deposited at 290C having resistivity of 250 .Q-cm. A 70 nm thick WCx film annealed for 8 hours at 400C resisted Cu diffusion. Sputtered WCx with resistivities ranging from 200 to 1000 .Q-cm were also tested as Cu diffusion barriers, with the incorporation of C resulting in smaller grain sizes relative to pure W films [VelO0, Wan0lb]. The onset of silicide

PAGE 72

63 formation above 700C, however, indicated the instability of the WCx/Si interface and also compromised barrier integrity against Cu diffusion. Kim [2003a] reported deposition of tungsten carbide (WCx) films by plasma assisted ALD (PAALD) using the bis(tert-butylimido)bis(dimethylamido)tungsten [CBuN)z(Me2N)2]W precursor w i th H 2 and N2 carrier gas. The films were deposited on SiOi/Si substrates at 250C, and had growth rates ranging from 0.4 to 0.7 Alcycle, 100 % conformality in 0.15 m features with a 15:1 aspect ratio, and resistivities ranging from 295 to 22000 Q-cm. The films contained 63 at.% W, 28 to 42 at.% C, 2 to 7 at.% N, and 1 to 6 at.% 0. Cu barrier integrity tests were not reported, however. 2.2.3 Ternary Refractory Metal-Based Barriers The addition of a third element to a binary refractory metal nitride matrix tends to further disrupt the microstructure, increasing the likelihood of nanocrystalline or amorphous phase formation [lst00, Jai99, Ram00]. C Si and B are the most examined third elements added to metal nitride films. 2.2.3.1 Refractory Metal Carbonitrides (M-C-N) The addition of C to refractory nitrides has two general purposes: to increase the likelihood of amorphous film deposition, and to decrease film resistivity relative to the binary nitride. The C and N intermix on the interstitial sublattice of the host metal, and lattice structures are similar to those for the nitrides and carbides. The advantages of adding C to the N on the non-metal sublattice are tempered by the decreased ability of C (relative to N) to chemically repel Cu diffusion. The composition and microstructure of the films must therefore be carefully controlled to deposit low resistivity films with high stability against Cu diffusion. Free carbon (C not bound to W), in particular, should be

PAGE 73

64 minimized, as it has minimal impact on Cu diffusion and detrimentally affects film resistivity. 2.2.3.1.1 TiCN One report of a TiCN Cu barrier, deposited by MOCVD of tetrakis(dimethylamino)titanium (TDMAT), has been given [Eiz94a]. While the films resisted Cu in-diffusion after annealing at 600C, they had high resistivities, ranging from 3000 to 20000 .Q-cm for 200 A films. 2.2.3.1.2 TaCxNy films deposited by sputtering of a TaC target in an Ar/N2 atmosphere were also tested as Cu diffusion barriers [Sun0la]. TaCxNy was found to have superior thermal stability to the respective binary phases, and its resistance to Cu diffusion was greater than TaC due to stuffing of the grain boundaries with N. The films had relatively low resistivity (-300 .Q-cm), and prevented Cu diffusion after a 30 min anneal at 600C. Jun [1996] used pentakis(diethylamido)tantalum to deposit TaCxNy films by MOCVD, but resistivity was high (;?: 6000 .Q-cm) and the barrier failed after annealing for 1 hour at 500C. MOCVD of TaCxNy films was also done using a mixture of pentakis(dimethylamido)tantalum and pentakis(diethylamido)tantalum [HosOO]. These films prevented Cu diffusion after a 30 min anneal at 500C, but had high resistivity (;?: 4000 .Q-cm). 2.2.3.1.3 The bulk resistivity of P-WCx is somewhat lower than P-WNx, and as expected, the resistivity of P-WNxCy is reportedly lower than that for P-WNx. Carbon and nitrogen in these films can deposit at interstitial lattice sites (i.e., bind to W) or as free C

PAGE 74

65 or N. Given a choice, free N is more desirable, because although it increases film resistivity (as does free C), it also repels Cu diffusion. Free C also increases resistivity, but is less effective at preventing Cu diffusion. The difficulty, then, is to minimize or eliminate deposition of free C, while encouraging some N to deposit interstitially and allowing some to deposit as free N. This should give minimum film resistivity and maximum Cu resistance. WNxCy films were deposited by non-reactive sputtering of W-C and W-N targets, with the resulting films being amorphous, but maintaining some short-range order [Vie02]. These films were not tested for Cu barrier reliability, however. MOCVD was used to deposit WNxCy as a barrier for Cu metallization in a patent application [FukOO]. Amorphous WNxCy films were deposited by reacting a gas containing W, such as WF6 W[N(CH3)]6, or W[N(C2H5)]6, with a hydrocarbon gas, such as C~, and a nitrogen supply source, such as a nitrogen plasma, at 360C. X-ray diffraction indicated a peak between 36 and 38 20 and a second position between 42 and 44 20, both of which are indicative of FCC WNxCy deposition. While resistivities down to 275 .Q-cm were reported, details of Cu barrier testing were not included. Several studies have reported ALD deposition of WNxCy. The deposition used sequential reactions of WF6 NH3 and Et3B [Ele03, Kim03d, Li02, Li03, Pet02, Smi02]. The lowest reported resistivity was 210 .Q-cm for a film deposited at 350C [Ele03, Li02]. One group reported that a 120 A thick WNxCy film was stable for a 30 min anneal up to 700C [Kim03d]. These ternary films had high density and excellent adhesion to copper, but contained some amount of F (0.5-1 at.%) [Ele03, Li03].

PAGE 75

66 2.2.3.2 Other Ternary Refractory Metal Compounds Several ternary silicide systems, including TiSixNy [Bai96, Chi0l, EisOOb, Jos02], TaSixNy [Lee99, Som97], and WSixNy [Bla97, HirOl], have been examined for diffusion barrier applications. While the addition of Si promotes amorphous film growth and improves Cu adhesion [HarOl], it also increases film resistivity and can decrease failure temperature [Kal00]. The ratio of metal/Si must be greater than 1.67 to ensure stability with Cu [Nic95]. The failure mechanism in these films tends to be grain boundary diffusion after barrier crystallization [Nic95], and the stability of these films is questionable due to the potential for Si out-diffusion into and reaction with the neighboring Cu layer. These barriers are also reported to have poor adhesion to low-k dielectrics [Pet03], and the majority of research on ternary silicide barriers has relied on PVD methods, which will have limited use in the future. 2.2.3.2.1 TiSiN TiSiN films were deposited by reactively sputtering a Ta-Si target in an Ar/N 2 gas mixture [Iij95, Rei94]. Reid et al. [1994] deposited amorphous films with a stoichiometry of Tio_34Sio.23No.4 3 and resistivity of 680 .Q-cm. These films resisted Cu diffusion after a 30 min anneal at 650C. Iijima et al. [1995] deposited amorphous films with a stoichiometry of Tio.31Sio.19No.so and resistivity of 500 .Q-cm. These films resisted Cu diffusion after a 30 min anneal at 600C. LPCVD of TiC14+Si~+NH3 +Hi/Ar at 500C was also reported [Bla97]. As-deposited films were microcrystalline, with a composition of Tio.41SiomNo.46, Chlorine and oxygen contamination levels ranged from 3 to 5 at.%. This film prevented Cu diffusion after a 600C anneal for 1 minute

PAGE 76

67 MOCVD of TiSiN films has been reported with multiple precursor schemes, including TDMAT +Si~+ NH3 [Chi0l, Jos02] and TDEAT +Si~+ NH3 [Bai96]. A 50 A thick TiSiN film had good wettability to Cu and low resistivity (350 .Q-cm) [Chi0l]. The Si was suggested to improve Cu adhesion when thin oxidized l ayers are present on the barrier, and also to reduce Cu agglomeration [Pet03]. TiSiN films deposited by a combination of MOCVD and plasma assisted CVD, however, allowed Cu to diffuse through after annealing at 500C [Jos02]. Amorphous TiSiN films deposited at 400C contained -3 at.% C and 0, and had high resistivity (>2000 .Q-cm) [Bai96]. Metalorganic ALD of TiSiN films was reported using TDMAT, NH3 and Si~ at 180C [MinOO]. The film contained small crystallites in an amorphous phase, and film stoichiometry was Ti0 .32Si 0.18N 0 .50, but barrier film resistivity and Cu testing were not reported. 2.2.3.2.2 TaSiN Several reports of TaSiN thin films as Cu diffusion barriers have been given. TaSiN films were deposited by reactive sputtering of a Ta-Si t arget in N2 [HarOl] and Ar/N2 [Kim97a, Kol91, Lee99] gases. This ternary layer had better barrier performance and better adhesion to Cu than a Ta/TaN dual layer barrier, but details of the barrier s performance and electrical properties were not given. Lee et al. [ 1 999] deposited films with Tao _43Sio.Q4No.s3 stoichiometry, with a failure temperature of 825C (decomposing to Cu3Si +TaNx + TiSii above this temperature) but the film had high resistivity (1419 .Q-cm). Kolawa et al. [1991] deposited 100 nm thick amorphous films with a stoichiometry of Tao 36Sio.1~0 .soThe films had high thermal stability, prevent ing interdiffusion between neighboring Cu and TiSii layers up to 900 C, had O content

PAGE 77

68 below 3 at.%, and had resistivity of 625 .Q-cm. Kim et al. [1997a] varied the N content of the TaSiN films to determine the impact on barrier performance. Films with N content greater than 40 at. % resisted Cu diffusion after an 800C anneal, while those with lower N levels failed after a 700C anneal. Information on film resistivity was not provided, however. LPCVD of TaC15+SiHi+NH3+H2/ Ar at 500C was also used to deposit TaSiN [Bla97]. As-deposited films were microcrystalline, with a stoichiometry of Tao.35Sio.11N0 54, and contained Cl and O levels ranging from 3 to 5 at.%. The films failed to prevent Cu diffusion after a 1 minute, 600C anneal, however. 2.2.3.2.3 Several reports of WSixNy thin films as Cu diffusion barriers have been given. Reactive sputtering of a W-Si target in N 2/Ar gas was used to deposit amorphous WSi0 6N films [Min96, Shi97], which crystallized after annealing at 850C [Shi97]. Film resistivity was 430-450 .Q-cm, and the WSio.~ film blocked Cu diffusion after a 30 min anneal at temperatures up to 600C. WSixNy films were also deposited by N2 plasma nitridation of sputtered WSix thin films [Hir98, Hir99, HirOl]. As-deposited films were amorphous, and their effectiveness decreased with increasing film crystallinity. A WSiN(6 nm)/WSix(14 nm) bilayer resisted Cu diffusion after annealing for 1 hour at 400C [Hir99]. Electrical properties of the barrier film were not reported, however. WSixNy thin films were inadvertently formed by annealing a W/WNx/poly-Si structure at 800 C [Nak97]. The resulting W/WSixN/poly-Si structure was stable up to 950C, above which silicidation of the W

PAGE 78

69 layer occurred. The ternary film was not tested as a Cu diffusion barrier, however, and bulk resistivity of this film was not reported. WSixNy films were also deposited by LPCVD of WCl6+Si!LJ+NH3+Hi/Ar at 500C [Bla97]. As deposited films had a stoichiometry of Wo.54Sio.12No.34, contamination levels of Cl and O ranging from 3 to 5 at.%, and high resistivity (1000 .Q-cm). These layers prevented Cu diffusion after a 600C anneal for 1 minute, and crystallized at 900c. Amorphous WSixNy thin films were also deposited by PECVD of WF6+ N2 + H2 + Si!LJ at 380 to 400C [Eck03]. While film resistivity was low (308 .Q-cm), layer uniformity and thermal stability was poor, as films crystallized after a 1 hour anneal at 600C. Amorphous WSixNy thin films were also deposited by PECVD of WF6+Sii~+NH3 at 350C [Gok99] Films with compositions of W1.21Si0.1~ and W o.%S io. ~ were deposited, which crystallized after annealing at 800C for 30 min Film resistivities below 200 .Q-cm were reported after annealing at 450C, but film performance with Cu was not reported. 2.2.3.2.4 Several reports of WBxNy thin films as Cu diffusion barriers have been given. WBxNy films with a variety of stoichiometries were deposited by reactive sputtering of a W-B target in Ni/Ar gas [Rei95] A Wo. ~o.2oNo.16 film showed the best resistance to Cu diffusion, preventing it after a 30 min anneal at 800C. This stoichiometry also had a low resistivity of 220 .Q-cm. WBxNy films were also deposited by reactive sputtering of W and W2Bs targets in N2/Ar gas, with substrate temperature ranging from 25 to 500C [Lee0lb, ParOO]. Film stoichiometries ranged from W o .wBo.osNo.os to

PAGE 79

70 Wo.s1Bo.osNo.3s, with resistivities as low as 140 .Q-cm obtained for the stoichiometry with the lowest N content. As-deposited films were amorphous, crystalliz i ng after a 700C anneal [ParOO], and resisted Cu diffusion after a 30 min anneal at 800C [Lee0lb] WBxNy films deposited by reactive sputtering of a W 2 B5 target in Nz/ Ar gas at 35C were used to form Schottky contacts to GaAs [Kim98b]. WBxNy films were formed by ion implantation of BF2 + into WNx thin films, which were deposited by PECVD of WF6 + NH3 + H 2 on Si (100) at 300C [Kim97b, Kim99a]. The ternary film was amorphous, and resisted N out-diffusion after annealing at 800C. As-deposited ternary bulk film resistivity was 200 .Q-cm, which was slightly higher than the value for the binary WNx film, and the ternary film resisted Cu diffusion after annealing at temperatures up to 750 C. Deposition of WBxNy thin films by PECVD of WF6 + NH3 + B1oH14 + H2 on Si (100) at 350C has also been reported [Kim97b Kim98a, Kim02a]. As-deposited film stoichiometries ranged from Wo.9oBo.osNo.os to Wo.JsBo.42No.20, depending on the B1oH1 J NH3 ratio, with the optimal film stoichiometry being Wo.~o.2sNo.29. Resistiv ity ranged from 100 to 844 .Q-cm, with a value -700 .Q-cm for a 2000 A thick film with optimal stoichiometry. Films with optimal stoichiometry remained amorphous even after annealing at 800C, and resisted Cu-Si interdiffusion up to this temperature. Tungsten rich as-deposited films were polycrystalline, and B atoms reportedly out-diffused even faster during annealing that did the N atoms. Both the Wo.wBo.osNo.os and Wo.soBo.1sNo.os stoichiometries failed to resist Cu diffusion after annealing above 600C

PAGE 80

71 2.3 Justification for WNx (and WNxCy) as the Barrier Material The microelectronics industry already uses Ti, Ta and W in one form or another (such as TiN barriers, TatraN barriers and Ta205 capacitors, and W plugs) for memory and processor devices [Jai99] TiN is used as a barrier for Al based metallization but TiN is unsuitable for Cu metallization, as mentioned earlier. Cu diffusion barrier research has focused on TaN and WNx, with WNx appearing to have technological advantages over TaN as a Cu diffusion barrier. WNx is known to be an effective diffusion barrier against copper penetration at temperatures up to 750 C [Pok91], and adhesion of CVD Cu to WNx is stronger than that to TaN [lva99]. In addition, WNx outperforms TaN as a liner material for seedless electrochemical deposition (BCD) as ECO Cu shows stronger adhesion to WNx film layers [ITR02, ShaOl, Sin02a]. A suitable liner material would enable seedless ECO of Cu Elimination of the Cu seed layer, typically deposited by sputtering would increase throughput by removing a process step as well as remove the complications associated with sputtering in ever-shrinking device features. The combination of these steps would result in cost savings per wafer, higher wafer throughput, and higher quality devices for a WNx based system as compared to its TaN counterpart [Gal99]. In addition to enable electroless plating of Cu onto the barrier materials, TaN barrier layers require a pre-deposition HF etch while WN x barriers do not. Tungsten oxide, present on the WNx surface due to air exposure, is readily dissolved in the electroless plating solution hence the pre-deposition etch step is unnecessary [Wan03]. There are also processing advantages for WNx, During chemical mechanical polishing (CMP) TaN is removed at a rate 18 times slower than Cu, while WNx is

PAGE 81

72 removed 1.5 times faster [Gal99]. One report gives WNx removal rates of -100 Nm.in for CMP pressures below 0.5 psi, while comparable removal rates for TaN require polishing pressures close to 2 psi [Tak:02]. The slow removal rate of TaN means that the pad must be in contact with the Cuff aN surface for an extended period of time to planarize the wafer. This extended contact time leads to dishing of the copper, where excessive copper is removed from the device. Dishing of both the Cu and dielectric layers impacts line resistance and planarity of the device [Jai99]. A time-consuming two-step CMP process, which includes a slurry and pad changeout, is required to minimize copper dishing during planarization of TaN. From the above discussion, it is evident that WNx films are the most promising candidates for diffusion barrier materials. The resistivities for these films are reasonably low, and WNx has shown good adhesion to Cu and other potential neighboring materials. CVD and ALD routes to low-resistivity WNx have been established, although the precursors must be optimized to minimize halide contamination and to control carbon levels. Recent studies have examined WNxCy for diffusion barrier applications, due to its ability to deposit in amorphous form, its lower resistivity relative to WNx, and its excellent adhesion to Cu. A review of the properties of WNx, along with a brief discussion of WNxCy films, follows. 2.4 WNx Film Properties The four commonly seen phases for tungsten nitride are BCC WNx, FCC ~-WNx, HCP WNx, and SHP-WN [Gui93, Nak87, Wri89]. The structure and properties of these phases are discussed in more detail in Chapter 3. FCC ~-WNx is the desired phase for Cu diffusion barriers because it has the lowest bulk resistivity (-50 .Q-cm) of the

PAGE 82

73 various tungsten nitride phases [Lee93]. The structure of FCC WNx is NaCl type, with W atoms at FCC sites and N atoms at octahedral interstitial sites. The FCC P-WNx phase, with x == 0.5 (also called P-WNo s or P-W2N) has lattice constant a=4.126 A, hence it has a lattice mismatch of -24% with Si (a=5.431 A) and -27 % with GaAs (a=5.653 A) [JCP88]. The P-WN0 5 stoichiometry is described as a defective NaCl-type structure, with half of the octahedral interstitial sites filled with N and half of them vacant [Hon00]. Bonding in metal nitrides is complex, with both metallic (valence electrons delocalized; non-directional bonding) and covalent (valence electrons shared; highly directional bonding) characteristics, due to the combination of metal-metal and metal-nonmetal interactions [Tot71]. Hones et al. [2000] described metal-nonmetal bonding in NaCl type structures as having n-like and cr-like bands due to overlap of the d and 2p orbitals from the metal and nonmetal, respectively. These bands (called pdn bands) have ionic character, with bonds having more ionic character when this band is populated and more covalent-metallic character when the band is depopulated [HonOO]. Shen et al. [2000a], using transmission electron diffraction (TED), found W-N and W-W nearest neighbor distances of 2.08 and 2 92 A, respectively, in sputter deposited P-W2N films. The P-W2N lattice parameter (4.126 A) represents the distance between centers of W atoms at the comers of an FCC unit cell, while the 2.92 A W-W correlation represents the distance between centers of the comer and face centered W atoms in the FCC unit cell (Figure 2-1 ) The theoretical density of FCC P-WN0 5 based on 4 tungsten and 2 nitrogen atoms per unit cell, is 18.0 g/cm3 while experimentally determined densities range from 8.0-17.9 g/cm3 [Bos91, Gal97, Hec02, Mar93, Sam80, Sot03].

PAGE 83

74 ) 4.126 A l fJ 2.92A Figure 2-1. Distances between W atoms in the FCC face of the J3-W 2N phase. Shen and Mai [2000b] have reported that binding energy shifts (per XPS) for W and N in polycrystalline J3-WN0 5 films indicate an ionic bonding character in the films (consistent with W&+ -~). The apparent binding energy difference between the metal and N atom, from XPS, gives an idea of the degree of charge transfer. A large binding energy difference infers greater bond ionicity [Pri95] Moreover, adding more nitrogen to the films has been suggested to cause a decrease in the number of free electrons provided by W in the solid film [Sot03], which is consistent with increasing film resistivity with nitrogen content. At lower N levels, the metal's band structure is retained, so that metal nitrides show metallic properties [Muk93]. Matsuhashi and Nishikawa [1994] reported FCC J3-WNx films to have a work function before annealing of 5.0 eV, where work function is a measure of the energy required to remove an electron from the material to a state of rest outside the material (i.e., at the vacuum level). Transmission electron diffraction (TED) intensities were also used to determine the concentration of W-N neighboring pairs in the J3-WNx films. Results showed that the concentration of W-N neighboring pairs did not increase when bulk N content was increased above x-0.5 [Shen00b]. This indicates that excess Nin the films migrated to

PAGE 84

75 the polycrystalline grain boundaries rather than filling the remaining vacant interstitial sites in the FCC lattice. Shen et al. [2000c] reported no SHP 8-WN formation, even for WNx with x=l.22, which shows that excess N migrated to the grain boundaries in the polycrystalline FCC films. Other reported properties for the binary nitride include a compressive stress of 4.3 0.5 GPa, nanohardness of 30 GPa, and a Poisson's ratio assumed to be 0.25 for films with the WN0 6 stoichiometry [Hon00]. In addition, depending on the WIN ratio thick WNx films were reported with refractive indices and absorption coefficients ranging from 3 20 to 4.00 and 1.20 to 3.95, respectively as determined by ellipsometry [Boh90]. Oxygen impurities tend to form solid solutions with metal carbides and nitrides [Tot71], and oxygen is reported to oxidize the surfaces of polycrystals [Par96]. Impurities such as 0, N and C are reported to enhance the stability of diffusion barrier films [Cha97c, Cha99, Lee94a]. Various stoichiometries of tungsten oxynitride have been reported. The JCPDS powder diffraction standard indicates two phases of W(N,O)x, one with x=0.62 and a=4.138 A, and the other with x= 0.57 and a=4.126 A [JCP88]. A WN1.34Oo.42 phase with a=4.153 A has also been reported [Sel95]. 2.5 WNxCy Film Properties Little information is available in the literature on the structure and properties of WNxCy films In going from nitrides to carbides, electron density increases around the metal atomic sphere and decreases around the non-metal atomic sphere, consistent with more ionic bond character for the nitrides and more covalent bond character for the carbides [Gub94] This is also consistent with the higher resistivity of 13-WNx relative to 13-WCx [Kim03a, Nak87]. Adding carbon to the nitride phase therefore leads to a

PAGE 85

76 decrease in film resistivity. The FCC WNxCy phase, with N and C intermixing on the interstitial sublattice of FCC W, has been predicted by thermodynamic analysis [Fri99b, Hua97, Jon93], and has been studied for diffusion barrier applications. A detailed experimental analysis of the composition, structure and electrical properties of these films has not been reported, however. WNxCy film deposition was reported using WF6 +NH 3 +B(C 2 H5 )3 at 350 C [Ele03 Li03]. The film composition was Wo.ssNo.1sCo.3o, and films had a nanocrystalline P-WCx or P -WNx cubic structure in an amorphous matrix along with low resistivity (210 to 400 .Q-cm) The films had F, 0 and B impurities below 0.5 at.% H con tent < 4 at.%, density of 14 g/cm3 and resistivity of 300 to 400 .Q-cm [Li03]. Another report using the same chemistry gave a film composition of Wo.s1No_13Co.3o and resistivity ranging from 600 to 900 .Q-cm, but deposition temperature was not given [Smi02]. Kim et al. [2003d] also reported deposition of WNxCy films using WF6 +NH3+ B(C2H5 )3 at 350C. Film composition from RBS was 48 at.% W, 32 at.% C and 20 at.% N, and film density was 15.37 g/cm3 The film had low resistivity (350 .Q-cm), and had an electron diffraction pattern that closely matched those for P-WNo s and P -WCo 6 One HR-TEM lattice fringe spacing was 2.39 A, which was between the i nterplanar spacings for (111) P-WNo s (2.38 A) and P -WC0 _6 (2.43 A). The other lattice fringe spacing was 2.08 A, which was between the interplanar spacings for (200) P-WN0 5 (2.06 A) and P-WCo. 6 (2.11 A). These values support the presence of a ternary WNxCy solid solution. Moreover a 12 nm WNxCy film prevented Cu diffusion during annealing up to 100c.

PAGE 86

77 Vlakhov et al. [1995] deposited W, WC and WNxCy films to study their electrical resistivities, and stated that bulk tungsten nitrides, carbides, and carbonitrides were good superconductors [Vla95]. The W films were deposited from both W(CO)6 and WC16, the WC films were formed by annealing the W films in a carbon-containing atmosphere, and the WNxCy films were deposited by co-reacting W(CO)6 +NH3 + CH3COCH3 [Vla95]. The room temperature (300 K) resistivities of W, WC and WNxCy films with thickness of 0.4 m were compared. The Wand WC films had similar resistivities, ranging from 109 to 191 .Q-cm, while the WNxCy films had a much higher resistivity of 651 .Q-cm. 2.6 Amorphous WNx Film Deposition Incorporation of smaller solute atoms into a metal matrix will cause a crystalline solid solution to become unstable above a certain solute concentration, fostering amorphous film growth [Ell90]. To enable amorphous film deposition in binary systems, the atomic radii of the two elements must differ by more than 10 % [Ell90]. The atomic radius of N is 0.74 A, while that for W is 1.394 A, hence N is 47% smaller than W, and so WNx should be able to deposit in amorphous form. Moreover nitrogen can act as a roadblock to trap diffusing W on the film surface, or can serve as a nucleation site for lattice defects. The trapped W and permanent defects caused by N incorporation prevent crystallization, and amorphous growth occurs [She00b]. Several reports of amorphous WNx deposition by a variety of deposition techniques have been given. These include reactive sputtering, LPCVD, PECVD, MOCVD, annealing Win NH3, PLO and ion implantation. FCC '3-WNx films with x < 0.5 were typically amorphous, while those with x ::::: 0.5 contained polycrystalline

PAGE 87

78 P-WN0 5 and those with x > 0.5 contained P-WNo.s and additional N at the grain boundaries [SheOOe,f]. Depending on their composition, amorphous WNx films have been reported to crystallize at temperatures ranging from 480-620C [Suh0l]. Additional contamination from C will likely increase the crystallization temperature of WNxCy films. 2.7 Demonstrated Uses of WNx WNx as a Diffusion Barrier Kilbane and Habig [1975] first reported growth of WNx thin films by reactively sputtering a W target in a mixture of Ar/N2 gas. Reichelt and Bergmann [1975] also studied deposition of WNx thin films by RF sputtering. Ten years later, two reports were published describing the use of sputter deposited WNx films as diffusion barriers [Aff85, Kat85]. Since then, many studies of WNx as a diffusion barrier separating Cu and Si have been reported. References for some of these reports are listed in several review articles [Jai99, KalO0, Kim03b]. The use of WNx as a diffusion barrier on low-k materials has also been examined. The interaction between WNx and the low-k material hydrogen silsesquioxane (HSQ) has been studied [Zen00a]. Retention of H in the HSQ film is very important, as the dielectric constant (k) of the film is a strong function of H content (as Si-H). A W 2 N barrier film deposited by PECVD of WF6+N2+H2 prevented the release of H (as H2) from HSQ more effectively than its TaN counterpart deposited by PVD [Zen00a]. Adhesion strength between WNx and two other low-k dielectrics, an aromatic hydrocarbon (SiLK, from Dow Chemical) and a polyarylene ether-based polymer (Flare 2.0, from Allied Signal), has also been studied [Lan00]. TiN, TaN and WNx all had similar

PAGE 88

79 adhesion strengths to these low-k materials, suggesting that a similar bonding mechanism exists between the barriers and these low-k films [LanOO]. These barriers are expected to have stronger adhesion than Cu to the polymer layers, because Cu, unlike the refractory metals, has fewer unfilled ct-orbitals available for bonding to the surface [LanOO]. WNx films have also been tested to separate Cu from the low-k material fluorine doped silicon oxide (SiOF) [Lee98c]. WNx had poor adhesion to SiOF layers which had not been pre-treated with 02 plasma at 300C, and failed to prevent Cu-SiOF intermixing. When the SiOF layer was pre-treated, however, WNx had good adhesion to and prevented Cu diffusion into SiOF for a 30 second anneal at 900C. Surface oxidation and densification of the SiOF, caused by the plasma pre-treatment, are believed to enhance barrier capability on pre-treated SiOF [Lee98c]. WNx films have also been tested as diffusion barriers between the Ni and W layers in an InxGa1-xAs/Ni/W contact structure for use with GaAs [Uch97]. These reactively sputtered WNx films prevented out-diffusion of In from the lnxGa1-xAs layer, and also lowered the contact resistance from ( 4 to 1 .Q-cm). WNx formed by nitridation of an amorphous CVD-W layer using N2 plasma was used as a barrier in the multilayer structure Cu/amorphous-WNxfamorphous-W/p+n-Si diode [Cha99]. The structure maintained its integrity after annealing at 725C. The same group determined the upper limit for a Cu/amorphous-WNxfamorphous-W/Si multilayer structure to be 750C [Cha97a, Cha97c].

PAGE 89

80 WNx as Gate Electrode in MOSFET Devices Historically, metal oxide semiconductor field-effect transistor (MOSFET) devices have used SiO 2 as the gate dielectric. Shrinking device sizes are forcing a decrease in SiO 2 layer thickness. Devices formed with SiO2 layers less than 3 run thick suffer from high leakage current due to direct tunneling of electrons through the oxide, hence an alternative material with higher permittivity is required to prevent tunneling [Lee0lc]. Ta2Os is a potential alternative due to its high dielectric constant (20-25) relative to SiO 2 (3.9) [May90, Par98a]. New contact materials, including WNx, are being investigated for devices with Ta 2 O5 gates. WNxffa 2 O5 structures had lower leakage current than TiNffa2O5 structures after annealing at 900C [Lee0lc]. In addition, the WNxffa2Os structure had no interfacial reaction, while the TiNffa2Os interface degraded after annealing. The superior performance of WNx is reportedly due to diffusion of N from WNx into the Ta2Os layer, which suppresses Ta and O out-diffusion [Lee0lc]. Diffusion of N into Ta2Os did not occur for the TiN/fa2 O5 structure. WNx has also been used as a diffusion barrier and W source in poly-Si gate stack structures [Gal00, Yan02]. Without a barrier, W reacts readily with Si, forming a highly resistive WSix phase [Kas94]. Kasai et al. [1994] formed a W/WNxfPoly-Si gate structure with a 50 A thick layer of WNx. The sheet resistance of W in this gate structure was low (1.6 Q/ ) compared to one without the barrier layer (18 Q/ ), which suffered from silicide formation. First, PECVD WNx was deposited on poly-Si and annealed to form a W-Si-N layer. Then, the structure was annealed at temperatures up to 1000C in either an Ar or N2 atmosphere to convert some of the remaining WNx layer to W. The newly formed layer of W can then be used as the contact metal for the gate. While

PAGE 90

81 as-deposited WNx films had sheet resistance of 19.2 Q/ the W layer produced by annealing had sheet resistances as low as 1.28 Q/ [Gal00]. Sputtered WNx was tested in a similar manner; formation of WSii was suppressed due to formation of a Si)N4 layer at the WNx-Si interface after a rapid thermal anneal (RTA) at 1000C [Yan02]. Kang et al. [2001,2002] used a sputtered, 100 A thick WNx film as a barrier layer in a W/WNxfpoly Sii-xGex gate structure for use in a CMOS-FET. Takagi et al. [1996] used a WIWNx polysilicon gate electrode to produce a CMOS device. The WNx layer suppressed silicide formation after a 30 second anneal at 950C, and the specific contact resistance for a W/WNxfpoly-Si structure was 10-7 .Q-cm2 [Tak96]. In addition, the sheet resistance of WNx deposited on poly-Si was reported to be an order of magnitude lower than that for WSix [Cho02]. WNx as a Gate Electrode in MESFET Devices Rectifying (Schottky) contacts for GaAs self-aligned gate field-effect transistors (SAGFETs) must have good thermal stability, good adhesion to GaAs, and high Schottky barrier height [Lee95, Pac91]. WNx Schottky contacts formed by PECVD of WF6+NH3+H2 on GaAs maintained their interface integrity after a 30 second, 1000C RTA [Lee95]. Yu et al. [1988] sputter deposited thin WNx films for use as a Schottky contact to GaAs, while Boher et al. [1990] deposited thick WNx films. Kim [1994] used PECVD of WF6 + NH3 + H2 at 350C to deposit WNx Schottky contacts to GaAs. WNx nucleated easily on GaAs and had higher thermal stability than W films, maintaining the W2N stoichiometry and blocking As out-diffusion during a RTA at 1000C for 30 seconds [Kim94]. Grain boundaries in the PECVD WNx films were stuffed by excess N, preventing Ga and As out-diffusion, and this Schottky contact structure was stable up to

PAGE 91

82 850C. Paccagnella et al. [1991] examined the effect of GaAs pre-treatment on the performance of WNxfGaAs Schottky diodes. Schottky banier height was highest for WNxfGaAs diodes annealed at 800C, where GaAs was pretreated by H2 plasma [Pac91]. Sputter deposited WNx has also been used as a gate metal in GaAs SAGFETs [Nag94, Gei86, Uch86]. In particular, WN0 _04 was found to produce Schottky baniers with excellent stability after annealing for 20 min at 810C [Gei86]. WNx films have also been deposited onto 4H-SiC to form Schottky contacts with high temperature stability [Pec97, Kak99]. The contacts, deposited by room temperature magnetron sputtering of a W target in Ar/N2 gas, remained rectifying up to 1200C despite formation of the W5Sh and W 2C phases [Pec97]. WNx as an Ohmic Contact in HBT Devices Thermally stable, low resistance ohmic contacts are critical to produce high speed, high frequency devices such as heterojunction bipolar transistors (HBTs). Park et al. [1998b] deposited the Au!Ptffi/WNx metal contact structure onto n-InGaAs, which is a cap layer used for AlGaAs/GaAs HBTs. The minimum contact resistivity for the Au/Pt/Ti/WNxfn-InGaAs structure was 9.5 x 10-8 .Q-cm2 which was obtained after annealing at 400C [Par98b]. In addition, the morphology of the contact remained smooth over a wide annealing range. WNx as a Top Electrode in DRAM Devices WNx films have also been investigated as top electrodes for Ta205 dynamic random access memory (DRAM) capacitors [Cho0l, Mat94]. Cho et al. [2001] deposited both TiN and WNx by reactive sputtering in Ar/N2 ambient to test their stability as a diffusion banier at the interface in a W /Ta205 gate structure. The structures with WNx

PAGE 92

83 barriers had higher thermal stability, as they were more resistant to Ta and 0 out-diffusion than TiN. The increased stability was presumably due to migration of N from the WNx film to the WNxlfa2Os interface, which prevented Ta and O from diffusing out of the Ta2 O5 Suzuki et al. (1998] used WF6 with two different N sources, NH3 and NF3 to deposit WNx thin films by PECVD. When tested as top electrodes in DRAM capacitors, these films had leakage currents that were an order of magnitude lower than PVD TiN top electrodes. Matsuhashi and Nishikawa [1994] tested sputtered WNx, TiN and TaN as top electrodes in Ta2 O5 DRAM capacitors. After annealing for 30 min at 800C, the WNxlfa2 O5 structure exhibited lower leakage current than structures with TiN or TaN electrodes [Mat94]. Kim et al. (2001] used a W/WNxlPoly-Si gate stack to produce polysilicon based DRAM devices. WNx as an X-ray Absorber Mask The use of amorphous WNx films in absorber masks for X-ray lithography has also been demonstrated [Lee97, Lee98b]. WNx has good stress controllability, strongly attenuates x-rays, and has a coefficient of thermal expansion (CTE) very close to common x-ray membrane materials such as SiC, BN, SiNx, and CN, making it a desirable x-ray mask material [Lee98b]. WNx masks were grown by reactive sputtering of a W target in an Ar/N 2 atmosphere on room temperature indium tin oxide (ITO) coated Si substrates. Film microstructure depended on the nitrogen content of the films, with 20 at. % N being the upper limit for the amorphous phase. As N levels increased above this, film microstructure shifted to polycrystalline f3-WN0 5

PAGE 93

84 WNx as a Liner for Cu Deposition Sputtered WNx films were also shown to be good liner materials for seedless electrochemical deposition, due to strong Cu-WNx adhesion and high Cu nucleation density on the WNx surface [ShaOl]. ECD Cu was deposited on air-exposed WNx films, and the composition of the ECD bath was controlled to strip off surface oxides [ShaOl]. Use of TiN as a seed layer for electroless Cu deposition resulted in only sporadic Cu coverage on the barrier layer [Mur95], while electroless Cu deposition occurred readily on WNx films [Wan03]. For the same bath temperature (70C), electroless Cu deposition occurred more rapidly on WNx than on TaN [Wan03]. Moreover, Cu deposited by CVD had low via resistance on WNx, but high via resistance with Ta based barriers [Jai99]. Other WNx Uses B-WNo.s has found use in bulk form (i.e., powder) as a catalyst for quinoline hydrodenitrogenation [Abe93], n-heptane isomerization [Sel95], deamination of 2-octylamine and alcohol dehydration [Lee92, Luc96]. Reactively sputtered amorphous WNx films have also been deposited and used as a means to form equiaxed, low resistivity W films in W/poly-Si gate structures [Lee03]. Amorphous WNx films annealed at 1273 K released nitrogen to form BCC W with a resistivity of 12 .Q-cm, which was similar to the value for pure, sputtered W The presence of a small amount of residual Nin the films at 1273 K suppressed silicide (WSix) formation, enabling the W film to retain its low resistivity. WNxEtching Several etching studies have been done on WNx films. Lee [98b] first reported use of SF6 + Ar + N2 in an inductively coupled plasma (ICP) etching system to

PAGE 94

85 anisotropically etch WNx thin films. Etch rates ranged from 4000 to 14000 .A/min. Vijayendran et al. [1999] used an NF:vN2 reactive ion etch (RIE) to remove WNx films. Reyes-Betanzo et al. [2002] used SF6 + Ar to anisotropically etch W and WN x films. 2.8 WNx Deposition Techniques A variety of techniques to deposit WNx have been tested in order to optimize key film properties, which include film resisitivity and deposition temperature. Ideal film resistivity was reported to be 500 .Q-cm, [Ele03], while a temperature ceiling of 400C has been given for IC production by numerous reports [EisOOb, Ele02, Hau00, Kim03b, Les02, Sun0lb]. Depositing the barrier at the absolute minimum temperature, however, may not be the best solution, as other processes occurring during IC processing will typically take the device temperature up to -400C. If the barrier is deposited at much lower temperature than the processing temperature ceiling, significant shift in barrier structure can occur during subsequent high temperature processing, leading to Cu diffusion and potential barrier failure. The reported deposition techniques include reactive sputtering, annealing of Win NH3, plasma nitridation of W, LPCVD, PECVD, MOCVD, and ALO. Reports of each of these techniques to deposit WNx are discussed below. 2.8.1 Annealing W in NH3 Deneuville et al. [1989] reported formation of WNx by annealing sputtered W films on Si in an NH3 atmosphere from 500 to 1100C The N/W ratio increased with temperature, going from 0.37 to 1.85 across the aforementioned temperature range. While several polycrystalline structures were postulated based on film composition XRD results were not given to support them. High temperature and pressure, along with long

PAGE 95

86 anneal times, are required to form WNx in this manner, making this an undesirable deposition technique for IC manufacturing. 2.8.2 Plasma Nitridation / Ion Implantation of W W films, deposited by CVD of WF6 + Si~ + H2 were nitrided to form thin WNx films using an in-situ N2 plasma treatment [Yeh96]. This nitride layer prevented WA112 formation after annealing at 550C, where WA112 formation can degrade contact structures using W plugs between Al and Si. WNx films were also formed by W ion implantation into W metal substrates [Zha99]. Barrier testing with Cu was not done for either of these films, however. W ion implantation was also done on W films deposited by PECVD of WF6 + H2 to form amorphous WNx films [Kwo95]. These films resisted Cu penetration after a 30 min anneal at 800C, while polycrystalline PECVD WNx films failed. The nitridation technique is directional in nature, however, meaning that good conformality will be difficult in highly aggressive future device topographies. Chang [1997a-c, 1999] reported Wion implantation into W films deposited by CVD of WF6 + Si~. The films had resistivity of 198 .Q-cm and resisted Cu diffusion for annealing temperatures below 700C, but had poor adhesion to the underlying Si substrate. 2.8.3 Pulsed Laser Deposition Soto et al. [2003] first reported pulsed laser deposition of tungsten nitride films from a W target in the presence of N 2 Films were deposited on n-type Si (111) and on Corning glass slide substrates, at N 2 pressures ranging from lxl0-8 to lxl0-1 Torr. Films contained W, N and some 0, believed to originate from W target contamination, and had higher density than films produced by DC magnetron sputtering. This technique, like

PAGE 96

87 plasma nitridation, is also directional, making its ability to deposit highly conformal films in future device generations questionable. 2.8.4 Reactively Sputtered WNx Deposition Typically, sputtered WNx films are formed by sputtering a W target in a Ni/Ar atmosphere. A variant of this, known as nitrogen ion beam sputter deposition (IBSD), has also been reported [Eiz94b, Gal93], where a nitrogen ion beam, rather than an Ar+ ion beam, is used to liberate W atoms from the W target. Nitrogen ions or radicals backscattered from the target are then incorporated with W at the substrate to deposit WNx. Both variants enable amorphous film deposition due to the decreased mobility of W in the presence of N on the substrate surface N can inhibit W mobility on the substrate surface by blocking W diffusion or by trapping W. Decreased W diffusivity leads to defect formation, failure of crystal growth and amorphization of the growing film [SheOOb]. Sputtering can be done at low deposition temperature (at or near ambient), which protects temperature sensitive components from thermal damage during the barrier deposition process. Many reports of sputtered WNx deposition have been given in the literature. The key properties of these films are summarized in Table 2-1. While sputter deposited WNx films are generally contaminant free, the major drawback of sputtering, as mentioned in Chapter 1, is poor conformality in small, high aspect ratio device features. 2.8.5 LPCVD WNx Deposition LPCVD of WNx has been researched as a possible alternative to sputtering for smaller device features with high aspect ratios. Many reports of LPCVD WNx films have

PAGE 97

N O" 0 0 ::, ()Q < 0 ::, Table 2-1. Key properties of sputtered WNx films reported in the literature ::, .... ::r 0 i::::::-: '"1 5:1) i Sputter Dep. Resistivity Contam. Conformality Dep. Rate Target/ Gas Temp. Range (%) Levels (Nmin) Adhesion Composition (OC) (Q-cm) (at % ) i---3 ::r 0 :,,;" 0 '-< Excellent "Cl a "Cl 0 ::l0 rn 0 ...... .... ::r 0 rn on Si (scotch tape test) W (target)+ 25-400 40-11000 NIA Ar<1% 17-700 Poor on Ar/N 2 gas 01-12% GaAs above 21 at.% N, 0 Poor on ::ti 8 quartz rn rn i:: ~ -N W (target)+ 20-600 200-540 NIA NIA 9-90 NIA Wion beam 0 0. ::, i---3 W (target)+ 25 132 15685 NIA NIA 12-63 NIA Ne/N 2 0 N I Failure Temp. with Cu (OC) 600-900 NIA NIA References [Aff85 Bak02, Bou97 Cha89 ChoOl Cho02 Gei86 Gic8 3, Hon03, Hub88 Kak99 Kat85 Kil75 Kol87 Lee90 Lee97 LeeOOb, Lee03 Lin90 L i p03 MigOI Moi98 Pac91 Pec97 Pok91 Rei75 Rey0 2, ShaOI, SheOOaf So88 Sot03 Suh99 SuhOI Tak97a Uch86 U e k96 Vu90 Yu88] [Bos91 Eiz94b Gal93 Z h a 99] [Hub88] 00 00

PAGE 98

89 As evident from Table 2-2, LPCVD reactions are based on halide chemistry, and include the reduction of WF 6 WO 3 or WC16 by NH 3 or NF 3 in a H2 atmosphere [Chi93, Gon02, Mar93, Nag93, Nak:87, Sak96, Suz98, Vol85]. Nagai and K.ishida [1993] reported that use of WC16 as the tungsten precursor for reaction with NH3 and H2 was more thermodynamically favorable than using WF 6 as the Gibbs free energy change was larger for the chloride precursor reaction. Marcus et al. [1993] have reported high LPCVD deposition rates, ranging from 1800-4500 A/min. One report (not included in the table) of the "indirect" deposition of ~-WNx has been given, where this phase formed after annealing a metastable W3Ns phase [Zha97]. Reacting WNCh with ZnN2 produced the W 3 N5 phase at 400C, and subsequent annealing at 600C resulted in formation of the ~-WNx phase. The halide precursors used for LPCVD can create difficulties during barrier deposition and subsequent device processing. WF 6 is reported to consume Si during the reaction, forming SiF4 which leaves vacancies on the Si substrate [Kim91, Kim92, Lai98a]. In addition, there are concerns that residual F and Cl in the barrier contribute to corrosion of metal interconnects [Fal98, Kel99, Raa93]. Contamination levels of 0.1-0.9% F have been reported for LPCVD films [Mar93]. Moreover, reactive by-products (e.g., hydrogen halides) resulting from LPCVD are a material handling concern [Cur92, Kim91]. Adduct formation is another difficulty associated with LPCVD. Adducts are compounds produced by chemical addition of two or more reactants. These adducts can settle onto the substrate and barrier film during deposition, resulting in dislocations and/or pinholes in the film that can lead to barrier failure. Gas phase adducts such as

PAGE 99

* ff O" C: ;;;: ,:0 I $3 z '8 Jt Er ::ti a g-(") 0 0.. 5 ..... a: rJ'l .g Table 2-2. Key properties of LPCVD WNx films reported in the literature LPCVD Dep. Resistivity Contam. Temp Conformality Dep Rate Deposition Range Range ( % ) Levels (Almin) Chemistry ( OC) (Q-cm) (at % ) F 1800-450 WF6+NH3 450-700 900-2800 NIA 0.1-0. 9 % 0 WF6 + NH 3 + 400-700 NIA NIA NIA NIA H 2 WF6+NH3 + H 2 by photo assisted 600 800-1000 NIA NIA NIA LPCVD (PACVD) 45-100% WF6+NH3 + 325-450 200-15000 (Diam: 0.2m, NIA 400 s~ AR: 3.0-9.8) WC16+ NH3 + 500-900 NIA NIA NIA NIA H2 WO 3 +NH3 700-1000 NIA NIA NIA NIA Failure Temp. Adhesion with Cu References ( OC) Cl) C: N 1.0 00 rJ'l p.) 1.0 0\ ....... Good [Mar93 (scotch NIA tape) Wan03] p.) ::, 0.. ::r NIA NIA [Nak87] p.) < 0 O" 0 0 ::, NIA NIA [Par97a] 1.0 ca 0 "C) 0 0 0.. 0.. C: ::i. Excellent ::, Otl (scotch 450 [Gon02] tape) after Ar plasma t; n 8 ca NIA NIA [Nag93, NagOO] p.) (") ..... 0 ::, rJ'l ,...., NIA NIA [Vol85]* z 00 .....J

PAGE 100

91 2.8.6 PECVD WNx Deposition The same halide chemistries used for LPCVD are also used for PECVD, whereby the plasma assists fragmentation of the reactants, lowering reaction temperatures considerably. After fragmentation by the plasma, the precursor fragments travel to the heated substrate, react on its surface and deposit a film. Many reports of PECVD WNx films have been given in the literature. The key properties of these films are summarized in Table 2-3. Films deposited by PECVD show excellent adhesion, with resistivities as low as 70 .Q-cm. Growth rates are moderate, ranging from 300-400 A/min [Gal97], and conformality up to 70 % has been reported for trench holes with diameter~ 0.30 m [Kim93]. PECVD films suffer from reduced conformality in small diameter, high aspect ratio trench features, however, due to the directional nature of the plasma [JohOOa, Tsa96]. Despite the low resistivity and deposition temperatures that PECVD offers, its inability to deposit highly conformal films in high aspect ratio features makes its use in future barrier deposition questionable. Moreover, PECVD films, like LPCVD films, suffer from halide impurities. Concerns about halides contaminating the barrier films also exist for PECVD, since the same reactants are often used for both LPCVD and PECVD. Contamination levels up to 4 at.% F, which are higher than those for LPCVD, have been reported for PECVD films [Lee93]. Another difficulty associated with PECVD is gas phase adduct formation; adducts such as (WF6)x(NH3)x and have been reported during WNx PECVD [Lu98b, Suz98]. These adducts can contaminate films grown by PECVD and deposit particles on the film surface, both of which can

PAGE 101

Table 2-3. Key properties of PECVD WNx films reported in the literature PECVD Dep Resistivity Contam. Temp Con formality Dep. Rate Deposition Range Range (%) Levels (Nmin) Chemistry (OC) (Q-cm) (at%) F: 1 at.% WF6+ N2+ 275-475 190-240 36-70% (0.3 C : -1 at.% 600-1800 H 2 m 3:1 AR), 0 : ~3. 4 at.% 33% (0.15 m, 9:1 AR), WF6+NH3 + 70% (2 m F: -2 at.% H2 25-620 90-550 2:1 AR), 0 : -2 at.% 210-400 60% (0.35 m, 3 5:1 AR) WF6+NF3 350 175 NIA NIA NIA Failure Adhesion Temp. with Cu (0C) Good (Scotch 400-700 tape) NIA 500-800 NIA NIA References [Eck02, GanOO, Hec02, Jin99, Lai98a, LiOO, Lu98b, Vij99, ZenOOa,b] [Bai03, Gal97, lva99, Kim91, 92, 93, Kwo95 Lee93 94a-c, Lee98c, Lin98a-b, Lin99, Meu98, Par97b, Par98a] [Suz98] \0 N

PAGE 102

93 cause premature device failure. The difficulties associated with sputtering, LPCVD and PECVD have shifted attention to the use of metalorganic precursors for WNx barrier deposition. At present, several metalorganic WNx precursors have been reported in the literature. Our research has focused on the thermal decomposition of novel single-source precursors via MOCVD. 2.8.7 MOCVD WNx Deposition MOCVD of WNx has many desirable properties. Films deposited by metalorganic complexes such as (1BuN)2 W(NHtBu)2 or W(CO)6+NH3 do not contain halides, which can etch subsequent metallization layers when incorporated into the barrier film By varying the structure of the MOCVD precursor molecule, the precursor can be optimized so that it dissociates in a "clean" fashion (i.e., with minimal oxygen and carbon contamination) and at relatively low deposition temperatures. Temperatures down to 200C have been demonstrated for MOCVD film deposition [Chi93, Kel99]. The presence of impurities (0, N, C) is reported to improve stability in contact structures, because the impurities tend to stuff grain boundaries and inhibit diffusion [S088]. Oxygen's "stuffing" effectiveness, however, is lower for Cu than Al, because the reactivity of Cu with O is less than that for Al with O [Kim99c]. Carbon has also been reported to improve thermal stability of W barriers and foster growth of smaller grains [Wan0lb]. Although these impurities may stabilize the film and decrease polycrystal size, they may also act as electron scattering sites, increasing film resistivity. The challenges for MOCVD of refractory nitride materials include minimizing carbon and oxygen contamination, minimizing film resistivity, and lowering the deposition temperature.

PAGE 103

94 Kelsey et al. [1999] reported the use of W(CO)6+NH3 to deposit WNx films. The metalorganic precursor was fed from a solid source delivery system, and was mixed with NH3 in the reactor, over a deposition temperature range of 200 to 350C. The deposited films were polycrystalline P-W2N, with growth rates ranging from 40 to 100 .A/min. Resistivities as low as 123 .Q-cm were observed, which coincided with carbon and oxygen contamination levels below 5%, and step coverage was greater than 90% on 0.25 m feature sizes. Barrier integrity tests for a sample with an annealed metallization overlayer have not been reported for films grown by this precursor scheme, however. The merits of growth with W(CO)6 are tempered by the possibility that decomposition of the precursor's CO byproduct can cause oxygen contamination in the films. Typical strategies for CVD of multi-element barrier materials involve the use of co-reactant precursors. Co-reactant deposition uses a separate precursor for each element desired in the film; hence, bonds between these elements must be formed by intermolecular processes during deposition. In contrast, a single-source precursor already has bonds established between the elements that will comprise the film prior to deposition. This approach is particularly useful when the bond strengths in the individual precursor candidates, and thus decomposition temperatures, are quite different. Another advantage to using a single source precursor include elimination of the need for multiple reactants so that flow ratio control is eliminated as a process variable. In addition, gas phase adducts, which have been reported for many LPCVD processes, are no longer a concern, because multiple gas phase reactants are not being introduced to the reactor. Our research involves the use of novel single-source metalorganic precursors to deposit WNx thin films.

PAGE 104

Ta b le 2 4. Key properties o f M O CVD WNx films re p orte d in the literature Dep. Resistivity Contam Temp. Conformality M O CVD Deposition Chemistry Range Range ( % ) Levels (OC) (.Q-cm) (at%) 450 85 C : 6 at. % CBuNh W(NH1Buh -650 620 7000 (Diam: 0 .40 0: 3 at.% m, AR : 3) 200-90 C : < 5 at.% W(CO)6+ NH3 35 0 123 (Diam : 0 25 0:<5 m,AR:4) at. % CH3 (CH 2hCH(CH3 )NCW(CO)5 250 NIA NIA NIA +N 2 + NH3 -400 Failure Dep. Rate (.A/min) Adhesion Temp with Cu (0C) 20100 Poor NIA 40-100 NIA NIA NIA NIA NIA References [C h i92, Chi93 Cur9 2 Nug80, Tsa96] [Kel99 ] [GorOO] \0 Vt

PAGE 105

96 Nugent et al. [1980] first reported synthesis of the (1BuN)2 W(NHtBu)2 molecule, which would later be the first single-source molecule tested for MOCVD of WN x Currie et al. [1992] used this molecule to deposit a W/W2N composite coating. The precursor was dissolved in hexane solvent, and the liquid mixture was used to coat the substrate. The solvent was then evaporated off and the precursor-coated substrate was subjected to heat treatment up to 1200C. The resulting film was a composite mixture of a-W/f3-W2 N. No growth rates or resistivities were reported for these films, however. Chiu et al. (1993] reported use of the same molecule as a single-source MOCVD precursor to deposit WNx films. The precursor was transported to the reactor by a solid source delivery system, and the depositions were done by thermal decomposition over a temperature range of 450 to 650C. The resulting films were polycrystalline f3-W2N, with growth rates ranging from 20 to 100 A/min, and lattice parameters ranging from 4.14 to 4 .18 A. Tsai et al. [1996] later reported that these films were nitrogen rich wit h additional N expanding the lattice parameter above the 4.126 A standard for f3-W zN films. These films had resistivities ranging from 620 to 7000 .Q-cm w i th step coverage for a 0.40 m diameter feature up to 85%. Resistivity increased with an increase i n the N/W ratio of the film, while the N/W ratio increased with decreasing deposition temperature Oxygen and carbon contamination levels of 3 and 6 at.%, respectivel y were given for these films. Crane et al. (2001] later studied the reaction mechanism for WNx deposition from this precursor. Barrier integrity tests with an annealed metallization overlayer have not been reported, however, for films grown by this precursor.

PAGE 106

97 Gordon et al. [2000] reported deposition of WNx thin films from tungsten (0) pentacarbonyl 1-methylbutylisonitrile (CH 3 (CH2)2CH(CH3)NCW(CO)s). A nebulizer vaporized the liquid phase precursor, and this vapor was conveyed to the reactor by a carrier gas mixture containing N 2 and NH3 Substrate temperatures ranged from 250 to 400C, and the reactor pressure was atmospheric. Films were amorphous per XRD, and Rutherford backscattering (RBS) showed the film stoichiometry to be near Wo.61No.33 One report of indirect P-W2 N deposition (not included in the table) from a MOCVD precursor has been given [Zha97]. The MOCVD precursor [WNCIJ-NCCH3]4-2CH3CN was used to deposit a metastable W3N5 phase, which was then annealed to form P-W2N. Key properties from several reports of MOCVD WNx films are summarized in Table 2-4. 2.8.8 ALD WNx Deposition The use of ALD in mass production is suggested to be necessary by the 65 nm node for diffusion barriers [Bey02, Han03] and by the 45 nm node for high-k gate dielectrics and the first few metal layers [Han03, Cha04]. In the front end, step coverage is not an issue, but ALD is very useful to deposit ultrathin films of high-k dielectric materials. In the back end, ALD is useful due to its excellent conformality, which is required to deposit seed and barrier layers in high aspect ratio vias and trenches. ALD of WNx films by sequential reaction of adsorbed WF6 with NH3 on Si(lOO) and SiO2 substrates has been reported [Ele02, KlaOOa-b]. Deposition of a WNx monolayer occurred by 2 half-cycles. In the first half cycle, WF6 was introduced into the reactor. After monolayer adsorption onto the substrates, any remaining WF 6 in the reactor atmosphere was pumped out. In the second half cycle, NH3 was introduced into

PAGE 107

98 the reactor to convert the adsorbed WFx monolayer into W2N. Klaus et al. [2000a,b] reported a linear deposition rate of 2.5 A/cycle for reaction temperatures ranging from 600 to 800 K, but the films had very high resistivity (4.5 x 106 .Q-cm). Elers et al. [2002] reported a deposition rate of 0.42 A/cycle at 350C (623 K), along with high resistivity (4500 .Q-cm) and 2.4% F content. Conformality data were not presented for these films. Nitride growth with ammonia by ALD is reported to have a low growth rate, which is due either to a small number of NHx groups left on surface after NH3 pulse or due to adsorption site blocking by readsorbed reaction byproduct HCl [Les02]. Sim [2003] used WF6 and a pulsed plasma NH3 to deposit WNx thin films by ALD at 350C on Si and non-Si substrates. The pulsed plasma NH3 provides reactive radicals such as Nlt', NHt, NH/ and Ir, which prevent fast reaction of WF6 with Si. The deposition rate was 2.2 A per cycle (on both Si and TEOS), and a 220 A thick barrier prevented Cu diffusion after a 30 min anneal at 700C. ALD of WNx thin films using the metalorganic precursor bis(tert-butylimido)bis(dimethylamido)tungsten (tBuNh(Me 2Nh)W and NH3 was recently demonstrated in the temperature range of 250 to 350C [Bec03a, Bec03b]. The growth rate for WNx films was 1 A per cycle, with no incubation period, and films were deposited on silicon, glass quartz, glassy carbon, stainless steel, aluminum, gold and copper substrates. The W:N stoichiometry of the films was shown to be -1:1 per RBS, and the films adhered very well to Cu deposited on their surface by ALD. The films resisted Cu in-diffusion up to 600C, but converted to pure W after a 30-minute anneal at 725C. Step coverage was 100 % in a feature with an aspect ratio greater than 200: I.

PAGE 108

99 A mechanism for the ALD reaction was proposed, in wh i ch i t was suggested that NH3 did not react with the metalorganic precursor, but instead catalyzed the removal of the alk y l groups on the precursor itself. Pyridine was used in place of NH3 and WN films wer e again deposited. 2.8.9 ALD WNxCy Deposition ALD of a WNxCy ternary solid solution onto Cu and SiO 2 substrates by sequent ial reaction of WF 6 NH 3 and Et 3 B has also been reported [Pet02, Smi02]. A 70 A thin film was deposited with a growth rate of 0.8 Alcycle and an AES composition of 57 % W, 30% C and 13% N. Conformality on 0.25 m vias was 100% on the feature bottom and sidewalls, while film resistivities ranged from 600 to 900 .Q-cm. Cu seed deposition by PVD was performed ex-situ before bias temperature stress (BTS) testing was performed on the samples. WNxCy had marginal adhesion on SiO 2 but via resistance with this barrier was about half that of PVD Ta [Smi02] The same chemistry was used to depos i t FCC ~-WNxCy nanolaminates at 350 C, which had res i stivity of 210 .Q-cm and excellent adhesion to Cu [Ele03, Li02]. Li et al. [2003] also deposited WNxCy films with WF 6 + NH3 + (Et) 3 B varying deposition temperature from 275 to 325C. The films had a growth rate of 0.8 t o 0.9 Alcycle, and step coverage in 130 nm features with a 40: 1 aspect ratio was > 90 % Kim et al. [2003d] reported deposition of WN x C y films us i ng ALD of WF6+NH3+TEB at 350 C. Film composition from RBS was 48 at. % W, 32 at. % C and 20 at.% N, and film density was 15.37 g/cm3 The film had low resistivity (350 .Q-cm ) and had an electron diffraction pattern that closely matched those for ~-WN0 5 and ~-WCo.6A 12 nm film survived a Cu anneal up to 700 C.

PAGE 109

100 2.9 Conclusions Continuous scaling down of barrier thickness will make material selection a critical issue. In thicker barrier films, deposition of an equiaxial, small-grained microstructure provides good barrier performance, because the diffusion path is very tortuous, hence Cu cannot penetrate through the barrier. Barrier thickness at the 100 nm node is 12 nm (120 A), however, and will shrink to 7 nm (70 A) by the 65 nm node in 2007 and eventually to 2.5 nm (25 A) at the 22 nm device node in 2016 [ITR02]. By 2016, the barrier layer thickness will correspond to 5 monolayers of barrier material. Barrier layers will become so thin in future device nodes that manipulating the microstructure (i.e., amorphizing the film) alone will not be enough to prevent Cu penetration, as any defects in the film will be weak points for Cu diffusion. The addition of an element such as nitrogen, which chemically repels Cu and stabilizes the metal against silicide formation, will be essential to bolster barrier resistance at weak points such as defects. Ru, for example, is a promising barrier and direct ECO candidate, but it does not chemically repel Cu, readily forms silicides, and could be susceptible to Cu diffusion through defects in film structure. The stuffed barrier concept, therefore, should likely be extended down as barrier layers are thinned Along with stuffed barrier materials, the future barrier deposition technique will most likely be ALO, due to the inability of PVD and CVD methods to produce ultra-thin, highly conformal films.

PAGE 110

CHAPTER3 EQUILIBRIUM MODELS FOR WNx-BASED BARRIERS 3.1 Motivation and Method The CVD process involves a gas phase, containing one or more species, interacting with a solid substrate, which may catalyze a reaction(s) leading to film formation. The CVD process is complex, with transport phenomena (energy, mass and momentum) and reaction kinetics controlling the rate of approach to thermodynamic equilibrium to determine which phase(s) will form and their extent. Transport phenomena govern the ability of the precursor(s) and any intermediate(s) to reach the reaction surface (substrate), while reaction kinetics address the ability of and pathway for the precursor(s) and intermediate(s) to react and lead to deposition of a film on the substrate surf ace. The parameters that affect transport phenomena in the reactor are the applied boundary conditions (e.g., susceptor temperature, reactor pressure, reactor inlet flowrates), the system geometry, and the thermophysical properties [Vos78]. Understanding the available kinetic pathways in the system is very important, as kinetic barriers (pathways with high activation energies) can prevent the growth of certain solid phases in favor of others or the incorporation of impurities. Kinetic limitations may prevent concentrations at the growth surface from being at their equilibrium values, and should be taken into consideration when predicting the phases expected to form in the film. Chemical equilibrium studies are useful to determine which solid phase(s) and gas species would form during CVD at specific process conditions (temperature, pressure and 101

PAGE 111

102 inlet compositions), barring any kinetic or transport constraints. Depending on the conditions in the reactor system, the process may or may not be close to equilibrium. While transport and kinetic constraints must be taken into account to form a complete CVD model, transport effects and especially kinetic pathways in the reactor are very difficult to unravel. Computing the thermodynamic equilibrium state in the reactor system is therefore a logical first step toward forming a complete CVD model. Thermodynamic modeling is an important tool used to estimate the equilibrium set of phases and their composition for a system at specific reaction conditions Experimentally, achieving thermodynamic equilibrium may be impractical, due to kinetic constraints, transport issues, and short time scales for reactant exposure to the substrate. To approach true thermodynamic equilibrium in the reactor, temperature should be maximized to overcome kinetic constraints, and the reactant exposure time should be maximized. If equilibrium conditions are assumed for a given process, the film provides a record of the growth condition at the interface, assuming the temperature is sufficiently low to quench the film Then, the equilibrium composition of the solid phase(s) formed during the reaction can be determined ex-situ, by a combination of characterization techniques such as XRD and XPS. Exclusion of certain solid phases evident from these techniques, such as surface oxides, may be needed if they form due to post-growth air exposure. The equilibrium composition of the vapor, however, must be determined in-situ, due to the disappearance of many high temperature species upon cooling to room temperature [Vos78].

PAGE 112

103 CVD tends to be a non-equilibrium process due to kinetic limits and short residence times of vapor species. Although CVD is a non-equilibrium process, equilibrium thermodynamic calculations can give us a preliminary idea of solid, liquid and gas phase composition in a multi-component system [Raz95], and hence can be used as a guide in choosing process conditions [Smi82]. In addition, an equilibrium thermodynamic model may be useful when rates of change are rapid (e.g., at high temperature). At high temperature, a mass-transfer limited regime typically exists If the mass transfer coefficients of the various reactants through the boundary layer are similar, then the reactant concentrations at the growth surf ace are equal to that in the bulk, and the surface conditions are a good equilibrium approximation. This utility diminishes at lower temperature, as reactor kinetics limit the rate [Smi82]. When performing a thermodynamic equilibrium analysis care must be taken to identify species whose formation is both thermodynamically and kinetically favorable. Each of the proposed species and phases should have a kinetic pathway to form and take part in the equilibrium. Several experimental variables are useful and necessary to carry out a thermodynamic equilibrium analysis. For a system operating at constant pressure, these variables include the system's temperature, pressure, and elemental makeup. Once these variables are specified, a list of potential species and phases present in the system is drafted. This list must then be checked and modified to include only those species and phases whose formation is kinetically possible. Finally, a set of equilibrium variables (e. g., temperature, pressure, moles of elements) must be specified to generate an equilibrium CVD phase diagram.

