<%BANNER%>

Chemical Vapor Deposition of Tungsten-Based Diffusion Barrier Thin Films for Copper Metallization

Permanent Link: http://ufdc.ufl.edu/UFE0041042/00001

Material Information

Title: Chemical Vapor Deposition of Tungsten-Based Diffusion Barrier Thin Films for Copper Metallization
Physical Description: 1 online resource (146 p.)
Language: english
Creator: Kim, Do
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: aacvd, barrier, copper, metallization, mocvd, tungsten
Chemical Engineering -- Dissertations, Academic -- UF
Genre: Chemical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The ternary material WNxCy was investigated for Cu diffusion barrier application. Thin films were deposited from tungsten diorganohydrazido(2-) complexes Cl4(CH3CN)W(NNR2) (1: R2=-(CH2)5-; 2: R2=Ph2; 3: R2=Me2) using metal-organic aerosol-assisted CVD. The films deposited from these novel precursors were characterized for their composition, bonding state, structure, resistivity, and barrier quality. WNxCy films from 1, 2 and 3 were successfully deposited in the absence and the presence of NH3 in H2 carrier in the temperature range 300 to 700 degreeC. All WNxCy films contained W, N, C, and a small amount of O as determined by XPS. The Cl content of the film was below the XPS detection limit (~ 1 at. %). The chemical composition of films deposited with 1 in H2/NH3 exhibited increased N levels and decreased C levels over the entire temperature range of this study as compared with to films deposited 1 in H2. As determined by XPS, W is primarily bonded to N and C for films deposited at 400 C, but at lower deposition temperature the binding energy of the W-O bond becomes more evident. The films deposited at 400 degreeC were X-ray amorphous and Cu/WNxCy/Si stacks annealed under N2 at 500 degreeC for 30 min maintained the integrity of both the Cu/WNxCy and WNxCy/Si interfaces. Comparison of films deposited from 2 with H2 only and H2/NH3 shows that the best films, in terms of composition, resistivity, surface roughness, and microstructure, are deposited using H2/NH3 carrier. The microstructure of films deposited with NH3 was X-ray amorphous below 450 degreeC. XPS measurements revealed that W is primarily bonded to N and C for films deposited between 300 and 700 degreeC. An Arrhenius plot of growth rate was consistent with surface reaction limited growth and the activation energy was lower for growth in the presence of NH3. It was observed that the surface roughness improved with added NH3. Samples annealed at higher temperature showed evidence of failure only when annealed at 700 degreeC. These results support the conclusion that WNxCy thin film deposited from 2 is a viable Cu diffusion barrier material. As anticipated, the film N content was higher for films deposited from 3 with added NH3 as compared to those deposited from 1 and 2. The films deposited with NH3 in H2 carrier at 400 degreeC had the highest N content of all films (27 at. %). An amorphous film microstructure was observed for films deposited below 500 degreeC. The apparent activation energy for the film growth in the kinetically controlled growth regime was 0.31 eV. The observation of AFM monograph indicates that the surface roughness improved with added NH3. Film growth of WNxCy by metal-organic aerosol-assisted CVD using 1, 2, and 3 highlights the importance of precursor selection, co-reactant selection (H2 only, H2/NH3, N2 only, and N2/NH3), and operating parameters (deposition temperature, pressure, and flow rate) on film properties and barrier performance. Preliminary material characterization and diffusion barrier testing reveals that films deposited using 2 with NH3 in H2 carrier is most promising for diffusion barrier applications.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Do Kim.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Anderson, Timothy J.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0041042:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041042/00001

Material Information

Title: Chemical Vapor Deposition of Tungsten-Based Diffusion Barrier Thin Films for Copper Metallization
Physical Description: 1 online resource (146 p.)
Language: english
Creator: Kim, Do
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: aacvd, barrier, copper, metallization, mocvd, tungsten
Chemical Engineering -- Dissertations, Academic -- UF
Genre: Chemical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The ternary material WNxCy was investigated for Cu diffusion barrier application. Thin films were deposited from tungsten diorganohydrazido(2-) complexes Cl4(CH3CN)W(NNR2) (1: R2=-(CH2)5-; 2: R2=Ph2; 3: R2=Me2) using metal-organic aerosol-assisted CVD. The films deposited from these novel precursors were characterized for their composition, bonding state, structure, resistivity, and barrier quality. WNxCy films from 1, 2 and 3 were successfully deposited in the absence and the presence of NH3 in H2 carrier in the temperature range 300 to 700 degreeC. All WNxCy films contained W, N, C, and a small amount of O as determined by XPS. The Cl content of the film was below the XPS detection limit (~ 1 at. %). The chemical composition of films deposited with 1 in H2/NH3 exhibited increased N levels and decreased C levels over the entire temperature range of this study as compared with to films deposited 1 in H2. As determined by XPS, W is primarily bonded to N and C for films deposited at 400 C, but at lower deposition temperature the binding energy of the W-O bond becomes more evident. The films deposited at 400 degreeC were X-ray amorphous and Cu/WNxCy/Si stacks annealed under N2 at 500 degreeC for 30 min maintained the integrity of both the Cu/WNxCy and WNxCy/Si interfaces. Comparison of films deposited from 2 with H2 only and H2/NH3 shows that the best films, in terms of composition, resistivity, surface roughness, and microstructure, are deposited using H2/NH3 carrier. The microstructure of films deposited with NH3 was X-ray amorphous below 450 degreeC. XPS measurements revealed that W is primarily bonded to N and C for films deposited between 300 and 700 degreeC. An Arrhenius plot of growth rate was consistent with surface reaction limited growth and the activation energy was lower for growth in the presence of NH3. It was observed that the surface roughness improved with added NH3. Samples annealed at higher temperature showed evidence of failure only when annealed at 700 degreeC. These results support the conclusion that WNxCy thin film deposited from 2 is a viable Cu diffusion barrier material. As anticipated, the film N content was higher for films deposited from 3 with added NH3 as compared to those deposited from 1 and 2. The films deposited with NH3 in H2 carrier at 400 degreeC had the highest N content of all films (27 at. %). An amorphous film microstructure was observed for films deposited below 500 degreeC. The apparent activation energy for the film growth in the kinetically controlled growth regime was 0.31 eV. The observation of AFM monograph indicates that the surface roughness improved with added NH3. Film growth of WNxCy by metal-organic aerosol-assisted CVD using 1, 2, and 3 highlights the importance of precursor selection, co-reactant selection (H2 only, H2/NH3, N2 only, and N2/NH3), and operating parameters (deposition temperature, pressure, and flow rate) on film properties and barrier performance. Preliminary material characterization and diffusion barrier testing reveals that films deposited using 2 with NH3 in H2 carrier is most promising for diffusion barrier applications.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Do Kim.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Anderson, Timothy J.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0041042:00001


This item has the following downloads:


Full Text

PAGE 1

1 CHEMICAL VAPOR DEPOSITION OF TUNGSTEN BASED DIFFUSION BARRIER THIN FILMS FOR COPPER METALLIZATION By DOJUN KIM A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQ UIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 D ojun K im

PAGE 3

3 To my parents Hwakyum Kim and Hyosun Kim

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank my research advisor Dr. Tim othy J. Anderson, for his full support and excellent guidance through my four year s study in University of Florida. I would like to thank the members of supervisory committee Dr. Lisa McElwee White, Dr. Valentin Craciun, and Dr. Fan Ren who provided me with valuable comments in my work. The work on maintenance of CVD and ALD systems would not have been successful if there was no help and support of Dennis Vince (Chemical Engineering, UF), Jim Hinnant (Chemical Engineering, UF), and Rob Holobof (A&N C orporation). The excellent facilities and helpful staffs for material characterization s at Major Analytical Instrumentation Center (MAIC) were highly helpful for me to obtain valuable results. I would like to thank Eric Lambers (XPS/AES) Kerry Siebe i n ( TEM/EDS) for their assistance. I also would like to thank Dr. Ivan Kravchenko ( S putter deposition system ) at Nanofabrication Facilities. I also would like to thank Dr. Khalil Abboud (Structural chemistry of X Ray diffraction) at Department of Chemistry. I give my thanks to my colleagues of research project for their assistance : Oh Hyun Kim, Jooyoung Lee, Dr. J rgen Koller, Dr. Lii Cherng Leu, Dr. Kee C han Kim, Dr. Hiral M. Ajmera, Dr. Michael June and Christopher O Donohue Especially, my fellow doctor al student, Oh Hyun Kim was an excellent collaborator for my research in many ways. T he last but not the least, I would like to give my s incerest thanks to my parents, Hwakyum Kim and Hyosun Kim, for their unconditional love and support I am very gratef ul to my wife, Sora Park, for her love and care in my life. My lovely son, Jinho Kim and my lovely daughter, Katherine Nayoun Kim, always ma de me happy to work hard er and harder during my studies.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ........... 4 LIST OF TABLES ................................ ................................ ................................ ...................... 8 LIST OF FIGURES ................................ ................................ ................................ .................... 9 ABSTRAC T ................................ ................................ ................................ ............................. 14 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ............. 16 2 LITERATURE REVIEW ................................ ................................ ................................ ... 19 2.1 Diffus ion Mechanism in Cu M etallization ................................ ................................ 19 2.2 Ta/TaN Bilayer Structure as a Diffusion Barrier ................................ ...................... 21 2.3 Chemical Vapor Deposition of Tun gsten Based Diffusion Barrier ........................... 23 2.3.1 Tungsten Nitride as a Diffusion Barrier ................................ ........................ 23 2.3.2 Tungsten Carbonitride as a Diffusion B arrier ................................ ................ 25 2.4 Atomic Layer Deposition of Tungsten Based Diffusion Barrier ............................... 29 2.4.1 Tungsten Nitride as a Diffusion Barrier ................................ ........................ 29 2.4.2 Tungsten Carbonitride as a Diffusion Barrier ................................ ................ 31 3 EXPERIMENTAL PROCEDURE ................................ ................................ ..................... 40 3.1 Precursor Synthesis ................................ ................................ ................................ .. 40 3.2 Film Growth ................................ ................................ ................................ ............ 40 3.3 Film Characterization s ................................ ................................ ............................. 40 3.4 Diffusion Barrier Testing ................................ ................................ ......................... 41 4 DEPOSITION OF WN x C y FROM Cl 4 (CH 3 CN)W(N pip) AS A SINGLE SOURCE PRECURSOR ................................ ................................ ................................ .................... 47 4.1 X ray Crystallographic Study of Cl 4 (CH 3 CN)W(N pip) ................................ ........... 47 4.2 Preliminary P recursor S creening ................................ ................................ .............. 47 4.3 Film S truc ture ................................ ................................ ................................ .......... 48 4.4 Chemical C omposition ................................ ................................ ............................. 49 4.5 Chemical B onding S tates ................................ ................................ ......................... 50 4.6 Lattice P arameter ................................ ................................ ................................ ..... 51 4.7 Average Gr ain S ize ................................ ................................ ................................ .. 52 4.8 Electrical R esistivity ................................ ................................ ................................ 52 4.9 Film G rowth R ate ................................ ................................ ................................ .... 53 4.10 Diffusion B arrier Testing ................................ ................................ ......................... 53 4.11 Conclusions ................................ ................................ ................................ ............. 54

PAGE 6

6 5 DEPOSITION OF WN x C y FROM Cl 4 (CH 3 CN)W(N pip) : EFFECT OF NH 3 ON FILM PROPERTIES ................................ ................................ ................................ .................... 67 5.1 Film Structure ................................ ................................ ................................ .......... 67 5.2 Surface Morphology ................................ ................................ ................................ 68 5 3 Chemical C omposition ................................ ................................ ............................. 68 5.4 Chemical Bonding States ................................ ................................ ......................... 70 5.5 Film Growth Rate ................................ ................................ ................................ .... 72 5.6 Electrical Resistivity ................................ ................................ ................................ 73 5.7 Diffusion Barrier Testing ................................ ................................ ......................... 73 5.8 Conclusions ................................ ................................ ................................ ............. 74 6 DEPOSITION OF WN x C y FROM Cl 4 (CH 3 CN)W(NNPh 2 ) AS A SINGLE SOURCE PRECURSOR ................................ ................................ ................................ .................... 83 6.1 Film Structure ................................ ................................ ................................ .......... 83 6.2 Lattice Parameter and Average Grain Size ................................ ............................... 83 6.3 Chemical Composition ................................ ................................ ............................. 84 6.4 Chemical Bonding States ................................ ................................ ......................... 85 6.5 Film Growth Rate ................................ ................................ ................................ .... 87 6.6 Electrica l resistivity ................................ ................................ ................................ 87 6.7 Diffusion Barrier Testing ................................ ................................ ......................... 88 6.8 Conclusions ................................ ................................ ................................ ............. 89 7 DEPOSITION OF WN x C y FROM Cl 4 (CH 3 CN)W(NNPh 2 ): EFFECT OF NH 3 ON FILM PROPERTIES ................................ ................................ ................................ .......... 97 7.1 Film Structure ................................ ................................ ................................ .......... 97 7.2 Chemical Compo sition ................................ ................................ ............................. 97 7.3 Chemical Bonding States ................................ ................................ ......................... 99 7.4 Surface Morphology ................................ ................................ .............................. 100 7.5 Film Growth Rate ................................ ................................ ................................ .. 101 7.6 Electrical Resistivity ................................ ................................ .............................. 101 7.7 Diffusion Barrier Testing ................................ ................................ ....................... 102 7.8 Conclusions ................................ ................................ ................................ ........... 103 8 DEPOSITION OF WN x C y FROM Cl 4 (CH 3 CN)W(NNMe 2 ): EFFECT OF NH 3 ON FILM PROPERTIES ................................ ................................ ................................ ........ 114 8.1 Film Structure ................................ ................................ ................................ ........ 114 8.2 Chemical Composition ................................ ................................ ........................... 114 8.3 Chemical Bonding States ................................ ................................ ....................... 116 8.4 Surface Morphology ................................ ................................ .............................. 118 8.5 Film Growth Rate ................................ ................................ ................................ .. 118 8.6 Electrical Resistivity ................................ ................................ .............................. 118 8.7 Conclusions ................................ ................................ ................................ ........... 119 9 REACTOR MODELING USING CFD SOFTWARE ................................ ...................... 125

PAGE 7

7 9.1 Description of th e Raman Assisted CVD reactor ................................ .................... 125 9.2 Multiphase Flow Simulation of the Raman Assisted CVD reactor ......................... 126 10 CONCLUSION S AND FUTURE WORK ................................ ................................ ........ 137 10.1 Ru WN x C y for Diffusion Barrier and Cu Direct Plate Applications ........................ 137 10.2 WN x C y for Realistic Diffusion Barrier Tes ting ................................ ....................... 138 LIST OF REFERENCES ................................ ................................ ................................ ........ 140 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ... 146

PAGE 8

8 LIST OF TABLES Table page 2 1 Precursors used for film growth of WN x by CVD ................................ .......................... 35 2 2 Precursors used for film growth of WN x C y by CVD ................................ ...................... 35 2 3 Precursors used for film growth of WN x by ALD ................................ ........................... 36 2 4 Precursors used for film growth of WN x C y by ALD ................................ ....................... 36 4 1 Crystal data and structure refinement for Cl 4 (CH 3 CN)W(N pip) ( 1 ) .............................. 56 4 2 Selected bond distances () and angles () for Cl 4 (CH 3 CN)W(N pip) ( 1 ) ...................... 57 4 3 Reported binding energy (BE) values ................................ ................................ ............ 58 9 1 Boundary conditions for CVD reactor ................................ ................................ .......... 129

PAGE 9

9 LIST OF FIGURES Figure pag e 1 1 The device delay as a function of device generation. Adopted from M. T. Bohr, IEEE Internat ional Electron Devices Meeting (1995) 241 242 ................................ ...... 18 1 2 SEM cross sectional images: A) Cu deposition without Cu diffusion barrier; B) Cu deposition with Cu diffusion barrier. ................................ ................................ .............. 18 2 1 Microstructure of Cu diffusion barrier materials: A) single crystal; B) polycrystalline; C) polycrystalline columnar; D) nano crystalline; E) amorphous. ltrathin diffusion barrier/liners for 385 ................. 37 2 2 Diagram showing the applications of metals and nitrides in modern semi conductor films: Current research efforts and applications for semiconductor device 2261. ................................ ............ 38 2 3 Simplified processing steps in dual damascene structure for Cu metallization ............... 39 3 1 The diorganohydrazido(2 ) tungsten complexes Cl 4 (CH 3 CN)W(NNR 2 ) ( 1 : R 2 = (CH 2 ) 5 ; 2 : R 2 = Ph 2 ; 3 : R 2 = Me 2 ) ................................ ................................ ............... 43 3 2 Schematic diagram of the aerosol assisted CVD system. ................................ ................ 44 3 3 Proce ss flow on film properties. (MAIC, http: \ \ maic.mse.ufl.edu, October, 2008) ......... 45 3 4 Process flow on diffusion barrier testing. (MAIC, http: \ \ maic.mse.ufl.edu, October, 2008). ................................ ................................ ................................ ............................ 46 4 1 Thermal ellipsoids diagram of the molecular structure of Cl 4 (CH 3 CN)W(N pip) ( 1 ) Thermal ellipsoids are drawn at 50% probability. H atoms are omitted for clarity ........ 59 4 2 XRD spectra for films deposited on Si(100) in H 2 carrier: A) 300 C, B) 700 C, C) between 300 and 700 C, and D) standard diffraction plots for W 2 N and WC 1 x ..... 60 4 3 Variation in chemical composition of W, N, C, and O content in the films with deposition temperature. Data are measured by XPS after 10 min Ar + ion sputter. .......... 61 4 4 Change of binding energies in A) W 4 f B) N 1 s C) C 1 s and D) O 1 s with deposition temperature. Data are from XPS after 10 min Ar + ion sputter. ...................... 62 4 5 Change in lattice parameter with deposition temperature for polycrystalline films deposited from 1 based on WN x C y (111) diffraction peaks. ................................ ......... 63

PAGE 10

10 4 6 Change in average grain size with deposition temperature for polycrystalline films dep osited from 1 based on WN x C y (111) diffraction peaks. ................................ ......... 63 4 7 Change in film resistivity with deposition temperature. Data are measured by four point probe. ................................ ................................ ................................ ................... 64 4 8 Change in growth rate with deposition temperature. Thickness measured by cross sectional SEM. ................................ ................................ ................................ .............. 64 4 9 The performance of diffusion barrier by XRD meas urement for Cu/WN x C y /Si stacks before and after annealing at 500 C. ................................ ................................ ............. 65 4 10 SEM images of Si surface after etch pit test A) before annealing and B) after annealing at 500 C. ................................ ................................ ................................ ...... 65 4 11 The performance of diffusion barrier by AES depth profile for Cu/WN x C y /Si stacks after annealing at 500 C. ................................ ................................ .............................. 66 5 1 XRD spectra for films deposited on Si(100) with NH 3 : A) between 300 and 700 C; B) standard diffraction patterns for W 2 N and WC 1 x ................................ ............... 76 5 2 Surface morphology of films deposited on Si(100) substrate at various temperature: A) 300 C without NH 3 ; B) 600 C without NH 3 ; C) 300 C with NH 3 ; D) 600 C with NH 3 ................................ ................................ ................................ ...................... 77 5 3 Variation in chemical composition of A) W, B) N, C) C, and D) O content in the films with deposition temperature with and without added NH 3 Data are measured by XPS after 10 mi n Ar + ion sputter. ................................ ................................ .............. 78 5 4 Change of binding energies in A) W 4 f B) N 1 s C) C 1 s and D) O 1 s with deposition temperature in the presence of NH 3 Data are from XPS after 10 min Ar + ion spu tter. ................................ ................................ ................................ ..................... 79 5 5 Change in growth rate with deposition temperature for films deposited with and without added NH 3 Thickness was measured by cross sectional SEM. ......................... 80 5 6 Change in film resistivity with deposition temperature with and without added NH 3 Data are measured by four point probe. ................................ ................................ ......... 80 5 7 The performance of diffusion barrier by XRD measurement for Cu/WN x C y /Si stacks before and after annealing at 500 C. ................................ ................................ ............. 81 5 8 TEM cross sectional images of Cu/WN x C y /Si stacks: [A) and B)] before annealing and [ C) and D)] after annealing at 500 C. ................................ ................................ ..... 82 6 1 XRD spectra for films deposited on Si(100) at various temperatures: A) 300 C, B) 700 C, C) between 300 and 700 C, and D) standard powder diffr action pattern for W 2 N and WC 1 x ................................ ................................ ................................ ...... 90

PAGE 11

11 6 2 Change in A) lattice parameter and B) average grain size with deposition temperature for polycrystalline films deposited from 2 The estimates a re based on position and shape of diffraction peaks. ................................ ................................ ......... 91 6 3 Variation of W, N, C, and O content in the films deposited from 2 Data are from XPS measurements after 10 min Ar + ion sputter. ................................ ........................... 91 6 4 Change of binding energies in A) W 4 f B) N 1 s C) C 1 s and D) O 1 s with deposition temperature. Data are from XPS after 10 min Ar + ion sputter. ...................... 92 6 5 SEM images of films grown on Si(100) substrate: A) cross sectional view of film grown at 300 C; B) cross sectional view of film grown at 700 C; C) surface morphology of film grown at 300 C; D) surface morphology of fi lm grown at 700 C. ................................ ................................ ................................ .......................... 93 6 6 Change in growth rate with deposition temperature for films deposited from 2 Thickness measured by cross sectional SEM. ................................ ................................ 94 6 7 Change in film resistivity (four point probe) with deposition temperature for films deposited from 2 ................................ ................................ ................................ ........... 94 6 8 Cross sectional TEM images of Cu/WN x C y /Si stacks: [ A) and B ) ] before annealing and [C) and D ) ] after annealing at 500 C. ................................ ................................ ..... 95 6 9 EDS depth profile of Cu/WN x C y /Si stacks annealed at 500 C. ................................ ...... 96 6 10 The performance of diffusion barrier by XRD measurement for Cu/WN x C y /Si stacks before and after annealing at 500 C. ................................ ................................ ............. 96 7 1 X RD spectra for films deposited on Si(100) w ith NH 3 : A) 300 C, B) 700 C, C) change in XRD spectra, and D) standard diffraction plots for W 2 N and WC 1 x ..... 105 7 2 XPS spectra for films deposited on Si(100) with NH 3 Note that Cl peaks are evident as a function of growth temperature. ................................ ................................ ............ 106 7 3 Comparison of W, N, C, and O content in the films deposited in the presence and absence of NH 3 Data are measured by XPS after 10 min Ar + ion sputter. ................... 107 7 4 Change of binding energies in A) W 4 f B) N 1 s C) C 1 s and D) O 1 s with deposition temperature. Data are from XPS after 10 min Ar + ion sputter. .................... 108 7 5 Surface morphology of films grow n on Si(100) substrate: A) film grown at 300 C without NH 3 ; B) film grown at 700 C without NH 3 ; C) film grown at 300 C with NH 3 ; D) film grown at 700 C with NH 3 ................................ ................................ ..... 109 7 6 SE M images of fi lms grown on Si(100) substrate: A) cross sectional view of film grown at 300 C; B) cross sectional view of film grown at 700 C. ............................. 110

