Measurement and modeling of the effects of pulsed laser deposited coatings on cathodoluminescent phosphors

MISSING IMAGE

Material Information

Title:
Measurement and modeling of the effects of pulsed laser deposited coatings on cathodoluminescent phosphors
Physical Description:
x, 322 leaves : ill. ; 29 cm.
Language:
English
Creator:
Thomes, William Joseph, 1974-
Publication Date:

Subjects

Subjects / Keywords:
Materials Science and Engineering thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Materials Science and Engineering -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 2000.
Bibliography:
Includes bibliographical references (leaves 310-321).
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by William Joseph Thomes.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 024909653
oclc - 45805680
System ID:
AA00018871:00001

Full Text









MEASUREMENT AND MODELING OF THE EFFECTS OF
PULSED LASER DEPOSITED COATINGS ON
CATHODOLUMINESCENT PHOSPHORS
















By

William Joseph Thomes, Jr.


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

UNIVERSITY OF FLORIDA


2000
































This work is dedicated to the memory of my grandfather,
Gilbert Walter Thorpe,
who passed away in the course of my graduate studies.















ACKNOWLEDGMENTS


There are so many people whom I wish to thank that it is hard to decide where to

begin. First, I would like to thank my fiance, Cynthia, for her loving support all these

years. She has seen me through the ups and downs of graduate school and has always

been there to pick me up when I was down or to share in my joy. She also provided

invaluable editorial advice and revisions for this dissertation. I would also like to thank

my family for their love throughout my life. Even when I was not sure that it was

possible to make it all the way to the end, they believed in me and provided the

encouragement to stick it out no matter what happened.

Dr. Paul Holloway deserves a special word of thanks for providing guidance

during my graduate education. His door was always open, and he never hesitated to

discuss my work with me and give me some direction when needed. He is truly one of

the greatest teachers that I have been fortunate enough to come across in my educational

pursuits.

Special thanks are also due to Dr. Carl Seager and Dr. Dave Tallant for providing

me with the opportunity and the funding to come to Sandia National Laboratories to

conduct many of my experiments with them. I am especially grateful to Carl, who, in

addition to providing me access to all of the equipment in his labs so that I could measure

my samples, was a mentor and a friend during my stay in New Mexico. Without his help








and guidance, much of the work contained in this dissertation would not have been

possible.

There are many others who made my stay in New Mexico a memorable one.

Although I cannot mention them all in the space allotted to me, I would like to point out a

few. John Hunter was everything I could have asked for in a friend, and never once did

he hesitate to allow me to stay at his house a little longer when I said, "I think they want

me to stay one more month." I would also like to thank my many other friends out in

New Mexico, including John, Shasta, Jerry, and Shawna. Thanks are due to Jonathan

Campbell for sedimenting all of my phosphor screens, helping me learn all the quirks of

the lab equipment, introducing me to everyone at the lab, and teaching me about all of the

fascinating things that New Mexico has to offer (including green chiles). I'd also like to

thank Guild, who always seemed to be in a good mood, for his help with the SEM work.

Many thanks go to those in Florida who have made these past four years

memorable ones. I wish to thank the entire Holloway group past and present, especially

Billie, Jay, Bo, Chris, Sean, Troy, Loren, Caroline, Bill, Mark, and Ludie (without whom

we might all be lost). I am grateful to the following for not only providing me with their

friendship, but also for performing measurements on my samples: Eric Lambers for AES

measurements, Wish Krishnamoorthy for TEM analysis, and Dr. Kumar for help with the

PLD of the coatings used in this work.

I'd also like to thank all of my friends in Cuong Nhu, especially Chi-Wah, who

suffered with me for many years under some of the best senseis in the martial arts. I will

never find another set of instructors who provide the type of training that I found in

Sensei Mark's and Sensei Joyce's classes.









Although I have made every attempt to include all those who deserve

acknowledgement in these pages, I am sure that there are some who have not been

mentioned. I regret the omission, but after many long days of putting this dissertation

into its final form, it's surprising that more of you are not missing. Please know that each

of you is appreciated.















TABLE OF CONTENTS


ACKN OW LED GM ENTS.................................................................................................. iii

ABSTRA CT ............................................................................................................................ix

CHAPTERS

1 INTROD UCTION ........................................................................................................... 1

2 LITERATURE REVIEW ................................................................................................ 5

Introduction.................................................................................................................. 5
Field Em mission D display (FED )..................................................................................... 6
Cathodolum inescence (CL).......................................................................................... 8
Phosphors ................................................................................................................... 14
Y203:Eu................................................................................................................. 15
Y2SiOS:Tb ............................................................................................................. 17
Efficiency ................................................................................................................... 18
Charging..................................................................................................................... 19
Cathodolum inescent Degradation.............................................................................. 25
Coatings for Phosphors .............................................................................................. 31
Pulsed Laser Deposition (PLD ) ................................................................................. 35
M otivation for this W ork ........................................................................................... 38

3 EXPERIM ENTAL PROCEDURES ............................................................................. 61

Introduction............................................................................................................... 61
Sedim entation of Phosphor Screens........................................................................... 61
Pulsed Laser D position (PLD ) of Coatings.............................................................. 63
Characterization Techniques...................................................................................... 66
Steady-State Cathodolum inescence ...................................................................... 67
Pulsed Cathodolum inescence................................................................................ 69
Ellipsom etry .......................................................................................................... 71
Transm mission Electron M icroscopy (TEM )........................................................... 74
A uger Electron Spectroscopy (AES)..................................................................... 76
Scanning Cathodoluminescence (CL) in a Scanning Electron
M icroscope (SEM ) ............................................................................................. 77








4 EXPERIM ENTAL RESULTS...................................................................................... 88

Introduction........................................................................................................ ........ 88
Cathodoluminescent Spectra...................................................................................... 88
Beam Energy Effects on Luminescence.................................................................... 92
Pulsed Cathodoluminescence..................................................................................... 93
Coating Thickness and Uniform ity............................................................................ 97

5 M ODELING OF RESULTS....................................................................................... 174

Introduction.............................................................................................................. 174
Energy Loss in Dead Layer...................................................................................... 174
Incident Angle Contributions and Coating Uniform ity........................................... 177
Path Length of Electrons in the Coating.................................................................. 184
Calculation of the Cathodoluminescence from Coated Phosphors.......................... 185
New Energy Loss Equation...................................................................................... 188
Backscattering Coefficients ..................................................................................... 190
Index of Refraction .................................................................................................. 192

6 DISCUSSION ............................................................................................................. 233

Introduction.............................................................................................................. 233
Initial Uniform Coating M odel ................................................................................ 233
Validity of the Energy Loss M odel.......................................................................... 236
Backscattering Coefficient....................................................................................... 241
Scattering ................................................................................................................. 242
Surface Recom bination and Charging..................................................................... 243
Surface Segregation ................................................................................................. 244
Surface Roughness................................................................................................... 245

7 SUM M ARY AND CONCLUSIONS.......................................................................... 247

APPENDICES

A MATHCAD DATA AVERAGING PROGRAM...................................................... 252

B LUMINESCENCE DUE TO THE INCIDENT ANGLE OF THE ELECTRONS.... 256

C NONUNIFORM COATING SHAPE ........................................................................ 260

D STOPPING POW ER .................................................................................................. 263

E MATHCAD PROGRAM FOR MgO (4 min) / Y203:Eu........................................... 265

F MATHCAD PROGRAM FOR MgO (8 min) / Y203:Eu........................................... 270








G MATHCAD PROGRAM FOR A1203 (1.2 min) / Y203:Eu....................................... 275

H MATHCAD PROGRAM FOR A1203 (2.4 min) / Y203:Eu....................................... 280

I MATHCAD PROGRAM FOR MgO (4 min) / Y2SiO5:Tb ......................................... 285

J MATHCAD PROGRAM FOR MgO (8 min) / Y2SiO5:Tb......................................... 290

K MATHCAD PROGRAM FOR A1203 (1.2 min) / Y2SiOs5:Tb ................................... 295

L MATHCAD PROGRAM FOR A1203 (2.4 min equiv.) / Y2SiO5:Tb ......................... 300

M MATHCAD PROGRAM FOR A1203 (5 min) / Y2SiO5:Tb...................................... 305

B IB LIO G R A PH Y ........................................................................................................... 310

BIOGRAPHICAL SKETCH ........................................................................................... 322















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

MEASUREMENT AND MODELING OF THE EFFECTS OF PULSED LASER
DEPOSITED COATINGS ON CATHODOLUMINESCENT PHOSPHORS

By

William Joseph Thomes, Jr.

August 2000


Chairman: Dr. Paul Holloway
Major Department: Materials Science and Engineering


Coatings have been shown to reduce the degradation of field emission display

phosphors and, in certain cases, improve the low voltage luminescence. To study the

energy loss mechanisms in coatings and to predict their impact on cathodoluminescence,

coatings of MgO and A1203 were pulsed laser deposited onto sedimented screens of

Y203:Eu and Y2SiO5:Tb. The thickness of the pulsed laser deposited coatings (which

were characterized by ellipsometry, transmission electron microscopy, and Auger

electron spectroscopy sputter profiles) were varied from 7.5 to 50 nm by changing the

deposition time and oxygen background pressure. A Si shadow mask was used to cover

half of the sedimented powder during deposition. This allowed for comparison of coated

with uncoated powder that experienced the same processing. Cathodoluminescence data

were collected for beam energies from 0.8 to 4 keV at a constant current density of

0.16 1A/cm2.








The coating thickness on the particles was modeled by assuming a uniform

deposition flux over a spherical powder surface. Spatially resolved electron energy loss

was calculated to predict the cathodoluminescence intensity versus beam energy and

incident angle relative to the local surface normal. A modified Bethe stopping power

equation was initially used to predict the luminescence from the coated phosphors. This

was found to overestimate the attenuation of the luminescence at beam energies below

about 3 keV. To provide a more accurate fit to the low energy region, a new energy loss

equation based on a form similar to a Makhov power loss equation was introduced. By

using the new energy loss equation, the cathodoluminescence intensity of the coated

phosphors could accurately be predicted for all energies up to 4 keV.

The model developed in this work was used to fit luminescence losses from as

low as 4.71% (at 4 keV) for a 10 nm coating of A1203 on Y203:Eu to as large as 27.4%

(at 4 keV) for a 43 nm MgO coating on Y2SiO5:Tb. No change in the surface

recombination rate was observed. The coatings were also found to have no effect on the

decay states under pulsed excitation, except those attributed to beam energy loss.















CHAPTER 1
INTRODUCTION


Cathode ray tubes (CRTs) have dominated the visual display market since their

invention in the 1920s.1 However, due to their large size, they are not suitable for use in

many modem applications that require compact screens. To fill this role, a class of

devices known as flat panel displays (FPDs) has been invented. These displays are very

thin in comparison to the CRT, which makes them the ideal display media for portable

electronic devices. So far, liquid crystal displays (LCDs) have led the market in the field

of FPDs.2 However, new types of FPDs are being developed to improve upon the LCD.

At the forefront of the new FPDs is the field emission display (FED). It offers

advantages in terms of its wide operating temperature range, wide viewing angle, fast

response time, low power consumption, high brightness, high durability, low weight, and

scalability.2-12 For these reasons, the FED is likely to challenge the LCD for dominance

in the FPD market.

The FED excites luminescence from phosphors deposited on a glass faceplate,

similar to a CRT. However, unlike a conventional CRT that relies on just three electron

guns that create beams that are rastered across the phosphor screen, the FED uses

thousands of tiny emitters behind each phosphor pixel. The depth of a CRT is

determined by the distance necessary for deflection of the electron beam over the entire

screen surface. The FED removes this limitation due to the fact that the emitters are








located directly behind the pixels and can be turned on and off, unlike scanning in the

CRT, in which the electron gun is continuously on. Therefore, the device can be made

very thin (the typical spacing between the emitter array and phosphor screen is around

10 mm).

The proximity of the emitters to the pixels also allows much lower voltage to be

used in the FED (4 or 5 keV, compared to 25 keV in a CRT). The problem with using a

lower voltage, however, is a reduction in the luminescence intensity from the phosphors.

To counteract this loss in intensity, the current from the emitters is increased. The

increased current helps regain the lost luminescence, but it also leads to more rapid

degradation. 13

Due to the operating conditions of FEDs (low voltages and high currents),

phosphor degradation is one of the limiting factors in the lifetime of the device. Coatings

have been shown to greatly improve the degradation resistance of the phosphors.14 One

of the drawbacks to using coatings is that they also cause a reduction in the

cathodoluminescence from the phosphor. Under certain circumstances, however,

coatings have been shown to improve the low-voltage efficiency of phosphors.15 If the

efficiency of the phosphors is increased at low voltages, then the power consumption of

the device and the rate of degradation can be reduced.

The focus of this work has been on examining the effects on cathodoluminescence

of wide-bandgap oxide coatings, specifically MgO and A1203, deposited by pulsed laser

deposition onto two commercially available phosphors, Y203:Eu and Y2SiOs:Tb, used in

FEDs. A review of the literature available on selected aspects of the cathodoluminescent

process is given in Chapter 2. Also included in the chapter is a discussion of the








mechanisms behind the degradation process and how the use of coatings can slow or

eliminate this loss. Near the end of the chapter, the characteristics and merits of pulsed

laser deposition are presented, along with the reason for choosing this technique to

deposit the coatings.