PAGE 113

104 The general criterion for equilibrium in an isolated system is that G', the overall Gibbs energy, be minimum [DeH93]. This is described mathematically as [DeH93, Kat97]: c = f G'a = minimum (3-1) a=l where G'a is the total (not per mole) Gibbs energy of the a phase, and p is the total number of phases in the system. If G' is minimum at equilibrium, dG' must be zero: dG' = f dG'a = 0 Cl=} (3-2) The expression for the change in overall Gibbs energy (G') for an arbitrary change in state in a system is [DeH93]: , C dG =-SdT+VdP+Lkdnk k = l (3-3) where k is the chemical potential, defined as the change in overall Gibbs energy with a change in the amount of component k present, holding temperature, pressure and composition of other components constant: (3-4) In a system with two components (1 and 2) and two phases ( a and '3) at equilibrium, for example, which is constrained to constant temperature and pressure (so that dT and dP are zero), the above expression for dG' becomes [DeH93]: dG' = dG'0+dG'~ = /dn/ + /dn/ + 2 dn2 a+ /dn/ = 0 (3-5)

PAGE 114

105 where 1 a is the chemical potential of component 1 in the a phase. In a two-phase non-reacting system, for example, dn 1 a = -dnl, so that the number of moles of each component is conserved Substituting this into the above equation yields: dG = (/-/)dn/ + (/ -/)dn/ = 0 (3-6) For the above equation to be valid, it is seen that the coefficients on dn1 a and dn2 a must be zero at equilibrium, since transport of components 1 and 2 between the a and P phases can occur (i.e. dn 1 a and dn t are non-zero). This implies that the chemical potential for a given component in the a phase must equal its chemical potential in the P phase at equilibrium. This result is general and can be applied to any number of components and phases, and is the criterion used to construct phase diagrams. Before setting up a phase diagram though the basis for Gibbs energy expressions (e.g. G' a and G'13) for different phases must be established. To calculate a thermodynamic equilibrium expressions describing the Gibbs energy of each phase must first be defined. If the system is assumed to be maintained at constant pressure (as in our experimental system) the effect of temperature on Gibbs energy can be examined. To begin, write the expressions for Gibbs energy, enthalpy and entropy for a simple single component system at constant pressure [DeH93] : G =H-TS T z C LlS = f-P dT r. T I (3-7) (3-8) (3-9)

PAGE 115

106 (3-10) This expression gives the Gibbs energy for a single component in terms of two experimentally measured variables, constant pressure heat capacity (Cp), which is typically measured by calorimetric methods, and temperature (T). Beginning with a list of values for the heat capacity as a function of temperature for a pure component, an equation may be fitted to this data to obtain a temperature dependent expression for heat capacity. Some common models for the heat capacity expression are shown in Equations 3-11 to 3-14 [Gur89]: C Cp(T) = a+ bT--2 T Cp(T) =a+ bT--;-+ dT2 T Cp(T) =a+ bT--;-+ dT2 + eT3 T Cp(T) =a+ bT--;-+ dT2 + eT3 + ff4 T (3-11) (3-12) (3-13) (3-14) Fitting experimentally determined heat capacity data to one of the heat capacity expressions above, and substituting this expression into Equation 3-10, enables calculation of Gibbs energy changes. For pure component phases, the Gibbs energy expression can often be determined experimentally by such techniques as calorimetry, gas phase equilibria measurements or electromotive force (EMF) measurements [Sau98]. Thermodynamic assessment, which is an alternative, stepwise approach, may be used when experimentally determined Gibbs

PAGE 116

107 energies are unavailable. The first step is to represent the Gibbs energy for the component(s) of interest by a power series in terms of temperature, as follows [Din91]: n 0Gf = a+ bT+cTlnT+ LdnTn (3-15) where O G f' is the Gibbs energy of a pure component and n typically has the values of 2, 3 and -1. In addition, a pressure dependent term may be added to the series if warranted. The values of n are determined during the procedure of thermodynamic optimization to obtain the best possible fit between the calculated and experimental compositions. Assessment involves modifications to the coefficients in the Gibbs energy expressions, so that the shape and position of phase boundary regions on the diagram, along with the thermodynamic properties (heat capacity, enthalpy, etc ) calculated from the Gibbs energy, are consistent with experimental observations, where available. The coefficients of the Gibbs energy expressions have physical meaning corresponding to such thermodynamic properties as enthalpy, entropy, and heat capacity, hence any expression resulting from an optimization should be validated against experimental values (when available) for these properties. If experimental property data are unavailable, the optimization can still be done, but equilibrium phase and composition data must be used to delineate the phase boundary regions, so that the resulting phase diagram is as accurate as possible. Estimation of the coefficients may be done by trial-and-error or mathematical methods [Kat97]. In this study, both the PARROT module within ThermoCalc and trial-and-error methods were used to optimize coefficients in the Gibbs energy expressions of interest.

PAGE 117

108 Multi-component mixtures have a more involved expression for Gibbs energy. The general expression for the Gibbs energy per mole of a solution phase containing i components is shown below: Gq:, mi x = "x. G0 + RT" x. ln x + oxs_ .::., I I .::., I I IIllX i i (3-16) where Xi is the mole fraction of component i, R is the gas constant and T is absolute temperature. The Gq:, mix term is the sum of three terms on the right hand side. The first term represents the contribution to Gibbs energy from pure components (also called the mechanical mixing term), with Gi being the Gibbs energy of pure component i per mole at its reference state (temperature pressure and phase). The second term represents the Gibbs energy change upon mixing to form an ideal solution, and for a solid solution assumes that components may mix on any position available in the phase without preferential site occupation [Sau98]. The third term, called the excess Gibbs energy of mixing represents non-ideal mixing interactions. The excess Gibbs energy can take several different forms, depending on which model one chooses to follow. The regular solution model for example, assumes that any non-ideal interactions between constituents in the solution depend only on temperature, and are not composition dependent. The excess Gibbs energy for a binary mixture of A and B using the regular solution model takes the following form: (3-17) where L is a temperature dependent parameter describing A-B interaction [Sau98]. A positive value for L indicates repulsive interaction between the components, while a negative value indicates an attractive interaction [Swa72]. The next level of complexity is for the interaction energies to incorporate composition dependence. An example is the

PAGE 118

109 sub-regular solution model, which assumes that interaction energies change linearly with composition [Sau98]: (3-18) where LA is the parameter describing A-B interaction on the A side of the composition range. More complex expressions for excess Gibbs energy are available, and are reviewed in detail elsewhere [Sau98]. Gibbs energies for some phases (e.g., unstable phases, multi-component phases) cannot be easily determined by calorimetric methods, and are determined instead by optimizing the diagram to fit experimentally observed equilibrium composition data. Once a Gibbs energy expression has been defined for each component included in the equilibrium, the overall Gibbs energy of the system must be minimized at every point over the entire temperature and composition range. Minimizing the Gibbs energy at every point then leads to equilibrium compositions at different conditions, the basis for the phase diagram. Tie lines and regions made up by them on the phase diagram represent two-phase equilibria resulting from the two phases having equal component chemical potentials at a given set of conditions. Use of a metalorganic precursor to deposit WNx films can lead to appreciable C contamination into the films, especially at higher deposition temperature, and can cause unintentional deposition of W-C-N solid solution or line compounds, along with free C in the film. The presence of C in our metalorganic precursor ligands and solvent makes the formation of a ternary W-C-N solid phase, rather than the binary WNx, highly probable. Recent reports, however, indicate interest in the FCC J3-WNxCy phase as a potential barrier material, as detailed in Chapter 2 [Ele03, Li02, Smi02]. While one

PAGE 119

110 benefit of depositing P-WNxCy films is lower resistivity than the binary nitride, the presence of free carbon in the films can increase film resistivity, negating this benefit. Moreover, as mentioned in Chapter 2, free C regions in the film were reported to be facile paths for Cu to penetrate the barrier [Ima97], due to carbon's inability to chemically repel Cu. An equilibrium phase diagram for the ternary W-C-N system at low temperature and pressure, consistent with MOCVD process conditions, would be a useful guide to select deposition conditions for P-WNxCy that minimize free carbon in the films. The elements contained in precursor, solvent, and carrier gas molecules used in this work are W, C, Cl, Hand N, with solid and gas phases being in contact during the CVD reaction, while purposely avoiding a liquid phase. Modeling of the experimental results from this work therefore requires inclusion and analysis of the W-C-N ternary system. While the experimental data from this work indicate the presence of ternary FCC P-WNxCy polycrystals in the films, the deposition mechanism that gives rise to them involves solid-gas equilibrium in the W-C-Cl-H-N system. This analysis must therefore model W-C-Cl-H-N solid-gas equilibrium, adjusting parameters for the ternary FCC P-WNxCy phase within this system to ensure its formation at our reactor conditions. From the results of this analysis, the equilibrium phase diagram for the ternary W-C-N system can then be predicted. Before the 5-component analysis can begin, models for the binary W-N and W-C systems must first be reviewed. Existing W-N and W-C phase diagrams, along with new literature data, will be analyzed to determine if the solution model parameters need to be revised. Once the binary analyses are complete, a degrees of freedom (DOF) analysis for the 5-component system will be done to determine the effect of varying specific conditions in our reactor on the predicted

PAGE 120

111 equilibrium. The ternary W-C-N phase diagram will then be predicted, and lastly, the B-WCo.sB-WNo s pseudobinary system will be discussed. 3.2 Constraints To perform an equilibrium analysis, several calculational constraints must be addressed. These include mandatory constraints that are independent of the particular system of interest, and specific constraints that define the parameters of a specific system. The mandatory constraints include conservation of elemental species in the system, and the non-negativity constraint, which dictates that the number of moles of any species in the system must be either positive or zero [Smi82]. The elemental species conservation constraint can be represented by a set of equations as follows: N Lakini = bk ;k = 1,2 ... E i=l (3-19) where aki is the subscript of the kth element in the molecular formula of component i, ni is the number of moles of component i, bk is the fixed number of moles of element k in the system, N is the number of components, and E is the number of elements. Specific constraints typically include temperature, pressure, and elemental moles, which must be given at the outset before the analysis can be done. The elemental species conservation constraints are linear in the mole numbers of the species present, while the chemical potential functions are not, meaning that any solution scheme will be iterative [Smi82]. 3.3 Degrees of Freedom The Gibbs phase rule defines the degrees of freedom for a system, F, as the number of independent variables required to describe the state of the system [DeH93]. For a non-reacting system, this is defined as F=C-P+2, where C is the number of components (in our case, elements) and Pis the number of phases. This relation is useful

PAGE 121

112 when examining a phase diagram or a system with a known number of phases. When running a thermodynamic analysis, however, the number of phases at a given set of conditions is typically an unknown, desired quantity. "True" degrees of freedom as defined by ThermoCalc [Sun95] is C+2, which represents the number of components (C) input to the system plus temperature and pressure. Finally, stoichiometric degrees of freedom (R) is the number of species (S, which includes molecules, line compounds, non-stoichiometric phases, gas species, etc.) being considered in an equilibrium analysis minus the number of components in the system, R=S-C [Smi82]. The value of R represents the number of independent chemical reactions relating the components in the system. Internal degrees of freedom are a separate issue, and do not affect the overall DOF for the system. In the FCC ~-WNxCy phase, for example, the values of x and y are internal DOF for the phase. Setting the values of x and y fixes the composition of the phase, but does not eliminate a DOF for the system. 3.4 Computational Methods Before starting a thermodynamic equilibrium analysis, the list of potential species in the solid, liquid, and gas phases must be assembled. In ThermoCalc, this list of species and their relevant thermodynamic data are assembled into a thermodynamic database (tdb) file. To begin the analysis, the "true" degrees of freedom must equal zero, hence the temperature, pressure, and elemental composition must be input to begin the calculation. Once these inputs are given, the value of R and the list of R independent chemical reactions among the specified elements in the tdb file must be determined. The general solution for the elemental species conservation equation above is of the form:

PAGE 122

113 R n = n~ + "v .. t I I L, ,J'-, J j = l i = 1,2 ... N (3-20) where ni is the final equilibrium mole number for species i, ni is the initial mole number of species i, and Vij and ~j are the stoichiometric coefficient and the extent of reaction parameter, respectively, for species i in the jth independent reaction among the species. After the list of R independent chemical reactions is determined, all values of Vij are known, and the optimization process focuses on solving for Ej. The computational method to solve for the equilibrium composition in a complex system has its origins in work done by Brinkley [1946] and White [1958]. This stoichiometric algorithm method is outlined in section 4.3.3 of Smith and Missen [1982]. ThermoCalc, the software used for this thermodynamic modeling, does this procedure automatically The sublattice model will be used to describe the solid phases in the W-C-Cl-H-N system. This model is phenomenological, and considers a given solid to be composed of separate, interlocking sublattices. Each sublattice has certain elements that may mix on it. Vacancies may also be added to each sublattice to model homogeneity ranges in the solid phases. This model is very useful to describe metallic systems with interstitial atoms, such as metal carbides and nitrides, where carbon and nitrogen occupy interstitial sites in the metal lattice The metal atoms are treated as one sublattice, while the interstitial carbon and nitrogen atoms are treated as a second interlocking sublattice. Since the sublattice model is a construct that uses different composition variables, the use of the sublattice model requires modifications to the Gibbs energy expression listed in Equation 3-16 above. A modified expression for the Gibbs energy of a solution phase, adapted for a two sublattice model (metal on one sublattice

PAGE 123

114 and interstitial atoms on the other) with mixing only taking place on the interstitial sublattice, is shown below: G
PAGE 124

115 3.5.1 Pure Components in the W-N System The pure, stable phase of tungsten at 298 K and 105 Pa is the body-centered cubic (BCC) form (also called a-W). This phase has a lattice parameter of 3.165 A [JCP88], a melting point of 3422C, density at 20C of 19.25 g/cm3 [Sha96] and a resistivity of 5.65 .Q-cm [Mur95]. Pure tungsten in face centered cubic (FCC) and hexagonally close packed (HCP) forms has been discussed in the literature, but both are considered to be metastable. A second element, such as interstitial C or N, is typically needed to stabilize the tungsten atoms into an FCC or HCP configuration. The stable phase of pure nitrogen at 298 K and 1 atm is N2 gas, which has a boiling point of 77.35 K. 3.5.2 Stable Solid Phases in the W-N System Several stable forms of solid tungsten nitride have been identified in the literature. Wriedt (1989] considered four solid phases to be stable in the W-N equilibrium system. These phases were BCC WNx (terminal solid solution), face-centered cubic (FCC) WNx (~-phase), simple hexagonal packed (SHP) WN (o-phase) and orthorhombic WN 2 (oRv-phase) [Gol67, Tak97a, Wri89]. Subsequent work excluded the latter phase in favor of a hexagonal close packed (HCP) WNx (Ott11) phase, which is referred to as HCP W 2 N when x = 0.5 [Fri99b, Hua97]. This is analogous to the HCP WCx phase, which was described as a random interstitial solution of carbon in HCP tungsten [Gus86]. Schematic diagrams of the reported stable solid phases in the W-N system are shown in Figure 3-1. The BCC WNx phase (Figure 3-1) is stable up to 3422C, and has a lattice parameter near 3.165 A The solubility of N in BCC WNx varies, as denoted by the subscript "x," and is reported to range from l.7x10-5 to 5x10-3 at. % N over the

PAGE 125

116 temperature range of 1200 to 2400C at 105 Pa pressure [Wri89]. Wriedt [1989] adopted the following equation from Jehn and Ettmayer [1980], which fits the concentration of nitrogen in BCC WNx to Sievert's Law (Equation 3-22): where PN2 is the nitrogen pressure (in Pa) and T is temperature (K). Since the composition dependence of BCC WNx on temperature and pressure has been thoroughly examined, the existing Gibbs energy expression for this phase [Gui93] will not be modified for this analysis. FCC ~-WNX HCPWNX SHPWN ---------... --Figure 3-1. Schematic of the reported stable solid phases in the W-N system (not to scale). Open circles are W atoms, while black circles are N atoms. The FCC B-WNx phase (Figure 3-1) has a defective NaCl structure, with W atoms at the FCC sites, N atoms and vacancies at the interstitial octahedral sites, and a lattice parameter of 4.126 A for x = 0.5 [Gol67, HonOO, JCP88, Tot71]. For the FCC P-WNo.s stoichiometry (used interchangeably with FCC P-W2N or FCC P-Wo 67No.33),

PAGE 126

117 half of the interstitial sites are vacant. The theoretical density for this phase and stoichiometry is 18.0 g/cm3 while experimentally determined densities range from 15.0 to 17.9 g/cm3 [Aff85, Bos91, Gal97, Sam80]. BCC WNx contains minimal nitrogen, which is required in appreciable amounts for this application to repel Cu diffusion through the film. Hence, FCC ~-WNx is the desired phase of tungsten nitride for its barrier potential. Furthermore, it has a relatively low bulk resistivity (-50 .Q-cm) [Lee93] despite considerable nitrogen content presumably because the metal atoms are arranged in the close-packed FCC configuration. Nitrogen atoms are assumed to locate randomly on the interstitial sublattice, as no evidence of N ordering on the interstitial sublattice has been reported. Due to a significant vacancy concentration on the interstitial sublattice, this phase has a substantial homogeneity range, which is denoted by the subscript x. A range of x values has been reported for the J3-WNx phase in the literature; this will be discussed in more detail in Section 3.7 below. The SHP o-WN phase (Figure 3-1) has a simple hexagonal packed structure similar to WC, with lattice parameters a=2.893 A and c=2.826 A [Sch54]. This phase is considered a stable stoichiometric line compound with formula WN [Gui93, Lee95, Sch54, She00a, Tot71, Zha99]. The calculated theoretical density for this phase is 16.0 g/cm3, while experimentally determined densities range from 12.1 to 15.9 g/cm3 [Sam80]. Experimental observations of this phase are discussed in more detail in Section 3.7 below. The hexagonal close-packed HCP WNx phase (Figure 3-1), referred to as the E phase, has lattice parameters a=2.89 A and c=22.85 A with a theoretical density of 12.0 g/cm3 for x=0.5 [Gui93, Pea67, Tot71]. An experimental density has not been reported

PAGE 127

118 for this phase, which contains a random interstitial solution of nitrogen in HCP tungsten For the HCP WN0 5 stoichiometry (HCP W 2N), half of the interstitial sites are vacant. Guillermet [1993] considered this phase to be metastable, but included it in their calculation of the W-N phase diagram. Huang [1997], however, considered this phase to be stable in her calculation of the Nb-W-C-N phase diagram, because the analogous HCP WCx phase is stable. A similar approach will be adopted for this analysis; hence this phase will be included in the assessment of the W-N phase diagram The previously reported Gibbs energy expression for this phase [Gui93] will not be modified, however, as no experimental data have been reported for this phase in the literature Several other tungsten nitride solid phases were identified but described as metastable [Wri89], hence they will not be included in this thermodynamic analysis. Liquid-solid phase equilibria data have not been reported for the W-N system. Guillermet (1993] modeled the W-N liquid phase using a substitutional solution model, which assumes that components can mix on any available position in the phase without any preferential site occupation (i.e., ideal mixing) [Sau98]. Additionally, the regular solution approximation was used, which assumes that interaction energy is independent of composition. 3.5.3 Liquid Solution in the W-N System The W-N liquid phase was previously modeled using a substitutional solution model [Fri99b Gui93 Hua97]. The regular solution approximation was applied, which assumes that component interactions are independent of composition. No experimental data have been reported for a W-N liquid phase but only inferred for liquid-solid phase

PAGE 128

119 equilibria data, hence the existing liquid model [Gui93] will be included in the present study without modification. 3.5.4 Gas Phase in the W-N System The gas phase species included in this analysis were N, N2, N3 and W, following previous work [Gui93]. The gas phase was treated as an ideal mixture, with Gibbs energies for the gas species taken from the SGTE database [Ans87]. 3.5.5 Equilibrium Modeling Basis Solid Phase Model Expressions describing the sublattice model are shown in Table 3-1, and are in the form (W)1 (N,Va)2 Tungsten (W) sits on its own sublattice, while nitrogen (N) and vacancies (Va) mix on the interstitial sublattice sites. The subscript on the interstitial sublattice, z, indicates the number of interstitial sites available per W atom, and depends on the structure of the metal (W) sublattice. For the BCC phase z = 3, while z = 1 for the FCC and SHP phases and 0.5 for the HCP phase [Gui93, Gus86]. Table 3-1. Sublattice Models for the Solid Phases in the W-N Binary System Solid Phase Sublattice Model Expression BCC solid solution (W)1(N,Va)3 FCC (y or r3) nitride (W)1(N,Va)1 HCP (E) nitride (W)1 (N, Va)o.s SHP (8) nitride (W)1CN)1 Use of the sublattice model to describe the BCC FCC, and HCP WNx solid solution phases requires modification of the standard expression for the Gibbs energy of mixing. This modification restricts mixing of N and Va to the interstitial sublattice, and assumes that all sites on the metal sublattice are filled. The modified expression for Gibbs energy of a solid solution phase, adapted for the two-sublattice model, is shown in Equation 3-23:

PAGE 129

120 Grp IGrp IGrp RT( 11 I 11 I) oxs m -YN W : N + Yva W:Va + Z YN nyN + Yva IlYva + mix (3-23) where G: is the Gibbs energy (per mole of formula units, where the formula contains l+z atoms) for the


PAGE 130

121 various solid phases of tungsten nitride (with all interstitial sites occupied by nitrogen), modeled with the ThermoCalc software by Guillermet [1993], are listed in Table 3-3. Table 3-2. Gibbs Energy Expressions (J/mol) for Pure Solid Phases of W [Gui93, Gus86]. Sublattice Gibbs Energy Expression Phase Model Layout T(K) 0G!~~. H~ER = GHSERW = -7646.311 + 130.4T 24.lTlnT .001936T 2 + 2.07x10 -7T3 + 44500T-1 p BCC 5 33x10 -II T4 + f V m (T, P)dP (for 298.15 :5 T :5 3695) (W)1(Va)3 0 a-W 0G!c~. H~ER = GHSERW = -82868.801 + 389.362335 T p 54TlnT + l.528621xl 033T-9 + fv m (T, P)dP 0 (for 3695 :5 T :5 6000) FCC (W)1(Va)1 0G~~a H~R = 19300 + 0.63T + GHSERW ~-W HCP (W)1(Va)o 5 0 G HCP HSER = 14750 + GHSERW E-W W:Va W SHP (W)1(Va)1 Not Reported &-W The !:J.. 0 term is the temperature dependent Gibbs energy of formation for BCC WN3, which is BCC WNx with all available interstitial sites (z=3 for BCC WNx) filled by AOGFCC AOGHCP AOGSHP N. The u w :N, u w : N o .5 and u w:N terms are defined similarly. HSER d w an H tER represent the enthalpy of W and N at the stable element reference (SER) conditions (298.15 Kand 105 Pa), i.e., solid BCC Wand N2 gas, respectively, at 298 .15 K and 105 Pa, with all subsequent Gibbs energy expressions referenced back to these pure component stable forms. For these elements in their stable reference state at 298.15 K, HSER HSER w and N are defined to be zero.

PAGE 131

122 Table 3-4 shows Gibbs energy models for the W-N solid solution phases. The first three solid phases are solid solutions, each with their own nitrogen homogeneity range on the interstitial sublattice. The fourth phase, SHP (8) nitride, is a line compound with a fixed nitrogen stoichiometry. Its Gibbs energy expression is listed in Table 3-3 rather than Table 3-4, because no mixing occurs on the interstitial sublattice in this phase. Table 3-3. General Gibbs Energy Expressions per mol of Formula for Solid W-N Phases. Phase BCCa-WNx FCC ~-WNx HCPE-WNx SHP8-WN Sublattice Model La out (W)1(N)3 (W)1CN)o.s (W)1(N)1 Gibbs Energy Expression oaBCC =HSER +3HSER +~oGBCC W : N W N WN3 0 GFCC = HSER + HSER + ~oaFCC WN W N WN oaH~P = HSER + 0.5HSER + ~oaHCP W N W N WN0 0 asHP HSER + HSER + ~oaSHP W : N W N WN Table 3-4. Mixing Gibbs Energy Expressions per mol of Formula for W-N Solid Solution Phases. Phase Sublattice Model Binary Gibbs Energy Expression Layout oaBCC OGBCC OGBCC BCC a-WNx (W)1(N,Va)3 m =yN W:N+Yva w :va+ 3RT[yN In YN + Yva 1n Yval +XS a:cc oaFCC OGFCC OGFCC FCC ~-WNx (W)1(N, Va)1 m = Y N W : N + Y Va W:Va + RT[yNlnyN + YvalnYval+xs G~cc oaHCP OGHCP OGHCP m = YN W : N +Yva W :Va + HCPE-WNx (W)1(N, Va)o.s 0.5RT[yNlnyN + YvalnYval+xs a:cp SHP8-WN (W)1(N, Va)1 -In the three solid solution phases, W occupies one sublattice, and does not undergo mixing, which is evident from a fixed W stoichiometry. The second sublattice contains a combination of N and Va. The Gibbs energy models for the three W-N solid solutions

PAGE 132

123 describe mixing of N and Va on the interstitial sublattice of metallic W. The different phase structures result from the variation in structure of the metallic W sublattice with temperature and N content. To account for the N homogeneity range, both N and Va are incorporated into the model for the interstitial sublattice. Table 3-5 shows the temperature dependent expressions for 0Gi: N and O L~: N v a reported by Guillermet [1993] The expressions are based on the regular solution model to address N-Va interactions on the interstitial sublattice, and the value for O L~~.va was assumed to be independent of temperature (i.e., strictly regular). A value for 0L~~.va was not reported. Table 3-5 Reported Gibbs Energy Parameters per mol of Formula for the Stable W-N Solid Phases [Gui93]. Sub lattice Phase Model Gibbs Energy (G) and Excess Gibbs Energy (L) Expressions Layout I!!,. 0ot ~ ~ = -268828.90 8 + 444.62002T BCC (W)1(Nh 72.7446506 Tln(T) + 1079965T 1 .025174346 8T2 a-WNx 0 L~. ~ Va = 83285 69 .977T I!!,. 00 = -40752.746 + 259.88468T -45.443666T lnT FCC (W)1(N)1 + 237083T 1 .004651912 3T 2 P-WNx 0L1;~,Va = -61824 HCP (W)1(N)o.s 1!!,.00~~~ = -12811.336 +186.68879T E-WNx -34.217887T ln(T) + 105601 T -t -.004509875 T2 SHP (W)1(N)1 I!!,. 00~~ = -66519.085 + 272.76021 T -45.64545Tl nT 6-WN + 392513TI .003530618 4T2 Note: above expressions are applicable over the temperature range 298.15g~6000 K.

PAGE 133

124 3.5.6 W-N Liquid Phase Model The Gibbs energy expression for the liquid phase in Equation 3-26 is based on the ideal mixing model, as no experimental data are available to assess any deviations from ideality for the liquid phase: (3-26) where XN is the mole fraction of N in the liquid phase and O Gtq and O G~ are temperature and pressure dependent Gibbs energies for pure liquid N and W, as reported previously [Din91,Gus85]. 3.5.7 W-N Gas Phase Model The molar Gibbs energy for each species in the gas phase is represented by Equation 3-27, which assumes ideal gas behavior and is both temperature and pressure dependent: 0 G ?as (T, P) HSE R =o G Gas (T) + RTlnP 1 m 1 m 1,m (3-27) where O G f: (T,P) is the Gibbs energy per mole of species i at temperature T and pressure P, and Hf!R is the enthalpy for the pure, stable form of species i at 298.15 K and 105 Pa. The O G ~= (T) term can be expanded as follows [Gus85]: 0 G ~= (T, P) H~!R =(a+ bT + cT In T + dT 2 + eT3 + gT-1 )i,m + RTlnP (3-28) An expression in the form of Equation 3-28 is generated for each species, and these are put into Equation 3-29 to get an overall Gibbs energy for the gas phase ( G ~as (T ,P) ): G~as (T, P) = I xi ( 0of: (T, P)Hf;R )+ RTI xi ln xi i i (3-29) where Xi is the mole fraction of species i in the gas phase.

PAGE 134

125 3.6 Previous Studies of the W-N Phase Diagram A phase diagram for the binary W-N system was previously reported, and is shown in Figure 3-2 [Gui93]. This diagram does not consider the presence of the gas phase at low pressure (:S 105 Pa). 4000 3500 3000 ,-... 2500 f3 s' 2000 1500 1000 500 I I I I I I I -,_ I~\~ \9. ,.,, !'-ti ,.,, Ill> 'pl 'F I I I I I ,-1~ \ '"0 '"' I I I I I I I I BCCWN, 2147 K I I I ,_ \~ I '"0 \ I'> I I I I I I I I I I I I I I I I I I ________ 109 Pa _________ Liquid 1264K t,() ~--~------x::.-=.===~tl) 556 K 0.0 0.1 0 2 0.3 0.4 0.5 mole fraction N 0 6 Figure 3-2. Binary phase diagram for the W-N system [Gui93] The gas phase was not included in this phase diagram. Adding the gas phase into Guillermet' s model while maintaining the pressure at 105 Pa (Figure 3-3), it is seen that the liquid phase is replaced by gas, and the FCC ~-WNx phase disappears. The formation of stable FCC ~-WNx in the presence of N2 gas has been reported, however, (Table 3-6), hence the model used to generate Figure 3-2 and 3-3 must be modified to account for this.

PAGE 135

4000 3500 3000 g 2500 t 2000 l 1500 1000 500 I BCCWNX SHPWN 0 0 0.2 126 Liquid+ N2 Gas N2Gas .. lllli I I 0.4 0.6 0.8 1.0 mole fraction N Figure 3-3. Binary phase diagram for the W-N system using Guillermet's model, with the gas phase included. In ddi h A oaFCC oLFCC d A oGSHP difi d. hi al a tion, t e u w : N w ,N.va an o w:N terms are mo 1e m t s an ys1s to reflect recently reported experimental data on solid-gas equilibria in the W-N system for the FCC and SHP solid phases. Existing Gibbs energy expressions for the BCC and HCP solid phases, as well as the liquid and gas phases, will not be modified, as no new experimental data have been reported for these. 3.7 W-N Optimization Results and Discussion FCC (3-WNx Stability Data As mentioned previously, the desired W-N solid phase for a barrier material is FCC B-WNx, but in the amorphous state. Various temperatures, pressures and deposition methods have been used to form the FCC B-WNx phase, and a range of x values in the

PAGE 136

127 resulting films has been reported. The temperature, pressure, and stoichiometry ranges reported in the literature must be reflected in the modified phase diagram. Many reports in the literature describe deposition of amorphous WNx films, with subsequent annealing at various temperatures and pressures leading to formation of one or more of the aforementioned W-N solid phases. In many cases, if the initial amorphous film was nitrogen rich (x in WNx > 0.5), FCC P-WNo s polycrystals formed with excess N migrating to the grain boundaries, which are disordered regions between the polycrystals in the films. Various techniques, including XPS [Nag93, Nak.87] and transmission electron diffraction (TED) [She00b], have been used to differentiate between N at the grain boundary and in the polycrystal. Since grain boundaries are inherently non-equilibrium regions in the film, only data that is available for polycrystalline materials will be used to optimize the phase diagram. Table 3-6 shows the representative temperature, pressure, and stoichiometry ranges for the polycrystalline FCC P-WNx phase, as reported in the literature, which were used in this assessment. Stable pressure and temperature ranges include conditions during deposition and during post-growth anneals. Operating outside of the stable temperature and pressure ranges for each of the respective references resulted in deposition of multiple phases or phases other than FCC As evident in Table 3-6 the upper and lower limits for x in P-WNx, which were used in this assessment, are 0.72 and 0.43, respectively. Schonberg [1954] first reported a stoichiometry range of 0.5:Sx:S0.72 for FCC P-WNx, formed by NH3 nitridation of Wat 800C. Although the pressure for this process was not specified, Banik et al. [1979] reported a pressure of lx108 Pa to form FCC P-WN05 by NH3 nitridation of W. The

PAGE 137

128 Table 3-6. Stable Existence Regions for FCC P-WNx Used in Assessment Stable FCC Value of x in Stable Pressure P-WNx Temperature Reference FCC P-WNx Range (K) Range (Pa) Deposition Method NIA (Assessment 0.43:Sx:S 0.72 556:sT:S2147* 105:SP:S109 of Lattice [Gui93] Parameter Data) x=0.5 473 7 Sputtering [Lin90] x=0.5 473
PAGE 138

129 to 109 Pa. Experimentally, the liquid phase is not observed at 105 Pa, as N prefers to exist in the gas phase as N 2 at high temperature, rather than forming a melt with W. Recent experimental studies have indicated that the stability range for formation of FCC P-WNx at low pressures (105 Pa and below) extends from 473 (200C) to 1123 K (850C) [She00b, Suh0l]. Hones et al. [2003] reported that the highest nitrogen content for FCC P-WNx was x=0 71, and increasing nitrogen content to xz0. 75 caused formation of a second phase, SHP &-WN. These reports indicate that the stability region for FCC P-WNx is larger than previously thought, and that previously accepted thermodynamic stability limits for this phase may have actually been kinetic limits on the formation of this phase due to experimental design. Previous thermodynamic assessments of the W-N system typically involved experimental formation of FCC P-WNx by NH3 (or N2) nitridation of BCC W at high temperature and pressure over long periods of time (typically several weeks) [Sch54, Ban79]. The formation of FCC P-WNx from BCC W and N2 gas was predicted to occur at 1100C and 108 Pa, with the high temperature and pressure required to dissociate N 2 [Ban79]. These processes required high temperature and pressure, owing to the need to break W-W bonds in the solid and N-H or N-N bonds in the gas before W-N bonds could form. After overcoming any strain energy in the solid and breaking the W-W bonds, nitride formation was then able to occur While formation of FCC P-WNx by these routes requires high temperature and pressure, this likely reflects a kinetic barrier to the NH3 (or N 2 ) nitridation process, rather than a thermodynamic stability limit. Newer deposition techniques such as sputtering PECVD, LPCVD, MOCVD and ion implantation have established low temperature and low-pressure kinetic routes to the formation of FCC P-WNx Sputtering, for example

PAGE 139

130 uses plasma energy to break the strong bond in gaseous N2 liberating atomic nitrogen, which is then free to react with W atoms emanating from the sputtered target to form tungsten nitride films. Sputtered films containing the FCC ~-WNx phase have been deposited at a temperature as low as 200C [SheOOb], and are reported to remain intact after annealing at temperature to 850C under UHV conditions (6.6x10-5 Pa) [SheOOb, Suh99, Suh0l]. Recent reports use high temperature vacuum anneals (typically 30 min) to establish temperature stability limits for the deposited phases. Although annealing for a longer time (days or weeks) would give more confidence that equilibrium existed, the data are only available for these short anneals. Hence, these high temperature anneals will be assumed to approximate equilibrium to assess the system at low pressure. Considering available experimental data in the literature, the most robust low-pressure phase diagram should contain an FCC ~-WNx solid solution with a temperature range of 473 to 1123 K (200 to 850C), a composition range 0.43:Sxg).72, and be stable at a total system pressure of 6.6x10 -5 Pa. Several other reports on the stoichiometry and temperature stability of ~-WNx phase were considered but excluded from this analysis. In addition to the stoichiometry range reported in Table 3-6, Hones et al. [2003] also reported the deposition of overstoichiometric (x>l) FCC ~-WNx films where additional N incorporated into tetrahedral holes in the interstitial sublattice. Substoichiometric films had increasing hardness with x, while hardness dropped for overstoichiometric films, which were partially delaminated from their substrates. The overstoichiometric films are therefore considered to be metastable and are not included in this assessment. One report of MOCVD grown FCC ~-WN1.o was given based on an XRD lattice parameter of 4.154 A

PAGE 140

131 and composition ratios from WDS [Chi93]. Reported AES concentrations were 12 at. % for both C and N, which did not agree with WDS values [Chi92]. The presence of C in these films suggests that formation of a B-WNxCy solid solution, where C intermixes with N on the polycrystal's interstitial sublattice [Hua97, Jon93], may have occurred. Incorporation of C on the interstitial sublattice expands the polycrystal lattice, resulting in a larger lattice parameter for B-WNxCy films relative to B-WNx, assuming a fixed number of occupied interstitial sites. Their reported lattice parameter expansion, interpreted as a binary B-WN1.o stoichiometry, may have actually been due to the formation of a ternary B-WNxCy solid solution, hence this data point will be excluded from the phase diagram analysis. A report of FCC B-WNo s deposition at 150C and 27 Pa has also been given [Bai03]. The film was deposited by PECVD, using WF6 NH3 and H2 on a SiO 2 substrate. Their XRD spectrum, however, suggests formation of W 30 rather than FCC B-W2N. The formation of this oxide was likely due to the presence of oxygen in the SiO2 substrate. Another report on PECVD grown B-WNx films indicated an upper stability limit of 1000C after annealing for 60 sec in a N 2 atmosphere [GalOO]. The annealing pressure was not specified, however, so this data point was omitted. Another report gave 1050C as the upper stability limit for FCC B-WN0 5 [Vie02]. While an XRD spectrum indicated peaks consistent with FCC B-WN0 5 after heating to 1050C, the heating process was done in a differential scanning calorimeter (DSC) under an N 2 + H2 atmosphere, at an unreported pressure. Since this temperature limit may reflect enhanced stability of FCC B-WNo.s due to the pressure of the N 2 + H 2 atmosphere, this data point will also be excluded from the analysis.

PAGE 141

132 SHP 6-WN Stability Data Table 3-7 shows representative temperature, pressure, and stoichiometry ranges for the polycrystalline SHP 6---WN phase, as reported in the literature, which were used in this assessment. Stable pressure and temperature ranges include conditions during growth and annealing. Lee [1995] formed the SHP o-WN phase by PECVD of WF6+NH3+H2 XRD results for their as-deposited films indicated SHP WN in equilibrium with FCC P-WN0 5 and P-W, but the SHP o-WN peak disappeared after annealing at 500C and 6.6x101 Pa. Shen et al. [2000a] also reported formation of SHP o-WN, with subsequent decomposition to P-WNo.s + N2 after annealing at 550C at 6.6x10 -5 Pa. Considering available information in the literature, the most robust low-pressure phase diagram should contain the SHP o-WN phase with an upper temperature limit of 823 K (550 C) at a total system pressure of 6.6x10-5 Pa. Table 3-7. Stable Existence Regions for SHP 8-WN Used for Assessment. Stable Stable Pressure Range SHP6---WN Temperature Deposition Reference Range (K) (Pa) Method O:sT:51264* 105:SP:5109 NIA [Gui93] (assessment) 473 7 Sputtering [Lin90] 500 20 6.6 xl0-1 Sputtering [Hon03] :5773 6.6 x101 PECVD [Lee95] WF6+NH3+H2 :5773 2 7 xl0-1 Sputtering [Uek96] :5823 6.6x10 -':SP:S7.9xl0-1 Sputtering [SheOOa] *Note that the gas phase was not included in the equilibrium calculation for this system at low pressure (:S 105 Pa). Several other reports on the stoichiometry and temperature stability of 6---WN phase were considered but excluded from this analysis. Schonberg [1954] reported the formation of SHP 6---WN in equilibrium with BCC W and FCC P-WN0 5 by NH3

PAGE 142

133 nitridation of W at 800C, but the pressure for this process was not specified. Schonberg also reported rapid decomposition of the SHP 8-WN phase during a 600C anneal in "vacuum." Banik et al. [1979] reported a pressure of lxl08 Pa to form FCC J3-WNo. s by NH3 nitridation of W. The pressure used by Schonberg is therefore assumed to have been substantially higher than atmospheric; hence this data point will be excluded in this low-pressure assessment. Zhang et al. (1999] reported the formation of SHP 8-WN by N ion implantation into a W surface at 3.4x10-3 Pa, but the substrate temperature was not specified. One report of 8-WN deposition by sputtering onto a room temperature (T=20C) GaAs substrate was also given [Gal93]. This report relied solely on SIMS data for the deposited film, and inferred the crystal structure based on WIN ratio. Since XRD analysis was not provided for this film, the data point will be excluded from this analysis. Assessed W-N Phase Diagram As mentioned above, the new literature data indicates that the FCC J3-WNx solid solution should have a temperature range of 473 to 1123 K (200 to 850C), a composition range 0.43:Sx:S0.72, and be stable at a total system pressure of 6.6x10-5 Pa. Moreover, the SHP cS-WN phase should have an upper temperature limit of 823 K (550C) at a total system pressure of 6.6x10 -5 Pa. The values for ~0G~~, 0L';, ~ ,va and ~0G~1: terms were modified through use of ThermoCalc' s Parrott optimization module to reflect the new data. Figure 3-4 shows the new W-N solid-gas phase diagram based on the assessment of the data listed in Table 3-6 and 3-7, at the lowest reported stable pressure for the FCC and SHP phases. The left side of the diagram contains a very narrow existence region for BCC WNx between O and 3695 K, which is not visible. Above 3695

PAGE 143

, 134 K, the left side of the diagram contains a narrow region representing a liquid W-N solution, which again is not visible. The right side of the diagram is N 2 gas over the entire temperature range. The FCC ~-WNx phase, which did not previously appear in equilibrium with the gas phase (Figure 3-3), now appears with a finite homogeneity range. The FCC B-WNo s phase/stoichiometry exists in equilibrium with BCC WNx and SHP WN at 473 K. The horizontal line at 1123 K represents a three-phase equilibrium 4000 ._ W-NLiquid 3695K N2Gas _. 3500 3000 ,-... 2500 '-' N2 Gas e 2000 k"' BCCWNX t FCC f3-WNX II) 1500 E-I 1123 K 1000 I I~ 823K 500 473K \/ 0 Cl) 0.0 0.2 0.4 0.6 0 8 1.0 mole fraction N Figure 3-4. Binary phase diagram for the W-N system at 6.6x10-5 Pa, using new data with the gas phase included. between BCC WNx, FCC B-WNx, and N2 gas. FCC P-WNo.s is reported to decompose under vacuum to BCC Wand N2 gas at 1123 K (850C) [Suh99, SuhOl], but a small concentration of dissolved N likely remains in the BCC W, as indicated by the BCC WNx phase on the left side of the diagram. The SHP 8-WN phase is known to decompose to FCC B-WNo s and N2 gas at high temperature under vacuum [Gol67, SheOOd], and this is

PAGE 144

135 indicated on the new diagram. The upper temperature limit for the SHP &-WN phase, previously 651 K (Figure 3-3), is now 823 K, reflecting enhanced stability for this phase per recent experimental reports. 4000 ---------------~ 3500 3000 2500 '-' W-NLiquid 2000 I ... BCCWNX E-o 1500 1000 3663 K 2325 K 1046K Cl) 0 +----~--~---'--.....---0.0 0.2 0.4 0.6 0.8 mole fraction N 1.0 Figure 3-5. Binary phase diagram for the W-N system at 105 Pa, using new data with the gas phase included. Increasing pressure in the W-N solid-gas system from 6.6x10 -5 to 105 Pa causes an expansion in the stability range of FCC P-WNx (Figure 3-5). The upper temperature limit for this phase increases from 1123 K at 6.6x10-5 Pa to 2325 K at 105 Pa. Interestingly, this upper temperature limit is close to 2147 K, the upper limit reported for the FCC P-WNx phase in the solid-liquid system at 105 Pa [Gui93]. In addition to the increased upper temperature limit, the upper solubility limit of N in the FCC P-WNx phase increases with increased pressure, increasing from x=0.72 to x=0.92, which is approaching the stoichiometry for SHP &-WN formation. The upper temperature limit

PAGE 145

136 for the SHP 6-WN phase also increased with pressure, rising from 823 Kat 6.6x10-5 Pa to 1046 K at 105 Pa. The new expressions for I!:,. 0GFCC oecc and I!:,. 0G8HP generated by this W:N W N,Va W : N assessment are listed in Table 3-8. To achieve an adequate fit for the model to the new experimental data, the parameter describing N-Va interactions on the FCC interstitial sublattice was given a temperature dependent expression, whereas this was previously assigned a constant value by Guillermet [1993]. Table 3-8. Modified W-N Gibbs energy expressions (J/mol of formula) generated by this work. Sub lattice Phase Model Gibbs Energy (G) and Interaction Parameter (L) Layout FCC~-WNx (W)1CN)1 = -124063.4-100.74T 0 L~~.va = -68362.39 9.01 T SHP6-WN (W)1(N)1 !!:,.0G~~ =-163975.23-63.38T Note: the above expressions are applicable over a temperature range of 298.15ST:s6000. If Equation 3-23 is written for the FCC phase, Equation 3-30 results: 0FCC=y I 0FCC +y 1 0 FCC +RT(y Ilny l+y Ilny I)+y Iy IoLFCC (3-30) m N W:N Va W:Va N N Va Va N Va W:N,Va To obtain the Gibbs energy for the FCC ~-WNo.s stoichiometry, for example, substitute YN1 = yv} = 0.5, the expression for G~~a from Table 3-2, and the expressions for and 0LFcc from Table 3-8. Simplifying gives Equation 3-31: G~CC = -73295.45 + 7.13T -12. 05TlnT -9.68x10 -4T2 + l.04x10 -7T3 + 22250 2.67x10 -11T4 T (3-31)

PAGE 146

137 This value is in J/mol of formula for FCC B-WNo.s (1 mole W + 0.5 mol N) To convert this Gibbs energy to J/mol of formula units for B-W2 N, this value must be multiplied by 2 A comparison of the enthalpies of formation for the FCC and SHP phases reported in the literature against those generated by this assessment is shown in Table 3-9 Lakhtin [1979] and Neumann [1934] gave Aflf values for FCC B-W2 N i n units of J/mol of formula (i.e 2 W atoms 1 N atom). Their values were divided by 2 to convert them to a J/mol of formula for FCC B-WNo s The new L\0Hf (298) values for FCC B-WN0 5 and SHP B-WN generated by this work, as shown in Table 3-9, indicate greater stability of these phases than previously reported. This reflects the ability of new deposition techniques to overcome kinetic barriers to form these phases, as discussed above. Table 3 9 Comparison of reported and calculated heats of formation (L\ 0H f ) for FCC B-WNo s and SHP B-WN phases. Phase & L\ 0H f (298) kJ/mol of Basis Reference Formula formula -18. 9 Assessment [Gui93] -33.0 Ab Initio [deB88] Calculation -36.0 Assessment [Lak79] FCC B-WNo s Calculation based -35.6 on experimental [Neu34] data of non-W metal nitrides -69.5 Assessment Thi s Work -15.0 Ab Initio [d e B88] Calculation SHPB-WN -24.0 Assessment [Gui93] -82.0 Assessment This Work

PAGE 147

138 Figure 3-6 plots the trend of Di. 0Ht{298) for nitrides of neighboring metals on the Periodic Table. The slope of the line connecting Di.0Ht{298) for ZrN, NbN and MoN is 158, while the slope for the line connecting HfN, TaN and our value for ~-WNo.5 is 150. Table 3-10. Enthalpies of Formation for Several Neighboring FCC Nitrides Compound FCCZrN FCCNbN FCCMoN FCCHfN FCCTaN FCCWNo. 5 FCCWNo.5 FCCWNo.5 Atomic Number Di. 0Ht (298), kJ per mole of ,__ 8 8 <'
PAGE 148

139 neighboring nitrides, it would appear that our new 1!J,. 0Ht{298) may be closer to the actual value than that suggested by Neumann [1934]. 3.8 Stable Solid Phases in the W-C System The same solid phase structures discussed previously for the W and the W-N system also apply to the W-C system, with C replacing N on the interstitial sites in the metal sublattice. The pure, stable form of Cat 298.15 Kand 105 Pa is solid graphite. The same solid phases described for the binary nitride exist for the binary carbide. The Gibbs energy model describing these phases is given in Table 3-11. The general Gibbs energy expressions for the solid W-C phases are given in Table 3-12, while the specific expressions for the W-C solid solution phases are given in Table 3-13. Gibbs energy parameters reported in the literature for the W-C solid phases are given in Table 3-14. The BCC WCx phase is stable at all temperatures up to the solidus (3422C), and has a lattice parameter of 3.158 A [Gol63]. The reported solubility of C in BCC WCx, based on experimental measurements, ranges from 0.05 at. % at 2273 K (2000C) to a maximum of 0.7 at.% at the eutectic temperature, 2983 K (2710C) [Geb66 Gol63] The relationship between temperature and carbon solubility in the BCC WCx phase was reported as: Inc = 4.67 15x103 max T(K) (3-32) where Cmax is maximum concentration in at. % and T is temperature in K. The FCC ~-WCx phase (also called ~-WCi-x) has a defective NaCl structure with a lattice parameter of 4 .2 36 A for ~-WCo.6 [Gus86, JCP88, Rud67 Tot71]. Toth [1971] reported this phase with 38 at.% C (x in ~-WCx is 0.61), while Gustafson [1986]

PAGE 149

140 reported C content ranging from 37 to 39 at. % (0.59:S x :S0.64). Pauleau [1992] found the maximum C content in B-WCx to be 40 at. % (x = 0.67), while Sara [1965] reported this phase with an "approximate" maximum of 41 at.% C (x=0.70). Overall, this phase has a reported C content ranging from 38 to 41 at. %, corresponding to a homogeneity range in B-WCx of 0.60 :S x :S 0.70 [Gus86, JCP88, Rud67, Sar65, Tot71] The SHP 8-WC phase has a hexagonal structure with lattice parameters a=2.906 A and c=2.837 A [Tot71] Gustafson [1986] considered this phase to be a stoichiometric line compound with formula WC, as did Toth [1971] Table 3-11. Solid Phases Considered in the W-C Binary Model Condensed Phase Sublattice Model Layout BCC solid solution (W)1(C,Va)3 FCC (y or B) carbide (W)1(C,Va)1 HCP ( E) carbide (W)1(C,Va)o.s SHP (8) carbide (W)1(C)1 Graphite (C)1 Table 3-12. -C General Gibbs Energy Expressions (J/mol of formula) for Solid W Phases. Phase Sublattice Model Gibbs Energy Expression Layout BCC a.-WCx (W)1(C)3 FCC B-WCx (W)1(C)1 HCPE-WCx (W)1(C)o.s SHPo-WC (W)1(C)1 oGBCC HSER + 3HSER + oGB W : C W C oaFCC HSER + HSER + oaFC C C W : C W C W 0 G = H~R + 0.5H~ER + 0G HCP WCo s 0 GSHP HSER + HSER + oas W : C W C W HP C While the HCP WCx phase is considered stable, three separate forms of this phase the low have been reported The high temperature form exists above 2750 K, temperature phase exists below 2350 K, and the third phase exists betw een these temperatures [Gus86]. Gustafson [1986] treated this as a single phase con random interstitial solution of carbon in HCP tungsten, with a stoichiometry taining a range of

PAGE 150

141 0.37 :'.S x :'.S 0.50. A similar approach will be taken here. Lattice parameters near a=3.00 A and c=4.73 A have been reported for the HCP W2C stoichiometry [JCP88, Gus86, Tot71]. Huang [1997] and Frisk [1999b] considered this phase to be stable for calculation of the Nb-W-C-N and Ta-W-C-N phase diagrams, respectively. Table 3-13. Mixing Gibbs Energy Expressions (J/mol of formula) for W-C Solid Solution Phases. Phase Sublattice Model Binary Gibbs Energy Expression Layout oaBee oaaee oaaee BCCWCx (W)1(C,Va)3 m =ye W :e+Yva W:va+ 3RT[yelnYe + YvalnYval+xs G!ee oaFee oaFee oaFee FCC P-WCx (W)1(C, Va)1 m = Ye W : e +Yva W :Va + RT[ye 1n Ye+ Yva 1n Yval +xs G~ee oaHeP oaHeP OGHCP HCP WCx (W)1(C, Va)o s m = Ye W : e + Y Va W :Va + 0.SRT[ye 1n Ye+ Yva 1n Yval +XS a:eP SHP6-WC (W)1(C, Va)1 -Table 3-14. Reported Gibbs energy parameters (J/mol of formula) for the stable W-C solid phases [Gus86]. Sub lattice Gibbs Energy (G) and Excess Gibbs Energy (L) Phase Model Layout Expressions BCCWCx (W)1(C)3 L\ 0G~~g = 3.751xl05 -35.87T FCC (W)1(C)1 L\ 00 = -17864 + 1.875 T P-WCx 0L~~.Va = -17.98T HCPWCx (W)1(C)o s L\ 00 = -13430 -13.3T 0L~~.v a = 8330 SHP6-WC (W)1(C)1 L\ 0G~~ = -58192 + 287.16T -46.48TlnT .0021435T 2 -9.2217x10 s T 3 + 590500T i Note: above expressions are applicable over temperature range of 298.15:'.ST:'.S:6000 K. 3.8.1 Liquid Solution in the W-C System The W-C liquid phase was previously modeled using a substitutional solution model [Fri99b, Gus86, Hua97]. The regular solution approximation was applied, which

PAGE 151

142 assumes that component interactions are independent of composition. No experimental data for a W-C liquid phase have been generated by our work, hence the existing liquid model will be included in the present study without modification. 3.8.2 Gas Phase in the W-C System The gas phase model for the W-C system is constructed in the same manner as that for the W-N system. The gas phase species included in this analysis were Ci, C2, C3, C4, C5 and W, following previous work [Jon96b]. The gas phase was treated as an ideal mixture, with Gibbs energies for the gas phase species taken from the SGTE database [Ans87]. 3.8.3 W-C Liquid Solution Model The Gibbs energy for the liquid phase is based on the ideal mixing model, as no experimental data are available to assess any deviations from ideality for this phase: (3-33) where Xe is the mole fraction of C in the liquid phase and 0G~q and 0G~ are temperature and pressure dependent Gibbs energies for the pure liquid C and W phases, as reported previously [Din91, Gus85]. 3.9 Previous Study of the W-C Phase Model The reported binary W-C diagram, which was modeled with ThermoCalc and assessed against experimental data, is shown in Figure 3-8 [Gus86]. The HCP, SHP and FCC phases are all evident in the diagram. While a small homogeneity range is evident for FCC ~-WCx between 2800 and 3000 K, this phase does not appear at the low temperature end of the phase diagram.

PAGE 152

143 5000~----------------, liquid 4000 0 3000 -U---~.:::7r""lcr.L:=r-------, WC + graphite 2000 1000 +------.------.---~----.-------r------1 0 0 0.2 0.4 0 6 0 8 1.0 Mole Fraction Carbon Figure 3-7. Binary phase diagram for the W-C system [Gus86]. The B-WCx phase was labeled "WC1-x" in the original diagram by Gustafson. As shown in Table 3-15, however, recent experimental results indicate the existence of a low-temperature stability region for FCC B-WCx. Qin et al. [1998] found B-WCx composition to vary from B-WCo.2 (B-Wo.s3Co.l7) to B-WCo.6 (B-Wo.62Co.Js) for films deposited by PECVD of W(CO) 6 + H 2 at 330C. Carim et al. [1989] formed the B-WCx phase at 4x10-3 Torr by sputtering alternating layers of Wand C onto a Si (111) substrate. The bulk sample temperature was -150C, but the temperature at the surface was not measured, and may have been slightly higher. Moreover, Pauleau and Gouy-Pailler [1992] reported formation of the B-WCx phase on polished stainless steel substrates by reactive sputtering of a tungsten target in an Ar/C~ gas mixture. The substrate temperature ranged from 200 to 250C. Srivastava et al. [1984] also reported deposition of B-WCx films by RF magnetron sputtering. These films were deposited on stainless steel substrates at 400 to 500C and a pressure of 2x10 2 Torr. Reniers [1996]

PAGE 153

144 reported formation of J3-WCx by reactive sputtering of an Ar/C~ gas mixture onto W foil. Vieira [2002] reported formation of J3-WCx by sputtering of a W-C target. This phase decomposed at 845C to form SHP WC+ HCP W 2 C, although the pressure for this transformation was not given. One report, which was not included in Table 3-15, gave an XRD spectrum with a broad peak near 20 = 38 [SunOlb]. While they reported deposition of SHP WC and HCP WC0 5 the peak location actually suggests FCC WCx deposition. Overall, these reports indicate the possibility of J3-WCx formation at much lower temperature than reported in the existing phase diagram [Gus86], which indicates the ability of sputtering, as a "kinetic helper," to form these phases at low temperature. This indicates a need for reassessment of the low temperature end of the W-C binary phase diagram; the currently available data, however, are inadequate to establish the temperature, pressure and composition boundaries of the FCC J3-WCx phase. Table 3-15. Stable Low Temperature Existence Regions for FCC J3-WCx Value of x in Stable Stable Pressure FCC FCC J3-WCx Temperature Range (Pa) J3-WCx Deposition Reference Range (K) Method 0.20 X 0.60 603 NR PECVD [Qin98] W(C0)6+H2 X = 0.67 473-523 NR Sputtering of W in [Pau92] C~Ar NR -423 0.53 Sputtering of W and [Car89] graphite NR 673-773 2.7 Sputtering of W in [Sri84] C2Hi/Ar NR 1118 NR Sputtering of W-C [Vie02] target NR NR NR Sputtering of W in [Ren96] C~Ar NR = not reported

PAGE 154

145 3.10 Equilibrium in the W-C-Cl-H-N System To model equilibrium in our reactor, all of the stable solid, liquid and gas phase species in the W-C-Cl-H-N system were included. Stable solid phases in the binary W-N and W-C systems were discussed above, but a C-N phase diagram, which would address the third binary side of the ternary W-C-N phase diagram, is unavailable. While carbon nitride (CNx) deposition has been reported, these phases are typically amorphous and/or metastable [JelOO, Nin02]. Hence, carbon nitride solid phases will be excluded from this analysis in favor of tungsten carbides, nitrides and carbonitrides with various lattice structures. In addition, other potential solid phases formed by addition of Cl and H to the ternary W-C-N subsystem will be included. While some CVD equilibrium models exclude the liquid phase from consideration due its presumed absence during reaction, all potential liquid and gas phase species from the Sub94 database will be included in this analysis for completeness. 3.10.1 Stable Solid Phases in the W-C-Cl-H-N System Thermodynamic properties for several solid phases, including tungsten chlorides of different stoichiometries and ammonium chloride (~Cl), were taken from the Sub94 database in ThermoCalc. Gibbs energy expressions for pure, solid W, graphite and diamond-like C phases were taken from the SGTE database [Din91]. In addition to these phases, several stable forms of tungsten nitride and carbide, as discussed above, as well as tungsten carbonitride, have been reported in the _literature. The solid species included in this analysis are listed in Table 3-16. As shown in Table 3-16, no evidence of new ternary W-C-N phases has been given, but mixing of C and N on the interstitial

PAGE 155

146 sublattices of the previously described W-N and W-C solid phases (Figure 3-1) has been reported [Fri99b, Hua97, Jon93, 96b ]. Implicit in the description of the stable HCP Table 3-16. Solid Phases Considered in the Model for W-C-Cl-H-N Solid Phase Data Source ~Cl_l [Lid85], [Sub 94 Database] ~Cl 2 [Lid85], [Sub 94 Database] WCh [Sub 94 Database] WCI] [Sub 94 Database] WCl4 [Sub 94 Database] WCls [Sub 94 Database] WC16 (al) [Sub 94 Database] WC16 (a2) [Sub 94 Database] WC16 (f3) [Sub 94 Database] Graphite [Din91] Diamond-like carbon [Din91] BCCa-W [Din91] FCC f3-W [Gus85] HCPE-W [Gus85] BCCa-WNx [Gui93] FCC J3-WNx This Work HCPc-WNx [Gui93] SHPo-WN This Work BCCa-WCx [Gus86] FCC J3-WCx [Gus86] HCPc-WCx [Gus86] SHPo-WC [Gus86] BCC a-WNxCy [Fri99b Hua97, Jon93, 96b] FCC f3-WNxCv [Fri99b, Hua97, Jon93, 96b] HCP -WN xCv [Fri99b Hua97, Jon93 96b ] SHP 0-WNxCl-x [Fri99b, Hua97, Jon93, 96b] carbonitride phase is the assumption that the HCP WNx phase is considered to be stable due to the stability of its HCP carbide analog. Jonsson [1993, 1996b] first predicted phase equilibrium in the W-C-N system, although no experimental data were available for assessment. Two W-C-N phase diagrams were generated at 2873 K; one excluded the gas phase from consideration, while the other included the gas phase at 1000 bar pressure. Both diagrams indicated the presence of ternary SHP WNxCi x and HCP

PAGE 156

147 WNxCy phase fields, but neither indicated an FCC ~-WNxCy phase field. None of these models included C-N or C-N-Va interaction parameters, but instead assumed ideal mixing on the interstitial sublattice. Several quaternary thermodynamic analyses involving the W-C-N subsys tem have also been performed, including the Ti-W-C-N, Nb-W-C-N and Ta-W-C-N systems [Fri99b, Hua97, Jon93, 96b]. These reports contained phase diagrams that included the BCC, FCC, SHP and HCP metal carbonitride phases. Jonsson, Huang and Frisk each described the metal-W-C-N quaternary system using the two sublattice model, with metal and W mixing on one sublattice and C, N and vacancies (where applicable) mixing ideally on a second, interstitial sublattice. To date interaction parameters describing C-N or C-N-Va interactions on the W interstitial sublattice have not been reported. As detailed in Chapter 5, our experimentally deposited films contain a mixture of FCC ~-WNxCy and amorphous carbon, but do not contain the BCC SHP or HCP carbonitride phases. Oxygen in the films is due to post-growth in-diffusion, and is therefore excluded from this equilibrium analysis. Our experimental data will therefore be used to estimate Gibbs energy and interaction parameter values for the FCC (W)(N,C,Va) phase while the published Gibbs energy values for the BCC, SHP and HCP carbonitride phases will not be modified. To examine the thermodynamic stability of ~-WNxCy the two-sublattice model will again be invoked The metal sublattice contains W atoms wh i le the interstitial sublattice contains a mixture of N C and Va. This model is represented by the notation (W)1(N C Va)z, where z is the number of available s i tes in the interstitial sublattice per W atom. For the FCC carbonitride the

PAGE 157

148 model talces the form (W)1(N,C,Va)1. The FCC B-WNx and FCC B-WCx phases are both reported to have a defective NaCl-type structure, with vacancies intermixing on the interstitial sublattice. For these phases, the most commonly reported stoichiometries are FCC B-WN0 5 and FCC B-WCo.6, respectively, indicating that some of the interstitial sites are vacant. A range of x values have been reported for both of these phases, as indicated in Table 3-6 and Table 3-15. The lattice parameter of the B-WNxCy polycrystals is highly dependent on the composition of the interstitial sublattice. Experimentally determined lattice parameters, along with the concepts of fractional site occupation and Vegard's law may be used to estimate the composition of these polycrystals [Kos93] Zhang [1994] used lattice parameter data to determine the content of C and N on the interstitial sublattice in ternary TiNxCi-x polcrystals, which had a rock salt (NaCl) structure. In addition, Vegard's Law was used to estimate the C and N content of carbonitrides with the general form (Nb,Ti,V)(C,N) [InoOl] and Zr(C,N) [Bin95]. Interaction parameters and phase equilbria generated with the Vegard's Law approximation were in good agreement with experimental data. 3.10.2 Liquid Solution in the W-C-CI-H-N System Thermodynamic properties for pure liquid W, C and N were talcen from the SGTE database [Ans87, Din91]. Thermodynamic properties for several other species in the liquid phase, including CCl4, C6HsCl, ~Cl and the tungsten chlorides, were talcen from ThermoCalc's Sub94 database. The W-N and W-C liquid phases were previously modeled using a substitutional solution model [Gui93, Gus86] and no new ternary W-C-N liquid species have been reported. The regular solution approximation, which

PAGE 158

149 assumes that component interactions are independent of composition, was applied for the W-N and W-C liquid phases. The substitutional, regular solution model will be extended to model the liquid phase in our system. The liquid phase species included in this analysis are summarized in Table 3-17. Table 3-17. Liquid Phase Species Considered in the W-C-Cl-H-N Model Liquid Phase Data Source CCl4 ThermoCalc Sub 94 Database C6HsCl ThermoCalc Sub 94 Database ~Cl ThermoCalc Sub 94 Database WCls ThermoCalc Sub 94 Database WC16 ThermoCalc Sub 94 Database C [Ans87, Din91] N [ Ans87, Din91] w [Ans87, Din91] 3.10.3 Gas Phase in the W-C-Cl-H-N System All of the 110 potential gas phase species in ThermoCalc's Sub94 database were included in the W-C-Cl-H-N calculation, but a truncated list of the major gas species is Table 3-18. Selected Gas Phase Species Considered in the W-C-Cl-H-N Model Species Data Source c~ Sub 94 Database N2 Sub 94 Database NH3 Sub 94 Database HCl Sub 94 Database H 2 Sub 94 Database C2H2 Sub 94 Database C2~ Sub 94 Database C2~ Sub 94 Database C3~ Sub 94 Database c6~ Sub 94 Database C~s Sub 94 Database HCN Sub 94 Database C_G Sub 94 Database NG Sub 94 Database WG Sub 94 Database