PAGE 12

12 7 7 Change in growth rate with deposition temperature fo r the films deposited in the presence and absence of NH 3 Thickness was measured by cross sectional SEM. ....... 110 7 8 F ilm resistivity as a function of deposition temperature for the films depos ited in the presence and absence of NH 3 ................................ ................................ ...................... 111 7 9 A) TEM image and B) EDS depth profile of a Cu/WN x C y /Si stack annealed at 500 C for 30 min. ................................ ................................ ................................ ....... 111 7 10 Change in XR D patterns with annealing temperature for Cu/WN x C y /Si stacks. ............ 112 7 11 Change in sheet resistance with annealing temperature for Cu/WN x C y /Si stacks. Da ta are measured by four point probe. ................................ ................................ ....... 112 7 12 Cross sectional TEM images of Cu/WN x C y /Si stacks: A) as grown and B) after annealing at 700 C. ................................ ................................ ................................ .... 113 7 13 C ross sectional SEM images of Cu/WN x C y /Si stacks: A) as grown and B) after annealing at 700 C. ................................ ................................ ................................ .... 113 8 1 XRD spectra for films deposited on Si(100) with NH 3 : A) 300 C; B) 400 C; C) change in XRD spectra; D) standard powder diffraction p attern for W 2 N and WC 1 x ................................ ................................ ................................ ......................... 120 8 2 XPS spectra for films deposited on Si(100) with NH 3 No Cl peaks detected. ............. 121 8 3 Variation in th e chemical composition of W, N, C, and O contents in the films with deposition temperature. Data are measured by XPS after 10 min Ar + ion sputter. ........ 121 8 4 Change of binding energies in A) W 4 f B) N 1 s C) C 1 s and D) O 1 s with deposition temperature. Data are from XPS after 10 min Ar + ion sputter. .................... 12 2 8 5 Surface morphology of films grown on Si(100) substrate: A) film grown a t 300 C with NH 3 ; B) film grown at 700 C with NH 3 ................................ ............................. 123 8 6 SEM images of films grown on Si(100) substrate: A) cross sectional view of film grown at 300 C; B) cross sectional view of film grown at 700 C. ............................. 123 8 7 Change in growth rate with deposition temperature for the films deposited from 3 Thickness was measured by cross sectional SEM. ................................ ....................... 124 8 8 Change in film resistivity (four point probe) with deposition temperature for the films deposited from 3 ................................ ................................ ................................ 124 9 1 Schematic photographs of A) CVD react or system that is interfaced to the Raman spectrometry ; B) nebulizer system; C) the impinging jet probe reactor. ....................... 130 9 2 Mesh design of CVD reactor using GAMBIT ................................ ......................... 131

PAGE 13

13 9 3 Color filled contours of static temperature (K) and contour line of static temperature (K) in the vicinity of the heater. ................................ ................................ ................... 132 9 4 Contours of velocity magnitude (m/s) and velocity vector colored by velocity magnitude (m/s) in the vicinity of the heater. ................................ ............................... 133 9 5 Contours of velocity magnitude (m/s) and volume fraction of solvent phase in multiphase flow model. ................................ ................................ ................................ 134

PAGE 14

14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillme nt of the Requirements for the Degree of Doctor of Philosophy CHEMICAL VAPOR DEPOSITION OF TUNGSTEN BASED DIFFUSION BARRIER THIN FILMS FOR COPPER METALLIZATION By D ojun K im December 2009 Chair: Tim othy J. Anderson Major: Chemical Engineering The ternar y material WN x C y was investigated for Cu diffusion barrier application. Thin films were deposited from tungsten diorganohydrazido(2 ) complexes Cl 4 (CH 3 CN)W(NNR 2 ) ( 1 : R 2 = (CH 2 ) 5 ; 2 : R 2 = Ph 2 ; 3 : R 2 = Me 2 ) using metal organic aerosol assisted CVD. T he films d eposited from these novel precursors were characterized for their composition, bonding state, structure, resistivity, and barrier quality. WN x C y films from 1 2 and 3 were successfully deposited in the absence and the presence of NH 3 in H 2 carrier in the t emperature range 300 to 700 C. All WN x C y films contained W, N, C, and a small amount of O as determined by XPS. The Cl content of the film was below the XPS detection limit (~ 1 at. %). The chemical composition of films deposited with 1 in H 2 /NH 3 exhib ited increased N levels and decreased C levels over the entire temperature range of this study as compared with to films deposited 1 in H 2 As determined by XPS, W is primarily bonded to N and C for films deposited at 400 C, but at lower deposition tempe rature the binding energy of the W O bond becomes more evident. The films deposited at 400 C were X ray amorphous and Cu/WN x C y /Si stacks annealed under N 2 at 500 C for 30 min maintained the integrity of both the Cu/WN x C y and WN x C y /Si interfaces.

PAGE 15

15 Compari son of films deposited from 2 with H 2 only and H 2 /NH 3 shows that the best films, in terms of composition, resistivity, surface roughness, and microstructure are deposited using H 2 /NH 3 carrier. The microstructure of films deposited with NH 3 was X ray amor phous below 450 C. XPS measurements revealed that W is primarily bonded to N and C for films deposited between 300 and 700 C. An Arrhenius plot of growth rate was consistent with surface reaction limited growth and the activation energy was lower for g rowth in the presence of NH 3 It was observed that the surface roughness improved with added NH 3 Samples annealed at higher temperature showed evidence of failure only when annealed at 700 C. These results support the conclusion that WN x C y thin film de posited from 2 is a viable Cu diffusion barrier material. As anticipated, t he film N content was higher for films deposited from 3 with added NH 3 as compared to those deposited from 1 and 2 The films deposited with NH 3 in H 2 carrier at 400 C had the high est N content of all films ( 27 at. % ) An amorphous film microst ructure was observed for films deposited below 500 C. The apparent activation energy for the film growth in the kinetically controlled growth regime was 0.31 eV. The observation of AFM monog raph indicates that the surface roughness improved with added NH 3 Film growth of WN x C y by metal organic aerosol assisted CVD using 1 2 and 3 highlights the importance of precursor selection, co reactant selection (H 2 only, H 2 /NH 3 N 2 only, and N 2 /NH 3 ), and operating parameters (deposition temperature, pressure, and flow rate) on film properties and barrier performance. Preliminary material characterization and diffusion barrier testing reveals that films deposited using 2 with NH 3 in H 2 carrier is most promis ing for diffusion barrier applications.

PAGE 16

16 CHAPTER 1 INTRODUCTION Statement of P roblem s The continuous challenges in microelectronic integrated circuits are increasing speed and improving reliability. The RC time delay hinder s further increasing of spe ed in integrated circuits (Figure 1 1) D evice dimensions continue to decrease on integrated circuit s, and the industry is transition ing from Al based interconnects to Cu based interconnects is required for multilevel metallization to minimize the RC time delay. C u based interconnects show greater resistance toward electromigration and 40% lower electrical resistivity ( Cu ~ 1.67 cm and Al ~ 2.65 cm) as compared to Al based interconnects [1 3] As a result of the high diffusivity of Cu in Si and SiO 2 ( D Cu ~ 2 10 5 cm 2 /s at 500 C) high priority has been placed on developing Cu diffusion barriers (Figure 1 2) [4] The presence of Cu in Si and SiO 2 results in serious degradation of device performance associated with contact resistance, barrier height, p n junctions, contact layers, and electrical connections [5] Therefore, an effective Cu diffusion barrier is required to block Cu t ransport and intermixing with adjacent dielectric materials for Cu interconnect technology. To provide excellent diffusion barrier performance characteristics, deposited fi lms need to possess certain properties such as good step coverage, low electrical resistivity, low deposition temperature and amorphous microstructure. Various transition metal nitrides have been investigated as Cu diffusion barriers including TiN, TiSi x N y TaN, TaSi x N y WN x WSi x N y and WB x N y [6 12] Ta/TaN bilayers deposit ed by physical vapor deposition (PVD) are the currently utilized Cu diffusion barriers in semiconductor device technology. However, limitations of PVD due to the directional nature of deposition cause problems upon scaling down the barrier thickness. Che mical vapor deposition (CVD) and atomic layer deposition (ALD) of TaN x thin

PAGE 17

17 films also ha ve difficulties in depositing conductive TaN due to the preferential formation of Ta 3 N 5 ( ~ 2 10 8 cm) [13 14] The binary phase material, tungsten nitride (WN x ), is a promising candidate for replac ing the prevailing diffusion barrier of Ta/TaN bilayer structure [15, 16] WN x film show s good thermal stability with Cu, acceptably low resistivit y when deposited by CVD, and reasonable chemical mechanical planarization (CMP) processing [10] The ternary phase material, tungsten carbonitride (WN x C y ), is also a promising candidate for diffusion barrier applications. WN x C y film has low electrical resistivity good adhesion to Cu, good r esistance to diffusion of Cu, and acceptable film growth on SiO 2 The efficacy of WN x C y film as a diffusion barrier has been demonstrated for films grown by both CVD and ALD [13, 17 19] It has been shown that X ray amorphous ternary phase materials such as TiSi x N y TaSi x N y WSi x N y and WB x N y hav e better performance as Cu diffusion barriers than binary phase materials due to higher recrystallization temperature and thus lack of grain boundaries, which can serve as Cu diffusion pathways [12, 20 22] Aerosol assisted CVD (AACVD) is a useful technique for growing films of refractory metal nitrides because aerosol assist ed delivery permits use of low volatility precursor s and thermally sensitive precursor s that decompose before sublimation can be used [23] Recently, we reported the synthesis of the diorganohydrazido(2 ) tung sten complexes Cl 4 (CH 3 CN)W(NNR 2 ) ( 1 : R 2 = (CH 2 ) 5 ; 2 : R 2 = Ph 2 ; 3 : R 2 = Me 2 ) and Cl 4 (pyridine)W(NNR 2 ) ( 4 : R 2 = Ph 2 ) by reacting 1,1 diorganohydrazines with tungsten hexachloride (WCl 6 ) followed by treatment with acetonitrile (CH 3 CN) or pyridine (C 5 H 5 N) [24] T he diorganohydrazido(2 ) tungsten complexes ( 1 3 ) w ere demonstrated to be single source precursor s for the metal organic CVD (MOCVD) of W N x C y thin films in the absence of NH 3 in H 2 carrier. T he effect of NH 3 in H 2 carrier was demonstrated on the properties of W N x C y thin films deposited from 1 2 and 3 The diffusion barrier testing was performed to investigat e the

PAGE 18

18 Cu diffusion barrier properties and the onset of failure process via formation of more resistive copper silicide (Cu x Si) Figure 1 1. The d evice delay as a function of device generation. Adopted from M. T. Bohr, Interconnect scaling the real limiter to high performance ULSI Proceedings of IEEE International Electron Devices Meeting (1995) 241 242 Figure 1 2. SEM cross sectiona l images: A) Cu deposition without Cu diffusion barrier; B) Cu deposition with Cu diffusion barrier B) A) Cu 3 Si x

PAGE 19

19 CHAPTER 2 LITERATURE REVIEW 2.1 Diffusion Mechanism in Cu M etallizatio n The substitution of Cu has been a recent technolog ical innovation for the standa rd Al Cu metal interconnects in order to reduce resistance and RC time delay in microelectronic integrated circuits [2, 3] The current Cu technolo gy shows improved current carrying capability by greater resistance toward electromigration and no device contamination by Cu migration. The success of the shift to Cu includes the development of an electroplating process for the Cu interconnects, dual dam ascene CMP, and an effective liner material for a Cu diffusion barrier and adhesion promoter It is required to establish a fundamental understanding of the predominant diffusion mechanisms for atomic mobility and associated diffusion phenomena in order t o identify an effective liner for Cu technology. The placement of chemically different atoms in close proximity causes atomic migration for the purpose of reducing the overall free energy and establish ing equilibrium. Typical reasons for atomic migration are the presence of concentration differences, existence of a negative free energy of reaction, application of an electrical field, availability of thermal energy, generation of a strain gradient, or a combination of some or all of these factors. Atomic migration could result in a diffu si ve flux. The net flow of atoms by diffusion is described by Fick s law. (2.1) where C is the atomic concentration, J is the atomic flux per unit area per second, and x is distance. The temperature dependence of the diffusion coefficient D takes the form of an Arrhenius relationship.

PAGE 20

20 (2.2) where D 0 is a constant, Q is the activation energy for diffusion, k is Boltzmann s constant, and T is the temperatu re in degrees Kelvin. There are three typical failure mechanisms in the Cu/liner material system. First, Cu diffuses along grain boundaries. Second, Cu (or substrate atoms) diffuses through bulk defects in the liner (vacancies and dislocations). Third, l oss of liner integrity results from a metallurgical or chemical reaction with the Cu and/or substrate. Lattice diffusion rates are proportional to the absolute melting temperature T m (2.3) where A is a proportionality co nstant that depends on a variety of factors, including lattice structure and type of material. Diffusion along grain boundaries has the highest diffusion rates (or largest A), which result from a large misfit between adjoining grains. Diffusion by disloc ations shows intermediate diffusion rates. Diffusion due to atom vacancy exchange has the lowest diffusion rate (or smallest A). This indicates that Cu barrier materials with highe r melting points could act as better Cu diffusion barriers. Also, the mic rostructure of Cu barrier materials plays an important role in the resulting diffusion barrier performance. Film microstructure s in Fig. 2 1 can be categorized as single crystal, polycrystalline, nano crystalline (i.e. polycrystalline with grain size belo w ~ 5 nm), and amorphous. Single crystal line materials are the ideal microstructure of Cu diffusion barrier s Lattice mismatch s with the underlying substrate and thermal budget limitations make it difficult to deposit liners in single crystal microstruct ure. Hence, amorphous phase Cu diffusion barriers are the most desirable for diffusion barrier applications. There are three basic requirements for diffusion barrier material s First, a viable barrier material must not react with Cu or the underlying sub strate under thermal,

PAGE 21

21 mechanical, and electrical stress conditions. Second, the density of the diffusion barrier must be as close to ideal as possible for the purpose of eliminating diffusion through voids, defects, or loosely packed grain boundaries. Th ird, the microstructure of diffusion barrier must have no grain boundary diffusion paths. 2. 2 Ta/TaN Bilayer Structure as a Diffusion Barrier The introduction of Cu interconnect technology results in t h e need for refractory metal nitride films in mod ern semiconductor technology (Figure 2 2 ) Since Cu rapidly diffuse s in Si, a diffusion barrier should be employed between the metals and dielectrics to prevent Cu transport and intermixing with adjacent dielectric materials. Even if Cu interconnects show l ow er electrical resistivity and great er res istance toward electromigration than Al interconnects the use of Cu interconnects req uires conducting layers in the metallization structure which enhance the Cu adhesion to dielectrics. E xcellent adhesion to the underlying layer or interconnect material is required to prevent reliability problem such as electromigration and gross delamination during the CMP process. A dditional requirement s for Cu diffusion barrier include amorphous microstructure, low electrical resistivity, high electromigration resistance, good step coverage, low deposition temperature ( 400 C), and minimal thickness [25] Cu interconnect metallization has introduced new concepts in integration schemes : dual damascene structures, CMP process and Cu electroplating. Figure 2 3 depicts the simplified processing steps for the barrier film process that is used for the Cu dual d amascene structures. The requirements are grouped into film properties and process compatibility. The film propert y requirements includes ultra low thickness ( < 100 ), low resistivity ( < 500 cm), low halide residues ( < 2 at. %), good step coverage ( > 90 %), and reasonable process rate (30 /min). The process compatibility requirements includes CMP compatible, low deposition temperature ( 400 C), good adhesion on the etch stopper, good a dhesion on SiO 2 and good adhesion on Cu.

PAGE 22

22 These requirements must be met by any new barrier material to be used in the Cu dual damascene structure [26] Ta/TaN bilayer structure has been used for diffusion barrier applications in dual damascene structures for current Cu interconnect metallization in the se miconductor industry. Ta shows a high melting point (2669 C) and good stablility with Cu. TaN shows high thermal, mechanical and chemical stability. TaN shows good adhesion to SiO 2 and low materials and Ta shows the lack of Cu Ta compound and better adhesion than Cu/TaN adhesion. T herefore, Ta/TaN bilayer structure has been used for Cu interconnect technology. Ta/TaN liner has very low in plane electrical resistivity, since phase Ta de posited on TaN surface is spontaneously formed with a resistivity in the range 15 to 60 cm. Although PVD TaN has been successful so far as a Cu diffusion barrier, due to the downscaling of device dimension s in microelectronic integrated circuits, future Cu interconnects require Cu diffusion barrier s deposited by CVD or ALD. T h e drawbacks of the present Ta/TaN bilayer stru c ture are both in process and material. PVD is a line of sight process indicating the limitation s of PVD due to the directional na ture of deposition This cause problems upon scaling down the barrier thickness T he application of PVD techniques is limited by concerns over their ability to provide good conformality in sub 100 nm device technology. Also, the u nderlying device layer o n the substrate can be damaged due to high energ y particles. Many researchers have attempted to grow TaN by CVD and ALD. TaN has many polymorphs with different film properties depending on N content: solid solution phase Ta, hexagonal Ta 2 N, hexagonal T aN, cubic TaN, hexagonal Ta 5 N 6 tetragonal Ta 4 N 5 and orthorhombic Ta 3 N 5 [27] T he growth of the insulati ng Ta 3 N 5 phase during grow th of TaN by CVD and ALD results in an increase in electrical resistivity of films [14]

PAGE 23

23 2. 3 Chemical Vapor Deposition of Tungsten Based Diffusion Barrier 2. 3 .1 Tungsten Nitride as a Diffusion Barrier In previous works, WN x has been deposited using PVD techniques such as reactive sputtering o f a W target under N 2 at mosphere However, this technique results in poor step coverage, a major disadvantage when applied to device structures with high aspect ratio features [28 31] T able 2 1 shows the halide and metal organic precurs ors used for film growth of WN x by CVD WF 6 precursor has been used with NH 3 coreactant to deposit WN x thin films for application as a barrier and glue layer for advanced metallization [32, 33] XRD data shows a consistent (111) orientation of W 2 N cubic structure at 450, 550, 650, and 700 C. The resistivity obtained for these films ranged from about 900 to 2800 cm. T he F content was detected from a maximum 0.9 at. % at 450 C to less than 0.1 at. % at 625 C [32] Addition of H 2 gas to the mixture of W F 6 and NH 3 facilitates the reaction of binary mixtures by breaking a W F bond of WF 6 or N H bond of NH 3 causing a decrease in activation energy for the reaction. XRD data shows only W 2 N was obtained without any diffraction line indicating WN. The N 1 s spec trum from XPS shows two peaks at 397 and 400 eV. The latter peak is ascribed to a N atom or molecule present in interstitial sites of W 2 N. The release of N is due to desorption of N 2 gas when it is heated to a high temperature [33] Another inorganic p recursor WCl 6 has been used for film growth of WN x with a mixture of NH 3 H 2 and Ar at temperatures of 500 to 900 C at 0.1 to 10 Torr. The temperature dependence of the Gibbs free energy shows a preferential reaction with WCl 6 in the temperature range of this study. XPS analysis shows three W 4 f 7/2 peaks at 31.5, 33.6, and 37.2 eV for films at 500 C Although no oxide peaks were observed by XRD, the surface of the film was contaminated with a small amount of oxide [34] The halide precursor s such as WF 6 and WCl 6 required high deposition temperatures ( > 450 C) and incorporated the halogen impurities during film growth. W(CO) 6 has been explored for film

PAGE 24

24 growth of WN x with low impurities a t low temperatures compared to other tungsten based precursors because the binding energy of W CO is low [16, 35, 36] WN x was deposited using W(CO) 6 and NH 3 in the temperature range 250 to 500 C. The film resistivity varied from 590 to 950 cm The growth rate varied from 3 to 1930 /min. Below 450 C, the growth regime shows an Arrhenius type dependence on the deposition temperature. The film growth was kinetically controlled with the activation energy of 1.00 eV. Sheet resistance mea surement and XRD analysis showed that the diffusion barrier (15 nm thick) blocked the diffusion of Cu up to 600 C for 1 h annealing [16] Both results showed that W 2 N film prevented diffusion of Cu up to 600 C, and started to fail at 620 C, while no barrier and the CVD W samples failed at 100 to 150 C and 525 to550 C. Barrier failure at 620 C is thought to be due to the diffusion of Cu via undesired grain boundaries [35] H 2 was premixed with precursor vapor at the reactor inlet and NH 3 as the N source was introduced directly into the chamber through a separate feedthrough. Film depo sited below 275 C was amorphous, while those deposited between 275 and 350 C were polycrystalline. Resistivity as low as 123 cm was obtained with corresponding step coverage better than 90 % in a nominal 0.25 m trench structure with aspect ratio of 4:1 [36] ( t BuN) 2 W(NH t Bu) 2 has been used for film growth of WN x as the single source precursor [37, 38] Polycrystalline WN x thin films were grown by low p ressure MOCVD using ( t BuN) 2 W(NH t Bu) 2 in Ar or H 2 carrier XRD studies showed that the films have cubic structures with the lattice parameter of 4.154 to 4.180 XPS showed the binding energies of the W 4 f 7/2 and N 1 s were 33.0 and 397.3 eV, respectively The secondary ion mass spectrometry ( SIMS ) compositional depth profiling indicated C and O levels were low in the films. Possible reaction pathways were suggested by detecting isobutylene, acetonitrile, hydrogen cyanide, and ammonia using gas chromatog raphy mass spectroscopy (GC MS) and nuclear magnetic resonance (NMR) [37]

PAGE 25

25 Annealing to 700 K cause d the loss of N content from the bulk deposited WN x layer as N 2 [38] Plasma enhanced CVD (PECVD) using WF 6 has been used for film growth of WN x N 2 was used as the N source and H 2 was used to remove F from halide precursor. F, O, and C present in the films were below 1% based on XPS. W 2 N films have good adhesi on to PVD Cu, CVD W, Si, SiO 2 and Si 3 N 4 as observed by tape peel tests. Despite higher step coverage for films deposited at 300 C, XPS indicated F impurity. Rapid thermal annealing (RTA) is used to treat the deposited films to reduce the F impurity le vel [39] W 2 N films were deposited at a wa f er temperature of 350 C on Si, SiO 2 and Ta 2 O 5 with and without an electron cyclotron resonance plasma formed SiO 2 (ECR SiO 2 ) top layer The resistivity of W 2 N films is 190 to 240 cm. XRD patterns are X ray amorphous [40] The resistivity for stoichiometric W 2 N, W rich W 2 N (x > 1.0) and N rich W 2 N (x < 1.0) is different. T he resistivity of W rich W 2 N is 145 cm and that of N rich W 2 N 3000 to 5000 cm. The decrease of N levels in W 2 N due to N 2 desorption is confirmed by AES [41] T he resistivity of as deposited films is 95 to 100 cm. In order to improve the adhesion strength of CVD W films W 2 N glue layer is in terposed between W and Si. T he number of vacancies at N lattice sites is reduced because N atoms occupy interstitial positions in the W lattice. The more adhesive contact is due to N interstitials due to the modification of the structural properties such as porosity and vacancies in the W 2 N [42] Diffusion barrier test results from SEM and XRD using Cu/WN x /SiCOH/Si stacks showed that W 2 N film s were stable up to 500 C. Above 600 C, WO 3 nanorods w ere grown from the sample surface due to the residual O in the films [15] 2. 3 .2 T ungsten Carbon itride as a Di ffusion Barrier Table 2 2 shows that the metal organic precursors that w ere used for film growth of WN x C y by CVD. [W( N t Bu)Cl 2 (H 2 N t Bu)] 2 [W(N t Bu)Cl 2 (TMEDA)] (TMEDA = N N N N