The experimental procedures used to deposit the coatings and collect the

cathodoluminescence from the phosphors are discussed in Chapter 3. This chapter also

includes a description of ellipsometry, transmission electron microscopy, Auger electron

spectroscopy, and scanning electron microscopy, all of which were used to characterize

the coatings.

Results of the cathodoluminescent measurements on the coated samples showed

that the coatings attenuated the luminescence from the phosphors. The effect was larger

at lower voltages due to the interaction volume of the electron beam. Thicker coatings

were found to produce a larger reduction in the cathodoluminescent intensity. These

results are discussed in Chapter 4.

A model was developed to explain the loss in luminescence based on a dead layer

approximation. The model relied on calculation of the energy loss of the electron as it

travels through the coating. Resultant energy after passing through the coating was then

used to find the luminescence. This model allows for calculation of the

cathodoluminescence over the entire energy range up to 4 keV. Spatially resolved energy

loss was found using a cosine dependent coating based on a uniform deposition flux onto

spherical phosphor powders. An improved energy loss equation was incorporated to

account for the extended range of the low energy electrons. Calculated results from this

model were found to accurately predict the cathodoluminescence from the coated







4

phosphors for all measured thicknesses. Derivation of the model is given in Chapter 5.

A discussion of the applicability of the model is presented in Chapter 6.

The conclusions from this work are given in Chapter 7. Appendices A through M

contain the various Mathcad programs used during the calculations.















CHAPTER 2
LITERATURE REVIEW


Introduction

Electronics have become an ever-increasing part of everyday human life. The

ability to process information is continually advancing as new and faster ways of

computing are developed. One important aspect of this trend is the ability of the machine

to transmit data to a human operator. Shown in Figure 2-1 are the many different means

of interfacing between a computer and a person. It can clearly be seen that visual

communication is currently one of the fastest means of data transfer, at around 300

Mb/sec. Many different types of visual signal media can be used to accomplish this goal.

Historically, cathode ray tubes (CRTs) have dominated the visual display market.

However, new types of displays have recently evolved. Most of these new technologies

have been designed to serve a more specific purpose that allows them to compete with

the CRT. For example, the development of smaller, more portable electronic devices

created a need for compact screens. Liquid crystal displays (LCD) have filled this role

thus far, with a 90% market share in 1995.2 However, in demanding applications such as

military uses, medical instruments, vehicles, and dusty environments, to name a few, the

weaknesses of the LCD (one of which is its inability to operate over wide temperature

ranges) can have severe consequences. It is in these areas that new devices are being

developed to challenge the dominance of the LCD. All of these new devices can be









lumped under the general heading of flat panel displays (FPDs) because of their slim

profile when viewed from the side. Each type of display has certain advantages and

disadvantages when used for different applications. Table 1 lists the various FPDs and

some of the attributes of each. Portable applications are not the only market for FPDs;

their strengths over other visual media will allow them to compete in all arenas of display

technology, from tiny head-mounted displays to huge billboard-size screens.

Among the various types of FPDs, field emission displays (FEDs) offer many

advantages, such as wide viewing angle, fast response time for video refresh rates and

"instant-on" capability, low power consumption, high brightness, durability, wide

operating temperature range, low weight, and scalability to a larger size. Each of these

attributes makes FEDs desirable for a wide variety of applications.2-12



Field Emission Display (FED)

Cathode ray tubes (CRTs), the displays found in all conventional televisions, rely

on cathodoluminescence for their operation. Figure 2-2 shows the basic design of a CRT.

An electron gun accelerates electrons toward a phosphor screen. These electrons are

deflected by means of plates or magnetic fields and are then raster scanned across the

screen. Red, green, and blue phosphors are deposited on the screen in small holes known

as pixels. (Actually, each hole that contains a single color phosphor is called a subpixel.

Three subpixels together, one of each color, make up a pixel.) In a full color display,

three separate guns are used, each specific to an individual color.

One of the biggest problems with CRTs is their depth, which is needed in order to

have sufficient distance of travel for deflection of the 20 keV electron beams across the






7

entire screen. FEDs remove this size restriction by using thousands of individual electron

sources behind each subpixel. Figures 2-3 and 2-4 show typical FED setups. Electrons

are extracted from very sharp tips by tunneling when high fields are present. These

electrons are then accelerated toward the faceplate, which contains the phosphors. The

emitter tips are usually located only about 10mm away from the phosphor screen, thus

allowing the display to be extremely thin.

Emitter arrays can consist of up to thousands of tips per pixel. They are usually

made from molybdenum, tungsten, platinum, or silicon2, although carbon has recently

been found to be an effective cathode material as well. Carbon emitters are not grown

like traditional tips and are either diamond thin films or are in the form of

nanotubes.4' 12, 16-18 Regardless of the type of material chosen for the emitters, electron

extraction from the tip is carried out in the same fashion. By applying an appropriate

bias, dependent upon the type of material and the specific device setup, electric fields can

be generated to cause tunneling of electrons out of the tip. This process is accurately

described by the Fowler-Nordheim tunneling theory for electrons.6'7,9, 19-21 Figure 2-5

depicts the field with and without applied bias. As can be seen from the picture, with an

applied bias, the potential energy barrier on the surface has been reduced in height and

has a finite width. The applied field is magnified at the tip because of its extreme

sharpness (its radius is approximately 50 nm). This results in more tunneling current at

lower voltages. Due to the finite barrier and large effective fields near the surface, it is

possible for the electrons to quantum mechanically tunnel out of the tip and enter the

vacuum as field-emitted electrons.22 For a more detailed discussion of the use of the

Fowler-Nordheim equation and suggestions for improving the description of electron








emission from sharp tips, the reader is referred to the previously-cited paper by Cutler

and associates.19

One of the challenges faced in applying this technology in a commercial flat panel

display is achieving uniform emission from the thousands of tips. To combat the

problem of nonuniform emission, resistors have been fabricated in series with the tips to

help control tip-to-tip nonuniformity in current.23 This has provided a partial solution to

the problem, but it is not a perfect solution. The best method would be to fabricate a

constant-current source to the tip to regulate current more effectively. A group from

Japan has reported doing this using silicon tips grown as part of a metal-oxide-

semiconductor field-effect-transistor (MOSFET structure).23 Using such a device

structure allows for uniform and precise control of electron emission. With these

advances in cathode technology, research attention is shifting to the improvement of

other aspects of the FED.

The phosphor screen is an integral part of the FED. In terms of the phosphors,

some of the hurdles that still need to be overcome in order to realize a commercially

viable full color display include: better low voltage efficiency, enhanced chromaticity,

reduced saturation, and lower degradation.3



Cathodoluminescence (CL)

When some materials are impacted with certain forms of energy, they will emit

photons in excess of thermal radiation. This process is known as luminescence, and it

can be categorized based on the type of excitation source. These categories include

photoluminescence, cathodoluminescence, chemiluminescence, triboluminescence, X-ray







9

luminescence, and electroluminescence, among others.24, 25 Photoluminescence refers to

excitation by photons, usually from a UV lamp. Cathodoluminescence is excitation from

a beam of energetic electrons, also known as cathode rays. Chemiluminescence results

from energy released during a chemical reaction. Triboluminescence is a result of

mechanical energy, such as friction or fracture, and can also be seen during processes

such as grinding. X-ray luminescence comes from X-ray excitation, as the name

suggests. Finally, electroluminescence is produced by an electric field resulting from an

applied voltage across the material. It should be noted however, that

thermoluminescence does not refer to thermal excitation from the ground state. Instead,

this term is used to describe the thermal stimulation of electrons from excited state traps

that can then recombine and produce luminescence. The applied heat does not actually

excite the electrons; this occurred during a prior excitation event. Rather, the heat only

gives the electrons the energy needed to surmount the energy barrier holding them in the

trapped state.25

In addition to the characterization of luminescence based on the different types of

excitation, there is a distinction made according to the length of the delay between the

excitation and emission of photons. Materials can exhibit either fluorescence or

phosphorescence. "Fluorescence" is the term used when the emission of photons occurs

within about 10-8 seconds after excitation. "Phosphorescence" pertains to any material

that displays luminescence for longer than this. These phosphorescent materials are

called phosphors, and they can display luminescence lifetime ranging from 10-7 seconds

to hours after the excitation source is removed.24









Phosphors have many technologically important roles, especially in the display

industry. They are responsible for converting incident electrons into light in devices such

as CRTs and FEDs. However, their luminescent decay rate must be considered in

determining their suitability in a particular application. In a display, the phosphor must

continue to luminesce long enough to display an image, but also extinguish quickly

enough to allow for fast video refresh rates. This is accomplished by choosing suitable

phosphor materials and processing, which allows the decay time of the phosphor to be

tailored.26 This is not usually a concern because most phosphors display a suitable

response time without any special processing. However, in applications such as the new

high-definition TVs, in which very fast video refresh rates are needed, the decay times of

various phosphors under different types of processing will need to be taken into

consideration.

When an electron beam enters a solid, it will undergo collisions with the host

atoms. These collisions are either elastic or inelastic. During elastic collisions, the

incident electron interacts with a nucleus of the atom and is deflected with little or no

energy loss. If the electron is deflected back toward the surface, it can be lost from the

solid as a backscattered electron. Inelastic collisions involve interaction between the

incident electron and the electrons of the atom. In these collisions, the incident electron

loses part of its energy to the atom. When the electrons in an atom return from the

excited state to their ground state, a host of signals can be produced. Among them are

Auger electrons, X-rays, secondary electrons, photons, and phonons (thermal effects such

as electroacoustic signals).24 For this study, the generated photons

(cathodoluminescence) are of primary interest.






11

The incremental energy lost (dE) over a distance of travel (ds) due to an inelastic

collision can be found from the Bethe equation14,24' 27-30:

dE _-785pZ l(1166EE [2.1]
ds A*E [ J

where A is the atomic weight in g/mol, Z is the atomic number, E is the electron energy,

p is the density in g/cm3, and J is the mean ionization potential in eV. The latter is the

average energy loss per interaction (for all possible energy loss processes), and for Z > 13

it can be expressed as:

J = 9.76Z + 58.5 [2.2]
Z0.19

These equations have been shown to be accurate for E > 6.34 multiplied by the mean

ionization potential.14'24, 28-30 Below this energy, the modification by Rao-Sahib and

Wittry needs to be considered30:

dE = -785*p*Z [2.3]
ds 1.2588*A*V*K V

This equation corrects for the low energy region while mimicking the behavior of the

original Bethe equation at high energy. This new equation can be used for energies as

low as approximately 500 V. Attempting to reduce this limit even further, Joy and Luo

suggested a new expression for the mean ionization potential using an energy dependent

term14, 29, 31:

J='- [2.4]
1+k -
[E








where J' is the new mean ionization potential, J is still the average energy loss per

interaction, and k is a fitting parameter. The constant k varies from 0.7 to 0.9, but it is

usually around 0.85.

The stopping power equations cited above provide a measure of how the electrons

lose energy in the solid, but they do not give an accurate indication of the interaction

volume. During the electron's trip through the lattice, it will also undergo elastic

collisions. These will not result in an energy loss, but rather will lead to a change in the

direction of the electron's travel. Equation 2.1 can be integrated to find the length of an

individual electron "random walk" trajectory. This is known as the Bethe range.

However, the interaction volume will be much smaller due to scattering. Typically, the

interaction volume is known as the Gruen, electron beam, or penetration range. It can be

expressed as:


Re = [2.5]


where p is the density, E0 is the electron beam energy, k' depends on the atomic number

and is a function of energy, and a depends on the atomic number and E. Various

equations have been given for the solution to this problem, but the one most widely

accepted was proposed by Kanaya and Okayama:

rf0.0276A [,667
\pZ08 )
'Re Jo [2.6]


measured in gm, where Eo is in keV, A is in g/mol, p is in g/cm3, and Z is the atomic

number.24 Figure 2-6 demonstrates how this interaction volume moves deeper into the

material as the beam energy is increased.









Every incident electron gives off energy to the lattice in multiple steps.

Therefore, it is possible for one electron to generate a multitude of secondary electrons.

The number of electron-hole pairs generated per incident electron is known as the

generation factor; it can be expressed as:

=E(1- [2.7]
E,

where y represents the fractional electron beam energy lost due to backscattered

electrons. E, is the ionization energy, the energy needed to form a single electron-hole

pair. The ionization energy is related to the bandgap of the material by:

E1 =2.8E9 +M [2.8]

where M is between 0 and 1 eV, depending on the material in question. For a more

detailed discussion, the reader is referred to Yacobi and Holt's work on the subject.24

Once the incident electron has transferred its energy to the solid, this energy can

be used to produce visible light. Luminescence from inorganic phosphors usually takes

place at an impurity, which is referred to as an activator.24' 32, 33 However, not all

impurities are activators. Those that do not lead to luminescence are referred to as

quenchers or killers.32 Since the activators are an imperfection in the crystal lattice, they

will usually exhibit energy levels between those of the conduction and valence bands of

the host lattice. These states can become populated either through direct excitation or

trapping of secondary electrons generated in the host lattice. When the activator returns

to its ground state through recombination of an electron and hole, the energy can be

released as luminescence.24,25, 32-36 This is known as a radiative transition. When the

atom relaxes without the emission of a photon, it is called a nonradiative transition.