PAGE 159

150 shown in Table 3-18 for brevity. The gas phase was treated as an ideal mixture, with Gibbs energies for the gas species taken from the SGTE database [Ans87]. 3.10.4 W-C-CI-H-N Solid Phase Model Dinsdale [1991] reviewed the thermodynamic data available in the literature for the condensed phases of pure W, C, and N. The Gibbs energy expressions therein are valid over a temperature range of 298.15:sT:S:6000 unless noted otherwise. Gibbs energy expressions for NH.iCl and the various stoichiometries of tungsten chloride were taken from ThermoCalc's Sub94 database. The two-sublattice model was used to describe the solid phases in the W-C-N system. Expressions describing the sublattice model are shown in Table 3-19 below, and are in the form (W)1(N,C,Va) z [Hua97]. W sits on its own sublattice, while N, C and Va mix on the interstitial sublatt i ce sites. The subscript on the interstitial sublattice, z is again the number of interstitial sites available per W atom. Table 3-19. Sublattice Models for the Ternary Solid Phases in the W-C-N System Solid Phase Sublattice Model Expression BCC a-WNxCv (W)1(N,C,Va)J FCC 8-WNxCv (W)1 (N ,C, Va)1 HCP E-WNxCv (W)1(N,C,Va)o.s SHPWNxC1-x (W)1(N ,C)1 The Gibbs energy models for the three W-C-N solid solutions describe mixing of N, C and Va on the interstitial sublattice of metallic W. The different phase structures result from the variation in structure of the metallic W sublattice with temperature and nonmetal content. To account for the homogeneity range vacancies are again incorporated into the model for the interstitial sublattice Use of the sublattice model to describe the BCC, FCC and HCP W-C-N solid solution phases requires modification of the standard expr e ssion for the Gibbs energy of

PAGE 160

-, 151 mixing. This modification restricts mixing of N, C and Va to the interstitial sublattice, and assumes that all sites on the metal sublattice are filled. The modified expression for Gibbs energy of the ternary solid solution phase, adapted for the two-sublattice model, is shown below: Gq, 1Gq, 1Gq, 1Gq, m Y N W : N + Y e W : e + Yva W:Va ( I I I I I I) xs + zRT YN lnyN + Y e lnye + Yva lnyv a + Gmix (3-34) where G : is the Gibbs energy (per mole of formula units) for the cp solution phase (cp=BCC, FCC or HCP), yN1, yc1 and yv/ are the site fractions of nitrogen and vacancies on the interstitial (I) sublattice, respectively, and Gt, N Gt, e and Gt, v a are the Gibbs energy for the cp phase if all interstitial sublattice sites were filled by N, C and Va, respectively The last term, a:x, is the excess term addressing interactions on the interstitial sublattice and can be expanded in the following form for the solid solution phases (
PAGE 161

152 If G:U is dependent on temperature and independent of composition (i.e., regular solution approximation), v = 0 and Equations 3-35 and 3-36 simplify to Equations 3-37 and 3-38, respectively: G xs_ I I o Lcp, O I I o Lcp, O rrnx Y N Y Va W : N Va + Y C Y Va W : C Va I I I O Lcp,O + Y N Y C Y Va W :N,C, Va (3-37) Gxs I Ior.;SHP mix = Ye YN W : C N (3-38) The regular solution approximation was used previously to model solid solution phases in the W-C-N system [Fri99b, Hua97], and is used in this work as well Values for 0L~c,va and 0L~N ,va were discussed above, while a value for 0~;.,e,va is addressed in this analysis to incorporate our experimental data for equilibrium in the W-C-Cl-H-N system. Table 3-20. Gibbs Energy Expressions for W-C-N Solid Solution Phases Phase Sublattice Model Ternary Gibbs Energy Expression Layout oaBCC OGBCC OGBCC OGBCC BCCWNxCy (W)1(N, C Va)3 m = Y N W : N + Y C W : C + Y Va W :Va + 3 RT[y NlnyN + YclnYc + YvalnYval +xs G!cc oaFCC OGfCC OGfCC OGFCC FCC (W)1(N, C, Va)1 m = Y N W : N + Y C W : C + Y Va W :Va + ~-WNxCy RT[yNlnyN +yclnYc +YvalnYval+xs a:cc oaHCP OGHCP OGHCP OGHCP m = Y N W : N + Y C W : C + Y Va W :Va + HCPWNxCy (W)1(N, C, Va)o.s 0.5RT[yN In YN + Y e ln Ye+ Yva In Yva l+xs G!CP oGSHP oGSHP oGSHP SHPWNxCl-x (W)1(N, C)1 m = Y N W:N + Y C W:C + RT[yNlnyN + Yelnyc] +xs G~

PAGE 162

153 3.10.5 W-C-Cl-H-N Liquid Phase Model The Gibbs energy per mole for the liquid phase ( G~ (T, P)) is based on the ideal mixing model, as no experimental data are available to assess any deviations from ideality for this phase : (3-39) where Xi is the mole fraction of species i in the liquid phase and O o:iq is the temperature and pressure dependent molar Gibbs energy for the pure liquid species i, as reported previously [Din91, Gus85]. Existing Gibbs energy expressions for species in the liquid phase will not be modified as no new experimental data have been reported. 3.10.6 W-C-Cl-H-N Gas Phase Model The W-C-Cl-H-N gas phase model is similar to that for the model for the W-N gas phase, described above although many more species are incorporated for this five-component system. Existing Gibbs energy expressions for species in the gas phase will not be modified as no new experimental data have been reported for them 3.11 Optimization of the FCC 13-WNxCy Gibbs Energy Several reports of tungsten carbonitride formation have been given m the literature, but an analysis of the polycrystal's structure and composition has not reported. Chitica [1996] reported WN xCy deposition by laser ablation of a W target in either CI-4 or C3Hs with small amounts of N2. The WNxCy phase had minimal N content, but considerable levels of carbon, which increased with pressure, rising from y = 0.08 to 0.34 over a pressure range of 0.05 to 0 2 Pa. The crystal structure was not indicated, however. Gogova et al. [1998] reported deposition of WN xCy films for use as barriers to Co

PAGE 163

154 diffusion. Two different methods were used to deposit the films; the first was to heat CVD-W thin films in NH 3 C~ and Ar at 1050C, and the second was to react W(C0)6 CH3COCH3 and NH3 at 400C. RHEED analysis of the films showed that the WNxCy was crystalline, but d-spacing values did not correspond to either SHP WC or BCC W. Neither a chemical analysis of the crystalline WNxCy phase before and after annealing nor the pressures for these processes were given, however. Wang et al. [2000] reported thin films containing the cubic B-WCx, HCP W 2C and SHP WN phases by W ion irradiation of a sputtered C film on a W substrate. The pressure during ion bombardment was 3.4x10 3 Pa, but substrate temperature was not measured. While they assigned a broad XRD peak near 20 = 37 to the cubic WCx phase, the peak's position suggests deposition of B-WNxCy; elemental analysis of the polycrystals to confirm this, however, was not reported. Moreover, although W4f and Cls XPS peaks were assigned to the suggested phases, a plot of the Nls peak was not shown. In addition, the phase described as SHP WN may actually have been SHP WNxCl-x The formation of carbonitride upon irradiation with W ions, but not with Ar+ ions, suggests a strong, attractive C-N interaction on the interstitial sublattice of W. This also infers greater stability of the carbonitride relative to the nitride or carbide for these deposition conditions. As mentioned in Chapter 2, a patent application reported the use of MOCVD to deposit B-WNxCy films [FukOO]. X-ray diffraction indicated a peak between 36 and 38 20 and a second position between 20= 42 and 44, both of which are consistent with FCC B-WNxCy deposition. ALD was also used to deposit FCC B-WNxCy films, as mentioned in Chapter 2 [Kim03d]. Bulk film composition from RBS was 48 at.% W, 32

PAGE 164

155 at.% C, and 20 at.% N, and the HR-TEM lattice fringe spacing was 2.39 A, which is between the interplanar spacings for the (111) orientation of P-WN0 .5 (2.38 A) and P -WCo. 6 (2.43 A). The other lattice fringe spacing was 2.08 A, which was between the interplanar spacings for the (200) orientation of P-WNo 5 (2.06 A) and P -WCo. 6 (2.11 A), respectively. In another AW study, P-WNxCy was deposited with a bulk composition of 55 at.% W, 15 at.% N, and 30 at.% C [Li02, Smi02]. While these studies support the presence of a ternary FCC P-WNxCy solid solution, a detailed compositional analysis of the P-WNxCy polycrystals has not been done. To begin the analysis of our experimentally deposited FCC P-WNxCy polycrystals, the modified expression for Gibbs energy for the FCC solid solution phase is shown below: FCC I FCC I FCC IGFCC Gm = YN Gw N + Y e Gw,c + Yva W:Va ( I I I I I I) xs + zRT y N lny N + y c lny c + y va lny va + G mix (3-40) where G xs_ I I o L
PAGE 165

156 (3-43) where Xi is the overall mole fraction of component i, and the superscripts M and I refer to the metal and interstitial sublattices, respectively. N1 is the total number of sites on the interstitial sublattice, while y/ and yv} are the site fraction of i atoms and vacancies, respectively, on the interstitial sublattice. Incorporation of vacancies into the sublattice model enables us to consider homogeneity ranges among the atoms mixing on a given sublattice. Vacancies will be incorporated into the interstitial sublattice, which contains nitrogen and carbon, but will not be considered for the tungsten sublattice. Equation 3-43 above can therefore be simplified, noting that YvaM = 0 and N1=zNM: (3-44) Moreover, assuming that all mixing occurs on the interstitial sublattice, and that tungsten is the only element on the metal sublattice, YwM = 1 and yw1 = YNM = YeM = 0. Therefore, expressions for xw, XN and xe in terms of z and interstitial site fractions can be written, h I l I I notmg t at Yva = YN -ye: (3-45) ( zy' J N XN -I I l+z(yN +Ye) (3-46) (3-47)

PAGE 166

157 The latter two equations can be solved simultaneously for YN1 and Ye1. given the overall mole fractions for C and Nin the polycrystal. Note, however, that the mole fraction of carbon in this expression is denoted xe1 which represents the mole fraction of carbon residing in the interstitial sites of the polycrystals. Experimentally, C incorporates into both the polycrystal's interstitial sites and at grain boundaries, as indicated by XPS results. Film composition was determined by AES, which is an area-averaged technique that cannot differentiate between C in the grain boundary and interstitial C. The value for xe provided by AES therefore is an "average" value, including both interstitial and grain boundary carbon, and can be described by the following equation: I GB Xe= Xe + Xe (3-48) where xeGB is the mole fraction of Cat the grain boundary. Any difference between xe1 and xe indicates leftover "free" carbon not residing in the polycrystal, which therefore resides at the grain boundary (denoted xcG8 ) Since xe, as determined by AES, cannot be directly related to ye1 another method is required to determine the C content in the ~-WNxCy phase The site fraction of C in the ~-WNxCy phase, ye1 may be estimated through the application of Vegard's Law, which represents a linear relationship between the lattice constants of end members in alloys. This Law requires the end members to have the same crystal structure. To find the lattice constant of an alloy mixture (aauoy) containing these three components mixing on the interstitial sublattice Vegard's law takes the form: (3-49) where a1, a2, and a 3 are the lattice constants for phases 1, 2, and 3, respectively The three end-member phases of interest are ~-W, ~-WNo 5 and ~-WCo.6The ~-W phase has an

PAGE 167

158 FCC W lattice structure, with all interstitial octahedral sites containing vacancies. This phase is not observed experimentally, but is useful for the application of Vegard's Law, and should not be confused with experimental reports of B-W, which actually have the W3O stoichiometry [Hag54]. The B-WNo.s and B-WCo 6 phases contain N and C atoms, at 1/2 and 3/5 of the available interstital octahedral sites in the FCC W lattice, respectively, while the remaining sites are vacant. Since the three phases are derived from the same FCC B-W lattice structure, Vegard's law may be applied to determine the shift in lattice parameter with interstitial site occupation. Vegard's law for the lattice parameter of the ternary W-C-N solid solution may be written: (3-50) where a13-WN, a13--wc and ~w are the lattice parameters for B-WN, B-WC and B-W, respectively, which are hypothetical FCC W phase compositions with all interstitial sites completely filled by N, C and Va, respectively. The term on the left, a13--WNxCy, is the experimentally determined lattice parameter of the deposited films (per XRD), while a lattice constant of 3.875 A for a13-w was calculated based on a W atomic radius of 1.37 A [Sha96]. Values of 4.377 and 4.477 A for ~WN and a13--wc, respectively, were extrapolations based on the calculated 3.875 A for ~wand reported lattice parameters of 4.126 and 4.236 A for a13--WNo.5 and a13--wco. 6 respectively Combining these values with Equation 3-50 gives : I I I I a~-WNxCy = 4.377(y N ) + 4.477(y c ) + 3.875(1-y N -y c ) (3-51) There is now one equation with 2 unknowns, YN1 and yc1, To solve this equation, an assumption regarding the site occupation of N must be made Per XPS results for film

PAGE 168

159 growth (Chapter 5), N is only present in the film's polycrystals. Nls XPS peaks at higher binding energy (above 400 e V), which indicate N at the grain boundaries, do not appear for any of our films deposited over the experimental temperature range. Films deposited with NH3 as a co-reactant to boost N content also lack higher binding energy Nls peaks. Therefore, all N detected by AES is assumed to reside in the polycrystals, with none at the grain boundaries. All N detected by AES is therefore considered to preferentially occupy interstitial sites in the polycrystals rather than grain boundaries. This assumption allows us to determine the fractional site occupation of N atoms (yN1 ) on the interstitial sublattice. Once this has been determined, Equation 3-51 is solvable. Before calculating YN1, though, the conversion from overall film mole fraction, x (w~ch is given by AES), to site fraction, y, must be addressed. To convert from mole fraction to site fraction for a two-sublattice model, the Equation 3-52 may be applied [Sau98]: (3-52) where Xj is the overall mole fraction of component j, N5 is the total number of sites on sublattice s, y/ is the site fraction of component j on sublattice s, and yv/ is the site fraction of vacancies on sublattice s. Denoting the metal sublattice "M" and interstitial sublattice "I," the equation for the overall mole fraction of N is : (3-53) This expression can be simplified by noting some properties of the (W)1(N,C,Va)z model. First, only W resides on the M sublattice, so YNM=O. Next, no vacancies are considered

PAGE 169

160 on the M sub lattice, so Yva M=O. Moreover, the number of sites on sub lattice M is related to those on the interstitial sublattice I through the factor z by the following expression: Lastly, since at first only nitrogen addition to the W sublatt ice is being considered: YN + Yva = 1 Now, Equation 3-53 can be rewritten: zy~ X N = 1+ zy~ Now, Equation 3-51 can be rewritten in terms of z and XN, which are knowns: Now, solve for yc1 the carbon site fraction in the ~-WNxCy polycrystals: (3-54) (3-55) (3-56) (3-57) (3-58) aP-WNxCy (A)-4.377 ( X N J3 875 (1-X N y c' J = +4.477 (y c 1 ) (3-59) z( 1 -x N ) z(l -x N ) Simplifying: a" -WN c 0.50,.,( XN J-3.875 I ., x v 1. z( 1-X N) Y e = 0.602 (3-60) This expression gives the site fr a ction of carbon in the ~-WNxCy polycrystals This can be converted to mole fraction of carbon in the polycrystal, xc1 by use of Equation 3-47. Experimental values for film composition and lattice parameter are shown in Table 3-21,

PAGE 170

161 while calculated polycrystal stoichiometries based on these data are shown in Table 3-22. These lattice parameters were derived from XRD spectra shown in Figure 5-31 in Chapter 5. As with most CVD deposition systems, there is some question about the reactor system's proximity to "true" equilibrium conditions. To minimize uncertainty regarding the reactor's proximity to equilibrium, only data for the highest deposition temperature (700C) was included in Table 3-22 above. Higher temperature enables more kinetic barriers to be overcome, and is therefore a better approximation to equilibrium than depositions at lower temperature. Film deposition with NH3 is mass transfer controlled across the entire 450 to 700C temperature range (Figure 5-36 in Chapter 5); in this regime, if the mass transfer rates of the various reactants are similar (as expected), the composition of reactants at the surface should be similar to that in the bulk gas phase. Thus, the film composition for deposition with NH3 at the highest temperature (700C) will be taken as the closest approach to equilibrium, and will be used in the Gibbs energy determination for the ternary B-WNxCy phase. Hence, the calculated B-WN0 _16C0 .43 (Wo.63No.10Co.21) stoichiometry will be taken as the experimental equilibrium value for deposition at 700C. The film composition was determined by AES, which has an accuracy of 5 at. %, hence the polycrystal stoichiometery is a rough estimate both due to compositional inaccuracy and to a lesser extent the use of Vegard's Law approximation. Lack of any literature data for the FCC W-C-N ternary phase made this necessary, however. Data from Rutherford Backscattering Spectroscopy (RBS) were collected to improve compositional accuracy, but the poor resolution provided by the

PAGE 171

162 specific RBS instrument used for measurements made interpretation of the RBS data impossible. Table 3-21. Experimental Bulk Composition and Lattice Parameter Data Used for Assessment of FCC ~-WNxCy Deposition Lattice Bulk Composition in MOCVD Deposition Parameter at Film (at.%) Chemistry Temperature Ambient (aC) Temp., (A) w N C 0 i-Pr + PhCN in H2 700 4.163 46.3 2.7 48.6 2.4 i-Pr+PhCN + 700 4.160 47.9 8.9 41.5 1.7 NH3 in H2 Table 3-22. Estimated Equilibrium Compositions for FCC ~-WNxCy Polycrystals MOCVD Deposition Deposition X in yin Normalized Deposition Chemistry Temp. (K) Pres. (Pa) ~-WNxCy ~-WNxCy Stoichiometry i-Pr+PhCN 973 4.67xl04 X = 0.04 y = 0.46 Wo.61No.02Co.31 in H2 i-Pr+PhCN 973 4.67x104 X = 0.16 y = 0.43 W o.63No.10Co.21 +NH3 inH2 Based on AES, XPS, and XRD data (Chapter 5), it was determined that films deposited at 700C, regardless of NH 3 addition, contained ~-WNxCy and amorphous C, along with minute amounts of WO 3 Oxygen incorporation is believed to occur by post-growth exposure to air, hence the oxygen content and oxide phase will excluded from consideration in the thermodynamic model. This leaves ~-WNxCy and amorphous carbon as the deposited phases to be considered for the model. Amorphous C, however, is inherently a metastable phase, so it will be approximated as being in its stable form (graphite) at experimental conditions for modeling purposes. Deposition of amorphous C rather than graphite may reflect kinetic constraints during deposition that prevent graphite which is considered stable at our experimental conditions, from forming. For purposes of the equilibrium model ~-WNxCy, graphite and gas will be assumed to exist

PAGE 172

163 in equilibrium during deposition in the W-C-Cl-H-N system, hence the final model should reproduce this equilbrium at 700C, with the appropriate stoichiometry in the carbonitride phase. To model the formation of FCC ~-Wo 6 3No.10Co.21 in equilibrium with graphite and a gas phase at 700C, several solid-phase parameters had to be modified. Modified low-temperature expressions for O G and O L~~v. along with a new value for 0L~~.c.va, were generated to accurately model the experimental system; the new values are shown in Table 3-23, and are compared to the original values reported by Gustafson [1986]. In addition, the Gibbs energy and interaction parameter expressions for several neighboring FCC transition metal carbides are listed for comparison; these expressions are truncated to the first three terms (a + bT + CTlnT), where for the Gibbs energy expressions, a=Ho (enthalpy at standard conditions) and -(b+c) = S0 (entropy at standard conditions). Comparing the new values for O G from this work to the original values reported by Gustafson [1986], it is apparent that the new value for S0 is only -8% larger than the existing one, while the new value for Ho is -30% larger. The differences between the values of Ho and S0 generated in this work for FCC WCx and those reported for FCC TiCx and NbCx are due to the inability of W to form a stable monocarbide in FCC form [Pie96]. The composition ranges for FCC TiCx and NbCx near 1000 K extend from Tio.6sCo.35 to Tio.s2Co.4s [Bin95] and Nbo.6oCo.40 to Nbo.soCo.so [Hua97], respectively. Similarly, that for FCC TaCx ranges from Tao_s3Co.47 to Tao.soCo.so [Fri98]. In fact, all of the Group IV, V and VI metals, with the exception of W, can form stoichiometric FCC carbides and nitrides [Pie96]. Stable, stoichiometric WC and WN only exist in SHP form, however. Since the FCC carbides of all Group IV, V and VI metals (except W)

PAGE 173

164 have homogeneity ranges extending up to the stoichiometric limit (Metalo.soCo.so), their minimum vacancy concentration is Yva::::O, As shown in Figure 3-7 and Table 3-15, the stoichiometry of FCC J3-WCx at maximum C content, however, is J3-Wo.60Co.40 (or J3-WC0.61), which corresponds to a minimum vacancy site fraction of Yva::::0.33. Since no ordering has been reported for C on specific interstitial sites in these FCC carbides, the higher minimum value for Yva in FCC WCx suggests that this phase should have higher positional entropy relative to the other Group IV, V and VI FCC carbides, which are not reported to be ordered. This is reflected in the higher S0 value generated by this work. Figure 3-8 shows a plot of the metal's atomic number (Ti=22, Nb=41 and W=74) vs. Ho for several FCC carbides. The solid circles are the reported Ho values for FCC TiCx and FCC NbCx, while the open circle is a calculated value for FCC W:C generated by extrapolation from the FCC TiCx and NbCx values. The open triangle data point is the value from the existing model for FCC WCx, while the open square is the new value generated by this work. Our new value for Ho falls well above the trend predicted by extrapolation from values for FCC TiCx and NbCx (solid line), but is slightly closer to the trend than the existing value from Gustafson [1986]. This indicates that FCC J3-WCx in our model is slightly more stable than reported previously by Gustafson. The original value reported by Gustafson [1986] for 0L~~.va in FCC WCx would be -17,980 J/mol at 1000 K, which is significantly less negative than values for the other Group IV, V and VI FCC carbides (Table 3-23). The higher minimum value for Yva in FCC J3-WCx means that the interaction parameter describing C-Va interactions in FCC W, 0L~~va, should actually have a larger negative value than that for the other Group IV, V and VI FCC carbides, reflecting stronger C-Va attraction in FCC J3-WCx, The

PAGE 174

165 new, low temperature value for 0L1;~va (-260,000 J/mol) indicates a much stronger C-Va attraction in FCC f3-WCx relative to the other carbides -20 ~--------------------, -40 -60 ,-... -80 0 a -100 .__, -120 ::C:o -140 -160 -180 t:,. Calculated [Gus86] o Calculated [this work] o Extrapolated from TiC and NbC values -200 ~---,,--------r------.----------1 20 40 60 80 Atomic Number of Metal in FCC Carbide Figure 3-8. Ho values for the FCC carbides of Ti [Jon96a], Nb [LeeOla] and W [Gus86, this work]. Modifications to the O G and O L1;~. v a values resulted in enhanced stability of the FCC f3-WCx phase at low temperature, consistent with recent experimental reports discussed in Table 3 -15. Although parameters addressing C-N and C-N-Va interactions on the FCC W interstitial sublattice have not been reported, they have been given for other refractory metals, as detailed in Table 3-24. Interestingly, C-N and C-N-Va interaction parameters in the Ti and Nb carbonitride systems are generally very attractive, having large, negative values. For FCC NbNxCy, the reported interaction parameter is variable, depending on the vacancy site fraction on the interstitial sublattice. As listed above the FCC NbCx stoichiometry corresponding to minimum C concentration near 1000 K is

PAGE 175

166 Nbo.60Co.40 [Hua97], which translates to a maximum vacancy site fraction of Yva ::=0.33. The stoichiometry corresponding to minimum N concentration in the analogous nitride, Table 3-23. Gibbs Energy and Interaction Parameter Expressions (J/mol of formula) for Several FCC Metal Carbides. Sub lattice Gibbs Energy (G) and Excess Phase Model Layout Gibbs Energy (L) Expressions Reference (J/mol) ~ 0G~~c = -156735 + 284 *T FCCNbCx (Nb)1(C)1 -46 *TlnT [Lee0la] 0L~~C,Va = -94050+ 23T = -163844 + 267 *T FCCTaCx (Ta)1(C)1 -45 *TlnT [Fri96] 0~~.va =-60408+4T 0G~~ = -168261 + 294 T FCCTiCx (Ti)1(C)1 -48 *TlnT [Lee0la] 0L~~.va = -52702-5T 0G~:~ = -224785 + 297 T FCC ZrCx (Zr)1(C)1 -48 *TlnT [Gui95] 01:i~.va =-41870-36T +6TlnT 0G = -42879 + 303 T FCC J3-WCx (W)1(C)1 -48 *TlnT [Gus86] 0L~~.va =-17.98*T 0G~, ~ = -55879 + 330 T FCC J3-WC x (W)1(C)1 -48 *TlnT This Work 0L~~.v a = -260000 FCC NbNx is Nbo.ssNo.45, which translates to a maximum interstitial vacancy site fraction of Yva ::=0.2. Assuming that the maximum vacancy fraction for the binary carbide would also be the maximum value in the ternary carbonitride the maximum value for the ternary interaction parameter is 0L~~c.N,va = -(0.33*( 312985)) == -104328 J/mol. The d . 1 J: o LFCC d o LFCC al h reporte ternary mteractton parameter va ues 1or n : c N ,va an Nb:c,N,va, ong wit

PAGE 176

167 the new value for 0L~~.N.va from this work, are plotted in Figure 3-9. The new value for 0L~~.N.va from this work fits the trend of the reported Ti and Nb values, as the goodness of fit (R2 value) for this line is 0.96. Table 3-24. Interaction Parameters for Several Refractory Metal Carbonitrides Compound Model FCC Fe1 (C,N)1 FCC Nb1(C,N)1 FCC Nb1(C,N,Va)1 FCC Ti1(C,N,Va)1 FCC W1(C,N,Va)1 Interaction Parameter Ex ression (J/mol) oy FCC LFe:C,N = -14545 oLFCC Nb:C,N =12.5922*T oLFCC Ti :C,N = -54176 oLFCC Nb: C,N,Va = -Yva*312985 oLFCC Ti : C,N Va = -121744 OT FCC LW:N,C,Va = -9()()()() Reference [Du91] [LeeOla] [Jon96a] [LeeOla] [Jon96a] This Work -85 ~-----------------~ ,-.. -90 0 I -95 i::: -110 .9 .... -115 2 .E 120 Nb(C,N,Va) 6 6 Ti(C,N, Va) -125 ....__-~-----~----~------1 20 40 60 80 Atomic Number of Metal in FCC Carbonitride Figure 3-9. C-N-Va interaction parameter trend based on reported values for Ti [Jon96a] and Nb [LeeOla] and the new value for W from this work.

PAGE 177

168 Our experimental XPS results also indicate attractive C-N interaction on the interstitial sublattice of the WNx polycrystals, as discussed in Chapter 5, but evidence of ordering is not present. Moreover, the ternary FCC TiNxCy and FCC HfNxCy phases are reported to have broader homogeneity ranges than their binary metal nitride and metal carbide counterparts [Bin95], which is further evidence of attractive C-N and C-N-Va interaction on the interstitial sublattice of these metals. 3.12 Degrees of Freedom (DOF) Analysis The CVD process involves a mixture of gaseous precursors and products contacting a solid substrate surface. For this reason, gas phase chemical equilibrium calculations are useful to estimate the makeup of the gas taking part in the reaction at the substrate surface. Equilibrium calculations can be done for a homogeneous gas phase system, which represents the scenario in a reactor system before deposition begins, and for a heterogeneous gas-solid phase system, which represents the equilibrium after deposition has begun. The mixture of precursor, solvent, and carrier gas examined in this study contain W, C, Cl, Hand N atoms. Both H2 and N2 were tested experimentally as carrier gases, but N2 did not afford low temperature deposition, and prevented deposition when added as a co-reactant to H2 carrier gas. N2 is assumed to be inert over the entire experimental temperature and pressure range, due to a large kinetic barrier to its dissociation/reaction, so this result was not surprising. N present in any equilibrium phases or species therefore originates from decomposition of the precursor and/or solvent. Since i-Pr + PhCN in H2 carrier gas provided the best experimental film results, this will be used as the basis for our thermodynamic analysis. As a base condition, the initial input values were the mole

PAGE 178

169 fractions of W, C, Cl, H and N present in a mixture containing 7 5 mg of the isopropyl precursor Cl4(CH3CN)W(NiPr), 10 mL of benzonitrile (PhCN) solvent, and 1000 standard cm3/min (seem) of H 2 carrier gas, as shown in Table 3-25. W-C-Cl-H-N equilibrium calculations were done for temperatures ranging from 300 to 1200 Kand pressures ranging from Oto lx105 Pa Temperature and pressure, the two easiest variables for us to manipulate experimentally, were varied to determine their impact on homogeneous and heterogeneous gas phase speciation The most important input atom ratios, along with temperature, were also varied to determine their effect on solid phase deposition at certain equilibrium conditions. Atom ratios considered in the DOF analysis are listed in Table 3-26, along with the values of these ratios in a basis mixture of i-Pr + PhCN + H 2 carrier gas, and the inputs for the calculation. The CVD phase diagrams generated by this DOF analysis are useful to determine which direction one must go experimentally, with respect to variable manipulation, to deposit the desired solid phase(s) Table 3-25. Elemental Composition for C4(CH3CN)WN-i-Pr in PhCN with 1000 seem H 2 Carrier Gas Amount Amount Amount Present in Present in Present in Total Input Mole Element i-Pr Precursor PhCN Carrier Gas (gmol/min) Fraction (gmol/min) (gmol/min) (gmol/min) w l.18xl0-0 -1.18 xl0-6 2.23x10 -:, C 5.90x10 -0 4.53x10-j 4 53x10-j 8.55x 10-.t Cl 4.72xl0-6 --4 72 xl0-0 8.90x10 -:, H l.18x10-:, 3.23x10j 4.46x10 -.t 4.78 xlo-.t 9.02x10 -1 N 2.36x10 o 6.46xl0-4 6.49xl0-4 l.22xl0-.t For a system containing 5 elements, there are 7 "true degrees of freedom (number of elements + 2) that must be satisfied before the equilibrium calculation can begin. For the homogeneous and heterogeneous specia t ion calculations the input moles

PAGE 179

170 of each of the five species were fixed to the basis value (Table 3-25). For isobaric calculations with variable temperature, the pressure was fixed and an initial temperature value was input to generate an initial equilibrium. From this initial equilibrium data point, the temperature was varied to determine its effect on the partial pressure of the gas phase species produced. For isothermal calculations with variable pressure, the temperature was fixed and an initial pressure value was input to generate an initial equilibrium Pressure was then varied to determine its effect on partial pressure of gas species produced. Table 3-26. Ratios and Inputs Used for Heterogeneous Solid Phase Equilibrium DOF Analysis Input Ratio Ratio in Basis Inputs for Calculation Mixture (status during mapping) Temperature (variable) Pressure (fixed) [ molesN ] moles W + N (fixed) moles N + moles W 0.998 moles W (variable) moles C (fixed) moles Cl (fixed) moles H (fixed) Temperature (variable) Pressure (fixed) [ molesC ] moles W + C (fixed) 0 999 moles W (variable) moles C + moles W moles Cl (fixed) moles H (fixed) moles N (fixed) For heterogeneous solid phase equilibrium (CVD phase diagrams), the atom ratios of interest, along with temperature, were varied to determine their impact on the solid phase deposited To vary the atom ratio, the sum of the two atoms of interest was set to a fixed value, as were the individual values for the remaining three atoms. By fixing the sum of the two atoms of interest, the ratio of atoms within this sum could be varied in a stepwise manner An initial value for one of the elements in the sum was supplied, which

PAGE 180

171 brings the degrees of freedom to zero so that the calculation may begin. After an initial equilibrium is calculated, the quantity of one element in the sum is varied, which automatically varies the value for the other element in the sum, since their total is fixed. In addition to element ratio, the temperature is also varied so that the CVD phase diagram can be mapped. Table 3-26 shows the input ratios studied, as well as the input values and their status (fixed or variable) during calculation of the CVD phase diagram 3.12.1 Homogeneous Gas Phase Speciation Homogeneous gas-phase equilibrium was calculated by considering all potential gas phase species (per ThermoCalc's SUB94 database) in equilibrium with solid W which has minimal vapor pressure at the temperatures of interest and has minimal impact on the gas species. Mole fractions for the precursor/solvent mixture in 1000 seem of H2 carrier gas are shown in Table 3-25. This is the basis concentration of reactants during deposition, without any co-reactants such as NH3 or N2 added. The values in this table serve as the initial conditions input to ThermoCalc for it to begin the equilibrium calculation. Figure 3-10 shows the variation of partial pressure with temperature for gas species in the Cl4(CH3CN)WN-i-Pr + PhCN + H 2 carrier gas system. The bulk of the gas phase is made up of H2. The dominant carbon containing species was C~, which is a stable and desirable reaction product in MOCVD. C~ is an effective gas phase carbon sink that limits the amount of carbon available to contaminate the film. The concentration of C~ was fairly steady over the temperature range examined. A variety of other C containing species are also present in the homogeneous equilibrium, as shown in Figure 3-10.

PAGE 181

172 The dominant equilibrium Cl-containing species in the gas phase was HCl, and its equilibrium concentration was unchanged with increasing temperature. Interestingly, the presence of Clz is not predicted in appreciable concentration. The predominance of HCl is due to excess H2 carrier gas, which is available to react with Cl ligands leaving the precursor to form HCl. Below 500 K, the dominant N containing species is NH3 while N 2 is dominant above 500 K. The NH3 level steadily decreases above 500 K, which is indicative of its tendency to decompose (presumably to N 2 and H 2 ) with increasing temperature. The concentration of N 2 increases up to 600 K and then levels off with increasing temperature, which is indicative of its high thermal stability in this temperature range. While N2 and NH3 are the principal species through which N escapes from the film, HCN and HNC are two other N-containing species in the homogeneous equilibrium through which N may desorb from the films. HCN first appears near 800 K, and overtakes NH3 in concentration above 1000 K. HNC appears above 1000 K, but its concentration remains below that for N2, NH3 and HCN up to 1200 K. The impact of overall system pressure on the gas phase equilibrium was also examined. Figure 3-11 shows the dependence of gas phase speciation on total system pressure as a function of temperature. The levels of H2, N 2 and HCl are essentially constant over the entire pressure range for the four temperatures shown. Increasing overall system pressure tends to decrease the variety of gas species present in appreciable concentration. Figure 3-11 indicates that an increase in total pressure causes an increase in the NH3 concentration, however. The concentration of NH 3 increases steadily with total system pressure at the four temperatures shown (Figure 3-1 la-d). The drop in NH3

PAGE 182

173 concentration with increasing temperature, discussed above, is agam evident when moving from Figure 3-1 la to d. The increase in NH3 concentration with pressure is expected, as increased pressure drives the following equilibrium to the left per Le Chatelier' s principle. 10' ...-------------------------, CH N 400 600 800 Temperature (K) Figure 3-10. Effect of temperature on homogeneous gas phase equilibria for a mixture of the i-Pr precursor in benzonitrile solvent, with H2 carrier gas flowrate of 1000 seem at 4.67x104 Pa (350 Torr). In this case, the equilibrium constant, K, is proportional to 11Ptotai2, hence increasing the total system pressure decreases the tendency of NH3 to dissociate. While the above equilibrium suggests that N2 and H2 should combine to form NH3 at higher pressure, a large kinetic barrier to the decomposition of N2, coupled with our relatively low operating pressures during CVD, make this unlikely during deposition.

PAGE 183

174 The concentration of CR. also increases with total system pressure; this increase is especially noticeable at 1200 K. This is presumably due to a shift in equilibria similar to that discussed above for NH3 Equilibria leading to formation of C2R. and C.JI8 for example, may be written: 2CR. C2R. + 2H2 4CR. C.Jis + 4H2 To form C2R., 2 moles of reactant (CR.) are converted to 3 moles of product (1 mole C2R., 2 moles H2). For both of these equilibria, the equilibrium constant, K, is proportional to l/P10ta1, hence increasing the total system pressure decreases the tendency of CR. to form products. HCN first appears in Figure 3-11 b at 800 K for pressures below 0.5 atm, but its concentration also decreases with increasing system pressure. Equilibria based on both N2 and NH 3 can be written which give rise to formation of HCN: 2CR. + N2 ;= 2HCN + 3H2 2CR. + 2NH3 ;= 2HCN + 6H2 To form HCN from CR. and N2, for example, 3 moles of reactant (2 moles CR., 1 mole N2) are converted to 5 moles of product (2 moles HCN, 3 moles H 2 ). For these equilibria, the equilibrium constant, K, is proportional to 11Ptotai2 and to l/P1otai4. respectively, hence increasing the total system pressure decreases the tendency to form HCN. As evident in Figure 3-1 ld at 1200 K, high temperature fosters formation of a large variety of C containing species. The concentration of all of these species, with the exception of CR., C28<, and C3Hs, decreases with increasing total system pressure. The increase in the concentration of CR., C28<, and C 3Hs with total pressure can again be

PAGE 184

175 explained by a shift in equilibrium to decrease the total number of moles at higher pressure. Formation of other C containing species from C2~ and C 3 H8 follows equilibrium behavior similar to that for C~. For example: C 2 ~ C2~+H2 4C3Hs 3C4Hs + 4H2 To form C 2~, 1 mole of reactant (C 2~) is converted to 2 moles of product (1 mole C2~, 1 mole H2). Likewise, to form C~8 4 moles of reactant (C 3Hs) is converted to 7 moles of product (3 moles C~8 4 moles H 2). For these equilibria, the equilibrium constant, K, is proportional to l!Ptotal and to 11Ptotai3, respectively, hence increasing the total system pressure favors formation of C2~ and C 3 H8 A comparison of the behavior for C~, C 2 ~ and C 3Hs with pressure results in the following equilibria: 2C~ 3C~ The total number of moles in the gas phase is conserved per these equilibria, so increasing total system pressure should not favor formation of C 2 ~ or C 3 H8 over C~, and vice versa. The concentration of each of these species increases with pressure due to decomposition of the other C containing species at higher pressure. In these equilibria, K is independent of pressure, hence increasing the total system pressure will not favor formation of reactants or products Homogeneous speciation calculations indicate that operation at lower temperature and higher pressure yield maximum C~ concentration which is the most abundant gas

PAGE 185

176 phase C sink. Increasing deposition temperature and decreasing pressure leads to the largest variety of C containing species in the gas phase. The concentrations of these 101 100 101 ,.... 10-2 s "' ._, "' "' -; -~ "' ._, "' "' 0 d:: -; -~ 10-3 10-4 I 10s 10-6 10-1 10-s 0 0 101 ID 10-1 10-2 10-3 10-4 10-s 10-6 101 Io-8 0.0 Hi CH4 N 2 HCI C 2 H 6 I 0 2 0 4 0.6 0.8 Total Pressure (atm) C CH4 N 0 2 0.4 0.6 0 8 Total Pressure (atm) 101 a b 100 CH 101 ,.... s "' ._, 102 J0-3 "' "' HCI 10-4 -; -~ 10-s J0-6 10-1 10-s 1.0 0.0 0 2 0.4 0 6 0 8 1.0 Total Pressure (atm) 101 d 100 CH 10-1 "' ._, 10-2 J0-3 "' "' 0 d:: 10-4 -; ~ 10s 10-6 101 Io-8 1.0 0.0 0.2 0 4 0 6 0 8 1.0 Total Pressure (atm) Figure 3-11. Effect of total system pressure on homogeneous gas phase equilibria for a mixture of the i-Pr precursor in benzonitrile solvent, with H2 carrier gas flowrate of 1000 seem. (a) 600 K (b) 800 K (c) 1000 K (d) 1200 K.

PAGE 186

177 various species, however, are orders of magnitude lower than that for C~, so operating in a region to maximize C~ content appears to be the best way to maximize the amoun t of C in volatile form. 3.12.2 Heterogeneous Gas Phase Speciation Heterogeneous gas phase speciation is shown in Figure 3-12. H 2 and C~ are the dominant H and C containing species, respectively, while the variety of carbon containing species with concentrations~ 10-s atm are considerably less for heterogeneous speciation relative to the homogeneous case. H 2 concentration is steady at lower temperature, but increases above 800 K, which suggests increased alkyl decomposition at higher temperature. The decrease in concentration for most C containing species at higher temperature, as well as the smaller variety of C containing species, is consistent with increasing C deposition in the films at higher temperature. The dominant Cl containing species is again HCl, but rather than having a constant concentration as in the homogeneous case, the HCl level drops rapidly below 425 K. This is due to the stability of solid ~Cl at low temperature, which serves as a solid phase sink for Cl. N 2 and NH3 behave in a similar fashion in both the homogeneous and heterogeneous equilibria. Interestingly, the C~, C 28
PAGE 187

178 deposit carbon into films at higher temperature ( discussed further in Chapter 5). Hence, operating below 800 K (527C) may help to decrease formation of HCN, which can subsequently decompose to deposit C in the films. 10' 100 ---10-1 ij .._, 10-2 0 !3 10-3 Cl) Cl) 0 10-4 10-s HCI ..... 0.. 10-6 10-1 10-s 400 H N2 600 800 1000 Temperature (K) ---H -----C2H2 ---HNc I \ 1200 CH3 Figure 3-12. Effect of temperature on heterogeneous gas phase equilibria for a mixture of the i-Pr precursor in benzonitrile solvent, with H2 carrier gas flowrate of 1000 seem at 4.67x104 Pa (350 Torr). The effect of total system pressure on heterogeneous gas phase speciation was also examined. Figure 3-13 shows species partial pressure vs. total system pressure at 600, 800, 1000 and 1200 K. H 2 N 2 and HCl levels were essentially constant over the entire pressure range explored. As in the homogeneous case, the concentrations of NH3 and C~ increase with total system pressure. While HCN concentration decreased with increasing pressure for the homogeneous case, its concentration was unchanged with pressure at 1000 and 1200 K. Moreover, HCN concentration increases with temperature, suggesting that this may be an important pathway for N and C desorption at higher temperature.

PAGE 188

179 In contrast to the homogeneous case, the concentration of many gas species increases with increasing pressure, especially at the two higher temperatures (Figure 3-13c-d). For several of these species (e.g., C2~. CJI8), this behavior runs contrary to Le Chatelier's principle, described above. This is evidence of the solid phases, especially the carbides, carbonitrides and graphite, taking part in the equilibrium and affecting speciation in the gas phase above it. Experimentally, both N 2 and H 2 were tested as carrier gases. Deposition temperatures 2': 800C were necessary to deposit films from the i-Pr precursor in N 2 carrier gas. Use of H 2 as the carrier gas, though, allowed for deposition down to 450C. This is most likely due the thermal conductivity difference between H2 and N2. The thermal conductivities of H 2 and N 2 at 500C are 404 and 53 W/m-K, respectively [lnc90]. Higher carrier gas thermal conductivities cause more upstream heating of the precursor. This improves the gas-phase thermal decomposition of the precursor into desired intermediates, which then react on the substrate surface to deposit WNx films. Addition of NH3 and/or N2 to the H2 carrier gas would appear to help suppress the evolution of N from the precursor and the films. Experiments have shown that addition of NH3 to the reactor system led to increased N content in the deposited films (Chapter 5). This result, however, is likely caused by reaction between NH3 and the precursor molecule/growing film, rather than a shift in the above equilibrium. Similar experiments with N2 (which is inert at our experimental conditions) in place of NH3 led to poor films with minimal nitrogen content. The N 2 co-reactant may have actually blocked surface sites rather than suppressing evolution of N, thereby inhibiting film deposition.

PAGE 189

I ...., ::I
PAGE 190

181 PhCN, HCl, and H 2 The presence of C 2 H 2 and C61-4 during deposition resulted from thermal decomposition of PhCN in the reactor. As discussed in Chapter 5, concentrations for PhCN decomposition products were adjusted to eliminate contributions from PhCN cracking in the RGA. The remaining signal was therefore due to thermal processes in the reactor. C2H2 is predicted to form during heterogeneous equilibrium for high temperatures (2: 1100 K), while C61-4 is not predicted to form in appreciable concentration. Experimental RGA results indicate the presence of PhCN and C~ at 873 K, however. This suggests that PhCN decomposition in our reactor is incomplete and does not occur near equilibrium, which is supported by the kinetically limited character of the carbon deposition process with PhCN solvent, as discussed in Chapter 5. The presence of HCl and HCN in the reactor is qualitatively consistent with heterogeneous equilibrium calculations. HCl is believed to be the major desorption route for Cl leaving the precursor molecules; RGA data and equilibrium calculations support this, and indicate that Ch is not present in significant concentration. While HCN seems to be an alternate pathway for N and C desorption at increasing temperature, literature reports suggest that it decomposes on surfaces containing W, and hence may be responsible for adding C and N to the films (Chapter 5 has more detail). Cl-4, the most abundant C containing species in the equilibrium model, may have been present in the reactor in significant quantity. Both Cl-4 and O have the same rn/z value (rn/z = 16) in our RGA, however, hence a peak at 16 is likely to be a convoluted signal, making the relative concentration of Cl-4 in the reactor unclear. Oxygen presumably emanates from fragmentation of background 0 2 or H 2 0 in the system. N 2 was the carrier gas for RGA studies, so the amount generated during reaction is unknown.

PAGE 191

182 NH3 may also have been present in the reactor in significant quantity, as a peak at m/z=l 7 was present during deposition. Since NH3 and [OHt have the same m/z value (mlz=l 7), however, this signal is likely to be convoluted, including signals for both NH3 and [OHt. The presence of [OHt is likely due to cracking of background H20 in the RGA. Heterogeneous equilibrium calculations, like the homogeneous ones, indicate that operation at lower temperature and higher pressure yields maximum C~ concentration. Again, increasing deposition temperature and decreasing pressure leads to the largest variety of C containing species in the gas phase, although the number of gas species is much lower relative to the homogeneous case, especially at higher temperature. The concentrations of these various species, as in the homogeneous case, are orders of magnitude lower than that for C~, however. Hence, operating conditions that maximize C~ concentration appear to be the best way to maximize the amount of C in volatile form, thereby minimizing the amount of C deposited in the film. Operation at lower temperature and higher pressure maximizes the C~ content in the system, and should therefore minimize the amount of C deposited in the films. 3.12.3 Heterogeneous Solid Phase Equilibrium The effect of varying two input atomic ratios on the solid phase equilibrium was also examined. The goal was to determine which values of these ratios would favor deposition of the FCC phase, without any second phase present. The first ratio considers changes to the number of N atoms relative W atoms in the input mixture. The plot in Figure 3-14 shows the effect of varying temperature and N/N+W on the deposited equilibrium solid phases. Note that the gas phase is in equilibrium with all solid phases

PAGE 192

183 shown on the diagram, but is not included on the diagram. Four solid phases are evident in the diagram: FCC J3-WNxCy, SHP WNxCt-x, graphite, and ~Cl. The roughly horizontal line extending from 770 to 828 K delineates two regions. Graphite is stable above this line, but does not appear below it. At the highest temperature, for all values of N/N+W, a mixture of FCC and graphite is stable. As temperature decreases toward the graphite demarcation line, a two-phase and a three-phase solid region appear. The two-phase solid region contains a mixture of SHP and graphite, and is present at higher values of N/N+W. The three-phase region consisting of FCC, SHP and graphite exists at lower values of N/N+W. Below the graphite demarcation line, graphite disappears, while the FCC and SHP phases remain. For N/N+W values above 0.5 and temperatures between 484 K and the graphite line, a single SHP phase region exists, while a two-phase SHP + ~Cl region exists below 484 K, with ~Cl existing in two different forms. Per JANAF [Lid85], ~Cl undergoes a second order transition at 457.7 K, which is evidenced by a change in heat capacity of the solid at this temperature This is reflected in the diagram, where ~Cl_l exists with SHP above this transition temperature, while ~Cl_2 exists with SHP below it. Moving below N/N+W = 0.5, the equilibrium shifts to a two-phase solid mixture of FCC and SHP. The small, triangular region between 550 and 770 K, with N/N+W :::; 0.04, contains a single FCC phase, which is the most desirable phase for barrier deposition. The operating window in which pure FCC phase material can be deposited at equilibrium is relatively narrow. This suggests that during low-temperature ( <770 K) equilibrium deposition conditions, the amount of available N input to the reactor should be decreased relative to our current input amount, to avoid formation of the high resistivity SHP phase. Even for depositions

PAGE 193

184 at 450C (723 K), however, FCC (presumably, based on inference from XPS data) + amorphous carbon is deposited. This suggests that kinetic constraints at this low temperature may cause a surface limited process that inhibits formation and/or desorption of C as a volatile species. Inadequate C~ production and/or desorption, for example, could lead to amorphous C deposition even at the lowest temperature. Experimentally, use of Cl4(CH3CN)WN-i-Pr + PhCN + H2 causes the minimum value of the N/N+W ratio to be -0.998, as shown in Table 3-26. By adding an N bearing co-reactant (such as NH3), this ratio can be increased essentially to 1. Conversely, if an N-free solvent were used in place of PhCN, the ratio above would reach a minimum value of 2/3, due to each i-Pr precursor molecule containing one W and two N atoms (imido N and N in CH3CN trans ligand). The trans ligand could be changed to one without N, which would decrease this ratio to 1/2. To lower the ratio any further, the number of W atoms in the precursor molecule must be increased or the single source chemistry would need to be abandoned in favor of a co-reactant scheme, with separate precursors providing the W and N atoms. The second ratio considers changes to the number of C atoms relative W atoms in the input mixture. The plot in Figure 3-15 shows the effect of varying temperature and C/C+W on the deposited equilibrium solid phases. Note again that the gas phase is in equilibrium with all solid phases shown on the diagram, but is not included on the diagram. Four solid phases are also evident in this diagram: FCC B-WNxCy, SHP WNxCi-x, graphite and ~Cl. The line extending from 518 to 438 K delineates two regions: ~Cl is stable below it (with SHP), but does not appear above it. Moving above this line, a single phase SHP field exists. The curve extending from C/C+W=O.l

PAGE 194

185 1200 FCC + graphite 1000 FCC '-' SHP+ graphite 800 ... :S .... 8. SHP 8 Cl.) 600 E-o SHP+NH4Cl_l FCC+SHP 400 SHP + NH4C1_2 0 .0 0.2 0.4 0.6 0.8 1.0 [ moles N ] moles N + moles W Figure 3-14. CVD phase diagram for variable temperature and N to W ratio; the gas phase is in equilibrium with all solids shown in the diagram, but is not indicated on the diagram. 1200 FCC + Graphite 1000 ,,...._ ::.i:: Graphite + SHP '-' 800 :S .... 8. 5 600 E-o SHP 400 0.0 0. 2 0.4 0.6 0.8 1.0 [ moles C ] moles C + moles W Figure 3-15. CVD phase diagram for variable temperature and C to W ratio; the gas phase is in equilibrium with all solids shown in the diagram, but is not indicated on the diagram. at T=1200 K to CIC+ W=l at T=828 K, with a discontinuity near T=l020 K, delineates the regions with and without graphite. Graphite appears to the right of this

PAGE 195

186 curve, in equilibrium with FCC at higher temperature and SHP at lower temperature The appearance of graphite at higher temperature and C content makes intuitive sense, as moving toward higher temperature and C levels in our system leads to greater C contamination in the films. The line that is slightly concave down and intercepts the axes on either side of the diagram at T=1020 K delineates regions with (above the line) and without (below the line) the FCC phase. At lower values of CIC+ W above 1020 K (747C), FCC exists as the single solid phase. This suggests that the amount of C in our system should be decreased significantly from the normal value of C/C+W = 0.999 (Table 3-26), and that operation at temperatures greater than 1020 K is required To decrease the C content, a change in precursor ligands and, more importantly, solvent, would be required. Operating at high temperature favors polycrystalline film deposition, however, which leads to poor barrier performance. Based on the CVD diagrams in Figure 3-14 and Figure 3-15, it appears that fixing the C input to current levels and lowering both the input N level and temperature is the likely route to equilibrium deposition of the FCC phase. Kinetic constraints on the formation and desorption of volatile C containing species, however, may make this difficult, as moving toward lower temperature, which is required for barrier deposition, inherently moves the reactor conditions further away from equilibrium. 3.13 Predicted W-C-N Ternary Phase Diagram Using the database generated for the W-C-Cl-H-N system, ternary phase diagrams for the W-C-N subsystem can be predicted. The predicted ternary equilibrium for the W-C-N system at 700 C (973 K) and lx105 Pa is shown in Figure 3-16. Four solid phases and one gas phase are present on this diagram. The four condensed phases

PAGE 196

187 are FCC BCC W, FCC ~WNxCy, SHP WNxC1x and graphite, while the gas phase is N2. Interestingly, the HCP WNxCy phase does not appear in the diagram, nor did it appear in either of the CVD phase diagrams. A sizable homogeneity range exists for the single phase field of FCC f3-WNxCy, which is similar to homogeneity ranges reported for the FCC phase in the Ti-C-N, Zr-C-N [Bin95] and Nb-C-N [LeeOla] systems. The FCC ~WNxCy region is broader at the C-rich end of its range, suggesting enhanced stability of this phase at lower x and higher y values. C 0.0 Graphite + N2 + 8-WN ~~=====;:::=:::::+--f-~~---,---r----r----r--40.0 W 0.0 0.1 0.2 0 5 0.6 0.7 0.8 0.9 1.0 N ~-WN x C y + 8 WN x Cl-x Figure 3-16. Predicted ternary equilibrium phase diagram for the W-C-N system at 973 Kand lx105 Pa.

PAGE 197

188 Three two-phase regions are present on the diagram. The triangular shaped region in the bottom left comer of the diagram extending from W to f3-Wo.sCo.20 to f3-W 0 _69N 0 _31 is a two-phase region containing the BCC Wand FCC f3-WNxCy phases in equilibrium. Another two-phase region exists between the single phase FCC f3-WNxCy field and the line connecting &-WN with &-Wo.soNo.49Co.01The third two-phase region exists between graphite and the FCC f3-WNxCy phase. Two three-phase regions are also present on the diagram. The triangular shaped region on the right of the diagram is a three-phase invariant region bounded by SHP &-WN, graphite, and N 2 gas. The other three-phase invariant region is bounded by graphite, f3-W o.62No. 13Co.2s and &-W o .soNo.49Co.01. In application to CVD, this diagram suggests that the W IC and WIN ratios should initially be made large to drive deposition into the two-phase W + FCC f3-WNxCy field, where W could be easily detected by XRD. The initial growth temperatures should be relatively high to produce films which are easy to characterize. It is noted that the boundary between the two-phase region and the single FCC f3-WNxCy field is only weakly dependent on temperature. Thus, the strategy would then be to increase the WIN and possibly W/C ratios, until the W metal phase disappears. The next step then would be to lower the processing temperature to foster deposition of an amorphous FCC f3-WNxCy film The predicted ternary equilibrium diagrams for the W-C-N system at lxl05 Pa are shown for temperatures of 500C (773 K) and 300C (573 K) in Figure 3-17 and Figure 3-18, respectively. Interestingly, the two-phase graphite + FCC f3-WNxCy region shrinks with decreasing temperature through growth of the two-phase FCC

PAGE 198

189 ~-WNxCy + 8-WN region. The increase in area of the FCC+ 8 two-phase region with decreasing temperature follows the trend on the binary W-N diagram, in which the FCC region narrows with decreasing temperature below 1046 K while the two phase FCC+ 8 region grows (Figure 3-5). The tested precursors were unable to deposit films at 300C, requiring higher temperature to trigger film growth. Hence, no experimentally grown film is available for comparison against this diagram. C 0.0 Graphite + N2 + 8-WN Figure 3-17. Predicted ternary equilibrium phase diagram for the W-C-N system at 773 Kand lx105 Pa.

PAGE 199

190 C 0.0 Graphite + N2 + 8-WN ~~~====~;:sr:.:::::;:~=-IIIIJ---.......--.......--....---....---~ 0.0 W 0.0 0.1 0.2 0.5 0.6 0.7 0.8 0.9 1.0 N ~-WNxCy + 8-WNxCl-x Figure 3-18. Predicted ternary equilibrium phase diagram for the W-C-N system at 573 Kand lxl05 Pa. 3.14 Predicted ~-WCo.s-f3--WNo.sPseudobinary Equilibrium Since FCC is the desirable structure for barrier film deposition, it is useful to understand how temperature affects the equilibrium composition of FCC B-WNxCy, The pseudobinary diagram in Figure 3-19 shows the effect of temperature on the FCC B-WNxCy for varying N and C content. The end members of this pseudobinary are the FCC B-WC0 5 and FCC B-WNo s phases. While the C and N content of this phase varies substantially, the end member stoichiometries were fixed to B-WCo.s and B-WNo.s to match the most desired stoichiometry for nitride barriers (P-WNo s = B-W 2N). As

PAGE 200

191 evident from Figure 3-19a, FCC B-WNxCy is stable at lxl05 Pa for temperatures below 1869 K. The pure nitride begins to decompose above this temperature, leading to formation of gas and FCC B-WNxCy. As temperature increases above 1869 K, the N content of FCC B-WNxCy decreases (due to formation of N2 in the gas phase) while C content increases. The trend of decreasing N and increasing C content for the FCC B-WNxCy phase with increasing temperature matches the trend for the deposited films. This behavior is consistent with decomposition of the binary nitride phase at higher temperature, rather than melting, which is reported for the binary carbide phase (far left of diagram). An expanded view of the left side of the pseudobinary diagram is shown in Figure 3-19b. On the far left side of the diagram, congruent melting of FCC B-WNxCy (with N content below 0.002) is predicted for temperatures at and above 3004 K. A single ternary liquid phase is present above 3004 K in this region of the diagram. A small triangular region from T=3004 to 3012 K with N content between 0.002 and 0.01 contains FCC+ liquid in equilibrium. Above 3012 K for N content above 0.002, liquid exists in equilibrium with a gas phase Between the FCC + gas and liquid + gas two-phase fields on the diagram is a narrow, three-phase equilibrium region containing FCC + liquid + gas. The temperature at which liquid forms in equilibrium with FCC and gas increases with increasing C content on the interstitial sublattice, reaching a maximum value near 3012 K. In addition to predicting the ternary W-C-N and pseudobinary tl-WNo s-B-WCo s equilibria, the site fraction of C, N and vacancies on the FCC W interstitial sublattice was also examined. The change in interstitial site occupation with

PAGE 201

192 4000 ----------------------, Liquid Liquid+ Gas a FCC + Liquid + Gas -3000f\::-------------~ .._, j J 2000 1000 0.00 t FCC 13-WCo.s 0 05 FCC+Gas FCC 0 .10 0 .15 0 20 mole fraction N 0 .25 0 30 1 FCC 13-WNo.s 3040-------------------, b 3020 .._, E3ooo i r4 2980 Liquid Liquid+ Gas FCC + Liquid + Gas FCC+ Gas FCC 2960 +--''----~----,---....._ _____ or FCC 13-WCo. s 0.05 0.10 0.15 0 .20 mole fraction N Figure 3-19. FCC J3-WN0 5 -J3-WC0 5 solid-liquid-gas pseudobinary diagram at lxl05 Pa. a) Full view. b) Zoom view to show three-phase regions. temperature is shown in Figure 3-20 below. The graph indicates complete occupation of all interstitial sites by N at temperatures below -680 K, with N occupation decreasing steadily above this temperature. Between 680 and 1100 K, escaping N atoms leave

PAGE 202

-------------------------------,,-,..,,=-, 193 behind vacancies, rather than incorporating C in place of N. At 1100 K, C begins to occupy interstitial sites on the sublattice, and increases to a maximum occupation of 0.66 near 2500 K. Experimentally C replaces N on the interstitial sublattice at temperatures far below 1100 K, which again may be due to the inability of C to form and/or desorb in a volatile species at our lower deposition temperatures. d.) 1.0 u .... .... ..... ,0 ::s Cl) 0.8 ca .... .::: c:/J .... 0.6 u u 0.4 i= Va 0 .::: 0.2 d.) .... 0 0 .... Cl) 500 1000 1500 2000 2500 3000 Temperature (K) Figure 3-20. Temperature dependence of C, N and Va site fraction on interstitial FCC W sublattice at 350 Torr.

PAGE 203

CHAPTER4 EXPERIMENTAL APPROACH FOR CVD 4.1 Substrate Preparation The majority of film depositions were done on Si substrates for several reasons. First, Si is inexpensive and readily available. Second, interdiffusion and reaction between Cu and Si is very rapid, even at low temperatures, malting Si a very conservative material on which to test a barrier. Third, future metallization schemes will look to extend Cu down to the contact level, meaning that Cu will be in intimate contact with Si. In addition to Si, however, barriers must also be tested on other materials, such as new, potential low-k and high-k dielectric materials. Depositions on these materials will be pursued in the future. Boron doped (p-type) Si substrates from Wacker-Chemitronic GMBH were used for the WNx growths. Substrates included both Si (100) and (111) orientations, which were cut from 150 mm wafers. Substrate resistivity ranged from 1 to 2 Q-cm. Substrate preparation prior to deposition included degreasing and etching steps. Grease remaining from the Si wafer production process must first be removed from the substrates, after which the thin native oxide layer (10 to 20 A) typically present on the Si surface (due to air exposure) must be removed [Jae88]. To degrease the substrates, they were immersed for three minutes per solvent in warm trichloroethylene (TCE), acetone, and methanol. The substrates were then placed in boiling deionized (DI) water for thirty seconds to remove any remaining solvent, and were then submersed for 2 minutes in buffered oxide 194

PAGE 204

195 etch (BOE) solution (6: 1 NH.iF:HF in water) to remove the native oxide layer. After etching, the substrates were wiped off with acetone, blown dry and placed onto the susceptor. The susceptor (with substrates) was loaded into the reactor before each run through a pneumatically controlled gate valve on the reactor head. 4.2 Solvent Tests After synthesis (details of which are given in Chapters 5 6 and 7), the solid precursor was dissolved in solvent and loaded into a Hamilton Samplelock 14.57 mm ID syringe. Syringes were normally filled to 10 mL. Due to the light sensitivity of the precursor/solvent mixture, syringes were covered with aluminum foil after being filled. Solvents tested included acetonitrile (MeCN), N-methylpyrrolidinone (NMP) dimethylacetamide (DMA), dimethylformamide (DMF), benzonitrile (PhCN) and o-dichlorobenzene (1,2-DCB). The structures of these solvents are shown in F i gure 4-1. N== Cl N== PhCN 1,2DCB MeCN 0)0 I -N -N \ O~N'-...,. DMA DMF NMP Figur e 4-1. Solvents tested for dissolution of the precursor

PAGE 205

196 4.3 Description of CVD System Components Film growths were conducted in a custom built chemical vapor deposition (CVD) reactor system (detailed PFDs are in Appendix A). The CVD system contains a 63 mm ID quartz tube reactor, capped on each end by stainless steel flanges. These flanges contain metal collars, which compress viton o-rings onto the quartz tube to vacuum seal the reactor system. To prevent overheating of these o-rings, the stainless steel flanges are encased in water-cooled jackets. The system was maintained at vacuum by a Fisher Scientific Maxima C roughing pump. A Pyrex cold trap upstream of the roughing pump was used to remove particulates from the effluent stream and prevent damage to the roughing pump. System pressures above 1 Torr were measured by an MKS Baratron Type 221A pressure gauge, while pressures below 1 Torr were measured with a Sensavac Series 315 Pirani gauge An MKS 253A throttle valve between the reactor and the cold trap was used to control operating pressure in the system. Depositions were conducted at pressures ranging from 150-760 Torr. Carrier and reacting gases, supplied from cylinders, were metered into the system via MKS 1159B Mass Flow Controllers. Gases available to the system included boron trichloride (BCh), diborane (B2Ilt;), nitrogen (N2 ), hydrogen (H2 ), ammonia (NH3) and helium (He). Lines were traced with heat tape when required. Heat tape temperature was controlled with Staco Energy Variable Autotransformers. A 32 mm OD cylindrical graphite pedestal-type susceptor was used to heat the substrates during deposition. The susceptor rests on an alumina encased type-S thermocouple (Pt-Rh), which gives a continuous temperature measurement. The susceptor has a small bore hole through its center, which extends from the bottom of the

PAGE 206

197 susceptor up to a point roughly 2 mm below its surface. The thermocouple was inserted into this hole, so that its top was very close to the susceptor surface for accurate temperature measurement. The thermocouple/susceptor assembly is shown in Figure 4-2. The susceptor is inductively heated by radio frequency (RF) coils, which are powered by a Westinghouse 7.5 KW RF Generator. Susceptor temperature is controlled manually by adjusting the power control knob on the RF Generator. Susceptor surface Small bore hole, containing thermocouple Substrate -----Thermocouple Figure 4-2 Graphite susceptor / alumina encased thermocouple assembly. A Cole-Parmer 74900-00 Syringe Pump was used to meter the precursor/solvent mixture to the nebulizer unit. A Swagelok 25 PSI poppet valve (SS-2C-KZ-25) was used to prevent the syringe contents from evaporating into the system's vacuum. A Cetac Technologies nebulizer unit was used to generate a droplet mist from the liquid precursor/solvent mixture. The nebulizer contains a piezoelectric quartz plate which vibrates continuously at a frequency of 1.44 MHz due to the application of an electrical potential. This rapid vibration causes liquid droplets contacting the plate to be vaporized

PAGE 207

198 into a mist. The diameter of liquid droplets in the mist formed by the nebulizer can be estimated with Equation 4-1, proposed by Lang [1962]: 1 D = 0.34(S1tTJ3 pF2 (4-1) where D is mean particle diameter (cm), T is the surface tension (dynes/cm), p is the liquid density (g/cm3 ) and Fis the excitation sound frequency (s-1). If one assumes the liquid/precursor mixture to have essentially the same physical properties as benzonitrile solvent at 25C, for example, the median droplet diameter produced by the nebulizer is -2.4 m. After nebulization, the mist is conveyed by carrier gas, which is introduced through the side of the nebulizer chamber, to a heated double tube impinging jet. The jet contained an inner tube, which conveyed the reactant/carrier gas mixture, and an outer tube, which was sealed by an o-ring against the reactor head. The annular space between the tubes contained heating tape, whose temperature was adjusted with an Autotransformer. The heating tape for the impinging jet and the nebulizer discharge line was typically kept at 50C. A schematic of the impinging jet is shown in Figure 4-3. A schematic of the nebulizer and reactor is shown in Figure 4-4. An Inficon Transpector 2 residual gas analyzer (RGA) was used during some experiments to determine composition in the gas phase. The device contains a quadrupole mass spectrometer, and is connected to the system by a 1/16 stainless steel capillary tube, which pulled a gas sample from very near the surface of the substrates inside the reactor. The RGA sensor required high vacuum for operation, and hence a turbo and roughing pump in series were connected to the sensor for this purpose.