PAGE 26

26 tetramethylethylenediamine) [W(N t Bu)Cl 2 (py) 2 ] (py = pyridine) and [W(N t Bu) 2 Cl(N {SiMe 3 } 2 )] have been used to deposit WN x C y in N 2 carrier at the deposition temperature of 550 C. Those compo unds can be used for film growth as single source precursors or dual source in the presence of NH 3 In all cases the Cl levels present in the films w ere less than 1 at. %. Film growth using NH 3 shows lower O level and no change in C content of the result ing films. XRD pattern of all the films indicated the formation of WN x C y SEM surface images of films suggest an island growth mechanism. T he films were uniform, adhesive, abrasion resistant, conformal and hard, being resistant to scratching with a steel scalpel [43] WH 2 ( i PrCp) 2 and WH 2 (EtCp) 2 ha ve been used for film growth of WN x C y in NH 3 /H 2 /N 2 carrier. Film growth was carried out on SiO 2 sub strates using N 2 carrier gas at temperature range 350 to 400 C. NH 3 (99.96 %) and H 2 (99.9999 %) were used as reactant. The W 4 f 7/2 and W 4 f 5/2 peaks at 31.6 and 33.8 eV are well matched with the WC x phase. The C 1 s peak located at 283.2 eV is well mat ched with the carbidic form. XRR and XRD analyses show no peaks indicating crystallization. The addition of NH 3 causes the O incorporation to decrease significantly. The l owest value of resistivity was 565 cm when no coreactant was used at 350 C. T his is correlated with the decrease of the C level present in the films. The addition of NH 3 causes an increase of the film resistivity because mobility is reduced by the scattering effect of incorporated N atoms [44] The tungsten isop r opylimido complex Cl 4 (CH 3 CN)W(N i Pr) has been used for film growth of WN x C y as a single source precursor [45, 46] The precursor structure was chosen so that the W N multiple bond of the precursor would survive while the ancillary ligands and the isopropyl imido substitu en t dissociated under CVD. Film microstructure at a temperature below 500 C w as X ray amorphous, with the minimum value of film resistivity (750 cm) and sheet resistance (47 / ) of this study occurring for CVD at

PAGE 27

27 450 C. Film growth rate varied from 10 to 27 /min within a temperature range of 45 0 to 700 C. The apparent activation energy for film growth in the kinetically controlled regime was 0.84 eV. C levels increased from 12 to 49 at. % in the temperature range 450 to 700 C. Fragmentation of li gands and solvent would leave C conta in ing moieties at the film surface, indicating C incorporation into the WN x film [45] WN x C y thin films w ere deposited using solutions of Cl 4 (CH 3 CN)W(N i Pr) in 1,2 dichlorobenzene (1,2 DCB). The results show the solvent affected deposition of C into the films in comparison with the films deposited with solutions of Cl 4 (CH 3 CN)W(N i Pr) in benzonitrile (PhCN). The increased N levels for films from PhCN solutions suggest that the nitrile (C N) group was a significant C source. The activation energy for film growth from PhCN solutions weas 0.70 eV, while that from 1,2 DCB solutions was 1.0 eV. This shift in activation energy upon changing the solvent is evidence for an alter native C depositi on process [46] The tungsten phenylimido complex Cl 4 ( Ph CN)W(NP h ) has bee n used for film growth of WN x C y as a single source precursor [47] Film growth rate s varied from 2 to 21 /min in the temperature range 475 to 750 C T he apparent activation energy for film growth in the kinetically controlled regime wa s 1.41 eV. Film microstructure was X ray amorphous below 500 C, with minimum film resistivity (225 cm) and sheet resistance (75 / ) observed for CVD at 475 C. Films deposited from Cl 4 (CH 3 CN)W(N i Pr) exhibited higher growth rates and higher N level in the same temperature range. These differen t results are due to the higher dissociation energy of the imido N C bond in Cl 4 ( Ph CN)W(NP h ) F ilms from Cl 4 (CH 3 CN)W(N i Pr) are superior to th ese from Cl 4 ( Ph CN)W(NP h ) for diffusion barrier application s due to lower amorphous deposition temperature, lower sheet resistance, and higher N level [45, 4 7] The tungsten isorpopylimido complex Cl 4 (CH 3 CN)W(N i Pr) has been used for film growth of WN x C y in NH 3 /H 2 carrier AES results initiated that film s deposited

PAGE 28

28 with NH 3 had higher N levels for low deposition temperature (450 550 C) along with decreas ed C and O levels as compared with films deposited without NH 3 Film microstructure was X ray amorphous for film deposited with NH 3 in contrast to polycrystalline for phase present in the films deposited without NH 3 at 500 C. An i ncrease in N level in the amorphous films would increase film resistivity because the film resistivity is higher for WN x phase relative to WC x phase and re placement of C by additional N causes electron scattering Film growth in the presence of NH 3 was mass transfer controlled across the entire temperature range (450 700 C) while film growth in the absence of NH 3 had a kinetic to mass transfer control transition point near 600 C [48] A mixture of t he tungsten allylimido complex Cl 4 (CH 3 CN)W(N C 3 H 5 ) and Cl 4 ( Ph CN)W(N C 3 H 5 ) has been used for film growth of WN x C y in the presence and absence of NH 3 [18, 19] Cl 4 ( Ph CN)W(N C 3 H 5 ) was not isolated but was produced in situ by the substitution of the acetonitrile ligand of Cl 4 (CH 3 CN)W(N C 3 H 5 ) with PhCN. The rapid rate of exchange of nitrile ligands in Cl 4 (CH 3 CN)W(N C 3 H 5 ) ensures that the precursor is completely converted to Cl 4 ( Ph CN)W(N C 3 H 5 ) before film growth starts. Films deposited from a mixture show X ray amorphous phase below 550 C. Film growth rate varied from 5 to 10 /min in the temperature rang e 450 to 650 C, and the apparent activation energy for film growth was 0.15 eV. The values of activation energy for film growth using Cl 4 ( R CN)W(N R ) [R = CH 3 Ph, and R = Ph, i Pr, allyl] against the N C plotted against the bond strengths for the amines R NH 2 is linear The linear relationship between activation energy for film growth using Cl 4 ( R CN)W(N R ) [R = CH 3 Ph, and R = Ph, i Pr, allyl] and the N C bond dissociation energy for the amines R NH 2 suggests that cleavage of the N C bond is the rate det ermining step in film growth. The strength of the N C imido bond has an effect on the amount of N incorporated in the film [18] Films deposited at 450 C with NH 3 as a coreactant showed 23 at. % N level which is higher than film

PAGE 29

29 growth without added NH 3 (4 at. % in N level). O incorporation remained below 6 at. % in the temperature range 450 to 750 C The films deposited below 500 C were X ray amorphous and t he X ray diffraction patterns suggest that either the mixture of W 2 N and W C 1 x or the solid solution WN x C y exist in the films The presence of Cl in the precursor raises the concern of Cl free films. XPS spectra show no Cl peaks were observed either the Cl 2 s (270 eV) or Cl 2 p 3/2 (199 eV), confirming that Cl level in the films was lower than the XPS detection limit (~ 1 at. %). An apparent activation energy for films with added NH 3 is 0.34 eV, as compared with the value of 0.15 eV for films without NH 3 The film resistivity for films depos ited with NH 3 exhibited higher film resistivity, with the lowest film resistivity of 1700 cm observed for films deposited at 550 C [19] 2. 4 Atomic Layer Deposition of Tungsten Based Diffusion Barrier 2. 4 .1 Tungsten N itride as a Diffusion Barri er T able 2 3 shows that the halide and metal organic precursors that w ere used for film growth of WN x by ALD. The WF 6 precursor has been used with NH 3 coreactant to deposit WN x thin films for application as a Cu diffusion barrier layer for advanced metall ization [26, 49 52] WN x on SiO 2 was deposited at 350 C. The growth rate was fairly high, saturating at a level of 0.42 /cycle. Even though t he F impurity was as low as 2.4 at. % in the film t he value of film resistivity is 4500 cm. Introduction of a third precursor between WF 6 and NH 3 pulses cause d improve d reduction of W and reduce d the formation of HF in order to reduce the resistivity of WN and avoid Cu pitting [26] WN x was deposited on Si and tetraethylothosilicate (TEOS)/Si substrates in the temper a ture range 200 to 400 C, synchronizing the NH 3 plasma (NH NH + NH 2 + NH 3 + and H + ) instead of NH 3 gas at the NH 3 exposure cycle s during ALD. T h e conventional ALD shows that a 22nm thick W layer is deposited and a 3 nm thick WN x layer appears on the top of this W layer durin g the 100 cycles exposing WF 6 and NH 3 AES depth

PAGE 30

30 profile s for films deposited by pulse plasma enhanced ALD show a uniformly distributed N concentration in the WN x films on Si and non Si surface s WF 6 either reacts with Si quickly due to the catalytic rea ction of Si, forming a thick W layer instead of WN x or does not adhere to the non Si surface s High resolution transmission electronic microscopy (HRTEM) reveals that WN x (22 nm thick) in the Cu/WN x /Si stack prevents Cu diffusion during the annealing pro cess at 700 C for 30 min [49] The deposition rate was about 3 /cycle at 350 C. There are two different growth regimes: one is the incubation regime and another is the linear and self limiting growth regime. Rutherford backscattering spectroscopy (RBS) revealed that WN x (22 nm thick) in the Cu/WN x /Si stack prevents Cu diffusion during the annealing process at 600 C for 30 min [50] Alternating exposures of NH 3 (A) and WF 6 (B) in an AB reaction sequence were used to deposit the WN x at the substrate temperature between 323 and 523 C Transmission Fourier transform infrared (FTIR) spectroscopy studies indicated that NH 3 and WF 6 surface reactions were complete and self limiting at deposition temperature over 32 3 C. AFM images exhibit a root mean square (rms) roughness of 0.61 nm. The rms roughness of initial SiO 2 on Si(100) was 0.25 nm. The XPS spectra indicate that the surface of WN x exhibited characteristic signals for W, C, N, F, and O. The XPS depth profiling reveals that the WN x films had a W to N ratio of ~ 3:1. The films also contained 5 at. % C and 3.6 at. % O. Glancing ang le XRD results indicate the films consisted of W 2 N crystallites with a diameter 11 nm and a (111) texture [51] Successive exposure to WF 6 and Si 2 H 6 (or NH 3 ) in an ABAB reaction sequence produced W (or W 2 N) deposit ion at substrate temperature 152 423 C (or 323 523 C). Between the WF 6 and Si 2 H 6 reactant exposures, the deposition chamber was purged with N 2 for several minutes. Si 2 H 6 serves only a sacrificial role to remove surface species without incorporat ion into the film [52] ( t BuN) 2 (Me 2 N) 2 W precursor has been used with NH 3 coreactant for film growth of WN x

PAGE 31

31 [53 55] WN x barrier films w ere deposit ed by ALD using ( t BuN) 2 (Me 2 N) 2 W and NH 3 in the temperature range 250 to 350 C. Film microstructure was X ray amorphous as deposited and 100% step coverage was obtained inside holes with aspect ratio greater than 40:1. RBS showed that O was not detected and C was less than the detectable limit ( < 0.5 at. %). WN x film (1.5 nm thick) proved to be good barriers to the Cu diffusion for temperature up to 600 C. Numerous crystals of Cu 3 Si were observed due to complete breakdown of the barrier for a sample an nealed at 650 C. RBS and XPS confirmed the loss of N in the annealed film at temperature s greater than 725 C indicating the WN x was converted to pure polycrystalline W [53] Films deposited above 350 C contained C in addition to W and N and their step coverage is not as good as that for films deposited within the range 250 to 350 C. The films deposited at 400 C were more conductive, 42 0 cm. No films were deposited at deposition temperature below 250 C ALD of Cu on the WN x could not be removed by adhesive tape applied to the Cu [54] ALD has been used to seal porous low material with silica (4 nm thick) and to add a WN x diffusion barrier (1.0 nm thick) a Co adhesion layer (1.0 nm thick) and a Cu seed layer (10 nm thick) Tape pull tests showed the Cu/Co/WNx/silica/low / Si stack has good adhesion. Samples annealed at 400 C for 30 min showed no agglomeration of Cu observed by SEM and no diffusion of Cu detected by RBS [55] 2. 4 2 Tungsten Carboni tride as a Diffusion Barrier Table 2 4 shows the halide and metal organic precursors that w ere used for film growth of WN x C y by ALD. The properties of WN x C y films deposited by ALD using WF 6 NH 3 and TEB as a source gases were characterized as a diffusion barrier for Cu metallization [13, 17, 56 63] ALD WN x C y wa s deposited in the temperature range 275 to 325 C by supplying WF 6 NH 3 and triethylboron [B(C 2 H 5 ) or TEB] in c yclic pulses. The growth rate wa s 0.8 /cycle; the WN x C y wa s conductive (300 400 cm) and dense (15.4 g/cm 3 ). XPS spectra indicate that C in ALD

PAGE 32

32 WN x C y is in the WC x phase, which is mor e conductiv e than the WN x phase [13] The films deposited at 313 C show resistivit ies of about 350 cm with densit ies of 15.4 g/cm 3 The chemical composition measured by RBS shows W, C, and N of 48, 32, and 20 at. %, respectively. TEM analysis shows that the as grown film was composed of a face centered cubic (fcc) phase with a lattice parameter simi lar to both WN 2 and W C 1 x with an equiaxed microstructure. Diffusion barrier test results show that ALD WN x C y film s (12 nm thick) deposited between Cu and Si failed after annealing at 70 0 C for 30 min The superior diffusion barrier performance is t he consequence of both the formation of films with equiaxed microstructure and high density [17, 60, 61] Film morpholo gy by AFM reveals island growth and fractal behavior of individual ALD WN x C y on the methyl terminated self assembled monolayers (SAMs) for film deposited at 300 C Initially, the film grows by deposition of WN x C y on the substrate defect sites. This depo sition causes increase d surface area and as a result film roughness increases. This situation continues until the film coalesces where the surface area is reduced and accordingly film roughness decreases. The film area and roughness become constant when the substrate is completely covered. Th is growth of ALD WN x C y is enhanced on N containing surfaces such as N 2 plasma treated SILK polymer films because of good binding states for TEB [56] TEM analysis reveals the island growth of individual ALD WN x C y nanocrystals on the PECVD SiO 2 during early stages of film growth. The capacitance voltage (C V) measurements after bias temperature stressing (BTS) reveal that WN x C y thin fil m (5.2 nm thick) acts a good diffusion barrier for Cu migration [57] WN x C y growth on SiC is similar to that on PECVD SiO 2 T h is is due to the presence of a C rich layer from TEB precursor decomposition [58] WN x C y was deposited at 300 C in a process sequence using WF 6 TEB, and NH 3 as precursors The bulk resistivity of WN x C y has low resisi tivity about 300 400 cm. XPS results show a ratio of

PAGE 33

33 W:N:O:C of 60:20:10:10 at. % throughout a very homogeneous layer. The rms roughness of 0.47 nm was determined for a WN x C y layer with AFM [59] WN x C y was deposited by introducing TEB as a reducing agent for W. WN x C y shows excellent film properties: good compat ibility with the Cu metal, strong adhesion on the Cu surface, and no pitting on the Cu surface. The growth rate is 0.08 nm/cycle and it remains constant in the temperature range 300 to 350 C (ALD window). XPS spectra indicate the chemical composition of W:N:C is 55:15:30. Boron (B) residues were below the XPS detection limit (0.5 at. %). F level s w ere below 2 at. % for films deposited from 225 to 400 C. XRD results show that the crystalline phase is W C 1 x The resistivity is as low as 210 cm, i ndicating that C is bound in the WC x phase [62] TEM image s show that the step coverage of WN x C y barrier film is nearly 100 % in one via in a via chain. AES analysis indicates that the chemical compo sition of W, C, and N is 57, 30, and 13 at. %. The resistivity was 600 to 900 cm [63] W 2 (NMe 2 ) 6 precursor has been used with NH 3 coreactant for film growth of WN x C y thin films between 150 and 250 C. NH 3 was used as a N source and Ar was used as the carrier and purge gas. At 180 C, surface limited growth was achieved with W 2 (NMe 2 ) 6 pulse lengths over 2.0 s. Shorter pulse of W precursor results in sub saturative growth and lower growth rate. The ALD window was detected at the deposition temperature between 180 and 210 C. XPS spectra indicate that W 4 f 7/2 binding energy was 31.5 eV, which is well matched with the binding energy of the WC x and WN x phases. The binding energy of C 1 s at 282 eV and N 1 s at 397.6 eV are consistent with C in carbides and N in nitrides. The binding energy of O 1s was 530. eV. Films deposited at 180 C exhibited a resistivity value of 810 50 cm. The resistivity of WN x films is sensitive to the W to N ratio. Further exposure of the same film to ambient atmosphere (an additional 30 days) caused an increase in film resistiv it y values over 10000 cm. XRD results for film deposited at 180 C indicates X

PAGE 34

34 ray amorphous microstructure. AFM analysis shows that t he rms roughness (2 m by 2 m area) was 0.9, 0.8, and 0.7 for films deposited at 150, 180, and 210 C, respectively [64] ( 5 C 5 H 5 ) W ( CO ) 2 NO precursor has been used with NH 3 coreactant to deposit WN x C y thin films by PEALD [65, 66]

PAGE 35

35 Table 2 1. Precursors used for film growth of WN x by CVD Technique Precursor Coreactant Reference CVD WF 6 NH 3 [32] CVD WF 6 NH 3 + H 2 + Ar [33] CVD WCl 6 NH 3 + H 2 + Ar [34] CVD W(CO) 6 NH 3 + Ar [16, 35] CVD W(CO) 6 NH 3 + H 2 [36] CVD ( t BuN) 2 W(NH t Bu) 2 H 2 or Ar [37] CVD ( t BuN) 2 W(NH t Bu) 2 [38] PECVD WF 6 NH 3 + H 2 +N 2 [39] PECVD WF 6 NH 3 + H 2 [40 42] PECVD W(CO) 6 NH 3 [15] Table 2 2. Pr ecursors used for film growth of WN x C y by CVD Technique Precursor Coreactant Reference CV D [W( N t Bu) (N t Bu)Cl 2 (H 2 N t Bu)] 2 NH 3 + N 2 [43] CV D [W(N t Bu) 2 Cl 2 (TMEDA)] NH 3 + N 2 [43] CV D [W(N t Bu) 2 Cl 2 (py) 2 ] NH 3 + N 2 [43] CV D [W(N t Bu) 2 Cl(N {SiMe 3 } 2 )] NH 3 + N 2 [43] CVD WH 2 ( i PrCp) 2 NH 3 + H 2 + N 2 [44] CVD WH 2 (EtCp) 2 NH 3 + H 2 + N 2 [44] CVD Cl 4 (CH 3 CN)W(N i Pr) H 2 [45, 46] CVD Cl 4 (PhCN)W(NPh) H 2 [47] CVD Cl 4 (CH 3 CN)W(N i Pr) NH 3 + H 2 [48] CVD Cl 4 (CH 3 CN)W (NC 3 H 5 ) H 2 [18] CVD Cl 4 (PhCN)W(NC 3 H 5 ) H 2 [18] CVD Cl 4 (CH 3 CN)W(NC 3 H 5 ) NH 3 + H 2 [19] CVD Cl 4 (PhCN)W(NC 3 H 5 ) NH 3 + H 2 [19]

PAGE 36

36 Table 2 3. Precursors used for film growth of WN x by ALD Technique Precursor Coreactant Refer ence ALD WF 6 NH 3 [26, 49 52] ALD ( t BuN) 2 (Me 2 N) 2 W NH 3 [53 55] Table 2 4. Precursors used for film growth of WN x C y by ALD Technique Precurs or Coreactant Reference ALD WF 6 NH 3 + (C 2 H 5 ) 3 B [13, 17, 56 62] ALD WF 6 NH 3 [63] ALD W 2 (NMe 2 ) 6 NH 3 + Ar [64] PE ALD ( 5 C 5 H 5 ) W ( CO ) 2 NO NH 3 [65] PEALD ( 5 C 5 H 5 ) W ( CO ) 2 NO NH 3 [66]

PAGE 37

37 Figure 2 1. Microstructure of Cu diffusion barrier materials: A) single crystal; B) polycrystalline; C) polycrystalline columnar; D) nano crystalline; E) amor phous. Adopted from A. Kaloyeros and E. Eisenbraun, Ultrathin diffusion barrier/liners for gigascale copper metallization Annu. Rev. Mater. Sci. 30 (2000) 363 385 A) B ) C ) D ) E )

PAGE 38

38 Figure 2 2. Diagram showing the applications of metals and nitrides in mo dern semiconductor devices. Adopted from H. Kim, Atomic layer deposition of metal and nitride thin films: Current research efforts and applications for semiconductor device processing J. Vac. Sci. Technol. B 21 (2003) 2232 2261 Cu diffusion barrier/adhesion promoter Cu seed layer Tungsten plug for via hole Diffusion barrier Metal gate electrode

PAGE 39

39 Figure 2 3. Simplified processing steps in dual damascene structure for Cu metallization

PAGE 40

40 CHAPTER 3 EXPERIMENTAL PROCEDURE 3.1 Precursor Synthesis T he diorganohydrazido(2 ) tungsten complexes Cl 4 (CH 3 CN)W(NNR 2 ) ( 1 : R 2 = (CH 2 ) 5 ; 2 : R 2 = Ph 2 ; 3 : R 2 = M e 2 ) w ere prepared as described in the literature [24] 3.2 Film Growth The each precursor was dissolved in benzonitrile (PhCN) in a concentration of 8.1 mg/m L 9.6 mg/m L and 7.4 mg/m L for 1 2 and 3 filled into a gas tight syringe and pumped into a nebulizer A quartz plate in the nebulizer vibrates at a frequency of 1.44 MHz generating a mist of precursor and solvent. Carrier gas flow s through the nebulizer assembly and transports the aerosol through the capillary tube from the syringe into a heated imping ing jet. The mixture of precursor and PhCN flows from the showerhead to reach the substrates on a heated graphite su s ceptor. A custom built vertical quartz cold wall CVD reactor system shown in Figure 3 1 was used to deposit the thin films on p type boro n doped Si (100) single crystal substrates with electrical resistivities in the range cm. A graphite susceptor was heated by radio frequency (rf) induction coils to maintain the substrates at the specific deposition temperature. The deposition te mperature was varied from 300 to 700 C in step s of 50 C. The operat ing pressure was maintained at 350 Torr using a mechanical roughing pump and pressure control valve. The H 2 (99.999 %, Airgas) carrier gas flow rate was 1 000 sccm (sccm denotes cubic ce ntimeters per minute at STP) the NH 3 (99.9999 %, Air Liquide) coreactant flow rate was 30 sccm and the d eposition time for all depositions was 150 min. 3.3 Film Characterization s Several methods were used to characterize the composition, chemical bondin g states, microstruture, surface morphology, growth rate, and electrical properties of the films. X ray

PAGE 41

41 photoelectron spectroscopy (XPS) was used to identify the chemical composition and the chemical bonding states of the elements in the film using a Perk in Elmer PHI 5600 ESCA system. XPS spectra were obtained by m onochromatic Mg ionizing radiation (1254 eV) with the X ray source operating at 300 W (15 kV and 20 mA). Prior to XPS measurement, Ar + ions were used to sputter as deposited samples for 10 min to remove residual surface contamination. X ray diffract ion (XRD) was used to identify the film microstructure with a Philips APD 3720 system, operating with Cu radiation (40 kV and 20 mA). XRD was performed from 30 to 80 2 with 0.02 step size Atomic force microscope (AFM) was used to measure the surface roughness with a Digital Instruments Dimension 3100 system, operating in tapping mode. AFM was performed with 2 Hz scan and with 512 by 512 resolution Cross sectional scanning electron microscop y (SEM) was used to measure the thickness of the film on a JEOL JSM 6335F to obtain the growth rate. T he sheet resistance of the film was measured by the four point probe method using an Alessi Industries four point probe to obtain film resistivity along with thickness from cross sectional SEM images. 3.4 Diffusion Barrier Te sting Cu ( 100 nm thickness ) was deposited by reactive sputtering using a Kurt Lesker CMS 18 S putter system at room temperature. Samples of WN x C y (15 20 nm thickness ) deposited by CVD at 400 C on the Si (100) single crystal substrates were loaded via a l oad lock system into the process deposition chamber with a base pressure of 3 10 7 Torr. The chamber pressure during deposition was 5 mTorr. The forward sputtering power for Cu was 200 W, while the WN x C y /Si stacks were rotated at 20 rpm during depositi on. Cu was deposited on top of WN x C y /Si s tack s to evaluate the ir performance as Cu diffusion barrier s The Cu/WN x C y /Si s tacks were then annealed in the CVD reactor at 500 600, and 700 C for 30 min/step

PAGE 42

42 Annealing was performed under N 2 (99.999 %, Prax air) to protect the Cu layer from oxidation. XRD and four point probe were used to investigate the onset of the failure process via the formation of Cu 3 Si. SEM imaging was used to reveal the Cu surface morphology. Cross sectional t ransmission electron mic roscope (TEM) imaging was used to detect the presence of Cu 3 Si in the WN x C y /Si interface. Energy d ispersive X ray s pectroscopy (EDS) qualitative analysis was used to identify the presence of Cu K signal in the WN x C y /Si interface. The cross sectional ima ge was taken by TEM using JEOL TEM 2010F to allow for high resolution imaging of multilayered interfaces. Prior to TEM imaging, f ocused i on b eam (FIB) was used to prepare samples for cross sectional TEM using FEI Strata DB 235 to allow for precise cross s ectioning in specific location. FIB was operated with a finely focused beam of Ga + ions.