Competition between these two types of transitions determines the intensity of the

emitted light.25

When an electron and hole radiatively recombine, the emitted photon will be

characteristic of the energy levels involved, which is determined by the electronic states

of the activator. The photon energy (hv) is given by:

hv = Ef -E, [2.9]

where Ef and Ei are the energy of the final and initial states, respectively. In the literature

on the subject, this energy is often referred to instead by its wavelength equivalent.

There is a simple relation relating the two:

1.2398 [2.101
E

where E is the photon energy in eV and X is in Rm.24, 36 For most materials, the

recombination of the electron and hole will be the rate- limiting step to the luminescent

process.37



Phosphors

Phosphors are responsible for converting incident energy into photons (light).

Without them, visual display media would not be possible. There are a plethora of

different materials that exhibit this behavior. The following is only a partial list of

phosphors that can be used to generate the three primary colors (red, blue, and green) of a

common display. For red light, there are Y202S:Eu, Y203:Eu, CaS:Eu2, SrS:Eu2, and

YVO4:Eu. For green light, there are ZnS:Cu:Au:Al, Zn2SiO4:Mn, Y2SiO5:Tb,

Gd2O2S:Tb, La2O2S:Tb, ZnO:Zn, Y3(AI,Ga)5012, Y2GeO5:Pr, Y202S:Pr, ZnGa204:Mn,

Y3AI5012:Ce, and SrGa2S4:Eu. For blue light, there are ZnS:Ag:CI, Zn2SiO4:Ti,






15

Y2SiO5:Ce, ZnGa204, and SrGa2S4:Ce.3,38 For this work, Y203:Eu and Y2SiO5:Tb were

chosen to be representative of the red and green groups, respectively. This allowed for

examination of not only the coating effects, but also the consequences of the underlying

host material. The attributes of each will be discussed below.



Y203:Eu

Y203:Eu shows a main emission peak in the visible spectrum at 611 nm (red)

under cathodoluminescent excitation. This results from a 5Do 7F2 transition of the Eu+3

site.25'39-43 Figure 2-7 shows such a transition. Eu is a rare earth ion and is

characterized by an incompletely filled 4f shell. The 4f 6 shell is shielded by the filled 5s2

and 5p6 shells.25'41 Shielding allows the Eu ion to retain its atomic character and have

energy levels close to that of a free ion.25, 44 As a result, the emission is very sharp at the

transmission wavelength.44 This is also the reason that Eu is usually able to keep its

characteristic luminescence, regardless of the host material. There are two possible

crystallographic configurations for the Eu atoms, either C2 or C3i.39-41 Figure 2-8 shows

these. The main emission from the phosphor is due to transitions in the C2 sites. Energy

absorbed into the C3i sites is effectively transferred into the C2 site for recombination.40

In the literature, C3i symmetry is often referred to as S6 symmetry.39

Of importance in the Y203 host is the charge transfer to the luminescent center.

The highest occupied levels of the ground state in the host lattice are the 2p orbitals of

oxygen, while the lowest unoccupied levels are a mixture of the 3s orbitals of oxygen and

the 4d orbitals of yttrium. When electrons are promoted to an excited state under electron

bombardment, they can be transferred into the 5D levels of Eu.25 The configurational








coordinate diagram can be used to represent such a transition. For a fixed central atom,

the diagram represents the potential energy (E) curves of the atom as a function of the

distance between it and its neighbors (R). Figure 2-9 shows such a diagram. The solid

lines in the figure represent the ground states and excited states of the activator, while the

dashed line represents another excited state of the activator known as the charge-transfer

state. The parabolic shape results from the assumption that the restoring force on the

neighboring atoms is proportional to their displacement from equilibrium. The charge-

transfer state is one which excited electrons can easily be fed into because of allowed

transitions from the ground state of the activator or from the host lattice. The

luminescent state is not directly accessible because the transitions are forbidden. In

Figure 2-9 (a), the minimum in the charge-transfer band causes excited electrons to be

effectively transferred into the luminescent states. This is the case for Y203:Eu. If,

however, the charge-transfer band had a minimum located closer to the ground state

bands, as shown in Figure 2-9 (b), then excited electrons would be effectively transferred

into the ground state through nonradiative relaxation. Due to the location of the charge-

transfer band in Y203:Eu, excited electrons populate the luminescent states, giving this

phosphor its high quantum efficiency.45

The efficiency of different FED phosphors are shown in Table 2.2 Y202S:Eu has

a larger efficiency than Y203:Eu, but it has other disadvantages in terms of degradation

that make it a less desirable material for FED applications. 14









Y2SiOs:Tb

Y2SiO5:Tb is a green phosphor with a main emission at 538 nm.46 This is a result

of a 5D4-7F5 electron transition on the Tb3 site, as shown in Figure 2-7. The other three

peaks seen in a typical Tb-doped spectrum are the result of transitions to the other 7Fn

states.47 As with the Eu atom in yittria, the Tb in Y2SiO5 is well shielded by electrons in

the outer shells. In a similar fashion, this usually makes the resulting spectrum invariant

to host lattice composition.44,48 It also makes the transitions exhibit little spread in the

emitted wavelength.44 Also similar to Y203:Eu, Y2SiOs:Tb has a charge-transfer band

that efficiently deposits excited electrons into the luminescent states, as shown in

Figure 2-9 (a). In the Y2SiOs:Tb system, this is the 4f75d band of Tb, and it results in a

relatively high quantum efficiency.25

One of the properties that makes Y2SiO5:Tb favorable over other green phosphors

is its saturation behavior. In a FED application, as compared to a CRT, lower voltages

are used (hundreds to a few thousand volts versus 10 to 20 kV). To regain brightness,

higher currents are employed (up to 1000 A/cm2 versus 0.5 A/cm2).11.43 Most phosphors

will begin to show saturation effects as the current is increased, especially at these low

voltages.3 This is speculated to result from ground state depletion.49'50 As the ground

state of the activators becomes depleted, the probability of exciting an electron into a

luminescent state is reduced. The result is an increase in the nonradiative recombination

rate, which leads to a reduction in the efficiency of the phosphor (see section on

efficiency in this chapter). However, this is not as prevalent in Y2SiOs:Tb because of the

fast decay rate of the excited state. Because the excited electrons recombine so quickly,

they refill the ground state and are thus available for excitation again.49' 50 Because









Y2SiO5:Tb shows little or no saturation effects, it is gaining considerable interest in the

display market.



Efficiency

A phosphor is "a solid which converts certain types of energy into

electromagnetic radiation over and above thermal radiation."25 In other words, when an

electron beam strikes a phosphor, it can impart some of its energy to exciting electrons

into higher energy levels. The rest of the energy is lost in collisions with the host

material. The excited electrons can then recombine with a hole in a radiative or

nonradiative transition. Radiative transitions give off electromagnetic radiation, whereas

nonradiative transitions transfer energy to Auger electrons, ionizations, or phonons.25

Efficiency is defined as the brightness per unit of input power and is measured in

lumens per watt.3,25,39,51 In a phosphor, charge transfer to a luminescent center is

critical to achieving high efficiency. If the electron and hole are not effectively

transferred to the luminescent impurities, then they will have a high probability of

recombining nonradiatively. This is one of the major problems in FEDs, where lower

voltages are used, as compared to CRTs. As the accelerating voltage is decreased, the

efficiency also decreases. 11, 35, 51-53 Table 3 and Figure 2-10 depict this situation. From

the table, it is apparent that the reduced efficiency at low voltage is one of the primary

disadvantages to FED operation. In order to regain high brightness, larger currents need

to be used. These larger currents, in turn, lead to saturation effects and enhanced

phosphor degradation and thus short device lifetime.3









Many researchers have attempted to devise ways of improving efficiency in this

low voltage (under 4 kV) range. One method involved placing the phosphor material

directly on top of the gate electrode, as seen in Figure 2-11. Other researchers have tried

using quantum confinement to enhance luminescence by spatially confining electrons and

holes. This method usually requires phosphor particles smaller than 10 nm in diameter,

as compared with the 5 nm phosphors currently in production.54-59

The problem with most of the aforementioned approaches is that they tend to be

very difficult to implement. The fundamental problem to solve in order to improve

efficiency is that of how to transport charge to luminescent centers while decreasing the

probability of the charge reaching nonradiative sites. In order to realize this goal,

charging in the phosphor during electron bombardment needs to be controlled.



Charging

Surface recombination refers to the return of excited electrons to the ground state

by recombination with an available hole at the phosphor surface. Due to the disruption in

the phosphor lattice at the surface, there will be a large number of defects in this region.

These defects provide efficient nonradiative pathways for electron-hole recombination.

In order to alleviate this problem, surface passivation is often used to regain

luminescence. Surface passivation involves coating the surface with another material so

that charge is effectively radiated and not allowed to reach the nonradiative surface

states.60-63 These surface states become critical when considering charging of the

phosphor. This is especially important in a FED environment, where low voltages are

used. As the accelerating voltage of the incident electrons is decreased, the generation of








electron-hole pairs will occur closer to the surface due to the lower penetration of the

incident electrons.50 Figure 2-6 shows a schematic of the electron beam penetration into

the phosphor at various beam energies. From the figure, it can be seen that at lower

voltages the electrons are localized closer to the surface. Therefore, surface states will

have a larger influence on the generated secondary electrons at these lower electron beam

voltages.

While surface recombination is possible, other effects can occur upon electron

bombardment of phosphors. Yoo and Lee attributed decreased luminescence on the basis

of trapped electrons on the surface. They postulate that low energy incident electrons are

trapped on the surface. These trapped electrons create an electrostatic potential barrier

for subsequent primary electron bombardment and therefore decrease luminescence.63

Ozawa further contributes to this explanation with what he calls secondary bound

electrons. When an incident electron enters a material, it creates secondary electrons in

the near surface region during interaction with the host. If these electrons are within the

mean free path of secondary electrons, then they can be ejected from the material. The

secondary electron coefficient (5) is the number of ejected secondary electrons per

incident electron. Typical values range from 1.5 to 3 for a primary beam energy of about

1 keV, meaning that more secondaries escape the surface than enter it. These leave

behind holes, which give the surface a net positive charge. If the secondaries do not have

sufficient energy to escape or reenter the crystal, then they can become trapped a short

distance above the surface. An electron cloud is thus produced above the surface, known

as a space charge, and it acts as a barrier for subsequent incident electrons.35






21

Bennewitz and associates studied the surface potential by looking at desorption of

F' ions from a CaF2 sample under 1 keV electron bombardment. In their work, they

point out that the positive surface potential is a result of secondary electron loss from the

near surface region. Therefore, it does not depend strongly on the primary beam current,

but only on the secondary electron distribution. The primary beam current will influence

this potential when there are current pathways to the surface from other pathways, such

as desorption of positive ions from the surface, or a leakage current between the

irradiated spot and the sample holder.64

Seager et al. have postulated that instead of trapping electrons outside the surface,

internal fields lead to charging during electron beam bombardment.52 Utilizing

secondary electron emission energies and carbon Auger peak shifts during irradiation

with electron beams of energies from 0.5 to 5 keV, they were able to provide evidence of

surface charging. Four phosphors were studied: ZnS:Ag, SrGa2S4:Eu, ZnO:Zn, and

Y203:Eu. All were powders sedimented onto -1 cm2 Au foil. Particle diameters varied

from 1 to 10 imn, and the layers were about 5 to 10 particles thick. When the researchers

changed the potential of the plate that the samples were mounted on, surface potential on

the front of the sample (closest to the beam) varied. Figure 2-12 shows results for

Y203:Eu, which is representative of the other phosphors. The figure shows that the

energy of the C Auger electrons changed linearly with application of the applied bias.

This suggests that the backing plate bias led to a change in the surface potential that

altered the escape energy of the Auger electrons. Similar shifts were seen for the

secondary electrons. Thus, it is possible to change the potential of the phosphor with

application of a backing plate bias.









Seager et al. also used dual electron beams to irradiate the sample during

excitation. The second electron flood beam was used to control surface potential of the

phosphor. Their results led them to believe that secondary electrons are not recaptured

and used to build an electron cloud, as postulated by Ozawa, Yoo, Lee, and others; rather,

an internal mechanism is responsible for bringing the phosphor into a steady state

condition. They postulated that this internal mechanism is a buildup of positive charge

near the surface, with a corresponding negative charge near the end of range of the

incident electrons.

One problem with the experiments of Seager et al. is the fact that the potential of

the front surface of the phosphor was unpinned. This made application of a known field

within the phosphor difficult. To circumvent this problem, Seager applied a metallic

mesh to the front surface of the phosphor to pin the surface potential at electrical ground.

Electric fields were then applied across the phosphor layer to either enhance or retard the

cathodoluminescence, depending on polarity.65 Two phosphors -- Y203:Eu and

Y2SiO5:Tb -- were sedimented onto brass plates. Total thickness of the phosphor layer

was around 25 pm. Metal mesh with square holes (7.5 gm) was attached to the top

surface of each sample. These grids allowed electric fields to be induced within the

phosphor layer while still allowing incident electrons to reach the phosphor surface.

Figure 2-13 shows this sample configuration. It was found that the maximum voltage

that could be applied before breakdown was between 400 and 450 V. This voltage is

important because it gives a measure of the dielectric strength of the material.

