AN ABSTRACT OF THE THESIS OF
Dong Li for the degree of Doctor of Philosophy in
Chemistry presented on April 6, 1999.
Title:
Phos hor Develo ment: S nthesis Characterization and Chromatic Control
Redacted for privacy
Abstract approved:
/ Douglas A. Keszler
Selected phosphors of importance for continued development of flat-panel
electroluminescent displays (ELDs) and field-emission displays (FEDs) have been
synthesized and studied.
By using ceramic-processing methods, solid-solution
techniques, and codopants, the luminous efficiencies and chromaticities of several
phosphors have been greatly improved. On the basis of these results, a trichromatic
phosphor system has been proposed for upgrading ELD performance from monochrome
to full-color capabilities.
The maximum emission wavelength of Y2Si05:Ce3+ has been shifted by adding Sc to
form the solid solution Y2..ScSiOs (0
x 5_ 0.75.) Structures in the solid-solution series
were refined by the Rietveld method. The shift of the emission band is associated with
the distribution of Sc on the two crystallographic types of Y sites and the nephelauxetic
effect.
Eu2+-doped Mgi.CaxS luminescent powders were prepared by flowing H2S(g) over
nanoparticulate Mgi_xCax0 that was prepared by the combustion method. The resulting
materials were characterized by X-ray diffraction, scanning electron microscopy, and
luminescence spectroscopy. The phosphor powder is characterized by faceted grains
with a median particle size near 2p.m and a high cathodoluminescent efficiency: 8.07
lm/w at 2000V.
GdNb04:Bi and YNbO4:Bi phosphors have been prepared by the sol-gel method.
Compared to Y2SiO5:Ce, GdNbO4:Bi has a similar luminous efficiency, while YNbO4: Bi
exhibits a higher efficiency. The relatively high efficiencies of these niobates are partly
associated with their long wavelength emission tails.
Photoluminescence emission and excitation spectra, decay curves, and thermal
quenching have been measured for SrS:Cu phosphors coactivated with different 1+ and
3+ ions. The shapes and positions of peaks in the emission spectra depend solely on the
local surroundings of the Cu atoms.
The luminescence mechanism is explained by
considering the nature of the isolated Cu coordination environments that result from the
method of charge compensation.
(Zn, Ga)S:Mn solid-solution phosphors were synthesized by using high-temperature
solid-state reactions.
chromatic
control
The emission colors range from yellow to deep red. Explicit
can be
achieved by
adjusting the
concentration
of Ga.
Photoluminescence emission and excitation spectra are reported for measurements taken
at both liquid nitrogen and room temperature.
relaxation on isolated Mn2+ centers.
The emission has been assigned to
Phosphor Development:
Synthesis, Characterization, and Chromatic Control
by
Dong Li
A THESIS
submitted to
Oregon Sate University
In partial fulfillment of
The requirements for the
Degree of
Doctor of Philosophy
Completed April 6, 1999
Commencement June 1999
Doctor of Philosophy thesis of Dong Li presented on April 6, 1999
APPROVED:
Redacted for privacy
Major rofessor, representing Chemistry
Redacted for privacy
er of Department of Chemistry
Redacted for privacy
Veanof4r
Gra
School
I understand that my thesis will become part of the permanent collection of Oregon State
University libraries.
My signature below authorizes release of my thesis to any reader
upon request.
Redacted for privacy
ACKNOWLEDGMENTS
I would like to take this opportunity to express my sincere thanks to my major
professor, Dr. Douglas A. Keszler, for his guidance, assistance, and encouragement
throughout my entire program. Without his expertise and direction, my Ph.D. thesis may
have never been completed.
I would also like to acknowledge my gratitude to my committee members, Prof.
Arthur Sleight, Prof. John Loeser, Prof. Michael Lerner, and Prof. Joe Karchesy for their
ongoing support of my doctoral program.
Next, I would like to extend thanks and appreciation to Prof J. F. Wager, Paul
Kier of ECE department for teaching me EL principle and SrS:Cu thin film deposition, to
Prof. J. Tate and V. Dimitrova of physics department for (ZnGa)S:Mn thin film
deposition, to Prof. W. Warren and J. Janesky of physics department for performing solid
state NMR, to Dr. D. Tuenge, Dr. S. Sun, and S. Moehnke of Planar for their help and
advice.
I am especially grateful to my friends and colleagues, all of whom have
contributed to my success and made my years at Oregon State University a fun time to
remember: Dr. Jun-Ming Tu, Dr. Anthony Diaz, Dr. Ki-Seog Chang, M.S. Ken
Vandenberghe, M.S. Stephen Crossno, Dr. Judith Kissick, Dr. Cho, Greg Peterson, Ben
Clark, Sangmoon Park, Jennifer Stone, and Melissa Harrington for their helpful
discussions and fun in the research group.
Special thanks go to Ju-Zhou Tao in Prof Sleight's group and Engelene
Chrysostom in Prof Nibler's group for their help on my research.
I also wish to thank all the faculty and staff of Department of Chemistry for their
continued assistance.
I would further like to acknowledge my gratitude to the faculty and classmates in
Department of Chemistry, Beijing University and my motherland China.
Finally, I would like to acknowledge my deep debt of gratitude to my parents, my
wife, other family members and friends for all the support and encouragement they have
given me throughout my studies.
TABLE OF CONTENTS
CHAPTER 1: INTRODUCTION
1
Luminescence Process
3
Phosphor Characteristics in Display Function
4
FED Emission Display
9
AC Thin-Film Electroluminescence (ACTFEL)
12
This Work
16
References
18
CHAPTER 2: LUMINESCENCE OF Y2_xScxSi05:Ce3+
19
Abstract
20
Introduction
20
Experimental
20
Results
22
Crystal Structure
Rietveld Refinement
Photoluminescence
22
26
29
Discussion
32
Conclusions
35
References
36
CHAPTER 3: Mgi_xCaxS:Eu (05.0.2) LUMINESCENCE MATERIALS FROM
COMBUSTION PRECURSORS
37
Abstract
38
Introduction
38
Experimental
39
Results and Discussion
40
Characterization of Oxide
Characterization of Sulfide
Photoluminescence
41
43
50
Summary
54
Acknowledgements
54
References
54
TABLE OF CONTENTS (Continued)
CHAPTER 4: NIOBATE PHOSPHORS PREPARED BY SOL-GEL METHOD
56
Abstract
57
Introduction
57
Experimental
58
Results and Discussion
60
Conclusion
67
References
68
CHAPTER 5: LUMINESCENCE OF SrS: Cu
69
Abstract
70
Introduction
70
Experimental
72
Results
72
Excitation
Photoluminescence decay time
Thermal Quenching
Concentration Quenching
Discussion
Excitation
Emission and Quenching
Decay Time (tentative explanation)
Concentration Quenching
77
81
84
87
88
90
91
93
93
Conclusion
94
References
95
CHAPTER 6: PHOTOLUMINESCENCE OF (Zn,Ga)S: Mn
97
Abstract
98
Introduction
98
Experimental
99
Results and Discussion
Lattice Constants
Photoluminescence
100
100
103
TABLE OF CONTENTS (Continued)
Analysis of excitation band
112
Summary
115
References
116
CHAPTER 7: CONCLUSION
118
BIBLIOGRAPHY
119
LIST OF FIGURES
Page
Figure
1.1.
Principal types of electronic information displays
2
1.2.
Configurational coordinate diagram illustrating luminescence process
4
1.3.
Luminosity curves for C.I.E. standard observer
5
1.4.
Color matching functions for CIE 1931 standard observer
6
1.5.
Electron emission process
10
1.6.
Field emission display structure
11
1.7.
Types of field emitters.
12
1.8.
Standard ACTFEL structure
13
1.9.
ACTFEL energy-band diagram illustrating fundamental physics
14
1.10.
Energy band bending (a) without space charge, and (b) with space charge
15
2.1.
Drawing of the structure of Y2SiO5. The red balls represent 05 atoms, the
green and black balls represent Y1 and Y2 atoms respectively. SiO4 are
designated by blue tetrahedra
23
2.2.
Labeled drawing of Yl- and Y2 centered polyhedra in Y2SiO5
24
2.3
Observed (+ line) and calculated (solid line) powder diffraction pattern of
Yi.25Sc0.75Si05:2%Ce. Difference curve appears at the bottom of the figure
26
2.4.
Cell Volume/Z (A3) of Y2_xScSi05
28
2.5.