PAGE 43

43 1 2 3 Figure 3 1. T h e diorganohydrazido(2 ) tungsten complexes Cl 4 (CH 3 CN)W(NNR 2 ) ( 1 : R 2 = (CH 2 ) 5 ; 2 : R 2 = Ph 2 ; 3 : R 2 = Me 2 )

PAGE 44

44 Figure 3 2 Schematic d iagram of the aerosol assisted CVD system

PAGE 45

45 Figure 3 3. Process flow on film properties. (MAIC, http: \ \ maic.mse.ufl.edu, October, 2008) XRD CVD H 2 (or NH 3 ) CVD H 2 (or NH 3 ) Si WN x C y Si WN x C y FE SEM XPS AFM 4PP AES Material Characterization Diffusion Barrier CVD

PAGE 46

46 Figu re 3 4. Process flow on diffusion barrier testing. (MAIC, http: \ \ maic.mse.ufl.edu, October, 2008)

PAGE 47

47 CHAPTER 4 DEPOSITION OF WN x C y FROM Cl 4 (CH 3 CN)W(N pip) AS A SINGLE SOURCE PRECURSOR 4.1 X ray Crystallographic Study of Cl 4 (CH 3 CN)W(N pip) Single crystals suitable for X ray diffraction were obtained from c ompound 1 and subjected to X ray crystallographic structure determination (Table 4 1). The solid state structure of 1 reveals the W metal center in a distorted octahedral geometry ( Figure 4 1). Four Cl atoms occupy the basal positions with the W Cl bonds averaging 2.34 (Table 4 2) which is within the expected range for W (VI) Cl bonds [67] The diorganohydrazido(2 ) ligand is strongly bound to the central metal atom as indicated by a short W(1) N(1) distance of 1.752(3) The short N( 1) N(2) bond distance (1.265(4) ) within the hydrazido ligand suggests a high degree of delocalization and multiple bond character throughout the W(1) N(1) N(2) unit. This phenomenon has been reported in the literature for other hydrazido complexes of W with 5 C 5 Me 5 )Cl 3 (NNPh 2 ) [68] (W N 1.769(2) N N 1.296(3) ), cis [WCl 3 (NNH 2 )(PMe 2 Ph) 2 ] [69] (W N 1.752(10) N N 1.300(17) ) and (CH 3 CN)Cl 4 W(NNMe 2 ) [70] (W N 1.769(5) N N 1.271(8) ). The piperidyl unit can be found in a typical chair l ike conformation with the C(1) C(5) N(2) N(1) unit exhibiting a trigonal planar arrangement as evidenced by the sum of the bond angles (359.7) around N(2). The remaining coordination site is occupied by a neutral acetonitrile solvent molecule. The W(1) N(3) bond distance of 2.237(3) is significantly shorter than those reported for related tungsten imido compounds [18] suggesting a decreased trans influence of the diorganohydrazido(2 ) ligand compared to the imido moiety. 4.2 Preliminary P recursor S creening Multiple spectroscopic techniques were applied to evaluate the viability of 1 as a precursor for WN x C y deposition [70] Kinetic data obtained by 1 H NMR spectroscopy confirmed an

PAGE 48

48 expected weak bond between the acetonitrile lig and and the metal center. Positive ion chemical ionization (CI) mass spectrometry was performed to obtain some insight into the fragmentation behavior of 1 The absence of molecular ion peaks is in good agreement with the acetonitrile ligand being labile Mass envelopes containing the piper i dine moiety ([pip] + and [H 2 pip] + ) were observed in high abundance, suggesting that cleavage between the two hydrazido N atoms is facile under high energy conditions (ionization during MS or pyrolysis during CVD). Ther mal behavior studies of 1 via thermogravimetric analysis (TGA) showed a drop in mass corresponding to loss of the acetonitrile ligand at approximately 100 C. 4.3 Film S tructure The XRD spectra in Figure s 4 2 A and 4 2 B show amorphous and polycrystalline microstructure for films deposited at 300 and 700 C, respectively. The four characteristic polycrystalline peaks exhibit locations that are consistent with WN x C y The XRD spectra contain no evident characteristic peaks observed up to 45 0 C, indicating that amorphous films were deposited from 300 to 45 0 C. The polycrystalline peaks appearing in spectra of material deposited from 5 0 0 to 700 C indicate no preferred crystal orientation. Primary peaks at 37.24 and 42.98 2 are consistent with (111) and (200) orientation, while the other peaks at 62.58 and 74.98 2 are from (220 ) and (311), respectively. The XRD spectra in Figure 4 2 C show the evolution o f film crystallinity with increasing deposition temperature. As the deposition temperature increases to 700 C, the (111) and (200) WN x C y peaks sharpen further. All of the films show three additional sharp peaks at 33.08, 61.76 and 69.14 2 indicati ng Si(200) Si(400) and Si(400) radiation, respectively. Additionally, all films displayed one peak at 65.99 2 representing W L radiation. This W peak comes from deposition on the target due to evaporation of the W filament [71] The ability to deposit amorphous films of WN x C y at low

PAGE 49

49 temperature is highly significant for diffusion barrier applications since the formation of polycrystalline films facilitates diffusion of Cu to the underlying Si via the gra in boundaries. 4.4 Chemical C omposition XPS results for chemical composition (Figure 4 3) show that W, N, C and O were present in the films. No Cl contamination in the films w as observed within detection limit of XPS (~ 1 at. %). The W level is high est between 450 and 500 C, while the N, C and O levels are fairly steady in this range. From 300 to 400 C, the C level is below 10 at. %, with the lowest level of 6 at. % for depositions at 400 C. Between 500 and 700 C, the C level increases gradual ly from 15 to 67 at. %. The overall trend for C content is consistent with the faster decomposition of hydro carbon groups in both the precursor and the solvent as the growth temperature increases, leading to C incorporation into the film. As the deposit ion temperature increased from 300 to 400 C, the N level increased from 10 to 18 at %. However above 500 C, N levels start to decrease, as a consequence of increased C concentration in this range. When the deposition temperature reaches 700 C, the N level has declined to 5 at. % due to the steep rise in C levels at high deposition temperatures. Typically, refractory metal nitride diffusion barriers show a decreasing tendency of N incorporation with increasing deposition temperature due to N 2 desorpti on [29, 33, 38, 72] Films deposited at 300 C s how over 2 0 at. % of O which decreased drastically to 14 at. % at 450 C. As the deposition temperature increased from 450 to 700 C, the O level decreased gradually to 5 at. %. From XRD spectra in Figure 4 2 C the polycrystalline microstructure becomes evident for depositions performed at 500 C. As the deposition temperature increases, the average grain size increases as well. As the film starts to crystallize, the microstructure gets denser, which inhibits post growth of O interdiffusion into the l attice of the film [73] This result comes from the densification of film by polycrystal grain growth between 500 to 700 C.

PAGE 50

50 4.5 Chemical B onding S tates XPS was used to measure the binding energy (BE) of atoms in the films. The W 4 f photoelectron line is a doublet due to spin orbit splitting into W 4f 7/ 2 and W 4f 5/2 while the N 1 s C 1 s and O 1 s photoelectron lines show a single peak. Figure 4 4 A indicates the evolution of XPS spectra in the W 4 f BE region as the deposition temperature changes. The major W 4 f7/2 and W 4 f 5/2 peaks at 400 C are at 31. 6 and 33.5 eV, which are close to values for WC x and WN x (Table 4 3). These two values are higher than those for metallic W and lower than those for WO 3 These BE values are consistent with W in the WN x C y chemical bonding state. As the deposition temp erature rises to 700 C, the BE increases to 31.8 and 33.8 eV, which correspond to WC x and WN x as well. This slight increase in BE comes from more carbon laden samples at the higher temperature. For material deposited at 300 C, the BE more closely resemb les that of WO 3 but shifts toward values for WC x or WN x as the deposition temperature rises to 350 C This indicates a chemical bonding state from WO 3 dominant to WN x C y as the deposition temperature increases from 300 to 700 C. Figure 4 4 B illustrates the change of the XPS spectra in the region of the N 1 s BE as the deposition temperature changes. The N 1 s peaks are observed at 397.3 eV, which is the reported value for WNx (Table 4 3). Hence, the N in the films is all bound in the WN x polycrystals. N at the grain boundary can be excluded due to a single N 1 s peak without a second peak near 399 eV. The maximum N intensity is seen in the spectra of films deposited at 400 C, which is consistent with the high N level in those films. Figure 4 4 C indica tes the evolution of the XPS spectra in the region of the C 1 s BE as the deposition temperature changes. As the deposition temperature increases to 700 C, the C 1 s peaks exhibit higher intensity. At lower deposition temperature, the C 1 s peaks occur at 283.1 eV, while at higher temperature, the C 1 s peaks are shifted to higher BE. Between 300 and 600 C, the BE is lower than reported for amorphous C, while over 600 C, the BE is higher than that of WC x

PAGE 51

51 Deconvolution of the broad C 1 s peak at 700 C in Figure 4 4 C yields two separate peaks, which are at 283.1 and at 284.5 eV. T he peaks at the lowe r BE are due to C in the WN x C y bonding states while t he peaks at high er BE are consistent with amorphous C present outside of the WN x C y nanocrystals. For deposition at temperatures higher than 650 C, a small portion of amorphous C starts to show up with WC x in the film. As the deposition temperature increases to 700 C, much more amorphous C exists with WC x in the film. Figure 4 4 D indicates the depende nce of the XPS spectra in the region of the O 1 s BE as the deposition temperature changes. The O peaks were observed near 530.5 eV, which is consistent with the presence of WO 3 (Table 4 3). O levels are at the maximum in the low temperature film growth d ue to film crystallization and C incorporation. As the deposition temperature increases up to 700 C, the film density is increased by film crystallization and the grain boundaries are infiltrated by C incorporation. 4.6 Lattice P arameter The lattice pa rameter was determined by XRD using the 2 position of the WN x C y (111) diffraction peaks. The WN x C y peak position was calibrated to the Si(400) diffraction peak at 69.14 2 The dashed line at 4.126 in Figure 4 5 indicates the value of the standard lattice parameter for W 2 N and the dash ed dot line at 4.236 in Figure 4 5 indicates the value of the standard lattice parameter for WC 1 x The lattice parameter in Figure 4 5 shows the increasing tendency as deposition temperature increases from 500 to 650 C. The change in lattice parame ter shows a composition change in polycrystals. The main reason that the lattice parameter increases between 500 and 650 C is that C is incorporated into the WN x C y polycrystals, not at the grain boundary. As the indicated by the chemical composition i n Figure 4 3, the C level continues to increase with deposition temperature over 500 C. Between 650 and 700 C, the lattice parameter decreases. At this range, the C level increases and W decreases

PAGE 52

52 with almost no compositional change in N and O. The de crease in lattice parameter between 650 and 700 C indicates the limitation of solubility for C in WN x C y polycrystals [45] These results are consistent with the chemical bonding states in Figure 4 4 C where the C 1 s peak in this range is shifted to higher BE and has a broad peak that can be deconvoluted into two sep arate peaks. The lattice parameter decreases between 650 and 700 C because C exists at the grain boundary, not in the WN x C y polycrystals. 4.7 Average G r ain S ize Average [71] The most dominant WN x C y (111) diffraction peak of the four characteristic polycrystalline peaks was used to Average grain size (Figure 4 6 ) increases between 500 and 6 0 0 C, varying from 31 to 47 As seen in Figure 4 2 C the films were X ray amorphous between 300 and 500 C, which places a limit of 31 on the maximum grain size. The overall tendency of polycrystal grain size increases with the deposition temperature, varying from 500 to 700 C. 4.8 Electri cal R esistivity The variation of film resistivity with deposition temperature is shown in Figure 4 7 The lowest resistivity is 0.9 m cm at 5 50 C and the values of film resistivity fluctuate with the interplay of polycr ys tal grain growth, C content, O content and film thickness between 500 and 700 C As shown in Figure 4 4, an increase in the amorphous C level as the deposition tem perature rises from 650 to 700 C results in an increase in electron scattering, which causes the film resistivity to increase. The highest film resistivity is 9 .4 m cm for films deposited at 350 C. The high N level in those films is consistent with in creased film resistivity in the WN x C y polycrystal structures, due to the higher resistivity for W 2 N relative to WC 1 x

PAGE 53

53 4.9 Film G rowth R ate The growth rate is in the range 2.7 to 29.4 /min, as determined by cross sectional SEM. For films deposit e d between 650 and 700 C, the growth rate increased drastically suggesting a change in the growth mechanism at these temperatures. This observation was also confirmed by the formation of WC x instead of WN x C y Figure 4 8 is consistent with the presence of two growth regimes. The region with the shallow slope is a mass transfer limited growth regime between 450 C and 600 C. The region with the steep slope is a kinetically controlled growth regime between 300 C and 450 C. The apparent activation energ y calculated for the activated process is 0.28 eV. 4.10 Diffusion B arrier T esting XRD measurement, detection of etch pit s, and AES depth profile were used to detect Cu transport through the film. XRD measurement w as used to search for formation of Cu 3 S i which occurs after barrier failure on Si substrates. As shown in Figure 4 9 the XRD data show no Cu 3 Si peaks in the region of 44 to 46 2 Before annealing, only the Cu(111) peak is observed. After annealing, there are three Cu peaks observed: Cu(111) at 43.44 Cu(200) at 50.80 and Cu( 220) at 74.42 2 [74] Cu r ecrystallization upon annealing resulted in increase in the intensity of the Cu related textures due to nucleation of the new grains or growth of preexisting ones in the Cu/WN x C y stacks [71] The e t c h pit test was also used to search for Cu 3 Si on the Si surface which would be evident if there was Cu transport through the barrier Figure 4 10 shows results of the etch pit test on samples before and after annealing. However, there are no etch pit s observed in either sample, indicating that no Cu was interdiffused or intermixed, and the barrier did not fail upon annealing under N 2 at 500 C for 30 min. If there w ere Cu transport through the barrier, defects caused by the form ation of Cu 3 Si on Si substrate would be shown as

PAGE 54

54 inverse pyramidal shaped etch pits after etch pit test. AES depth profile in Figure 4 11 shows only negligible background signals for Cu where there is no Cu transport The sharp interface between WN x C y an d Si indicate that there is no detectable Cu signal at this interface. Thus, WN x C y deposited at 400 C is a viable Cu diffusion barrier material to prevent Cu transport and intermixing in Si during annealing under N 2 at 500 C for 30 min. 4.11 Conclusio n s The tungsten piperidyl hydrazido complex Cl 4 (CH 3 CN)W(N pip) ( 1 ) was used as a single source precursor for film growth of WN x C y to i nvestigate the film properties for diffusion barrier applications. XRD results suggest that films deposited at temperature below 500 C are X ray amorphous and films deposited at higher temperature are polycrystalline. The XPS of the W 4 f bonding state is consistent with W is present in WN x C y and WO 3 For films deposited at the low end of the temperature range, WO 3 predomina tes and as the deposition temperature increases, WN x C y becomes the dominant W species. XRD results, however, do not indicate any WO 3 peaks. The XPS data on the N 1 s bonding state suggest that N is present in WN x while results on the C 1 s bonding state i ndicate that C is present in WC x and amorphous C. For depositions at temperature higher than 650 C, amorphous C coexists with WC x The XPS data on the O 1 s bonding state suggest that O is present in WO 3 and the O levels are the highest for growth temper atures below 400 C. As the deposition temperature varies, the film growth rate changes from 2.7 to 29.4 /min, with the transition from a kinetically controlled growth regime to a mass transfer controlle d growth regime occurring near 550C. Film resisti vity changes with the interplay of polycry s tal grain growth, C content, O content and film thickness. The WN x C y films were evaluated to determine their suitability as Cu diffusion barriers. WN x C y deposited from 1 is a viable Cu diffusion barrier materia l to prevent diffusion of Cu into Si after annealing

PAGE 55

55 under N 2 at 500 C for 30 min. Further diffusion barrier testing and film characterization are underway.

PAGE 56

56 Table 4 1 Crystal data and structure refinement for Cl 4 (CH 3 CN)W(N pip) ( 1 ) Empirical formula C 7 H 13 Cl 4 N 3 W Formula weight 464.85 Temperature (K) 173(2) Wavelength () 0.71073 Crystal system Monoclinic Space group P 2 1 / n Unit cell dimensions () a = 9.7113(15) = 90 b = 14.874(2) = 92.777(3) c = 9.7939(15) = 90 Volume ( 3 ) 1413.0(4) Z 4 Density (Mg/m 3 ) 2.185 Absorption coefficient (mm 1 ) 8.906 F (000) 872 Crystal size (mm 3 ) 0.19 0.11 0.02 range for data collection () 2.49 27.50 Index ranges h k l Reflections collected 9391 Independent reflections 3245 [ R int = 0.0546 ] Completeness to = 24.60 (%) 99.9 Absorption correction Integration Max and min transmission 0.8420 and 0.2825 Data/restrain ts/parameters 3245/0/137 Goodness of fit on F 2 1.155 Final R indices [ I > 2 ( I )] R 1 a = 0.0203, wR 2 b = 0.0541 Largest diff. peak and hol e ( 3 ) 1.309 and 1.035 a R 1 = (|| F o | | F c ||)/ | F o | b wR 2 = [ [ w ( F o 2 F c 2 ) 2 ]/ [ w ( F o 2 ) 2 ] ] 1/2 S = [ [ w ( F o 2 F c 2 ) 2 ]( n p )] 1/2 w = 1/[ 2 ( F o 2 ) + ( mp ) 2 + np ], p = [max( F o 2 ,0) + 2 F c 2 ]/3.

PAGE 57

57 Table 4 2 Selected bond distances () and angles () for Cl 4 (CH 3 CN)W(N pip) ( 1 ) W ( 1 ) N ( 1 ) 1.752(3) N ( 1 ) W ( 1 ) Cl ( 1 ) 92.03(9) W ( 1 ) N ( 3 ) 2.237(3) N ( 1 ) W ( 1 ) Cl ( 2 ) 96.70(8) W ( 1 ) Cl ( 1 ) 2.3609(9) N ( 1 ) W ( 1 ) Cl ( 3 ) 95.00(9) W ( 1 ) Cl ( 2 ) 2.3252(8) N ( 1 ) W ( 1 ) Cl ( 4 ) 98.17(8) W ( 1 ) Cl ( 3 ) 2.3563(9) N ( 3 ) W ( 1 ) N ( 1 ) 178.24(11) W ( 1 ) Cl ( 4 ) 2.3444(8) W ( 1 ) N ( 1 ) N ( 2 ) 176.6(3) N ( 1 ) N ( 2 ) 1.265(4) N ( 1 ) N ( 2 ) C ( 1 ) 120.7(2) N ( 2 ) C ( 1 ) 1.461(4) N ( 1 ) N ( 2 ) C ( 5 ) 120.3(3) N ( 2 ) C ( 5 ) 1.473(4) C ( 1 ) N ( 2 ) C ( 5 ) 118.7(3)

PAGE 58

58 Table 4 3 Reported binding energy (BE) values W 4 f 7/2 W 4 f 5/2 N 1 s C 1 s O 1 s Ref. Metallic W 31.2 31.7 33.4 [75 77] WN x 32.7 33.6 33.3 35.8 396.2 398.2 [76 80] N at grain boundary 399.2 400.0 [33, 34] WO 3 35.5 36.7 37.6 37.8 528.2 531.6 [33, 76 78, 81] W C x 31.6 32.3 33.7 33.9 279.7 283.8 [75 77, 82] Amor phous C 284.2 285.2 [76, 77, 83, 84]

PAGE 59

59 Figure 4 1 Thermal ellipsoids diagram of the molecular structure of Cl 4 (CH 3 CN)W(N pip) ( 1 ) Thermal ellipsoids are drawn at 50% probability. H atoms are omitted for clarity

PAGE 60

60 A ) B ) C ) D ) Figure 4 2 XRD spectra for films deposited on Si(100) in H 2 carrier : A) 300 C, B) 700 C, C) between 300 and 700 C, and D) s tandard diffraction plots for W 2 N and WC 1 x

PAGE 61

61 Figure 4 3 Variation in chemical composition of W, N, C and O content in the films with deposition temperature. Data are measured by XPS after 10 min Ar + ion sputter.