Initial results suggested that internal fields could be used to sweep generated

charges away from or toward the surface. This would, in turn, lead to an increase or








decrease in cathodoluminescence because of losses to surface recombination.65

However, further experimentation led to the conclusion that the results seen could be

caused by surface potential fluctuations. Instead of changing the internal fields of the

material, these fluctuations caused the energy of the incident electrons to be altered. The

most likely explanation for the unpinned potential of the surface is that there was an air

gap between the phosphor and grid as a result of the roughness of the powder.66 Thus,

intimate contact is needed if the internal fields of a material are to be altered in a

controlled fashion. Coatings applied directly to the phosphors could provide such a

situation.

Recently, Pantano et al. published an explanation of charging that takes into

account the internal fields, with the addition of allowing for a negative surface potential.

They rely on a model proposed by Cazaux67 which depends on the stored charges per unit

area, q+ and q-, and the charge densities, p+ and p.:


q =J08t P -
ds

qt J0(1-rnt)
q_ = Jo(1-Tlt) P_= -
dp

where Jo is the incident primary-beam current, 8 is the secondary-electron emission

coefficient, T" is the backscatter coefficient, t is the time, ds is the thickness of the

positively charged surface layer, and dp is the penetration depth of the incident

electrons.68 If 1 ri ) 8, the total negative charge q- will exceed q+ in time and the surface

will charge up negatively. The negative surface charge will then act as a potential barrier

for incoming electrons. This can only occur for 8 < 1, such as for beam energies below








about 1 kV. The reduced energy of the incoming electrons will limit the luminescence

from the material. If 8 > 1 il, the positive surface charge q+ will exceed the negative

end of range charge q.. This process will be self-regulating because as the positive

surface potential increases, the low energy secondary electrons will not have enough

energy to escape. The result is a limited, but stable, positive surface.68 Under such

conditions, the excited electrons generated within the phosphor are swept toward the

surface, where there is a high probability of nonradiative recombination.

One possible method from the presented models above for affecting the

luminescence is to change the velocity of the incoming electrons. Yoo and Ozawa claim

that charge buildup on the surface produces such a velocity change.35,63 Internal fields

could also build up similar surface fields if 8 < 1, as shown in the models by Seager and

by Pantano et al. As a consequence, incident electrons are repelled by an electrostatic

potential at the surface, which lowers their velocity. Due to their lower incident energy,

the electrons that make it into the material have less energy with which to excite

luminescence. Based on calculations of the dielectric strength of the phosphors, as

discussed below, however, this does not seem to be a reasonable model.

Dielectric strength is a measure of the maximum electric field that can be applied

across a dielectric before breakdown. The equation is as follows:

(V/l

where represents the electric field, V represents the voltage, and d represents the

distance. Most materials have a dielectric strength of 105 to 107 V/cm.69, 70 For

Y203:Eu, it has been reported that breakdown occurred at 450 V over a 25 gim film.65

This would correspond to a dielectric strength of 1.8* 105 V/cm, which is within the









expected range. The value calculated is lower than the bulk dielectric strength because

the numbers were taken from a powder sample. Therefore, the number has a larger

contribution from the lower surface dielectric strength than from the higher bulk

dielectric strength. Given the above dielectric strength and that the penetration depth of

the electrons within the phosphor surface is on the order of 1000 A (probably an

overestimate), breakdown would be expected at around 1.8 V. Even if one assumed a

dielectric strength one order of magnitude larger, breakdown would be expected at 18 V.

When this is compared to the energy of the incident electrons (500 4000 V), it does not

seem possible that these small voltages could be affecting the velocity of the incident

electrons enough to cause a large decrease in luminescence.

Considering the above argument, the most plausible explanation of charging is

that internal fields are causing excited secondary electrons to be swept toward the

surface, where they have a high probability of nonradiative recombination or emission.

These internal fields are a result of a positive surface charge due to secondary electron

emission and a negative charge region near the end of range of the primary electrons. In

order to reduce charging in phosphors, the internal fields need to be minimized or

removed so that generated secondary electrons remain in the bulk of the phosphor, where

they have a higher probability of radiative recombination. This, in turn, will lead to a

higher overall efficiency.



Cathodoluminescent Degradation

One of the main problems that needs to be overcome in order to make an FED

device for mass production is that of increasing FED lifetime. This is especially









important for a low voltage device. As the accelerating voltage of the electrons is

reduced, the brightness of the device decreases rapidly, due to a loss in efficiency. To

attempt to regain the original luminescence, higher currents are used.43 In turn, these

higher currents cause operational problems in the device. Besides saturation, as

discussed earlier, degradation is greatly enhanced due to the high current densities

employed. When the lifetime of the device is considered, the degradation of both the

phosphor and field emitters needs to be taken into account. Both can lead to a reduction

in the cathodoluminescence seen from the FED.

In a CRT, it is relatively easy to maintain a good vacuum over the lifetime of the

device. This is due to two reasons. First, the relatively small surface-to-volume ratio of

CRTs makes it easy to initially pump the device. Second, there is a large area over which

to apply a getter material, giving a large ratio of active getter area to system volume. The

FED does not have either of these benefits. Due to its closely spaced anode and cathode,

it has a large surface-to-volume ratio. This makes it difficult to initially pump the device

due to the conductance between the plates. There is also little area over which to apply

the getter material.71 Some improvements to the pumping could be realized by

increasing the gap between the plates, but this creates problems with the focusing of the

electrons and would require the insertion of a focusing grid into the device. Placing the

getter between the anode and cathode could help with the pumping, but this too would

cause problems with the actual operation of the display.71 New techniques are being

explored for gettering, such as placing a non-evaporable getter around the sides of the

package or attaching a getter to the back with small holes to allow pumping. Although








these might help in the maintenance of the vacuum within the FED, the role of gases

needs to be considered when examining the device operation.

There are two main sources of background gas in a FED device, those resulting

from desorption from the device structure and those released from the phosphor due to

electron beam impingement. Gases released from the phosphors depend on the specific

material in question and will be discussed later in this section. All other gases are mainly

a result of outgassing from the structural components of the device, such as the spacers,

cathode, and black matrix. Other sources of background gas are present -- for example,

permeation -- but these usually contribute a negligible amount to the overall composition

of gases in the device. The main outgassing products inside a FED are H2, H20, CO,

CO2, and hydrocarbons such as methane, ethane, and propane.72'73 Depending on the

relative amount of these gases and the specific material, various processes can lead to

decreased cathodoluminescence. If the environment is dominated by carbon-containing

gas molecules, then carbon deposition on the phosphor surface will most likely

result.74-76 Under steady-state conditions, gas molecules are constantly physisorbing on

and desorbing off the phosphor surface. When an electron beam is present, it can impart

energy to these molecules and cause them to crack (i.e., break the bonds holding them

together) to atomic species. The carbon is then free to form bonds with the surface, while

the other components combine with each other and are released back to the gas phase.

The rate at which this will occur depends on the residual vacuum pressures of the specific

atoms and molecules. As the carbon layer grows, it will substantially reduce the

luminescence from the phosphor particle in the electron beam exposed area.74-76 After






28

exposure in such an environment, a dark spot is usually present that can be seen with the

naked eye.

If hydrogen and water, rather than carbon compounds, dominate the FED

environment, then degradation is still present, but it often will be a result of a change in

the internal efficiency of the phosphor.77 Although different materials degrade in

different ways and at different rates, certain generalizations can be made for various types

of phosphors. Some of the most efficient high voltage phosphors currently known

contain sulfur. Examples include Y202S:Eu for red, ZnS:Ag:Cl for blue, and

ZnS:Cu:Al:Au for green. Under the electron beam, dissociation of residual gas

molecules can occur. These atoms can then form bonds with the atoms in the outer layer

of the phosphor. Due to the volatile nature of the sulfur compounds such as SOx, H2S,

etc., they desorb back into the vacuum. As a result, sulfur is leached from the

surface.13, 78-83 It would be expected that this phenomenon would be limited to the near

surface region, however, the lack of sulfur causes a diffusion gradient that brings more

sulfur to the surface region.68 As the sulfur is removed, the particle surface is converted

into its oxide equivalent. Swartz et al. showed that when ZnS is subjected to an electron

beam, its surface is converted into the non-luminescent ZnO.13,78'80,82 Trottier

demonstrated that Y202S:Eu is transformed into the less efficient Y203:Eu on the surface

under an electron beam. In addition to the decrease in luminescence, a peak shift was

observed for Y202S:Eu that was due to the emission from the Y203:Eu layer.13, 14 These

types of degradation reactions have been termed electron-stimulated surface chemical

reactions (ESSCR) because of their mechanism.13' 14








It would seem, based on the mechanism discussed above, that oxide-based

phosphors would not be susceptible to electron beam degradation. However, this is not

the case. Oxides usually degrade because of electron-stimulated desorption on the

surface. This causes the introduction of lattice defects in the material.68, 84 Because they

are a disruption in the lattice symmetry, defects tend to increase the nonradiative

recombination rates. As the rate of nonradiative transitions increases, the efficiency of

the phosphor decreases and, thus, so does the cathodoluminescent brightness. Although

this is fundamentally a surface phenomenon, the electron interaction proceeds deep into

the material. Once the outermost surface is changed, an activity gradient exists to drive

the diffusion of species in subsurface regions. Thus, the damage continues into the

phosphor. Compounds most affected by this are ones in which the cation and anion have

Pauling electronegativity differences greater than 1.7, a category that includes many

oxides.68 In a way, this is similar to the sulfide case described above. Even though

ESSCR causes the majority of the degradation in sulfide-based materials, defect

introduction is occurring simultaneously to further decrease the luminescence.

In all of the scenarios discussed, the electron beam caused the degradation to take

place. It seems natural, therefore, that the length of exposure determines the extent of

degradation. By convention, this is measured as the coulomb load to which the phosphor

has been exposed. It can be expressed as:

Coulomb load (Coulombs/cm2) = (I t) / A

where I is the sample current (Coulombs/sec), t is the time of exposure (sec), and A is the

area of the beam spot (cm2). For this reason, this type of degradation is often referred to

as Coulombic aging.85-87 There have been different formulas presented to represent the






30

loss in luminescence with Coulomb load, but the one most widely referenced was derived

by Pfanhl at Bell Labs. It simply states that the intensity as a function of dose can be

expressed as:
I(N)=- O

(I+ c'N)

where I is the cathodoluminescent intensity at any given dose N (number of

electrons/cm2), Io is the initial cathodoluminescent intensity, and C' is the burn parameter

(cm2).14,71 The bum parameter C' will often be replaced by a quantity l/Qso% in which

Q50% is equal to the charge dose, for which the initial intensity of the phosphor is

halved.88 Not all phosphors behave in such an ideal way. Sometimes a phosphor will

show an initial rapid decrease in luminescence that is followed by a slight rise and

eventual inflection and further decrease in intensity. To account for this rise, Cappels and

associates have derived an equation that uses two decay curves similar to the equation

above.89 In essence, their equation is just the superposition of two simple inverse decay

curves.

Holloway, et al. modeled the degradation they observed in sulfide-based

phosphors based on an ESSCR sulfur removal rate. Their model predicted that the

concentration of sulfur on the surface is exponentially dependent on the dose. It is a

more in-depth study of the exact mechanisms behind the degradation process. Their

equation is as follows:

1 e(K'PN)
e
10






31

where K' is a constant, Prg is residual gas pressure, and N is the electron dose. However,

they pointed out that within experimental error, the data could be made to fit their

equation or Pfahnl's.13

At the beginning of this section, it was mentioned that the electron beam can not

only degrade the phosphor, but can lead to degradation of the emitter tips as well.

Usually, this is an indirect contamination due to volatile species being ejected from the

phosphors. It is well established that sulfide-containing phosphors "poison" the emitter

tips during excitation with an electron beam.11, 90-93 This is suspected to be a result of

the ejected volatile species traveling through the vacuum and reacting with the emitters.

Upon reaction, these species can change the work function of the tip material. Due to the

change in work function, the emission characteristics will be altered.3,94 Therefore, not

only a sulfide, but any phosphor has the potential to degrade the cathode emitter tips.

Consequently, the more volatile species have a much larger effect on field emission from

the cathode than their more stable counterparts. For this reason, oxides have received

considerable attention. However, due to efficiencies lower than sulfides or oxysulfides,

they represent a trade-off in performance.



Coatings for Phosphors

Key issues in designing phosphors for low voltage FED applications are improved

efficiency and reduced degradation. Many researchers have attempted to use coatings to

accomplish these goals.

In CRT manufacturing, an aluminum backing layer is applied to the phosphor

screen. Usually, a lacquer film is first deposited to provide a planer surface on which to








deposit the aluminum film. Then the lacquer is baked out. The aluminum reflects

generated light back toward the front of the device to improve brightness and contrast.