Emission spectra of Y2SiO5:2 %Ce at 4 K and 298 K. -kexc = 300 nm
29
2.6.
Thermal quenching of Y2SiO5:2 %Ce
30
2.7.
Emission spectra of Y2,ScxSi05:2%Ce at room temperature. Xexc = 300 nm.
30
2.8.
Comparison of thermal quenching of Y2SiO5:2 %Ce to Y1.75Sco.25Si05:2%Ce.
33
Xexc 300nm, monitoring the emission peak, respectively
2.9.
Excitation spectra of Y2,ScxSi05:2%Ce3+. Xem=450nm
34
LIST OF FIGURES (Continued)
Page
Figure
2.10.
Excitation spectra of Y2_,,Sc,,Si05:2%Ce3+. Y2: Y2Si05, Y1.25:)(1.25SC0.75Si05,
35
20: Xem=520nm, 450: kem=450nm
3.1.
X-ray powder pattern of Mg0.9Cao.10 prepared by combustion synthesis.
The line is diffraction pattern of MgO...
3.2.
Plot of the Fourier size coefficient against (a). Crystallite size distribution and
(b). Median crystallite size of Mg0.9Ca43.10
3.3
42
43
SEM micrographs of Mg0.96Ca0.04S: 0.1%Eu2+ prepared by sulfate method
(upper) and combustion method (lower)
44
3.4.
Particle size distribution of Mg0.9Cao.IS determined by light scattering
45
3.5.
X-ray powder diffraction pattern of Mgo.8Cao.2S:0.1%Eu
46
3.6.
Lattice constants of Mgi.CaS
47
3.7,
Representative Williamson-Hall plots for Mgo.9Ca0.iS: 0.1%Eu2+
49
3.8.
Emission Spectra of Mgi.xCaxS :0.1%Eu. kexe=315nm
50
3.9.
Excitation spectra of Mgi.,,C aS : 0.1%Eu
51
3.10.
omparison of luminous efficiency of Y203:4%Eu and Mg0.9Cao.IS:0.1%Eu
prepared by combustion and sulfate/H2S method
53
4.1.
Emission of YNbO4 under 254nm and 304nm excitation
60
4.2.
Corrected emission and excitation spectra of YNbO4 and YNbO4:Bi
61
4.3.
Emission of GdNbO4 under 254nm and 307nm excitation
62
4.4.
Eexcitation and emission of GdNBo4 and GdNbO4:Bi
62
4.5.
Relative brightness as a function of the bismuth concentration..
63
4.6.
Dependence of luminous efficiency on voltage
65
4.7.
Schematic representation of the energy level scheme of a free s2 ion
66
5.1.
Emission spectra of SrS doped samples, kexc=3 lOnm
73
LIST OF FIGURES (Continued)
Page
Figure
5.2.
Effect of different excitation at 77K
74
5.3.
Effect of temperature on emission of SrS:0.2%Cu, X_exc:=_310nm
75
5.4.
Effect of temperature on emission of SrS:0.2%Cu,Y, X
5.5.
Effect of temperature on emission of SrS:Cu(Ga2S3) film (N8-018),
Xexc310nm
76
5.6.
Emission of samples with x%Cu, 1%Na, Xexe=310nm
77
5.7.
Effect of temperature on excitation of SrS:0.2%Cu, Xem=525nm at both 298K
and 230K, Xem=520nm at 180K, and Xem=515nm at 78K
78
Effect of temperature on excitation of SrS:0.2%Cu, 0.2%Y; Xem500nm at
77K; 495nm at 130K; 485nm at 180K; 475nm at 230K; and 470nm at 298K.
79
Effect of codopants on excitation of SrS:0.2% Cu. Xem=525nm for 0.2%Cu
sample. Xem475nm for 0.2%Cu,0.2%Y sample. Xem=530nm for 0.2%Cu,
0.2%Li sample
80
5.8.
5.9.
310nm
310 nm
75
5.10.
Effect of La3+ on the emission of SrS:Cu,
5.11.
Decay curve for SrS:0.2%Cu, band emission
82
5.12.
Decay curve for SrS:0.2%Cu, 0.2%Y, band emission
82
5.13.
Decay curves for SrS:0.2%Cu and with codopants, Xem =550nm
83
5.14.
Decay curves for SrS:0.2 %Cu with codopants Xem = 450nm
83
5.15.
Quenching of SrS:0.2%Cu and SrS:0.2%Cu, 0.2%Y by monitoring
specific wavelengths noted in legend, Xexe = 310 nm
85
Quenching of SrS:0.2%Cu and SrS:0.2%CuY by monitoring
integrated emission signal with excitation wavelength noted
in legend (BEX = band excitation.)
86
5.17.
Comparison of thermal quenching of thin film and powder phosphors..
86
5.18.
Conxentration quenching of SrS:x%Cu,x%La and SrS:x%Cu,1%Na.
87
5.16.
X
81
LIST OF FIGURES (Continued)
Page
Figure
5.19.
Cu+ energy levels in alkali halide host NaF
88
5.20.
Schematic configuration coordinate diagram for SrS:Cu with
both green and blue emission centers
92
5.21.
Photoluminescence of SrS:0.2%Cu,x%Na,y%Y
94
6.1.
Structure of ZnS and ZnGa2S4
101
6.2.
Cell volume of Zni-3,e2GaS
103
6.3a.
Emission of Zno-3x/2)GaxS: 1%Mn (Room Temperature). ?t,exe=335nm
104
6.3b.
Emission of Zno-3x/2)GaxS: 1%Mn (78 K). kexc335nm
105
6.4.
C. I. E. diagram of Zn(1-3,e2)GaxS:1%Mn phosphors
106
6.5.
Thermal quenching curve of (ZnGa)S:Mn by monitoring the emission peak..
107
6.6.
Band excitation spectra of Zno-3,a2)GaxS: 1%Mn at 300 K by monitoring
the emission peak, respectively
108
Excitation spectra of Zno-3,d2)GaxS: 1%Mn at 78 K by monitoring
the emission peak, respectively
108
6.8.
Excitation Spectra include both band gap and Mn excitation
109
6.9.
Spectra of pure and Mn doped Zno.7Gao.2S
110
6.10.
Effect of band gap excitation (335nm) and Mn intraionic excitation (430nm)
at 78 K
111
6.11.
Decay time of Zno-3x/2)GaxS:1%Mn2+ at 300 K
112
6.12.
Energy level diagram of Mn2+ in ZnS with B=0.0741eV, C=0.410eV,
and 6=0.043. The observed band energies (peak) are marked 0
113
6.7.
LIST OF TABLES
Page
Table
1.1.
Addressing methods
2
1.2.
CIE coordinates for phosphors in tri-chromatic system
7
1.3.
Summary of radiometric and photometric quantities
9
1.4.
TFEL phosphor performance summary
16
2.1.
Crystallographic data and atomic parameters for Y2SiO5 (signal crystal
resultes)
23
2.2.
Selected inter-atomic distances (A) and Angles (°) for Y2SiO5
25
2.3.
Selected crystal data and interatomic distances for Y2,ScxSi05
27
2.4.
Chromaticity, Excitation, and Emission Wavelength (nm) of Y2.Sc.Si05.....
31
2.5.
Luminous efficiency (lm/w)
31
3.1.
Characterization of Mg i.CaS
46
3.2.
Spectroscopy parameter of Mgi,CaxS
52
3.3.
Luminous efficiency (lm/w) at different voltage
53
4.1.
Luminous efficiencies of some phosphors
64
4.2.
Chromaticities of some phosphors
64
5.1.
Chromaticities of selected samples
76
5.2.
Decay times of SrS: 0.2%Cu, x%A(p,$)
84
6.1.
Lattice constants and lifetime of Zn1-3x/2GaxS: 1%Mn
102
6.2.
Emission peaks and chromaticity of Zni-3,d2GaxS: 1%Mn
104
Phosphor Development:
Synthesis, Characterization, and Chromatic Control
CHAPTER 1
INTRODUCTION
During the last ten years or so, dramatic changes have occurred in our everyday
lives as a result of the introduction and evolution of a vast array of electronic and
photonic technologies, and the information display has emerged as a key component in
the evolutionary process. In the past, displays were added to various systems almost as
an afterthought.
Today, systems are increasingly being designed around the display
device, especially the flat-panel configuration.
Flat Panel Display (FPD) technologies were proposed in the 1950s as
replacements for cathode ray tube (CRT) displays. Although this market objective has
not been fully achieved, several new families of products have been created, including
notebook computer displays, hang-on-the-wall I-IDTVs, and virtual reality systems.