PAGE 62

62 A ) B ) C ) D ) Figure 4 4 Change of binding energies in A) W 4 f B) N 1 s C) C 1 s and D) O 1 s with deposition temperature. Data are from XPS af ter 10 min Ar + ion sputter.

PAGE 63

63 Figure 4 5 Change in lattice parameter with deposition temperature for polycrystalline films deposited from 1 based on WN x C y (111) diffraction peaks. Figure 4 6. Change in average grain size with deposition temperature for polycrystalline films deposited from 1 based on WN x C y (111) diffraction peaks.

PAGE 64

64 Figure 4 7 Change in film resistiv ity with deposition temperature. Data are measured by four point probe. Figure 4 8 Change in growth rate with deposition temperature. Thickness m easured by cross sectional SEM.

PAGE 65

65 Figure 4 9 The performance of d iffusio n barrier by XRD measurement for Cu/WN x C y /Si st acks before and after annealing at 500 C. Figure 4 10 SEM images of Si surface after etch pit test A) before annealing and B) after annealing at 500 C. B) A)

PAGE 66

66 Figure 4 11. The performance of d iffus ion barrier by AES depth profile for Cu/WN x C y /Si st acks after annealing at 500 C

PAGE 67

67 CHAPTER 5 DEPOSITION OF WN x C y FROM Cl 4 (CH 3 CN)W(N pip) : EFFECT OF NH 3 ON FILM PROPERTIES 5 1 Film Structure Fig ure 5 1 A show s the progression of XRD patterns with increasin g deposition temperature for films deposited in NH 3 / H 2 atmosphere The XRD patterns have been compressed to include the results from all nine growth runs, and thus the resolution is diminished in the figure. An analysis of the original data, however, rev eals four reflections, which w ere calibrated to the Si(400) diffraction pea k. T he reflections at 62.02 and 75.48 2 show low intensity, as compared with the primary ones at 37.48 and 42.78 2 T he peak positions in the se patterns are well matched with a two phase mixture of WN 2 and W C 1 x phases or their solid solution WN x C y with the same crystal structure. Both standards exhibit ed a face centered cubic (fcc) structure with similar lattice parameter ( WN 2 : 4.124 W C 1 x : 4.236 ). In addition to the WN x C y peaks, t hree sharp single crystal peaks were detected at 33.08, 61.76 and 69.14 2 and associated with Si(200) Si(400) and Si(400) radiation, respectively T he XRD spectra indicate that amorphous films were deposited from 300 to 45 0 C while polycrystalline materials were deposited at and above 5 0 0 C. The relative intensities of the four characteristic reflections are consistent with random grain orientation and as expected their variation with growth temperature follow the sam e pattern as the film thickness Note that the growth rate data is proportional to film thickness since the growth time is constant for all runs. Primary peaks at 37. 48 and 42. 7 8 2 are assigned to the (111) and (200) orientations while the other two reflections at 62.02 and 75.48 2 are attributed to (220) and (311) orientation s respectively. As the deposition temperature was increased from 500 to 600 C, the peak intensit ies increased, primarily as a result of increased film thickness and possibly changes in crystallinity. The intensities then successively decreased for the 650 and 700 C deposited

PAGE 68

68 samples. This result is in contrast to that of the previous work using only H 2 as the carrier gas in which the growth rate continued to increase with temperature (Fig ure 4 8 ) The decreased intensity when NH 3 is present is likely a result of precursor depletion from parasitic gas phase reactions T ransamination with NH 3 has been p ostulated to remove the hydrocarbon group in the precursor, changing the rate determining step [85] and thus the growth rate of the films deposited in the presence of NH 3 is less than that for H 2 only. 5 2 Surface M orphology The root mean square (rms) roughness of the surface of film s deposited at 300 C in the absence of NH 3 was determined by AFM to be 0.99 nm, with a rise to 17.17 nm for deposition at 600 C. From the AFM micrographs in Fig ures 5 2 C and 5 2 D the rms roughness of the film surface for films deposited in the presenc e of NH 3 was 0.81 nm at 300 C and 1.28 nm at 600 C, indicating addition of NH 3 results in films with smoother surfaces. The increase in surface roughness with the increase in the deposition temperature up to 600 C is accompanied by increased crystallin ity and grain size (Fig ure 5 1). The decrease in roughness is consistent with an amorphous microstructure and more facile migration of absorbed species on the surface at higher deposition temperatures [86] 5 3 Chemical C omposition Despite t he presence of Cl in the precursor n o p eaks were observed for either Cl 2 p 3/2 or Cl 2 s at 199 and 270 eV, respectively ruling out Cl contamination in the films within the detection limit of XPS ( ~ 1 at. %) The absence of Cl signals is consistent with prior computational results on the relate d imido complex Cl 4 (CH 3 CN)W(N i Pr) for which a mechanistic pathway was found for reaction of the H 2 carrier gas with W Cl bonds to produce HCl in the gas phase [87] Fig ure 5 3 A shows the W levels hav e their highest value between 450 and 500 C Between 650 and 700 C the measured W levels in the films deposited with NH 3

PAGE 69

69 are higher than that in films deposited without NH 3 This difference is related with the chemical bonding states of C 1 s in Fig ure 5 4 C Between 650 and 700 C, C 1 s binding energy ( BE ) shifted from lower BE to higher BE, indicating that amorphous C is more dominant than WC x The increased extent of amorphous C formation dilutes the amount of W deposited. Fig ure 5 3 B shows that th e N levels increased over the entire temperature range after the addition of NH 3 The highest N levels for films deposited with NH 3 (24 at. % at 400 C) is greater than that of films without NH 3 (18 at. % at 400 C). The NH 3 coreactant is used as an addi tional N source, which allows deposition of high N level films compared to depositions without NH 3 However, the increase in the flow rate of NH 3 shows no significant variation in N levels [88] As deposition temperature increases up to 700 C the N levels drop gradually F or films deposited at 700 C the N levels in single source deposition indicate an N concentration without including NH 3 is 5 at. % while for films deposited with NH 3 the N increased to 12 at. %. For refractory m etal nitride diffusion barrie rs the N levels generally decrease with increasing deposition temperature in film deposit ed without NH 3 and with NH 3 It has been suggested that t he higher deposition temperature increases the rate of N desor ption as N 2 gas, evidenced by Fig ure 5 3 B [40] The results shown in Fig ure 5 3 C point to lower C levels in films deposi ted with NH 3 than that in single source deposition from 500 to 700 C The decrease in the C leve ls is attributed to increase d competition from N when NH 3 is present Addition of NH 3 seems to have no significant effect on the C levels in the films deposi ted below 500 C likely due to the lower reactivity at lower temperature Figure 5 3 D shows that deposition with NH 3 has lower O levels than deposition without NH 3 As the deposition temperature increased from 450 to 700 C, the O levels in the presence and absence of NH 3 decreased gradually to 5 at. %. The low O incorporation is consistent with dense WN x C y films at higher deposition temperature.

PAGE 70

70 C rystallization occurs to a greater extent at higher temperature reducing diffusion of O from air into the films [89] 5 4 Chemical B onding S tates The values of the BE relative to the emitted the kinetic energy ( KE ) determined by XPS were used to identify the elemental chemical bonding states Fig ure 5 4 A displays the evolution of XPS pattern s for the W 4 f BE as deposition temperature is varied for films deposited with NH 3 For films deposited at 300 C the W 4 f BE is higher than WC x and WN x phases. The major W 4 f 7/2 and W 4 f 5/2 peaks for films deposited at 30 0 C are at 3 6 8 and 37.7 eV, which are close to th ese in the W O 3 phase. These values for W 4 f 7/2 and W 4 f 5/2 peaks agree well with the reported values of 35.5 36.7 eV and 37.6 37.8 eV for the WO 3 phase [33, 76, 78, 90] A s the deposition temperature increase d from 300 to 35 0 C, the W 4 f BE shifted f rom the higher BE (WO 3 ) to the lower BE ( WC x and WN x ). The major W 4 f 7/2 and W 4 f 5/2 peaks for films deposited at 35 0 C are at 31. 7 and 33. 6 eV, which are close to these of the WC x and WN x phases These values for W 4 f 7/2 and W 4 f 5/2 peaks agree well wi th the reported values of 32.7 33.6 eV and 33.3 35.8 eV in WN x phase. Also, these values for W 4 f 7/2 and W 4 f 5/2 peaks agree well with the reported values of 31.6 32.3 eV and 33.7 33.9 eV in WC x phase [76, 78, 80, 90] This indicates that the peak shift of W 4 f occur s at lower temperature than for films deposited without NH 3 The increased N levels in the films as shown in Fig ure 5 3 B is believ ed to be responsible for this shift The major W 4 f 7/2 and W 4 f 5/2 peaks for films deposited at 400 C are at 31. 7 and 33. 7 eV, which are close to the reported values for WC x and WN x phases These two values are higher than metallic W and lower than WO 3 [76, 90] From 350 to 700 C the major W 4 f 7/2 and W 4 f 5/2 peaks correspond to WC x and WN x phases, indicat ing that the chemical b o nding state in W changes from a dominant WO 3 phase to the mixture of WN 2 and WC 1 x phases or WN x C y single solid solution as depositio n temperature increases to 700 C.

PAGE 71

71 T he evolution of XPS pattern s for the N 1 s BE with deposition temperature for films deposited with NH 3 is summarized in Fig ure 5 4 B The N 1 s peak located at 397.3 eV is close to the reported value for WN x phase This value agrees well with the reported values of 396.2 398.2 eV in WN x phase [76, 78, 80, 90] It appears that the N in the film is bound to W in the WN x phase. The intensity of this N 1 s peak is much higher over the entire temperature range, as compared to those for films deposited without NH 3 Films deposited at 400 C have the highest intensity of N indicating the highest N levels in the films as sh own in Fig ure 5 3 B A single N 1 s peak indicates that N has the same metal nitride bonding state over the entire temperature range, irrespective of the other components in the film. From 300 to 700 C there is only a single N 1 s peak near 397.3 eV (i.e. no second N 1 s peak near 400.0 eV ) XPS pattern s for C 1 s BE are shown in Fig ure 5 4 C over the range of deposition temperature for films deposited with NH 3 As shown in this figure the BE of C 1 s peak located at around 283.2 eV and correspond ing to W C x phase is evident in films grow n up to 650 C. For films deposited at 700 C, the bonding states of C 1 s shifted from lower BE to higher BE. Deconvolution of the broad C 1 s peak for films deposited at 700 C using a Gaussian Lorentzian function with ba ckground subtraction yields two separate peaks. The BE of the C 1 s peak located at 284.7 eV corresponds to amorphous C present outside of the WN x C y nanocrystal line regions while the BE of the C 1 s peak located at 283.4 eV corresponds to WC x in the WN x C y nanocr y stals The former value for C 1 s peak agrees well with the reported values of 284.2 285.2 eV for the amorphous C phase, while the latter value for C 1 s peak agrees well with the reported values of 279.7 283.8 eV for the WC x phase [75, 76, 82 84, 90] Amorphous C begins to appear with the WC x phase in the film deposited at 700 C in the presence of NH 3 while

PAGE 72

72 amorphous C appears at 650 C in the absence of NH 3 T his i ndicates the addition of NH 3 promotes metal carbide bonding at higher temperature (650 C). O levels were also probed by XPS. Fig ure 5 4 D traces the evolution of the O 1 s BE with deposition temperature for films deposited with NH 3 The XPS pattern in O 1 s BE over the entire temperature range is similar to that measure d in films deposited with no added NH 3 The O 1 s peaks were near 530. 4 eV, which is close to the reported value for WO 3 phase This value for the O 1 s peak agrees well with the reported v alues of 528.2 531.6 eV for the WO 3 phase [33, 76, 78, 90] As deposition temperature increases to 700 C, the peak intensity of O 1 s decreased as a result of film crystallization and C incorporation block ing uptake along grain boundary. 5 5 Film G rowt h R ate The growth rate in the presence of NH 3 was low, in the range 0.6 to 4.2 /min as compared to the range 2.7 to 29.4 /min for films deposited in the absence of NH 3 Figure 5 5 shows the variation of growth rate with deposition temperature for films deposited with and without NH 3 Both plots reveal a transition from a kinetically controlled growth regime to a mass transfer controlled one Films deposited with NH 3 had a transition point near 450 C while films deposited without NH 3 had a slightly hi gher transition temperature near 500 C Th ese difference s in activation energy, transition temperature, and absolute growth rate are consistent with a change in deposition mechanism due to the addition of NH 3 It is also noted that t he growth rate for f ilms deposited with NH 3 was decreased at higher temperature ( 650 to 700 C ), likely a result of precursor depletion in the gas phase near the substrate surface. T ransamination with NH 3 has been postulated to remove the hydrocarbon group in 1 changing the rate determining step, and thus the growth rate of the films deposited with NH 3 is different from that without added NH 3 [85]

PAGE 73

73 5 6 Electrical R esistivity The f ilm resistivity was determined from measurement of the sheet resistance ( four point probe) and film thickness ( cross sectiona l SEM ) The effect of growth temperature on the film resistivity for films deposited with NH 3 is shown in Fig ure 5 6 Films deposited at 30 0 C show the lowest film resistivity (290 cm ) and the values of film resistivity increase with the interplay of grain boundary density film microstructure, film density, chemical bonding states and film thickness over the entire temperature range. At lower deposition temperature, an increase in N levels with WN x bonding states results in a decrease in film resis tivity and at higher deposition temperature, an increase in film thickness results in an increase in film resistivity. Films deposited at 60 0 C show the highest film resistivity (5450 cm ). The resistivity for films deposited with NH 3 at lower deposition temperature (350 450 C) was lower than those deposited without NH 3 T h is is attributed to increase in the film N content since the resistivity of WN x is lower than WO 3 At highe r deposition temperature (500 700 C), the film resistivity when deposit ed with NH 3 was higher due to the decrease in C content since the resistivity of WN x (4000 cm ) is higher than of WC x (300 400 cm ) [59] The formation of WC x is an important factor in decreasing film resistivity. T h is observation was confirmed by the XPS results for the bonding states of C in the films, as shown in Fig ure 5 4 C Hence, the proper combination of WN x and WC x is important in formation of ternary based refractory metal nitrides for diffusion barrier appl ications. 5 7 Diffusion B arrier T esting XRD measurement s were used to confirm the formation of Cu 3 Si that occurs after barrier failure for Si substrates As shown in Fig ure 5 7 the XRD patterns show no reflections attributable to Cu 3 Si Generally, the standard diffraction peaks of Cu 3 Si appear near 44 .00 to 46 .00 2 with barrier failure After annealing there are three peaks clearly observed which are

PAGE 74

74 assigned to Cu(111) at 43.44 2 Cu(200) at 50.80 2 and Cu(220) at 74.42 2 while patterns on samples b efore annealing only indicate Cu(111). Cu r ecrystalli zation upon annealing resulted in an increase in the intensity of the Cu related reflections due to nucleation of the new grains or growth of preexisting ones in the Cu/WN x C y /Si stacks [71] It is noted that for metall ization applications, the Cu(111) texture is preferred since it shows a higher resistance to electromigration [91] In t ypical diffusion barrier test results the Cu XRD peak intensit ies decrease as silicide peaks are detected along with the evolution of Cu peaks near 44.00 to 46.00 2 [74] Cross sectional TEM images were taken to observe the quality of the Cu/WN x C y and WN x C y /Si interfaces. As shown in Fig ure 5 8, cr oss sectional TEM images reveal that there is no Cu transport through WN x C y before and after annealing under N 2 at 500 C for 30 min. Both Cu/WN x C y and WN x C y /Si interfaces are clearly defined without any evidence of intermixing between the layer s after an nealing. The XRD patterns and cross sectional TEM images reveal no failures of diffusion barrier in Cu/WN x C y /Si stacks. Hence, WN x C y is a promising Cu barrier material 5 8 C onclusions T he tungsten piperidylhydrazido complex Cl 4 (CH 3 CN)W(N pip ) ( 1 ) was u sed to deposit WN x C y with NH 3 to investigate the effect of this coreactant on the film properties for diffusion barrier applications. The deposited films show higher N levels with lower C incorporation as compared to films deposited without NH 3 XRD resu lts suggest that film s deposited at a lower deposition temperature (below 5 0 0 C) w ere amorphous with crystallinity evolving at higher deposition temperature. The XPS W 4f bonding state indicate s that most of the W is present as a mixture of WN x and W C x p hases or a WN x C y single solid solution XPS results for both the W and O indicate WO 3 is present at low deposition temperature (300 C ) in the amorphous state (XRD results) and as deposition temperature increase s WN x C y becomes the dominant W phase

PAGE 75

75 rathe r than WO 3 XPS spectra of the O 1 s bonding state show low O incorporation at higher temperature, which produces films with higher density. An examination of the XPS N 1 s bonding state indicate s that N is present in the WN x phase. XPS spectra show films deposited at 400 C have the highest N levels The XPS C 1 s bonding state results suggest that C is present as WC x and amorphous C The C 1 s BE shifted from lower energy (283.1 eV) to higher energy (284.5 eV) for films deposited at 70 0 C, indicat ing th at amorphous C coexists with WC x XPS observation of the O 1 s bonding state indicate s that O is present as WO 3 XPS s pectra also show lower O incorporation at higher temperature, which produces films with higher density. The film growth rate with NH 3 ad dition varied in the range 0.6 to 4.2 /min over the entire temperature range of study. Large f ilm resistivity changes were observed and can result from various reasons including grain boundary density film microstructure, chemical bonding states and fi lm thickness. Films deposited at 30 0 C show the lowest film resistivity (290 cm ) while the f ilm resistivity was lower at low deposition temperatures as compared to films deposited without NH 3 WN x C y thin films of 20 nm thickness were tested for barr ier performance The results show that WN x C y films are viable diffusion barrier s to prevent Cu interdiffusion and intermixing with Si after annealing under N 2 at 500 C for 30 min.

PAGE 76

76 A) B ) F igure 5 1 XRD spectra for films deposited on Si(100) with NH 3 : A) between 300 and 700 C ; B) standard diffraction patterns for W 2 N and WC 1 x

PAGE 77

77 F igure 5 2 Surface morphology of films deposited on Si(100) substrate at various temperature : A) 300 C without NH 3 ; B) 600 C without NH 3 ; C) 300 C with NH 3 ; D) 600 C with NH 3 A) B) D) C)

PAGE 78

78 A) B ) C ) D ) F igure 5 3 Variation in chemical co mposition of A) W, B) N, C) C and D) O content in the films with deposition temperature with and without added NH 3 Data are measured by XPS after 10 min Ar + ion sputter.

PAGE 79

79 A) B ) C ) D ) F igure 5 4 C hange of binding energies in A) W 4 f B) N 1 s C) C 1 s and D) O 1 s with deposition temperature in the presence of NH 3 Data are from XPS after 10 min Ar + ion sputter.

PAGE 80

80 F igure 5 5 Change in growth rate with deposition temperature for films deposited with and without added NH 3 Thickness was measured by cross sectional SEM F igure 5 6 Change in film resistivity with deposition temperature with and without added NH 3 Data are measured by four point probe.

PAGE 81

81 F igure 5 7 The performance of d iffusion barrier b y XRD measurement for Cu/WN x C y /Si st acks before and after annealing at 500 C.

PAGE 82

82 F igure 5 8 TEM cross sectional images of Cu/WN x C y /Si stacks: [A) and B)] before annealing and [C) and D)] after annealing at 500 C. A) B) C) D) Si WN x C y Cu Si WN x C y Cu Si Cu WN x C y Si WN x C y

PAGE 83

83 CHAPTER 6 DEPOSITION OF WN x C y FROM Cl 4 (CH 3 CN)W(NNPh 2 ) AS A SINGLE SOURCE PRECURSOR 6.1 Film S tructure Figure 6 1 displays XRD patter ns for films grown in the temperature range 300 to 700 C Th e XRD spectra indicate amorphous film deposition below 500 C ( Figure 6 1 ) while the films grown above 500 C yield f our characteristic peaks positions that are consistent with WN x C y (Figure 6 1D) The p rimary peak at 37.20 2 is consistent with the (111) orientation, while additional peaks at 43.66, 63.24, and 75.38 2 are consistent with (200), (220) and (311) orientation, respectively. As the deposition temperature increases from 500 t o 700 C, the reflection associated with the (111) orientation sharpens, likely a result of increasing grain size. Since the f ilm grown at and above 500 C produce a polycrystalline microstructure rapid diffusion of Cu along grain boundaries the underlyin g Si renders these films impractical for barrier applications This is not an issue since d eposition above 400 C is undesirable as many low materials have weak thermal stability [27, 92] T hus this precursor is capable of depositing amorphous films at deposition temperature 400 C 6 .2 Lattice P arameter and Average Grain Size Analysis of the diffraction patterns for the polycrystalline films was performed to estimate the lattice parameter and average grain size. T he standard lattice parameter s for face centered cubic (fc c) W 2 N ( 4.126 ) and WC 1 x (4 .236 ) are shown in Figure 6 2A along with the estimated lattice parameter of the deposited films as a function of temperature The r ock salt structure for W 2 N and WC 1 x consist s of W located o n the fcc positions and N and C located o n the octahedral interstitial sites [93] The 2 position of the WN x C y (111) diffraction peaks can vary due to the change in chemical composition in the films. If the dominant peak position

PAGE 84

84 is between 37.0 1 2 ( WC 1 x ) and 37.7 7 ( W 2 N ) according to Bragg s law, the lattice parameter will be between 4.126 and 4.236 indicating that N, C, and vacancies are mixed o n the interstitial sublattice. T his is the case for films deposited at and above 500 C. The l attice parameter increase s from 4.15 to 4.20 between 500 and 650 C indicat ing that C composition increases with growth temperature The l attice para meter a t 700 C, however, decrease s to 4.18 suggesting that the incorporation of C into the interstitial sublattice results in expanding the lattice in the film structure [18] The most dominant WN x C y (111) diffraction peak was used to estimate the average grain size using Sherrer s formula from the experimental FWHM of this reflection As shown in Figure 6 2B, the a verage grain size for the polycrystalline phase increases with temperature in th e range 500 to 70 0 C, varying from 25 to 55 ( Fig ure 6 2 B) A p olycrystalline phase with grain size below 50 is generally categorized as nanocrystalline [2] Increasing grain size with deposition temperature is common and is often attributed to higher surface diffusivity of absorbed species leading to form ation of stable nucle i and subsequent growth, then decreased because an increase in the C level s on the film surface inhibits surface diffusion, nucleation and growth [18] 6.3 Chemical C omposition The presence of Cl in 2 is of interest due to the need for deposition of Cl free thin films for diffusi on barrier applications As a first test, n o peaks from Cl were observed within the detection limit of XPS ( ~ 1 at. %) for either Cl 2 p 3/2 or Cl 2 s at 199 and 270 eV, respectively Figure 6 3 shows the elementa l composition in the films deposited from 2 a s a function of temperature. T he W level is highest for the films grown at 450 and 500 C varying from 54 to 56 at. %, while N C and O levels are fairly constant for T 500 C As deposition temperature increases from 500 to 700 C, the W level decre ases to 28 at. %. As deposition temperature