PAGE 208

I .I : J I I I .. I----! 2 ,_----1 -:-:-_-199 -~--Precursor Inlet 1 I I I , .. --------Heating tape in annular space --. I ::::,--,-, t : /:/_I :::::;:;:__: I I I l J I I l I I I l I I 1 I I : -!":::-::::::: r > r I I I -1 I I __ I --i -,:,_, __ l I 1 I I I r_--I'-: ,::::,::.,r -I"'--, I i I I ,--1-:-:-_-__ ,.::::--.--,u-1 -1 1:., __ I-, ... --r 0 0 --I _..,_ ____ \.-,,~ ":!"~ -.-:'i_,,,~~ ~-------_, \.~ -''-'-'-''.l "-' \.-:, "-' Perforated Showerhead Plate Figure 4-3 Schematic of the double tube impinging jet with det ai led inlay o f the perforated showerhead plate.

PAGE 209

::n ca en (') ::r 0 8 la (') g, -::r 0 ::, g. C N P-ea RGA 1----Capillary Sample Tube Gate Valve (for sample loading) Graphite Susceptor Heated Transfer Tube i,.1 Impinging Jet Quartz Tube Vibrating Quartz Plate Water Cooled Flange ~RF Coils \/ Cable to Power Supply Dissolved Precursor from Syringy Pump Precursor "Aerosol" Carrier Gas to Nebulizer I Note: Not to Scale I N 8

PAGE 210

201 4.4 Start-Up The time lag between substrate removal from the etch solution and achievement of reactor vacuum was minimized ( <5 minutes) to limit the thickness of the native oxide (SiO2) layer reforming on the substrates, which was estimated to range from 1-8 A for exposures less than 10 minutes [Gan92 Phi92]. Once the substrates were loaded the reactor was sealed and brought to ultimate vacuum (usually 600-900 mTorr). Vacuum was pulled on the system for 5 minutes, and inert gas was then introduced to bring the system up to its operating pressure. Once at the operating pressure, the substrates were typically reduced in a H 2 atmosphere at 1000C for 10 minutes ("H2 bake") to remove any remaining surface oxides. Upon completion of the H 2 bake, the system was brought to the reaction temperature. Once reaction temperature and pressure were stabilized the syringe pump and nebulizer were started. Deposition time ranged from 0.5 to 2 5 hours with syringe pump flowrate set to 4.0 mIJhr. The goal of the depositions was to thermally decompose the precursor molecule in a manner that cleaved off all leaving groups in a clean fashion, leaving behind only W and N in the film 4.5 Copper Deposition The two means of Cu deposition available to us were ECD and sputtering A prior study was used as the framework for our Cu ECD setup [Sha0l]. The ECD solution contained copper sulfate pentahydrate [CuSO 4-5H2O], ammonium ci t rate ((~)2HC6HsO1), and aqueous ammonia [NH 3 (aq)], all of which were purchased from Fisher Scientific. The molar concentration of these species in the bath was 0 08 M CuSO4H 2 O, 0.08 M ~hHC6HsO1, and 0.175 M NH 3 (aq), per a published ECD procedure [Sha0l]. The beaker of ECD solut i on which contained a magnetic stirrin g

PAGE 211

202 bar, was placed on a stirring plate to get good mixing during the plating process. Good mixing ensured intimate contact between the film surface and fresh electroplating solution. Gator clips were clamped on the substrates, and a Gelman Instrument Company constant voltage power supply was used to provide power to the conductive surface of the substrates. Barrier film samples were also coated with Cu by RF magnetron sputtering at 5 mTorr pressure in Ar. Michael Jones in Dr. Norton's group ran the depositions. The sputter deposited Cu film thickness was 100 nm. Annealing of some samples with sputtered Cu was done in-situ in the sputter chamber. 4.6 Analysis Techniques The following is a brief description of the different characterization techniques used to analyze the films in this study. AES Auger electron spectroscopy (AES) is a three-electron process, which can be used to determine sample composition. An electron beam is used to eject a core electron from an atom, forming a hole in the subshell from which ejection occurred. This hole is then filled by an electron from a higher energy (outer) shell. The higher energy electron must release energy as it drops to fill the lower energy hole. The energy may be released either as an x-ray or by ejected a third electron from one of the outer electron shells in the atom. The energy of this third electron, called the Auger electron, is very specific to the atom and energy level from which it originated. By comparing Auger electron energies for a given sample against standards, the element and percent composition can be determined.

PAGE 212

203 Film composition was checked by Auger electron spectroscopy (AES) using a Perkin-Elmer PHI 660 Scanning Auger Multiprobe. The AES system had a nominal sputter rate of 100 A/min. It is important to note that calculation of film concentrat i ons using AES involves taking the ratio an element's peak height in the experimentally measured film to its peak height for the pure element standard. This is shown mathematically in Equation 4-2 [Bri90, Fit95]: (4-2) where XA is the concentration of element A in the film, IA is the measured intensity of the Auger electron from component A in the film matrix, IAstd is the intensity for A's pure element standard, and F AB is a matrix dependent factor, which would take into consideration interactions between the elements in the film matrix, and adjust the measured intensities accordingly. If this F AB is unknown, as is the case for our films, it is not possible to get highly quantitative values for elemental concentrations in the films. The value in using AES for multi element films where values for F AB are unavailable is therefore to compare elemental trends in films with similar properties which should have similar values for F AB Trends in film content vs. deposition temperature, for example, can give insight into the impact of experimental conditions on film compositions without giving exact compositions. Thus, the concentration measurements do not take into account any shifts in sensitivity factor due to m a trix effects.

PAGE 213

204 In addition to determination of film composition, the AES system was used for some depth profiling studies. The three-point AES depth profiling method was used to determine the change in elemental content with depth into the films. AFM Film roughness was analyzed by atomic force microscopy (AFM) using a Digital Instruments Nanoscope ill in tapping mode. Kee-Chan Kim assisted with some AFM measurements. CS-SEM Samples were cleaved to expose the film's cross section at the middle of the substrate, and were then mounted vertically onto modified SEM stub pieces. Conductive carbon paint was used to mount the samples, and to close the electrical connection between the stub piece and the film surface. Film thickness was estimated by cross-sectional scanning electron microscopy (CS-SEM) on a JEOL JSM-6400, with growth rate calculated by dividing film thickness by deposition time. Four-Point Probe Four point probe was used to determine the resistivity of our sample films. The 4-point probe sensor head has 4 metal contacts (points) where the two outside tips pass a current and the two inner tips measure the voltage between them. Once measurements were taken, the sheet resistance was determined by Equation 4-3: R, =4.532*(~J (4-3)

PAGE 214

205 where Rs is sheet resistance (Q / square), V is voltage (Q), and I is current (Amperes). Film resistivity is determined by multiplying sheet resistance by film thickness (Equation 4-4): (4-4) where p is film resistivity (Q*cm) and t is film thickness (cm). Film resistivity was measured with an Alessi Industries four-point probe. SIMS Secondary ion mass spectrometry (SIMS) depth profiles were acquired with a Perkin-Elmer PHI 6600 SIMS system using a 5ke V Cesium primary ion beam and negative secondary acquisition. Dr. Margarida Lambers operated the SIMS system and provided the depth profile data to us. After the films were sputtered by the Cs beam, the crater depths were measured with an Alpha-Step 500 stylus profilometer. TEM Young-Woo Heo from Dr. Norton's group prepared samples for transmission electron microscopy (TEM) using an FEI Strata DB235 Focused Ion Beam (FIB) system. After the samples were prepared, he used a JEOL 2010F TEM to collect diffraction patterns and micrograph images. XPS X-ray photoelectron spectroscopy (XPS) uses monoenergetic soft X-rays to eject electrons from the various electron shells of the atoms under examination. The kinetic energy (KE) of the ejected electron (called a photoelectron) is directly dependent on the x-ray excitation energy and the binding energy (BE) which holds the electron to the atom, per Equation 4-5:

PAGE 215

206 BE = hv KE cJ>sr (4-5) where hv is the energy of the incoming x-ray (h is Planck's constant, 41.3x10-16 eVs and v is frequency in s-1), and cJ>sr is the work function of the spectrometer [Mou95, Dia73, Ole97, Bar94]. The kinetic energy of ejected photoelectrons is measured with the spectrometer, so for known values of hv and cJ>sr, the BE can be calculated. A Mg anode x-ray source with hv =1253.6 eV was used for this analysis. Photoelectron BE is very specific to the parent atom and subshell from which the ejected electron emanates. While the magnitude of the BE can be used to determine the atom and subshell from which the photoelectron emanates, small shifts in the BE can be used to determine the bonding state of the parent atom. Bonding states for the parent atoms can be deduced by comparison of BE values to those published in the literature. Determination of bonding states from BE shifts must be reconciled with sample charging, which occurs when there is an imbalance between the number of photoelectrons ejected from the sample and the number of electrons introduced to the sample by conduction. Positive charging of the sample retards outgoing electrons and causes peaks to shift to higher binding energy, while negative charging causes peaks to shift to lower binding energy [Mou95]. Photoelectron lines are typically 0.5 eV wider for insulators than for conductors, since line width depends on the lifetime of holes generated by the photoionization process [Mou95]. The conductor is able to fill holes much more quickly and therefore has a narrower linewidth. This is true in the case of WNx vs WO3 where WNx (a conductor) has a narrower peak relative to WO3 (a semiconductor). A more detailed description of XPS fundamentals is available elsewhere [Bar94, Mou95, Ole97].

PAGE 216

207 Bonding in the films was examined by X-ray photoelectron spectroscopy (XPS) using a Perkin-Elmer PHI 5100 ESCA System with a Mg anode (hv=1253.6 eV). The XPS system had a nominal sputter rate of 20 Afmin. Deconvolution of XPS spectra was done using RBD Analysis Suite software. XRD After growth, film structure was examined by room temperature X-ray diffraction (XRD) on a Philips APO 3720 diffractometer operating from 5-85 20 degrees with Cu Ka radiation. The term diffractometer refers to an instrument used to study crystalline (and non-crystalline) materials by measuring the way they diffract x-rays of known wavelength [CulOl] The XRD unit used for analysis used the Bragg-Brentano 0-20 method to collect spectra. The governing equation used for XRD is Bragg's Law (Equation 4-6): nA = 2dsin0 (4-6) where n is the number of whole wavelengths, A is the x-ray wavelength (1.5406 A for Cu Ka), dis the interplanar spacing (A) and 0 is the Bragg angle (degrees). Several potential sources of error are associated with this method of XRD including the flat substrate error, the specimen transparency error, and specimen displacement error [Jen96]. To limit the effects of these error sources on the location of peak maxima, all peaks from the deposited film were calibrated to the Si(400) peak position, which was shifted (when necessary) to its standard value (20 = 69.13). XRD spectra were used to estimate lattice parameter, calculate grain size, and to observe texturing. Since the (111) orientation of ~-WNx (or ~-WNxCy) is typically the dominant

PAGE 217

208 peak in the film's spectrum, it was used to estimate grain size and lattice parameter in the films. The lattice parameter of a cubic phase is shown in Equation 4-7: (4-7) where a is the lattice parameter (A) and h, k, and 1 are the plane indices [Cul78]. The Scherrer equation (Equation 4-8) may be used to estimate the average crystallite size (t, A) [Cul78, NagOO]: 0.9l t=---(Bcos0) (4-8) where B is the full width at half maximum (FWHM) for the selected diffraction peak. XRR X-ray reflectivity measurements were taken to estimate density in some film samples. XRR data were collected and analyzed by Dr. Valentin Craciun, using a Philips MRD X'Pert system. 4.7 Precursor Screening Procedure After synthesis of a new precursor by Dr. McElwee-White's group, the candidates were screened by a two-part procedure. This procedure involved precursor composition and mass spectrometry testing by the chemistry group and CVD testing of these precursors by our group. The stepwise screening procedure is outlined below. Part I -Feasibility Study Synthesize precursor. Collect mass spectral data (both positive ion EI and negative ion Cl). Analyze spectral data to determine if precursor molecule may cleave in a manner consistent with objectives.

PAGE 218

209 If promising, run solvation tests to see which solvent may be used for delivery to the nebulizer, and to see if adequate transport from nebulizer to reactor occurs. Use dissolved precursor to grow film at mid-range growth temperature (-600C). Check film crystallinity (or lack thereof) with XRD. If XRD indicates possible presence of WNx film, run AES. If AES results reinforce that Wand N are present, continue to deposition and analysis section Part II -Film Deposition and Analysis Vary film growth temperatures to minimize carbon and oxygen contamination optimize nitrogen concentration and grow amorphous films. Analyze films for crystallinity with XRD. Analyze films with AES for compositions. Using AES data determine nitrogen carbon and oxygen content of films as a function of deposition temperature. Determine film thickness by X-SEM, and calculate growth rate. Determine activation energy for film growth Analyze film resistivities as a function of contaminant concentration and reaction temperature Run SIMS to check for Cl contamination in films. Check film conformality by depositions on patterned wafers. Test barrier integrity by a Cu deposition and anneal process

PAGE 219

CHAPTER5 EVALUATION OF Cl4(CH3CN)WN-i-Pr AS A SUIT ABLE WNx PRECURSOR lmido complexes are attractive candidates for MOCVD of tungsten nitride since the W-N multiple bond in the precursor is likely to survive the deposition process and facilitate incorporation of nitrogen into the film. Additionally, the linear geometry of the multiple bond keeps the alkyl or aromatic group of the imido ligand away from the metal, thus deterring any cyclometallation process. Moreover, by changing this group, the N-C homolytic bond strength of the imido moiety is tunable, which should allow for a degree of control over carbon impurities and other film characteristics. 5.1 Synttesis of Isopropyl [C'4(CH3CN)WN-i-Pr] Precursor The first precJrsor examined in detail was C4(CH 3CN)WN-i-Pr, which contains an isopropyl group singly bound to the imido nitrogen. To prepare the precursor, an orange slurry of tungsten oxychloride (WOC14 3.41 g, 10 mmol) and isopropyl isocyanate (i-PrNCO, 1.57 mL, 16 mmol) in 60 mL of heptane was heated while stirring at 100C for 2 days. After this time, a dark red solution containing red crystals was obtained. The solvent was removed in vacuo and the solid was redissolved in 20 mL of acetonitrile. After stirring for 2 hours at room temperature, the brown solution was filtered and the solvent evaporated to yield 3.7 g of brown crystals. The yield of the crystals from the solution was 88%. 1H NMR (CDCh): 8 1.67 (d, 6H), 2.49 (s, 3H), 7.07 (m, lH). 13C NMR (CDCh): 8 3.35, 23.12, 68.52, 118.37. IR (CH2Ch): 2319, 2282, 1462, 1452, 1386, 1366, 1310, 1271, 1154, 1108, 1075 cm -1 210

PAGE 220

211 The solid precursor had low vapor pressure, making solid source delivery to the reactor impractical. To overcome this inherently low vapor pressure, the single-source precursor was dissolved in a solvent and the precursor/solvent mixture was subsequently vaporized in a nebulizer for easy delivery to the reactor. 5.2 Solvent Selection Choice of an appropriate solvent to convey the precursor compound is essential. The solvent's volatility must be adequate for the solvent/precursor mixture to be nebulized. If the solvent's volatility is too high, it evaporates in the nebulizer and leaves the dissolved solid precursor behind. Several solvents, including benzonitrile (PhCN), acetonitrile (MeCN), N-methyl pyrrolidinone (NMP), dimethylformamide (DMF) and dimethyl acetamide (DMA) were tested with the i-Pr precursor. Benzonitrile (PhCN), with a mid-range vapor pressure (-1 Torr at 25C), was the most commonly used solvent because the other solvents had various drawbacks. Depositions using MeCN, which has a high vapor pressure (-89 Torr at 25 C), resulted in no film deposition, because the solvent and precursor disengaged in the nebulizer. Solid precursor remained in the nebulizer chamber while the MeCN evaporated into the reactor, hence no precursor transport to the substrate surface occurred. Use of N-methyl pyrrolidinone (NMP) solvent (-0.33 Torr at 25 C) at 550C resulted in the incorporation of iron into the deposited films The source of the iron was the impinging jet, which appeared to be oxidized after the run was complete. The NMP solvent, either with or without the precursor, reacted with the iron in the steel impinging jet. This iron was then transported to the substrate surface and incorporated into the films. Per AES, the percentage of iron in the films after deposition with NMP solvent ranged from 10 to 17 at. % Use of

PAGE 221

212 dimethyl formamide (DMF) solvent (-3.7 Torr at 25C) at 550C also caused iron incorporation in the deposited films, for reasons similar to NMP. These films contained 6 to 7 at.% iron. Films grown from the i-Pr precursor in dimethyl acetamide (DMA) solvent (-1.3 Torr at 25C) at 550C yielded high carbon contamination (-50 at. % ), believed due to substantial solvent decomposition at this temperature. 5.3 Precursor Mass Spectral Pre-Screen Because benzonitrile was typically used as the solvent with the Cl4(CH3CN)W(NiPr) precursor, exchange of the acetonitrile ligand with benzonitrile occurs to some extent in the solutions used for MOCVD. For completeness, authentic samples of the Cl4(CJ{5CN)W(NiPr) compound were synthesized for analysis. The synthesis was done by slurrying WOC14 (1.025 g, 3.00 mmol) in a solution of i-PrNCO (0.435 g, 5.11 mmol) in heptane (35 mL) in a sealed pressure vessel. The mixture was heated for two days at 110 C. The solvent was removed from the resulting dark orange-red solution in vacuo. The orange residue was dissolved in a minimal amount of C6H5CN (approximately 5 mL). The resulting solution was stirred for two hours and then added to stirring pentane in a nitrogen glove box to precipitate the dark orange product. The solid was filtered and washed with 25 mL of additional pentane to afford 0.829 g (58.4 % yield) of the product. The desired isopropyl imido precursor structure and its PhCN derivative are shown in Figure 5-1. Although the precursor solutions were prepared with Cl4(CH3CN)W(NiPr) (1), the CVD process uses benzonitrile as the solvent. We have demonstrated by NMR experiments that exchange of the acetonitrile ligand of la with benzonitrile to yield Cl4(PhCN)W(NiPr) (lb) occurs to some extent in the solutions used for MOCVD.

PAGE 222

213 Therefore, mass spectrometry (MS) experiments were carried out with both la and lb. These experiments yielded information about the fragmentation characteristics of the complexes under gas phase ionization conditions. The utilization of mass spectral data to predict possible mechanisms for CVD processes has been previously established [Int89]. Care must be taken in extrapolating any fragmentation information from mass spectral data to the CVD process, however, since the latter is thermal in nature [Lew94]. Nevertheless, MS data have provided insights into probable fragmentation patterns of the tungsten imido precursors. y y N N c,, ,, . fil . \ ,c, c,, ,,, . fil .. \ ,c, c,/ t "'c, c,~ t "'c, N N Ill Ill C C I I CH3 Ph a b Figure 5-1 Schematic of the isopropyl imido precursor. a) With the desired acetonitrile ligand. b) With the benzonitrile ligand. All mass spectral analyses were performed using a Finnigan MAT95Q hybrid sector mass spectrometer (Thermo Finnigan, San Jose, CA). Electron ionization (El) was carried out in positive ion mode using electrons of 70 eV potential and a source temperature of 200 C. Negative ion electron capture chemical ionization (NCI) used methane as the bath gas at an indicate pressure of 2x10 -5 Torr, an electron energy of 100 volts and a source temperature of 120C. All samples were introduced via a controlled temperature probe with heating and cooling enabling temperature control down to 35 C. The mass resolving power (m/&n) was 5000 full width-half maximum (FWHM).

PAGE 223

214 Figure 5-2 shows the positive ion electron-impact ionization (El) and negative ion electron capture chemical ionization (NCI) mass spectra of la. The molecular ion could not be observed in either case. The EI spectrum displays more fragmentation than the CI spectrum, due to the more energetic nature of the EI process relative to the CI process. The base peak in the EI spectrum occurs at m/z=348 and is assigned to the [ChW(NiPr)t fragment. The isotopic pattern is consistent with a moiety containing three chlorines. The only organic ion observed in the EI spectrum occurs at mlz=41, corresponding to the fragment [CH3CNt. Another tungsten fragment at mlz=306 demonstrates a high relative abundance (-80% ). This ion corresponds to the [ChW(NH)t fragment. The only other significant peaks occur at mlz=291 and 326. These represent the loss of both the nitrile and imido ligands to form [Ch Wt and [Cl4 Wt, respectively (abundances for both < 30% ). Only two peaks of consequence are observed in the NCI experiment (Figure 52b) The first occurs at m/z=383 (-42% abundance), and involves the loss of the acetonitrile ligand to form [Cl4W(N;Pr)r. The base peak of the NCI spectrum appears at mlz=340, corresponding to the formation of [Cl4 WNr. Thus, the NCI spectrum indicates cleavage of both the nitrile ligand and the isopropyl moiety to afford a tungsten nitrido fragment. This tungsten nitrido fragment is desirable, as it will ostensibly lead to deposition of W and N in the films. The benzonitrile complex lb was also subjected to similar MS analysis. Again, the positive ion EI spectrum contains a base peak at m/z =348, corresponding to the [Ch W(NiPr)t fragment. All of the other tungsten-containing fragments observed for the acetonitrile complex were also found in the spectrum of la. Interestingly, the NCI

PAGE 224

215 experiment again exhibited the base peak at mlz=340 ([Cl 4 WNr); however, only a small amount (-1% relative abundance) of the [Cl 4W(N;Pr)r fragment was observed. Table 5-1 summarizes the fragments and the relative abundances observed for la and lb. The mass spectral data collected for the tungsten imido complexes offer some perspectives about these systems as CVD precursors. Most importantly, the cleavage of the isopropyl moiety from the imido group is relevant due to the necessity of this process in forming tungsten nitride films. Furthermore, the absence of a molecular ion in either the EI or CI mass spectra suggests facile loss of the nitrile trans ligand. Observation of the [WC4t and [WCht fragments may be problematic since this indicates loss of the W-N bond; however, similar fragments were not present in the NCI spectrum Thus, these fragments may be a result of the more energetic EI process. Based on these results and the deposition of tungsten nitride from the imido complexes, there appears to be a correlation between the mass spectral data and behavior under CVD conditions. Table 5-1. Summary of Relative Abundances for Positive Ion EI and Negative Ion NCI M S f T lmid P 1 d lb a ass ipectra o ungsten o recursors aan EI Fragments mlz la lb [ChW(N1Pr)t 348 100 100 [C4Wt 326 26 41 [Cl 3W(NH)]+ 306 78 90 [ChWt 291 30 52 [CJisCNt 103 n/a 97 [CH3CNt 41 24 n/a CI Fragments m/z la lb [Cl4 W(N1Pr)r 383 42 1 [Cl4WN]340 100 100 a Relative abundances were adjusted by summing the observed intensities for the predicted peaks of each mass envelope and normalizing the largest sum to 100%

PAGE 225

100 G) u 75 C ca "C C :::, .c 50 < ca 'ii a: 25 0 100 G) u C 75 ca "C C :::, .c < 50 :; 'ii a: 25 0 216 348.0 [Cl3 W ( N1Pr)]+ a 306.0 41.0 [Cl3WNH]+ [CH3CN]+ 291.0 326.0 [Cl3W]+~ [Cl4W]+ 100 200 300 m/e 400 500 600 340.0 [Cl4WN] b 383.0 [Cl4 W(N1Pr)] 320 330 340 350 360 370 380 390 400 m/e Figure 5-2 Mass spectra of C l 4(CH3CN) W (NiPr). a) Posit i ve ion electron-impact ionization (El) (b) N e gative ion electron capture chemical ionization (NC I )

PAGE 226

217 5.4 Film Structure 5.4.1 XRD Results The XRD plot in Figure 5-3a indicates amorphous film deposition at 450C, evidenced by a flat profile without any characteristic FCC J3 -WNx peaks. Figure 5-3b depicts polycrystalline film deposition at 700C. Four characteristic peaks are evident, with relative peak intensities indicating that no preferred crystal orientation exists. Although the relative peak intensities in Figure 5-3b are consistent with the pattern for polycrystalline J3-WN0 5 the 20 peak positions lie between the standard values for J3-WN0 5 and J3-WC0 6 which are shown in Figure 5-3c. Peak positions between these standard values suggest that carbon is mixing with nitrogen and vacancies on tungsten's interstitial sublattice to form '3-WNxCy polycrystals. For the spectrum in Figure 5-3b, primary reflections at 37.42 and 43.12 20 degrees are consistent with (111) and (200) J3 -WNxCy growth planes, while additional reflections at 62.52 and 75.22 20 degrees indicate (220) and (311) planes, respectively. No peaks consistent with the hexagonal WN or WC (or WNxCi-x) phases or the BCC W phase were evident for any of the films. Figure 5-4 shows the evolution of film crystallinity with deposition temperature. At the two lowest deposition temperatures (450 C and 475C), the characteristic '3-WNx peaks are not observed. At 500C, a broad peak appears near 37 .94 20 degrees, indicating polycrystalline J3 -WNxCy (111) growth. This is similar to a previous report that amorphous WNx films transform into polycrystalline J3-W2N between 450 and 500C [Gal97]. As the deposition temperature increases to 550C, this peak sharpens and a broad peak at 44.42 20 degrees appears, indicating J3 -WNxCy (200) growth. Films deposited between 475 and 600 C displayed two additional peaks at 33.03 and 61.67 20

PAGE 227

218 degrees representing Si (200) Ka and Si (400) KP radiation, respectively. Broad peaks emerge at 63.00 and 75.82 20 degrees for growth at 650C indicating P-WNxCy (220) and (311) growth The peaks sharpen further as the temperature approaches 700C indicating polycrystalline grain growth. The formation of polycrystalline films is highly undesirable, since grain boundaries are facile paths through which Cu can diffuse to the underlying Si. Hence, deposition below 500 C with this precursor is required to grow an amorphous film. ,__500 !!) a .... Si ( 400 )--. ;3400 >, 300 :E $ 200 0 ~ 100 = 0 3 0 40 5 0 60 70 8 0 2 0 D e gr ees ,__ 500 !!) f3-WNx C y f3 WN C b a ::> 400 (111) X y Si ( 400 ) [ 3 00 't (2 00 ) :E I f3WN xCy $ 2 00 (220 ) >, ... f3-WNx C y t ~ 100 .s 0 ( 311) 3 0 40 5 0 6 0 70 8 0 20 D egrees 2'100 I C a ~ -WN05 :..-~-WCo 6 ~-W N o s ::> 8 0 ~ WN05 [ 60 (111) I (2 00 ) (22 0 ) I ~-WN o s ~-WCo6 (311) :E ~ -WCo. 6 I ~ WC o 6 1 < 40 I '-' ( 1 1 1 ) I I ( 200) (22 0 ) (311) I >, I ~I ... I ;;; 2 0 I = I I I .s 0 3 0 40 50 60 7 0 80 2 0 D egr ees Figure 5-3 XRD spectra for films grown on Si (100) in a H2 atmosphere. a) 450 C b) 700 C c) Standard powder diffraction plots for P -WNo. s and P WC0 6

PAGE 228

,-.._ "' .... .... 2500 219 T=675 C 8 2000 ~-..,,,-~-......,--.:T..._=..,;65..,0...,C.;....."'""~""'"-~ j T=600C 1500 .._, ;:... .... .... "' c: 1000 B ..s 30 T=550 C T=500 C T=475 C T=450 C 40 50 20 Degrees 60 70 80 Figure 5-4 Change in XRD pattern with deposition temperature for films grown with lb on Si (100) in a H2 atmosphere. 5.4.1.1 Lattice Parameter The dependence of lattice parameter on deposition temperature was determined by XRD using the 20 position of the B-WNxCy (111) diffraction peak, with peak position calibrated to the Si (100) diffraction peak. Figure 5-5 shows the change in lattice parameter with deposition temperature. The dashed line at 4.126 A represents the standard lattice parameter value for B-WNo.s, while that at 4.236 A is for B-WCo.6Error bars indicate uncertainty in lattice parameter ( 0.002 A) due to X-ray Ka line broadening [JCP88]. Figure 5-5 shows the lattice parameter increasing from 4.11 to 4.17 A as deposition temperature increases from 500 to 650C, and then decreasing to 4.16 A at 700 C. Shifts in the lattice parameter reflect either a composition change in the polycrystals or a change in the residual film stress Since AES gives bulk film concentrations, compositions in the bulk polycrystals and at the grain boundaries cannot

PAGE 229

220 be resolved. Shifts in lattice parameter cannot therefore be attributed solely to film composition without considering the possible influence of residual film stress. But, if we assume compositional variation dominates the shift in lattice parameter, the likely cause for lattice expansion between 500 and 650C would be incorporation of additional C into the polycrystals Lattice dilation for W-C mixtures relative to pure W has been reported, with lattice parameter increasing by more than 3 % with C content of 22.5% [Pau92]. 4 .24 ,----------------------, 4 .22 4.20 v:, bO Cl < 4 18 '-" 2 OJ 4 16 C-:1 4.14 A. OJ -~ 4.12 ... ...:I 4 10 -7------f3-WC0.6 4 .08 ...__-..--------,-------r------r-----.-----' 500 550 600 650 700 Deposition Temperature (0C) Figure 5-5 Lattice parameters for films grown from lb based on f3-WNxCy (111) diffraction peak. As shown in Figure 5-10, only the C content increases continuously with temperature in this range The existence of a continuous FCC f3-WNxCy solid solut i on has been predicted [Jon93], hence as x decreases the lattice parameter of f3-WN xCy polycrys t als would increase from a= 4.126 A at x = 0.5 (f3-WN0 5 ) to a= 4 236 A at x = 0 (f3-WC0 6). Above 650 C, the lattice parameter decreases. Again, this may indicate a change in film stress or a change in film composition. If we again assume that compositional variation

PAGE 230

221 dominates, we note that N and O levels are essentially unchanged in this temperature range, while C levels continue to increase with temperature. Lattice parameter decrease may indicate a solubility limit for C in the P-WNxCy polycrystals at 650C, with additional C incorporated above 650C preferring to reside in grain boundaries rather than bulk polycrystals. 5.4.1.2 Average Grain Size Two main causes of peak broadening in XRD thin film spectra have been reported: non-uniform film stress and the presence of small crystal grains. If the film stress is determined to be uniform, peak broadening may be attributed simply to the presence of small crystal grains. Likewise, if the presence of small crystal grains is ruled out, peak broadening may be attributed simply to non-uniform film stress. If we assume only uniform film stress, the average crystallite size (t) may be estimated using the Scherrer equation (Equation 5-1) [Cul78, NagOO]: 0.9). t=---(Bcos0) (5-1) where A is the X-ray wavelength (1.5406 A for Cu Ka), B is the full width at half maximum (FWHM) for the selected diffraction peak and 0 is the Bragg angle for that peak. The dominant diffraction peak for the films, P-WNxCy (111) was used as the reference peak for FWHM determination As depicted in Figure 5-6, grain size increased with deposition temperature, varying from 32 to 67 A over the 500-700C temperature range. Below 500 C, the films were X-ray amorphous, inferring that amorphous films had a maximum grain size near 30 A In comparison, Leedy et al. [2000b] reported grain sizes of 50 to 100 A for sputter-deposited WNx films, while Jun [1996] reported an average grain size of 30 A for TaC xNy films deposited by MOCVD at

PAGE 231

222 350C. The incorporation of C into W barrier films results in much smaller grain sizes relative to films containing pure W metal [WanOlb]. The presence of N and O has also been reported to hinder grain growth, resulting in smaller grain sizes [SheOOb]. The levels of N and O drop off with deposition temperature, though, and C becomes the dominant contaminant in films deposited at and above 500C. Thus, as deposition temperatures increase, a competition between increased grain growth due to higher surface diffusivity and decreased grain growth due to increased C content occurs. At lower growth temperatures, enhanced surface diffusivity due to temperature dominates the grain growth process, causing grains to increase with deposition temperature. Above 675C, though, grain size levels off, indicating that large scale C contamination is retarding the grain growth that would normally occur with increased temperature. 450 500 550 600 650 700 Deposition Temperature (0C) Figure 5-6 Average grain size for polycrystalline films based on the ~-WNxCy (111) diffraction peak. Error bars reflect uncertainty of 0.1 20 degrees in FWHM measurements.

PAGE 232

223 5.4.2 TEM Transmission electron microscopy (TEM) may be used to determine the presence and size of crystallites in the film The transmission electron diffraction (TED) image in Figure 5-7a shows that a film deposited with lb at 450 C on Si (100) had an amorphous structure, as evidenced by the blurred pattern without rings or spots, which is consistent with XRD measurements In contrast Figure 57b shows diffuse rings for the film deposited at 700 C, indicating that this film had some degree of crystallinity in agreement with XRD measurements. Figure 5-7 TED patterns for films deposited with i-Pr on Si (100) a) 450C. b) 700 C. Figure 5-8 TEM micrographs for film deposited on Si (100) w i th lb at 450 C.

PAGE 233

224 TEM micrographs for the film deposited at 450C are shown in Figure 5-8. These images indicate a -60 A bright interface layer between the Si (100) substrate and the deposited film. This interface contained a mixture of WSi2, WO3, SiO2 and Si 0, which is discussed further in the XPS analysis section below. TEM micrographs for the film deposited at 700C are shown in Figure 5-9. Polycrystal grains are evident in the micrograph on the left, in agreement with XRD results, along with some roughness at the top of the film. These images also indicate a bright interface layer (-50 A thick) between the Si (100) substrate and the deposited film. The edges of this interface layer are not as sharply defined as for the 450C film, which may point to increased interface roughness for the higher deposition temperature. This agrees with SIMS depth profile results (Figure 5-21 and Figure 5-22), which indicate a longer overlap time for the W and Si signals in the 700C film relative to the 450C one. Figure 5-9 TEM micrographs for film deposited on Si (100) with lb at 700C.

PAGE 234

225 5.5 Film Composition 5.5.1 AES Auger results in Figure 5-10 indicate that W, N, C and O were present in the deposited films, while Cl was not detected. Between 450C and 500C, the C level is fairly constant at -12 at. %. Between 500 and 650C, C content increases steadily from 12 to 43 at. %, and then increases mildly to 49 at. % at 700C. The overall trend reflects the increasing tendency of the hydrocarbon groups present in the precursor ligands and the solvent to decompose with increasing deposition temperature. Fragmentation of ligands and solvent leaves carbon-containing moieties at the film surface, resulting in C incorporation into the growing film. 100-,----------------------, i::: ..... 40 .:i i::: 8 20 u 400 450 500 550 600 650 700 750 Deposition Temperature (C) Figure 5-10 Variation of W, N, C and O content in films from lb with deposition temperature. Data are from AES after 2.0 minutes sputter. The N content in films grown at 450C was 8 at.%, and this rose slightly to 11 at.% at 500C. Above 500C, the N levels decrease, dipping to -3 at.% at 700C. The higher N levels at and below 500C reflect the stability of the W-N multiple bond in the

PAGE 235

226 precursor molecule, which likely endures at deposition temperatures up to 500C, inhibiting release of N into the gas phase during deposition. The drop in N above 500C may indicate decomposition of the W-N multiple bond in the gas phase and/or increased N desorption from the film (to form N2 gas) at higher temperature. Substantial N release from WNx films at higher temperature (-700C) has been reported [Kat85], consistent with AES results in this range. Lack of a load-lock and vacuum transfer system necessitated a vacuum break after barrier film deposition. Oxygen contamination resulted from post-growth exposure of the film samples to air. Incremental AES sputtering showed a steady decrease in 0 levels with increasing depth into the WNx films. The O concentration was highest at 450C, reaching 22 at. %, and decreased slightly to 18 at. % at 475C. The amorphous films deposited below 500C had low density and high porosity, allowing substantial amounts of oxygen to penetrate into the film lattice. X-ray reflectivity (XRR) indicated that the film deposited at 450C, for example, had a density of 8.6 g/cm3 High oxygen concentrations (-20%) attributed to air exposure have also been reported for porous TiN, TiC and TiCN barriers [Eiz94a, Par96, Wan0la]. As the deposition temperature was increased from 475 to 500C, the O content drops sharply down to 7 at. %, while the C and N levels are fairly steady. This indicates crystallization of the film in this temperature range. As the film crystallizes, its microstructure densifies, thereby inhibiting post-growth oxygen diffusion into the lattice [Jos02, Pok91]. XPS results for oxygen in the films are consistent with WO3 which has considerably higher thermodynamic stability than f3-WNx or f3-WCx. For example, reported values of the Gibbs energy of formation (~G0r) at 700C for the WO3, f3-WN0 5 and f3-WC0 6 phases

PAGE 236

227 are -592 kJ/mol, +17 kJ/mol, and -11 kJ/mol, respectively [Lid85, Lak79, Gus86]. The experimental observation of lower levels of oxygen at higher deposition temperatures is consistent with post-growth oxygen contamination. As deposition temperature increases above 500C, the O concentration drops further, falling to 2 at. % at 700 C. This results from film densification (by polycrystal grain growth) and increased C levels at higher deposition temperature which stuffs the grain boundaries and blocks O in-diffusion. Porosity of amorphous films grown below 500C may be problematic, as defects in the film can degrade the barrier's resistance to Cu diffusion. A previous report, however indicates that diffusion barrier performance depends more strongly on film microstructure than film density [Kim99c] 30 ...--------------------. 25 -AW IN --W IC ......... W I O 20 0 .:;j 15 10 5 0 450 500 550 600 650 700 Deposition Temperature (0C) Figure 5-11 Variation of tungsten to nitrogen (WIN), tungsten to carbon (W/C) and tungsten to oxygen (W/O) ratios with deposition temperature for films from lb. Data are from AES after 2 0 minutes sputter. Figure 5 -11 shows the ratios of W to C N and O in the deposited films. The WIN ratio was quite high ranging from 6 to 18, which is much higher than the des ir ed ratio of

PAGE 237

228 2 for P-W2 N. The trend followed by the WIN ratio is expected, since higher deposit i on temperature enhances N desorption from the films to form N2 gas, causing WIN to increase. Although the general trend of N content with deposition temperature is discernible from the AES data, exact measurement of the N level is difficult. Preferential incorporation of C and removal of N by Ar+ sputtering during AES ana l ysis has been reported to cause artificially high C and low N concentration readings [Ing82]. Since the AES data were collected after 2.0 minutes of sputter, N atoms were likely sputtered and replaced by C atoms, causing the WIN and W/C ratios to be artificially high and low, respectively. Moreover, Lee et. al [1994a] reported P-W2N growth per XRD while their AES results showed only 12 at. % N. This was attributed to a shift in the sensitiv i ty factor for N in the WNx films away from the value for pure elemental N. In addition, Alay et al. [1991] studied the N content of sputter deposited amorphous WNx fi lms with a variety of characterization techniques. For a given WNx film, the N content (per AES) was -13%, while Rutherford backscattering spectrometry (RBS) indicated -33%. Since RBS does not require an elemental standard for calibration and is considered accurate to -1 at. % [Ohr92], these results may indicate that our actual N levels, especially at lower deposition temperature, are considerably higher than indicated by AES. The decrease in W/C with increasing deposition temperature reflects the enhanced decomposition of hydrocarbon moieties from the precursor and sol vent with temperature. The W/C ratio drops to 1 at 700 C due to large-scale alkyl decomposition. The W/O ratio increases from 3 to 10 between 450 and 500C due to the aforementioned film crystallization and then rises from 10 to 27 between 500 and 650 C This increase is related to the jump in C content in this temperature range (Figure 5-10), where C further

PAGE 238

229 stuffs the grain boundaries, physically blocking oxygen in-diffusion. The WIN and W/0 ratios reached their maximum values at 650C, and decreased above this temperature. The decrease in these ratios above 650C does not reflect an increase in C and O content in the films, but rather a decrease in the W content of the film, due to massive C contamination. 5.5.2 AES Depth Profiling The three-point AES depth profiling method was used to determine the change in elemental content with depth into the films. As evident in Figure 5-12, films deposited at 450C contained significant levels of 0, due to high film porosity enabling substantial 0 in-diffusion after air exposure. Note that due to use of the three-point depth profiling method, the y-axis in Figure 5-12 is not concentration in film, but is proportional to the area under the curve for the elemental signal. The oxygen level initially dropped with increasing sputter depth, but increased at the barrier-Si interface. This suggests that the barrier film near the interface is very porous and defective, which may be a consequence of film growth dynamics during the nucleation process on Si. HCl, formed by reaction of H2 carrier gas (or Hon the precursor ligands) with Cl from the precursor, may etch and roughen the Si interface before and during the nucleation process. Moreover, steric effects may inhibit the ability of the precursor to uniformally cover the Si surface, all of which can lead to voids and stresses at the interface. The defects would leave substantial amounts of exposed W with dangling bonds, which readily reacts with O diffusing in from the film surface. The makeup of this interface layer is discussed in more detail in the XPS film analysis section below. After the nucleation stage, the films are growing on WNx (or WNxCy) that has already deposited, and this decreases the concentration of

PAGE 239

230 defects (porosity), leading to lower O levels in the middle of the film. The O level is highest at the film surface due to direct contact with air. Apart from 0, a high level of C was evident on the film surface, indicating the presence of both deposited C and adventitious C. After sputtering off the initial film layers, the C level dropped to a steady level, with magnitude similar to that for N. The W level is initially steady, and drops gradually at the approach to the Si interface, due to increasing O levels. The O and Si signals overlapped briefly, but the W and Si signals overlapped some additional time. This suggests that the barrier-Si interface may be rough, which could be caused by HCl etching during deposition. In addition, the W-Si overlap may indicate that some amount of WSix may be present at the interface 40000 ~-------------------~ ,,-., en ..... ..... 30000 .= ..... 20000 "-' 0 Si ..... en = 10000 Q) ..... 0 5 10 15 20 Sputter Time (min) Figure 5-12 Three-point AES depth profile for film deposited from lb at 450 C. Nominal AES sputter rate was 100 Almin Figure 5-13 shows the depth profile for a film deposited at 500 C As expected the O level decreases relative to the 450 C film, due to crystallizat i on (and densification)

PAGE 240

231 of the film at 500C, which decreases the degree of O in-diffusion. The O level still increases near the barrier-Si interface, however, for reasons discussed above, but the 0 "bump" near the interface is narrower, indicating that the defect-laden layer at the interface thins with increasing deposition temperature. Increased deposition temperature improves surface diffusion of depositing species, which would cause defect (and void) density to decrease. The C signal increases slightly relative to the 450C film, due to increased C contamination at higher temperature. Overlap of the W and Si signals occurs for longer sputter time relative to the 450C sample, and may indicate increased roughening of the Si surface (due to HCl etching) and/or increased WSix formation with higher temperature. The increased sputter time relative to the 450C sample reflects increased film thickness and higher film density for the 500C sample. 30000 --25000 Cl.) .... .... = 0 20000 !:I .... 15000 '-" :>-. .... 10000 .... en = (I.) .... = -5000 0 0 10 20 30 Sputter Time (min) Figure 5-13 Three-point AES depth profile for film deposited from lb at 500C. Nominal AES sputter rate was 100 Almin.

PAGE 241

232 30000 w /""'-. 25000 r.l.l ..... ..... i= 20000 b .... 15000 '-' >, ..... 10000 .... r.l.l i= d) ..... i= 5000 0 0 20 40 60 80 Sputter Time (min) Figure 5-14 Three-point AES depth profile for film deposited from lb at 700C. Nominal AES sputter rate was 100 A/min. Figure 5-14 shows the depth profile for a film deposited at 700C. Oxygen levels through the film are decreased relative to the two lower temperatures, due to densification by polycrystal grain growth and plugging of voids/grain boundaries by additional C in the films A very small O "bump" at the barrier-Si interface is visible, indicative both of a thinner defective region and decreased O in-diffusion. The N level is steady, but decreased relative to the 450 and 500 C films, due to increased N desorption at higher deposition temperature. Carbon levels are higher relative to the lower temperature films, which again reflect the increased ability of the ligands and solvent to decompose at higher temperature. Overlap of the W and Si signals occurs for longer sputter time relative to the lower temperature samples, which again may indicate increased roughening of the Si surface due to HCl etching and/or increased WSix formation with higher temperature The i ncreased sputter time relative to the lower temper ature films

PAGE 242

233 again reflects increased film thickness and possibly higher film density for the 700C sample. 5.5.3 XPS Standard binding energies for elements present in the films from this study are shown in Table 5-2 below. The sensitivity factor relates to an atom's ability to emit a photoelectron when exposed to X-ray radiation, and is relative to the Fis line, which is l.0eV. Tabl 5 2 S d d El e tan ar ementa a ues or 1m 1 XPS V 1 Fl C Element Binding Energy ( e V) W4f 31.32 Cls 284.5 (C in Graphite) Nls 398.5 (N in BN) Ols 531 0 (0 in A}iO3) *References: [Wag90 Cri00] 5.5.3.1 XPS Literature Data omponents Sensitivity Factor (S) 2.75 0.25 0.42 0.66 XPS survey spectra for W typically show peaks at many binding energies, with each representing photoelectrons from a specific energy level in the W atom. The W 4f photoelectron line has two peaks (W4f712 and W4f512), while the Nls, Cls and Ols lines each have one. Binding energy ranges for the f and d subshells of some potential W compounds present in our films are shown in Table 5-3. As seen in Table 5-3 reported W4f712 binding energies for metallic W range from 30.9 to 32.7 eV, while W4f512 binding energies range from 33.2 to 34.8 eV. XPS data for two forms of tungsten nitride, the FCC ~-WNx and SHP WN forms, have been reported. W4fm and W4fs12 values for ~-WNx, the desired phase, range from 30.7 to 33.6 and 33.3 to 35.8 eV, respectively. Binding energies for the analagous tungsten carbide phases (FCC and SHP) have also been reported, with SHP WC having a W4f712 ranging from

PAGE 243

234 31.4 to 32.8 eV. FCC WC1-x has been reported with W4fm BE values ranging from 31.2 to 31.9 eV [Ji0l, Pau92], which are close to the value for metallic W, since Wand C do not have a strong covalent or ionic bond in this phase [Nak87]. The addition of Oto WC films has been reported to increase the W4f binding energies, due to formation of a WOxCy compound [Chi97, Ji0l]. Katrib [1994] reported a W4fm BE value for WOxC y T bl 5 3 L"t t B" di E a e -1 era ure m ng nergy a a or ec ron use sm Dt t: El t Sbhll. WC d ompoun s Binding Energy ( e V) for Subshell Compound W4fm W4fs12 W4d312 W4ds12 References [Ala91, Cha97b, Cha99, Col76, 30.9-33.2-242.6-CriOO, GirOO, Gru87 JiOl, Kak99, Metallic W 32.7 34.8 255.6 244.2 Kat95, LutOl, McG73, Meu98, Mou95, NagOO, Nak:87, Pau92, SheOOc-f, Wag90, Yeh96, Zha99] 30.7-33.3-[Ala91, Chi93, Col76 Lee93, WNx 33.6 -244.9 Meu98, NagOO, Nak:87, SheOOb-d 35.8 Tak97a,Zha97,Zha99] FCC 31.2-[JiOl, Kou02, Pau92] B-WCx --31.9 WO4 -37.6 --[Col76] [Bla86, Cha88, Cha97b, Cha97c 34.9-35.2-259.9-247.4-Cho02, Col76, CriOO, GirOO, GogOO, WO3a 38.0 38.7 260.0 248.4 Gru87 Kat95, Lec96, Lee98b, LutOl, McG73 NagOO, Nak:87, Rus99, Sar80, SheOOdf, Wag90, Zha99] W2Ol 34.5-36.7-[Kat95, JiOl NagOO] -35.7 37 8 31.1-33.1-[Cha88, Col76, JiOl, Kat95, LutOl, WO/ 34.8 36.9 256.1 243.2 Mon99, NagOO, Rus99, Sar80, Wag90] WO 32.9 35 --[NagOO] WOxCv 31.5 -[Kat94] WC16 37.1 --248.5 [McG73] WOC4 37.4 ---[McG73] WSiz 31.5 33.5 --[Cho02] aNote, this includes references to "surface tungsten oxide," which are taken to mean WO3 "Note, this includes references to "W5+," which are taken to mean WO2 _5 or W 2 O5 cNote, this includes references to "W4+," which are taken to mean WO2 dNote, this includes reference to "W2+," which is taken to mean WO of 31.5 eV. Bonding between Wand O in tungsten oxide compounds exhibits substantial electron transfer from metal to nonmetal, similar to the bonding in tungsten nitrides [Col76]. This electron transfer causes an increase in binding energy for W in oxides

PAGE 244

235 relative to the metallic state, with the increase frequently on the order of 1 eV per un i t change in oxidation state [Rig75]. Several WOx compounds, with x ranging from 1 to 4, have been reported in the literature, as indicated in Table 5-3. Binding energy for N atoms is strongly dependent on their environment in the WNx film. Care must be taken when determining WNx film compositions from XPS data, as large shifts in sensitivity factor for Nin compound materials have been reported [Lee93]. Nitrogen atoms are typically found either at octahedral i nterstit i al sites within the WNx crystal lattice or at the grain boundaries. T bl 5 4 L. a e -. 1terature m ng nergv ff di E D f: El ata or ectron Sbhll"NC u s e s m d ompoun s. Nitrogen Bonding Nls Binding References Environment Energy Range ( e V) [Cha97b-c Cha99 Chi93 Col76, Interstitial N in HonOO, Kak:99, Lee90, Lee93 13-WN x 396.7-398.3 LeeOOb, Meu98, Mou95, Nag93 Nak:87, SheOOb-c Yeh96 Zha97, Zha99] Interstitial N in SHP 398.2 [Zha97] WN W(O,N) x 398.1-399.3 [Zha99] Nin l3-C 3 N 4 398 5 [Wei99] Nin N-spjC 398.5-398.8 [JelOO, Nin02] Free N (grain 399 2-400.1 [Cha97b-c, Cha99 Nag93 Nak:87] boundary N) NinN-sp..:C 400.0-400.9 [JelOO, Nin0 2 ] Nin Al(N O) 400.1 [Si m94] Oxynitride Nin CNx 400.2 [Wei99] N in TiAl x Oxynitride 400.3 [Sim94] N in an N-0/NN 402.0-402.9 [CriOO, LeeOOb, Wei99] surface species "Dissolved" N 403.3-404.5 [Sim94] For WNx films Nls binding en e rgies ranging from 396.7 t o 3 98 .3 e V which ar e below the standard Nls value (398.5 eV) are typically reported as i nter st i t i al N (N bound to W). The drop in BE is c a us e d by e l e ctron trans fe r from W to N du e to the Wo+~

PAGE 245

236 bond having some ionic character [She00b]. Shifting of the N peak to BE values above the Nls standard value may be due to several different phenomena. N present at grain boundaries in the film would have higher BE because electron transfer from W to N at the grain boundary does not occur A lack of electron transfer means that Nls core electrons (for N at the grain boundary) do not decrease in binding energy [Cha97b-c, Cha99, Nag93, Nak87]. Another cause may be the reaction of O (after air exposure) with the WNx lattice to form a complex bivalent oxynitride such as W(N,O)x [Lai98b, Sim94 Zha99]. Oxygen atoms are more electronegative than N atoms withdrawing electrons more readily from the metal. This causes increased ionicity of the metal-oxide bond relative to the metal-nitride bond [Gub94] The addition of O to WNx films causes electrons to be withdrawn more strongly from W by 0, decreasing the net electron transfer from W to N and thereby increasing the Nls binding energy. One report from Prieto et al. [1995] suggests that Nls BE values are lower for oxynitrides relative to nitrides, but this conflicts with many other reports describing higher oxynitride BE values relative to nitrides [Lai98b, Sim94, Zha99]. Several lines in the standard XRD powder diffraction spectra for ~-WNo .5 and the oxynitride W(N,O)u3 overlap exactly due to similarities in their lattice structures, with O and N atoms sitting on interstitial sites in the FCC W lattice [JCP88, Sel95]. Hence, the presence of the oxynitride typically cannot be ruled out if some amount of O is present in the film and an Nls peak is present at higher BE. XRD spectra rule out the presence of another oxynitride (W(O,N)1.6J), however, which has a hexagonal structure [Che97 JCP88]. A third cause for increased Nls binding energy is the reaction of N with O near the film surface [Cri00, Lee00b, Wei99]. Nls BE values of 401.4 eV and 402 eV respectively, have been reported for N that has reacted with Oat the

PAGE 246

237 film surface [Cri00, Lee00b], while another report assigned a BE of 402 88 eV to surface N-O/N-N species [Wei99]. A fourth cause may be expansion of the WNx lattice due to C addition to form WN xCy, As discussed earlier, substitution of C for some N on the interstitial sublattice increases the film's lattice parameter relative to that for the nitride. This lattice expansion would strain the W&+ -Nobond, decreasing the metal to nonmetal electron transfer. The decreased electron transfer would cause the Nls binding energy to increase relative to its value in pure FCC B-WNx Yet another possibility to account for increased BE is the formation of N-C bonds in the film. Ning et al. [2002] reported that Nls peaks at 398.6 and 400.0 eV were consistent with formation of N-sp3C and N-sp2C, respectively. Another report suggested that ls peaks at 398.5 and 398.8 were for N-sp3C bonding while peaks at 400 and 400.9 eV were for N-sp2C bonding and a peak at 399.9 eV was for N bound to H [Jel00]. The presence of B-C3N4, which has a hexagonal structure homologous to Si]N 4 was suggested, as this phase is considered to have a mixture of sp3 and sp2 C-N bonds [Jel00]. Wei [1999] reported that Cls and Nls binding energies of 2 85.96 and 400 .17 eV, respectively indicated the presence of CNx In addition, Cls and Nls bind i ng energies at 287.32 and 398.46 eV, respectively, represented B-C3N 4 XPS values for a variety of potential C-containing compounds in our films are listed in Table 5-5. While graphite and amorphous ("free") C have similar BE ranges t he presence or absence of graphite can be confirmed by XRD. If graphitic peaks do not appear on XRD, the C is present in amorphous ( free") form. Diamond-like C, like graphite, is discernible from XRD analysis. Other potential bonding states include C bound to H, 0, or N and C in the oxycarbide WOxCy

PAGE 247

238 XPS values for a variety of potential O-containing compounds in our films are listed in Table 5-6. 0 bound to W has been reported in oxides CWOx) of varying stoichiometry, oxynitrides (W(N,O)x), oxycarbides (WOxCy), and in compounds containing C or H. A BE value for O in WOxCy was not reported [Kat94], however. T bl 5 5 L a e 1terature m ng nernv B" di E ata or D f El ectron Sbhll. CC use sm d ompoun s Carbon Bonding Cls Binding References Environment Energy (eV) C in SHP WC or FCC [Bon80, Col76, Gir()(}, Kat94, Kou02, Lai98b, WC1x 282.6-283.8 Lec96, Lut0l, Mon99, Pau92, Sun0lb, Wan0lb, Xue98] C in WOxCv 283.2 [Kat94] Graphitic C 284.2-285.0 [Ala91, Bon80, Kat94, Kou02, Lut0l, Mou95, Pau92, Wag90, Xue98] Free (amorphous, 284.5-285.2 [Bla86, CriOO, Gir()(}, GogOO, Kat94, Kou02 Lec96, adventitious) C Lut0l, Pau92, Wei99] C in C=C 284.5 [Nin02] C in C-H 284.8-285.1 [Bou91, Pau92] "Diamond-like" C 285.5-286.8 [Mon99] C bound to 0 285.3-291.9 [Bon80, Bou91, Chi93 GogOO, Jam92, Lec96, Wag90, Wei99] C in CNx 286.0 [Wei99] C in sp2C-N 286.1 [Nin02] C in P-C3N4 287.3 [Wei99] C in spjC-N 287.5 [Nin02] T bl 5 6 L' a e 1terature m ng nergy B" di E ata or D f El ectron s d ubshel s in O Compoun s Oxygen Bonding Ols Binding References Environment Energy (eV) Oin WO3 530.1-531.6 [Chi93, Col76, CriOO, Gir()(}, GogOO, Gru87, Kat94, Lec96, Lut0l, Sar80 Zba99] Oin WO2 530.2-530.5 [Col76, Sar80] 0 bound to C 531.3-534.9 [Bou91, Jam92, Lec96] "Free" 0 531.6-531.7 [Ala91, CriOO] 0 in oxynitride 531.7 [Pri95] 0 in OH group 532.6 [Lec96] 0 in H2O 532.6-533 [Bla86, Gar90, Lec96, Wag90] 0 in WOx 532.9 [Zha99] 0 in SiO2 532.5-533.0 [Mou95, Wag90]

PAGE 248

239 5.5.3.2 XPS Film Analysis Deconvolution of XPS spectra was performed with combined Gaussian and Lorentzian curve fitting (90%-10%) and Shirley background subtraction using the RBD 200000 0 (Auger) W4f i Ols ,-._ \ Cls '):4d \ en .... a 150000 W4p Nls l0min .b sputter .0 100000 I-< < '-' >, 5 min .... en sputter i::: 50000 B i::: -0--1-----~------~---~--1000 800 600 400 200 0 Binding Energy (eV) Figure 5-15 XPS survey spectra for films deposited at 450 C a) As received. b) After 5 minutes sputter. c) After 10 minutes sputter. ,-._ en .... 400000 8 300000 b 200000 '-' 0 en 100000 i::: W 4 p ,..,._, Cls \ ~4d Nls ~---..A.-...A '---01s W4f o ---~---~---~---~--1000 800 600 400 2 00 0 Binding Energy (eV) Figure 5-16 XPS survey spectra after 10 min sputter fo r films deposited at 450 to 700 C.

PAGE 249

240 Analysis Suite software. Due to the presence of C both at the film surface and in the film bulk, quantification of charging to an exact value of BE shift was difficult. XPS survey spectra for films deposited at 450C are shown in Figure 5-15. These spectra indicate the presence of W, N, C and O in the films, while a peak for Cl is not present. Some amount of adventitious C is evident on the as-received spectra, as the Cls peak decreases in height after sputtering. The Ols peak endures after sputtering, however, due to the porosity of these films. The W4d512 and W4d312 BE values in the as received film were 248 and 260.5 eV, respectively, which are consistent with WO 3 (Table 5-3). The W4ds12 and W4d312 BE values after 5 and 10 min sputter were 245 and 257 eV, respectively, which is indicative of WNx deposition. Figure 5-16 shows survey spectra after 10 min sputter for films deposited at 450 to 700C. The Nls and Ols peaks decrease with increasing deposition temperature, while the Cls peak increases with deposition temperature, for reasons discussed above The W 4ds12 and W 4d312 BE values are similar for all temperatures after sputtering, near 245 and 257 eV, respectively, which indicates the presence of WNx at all of the temperatures. To view the most intense elemental peaks more closely, multiplexes were taken for the W4f, Nls, Cls and Ols lines. XPS multiplex spectra for films deposited at 450C are shown in Figure 5-17. Focusing first on the sputtered films, we see that the elemental peaks have similar positions after 5 and 10 minutes of sputtering. The spectra in Figure 5-17a taken after 10 minutes sputter shows two sets of overlapping W4f binding energy curves, which indicate W in two different bonding states. The W 4f 712 and W 4f 512 peaks at 35.7 and 37.8 eV fall into the range for WO3 (Table 5-3), with the 1.9 eV spin orbit

PAGE 250

241 splitting for these peaks being close to 2.0 eV reported previously for WO3 [Sar80]. This is supported by an Ols peak at 531.4 eV, which is consistent with WO3 (Table 5-3). As 100000 W4f7/.l W4 80000 .... = lOmin j 60000 sputter .... '---' 40000 0 sputte .... Cl) = B = 20000 -As dep 0 44 42 40 38 36 3 4 Binding Energy (eV) 14000 lOmin 12000 ,___ sputter 10000 f 8000 :E $ 6000 0 .... Cl) j 4000 .s 2000 290 288 286 284 Binding Energy (eV) a 32 30 Cls C 2 8 2 ----Cl) .... s .b ... .D < '---' 0 .... Cl) = (!) .... = ..... ,,-.._ Cl) .... s 10000 Nls lOmin 8000 6000 5 min 4000 sputter 2000 b 406 404 402 400 398 396 Bind i ng Energy (eV) 70000 ~----------~ 01s 60000 l0min 50000 sputter 5 min 40000 sputter .b 30000 < V '---' 1/ o 20000 .... I Cl) = Asdep I B 10000 I = 7-..... 0 ......_ ----d 5 3 6 5 3 4 532 5 3 0 5 2 8 Binding Energy ( eV ) Figure 5 -17 XPS spectra for films deposited from lb at 450 C for the a) W4f b) Nls c) Cls d) Ols BE lines. Dashed l i nes are deconvoluted peaks. mentioned earlier 0 in the films is due to post-growth air exposure where O diffuses into the film and oxidizes unsaturated W atoms Oxygen is reportedly mobile enough at

PAGE 251

242 ambient conditions to diffuse into surlace or near-surlace region of tungsten oxynitride, with chemisorption of oxygen to this surlace causing continuous formation of WO3 [Cho00]. The behavior of our films in air is likely similar to the oxynitride films. As seen for the sputtered films in Figure 5-17 a, the W 4f 712 and W 4fs1 2 peaks at 32.8 and 34.6 eV fall into the range for B-WNx, but are believed to actually represent B-WNxCy. BE values for B-WNxCy have not been reported in the literature, but several things point to this possibility. First, the single Nls peak at 398.4 eV is shifted above the standard values for B-WNx to a value similar to that for N-C bonding (Table 5-4). Second, a Cls BE value at 285.8 eV supports the presence of C-N bonding (Table 5-5), and is considerably higher than reported values for carbidic C. Third, the upward shift for W4f BE away from metallic or carbidic W (Table 5-3), coupled with nitridic peak positions for W4d, suggest W-N bonding. Since no metallic or carbidic W peaks were observed, the presence of metallic W or B-WCx clusters in the film is unlikely [Mon99]. These items therefore point toward formation of B-WNxCy, which is also supported by our XRD lattice parameter data for films deposited at higher temperature. Moreover, the upward shift in Nls and Cls binding energy in B-WNxCy suggests that these elements have an attractive interaction on the FCC W interstitial sublattice. In an analogous case, attractive N-C interactions were reported on the FCC Nb and FCC Ti interstitial sublattices [Lee0la]. The large FWHM value for BE of Nls electrons suggests a disordered or defective structure, which is consistent with amorphous (or nanocrystalline) films [Pri95]. The lack of a second Nls peak above 399 eV suggests that all N is bound to W, with none present in its "free" form.