Even though the aluminum is an effective energy barrier for the incoming electrons, it is

not a problem due to the high voltages used in CRTs.71,95

Many groups have looked at coating for protecting phosphor particles from

electron degradation. Some of the earlier work in this area was performed by Kingsley

and Prener. They used sol-gel processes to coat ZnS:Cu particles with non-luminescent

ZnS layers. Thickness of the layers was determined based on weight gain and size

distribution of the powder after processing. Figures 2-14 and 2-15 show results of their

work. As can be seen from the figures, the coatings did not change the slope of the

luminescence versus voltage curves; they only changed the turn-on voltage. The turn-on

voltage is found by taking an extrapolation from the linear portion of the luminescence

versus voltage curve. The voltage at which the linear extrapolation crosses the axis (i.e.,

when the intensity is zero) is the turn-on voltage. The fact that the coatings only changed

the turn-on voltage suggests that the coating did not alter the luminescent states in the

phosphor, but only caused a decrease in the incoming electron energy. (A layer that

exhibits this type of behavior is often referred to as a "dead layer" because the energy lost

in the layer is no longer available to excite luminescence.) Using this assumption,

Kingsley and Prener were able to model the decrease in luminescence with a Makhov

power loss equation with corrections for the spherical surface and surface

recombination.96 The Makhov power loss equation is given by:

P(x, jo,Vo)= joVo exp(-X2 XO9)






33

where P is the beam power per unit area, jo is the beam current density at the surface, Vo

is the accelerating potential, and X is a normalized depth in the material given by:


X(X,VO) = X-
X
- p *C*Vo


where p is the density, C and n are materials constants, and x is the depth in the

material.96

As noted above, sulfide-based coatings may be degraded under electron beam

exposure due to ESSCR at the surface. For this reason, other materials have been studied

as potential capping layers. Bechtel reports on the use of a phosphate coating to improve

the degradation characteristics of ZnS-based phosphors.97 Dmitrienko suggests the use

of stable oxides such as SiO2, MgO, or A1203, however, he notes that the optimal

thickness and deposition techniques still need to be determined.98

One of the first studies to show the protection offered by applying coatings to

phosphors was done by Trottier and Fitz-Gerald. They demonstrated that coatings of

TaSi2 and Ag deposited by pulsed laser ablation onto Y202S:Eu, a highly efficient red

phosphor, could be effective at reducing the degradation of this phosphor. Figure 2-16

shows the results from their work. The pulsed laser deposited coatings of TaSi2 and Ag

were compared against wet-chemistry coatings of SiO2 and phosphate. The pulsed laser

coating was better at slowing degradation and exhibited far less loss in brightness after

aging to 20 C/cm2. Of the two laser ablated coatings, TaSi2 protected the underlying

phosphor better than Ag. This is believed to be due to Ag migration or volatilization

under the electron beam. 14, 99

Recent interest in phosphor coatings has concentrated not only on their protective

properties, but also on their use to increase cathodoluminescent efficiency at low






34

voltages. The group of Yang and Yokoyama showed that coating CuxS on ZnS:Ag:Cl led

to an increase in the efficiency of the phosphor.100 However, sulfides are susceptible to

ESSCR in a FED environment. It seems that wide-band oxides are the preferred choice

for increasing the efficiency of the phosphors. 101 The group of Kominami et al. showed

that coating ZnS:Ag:Cl with In203 led to improved efficiency at voltages below 500 eV.

They attributed this to an increase in the conductivity of the phosphor, which allowed

charge to be dissipated. An increase in the phosphor lifetime was also reported.102

Villalobos and associates coated SiO2 on ZnS:Ag particles using a newly designed spray

coating process, in which the host particle is suspended in a liquid mixture and then

sprayed into a hot zone where the coating gels on the surface. The particles are then heat

treated as a final step. Using this technique, increases in both the brightness and

degradation resistance were reported. 103

Attempts to model these phenomena have been made by groups at the Georgia

Institute of Technology.15,93' 104 They used a 1-D discrete computer model to calculate

the efficiency of the phosphor with varying thicknesses of coating applied. A correction

term was added to account for surface recombination velocity and the band offset of the

coating compared to the host material. Experimental results were collected for SiO2

coatings on ZnS:Cu:AI deposited by sol-gel methods. If the pH was kept around 6.5 and

the silica concentration at 1 wt%, then enhancement of the cathodoluminescent efficiency

was possible below around 2 keV. However, the enhancement was only seen for coatings

less than 40 nm thick. Beyond this value, electron stopping in the coating was reported to

overshadow the surface recombination gains. 15,93, 104









It should be noted that processing of the phosphors will have an impact on their

cathodoluminescent response. Many commercially available phosphors have undergone

a series of growth, grinding, and annealing steps. Many researchers have shown that

improved cathodoluminescence is possible by tailoring these processes to produce a more

uniform and less defective particle surface.105' 106 The reader should be aware of this

when examining some reports on the effects of coatings. Some authors will apply

coatings to phosphors that have been grown in their own laboratory. The coated

phosphors will then be compared against commercially available materials of the same

composition. This is an unfair comparison because the commercial powders may have a

lower CL response than the laboratory-grown phosphors. Thus, it is difficult to separate

the effects of the coating from those of the phosphor processing.

Another issue to consider when coating phosphors is the consequences of the

coating on the ability to apply the phosphor to the anode faceplate. Certain materials will

cause aggregation of the phosphor particles in the slurry before screening, while others

will help them stay dispersed. In order to create screens with the highest possible

brightness, dispersion of the particles needs to be maintained, and thus the coatings must

help maintain this condition.107



Pulsed Laser Deposition (PLD)

Advances in PLD have made it an increasingly popular method for applying

coatings to materials. 108, 109 In this technique, short laser pulses are used to evaporate

material from a target. Several different things happen during the laser pulse. Lowndes

describes these as "rapid heating and vaporization of the target; increasing absorption by








the vapor until breakdown occurs to form a dense plasma; and absorption of the

remainder of the laser pulse to heat and accelerate the plasma."110 The particles in this

plasma undergo collisions, which in turn produce a highly directional expansion away

from the target. All of this is occurring in the Knudsen layer above the target."10 These

evaporated materials undergo gas phase collisions and are subsequently deposited onto a

substrate. By controlling the ambient gas mixture, background pressure, substrate

temperature, and laser energy and duration, films can be grown to desired specifications.

Due to the method of ablation, the flux of material leaving the target is strongly peaked in

the direction perpendicular to the target surface.110. 11 However, recent advances in the

use of target rotation and apertures have made uniform deposition over larger areas

possible. 112

Of particular interest are the advances being made in the application of PLD to

optical materials. An example includes the deposition of ultrathin (<10 nm) indium tin

oxide (ITO) films. ITO is one of the most widely used transparent conductive coatings in

electro-optic applications.113 PLD is particularly good for growing oxides due to the fact

that the material transfer from target to substrate is normally stoichiometric, meaning that

the deposited film has the same composition as the target. 110, 114 This is a result of the

high initial rate of heating and highly nonthermal target erosion. 110 Large bandgap

semiconductors and dielectric materials usually have large optical absorption at short

wavelengths. Therefore, large amounts of the laser energy are deposited in a small

volume close to the surface. As a result, the sublimation temperature of the material is

attained at depths greater than the thermal diffusion distance of the constituents. Target






37

components are not able to segregate, therefore they leave the target in proportion to their

bulk concentrations. This results in deposits with the same composition as the target.39

By varying the conditions under which thin film phosphors are deposited, their

optical properties can be improved. Y203:Eu is a good example of these effects. It has

been shown that by increasing the oxygen partial pressure during deposition, rougher

films are produced. This leads to an increase in brightness due to forward scattering by

anomalous diffraction.39' 115 Microcrystallites of f-IV compounds such as CdTe and

CdS have been grown in a similar fashion. 116 Temperature of the substrate can also

affect film properties. Models suggest that higher substrate temperature will lead to

rougher films. 117 The type of substrate will also determine whether amorphous,

polycrystalline, or epitaxial single-crystal films are grown.110 In one experiment,

Y203:Eu was grown on bare (100) Si wafers and on diamond-coated Si wafers. The

diamond coating was prepared using a hot filament chemical vapor deposition. The

phosphor films were grown under identical conditions on the two substrates. There was a

substantial improvement in luminescent brightness from the diamond-coated substrate.

This was attributed to the higher roughness of the surface; see Figures 2-17 and

2-18.41,118

Sapphire substrates have also been used to improve the brightness of Y203:Eu

thin films. The increase in brightness is believed to be due to the low absorption and low

refractive index of red light in sapphire and to the improved growth of grains with

unidirectional orientation on the (0001) sapphire substrates.119

Finally, Y203:Eu films have been epitaxed onto LaA103 substrates. These are

good substrates because the lattice mismatch is only 0.8% with an orientation of









[110]Y20311[100]LaA103 and [-110]Y203||[010]LaAlO3. Z-contrast scanning

transmission electron microscopy (STEM) was used to demonstrate the absence of

precipitates of Eu in the deposited films.120 This is advantageous in a luminescent film

because the activators are thus spread out.

Besides its use to deposit phosphor thin films, PLD has also been used to coat

phosphor powders to improve luminescence and reduce degradation. Fitz-Gerald

deposited Y203:Eu onto powders of SiO2 and A1203. The powders were agitated in a

fluidized bed setup so that the deposited coating would be uniformly distributed over the

powder surface.121-123 The photoluminescent spectra showed the 612 nm peak from the

5Do-7F2 transition in Y203:Eu; so the coatings were effectively deposited onto the core

particles.122 Transmission electron microscopy (TEM) was used to confirm that the

coatings were continuous.123 Next, coatings of TaSi and Ag were deposited on

Y202S:Eu, a highly efficient red phosphor, to help reduce degradation under electron

beam excitation. As discussed in the previous section, degradation curves clearly showed

the advantage of these two coatings for protecting the phosphor.14, 99



Motivation for this Work

Coatings have been shown to be effective in slowing the degradation of

phosphors, especially at the low voltages and high currents of a FED environment.

Improvements in the efficiency of cathodoluminescence at low voltages have also been

demonstrated. Thus far, most of the work on coatings has focused on putting a coating

around the entire phosphor particle, using such techniques as sol-gel processing or spray

pyrolysis. These steps add considerably to the cost of making the powder and must be








monitored closely to ensure proper results. In this work, the commercially available

phosphors Y203:Eu and Y2SiO5:Tb were sedimented onto Mo substrates and coated using

pulsed laser deposition. This is a much simpler process that can be applied to large area

screens using existing coating techniques.

To ensure that all results observed were due to the coating and not the processing

of the screen (such as deposition, CL geometry, etc.), half of the screen was masked

during deposition. This allowed side-by-side comparison of coated and uncoated

phosphor material. Coatings of MgO and A1203 were deposited at various thicknesses to

examine their effects on cathodoluminescence. These materials were chosen because

they both have large work functions compared to the host materials. This should lead to

enhanced efficiency due to repulsion of electrons from the particle-coating interface and

thus a lower surface recombination velocity.

Phosphor powders were chosen for this study instead of thin films because they

represent real world devices. Thin films provide an excellent surface on which to study

the basics of electron interactions, but they can introduce other problems. Lattice

mismatch between the luminescent film and substrate can lead to a drastic reduction in

cathodoluminescence. This can be alleviated in some instances by the careful choice of

substrate. (For example, LaAIO3 for growth of Y203:Eu.124) However, this is often a

difficult challenge. There is also the problem of a much lower intensity due to greater

internal reflection in the thin film.39 Because the initial intensity from the thin films is

low, detecting changes in the brightness becomes more difficult. Because phosphor

powders are used in FED devices and offer high initial brightness, they were chosen for

this study.
























Direct
Connection
to Brain


I ........Sens -- Visual (300 Mb/oec)

Audio (20 kb/ec)

Touch (10 b/sec)


Smell (4 b/sec)

Heat (0.1 b/sec)






Figure 2-1 The different means of interfacing between a human and a computer. Also
shown are the speeds of each type of transfer.6



















Table 1

The different types of flat panel displays and the advantages of each.


Teehnnlnov


Tvne


Size


Power


Advantages


Barriers


lat----. e.L _Rugged,___high-_
AMEL emissive small moderate Rugged, high- Pixel size limit, lack
resolution, high- of suitable phosphor
brightness for full-color.
OEL emissive small low High luminous Lithography
efficiency, low incompatibilities,
drive voltage instability, short
lifetime, limited
temperature range
Plasma emissive large high Mature High-power, slow
technology, high refresh rates, limited
______ ___brightness military market
Laser light large, moderate High brightness, Still in early stages
Projection valve scaleable display size is of development, no
scaleable full-scale prototype
MEMS light moderate Some Development of new
Projection valve large, commercially technologies costly,
scaleable available (TI- digital artifacts may
____________ ________ DMD), scaleablebe a problem
Reflective light small to very low Very low power, Limited temperature
LCD valve medium bistable range.
manufacturability,
durability


FED


emissive


small to
medium


low


Low power, mign
brightness, wide
viewing angle


Manuracturaoility,
lifetime


Reproduced from reference 8.
The acronyms in the table stand for the following:
AMEL Active Matrix Electroluminescent
OEL Organic Electroluminescent
LCD Liquid Crystal Display
FED Field Emission Display






42

















/^I"


c/t Electron
_Beam





Electron Gun Vertical
Deflection Horizontal
Plates Deflection
Plates

Phosphor
Screen


Figure 2-2 Typical cathode ray tube (CRT), consisting of an electron gun and deflection
plates to raster the electron beam across the phosphor screen.6

























Anode --
Assembly

'"I^i rr "J 'I~ .I


Spacer







Cathode
Assembly

Baseplate Row Electrodes Column Electrodes



Figure 2-3 A typical FED setup, consisting of the cathode assembly, which houses the
field emitter tips, and the anode assembly, a glass plate on which the
phosphors are deposited. Spacers are used to separate the two halves of the
device.























light
B i I l T T Transparent Conductor (-1000V),



EmttF a plate+
... .... ........ ....... .................tte... ....