PAGE 85

85 increase s from 300 to 55 0 C, the N level varies from 10 to 1 4 at %. The N level in films deposited above 5 5 0 C h owever s hows a gradual decrease accompanied by an increased C level T he N level for films de posited at 400 C has the highest value (14 at. %) The C level is lowest for films deposit ed at 400 C (14 at. %), while at lower deposition temperature ( 400 C), the C level varies from 30 to 24 at. %. As deposition temperature increases up 700 C, the C level increases to 62 at. %. The overall trend for C content is due to the faster decomposition of hydrocarbon groups in the precursor and the solvent (PhCN) as the deposition temperature increases, leading to C incorporation into the film As the deposition temperature increase s from 30 0 to 700 C, the O level decrease s gradually to 5 at. %. The low O incorporation indicates that WN x C y films have dens e microstructure at higher deposition temperature. 6 4 Chemical B onding S tates XPS was used to give information for the bonding states in films and the results are summariz ed for each elements in Figure 6 4A D. The evolution of XPS pattern s in the BE of W 4 f as the deposition temperature varies i s shown in Fig ure 6 4 A For films deposited at 300 C t he major W 4 f 7/2 and W 4 f 5/2 peaks are at 3 1 5 and 33.7 eV, which are close to WN x and WC x phases. These values for W 4 f 7/2 and W 4 f 5/2 peaks agree well w ith the range in reported values of 32.7 33.6 eV and 33.3 35.8 eV in the WN x phase [76, 78 80, 90] Also, these values for the W 4 f 7/2 and W 4 f 5/2 peaks agree well with the reported range s of 31.6 32.3 eV and 33.7 33.9 eV for the WC x phase [75, 76, 82, 90] A s deposition temperature increases from 300 to 70 0 C, t he major W 4 f 7/2 and W 4 f 5/2 peak s are observed at bonding energies of 3 1 5 31.6 and 33.5 33.8 eV which agree well with WN x and WC x phases. T he XPS results indicate that W is primarily bonded to C and N T he evolution of XPS pattern s in BE of the N 1 s as the deposition temperature varies is shown in Fig ure 6 4 B For films deposited at 300 C t he major N 1 s peak is located at 397.4 eV, which is cont aine d within the reported range of 396.2 398.2 eV in WN x

PAGE 86

86 phase [76, 78 80, 90] A s deposition temperature increases from 300 to 70 0 C, t he major N1 s peak s hows a slight shift for higher BE at 397.3 397.5 eV which agrees well with WN x phase. A ll N in the film is bound in the WN x phase not in the N at the grain boundaries The intensity of N is higher at lower deposition temperature (< 400 C). A single N 1 s peak indicates tha t N has the same metal nitride bonding state over the entire temperature range, irrespective of the other co nten ts in the films. T he evolution of XPS pattern s in BE of C 1 s as the deposition temperature varies is shown in Fig ure 6 4 C For films deposited at 300 C t he major C 1 s peak is at 283.1 eV, which is close to WC x phase. As deposition temperature increases to 600 C, the BE of C 1 s peak is at 283.1 283.2 eV, which agrees well with WC x phase. Deconvolution of the broad C 1 s peak for films depos ited at 700 C using Gaussian Lorentzian function with background subtraction yields two separate peaks associated with W C and C C. The BE of the C 1 s peak located at 284.4 eV agrees well with C C bonding outside of the WN x C y nanocrystals, while the BE of C 1 s peak located at 283.8 eV agrees well with W C bonding in the WN x C y The former value for C 1 s peak agrees well with the reported range s of 284.2 285.2 eV in amorphous C phase, while the latter value for the C 1 s peak agrees well with the reported range s of 279.7 283.8 eV in WC x phase [75, 76, 82 84, 90] The presence of the bulky phenyl ligand s in 2 is of interest since thermal decomposition would lead to more C in the films. For diffusion barrier applications, low electrical resistivity is sig nificant and the WC x phase is more conductive than WN x phase [13] T he evolution of XPS pattern s in the BE of O 1 s as a function of deposition temperature is sho wn in Fig ure 6 4 D For films deposited at lower deposition temperature ( 400 C), t he major O 1 s peak is at 530.2 eV, which is within the reported range of 528.2 531.6 eV associated with WO 3 phase [33, 76, 78, 81, 90] As evidenced by Fig ure 6 4 D the peak intensity of O 1 s decreased with increasing deposition

PAGE 87

87 temperature to 700 C IN the range of higher deposition temperature film crystallization (> 500 C) and C incorporation (W C and C C phases) are important factors to block g rain boundary regions in the films. T he XPS results indicate that W is primarily bonded to N and C for films deposited over the entire temperature range (300 700 C) 6 5 Film G rowth R ate The growth rate was determined from the measured film thickness b y c ross sectional SEM images Fig ures 6 5A and 6 5B shows images for films grown at the lowest (300 C ) and highest (700 C ) growth temperature. Fig ures 6 5C and 6 5D display t he SEM images for surface morphology indicat e that films deposited at 700 C a re polycrystalline with a rough surface As shown in the Arrhenius plot in Figure 6 6, t he growth rate var ied from 1 .0 to 25.4 /min. The growth rate at 700 C increased drastically suggesting a change in growth mechanism at this temperature. This obser vation is cons is t e nt with the shift in binding energy in XPS results indicating amorphous C co exists W C bonding states. The Arrh en ius plot indicates the presence of two growth regimes and their slopes consistent with mass transfer limited growth at high er temperature and surface reaction limited at lower temperature The apparent activation energy calculated for the activated process is 0.49 eV in the surface reaction lim i ted growth regime. 6 6 Electrical resistivity Film resistivity varied from 0.6 to 7.9 m cm (Fig ure 6 7). The lowest resistivity of 0. 6 m cm is obtained for films deposited at the lowest temperature ( 300 C ) The value for films deposited at 300 and 35 0 C agrees well with the reported value of 0. 3 to 0.4 m cm in WN x C y phase. WN x C y phase s hows a much lower film resistiv it y than WN x (~ 4 5 m cm ) [13] There are several factors t hat influence the WN x C y resistivity, includ ing the distribution of W N and

PAGE 88

88 W C bonding states, the presence of W O bonding state, film microstructure, and vacancies in sublattice. The interplay of these factors can lead to a complex variables of the film resistivity with deposition temperature The low f ilm resistivity at low temperature is li kely due to higher W C phase. The formation of WC x in WN x C y ( 0.3 0. 4 m cm ) lead s much more conductive films as compared with WN x (4 .0 m cm ) [59] 6 7 Diffusion B arrier T esting C ross sectional TEM images show that there is no diffusion of Cu into WN x C y and Si before and after annealing under N 2 at 500 C for 30 min ( Fig ure 6 8) Both Cu/WN x C y and WN x C y /Si interfaces are clearly observed without any intermixing between the layer s after annealing, which indicates WN x C y thin films block the Cu diffusion. T he EDS depth pro file shows that the Cu K peak decreases sharply at the Cu/WN x C y interface ( Fig ure 6 9) Al though it is clear that Si K and W L peaks are present in the films, it is impossible to separate two overlapping peaks due to the limitation of EDS. The EDS dep th profile h owever, clearly shows that no Cu diffusion is observed between Cu/WN x C y and WN x C y /Si interfaces after annealing. Cross sectional TEM images and EDS depth profile s reveal no onset of failure in the Cu/WN x C y /Si stacks. XRD measurements show th at there is no formation of Cu 3 Si that occurs after failure for Si substrates either before or after annealing under N 2 at 500 C for 30 min (Fig ure 6 10). T he XRD patterns show no reflections attributable to Cu 3 Si. B efore annealing there is only one pea k clearly observed, which is assigned to Cu(111), while diffraction patterns after annealing evidence Cu(111), Cu(200), and Cu(220). Cu recrystallization upon annealing cause s an increase in the intensity of Cu texture due to grain growth in the Cu/WN x C y / Si stacks. The enlargement of Cu grains results in the reduction of the density of Cu grain boundaries, which contributes to lower film resistivity due to low electron scattering. It is noted that for Cu interconnect technology, the Cu(111) texture is pr eferred since it shows greater resistance toward

PAGE 89

89 electromigration. In summary, t he cross sectional TEM images, EDS depth profile s and XRD measurements reveal no onset of failure of the diffusion barrier in the Cu/WN x C y /Si stacks. 6.8 Conclusion s Cl 4 (CH 3 C N)W(NNPh 2 ) ( 2 ) was evaluated as a single source precursor for CVD of WN x C y XRD patterns show that films deposited below 500 C were amorphous while polycrystalline films were grown b etween 500 and 700 C T he l attice parameter of the polycrystalline fi lms varie d from 4.15 to 4.20 while the average grain size increased from 25 to 55 over the temperature range of growth Examination of t he XPS W 4 f bonding state indicates that most of the W is present as a mixture of WN x and WC x or a WN x C y single so lid solution. XPS measurements revealed that W was predominantly bonded to N and C, with C portion increasing with growth temperature. This was attributed to in part to decomposition of the solvent T he amount of W bonded to O, however, was limited. Th e XPS N 1 s bonding state indicates that N is present in tWN x while the XPS C 1 s bonding state indicate s that C is present in WC x and as amorphous C A large variation in film resistivity was measured and is due to the interplay of the combination of W N and W C bonding states, the presence of W O bonding state, film microstructure, and film thickness The results show that WN x C y films are viable Cu barrier materials to prevent diffusion of Cu into Si after annealing under N 2 at 500 C for 30 min. Theref ore, WN x C y is a viable Cu barrier material to prevent diffusion of Cu into Si for Cu interconnect technology

PAGE 90

90 A ) B ) C ) D ) Fi g ure 6 1 XRD sp ectra for films deposited on Si (100) at various temperatures : A) 300 C, B) 700 C, C) between 300 and 700 C, and D) standard powder diffraction p attern for W 2 N and WC 1 x

PAGE 91

91 A ) B ) Fig ure 6 2. Change in A) lattice parameter and B) average grain size with deposition temperature for polycrystalline films deposited from 2 The estimates are b ased on position and shape of diffraction peaks. Fig ure 6 3. Variation of W, N, C and O content in the films deposited from 2 Data are from XPS measurements after 10 min Ar + ion sputter.

PAGE 92

9 2 A ) B ) C ) D ) Fi g ure 6 4. Change of binding energies in A) W 4 f B) N 1 s C) C 1 s and D) O 1 s with deposition temperature. Data are from XPS after 10 min Ar + ion sputter.

PAGE 93

93 Fig ure 6 5 SEM images of films grown on Si (100) substrate : A) cross sectional view of film grown at 300 C ; B) cross secti onal view of film grown at 700 C ; C) surface morphology of film grown at 300 C ; D) surface morphology of film grown at 700 C A) B) WN x C y WN x C y Si Si C) D)

PAGE 94

94 Fig ure 6 6. Change in growth rate with deposition temperature for films deposited from 2 Thickness measured by cross sectional SEM. Fig ure 6 7. Change in film resistivity (four point probe) with deposition temperatur e for films deposited from 2

PAGE 95

95 Fig ure 6 8 C ross sectional TEM images of Cu/WN x C y /Si stacks: [A) and B ) ] before annealing and [C) and D ) ] after annealing at 500 C. WN x C y A) B) Cu C) D) Si WN x C y Si Cu Cu WN x C y Si Si WN x C y

PAGE 96

96 Fig ure 6 9. E DS depth profile of Cu/WN x C y /Si stacks annealed at 500 C. Fig ure 6 10 The performance of diffusion barrier by XRD measurement for Cu/WN x C y /Si stacks before and after annealing at 500 C. A) B) Cu WN x C y Si

PAGE 97

97 CHAPTER 7 DEPOSITION OF WN x C y FROM Cl 4 (CH 3 CN)W(NNPh 2 ): EFFECT OF NH 3 ON FILM PROPERTIES 7 .1 Film S tructure The XRD spectra in Fig ure 7 1 are used to identify the crystalline phase s and to measure the lattice parameter and polycrystalline grain size. T h e peak positions in the X ray diffraction patterns are well matched with a polycrystal mixture of W 2 N and WC 1 x phases or the solid solution WN x C y The XRD spectra show that amorphous films were deposited from 300 to 4 5 0 C, while the polycrystalline materials deposited at temperature from 50 0 to 700 C. The se four observed characteristic peaks with primary peaks at 37.62 and 43.18 2 are consistent with (111) and (200) orientation s, respectively. T wo additional reflections at 62.74 and 75.72 2 are attributed to the (220) and (311) orientation s respectively. Comparing t he observed relative peak intensities to those of the standard powder diffraction intensities [Figure 7 1 D)] indicate that no preferred crystal orientation (texture) exists. As the deposition temperature was increased to 700 C, the peak intensity increased. F ilm depositio n with NH 3 however, resulted in decrease of intensity in XRD peaks as compared to film deposition without NH 3 7 2 Chemical C omposition The measured photoelectron intensities of XPS are used to identify unknown elements and measure the atomic concentrati on The XPS spectra in Fig ure 7 2 indicate that W N, C and O were present in the films. The AES spectra shown in Fig ure 7 3 were taken at the same time as the XPS spectra. In AES, the ejected electrons are not the primary ionized electrons but the seco ndary ionized electrons, which are produced by the decay of ionized atoms from exited states to lower energy states [94] Despite the presence of Cl in the precursor, no peaks were observed for either Cl 2 p 3/2 or Cl 2 s at 199 and 270 eV, r espectively, ruling out Cl contamination in the films within the detection limit of XPS (~ 1 at. %). The absence of Cl signals is consistent

PAGE 98

98 with prior computational results on the related aryl and alkylimido complex Cl 4 (CH 3 CN)W(NR), ( R = Ph, i Pr, C 3 H 5 ) for which a mechanistic pathway was found for reaction of the H 2 carrier gas with W Cl bonds to produce volatile HCl in the gas phase [87] Figure 7 3 A shows that at lower deposition temperatures ( 400 C), the measured W levels in the films in the presence of NH 3 are higher than for films deposited without NH 3 Between 300 and 600 C, the W levels are between 47 and 51 at. %. The N levels, shown in Figure 7 3 B increased over the temperature ra nge 400 to 700 C after addition of NH 3 The N levels have their highest value (19 at. %) for films deposited at 450 C as compared with th o se of films grown without NH 3 (14 at. %). As the deposition temperature increases to 700 C, the N levels graduall y drop. For films deposited at 700 C, N levels without NH 3 are 4 at. %, while for films deposited without NH 3 the N level is 11 at. %. It is postulated that the higher deposition temperature increases the rate of N desorption as N 2 gas, as shown in Fig u re 7 4 B C competes with N for bonding with W, and as shown in Figure 7 3 C the C levels in films deposited with NH 3 than are lower than th o se from the single source only deposition from 300 to 400 C. The decrease in the C levels is due to increased com petition from N when NH 3 is present. For films deposited at 400 C, C levels without NH 3 show 24 at. %, while films deposited without NH 3 shows 15 at. %. Ternary phase metal carbonitride barrier materials show that the C levels lower the resistivity beca use W C phase has lower resistivity than W N phase. The Figure 7 3 D shows that deposition with NH 3 has lower O levels than deposition without NH 3 between 450 and 700 C. As the deposition temperature increased, the O levels in the presence and absence o f NH 3 decreased, with the O levels reaching at 5 at. % in films deposited at 700 C. The low O incorporation is consistent with a dense film microstructure [89]

PAGE 99

99 Crystallization and th us denser films at higher temperature results in reduction of O adsorptio n and reaction from air exposure post growth 7 .3 Chemical B onding S tates The binding energ ies (BE) from XPS measurement are used to identify the chemical bonding states of the elements in the films. This is accompanied by measuring the kinetic energy of emitted elements and relating it to the binding energy according to : where E is kinetic energy of the ionized electron, hv is incident radiation, and E b is the binding energy of the electron [94] This method was used to investigate the chemi cal bonding states for four element s in the films (W, N, C, and O) T he photoelectron line of W 4 f is a doublet due to two spin orbit states, 4 f 7/2 and 4 f 5/2 while the photoelectron lines of N 1 s C 1 s and O 1 s are singlet s The evolution of XPS patterns for W 4 f BE with deposition temperature for films deposited in the presence of NH 3 is summarized in Figure 7 4 A The major W 4 f 7/2 and W 4 f 5/2 peaks are at 31.7 and 33.7 eV, which are close to the values for WC x and WN x re spectively, over the entire deposition temperature range of this study (300 700 C). The values for the W 4 f 7/2 peaks agree well with the reported range s of 32.7 to 33.6 eV for WN x and 31.6 to 32.3 eV for WC x The values for W 4 f 5/2 peaks agree well wit h the reported ranges of 33.3 to 35.8 eV for WN x and 33.7 to 33.9 eV for WC x [76, 78, 80, 90] The evolution of XPS patterns shown in Fig ure 7 4 A indica tes W bonding state is dominant in the physical mixture of WN 2 and WC 1 x or WN x C y solid solution. The evolution of XPS patterns for N 1 s BE with deposition temperature for films deposited in the presence of NH 3 is summarized in Figure 7 4 B The major N 1 s peak i s at 397.3, which is close to the values for W N x The peak position of this BE remained constant over the deposition

PAGE 100

100 temperature range of this study (300 700 C) and the BE associated with the N at the grain boundary at 400.0 eV is absent The value for N 1 s peak agree s well with the reported rang e of 396.2 to 398.2 eV for WN x [76, 78, 80, 90] Th e XPS pattern shown in Fig ure 7 4 B indicates that the N in the film is bound to W in the WN x O n ly a single N 1 s peak is located at near 397.3 eV without a second N 1 s peak near 400.0 eV. The evolution of XPS patterns for the C 1 s BE with deposition temperature for films deposited in the presence of NH 3 is summarized in Fig ure 7 4 C The major C 1 s pea k observed for T 600 C is at 283.3 eV which is close to the value for WC x Th is value for the C 1 s peak agree s well with the reported range of 279.7 to 283.8 eV for WC x [75, 76, 83, 84, 90] For films deposited above 600 C, the bonding state of C 1 s is shifted from lower to higher BE. The higher BE value of the C 1 s peak is located at 284.7 eV, which is close to the value for amorphous C of 284.2 to 285.2 eV for WC x [75, 76, 83, 84, 90] This peak shift indicates that WC x in the WN x C y nanocrystals coexists with amorphous C. The evolution of X PS patterns for O 1 s BE with deposition temperature for films deposited in the presence of NH 3 is summarized in Fig ure 7 4 D The major O 1 s peak is at 540.4 eV, which is close to the value for WO 3 This value remained extra over the entire temperature ra nge of this study (300 700 C). The value for O 1 s peaks agree well with the reported range of 528.2 to 531.6 eV for WO 3 [33, 76, 78, 90] As deposition temperature increase s to 700 C, the peak intensity of O 1 s is decrease s due to film crystallization and C incorporation. 7 .4 Surface M orphology The root mean square (rms) surface roughness of the film deposited at 300 C without NH 3 was determined by AFM to be 5.0 nm, while that of films deposited at 700 C with NH 3 was 87.4 nm ( Figures 7 5 A and 7 5 B) The surface roughness shown in Fig ure s 7 5 C and 7 5 D shows the value of rms roughness was 1.1 nm at 300 C and 5.7 nm at 700 C, indicating the

PAGE 101

101 addition of NH 3 results in films with smoother surfaces. The increase in surfa ce roughness is accompanied by i ncreased crystallinity in film microstructure as the deposition temperature increase s to 700 C. The decrease in roughness is due to an amorphous microstructure and more facile migration of absorbed species on the surface [86] 7 .5. Film G rowth R ate Cross sectional SEM images as ex emplified in Fig ure 7 6 were used to measure the film thickness. The growth rate in the presence of NH 3 was low, in the range 7.3 to 14.3 /min as compared to the range 1.0 to 25.4 /min for films deposited in the absence of NH 3 The Arrhenius plot in th e presence of NH 3 reveals one growth regime while the plot in the absence of NH 3 reveals a transition from a kinetically controlled growth regime to a mass transfer controlled one. T hese differences in growth rate and transition in growth regime are consi stent with a change in deposition mechanism due to the addition of NH 3 T he difference in growth rate indicates a shift in deposition mechanism due to the addition of NH 3 Transamination with NH 3 has been postulated to remove the hydrocarbon group in 1 changing the rate determining step of this study. For films deposit ed at 700 C using a single source the growth rate increased drastically suggesting a change in the growth mechanism with increasing temperatures. This observation was also confirmed by the formation of WC x indicating most of C exists with C C bonding states with small portion of W C bonding states. 7 .6 Electrical R esistivity The f ilm resistivity was determined from the measured sheet resistance (four point probe) and films thickness (c ross sectional SEM). T h e effect of growth temperature on the film resistivity for films deposited with NH 3 is shown in Fig ure 7 8 Films deposited at 400 C show the lowest film resistivity (1 9 m cm ) and the values of film resistivity vary with the int erplay of grain size film microstructure, film density, metal to non metal ratio, and film thickness. A t

PAGE 102

102 lower deposition temperature, increase in N levels with W N bonding states result in decrease in film resistivity. However, after addition of NH 3 a decrease in C levels with W C bonding states cause s an increase in the film resistivity. Also, over 450 C, the film microstructure change s from amorphous to polycrystalline, which causes an increase in film resistivity. W C bonding is an important fact or in decreasing film resistivity as the both sensitivity of WC x is considerably lower than WN x Hence, the proper combination of W N and W C bonding states is significant in formation of ternary based metal nitrides for diffusion barrier applications. 7 7 Diffusion B arrier T esting Diffusion barrier testing was performed to evaluate the performance for Cu interconnects on Si Cross sectional TEM images and the EDS depth profile were used to observe Cu/WN x C y and WN x C y /Si interfaces after annealing The T EM images shown in Figure 7 9 A reveal that there is no Cu diffusion through WN x C y after annealing under N 2 at 5 00 C for 30 min. Both Cu/WN x C y and WN x C y /Si interfaces are clearly observed without any evidence of Cu transport and intermixing between the la yers. T he EDS depth profile shown in Fig ure 7 9 B indicates that the Cu K signal decreases sharply at the Cu/WN x C y interface, demonstrat ing that there is no Cu diffusion observed either before or after annealing under N 2 at 5 00 C for 30 min. Even if a small trace of Cu transported to the Cu/WN x C y interface, the Cu K signal was not detected when the scan moved into the Si substrate. C ross sectional TEM images and the EDS depth profile s reveal no onset of failure in Cu/WN x C y /Si stacks. XRD was also employed to identify the phase of Cu related textures before and after annea ling Cu/WN x C y /Si stacks (Fig ure 7 1 0 ). Only the Cu peak at 43.46 2 for the (111) orientation was observed for the as deposited sample. After annealing at 500 C three Cu peaks at 43.46, 5 0.96 and 74 .60 2 were observed, for the (111), (200), and (222) orientation s respectively. As the annealing temperature increas es to 600 C, the gradual increase in the

PAGE 103

103 intensity of Cu peaks indicates grain growth of Cu, evidenced by the decrease in value of full width half maximum (FWHM). However, after annealing at 700 C the appearance of new Cu related peaks at 37.14 and 57. 24 2 indicates the formation of Cu 3 Si. The decreasing intensity of three Cu peaks and the decreasing thickness of Cu films were due to the diffusion of Cu in Si. Four point probe was employed to measure the change in sheet resistance before and after anne aling Cu/WN x C y /Si stacks (Fig ure 7 1 1 ). As the annealing temperature increases to 600 C, the decrease in the sheet resistance indicates enlargement of Cu grain s The increase in grain size reduces the Cu grain boundaries which contributes to the decrea se in resistivity due to the lower electron scattering. However, after annealing at 700 C the rapid increase in the sheet resistance is consistent with the formation of Cu 3 Si. The cross se ctional TEM images shown in Figure 7 1 2 indicate Cu/WN x C y and WN x C y /Si interfaces were clearly observed without Cu transport and intermixing in the layers before annealing, whereas the Cu 3 Si crystallite exists in WN x C y /Si interface after annealing at 700 C The SEM images shown in Fig ure 7 1 3 B indicate a deterioratio n of Cu surface morphology after annealing at 700 C The color change in the surface from reddish yellow to dark grey indicates a decrease of thickness in Cu layer, the increase in surface roughness of Cu, and the formation of Cu 3 Si, which are all indica tions of the transport and intermixing of Cu in Si. 7.8 Conclusion s The tungsten diphenylhydrazido complex Cl 4 (CH 3 CN)W(NNPh 2 ) ( 2 ) was used to deposit WN x C y with NH 3 coreactant to investigate the effect of this coreactant and the onset of failure process o n the film properties for diffusion barrier applications. The N levels in the films in the presence of NH 3 were higher than th ose in the absence of NH 3 The result shown in the XRD patterns suggests that film microstructure was amorphous for films deposi ted at a lower deposition temperature (below 450 C). The XPS W 4 f bonding state indicates that most of the

PAGE 104

104 W is present in the carbide and nitride mixture or a WN x C y single solid solution. The dominant W bonding state is WN x C y rather than WO 3 from 300 t o 700 C. An investigation of the XPS N 1 s bonding state indicates that N is present as the nitride. XPS spectra show the highest N levels for films deposited at 450 C. An examination of the XPS C 1 s peak indicates that C is present as the carbide. How ever, for films deposited over 600 C, the BE in C 1 s shifted from the lower energy to higher energy, indicating that the W C phase coexists with a C C phase. An observation of XPS O 1 s indicates that O is present as WO 3 or O in the WN x C y XPS spectra sh ow lower O incorporation at higher temperature, which produces films with higher density. AFM micrographs indicate that addition of NH 3 cause s deposit ion of films with smoother surface as compared to th ose from single source deposition. The growth rate with added NH 3 varied in the range 7.3 to 14.3 /min over the entire deposition temperature of study. A large variation of film resistivity is due to the interplay of various reasons such as grain size film microstructure, film density, metal to non metal ratio, and film thickness. F ilms deposited in the absence of NH 3 have lower film resistivity than that of films deposited in the presence of NH 3 Optimal combination of WN x and WC x phase is important in formation of ternary phase materials for diffusion barrier applications. The diffusion barrier test results support the conclusion that WN x C y deposited from 2 is a viable Cu diffusion barrier material for Cu interconnect technology.