PAGE 252

243 The additional Cls peak at 284.6 eV falls into the range for either graphitic or free (amorphous) C (Table 5-5). Since XRD results do not indicate the presence of graphite in the films, this Cls value represents free (amorphous) C. A small, additional Ols peak at 532.4 eV is likely due to the presence of a suboxide (WOx), While the suboxide W4f peaks are convoluted with the WO3 and P-WNxCy peaks, their exact location is unknown due to lack of an additional shoulder for them in the W4f spectra When compared to the sputtered films, the as received films in Figure 51 7 have significant BE shifts and additional peaks. The surface region of the as-received films contain significant amounts of WO 3 evidenced by the large W4f712 and W4fs 1 2 peaks at 37.4 and 39.4 eV, respectively, which corroborate the W4d BE values i n the as-received survey spectrum (Figure 5-15). An Ols peak at 531.9 eV also supports the presence of WO 3 A significant amount of non-conductive WO 3 at the film surface leads to sample charging, which causes the shift in BE to higher values [Gru87, NagOO]. The W4f712 BE, in particular, increases by 1.7 eV from 35.7 eV after 10 minutes sputter to 37.4 eV at t he surface. The area under the WO 3 peak decreases after sputtering but does not disappear completely, which is consistent with a previous report on ambient air oxidation of a tungsten oxynitride surface [Cho00] Decreasing the as-received BE values for W, C and Oby 1.7 eV leads to a close match to peak positions after 5 min and 10 min sputter. With this in mind, the W4f712 and W4fs12 peaks at 35.2 and 37.1 eV would suggest P-WNxCy that has been shifted upward by 1.7 eV. The corresponding Nls value for nitridic Nat 401.3 eV would be 399.6 eV after a downward shift of 1.7 eV, however, which is still higher than that for Nls in P-WNxCy, This suggests that a tungsten oxynitride (W(O,N)x) or oxycarbonitride

PAGE 253

244 (W(O,C,N)x) is present at the surface before sputtering. This is supported by an Ols peak at 533.5 eV, which would be 531.8 eV after a downward shift of 1.7 eV, consistent with 0 in oxynitride (or possibly oxycarbonitride). A second Nls peak at 403.0 eV, which would be 401.3 eV after a 1.7 eV shift, is close to the value of 402.0 eV reported for Nin a surface N-O/N-N compound [Wei99]. The two Cls peaks at 286.2 and 287.4 eV, when shifted down by 1.7 eV, are 284.5 and 285.7 eV, consistent with free C and C in ~-WNxCy, respectively. Table 57 summarizes the BE values for the elemental peaks shown in Figure 5-17 and the compounds to which these BE values correspond (based on comparison of these values to Table 5-3 through Table 5-6) Question marks in Table 57 indicate that exact location of the corresponding W4f peaks for the WOx suboxide is unknown, due to lack of additional shoulders in the W4f spectra Tabl 5 7 C e -d Fl D ompoun s m 1 ms df epos1te rom th 450 c at Sputter Time W4f712 W4fs,2 Nls Ols Cls Compound (min) (eV) (eV) (eV) (eV) (eV) 0 37.4 39.4 -531.9 WO3 0 35.2 37.1 401.3 533 5 287 4 8-W(N,C,O)x 0 403.0 N-0/N-N ----surf ace species 0 ---286.2 FreeC 5 33.1 35.0 398.7 -286.3 13-WNxCv 5 36.0 38.4 -531.7 WO3 5 ----284.6 FreeC 10 32.8 34.6 398.4 -285.8 8-WNxCv 10 35.7 37.8 -531.4 WO3 10 ? ? -532.4 WOx 10 ----284.6 FreeC The interface layer evident in the TEM and AES Depth Profiling sections was also examined with XPS to determine the bonding states in this layer. The film was

PAGE 254

245 sputtered until the interface layer was exposed, and the multiplex was then collected. XPS spectra in Figure 5-18 indicate the presence of W, Si and O in the interface layer, 1000 800 a ::i 600 E 400 -~ 200 <: !l .s 0 a b 42 40 38 36 34 32 30 28 108 106 104 102 100 98 96 Binding Energy (eV) Binding Energy (eV) 6000 ..-----------------, Ols 0 C 536 534 532 530 528 Binding Energy (eV) Figure 5-18 XPS spectra for the film-Si subsrate interfacelayer for film deposited from lb at 450C. a) W4f b) Si2p312 c) Ols BE lines. Dashed lines are deconvoluted peaks. while the Cls and Nls BE lines were absent. The W4f712 and W4fs12 BE values at 34.4 and 36.7 eV, coupled with an Ols BE at 530.8, are consistent with WO 3 (Table 5-3 and Table 5-6). The Si 2p 312 and Ols BE values at 103.5 and 532.9 eV, respectively, indicate SiO2 (Table 5-6). W4f712 and W4fs12 BE values at 31.5 and 33.5, coupled with the Si 2p312 peak at 99.5 eV, suggests WSi2, where the Si 2p31 2 BE value is consistent with silicide bonding [Mou95] This indicates that some W reacts with the Si substrate. A mixture of WSi2 and metallic Wis also possible, however, as their W4f peaks overlap. The small

PAGE 255

246 peak at 101.4 eV is close to a reported BE of 101.7 eV for SiO [Ngu89], which may form due to the reduction of SiO 2 at the interface by the XPS Ar+ sputter beam. The corresponding Ols BE peak for SiO is likely convoluted under the two larger Ols peaks in Figure 5-18. Table 5-8 summarizes the BE values for the elemental peaks shown in Figure 518 and the compounds to which these BE values correspond (based on comparison of these values to Table 5-3-Table 5-6). The question mark in Table 5-8 indicates that exact location of the corresponding Ols peaks for the SiO suboxide is unknown, due to the likely convolution of this peak under the two larger Ols peaks for SiO 2 and WO 3 Table 5-8. C d Fl s b rf i D ompoun sat 1 m-1 su strate mte ace or h lb at 450C epos1t1on wit W4f112 W4fs12 Si2p312 Ols Compound (eV) (eV) (eV) (eV) 31.4 33.4 99.5 WSii 34.4 36.7 -530 8 WO3 -101.4 ? SiO --103.5 532.9 SiO2 The bulk of the O in the films appears to be from post-growth in-diffusion, as evidenced by the decrease in O content with increased crystallinity and C content. The presence of SiO2 at the interface suggests, however, that a small amount of background oxygen (from 02 or H2O) in the reactor may be causing reformation of the Si02 layer before film deposition. The source of the background O is unclear, but may be from an impure N2 gas source (used during the reactor heating cycle before the H 2 bake) or from a small air leak in the system. This background level is believed to be small, however, because the presence of considerable amounts of O during reaction will favor WO3 formation, especially at higher deposition temperature. Deposition of WO3 films in our

PAGE 256

247 system, evidenced by high intensity WO3 XRD peaks, has occurred on several occasions due to large-scale air leaks during deposition at a variety of temperatures. The interface layer thickness was similar for deposition at 450 and 700C, but the content of the layer changes with increasing deposition temperature, as indicated in the AES depth profiles. The decrease in O level at the interface for deposition at 700 C (Figure 5-14) may be due to higher temperature shrinking incubation time, which leaves less time for background O to react with the Si substrate before film deposition beg i ns. Lower O content at the interface for the 700C film also suggests that the relative amount of WSi 2 is higher for this interface compared to that deposited at 450 C. This is supported by a larger overlap region for the W a nd Si signals in the SIMS spectra for the 700C film relative to the 450 C one (Figure 5-21 and Figure 5-22). Nucleation of the films presumably occurs on SiO2, with initial deposition of W on the oxide leading to formation of both WSiz and WO3 After nucleation of the initial interface layer is complete, deposition of the WN xCy films begins. One report of sputtered WNx films on Si (100) found an interface layer with a multilayered structure consisting of SiOz/SiOxN/nano-WSiz [Cho02]. An oxynitride layer is not evident in our interface layer, however, as N is not present. Figure 5-19 shows multiplex spectra for films after 10 min sputter. The peak assignments are similar to those for the 450 C film after 10 minutes sputter, and are summarized in Table 5-9. Again, question marks in Table 5-9 indicate that exact location of the corresponding W 4f peaks for the WO x suboxide is unknown, due to lack of additional shoulders in the W4f spectra. Trends indicate that O and N levels generally decrease with increasing temperature. The Nls peak intensity increases from 450 to

PAGE 257

248 500C, however, which is consistent with AES results indicating higher N levels at 500C. The Nls peak intensity drops at 600C, and then disappears into the background noise at 700C, although the B-WNxCy peak is still evident. The disappearance of this peak contrasts with AES results, which indicate that a small amount of N (-3 at.%) remains in the films at 700C. This discrepancy may be due to large shifts in XPS sensitivity for Nin compound materials, as has been reported [Lee93]. XRD results for higher temperature films indicate B-WNxCy formation. While films at 450C are x-ray amorphous, they likely contain FCC B-WNxCy in unit cell form, because the W4f BE values for B-WNxCy have minimal shift when going from an amorphous film (below 500C) to a polycrystalline film (at and above 500 C). This infers that the bonding state doesn't change for B-WNxCy with increasing temperature, but rather that the unit cells grow with increasing temperature until they are large enough (at 500C) to diffract x-rays. In contrast the C content increases with temperature, with the peak for free C growing significantly when going from 600 to 700 C. In addition the Cls peaks shift to slightly higher BE at 700C, which may be due to some charging, as film resistivity jumps at 700C. Interestingly, no W4f or Cls peaks consistent with carbide compounds were evident for any of the films, even at the highest temperature (Table 5 3 and Table 55). This may be due to the presence of some N in the films at all deposition temperatures, which may alter the electron distribution around the W atoms enough to force an upward shift in BE, irrespective of the interstitial site fraction of C. Figure 5-20 shows the area under the W4f712 XPS curve for films deposited from 450 to 700C. The area under the W4f712 curve for B-WNxCy increases from 450 to 500C, and then drops above this. As mentioned above, the film crystallizes at 500 C,

PAGE 258

249 which densifies the film and decreases oxygen's ability to in-diffuse and form WO3 on the surface of P-WNxCy nano-or polycrystals. This is reflected by the increase in area 200000 18000 W4f Nls 180000 16000 700 C v-, .... ,, ,.. .J ,,,,,,,~ 160000 v "''V .-,v.1 , v ,v...,~.~J\r..1 700 C 14000 --140000 --12000 600C a a 120000 >, 10000 1a ti 100000 ti :E :E 8000 < 80000 < 500 CJ\.. '-' '-' >, >, .... .... 6000 ;;; ;;; 60000 i::: ..s ..s 4000 40000 20000 2000 r450 C 1/-z_ \ 450 C 0 ___ 2:::::~-~ 0 a b 42 40 38 36 34 32 30 404 402 400 398 396 Binding Energy (eV) Binding Energy (eV) 25000 25000 Cls 01s 20000 20000 100c 100c 600C a 15000 a 15000 600C l::I l:I 500 C i! 10000 :E 10000 $ 500 $ 0 0 ;;; ;;; C C 5000 5000 .s .s 0 450c 0 450 c C d 290 288 286 284 282 538 536 534 532 530 528 Binding Energy (e V) Binding Energy (eV) Figure 5-19 XPS spectra for films deposited from lb at 450 to 700C after 10 min sputter. a) W4f b) Nls c) Cls and d) Ols BE lines. Dashed lines are deconvoluted peaks.

PAGE 259

250 under the ~-WNxCy curve at 500C. The drop in ~-WNxCy area above 500C follows the trend of decreasing N content with increasing temperature. Decreasing levels of N decrease the ability to form ~-WNxCy at higher temperature. The area under the curve for WO 3 decreases steadily with increasing temperature. This may indicate a shift in film composition with increasing depth into the film, or it may reflect a change in film structure due to sputtering by Ar+ ions. Katrib et al. [1995] reported that Ar+ sputtering of a mixture of WO2+WO 3 caused formation of metallic W Although metallic W is not formed in our films, the drop in WO 3 content may indicate preferential sputtering of O atoms from WO 3 in the film. Sputtering of O atoms from WO 3 would lead to an apparent decrease in WO 3 content and possibly an increase in WOx content. Since the pattern for WO 3 is not evident on the XRD spectra, any WO3 present in the film is likely to be highly disordered (e.g., nanocrystallites), rather than existing as large crystallites. Tabl 5 9 C e ompoun s m 1 ms rom d Fl f lb Af 10 Mi ter nutes s iputter Deposition W4fm W4fs12 Nls Ols Cls Compound Temp. (C) (eV) (eV) (eV) (eV) (eV) 450 32.8 34.6 398.4 -285.8 r3-WNxCv 450 35.7 37.8 -531.4 WO3 450 ? ? -532.4 WOx 450 ----284.6 FreeC 500 32.7 34.8 398.4 -286.0 r3-WNxC v 500 35.6 37.7 531.4 WO3 500 ? ? -533.1 WOx 500 ----284.3 FreeC 600 32.9 34.8 398.5 -286.0 r3-WNxCv 600 35.4 37.1 -531.3 WO3 600 ? ? -533 0 WOx 600 ----284.4 FreeC 700 33.0 35.0 --286.3 r3-WNxCv 700 35.9 37.9 -531.3 WO3 700 ? ? -532.7 WO x 700 ---284.8 FreeC

PAGE 260

140000 Q) 120000 t ::s u 100000 ...s80000 Q) .s 60000 .... Q) "'C d 40000 20000 0 400 450 251 500 550 600 650 Deposition Temperature (C) 700 750 Figure 5-20 Area under the W4f712 peak for films after 10 minutes sputter. The formation of oxide is believed to be due to post-growth reaction of the films with ambient air. Potential oxygen sources in air include water and 02. Vu et al. [1990] reported values for activation energy and pre-exponential factor for the oxidation of Wo.soNo.20 in wet and dry air over a temperature range of 450 to 575C. The rate constant for oxide formation was slightly higher in wet air than dry air, indicating that water is the more favorable oxygen source in this temperature range. Oxide formation still occurred in dry air, however, suggesting that both 0 2 and H 2 O participate in the oxidation process with ambient air. Extrapolating Vu's kinetic data down to room temperature (25C) gives a negligibly small value for oxide thickness. Since our low temperature amorphous films are very porous, significant O was able to diffuse down into the bulk of the films. A thin, amorphous WO3 layer (which is invisible to XRD), consistent with Vu's extrapolated data, may have formed by in-diffused O oxidizing the surface of ~-WNxCy nanocrystals in the film. This is consistent with a previous report, which indicated that

PAGE 261

252 oxidic phases formed by exposure of tungsten-containing films to air were amorphous [Cho00] AES analysis after 1 minute of sputter indicates substantial drop in oxygen content for all deposited films, further supporting this assertion. Overall results indicate that W bonding in the bulk films may be a mixture of WNxCy, WO3 and WOx. The presence of some oxynitride in the film cannot be ruled out by XRD due to exact overlap of ~-WNx and Wo.1s(N,O) spectra [JCP88]. XPS results, however, exclude this possibility, as no additional W and N peaks consistent w i th oxynitride in the bulk films are present. Carbon is typically present in two bonding states, as amorphous (free) C and as C bound to W (and possibly to N) in ~-WNxCy form. While the formation of CNx at the grain boundaries is possible, the single Nls peak and upward shift for W4f BE away from carbidic W suggest that all N is preferentially located at the polycrystal's interstitial sublattice, along with some C, which causes the upward shift in Nls BE. Although the lattice parameter for the films increases up to 675C, it levels off above this temperature. This may indicate a solubility limit for C in the ~-WNxCy polycrystals. Above 675C, additional C contamination goes to the grain boundaries, which is confirmed by a sharp increase in film resistivity and the Cls peak for free Cat 700C. As mentioned previously, ~-WCx films with interst i tial C tend to have low resistivity due to little covalency or ionicity in the W-C bonds. In contrast, when W films contain substantial free C rather than interstit ial C, the free C regions scatter electrons, increasing film resistivity. As the amount of free C in the grain boundaries increases, the resistivity of the films also increases. While the W4f peak representing ~-WNxCy remains for the 700 C films, the corresponding Nls peak disappears. AES results contrast with XPS, though showing that some N remains i n the

PAGE 262

253 films at 700C, supporting the possibility of ~-WNxCy remaining in the film at this temperature. This suggests, however, that ~-WNxC y present in the films at higher temperature is N deficient, with values of x being well below 0.33. 5.5.4 SIMS Depth Profiling SIMS depth profile results are shown in Figure 5-21 and Figure 5-22. Since ~-WNx and ~-WNxCy film standards were unavailable for calibration the SIMS data are not quantitative in nature, as the sputter rates and ionization probabilities are matrix dependent. The change in the SIMS profiles with depth into the film, along with the change in profiles with film deposition temperature, can still give us some perspective, however. Figure 5-21 indicates a SIMS depth profile for a film deposited at 450 C These results reinforce XPS data suggesting W-0 and W-N bonding in the films because WO and WN fragments were sputtered from the films during SIMS analysis. In addition to the presence of W, N, C and 0, SIMS analysis also shows significant amounts of H, which cannot be detected by AES or XPS, and Cl, which can be detected by AES and XPS if present in sufficient quantity (-1 at. %). This suggests that some amount of Cl from the precursor is depositing in the films, although the amount is less than 1 at. % This small amount of Cl may still be troublesome for device applications, as Cl can etch substrates and corrode subsequent metal layers. H in the films originated e i ther from the precursor ligands, the solvent, or the H 2 carrier gas At low temperature, the abi l ity t o form volatile HCl is lower, meaning that more Cl may remain at the film surface without desorbing. At higher temperature, volatile HCl forms more readily and this is reflected by decreased Cl content at the higher deposition temperature. Another difference between the films at 450 and 700 C involves the Si levels. The Si signal increases

PAGE 263

,-._ u 0 en en ..... s= ::, 0 u .._, >-, ..... .... en s= 0 ..... 106 105 104 103 102 101 100 0 200 254 Cl 400 Sputter Time (sec) 600 Figure 5-21 SIMS depth profile for film deposited with lb at 450C. 106 105 ,-._ u Cl 0 en 104 en ..... s= ::, 0 0 u 10 3 .._, >-, ..... .... en s= 0 ..... 102 10' Si 100 0 200 400 600 Sputter Time (sec) Figure 5-22 SIMS depth profile for film deposited with lb at 700C. 0 WO C sharply at the barrier-Si interface in the 450 C film, while it increases much more gradually in the 700 C film. This suggests either substantial Si roughness at the interface

PAGE 264

255 and/or silicide (WSix) formation at the higher temperature, consistent with AES depth profile results. The levels of O and WO are significantly higher in the low temperature film, consistent with increased O in-diffusion into these porous films. In addi t ion, the C level is higher in the 700C film, consistent with increased ligand and solvent decomposition at the higher temperature. 5.6 Film Growth Rate (XSEM) Growth rates were estimated by dividing the total film thickness (from X-SEM) by deposition time. Figure 5-23 gives two sample X-SEM photos. The surface of the deposited films appeared to be fairly smooth, and film thickness varied from -1500 A for the deposition at 450c up to -4000 A for deposition at 100c. Figure 5-23 Cross-section SEM photo depicting film thickness for growth from lb for 150 minutes on a Si (100) substrate. a) Deposition at 450C. b) Deposition at 700C. These film thicknesses correspond to deposition rates ranging from 10-27 .A/min. The growth rate of the films (as determined by X-SEM) varied with temperature, as shown in Figure 5-24.

PAGE 265

256 Deposition Temperature (C) 700 650 600 550 500 450 3.6 3.4 3.2 t-f+-.f--......... 3.0 c., 2.8 i::: -2.6 2.4 2 2 2.0 1.1 1.2 1.3 1.4 1 ooorr (K 1) Figure 5-24 Plot of film growth rate on Si (100) vs. inverse temperature. Error bars indicate uncertainty due to deposition temperature variation ( +/-10C) and thickness measurement from XSEM photos. From this figure, two growth regimes are apparent. The region with shallow slope shows a weak dependence on temperature that is indicative of the mass transfer controlled regime Mass transport to the surface is the rate-determining step in this regime, where high substrate temperature enables immediate reaction of species arriving at the surface. The region with a steep slope indicates that growth rate increases exponentially with temperature, which is indicative of the kinetically controlled regime. In this regime, the rate-determining step for film growth is reaction on the substrate surface. Ample reactants are available at the substrate surface, and growth rate is determined by the reaction rate. Growth rate in the kinetically controlled regime can be estimated by the Arrhenius expression for growth rate (Equation 5-2): G = Ae-E,tRT (5-2)

PAGE 266

257 where G is growth rate (Ns) and Ea is the apparent activation energy (eV). From this expression, we can calculate Ea for the precursor molecule by plotting ln G vs. 1/f according to Equation 5-3: lnG=-(i )(~)+lnA (5-3) where A is a constant, Ea/R is the slope of the line in the kinetically controlled regime, and Ea is the apparent activation energy for film growth. The calculated Ea for film growth from the lb precursor was 0.84 eV in the kinetically controlled regime, compared to 0.9 eV reported for bis(tert-butylimido)bis(tert-butylamido)tungsten [Tsa96]. This value falls within the typical activation energy range for CVD growth in the kinetic regime, which ranges from 0.5 to 1 eV [Raa93]. In the kinetically controlled regime, a change in temperature of 1 % yielded a 10% change in growth rate. The kinetically controlled regime appears to have an upper temperature limit near 600 C. Above 600 C, the region's shallower slope indicates mass transfer controlled growth. Depositions at 400 and 750 C were done to establish the temperature boundaries for film growth with the precursor. Although the results of these growths were not included in Figure 5-24 they do provide some useful information. Films deposited at 400C were not visible by X-SEM, but AES depth profiling indicated a nominal film thickness of -100 A This thickness value was used to estimate Ea for the kinetic regime. The growth rate at 750 C dropped to 10 A /min, with black particulates visible on the substrates after deposition. Particulate formation and reduced growth rate result from gas phase decomposi t ion of the precursor. At this high temperature, upstream heating caused precursor decomposition and reaction in the gas phase. Particulates on the substrates confirmed a gas phase reaction, while the drop in film deposition rate is consistent with reactant depletion.

PAGE 267

258 28 27 G =0.0001 *T1.77 ,-... = 26 ....... v., lj 25 v., 00 = < 24 '-" c:, 23 22 860 880 900 920 940 960 980 Deposition Temperature ( K) Figure 5-25 Temperature dependence of film growth rate on Si ( 100 ) in the diffusion controlled regime. 30 I I I I I ,-... I = I I a I I I ....... 25 I a I g v., 00 = < 20 I I '-" I Q) I I ... I C':S I Different ial I I I ..c:: I Operation I ... 15 I I .. 0 Starved I -C, I Operation I I I I 10 .. 5.0e-10 l.Oe9 l.5e-9 2.0e-9 2 .5e-9 Precursor Concentration in Gas Phase (mol/cm3 ) Fig ure 5-26 Plot of film growth rate on Si (100) vs precursor concentration in the gas phase. Error bars indicate uncertainty due to deposition temperature variation(+/10C) and thickness measurement from XSEM photos. Figure 5 25 shows a plot of d eposition temperature vs. film growth rate in the mass transfer controlled regime. Experiments indicated growth rate in the mass transfer

PAGE 268

259 controlled regime had a temperature dependence of Ti.78 This is similar to a reported value of T1.7 for diffusion controlled film growth in an impinging flow reactor [Hua03]. Figure 5-26 indicates dependence of film growth rate on the concentration of i-Pr precursor in the gas phase. Below -l.2xl0-9 mol/cm3 the growth rate appeared to have a linear dependence on concentration, indicative of operation in the starved regime (Chapter 1). Above this concentration, the growth rate was steady, which indicates operation in the differential regime. 5. 7 Film Incubation Time and Morphology The orientation of the underlying Si substrate also affected film deposition rate. Growth from the i-Pr precursor at 650C on Si (100) yielded an incubation time of -50 minutes, while an incubation time of -30 minutes was evident for growth on Si (111) with a 4 miscut. Initial nucleation for the deposited films is therefore slower on Si (100) than on miscut Si (111), which is caused by a difference in surface energy between the Si (100) and miscut Si (111) surfaces. Several factors give rise to this difference in surface energy, including the number of dangling bonds, the density of surface sites, and the number of surface features (such as steps kinks and ledges) [Vos78]. The number of dangling bonds on the surface, for example, is proportional to that surface's energy where more dangling bonds mean higher surface energy. Si crystals have a diamond cubic structure, with each Si atom tetrahedrally bonded to four other Si atoms. When cleaved along the (100) plane, each Si atom on the substrate surface retains two of the four bonds to other Si atoms, while two of the bonds are left dangling (unbonded). When cleaved along the (111) plane, each Si atom retains three bonds to neighboring Si atoms, while one bond is left dangling. Since Si (100) has 2 dangling bonds per surface atom it

PAGE 269

260 has a higher surface energy than the Si (111) face. But, when Si(ll 1) is miscut, the number of dangling bonds, along with steps, kinks and ledges, increases, due to the formation of vicinal surfaces. Hence, the 4 miscut Si (111) has higher surface energy than Si (100). The substrate's surface energy plays a large role in nucleation and film growth, and takes part in two of the three interface energies, which together determine the feasibility of film deposition. The three energies are Ysv (substrate-vapor interface energy), Y s f (substrate-film interface energy) and Yfv (film-vapor interface energy). The necessary criterion between these three energies for film deposition to occur is shown in Equation 5-4: Ysv > Ysf + Yfv (5-4) Film deposition will not occur unless this criterion is met; hence deposition is more likely for a system with a larger value of Ysv The relationship between the two terms on the right side of Equation 5-4 determines how the film will nucleate on the substrate surface (i.e., island formation vs. film wetting). Since Si (100) has lower surface energy (Ysv) than 4 miscut Si (111), nucleation should occur more readily on miscut Si (111). This is consistent with growth data, which indicate lower incubation time (-30 minutes) on miscut Si (111) compared to the incubation time (-50 minutes) for deposition on Si (100) as shown in Figure 5 -2 7. Once a layer of film has deposited on the Si substrate, the film continues to grow on itself at essentially the same rate regardless of substrate orientation, evidenced by similar slopes for the film thickness vs. deposition time lines in Figure 527.

PAGE 270

261 The variation in film roughness with deposition time at 650C on Si (100) is also shown in Figure 5-27. RMS roughness was highest (470 A) for the 30 min deposition, indicating that some deposition and/or etching had occurred by this time. After 45 min of deposition, RMS roughness dropped substantially to a local minimum (100 A), suggesting that nucleation of the interface layer was near completion, which supports an incubation time of -50 min. At the 60 min point, film roughness increased to 190 A, suggesting a Stranski-Kranstanov (S-K) growth mechanism [Ohr92], which may be caused by strain energy associated with defects and pores in the interface layer. As growth continued to 120 minutes, the RMS roughness dropped slightly to 170 A, suggesting that S-K growth continued through this time. Between 120 and 150 minutes, roughness dropped considerably to 40 A, which may indicate a shift toward layer-by-layer growth when film thickness increases above 2000 A. AFM micrographs supporting the roughness results are shown in Figure 5-28. 6000 500 ... Si(lOO) 5000 \ A Si(l 11) ... \ 400 ,-.. \ ,-.. o< \ o< 4000 '-' '-' \
PAGE 271

262 .. ... Figure 5-28 AFM micrographs of film surface for growth at 650C on Si (100). a) 30 min growth. b) 45 min growth. c) 60 min growth. d) 150 min growth AFM roughness measurements are supported by SEM micrographs of the films at different times during the growth cycle. Figure 5-29 shows films deposited at times varying from 30 to 150 minutes. For the 30 minute sample, a very disordered nucleation pattern is evident on the underlying Si (100) substrate which may be the initial stages of interface layer deposition. As deposition time increases to 45 minutes (near the incubation time) the film is still very disordered, but its coverage of the Si substrate has increased substantially. After 60 minutes, islands appear to have formed on the underlying nucleation layer of the film in a manner consistent with S-K growth. As growth time increases to 90 and 120 minutes, the average size of the islands increases, indicating that the island/film interface energy is decreasing which leads to islands with less curvature. Finally, after 150 minutes a shift to layer growth is apparent as the islands have disappeared in favor of a very smooth surface

PAGE 272

263 Figure 5-29 Morphology of films deposited at 650C. a) 30 min growth. b) 45 min growth. c) 60 min growth. d) 90 min growth. e) 120 min growth. f) 150 min growth. Cross-section SEM micrographs in Figure 5-30 below further underscore the change in film growth with deposition time Film thickness is minimal after 60 minutes, and increases substantially for 90 minutes of deposition. The thickness continues to increase with deposition time, and the film thickness and roughness are highest and lowest, respectively, after deposition for 150 minutes.

PAGE 273

264 Figure 5-30 XSEM pictures of films deposited with lb at 650 Con Si (100) substrates. a) 60 min growth. b) 90 min growth. c) 120 min growth. d) 150 min growth. 5.8 Film Electrical Properties Film resistivity was calculated using Equation 5-5: (5-5) where p is resistivity (Q-cm), Rs is sheet resistance (Q/O) from 4-point probe, and t is film thickness (cm) from X-SEM. Sheet resistance of the films was checked with 4-point probe. The variation of film resistivity with deposition temperature is shown in Figure 5-31. Deposition at 450C produced films with the lowest resistivity value, 750 .Q-cm, despite C and O contamination levels of 13 and 22 at.%, respectively. This is slightly higher than 620 .Q-cm reported for film growth from the single-source precursor bis(tert-butylimido )bis(tert-butylamido )tungsten at 650C [Tsa96].

PAGE 274

265 Resistivity climbs steadily from 750 .Q-cm to 4800 .Q-cm as deposition temperature increases from 16000 14000 ,....._ s 12000 (.) I a ::::!.. 10000 '-' 0 8000 .::: ... ell -ell 6000 (1) s 4000 2000 0 450 500 550 600 650 700 Deposition Temperature ( C) Figure 5-31 Variation of film resistivity with deposition temperature. 450-550C The jump in resistivity is most likely due to the combined effects of increasing C levels and polycrystal formation. Contamination by free C increases film resistivity by scattering electrons as they travel through the film, and C levels increased from 13 to 20 at.% over this temperature range In addition, transitioning from an amorphous to polycrystalline film structure causes formation of grain boundaries which scatter electrons and increase film resistivity. Once polycrystal grains have formed, the film resistivity is typically lowered by grain growth, which decreases the number of grain boundaries. Between 550 and 675 C, the resistivity increases gradually from 4800 t o 5500 .Q-cm, despite grain growth in this temperature range, evident in Figure 5-4. This can be attributed to increasing C contamination through this temperature range which is shown in Figure 5 -10. In this range, interplay between grain growth which decreases

PAGE 275

266 resistivity, and increasing free C content, which increases resistivity, led to the gradual increase in film resistivity. Between 675 and 700C, film resistivity jumps to 15000 .Q-cm, and is due to massive C concentration, which overtakes the W concentration near 700C. In Figure 5-6, the grain size levels off between 675 and 700C, which is the same temperature range where the film resistivity increases greatly. Figure 5-10 indicates that C content continues to rise in this temperature range. This may indicate that the additional free carbon is depositing in the grain boundaries, making it difficult for current to pass between the polycrystal grains. Decreasing the deposition temperature decreases the amount of C present in the film and therefore should decrease the film resistivity. This trend is additional incentive to move to lower deposition temperatures. Film resistivities calculated with Equation 5-5 depend on both the sheet resistance and thickness of the analyzed films. As deposition temperature increases, both the film resistivity and the film thickness increase. To decouple the impact of film thickness from carbon contamination on the film's electrical properties, the sheet resistance was plotted as a function of deposition temperature in Figure 5-32. The sheet resistance increases with deposition temperature from 47 .Q/O at 450C to 184 .Q/O at 500C, decreases to 145 .Q/O at 675C, and increases sharply to 371 .Q/O at 700C. The increase in sheet resistance up to 550C is attributed to polycrystal formation and increased C contamination. Between 550 and 675 C, the decrease in sheet resistance is due to grain growth, which overcomes any increase caused by additional C contamination. At 700C, overwhelming C contamination dominates the sheet resistance, driving it up sharply.

PAGE 276

267 400 --350 (1) 300 &
PAGE 277

268 amorphous WNx films transform into polycrystalline ~-W2N between 450 and 500C [Gal97]. As the deposition temperature increases to 600C, the (111) peak sharpens and shifts to a lower 20 value, consistent with ~-WNxCy formation, in which N and C intermix on the FCC W interstitial sublattice [Bch03c]. The (111) peak positions for the 1400 Si (400) ,-._ 1200 en ..... 13-WN, C Y 13-WN, C y (200) 700 o c (220) ;:J 1000 b 800 ..... '--' 600 >-. .-:::: en s:= 2 400 s:= -200 0 30 40 50 60 70 80 20 Degrees 1600 b ,-._ 1400 en ..... .... s:= ;:J 1200 b 1000 .... 800 '--' 0 600 .... en s:= 2 400 s:= -200 0 30 40 50 60 70 80 20 Degrees Figure 5-33 XRD spectra for films grown on Si (100) in a H2 atmosphere. a) Without NH 3 b) With NH3.

PAGE 278

269 binary P-WN0 5 and P-WC0 6 phases are 37.74 and 36.98 20 degrees, respectively, hence a peak location between them suggests a ternary P-WNxCy phase [JCP88]. Another broad peak at 44.42 20 degrees also appears at this temperature, suggesting P-WNxCy (200) formation. Films deposited at 500 and 600C displayed two additional peaks at 33.03 and 61.67 20 degrees, representing Si (200) Ka and Si (400) KP radiation, respectively Increasing the deposition temperature to 700C results in further sharpening of the ~-WNxC y (111) and (200) peaks; a shift to lower 28 value for these peaks is consistent with decreasing x and increasing y in the ~-WNxCy polycrystals. The emergence of two broad peaks at 63.00 and 75.82 28 degrees at this temperature indicate P-WNxCy (220) and (311) formation. No XRD peaks consistent with graphite, metallic W or W03 polycrystals were present in the spectra. Deposition with NH3 at 450C resulted in deposition of an X-ray amorphous film, as had been observed without NH 3 Material grown with NH 3 at 500C was also amorphous, in contrast to polycrystalline films grown without NH3 at 500C This phenomenon may be due to surface site blocking and insufficient surface diffusion of NH 3 at lower deposition temperature. Addition of NH 3 resulted in similar spectra for films grown at 600 and 700 C, but with sharpened peaks (Figure 5-33b) indicating enhanced grain growth relative to deposition without NH3. In addition the peaks at 63.00 and 75.82 28 degrees appear for the 600C deposition with NH3 while they do not appear until 700 C for films deposited without NH3 Carbon is reported to decrease grain size when added to refractory metal films [WanOlb]. Thus, the increase in polycrystal grain size is likely due to the decrease in C content (Figure 5-34c) accompanying NH3 addition at these higher temperatures

PAGE 279

5.9.2 Film Composition 5.9.2.1 AES Results 270 Figure 5-34 depicts the impact of NH3 addition on the concentration of W, N, C and O in the films. Addition of NH3 dramatically i ncreased the N levels over the entire deposition temperature range, as depicted in Figure 5-34b. The concomitant decrease in W is attributed to the increased concentration of N in the films. The addition of NH3 increased the N content from 8 to 24 at. % for deposition at 450C, and levels increased from 11 to 29 at. % at 500C. As the deposition temperature increases to 600 and 700 C, N levels drop off. Metal nitride barriers typically show a decrease i n N content w it h increasing deposition temperature, because higher temperatures impart more energy t o the film lattice, inducing N to desorb as N2 gas. Nitrogen loss from WNx films has been reported at temperatures above 700C [Kat85] consistent with AES results in t his range. Similar experiments replacing NH3 with N 2 an inert species at these temperatures, yielded poor film coverage and lower nitrogen levels than films deposi t ed without a co-reactant. Figure 5-34c dep i cts the change in C concentration with deposition tempe r atu r e for films deposited with and without NH3 Addition of NH3 appears to ha v e l i ttle i mpact on C content at the two lower deposition temperatures Th i s may reflect a change in the reaction kinetics of C removal processes between low and high temperature. Enhanced NH3 decomposition at increased temperatures may improve produc t ion of gaseous carbon species (such as CH.i), which scavenge C from the film.

PAGE 280

271 80 80 i a i b (.) Tungsten (.) N i trogen s s 8 60 0 60 8 tt 40 u: 40 .!3 .!3 ---Without := := ---------With NH3 a a .... -.. 20 C 20 ---.. C .., ....................................... 8 (.) -Without NH3 C C 0 0 u .... u --With NH3 0 0 400 450 500 550 600 650 700 750 400 450 500 550 600 650 700 750 Deposition Temperature ( C) Deposition Temperature (C) 80 80 i C i d (.) Carbon (.) Oxygen s 60 s 60 8 8 .. u: 40 ,,. u: 40 .!3 / .9 / = / = 0 / 0 := / .:i -+-Without ~ i 20 / fl -~ = 20 -+With NH3 .., .. .., (.) --..-Without~ (.) = = 0 0 u --&--With NH3 u 0 0 400 450 500 550 600 650 700 750 400 450 500 550 600 650 700 750 Deposition Temperature ( C) Deposition Temperature ( C ) Figure 5-34 Variation of content in the films with deposition temperature. a) W b) N c) C and d) 0 Data are estimated from AES spectra taken after 2 min sputter time. Figure 5-34d depicts the change in O concentration with deposition temperature for films deposited with and without NH3. Oxygen contamination results from post-growth exposure of the films to air, and high O levels at low deposition temperatures are likely due to film porosity [Bch03a]. Oxygen content drops substantially for deposition without NH 3 between 450 and 500 C. Film crystallization at higher deposition temperature causes an increase in film density, which inhibits diffusion of O into the films upon exposure to air. The -12% drop in O concentration for films deposited with NH 3 at 450C relative to those without might suggest that addition o f NH 3 leads to a denser amorphous film structure, which allows less O in-diffusion. However,

PAGE 281

272 this possibility was ruled out by density measurements (per XRR) of 8.3 and 8.6 g/cm3, respectively, for films deposited at 450C with and without NH3 While these values are similar to those obtained for WNx deposited with WF6 and NH3 at 450C [Mar93], they indicate that film density actually decreases with NH3 addition. Oxygen content was lower in these films, however, despite their lower density, suggesting that microstructure has greater impact than density on the film's resistance to post-growth oxygen incorporation. This is consistent with a previous report, which indicates that diffusion barrier performance depends more strongly on film microstructure than film density [Kim99c]. As the deposition temperature increases above 500C, the O levels in both films decrease, becoming roughly equal near 2 at. % for deposition at 700C. Similar 0 concentrations at 700C reflect the impact of polycrystal size and C contamination on 0 diffusion. Films deposited without NH3 have higher C content at 700C, and this C is expected to stuff grain boundaries, preventing O diffusion into the film. Films deposited with NH3 have lower C contamination, but larger polycrystals, which resist O migration. Figure 5-35 shows an AES depth profile for a film deposited at 450C with NH3 The N level is elevated and the O level is reduced in this film compared to that deposited without NH3 (Figure 5-12). The O level still rises at the barrier-Si interface, however, indicating that NH3 addition does not eliminate the formation of a porous/defective layer at the interface. This layer may be thinner when NH3 is added, however, which is evidenced by the narrower O bump at the interface. The C level is somewhat reduced throughout the film when NH3 is added. As in the film deposited without NH3 the W and O signals overlap with the Si signal for a similar amount of time.

PAGE 282

273 35000 30000 --Cl) ..... ..... i:: ;:J 25000 :>-. 1-, ro !J 20000 ..... '-' 15000 :>-. ..... ..... Cl) 10000 i:: ..... .E 5000 0 0 5 10 15 20 25 30 Sputter Time (min) Figure 5-35 Three-point depth profile for film deposited with NH3 at 450C. Nominal AES sputter rate was 100 Nmin 5.9.2.2 XPS Results The W4f, Nls, Cls and Ols photoelectron lines were also used to examine the bonding states in films deposited with NH3 Again, literature BE values for potential compounds in the films are listed in Table 5-3 through Table 5-6. XPS results for films deposited at 4 different temperatures without NH3 after 10 minutes sputter were discussed above. Figure 5-19, along with Table 5-9, summarize those results. Figure 5-36 shows the effect of NH3 addition on the W4f and Nls binding energies. The principal W4f712 and W4f512 peaks for 450C are at 33.6 and 35.6 eV, respectively, corresponding to the upper limit for WNx (Table 5-3). Again, this upward shift may be due in part to an attractive N-C interaction on the interstitial sublattice of W in the 13-WNxCy bonding state. Although these BE values are stable over the entire temperature range, they are shifted to -1 e V higher relative to films deposited without

PAGE 283

180000 160000 1 40000 ,,__ 1 20000 Cl} ..... 2 :::, b 1 00000 '-' 80000 550 C c 60000 ;;i s::: s::: 40000 .... 20000 0 42 40 38 3 6 34 B i ndin g Energy ( e V ) 12000 I 0000 0 -l-..::!4:::!:50~0c~--290 288 286 284 Binding Energy (eV) 274 W4f Nls 700 C 30000 ,,__ 600 C Cl} ..... :::, 20000 b 5 50 C '-' >-, ..... ;;i s::: 10000 500 C ..... s::: .... 450 C 0 a b 32 30 406 404 4 0 2 400 398 396 394 B i nd i n g Energy ( e V ) C 282 ,...._ !!l a ;:i 7 00 C 30000 -1-....:....:....:..........::.._..........:c >. 20000 550 C g I 450 C 0 -1-~:..:.... ...;;;..,= 538 536 534 532 B inding Energy (eV) Ols d 530 528 F i g ur e 5 3 6 V ariatio n of b i ndin g e n ergies w ith d epositi on tem p erature for films d e po s it e d with lb+ NH 3 a) W4f b ) Nls c) Cls a nd d ) Ols. Dat a are fr om XP S after 10 min s putter Dash e d lines ar e decon v oluted p eaks

PAGE 284

275 NH3. This shift is due in part to increased nitrogen content in the films, which suggests greater occupancy of N sites relative to films deposited without NH3 (Figure 5-17 a), and therefore more complete W-N bonding. Higher numbers of W-N bonds raise the partial positive charge at W and thereby increase the W4f BE. In addition, some sample charging may occur, which would also cause an increase in BE. Like the films deposited without NH3, a W4f shoulder was observed for films deposited at lower temperatures, indicating the presence of WO3 The Nls BE values are steady over the entire temperature range, and are centered around 398.8 eV, above the reported range for WNx values (Table 5-4) Again, the lack of a second Nls peak above 399 eV suggests that all N is bound to W, with none present in its "free" form. This increase in Nls BE is unusual, considering the increased W-N bonding that should occur for higher N content, and may be attributed to sample charging. Again, the upward shift in Nls BE may also be due in part to an attractive N-C interaction on the interstitial sublattice of W in the B-WNxCy bonding state. In addition to BE, we note that the Nls peaks endure through the entire temperature range for films deposited with NH3. Although N content does drop for these films above 500C, AES results in Figure 5-34b indicate a substantial increase in N for these films relative to those grown without NH3 This is consistent with competition between increased rates of nitrogen release from the film into the gas phase and higher reactivity of NH3 with the growing film at higher temperature. Again, while films deposited with NH3 at 450C are x-ray amorphous, they likely contain B-WNxCy in unit cell form, because the W4f BE values for B-WNxCy have minimal shift when going from an amorphous film (below 600 C) to a polycrystalline film (at and above 600 C).

PAGE 285

276 Comparing Figure 5-36c to Figure 5-19c shows that the Cls peaks have lower intensity for films deposited with NH 3 consistent with the lower C content shown in Figure 5-34c. For the film deposited at 450C, the Cls peak at higher BE is very small, suggesting that the interstitial sites are almost completely filled by N, leaving little room for C to intermix. In addition, the convoluted peak at 700C is narrower than that from films without NH 3 with a small peak at lower BE and a small shoulder at higher BE. This suggests that NH 3 addition not only lowers the C content at higher temperature, but the additional N displaces C into the grain boundary region, leaving a smaller amount present in the B-WNxCy bonding state (higher BE). The additional, displaced C at the boundaries of the B-WNxCy unit cells (or nanocrystals) in the amorphous film deposited at 450C could help to plug these regions, inhibiting O in-diffusion. Hence, although films deposited at 450C with NH 3 have lower density, their microstructure may enhance resistance to O intrusion in the films. Comparing Ols lines in Figure 5-36d and Figure 5-19d indicates that O levels are highest in the low temperature films, and decrease with increasing deposition temperature due to film crystallization (which increases density) and additional C incorporation (which stuffs grain boundaries). The O peaks for deposition without NH 3 were centered near 531.3 eV, consistent with WO 3 formation (Table 5-6). The peaks were broad, with small shoulders near 533 eV caused by WOx [Zha99], which forms during Ar+ ion sputtering of WO 3 in the XPS chamber. Similar behavior, with lower intensity, was evident for films deposited with NH 3 although the peaks are shifted to slightly higher BE (-531.6 and -533.3 eV, respectively), which may be due to charging effects. For

PAGE 286

277 deposition at 450C, the peak intensity is noticeably lower for deposition with NH 3 than without, and is attributed to free C, which is displaced from interstitial sites by additional N, stuffing pores in the film. Tabl 5 10 C e -ompoun s m 1 ms own wit 3 ter nu es d Fl Gr h NH Af 10 Mi t S putter Deposition W4f712 W4fs12 Nls Ols Cls Compound Temp. (C) (eV) (eV) (eV) (eV) (eV) 450 33.6 35.4 398.8 -286.0 B-WNxCv 450 36.6 38.9 -531.9 WO3 450 ? ? -533.6 WOx 450 --284.8 FreeC 500 33.6 35.5 398.8 -286.6 '3-WNxCv 500 36.5 38.4 531.6 WO3 500 ? ? -532.9 WOx 500 ---284.8 FreeC 550 33.6 35.7 398.7 -286.8 '3-WNxCv 550 36.2 38.7 531.5 WO3 550 ? ? -533.5 WOx 550 ----284.6 FreeC 600 33.5 35 4 398.8 -286.9 '3-WNxCv 600 36.4 38 6 -531.6 WO3 600 ? ? -533.3 WOx 600 ----284.9 FreeC 700 33.6 35.5 398.9 -286.4 '3-WNxCv 700 36 2 38 0 -531.6 WO3 700 ? ? -532.9 WOx 700 ----284.8 FreeC The peak assignments are summarized in Table 5-10. Question marks in Table 510 indicate that exact location of the corresponding W4f peaks for the WOx suboxide is unknown, due to lack of additional shoulders in the W4f spectra. 5.9.3 Growth Rate Growth rates were estimated by dividing the total film thickness (as measured X-SEM) by deposition time. Figure 5-37 depicts two sample X-SEM photos. The surface of the films was smoother for those deposited with NH 3 than those without, with AFM indicating a root mean square (RMS) roughness of 3.8 and 7.1 nm for films

PAGE 287

278 deposited at 450C with and without NH3 respectively. Film thickness for depositions with NH3 varied from 2600 A for the deposition at 450C up to 3400 A for deposition at Figure 5-37 Cross-section XSEM photo of films deposited with NH3 a) 450C. b) 100c. 3.4 3.2 3.0 Cl 2.8 i:: -2.6 2.4 2.2 Deposition Temperature (C) 700 650 600 550 500 450 I -.. -.. .. _ I --AWith NH3 Without NH3 2.0 .....___....---,---,-------r-----.-------,r-------r----' 1.05 1.10 1.15 1.20 1.25 1.30 1.35 1000/f (K-l) Figure 5-38 Dependence of film growth rate (G) on temperature for films deposited with and without NH3 Error bars indicate uncertainty due to deposition temperature variation(+/10C) and thickness measurement from X-SEM photos.

PAGE 288

279 700C. These film thicknesses correspond to deposition rates ranging from 17-23 Almin, compared to rates of 10-27 Almin for depositions without NH3 [Bch03a]. The growth rate of the films (as determined by X-SEM) varied with temperature, as shown in Figure 5-38. Interestingly, the deposition rate with NH3 was enhanced at lower temperature and reduced at higher temperature relative to deposition without NH3. In addition, deposition with NH3 was mass-transfer controlled across the entire temperature range, while deposition without NH3 had a kinetic to diffusion control transition point near 600C. This indicates a shift in deposition mechanism with the addition of NH3 as a co-reactant. 5.9.4 Film Resistivity Film resistivity was again calculated using Equation 5-5 above. The variation of film resistivity with deposition temperature is shown in Figure 5-39, which indicates that resistivity with NH3 goes through a maximum value of 55,000 .Q-cm at 500C. The temperature dependence of resistivity for films deposited with NH3 followed a trend consistent with N content, shown in Figure 5-34b. Additional N in the amorphous films would increase film resistivity in two ways. First, the added N would increase resistivity in the ~-WNxCy nanocrystals, since the resistivity is higher for ~-WNx relative to ~-WCx. Second, additional N would displace C from the nanocrystals into the "free" state. Increased free C increases electron scattering, thereby raising film resistivity. As the deposition temperature increases from 500 to 600C, the resistivity drops sharply, caused by a decrease in N content along with film crystallization in this temperature range. A resistivity decrease is expected with decreasing N content in WNx films

PAGE 289

280 [Lee93]. At the highest temperature, resistivity for films deposited with NH3 dropped below that for films grown without NH3 This is due to lower C content, larger polycrystal grain size (not shown), and decreased thickness for films grown with NH3. 60000 .----1 ----With NH3 50000 k"/ \ ,___ Without NH3 e \ u 40000 \ I Cl \ :::t \ ..._, o 30000 \ : ~ \ \ ~ 20000 \ "' Q) \ 10000 \ \ ---. 0 400 450 500 550 600 650 700 Deposition Temperature (C) Figure 5-39 Film resistivity for samples grown with and without NH3. 5.9.5 Sheet Resistance 750 To decouple the impact of film thickness from the electrical properties, the sheet 2500 ~---------------------With NH3 o' 2000 l;; ;::l C" ... --~ Without NH3 a 1500 ..._, Q) g 1000 !S "' ;;; 500 ] VJ 0 400 450 \ \ \ \ \ 500 \ \ \ \ 550 600 650 Depos ition Temperature (0C) 700 750 Figure 5-40 Sheet resistance for samples grown with and without NH3

PAGE 290

281 resistance as a function of temperature is plotted in Figure 5-40. The trend is similar to that for film resistivity, for similar reasons. A difference in Rs at 700C is evident, confirming that the electrical properties of the two types of films differ at this temperature. Again, the lower Rs value for the film deposited with NH3 at 700C is due to lower C content and larger polycrystal grain size. 5.10 Conclusions for Use of NH3 and N2 with i-Pr Deposition of WNx (or WNxCy) thin films from C4(CH3 CN)W(N i Pr) m the presence and absence of an NH3 co-reactant was demonstrated. AES results indicated that films grown with NH3 had significantly higher levels of N (especially for low deposition temperatures), along with decreased levels of C and O as compared to films grown without NH3. XRD and AES results suggest that increased crystallinity for films deposited at 600 and 700C with NH3 was due to decreased C content. XPS W4f binding energies suggested the presence of W in both the ~-WNxCy and W03 environments, with 0 incorporation occurring post-growth. Moreover, due to the presence of O in the films, the possibility that some W(N,O)x exists in the films cannot be ruled out, as the Nls binding energies and XRD pattern for the oxynitride are very close to those for WNx, While XPS indicates the presence of W03 XRD results do not show any W03 peaks. Hence, any W03 present in the film is likely in the form of nanocrystallites, which are not detectable by XRD. The overall film structure is amorphous at lower temperatures, and the polycrystalline films deposited at higher temperatures likely contain small W03 nanocrystallites embedded between larger WNxCy polycrystals. XPS also indicated a lack of grain boundary N in films regardless of NH3 addition. Although some evidence

PAGE 291

282 of charging was apparent from the XPS results, quantification of charging through indexing of the adventitious C peak is difficult because the films contain C throughout their bulk. Adding NH3 as a co-reactant changed the mechanism for film deposition, as is reflected by the mass transfer controlled nature of film growth over the entire temperature range. Film resistivity at lower temperature was significantly higher for films deposited with NH3, and this was attributed to additional N in B-WNxCy and additional C at the grain boundary. Addition of N2 to lb during deposition prevented film deposition from occurring. This was likely due to N 2 site blocking on the substrate surface, as N 2 is inert over our experimental temperature range, and would not take part in the reaction. 5.11 Effect of Solvent Change on Carbon Content Potential sources of the C contamination in the films grown from Cl4(CH3CN)W(NiPr) (la) (Figure 5-1) are the alkyl moieties in the precursor itself and the benzonitrile (PhCN) solvent. Use of an alternative solvent in the volatilization process, then, could provide a means to differentiate these possibilities. If the C incorporation is dominated by decomposition of la, the amount of C will be insensitive to the nature of the solvent. If the C derives from the solvent instead, the level of contamination will depend on the solvent. To address this question, films were deposited from solutions of la in 1,2-dichlorobenzene (1,2-DCB) for comparison to films derived from lb in PhCN. 1,2 DCB was chosen because it has a similar vapor pressure to PhCN and is unable to engage in substitution chemistry with the acetonitrile ligand of precursor la. During the depositions, the composition of the gas near the substrate surface was

PAGE 292

283 measured using an lnficon Transpector 2 residual gas analyzer (RGA). N2 was used as the carrier gas for RGA analyses, and total pressure in the RGA was -6.5x10-6 Torr. Table 5-11. Potential solvent-derived species in the reactor and their decomposition d t pro UC S. Major Reported Molecule Decomposition Decomposition Decomposition Reference (source) Type Products Temperature (oC) PhCN Homogeneous HCN, C68763C [Bro79, C68-73C [Car74] (112), (111) (98.5%) HCN Homogeneous H,CN 2427-3327C [Sze84] (dee.) Heterogeneous on N2, H2, C(ads), 27-1127C [Cai70, W(lO0) N(ads), WC Pea86a] a dee= derived from primary or secondary decomposition of one or both solvents. In the deposition temperature range of this study (450 to 700C), the solvent molecules likely undergo some degree of homogeneous decomposition before reaching the substrate surface, so homogeneous and heterogeneous C deposition processes must be

PAGE 293

284 considered. Species that are potentially present in the reactor (including previously reported primary and secondary solvent decomposition products) are listed in Table 5-11. Data on heterogeneous decomposition for PhCN and 1,2-DCB are not available. Table 5-11 indicates that several primary and secondary gas-phase decomposition products from each solvent may be present in the reactor, along with the solvents themselves. Each of these species is potentially capable of depositing C. Furthermore, all of the decomposition products, with the exception of HCN and HCl, are common to both PhCN and 1,2-DCB. Low concentrations of HCN and HCl that derive from la, however, are presumably present in the reactor regardless of the solvent. This is due to reaction of the CH3CN ligand or the four chlorines, respectively, with the H2 carrier gas. A study of C deposition dependence on solvent should therefore focus on the concentration of the various carbon-containing species near the substrate surface, as measured by ROA, to identify which species are the most likely sources of C deposition. 5.11.1 Film Structure 5.11.1.1 XRDResults XRD spectra for films deposited with PhCN (Figure 5-4) and 1,2-DCB (Figure 541) indicate ~-WNxCy polycrystal formation down to 500 and 525C, respectively. Below 500C, films grown with PhCN were amorphous, while those grown with 1,2-DCB at and below 500C were consistent with tungsten oxide formation. Lower levels of N and C make these films less resistant to O in-diffusion, enabling oxide formation to occur. Above 500C, films from 1,2-DCB solutions were polycrystalline and contained substantial C, which enabled them to resist O in-diffusion and prevented oxide formation.

PAGE 294

285 3000 ~-WNxC y ~-WNxCy ,j,(111) ,I (200) T=700 C 2500 T=675 C ---Cl.) T=650 C ..... ..... s::: 2000 ;>,, a b T=600 C ..... ..0 < 1500 T=57 C "-' ;>,, ..... T=550 C ..... 1000 Cl.) T=525 C s::: ..... wax T=500 C s::: 500 T=475 C T=450 C 0 30 40 50 60 70 80 20 Degrees Figure 5-41 Change in XRD pattern with deposition temperature for films grown with la+ 1,2-DCB on Si (100) in a H2 atmosphere. 5.11.1.2 Lattice Parameter and Grain Size Lattice parameter and grain size were determined using the 20 position and full width half maximum (FWHM) of the B-WNxCy (111) diffraction peak as described previously [Bch03c]. The values of the lattice parameter in films deposited at higher temperature (i.e 650C) were similar regardless of the solvent used as shown i n Figure 5-42a. A noticeable deviation occurred at 550 and 575C with films from 1 ,2-DCB solutions having a higher lattice parameter at these temperatures. Although films from 1,2-DCB solutions had lower N content than those from PhCN so l utions at these temperatures, they had substantially higher C content. Decreasing x and increas i ng y in the ~-WNxCy polycrystals causes lattice expansion since B-WNo s and B-WC0 6 have lattice parameters of 4 1 2 6 and 4.236 A respectively Some of the addi t ional

PAGE 295

286 carbon was incorporated into the polycrystals of films grown from solutions of 1,2-DCB, causing lattice expansion, while the remainder presumably deposited at the grain boundary. The drop in lattice parameter at 615C, which accompanies a drop in C content (Figure 5-43), indicates decreased C incorporation into the ~-WNxCy polycrystals at this temperature. The discontinuity in the C incorporation process suggests a change in the C deposition mechanism near this temperature. 4 .24 80 a ,..._ b "' ,..._ s "' 4.22 70 s 0 l,2DCB 0 1,2DCB ti ti "' -0PhCN Oil -0PhCN 4.20 60 '-' ';:;' 4 .18 II) N v5 50 II) = 4.16 ca la 0 40 ll; II) II) 4.14 Oil u s II) 30 "' 4 12 ;;,. ...l
PAGE 296

287 5.11.2 Film Composition 5.11.2.1 AES Results Elemental compositions of films deposited with PhCN and 1,2-DCB as determined by AES are shown in Figure 5-43. Since standard film samples for calibration of our elemental concentrations are unavailable, the error bars on the AES compositions are several atomic percent. Despite the uncertainty in the concentration values, the AES data serve to identify trends in film composition with deposition temperature. The change in C content with deposition temperature was significantly different for the two solvents. The C levels were independent of temperature below 500C, with films from PhCN and 1,2-DCB solutions containing -12 and -5 at.% C, respectively. For films grown from both solvents, an increase in C content above 500C coincided with the onset of film crystallization, which suggests that polycrystals and/or grain boundaries may promote C deposition. Above 500C, the C content increased steadily for films from PhCN solutions. For films from 1,2-DCB solutions, C levels increased between 500 and 600C, and then dropped, reaching a constant value near 50 at.% up to 700C. The decreased C level above 600C reflects a change in C deposition behavior, but the cause of the change is unclear. This effect may be due to a side reaction in the gas phase and/or on the reactor walls in this temperature range, or due to a competing reaction on the film surface. Although trends in N content are evident, exact measurement of the N level by AES is problematic, due to preferential incorporation of C and removal of N by Ar+ sputtering [Ing82]. The films from 1,2-DCB solutions had lower N content than those from PhCN, especially at lower temperature, with maximum levels of 4 and 11 at. %

PAGE 297

288 respectively, at 500C. This suggests that the CN group, produced by PhCN decomposition, is deposited on the film surface, leaving adsorbed C and some N Direct decomposition could occur via 111 bonding (through the N cr lone pair) of PhCN to the film surface [Mro0l], with subsequent cleavage of the C-Ph bond. An alternative is a stepwise process that would entail homogeneous decomposition of PhCN to HCN, which then adsorbs onto the film surf ace. Since homogeneous decomposition of HCN requires temperatures well above our experimental range [Sze84], HCN decomposition in this system would be heterogeneous. If the CN moiety contributes to N levels in films from PhCN solutions, decreased N levels would be expected for films from 1 ,2-DCB solutions, in which the solvent contains no N. An alternative explanation involves selective N etching by additional HCI produced by 1,2-DCB decomposition. The N level in the films from 1,2-DCB solutions is relatively constant (-2 at.%) at and above 650C, however, which makes HCI etching (which should accelerate with temperature) unlikely. Moreover, HCI etching of WNx has not been reported. Chemical etchants for tungsten nitride typically involve the use of H2O2 [lva99] and are not selective for N. The N content for films grown from both PhCN and 1,2-DCB solutions is similar at the highest deposition temperature. While low temperature deposition may enable N addition to the films from PhCN solutions, increasing deposition temperature enhances desorption of this N as N 2 At the highest temperature, most of this N in the films grown from PhCN solutions has desorbed, leading to similar N levels as the films deposited from 1 ,2-DCB solutions, whose N is derived entire I y from the precursor.

PAGE 298

289 80 ~-----------------, 60 tt s 40 .,, j 20 .... l,2DCB -o-PhCN w a 0 +--~-~--r---,----r-----,-----1 400 450 500 550 600 650 700 750 Deposition Temperature ( C) 80 ..-------------------, 60 tt s 40 .,, C, 8 20 8 .._ 2DCB -+ PhCN C C 400 450 500 550 600 650 700 750 Deposition Temperature ( C) 80 ..-----------------, 60 s 40 ., i o 20 8 .. 1.2DCB -oPhCN N b o L___;~=!=~~~~~;ae;;elil----_:__i 400 450 500 550 600 650 700 750 Deposition Temperature (0C) 80 ..----------------, j 60 tt s 40 ., j 20 ... l ,2DCB -o-PhCN 0 400 450 500 550 600 650 700 750 Deposition Temperature ( C) Figure 5-43 Variation of film content with deposition temperature and solvent. a) W b) N c) Cd) 0. Data are from AES after 2.0 minutes sputter. Oxygen in the films originated from post-growth exposure to air. Substantial levels of C and N, which help the film's polycrystals to resist post-growth 0 in-diffusion, were deposited with PhCN solutions, even at the lowest temperature. Films from 1,2-DCB solutions deposited at lower temperature contained lower C and N levels, leaving a significant amount of bare W, which is susceptible to oxidation. The O level in the films from 1,2-DCB solutions was significantly lower above 500C, due to film crystallization and increased C levels, which make the film denser and enable it to resist 0 in-diffusion. Chlorine was not detected in the films by AES for either solvent, placing an upper limit of -1 at. % on the Cl content. Although use of 1,2-DCB increases the amount of Cl

PAGE 299

290 in the reactor, the decomposition of chlorinated aromatics reportedly accelerates in a H2 atmosphere relative to an inert one [SerOl], producing gaseous HCl as the principal chlorine-containing product. This likely occurs in the reactor under deposition conditions, causing minimal Cl incorporation into the films. 5.11.2.2 XPS Results XPS analysis indicated that polycrystalline films deposited from PhCN solutions 35000 4000 W4f Nls ,-.._ 30000 ,-.._ 700 C "' "' .... .... 3000 a 25000 a :::i :::i 20000 b j 2000 15000 600C .._.. .._.. 1000 >-. 10000 q .... 'cil "' C 5000 C .... .... 0 C C -..... 0 a b 42 40 38 36 34 32 30 406 404 402 400 398 396 394 Binding Energy (eV) Binding Energy ( e V) 12CXX) 5CXX> ~10000 4CXX) a a 100c :::i 8CXX) :::i i 700 C 3CXX) 6000 ti 4CXX) 2CXX> ,._, ,._, 0 >-. .<;:: l ;;; 2CXX> "' r= r= 600C .s 0 .s 0 C d 290 288 286 284 282 538 536 5 3 4 532 530 528 Binding Energy (eV) Binding Energy (eV) Figure 5-44 XPS spectra for films deposited at 600 and 700C with 1,2-DCB after 10 min sputter. a) W4f b) Nls c) Cls and d) Ols BE lines. Dashed lines are deconvoluted peaks.

PAGE 300

291 contained W in the nitride/carbonitride (B-WNxCy) bonding state, as well as a small amount of oxide (WO 3 ) [Bch03b]. Carbon was present at interstitial lattice sites in the B-WNxCy polycrystals and also in amorphous form at the polycrystal (or nanocrystal) boundaries, while N preferentially occupied interstitial sites in the B-WNxCy poly-(or nanocrystals). Oxygen was present in the WO 3 bonding state. Films from 1,2-DCB solutions had similar XPS results, as shown in Figure 5-44. The results are summarized in Table 5-12. Question marks in Table 5-12 indicate that exact location of the corresponding W4f peaks for the WOx suboxide is unknown, due to lack of additional shoulders in the W4f spectra. Table 5-12. Compounds in Films Grown with i-Pr in 1,2-DCB After 10 Minutes s ;putter Deposition W4f112 W4fs12 Nls Ols Cls Compound Temp. (C) (eV) (eV) (eV) (eV) (eV) 600 32 2 34.1 397.9 -286.5 B-WN xCv 600 34.9 36.8 -530 8 WO3 600 ? ? -532.1 WOx 600 ----284.6 FreeC 700 32 1 34.0 397.9 -286.0 J3-WNxCv 700 35.0 37.0 -530.8 WO 3 700 ? ? -532.6 WOx 700 ----284.4 FreeC 5.11.3 Carbon Deposition Rate To determine the C deposition rate, the number of atoms deposited in each film was first estimated using an atom balance approach with W, N C and O atoms situated in the film based on XPS results. The analysis was restricted to polycrystalline films which contain both polycrystals and grain boundary regions because they have relatively low porosity and steady oxygen levels. Using a basis of 100 atoms, for example, the number of W atoms can be calculated from the AES fraction. From XPS data, the

PAGE 301

292 location of W would be either in B-WNxCy polycrystals or in WO3. The number of W atoms bonded in WO 3 assumed to be 1/3 the number of oxygen atoms, was subtracted to give the number of W atoms in the B-WNxCy polycrystals. The (W)1(N,C,Va)1 two-sublattice model [Hua97, Sau98], with N, C and vacancies (Va) intermixing on the interstitial FCC W sublattice, was coupled with Vegard's law [Kos93] and lattice parameter data [Bch03c] to approximate the C content and density of the polycrystals. The grain boundaries were assumed to contain any remaining C in amorphous form [Bch03b], with an assumed density of 3 g/cm3 [Koh95], along with a small amount of amorphous W03 with an assumed density of 6.4 g/cm3 [Sch87] The volume of the 100 atom basis set was calculated by summing the volume contributions from the B-WNxCy polycrystals, the amorphous WO 3 and the amorphous carbon. The density of W atoms in this basis set, Pw,basis (W atoms/cm3), was calculated by dividing the number of W atoms in the 100 atom basis by the volume of the basis set. The total number of W atoms CWT) deposited was then estimated with Equation 5-6: WT =(a *t)*(l-VF)*Pw,basis (5-6) where a is the area of the film (cm2 ) and t is film thickness (cm). The area of the film was assumed to be 2 cm2 which is the typical total cross sectional substrate area input to the reactor, and t was determined by XSEM. A film void fraction (VF) of 0.10, per a previous report, was included to account for any defects/pores in the film structure [She00c]. The total number of carbon atoms (CT) in each film was estimated with Equation 5-7 below: C = ( WT )*%C T %W (5-7)

PAGE 302

293 where% Wand% Care the atomic% of Wand C, respectively, as determined by AES. The area-averaged nature of AES ensures that W and C contents are measured regardless of location in the film's microstructure. Dividing CT by deposition time gives the C deposition rate, which can be estimated by an Arrhenius expression in the kinetically controlled regime (Equation 5-8): D = Ae-E,tRT (5-8) where Dis C deposition rate (atoms/min), Ea is the apparent activation energy (eV) for C deposition, and A is a constant. Figure 5-45 shows a plot of ln D vs. 1/T for films deposited with PhCN and 1,2-DCB. A line with a steep slope indicates a kinetically controlled regime (slope = -EalR), whereas a line with small slope indicates mass transfer control. If the rate of carbon deposition shows an exponential dependence, the process is kinetically controlled. If the rate does not show this dependence, the carbon deposition process is assumed to be mass-transfer controlled. Error bars on Figure 5-45 address several sources of uncertainty in the calculation of carbon deposition rate. Uncertainty in film thickness was due to error in thickness measurement from X-SEM photos ( 0.02 m), roughness on the film surface ( 0.01 m) as determined by AFM, and fluctuations in film growth rate due to experimental temperature variation { 10C) Uncertainty in film density was due to lattice parameter shifts from X-ray line broadening ( 0.002 A). Uncertainty in C deposition rate was due to uncertainty in the number of W and C atoms from AES ( 5 at. % each). Figure 5-45 indicates C deposition rate to be kinetically controlled over the entire temperature range for PhCN, suggesting that a single process dominates C deposition at

PAGE 303

294 the surface of these films The apparent activation energy for C deposition from PhCN was 0.70 0.10 eV. Deposition Temperature (0C) 700 675 650 625 600 575 550 525 500 39 ....,.1,2 DCB -0PhCN 38 37 Cl .E 36 35 34 LOO 1.05 1.10 1.15 1.20 1.25 1.30 10oorr Figure 5-45 Dependence of carbon deposition rate on growth temperature for polycrystalline films grown with PhCN and 1,2-DCB solvent. Figure 5-45 indicates kinetically controlled C deposition at and below 600C for films from 1,2-DCB solutions, with an apparent activation energy for C deposition of 1.0 0.14 eV. The drop in C level for films from 1,2-DCB solutions above 600C was accompanied by a drop in film thickness, as measured by XSEM. The region between 600 and 625C appears to be a transition zone between different C deposition processes, but the mechanisms for these processes are unclear. The film growth rate from 1,2-DCB solutions at and below 600C was exponentially dependent on temperature, reaching a maximum of 43 Nmin at 600 C. The growth rate dropped to 22 Nmin at 650C. This behavior contrasts with that for films deposited from PhCN solutions, whose thickness increased steadily from low to high temperature.