FPDs are categorized on the basis of the physical mechanisms of light production
and pixel activation (addressing methods.) Currently the FPD market is dominated by
liquid-crystal displays (LCDs.)
Each type of display has its own advantages and
disadvantages, allowing utilization of a selected type in a specific market. The principal
electronic information displays and addressing methods are summarized in Figure 1.1 [1]
and Table 1.1 [2], respectively.
2
Projection
Virtual
CRT, LCD
HMD
Direct View
Nonemissive
Emissive
Incandescent
LCD,
Luminescent
Electrochromi
c, etc.
U
CL, FED,
EL
PDP
VFD
High Field: ACTFEL, DCEL
AC matrix
Low Field: OLED, Inorganic
DC matrix
Figure 1.1. Principal types of electronic information displays.
Table 1.1. Addressing methods.
Cost in
Cost in
display
Electronics
structure
Scan
Gun(CRT)
Grid
Grid(VFD, etc)
Shift
Shift channel(PDP)
Matrix
Column and Row lines
Direct (active matrix)
Each pixel
Low
Low
High
High
There have been a number of developments in emissive displays.
The present
emphasis is on electroluminescent displays (ELDs) and field emission displays (FEDs)
for small-to-medium-sized screens and plasma display panels (PDPs) for large screens.
Phosphors provide the means to achieve a full-color gamut in each of these devices, and
3
they are an essential component of each system. While suitable phosphors have been
available for CRT applications for many decades, they have not been fully optimized for
operation under the specific excitation conditions of ELDs, FEDs, and PDPs.
In this
study, the focus is on development of phosphors for ELDs and FEDs. Before discussing
the specific attributes of these displays, it is essential to briefly review some general
concepts common to all types of phosphors.
Luminescence Process
The luminescence process is commonly considered in the context of the
configuration-coordinate diagram (Figure 1.2.) This type of diagram is often used to
describe transitions between electronic-energy levels coupled to lattice vibrations.
Although the model involves a one-dimensional displacement of a single vibrational
mode, which is a great oversimplification of the true situation, it is still quite useful in
explaining some aspects of the optical spectra of ions in solids.
In a configurational coordinate diagram (Fig. 1.2,) ground and excited electronic
states are represented as harmonic-potential wells with an ordinate of energy (E) and an
abscissa of distance (R), the displacement of the luminescent ion from its equilibrium
ground-state position Ro. The optical absorption generally brings the electron from the
lowest vibrational level of the ground state to a high vibrational level of the excited state.
The system returns first to the lowest vibrational level of the excited state, giving up the
excess energy to the surroundings. Because of the redistribution of electron density in
the excited state, the configurational coordinate changes by some value AR; this process
is called relaxation. The system can return to a high vibrational level of the ground state
by emission of radiation, and once again, relaxation occurs to the lowest vibrational level.
Because of these relaxation processes, the emission occurs at a lower energy than that of
the absorption. If the temperature is sufficiently high or the value of AR is large, the
electron can also return to the ground state through the crossing point of the two
parabolas.
At this crossover point, an electron in the excited state can enter a highly
4
excited vibrational level of the ground state and subsequently decay nonradiatively to the
ground vibrational level. More details can be found in the literature [3, 4, 5].
relaxation
Excite
State
Cross
Point
E
Ground
state
RO
-I.
AR 4--
Figure 1.2. Configurational coordinate diagram illustrating luminescence process.
Phosphor Characteristics in Display Function [6, 7, 8]
a. Luminosity Curve
A luminosity (or sensitivity) curve is a plot of the spectral sensitivity of the human
eye. Thus, it is critically important that the spectral output of a phosphor correlates well
with the eye's response. In the photopic range (vision at high levels of illumination,) the
brightest part of a spectrum with equal amounts of radiant flux per unit wavelength
interval occurs at about 555 nm.
In the scotopic range (vision at low levels of
illumination,) there is a shift of about 40 nm to shorter wavelengths (Figure 1.3.)
5
I
ti
0.8
Scotolt ;
P hot° pic
0.6
0.4
0.2
o
350
400
450
500
550
600
650
700
750
Wavelength (nm)
Figure 1.3. Luminosity curves for C.1E. standard observer.
b. Color [9]
A perceived color from a self-emitting object is typically described by its hue,
brightness, and saturation. Hue is the attribute that we denote by red, yellow, and green;
thus hue is determined by the dominant wavelength. Brightness or luminance is the
attribute that we use to describe the perceived intensity of light. Saturation refers to the
degree a given color differs from white (or gray,) providing a description of the purity of
chromatic color.
The three attributes only describe the appearance of color in a qualitative manner.
If we would like to compare different colors in a quantitative way, we employ the CIE
system developed by the Commission Internationale de I'Eclairage (International
Commission on Illumination.) The CIE system is based on trichromatic theory. It deals
with three fundamental aspects
the object, the light source, and the observer
for a
given color experience. In simple terms, CIE tristimulus values (X, Y, Z) are used to
denote the quantities of primary colors in any arbitrary sample color. The tristimulus
6
value of a body with spectral output of radiance S(X) can be calculated by using the
color-matching functions x(X), y(X), z(X) shown in Figure 1.4.
380
430
480
530
580
630
680
780
730
Wavelength(nm)
Figure 1.4. Color matching functions for CIE 1931 standard observer.
Tristrimulus values X, Y, and Z are calculated according to:
X = J S(X.)x(X)(1X.
Y = J S(X)y(X)dX,
Z = J S(X)z(X)R.
Chromaticity coordinates x, y, and z are then derived by normalizing each of the
tristimulus values:
x = X/(X+Y+Z)
y = Y/(X+Y+Z)
z = ZI(X +Y +Z)
Because x + y + z = 1, we consider only the pair x, y, which is commonly plotted on a
CIE chromaticity diagram to give a quantitative rendition of the color.
In display
applications, the production of saturated white light is required. On the basis of a triphosphor system, the chromaticity requirements for each phosphor have been derived [8]
(Table 1.2.)
7
Table 1.2. CIE Coordinates for phosphors in tri-chromatic system
Red
Green
Blue
White
x
>0.55
<0.3
<0.15
=0.33
y
<0.30
>0.5
<0.15
=0.33
These colors must also be balanced in the brightness ratio 22-27% red, 65-68% green,
and 8-10% blue to match the luminosity curve.
c. Resolution and Sharpness
Resolution is the extent to which a picture medium can reproduce fine detail. For
a video system, resolution is expressed in terms of the maximum number of lines per inch
(ipi) and therefore is determined by pixel format. Sharpness of an image is characterized
in terms of the edge characteristics. The more clearly defined the line that separates dark
areas from light ones, the greater the sharpness of the picture. Resolution and sharpness
are somewhat interrelated and both are general characteristics of picture definition.
d. Contrast Ratio
The contrast ratio, Cr, specifies the observable difference between brightness of a
pixel that is switched on (L.) and its corresponding off state (Loff)
Cr = Lon/Loff
Because the body color of common phosphors is commonly white, room lighting can be a
major component of unexcited screen brightness. By applying neutral density filters,
pigments, and black matrices in the construction of a screen, contrast can be greatly
improved.
e. Gray Scale
To approach photographic film quality, a display requires more than 256 levels of
brightness.
8
f. Efficiency
There are several definitions of efficiency for luminescence.
In the case of
photoluminescence, an important term is quantum efficiency (q), which is defined as the
ratio of the number of emitted quanta to the number of absorbed quanta.
For
cathodoluminescence and electroluminescence, energy efficiency (i) and luminous
efficiency (L) are often used. Energy efficiency is the ratio of emitted and absorbed
power. The highest energy efficiency obtained to date in a phosphor is about 25% for
ZnS:Ag. For CRT phosphors, an empirical relationship has been formulated to allow
correlations between selected materials properties and energy efficiency:
rl = (1-r) * (hve) * S * q / (I3Eg)
r: fraction of primary beam energy that is not absorbed (due to back scattering).
hve: mean energy of emitted photon.
13Eg: beam energy required to produce an electron-hole pair (Eg is band gap).
S: efficiency of transfer of e-h pair energy to the luminescent center.