j \;/ \ / \ .\;/ Electrons i j \;
I, t /^ xtracionti u/ u/ If I II



IEmitter Baseplate Emitter & Elerodes


Figure 2-4 Schematic cross section of a typical FED. Not drawn to scale.9


Ph




















Enery


-Fex


Total Potential


Figure 2-5 Potential energy curves of an electron near a metal surface. The "Image
Potential" is with no external field. An externally applied field that is
electrically negative to the surroundings can be represented by the "Applied
Potential" curve. The resultant "Total Potential" will then be the
superposition of the two curves. Due to the lower energy barrier at the
surface, electrons in the metal have a finite probability of quantum
22
mechanically tunneling into the vacuum.
























ary electron


beam



phosphor




Electron penetration in a phosphor at various beam energies. Incident beam
energy is increasing from A to B to C. This also demonstrates how the
interaction volume moves closer to the surface as the beam energy is
decreased. When surface recombination is considered, the interaction
volume in A will clearly have a stronger influence from the surface than that
in C. As a result, surface effects become more pronounced at lower beam
voltages .35


prim


Figure 2-6
























Erouc2 --
35 -



Esmxr+




25 -
-.-'O3

20 SD



CP)



10 C

o 3 2 H, --'

_H4


Fa7

W* -7F. -7F

Ce03 Gd3 Tm3 Tbh3 Eu3'




Figure 2-7 Expected transitions in various rare-earth dopants used for FED phosphors.47










, a a a 0 0 000*l


0,

Y


~ IN WW m


C2


S6


Figure 2-8 The two different crystallographic configurations of Eu in a Y203 host lattice.
The S6 symmetry is often referred to as C31 symmetry.4


















CT E
c, E c!
%, I % %I


a_



Figure 2-9


\,TH h T R
_ R b i R



The dashed curves represent the charge-transfer (CT) state. In (a), the CT
state helps feed the emissive 5D levels because of its position relative to the
7F ground state. This is the situation found in Y203:Eu and Y2SiO5:Tb. In
(b), the offset of the CT state causes the electrons to be returned to the ground
state, thus reducing the luminescence of the material.25















Table 2


Composition, color and efficiency
FED phosphors


at low-voltage operation of selected


Composition Color Efficiency (Im W )

500 V
ZnO:Zn Green 10.7
Gd202S:Tb Green 7.9
YAI O2:Tb Green 2.0
Y202S:Eu Red 3.5
Y20:Eu Red 2.2
YVO4:Eu Red 0.4

300 V, 131 iA
CaS:Ce Green 3.10
SrGa2S4:Eu Green 3.00
La2O2S:Tb Green 5.20
Gd2O2S:Tb Green 3.52
Y202S:Eu Red 2.20
Y203:Eu Red 1.57
Y2SiO.:Ce Blue 0.25
Y2SiO5:Tb Green 1.05
LaOBr:Tb Green 1.95
LaOCI:Tb Green 0.36
LaOBr:Tb Blue 0.54
LaOBr:Tm Blue 0.17
Reproduced from reference 2.























Table 3

Classification of FED structures
Anode Anode-cathode
Type voltage separation Advantages Disadvantages
(kV) (Mnm)

High CRT phosphors 8 *Focussing electrodes required
-voltage 4-10 2-3 processes applicable *Spacers with a high aspect ratio
-High efficiency *Breakdown

Medium 1-4 0.2-0.8 *Simple structure -Spacers with a relatively higk
-voltage 'Fair efficiency aspect ratio
Low -Simple structure -'Limited efficiency
-voltage 0.4-1 0.1-0.2 'High reliability
*Low cost

Reproduced from reference 11.












































500 1000 1500 2000 2500 3000 3500 4000 4500
acceleration voltage [volts]


Figure 2-10 Efficiency versus acceleration voltage for Y203:Eu.38

























faceplate


Electrons


Phosphor


Figure 2-11 FED with phosphor on the gate electrode to double the light output from the
device.6
device.






























2 3 0 1iiiiiiiiiiiiiiiiiiiiiiiiiiii
-40-30-20-10


0 10


20 30 40


-COLLECTION BIAS ( V )


Figure 2-12 Carbon Auger peak shifts after changes in the bias of the sample mounting
plate.52


310


0

ei

w
z
w
01


w
z

w
a

U
,=x:


300


290


280


270


260


250


240
















- m m


- 0 0 a U m


a an .a


PHOSPHOR
LAYER


II


BRASS BACK
PLATE


m


Figure 2-13 Setup for metal mesh placed on the phosphor surface for taking internal field
measurements under electron excitation.65


MESH


v


E.I




















S02 0.127p

>- 14 t T
a: .:^ /
I.-








0 4 8 12 16 200.254









ACCELERATING POTENTIAL (kV)



Figure 2-14 Luminescence versus accelerating potential for ZnS-coated (0.127 lHtm,
0.254 Inm, or 0.389 r thick) ZnS:Cu. The similar slope of the lines
suggests that the coating did not alter the luminescent states of the
phosphor, but only decreased the incoming electron energy.z
LU
rO O-389F
LU 6
Q

4-

2-


00 4 8 12 16 20
ACCELERATING POTENTIAL Mk)



Figure 2-14 Luminescence versus accelerating potential for ZnS-coated (0.127 g.m,
0.254 g~m, or 0.389 g~m thick) ZnS:Cu. The similar slope of the lines
suggests that the coating did not alter the luminescent states of the
phosphor, but only decreased the incoming electron energy. 9






57




20 -|1 i-|-| --- |-1 -\-1 -1

18 0.0508AF(

16- 0
S0.0847 0

| 14- o84
49

12-
S0. 188,

8- o



C.,
co
S4- -


0.50opt
2-
A/
0 A,, / I I I A I ~ -



0 4 8 12 16 20
ACCELERATING POTENTIAL (kV)



Figure 2-15 Luminescence versus accelerating potential for ZnS-coated (0.0508 pnm,
0.0847 gm, or 0.188 Lim thick) ZnS:Cu. The similar slope of the lines
suggests that the coating did not alter the luminescent states of the
phosphor, but only decreased the incoming electron energy.96















1

0.9

. 0.8
*
S0.7

I o.6

d 0.5


0.4

0.3


0 5 10 15 20 25
Coulomb Load (C/cm2)


Figure 2-16


Cathodoluminescence degradation of Y202S:Eu with various coatings. The
SiO2 and phosphate coatings were applied by wet chemistry techniques,
while the Ag and TaSi coatings were deposited by pulsed laser ablation in a
fluidized bed setup. "Original" refers to the uncoated powder.14
















































Figure 2-17 Atomic force microscopy (AFM) images of Y203:Eu film grown on bare Si
substrates and on diamond-coated Si substrates. The images clearly show
the higher roughness of the films on the diamond-coated material.41











100 ..
p Eu:Y203 Film on Diamond
90 "-i- Eu:Y203 Film on Silicon

o? 80

Z 70
Co
S 60

~g 50

"0 40
N
-1 30
20
z
10

0 I ..I I.....II
300 400 500 600 700

DepositionTemperature (C)


Figure 2-18 Photoluminescence from Y203:Eu films grown on bare Si and diamond-
coated Si substrates. The curves show the higher luminescence attained
with the diamond-coated substrate. This is attributed to the higher
roughness of these films.41














CHAPTER 3
EXPERIMENTAL PROCEDURES


Introduction

The effects of coatings on field emission display (FED) phosphors were studied in

this work. Coatings of A1203 and MgO were deposited by pulsed laser deposition (PLD)

onto screens of Y203:Eu and Y2SiOS:Tb. These screens were prepared by sedimentation

from a phosphor slurry. During deposition of the coatings, a Si shadow mask was used to

provide coating on only half of the sample. This allowed for direct comparison of coated

and uncoated phosphor material.

Many different types of characterization techniques were used to measure the

samples. Cathodoluminescence measurements were taken under steady-state and pulsed

conditions to examine the response of the phosphors to an electron beam. Ellipsometry,

transmission electron microscopy (TEM), and Auger electron spectroscopy (AES) were

used to determine the thickness of the coatings. Finally, AES and a scanning electron

microscope (SEM) with a fiber optic attachment used for measuring CL maps were used

to examine coating uniformity.



Sedimentation of Phosphor Screens

The phosphor screens of Y203:Eu and Y2SiO5:Tb were prepared by sedimentation

of the powder from a slurry. The apparatus consists of plastic beakers with a hole drilled








in the bottom. Attached through this hole is a connector that allows for a tube to be

inserted onto the bottom of the beaker. A vise-style clamp on this tube controls the flow

of the liquid out of the container. Samples are held down on the bottom of the vessel by

spring clips that are attached along the bottom edge.

The first step in the sedimentation process is preparing the slurry mixture. The

phosphor is weighed and added to a magnesium nitrate hexahydrate and isopropyl

alcohol mixture. This is then placed in a sonicator for 30 minutes to ensure complete

dissociation of the phosphor throughout the slurry. Natural heating also takes place

during the agitation, which helps hold the powder in suspension. At this time, the

substrates are cleaned by rubbing them with a methanol-soaked swab and then immersing

them in methanol and placing them in the sonicator for about five minutes. After

cleaning, the substrates are attached to the bottom of the beaker. The hose is then

attached and clamped off. With all of this in place, the slurry is poured into the beakers.

It takes many hours for the phosphor to settle out of solution and coat the samples.

Usually, the beakers are covered and left overnight. In the morning, the liquid is slowly

drained from the beaker by loosening the clamp on the hose. The samples are then air

dried in the container for several hours.

Both of the phosphors used for this work, Y203:Eu and Y2SiOs:Tb, were standard

Nichia powders. These were chosen because they were readily available and would be

typical for use in a FED. Molybdenum was chosen as a substrate material for all of the

samples because of its good electrical conduction and high melting temperature.

(Originally, gold was used as a substrate material because it is inert to most materials and








has a very high electrical conductivity. Unfortunately, the gold would not endure the

750 C temperature needed during the pulsed laser deposition of the coating materials.)



Pulsed Laser Deposition (PLD) of Coatings

Since PLD has been shown to be particularly good for growing oxides due to the

fact that the material transfer from target to substrate is generally stoichiometric110' 114, it

was chosen as the means of depositing the A1203 and MgO coatings for this work. The

apparatus used consists of a stainless steel chamber pumped by a Pfeiffer Balzers TPU

450 H corrosive turbomolecular pump with a TCP305 controller and backed by a MD41

diaphragm pump. A Leybold Trivac mechanical pump model # D4A was used for

roughing the chamber. To control background pressure, a gate valve was used between

the pump and chamber to adjust the pumping speed. Precision leak valves provide

further control of pressure through backfilling of the chamber with a variety of gasses.

Ablation energy comes from a KrF (X=248 nm) 1 Watt Lambda Physik Lasertechnik

LPX300 excimer laser. Pulse width was set at 10 nanoseconds and laser energy kept at

about 350 mJ for all depositions. Pulse frequency was variable between 1 and 100 Hz,

but 10 Hz was chosen for all experiments. Laser light enters through a viewport on the

side of the chamber. Control of the laser dimensions and position is accomplished with

collimating and focusing lenses located on an optics bench positioned between the laser

and vacuum chamber.

Directly across from the laser entry viewport inside the chamber is the target

holder. The holder is positioned so that the laser strikes at 45 from the surface normal,

which in turn generates a plume perpendicular to the target surface.110, 111 The target








holder is connected to a rotary motion feedthrough to allow target rotation during

deposition, thus allowing better utilization of the targets. Target rotation also helps

minimize deposition of large particulates from the target onto the growing film.125

Substrates are mounted on a sample holder located directly across from the target.

Target-to-substrate distance was set such that the samples were located near the end of

the plasma plume, which was around 3 centimeters in the chamber used for this work.

Plasma expansion models suggest that this is the optimal position for the substrate.114

The sample was not rotated during deposition. Heating of the samples was accomplished

using a quartz lamp located within the sample holder. This allowed for a range of 150C

to 1000C, although 750C was used throughout these experiments. A stainless steel

heater plate was used to transfer heat to the substrates, which were mounted to this plate.

A thermocouple attached to the plate provided accurate control of the deposition

temperature throughout the run.

The A1203 and MgO targets were made from their powders. First, approximately

five grams of the powder was placed in a hardened stainless steel die. The die used was

designed to produce targets an inch and a half in diameter. A hydraulic press was used to

compress the powder in the die to around 120 psi for 4 minutes. The green powder

compacts of these two materials were found to hold together very well. Both materials

came out of the die without excessive force, so no lubrication of the die was necessary.

Had such lubrication been used, it could have been a source of contamination for the

target. For similar reasons, no flux was used to help hold the compacts together, although

a flux is occasionally added to increase the sintering of the powder in the compact.

During firing of the compacts, the flux is supposed to evaporate, but if it does not, it can








also be a source of contamination. The compacts were then placed in a tube furnace at

1200C for 12 hours. This anneal is used to greatly increase the density of the target.

Finished targets were then silver pasted onto holders used to attach them to the rotary arm

of the target holder.

Phosphor screens were attached to the stainless steel heating plate of the sample

holder by silver paint. A piece of Si was placed over half of the screen to act as a shadow

mask. A spring clip attached to the side of the sample holder was used to hold the mask

in place. The target and substrate were then inserted into the vacuum chamber. Once

sealed off, the chamber was evacuated to a base pressure of around 1 105 Torr.