PAGE 304

295 Above 625C, film growth and C deposition rates for films from 1,2-DCB solutions appear to be mass transfer controlled, with both increasing slightly up to 700C. Kinetically controlled C deposition for films from 1,2-DCB solutions has a higher apparent activation energy than the value obtained for films deposited from PhCN solutions. This difference is consistent with a change in the C deposition process between the two solvents, which is expected, assuming that the nitrile carbon group (which is present in PhCN but absent in 1,2-DCB) controls C deposition in films from solutions of the former. Residual gas analysis indicated the presence of PhCN and 1,2-DCB molecular ions, as well as various fragments associated with solvent decomposition. While a full analysis of the fragmentation pathways is beyond the scope of this dissertation, Figure 546 shows partial pressures for several products whose m/z ratios correspond to known solvent decomposition products. The measured partial pressures of the various fragments were adjusted to compensate for contributions from electron impact (El) fragmentation in the RGA [NIS03]. The RGA data are included to show which species were present in the gas phase for depositions with the two solvents. The species concentrations are generally insensitive to temperature, however, which reflects the large excess of gas phase carbon-containing species relative to the amount of any particular specie(s) removed from the gas by deposition on the substrate. For deposition with PhCN, partial pressure readings at m/z = 2, 26, 27, 36 and 76 are consistent with formation of H2 C2H2 HCN, HCl and C6~, respectively [Las96, NIS03]. Significant quantities of HCN (m/z = 27), the principal carbon-containing fragment, along with C~ (mlz = 76), are consistent with PhCN decomposition in the

PAGE 305

296 reactor. The other carbon-containing product, C2H2 (ml z = 26) is also a potential C source. The formation of C 6Hi and H 2 by benzene (C 6~) adsorption on lr(lOO) [JohOl] and Mo(l 10) [Liu88] has been reported, and may explain why C6~ was not observed with the RGA. For deposition with 1,2-DCB, peaks are again observed at mlz 2, 26 and 36 along with readings at m/z = 50 and 75 which are consistent with formation of CJI2 and C6H3, respectively [NIS03, SchOl]. Figure 5-46 also indicates high partial pressures of H2, which are produced by solvent and ligand decomposition. 1,2-DCB solutions also give rise to a partial pressure of C 6 H 3 the principal carbon-containing fragment, which is almost an order of magnitude higher than its PhCN counterpart, C6Ri Another carbon-containing fragment, CJI2 also has relatively high partial pressure. These fragments are likely contributors to C deposition from 1 ,2-DCB. Similar levels of C 2 H 2 are present for 1 ,2-DCB and PhCN solutions. Surprisingly the HCI level with 1,2-DCB was just slightly higher than for films from PhCN solutions ; however this may not be the case for depositions utilizing H 2 as the carrier gas, which should facilitate HCI formation. The H 2 level initially decreases with increasing temperature reaching a minimum at 600C, and then increases above this temperature. The increase in H 2 above 600 C suggests a change in reaction mechanism at higher temperature. Results for RGA runs with neat PhCN and 1 ,2-DCB were similar to those containing precursor 1, except that in both cases, the concentration of the principal decomposition product (HCN and C 6 H 3 respectively) was higher than the molecular solvent ion. This suggests that the primary decomposition products may

PAGE 306

297 interact with the growing films and deposit C, causing their concentration to drop relative to that for neat solvent. le-7 le-7 a~ b mlz= 2 (H22. .--. .--. I:: I:: 0 0 f--4 f--4 m/z = 146 (1,2-DCB) '-' '-' 0 =3 le-8 ::I le-8 mlz = 75 (CJ{3 ) "' "' "' "' 0 V V V V V m/z = 27 (HCN) i::I::: mlz ; 50 (C4Hz) .; .; ~ -~ le-9 mlz= 36 (HCU le9 "8 :=:= ] m/z = 36 (HCI) ... i "' "' ::I m/z = 76 (C6H4 ) ::I ;a' ;a' < < le-10 le-10 450 500 550 600 650 700 450 500 550 600 650 700 Deposition Temperature {0C) Deposition Temperature {0C) Figure 5-46 Dependence of adjusted partial pressure on temperature for selected gas phase species. a) Deposition with PhCN. b) Deposition with 1,2-DCB. 5.11.4 Conclusions On Effect of Solvent Change Carbon and nitrogen levels in the deposited films were highly dependent on the nature of the solvent used. The activation energy for C deposition in WNx films grown from la in PhCN solutions was 0.70 0.10 eV, while that for 1,2-DCB solutions was 1.0 0.14 eV This shift in activation energy for C deposition upon changing the solvent is evidence for an alteration in the C deposition process. Higher N content for films deposited with la in PhCN solution relative to those from 1,2-DCB solutions suggests that CN moieties derived from the solvent decompose on the film surface, leaving adsorbed C and N RGA results indicated substantial levels of HCN, along with some C6R., consistent with decomposition of the PhCN solvent in the reactor. The C deposition source for films from 1 2-DCB solutions is unclear, although deposition behavior suggests a kinetically controlled surface process at and below 600C. Films

PAGE 307

298 deposited from 1,2-DCB solutions above 600C had decreased growth rate and C levels, which may be due to a change in C deposition process at the higher temperatures. 5.12 Conf ormality Tests Figure 5-47 indicates that for deposition with lb at 450C on Si (100) with 2 m diameter features and AR= 2.75, sidewall and bottom conformality were 90 and 50%, respectively. Deposition at 500C resulted in sidewall and bottom conformalities of 45 and 25%, respectively. As expected, film conformality improves at lower deposition temperature, for reasons discussed in Chapter 1. The conformality values at 450C, while better than those at 500C, are inadequate for industrial applications, however, especially considering that the tested feature diameter (2 m) was much larger than the current feature diameter used in IC production (0.1 m). In addition, film roughness in and around the feature is larger than that for deposition on flat Si (100) substrates. The reason for this is unclear. Figure 5-47 Conformality for films deposited with i-Pr on patterned Si (100) substrates. a) 450C. b) 500C. 5.13 Adhesion Tests Adhesion to Si (100) substrates for films deposited with and without NH3 was tested using the Scotch tape method. In this method, a small grid is scribed onto the film surface, and Scotch tape is then pressed directly onto the film. The tape is then pulled off

PAGE 308

299 of the film, and is inspected to determine if the film delaminated and stuck to the tape. Although this method does not give quantitative data on the strength of a film's adhesion to the underlying substrate, it is a quick way to determine if the films are prone to easy delamination. All of the films tested in this manner, which were deposited with and without NH3 passed this test, as summarized in Table 5-13 below. This indicates that all of the films had good adhesion to the underlying substrate, with the film-substrate adhesion strength being higher than the film-Scotch tape adhesion strength. Table 5-13. Scotch tape test results for films deposited with and without NH3 on Si (100) substrates Deposition Chemistry Deposition Temperature Scotch Tape Test Results (aC) 450 Passed 500 Passed lb+PhCN +H2 550 Passed 600 Passed 700 Passed 450 Passed 500 Passed lb+ PhCN + H2 +NH3 550 Passed 600 Passed 700 Passed 5.14 Conclusions on Use of Cl4(CH3CN)WN-i-Pr to Deposit WNx Repeatable deposition of films using the i-Pr precursor with and without NH3 was achieved. Use of 1,2-DCB in place of PhCN solvent resulted in significantly different C incorporation behavior. The PhCN solvent appears to contribute both C and N to the films, which typically contain FCC ~-WNxCy and amorphous C, with some amount of W03 depending on deposition temperature. All of the films had good adhesion to the Si (100) substrates, which is vital to ensure good device lifetime.

PAGE 309

300 The minimum film deposition temperature with lb was 450C, which is higher than the current temperature ceiling for IC production ( < 400C). Films deposited with the i-Pr precursor at the lowest achievable temperature were very porous, enabling them to absorb significant amounts of O from the air. While addition of NH3 at low temperature decreased the O content, these films were also too porous, and had very high resistivity. Chlorine present in the films (per SIMS) is also a concern, as it can corrode subsequent metallization layers and etch Si substrates. In addition, a -60 A thick interface layer containing a mixture of WSh, WO3 SiO2 and SiO was present between the film and the Si substrate. This layer alone is almost as thick as the current required barrier thickness (100 A), and its presence makes the deposition of ultra-thin (100 A), high quality WNx films by MOCVD using lb in our reactor unlikely. The overall film thickness was -1500 A, which is much larger than the currently required barrier thickness. While decreasing deposition time would decrease the film thickness, it would not result in a barrier film with appropriate properties, because the properties of the bulk film and the interface layer are not the same. Moreover, a significant incubation time on the Si substrates would make it difficult to rapidly deposit a barrier film with appropriate thickness in a reasonable amount of time. Lastly, conformality of films deposited at 450C on large diameter features was mediocre, suggesting that deposition on state-of-the-art feature sizes would be poor. While we do have state-of-the-art substrates with 100 nm features (provided by Intel), the recommended temperature ceiling for these substrates is -400C due to the presence of a low-k dielectric material on the substrate surface. A minimum deposition

PAGE 310

301 temperature of 450C was required for deposition with the i-Pr precursor, hence the use of these substrates will be deferred to future ALD studies. ALD using this precursor may be a promising way to deposit denser films at low temperature. The increase in N content due to NH3 addition during CVD suggests that a stepwise i-Pr/NH3 pulse scheme may enable ALD deposition. ALD should improve smoothness and conformality of the films, and enable ultra-thin film deposition. The Cl content of the precursor would still be an issue during ALD, however. Modification of this precursor to eliminate Cl, as well as boosting its N content, would be preferable. Finally, further investigation of solvent schemes should be carried out, to determine if another solvent with higher thermal stability than PhCN could be used to limit the solvent's participation in the deposition process.

PAGE 311

CHAPTER6 EVALUATION OF Cl4(PhCN)W(NPh) AS A SUIT ABLE WNx PRECURSOR MOCVD of thin WNxCy films from benzonitrile solutions of the single-source imido precursor Cl4(CH3CN)W(NiPr) (la) as a mixture with its benzonitrile derivative lb was discussed in Chapter 5. The simple synthesis of la,b was adapted to produce a series of related complexes containing different alkyl or aryl substituents on the imido ligand. This feature allowed variation of the N-C bond dissociation energy, which is important since this bond must be cleaved during the CVD process. The N-C (imido N to 2 alkyl) bond of la,b will be relatively weak, compared to the N-aryl bond of the phenylimido complex Cl4(PhCN)W(NPh) (2b). These precursors are shown schematically in Figure 6-1. WNx (and WNxCy) films were deposited from 2b, and a comparison of material deposited from la,b and 2b was performed to determine if the imido N-C bond strength has an effect on the MOCVD process. 1a R = CH3 1b R = Ph y N c1,,, Ill .,,c1 w c1_........1........_c1 N Ill C I R 2a R = CH3 2b R = Ph Figure 6-1. Structures of the Cl4(RCN)W(NiPr) and Cl4(RCN)W(NPh) precursors. 302

PAGE 312

303 6.1 Film Growth Studies Experiments were conducted in the CVD reactor system described in Chapter 4. P-type boron doped Si (100) substrates with resistivity of 1-2 Q-cm were used for the film growths. Growths were conducted for a fixed time of 150 minutes at temperatures ranging from 450-800C. The system was maintained at vacuum by a mechanical roughing pump with the operating pressure fixed at 350 Torr. Hydrogen (H2) carrier gas was used for the depositions 6.2 Synthesis of C1'(PhCN)WNPh The complex Cl4(OEt 2 )W(NPh) (3) was synthesized as previously described by Schrock et al. [Ped82]. The benzon itrile complex Cl4(PhCN)W(NPh) (2b) was no t isolated, but was produced in situ by the substitution of the ether ligand of 3 with benzonitrile following solvation in the 10: 1 PhCN/Et 2 O co-solvent utilized for the deposition experiments The acetonitrile complex Cl4(CH3CN)W(NPh ) (2a ) was prepared as previously reported [Nie83]. 6.3 Solvent Selection Deposition of thin films by MOCVD requires transport of the solid phenylimido precursor 2b to the reactor in the vapor phase. Previous tests with similar complexes in a solid source delivery system resulted in minimal precursor transport, due to the low vapor pressure of the compounds. Transport difficulties were overcome by using a nebulizer to generate an aerosol of the precursor/solvent mixture, which is conveyed by carrier gas to the reactor. Although benzonitrile is an appropriate solvent for deposition from isopropylimido complexes la,b, poor solubility of the phenylimido complexes in benzonitrile necessitated a co-solvent mixture of 10 : 1 benzonitrile:ether to achieve the

PAGE 313

304 same precursor concentration (7 .5 mg/mL) previously used with la,b. To determine the impact, if any, of the co-solvent on film composition, acetonitrile was tested in place of ether. AES results indicated similar film compositions regardless of the co-solvent used. 6.4 Precursor Mass Spectral Pre-Screen All mass spectral analyses were again performed using a Finnigan MAT95Q hybrid sector mass spectrometer (Thermo Finnigan, San Jose, CA). Electron ionization (El) was carried out in positive ion mode using electrons of 70 eV potential and a source temperature of 200C. Negative ion electron capture chemical ionization (NCI) used methane as the bath gas at an indicate pressure of 2x10-5 Torr, an electron energy of 100 volts and a source temperature of 120C. All samples were introduced via a controlled temperature probe with heating and cooling enabling temperature control down to 35C. The mass resolving power (m/~m) was 5000 full width-half maximum (FWHM). As noted before, care must be taken in using mass spectral data to predict CVD behavior since the latter is thermal in nature [Lew94]. Nevertheless, mass spectrometry does provide insights into the relative fragmentation characteristics of various precursors [lnt89]. In order to provide a direct comparison of mass spectral data for a phenylimido complex and its isopropyl analogue, the acetonitrile adduct Cl4(CH3CN)W(NPh) (2a) was prepared as a model for 2b. Precedent for the modeling of 2a by 2b is provided by the virtually identical ion fragmentation patterns for the isopropyl imido complexes la and lb [Bch03a]. Mass spectrometry results for 2a could then be compared to our previously reported data for the isopropyl imido complex Cl4(CH3CN)W(NiPr) (la) [Bch03a]. Accordingly, 2a was analyzed using both positive ion electron-impact (El) and negative ion electron-capture chemical ionization (NCI) methods.

PAGE 314

305 The EI and NCI mass spectra of 2a are depicted in Figure 6-2, and the relat i ve abundances of the observed peaks of the phenylimido and isopropylimido complexes 2a and la are summarized in Table 6-1. As observed in the mass spectra of the isopropyl precursor, no molecular ion was detected for 2a using e i ther ionization method. The base peak in the EI spectrum occurs at m/z=382 and corresponds to the fragment [Ch W(NPh)t. The highest mass peak observed in the EI spectrum was [Cl4 W(NPh)t at m/z=417 (9% abundance). Interestingly, although the high mass envelopes correspond to fragments in which acetonitrile is lost, only a small amount (-1 % abundance ) of the [CH3CNt ion was detected at mlz =41 i n the EI spectrum of 2a. The presence of the [Cl4Wt and [ChWt fragments suggests cleavage of the W-N bond occurs in the gas phase. Furthermore, observation of the [Pht fragment (m/z=77) indicates that the crit i cal N-Ph bond is broken to some extent under EI conditions; however, there is no e v idence of a metal nitrido fragment in the resulting spectrum. Moreover the base peak in the NC I spectrum corresponds to [Cl4W(NPh)r (m/z =417) while the mass envelope of the nitride fragment [Cl 4 WNr (mlz =340) has a relative abundance of 4%. The presence of the fragment [Cl5W(NPh)r suggests that the nitrile ligand of 2a is removed during the process of heating the condensed phase sample to afford [W(NPh)Cl 4h prior to ionization. The mass spectral data for phenylimido complex 2a and isopropylimido complex la show some similarities. For both, the most prevalent ion on the high mass end of the EI spectrum corresponds to [Ch W(NR)]+, although 2a does a l so exhibit lower abundan c e peaks from [Cl 4W(NPh)t. This coupled with the lack of molecul ar ion signals i s consistent with high lability of the nitrile ligand in both complexes Mo r eo ve r th e

PAGE 315

306 presence of the [Cl5W(NPh)r fragment in the NCI spectrum suggests that the !ability of the nitrile ligand results in partial conversion of 2a to the dimer [W(NPh)Cl4h prior to ionizations. The observation of this chloride transfer process in 2a, but not la, is consistent with the greater electron withdrawing nature of the phenyl substituent, as compared to isopropyl. Table 6-1. Summary of relative abundances for positive ion EI and negative ion NCI mass spectra of tungsten imido complexes Cl4(CH 3 CN)W(NPh) (2a) and Cl4(CH3CN)W(NiPr) (la). Complex EI Fragments NCI Fragments m/z Abundancea,0 [Cl4 W(NPh)t 417 9 [ChW(NPh)t 382 100 [Cl4Wt 326 7 [ChWt 291 15 [PhNt 91 <1 [Pht 77 22 [Cs~t 64 1 Cl4(CH 3 CN)W(NPh) (2a) [C~3t 51 10 [CH3CNt 41 1 [C3H2t 38 1 [ClsW(NPh)r 452 23 [Cl4 W(NPh)r 417 100 [Cl4WNr 340 4 [Ch W(N'Pr)t 348 100 [C4Wt 326 26 [ChWNH]+ 306 78 [ChWt 291 30 Cl4(CH 3 CN)W(N1Pr) (la) [CH3CNt 41 24 [Cl4 W(N'Pr)r 383 42 [Cl4WN] 340 100 aRelative abundances were adjusted by summing the observed intensities for the rredicted peaks of each mass envelope and normalizing the largest sum to 100%. Values for la are from Chapter 5.

PAGE 316

Cl) u C l'CI "C C ::, .0 < -~ 'tu 'ii 0:: Figure 6-2. 307 382 100 77 [Ph]+ 80 / a=[C3HJ [Cl3W(NPhW 60 b b=[C4HJ '1t c=[C5HJ 51 d=[C6H5NJ 40 (Cl,WJ a I [Cl4W(NPh)J+ '1t C d .,,I w 291 20 38 '1t [Cl3W] 64 326 91 0 50 100 150 200 250 300 350 400 450 m/z 417 100 [Cl5W(NPh)J80 [Cl4W(NPh)J/ 452 20 O 50 100 150 200 250 300 350 400 450 500 m/z Positive ion electron-impact ionization (EI) and negative ion electron capture chemical ionization (CI) mass spectra of 2a. The most notable difference in the spectra of 2a and la concerns the fragments [ChWNHt and [Cl4WN] Since these ions are derived from cleavage of the N-R bond

PAGE 317

308 of the imido moiety, they are critical to the CVD process. As shown in Table 6-1, [Ch WNH]+ appears in the EI spectrum of la with a relative abundance of 78%; however, this fragment is not present in the spectrum of the phenylimido complex 2a. Nevertheless, the presence of [Pht indicates the N-Ph bond is broken to a certain extent under EI conditions. The ion [PhNt (m/z=91) was observed in very small relative abundance ( <1 % ), and its subsequent fragmentation may be responsible for the small clusters of peaks centered at mlz=64, 51 and 37 [Cro67]. Even more striking is the fact that the [Cl4 WNr fragment is the base peak in the NCI spectrum of la, but only accounts for 4% relative abundance in the phenylimido complex 2a. In relation to the use of Cl4(PhCN)W(NPh) (2b) as a precursor for tungsten nitride deposition, the mass spectral data of 2a and la suggest that the N-Ph bond is more difficult to break than the N-iPr bond. This is consistent with the homolytic bond strength of the two N-R moieties [Ben76]. If N-Ph bond cleavage were involved in the rate determining step, 2b would be expected to require higher deposition temperatures relative to the isopropyl system. The stronger N-Ph bond may also affect growth rate and composition of the deposited films. Additionally, these data suggest that replacing the phenyl moiety with a group that will cleave more readily (e.g., allyl) could decrease the deposition temperature and improve the compositional characteristics of the WNx films. 6.5 Film Structure 6.5.1 XRD The films typically had a smooth, shiny metallic appearance with color varying from black to gold, depending on the deposition conditions. The desired WN x film

PAGE 318

309 structure and stoichiometry is face centered cubic (FCC) B-W 2N (B-WNo s), since this phase has the lowest resistivity. The X-ray diffraction (XRD) spectrum in Figure 6-3a is consistent with amorphous film deposition from 2b at 475C, as evidenced by a flat profile with no characteristic B-WNx peaks. In contrast, the XRD spectrum shown in Figure 6-3b indicates polycrystalline film deposition at 750C. Four characteristic peaks are evident, with relative peak intensities indicating that no preferred crystal orientation exists. Although the relative peak intensities in Figure 6-3b are consistent with the pattern for polycrystalline B-WN0 5 the 20 peak positions lie between the standard values for B-WNo.s and B-WCo .6, which are shown in Figure 6-3c. Peak positions between these standard values suggest that C is mixing with N and vacancies on tungsten's interstitial sublattice to form B-WNxC y polycrystals. For the spectrum in Figure 6-3b, primary reflections at 37 .13 and 43.08 20 degrees are consistent with (111) and (200) B-WNxCy growth planes, while additional reflections at 62.73 and 74.98 20 degrees indicate (220) and (311) planes, respectively. No peaks arising from the hexagonal WN or WC phases, or the body centered cubic (BCC) a-W phase were evident for any of the films. Figure 6-4 illustrates the evolution of film crystallinity with deposit i on temperature for growth from 2b. The trend of increasing crystallinity with deposition temperature is similar to that observed for the isopropyl i mido precursor lb [Bch03a]. At the lowest deposition temperature (475 C) the characteristic B-WNx peaks are not observed, indicating that the film is amorphous.

PAGE 319

310 ,-... 500 Z3 Si (400) Ka a :::> 400 b 300 ..... Si (200) 200 / '-' ;;,-. ..... ..... 400 B-WN xCy (200) b 300 I ..... B-WNxCy '-' 200 (220) 0 100 ..... 80 60 b 40 '-' 0 ..... 20
PAGE 320

311 consistent with WNx (or WNxCy) are seen at all deposition temperatures above this. Reproducible deposition of tungsten oxide at a single temperature suggests that an air leak in the reactor system is unlikely to be the source of 0. Microstructure dependent, post-growth O incorporation is a likely cause of oxide formation, especially for low temperature, "clean" tungsten films with low C and N levels. This is due to lower contamination levels, which make these films less resistant to O in-diffusion and reaction. The existence of a second O incorporation pathway involving the Et 2 0 co-solvent, however, cannot be ruled out. Diethyl ether is known to undergo both homogeneous and heterogeneous decomposition at temperatures near 500C [Fle36, Fou79]. Although oxide formation still occurs for films grown at 500 C from Cl4(PhCN)W(NPh) (2b) in a benzonitrile/acetonitrile mixture (no Et20 present), the size of the oxide crystallites is larger when the ether complex C4(Et 2 0)W(NPh) (3) is used to generate precursor or when Et 2 0 is the co-solvent. Even though the film is amorphous at 475C, indicating that inadequate thermal energy is available to produce oxide polycrystals, the possibility of O incorporation similar to that at 500 C cannot be ruled out. High levels of O in the amorphous film, as demonstrated by AES data, support this possibility. A broad peak near 38.03 28 degrees, which is above the standard peak position for ~-WN0 5 (111), appears for deposition at 525C, and indicates the presence of N-deficient polycrystalline ~-WNx (111). A broad peak at 44.03 20 degrees for deposition at 550C indicates the first appearance of ~-WNx (200). As the te mperature increases to 600 C, film peak positions shift below the standard ~-WN0 5 positions indicating that f3-WNxCy deposition is occurring. Broad peaks emerge at 62.93 and

PAGE 321

312 75.58 20 degrees for deposition at 700C indicating '3-WNxCy (220) and (311) growth The peaks sharpen further as the temperature increases to 750C, due to polycrystalline grain growth. Peak sharpness decreased somewhat for the film deposited at 800 C (not shown) which is likely due to a decrease in film thickness as measured by XSEM. Decreased film thickness may be caused by rapid desorption of reactants from the film surface due to high temperatures or by gas phase reactions, which deplete the amount of reactant available at the film surface. Some films deposited between 475 and 600 C displayed additional peaks at 32.98 and/or 61.68 20 degrees representing reflections from the underlying Si consistent with Si (200) Ka and Si (400) K'3 radiation, respectively. T=700 C T=675 C T=650 C 2000 T=625 C T=600 C T=575 C ti'.) 1000 _,....._,,.,,,,,.,....,.,,_.,_,.___,~...!.T.;:.=5~5~0:_0zC_,......,.;.,,,'I' 'E T=525C WO x -----+ T=500 C '------_ .. ,._.,.,,.,. T=475 C ,._Si(400) Ka. ~-WN x C y ~311) 0-+-----r-------.-----~----~---~ 30 40 50 60 70 80 20 Degrees Figure 6 4. Change in XRD pattern with deposition temperature for WNxCy grown from 2b on Si (100) in a H 2 atmosphere.

PAGE 322

313 At the highest deposition temperature, minimal N and high C levels suggest that the stoichiometry of the P-WNxCy polycrystals may be approaching P-WCy deposition. A shift in the position of the peaks toward P-WCo.6 at higher deposition temperature provides evidence of this trend. 6.5.2 Lattice Parameter The dependence of lattice parameter on deposition temperature in Figure 6-5 was determined by XRD using the 20 position of the (111) P-WNxCy diffraction peak, with peak position calibrated to the Si (400) diffraction peak. The standard P-WNo.s (111) and P-WC0 .6 (111) peak positions are 37.735 and 36.977 20 degrees (Figure 6-3c), respectively, and correspond to standard lattice parameter values of 4.126 and 4.236 A. The position of the (111) reflection peak can vary as a result of a change in composition or a change in the film's residual stress. 4.24 j 4.22 g V} 4.20 0() = $ 4.18 I-. 4.16 J 4.14 0 -~ 4.12 ...:l 4 .10 500 600 700 800 Deposition Temperature (0C) Figure 6-5. Lattice parameters for films grown from lb and 2b, based on the P-WNxCy (111) diffraction peak. The dashed line at 4.126 A represents the standard lattice parameter value for P-WNo.s, while the dash-dot line at 4.236 A is that for P-WCo.6Error bars indicate uncertainty in lattice parameter( 0.002 A) due to X-ray Ka line broadening.

PAGE 323

314 Assuming that compositional variation is responsible for the peak shift and, thus, lattice parameter change, trends in the relative concentrations of N, C, and Va on the interstitial sublattice can be suggested by coupling the peak shift with the compositions measured by AES (Figure 6-7). If the (111) peak position is higher than 37.735 20 degrees, corresponding to a lattice parameter below 4.126 A, then minimal C, a deficiency of N, and an excess number of vacancies exist in the polycrystals. This is the case for depositions between 525 and 575C. A peak position between 37.735 and 36.977 20 degrees, which corresponds to a lattice parameter between 4.126 and 4.236 A, suggests mixing of N, C and Va on the interstitial sublattice This is the case for depositions at and above 6OOC. As mentioned above, however, P-WNxCy polycrystals may be forming even at the lower temperatures, with C playing a role in the lattice parameter increase at lower temperatures. While an explicit relationship between x and y in P-WNxCy cannot be determined from XRD and AES data without some assumptions (e.g., use of Vegard's law), insight into the relation between C and Va on the interstitial sublattice is possible. Nitrogen prefers to occupy interstitial lattice sites in the polycrystals, and only resides at the grain boundary when its concentration in the film exceeds the P-WNo s stoichiometry. Low N levels in the 2b films imply that all N resides at interstitial sites. Essentially constant, low N levels (Figure 4) indicate that changes in lattice parameter will likely depend on the ratio of C to vacancies on the interstitial sublattice. Values for lattice parameter in the films from 2b tend to follow C content over the entire deposition temperature range. This likely reflects an increase in the ratio of C to vacancies on the interstitial sublattice with deposition temperature. The lattice parameter

PAGE 324

315 for the films from 2b is 4.10 A for 525c deposition, drops to 4.09 A for 550c, and then increases to 4.14 A for 600C. The C content follows the same trend (Figure 6-7), decreasing from 525 to 550C, and then increasing up to 600C. The lattice parameter increases only slightly from a value of 4.14 A at 600c to 4.15 A at 650c, consistent with a small increase in C levels as measured by AES. This behavior is consistent with a decrease in compressive film stress, coupled with grain growth. Above 650C, the lattice parameter again increases with C content, reaching 4.19 A at 750C. 6.5.3 Polycrystal Grain Size Grain size (t) was estimated using the Scherrer equation (Equation 6-1) [Cul78, Nag00]. The two main causes of peak broadening in XRD thin film spectra are non-uniform film stress and the presence of small crystal grains. If we assume uniform film stress, the average crystallite size (t) may be estimated using the Scherrer equation: 0.9). t=-(Bcos8) (6-1) where A is the X-ray wavelength (1.5406 A for Cu Ka), B is the full width at half maximum (FWHM) for the selected diffraction peak and 8 is the Bragg angle for that peak. The dominant (111) diffraction peak for the films was used as the reference peak for FWHM determination. As depicted in Figure 6-6, grain size for the films from 2b increased with deposition temperature, varying from 35 to 67 A over the 525 to 750C temperature range. Below 525C, the films were X-ray amorphous, hence the maximum grain size for amorphous films was below 35 A. Interestingly, the grain size is essentially constant between 550 and 600C, which coincides with the transition region

PAGE 325

316 between kinetic and diffusion controlled growth. As the deposition temperature increases, a competition likely exists between increased grain growth due to higher surface diffusivity and decreased grain growth due to increasing C concentration on the film surface, which inhibits surface diffusion. Interplay between these phenomena cause grain growth in regions with small shifts in C content and leveling off of grain size in regions with larger shifts in C content. --8 0 !J en 0.0 = < '-' Q.) N .... Cl) = a 1-, C) Q.) 0.0 1-, Q.) > < Figure 6-6. 6.6.1 AES 80 g b 70 60 50 40 30 ~-~----~-------------< 500 600 700 800 Deposition Temperature (C) Change in average grain size with deposition temperature for polycrystalline films grown from lb and 2b based on the FWHM of the BWN xCy(ll 1) diffraction peak. Error bars reflect uncertainty in FWHM measurements 6.6 Film Composition AES results for films deposited from the phenylimido complex 2b indicate the presence of W, N C and O (Figure 6-7). No chlorine was detected in the films by AES or XPS placing an upper limit of -1 at. % on Cl content. HCl is the thermodynamically

PAGE 326

317 favored gas phase chlorine-containing species, and was observed by residual gas analysis during deposition from lb. It is assumed that HCl is the dominant gas phase chlorine-containing species for deposition from 2b as well. Neither Ch nor chlorinated hydrocarbons were detected in the reactor effluent, leading to the conclusion that Cl is lost from the precursor as HCI. From 475 to 500C, the C level is constant at approximately 3 to 5 at. %. The C content jumps from 5 to 14 at. % between 500 and 525C. Although XRD indicates B-WNx polycrystalline deposition at 525C, results for growth rate and sheet resistance at this temperature show strong deviation from the trends evident between 550 to 750C. Given the proximity in deposition temperature of the C spike (525C) to the anomalous tungsten oxide formation seen in the XRD (500C), the two phenomena may be related. Above 550C, the C content rises steadily from 9 to 22 at. % at 750C. The increase in C content from lowest to highest deposition temperature reflects the increasing tendency of the hydrocarbon groups present in the precursor ligands and the solvent to deposit in the films at higher growth temperature. The initial N content of films grown at 475C was 1 at. %. The N content increased to a maximum value of 3 at. % at 525C, and then decreased with increasing temperature, dropping below 1 at. % above 700C. Although metal nitride barriers typically exhibit low N content at higher deposition temperatures (due to desorption of N2 gas), the films deposited from 2b were N-deficient throughout the temperature range studied. This N deficiency in the films contrasts with XRD results in Figure 3, which indicate B-WNx polycrystal growth at lower temperatures. This may indicate that B-WNxCy polycrystal formation begins at temperatures below 600C, with C filling the excess vacancies present in the polycrystals due to N deficiency.

PAGE 327

318 100 100 80 80 G]
PAGE 328

319 minutes of sputter, artificially high C and low N compositions may have been observed. In addition, the lack of a standard film sample for calibration of elemental concentrations means that AES data may vary up to several atom percent from the actual values. Despite the error bars on the concentrations, the AES data serve to identify trends in film composition with deposition temperature. The slight O contamination in the film samples deposited at higher temperatures likely resulted from post-growth exposure of the film samples to air. The higher 0 levels in films deposited at and below 500C may be influenced by the presence of the Et20 co-solvent during growth and exposure of the films to air after growth. Incremental AES sputtering showed a steady decrease in O levels with increasing depth into the WNx films. The O concentration was highest at 475C, reaching 15 at. %, and then decreased slightly to 11 at. % at 500C. This behavior is consistent with low density and high porosity in the amorphous films deposited below 525C, which allow substantial amounts of O to penetrate into the film lattice. High O concentrations (-20%) attributed to air exposure have been reported for porous TiN, TiC and TiCN barriers [Eiz94a, Par96, WanOla]. XPS results for O in the films are consistent with W03, which has considerably higher thermodynamic stability than ~-WNx or ~-WCx. For example, values of the Gibbs energy of formation (~G0r) at 750C for the W03, ~-WNo.s and ~-WCo.s phases are -579 kJ/mol, +21 kJ/mol, and -8.5 kJ/mol, respectively [Gus86, Lak79, Lid85]. The experimental observation of lower levels of O at higher deposition temperatures is consistent with post-growth O contamination, as oxide formation is thermodynamically favored to occur if an O leak into the reactor occurred during growth. As the deposition temperature rises from 500 to 525C, the O content drops sharply to 4

PAGE 329

320 at. %, while the C and N levels are moderately steady. This behavior is consistent with the change in crystallinity observed by XRD. As the film crystallizes, it becomes more dense [Pok91], thereby inhibiting post-growth O diffusion into the lattice and decreasing the density of adsorption sites. As deposition temperature increases above 525C, the 0 concentration drops further, falling below 1 at. % above 700C. This drop in O levels likely results from film densification (by polycrystal grain growth) and increased C levels at higher deposition temperature stuffing the grain boundaries. Porosity of amorphous films grown below 525C may be problematic for diffusion barrier applications, since defects in the film may degrade the barrier's resistance to Cu diffusion. A previous report, however, indicates that diffusion barrier performance depends more strongly on film microstructure than film density [Kim99c]. In addition, impurities such as 0, N and C have been reported to enhance the stability of diffusion barrier films [Cha99]. 6.6.2 Film Growth Rate (XSEM) Growth rates were estimated by dividing the total film thickness (from X-SEM or AES sputter profile for the sample grown at 425C) by deposition time. Figure 6-8 depicts X-SEM photos for films grown at the lowest and highest growth temperature, corresponding to deposition rates of 2 and 21 A/min An Arrhenius plot using the measured growth rates (Figure 6-9) clearly delineates two growth regimes. The region with the shallow slope above 550C exhibits a weak dependence on temperature that is indicative of mass transfer controlled growth. The region with a steep slope below 550C suggests that growth is controlled by a kinetic process. In this regime, the rate-determining step for film growth is presumably reaction on the substrate surface. As observed in the XRD and AES studies, the growth rates for films grown at 500 and

PAGE 330

321 525C appear to be anomalous. This may be due to a decrease in the film density causing the measured thickness change to overestimate the reaction rate. Excluding the data points at 500 and 525 C a fit of the data gives an apparent activation energy for growth of films from 2b in the kinetic regime as 1.41 0.58 eV. Figure 6-8 Cross-sectional SEM photo depicting thickness of films grown from 2b on a Si (100) substrate. a) 475C b) 750C. Deposition Temperature (C) 800 750 700 650 600 550 500 450 4 ! 3 .. I: .. I I I I',- 2 c., ..e 1 0 0 9 1.0 1.1 1.2 1.3 1.4 1000/T(K1 ) Figure 6-9. Plot of film growth rate (G, Afmin) from 2b on Si(lO0) vs. inverse temperature. Error bars indicate uncertainty due to deposition temperature vari a tion (+/-10 C) and thickness measurement from X-SEM photos. The line fit for the kinetic regime includes data points for 450 475 and 550C and excludes points at 500 and 525C.

PAGE 331

322 Depositions at 450 and 800C were done to establish the temperature boundaries for film growth from the phenylimido complex 2b. Films deposited at 450C were not visible by X-SEM, but AES depth profiling indicated a nominal film thickness of -100 A. This thickness value (triangular symbol in Figure 6-9) was used to estimate the value of the apparent activation energy for film growth in the kinetic regime. The growth rate at 800C dropped to 13 .A/min (square symbol in Figure 6-9), with a high degree of surface roughness visible on the substrates after deposition. These results may be explained by either premature gas phase decomposition of the precursor, due to upstream heating at the higher temperature or by high desorption rate for a reactant species. 6.7 Film Electrical Properties 6.7.1 Film Resistivity Film resistivity was calculated with Equation 5-5, as discussed in Chapter 5. Sheet resistance measurements were collected using 4-point probe, and film thickness measurements were taken from X-SEM images. The variation of film resistivity with deposition temperature is shown in Figure 6-10. A deposition at 475C produced films with the lowest resistivity value (225 .Q-cm) despite C and O contamination levels of 3 and 15 at.%, respectively. This is lower than the 620 .Q-cm reported for films grown from (1BuNH)2W(N1Bu)z at 650C [Chi93, Tsa96], and also lower than the 750 .Q-cm value observed for films grown from lb. In the temperature range where anomalous growth was observed, 500 to 525C, the resistivity sharply increases, e.g., 73,000 .Q-cm at 500C. This dramatic rise is consistent with deposition of tungsten oxide at 500C. Interestingly, the resistivity

PAGE 332

323 100000 ~---------------~ 880000 (.) ~60000 '--" 0 :~40000 ~ "' 0 ~20000 0 400 500 600 700 Deposition Temperature (0C) lb -2b 800 Figure 6-10. Variation of film res1st1v1ty with deposition temperature for films deposited with lb and 2b. climbs even higher to 87,000 .Q-cm for deposition at 525C, despite XRD data consistent with polycrystalline ~-WNx, Moving to 550C, the film resistivity drops to 4,750 .Q-cm, then fluctuates somewhat with increasing growth temperature. Changes in the film resistivity above 550C occur due to the interplay of polycrystal grain growth, C content and film thickness. Increased grain growth generally causes a resistivity decrease, while increased free C content and film thickness are associated with increased resistivity. 6.7.2 Film Sheet Resistance Film resistivities calculated with Equation 5-5 depend on both the sheet resistance and thickness of the analyzed films. As deposition temperature increases above 550C, the film thickness also increases. To decouple the impact of film thickness from the film's electrical properties, the sheet resistance was plotted as a function of deposition temperature (Figure 6-11). The sheet resistance increases sharply with deposition temperature from 75 Q/0 at 475 C to 1600 Q/D at 500 C, then rises further to

PAGE 333

324 2000 .Q/D at 525 C. This increase in sheet resistance for the less conductive films grown at 500 and 525C indicates that these higher resistivity values (Figure 6-10) are not due solely to higher film thickness and is consistent with oxide formation and the spike in C content determined by AES (Figure 6-7). As the deposition temperature is increased to 550 C the sheet resistance decreases to 279 .Q/D concomitant with lower O and C content. The sheet resistance, like the resistivity, fluctuates somewhat for deposit i on temperature above 550C, but eventually reaches a constant value near 210 .Q/D at 750c. 2500 ~----------------~ ---0-lb e 2b 400 500 600 7 00 800 Deposition Temperature (0C) Figure 6-11. Variation of film sheet resistance w i th deposition temperature for films deposited from lb and 2b. 6.8 Conclusions on Use of Cl4(CH3CN)WNPh to Deposit WNx In terms of their decomposition chemistry, the most significant diffe r ence between isopropylimido complex lb and phenylimido complex 2b is the dissociation energy of the N-C bond in the imido ligand. Based on data from organ i c model compounds, the N-C bond of isopropylimido complex lb is expected to be approximately 20 kcal/mol weaker than the analogous bond in 2b [ Ben76]. Since

PAGE 334

325 cleavage of this bond is necessary for deposition of WNx, one would expect there to be differences in film structure, film composition, and l attice parameter between films grown from the two precursors. Amorphous film growth occurs below 500C for lb and 2b. At 500C, a broad B-W2N (111) polycrystalline peak appears for films from lb, while polycrystalline oxide peaks appear for the material from 2b Evidence of polycrystalline WNx deposition from 2b first appears at 525C. The anomalous film characteristics of material grown from 2b at 500C appear to be linked to the presence of the Et 20 co-solvent (necessary because the solubilities of lb and 2b differ). The maximum deposition temperature for films deposited from the isopropyl complex was 700 C. Above this temperature, black particles were deposited on the substrate and susceptor, which subsequently compromised film quality. In contrast, deposition from phenylimido complex 2b w a s possible up to 750 C. The higher temperature limit for 2b cou l d be due to the enhanced N-C bond strength in its imido ligand. Carbon levels in the films from 2b were considerably lower than those from lb throughout the temperature range (Figure 67). From 550 to 750 C, films from both precursors exhibited a gradual rise in C content, with the isopropyl imido complex lb affording a steeper rise in C as growth temperature increased. Maximum N content for films from the two precursors occurred near their respective crystallization temper a tures (500C for lb and 525C for 2b). Throughout the temperature range studied, the N content in films deposited from 2b was significantly lower than those grown from lb. The fact that films from phenylimido complex 2b contained lower levels of both N and C than those from isopropylimido complex lb suggests that the phenylimido moiety i s

PAGE 335

326 more likely to dissociate intact than the isopropylimido ligand. This is consistent with the higher N-C bond strength in the phenylimido group. Corroborating evidence for this effect can be found in the mass spectra of la and 2a. Facile dissociation of the isopropyl group from la is indicated by the observation of [ChWNHt at 78% abundance in the EI spectrum and detection of [Cl4WNr as the base peak in the NCI trace. Loss of the phenyl moiety from 2a to yield the same ions does not occur under EI conditions and the NCI spectrum contains [Cl 4 WN]at only 4% abundance Moderate amounts of [Cl4 Wt and [Ch Wt were detected in the mass spectra of 2b under EI conditions, as well as fragments a-d (Figure 6-2), which are consistent with what has previously been observed upon generation of NPh + from phenyl azide [Cro67]. Although the mass spectral evidence for NPh loss consists of low intensity mass envelopes, their presence is significant. Because of the difference in conditions between mass spectrometry (ion chemistry) and MOCVD (thermal decomposition), the data in Figure 6-2 and Table 6-1 do not rule out loss of NPh as a major heterogeneous process during deposition Ideally, cleavage of the N-C bond in 2b to release a phenyl group should occur during CVD of WNx films. However, the high N-C dissociation energy would slow this process, allowing cleavage of the W-N bond to compete. Isopropylimido complex lb, with its weaker N-C bond, would be more likely to release the isopropyl moiety and leave the imido N in the growing film. For both precursors, 0 levels were highest for films deposited at the lower end of the temperature range (~ 500C). AES indicated a decrease in O content with increasing sputter depth into the films; hence, high O levels were attributed to post-growth exposure of the films to air. The low density and high porosity of the amorphous structures grown

PAGE 336

327 at lower temperatures allow more rapid diffusion of O from the air into the films and provide a large surface to volume ratio for adsorption. High O levels in the films from 2b, however, may be due to a combination of post-growth exposure to air and the presence of Et2O during deposition. Increased C content in the films grown at higher temperatures is believed to inhibit post-growth oxidation of the films by stuffing grain boundaries. Lattice parameter shifts were also dependent on the precursor. Figure 6-5 shows the lattice parameter increasing with temperature for films from both lb and 2b at temperatures up to 650C, with films from lb having higher lattice parameter values throughout this range. This is expected, due to higher C content in films from the isopropylimido complex lb. Just above 675C, though, the lattice parameter in the films from 2b becomes larger than that of the films from lb. Polycrystalline films deposited by lb have adequate N, even at high temperatures, to be considered primarily P-WNx into which C is incorporated to form P-WNxCy. This P-WNxCy, although N-deficient at higher temperatures contains sufficient N to prevent a shift from formation of P-WNxCy toward P-WCy, and leads to an upper limit in lattice parameter for films from lb near 4.16 A Thus, the decrease in lattice parameter in films from lb deposited above 650C was attributed to a solubility limit for C in P-WNxCy polycrystals with excess C residing at the grain boundaries above 675C. The films from 2b, however, contain so little N that W is effectively left "bare" during deposition, enabling formation of carbide by reaction with depositing C at high temperature. Minimal N coupled with -20 at. % C above 675C may mean that a shift away from deposition of P-WNxCy toward P-WCy is occurring at high temperature with

PAGE 337

328 the 2b precursor. The lattice parameter of 4.19 A for films grown from 2b at 750C indicates that the composition of the B-WNxCy polycrystals is nearing B-WCo 6 (i.e., xis approaching zero and y is approaching 0.6), which has a lattice parameter of 4.236 A. Although B-WNx films are reported to outperform B-WCy for diffusion barrier applications, recent reports indicate interest in B-WNxCy as a potential barrier material. Thin films (-250 A) of B-WNxCy have been deposited by ALO onto various substrates (e.g., SiO2 Cu, SiC and SiLK) by sequential reaction of WF6, NH3 and BEt3 [Li02, Smi02]. Addition of C to B-WNx was reported to lower film resistivity. In addition, B-WNxCy with a composition ratio W:N:C 55:15:30 exhibited excellent adhesion to Cu. Hence, although films deposited from 2b have low N content ( compared to typical WNx films), this precursor may be of interest for B-WNxCy deposition. While we have demonstrated the ability of an NH 3 co-reactant to dramatically increase the N content for films deposited from lb, as discussed in Chapter 5, a similar approach with 2b did not result in increased N content in the deposited film. Nitrogen levels in the films from 2b are much lower than films from lb at all temperatures. This trend in the films grown from the phenylimido precursor can again be attributed to the high N-C bond dissociation energy in the imido ligand of 2b, which causes less N incorporation into the films, leaving open the possibility of B-WCy formation. The XPS binding energy values for Cls and W4f712 energy levels for films deposited from 2b at 650C were 284.6 and 33.0 eV, respectively, while those for deposition at 750C, were 283.6 and 32.1 eV. These data are consistent with a shift toward formation of the B-WCy phase for the film deposited from 2b at 750 C [Kou02]. The XPS results for films from lb did not indicate this shift. This suggests that adequate

PAGE 338

329 N remained in the polycrystals to prevent a shift to carbide formation, forcing additional C to deposit at the grain boundaries. Grain growth behavior was similar for the two precursors. Grain sizes for films grown from lb and 2b increased from lowest to highest deposition temperature, with those from 2b generally having smaller size grains than those from lb. Above 600C, where both precursors are operating in the diffusion controlled regime, the films from 2b required a -50C higher temperature to deposit grains of similar size to their lb counterparts. This observation can again be traced to the difference in the strengths of imido N-C bonds in the precursors, with the weaker isopropyl imido bond affording higher growth rate and grain size for a given deposition temperature. Deposition rates from the phenylimido complex 2b films were lower than those from the isopropylimido complex lb. Film deposition rates ranged from 10 to 27 A/min for films from lb, while the range was 2 to 21 A/min for films from 2b. The calculated Ea for film growth from 2b in the kinetically controlled regime was 1.41 0.58 eV, as compared to 0.84 eV for lb and 0.9 eV for (1BuNH)2W(N1Buh [Chi93, Tsa96]. The higher activation energy for 2b is consistent with cleavage of the stronger imido N-C bond before or during the rate-determining step of the deposition process. Interestingly, the Ea for 2b is above the typical activation energy range for CVD growth in the kinetic regime, which is 0.5 to 1 eV [Raa93]. With the exception of the anomalous results for growth at 500 and 525C, comparison of the sheet resistances, which negate the impact of film thickness on electrical properties, shows that the films from lb and 2b have similar electrical properties when deposited at or below 675C. Above 675C, lb films have higher sheet

PAGE 339

330 resistance than 2b films, consistent with the high C levels in the lb films, which scatter electrons flowing through the material. The tungsten imido complex Cl4(CH3CN)W(NPh) 2b was tested to determine its suitability as a single-source precursor for low temperature growth of ~-WNx thin films. Comparison of the film growth properties of 2b to those of its isopropylimido analogue lb allowed evaluation of the effect of the imido N-C bond dissociation energy on film growth and properties. Films deposited from 2b were deficient in N compared to those from lb, consistent with a tendency of the stronger imido N-C bond of 2b to result in dissociation of intact NPh fragments during deposition. Since its films contain more N and have lower amorphous deposition temperatures and sheet resistances, the isopropyl imido precursor lb is superior to 2b for barrier deposition.

PAGE 340

CHAPTER 7 EVALUATION OF Cl4(CH3CN)WNCH2CH=CH2 AS A SUIT ABLE WNx PRECURSOR MOCVD of thin WNxCy films from benzonitrile solutions of the single-source imido precursors Cl4(CH 3 CN)W(NiPr) (la, as a mixture with its benzonitrile derivative lb) and C4(PhCN)W(NPh) (2b) were discussed in Chapters 5 and 6, respectively. Based on mass spectral data of the precursors and film properties from material grown with the isopropyl imido and phenyl imido complexes, the isopropyl imido precursor la,b enables increased N incorporation relative to the phenyl imido precursor 2b, presumably due to the weaker N-C imido bond of the former (Chapter 5 and 6 have more detail). By this reasoning, a more promising precursor should contain a weaker N-C imido bond than that in either of the previously discussed complexes. Deposition tests were therefore undertaken with the allyl imido precursor Cl4(RCN)W(NCH 2 CH=CH 2 ) (3a, R=CH 3 and 3b, R=Ph). 7.1 Film Growth Studies Experiments were conducted in the CVD reactor system described in Chapter 4. P-type boron doped Si (100) substrates with resistivity of 1-2 .Q-cm were used for the film growths. Growths were conducted for a fixed time of 150 minutes at temperatures ranging from 425 to 675 C. The system was maintained at vacuum by a mechanical roughing pump, with the operating pressure fixed at 350 Torr Hydrogen (H2 ) carrier gas was used for the depositions. 331

PAGE 341

332 All mass spectral analyses were again performed using a Finnigan MAT95Q hybrid sector mass spectrometer (Thermo Finnigan, San Jose, CA). Electron ionization (El) was carried out in positive ion mode using electrons of 70 eV potential and a source temperature of 200C. Negative ion electron capture chemical ionization (NCI) used methane as the bath gas at an indicate pressure of 2x10-5 Torr, an electron energy of 100 volts and a source temperature of 120 C. All samples were introduced via a controlled temperature probe with heating and cooling enabling temperature control down to 35 C. The mass resolving power (m/Am) was 5000 full width-half maximum (FWHM). 7.2 Synthesis of Allyl Clt(CH3CN)WNCH2CH=CH2 Precursor Standard Schlenk and glovebox techniques were employed in the synthesis of Cl4(CH3CN)W(NCH2CH=CH 2 ) (3a). Tungsten oxychloride was prepared by a slightly modified literature method [Ped82]. Complex 3a was prepared in an analogous fashion to other previously reported tungsten imido precursors [Bch03a]. The benzonitrile complex Cl4(PhCN)W(NCH2CH=CH2) (3b) was not isolated, but was produced in situ by the substitution of the acetonitrile ligand of 3a with benzonitrile, which was utilized as the solvent for the deposition experiments. Tungsten(VI) chloride was purchased from Strem Chemical Company. Allyl isocyanate and anhydrous heptane were used as purchased from Aldrich Chemical Company. Toluene and hexane were purged with N 2 and dried by passing the degassed solvent through a column packed with activated AhO3 [Pan96]. Robertson Microlit in Madison, NJ performed the elemental analyses. To prepare Cl4(CH 3 CN)W(NCH 2 CH=CH 2 ) (3a), tungsten oxychloride (2.047 g, 5.99 mmol) was slurried in a solution of allyl isocyanate (0.513 g, 6.17 mmol) in heptane (60 mL) in a sealed pressure vessel. The mixture was heated for 18 h at 110 C. The

PAGE 342

333 solvent was removed from the resulting dark red solution in vacuo. The reddish brown residue was dissolved in a minimal amount of CH3CN (approximately 10 mL). The resulting solution was stirred for two hours and the solvent removed under reduced pressure. The resulting brown residue was washed with 5x10 mL of toluene and the extracts concentrated to approximately 5 mL. Hexane was added to precipitate the product. The orange-brown solid was filtered and washed with pentane to afford 1.59 g (63 % yield) of the imido complex. 1H NMR (CDCh, 8): 7.57 (d, J = 5.6 Hz, 2H, NCH2CHCH2); 6.07 (dddd, J = 5.6, 10.2, 16.9 Hz, lH, NCH2CHCH2); 5.73 (d, J = 16.9, lH, NCH2CHCHH); 5.60 (d, J = 10.2 Hz, lH, NCH2CHCHH); 2.50 (s, 3H, CH3 CN). 13C NMR (CDCh, 8): 129.8, 121.9 (CT!iCHCH2N); 118.9 (CH3CN); 68.3 (CH2CHCH2N); 3.5 (CH3CN). Anal. Calculated for WCsHsN2Ck C, 14.24%, H, 1.91 %, N, 6.64%. Found: C, 14.51 %, H, 1.86%, N, 6.43%. 7.3 Solvent Selection For the isopropyl imido system, lb was generated in situ following solvation of la in the benzonitrile used as the solvent for the CVD process. In the case of the phenyl imido compound, the solid precursor employed was the ether adduct Cl4(OEti)W(NPh) (4); however, the limited solubility of this (and other phenyl imido precursors) in benzonitrile necessitated the use of a 10: 1 benzonitrile/ether co-solvent mixture. Accordingly, compound 4 was converted to the benzonitrile adduct Cl4(PhCN)W(NPh) (2b) in situ prior to deposition. Much like the isopropyl imido precursor la, the allyl complex 3a is soluble in PhCN allowing this solvent to be used for deposition experiments

PAGE 343

334 7.4 Precursor Mass Spectral Pre-Screen We have previously shown that mass spectrometry of the tungsten imido precursors Cl4(CH3CN)W(NiPr) (la) and Cl4(CH 3 CN)W(NPh) (2a) affords insights into possible fragmentation patterns which arise during the CVD process [Bch03a, Bch03c]. Nevertheless, since mass spectrometry is an ionic process, care must be taken in applying the fragmentation patterns observed to a thermal process such as CVD [Lew94]. As in previous studies, we subjected the allyl imido precursor 3a to both positive ion electron-impact (El) and negative ion electron-capture chemical ionization (NCI) processes. Table 7-1 summarizes the major fragments observed in the EI and NCI spectra of precursor 3a. As with the isopropyl and phenyl imido complexes la and 2a, neither method detected a molecular ion. Instead the highest mass envelopes in the EI and NCI spectra occurred at m/z=346 and 381, corresponding to [ChW(NC3Hs)t and [C4W(NC3Hs)r respectively. The [ChW(NC3Hs)t fragment was also the base peak of the EI spectrum. A high abundance (95%) peak at m/z=41 corresponds to both the acetonitrile fragment [CH3CNt and the allyl fragment [C 3 H5t from the imido moiety. Importantly, the base peak of the NCI spectrum corresponds to the mass envelope of the nitride fragment [Cl4 WNr (m/z=340). A small amount of the protonated nitrido complex fragment [Ch WNHt was detected in the EI spectrum, but at low relative abundance (-12%). As observed with precursors la and 2a, the EI spectrum of the allyl imido complex 3a also exhibited fragments corresponding to loss of the imido N. Accordingly, mass envelopes at mlz=256 (27% abundance) and 291 (58% abundance) are assigned to the fragments [Ch Wt and [Ch Wt respectively.

PAGE 344

335 Cleavage of the W-N multiple bond to afford the tungsten chloride fragments [Clz wt and [Ch wt seems problematic in regards to CVD of tungsten nitride; however, the high-energy nature of EI mass spectrometry most likely accounts for this observation. The lack of any molecular ion in either mass spectrum is consistent with !ability of the nitrile ligand. More importantly, the observation of the nitride fragments [Cb WNH]+ and [Cl 4 WN]indicates that the critical imido N-C bond is broken under the ionization conditions. Relative to the isopropyl and phenyl imido precursors la and 2a, the allyl complex 3a shows some similarities and important differences. In each case, the base peak of the EI spectrum corresponds to the loss of CH3CN and one chloride ligand (i.e., [CbW(NR)t). The abundance of the protonated nitrido fragment [ChWNH]+ in the EI spectrum of the allyl imido complex 3a was only about 12% as compared to the 78% relative abundance of the same mass envelope in the spectrum of the isopropyl imido precursor la. Strikingly, this fragment is not observed at all in the EI spectrum of the phenyl precursor. Additionally, the fragment corresponding to the loss of the nitrile ligand (i.e., [Cl4 W(NR)r) was observed in the NCI spectra of all three precursors The relative abundance of this fragment is consistent with the relative strength of the N-C imido bond for the three precursors. Interestingly, this fragment was the base peak for the NCI spectrum of the phenyl precursor 2a, while its relative abundance was lowest for 3a (5%) and in-between (42%) for la. In contrast, the nitrido fragment [C4WNr was the base peak in the NCI spectra of the isopropyl and allyl complexes la and 3a, while this mass envelope only accounted for 4% relative abundance in the spectrum of 2a

PAGE 345

336 The greater abundance of the nitrido fragment [Cl 4 WN]in the NCI spectra of la and 3a relative to 2a indicates that the N-C bond of the imido ligand is more easily broken for the ally! and isopropyl imido precursors than for their phenyl imido analogue. This conclusion is supported by the absence of [Ch WNHt in the EI spectrum of 2a and its presence in the spectra of la and 3a. These data correlate well with the homolytic C-N bond dissociation energies reported for the corresponding amines (C 3 H5NH2=73 kcal/mo!, iPrNH 2 =84 kcal/mo!, and PhNH 2 =105 kcal/mo!) [Ben76, Luo94]. The smaller relative abundance of [Ch WNHt observed for ally! imido complex 3a versus the isopropyl imido complex la cannot be easily explained within the context of C-N bond strengths. Table 7-1. Summary of relative abundances for positive ion EI and negative ion NCI mass spectra of tungsten imido complexes Cl4(CH3CN)W(NCH2CH=CH2) (3a) EI Fragments NCI Fragments m/z Abundancea [Cl3W(NC3Hs)]+ 346 100 [Cl4Wt 326 34 [ChWNH]+ 306 12 [ChWt 291 58 [CliWt 256 27 [CH3CNt or [C3Hst 41 95 [Cl4W(NC3Hs)] 381 5 [Cl4WNr 340 100 aRelative abundances were adjusted by summing the observed intensities for the predicted peaks of each mass envelope and normalizing the largest sum to 100% 7.5 Film Structure 7.5.1 XRD The X-ray diffraction (XRD) spectra in Figure 7-1 indicate amorphous and polycrystalline film deposition at 450 and 650 C, respectively, from the 3b complex. The polycrystalline film has peak locations consistent with polycrystalline~ -WNxCy.