To account for the spectral sensitivity of the human eye (luminosity curve,) luminous
efficiency is used. Luminous efficiency (1m/W) is defined as the ratio of the luminous
flux (1m) emitted by the material and the absorbed power (W.) By definition, one lumen
is the amount of light (luminous flux) of monochromatic radiation with a frequency of
540x1012 Hz (555 nm in air) and a power of 1/683 watt. Another useful definition is the
luminous efficacy, which is the ratio between luminous flux emitted for each watt of
emitted radiant power. The luminous efficacy at other wavelengths is calculated with the
following equation
La, = Lin * VA, = 683 * VA, (lm/W)
where V?, is the luminosity curve (photopic). Related definitions are given in Table 1.3.
9
Table 1.3. Summary of Radiometric and Photometric quantities.
Photometric Quantities
Radiometric Quantities
Radiant flux 4)e, watt(w)
x
Luminous flux 4)v, lumen(lm)
Radiant intensity Ie,watt/steradian luminous
Luminous intensity L
(w/sr), (point source, independent efficiency
Lumen/steradian or candela
of direction)
(1m/sr or cd)
Radiance Le, w/steradianmeter2
Luminance Lv, candela/meter2 or
(w/sr.m2)
for extend source by
Lumen/steradian.meter2
perceived area
(1m/srm2 or cd/m2)
Irradiance Ee
Illuminance Ev
watt/meter2 (w/m2)
lumen/meter2 (lm/m2)
Field Emission Display [10, 11, 12]
A number of attempts have been made to construct a flat CRT that can overcome
the dimensional and mass limitations of the traditional CRT. In principle, the Field
Emission Display (FED) can fulfill these objectives.
The FED operates in manner similar to that of a CRT, i.e., by electron excitation
of a phosphor. Whereas a CRT might be operated at 10-20 kV, a FED is projected to
operate at voltages lower than 5 kV. Contrary to the thermionic emission (thermal
excitation of electrons) utilized in a conventional CRT, the FED cathode emits electrons
by field emission. Because this process involves no thermal excitation, the electron
source is referred to as a cold cathode. Application of an electric field (5x107 V/cm) at
a metal-vacuum interface of the cathode leads to band bending and electron emission
through tunneling (Figure 1.5.)
There are two basic characteristics of FED technology. The heart of the new
technology is the Field Emission Array (FEA,) which provides a uniform large-area
source of electrons. FEAs also can be addressed at low voltages, making the technology
quite practical. The most widely used FEA is composed of a Spindt tip. In this type
10
Vacuum level
EF
Partly filled
conduction band
4
Metal
Vacuum
Figure 1.5. Electron emission process.
of FEA, electron emission is achieved by using a sharp tip (nanometer scale) to enhance
the field. The emission process is described by the Fowler-Nordheim equation:
J = (RV)2 /
x exp ( 4)1 5
(RV))
J: current density
V: voltage between gate and cathode
13: field enhancement factor (accounts for electrode spacing and tip sharpness)
4): work function
To reduce addressing voltage (V), the work function, radius of emitter tip, and
electrode spacing must be minimized.
The work function, however, must be large
enough to limit thermionic emission during normal operation.
While the Spindt tip
exhibits fundamental advantages, research continues on edge emitters and planar emitters
because of their low fabrication costs. Because of their small work function (-1eV),
diamond films have also been promoted as an effective source of electrons for FEDs.
11
To avoid arcing and the formation of plasmas, the FED is fabricated as a so-called
vacuum device.
Spacers must be placed between the faceplate and the baseplate to
support the structure (Figure 1.6.) These spacers must be mechanically strong, invisible
to the viewer, and resistant to damage by high electric fields. Use of these spacers,
however, presents special problems in the manufacturing process, as a high vacuum must
be maintained throughout.
Black Matrix
Transparent Conductor
Extraction
Grid
Emitter
Baseplate
Emitter Electrodes
Figure 1.6. Field emission display structure
Because of the thin profile of the FED, the operating voltage is maintained below
10 kV, and voltages much lower than 5 kV are desired. These low voltages place unique
restrictions on the properties of phosphors. Because electron-penetration depths decrease
to about 20 nm at these voltages, efficiencies are much lower than those achieved at high
operating voltages. To improve efficiencies, high current densities (-100p.AJcm2) are
employed.
Under these conditions, however, phosphor saturation and degradation
become major problems.
One alternative structure for the field emitter is the tetrode design (Figure 1.7.)
With this structure providing a collimated source of electrons, the vacuum gap can be
increased, allowing voltages higher than 5 kV. At sufficiently high voltages, standard
aluminized CRT phosphors can be used.
12
Diode
Triode
Tetrnde
IN=
Figure 1.7.
1.7. Types of field emitters.
AC Thin-Film Electroluminescence (ACTFEL) [15, 16, 17]
Electroluminescence (EL) occurs in two forms - injection and high field. In
injection EL, light energy is released upon recombination of minority and majority
carriers across the band gap in a single crystal solid. This process is the basis of light
emitting diodes.
In high-field EL, the emission arises from impact excitation of
luminescent centers with charge carriers that have been accelerated by high-electric fields
within the solid.
EL was first reported by Destriau in 1936, when he used oil as a dielectric and
water as the transparent electrode to study the luminescence of ZnS:Cu powder. During
the 1950s, following the development of transparent conductive thin films based on
Sn02, many studies were performed on AC EL devices with powder phosphors. These
devices were characterized by the absence of a clearly defined threshold, low brightness,
and limited operational lifetimes in the range of 2000 to 5000 hours. These problems
were largely overcome in the 1970s with the development of AC thin film EL (ACTFEL)
devices operating with ZnS:Mn. One of the more attractive features of TFEL technology
lies in the fact that the device has a very steep luminance vs. voltage slope above the
threshold voltage. And mean-times-to-failure of greater than 140,000 hours have been
demonstrated for production displays.
The basic ACTFEL device structure is illustrated in Figure 1.8. The phosphor
layer is encapsulated by insulators, and charge is generated at the interface of the
13
phosphor and insulator layers.
In a conventional device, the substrate is glass.
Conductor 1 and insulator 1 are transparent, allowing transmission and viewing of light
through the substrate. An inverted structure has been developed for use in active-matrix
EL (AMEL) and "color by white" devices. In this configuration, the display is viewed
from the top of the stack, and insulator 2 and conductor 2 are transparent. The insulator
layers shield the phosphor layer from extreme currents, so they must exhibit high
dielectric strength and "self-healing," which ensures pinholes and other imperfections do
not propagate. The sealing layer is also important, since the other layers can be sensitive
to moisture.
Seal Material
Conductor 2
Insulator 2
Phosphor
Insulator 1
Conductor 1
Glass or Si wafer substrate
Figure 1.8. Standard ACTFEL structure.
An energy-band diagram is employed in Figure 1.9 to summarize the physical
processes that are important for light emission in an ACTFEL device. A high-voltage
AC signal is applied to the conductors. This external voltage causes an electric field, Fext,
to develop across the device. Prior to breakdown, very few free carriers are present
within the phosphor layer. Carriers are available, however, in deep traps associated with
surface defects and impurities at the interfaces between the phosphor and the insulators.
Additional carriers are also available from deep traps associated with crystal defects and
14
impurities within the phosphor layer. The applied electric field results in energy-band
bending, which reduces the energy barrier between carriers in traps at the interface and
the conduction band.
When the electric field approaches
MV/cm, many carriers
tunnel from the phosphorinsulator interface states and accelerate in the high electric
field (process 1.) The threshold voltage for this process corresponds to the voltage for
initiation of electron emission in the absence of any internal field. Those accelerated
electrons with sufficient energy can impact excite a luminescent center (process 2,)
leading to the generation of light.
In process 3, non-radiative decay may also occur on
an excited activator center.
Fext
2.
Fpol
4.
anode
mn2+
3.
Figure 1.9. AC7FEL energy-band diagram illustrating fundamental physics.
In carrier multiplication (process 4), high-energy electrons impact atom sites and
impurities to create electron-hole pairs. Finally, the free carriers are trapped in the anode
15
at the phosphor-insulator interface (process 5.) This sets up an internal electric field, Fp01,
opposing the carrier driving electric field.
An alternating external applied voltage results in an alternating field, which is
opposite to that shown in Figure 1.9. This field is aided by the polarization field created
by the previous voltage transient, so conduction occurs at an externally applied voltage
lower than the threshold voltage. This voltage is called the turn-on voltage.
All carriers available for creating light are trapped between the two insulators.
The light emitted is directly related to the flux of electrons traversing the semiconducting
phosphor layer. Because most of the electrons originate from the interface states, the
excitation process depends strongly on the interface conditions, such as the nature of the
dielectric materials and the methods employed for manufacturing the device.