Temperature of the substrate was then raised to 750C. At this point, the gate valve was

partially closed and the precision leak valve opened to adjust the background pressure of

oxygen in the system. Oxygen was fed to the leak valve from ultra-high purity

compressed gas cylinders. A MKS model 247C mass flow controller set at 80 sccm was

used between the gas cylinder and leak valve. Background pressure of 50 mTorr was

used for the majority of the experiments, with 200 mTorr used for the remainder (as

noted later in the experimental results section).

Contrary to most other deposition techniques, such as electron beam, thermal

evaporation, and sputtering, PLD does not show a strong decrease in deposition rate with

increasing pressure above a few mTorr. This is due to the inherent quasi-free jet

expansion of the plume during pulsing. The particles in the plume impart momentum to

the background gas molecules. Therefore, the background gas and plume travel together

toward the substrate. This greatly reduces the scattering of deposition species due to






66
collisions with gas molecules. The result is a very long mean-free path for the particles,

which allows deposition at higher pressures."1

Control of the coating thickness was achieved by closely monitoring the

deposition time. Materials ablate at different rates, so the coating time had to be adjusted

for each of the material systems used. For the system used for this work, it was found

that the deposition rate at 50 mTorr was around 65 A/min for A1203 and 18.2 A/min for

MgO.



Characterization Techniques

Many different characterization techniques were used to determine the coatings'

thickness and effect on the luminescence of the phosphor. The techniques include

steady-state and pulsed cathodoluminescence (CL), ellipsometry, transmission electron

microscopy (TEM), Auger electron spectroscopy (AES), and scanning CL in a scanning

electron microscope (SEM).

Figure 3-1 shows the energy distribution of electrons emitted from a sample

surface under electron bombardment. Electrons in various energy regions on this curve

are measured in some of the aforementioned characterization techniques. The secondary

electrons are used for secondary electron (SE) imaging in the SEM. The Auger electrons

are measured in AES. The elastically backscattered electrons are used for diffraction

analysis in the TEM.








Steady-State Cathodoluminescence

Cathodoluminescence (CL) is the process of generating light by electron beam

excitation of a phosphor material. Incident electrons impart some of their energy to the

phosphor material through electron excitation. These excited electrons can then

recombine with an available hole (the absence of an electron) to produce visible

light.25, 35

Cathodoluminescence measurements were carried out in a stainless steel vacuum

chamber pumped with a CTI-Cryogenics cryopump and a Varian ion pump. Roughing

was accomplished using a rotary-vane mechanical pump. Base pressure for this

combination of pumps was between 5* 10-9 and 1 108 Torr with no bake-out. Although

baking the chamber would have allowed for a lower pressure by removing some of the

desorbed gas on the walls, it was not an option due to a fiber optic feedthrough.

However, these pressures are more than adequate for taking CL measurements. Electron

excitation was from a Kimball Physics EFG-7F electron gun with a EGPS-7H power

supply and RGDU-3C raster generation unit. The electron gun was capable of energies

up to 5 keV and currents from 0.01 to 500 giA. Focus was set to maintain a spot size of

approximately 2 mm.

Phosphor luminescence was collected using a fiber optic connected to an Ocean

Optics S2000 spectrometer. This spectrometer uses a diffraction grating and a charge

couple device (CCD) array detector, which allows luminescent spectra to be collected

over the entire visible wavelength range. To further improve wavelength resolution, a

homemade slit was fashioned on the end of the fiber before it entered the spectrometer.

This greatly reduced the broadening of the luminescent peaks. This is especially helpful








in materials with closely spaced peaks that can become indistinguishable due to peak

overlap.

To improve signal-to-noise in the collected spectra, the raster unit was utilized to

redirect the electron beam off-axis. Due to the electron gun design, the filament used to

generate electrons within the gun is visible through the exit aperture. This causes a bright

white spot of light on the sample directly in front of the gun. Using a mechanical

manipulation system, this filament light can be redirected away from the area of interest

on the sample. This greatly reduces the background signal detected by the spectrometer.

By setting the appropriate bias on the raster unit, voltage was applied to deflection plates

on the electron gun to redirect the electron beam to the center of interest on the sample.

Accurate current measurement requires the collection of secondary as well as

primary electrons. For this reason, a +100 V bias from a Fluke 343A DC Voltage

Calibrator power supply was applied to the electrically isolated sample carousel. A

Keithley 619 Electrometer/Multimeter was connected in series between the power supply

and the carousel to measure the current. Background subtraction was used to account for

any stray currents present in the system. A switch was installed between the meter and

carousel to allow the carousel to be shorted to ground. This is necessary during

measurement of the luminescence because the sample bias will impart extra energy to the

incoming electrons.

The sample carousel held samples perpendicular to the incident electron beam.

Samples were inserted such that the uncoated half of the powder screen was above the

coated half, instead of beside it. In this arrangement, the electron beam could be moved

from an uncoated to a coated area by adjusting the vertical position of the carousel.








Using only vertical repositioning assured that the electron beam-to-sample and sample-

to-fiber alignment would not be changed. It is critical to keep these alignments constant

because changing them will result in a change in the measured luminescence.



Pulsed Cathodoluminescence

Pulsed cathodoluminescence measurements are used to study the decay of the

luminescence from a phosphor under a pulsed electron beam. The same chamber was

utilized as in the steady-state CL measurements above. The Kimball Physics electron

gun has a fast pulsing option. A grid located inside the gun is used to blank the beam.

By applying a large negative bias to the grid, a potential barrier for the electrons is

established. The magnitude of this potential barrier, adjusted through the grid voltage,

determines how many electrons can make it past the grid. If an electron has enough

energy to surmount this barrier, then it will proceed through the gun as normal. Pulsing

is accomplished by sending a positive voltage pulse along the grid supply line to reduce

the grid retarding voltage. By selecting the appropriate grid bias, the positive pulse can

be used to turn the electron beam on and off. Thus, the characteristics of the positive

pulse determine the electron beam characteristics. Kimball Physics supplies a box that

can be installed in the grid voltage line for such an operation. This box allows the pulse

generator to be connected to the grid line and prevents the grid voltage from being sent

back through the line to the pulse generator. Due to the design of the electron gun, the

grid voltage (typically 100 to 250 V) is added to the accelerating voltage (typically 0.5 to

4 keV) before being transmitted down the grid line. During operation, it was noticed that








the pulsing box would not give consistent pulsing. To correct this, the circuit was

redesigned and a new box built to provide a stable, low noise operation.

A Hewlett Packard (HP) 8112A Pulse Generator was used to provide the positive

voltage pulses to turn the electron gun on. Pulse width was set to 10 is with a delay of

65 ns and a period of 100 ms. Amplitude was adjusted to produce a 16 volt turn-on

pulse.

Luminescence was collected using a fiber optic mounted inside the vacuum

system. The other end of the fiber was connected to a Hamamatsu Photomultiplier Tube

(PMT). A 1150 V power supply was used to power the PMT. A box was used to house

the PMT with a fitting for the fiber on the outside. During operation, a thick black piece

of fabric was wrapped around the box to reduce the amount of stray light. A holder was

designed for the inside of the box to allow filters to be placed between the fiber and PMT.

Filters were used to select specific wavelength regions. This gave the ability to study the

decay of single peaks (different decay states) of the phosphors. Filters used included a

535 nm bandpass and Coming glass 2-73, 3-70, and 5-58 highpass filters. Figures 3-2 to

3-4 show the transmission of these Coming filters. The output of the PMT was

connected through a "T" junction to an oscilloscope and a HP 3478A Multimeter. The

multimeter was used to look at the voltage from the PMT due to background light in the

chamber, most of which was due to filament light from the electron gun. During actual

measurement, the multimeter was disconnected to reduce the amount of noise in the

signal. The oscilloscope used was a 500 MHz 2 GS/s Hewlett Packard Infinium

Oscilloscope Model 54825A. A built-in disk drive was used to save files and transfer

them to another computer.








Sweep rate on the oscilloscope was chosen to ensure that the entire decay pulse

was measured. This meant making sure that the after-pulse luminescence returned to

background level. Once the sweep rate was set, the sample averaging was adjusted to

give 25044 data points per scan. To further improve signal-to-noise, 3000 scans were

collected for every set of conditions. The oscilloscope automatically averaged these as

they were collected. Therefore, every data file consisted of a 25044x2 matrix of time and

intensity. Comparison of multiple files of this size is very computer processor intensive,

so further averaging was done. Files were loaded into Mathcad, where a sliding time

average was applied. The program takes a preset number of data points and finds an

average, which it assigns to that interval. It then moves on to the next interval and

repeats the process. In this fashion, the entire data set is averaged to produce a more

manageable size. See Appendix A for a more detailed description of the program.

Typically, the data interval was set for 25 points and the last 44 points discarded, which

produced a 1000 point file.



Ellipsometry

Ellipsometry, also known as polarimetry and polarization spectroscopy, can be

used to obtain the thickness and optical constants of thin films.126 Some of the

advantages of this technique over others is that it is simple to operate, nondestructive, and

requires no vacuum system. Measurements can be taken in a vacuum system, in air, or in

a liquid.36 Figure 3-5 shows the typical experimental arrangement. Collimated

monochromatic light is passed through a polarizer (Glan-Thomson calcite prism) and a

quarter-wave compensator (mica plate with 45 retardation) to give elliptically polarized








light. This is reflected off the sample surface into a second polarizer (Glan-Thomson

calcite prism) that acts as an analyzer. The polarizer and analyzer are then rotated to find

the maximum extinction of the reflected light. 127

It is important to know the angle of incidence of the incoming light and its

wavelength and to keep these constant throughout the measurement. Once the initial set

of polarizer and analyzer settings is found, the two are rotated to find a second set of

extinction conditions. The polarizer is adjusted to 90 plus the original polarizer angle.

The analyzer is adjusted to 180 minus the original analyzer angle. Both are then rotated

to find maximum light extinction. In order to ensure accurate measurements, the new

angles should not differ from the calculated positions by more than four degrees.128 With

these data, the values of psi (T) and delta (A) can be determined from the relationships:

S_180 -(A2 -Al)
2

A= 3600 -(P + P)

where A2, A&, P2, and P, are the second and first set of analyzer and polarizer angles,

respectively. 127, 128

Light is an electromagnetic wave and therefore must obey Maxwell's equations.

As a consequence, there are certain relations that must be obeyed when light encounters a

boundary between two media. First, the angle of incidence must equal the angle of

reflectance. Second, in the case of one material on top of another, Snell's Law must

hold:


n, sin 01 = no sin 00









where niand no are the complex indices of refraction in material 1 and material 0 and 01

and 02 are the angles from the surface normal in material 1 and material 0.36 Third, the

Fresnel reflection coefficients are given by:

no Cos 00 n, Cos 01
no cos 0o + n, cos 01

n, cos 00 no cos 0

no cos0o + n, cos01

where s refers to the light vector component perpendicular to the plane of incidence and p

the parallel component.127 The plane of incidence is defined by the incident and reflected

beams and the surface normal.36 It is these last two relations that are important for the

ellipsometry measurement. They are related to the T and A values obtained earlier by the

following equation:

rp = (tan )T .
r,

These equations can also be related to the reflection coefficients through the following:

R rol,, + rlsub^,se-2i
gs +
I + rol, rl,, e~-2i

R rol.p + rlsubPe-2
1 + ro.p'rsub,pe -2i

where the subscripts 0, 1, and sub represent the measurement medium, film, and

substrate, respectively. The film thickness, di, can be found from the equation for the

phase angle, 13:


P=27j i cos0.









It is this equation that relates the index of refraction to the thickness, as well as to the

phase changes due to reflection at the interface.36' 126 There are many computer

programs available that use these relations to determine film thickness or index of

refraction from the I and A values found from the measured analyzer and polarizer

settings. Due to time and space constraints, a more detailed discussion of the equations

will not be attempted here. However, it is possible to analyze multiple films using this

technique, although such analysis is very complex.127, 129

For the current work, two different types of ellipsometry equipment were

utilized. The first was a Gaertner Scientific Corp. ellipsometer with a HeNe laser

(632.8 nm) and manual polarizer and analyzer. For this apparatus, the angle was set at

70, and the sample positioned to give reflection into the detector. Once this was set,

polarizer and analyzer readings could be collected. The second piece of ellipsometry

equipment used was a fully automated rotating analyzer ellipsometer. In this

arrangement, a fixed polarizer is used and the analyzer is rotated to determine extinction

values. Advantages of the latter include multiple wavelength capabilities and more

accurate measurement of the analyzer and polarizer relative positions. The computer

program also allows for multiple angles to be used to further improve analysis. Built-in

data libraries for most elements and compounds make data analysis much simpler and

more precise.



Transmission Electron Microscopy (TEM)

TEM is a powerful technique for obtaining information about the atomic structure

of a material. The main requirement is that the sample is thin enough to transmit






75

electrons.36, 126 Figure 3-6 shows a typical TEM setup. For the current work, TEM was

used to determine coating thickness from the Si shadow masks used during deposition.

One of the most difficult aspects of using TEM is sample preparation. Since coating

thickness was desired in the present study, cross-sectioned samples were prepared. This

is done by first thinly slicing the sample using a diamond saw. Two of these thin slices

are then laid flat with the coated sides touching. Wax is used to hold the sample together.

In order to obtain a region thin enough for electrons to propagate through, the sample is

placed in an ion-mill. This machine uses energetic ions to sputter thin the sample. Once

a hole appears, the sample is ready for analysis. By looking at regions on the periphery

of the hole, films sufficiently thin for analysis yield information about the sample.