PAGE 346

400 --en ..... .... i:::::: :::> 300 ti .... 200 '-' ..... 100 .... en i:::::: ..... i:::::: -0 30 40 337 Si (400) 50 60 2 8 Degrees a 70 80 400-,---..--------------.----.---r---~ 30 Si (200) ~-WNxCy _.-(111) Si (400) k~ 40 ~-WNxCy .,-(200) 50 60 2 8 Degrees b 70 80 100~--~~-------------~ -en ..... .... ;5 80 i 60 .... -._, 40 0 .... en 20 B -WNo. (111) B -WC0.6 (111) I I I I I B-WCo. 6 I.(200) I I I I B WN 0 5 : (200) B-WN0. 5 (220\ I I I I B-WCo. 6 I (220)----..i C B -WN0. 5 (311) B-WCo. 6 (311) I \t: I I I i 0 +------'-~____......___--.-___ .,............__.__~ _.____.___---i 30 40 50 60 70 80 2 8 Degrees Figure 7-1. XRD spectra for films grown with 3b on Si (100) in a H2 atmosphere. a) 450 C. b) 650 C. c) Standard powder diffraction plots for ~-WNo.s and ~-WCo .6-

PAGE 347

338 Four characteristic peaks are evident, indicating that no preferred crystal orientation was present in the films. Primary reflections at 37.18 and 43.33 28 degrees are consistent with (111) and (200) -WNxCy growth planes, while additional reflections at 62.88 and 74.88 28 degrees indicate (220) and (311) planes respectively. ~-WNxCy ~-WNxCy Si (400) 5000 (111) ;( (200) T= 650 C Ka ,_ T= 625 C Cfl .... ..... T= 600 C = 4000 T= 575 C tJ 3000 ..... T= 550 C -e '-" T= 525 C ;:>-. .... 2000 ..... Cfl Si (200) Ka = T= 500 C .... = 1000 >---( T= 475 C T=450 C 0 30 40 50 60 70 80 20 Degrees Figure 7-2. Change in XRD pattern with deposition temperature for films grown with 3b on Si (100) in a H2 atmosphere. Figure 7-2 shows the evolution of film crystallinity with deposition temperature For deposition at and below 525 C, the characteristic -WNx peaks are not observed. At 550C, a broad peak appears near 37.63 28 degrees, indicating polycrystalline -WNxCy (111) growth. As the deposition temperature increases to 575C, this peak sharpens and a broad peak at 44 08 28 degrees appears, indicating -WNxC y (200) growth. The peaks sharpen further as the temperature approaches 650 C indicating polycrystalline grain growth. Some of the films displayed two additional peaks at 32.98

PAGE 348

339 and 61.63 28 degrees, representing Si (200) Ka and Si (400) KB radiation, respectively. Broad peaks emerge at 63.33 and 75.43 28 degrees for growth at 650C, indicating B -WNxCy (220) and (311) growth. Since the formation of polycrystalline films is highly undesirable, deposition below 550C with this precursor is required to grow an amorphous film. 7 .5.2 Lattice Parameter The dependence of lattice parameter on deposition temperature in Figure 7-3 was determined by XRD using the 28 position of the (111) B-WNxCy diffraction peak, with peak position calibrated to the Si (400) diffraction peak. The standard B-WNo.s (111) and B-WC0 6 (111) peak positions are 37.735 and 36.977 28 degrees (Figure 7-lc), respectively, and correspond to standard lattice parameter values of 4.126 and 4.236 A. The position of the (111) reflection peak can vary as a result of a change in composition or a change in the film's residual stress. Assuming that compositional variation is responsible for the peak shift and, thus, lattice parameter change, trends in the relative concentrations of N, C, and Va on the interstitial sublattice can be suggested by coupling the peak shift with the compositions measured by AES (Figure 7-5), as discussed in Chapter 6. If the (111) peak position is higher than 37.735 28 degrees, corresponding to a lattice parameter below 4.126 A, then minimal C, a deficiency of N, and an excess number of Va exist in the polycrystals. This is true for deposition with lb at 500C. A peak position between 37.735 and 36.977 28 degrees, which corresponds to a lattice parameter between 4.126 and 4.236 A, suggests mixing of N, C, and Va on the interstitial sublattice. This is the case for all polycrystalline films deposited from 3b, and for films deposited at and above 550C with lb. As mentioned in Chapter 6, however, B-WNxCy

PAGE 349

340 polycrystals may be forming even at the lower temperatures, with C playing a role in the lattice parameter increase at lower temperatures. :::: ....----0---,--.-~-b--,--.--,--.---,--.-_-,_-, -7---.---,--.--.---.--.---,--,--.-_-,_-, S _._ 2b 4.20 -o-3b J3-WCo. 6
PAGE 350

341 650C to a maximum of 4.17 A. This lattice parameter value coincides with the maximum lattice parameter for films deposited from lb, which also occurs at 650C. 7.5.3 Polycrystal Grain Size Grain size (t) was estimated using the Scherrer equation, as discussed in Chapter 6. The dominant (111) diffraction peak for the films was used as the reference peak for FWHM determination. As depicted in Figure 7-4, grain size for the films from 3b increased with deposition temperature, varying from 34 to 48 A over the 550 to 650C 80 -,---------------------, j 70 Vl 00 i:: 60 < ..._, (!) N en 50 = ..... CJ 40 30 > < -0-lb ______._ 2b -<>3b 20 ~---,-----~------,--------i 500 600 700 800 Deposition Temperature (C) Figure 7-4. Change in average grain size with deposition temperature for polycrystalline films grown from lb, 2b and 3b based on the FWHM of the ~-WNxCy(111) diffraction peak. Error bars reflect uncertainty in FWHM measurements. temperature range. Below 550C, the films were X-ray amorphous, hence the maximum grain size for these films was below 34 A. The grain size increases most rapidly between 550 and 575C, with grain growth slowing above this temperature. As mentioned in Chapter 5, increasing deposition temperature causes a competition between increased grain growth due to higher surface diffusivity and decreased grain growth due to

PAGE 351

342 increasing C concentration on the film surface, which inhibits surface diffusion. Interplay between these phenomena cause grain growth in regions with small shifts in carbon content and leveling off of grain size in regions with larger shifts in carbon content. 7.6 Film Composition Auger results in Figure 7-5 indicate that W, N, C and O were present in the films deposited from 3b, while Cl was not detected. The lowest C level, 6 at. %, occurs at the lowest deposition temperature, 450C. Carbon content increases with deposition temperature from 6 to 38 at. % between 450 and 600C. Above 600C, C content levels off near 40 at. %. As in the case for the lb and 2b precursors, the overall trend in C content for 3b reflects the increasing tendency of the hydrocarbon groups present in the precursor ligands and the solvent to decompose with increasing deposition temperature. The N content in films grown at 450C was 4 at.%, and this rose to a maximum of 11 at. % for deposition at 500C. Above 500C, the N levels decrease, dipping to 2 at. % at 650C. The higher N levels at lower temperature reflect the stability of the W-N multiple bond in the precursor molecule, which likely endures at deposition temperatures up to 500C, inhibiting release of N into the gas phase during deposition. The drop in N above 500C may indicate decomposition of the W-N multiple bond in the gas phase and/or increased N desorption from the film (to form N2 gas) at higher temperature. Again, 0 contamination resulted from post-growth exposure of the film samples to air. Incremental AES sputtering showed a steady decrease in O levels with increasing depth into the films. The O concentration was highest at 450C, reaching 16 at. %, and decreased slightly to 11 at. % at 525C. Amorphous films deposited below 550C had low density and high porosity, which allowed substantial amounts of oxygen to penetrate

PAGE 352

343 into the film lattice. As the deposition temperature was increased to 550C, the O content dropped sharply to 4 at. %. This was consistent with crystallization of the film in this 100 100 w N ii 80 80 -lb C 20 u 0 u -+2b 3b 0 0 400 500 600 700 800 400 500 600 700 800 Deposition Temperatur e (0C) Deposition Temperature (0C) 100 100 C 0 80 ~lb 80 lb -+2b -<>3b 5 5 C C 0 40 0 40 g a C C 8 8 C C 0 20 0 20 u u 0 0 400 500 600 700 800 400 500 600 700 800 Deposition Temperature ( C) Deposition Temperature (0C) Figure 7-5 Comparison of W, N, C and O content in the films grown from lb, 2b and 3b. Data are from AES measurements after 2.0 minutes sputter. temperature range. As the film crystallized, its microstructure densified, thereby inhibiting post-growth oxygen diffusion into the lattice [Pok91, Jos02]. As deposition temperature increased above 550C, the O concentration dropped further, steadying out near 3 at. % at the highest deposition temperatures. This resulted from further film densification (by polycrystal grain growth) and increased C levels at higher deposition

PAGE 353

344 temperature, which stuffed the grain boundaries and blocked O in-diffusion. Again, porosity of amorphous films may be problematic, as defects in the film can degrade the barrier's resistance to Cu diffusion. 7. 7 Film Growth Rate (XSEM) Growth rates were estimated by dividing total film thickness (from X-SEM) by deposition time Figure 7-6 depicts X-SEM photos for films grown at the lowest and highest growth temperatures 450 and 650C, respectively, which have corresponding deposition rates of 5 and 10 A/min. An Arrhenius plot using the measured growth rates (Figure 77) indicates one growth regime, evidenced by a single line without any break points. The apparent activation energy (Ea) for deposition with 3b was 0.15 eV, which is substantially lower than the typical activation energy range for CVD growth in the kinetic regime (0.5 to 1.0 eV) [Raa93]. Figure 7-6. Cross-sectional SEM photo depicting thickness of films grown from 3b on a Si (100) substrate a) 450C. b) 650C Deposition with 3b at temperatures below 450 C did not result in significant film growth. This suggests that while cleavage of the N-C bond in the imido moiety may be the rate-determining step for deposition from this family of precursors, another gas-phase or surface process dictates the minimum temperature required to initiate

PAGE 354

345 deposition. Deposition at 675C resulted in deposition of black particles on the substrate and susceptor surface, indicating that gas-phase nucleation was occurring. For this reason, 650C was deemed to be the upper temperature limit for deposition from this precursor. Deposition Temperature (C) 650 600 550 500 450 3.0 2.5 2.0 d 1.5 i:: -1.0 0.5 0 0 1.05 1.10 1.15 1.20 1.25 1.30 1.35 1 ooorr (K-1 ) Figure 7-7. Plot of film growth rate on Si (100) vs. inverse temperature for deposition from 3b. Error bars indicate uncertainty due to deposition temperature variation(+/10C) and thickness measurement from XSEM photos. 7 .8 Film Electrical Properties 7.8.1 Film Resistivity Film resistivity was calculated with Equation 5-5, as discussed in Chapter 5. Sheet resistance measurements were collected using 4-point probe, and film thickness measurements were taken from X-SEM images. The variation of film resistivity with deposition temperature is shown in Figure 7-8. Deposition at 450C produced films with the lowest resistivity value (287 Q-cm) despite C and O contamination levels of 6 and

PAGE 355

346 16 at.%, respectively. This is slightly higher than the 225 Q-cm value for a film deposited at 475C with 2b, and lower than the 750 Q-cm value observed for films grown at 450C from lb. The lower resistivity for films deposited at 450 C from 3b relative to lb was likely due to decreased N content in films from 3b. 100000 -.--------------------, ,_ 80000 s (.) 60000 ..._, c :~ 40000 ....
PAGE 356

347 increases with deposition temperature from 36 Q/0 at 450 C to 172 Q/0 at 475 C, and is fairly steady to 600C. Above 600C, the sheet resistance increases again, reaching a maximum value of 274 Q/0 at 650 C. 2500 --0lb ,--... 2000 _._2b 3b ::, c:r
PAGE 357

348 properties and deposition behavior for the three precursors. Table 7-2 summarizes some of the main differences in deposition behavior for the three precursors. T bl 7 2 C a e -. fD ompansono epos1t1on e av1or or t e ee ecursors Bh. f hThr Pr Deposition Deposition Approx. Film Precursor Temp. Rate Crystallization Ea Range (QC) (Almin) Temperature (eV) (QC) C4(PhCN)W(NCH2CH=CH2), 450-650 5-10 550 0.15 3b Cl4(PhCN)W(N1Pr), lb 450-700 10-27 500 0.84 C4(PhCN)W(NPh), 2b 475-750 2-21 525 1.41 The Ea for film deposition varied significantly for the three precursors, following a trend consistent with the strength of the N-C bond in the imido moiety of the precursor. Film deposition from 2b, with the strongest N-C bond, yielded the highest value for Ea (1.41 eV), while that from lb yielded an intermediate value (0.84 eV). Deposition from 3b, which presumably has the weakest N-C bond, yielded the lowest Ea (0.15 eV), which is well below the typical activation energy range for CVD growth in the kinetic regime (0.5 to 1 eV) [Raa93]. A plot of the Ea values for deposition with the three precursors against the N-C imido bond strengths for the analogous amines is linear (Figure 7-10), with a goodness of fit (R2 ) of 0.96. The linear relationship suggests that cleavage of the N-C imido bond is the rate-determining step in film growth from the lb, 2b and 3b complexes. The strength of the bond in 3b is so low, though, that film growth borders on being mass-transfer controlled. While the typical film growth temperature dependence in the mass-transfer controlled region is -T1.7-1.s (Chapter 5), the temperature dependence for film growth from 3b was slightly higher (-T2 1), indicating that the rate-determining step for film growth from 3b has a very weak kinetic barrier. Changing the imido ligand from the isopropyl to the ally! complex may lower the N-C imido bond energy to where

PAGE 358

349 loss of the imido alkyl is no longer the rate-determining step at the surface. Weakening this bond any further would likely eliminate the kinetic barrier posed by this bond, causing a shift to mass-transfer controlled growth over the entire deposition temperature range. lmido moieties with higher N-C bond energies, such as those in lb and 2b, present a substantial kinetic constraint on film growth at lower temperature. 1.6 1.4 ,-._ > 1.2 Q.) '-' ..c: .... 1.0 0 C, 0.8 s -.... ll,. 0.6 1-. .8 .. 0 4 0.2 Allyl-NH2 0.0 2.5 3.0 3.5 4.0 4.5 5.0 Corresponding Amine Bond Dissociation Energy (eV) Figure 7-10. Variation of apparent film growth activation energy (Ea) with predicted N-C imido bond energies for the lb (i-Pr), 2b (NPh) and 3b (Ally!) precursors. Overall, deposition with 3b resulted in a substantially lower maximum growth rate than for deposition with lb or 2b. Deposition rates ranged from 10 to 27 A/min for films from lb, while the range was 2 to 21 A/min for films from 2b and 5 to 10 A/min for films from 3b. At 450C, the deposition rate with 3b (5 A/min) was half of the value for deposition with lb (10 A/min). The shelf life of the 3b precursor was poor relative to lb and 2b, as evidenced by a discoloration of the precursor/solvent mixture after sitting in the syringe for more than 1 to 2 days. This color change was accompanied by a negligible growth rate and/or a shift

PAGE 359

350 in film composition away from the general trend. The decrease in deposition rate for films from 3b may be due to a lower concentration of precursor available, due to decreased stability of the 3b precursor. Alternatively, cleavage of the N-C imido bond may occur to a substantial degree in the gas phase for 3b, in contrast to lb and 2b, leading to oligomerization or some other gas phase process that consumes the remaining W-based moiety. The small kinetic barrier posed by the N-C imido bond in 3b, whose deposition behavior borders on being mass-transfer controlled means that adequate energy may be available in the gas phase to cleave this bond prior to the precursor's arrival at the deposition surface Film crystallization occurs for growth at 550C with 3b compared to 500 and 525C for lb and 2b, respectively. This may be due to higher C levels in the films from 3b relative to those from lb and 2b. Higher C content at a given deposition temperature decreases surface diffusivity, thereby suppressing crystallization in the film. The maximum deposition temperature for films deposited from 3b was 650 C, as compared to 700 and 750 C for lb and 2b, respectively. For all three precursors, deposition above the respective maximum growth temperature resulted in formation of black particles on the substrate and susceptor, which subsequently compromised film quality. The magnitude of the maximum deposition temperatures for the three precursors reflects the stability of the N-C bond in the imido group for each precursor The precursor containing the weakest bond, 3b has the lowest maximum deposition temperature, while the precursor with the strongest bond 2b, has the highest, with the lb precursor falling in between. As mentioned above, the weaker N-C imido bond in 3b did not lead to a decrease in the minimum deposition temperature relative to growth with lb.

PAGE 360

351 The strength of the N-C imido bond has a strong effect on the amount of N incorporated into the film. If the N-C imido bond strength is relatively high, as with 2b, the imido-N group stays intact, causing cleavage of the W-N bond, which leaves the films deficient in N If the bond strength is relatively low, as with lb and 3b, the alkyl group cleaves from the N more easily, leading to substantial N in the film. The amount of N incorporated into films from lb and 3b is similar at all temperatures except the lowest one ( 450C). The value for N content at 450 C is lower for 3b ( 4 at.%) than for lb (8 at.%). The reason for this is unclear, as one might expect the weaker N-C bond in the allylimido ligand to enhance N incorporation by decreasing the likelihood for the entire imido ligand to desorb intact from the reaction surface. If cleavage of the imido N-C bond occurs in the gas phase, however, some of the resulting nitrido complexes may be consumed in other processes before the W-containing moiety reaches the deposition surface. EI mass spectral data indicate 78% and 12% relative abundance for the [ChWNHt fragment from la (Chapter 5, Table 5-1) and 3a (Table 7-1), respectively, while this fragment did not appear for 2a (Chapter 6, Table 6-1). A lower relative abundance for [ChWNHJ+. coupled with higher relative abundances for the [Cl4Wt and [Ch Wt fragments and the presence of a [Ch Wt fragment for 3a (Table 7-1), are further evidence that loss of imido N is more likely for 3a than for la. This would lead to lower N levels in films from 3b relative to lb. These results suggest that an optimal window for N-C imido bond energy exists in these precursors, with high and low bond energy values detrimentally affecting N content. Maximum N content for films from lb and 3b occurred near 500C, while that for 2b occurred near 525 C. Throughout the temperature range studied, the N content in

PAGE 361

352 films deposited from 2b was significantly lower than those grown from lb and 3b. The fact that films from 2b contained lower levels of both N and C than those from lb or 3b suggests that the phenylimido moiety is more likely to dissociate intact than the isopropylimido or allylimido ones. This is consistent with the higher N-C bond strength in the phenylimido group relative to the others. Similar N levels for growth with lb and 3b at and above 475C suggest that N derived from PhCN decomposition may supplant the precursor as the principal N source for the films deposited above 450C. While deposition from lb in 1,2-DCB rather than PhCN solvent resulted in reduced N levels, N was still deposited in those films, indicating that the imido N was retained in the films to some degree. Testing the 3b precursor with a N-free solvent, such as 1,2-DCB, would be useful to determine if the ally! imido ligand causes a change in precursor-derived N content compared to films deposited from lb in 1,2-DCB. Carbon levels in the films from 3b were slightly higher than those from lb for most deposition temperatures and significantly higher than those from 2b (Figure 7-5) Above 500C, films from the three precursors typically exhibited a gradual rise in C content, with lb and 3b affording a steeper rise in C with increasing growth temperature compared to 2b. The fact that films from 2b contained lower levels of both N and C than those from lb or 3b suggests that the phenylimido moiety is more likely to dissociate intact than the isopropylimido or allylimido ones, consistent with the higher N-C bond strength in the phenylimido group. Higher C levels for 3b compared to lb were probably caused by the decreased stability of the allylimido ligand relative to the isopropylimido one

PAGE 362

353 Oxygen content for films deposited with all three precursors was essentially similar. For all three precursors, 0 levels were highest for amorphous films, which were deposited at the lower end of the temperature range. AES indicated a decrease in 0 content with increasing sputter depth into the films; hence, high O levels were attributed to post-growth exposure of the films to air. The low density and high porosity of the amorphous structures grown at lower temperatures allow more rapid diffusion of O from the air into the films and provide a large surface to volume ratio for adsorption. Oxygen content drops considerably for films from each of the three precursors above their respective crystallization temperatures, due to film densification upon crystallization In addition to crystallization at higher deposition temperature, increased C content in the films grown at higher temperature is believed to further inhibit post-growth oxygenation of the films by stuffing grain boundaries. Lattice parameter shifts were also dependent on the precursor. Figure 7-3 shows lattice parameter generally increasing with temperature for films from lb, 2b and 3b. Lattice parameter was slightly higher for films deposited with 3b relative to lb, and significantly higher than for 2b at deposition temperatures :S 650C. Higher C content in films from lb and 3b compared to 2b is consistent with lower lattice parameter for films from 2b, as less C is available to expand the polycrystalline lattice in films from 2b. Since N levels are similar for polycrystalline films from lb and 3b, differences in lattice parameter are likely caused by differing C content on the interstitial sublattice. Films from 3b have slightly higher C content than those from lb, which may explain the slightly higher lattice parameter in films from 3b. Interestingly, the lattice parameter for films deposited with both lb and 3b is 4.17 A at 650 C. Polycrystalline films deposited

PAGE 363

354 by lb and 3b, then, have adequate N, even at high temperatures, to be considered primarily ~-WNx into which carbon is incorporated to form ~-WNxCy. The ~-WNxCy polycrystals deposited with lb and 3b, although N-deficient at higher temperatures, contain sufficient N to prevent a shift from formation of ~-WNxCy toward ~-WCy This is not the case for films from 2b, however, which experience a shift toward ~-WCy for deposition at temperatures above 675 C, as discussed in Chapter 6. The difference in lattice parameter for films from lb and 3b at 450C suggests that the imido ligand takes part in the deposition of C into the ~-WNxCy polycrystals, potentially by interacting with the W atom during deposition. Grain growth behavior was similar for the three precursors, with gram size increasing from lowest to highest deposition temperature. The smaller grain size for films from 2b relative to lb was attributed to the difference in the strengths of imido N-C bonds in the precursors, where the weaker bond in lb afforded higher growth rate and grain size for a given deposition temperature While the same bond is presumably weakest in 3b, the estimated grain size was surprisingly lower than that for films deposited with lb and 2b. The reason for this is unclear, but may be related to differences in the decomposition mechanism of 3b relative to lb and 2b which cause a decreased growth rate for films from 3b. As mentioned in Chapter 2 recent reports indicate interest in ~-WNxCy as a potential barrier material Although films deposited from 3b had relatively low N content (compared to typical WN x films), this precursor may be of interest for ~-WNxCy deposition. We have demonstrated the ability of an NH3 co-reactant to dramatically increase the N content for films deposited from lb, as discussed in Chapter 5, and a

PAGE 364

355 similar approach with 3b may be possible. A similar approach with 2b did not result in increased N content in the deposited film and may be due to the stronger N-C bond in the imido group of 2b inhibiting reaction with NH3 If this is the case, the decreased energy of the N-C bond in the imido group of 3b relative to lb may make 3b even more reactive with NH3 than its lb analogue. With the exception of the anomalous results for growth at 500 and 525C with 2b, a comparison of the sheet resistance, which negates the impact of film thickness on electrical properties shows that films from lb, 2b and 3b have similar electrical properties when deposited at or below 625 C. The sheet resistances for films from lb and 3b are 47 and 36 Q/0, respectively, for deposition at 450 C, rising above these values for increasing deposition temperature. The sheet resistance for films from 3b at 650C is 274 Q/0 compared to 139 and 127 Q/0 for films from lb and 2b, respectively The higher value for films from 3b is related both to elevated C content and smaller average grain size at 650 C compared to films from the other precursors Films from lb have similar C content ( 43 at.% ) to those from 3b ( 40 at.%) at 650C, but have a larger average grain size (60 A) compared to films from 3b (48 A), which tends to lower sheet resistance. Films from 2b have a slightly larger grain size (52 A) than those from 3b and also substantially lower C content (13 at.%), both of which tend to decrease sheet resistance. Comparison of the film growth properties of 3b to those of lb and 2b allowed evaluation of the effect of the imido N-C bond dissociation energy on film growth and properties. Films deposited from 2b wer e deficient in N compared to those from lb and 3b, consistent with a tendency of the stronger imido N C bond of 2b to result in

PAGE 365

356 dissociation of intact NPh fragments during deposition. As mentioned above, an optimal window for N-C imido bond energy appears to exist in these precursors. If the energy of this bond is too high (as in 2b), the W-N bond cleaves, leaving the films very deficient in N. If the bond energy is too low (as in 3b), N-C bond cleavage may occur in the gas phase, leading to processes that cause loss of imido N before final reaction at the substrate surface. A moderate N-C imido bond energy (as in lb) combines better N retention in the films with a substantial growth rate, due to enhanced precursor stability (relative to 3b) coupled with greater likelihood of N-C imido bond cleavage (relative to 2b). As discussed in Chapter 6, films from the isopropylimido precursor lb are superior to those from the phenylimido precursor 2b for barrier applications because they can be deposited at a lower minimum temperature (450C), contain more N, and have lower sheet resistance. Moreover, the isopropylimido precursor lb appears to be superior to the allylimido precursor 3b, due to higher growth rate and N content at their mutual lowest deposition temperature (450C). A variety of other precursors, listed in Appendix B, have also been tested for MOCVD of WNx. These precursors, however, did not prove to be promising

PAGE 366

CHAPTERS COPPER BARRIER TESTS 8.1 Copper Deposition Results Barrier films deposited from the i-Pr complex lb in PhCN solvent at the lowest temperature ( 450C) were used for Cu barrier testing. BCD and sputtering, the details of which are discussed in Chapter 4, were used to deposit Cu on the barrier films. Vacuum had to be broken after barrier film deposition and prior to Cu deposition, due to the nature of our CVD reactor set-up. This vacuum break enables large scale O in-diffusion into our barrier films, especially for the highly porous films deposited at low temperature. BCD Cu deposited by the bath chemistry detailed in Chapter 4 is reported to enhance the removal of surface oxides from the barrier layer, which form due to exposure of the barrier films to air [Sha0l] While some preliminary BCD Cu results are shown below, the properties of these films during the Cu removal process (discussed below) made them unusable for barrier integrity analyses. Moreover, due to difficulties with Cu oxidation during anneals in sweeping N2 flow, only the sputtered Cu/barrier/Si stack samples which had been annealed in-vacuo after Cu sputter deposition were examined in detail. 8.2 Electrical Measurements Sheet resistance of the deposited Cu films was measured by 4-point probe. Measurements indicated that as-deposited BCD Cu had a sheet resistance of 0.29 Q/0, while as-deposited sputtered Cu had a sheet resistance of 0.54 Q/0 The difference in sheet resistance was likely due to film stress and voids in the sputtered Cu, which can 357

PAGE 367

358 increase resistivity. Sputtered Cu films deposited at room temperature suffer from film stress and voids due to the inability of Cu atoms on the film surface to diffuse at low temperature. The film resistivity can be lowered by sputtering at elevated substrate temperature or by annealing the film stack. As discussed in Section 8.4.2, sheet resistance for sputtered Cu films decreased to a value of 0.24 Q/0 after annealing the film stack in-vacuo at 500C. 8.3 Scotch Tape Tests Scotch tape tests were conducted on as-deposited ECD and sputtered Cu/barrier stacks. The sputtered Cu had poor adhesion to the barrier layer, and came off completely onto the Scotch tape, leaving the barrier intact on the Si (100) substrate. This reflects poor adhesion at the sputtered Cu/barrier interface, likely caused by a significant quantity of oxide at this interface. The ECD Cu had strong adhesion to the barrier, however, which may reflect the ability of the electrochemical bath to reduce/remove the oxide at the barrier surface before Cu deposition begins. While the ECD Cu had good adhesion to the barrier, the ECD Cu and barrier layers came off onto the Scotch tape together. This indicates that the barrier/Si adhesion strength was lower than the ECD Cu/barrier adhesion strength. This contrasts with Scotch tape tests indicating good barrier/Si adhesion strength (discussed in Chapter 5), and suggests that the electrochemical bath deposition process weakens the barrier/Si interface. The weakening of this interface may be due to removal of oxygen from the bulk of the barrier film (due to the barrier's porosity), rather than just the removal of surface oxides. 8.4 Barrier Integrity Tests After Cu deposition the samples were annealed at various temperatures for 30 minutes to determine the maximum temperature at which the barrier could repel Cu from

PAGE 368

359 diffusing into the Si substrate. Annealing of the samples was first attempted in a Bio-Rad SKZ-1 Seiwa Optical microscope with a built-in bell jar chamber and heater assembly. N2 gas was flowed through the chamber during annealing to minimize the amount of oxygen in the chamber and to sweep out any Cu present in the vapor above the sample. Significant oxidation of the Cu during annealing in the bell jar, however necessitated an in-vacuo anneal of the sputtered Cu/barrier/Si stack samples in the sputter chamber after Cu deposition. Unlike sputtered Cu, which is deposited in a vacuum chamber, BCD Cu was deposited in ambient air making an in-situ vacuum anneal of the BCD Cu/barrier/Si stacks impossible. Moreover, attempts to etch BCD Cu from the stack prior to SIMS depth profiling resulted in dissolution of both the Cu and the barrier layers into the etchant solution. For this reason, sample stacks containing BCD Cu were not useful for SIMS depth profiling analyses. The barrier integrity study was therefore focused on the sputtered Cu/barrier/Si stack samples that were annealed in-vacuo, as only Cu dissolved into the etchant from these samples. 8.4.1 XRD XRD was done as a quick first check to determine if massive barrier failure had occurred. The presence of CuSi x compounds on the XRD spectra would indicate large-scale Cu penetration into the Si substrate meaning that the barrier layer failed to prevent diffusion. Figure 8-1 shows XRD spectra for BCD Cu/barrier/Si stacks, while Figure 8-2 shows spectra for sputtered Cu/barrier/Si stacks. Note that the Si (400) peak in the as-received spectrum of Figure 8-1 is not visible due to higher BCD Cu film thickness relative to the sputtered Cu layer shown in the as-received spectrum in Figure 8 2 The spectrum for the as received BCD stack (Figure 8-1) shows three Cu orientations deposited by BCD indicating polycrystalline Cu deposition.

PAGE 369

12000 CuO ,;;-10000 (111) -.... ; j .... '-" CUzO 8000 (111) 6000 360 CuO CuO (200) (202) Cu_. (111) Cu20 (220) ..-Cu (200) Si_. (400) 400c anneal 300 C anneal CUzO 'i (311) Cu (220) As J rec'd 0+-----~---~---~---~----1 30 40 50 60 70 80 20 Degrees Figure 8-1. XRD spectra for ECD Cu, before and after annealing in sweeping N2 flow. The most intense peak is for Cu (111) at 43.50 20 degrees, which is the desired phase from an electromigration and resistivity standpoint, but other peaks for Cu (200) and (220) at 50.65 and 74.50 are also present. The as-received spectrum in Figure 8-2 for sputtered Cu indicates only two peaks at 43.50 and 50.70 20 degrees, consistent with the Cu (111) and (200) orientations. While the Cu (111) and Cu (200) typically appear together in deposited Cu films, maximizing the intensity ratio of Cu (111) to Cu (200) is desirable. The intensity ratio in the as-received ECD Cu film is -2.8, while that for the as-received sputtered Cu was -2.2 (Figure 8-2), suggesting that ECD is superior to sputtering for preferential deposition of Cu (111) on our barrier films. The spectra in Figure 8-1 for samples after annealing do not indicate the presence of any peaks consistent with CuSix compounds, hence massive ECD Cu in-diffusion into the underlying Si after annealing up to 400 C is unlikely. While silicides are not present, spectra for the annealed stacks do indicate the presence of several Cu and CuO x peaks.

PAGE 370

361 The ECD Cu stack sample annealed at 300C still shows the Cu (111), (200) and (220) peaks, but contains oxide peaks as well. Peaks at 35.75 and 39.25 20 degrees are consistent with the (111) and (200) growth planes of CuO, respectively, while peaks at 36.65, 42.55, 61.60 and 74.30 20 degrees are consistent with the (111), (200), (220) and (311) growth planes of Cu20, respectively. The ECD Cu stack sample annealed at 400C does not contain any peaks consistent with pure Cu. This stack still contains the Cu20 (111) and (220) orientations, but the other Cu 2 0 peaks have disappeared in favor of CuO. One new CuO peak appears at 49.10 20 degrees, consistent with the (202) orientation. While the annealed ECD Cu spectra contain peaks consistent with Cu, CuO and Cu 20, the annealed sputtered Cu spectra indicate the presence of only CuO and Cu20 polycrystals. The sputtered Cu sample annealed at 300C has one peak at 36.65 20 degrees, which indicates the Cu 2 0 (111) growth plane. The sputtered Cu sample annealed at 400C has peaks at 35.75 and 39.05 20 degrees, which indicate the CuO (111) and (200) growth planes. Several of the ECD and sputtered Cu spectra display additional peaks at 32.70 and 61.85 20 degrees, representing reflections from the underlying Si consistent with Si (200) Ka and Si (400) K~ radiation, respectively. The increase in CuO intensity and concomitant decrease in Cu 2 0 intensity when increasing annealing temperature from 300 to 400C indicates the increased stability of CuO relative to Cu20, assuming that adequate O is available to make a 1:1 stoichiometric oxide. The presence of Cu peaks on the ECD Cu samples after annealing at 300C, but not on the sputtered Cu samples annealed at the same temperature, was due to increased thickness of ECD Cu relative to the sputtered Cu layer. The thicker ECD Cu layer was not consumed by oxide formation as quickly as the thinner sputtered Cu layer.

PAGE 371

2500 ,-.,
PAGE 372

363 the bell jar, because Cu is not consumed by reaction with atmospheric oxygen in this case, leaving more Cu available to potentially diffuse into the barrier. None of the spectra indicate the presence of any peaks consistent with CuSix compounds, even after 5000 Cu Cu (200) (111) ,. ,-.. 4000 Cl.l .... .... s::: 3000 ti .-si c200) Si (400) Kl}+ .... '-' 2000 ;>-. .... .... Cl.l s::: Q) .... 1000 30 40 50 60 70 80 28 Degrees Figure 8-3. XRD spectra for as-received and annealed sputtered Cu/barrier/Si stacks, with annealing temperatures listed on the right side of the diagram. Annealing was performed in vacuo in the sputter chamber. annealing at 500C, hence massive in-diffusion of Cu into the underlying Si after annealing in vacuo is unlikely. The spectra do indicate the presence of polycrystalline Cu, without any oxide peaks being present. Grain growth occurred for the Cu films after annealing in vacuo, as is evident by an increase in intensity and sharpness of the Cu (111) peak when comparing spectra for the as-received films to the annealed ones. The difference between XRD spectra for films annealed in vacuo and in atmosphere highlights the importance of annealing in vacuo to prevent degradation and atmospheric oxidation of the Cu layer during annealing. The presence of O in the barrier and/or Cu layers can be eliminated by depositing the Cu on the barrier layer without a vacuum

PAGE 373

364 break ( discussed further in Chapter 9), followed by subsequent annealing of the stack in-vacuo. Eliminating O from the barrier and Cu layers would give the most robust test of a barrier's ability to resist Cu diffusion. 8.4.2 Electrical Measurements Four-point probe measurements were attempted on the samples that were annealed in the bell jar, but the film surface, which was presumably copper oxide, was too resistive to give a sheet resistance measurement. Measurements were obtained for sample stacks containing sputtered Cu that were annealed in vacuo, however. The results are shown in Figure 8-4. The results indicate that the sheet resistance of the as-deposited sputtered Cu film drops after the in vacuo anneal. This is consistent with XRD spectra, 1.0 -r-----------------------, -Q.) 0 8 :::l er ell a 0 6 ..._, Q.) u ro .. ell 0.4 ..... ell Q.) p:: .. Q.) Q.) 0.2 ..c: Cf) 0.0 0 100 200 300 400 500 Annealing Temperature (C) Figure 8-4. Variation of Cu sheet resistance with annealing temperature for a sputtered Cu/barrier/Si stack annealed in vacuo. As-deposited Cu sample shown with annealing temperature of 25C which indicate grain growth in the Cu film after annealing. Increasing grain size in the Cu films tends to decrease the resistance to current flow, which is indicated by the drop in sheet resistance for annealed films relative to the as-deposited one. Some variation in

PAGE 374

365 sheet resistance for the annealed samples is evident in the plot, but the general trend indicates a decrease in sheet resistance with higher annealing temperature. If large-scale Cu penetration through the barrier was occurring, one would expect an increase in sheet resistance as the Cu moves into the barrier and Si substrate. 8.4.3 Depth Profiling Analysis AES depth profiling was done as a first pass to determine if the barrier prevents large-scale Cu diffusion. Figure 8-5 shows a depth profile of a Cu/barrier/stack annealed at 300C in the bell jar assembly. After nearly 40 minutes of sputtering, the Cu signal disappears to the noise level. The Si signal rises above the noise level after slightly more than 40 minutes of sputtering, suggesting that the barrier prevented Cu from intermixing with Si during the 300C anneal. Significant O contamination is evident in the Cu and barrier layers, with the Cu layer's O contamination coming from ambient oxygen during the anneal. Since the consumption of Cu by oxide formation could presumably slow the advance of Cu through the barrier, samples which were annealed in vacuo need to be checked by depth profiling. In addition to testing of samples annealed in vacuo, depth profiling by SIMS should be done to get a more conservative analysis of the barrier's ability (or lack thereof) to prevent Cu diffusion, since the elemental detection limits for SIMS are much lower than those for AES. To prevent knock-in effects from ion bombardment during SIMS, any Cu on the samples after annealing must be stripped off prior to SIMS depth profiling [Wan0lb, Bai90]. A variety of etchant recipes for the removal of Cu deposited on WNx barrier layers have been reported in the literature. One report suggested a Cu etchant containing H3PO4:HNO3:CH3COOH:H2O with a ratio of 80:5:5:10 [Lee98c],

PAGE 375

366 while other reports suggested 10% nitric acid (HN03 ) [lva99] or "diluted" HN03 in water [Vij99]. Due to its simplicity, the 10% HN03 in water formulation was used. 300000 ,,-.... 250000 Cl) ..... a Cu 200000 h ..... 150000 ..__, 0 100000 ..... Cl) s::: Q.) ..... s::: 50000 0 25 30 35 40 45 50 Sputter Time (min) Figure 8-5. AES depth profile for sputtered Cu/barrier/Si stack after annealing m sweeping N2 at 300C for 30 minutes SIMS depth profiles are required both to test barrier integrity and to determine the apparent diffusivity and the apparent activation energy for diffusion in the films. Due to the current inoperability of the SIMS system, however, these measurements must be deferred to future work. 8.5 Conclusions Regarding Cu Testing of Deposited Barrier Layers ECD Cu had superior adhesion to the barrier layer compared to sputtered Cu, and may have been due to reduction of oxides on the barrier surf ace before ECD deposition initiated. The as-deposited ECD Cu layer also had lower sheet resistance than as-deposited sputtered Cu. Annealing of ECD samples had to be done in the bell jar, as the ECD process is done at ambient conditions, making a post Cu deposition in vacuo

PAGE 376

367 anneal of ECD Cu impossible. Annealing in the bell jar did not cause any Cu silicide peaks to appear on the XRD spectra, but did cause large-scale oxidation of the ECD and sputtered Cu samples, making this annealing process impractical if further barrier integrity testing with 4-point probe is desired. To prevent oxidation of Cu by ambient oxygen during annealing, some sputtered Cu/barrier/Si stacks were annealed in vacuo. For the sputtered Cu/barrier/Si sample stacks annealed in vacuo, neither XRD nor 4-point probe measurements were consistent with large-scale Cu diffusion through the barrier to the underlying Si even after a 30 minute anneal at 500C. Due to the possibility of knock-in effects during SIMS analysis, the Cu layer should be etched off of the barrier before a SIMS depth profile is collected. Since the ECD Cu and barrier came off together during the etch step, a study of ECD Cu/barrier integrity (or lack thereof) by SIMS depth profile analysis was impractical. Future SIMS depth profiles are still required, however, and await repairs to the SIMS system. Due to the high porosity and high O content in the films deposited from lb at 450C, further work to deposit dense, amorphous barrier films is necessary, as O may have enhanced the ability of these barrier films to resist Cu diffusion. As discussed further in Chapter 9, in situ Cu deposition and annealing capability will eliminate the occurrence of barrier oxidation prior to Cu deposition tests, giving a truer test of the barrier's integrity.

PAGE 377

CHAPTER9 RECOMMENDATIONS FOR FUTURE WORK In this study, we have shown a correlation between precursor structure and deposited film properties, and how modifications to the precursor structure can lead to desirable ( or undesirable) shifts in these properties. Overall results for the three precursors studied in depth however, indicate that these precursors, deposited by solvent/nebulizer based MOCVD, are unable to deposit clean, dense, low resistivity, highly conformal barrier films with excellent barrier integrity at low temperature ( <400C). Continued synthesis and testing of novel precursors, which lead to industrially applicable diffusion barrier materials, will require significant future research and development. Several different avenues of precursor optimization and barrier research that should be pursued, and these are discussed in some detail below. 9.1 Control Film Composition Contamination levels in the deposited films must be reduced in order for the deposited films to be industrially applicable. Contamination by 0, Cl and free C can have detrimental effects on film properties, such as increased film resistivity, poor inter-layer adhesion, and corrosion of the metallization layers. In addition, the amount of N (and C) in ~-WNx (or ~-WNxCy) must be controlled to ensure good film resistivity coupled with excellent resistance to Cu in-diffusion. To decrease contaminant levels, the precursor's molecular structure must be modified. The precursor structure should continue to exclude oxygen, and its R-groups 368

PAGE 378

369 should resist homogeneous and heterogeneous decomposition at the reaction conditions. If the R-groups separate from the W-N fragment and remain intact in the gas phase, the amount of free C from the R-groups depositing into the growing film decreases. If these R-groups decompose during the reaction, though, the amount of free C deposited into the film will be significant. In addition, Cl should be eliminated from the precursor as the Cl ligands can cause significant Cl incorporation into the barrier films, which can corrode subsequent metallization layers deposited on the barrier. Other N bearing ligands should be used to replace the Cl atoms if possible, both to eliminate Cl and to boost N levels in the film. The addition of N-bearing co-reactants (such as NH3 ) to the reactor should also be pursued with any new precursors, as this showed promise in boosting N levels for films deposited with the isopropyl imido (lb) precursor. Apart from the precursor's structure, the solvent selection issue should be revisited, as benzonitrile has been shown to cause significant amounts of unintentional C and N incorporation into the films. Another solvent which has desirable vapor pressure properties coupled with minimal operator exposure concerns, should be selected so as to minimize C incorporation We have discussed the use of mesitylene [1,3,5-trimethylbenzene] as a possible solvent, and this should be pursued experimental} y. If intentional deposition of WN x C y films is desired in the future, the amount and location of C in the film must be controlled. A thorough study of C deposition and positioning in WNxCy films has yet to be done however. To deposit C into the film in a controlled manner, a study using C 2Ri as a co-reactant may be done. This C 2Ri co-reactant would enable us to independently control the C level in film so that the

PAGE 379

370 effect of varymg C content (at a given deposition temperature) on film structure, resistivity, and barrier integrity may be observed. These films should be examined with XRD, AES, RBS (if available) and XPS to determine their structure, content and elemental bonding states. To limit contamination in the films caused by background contaminants during the reaction, a load-lock chamber should be added to the reactor. After loading the substrate and susceptor into this chamber, the loading door would be closed and vacuum would be pulled on the chamber. Then, the gate valve on the reactor would be opened, and a magnetically coupled loading arm would be used to move the susceptor into the reactor, which would be maintained at vacuum and never exposed to atmosphere (except during cleanout). A glove box would be required around the thermocouple port on the reactor to prevent any air leaks during vertical adjustment of the thermocouple. By maintaining the reactor at vacuum between runs, the background contaminant levels (especially 0) in the reactor during deposition will be minimized, thereby improving properties of the deposited film This should minimize or eliminate the interface layer between the barrier film and the Si substrate, which was discussed in Chapter 5. After deposition, barrier integrity must be tested by deposition of a Cu layer, followed by subsequent annealing of the Cu/barrier/substrate stack. While a load-lock chamber would minimize contamination of the barrier film during deposition exposure of the barrier layer to atmosphere after deposition can cause undesirable oxidation of the surface (and potentially bulk) of the film. To prevent oxidation of the barrier film prior to Cu deposition the vacuum break step should be eliminated. Adding a Cu deposition reactor with a transfer chamber connected in vacuo to the barrier deposition system,

PAGE 380

371 should make this possible. After Cu deposition, annealing of the Cu/banier/substrate stack could be done in the Cu reactor before exposing the sample to air for ex situ characterization. 9.2 Decrease Deposition Temperature The deposition temperature for the MOCVD precursor should be decreased below 400C, as discussed in Chapter 1. To enable a lower deposition temperature, modifications to the precursor's structure, coupled with the possibility of co-reactant addition (especially for ALD), should be examined. The precursor structure should be altered to contain halide-free ligands that can dissociate at low temperature ( <400C) NH3 is the obvious choice for a N-containing co-reactant, and is typically used for co-reactant CVD and ALD of WNx (and WNxCy) films. Reactivity of NH3 drops significantly at lower temperature hence an alternative low-temperature N source should be investigated. Plasma cracking of an NH3 or N 2 co-reactant, for example, before they are introduced to the reactor, may help to drop deposition temperature. Alternatively, other less stable N bearing co-reactants, such as hydrazine (N 2~), for example, may also be promising as a low-temperature N source, and should be tested with the novel CVD and ALD precursors generated by further study Lab safety concerns have prevented us from testing N2~ as a co-reactant with existing precursors so collaborating with a group that has N2~-based CVD (or ALD) capability may be useful. 9.3 Decrease Barrier Thickness The current industry goal is to deposit a stable diffusion barrier layer with a thickness of 50 A. Deposition rates and precursor chemistry must be controlled to grow thin conformal films that adequately prevent Cu-substrate intermixing. If a good, conformal barrier with 50 A thickness is achieved, this should be valuable for several

PAGE 381

372 generations of device feature sizes. The use of ALD should enable deposition of ultra-thin barrier films to meet the needs of future device generations. Moreover monolayer by monolayer growth that is characteristic of ALD should enhance barrier film density, decreasing the number of voids or pores through which Cu may diffuse. This would also enhance the ability of the barrier to resist O in-diffusion, if a vacuum break is necessary before Cu deposition. To ensure barrier deposition by ALD, precursor modifications will be necessary so that film growth is stepwise and self-limiting. We have considered the use of an N-chloro tungsten imido precursor Cl4(CH3CN)WNC1 for ALD, but moving toward an alkyl ligand based chemistry would be preferable to prevent Cl incorporation. 9.4 Further Conformality Testing Although conformality tests have been performed on 2 m diameter trench features with an AR of 2 75, the conformality of films grown by MOCVD precursors should be tested on state-of-the-art device feature sizes (diameter :S 0.10 m). While we do have test wafers with 0.10 m diameter features in hand these wafers contain low-k dielectric material and shallow doped junctions, which must be processed below 400C to prevent damage to these components. The deposition temperature must be lowered (by precursor modification, use of ALD, etc ) before these test wafers may be used in our reactor. 9.5 Further Adhesion Tests Good adhesion is essential so that gaps between the barrier and substrate (where electromigration or diffusion may easily occur) do not form Scotch tape tests were conducted to determine if the barriers adhered better to the Si substrates than to the tape

PAGE 382

373 but this is simply a qualitative test. A nanoindentation or scratch test should be done to get a quantitative idea of adhesion strength between the barrier and the substrate layers. 9.6 Further Barrier Integrity Analyses To enhance sensitivity of the barrier integrity tests, SIMS depth profiling of the barrier films after annealing with Cu in-vacuo should be performed. Another sensitive test involves deposition of the barrier films on pre-formed p-n junctions on Si substrates, which may be provided by Dr. Norton's group. After barrier deposition, the Cu would be deposited, and the Cu/barrier/p-n junction stack would be annealed to a target temperature. After annealing, the Cu and barrier would be stripped, and leakage current across the p-n junction should be measured. Cu that has penetrated to the Si will cause a significant leakage current, making this a very sensitive electrical technique for detecting barrier failure. While this technique is very sensitive test of Cu diffusion, it is qualitative in nature, where the barrier film's ability to resist Cu in-diffusion is judged on a pass/fail basis. Typically, the appearance of Cu in the underlying substrate is interpreted as barrier failure. Quantitative measurements of copper flux through the diffusion barrier are lacking. By measuring the flux of Cu through a barrier layer and the Cu concentration at the barrier-substrate interface, Fick's first law may be solved for the true apparent diffusivity of the barrier film This would enable quantitative comparisons of various barrier materials, and allow determination of the impact of chemical and microstructural differences on barrier behavior. To measure Cu flux through the barrier, the diffusing Cu must be "captured" in a medium that will allow easy composition analysis. In addition to easy compositional analysis, the capture medium should readily absorb Cu and be well mixed so that the Cu concentration in this medium can be assumed constant. A solid

PAGE 383

374 capture medium presumably would contain concentration gradients, making concentration determination at the barrier-substrate interface difficult. Si, for example, has very low Cu solubility, with Cu preferentially populating defect sites in the Si lattice. This would make measurement of an average Cu concentration in the Si difficult, due to the dependence of Cu concentration on defect concentration. A new technique to capture copper diffusing through a barrier may be set up, wherein a stack of low melting point metal/barrier/Cu would be deposited inside a capped "crucible" made of an impermeable material (e.g., Si]N4). A known quantity of low melting point metal would be the capture medium. During annealing, the metal should liquefy to yield a medium with high Cu diffusivity (-10-5 cm2/s), allowing thorough mixing and uniform impurity (i.e., Cu) concentration in the melt. The Cu concentration at the barrier-metal interface can therefore be approximated as bulk concentration in the metal. After annealing for a set time, the stack would be quenched, and the concentration in the capture medium can be checked by a variety of methods, such as Rutherford backscattering spectrometry (RBS). With the Cu concentration in a known quantity of low melting temperature metal determined, the flux of Cu through the barrier, and the true apparent diffusivity, can be calculated. This work would need to address several potential issues before implementation, including de-wetting of the Cu from the crucible, barrier decomposition or dissolution into the liquid metal, and uniformity of crucible heating. Quantitative barrier diffusivity studies could be coupled with an in-depth analysis of diffusion behavior through the grain boundaries in polycrystalline WNx films. Depositing amorphous WNx films with x<0.5 by a clean method (such as sputtering),

PAGE 384

375 followed by annealing, typically leads to polycrystal ~-WNx films without Nat the grain boundary. Increasing N content above x=0.5 should lead to polycrystalline ~-WN0 5 films with N at the grain boundaries. Quantitative diffusivity through polycrystalline films with and without N at the grain boundaries could be measured using the stack cell mentioned above. The change in diffusivity with annealing temperature may then be used to determine apparent activation energies for diffusion through grain boundaries with and without N. This should give insight into nitrogen's ability to chemically repel Cu diffusion at the grain boundaries. XPS may also be used to study Cu-N interaction at the grain boundary. XPS can be used to differentiate between N in the polycrystals and N at the grain boundary. If significant repulsion exists between grain boundary N and Cu, a shift in Nls BE for grain boundary N may occur after the barrier is exposed to Cu. Experimental evidence of this type indicating N-Cu repulsion would be very useful in making a case for the continued use of nitride barriers in future device generations, rather than changing over to elemental metal barriers such as Ru 9.7 Deposition on Alternate Substrate Materials Diffusion barrier layers will need to be deposited on a variety of substrates on the IC device. Ideally, one barrier material could be used at all points on the IC where a barrier is required. If this material could be deposited on all necessary surfaces by a single precursor, this would save considerable process time caused by precursor/material changeout. While we have studied deposition from a variety of tungsten imido precursors on Si substrates the ability of these precursors to deposit on other potential IC substrate materials should be tested. Several potential materials that should be

PAGE 385

376 investigated include Si02 (inter-level and gate dielectric) Si]N4 (etch stop layer), new low-K dielectric materials such as SiCOH, HSQ and polyimide, and new high-K dielectric materials such as HfSiOxNy, In addition, the ability of these precursors to deposit a barrier layer on Cu should be tested, as this should be the future metallization material at all levels on the IC device. 9.8 Testing of Films for X-ray Absorption Mask Applications Since WNx films have been reportedly used as X-ray absorption masks for lithography applications our deposited films should be tested to determine their applicability for this use. Collabor~tion with a lithographic mask maker for example may enable this type of testing to be done. 9.9 Deposition of Alternate Barrier Materials Other barrier materials in addition to WN x should be considered for future studies. Current industrial interest in ALD of TaN and WNxCy, coupled with minimal precursor routes to these materials, makes precursor development for these schemes a potentially hot area 9.10 System Modifications Several modifications to the existing MOCVD system would be useful to improve its operation. In addition to the load-lock and in-situ Cu deposition chambers discussed above, several more modifications to th e MOCVD system should be made to improve operation First some valves in the system such as the H 2 NH3 and B 2Ht; source valves should be automated and interlocked to a central control panel. This would enable immediate shutdown o f flammable/hazardous process gases in the event of a safety incident (e.g., fire) in the lab. Another safety issue to b e resolved is replacement of the H 2 g a s alarm in the lab. Since the existing alarm is not functioning a new one should be

PAGE 386

377 purchased and installed to ensure that any H2 leaks from the system are immediately evident. Apart from the deposition system, one modification to the residual gas analysis (RGA) cart in the lab is necessary. The roughing and turbo pumps on the RGA cart are unable to adequately pump H2 carrier gas, leading to an overpressure in the RGA sensor when H2 is used in the reactor. These pumps should be replaced with higher capacity models that can handle the expected H2 flowrate from the capillary sample tube to the RGA cart. Work 1s currently underway to construct a new ALD system for barrier deposition, which will have in-situ Cu deposition capability. The system will contain separate barrier and Cu deposition reactors, connected by an in-vacuo transfer chamber with associated linear magnetic feedthroughs. The barrier chamber will have ALD capability, but can be easily changed over to CVD operation by leaving all reactant gases flowing simultaneously, rather than having separate pulses isolated by purge and vacuum steps. The Cu reactor will be run in CVD mode, based on existing Cu CVD chemistries, but will have the potential for operation in ALD mode, if future research on Cu ALD is desired.

PAGE 387

REFERENCES [Aal03] Aaltonen, T., P. Alen, M. Ritala and M. Leskela, Chemical Vapor Deposition 9(1) (2003) 45-49. [Abe93] Abe, H., T.-K. Cheung and A. T. Bell, Catalysis Letters 21(1-2) (1993) 11-18. [Aff85] Affolter K., H. Kattelus and M.-A. Nicolet, Materials Research Society Symposium Proceedings, Vol.47 (1985) 167-173. [Ala91] Alay, J. L., H. Bender, G. Brijs, A. Demesmaeker and W. Vandervorst, Surface and Interface Analysis 17(6) (1991) 373-382. [And99] Andricacos, P. C., Electrochemical Society Interface (1999) 32-37. [Ans87] Ansara, I. and B. Sundman (1987). Proceedings of the International CODATA Conference (1987): lOth(Comput. Handl. Dissemination Data) Vol.1986, Ottawa, North Holland: 154-158. [Aoy99] Aoyama, T. and K. Eguchi, Japanese Journal of Applied Physics, Part 2 38(10A) (1999) 1134-1136. [Asm69] Asmus T. W. and T. J. Houser, Journal of Physical Chemistry 73(8) (1969) 2555-2558. [Bai90] Bai, P., G.-R. Yang, L. You, T.-M. Lu and D. B. Knorr, Journal of Materials Research 5(5) (1990) 989-997. [Bai96] Bai G., S. Wittenbrock, V. Ochoa, R. Villasol C. Chiang, T. Marieb, D. Gardner, C. Mu, D. Fraser and M. Bohr (1996). Materials Research Society Symposium Proceedings, Vol.403: 501-506. [Bai03] Bain, M. F., B. M. Armstrong and H. S. Gamble, Journal of Materials Science: Materials in Electronics 14(5-7) (2003) 329-332. [Bak02] Baker, C. C and S. I. Shah Journal of Vacuum Science Technology A 20(5) (2002) 1699-1703. [Bal75] Balluffi, R. W. and J.M. Blakely, Thin Solid Films 25(2) (1975) 363392. 378

PAGE 388

379 [Ban79] Banik, G., P. Ettmayer, A. Vendl and R. Kieffer, High Temperatures High Pressures 11 (1979) 349-352. [Bar94] Barr, T. Modern ESCA: The Principles and Practice of X-Ray Photoelectron Spectroscopy. Boca Raton, CRC Press, Inc., (1994) [Bau97] Baumann, J., C. Kaufmann, M. Rennau, T. Werner and T. Gessner, Microelectronic Engineering 33 (1997) 283-291. [Bau98] Baumann, J., J. Kriz, K. Gottfried, J. Brauer, A. Ehrlich, C. Kaufmann and T. Gessner (1998). Advanced Metallization and Interconnect Systems for ULSI Applications in 1997,Vol.ULSI XIII, San Diego, CA, MRS: 117-123. [Bch03a] Hehir, 0. J., S. W. Johnston, A. C. Cuadra, T. J. Anderson, C. G. Ortiz, B. C. Brooks, D. H. Powell and L. McElwee-White, Journal of Crystal Growth 249(1-2) (2003a) 262-274. [Bch03b] Hehir, 0. J., T. J. Anderson, B. C. Brooks and L. McElwee-White (2003b). Electrochemical Society Conference: Chemical Vapor Deposition XVI (CVD-VXI) and EUROCVD 14,Vol.2003-08, Paris, France, The Electrochemical Society: 424-431. [Bch03c] Hehir, 0. J., K. M. Green, M. S. Hlad, T. J. Anderson, B. C. Brooks, C. B. Wilder, D. H. Powell and L. McElwee-White, Journal of Organometallic Chemistry 684(1-2) (2003c) 338-350. [Bec03a] Becker, J. S. and R. G. Gordon, Applied Physics Letters 82(14) (2003a) 2239-2241. [Bec03b] Becker, J. S., S. Suh, S. Wang and R. G. Gordon, Chemistry of Materials 15(15) (2003b) 2969-2976. [Ben76] Benson, S. W. Thermochemical Kinetics. New York, Wiley lnterscience, (1976) [Bey02] Beyer, G., A. Satta, J. Schuhmacher, K. Maex, W. Besling, 0. Kilpela, H. Sprey and G. Tempel, Microelectronic Engineering 64(1-4) (2002) 233-245. [Bha80] Bhattacharya, A. K., Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases 76(1) (1980) 126-134. [Bin95] Binder, S., W. Lengauer, P. Ettmayer, J. Bauer, J. Debuigne and M. Bohn, Journal of Alloys and Compounds 217(1) (1995) 128-136.

PAGE 389

380 [Bla86] Blanchard, L., J. Grimblot and J.P. Bonnelle, Journal of Catalysis 98 (1986) 229-234. [Bla97] Blanquet, E., A. M. Dutron, V. Ghetta, C. Bernard and R. Madar, Microelectronic Engineering 37/38 (1997) 189-195. [Boh90] Boher, P., P. Houdy, P. Kaikati and L. J. Van Ijzendoorn, Journal of Vacuum Science Technology A 8(2) (1990) 846-850. [Bon80] Bonzel, H.P. and H.J. Krebs, Surface Science 91 (1980) 499-513. [Bos91] Bosseboeuf, A., A. Fourrier, F. Meyer, A. Benhocine and G. Gautherin, Appl. Surf. Sci. 53 (1991) 353-357. [Bou91] Bou, M., J.M. Martin, T. Le Mogne and L. Vovelle, Appl. Surf. Sci. 47 (1991) 149-161. [Bou97] Boukhris, L. and J.-M. Poitevin, Thin Solid Films 310 (1997) 222-227. [Bri46] Brinkley, S. R., Jr., Journal of Chemical Physics 14 (1946) 563. [Bri90] Briggs, D. and M. P. Seah. Practical Surface Analysis. New York, John Wiley and Sons, (1990) [Bro79] Brooks, C. T., S. J. Peacock and B. G. Reuben, Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases 75(3) (1979) 652-662. [Cai70] Caillet, M., Y. L. Lagarde and J. Besson, Comptes Rendus des Seances de l'Academie des Sciences, Serie C: Sciences Chimiques 270(23) (1970) 1867-1869. [Car74] Cartier, P. G. and R.R. Rye, Journal of Catalysis 32(1) (1974) 88-98. [Car89] Carim, A.H., A. F. De Jong and P. Houdy, Thin Solid Films 176 (1989) LI 77-L182. [Cha82] Chamberlain, M. B., Thin Solid Films 91(2) (1982) 155-162. [Cha88] Chappell, P. J. C., M. H. Kibel and B. G. Baker, Journal of Catalysis 110 (1988) 139-149. [Cha89] Charai, A., S. E. Hoernstroem, 0. Thomas, P. M. Fryer and J.M. E. Harper, J. Vac. Sci. Technol., A 7(3, Pt. 1) (1989) 784-9. [Cha97a] Chang, K. M., I. C. Deng, T. H. Yeh, C. W. Shih and S. C. Chien (1997a). Electrochemical Society Conference Proceedings,Vol.97 25: 1559-1565

PAGE 390

381 [Cha97b] Chang, K. M., T. H. Yeh and I. C. Deng, Journal of Applied Physics 81(8) (1997b) 3670-3676. [Cha97c] Chang, K. M., T. H. Yeh, I. C. Deng and C. W. Shih, Journal of Applied Physics 82(3) (1997c) 1469-1475. [Cha99] Chang, K. M., I. C. Deng, T. H. Yeh, K. D. Lain and C. M. Fu, Japanese Journal of Applied Physics 38(Part 1, No. 3A) (1999) 13431351. [Cha04] Chappell, J., PVD Still in for the Long Haul, Electronic News (2004) [Che97] Chen, F., J. Aitken and S. Parasher (1997). Materials Research Society Proceedings, Vol.453: 121-126. [Che99a] Chen, X., H. L. Frisch, A. E. Kaloyeros, B. Arkles and J. Sullivan, Journal of Vacuum Science Technology B 17(1) (1999a) 182-185. [Che99b] Chen, X., G. G. Peterson, C. Goldberg, G. Nuesca, H. L. Frisch, A. E. Kaloyeros, B. Arkles and J. Sullivan, Journal of Materials Research 14(5) (1999b) 2043-2052. [Chi92] Chiu, H. T. and S. H. Chuang (1992). Materials Research Society Symposium Proceedings: Chemical Vapor Deposition of Refractory Metals and Ceramics II,Vol.250: 317-322. [Chi93] Chiu, H. T. and S. H. Chuang, Journal of Materials Research 8(6) (1993) 1353-1360. [Chi96] Chitica, N., A. Lita, G. Marin, I. N. Mihailescu, M. Popescu, C. Grivas and A. Hatziapostolou, Journal De Physique IV 6(C4 Rayons X et Matiere) (1996) C4-461 C4-465. [Chi97] Chitica, N., E. Gyorgy, A. Lita, G. Marin, I. N. Mihailescu, D. Pantelica, M. Petrascu, A. Hatziapostolou, C. Grivas, N. Broll, A. Comet, C. Mirica and A. Andrei, Thin Solid Films 301 (1997) 71-76. [Chi98] Chin, B., P. Ding, B. Sun, T. Chiang, D. Angelo, I. Hashim, Z. Xu, S. Edelstein and F. Chen, Solid State Technology 41(7) (1998) 141-147. [Chi0l] Chin, B., G. Yao, P. Ding, J. Fu and L. Chen, Semiconductor International 24(5) (2001) 107. [Chi02] Chin, Y. L., B. S. Chiou and W. F. Wu, Japanese Journal of Applied Physics, Part 1 41(5A) (2002) 3057-3064.

PAGE 391

382 [Cho98] Cho, S.-L., K. B. Kim, S. H. Min and H.-K. Shin (1998). Materials Research Society Conference Proceedings, Vol.514, San Francisco, CA, Materials Research Society: 531-536. [Cho99] Cho, S. L., K. B. Kim, S. H. Min, H. K. Shin and S. D. Kim, Journal of the Electrochemical Society 146(10) (1999) 3724-3730. [Cho00] Cho, D.-H., T.-S. Chang and C.-H. Shin, Catalysis Letters 67(2-4) (2000) 163-169. [Cho0l] Cho, I. H., J. S. Park, D. K. Sohn and J. H. Ha, Japanese Journal of Applied Physics 40(8) (2001) 4854-4861. [Cho02] Choi, I. S., J.C. Park, J. T. Choi, H.J. Kim and S. Y. Lee, Applied Physics Letters 80(23) (2002) 4339-4341. [Chy03] Chyan, 0., T. N. Arunagiri and T. Ponnuswamy, Journal of the Electrochemical Society 150(5) (2003) C347-C350. [Col76] Colton, R. J. and J. W. Rabalais, Inorganic Chemistry 15(1) (1976) 236-238. [Coo89] Cooke, M. J. and G. Harris, Journal of Vacuum Science & Technology, A: Vacuum, Surfaces, and Films 7(6) (1989) 3217-3221. [Cra0l] Crane, E. L., H. T. Chiu and R. G. Nuzzo, Journal of Physical Chemistry B 105 (2001) 3549-3556. [Cri00] Crist, V. Handbook of Monochromatic XPS Spectra: The Elements and Native Oxides. New York, John Wiley & Sons, Ltd., (2000) [Cro67] Crow, W. D. and C. Wentrup, Tetrahedron Lett. (44) (1967) 4379-&. [Cul78] Cullity, B. D. Elements of X-ray Diffraction. Reading, MA, Addison Wesley Publishing Company, (1978) [Cul0l] Cullity, B. D. and S. R. Stock. Elements of X-ray diffraction: Third Edition. Upper Saddle River, NJ, Prentice Hall, (2001) [Cur92] Currie, A. L. and K. E. Howard, Journal of Materials Science 27 (1992) 2739-2742. [Dan02] Dance, B., Semiconductor International 25(4) (2002) 40. [deB88] de Boer, F. R., R. Boom, W. C. M. Mattens, A. R. Miedema and A. K. Niessen. Cohesion in Metals: Transition Metal Alloys. Amsterdam, North-Holland Physics Publishing, (1988)

PAGE 392

383 [DeH93] DeHoff, R. T. Thermodynamics In Materials Science. New York, McGraw-Hill, Inc., (1993) [Den89] Deneuville, A., M. Benyahya, M. Brunel, J.C. Oberlin, J. Torres, N. Bourhila, J. Paleau and B. Canut, Appl. Surf. Sci. 38 (1989) 139-147. [Dia73] Dianis, W. P. and J.E. Lester, Analytical Chemistry 45(8) (1973) 14161420. [Din91] Dinsdale, A. T., Calphad: Computer Coupling of Phase Diagrams and Thermochemistry 15(4) (1991) 317-425. [Du91] Du, H. and M. Hillert, Zeitschrift filr Metallkunde 82 (1991) 310-316. [Eck02] Ecke, R., S. E. Schulz, M. Hecker and T. Gessner, Microelectronic Engineering 64 (2002) 261-268. [Eck03] Ecke, R., S. E. Schulz, M. Hecker, N. Mattern and T. Gessner, Microelectronic Engineering 70(2-4) (2003) 346-351. [Eis00b] Eisenbraun, E., A. Upham, R. Dash, W. Zeng, J. Hoefnagels, S. Lane, D. Anjum, K. Dovidenko, A. E. Kaloyeros, B. Arkles and J. Sullivan, Journal of Vacuum Science & Technology B 18(4) (2000b) 2011-2015. [Eiz94a] Eizenberg, M., K. Littau, S. Ghanayem, A. Mak, Y. Maeda, M. Chang and A. K. Sinha, Applied Physics Letters 65(19) (1994a) 2416-2418. [Eiz94b] Eizenberg, M., F. Meyer, A. Benhocine and D. Bouchier, Journal of Applied Physics 75(8) (1994b) 3900-3907. [Eks0l] Ekstrom, B. M., S. Lee, N. Magtoto and J. A. Kelber, Appl. Surf. Sci. 171(3-4) (2001) 275-282 [Ela0l] Elam, J. W., C E. Nelson, R. K. Grubbs and S. M. George, Surface Science 479(1-3) (2001) 121-135. [Ele02] Elers, K.-E., V. Saanila, P. J. Soininen, W.-M. Li, J. T. Kostamo, S. Haukka and J. Juhanoja, Chemical Vapor Deposition 8(4) (2002) 149153. [Ele03] Elers, K.-E., V. Saanila, W.-M. Li, M. Soininen, J. T. Kostamo, S. Haukka, J. Juhanoja and W. Besling, Thin Solid Films 434(1-2) (2003) 94-99 [Ell90] Elliott, S. R. Physics of Amorphous Materials: Second Edition. New York, John Wiley & Sons, (1990) [Ete93] Etemad-Rad, S. T. and E. Metcalfe, Fire and Materials 17 (1993) 33-37.

PAGE 393

384 [Fal98] Faltermeier, C. G C. Goldberg, M. Jones, A. Upham, A. Knorr, A. Ivanova, G. Peterson, A. E. Kaloyeros and B. Arkles, Journal of the Electrochemical Society 145(2) (1998) 676-682. [Fit95] Fitzgerald, S., USA Microscopy and Analysis (1995) 23-25. [Fix93] Fix, R., R. G. Gordon and D. M. Hoffman, Chemistry of Materials 5(5) (1993) 614-619 [Fle36] Fletcher, C. J.M. and G. K. Rollefson, Journal of the American Chemical Society 58 (1936) 2129-2135. [Fle98] Fleming, J. G., E. Roherty-Osmun, P. M. Smith, J. S. Custer, Y. D. Kim, T. Kacsich, M. A. Nicolet and C. J. Galewski, Thin Solid Films 320(1) (1998) 10-14. [Fou79] Foucaut, J. F. M., R., International Journal of Chemical Kinetics 11(7) (1979) 789-798. [Fri91] Frisk, K., Calphad: Computer Coupling of Phase Diagrams and Thermochemistry 15(1) (1991) 79-106. [Fri96] Frisk, K. and A. F. Guillermet, Journal of Alloys and Compounds 238 (1996) 167-179. [Fri98] Frisk K., Journal of Alloys and Compounds 278 (1998) 216-226. [Fri99a] Friedrich, L. J., S K. Dew, M. J. Brett and T. Smy, Journal of Vacuum Science Technology B 17(1) (1999) 186-193. [Fri99b] Frisk, K., Zeitschrift fi.ir Metallkunde 90(9) (1999) 704-711. [Fuk00] Fukiage, N. Semiconductor device and fabrication thereof. 2000 PCT Int. Appl., Japan, EPl 170784 [Gal93] Galanin, S. G., G.D. Kuzentsov, A. B. Tolstoguzov and E. Y. Chernyak, Physics and Chemistry of Materials Treatment 27(5) (1993) 489-493. [Gal97] Galewski, C. J. P. N. Gadgil, L. D Matthysse, C. A. Sans and V. S. Dharmadhikari (1997). Materials Research Society ULSI XII: 277-282. [Gal99] Galewski, C. and T. Seidel, European Semiconductor 21(1) (1999) 3132. [Gal00] Galewski, C. J., A. R. Londergan, C. A. Sans, T. E. Seidel, D. Clarke and Q. Zhu, European Semiconductor 22(Part 2) (2000) 28-33.