Space charge has very important effects on the operation of EL devices. As
shown in Figure 1.10, space charge in the phosphor layer will cause the energy bands to
curve and the electric field to vary. If space charge is assumed to be positive, the field
near the source (the cathode interface) is enhanced, while the field near the collector (the
anode interface) is reduced. The existence of space charge lowers the barrier for electron
tunneling, improves carrier injection, and allows the energetic electrons to cool before
hitting the interface, reducing insulator damage and electron leakage into the insulator
layer.
(a)
Figure 1.10. Energy band bending (a) without space charge, and (b) with space
charge.
16
The most widely used EL phosphors are listed in Table 1.4.
Table 1.4. TFEL Phosphor Performance Summary
Emission
Materials
CIE X
CIE Y
Color
Luminance
Efficiency
(Lao)
(Eat))
(cd/m2)@60HZ
(Luin/Watt)
Yellow
ZnS: Mn
0.50
0.50
400
3-6
Red
ZnS: Mn
0.65
0.35
70
0.8
CaS: Eu
0.68
0.31
12
0.2
CaSSe: Eu
0.66
0.33
25
0.25
MgZnS: Mn
0.63
0.36
5.6
ZnS: Mn
0.47
0.53
160
1.0
ZnS: Tb
0.30
0.60
100
0.6-1.3
MgZnS:Mn
0.48
0.51
6.3
SrS: Ce
0.28
0.53
110
0.8-1.6
CaGa2S4: Ce
0.15
0.19
10
0.03
SrGa2S4: Ce
0.15
0.10
5
0.02
SrS: Cu
0.17
0.16
28
0.2
SrS: Ce/ZnS:Mn
0.46
0.50
470
1.5
SrS:Cu/ZnS:Mn
0.45
0.43
240
(Filtered)
Green
(Filtered)
Blue-
green
Blue
White
This Work
The major purpose of this work is to study and improve the luminance and
chromaticity of FED and EL phosphors.
To achieve these goals, solid-solution
17
techniques and ceramic processing methods have been employed to synthesize new
compositions.
As mentioned previously, FEDs require cathodoluminescent phosphors operating
at or preferably below 5 kV. As noted, low excitation voltages and attendant high current
densities in these devices lead to small electron-penetration depths. This results in low
intrinsic luminous efficiency, saturation, and severe degradation. On the basis of current
models, a FED phosphor should have a low surface recombination rate, a deep
penetration depth, a long diffusion length, a fast decay time, high conductivity, and a high
activator concentration.
In Chapter 2, results are presented on controlling the
chromaticity of the blue-emitting phosphor Y2SiO5: Ce, which is currently the material of
choice for a blue FED phosphor.
In Chapter 3, the combustion synthesis and
cathodoluminescence of the red FED phosphors Mgi_xCaS: Eu are described.
YNbO4:
Bi has previously been examined as a potential blue FED phosphor. Results from the solgel synthesis of this material are discussed in Chapter 4.
Although EL displays have been a commercial success as monochromatic
devices, greater utility would be achieved with the development of a full-color device.
Such a display has not been realized because of the absence of an efficient red-green-blue
(RGB) phosphor set.
Recent developments in the production of efficient blue EL
phosphors in the SrS:Cu, Ag system has stimulated our interest in developing a true RGB
set. Results on the production of green and red luminescence in the systems SrS:Cu and
ZnGaS:Mn, respectively, are presented in Chapters 5 and 6.
18
References
1.
Frederic J. Kahn, SID seminar lecture notes, M-1/3 (1997).
2. Lawrence E. Tannas, jr. SID seminar notes, M-1/3 (1995).
3.
G. Blasse, B. C. Grabmaier, Luminescent Materials. Springer-Verlag 1994.
4. D. R. Vij, Luminescence of Solids, Plenum, 1998.
5.
B. Henderson and G. F. Imbusch, Optical Spectroscopy of Inorganic Solids, Oxford
Scientific Publications, 1989.
6. J. C. Whitaker, Electronic Display, Technology, Designs, and Applications. McGraw-
Hill, Inc. 1994.
7.
Aron Vecht, SID Seminar lecture notes, F-2/3 (1996).
8.
C. J. Summers, SID Seminar lecture notes, M-5/1 (1997).
9.
S. J. Williamson and H. Z. Cummins, Light and color in nature and art, John Wiley &
Sons Inc., 1983.
10. B. R. Chalamala, Y. Wei, B. E. Gnade, IEEE spectrum, April, pp. 42-51 (1998).
11. R. T. Smith, Information Display Vol 14, No.2, pp.12-16 (1998).
12. R. K. Ellis, SID Seminar lecture notes, M-8/3 (1997).
13. I. Brodie, C. A. Spindt, Advances in electronics and electron physics, Vol. 83, pp. 2108 (1992).
14. V. T. Binh, N. Garcia, S. T. Prucell, Advances in electronics and electron physics,
Vol 95, pp. 63-150 (1996).
15. Y. A. Ono, Electroluminescent display, World Scientific Publishing Co., 1995.
16. C. N. King, 1988 SID Seminar lecture notes, Vol. 1 (1988) 2A/1.
17. J. F. Wager and P. D. Keir, Ann. Rev. Mater. Sci., Vol. 27, pp. 223-248, (1997).
19
CHAPTER 2
LUMINESCENCE OF Y2_xScxSi05: Ce3+
Dong Li and Douglas A. Keszler*
Department of Chemistry and Center for Advanced Materials Research
Gilbert Hall 153, Oregon State University, Corvallis, OR 97331-4003
20
Abstract
The wavelength of maximum emission of the phosphor Y2Si05:Ce3+ has been red
shifted by substituting Sc for Y in the series Y2_xScxSi05 (0
x s 0.75). The crystal
structure of Y2SiO5 has been refined by using single crystal X-ray data, and the structures
of Y2,ScxSi05 (x = 0, 0.25, 0.5, 0.75) have been refined from powder X-ray data by
using the Rietveld method. On the basis of the structural results, the shift in the
luminescence band is associated with the preferred occupation of one of the available Y
sites by Sc and an attendant increase in the nephelauxetic effect, which leads to a lower
energy for the Ce3+ emission level.
Introduction
The luminescence of the phosphor Y2SiO5:Ce has been extensively studied,
because of its efficient luminescence under cathode ray [1], UV [2], and X-ray excitation
[3].
Due to its reasonable light output and high stability, the phosphor is also the
preferred blue component in full-color, flat-panel field-emission display [4].
The
maximum emission of the phosphor, however, occurs at 400 nm, a wavelength where the
sensitivity of the human eye is rather low. To improve the brightness of the material, it is
desirable to red shift the emission band. This goal may be achieved by strengthening the
Ce-O interactions through a decrease in the unit-cell volume. This desired effect has
been realized by producing materials in the solid-solution series Y2_xScxSi05 (0 < x
0.75).
In this contribution, we describe the results of structural measurements and
luminescence studies as they relate to the red shift in the emission of the phosphor.
Experimental
Powder samples of Y2,ScxSi05:2at% Ce3+ (x = 0, 0.25, 0.5, 0.75) were prepared
by mixing stoichiometric quantities of Y203 (AESAR, 99.99%), Sc203 (Bolder Scientific,
99.99%), SiO2 (CERAC 99.99%), and CeO2 (Aldrich, 99.9%) with 4wt% LiF (CERAC,
99.9%). The mixtures were fired at 1673 K for 20 h in a 4% Hz, 96% Ar atmosphere.
21
A small, Czchokralski-grown boule of Y2SiO5 was kindly provided by Professor
Bruce Chai; a crystal of approximate dimensions 0.2 x 0.1 x 0.1 mm was extracted from
the boule and mounted on a glass fiber with epoxy for structure determination.
All
measurements were made with graphite monochromated Mo Ka radiation on a Rigaku
AFC6R diffractometer. Unit-cell parameters and an orientation matrix for data collection
were obtained from a least-squares refinement with 21 reflections in the range 30
20 <-
36°. Laue symmetry 2/m was observed. Intensity data were collected over the range of
indices 1 5 h
20, -1
k < 9, -14
1
14 by using the co-20 scan technique to a
maximum of 20 = 60°. From 1715 measured reflections, 1216 had F.2
3cy(F.2). The
structure was solved and refined by using programs from the PC-version of the TEXSAN
crystallographic software package (Molecular Structure Corporation,
1997).