PAGE 105

105 A ) B ) C ) D ) Fig ure 7 1 X RD sp ectra for films deposited on Si (100) with NH 3 : A) 300 C, B) 700 C, C) change in XRD spectra and D) standard diffraction plots for W 2 N and WC 1 x

PAGE 106

106 Fig ure 7 2. XPS sp ectra for films deposited on Si (100) with NH 3 No te that Cl peaks are evident as a function of growth temperature

PAGE 107

107 A ) B ) C ) D ) Fig ure 7 3. Comparison of W, N, C and O content in the films deposited in the presence and absence of NH 3 Data are measured by XPS after 10 min Ar + ions sputter.

PAGE 108

108 A ) B ) C ) D ) Fig ure 7 4. Change of binding energies in A) W 4 f B) N 1 s C) C 1 s and D) O 1 s with deposition temperature. Data are from XPS after 10 min Ar + ions sputter.

PAGE 109

109 Fig ure 7 5. Surface morphology of films grown on Si (100) substrate : A) film grown at 300 C without NH 3 ; B) film grown at 700 C without NH 3 ; C) film grown at 300 C with NH 3 ; D) film grown at 700 C with NH 3 A) B) C) D)

PAGE 110

110 Fig ure 7 6. SE M images of films grown on Si (100) substrate : A) cross sectional view of film grown at 300 C; B) cross secti onal view of film grown at 700 C Fig ure 7 7. Change in growth rate with deposition te mperature for the films deposited in the presence and absence of NH 3 Thickness was measured by cross sectional SEM. A) B) WN x C y Si Si WN x C y

PAGE 111

111 Fig ure 7 8. F ilm resistivity as a function of deposition temperature for the films deposite d in the presence and absence of NH 3 A) B) Fig ure 7 9. A) TEM image and B) EDS depth profile of a Cu/WN x C y /Si stack annealed at 500 C for 30 min Si WN x C y Cu

PAGE 112

112 Fig ure 7 1 0. Change in XR D patterns with annealing temperature for Cu/WN x C y /Si stacks. Fig ure 7 1 1. Change in sheet resistance wi th annealing temperature for Cu/WN x C y /Si stacks Data are measured by four point probe.

PAGE 113

113 Fig ure 7 1 2 Cross sectional TEM images of Cu/WN x C y /Si stacks: A) as grown and B) after annealing at 700 C. Fig ure 7 13. C ross sectional SE M images of Cu/WN x C y /Si stacks: A) as grown and B) after annealing at 700 C. A) B) WN x C y Cu Si Cu Cu 3 Si Si WN x C y SiO 2 A) B)

PAGE 114

114 CHAPTER 8 DEPOSITION OF WN x C y FROM Cl 4 (CH 3 CN)W(NNMe 2 ): EFFECT OF NH 3 ON FILM PROPERTIES 8 .1 Film S tructure The progression of XRD patterns in Fig ure 8 1 C are used to identif y the crystalline phase for films deposited in the presence of NH 3 in H 2 carrier with increasing deposition temperature. The XRD pattern of the film deposited below 500 C indicates that it is X ray amorphous, while the films deposited at higher depositio n temperature ( 500 C) are polycrystalline The XRD patterns shown in F igure 8 1C have been compressed to include the results from all nine growth run s of this study, and thus the resolution is decreased in the figure. A n analysis of original data however, reveals f our reflections at and over 500 C, which were calibrated to the Si(400) diffraction peak. The primary peaks at 37.50 and 42.92 2 show relatively high intensity, as compared with the reflections at 61.58 and 75.40 2 All four peaks are between the st andard diffractions of W 2 N and WC 1 x Both standards exhibited a face centered cubic (fcc) structure with similar values of lattice parameter ( W 2 N : 4.124 and WC 1 x : 4.236 ). XRD results suggest the existence of a two phase mixture ( W 2 N or WC 1 x phases) or the presence of their solid solution (WN x C y ). The crystallinity of films deposited in the absence of NH 3 increases with deposition temperature [18] However, the films deposited in the pre sence of NH 3 show a different trend, suggesting that NH 3 coreactant in H 2 carrier gas alters the growth mechanism of CVD [95] A t ransamination reaction with NH 3 has been postulated to remove the allyl substitut e on the imido group in the precursor, changing the rate determining step [85] 8 .2 Chemical C omposition The measured photoelectron intensities of XPS were used to identify elements present in the films and measure the ir atomic concentration XPS spectra in Fig ure 8 2 indicate that W N,

PAGE 115

115 C and O were identified in the films. No Cl contamination in the films w as observed within the detection limit of XPS (~ 1 at. %). Figure 8 3 shows the variation in the chemical composition of W, N, C, and O contents in the films with deposition te mperature. T h e W level is constant (50 at. %) for runs between 300 and 600 C. O v er 600 C, the W level drops gradually because amorphous C starts to coexist with W C. As the deposition temperature increased from 300 to 4 5 0 C, the N level increased fro m 24 to 27 at %. N levels above 45 0 C however, start to decrease, as a consequence of increased C concentration in this range. When the deposition temperature reaches 700 C, the N level has declined to 13 at. % due to the steep rise in C levels at hig h deposition temperatures. It has been suggested that the higher deposition temperature increases the rate of N desorption as N 2 gas [29, 33, 38, 72] Typical refractory metal nitride diffusion barriers show the decreasing tendency of N level with increa sing deposition temperature because higher thermal energy in the lattice structure of the film comes from a higher temperature. From 300 to 4 5 0 C, the C level is below 10 at. %, with the lowest level of 7 at. % for film growth at 3 00 C. As deposition t emperature increase up to 550 C the C level increases gradually, while b etween 6 00 and 700 C, the C level increases drastically from 1 9 to 46 at. %. The overall trend for C content is consistent with the faster decomposition of C H groups in both the precursor and the solvent as the growth temperature increases, leading to C incorporation into the film. This is well matched with the pyrolysis of PhCN around 600 C Films deposited at 300 C show 30 at. % of O which decreased drastically to 1 0 at. % at 50 0 C. As the deposition temperature increased from 50 0 to 700 C, the O level decreased gradually to 4 at. %. From XRD spectra in Fig ure 8 1 C the polycrystalline microstructure becomes evident for depositions performed at 500 C. As the film start s to crystallize, the film microstructure gets

PAGE 116

116 denser by polycrystal grain growth which prevent interdiffusion of O into the lattice of the film after film growth [73] 8 .3 Chemical B onding S tates The binding energy (BE) of XPS are used to identify the chemical bonding states of the elements present from any variation in the determined BE from measurement of kinetic energy (KE) emitted from the elements: where E is kinetic energy of the ionized electron, hv is incident radiation, and E b is the binding energy of the electron [94] They were used to investigate the chemical bonding states for four atoms in the films. T he photoelectron line of W 4 f is a doublet due to two spin orbit states, 4 f 7/2 and 4 f 5/2 while the photoelectron lines of N 1 s C 1 s and O 1 s are si nglet s Figure 8 4 A displays the evolution of XPS patterns in BE of W 4 f as deposition temperature increases. T he major W 4 f 7/2 and W 4 f 5/2 peaks are at 3 1 4 and 33. 8 eV, which are close to WN x and WC x in the temperature range of this study ( 300 70 0 C ) These values for W 4 f 7/2 and W 4 f 5/2 peaks agree well with the reported range of 32.7 to 33.6 eV and 33.3 to 35.8 eV for WN x [76, 78 80, 90] while t hese values for W 4 f 7/2 and W 4 f 5/2 peaks agree well with the reported range of 31.6 to 32.3 eV and 33.7 to 33.9 eV for WC x [75, 76, 82, 90] respectively. T h e major W 4 f 7/2 and W 4 f 5/2 peaks correspond to WN x and WC x These results indicate that a chemical bonding state in W is the mixture of WN 2 and WC 1 x or one single solution of WN x C y XPS pattern s for N 1 s BE are shown in Figure 8 4 B over the range of deposition temperature for films deposited w ith NH 3 This value for the N 1 s peak agrees well with the reported range of 396.2 to 398.2 eV for WN x [76, 78 80, 90] T he major N1 s peak is at 397. 3 to

PAGE 117

117 397.5 eV which agrees well with WN x in the temperature range of this study (300 700 C) A ll N in the films is bound in the WN x A single N 1 s peak shows the metal nitride bonding state, regardless of the other co nten ts in the films. The results shown in Figure 8 4 B suggest that N at the grain boundary can be ruled out due to a single N 1s peak without a second peak near 399 eV. Films deposited at 450 C have the highest intensity of N, indicating the highest N levels in the films, as shown in Figure 8 3. The evolution of XPS patterns for the C 1s BE with deposition temperature for films in the presence of NH 3 is summarized in Figure 8 4 C. U p t o 600 C, the BE of the C 1 s peak located at 283. 0 283. 3 eV agrees well with WC x Deconvolution of the broad C 1 s peak for films deposited from 650 to 700 C using Gaussian Lorentzian function with background subtraction yields two separate peaks (W C a nd amorphous C). The BE of C 1 s peak located at 284.4 eV agrees well with an amorphous C phase present outside of the WN x C y nanocrystals, while the BE of C 1 s peak located at 283. 7 eV agrees well with a W C phase in the WN x C y The former value for C 1 s peak agrees well with the reported range of 284.2 to 285.2 eV for amorphous C, while the latter value for C 1 s peak agrees well with the reported range of 279.7 to 283.8 eV for WC x [75, 76, 82 84, 90] O levels were also probed by XPS in the temperature range of this study ( 300 70 0 C ). T he major O 1 s peak is at 530.2 530.3 eV, which is close to WO 3 This value for the O 1 s peak agrees well wi th the reported range of 528.2 to 531.6 eV in WO 3 [33, 76, 78, 81, 90] As evidenced by Figure 8 4 D, the peak intensity of O 1 s decreased with deposition temperatu re increases to 700 C O levels are l ower in films grown at high temperature due to great extent of film crystallization and C incorporation.

PAGE 118

118 8 4 Surface Morphology The root mean square (rms) roughness of the film surface was determined by AFM to be 1.2 3 nm for films deposited at 300 C in the presence of NH 3 with an increase up to 3.47 nm for deposition at 700 C (Figures 8 5A and 8 5 B). As the deposition temperature increases up to 700 C, the increase in surface roughness is accompanied by the incre ase in crystallinity, while the decrease in surface roughness is consistent with an amorphous microstructure (Figure 8 1). The AFM micrographs indicate that films with smoother surface are due to deposition at lower temperature and addition of NH 3 8 5 F ilm G rowth R ate The growth rate is in the range 1.6 to 32.0 /min, as determined by cross sectional SEM (Figure 8 6) For films deposit ed at 700 C, the growth rate increased drastically suggesting a change in the growth mechanism at these temperatures. Figure 8 7 is consistent with the presence of two growth regimes below 700 C The region with the shallow slope is a mass transfer limited growth regime between 4 0 0 C and 6 5 0 C. The region with the steep slope is a kinetically controlled growth regime between 300 C and 4 0 0 C. The apparent activation energy calculated for the activated process is 0. 31 eV. 8 6 Electrical R esistivity The variation of film resistivity with deposition temperature is shown in Fig ure 8 8 The lowest resistivity is 3.7 m cm at 300 C and t he highest film resistivity is 1 9 .4 m cm for films deposited at 70 0 C. T he values of film resistivity fluctuate with the interplay of polycr ys tal grain growth, C content, O content and film thickness in the temperature range of this study. The high N level in those films is consistent with increased film resistivity in the WN x C y polycrystal structures, due to the higher resistivity for W 2 N relative to WC 1 x As shown in

PAGE 119

119 Fig ure 8 4, an increase in the amorphous C level as the d eposition temperature rises from 650 to 700 C results in an increase in electron scattering, which causes the film resistivity to increase. 8.7 Conclusion s T he tungst en dimethylhydrazido complex Cl 4 (CH 3 CN)W(NNMe 2 ) ( 3 ) was used to deposit WN x C y with NH 3 to investigate the effect of NH 3 on the film properties for diffusion barrier applications. T h e deposited films show higher N levels with lower C incorporation as compared to films deposited without NH 3 XRD results suggest that films deposited below 50 0 C were X ray amorphous with crystallinity evolving at higher deposition temperature. The XPS W 4 f bonding state indicate s that most of the W is present as a mixture of WN x and W C x phases or a WN x C y single solid solution XPS results for the W indicat es WN x C y is the dominant W phase in the temperature range of this study. XPS spectra of the O 1 s bonding state show low O incorporation at higher temperature, which produces films with higher density. An examination of the XPS N 1 s bonding state indicate s that N is present in the WN x phase XPS spectra show films deposited at 4 5 0 C have the highest N levels The XPS C 1 s bonding state results suggest that C is present as WC x and amorphous C The C 1 s BE is shifted from lower energy (283.1 eV) to highe r energy (284.5 eV) for films deposited at 70 0 C, indicat ing that amorphous C coexists with WC x XPS observation of the O 1 s bonding state indicate s that O is present as WO 3 XPS s pectra also show lower O incorporation at higher temperature which produ ces films with higher density The film g rowth rate with NH 3 addition varied in the range 1 .6 to 3 2 /min in the temperature range of 300 to 700 C T he values of film resistivity fluctuates with the interplay of polycr ys tal grain growth, C content, O co ntent and film thickness in the temperature range of this study. The high N level in those films is consistent with increased film resistivity in the WN x C y polycrystal structures, due to the higher resistivity for W 2 N relative to WC 1 x Film resistivity varied in the range 3.7 m cm (300 C ) to 19.4 m cm (700 C ).

PAGE 120

120 A ) B ) C ) D ) Fig ure 8 1 XRD spectra for films deposited on Si(100) with NH 3 : A) 300 C; B) 400 C; C) change in XRD spectra ; D ) standard powder diffraction p attern for W 2 N and WC 1 x

PAGE 121

121 Fig ure 8 2 XPS sp e ctra for films deposited on Si (100) with NH 3 No Cl peaks detected. Fig ure 8 3 Variation in the chemical composition of W, N, C, and O contents in the films with deposition temperature. Data are measured by XPS after 10 min Ar + ion sputter.

PAGE 122

122 A ) B ) C ) D ) Fi gure 8 4 Change of binding energies in A ) W 4 f B ) N 1 s C ) C 1 s and D ) O 1 s with deposition temperature. Data are from XPS after 10 min Ar + ion sputter.

PAGE 123

123 Fig ure 8 5 Surface morphology of films grown on Si(100) substrate : A ) film grown at 300 C with NH 3 ; B ) film grown at 700 C with NH 3 Fig ure 8 6 SEM images of films grown on Si(100) substrate : A ) cross sectional view of film grown at 300 C; B ) cross secti onal view of film grown at 700 C A ) B ) A ) B ) WN x C y WN x C y Si Si

PAGE 124

124 Fig ure 8 7 Change in growth rate with deposition temperature for the films deposited from 3 Thickness was measured by cr oss sectional SEM. Fig ure 8 8. Change in film resistivity (four point probe) with deposition temperature for the films deposited from 3

PAGE 125

125 CHAPTER 9 REACTOR MODELING USING CFD SOFTWARE 9 .1 Description of the Raman Assisted CVD reactor To better und erstand the decomposition mechanism of the tungsten dimethy lhydrazido complex Cl 4 (CH 3 CN)W(N NM 2 ) ( 3 ) a set of preliminary experiment was performed in the probe CVD reactor shown in Fig 9 1 This unique system is interfaced with an in situ Raman spectromet er (Ramanor U 1000, Jobin Yvon) which includes double additive monochromator and uses the 532.08 nm line of Nd:YAG solid state laser as the light source. As described in detail elsewhere [96, 97] this CVD reactor is an up flow impinging jet cold wall reactor that was custom built to quantitative study of the gas phase decomposition kinetics. The i nlet to the reactor consists of three concentric tube s center, annulus and sweep flows. Ea ch inlet line is packed with 3 mm glass beads to provide parallel flow inlet boundary condition Complex 3 which was tested as a metal organic precursor for tungsten based diffusion barrier material is introduced through a center line and N 2 carrier gas was input into the outer tow lines to prevent wall deposition. Based on the previous study [9 6] that developed and validated a steady state, two dimensional mass transport model using results from a CH 4 tracer experiment, conditions are selected such that recirculation flow was not p resent in the reactor. In a preliminary experiment, aerosol ass isted CVD (AACVD) for complex 3 was adopted because this technique has less strict volatility limitation in selecting precursors. The solid precursor 3 was dissolved in benzonitrile (PhCN) to the concentration 7 4 mg/mL (0.0174 mol/L) and then pumped into a nebulizer from a syringe. A piezoelectric material in the nebulizer vibrates at a frequency of 1.44 MHz which generates a mist of precursor 3 and PhCN and the mist is then transported to the reactor with a carrier gas. The N 2 (99.999%, Airgas) carri er gas flow velocity was 0.025 m/s, and the mixture of precursor and solvent was injected at

PAGE 126

126 a rate of 0.5 ml/h P ure N 2 gas was also delivered to the annulus and sweep line s after same flow velocity (0.025 m/s) and sufficient time was allowed to reach st eady state before measurements were taken Then the 1W Nd:YAG solid state laser line (532.08 nm) was used to excite complex 3 and several vibrational Raman excitation line s were observed 9 2 Multiphase Flow Simulation of the Raman Assisted CVD reactor Simulations on flow and thermal patterns in the reactor were performed using FLUENT computational fluid dynamics (CFD) packages Equations of conservation for mass, momentum, and energy were solved with geometry and boundary conditions specific to this reactor geometry. The conservation of m ass can be written as follows. (9 1) T his equation describes the time rate of change of the fluid density at a fixed point in the space. The vector is the mass flux, and its divergence is the net rate of mass efflux per unit volume. The conse rvation of momentum can be written as: (9 2) This equation describes the rate of increase of momentum per unit volume. T h e term is the rate of momentum addition by conservation per unit volume. The tw o terms are the rate of momentum addition by molecular transport per unit volume. The term is the external force of gravity on the fluid per unit volume. T he conservation equation for momentum is equivalent t o Newton s second law of motion: the statement of mass x acceleration = sum of forces. The conservation of energy can be written as:

PAGE 127

127 (9 3) This equation describes the rate of increase of kinetic energy per unit volume. Two major energy terms commonly used in computational fluid mechanics are kinetic energy associated with observable fluid motions of the molecules, plus the energy of interaction between molecules. The term is the rate of addition of kinetic e nergy by convection per unit volume. The term is the rate of work done by pressure of the surroundings on the fluid. The term is the rate of reversible conversion of kinetic energy into internal energy. Th e term is the rate of work done by viscous forces on the fluid The term is the rate of irreversible conversion from kinetic to internal energy T he term is the rate of work by gravity on the fluid. FLUENT use s a Finite Volume (FV) method to convert the governing equations to algebraic equations Algebraic equations can be solved numerically in order to solv e these coupled con s ervation partial differential equations ( PDE ) The solution domain is subdivide d into a finite number of contiguous control volumes (CV). Then, the conservation equations are applied to each CV. This CV technique has two parts f or integrating the governing equations a bout each CV and yielding discrete equations to conserve each quan tity on a CV basis. The mesh was generated using the G AMBIT (version 2.2.30) with cylindrical coordinates in two dimensional format. An unstructured quadruple grid was employed. Three types of boundary conditions were assigned during the mesh design step: inlet flow velocity at reactor inlet, outflow type at re actor outlet, and surface wall temperature at heater surface. The segregator solver was used, where the governing equations (momentum, continuity, and scalar)

PAGE 128

128 are solved sequentially rather than in a simultaneous way. Note that sequentially means segr egated from one another The simulation of the behavior of the flow and thermal pattern in the reactor was performed using F LUENT (version 6.2.16). The operating pressure was fixed at 101325 Pa, that is an atmospheric pressure growth condition for the CVD reactor, and the gravitational acceleration was turned on to the minus Y axis direction ( 9.8 m/s). Boundary conditions used in this simulation are summarized in the following table (Table 9 1). The i nlet velocity was 0.025 m/s while the o utlet boundary condition uses a typical outlet pressure of reactor, which is 760 Torr. T he t emperature at the heater surface i s set at 1200 K and there is no heat flux through the side walls of internal reactor. Figures 9 3 and 9 4 show the calculated contour s of static temperature and velocity magnitude from the CFD simulation Figure 9 5 shows the calculated contour s of velocity magnitude and volume fraction of the secondary phase in the multiphase flow model.