PAGE 394

[Gan92] [GanOO] [Gar90] [Gat96] [Geb66] [Gei86] [Gic83] [GirOO] [Gog98] [GogOO] [Gok99] [Gol63] [Gol67] [Gon02] [Goo86] [GorOO] 385 Ganem, J.-J., S. Rigo, I. Trimaille and G.-N. Lu, N[Uclear Instruments and Methods in Physics Research B64 (1992) 784 r788. Ganguli, S., L. Chen, T. Levine, B. Zheng and M. fhang, Journal of Vacuum Science Technology B 18 (2000) 237-24l Garnich, F., P. C. Yu and C. M. Lampert, Solar Energy Materials 20 (1990) 265-275. Gates, S. M., Chemical Reviews 96(4) (1996) 1519-1532. Gebhardt, E., E. Fromm and U. Roy, Zeitschrift flir Metallkunde 57(10) (1966) 732-736. Geissberger, A. E., R. A. Sadler, F. A. Leyenaar and M. L. Balzan, Journal of Vacuum Science Technology A 4(6) (1986) 3091-3094. Gicquel, A., M. P. Bergougnan and J. Amoroux (1983). Electrochemical Society Proceedings of the Symposium on Silicon Nitride Thin Insulating Films,Vol.83-8,,: 442-452. Giraudon, J.-M., P. L. Devassine, J.-F., L. Delannoy, L. Leclercq and G. Leclercq, Journal of Solid State Chemistry 154(2) (2000) 412-426. Gogova, D., K. Gesheva and A. Veneva, Materials Letters 35(5-6) (1998) 351-356. Gogova, D., K. Gesheva, A. Kakanakova-Georgieva and M. Surtchev, European Physical Journal 11 (2000) 167-174. Gokce, 0. H., S. Amin, N. M. Ravindra, D. J. Szostak, R. J. Paff, J. G. Fleming, C. J. Galewski, J. Shallenberger and R. Eby, Thin Solid Films 353(1,2) (1999) 149-156. Goldschmidt, H. J. and J. A. Brand, Journal of the Less-Common Metals 5 (1963) 181-194. Goldschmidt, H.J. Interstitial Alloys. New York, Plenum Press, (1967) Gonohe, N., Materials Transactions 43(7) (2002) 1585-1592. Goodman, C.H. L. and M. V. Pessa, Journal of Applied Physics 60(3) (1986) R65R81. Gordon, R. G., S. Barry, R. N. R. Broomhall-Dillard, V. A. Wagner and Y. Wang (2000). Mater. Res. Soc. Symp. Proc.,Vol.612: D9.12/1-

PAGE 395

386 D9.12/6. [Gru87] Grunert, W., E. S. Shpiro, R. Feldhaus, K. Anders, G. V. Antoshin and K. M. Minachev, Journal of Catalysis 107 (1987) 522-534. [Gub94] Gubanov, V. A., A. L. lvanovsky and V. P. Zhukov. Electronic Structure of Refractory Carbides and Nitrides. Cambridge, Cambridge University Press, (1994) [Gui93] Guillermet, A. F. and S. Jonsson, Zeitschrift ftir Metallkunde 84(2) (1993) 106-117. [Gui95] Guillermet, A. F., Journal of Alloys and Compounds 217(1) (1995) 6989. [Gup95] Gupta, D., Materials Chemistry and Physics 41(3) (1995) 199-205. [Gur89] Gurvich, L. V., I. V. Veyts and C. B. Alcock. Thermodynamic Properties of Individual Substances. New York, Hemisphere Pub. Corp., (1989) [Gus85] Gustafson, P., International Journal of Thermophysics 6 (1985) 395409. [Gus86] Gustafson, P., Materials Science and Technology 2 (1986) 653-658. [Hag30] Hagg, G., Z. physik. Chem. 7 Abt. B (1930) 339-362. [Hag54] Hagg, G. and N. Schonberg, Acta Crystallographica 7 (1954) 351-352. [Han03] Hand, A., Semiconductor International 26(5) (2003) 46-51. [HarOl] Hara, T. and K. Sakata, Electrochemical and Solid-State Letters 4(10) (2001) 077-079. [Hau93] Haukka, S., E.-L. Lakomaa and T. Suntola, Thin Solid Films 225(1-2) (1993) 280-283. [HauOO] Haukka, S., K.-E. Elers and M. Tuominen (2000). Materials Research Society Symposium Proceedings: Materials, Technology and Reliability for Advanced Interconnects and Low-k Dielectrics,Vol.612: D6.4/1-D6.4/6. [Hec02] Hecker, M., R. Hubner, R. Ecke, S. Schulz, H.-J. Engelmann, H. Stegmann, V. Hoffmann, N. Mattern, T. Gessner and E. Zschech, Microelectronic Engineering 64(1-4) (2002) 269-277. [Hei97] Heiser, T., S. McHugo, H. Heislmair and E. R. Weber, Applied Physics

PAGE 396

387 Letters 70(26) (1997) 3576-3578. [HilO0] Hillman, J., J. Hautala and S. Caliendo (2000). Advanced Metallization Conference San Diego, CA, Materials Research Society: 313-319. [Hir98] Hirata, A., K. Machida, S. Maeyama, Y. Watanabe and H. Kyuragi, Japanese Journal of Applied Physics Part 1 37(3B) (1998) 1251-1255. [Hir99] Hirata, A., K. Machida, N. Awaya, H. Kyuragi and M. Maeda, Japanese Journal of Applied Physics, Part 1 38(4B) (1999) 2355-2359. [HirOl] Hirata, A., T. Maeda, K. Machida, M. Maeda, K. Yasui and H. Kyuragi, Journal of Vacuum Science Technology B 19(3) (2001) 788-793. [Ho04] Ho, P., R. Rajagopalan, H. Chong, H. Chung and J. Yu, Semiconductor International 27(7) (2004) [Hom86] Homann, K. H. and J. V. Pidoll, Bunsen-Gesellschaft fur Physikalische Chemie 90(10) (1986) 847-854. [Hon00] Hones, P., M. Diserens, R. Sanjines and F. Levy, Journal of Vacuum Science Technology B 18(6) (2000) 2851-2856. [Hon03] Hones, P., N. Martin, M. Regula and F. Levy, Journal of Physics D: Applied Physics 36 (2003) 1023-1029. [Hos00] Hoshino, A., S. T., S. Hiiro, H. Machida, A. Ogura and Y. Ohshita (2000). Advanced Metallization Conference San Diego, CA, Materials Research Society: 403-408. [Hou65] Hou, K. C. and H. B. Palmer, Journal of Physical Chemistry 69(3) (1965) 858-862. [Hu98] Hu, C.-K. (1998). Materials Research Society Syposium Proceedings,Vol.511, Materials Research Society: 305-316. [Hua97] Huang, W., Zeitschrift fi.ir Metallkunde 88 (1997) 63-68. [Hua03] Huang, M., Study of TMGa-NH3-N2 systems using in situ Raman spectroscopy, PhD, University of Florida, Gainesville, FL,(2003) [Hub88] Huber, K. J. and C.R. Aita, Journal of Vacuum Science Technology A 6(3) (1988) 1717-1721. [Huo02] Huo J., R. Solanki and J. McAndrew, Journal of Materials Research 17(9) (2002) 2394-2398 [Hym98] Hymes, S., K. S. Kumar, S. P. Murarka, P. J. Ding, W. Wang and W. A.

PAGE 397

388 Lanford, Journal of Applied Physics 83(8) (1998) 4507-4512. [IBM97] IBM, IBM Research 35(4) (1997) [Iij95] Iijima, T ., Y. Shimooka and K. Suguro (1995). Materials Research Society Conference Proceedings,Vol.355,, Materials Research Society: 459-464. [lma97] lmahori, J., T. Oku and M. Murakami, Thin Solid Films 301(1-2) (1997) 142-148. [lnc90] lncropera, F. P. and D. P DeWitt. Fundamentals of Heat and Mass Transfer: Third Edition. New York, John Wiley & Sons, (1990) [Ing82] Ingrey, S., M. B. Johnson, R. W. Streater and G. I. Sproule, J. Vac. Sci. Technol. 20(4) (1982) 968-970. [Ino0l] Inoue, K., N. Ishikawa, I. Ohnuma, H. Ohtani and K. Ishida, ISU International 41(2) (2001) 175-182. [lnt89] Interrante, L. V., G. A. Sigel, M. Garbauskas, C. Hejna and G. A. Slack, Inorganic Chemistry 28(2) (1989) 252-257 [lst98] Istratov, A. A., C. Flink, H. Hieslmair, E. R. Weber and T. Heiser, Physical Review Letters 81(6) (1998) 1243-1246. [lst00] Istratov, A. A. and E. R. Weber (2000). Electrochemical Society Conference Proceedings, Vol.2000-27: 90-111. [ITR02] International Technology Roadmap for Semiconductors: Interconnect. Austin, TX, International SEMATECH, 2002.: 1-25. [lva99] lvanova, A. R., C. J. Galewski, C. A. Sans, T. E. Seidel, S. Grunow, K. Kumar and A. E. Kaloyeros (1999). Materials Research Society Symposium Proceedings; Advanced Interconnects and Contacts,Vol.564, San Francisco, CA: 321-326. [Jae88] Jaeger, R. C. Introduction to Microelectronic Fabrication. Reading, MA, Addison-Wesley Publishing Company, (1988) [Jai99] Jain, A., 0. Adetutu, B. Ekstrom, G. Hamilton, M. Herrick, R. Venkatraman and E. Weitzman (1999). Materials Research Society Symposium Proceedings, Vol.564: 269 280. [Jam92] Jama, C., 0. Dessaux, P. Goudmand, L. Gengembre and J. Grimblot, Surface and Interface Analysis 18 (1992) 751-756. [JCP88] Powder Diffraction File. JCPDS, International Center for Diffraction

PAGE 398

389 Data, Swarthmore, PA, 1982,1988. [Jeh80] Jehn, H. and P. Ettmayer Monatshefte fur Chemie (in German) 111 (1980) 1437-1440. [JelO0] Jelinek, M., J. Zemek, M. Trchova, V. Vorlicek, J. Lancok, R. Tomov and M. Simeckova, Thin Solid Films 366(1-2) (2000) 69-76. [Jel03] Jelenkovic, E. V., K. Y. Tong, W. Y. Cheung and S. P Wong, Semiconductor Science and Technology 18(6) (2003) 454-459. [Jen96] Jenkins, R. and R. Snyder. Introduction to X-ray Powder Diffractometry. New York, John Wiley & Sons, (1996) [Ji0l] Ji, H. X. and C. A. Amato-Wierda, Surface and Coatings Technology 148 (2001) 262-267. [Jin99] Jin, S., H. Li, H. Bender, I. Heyvaert and K. Maex, Institute of Physics Conference Series 164 (1999) 541-544 [Joh00a] Johnston, S. W., Chemical Vapor Deposition of Nitrides for Electronic Device Applications, PhD Dissertation, University of Florida, Gainesville,FL (2000a) [Joh00b] Johnston, S. W., C G Ortiz, 0. J. Bchir, Y. Zhang L. McElwee-White and T. J. Anderson, Chemical Vapor Deposition: CVD XV (15th), M. D. Allendorf and T. M. Besmann, Electrochemical Society, Pennington, NJ, 2000b. p. 268-276. [Joh0l] Johnson, K., B. Sauerhammer, S. Titmuss and D. A. King, Journal of Chemical Physics 114(21) (2001) 9539-9548. [Jon93] Jonsson S., Phase Relations in Quaternary Hard Materials, PhD Dissertation, Royal Institute of Technology, Stockholm, Sweden,(1993) [Jon96a] Jonsson, S., Zeitschrift ftir Metallkunde 87(9) (1996a) 713-720 [Jon96b] Jonsson, S., Zeitschrift fi.ir Metallkunde 87(10) (1996b) 788 795. [Jos02] Joseph S M. Eizenberg, C. Marcadal and L. Chen, Journal of Vacuum Science Technology B 20(4) (2002) 1471-1475. [Jos03] Josell D., D. Wheeler, C. Witt and T. P. Moffat, Electrochemical and Solid-State Letters 6(10) (2003) C143-C145. [Jun96] Jun, G.-C. S .L. Cho, K. B. Kim, H -K. Shin and D.-H. Kim (1996). Materials Research Society Symposium (Advanced Metallization for

PAGE 399

390 Future ULSI), Vol.427, San Francisco, CA, Materials Research Society: 349-354. [Jup97] Juppo, M., M. Ritala and M. Leskela, Journal of Vacuum Science & Technology, A: Vacuum, Surfaces, and Films 15(4) (1997) 2330-2333. [Kak99] Kakanakova-Georgieva, A., L. Kassamakova, T. Marinova, R. Kakanakov, 0. Noblanc, C. Amodo, S. Cassette and C. Brylinski, Appl. Surf. Sci. 151 (1999) 225-232. [Kal99] Kaloyeros, A. E., X. Chen, T. Stark, K. Kumar, S.-C. Seo, G. G. Peterson, H. L. Frisch, B. Arkles and J. Sullivan, Journal of the Electrochemical Society 146(1) (1999) 170-176. [Kal00] Kaloyeros, A. E. and E. Eisenbraun, Annual Review of Materials Science 30 (2000) 363-385. [Kan0l] Kang, S.-K., J. J. Kim, D. H. Ko, T. H. Ahn, I. S. Yeo, T. W. Lee and Y. H Lee (2001). Materials Research Society Symposium Proceedings, Vol.670: K5.9.l-K5.9.6. [Kan02] Kang, S.-K., J. J. Kim, B. G. Min, D.-H. Ko, H.B. Kang, C. W. Yang, T. H. Ahn, T. W. Lee and Y. H. Lee, Journal of the Korean Physical Society 40(1) (2002) 184-187. [Kap02] Kapur, P., J.P. McVittie and K. C. Saraswat, IEEE Transactions on Electron Devices 49(4) (2002) 590-597. [Kas94] Kasai, K., Y Akasaka, K. Nakajima, S. Suehiro K. Suguro, H. Oyamatsu, M. Kinugawa and M Kakumu (1994). IEEE Conference Proceedings IEDM 94,Vol. IEDM 94-497: 19.4.1-19.4.4. [Kat85] Kattelus, H.P., E. Kolawa, K. Affolter and M.A. Nicolet, J. Vac. Sci. Technol., A 3(6) (1985) 2246-54. [Kat94] Katrib, A., F. Hemming, L. Hilaire, P. Wehrer and G. Maire, Journal of Electron Spectroscopy and Related Phenomena 68 (1994) 589-595. [Kat95] Katrib, A., F. Hemming P. Wehrer, L. Hilaire and G. Maire, Journal of Electron Spectroscopy and Related Phenomena 76 (1995) 195-200. [Kat97] Kattner, U. R., JOM: The Journal of the Minerals, Metals and Materials Society 49(12) (1997) 14-19. [Kel91] Kelder, E. M ., Kinetics and Characterization of CVD-Grown BP, PhD

PAGE 400

391 Technical University of Delft, Delft, Netherlands,(1991) [Kel99] Kelsey, J. E., C. Goldberg, G. Nuesca, G. Peterson, A. E. Kaloyeros and B. Arkles, Journal of Vacuum Science & Technology B 17(3) (1999) 1101-1104. [Khi59] Khitrova, V. I. and Z. G. Pinsker, Kristallografiya 4 (1959) 545-53. [Kie51] Kiessling, R. and Y. H. Liu, Journal of Metals 3(Trans.) (1951) 639642. [Kil75] Kilbane, F. M. and P. S. Habig, Journal of Vacuum Science Technology 12(1) (1975) 107-109. [Kim91] Kim, Y. T. and S. K. Min, Applied Physics Letters 59(8) (1991) 929931. [Kim92] Kim, Y. T., C. W. Lee and S. K. Min, Applied Physics Letters 61(5) (1992) 537-539. [Kim93] Kim, Y. T., C. W. Lee and S. K. Min, Jpn. J. Appl. Phys. Part 1 Regul. Pap. Short Notes Rev. Pap. 32(12B) (1993) 6126-6131. [Kim94] Kim, Y. T. and C. W. Lee, Journal of Applied Physics 76(1) (1994) 542-545. [Kim97a] Kim, D. J., Y. T. Kim and J.-W. Park, Journal of Applied Physics 82(10) (1997a) 4847-4851. [Kim97b] Kim, Y. T., D. J. Kim, C. W. Lee and J.-W. Park (1997b). Proceedings of SPIE-The International Society for Optical Engineering (Multilevel Interconnect Technology), Vol.3214: 48-56. [Kim98a] Kim, D. J., Y. T. Kim and J. W. Park (1998). Advanced Interconnects and Contact Materials and Processes for Future Integrated Circuits,Vol.514, San Francisco, CA, Materials Research Society: 417422. [Kim98b] Kim, Y. T., C. W. Lee and D. J. Kim, Applied Physics Letters 72(12) (1998) 1507-1509. [Kim99a] Kim, D. J., Y. T. Kim and J.-W. Park, Journal of Vacuum Science & Technology B 17(4) (1999a) 1598-1601. [Kim99c] Kim, S. H., D.S. Chung, K. C. Park, K. B. Kim and S. H. Min, Journal of the Electrochemical Society 146(4) (1999c) 1455-1460. [Kim0l] Kim, I. G., N. S. Kim, J. H. Choy, B. H. Lee, Y. G. Sung, J. H. Kim, J.

PAGE 401

392 H. Cho, D. C. Kim, K. S. Kim, J. B. Choi, S. K. Choi, Y. W. Kweon, H. Y. Kwon, D. G. Park, J. S. Park and D. Y. Park (2001). IEEE: 28.5.128.5.4. [Kim02a] Kim, D. J., H. S Sim, S.-1. Kim, Y. T Kim and H. Jeon, Journal of Vacuum Science & Technology, A: Vacuum, Surfaces, and Films 20(1) (2002a) 194-197. [Kim02b] Kim, H., C. Cabral Jr. C. Lavoie and S. M. Rossnagel, Journal of Vacuum Science Technology B 20(4) (2002b) 1321-1326. [Kim03a] Kim, D.-H., Y. J. Kim, Y. S Song, B.-T. Lee, J. H. Kim, S. Suh and R. G. Gordon (2003a). Electrochemical Society Conference: Chemical Vapor Deposition XVI (CVD-VXI) and EUROCVD 14,Vol.2003-08, Paris, France, The Electrochemical Society: 491-502. [Kim03b] Kim H., (2003b) Atomic Layer Deposition of Metal and Nitride Thin Films: Current Research Efforts and Applications for Semiconductor Device Processing. Yorktown Heights, NY, IBM, (2003b) 121. [Kim03c] Kim, S. H., K. T. Nam, A. Datta, H. M. Kim, K. B. Kim and D. H. Kang, Journal of Vacuum Science Technology B 21(2) (2003c) 804813. [Kim03d] Kim, S.-H., S. S. Oh, K.-B. Kim, D.-H. Kang, W.-M. Li, S. Haukka and M Tuominen, Applied Physics Letters 82(25) (2003d) 4486-4488. [Kla00a] Klaus, J. W., S. J. Ferro and S. M. George, Journal of the Electrochemical Society 147(3) (2000a) 1175-1181. [Kla00b] Klaus, J. W., S. J. Ferro and S. M. George, Appl. Surf. Sci. 162-163 (2000b) 479-491. [Klu91] Kluwe, M., B. Kaimann, K. E. Lorber and H. Meier zu Kocker, Chemosphere 23(8-10) (1991) 1465-1471. [Kod94] Kodas T. T and M. J. Hampden-Smith, Eds. The Chemistry of Metal CVD,Weinheim, VCH (1994) [Koh95] Kohler, T., T. Frauenheim and G. Jungnickel, Physical Review B 52(16) (1995) 11837-11844. [Kol87] Kolawa, E., F. C. So, J. L. Tandon and M. A. Nicolet, Journal of the Electrochemical Society 134(7) (1987) 1759-1763

PAGE 402

393 [Kol91] Kolawa, E., P. J. Pokela, J. S. Reid, J. S. Chen and M. A. Nicolet, Appl. Surf. Sci. 53 (1991) 373-376. [Kos93] Kostikova, G. P. and Y. P. Kostikov, Inorganic Materials 29(8) (1993) 1005-1006. [Kou02] Koutzaki, S. H., J.E. Krzanowski and J. J. Nainaparampil, Metallurgical and Materials Transactions A 33A (2002) 1579-1588. [Kri0l] Krishnamoorthy, A., K. Chanda, S. P. Murarka, G. Ramanath and J. G. Ryan, Applied Physics Letters 78(17) (2001) 2467-2469. [Kwo95] Kwon, C. S., D. J. Kim, C. W. Lee, Y. T. Kim and 1.-H. Choi, Mater Res. Soc. Symp. Proc. 355(Evolution of Thin Film and Surface Structure and Morphology) (1995) 441-5. [Lai98a] Lai, K. K., A. W. Mak, T. Wendling, P. Jian and B. Hathcock, Thin Solid Films 332(1-2) (1998) 329-334. [Lai98b] Laidani, N., C. Dorigoni, A. Miotello, L. Calliari, G. Scarel and M. Sancrotti, Thin Solid Films 317 (1998) 477-480. [Lak79] Lakhtin, Y. M., Y. D. Kogan, T. M. Borovskaya and G. A. Solodkin, Metally (in Russian) 4 (1979) 176-184. [Lan62] Lang, R. J., Journal of the Acoustical Society of America 34(1) (1962) 6-8. [Lan00] Lanckmans, F., S. H. Brongersma, I. Varga, H. Bender, E. Beyne and K. Maex (2000). Materials Research Society Symposium Proceedings, Vol.612: Dl.4.1-Dl.4.6. [Las96] Laskin, A. and A. Lifshitz (1996) Twenty-Sixth Symposium (International) on Combustionffhe Combustion Institute, Vol.I: 669675. [Lau02] Laurila, T., K. Zeng, J. K. Kivilahti, J. Molarius and I. Suni, Journal of Applied Physics 91(8) (2002) 5391-5399. [Lec96] Leclercq, G. M. Kamal, J.M. Giraudon, P. Devassine, L. Feigenbaum, L. Leclercq, A. Frennet, J.M. Bastin, A. Lofberg, S. Decker and M. Dufour, Journal of Catalysis 158 (1996) 142-169. [Lee90] Lee, J. S., C. S. Park, J. Y. Kang, D. S. Ma and J. Y. Lee, Journal of Vacuum Science Technology B 8(5) (1990) 1117-1121.

PAGE 403

394 [Lee92] Lee, J. H., C. E. J. Hamrin and B. H. Davis, Catalysis Today 15(2) (1992) 223-241. [Lee93] Lee, C. W., Y. T Kim and S. K. Min, Applied Physics Letters 62(25) (1993) 3312-3314. [Lee94a] Lee, C. W., Y. T. Kim, C. Lee, J. Y. Lee, S. K. Min and Y. W. Park, Journal of Vacuum Science & Technology B 12(1) (1994a) 69-72. [Lee94b] Lee, C. W., Y. T. Kim and J. Y. Lee, Applied Physics Letters 64(5) (1994b) 619-621. [Lee94c] Lee, C. W., Y. T. Kim, S.-K. Min, C. Lee, J. Y. Lee and Y. W. Park, Mater. Res. Soc. Symp. Proc. 318(1nterface Control of Electrical, Chemical, and Mechanical Properties) (1994c) 335-40. [Lee95] Lee, C. W. and Y. T. Kim, Solid-State Electronics 38(3) (1995) 679682. [Lee97] Lee, D. H., C. K. Park, K. C. Song, Y. S. Jeon, T. H. Lee, C. Y. Jeong and J. Ahn, Japanese Journal of Applied Physics 36(Part 1, No. 12B) (1997) 7570-7574. [Lee98a] Lee, C.-Y. H., Tzuen-Hsi; Lu, Shing-Chi, Journal of Materials Science: Materials in Electronics 9(5) (1998a) 337-346 [Lee98b] Lee, H. G., S. Y. Lee, H. S. Moon, S. H. Kim, J. H. Sohn and J. Ahn, Japanese Journal of Applied Physics 37 (1998b) 6819-6823. [Lee98c] Lee, S., D. J. Kim, S.-H. Yang, J. Park, S. Sohn, K. Oh Y. T. Kim and J.-W. Park (1998c). Materials Research Society Symposium Proceedings, Vol.511: 347-352. [Lee99] Lee, Y.-J., B.-S. Suh, M. S. Kwon and C.-O. Park, Journal of Applied Physics 85(3) (1999) 1927-1934. [Lee00a] Lee, J.-H., J.-Y. Kim, S.-W. Rhee, D. Y. Yang D.-H. Kim, C -H. Yang, Y.-K. Han and C.-J. Hwang, Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films 18(5) (2000) 2400-2403. [LeeOOb] Leedy, K. D., M J O'Keefe, J G. Wilson, R. Osterday and J. T. Grant (2000). Materials Research Society Conference Proceedings, Vol.612, San Francisco, CA: D9.20.l-D9.20.6. [Lee0la] Lee, B.-J ., Metallurgical and Materials Transactions A 32A(10) (2001a) 2423-2439.

PAGE 404

395 [Lee0lb] Lee, C. W., Y. T. Kim, H. S. Lee, Y. K. Park, T.-H. Lee, Q. Chen and N. V. Richardson, Journal of the Korean Physical Society 39(6) (2001b) 1019-1022. [Lee0lc] Lee, J. W., C.H. Han, J. S. Park and J. W. Park, Journal of the Electrochemical Society 148(3) (2001c) G95-G98. [Lee03] Lee, K. S., Japanese Journal of Applied Physics 42(Part I, No. 6A) (2003) 3368-3371. [Les02] Leskela, M. and M. Ritala, Thin Solid Films 409(1) (2002) 138-146. [Lev98] Levine, T. E., L. Chen, S. Ganguli, A. Mak, J. Tseng and M. Chang (1998). Advanced Metallization and Interconnect Systems for ULSI Applications in 1997,Vol.ULSI XIII, San Diego, CA, MRS: 95-101. [Lew94] Lewkebandara, T. S., P.H. Sheridan, M. J. Heeg, A. L. Rheingold and C.H. Winter, Inorganic Chemistry 33(25) (1994) 5879-5889. [LiOO] Li, H., S. Jin, H. Bender, F. Lanckmans, I. Heyvaert, K. Maex and L. Froyen, Journal of Vacuum Science Technology B 18 (2000) 242-251. [Li02] Li, W. M., K. Elers, J. Kostamo, S. Kaipio, H. Huotari, M. Soininen, P. J. Soininen, M. Tuominen, S. Haukka, S. Smith and W. Besting (2002). IEEE International Interconnect Technology Conference, 5th Burlingame, CA, United States: 191-193. [Li03] Li, W.-M., M. Tuominen, S. Haukka, H. Sprey and I. J. Raaijmakers, Solid State Technology 46(7) (2003) 103-104,106. [Lid85] Lide, D. R., Ed. JANAF Thermochemical Tables, Third Edition. New York, American Institute of Physics ( 1985) [Lif96] Lifshitz, A., Y. Cohen, M. Braun-Unkhoff and P. Frank (1996). Twenty-Sixth Symposium (International) on Combustion, The Combustion Institute: 659-667. [Lim03] Lim, B. S., A. Rahtu and R. G. Gordon, Nature Materials 2(11) (2003) 749 -754. [Lin90] Lin, D., B. Yan and W Yu (1990). Materials Research Society Symposium Proceedings: Thin Film Struct. Phase Stab.,Vol.187: 161166.

PAGE 405

396 [Lin98a] Lin, J., A. Tsukune, T. Suzuki and M. Yamada (1998a). Advanced Metallization and Interconnect Systems for ULSI Applications in 1997,Vol.ULSI XID, San Diego, CA, MRS: 693-697. [Lin98b] Lin, J., A. Tsukune, T. Suzuki and M. Yamada, Journal of Vacuum Science & Technology, A: Vacuum, Surfaces, and Films 16(2) (1998b) 611-614. [Lin99] Lin, J., A. Tsukune, T. Suzuki and M. Yamada, Journal of Vacuum Science & Technology, A: Vacuum, Surfaces, and Films 17(3) (1999) 936-938. [Lip03] Lipsanen, A., P. Kuivalainen, E. Rauhala, A. Seppala and S. Franssila, Journal of Materials Science: Materials in Electronics 14(5/6/7) (2003) 315-318. [Liu88] Liu, A. C. and C M Friend, Journal of Chemical Physics 89(7) (1988) 4396-4405. [Liu98] Liu, Z. Q., W. J. Wang, T. M. Wang, S. Chao and S. K. Zheng, Thin Solid Films 325 (1998) 55-59. [Liu00] Liu, P. T ., T. C. Chang, J.C. Hu, Y. L. Yang and S M. Sze, Journal of the Electrochemical Society 147(1) (2000) 368-372. [Llo95] Lloyd J. R. and J. J. Clement, Thin Solid Films 262(1-2) (1995) 135141. [Lu98a] Lu, J.P., W Y. Hsu, Q. Z. Hong, G. A. Dixit, V. T. Cordasco, E. M. Zielinski, J. D. Luttmer, R. H. Havemann, L. K. Magel and H. L. Tsai (1998a). Materials Research Society Conference Proceedings,Vol.514: 507 -512 [Lu98b] Lu, J.P., W. Y. Hsu, J. D Luttmer, L. K. Magel and H. L. Tsai, Journal of the Electrochemical Society 145(2) (1998b) L21-L23. [Luc96] Lucy, T. E., Synthesis, Characterization and Reactivity of Tungsten Oxynitride, Master's Thesis, Virginia Polytechnic Institute and State University, Blacksburg,(1996) [Luo94] Luo, Y.-R. and J. L. Holmes, Journal of Physical Chemistry 98(1) (1994) 303-12. [Lut0l] Luthin, J. and C. Linsmeier Journal of Nuclear Materials 290-293 (2001) 121 125.

PAGE 406

397 [Mar93] Marcus, S. D. and R. F. Foster, Thin Solid Films 236(1-2) (1993) 330333. [Mar95] Maruyama, T. and T. Morishita, Journal of Applied Physics 78(6) (1995) 4104-4107. [Mas96] Masel, R. I. Principles of Adsorption and Reaction on Solid Surfaces. New York, John Wiley & Sons, (1996) [Mat94] Matsuhashi, H. and S. Nishikawa, Japanese Journal of Applied Physics 33(3A) (1994) 1293-1297. [Mat02] Matsui, Y., T. Hiratani, T. Nabatame, Y. Shimamoto and S. Kimura, Electrochemical and Solid-State Letters 5(1) (2002) C18-C21. [May90] Mayer, J. W. and S. S. Lau. Electronic Materials Science: For Integrated Circuits in Si and GaAs. New York, Macmillan Publishing Company, (1990) [McG73] McGuire, G. E., G. K. Schweitzer and T. A. Carlson, Inorganic Chemistry 12(10) (1973) 2450-2453. [Mer97] Mercier, M., S. Weber, A. Jacques, H. Hirabayashi, H. Ohkawa and M. Kinoshita, Diffusion and Defect Data--Solid State Data, Pt. A: Defect and Diffusion Forum 143-147(Pt. 2, Diffusion in Materials, Pt. 2) (1997) 1285-1290. [MerOl] Merchant, S. M., S. H. Kang, M. Sanganeria, B. van Schravendijk and T. Mountsier,Copper Interconnects for Semiconductor Devices, Novellus (2001 ), http://www.novellus.com/damascus/tec/tec_24.asp [Meu98] Meunier, C., C. Monteil, C. Savall, F. Palmino, J. Weber, R. Berjoan and J. Durand, Appl. Surf. Sci. 125 (1998) 313-320. [Mig0l] Migita, T., R. Kamei, T Tanaka and K. Kawabata, Appl. Surf. Sci. 169-170 (2001) 362-365. [Min96] Minarnihaba, G., T. Iijima, Y. Shimaoka, H. Tamura, T. Kawanoue, H. Hirabayashi, N. Sakurai, H. Ohkawa, T. Obara, H. Egawa, T. ldaka, T. Kubota, T. Shimizu, M. Koyama, J. Ooshima and K. Suguro, Japanese Journal of Applied Physics Part 1: Regular Papers, Short Notes & Review Papers 35(2B) (1996) 1107-1110. [Min00] Min, J.-S., J.-S. Park, H.-S. Park and S -W. Kang, Journal of the Electrochemical Society 147(10) (2000) 3868-3872.

PAGE 407

398 [Mon99] Monteiro, 0. R., M. P. Delplancke-Ogletree and I. G. Brown, Thin Solid Films 342 (1999) 100-107. [Mot98] Mori, H., J. Imahori, T. Oku and M. Murakami (1998). AIP Conference Proceedings, Vol.418 (Stress Induced Phenomena in Metallization): 475-480. [Mou95] Moulder, J. F., W. F. Stickle, P. E. Sobol and K. D. Bomben Handbook of X-ray Photoelectron Spectroscopy. Eden Prairie, MN, Physical Electronics, Inc. (1995) [Mro0l] Mrozek, M. F., S. A. Wasileski and M. J. Weaver, Journal of the American Chemical Society 123(51) (2001) 12817-12825. [Muk93] Mukerji, J., Chemistry of Advanced Materials: A 'Chemistry for the 21st Century' monograph, C. N. R. Rao, Blackwell Scientific Publications, Oxford, 1993. p. 169-196. [Mut95] Murarka, S. P. and S. W. Hymes Crit. Rev. Solid State Mat. Sci. 20(2) (1995) 87-124. [Nag93] Nagai, M. and K. Kishida, Appl. Surf. Sci. 70-1 (1993) 759-762. [Nag94] Nagaoka, M., K. Ishida, T. Matsunaga, K. Nishihori, T. Hashimoto, M. Yoshimura, Y. Tanabe, M. Mihara, Y. Kitaura and N. Uchitomi, Japanese Journal of Applied Physics 33(Part 1, No. IB) (1994) 767770. [Nag00] Nagai, M., N Hirano and S. Omi, Japanese Journal of Applied Physics 39(Part 1, No. 7B) (2000) 4558-4560. [Nak87] Nakajima, T., K. Watanabe and N. Watanabe, Journal of the Electrochemical Society 134(12) (1987) 3175-3178. [Nak97] Nakajima, K., Y. Akasaka, K. Miyano, M. Takahashi, S. Suehiro and K. Suguro, Appl. Surf. Sci. 117/118 (1997) 312-316. [Neu34] Neumann, B., C. Kroger and H. Kunz, Zeitschrift fiir anorganische und allgemeine Chemie 218 (1934) 379-401. [Ngu89] Nguyen, T. P. and S. Lefrant, Journal of Physics: Condensed Matter 1(31) (1989) 5197-5204. [Nic78] Nicolet, M.A., Thin Solid Films 52(3) (1978) 415-443. [Nic95] Nicolet, M.-A., Appl. Surf. Sci. 91(1-4) (1995) 269-276. [Nie83] Nielson, A. J. and J.M. Waters, Aust. J. Chem. 36(2) (1983) 243-251.

PAGE 408

399 [Nin02] Ning, J., S. Xu, J. W. Chai, L. K. Chen, A. See and J. Ahn, International Journal of Modem Physics B 16(6&7) (2002) 1127-1131. [NIS03] NIST Chemistry Webbook: Standard Reference Database Number 69, (2003 ), http:/ /webbook. nist. gov /chemistry [NorOl] Norman, J. A. T., Journal de Physique IV: Proceedings ll(Pr3, Thirteenth European Conference on Chemical Vapor Deposition, 2001) (2001) Pr3/497-Pr3/503. [Nov00] Novellus, !.,Benefits of Copper :Copper Technology is Here Today in Working Devices, Novellus, Inc. (2000), http://www.novellus.com/ damascus/boc/boc.asp [Nug80] Nugent, W. A. and R. L. Harlow, Inorganic Chemistry 19(3) (1980) 777-779. [Ohr92] Ohring, M. The Materials Science of Thin Films. Boston, Academic Press, Inc., (1992) [Ole97] Olefjord, I., Surface Characterization: A User's Sourcebook, D. Brune, R. Hellborg, H.J. Whitlow and 0. Hunderi, Wiley-VCR, Weinheim, 1997. p. 291-319. [Pac91] Paccagnella, A., A. Callegari, A. Camera, M. Gasser, E. Latta, M. Murakami and M Norcott, Journal of Applied Physics 69(4) (1991) 2356-2364. [Pan96] Pangborn, A. B., M. A. Giardello, R. H. Grubbs, R. K. Rosen and F. J. Timmers, Organometallics 15(5) (1996) 1518-1520. [Par95a] Park, K. C. and K. B. Kim (1995). Materials Research Society Symposium Proceedings, Vol.391, Materials Research Society: 211216. [Par95b] Paranjpe, A. and M. IslamRaja, Journal of Vacuum Science Technology B 13(5) (1995) 2105-2114. [Par96] Park, K. C., K. B. Kim, I. J. Raaijmakers and K. Ngan, Journal of Applied Physics 80(10) (1996) 5674-5681. [Par97a] Park, B. L., J.M. Ha, D. H. Ko, B. Y. Yoo, S. I. Lee, H. D. Lee, M. B. Lee, U. I. Chung, Y. B. Koh and M. Y. Lee (1997a). Materials Research Society ULSI XII: 283-288.

PAGE 409

400 [Par97b] Park, B. L., D. H. Ko, Y. S. Kim, J.M. Ha, Y. W. Park, S. I. Lee, H. D. Lee, M. B. Lee, U. I. Chung, Y. B. Koh and M. Y. Lee, Journal of Electronic Materials 26(2) (1997b) Ll-L5. [Par98a] Park, B. L., M. B. Lee, K. J. Moon, H. D. Lee, H. K. Kang and M. Y. Lee (1998a). Advanced Metallization and Interconnect Systems for ULSI Applications in 1997,Vol.ULSI XID, San Diego, CA, MRS: 163167. [Par98b] Park, S. H., I. H. Kim, T. W. Lee and M. P. Park (1998b). IEEE,Vol. 156, Institute of Physics, Great Britain, Adam Hilger, Ltd: 459-462. [ParO0] Park, Y. K., S.-1. Kim, Y. T. Kim and C. W. Lee, Journal of the Korean Physical Society 37(3) (2000) 324-327. [Pau87] Pauleau, Y., Solid State Technology (February 1987) (1987) 61-67. [Pau92] Pauleau, Y. and P. Gouy-Pailler, Journal of Materials Research 7(8) (1992) 2070-2079. [Pea67] Pearson, W. B. A Handbook of Lattice Spacings and Structures of Metals and Alloys. Oxford, Pergamon Press, (1967) [Pea86a] Pearlstein, K. A. and C. M. Friend, Journal of Physical Chemistry 90(18) (1986a) 4341-4343. [Pea86b] Pearlstein, K. A. and C. M. Friend, Journal of Physical Chemistry 90(18) (1986b) 4344-4347. [Pec97] Pecz, B., A. Sulyok, G. Radnoczi, 0. Noblanc, C. Amodo, S. Cassette and C. Brylinski (1997). Microscale Semiconductor Materials Conference, Vol.157, Oxford, England, IOP Publishing: 519-522. [Ped82] Pederson, S. F. and R.R. Schrock, Journal of the American Chemical Society 104 (1982) 7483-7491. [Pet02] Peters, L., Semiconductor International 25(9) (2002) 17. [Pet03] Peters, L., Semiconductor International 26(3) (2003) 50-52,54. [Phi92] Philipossian, A., Journal of the Electrochemical Society 139(10) (1992) 2956-2961. [Pie96] Pierson, H. 0. Handbook of Refractory Carbides and Nitrides: Properties, Characteristics, Processing and Applications. Westwood, NJ, Noyes Publications, (1996)

PAGE 410

401 [Pok91] Pokela, P. J., C. K. Kwok, E. Kolawa, S. Raud and M.A. Nicolet, Appl. Surf. Sci 53 (1991) 364-72. [Pri95] Prieto, P. and R. E. Kirby, Journal of Vacuum Science Technology A 13(6) (1995) 2819-2826. [Qin98] Qin, W., W. Shih, J. Li, W. James, H. Siriwardane and P. Fraundorf (1998) Materials Research Society Symposium Proceedings : Nanostructured Powders and Their Industrial Applications,Vol.520,, Materials Research Society: 217-222. [Raa93] Raaijmakers, I. J. and J Yang, Appl. Surf. Sci. 73 (1993) 31-41. [RamOO] Ramberg, C. E., E. Blanquet, M. Pons, C Bernard and R. Madar, Microelectronic Engineering 50 (2000) 357-368. [Raz95] Razeghi, M The MOCVD Challenge. Volume 2: A survey of GalnAsP-GaAs for photonic and electronic device applications. Bristol, Institute of Physics Publishing, (1995) [Rei75] Reichelt K. and G. Bergmann, Journal of Applied Physics 46(6) (1975) 2747-2751. [Rei92] Reid, J. S., E. Kolawa and M.-A. Nicolet, Journal of Materials Research 7(9) (1992) 2424-2428. [Rei94] Reid, J. S., X. Sun E. Kolawa and M.-A. Nicolet, IEEE Electron Device Letters 15(8) (1994) 298-300. [Rei95] Reid, J. S., R. Y. Liu, P. M Smith, R. P Ruiz and M. A. Nicolet, Thin Solid Films 262(1-2) (1995) 218-223. [ReiOO] Reid, J. D.,Fundamentals of Copper:Fundamentals of Electroplating, Novellus Systems (2000), http://www.novellus.com/damascus/foc/foc_e asp [Ren96] Reniers F., E. Silberberg, N. Roose and J. Vereecken, Appl. Surf. Sci 99(4) (1996) 379-392. [Rey02] Reyes-Betanzo, C. S. A Moshkalyov, A C. Ramos, J. A. Diniz and J. W. Swart, Journal of the Electrochemical Society 149(3) (2002) G 179-0183. [Rig75] Riggs, W M and M J. Parker, Methods and Phenomena: Their Applications in Science and Technology !(Methods Surf. Anal.) (1975) 103-158.

PAGE 411

402 [Rit90] Ritter, E. and J. W. Bozzelli, Hazardous Waste & Hazardous Materials 7(1)(1990) 103-115. [Rit03] Ritala, M. and M. Leskela (2003). Electrochemical Society Conference: Chemical Vapor Deposition XVI (CVD-VXI) and EUROCVD 14,Vol.2003-08, Paris, France: 479-490. [Ros0l] Rossnagel, S. M. and H. Kim, IEEE (2001) 3-5. [Rud67] Rudy, V. E. and J. R. Hoffman, Planseeberichte fuer Pulvermetallurgie 15(3) (1967) 174-178. [Rus99] Rusli, S. F. Yoon, H. Yang, J. Ahn, Q. F. Huang, Q. Zhang, Y. P. Guo, C. Y. Yang, E. J. Teo, A. T. S. Wee and A. C.H. Huan, Thin Solid Films 355-356 (1999) 174-178. [Rye69] Rye, R. R. and R. S. Hansen, Journal of Chemical Physics 50(8) (1969) 3585-3595. [Sak96] Sakata, Y., A. Muto, M. A. Uddin and K. Harino, Journal of Materials Chemistry 6(7) (1996) 1241-1243. [Sam80] Samsonov, G. V. and I. M. Vinitskii. Handbook of Refractory Compounds. New York, IFI/Plenum Data Company, (1980) [Sar65] Sara, R. V., Journal of the American Ceramic Society 48(5) (1965) 251-257. [Sar80] Sarma, D. D. and C. N. R. Rao, Journal of Electron Spectroscopy and Related Phenomena 20 (1980) 25-45. [Sau98] Saunders, N. and A. P. Miodownik. Calphad (Calculation of Phase Diagrams): A Comprehensive Guide. Oxford, Pergamon Press, (1998) [Sch54] Schonberg, N., Acta Chemica Scandinavica 8(2) (1954) 204-207. [Sch87] Schlotter, P., Solar Energy Materials 16(1-3) (1987) 39-46. [Sch0l] Scheuermann, F., E. Salzborn, F. Hagelberg and P. Scheier, Journal of Chemical Physics 114(22) (2001) 9875-9881. [Sel95] Sellem, S., C. Potvin, J.M. Manoli, R. Contant and G. Djega Mariadassou, Journal of the Chemical Society, Chemical Communications 3 (1995) 359-360. [SerOl] Serguchev, Y. A. and Y. V. Belokopytov, Kinetics and Catalysis 42(2) (2001) 174-181.

PAGE 412

403 [Sha89] Shamir, N., J. C. Lin and R. Gomer, Surface Science 214 (1989) 74-84. [Sha95] Shacham-Diamand, Y., V. Dubin and M. Angyal, Thin Solid Films 262 (1995) 93-103. [Sha96] Shackleford, J. F. Introduction to Materials Science for Engineers. Upper Saddle River, Prentice-Hall, (1996) [Sha97] Shacham-Diamand, Y. and V. Dubin, Microelectronic Engineering 33 (1997) 47-58. [Sha0l] Shaw, M. J., S. Grunow and D. J. Duquette, Journal of Electronic Materials 30(12) (2001) 1602-1608. [She00a] Shen, Y. G. and Y. W. Mai, Appl. Surf. Sci. 167 (2000a) 59-68. [She00b] Shen, Y. G. and Y. W. Mai, Materials Science and Engineering A288 (2000b) 47-53. [She00c] Shen, Y. G. and Y. W. Mai, Surface and Coatings Technology 127 (2000c) 239-246. [She00d] Shen, Y. G., Y. W Mai, W. E. McBride, Q. C. Zhang and D.R. McKenzie, Thin Solid Films (372) (2000d) 257-264. [She00e] Shen, Y. G., Y. W. Mai, D R. McKenzie, Q C. Zhang, W. D. McFall and W. E. McBride, Journal of Applied Physics 88(3) (2000e) 13801388. [She00f] Shen, Y. G., Y. W. Mai, Q. C. Zhang, D.R. McKenzie, W. D. McFall and W. E. McBride, Journal of Applied Physics 87(1) (2000f) 177-187. [Shi97] Shimooka, Y., T. Iijima, S. Nakamura and K. Suguro, Japanese Journal of Applied Physics, Part 1: Regular Papers, Short Notes & Review Papers 36(3B) (1997) 1589-1592. [Sim94] Simson, S., N. Heide and J. W. Schultze, Surface and Interface Analysis 22 (1994) 431 435. [Sim03] Sim, H. S S.-1. Kim and Y. T. Kim, Journal of Vacuum Science Technology B 21(4) (2003) 1411-1414. [Sin02a] Singer, P., Semiconductor International 25(5) (2002a) 46-53. [Sin02b] Singer, P., Semiconductor International 25(6) (2002b) 48-52.

PAGE 413

404 [Siv95] Sivaram, S. C. V. D T. a. P. D. o. E. M., 119, (1995). Chemical Vapor Deposition: Thermal and Plasma Deposition of Electronic Materials. New York, Van Nostrand Reinhold International Thomson Publishing, (1995) [Slo99] Sloan, D. W. P. M. Blass and J. M. White, Appl. Surf. Sci. 143 (1999) 142-152. [Sly61] Slysh, R. S. and C.R. Kinney, Journal of Physical Chemistry 65 (1961) 1044-1045. [Smi82] Smith, W. R. and R. W. Missen. Chemical Reaction Equilibrium Analysis: Theory and Algorithms. New York, John Wiley & Sons, (1982) [Smi02] Smith, S ., W.-M. Li, K.-E. Elers and K. Pfeifer, Microelectronic Engineering 64(1) (2002) 247-253. [So88] So, F. C. T., E. Kolawa, X. A. Zhao, E.T. S. Pan and M.A. Nicolet, J. Appl. Phys. 64(5) (1988) 2787-9. [Som97] Somatri-Bouamrane, R., N. Chevarier, A Chevarier, A. M. Dutron, E. Blanquet and R. Madar (1997). Proceedings of the Electrochemical Society, Vol.97-25 (Chemical Vapor Deposition),, The Electrochemical Society: 741-748. [Sot03] Soto, G., W. de la Cruz, F. F. Castillon, J A. Diaz, R. Machorro and M. H. Farias, Appl. Surf. Sci. 214(1-4) (2003) 58-67. [Sri84] Srivastava, P. K., T. V Rao, V D Vankar and K. L. Chopra, Journal of Vacuum Science Technology A 2(3) (1984) 1261-1265 [Suh99] Suh, B. S. Y. J. Lee, J. S. Hwang and C. 0. Park, Thin Solid Films 348 (1999) 299-303 [Suh0l] Suh, B. S., H.K. Cho, Y. J. Lee, L. W.J. and C. 0. Park Journal of Applied Physics 89(7) (2001) 4128-4133. [Sun92] Suntola T ., Thin Solid Films 216(1) (1992) 84-89. [Sun93] Suntola, T., Thin Solid Films 225 (1993) 96-98. [Sun95] Sundman B., (1995) Thermo-Cale: User's Guide. Stockholm, Department of Materials Science and Engineering, Royal Institute of Technology, (1995) 300.

PAGE 414

405 [Sun98a] Sun, B., T. Chiang, P. Ding and B. Chin (1998a). Advanced Metallization and Interconnect Systems for ULSI Applications in 1997,Vol.ULSI Xill, San Diego, CA, MRS: 137-141. [Sun98b] Sun, S. C. (1998b). AIP Conference Proceedings, 418 (Stress Induced Phenomena in Metallization) Tokyo, Japan, The American Institute of Physics: 451-456. [Sun0la] Sun, S. C., H. Y. Tsai and J. Wang (2001a). Advanced Metallization Conference Montreal, Canada, Materials Research Society: 587-591. [Sun0lb] Sun, Y. M., S. Y. Lee, A. M. Lemonds, E. R. Engbrecht, S. Veldman, J. Lozano, J.M. White, J. G. Ekerdt, I. Emesh and K. Pfeifer, Thin Solid Films 397 (2001 b) 109-115. [Suz98] Suzuki, T., A. Tsukune, 0. Tsuboi and M. Yamada (1998). Advanced Metallization and Interconnect Systems for ULSI Applications in 1997,Vol.ULSI Xill, San Diego, CA, MRS: 49-55. [Swa72] Swalin, R. A. Thermodynamics of Solids. New York, John Wiley & Sons, (1972) [Sze84] Szekely, A., R. K. Hanson and C. T. Bowman, Journal of Physical Chemistry 88(4) (1984) 666-668. [Tak96] Takagi, M. T., K. Miyashita, H. Koyama, K. Nakajima, K. Miyano, Y. Akasaka, Y. Hiura, S. Inaba, A. Azuma, H. Koike, M. Yoshimura, K. Suguro and H. Ishiuchi (1996). Technical Digest International Electron Devices Meeting: 455-458. [Tak97a] Takeyama, M. and A. Noya, Japanese Journal of Applied Physics 36(4A) (1997a) 2261-2266. [Tak97b] Takeyama, M., A. Noya and T. Fukuda, Journal of Vacuum Science & Technology, A: Vacuum, Surfaces, and Films 15(2) (1997b) 415-420. [Tak02] Takahashi, S., K. Tai, H. Ohtorii, N. Komai, Y. Segawa, H. Horikoshi, Z. Yasuda, H. Yamada, M. Ishihara and T. Nogami (2002). Digest of Technical Papers: 2002 Symposium on VLSI Technology Honolulu, Ill: 32-33. [Tim0l] Time Domain CVD,Fundamentals of Chemical Vapor Deposition, Time Domain CVD Inc. (2001), www.timedomaincvd.com [Tot71] Toth, L. E Transition Metal Carbides and Nitrides. New York, Academic Press, (1971)

PAGE 415

406 [Tra03] Traving, M., G. Schindler, G. Steinlesberger, W Steinhogl and M. Engelhardt, Semiconductor International 26(8) (2003) 73-78. [Tsa86] Tsai, C. C. J.C. Knights G. Chang and B. Wacker, Journal of Applied Physics 59(8) (1986) 2998-3001. [Tsa95] Tsai, M. H. S. C. Sun, C. P. Lee, H T. Chiu, C. E. Tsai S. H. Chuang and S. C. Wu, Thin Solid Films 270(1-2) (1995) 531-536. [Tsa96] Tsai, M. H S. C. Sun, H. T. Chiu and S H. Chuang, Applied Physics Letters 68(10) (1996) 1412-1414. [Uch86] Uchitomi M., M. Nagaoka, K. Shimada, T Mizoguchi and N. Toyoda Journal of Vacuum Science Technology B 4( 6) ( 1986) 1392-1397. [Uch97] Uchibori, C. J., Y. Ohtani and T. Oku, Journal of Electronic Materials 26(4) (1997) 410-414 [Uek96] Uekubo M., T. Oku K. Nii, M. Murakami, K. Takahiro, S. Yamaguchi, T Nakano and T. Ohta, Thin Solid Films 286 (1996) 170 175 [VelO0] Veldman S., A. M. Lemonds K. Kershen Y.-M. Sun, I. Emesh, K. Pfeifer J.M. White and J. G. Ekerdt (2000) Advanced Metallization Conference San Diego, CA, Materials Research Society: 307-312. [Vie02] Vie ira, M. T., A. Cavaleiro and B. Trindade Surface and Coatings Technology 151-152 (2002) 495-504 [Vij99] Vijayendran, A. and M. Danek (1999). Materials Research Society Conference Proceedings Vol.564,, Materials Research Society: 327334 [Vla195] Vlakhov E S., K. A. Gesheva G. I. Stoyanov and G.D. Beshkov Electronic & Electrical Engineering Research Studies Electronic Materials Series 1994, l(Electronic, Optoelectronic and Magnetic Thin Films) (1995) 456-459. [Vol85] Volpe L. and M. Boudart Journal of Solid State Chemistry 59(3) (1985) 332-347. [Vos78] Vossen J. L. and W. Kem, Eds. Thin Film Processes .New York, Academic Press (1978) [Vu90] Vu, Q T ., P J. Pokela C. L. Garden E. Kolawa, S. Raud and M.-A. Nicolet, Journal of Applied Physics 68(12) (1990) 6420-6423.

PAGE 416

407 [Wag90] Wagner, C. D. Practical Surface Analysis. Volume 1: Auger and X-ray Photoelectron Spectroscopy. New York, John Wiley and Sons, (1990) [Wan94] Wang, S. Q., MRS Bulletin 19(8) (1994) 30-40. [Wan00] Wang, D. Z., J. X. Chen, H. L. Zhang and N. K. Huang, Nuclear Instruments and Methods in Physics Research B 171 (2000) 465-469. [Wan0la] Wang, S. J., H. Y. Tsai and S. C. Sun, Journal of the Electrochemical Society 148(8) (2001a) C563-C568. [Wan0lb] Wang, S. J., H. Y. Tsai, S. C. Sun and M. H. Shiao, Journal of the Electrochemical Society 148(9) (2001b) G500-G506. [Wan03] Wang, Z., T. Ida, H. Sakaue, S. Shingubara and T. Takahagi, Electrochemical and Solid-State Letters 6(3) (2003) C38-C41. [Wei99] Wei, J., P. Hing and Z. Q. Mo, Wear 225-229 (1999) 1141-1147. [Whi58] White, W. B., S. M. Johnson and G. B. Dantzig, Journal of Chemical Physics 28 (1958) 751-755. [Whi91] White, F. M. Viscous Fluid Flow: Second Edition. New York, McGraw-Hill Inc., (1991) [Wit80] Wittmer, M., Applied Physics Letters 36 (1980) 456-458. [Wol86] Wolf, S. and R. N. Tauber. Silicon Processing for the VLSI Era. Sunset Beach, CA, Lattice Press, (1986) [Wri89] Wriedt, H. A., Bulletin of Alloy Phase Diagrams 10(4) (1989) 358-367. [Xue98] Xue, Z., S. H. Chuang, K G. Caulton and M. H. Chisholm, Chemistry of Materials 10 (1998) 2365-2370. [Yan02] Yang, J.-M., 1.-S. Choi, Y. S. Kim, J -C. Park, S.-M. Lee, T.-E. Hong, Y.-B Park, S.-Y. Lee and Y. W. Kim, Journal of Applied Physics 91(12) (2002) 9788 9793. [Yeh96] Yeh, W. K., K. Y. Chan, T. C. Chang, M. H Tsai, S. H Chen, M. C. Chen and M. S Lin, Japanese Journal of Applied Physics 35(Part 1, No. 2B) (1996) 1115-1119. [You89] Young, C. M. and K. J. Voorhees, Preprints of papers -American Chemical Society, Division of Gas and Fuel Chemistry 34(1) (1989) 280-287.

PAGE 417

408 [Yu88] Yu, K. M., J.M. Jaklevic, E. E. Haller, S. K. Cheung and S. P Kwok, Journal of Applied Physics 64(3) (1988) 1284-1291. [Zen00a] Zeng, Y., L. Chen and T L. Alford (2000a). Materials Research Society Conference Proceedings, Vol.612, Materials Research Society: D9. l l.1-D9.l l.6. [Zen00b] Zeng Y., S. W Russell A. J. McKerrow L. Chen and T. L. Alford Journal of Vacuum Science Technology B 18(1) (2000b) 221-230. [Zha94] Zhang, S. C D. Qin and L. C. Lim, International Journal of Refractory Metals & Hard Materials 12(6) (1994) 329-333. [Zha97] Zhang, Z. Synthesis and characterization of novel Group VI metal (Mo ,W) nitride and oxide compounds, PhD Dissertation, Iowa State University, Ames (1997) [Zha99] Zhang, H. L. D. Z. Wang and N. K. Huang, Appl. Surf. Sci. 150 (1999) 34-38.

PAGE 418

+>0 \0 "I1 1 > I ..... 0 ~"1 ::, 0 c:r C N 0 "1 0. V, V, 0 (") i" 0. "O ,a s C1Q I Channel 1 to : Reactor System : I FrDwa3>-I Curtain H2 I Channel 2 to :T Reactor System 1 I Fr Dwg a> i r Precursor 1 lL:lnJection Pump : (Syringe Type) I I I Channel 3 to 1 Reactor System: I Fr Dwg 3) .... ,:----.-----I I I Baratron Gauge Diffusion Pump Master Vacuum Valve Precursor Nebulizer Chamber Quick Flanges Automated Gate Valve Sample Loading Port "'Graphite Susceptor ,_, Dwg1 Cooling Water to Drain Air to Gate Valve I CFrDwg 6 I 0 LJ Trap To Roughing Pump r trl ~a a~ ::0 > j n a Cl) Cl)

PAGE 419

"T1 (IQ c:: c:a >I N tc c:: O" O" Y' s Cl> Cl> ::!:l 0 (") 0 ::s [ 0. Cl> Cl> 0 8. 0. "O .a 5 (IQ Source Arto ,\._vac "' oi C: C: "' .c u g t: Q) E "' I Mo Bubbler to Ch 2, Ch. 4 Disconnected I I 1 ____________ HCI, SiH2Cl2 How contr olle r TiCl4 Bubbler needle va l ve TiCl4 to Ch. 4 ToDwg3 Ch. 4 H2 I !To Dwg 3 Ch. 3 H2 1 IToDwg3 Ch. 2 H2 I !To Dwg3 Ch. 1/Makeup H2 To Dwg 3' +>-0

PAGE 420

'"r:1 .... OQ i:: > Dwg3 I u.) Channel 1/Makeup Mak eup/Ma k eup Channel 3 to n H2 / Inert Diff Press Reactor System =:,--FrDw 2 ::, [To Dwg 1 > I (1) -=:,-(1) p) fs" '""1 Cl> Mo Bubbler Outlet FrDw 2 0.. Channel 2 H2 / Ch 2 Makeup p) Inert D P Cl> Channel 2 to Cl> FrDw 2 0 Reactor System .p. I To Ow 1 () ...... .... ...... p) ..... SiCl4 Bubbler Ou tl et (1) 0.. row 2 'O ,a Channel 3 H2 / co.,-. j I BCl3 from Source .... D P ::, Inert OQ Fr Dwa 2 Channel 4 H2 / Ch. 4 Makeup I I ( FC )--.J Channel 4 to Inert D P R'-1 I I Fr Dwg 2 Reactor System CH4from To Dwa 1 CH4Flowto I Source 0 Ch. 2 y CH4Flowto TiCl4 to Ch. 4 Ch. 4 I FrDwg4) c:;; CH4Flow CH4Common Contr ol Fl o w

PAGE 421

"TI (JQ C: > I +"-0 I),) v., 0 '< -:::, "'1 v., 0.. I),) v., v., 0 o I),) ..... 0 0.. "O "O s (JQ Pur ge I F r Dwg 2) Box BCl3 t o Sys tem <) Regula t o r R egulator : I I><) I I><) I Dwg4 5%BCl3in H2Cylinder Source 10% HClin H2 (lormer1y Si Cl4 line) I S ource Methane ( CH4) S ource N H3Source I +"--N

PAGE 422

'T1 1 > I Vt < 0 ::, .... ti) (") -er 0 ~'"I ti) 0 (") 0 ::, 0.. e; '< (i f!; .... 0 '"I 0.. ti) 0 (") i 0.. 't:I .; .... ::, (JQ ... Reactor Outlet 4l :S -~r a: N.C Ch. 2 & 3 Valve 2 0 Reactor Outlet Bubbler Li ne CD Cl CD .._ C :::, a.. ...J Small Black Valve Middle Underhood From Roughing Pum Fr Dwo 1 Pump Outlet Vacuum \._ Ven t Hood Secondary Reactor Box !J ...J Fill Nozzle Vent Scrubber Drain SS TubetPVC Dwq5 To P u rge Vac Line < FrDwg2 To Vac Pump < FrOwg2 T o Bubbler Vac Line 'FrDwo2' Vent to Hood ti) ti) , CD I I 2 u en -u.)

PAGE 423

'T:1 ~@ > I Q'I n a. 0.. s 0 ::i. OQ c:,, (') '$. s 0.. c:,, c:,, 0 (') ~-0.. 'O .... 'O s OQ Purge to H2 Cab. [><].-----, I Fr Dwg 2) ; I B I ox I H2 Cabinet (inside) ; (Outside) Regulator Dwg6 H2 Source Note : Reverse Threaded I Outs ide t ---------------------------------------Ins ide Inert Source to Vacuum :To Dwq 2 l I Box Air to Gate Valve I l Inert ~binet I I Filter Regulator Source Ar VacAr N2 Cylinder He Cyl inde r ~To Dwg 1 I I t>
PAGE 424

APPENDIXB OTHER PRECURSORS TESTED Below is a list of precursors that have been tested to check their performance in growing tungsten nitride (WNx) thin films. Dr. McElwee-White's group at the University of Florida's Department of Chemistry generated the precursors. The precursors have been listed in order of promise, with the first precursor/solvent pair showing us the best results thus far. Structures of the precursor molecules have been included, as well as the delivery system type (solid source vs. nebulizer) and the experimental temperature range for each precursor. In some cases, especially those using solid-source delivery, flux of the precursor to the reactor was minimal. In other cases using nebulizer delivery, flux to the reactor was adequate, but film deposition was minimal. 415

PAGE 425

416 Table B-1. Other precursors tested for MOCVD of WNx Deposition Delivery Promising? Precursor Structure Temp Range Notes ( C) Method (Yes/No) 9 C I 'I,, '' ',c I 575-600 Nebulizer Yes Film composition . w Cl~t"'CI similar to that for i-Pr precursor 1 b N Ill r CH 3 -t c1,,, .. ,,,,c1 Purification during .. w 700-800 Nebulizer Yes c1~f ""c1 precursor synthesis was N an issue Ill f CH 3 fl) I N oo, 111 ,,d 650-700 Solid M inimal precursor w No oc'l"O Source transport caused by low I I vapor pressure fl) I N oc .. 111 .. ,,Cl 700 Solid Minimal precursor w Source No oc.,,,,l "N transport c a used by low Cl ~C vapor pressure 'cH3

PAGE 426

Table B-1. Continued Precursor Structure '><_ t )< NH2 /N~ N CJ -........ I / II / Cl w w cI,......-1 '--.. /I '---c1 NH2 "-.,. N / NH2 x + 'x y NH I >-N=W=N-< I NH A 417 Deposition Temp. Range (QC) 1000 850 700 800 500 Deli very Method Nebulizer Solid Source Solid Source Nebulizer Nebulizer Promising ? (Yes/No) No No No No No Notes Minimal deposition Minimal precursor transport caused by low vapor pressure Minimal precursor transport caused by low vapor pres sure M inimal deposition Minimal deposition

PAGE 427

418 Table B-1 Continued Precursor Structure Deposition Delivery Promising? Temp Range Notes (OC) Method (Yes/No) Cl "-. Cl Cl Ph-N:=W/ "-./ N-Ph 500 Nebulizer No Minimal I 'c1/ \ deposition Cl Cl 'f-.... Cl N~l~N .,)( w 500 Nebulizer No Minimal /I' deposition )(' NH2 Cl NH2'>( CH 3 I CH 2 I C }I,,, .. m ,,,,cl 550-650 Nebulizer No Minimal w deposition c1~k~c1 Ill f CH 3

PAGE 428

BIOGRAPIIlCAL SKETCH Omar James Bchir was born on October 31, 1974 in St. Petersburg, Florida. He is the son of Sara and Hachemi Bchir and has an older sister, Annissa. He graduated in 1992 from the International Baccalaureate (IB) program at St. Petersburg Senior High School, receiving both the high school and 1B diplomas. Following high school, he attended the Georgia Institute of Technology, graduating with a bachelor's degree in chemical engineering in March 1997. Upon graduation, he worked for more than two years as a process engineer for Fluor Daniel, Inc in Greenville, SC. In the summer of 1999, he left Fluor Daniel to start graduate school at the University of Florida, pursuing his doctoral degree in chemical engineering. Upon graduation, he will work as a senior engineer at Intel Corporation in Chandler, Arizona. 419

PAGE 429

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ,~tk Timothy dron Chair, Professor of Chemical Engineering I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Professor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. fq~ Paul H. Holloway Distinguished Professor of Materials Science and Engineering I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality as a dissertation for the degree of Doctor of Philosophy. F& /4 = Professor of Chemical Engineering

PAGE 430

This dissertation was submitted to the Graduate Faculty of the College of Engineering and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August2004 Pramod P. Khargonekar Dean, College of Engineering Kenneth Gerhardt Interim Dean, Graduate School

PAGE 431

J LO 1 7 80 .. 20 .04 B J0+ II llllllll~lviii~ii1i1i11i~ii1001111 3 1262 08554 2990