The
intensity statistics were indicative of a centrosymmetric space group. The systematic
absences h + k = 2n + 1 and 1 = 2n + 1 (for h01) are consistent with space group C2/c.
The Y atom was located by using the direct-methods program Sir92 [5] and the
remaining atoms were positioned following analysis of difference electron density maps.
All of the atoms occupy general positions. After full- matrix, least-squares refinement of
the model with isotropic displacement coefficients on each atom, an absorption correction
was applied by using the program DIFABS [6]. The equivalent data were then averaged
(Ri0t=0.063) and the final refinement resulted in the residuals R = 0.036 and 11,, = 0.036
for 919 observations. Relevant crystallographic data and atomic parameters are listed in
Table 1.
X-ray powder patterns were recorded for the compositions Y2,ScSiO5 (x = 0,
0.25, 0.5 and 0.75) with a Siemens D5000 X-ray diffractometer by using Cu Ka
radiation. Data were collected over the 20 range 10-90° by counting for 2 s at each step
of 0.02°. Structural refinements were made by using the program GSAS [7] and the
atomic parameters from the single-crystal structure determination as a trial model. Forty-
seven parameters, including scale factor, background, profile, asymmetry, zero point,
unit-cell constants, atomic positions, occupancy factors, and isotropic displacement
coefficients were refined.
22
Photoluminescence was measured at room temperature by using a procedure that
has been described elsewhere [8]. Cathodoluminescence was measured at the Georgia
Technology Research Institute.
Results
Crystal Structure
Two structure types are known for the materials Ln2Si05 (Ln = lanthanide.) For
large Ln (La-Tb), the monoclinic form, space group P2i/c, is found, while for small Ln,
(Dy-Lu, Sc), a different monoclinic form, space group C2/c, is adopted [9].
Y2Si05,
however, is dimorphic, crystallizing in the primitive cell at low temperatures and the Ccentered cell at high temperatures [10, 11]; the phase-transformation temperature is about
1250°C [12]. The high-temperature phase is used in phosphor applications.
Because the structure of Y2SiO5 has not been reported with high precision [13],
we have refined the atomic parameters from single-crystal X-ray data. The resulting
contents of a unit cell are depicted in Figure 2.1 and Table 2.1. Two crystallographically
distinct Y sites are present in the structure as possible substitutional sites for other +3
cations such as Ce3+ and Sc3±.
One site, Y1, is characterized as a 6-coordinate
polyhedron, while the other, Y2, is best described as a [6 + 1]-coordinate environment;
each has symmetry C1. As seen from the representative formula Y2Si040, the structure
also contains simple orthosilicate groups Slat and a unique 02- anion that is bound only
to the y3+ cations.
Each of the Y sites has two of the unique 02- anions in its
coordination sphere.
The coordination environments of the two Y sites are shown in
Figure 2.2.
Selected interatomic distances and angles are listed in Table 2.2. The
average Y1-0 distance, 2.27(3) A, is somewhat shorter than that of Y2 -O, 2.37(11) A.
23
Table 2.1. Crystallographic data and atomic parameters for Y2SiO5 (signal crystal
resultes).
a=14.407(1), b=6.727(1), c=10.417(1) A, P=122.19(1)°, space group C2/c (No. 15), Z=8
x
y
z
Beg*
Y1
0.85834(5)
0.87702(9)
0.16489(7)
0.57(1)
Y2
0.03705(5)
0.24483(9)
0.46536(6)
0.54(1)
Si
0.8183(1)
0.4100(2)
0.1926(2)
0.52(3)
01
0.7989(3)
0.5687(6)
0.0633(5)
0.84(7)
02
0.9092(3)
0.5008(7)
0.3571(4)
0.84(8)
03
0.7042(3)
0.3545(6)
0.1758(4)
0.80(8)
04
0.8801(3)
0.2141(6)
0.1794(5)
0.72(9)
05
0.9812(3)
0.0974(6)
0.6024(4)
0.62(8)
*Beg = (8/c2/3)X,XjUjiai*aj*aiaj
Figure 2.1: Drawing of the structure of Y2SiO5. The red balls represent 05 atoms, the
green and black balls represent VI and Y2 atoms respectively. Sias tetrahedra are
designated by blue tetrahedra.
24
Figure 2.2: Labeled drawing of Y 1- and Y2 centered polyhedra in Y 2SiO 5.
25
Table 2.2. Selected inter-atomic distances (A) and Angles () for Y2Si05.
Y1-01
2.228(4)
01-Y1-01
75.1(2)
Y1-01
2.276(3)
O1 -Y1 -03
94.9(2), 102.2(2)
Y1-03
2.279(6)
O1 -Y1 -04
160.1(1), 86.4(1)
Y1-04
2.294(4)
01-Y1-05
95.3(2), 118.4(2)
100.8(2), 166.4(1)
Y1-05
2.197(6)
03-Y1-04
96.3(2)
Y1-05
2.303(3)
03-Y1-05
156.5(1), 78.5(2)
04-Y1-05
80.7(2), 80.1(1)
Y2-02
2.328(4)
02-Y1-02
71.8(1)
Y2-02
2.334(4)
02-Y1-03
145.5(1), 74.0(1)
Y2-03
2.327(3)
02-Y1-04
61.4(1), 106.5(2)
132.9(1), 120.1(2)
Y2-04
2.610(4)
02-Y1-05
100.4(2), 129.6(1)
84.1(2), 151.8(1)
Y2-04
2.380(6)
03-Y1-04
151.9(2), 87.6(2)
Y2-05
2.207(5)
03-Y1-05
79.4(2), 83.9(1)
Y2-05
2.381(4)
04-Y1-04
71.9(2)
04-Y1-05
108.0(1), 72.5(1)
148.3(2), 75.3(2)
05-Y1-05
74.6(2)
Si-01
1.621(5)
01-Si-02
108.3(2)
Si-02
1.617(4)
01-Si-03
110.5(2)
Si-03
1.603(5)
01-Si-04
108.6(3)
Si-04
1.630(5)
02-Si-03
114.3(3)
02-Si-04
102.5(2)
03-Si-04
112.3(2)
26
Rietveld Refinement [14]
Powder X-ray data for Y2SiO5 have also been used to refine the structure. The
results obtained from the single crystal analysis have been used as a starting point.
Selected refinement results are summarized in Table 2.3. We find the unit cell volume
and bond distances to be equivalent to those of the single-crystal study (Table 2.2.) The
observed and calculated powder patterns of the Y1.25Scoi5Si05:2%Ce are shown in Figure
2.3.
Y1.25Sc0.75Si05:2% Ce3+, const=1
Lambda 1.5406 A, L-S cycle 1715
III
1.0
2-Theta, deg
I
1
2.0
11
1111
I I II 4
3.0
Hist 1
Obsd. and Diff. Profiles
II 111
IMq
4.0
I 11111111111114N I I= NI HI II
5.0
6.0
II 141414111NNIIIIMMINIONIMPOINNIIIIIMIN 11411IMIMI
7.0
8.0
9.0
X10E 1
Figure 2.3. Observed (+ line) and calculated (solid line) powder diffraction pattern
Of Y 1.25SCaT5S105:2%Ce. Difference curve appears at the bottom of the figure.