PAGE 129

129 Table 9 1 Boundary conditions for CVD reactor Boundary location Boundary type Specific condition Inlet flow Velocity inlet 0.025 m/s Outlet flow Outlet pressure 760 T orr Heater temperature Wall 1200 K Reactor wall Wall No heat flux

PAGE 130

130 Figure 9 1. Schematic photograph s of A) CVD reactor system that is interfaced to the Raman spectrometry ; B) n ebulizer system; C) the impinging jet probe reactor A ) B ) C )

PAGE 131

131 Figure 9 2. Mesh design of CVD reactor u sing G AMBIT

PAGE 132

132 Figure 9 3 Color filled contour s of static temperature (K) and contour line of static temperature (K) in the vicinity of the heater

PAGE 133

133 Figure 9 4 Contour s of velocity magni tude (m/s) and velocity vector colored by velocity magnitude (m/s) in the vicinity of the heater

PAGE 134

134 A) B) C ) D ) Figure 9 5. Contour s of velocity magnitude (m/s) and volume fraction of s olvent phase in multiphase flow model.

PAGE 135

135 E ) F ) G ) H ) Figure 9 5. Continued

PAGE 136

136 I ) J ) K ) L ) Figure 9 5. Continued

PAGE 137

137 CHAPTER 10 CONCLUSION S AND FUTURE WORK F ilm s grow n by CVD using 1 2 and 3 were used to investigate the film properties and diffusion barrier quality The results detailed i n chapters 3 8 indicate that tungsten based films have many positive properties important to diffusion barrier application. First, the work demonstrates that WN x C y is an effective Cu diffusio n barrier material to prevent the transport of Cu into Si. Fil ms deposited at 400 C using 1 2 or 3 are able to prevent Cu interdiffusion after annealing at 500 C for 30 min under N 2 atmosphere In particular, s amples annealed at higher temperature using 3 showed evidence of failure only when annealed at 700 C. Second, each precursor 1 2 and 3 yielded film growth at temperatures as low as 300 C indicating that facile precursor decomposition pathway and aerosol assisted metal organic CVD can be used at an acceptable deposition temperature ( < 400 C ) for diffu sion barrier applications. Third, WN x C y promotes the PVD growth of Cu with the preferred (111) orientat ion on a WN x C y /Si stack It is noted that for metallization applications, the Cu(111) texture is preferred since it shows a higher resistance to electr omigration Fo u rth, addition of C to WN x causes to lower the film resistivity because WC x phase has a lower resistivity than WN x It was also found that incorporation of NH 3 in the gas stream results in the deposition of higher resistivity films due to t he greater incorporation extent of N. Fi nally WN x C y films show good adhesion to Cu indicating the Cu/WN x C y /Si stack is thermally and mechanically stable after annealing at 500 C for 30 min 10 .1 Ru WN x C y for Diffusion Barrier and Cu Direct Plate App lications Several groups have attempted the growth of bilayer direct plate liner/ diffusion barrier materials for Cu integration without the need for a Cu seed layer. A mixed phase Ru W N x C y deposited by aerosol assisted metal organic CVD ( or metal organic ALD ) is proposed as a novel direct plate liner for advanced Cu metallization [66] From present study, it is know that WN x C y

PAGE 138

138 diffusion barriers an effective Cu barrier s Based on the combination of know n barrier properties and electrochemical properties, indicating a higher affinity to the direct plate process, a tungsten based diffusion barrier can be selected as a candidate material to combine with Ru for evaluation as an extendible direct plate liner technology. The Ru WN x C y mixture can be deposited u sing the precursor s from this study and Ru precursors supported by Dr. McElwee White in the D epartment of C hemistry at the University of Florida. By mixing both precursors in the solvent, The Ru:W overall composition in the films can be varied simply by c hanging the relative concentration of both precursors in PhCN T he current Cu barrier/seed stacks are a trilayer consisting of PVD Cu on top of Ta (adhesion layer)/TaN (diffusion barrier). Using Ru WN x C y mixture phase films, it may be possible to replace the traditional PVD Cu/Ta/TaN stack with a single layer of physical mixtures. That is the Ru should provide good adhesion since the WN x C y showed nucleate Cu (111) [66] 10 2 WN x C y for Realistic Diffusion Barrier Testing The diffusion barrier films tested in the present study were 15 to 20 nm in thick nes s. As features in interconnects continue to shrink to align with the International Technology Roadmap for Semiconductor (ITRS), the thickness of the diffusion barrier is required to be 2.9 nm in 2015 for metalliz ation. Thus, a study of how thin can the barrier layer be grown and still be effective should be made. Th e film thickness can be easily reduced by either reducing the reaction time or the precursor concentration in the solvent by aerosol assisted metal o rganic CVD. The diffusion barrier test results in the present study w ere characterized by XRD measurement, AES depth profiling, Secco etch test, cross sectional TEM imaging with the EDS analysis, sheet resistance measurement, and SEM surface imaging. The se techniques require significant Cu transport across the barrier to be effective It is suggested that e lectrical characterization such as triangle voltage sweep (TVS) techniques should be used to detect trace of Cu transport through the

PAGE 139

139 barrier film to detect the time barrier limit The effective capacitance of the structure is employed as a measure of the free charge of Cu ions that has diffused through the barrier into the adjacent dielectric. This result could then be correlated with the other metho ds to give a sense of their sensitivity.

PAGE 140

LIST OF REFERENCES [1] E. Kolawa, P. J. Pokela, J. S. Reid, J. S. Chen, M. A. Nicolet, Appl. Surf. Sci. 53 (1991) 373 376. [2] A. E. Kaloyeros, E. Eisenbraun, Ann u Rev. Mat er Sc i. 30 (2000) 363 385. [3] R. Rosenberg, D. C. Edelstein, C. K. Hu, K. P. Rodbell, Annu. Rev. Mat er Sci. 30 (2000) 229 262. [4] A. A. Istratov, J. Electrochem. Soc. 149(1) (2000) G21 G30. [5] M. Wittmer, Appl. Phys. Lett. 36 (1980) 456 458. [6] D. H. Kim, S. L. Cho, K. B. Kim, J. J. Kim, J. W. Park, Appl. Phys. Lett. 69 (1996) 4182 4184. [7] G. Llauro, R. Hillel, F. Sibieude, Chem. Vapor Depos. 4 (1998) 247 252. [8] M. H. Tsai, S. C. Sun, H. T. Chiu, C. E. Tsai, S. H. Chuang, Appl. Phys. Lett. 67 (1995) 112 8 1130. [9] T. Hara, M. Tanaka, K. Sakiyama, S. Onishi, K. Ishihara, J. Kudo, Jpn. J. Appl. Phys. 36 (1997) L893 L895. [10] N. Gonohe, Mater. Trans. 43 (2002) 1585 1592. [11] O. H. Gokce, S. Amin, N. M. Ravindra, D. J. Szostak, R. J. Paff, J. G. Fleming, C J. Galewski, J. Shallenberger, R. Eby, Thin Solid Films 353 (1999) 149 156. [12] D. J. Kim, H. S. Sim, S. I. Kim, Y. T. Kim, H. Jeon, 20 (2002) 194 197. [13] W. Koh, D. Kumar, W. M. Li, H. Sprey, I. J. Raaijmakers, Solid State Technol 48 (2005) 54 58. [ 14] R. Fix, R. G. Gordon, D. M. Hoffman, Chem. Mater. 5 (1993) 614 619. [15] S. Jeon, K. Yong, S. G. Park, S. W. Rhee, J. Vac. Sci. Technol. B 24 (2006) 1428 1431. [16] B. H. Lee, K. J. Yong, J. Electrochem. Soc. 151 (2004) C594 C597. [17] S. H. Kim, S. S. Oh, H. M. Kim, D. H. Kang, K. B. Kim, W. M. Li, S. Haukka, M. Tuominen, J. Electrochem. Soc. 151 (2004) C272 C282. [18] O. J. Bchir, K. M. Green, H. M. Ajmera, E. A. Zapp, T. J. Anderson, B. C. Brooks, L. L. Reitfort, D. H. Powell, K. A. Abboud, L. McElwe e White, J. Am. Chem. Soc. 127 (2005) 7825 7833.

PAGE 141

[19] H. M. Ajmera, A. T. Heitsch, O. J. Bchir, D. P. Norton, L. L. Reitfort, L. McElwee White, T. J. Anderson, J. Electrochem. Soc. 155 (2008) H829 H835. [20] E. Eisenbraun, A. Upham, R. Dash, W. Zeng, J. Ho efnagels, S. Lane, D. Anjum, K. Dovidenko, A. E. Kaloyeros, B. Arkles, J. Sullivan, J. Vac. Sci. Technol. B 18 (2000) 2011 2015. [21] Y. J. Lee, B. S Suh, M. S. Kwon, C. O. Park, J. Appl. Phys. 85 (1999) 1927 1934. [22] J. G. Fleming, E. Roherty Osmun, P. M. Smith, J. S. Custer, Y. D. Kim, T. Kacsich, M. A. Nicolet, C. J. Galewski, Thin Solid Films 320 (1998) 10 14. [23] L. McElwee White, Dalton Trans. (2006) 5327 5333. [24] J. Koller, H. M. Ajmera, K. A. Abboud, T. J. Anderson, L. McElwee White, Inorg. Ch em. 47 (2008) 4457 4462. [25] H. Kim, C. Cabral Jr., C. Lavoie, S. M. Rossnagel, J. Vac. Sci. Technol. B 20 (2002) 1321 1326. [26] K. E. Elers, V. Saanila, P. J. Soininen, W. M. Li, J. T. Kostamo, S. Haukka, J. Juhanoja, W. F. A. Besling, Chem. Vap or Depo s 8 (2002) 149 153. [27] H. Kim, J. Vac. Sci. Technol. B 21 (2003) 2232 2261. [28] F. C. T. So, E. Kolawa, X. A. Zhao, E. T. S. Pan, M. A. Nicolet, J. Appl. Phys. 64 (1988) 2787 278 9. [29] H. P. Kattelus, E. Kolawa, K. Affolter, M. A. Nicolet, J. Vac. Sci. Technol. A 3 (1985) 2246 2254. [30] M. Uekubo, T. Oku, K. Nii, M. Murakami, K. Takahiro, S. Yamaguchi, T. Nakano, T. Ohta, Thin Solid Films 286 (1996) 170 175. [31] K. Nakajima, Y. Akasaka, K. Miyano, M. Takahashi, S. Suehiro, K. Suguro, Appl. Surf. Sci. 117/118 (1997) 312 316. [32] S. D. Marcus, R. F. Foster, Thin Solid Films 236 (1993) 330 333. [33] T. Nakajima, K. Watanabe, N. Watanabe, J. Electrochem. Soc. 134 (1987) 3175 3178. [34] M. Nagai, K. Kishida, Appl. Surf. Sci. 70 1 (1993) 759 762. [35] B. H. Lee, K. J. Yong, J. Vac. Sci. Technol. B 22 (2004) 2375 2379. [36] J. E. Kelsey, C. Goldberg, G. Nuesca, G. Peterson, A. E. Kaloyeros, B. Arkles, J. Vac. Sci. Technol. B 17 (1999) 1101 1104. [37] H. T. Chiu, S. H. Chuang, J. Mater. Res. 8 (1993) 1353 1360

PAGE 142

[38] E. L. Crane, H. T. Chiu, R. G. Nuzzo, J. Phys. Chem. B 105 (2001) 3549 3556. [39] J. P. Lu, W. Y. Hsu, J. D. Luttmer, L. K. Magel, H. L. Tsai, J. Electrochem. Soc. 145 (1998) L21 L23. [40] J. Lin, A. Tsukune, T. Suzuki, M. Yamada, J. Vac. Sci. Techn ol. A 16 (1998) 611 614. [41] J. Lin, A. Tsukune, T. Suzuki, M. Yamada, J. Vac. Sci. Technol. A 17 (1999) 936 938. [42] Y. T. Kim, C. W. Lee, S. K. Min, Appl. Phys. Lett. 61 (1992) 537 539. [43] S. E. Potts, C. J. Carmalt, C. S. Blackman, T. Leese, H. O. D avies, Dalton Trans. (2008) 5730 5736. [44] A. C. Anacleto, N. Blasco, A. Pinchart, Y. Marot, C. Lachaud, Surf Coat Technol 201 (2007) 9120 9124. [45] O. J. Bchir, S. W. Johnston, A. C. Cuadra, T. J. Anderson, C. G. Ortiz, B. C. Brooks, D. H. Powell, L. McElwee White, J. Cryst. Growth 249 (2003) 262 274. [46] O. J. Bchir, K. M. Green, M. S. Hlad, T. J. Anderson, B. C. Brooks, L. McElwee White, J. Cryst. Growth 261 (2004) 280 288. [47] O. J. Bchir, K. M. Green, M. S. Hlad, T. J. Anderson, B. C. Brooks, C. B. Wilder, D. H. Powell, L. McElwee White, J. Organomet. Chem. 684 (2003) 338 350. [48] O. J. Bchir, K. C. Kim, T. J. Anderson, V. Craciun, B. C. Brooks, L. McElwee White, J. Electrochem. Soc. 151 (2004) G697 G703. [49] H. S. Sim, S. I. Kim, Y. T. Kim, J. Vac. Sci. Technol. B 21 (2003) 1411 1414. [50] C. W. Lee, Y. T. Kim, J. Vac. Sci. Technol. B 24 (2006) 1432 1435. [51] J. W. Klaus, S. J. Ferro, S. M. George, J. Electrochem. Soc. 147 (2000) 1175 1181. [52] J. W. Klaus, S. J. Ferro, S. M. George, Appl. Su rf. Sci. 162 (2000) 479 491. [53] J. S. Becker, R. G. Gordon, Appl. Phys. Lett. 82 (2003) 2239 2241. [54] J. S. Becker, S. Suh, S. Wang, R. G. Gordon, Chem. Mater. 15 (2003) 2969 2976. [55] P. de Rouffignac, Z. Li, R. G. Gordon, Electrochem Solid State Let t. 7 (2004) G306 G308. [56] A. M. Hoyas, J. Schuhmacher, C. M. Whelan, T. F. Landaluce, D. Vanhaeren, K. Maex, J. P. Celis, J. Appl. Phys. 100 (2006) 114903 6 114903 6 [57] K. S. Kim, M. S. Lee, S. S. Yim, H. M. Kim, K. B. Kim, H. S. Park, W. Koh, W. M. L i, M. Stokhof, H. Sprey, Appl Phys. Lett. 89 (2006) 081913 1 081913 3

PAGE 143

[58] A. M. Hoyas, C. M. Whelan, J. Schuhmacher, K. Maex, J. P. Celis, Microeletron. Eng. 83 (2006) 2068 2071. [59] H. Volders, Z. Tokei, H. Bender, B. Brijs, R. Caluwaerts, L. Carbonel l, T. Conard, C. Drijbooms, A. Franquet, S. Garaud, I. Hoflijk, A. Moussa, F. Sinapi, Y. Travaly, D. Vanhaeren, G. Vereecke, C. Zhao, W. M. Li, H. Sprey, A. M. Jonas, Microelectron. Eng. 84 (2007) 2460 2465. [60] S. H. Kim, S. S. Oh, K. B. Kim, D. H. Kang, W. M. Li, S. Haukka, M. Tuominen, Appl. Phys. Lett. 82 (2003) 4486 4488. [61] S. H. Kim, S. S. Oh, H. M. Kim, D. H. Kang, K. B. Kim, W. M. Li, S. Haukka, M. Tuominen, Mat. Res. Symp. Proc. 766 (2003) E10.9.1 E10.9.7. [62] K. E. Elers, V. Saanila, W. M. Li P. J. Soininen, J. T. Kostamo, S. Haukka, J. Juhanoja, W. F. A. Besling, Thin Solid Films 434 (2003) 94 99. [63] S. Smith, W. M. Li, K. E. Elers, K. Pfeifer, Microelectron. Eng. 64 (2002) 247 253. [64] C. L. Dezelah, O. M. El Kadri, K. Kukli, K. Arstila, R. J. Baird, J. Lu, L. Niinisto, C. H. Winter, J. Mater. Chem. 17 (2007) 1109 1116. [65] W. X. Zeng, X. D. Wang, S. Kumar, D. W. Peters, E. T. Eisenbraun, J. Mater. Res. 22 (2007) 703 709. [66] D. Greenslit, T. Chakraborty, E. Eisenbraun, J. Vac. Sci. Tec hnol. B 27 (2009) 631 636. [67] A. G. Orpen, L. Brammer, F. H. Allen, O. Kennard, D. G. Watson, R. Taylor, J. Chem. Soc. Dalton. (1989) S1 S83. [68] C. Redshaw, V. C. Gibson, W. Clegg, A. J. Edwards, B. Miles, J. Chem. Soc. Dalton. (1997) 3343 3347. [69] J Chatt, M. E. Fakley, P. B. Hitchcock, R. L. Richards, N. T. Luong Thi, J. Chem. Soc. Dalton. (1982) 345 3 52. [70] J. Koller, H. M. Ajmera, K. A. Abboud, T. J. Anderson, L. McElwee White, Inorg. Chem. (2007) 4457 4462. [71] B. D. Cullity, S. R. Stock, Ele ments of X ray diffraction: Third Edition; Third ed.; Prentice Hall: Upper Saddle River, NJ, 2001 [72] J. S. Reid, E. Kolawa, R. P. Ruiz, M. A. Nicolet, Thin Solid Films 236 (1993) 319 324. [73] P. J. Pokela, C. K. Kwok, E. Kolawa, S. Raud, M. A. Nicolet, Appl. Surf. Sci. 53 (1991) 364 3 72. [74] D. J. Kim, Y. B. Jung, M. B. Lee, Y. H. Lee, J. H. Lee, J. H. Lee, Thin Solid Films 372 (2000) 276 283.

PAGE 144

[75] R. J. Colton, J. W. Rabalais, Inorg. Chem. 15 (1976) 236 238. [76] C. D. Wagner, W. M. Riggs, L. E. Davis J. F. Moulder, G. E. Muilenberg, Handbook of X ray Photoelectron Spectroscopy; Perkin Elmer Corporation: Eden Prairie, MN, 1978. [77] J. F. Moulder, W. F. Stickle, P. E. Sobol, K. D. Bomben, Handbook of X ray Photoelectron Spectroscopy; J. Chastain and R C. King Jr., Eds.; Physical Electronics, Inc.: Eden Prairie, MN, 2000. [78] H. L. Zhang, D. Z. Wang, N. K. Huang, Surf. Coat. Technol. 115 (1999) 140 144. [79] Y. G. Shen, Y. W. Mai, Mater. Sci. Eng. B 76 (2000) 107 115. [80] J. S. Lee, C. S. Park, J. Y. Kang, D. S. Ma, J. Y. Lee, J. Vac. Sci. Technol. B 8 (1990) 1117 1121. [81] P. J. C. Chappell, M. H. Kibel, B. G. Baker, J. Catal. 110 (1988) 139 149. [82] G. Leclercq, M. Kamal, J. M. Giraudon, P. Devassine, L. Feigenbaum, L. Leclercq, A. Frennet, J. M. Bastin, A. Lofberg, S. Decker, M. Dufour, J. Catal. 158 (1996) 142 169. [83] V. Crist, Handbook of Monochromatic XPS Spectra: The Elements and Native Oxides; John Wiley & Sons, Ltd.: New York, 2000. [84] J. Luthin, C. Linsmeier, J. Nucl. Mater. 290 293 (2 001) 121 125. [85] Y. S. Won, Y. S. Kim, T. J. Anderson, L. McElwee White, Chem. Mater. 20 (2008) 7246 7251. [86] K. Takahashi, M. Nakayama, S. Tokoyama, T. Kimura, E. Tokumitsu, H. Funakubo, Applied Surface Science 216 (2003) 296 301. [87] Y. S. Won, Y. S Kim, T. J. Anderson, L. L. Reitfort, I. Ghiviriga, L. McElwee White, J. Am. Chem. Soc. 128 (2006) 13781 13788. [88] S. L. Cho, K. B. Kim, S. H. Min, H. K. Shin, S. D. Kim, J. Electrochem. Soc 146 (1999) 3724 3730. [89] S. H. Kim, D. S. Chung, K. C. Park, K. B. Kim, S. H. Min, J. Electrochem. Soc. 146 (1999) 1455 1460. [90] J. F. Moulder, W. F. Stickle, P. E. Sobol, K. D. Bomben, Handbook of X ray Photoelectron Spectroscopy; J. Chastain and R. C. King Jr., Eds.; Physical Electronics, Inc.: Eden Prairie, MN 1995. [91] A. Furuya, M. Tagami, K. Shiba, K. Kikuta, Y. Hayashi, IEEE Trans Electron Devices 49 (2002) 733 738.

PAGE 145

[92] K. E. Elers, V. Saanila, P. J. Soininen, W. M. Li, J. T. Kostamo, S. Haukka, J. Juhanoja, Chem. Vap or Depos 8 (2002) 149 153. [93] L. E Toth, Transition Metal Carbides and Nitrides; J. L. Margrave, Ed.; Academic Press: New York, 1971. [94] A. C. Westerheim, J. M. Bulger, C. S. Whelan, T. S. Sriram, L. J. Elliott, J. J. Maziarz, J. Vac. Sci. Technol. B 16 (1998) 2729 2733. [95] H. M. Ajme ra, A. T. Heitsch, O. J. Bchir, D. P. Norton, L. L. Reitfort, L. McElwee White, T. J. Anderson, J. Electrochem. Soc. 155 (2008) H829 H835. [96] J. Y. Hwang, C. Park, M. Huang, T. Anderson, J. Crst. Growth (2005) 521 530. [97] Y. S. Kim, Y. S. Won, H. Hagel in Weaver, N. Omenetto, T. Anderson, J. Phys. Chem. A (2008) 4246 4253.

PAGE 146

BIOGRAPHICAL SKETCH Dojun K i m was born in 1975 in Pusan, Korea. He entered Seoul National University in March 1994 and received his Bachelor of Science degree in c he mical e ngineering in February 1998. Following undergraduate, he started graduate school in Seoul National University and received his Master of Science degree in c hemical e ngineering in February 2000. Upon graduation, he worked for more than five years a s a process engineer for SK E ngineering & Construction (SKEC) in Seoul, Korea (2000 2005). The main role is process design, simulation, control, and consultation for petrochemical and refinery processes. Following his five year s industr y he started hi s doctoral studies in the Department of C hemical E ngineering at the University of Florida in August 2005. He joined the electronic materials processing group under the guidance of Dr. Timothy J. Anderson in December 2005. His research topic is chemical v apor deposition and atomic layer deposition of metal nitride thin films for diffusion barrier application. While he was working in SKEC and UF, he married Sora Park on February 23, 2002 and had a son, Jinho Kim, on October 28, 2003 (Seoul, Korea) and a da ughter, Katherine Nayoun Kim, on January 23, 2007 (Gainesville, FL). Upon graduat ion he plans to work as a s enior p rocess e ngineer in Intel s Portland Technology Development D ivision based in Hillsboro, OR