In the microscopy, electrons are scattered as they pass through the samples. It is

the nature of the scattering that determines the type of information that is obtained. There

are two types of scattering events: elastic and inelastic scattering. Elastic scattering is a

result of Coulombic interactions of the incoming electron and the potential field of the

ion cores, and it results in no loss in energy to the electron. This type of process is

known as Rutherford scattering and gives rise to diffraction patterns. The magnitude of

the interaction scales with the charge on the nucleus, and thus with the atomic number.

Inelastic scattering is the interaction between the primary electron beam and electrons in

quantum states around the nuclei or in the solid. Energy is transferred during the

scattering, giving rise to spatial variation in the intensity of the transmitted beam

dependent upon defects and heterogeneities.126 By examining the intensity of the

transmitted electrons, different layers and their interfaces are visible. From this, the

thickness of the coating layer can be determined.










Auger Electron Spectroscopy (AES)

AES is a very useful tool for looking at surface compositions. By utilizing ion

beam sputtering, composition versus depth is attainable. Figure 3-7 shows a typical AES

setup. Ultra-high vacuum (UHV) conditions are needed to reduce surface contamination

during analysis. Sample characterization takes place as follows. Energetic primary

electrons (-5 keV) are focused by an electron gun onto the sample surface. These

primary electrons lose energy as they traverse the sample. Similar to the process

described for the TEM, some of this lost energy goes into exciting ground state electrons

into empty quantum states of the atoms or into continuum energy states. The atoms have

two options for recombining an electron with an available hole and returning to the

ground state. These are to produce an (1) X-ray or (2) Auger electron; see Figure 3-8.

Both of these processes happen simultaneously, but this analysis is concerned only with

the Auger electrons. In addition, for de-excitation energies less than approximately

2000 eV, Auger electron emission dominates over X-ray emission.

Auger electrons get their energy from the atom when it relaxes back to its ground

state. Therefore, they are characteristic of the energy levels of that specific atom. Once

Auger electrons are ejected from the atom, they must make it through the material and be

ejected without energy loss. Figure 3-9 shows the mean free path of various atomic

species. Immediately obvious from this graph is the fact that none of the elements listed

have an Auger electron escape depth over 30 A. This is the reason that AES

characterizes the near surface region. With the use of computers, data can now be

collected directly in N(E) versus E mode and then manipulated to obtain the differential








dN(E)/dE versus E. Historically, the data was collected in differential form due to the

use of an ac modulation on the signal and detection with lock-in amplifiers, as shown in

Figure 3-7, but with modem detectors and computers, this is no longer necessary. Saving

the data in its non-differentiated form presents advantages in noise and processing.36

However, the differential form is best for viewing the data since it highlights the Auger

peaks.22, 36, 126

For the current work, a Perkin-Elmer PHI660 Scanning Auger Multiprobe system

was utilized. Electron beam energy was kept at 5.0 keV with a 25 uLA current. Pressure

in the system was around 3*10-8 Torr during sputter profiling. A 3x3 raster was used on

the 3 keV Ar ion gun during sputter analysis. Initial surface scans were collected from

coated and uncoated parts of the sample to find peaks for analysis during depth profiling.

The surface scans of the coated materials can give information about the uniformity of

the coating within a thickness range equal to small multiples of the escape depth of the

Auger electrons. Depth profiling was used to determine thickness of the coating on

powders as compared to complementary coatings on Si masks.



Scanning Cathodoluminescence (CL) in a Scanning Electron Microscope (SEM)

Uniformity is also thought to be important to the success of the coatings. A SEM

with a fiber optic attached to a spectrometer was used for producing CL maps of the

sample surface. The SEM is a useful tool for magnifying the sample surface (about O10X-

300,OOOX).36 To produce an image, a focused electron beam is rastered across the

sample surface. When the electrons enter the sample, they lose energy through inelastic

collisions, as discussed above. This inelastic energy loss is transferred to the host lattice






78

and gives rise to a multitude of different electron energies that leave the surface.

Figure 3-1 shows a plot of these different electrons. In addition, other signals can be

produced, such as X-rays, light, and heat. Due to the inherent roughness of a sedimented

powder screen, there are intensity variations in the CL map of the surface. In order to

provide a complementary image of the surface being studied, a secondary electron (SE)

image was collected. Uniformity of the coating could then be checked by comparison of

the two images. The SE image provided a picture of the different phosphor surfaces.

The corresponding surfaces on the CL image could then be located and studied.





















ELECTRON
YIELD


SECONDARY ELASTIC BACK
ELECTRONS SCATTERED
ELECTRONS
AUGER
\ ELE CTRNS


5 50
ELECTRON ENERGY (eV)


2000
2000


Energy distribution of electrons emitted from a sample surface under electron
bombardment. The secondary electrons are used in secondary electron (SE)
imaging in a scanning electron microscope (SEM). The Auger electrons are
used in Auger electron spectroscopy (AES). The elastic energy is used for
diffraction analysis in the transmission electron microscope (TEM).126


Figure 3-1






























: t I. I I I 1 4 1 I I I I I I I I i I
S- - --- -- - -f



| ^ ^ ~ ~Z~-I.ZZ~LZ 'Z.Z'r rF' l'i I il I I /I;S/ i -,'~", "
-- ... i Ii 1 i i I fmw i C i I
"5' --'" f


I w I L 1 I1 I W1"
...2:I l 0 l, i
10l l! lIl l


400 420 440 460 480 500 520 640 S0 W80 00 020 640 600 W
WAVE LENGTH MILUMICRONS


Figure 3-2 Transmission of light through a Coming glass 2-73 filter. Taken from data
sheet supplied with filter.


70T 720 740


- 240 260 "D 300 SM 040 SOO no





























101 -- -- -- T- -- ---.-- .- -- -- -- ,- -_ _ _ _ _ _ _
o -: - - - -4-- -4. - ,_ ~~^ ~_ ^ ^ ^ ^^._
-c----------z-_--.:-I- ---I- --I-- I-__



I I 4 \ I "I
TO---------------- -2- 5 S _ _ _ _
40 i <; t \ t
S\ -\ -
20!= = =/ = =s == E E
so /E... ,
1. 11"=^=\^^ == \5=====....=I==
+o / I#1 \ \ 5c=^= = I,== /
I-/L/ x \ ;\l // == ::::


2, 2w 3 3 20 U40 a 30 000 400 420 440 <60 490 500 520 540 560 80 600 620 40
WAVE LENGTH MILLIMICRONS


60 6O 700 720 740


Figure 3-3 Transmission of light through a Coming glass 5-58 filter. Taken from data
sheet supplied with filter.


*w wO










































Figure 3-4 Transmission of light through a Coming glass 3-70 filter. Taken from data
sheet supplied with filter.


















SUBSTRATE


LYZER

TELESCOPE


,. LIGHT SOURCE


DETECTOR V
(EYE OR MICROPHOTOMETER)


Figure 3-5 Schematic of a typical ellipsometer. Monochromatic light is passed through a
polarizer and a quarter-wave compensator to give elliptically polarized light.
This is reflected off the sample into the analyzer (a second polarizer). The
polarizer and analyzer are rotated to determine maximum extinction of the
reflected light. 126





















Electron gun -
Anode ---- L ^
Gun alignment coils
Gun airlock --f

1 st Condenser lens -
2nd Condenser lena -s
Beam tilt 0oils sl
Condenser 2 aperture --- ION
Objective lens GETTER
Specimen block PUMP

Diffraction aperture
Diffraction lens -
Intermediate lens __ |

1 ast Projector lens --
2nd Projector lens -- -


Column vacuum block
35 mm Roll film camera --
Focussing screen
Plate camera /
16cm Main screen I








Figure 3-6 A typical transmission electron microscope. The top portion of the
instrument above the specimen block is used to generate a focused high-
energy electron beam. The apertures and lenses below the specimen block
are used to select specific regions of the diffracted electron beam for imaging.
Most instruments allow for collection of generated images on a photographic
film.36


















































Figure 3-7 A typical Auger electron spectroscopy (AES) setup, based on a cylindrical
mirror electron energy analyzer (CMA).22











C harvwcstc X-ray
O -
huE .EK-E

0


C Ev
0
@amC

c

C
I0..


e~

UI !L E Edwao


EwE, ELI


n(E)


Figure 3-8 Schematic representation of the processes of X-ray fluorescence and Auger
electron production. A KLL transition is shown as an example. Initial
excitation comes from ejection of a core shell electron by an incident
electron. Both processes occur in a material under electron beam
bombardment.22


Auge ElKctro
Emisio
E -E -E-























Au


I I
t0 20


-I I I


I I -- I I I I I
s0 100 200 300 500 1000 2000
Electron Energy, eV


Electron escape depth as a function of initial kinetic energy. This is a
measure of the average distance the electron will travel before undergoing an
inelastic collision with the host lattice. It is also commonly referred to as the
electron's inelastic mean free path.22


0.



0


w
LU


Figure 3-9















CHAPTER 4
EXPERIMENTAL RESULTS


Introduction

Results are separated into four main categories. These include the steady-state

cathodoluminescence (CL) spectra, which show the luminescence over the visible

wavelength range and at multiple beam energies. Next, the effects of beam energy on

steady-state CL intensities at specific wavelengths are reported. Third, pulsed CL decay

curves at various beam energies and wavelengths are reported. Fourth, the coating

thickness and uniformity data are presented. A list of all samples is given in Table 4.



Cathodoluminescent Spectra

Figure 4-1 shows the cathodoluminescent spectra of Y203:Eu as reported in the

Phosphor Technology Center of Excellence: Low Voltage Phosphor Data Sheets.38 This

figure shows the relative radiant intensity (i.e., cathodoluminescence intensity) as a

function of wavelength at 1 keV and 1 gA/cm2. The intensity has been normalized to the

value of the 611 nm peak. Figures 4-2 through 4-5 depict the cathodoluminescent spectra

for Y203:Eu with coatings of MgO deposited for 4 or 8 minutes, and A1203 deposited for

1.2 or 2.4 minutes, respectively. These curves show the intensity as a function of

wavelength from 520 to 720 nm. Spectra are displayed for uncoated and coated samples









at beam energies of 0.8 keV, 1.4 keV, 2.5 keV, and 4 keV. Current density was kept

constant at 0.16 IA/cm2 for each measurement. From the data, it can be seen that the

coating reduces the cathodoluminescent intensity over the entire wavelength range.

However, no measurable change occurred in the wavelength dependence of the

luminescent peaks, as will be discussed later in this section. Another feature of these

graphs is that coatings of the same material deposited for longer times (which were

therefore thicker) led to a larger attenuation of the intensity at all beam energies from 0.8

to 4 keV. This can be seen by comparing the intensity ratio of coated to uncoated

phosphors at any given beam energy for the same coating material and time.

Figure 4-6 shows a typical cathodoluminescent spectra for Y2SiOs:Tb from 400 to

720 nm measured at 1 keV, as reported in the literature.49 The intensity has been

normalized to the main emission peak. Figures 4-7 through 4-10 are spectra of uncoated

and coated Y2SiOs5:Tb. Coatings of MgO, deposited for 4 and 8 minutes, and A1203,

deposited for 2.4 and 5 minutes, are presented. No spectra are reported for the 1.2 minute

A1203 coating on Y2SiO5:Tb because the data files were corrupted. The intensity as a

function of wavelength from 460 nm to 680 nm at beam energies of 0.8 keV, 1.4 keV,

2.5 keV, and 4.0 keV are shown. Similar to the Y203:Eu samples, the MgO and A1203

coatings caused a reduction in the cathodoluminescent intensity over the entire

wavelength range. Also, coatings of the same material deposited for longer times (which

were therefore thicker) attenuated the luminescence more strongly. Wavelength

dependence of the luminescent peaks was unaffected by the coatings.

No calibrated light source was available inside the chamber, so luminescence in

Figures 4-2 to 4-5 and 4-7 to 4-10 was plotted in arbitrary units. During measurement,








the integration time of the spectrometer was adjusted to provide a main peak intensity

signal that was between 80 and 90 percent of the maximum allowable intensity. This

helps increase the signal-to-noise ratio. To further improve this ratio, 64 scans were

collected and averaged. Sixty-four background spectra scans were also collected and

averaged before each run and were subtracted from the luminescent spectra in real time.

Averaging of the background spectra is needed in order to keep from introducing noise

during subtraction.

All spectra in a given figure have been corrected for any differences in integration

time and can therefore be directly compared against one another. However, care must be

taken when comparing curves from different figures because not all figures have been

normalized to the same overall integration time. This was necessitated by small changes

in sample positioning and light collection between experiments. In order to compare

curves from different figures, the ratio of the coated to uncoated spectra needs to be used.

Since all powders of each type of phosphor are the same, the luminescence of the

uncoated side should be the same for all samples of that phosphor. This is one of the

main reasons for masking half of the phosphor screen during deposition. Therefore, by

comparing the ratio of intensities at different beam energies, the effects of the coating can

be obtained. Using this method, it can be concluded from the figures that coatings of a

given material deposited for longer times show larger attenuation of the

cathodoluminescence intensity, as stated in the previous paragraph.

In order to better understand the beam energy and coating effects on the CL peak

positions and relative heights, spectra for beam energies from 0.8 to 4 keV were

normalized to the intensity at the main emission peak. These are shown in Figures 4-11




Full Text
xml version 1.0 standalone yes
Volume_Errors
Errors
PageID P2
ErrorID 7