27
Table 2.3: Selected crystal data and interatomic distances for Y2_xScx,SiO 5
Rwp
Crystal
Y2SiO5
X=0.25
X=0.5
X=0.75
0.036
0.133
0.125
0.0877
0.101
2.721
2.735
1.315
1.515
X2
V(A3)-GSAS
854.4(1)
855.16(2)
844.76(8)
827.94(6)
811.2(2)
V(A3)-POLSQ
854.4(1)
854.3(1)
841.5(2)
829.3(3)
817.0(2)
0
0
0.206(4)
0.414(4)
0.590(4)
Y1 -01(A)
2.228(4)
2.251(9)
2.203(9)
2.239(9)
2.151(8)
Y1 -01(A)
2.276(3)
2.296(7)
2.363(7)
2.248(6)
2.232(6)
Y1-03(A)
2.279(6)
2.264(8)
2.259(8)
2.248(6)
2.222(7)
Y1-04(A)
2.294(4)
2.295(7)
2.221(7)
2.205(6)
2.222(7)
Y1 -O5(A)
2.197(6)
2.235(7)
2.127(7)
2.108(6)
2.053(7)
Y1-05(A)
2.303(3)
2.287(8)
2.237(8)
2.267(8)
2.272(8)
avg. Y1-0
2.263(41)
2.271(25)
2.235(77)
2.219(58)
2.192(78)
Y2-02
2.328(4)
2.319(8)
2.283(9)
2.217(9)
2.146(8)
Y2-02
2.334(4)
2.333(8)
2.372(8)
2.324(7)
2.394(8)
Y2-03
2.327(3)
2.290(8)
2.305(8)
2.307(7)
2.319(8)
Y2-04
2 610(4)
2.565(8)
2.620(8)
2.659(7)
2.706(7)
Y2-04
2.380(6)
2.393(8)
2.412(7)
2.383(6)
2.374(7)
Y2-05
2.207(5)
2.178(8)
2.238(7)
2.202(7)
2.211(7)
Y2-05
2.381(4)
2.423(7)
2.432(8)
2.411(7)
2.412(7)
avg. Y2-0
2.37(12)
2.37(12)
2.38(13)
2.36(15)
2.37(18)
coordinated)
2.326(63)
2.323(86)
2.340(77)
2.307(85)
2.309(108)
Si-01
1.621(5)
1.646(8)
1.622(9)
1.638(9)
1.656(8)
Si-02
1.617(4)
1.607(9)
1.605(9)
1.641(8)
1.576(8)
Si-03
1.603(5)
1.646(9)
1.613(8)
1.589(7)
1.558(8)
Si-04
1.630(5)
1.646(8)
1.633(7)
1.627(7)
1.586(7)
Occupation
factor for
Sc inY1
avg. Y2-0 (6-
28
On substitution of Sc for Y in the series Y2_xSc,,Si05 (0 < x
0.75) a
substitutional solid solution is formed, as judged by the monotonic decrease in unit-cell
volume with increasing values of x (Table 2.3). To gain a higher accuracy on the
volume, the unit cell parameters were refined by the computer program POLSQ based on
the same set of x-ray powder diffraction measurements for GSAS. Both results from
GSAS and POLSQ are shown in Figure 2.4. The occupation factors of the two Y sites by
Sc were determined by refining the occupancy factors of the two Y sites to set
approximate Sc concentrations in each site. Occupancies were then refined by
constraining the composition to that of the nominal formulation. From each refinement, a
consistent preference by Sc for the smaller Y1 site is observed. This result is presented in
Table 2.3 as the occupancy factor for Sc on the Y1 site, where approximately 80% of the
Sc atoms enter this site for each composition. Because most of the Sc enters the Y1 site,
the average Y1-0 interatomic distances become smaller, whereas the average Y2-0
distances remain essentially unchanged. But the additional coordinated oxygen present at
a relative longer distance [04] moves a little further away (shaded gray in Table 2.3). So
the coordination for Y2 site looks more like 6 instead of 7 as the concentration of Sc
increases.
108
107
--POLSQ
41- 'GSAS
106
Y3 105
E
j 104
103
102
101
0
0.25
0.5
0.75
x
Figure 2.4: Cell volume/Z (43) of Y2,ScxSi05.
29
Photoluminescence
As shown in Figure 2.5, Y2SiO5:2at% Ce3+ exhibits a UV-to-blue luminescence
under UV excitation at room temperature.
The emission doublet for Y2SiO5:Ce
corresponds to transitions from the lowest 5d level to the 2F512 and 2F712 ground states.
The energy difference is about 1910 cm -1 at 4.2 K. At 4.2 K (Figure 2.5); the profile is
essentially the same as that of 298 K except the FWHM is about 300 cm -1 narrower. By
monitoring the signal of the integrated emission band as a function of temperature, the
quenching temperature is found to be approximately 290 K (Figure 2.6). On addition of
Sc in the series Y2_,,Sc.Si05, the maxima of the emission bands red shift, and the doublet
character becomes blurred (Figure 2.7). The chromaticity values associated with these
shifts are summarized in Table 2.4.
on
in
Ed 0.6
13
0.5
a)
N
740.4
4 0.3
0.2
0.1
0
19750
20750
21750
22750
23750
24750
25750
26750
27750
Wavenumber (cm-1)
Figure 2.5: Emission spectra of Y2Si05:2%Ce at 4 K and 298 K 2,,,c = 300 nrn.
30
1
0 0.5
cv
0
I
50
0
150
100
250
200
300
350
Temperature (K)
Figure 2.6: Thermal quenching of Y2SiO5:2 %Ce.
Y2Si05
0.9
Y1.75Sc.25S i05
0.8
171
Y1.5Sc.5S i05
Y1.25 Sc.75 S i05
0.7
Oq
0.6
'34
0.5
74 0.4
2
0.3
0.2
360
380
400
420
440
460
480
500
520
540
Wavelength (nm)
Figure 2.7: Emission spectra of Y2,ScASi05:2%Ce at room temperature.
= 300 nm.
31
Table 2.4: Chromaticity, Excitation, and Emission Wavelength (nm) of Y2_Sc,Si05.
Y2Si05
Y1.5Sco.sSi05
Y1.75Sco.25Si05
Y1.25Sco.75Si
05
Chromaticity
x
0.151
0.149
0.149
0.149
y
0.039
0.041
0.042
0.043
Emission
262 300
264 302 357
265 304 361
270 305
at 450nm
356
Excitation X
364
Emission
320 376
-
-
330 391
400 426
413
420
422
The data of cathodoluminescence are shown
in Table 2.5.
at 520nm
Emission X
The efficiency of
Yi.75Sco.25Si05:2% Ce3+ is higher than that of Y1.5Sco.5Si05:2% Ce3+.
Table 2.5: Luminous Efficiency (lm/w).
Voltage (v)
Y1.75Sco.25Si05
Y1.5Sco.5Si05
Y2SiO5
500
0.13
0.065
0.22
700
0.15
0.081
0.35
1000
0.27
0.15
0.52
2000
0.54
0.29
1.02
3000
0.70
0.45
1.14
4000
0.84
0.50
1.19
* Commercial sample, Data provided by PTCOE
32
Discussion
Why does the Ce3+ emission peak shift to the red as Sc is introduced into the
Y2Si05 host? As seen from the structure determinations, the small Sc3+ cation (r = 0.88
A) [15] prefers to occupy the small Y1 sites. By analogy, we expect the large Ce3+ ion (r
= 1.15 A) to occupy the large Y2 sites, so we should concentrate on the characteristics of
the Y2 sites in explaining the observations. From the structure analyses we find the 0
atom at the relatively longer distance (04) has been push further away.
So the
coordination number of Y2 is more like 6 instead of 7. This makes the Y2 site a little bit
smaller.
These results are consistent with an increase in the Ce-0 interaction and a
lowering of the energy of the excited state (nephelauxetic effect).
The lower CL efficiencies of the Sc samples relative to that of the commercial
powder of Y2Si05:Ce may arise for a variety of different reasons.
The commercial
sample has been optimized in terms of its preparation and Ce doping level; the Sc
samples have not been optimized. The Sc samples could exhibit a lower intrinsic Ce3+
emission efficiency, but the thermal quenching measurement on Y1.75Sco.25Si05 (Figure
2.8) does not quite support this assertion. The inclusion of Sc into Y2Si05 would be
inducing sufficient strain to give rise to the production of high levels of defects and grain
boundaries.
Such imperfections are asserted to be important sites for electron-hole
recombination, which leads to poor phosphor efficiency.
In any event, a CL
measurement should be preformed on a Y2Si05:Ce samples that is prepared in the same
manner as that of the Sc derivatives. In this way, we can determine whether the Sc
phosphors
have
intrinsic
cathodoluminescent
comparable to those of Y2Si05:Ce.
or
photoluminescent
efficiencies
33
1.25
AAA
1
II
A.
A
A
A
A
ear)
in
E. 0.75
-8
s
0.5
z0
Y2SiO5, 407nm
0.25
Y1.75Sc0.25S105, 414nm
0
0
50
100
150
200
250
300
350
Temperature (K)
2.8: Comparison of thermal quenching of Y2SIO5:2 %Ce
Yi.7.5Sco.25SiO5:2%Ce. 2e = 300nm, monitoring the emission peak, respectively.
Figure
to
Excitation spectra have been measured at room temperature by monitoring the
emission at 450 nm (Figure 2.9). Three bands corresponding to the crystal-field split d
levels are observed near 270, 305, and 360 nm. By monitoring the peak wavelength of
the Ce emission for each sample, we find that the 360-nm band broadens, and the relative
intensities of the 270-nm bands decrease as the concentration of Sc increases.
This
maybe associated with the change of coordination from 7 to 6 (regular octahedral.) So
there should be just two energy levels in d orbital. Excitation in any of the three bands
yields the same emission spectrum for a given sample. These measurements provide no
direct evidence for the existence of more than one Ce3+ emission center. Lammers and
Blasse have made a similar observation [16], explaining the result by the existence of the
very similar y3+ cooridnation environments in Y2SiO5 that lead to similar energy
separations between the ground 4f1 and excited 5d' levels.
34
0.8
=
on
E-
0.6
O.
la
r:?
7,4
0.4
,!,
o
Y2SiO5
Y1.75Sc0.25S105
Y1.5Sc0.5S105
Z
0.2
Y1.25Sc0.75Si05
240
260
280
300
320
340
360
380
400
Wavelength (nm)
Figure 2.9: Excitation spectra of Y2,ScxSi05:2%Ce3+. Aeni=450nm.
By determining excitation spectra through the emission at 520 nm, however,
much different results are obtained (Figure 2.10). Each of the excitation bands shifts to
the red with increasing Sc concentration. These observations are consistent with the
presence of an additional luminescent center, possibly Ce3+ ions occupying the small Y1
site.
35
0.8
Y2-520
Y1.25-520
Y2-450
Y1.25-450
250
270
290
350
330
310
Wavelength (nm)
370
390
410
Figure 2.10: Excitation spectra of Y2_,ScASiO5:2%Ce3+. Y2: Y2SiO5,
Y1.25: Y1.2sSco.7sSi05, 520: Aem=520nm, 450: Aem=-450nm.
Conclusions
The maximum emission of the phosphor Y2SiO5:Ce
wavelengths by forming the solid solution Y2,ScxSi05:Ce (0 < x
is shifted to longer
0.75.) On the basis of
the results obtained from Rietveld refinements and photoluminescence measurements, the
shift is associated with an increased Ce-O interaction that results from the preferential
occupancies of Ce and Sc between the two substituional sites.
While a potentially
valuable red shift has been observed in the materials, additional work is necessary to
establish the CL and PL efficiencies of the Sc derivatives.
36
References
1.
J. Shumulovich, G.W. Berkstresser, C. D. Brand le and A. Valentino. J. Electrochem.
Soc., 135 (1988) 3141.
2. J. Lin, Q. Su, H. J. Zhang and S. B. Wang. Mater. Res. Bull., V31 (1996) 189.
3.
A. Meijerink, W. J. Shipper and G. Blasse. J. Phys. D-Applied Physics, V24 (1991)
997.
4. B. R. Chalamala, Y. Wei, B. E. Gnade, IEEE spectrum, April pp42-51(1998).
5.
A. Altomare, G. Cascarano, C. Giacovazzo & A. Guagliardi, J. Appl. Cryst. Vol. 26
(1993) 343.
6.
N. Walker and Stuart, Acta Cryst. A39, (1983) 158-166.
7.
A. C. Larson and R. B. Von Dreele, General Structure Analysis System, University
of California (1985-1990).
8.
A. Akella and D. A. Keszler. Mater. Res. Bull., V30 (1995) 105.
9. J. Felsche. Structure and Bonding V13 (1973) 99.
10. Powder Diffraction File Card No. 36-1476.
11. Powder Diffraction File Card No. 41-4.
12. E. M. Rabinovich, J. Shmulovich, V. J. Fratello, and N. J. Kopylov, Am. Ceram.
Soc. Bull., Vol. 66 (1987) 1505.
13. B. A. Maksimov, V. V. Iljukhin, Ju. A. Khantanov and N. V. Belov. Soviet PhysicsCrystallography, V15 (1970) 806.
14. R. A. Young. "The Rietveld Method", International Union of Crystallography,
Oxford Science Publications, 1993.
15. Shannon and Prewitt, Acta Crystallogr. B25 (1969) 925. Shannon, ibid., A32 (1976)
751.
16. M. J. J. Lammers and G. Blasse. J. Electrochem. Soc., V34 (1987) 2068.
37
CHAPTER 3
0.2) LUMINESCENT MATERIALS
FROM COMBUSTION PRECURSORS
Mgi-xCaNS:Eu (0
x
Dong Li and Douglas A. Keszler*
Department of Chemistry and Center for Advanced Materials Research
Gilbert Hall 153, Oregon State University, Corvallis, OR 97331-4003
38
Abstract
Eu2+-doped Mgi_xCaxS (0
x
0.2) phosphor powders have been prepared by
using an Mgi_Cax0 precursor that was prepared via a combustion process. The products
have been characterized by X-ray diffraction, scanning-electron microscopy, light
scattering, and luminescence spectroscopy.
The powders exhibit well-crystallized
micron-size particles and a high-efficiency, bright luminescence.
Introduction
Simple, lanthanide-doped binary alkaline-earth sulfides have for some time been
regarded as efficient, high-brightness phosphors. For example, red emitting CaS:Eu2+ has
a cathodoluminescent energy efficiency of 16% [1], and SrS:Cu+, Ag+ is regarded as the
brightest blue-emitting electroluminescent (EL) phosphor [2] that is currently available.
Because of their high-performance characteristics, the materials have been considered for
applications in CRTs [3,4], flat-panel EL displays [5], and optically stimulated
luminescence dosimeters [6-8].
Although many studies have been described on the
development of CaS- and SrS-based phosphors [1-8], few reports have appeared on the
properties of doped MgS [9].
A presumed chemical instability and difficulties in
synthesis appear to have contributed to this lack of study.
Here, we describe results on (Ca,Mg)S:Eu luminescent powders that have been
synthesized by utilizing a combustion method [10-12]. Impetus for doing this work has
been derived from the reported energy efficiencies of Eu2+-doped materials, which
increase in the order BaS < SrS < CaS < MgS [3]. In addition, by adjusting the relative
concentrations of Ca and Mg in the solid solution Mgi,CaxS:Eu2+,
Casano and co-
workers have shown that a true, chromatically red phosphor can be produced. Up to
now, MgS-based phosphors have been synthesized by passing CS2(g) or H2S(g) over
MgCO3 or MgSO4 at elevated temperatures [9] with and without various promoters, such
as PC13 [3]. The chemical stability (especially with respect to H2O) and luminescence
efficiency of the product appears to be greater in those samples containing lower levels of
oxide impurity [3]. So, from the point of view of luminescence efficiency, color control,
39
and chemical stability, it is desirable to develop alternative means for the synthesis of
(Mg,Ca)S:Eu phosphors with prospects for improved performance characteristics.
Experimental
Stoichiometric compositions of Mg(NO3)2.6H20 (EM Science, GR),
glycine
(Sigma, 99% TLC,) and Ca(NO3)2.4H20 (EM Science, GR) were set from the following
equation:
91(1-x) Mg(NO3)2.6H20 (s) +x Ca(NO3)2.4H20 (s)
NH2CH2COOH (s)
+ 9Mgi_1Car0 (s)+14N2 (g) +200O2 (g) +( 79-18x)H20 (g)
Approximately 6 g of the mixture and 0.1 at% Eu as Eu203 (Kerr-McGee, 99.9%)
were placed in a 250-mL beaker and dissolved in 20 ml of H20(1) obtained from a
Millipore purification system. The beaker was covered by a watch glass, and then heated
on a hot plate in a fume hood to evaporate the water and initiate the reaction. When the
mixture reached the ignition temperature, a mild explosion occurred with a flame lasting
approximately three seconds.
The reaction was very rapid (within fifteen seconds)
resulting in the formation of white fluffy oxide powder. Because the product is a very
fine dust, the reaction should be performed in a fume hood, and a suitable dust mask
should be employed. Due to the high explosive feature, it is highly recommended the
reaction is performed in a furnace to provide better controlled heating atmosphere and
shield for any possible flying particles.
The resulting powder was transferred to a
graphite boat and inserted into a fused silica tube. The sample was then fired for 2 h at
1373 K in a stream of H2S(g) (Matheson Gas Products Inc., C. P.). H2S is a flammable,
highly poisonous gas. It is a severe irritant to the eyes and mucous membranes.
Outgoing gas can be effectively trapped with an aqueous NaOH + H202 solution.
Particle sizes, crystallite sizes, and unit-cell parameters were measured by
scanning electron microscopy, light scattering, and X-ray diffraction. SEM micrographs
for MgS prepared by sulfate/H2S method were obtained at 30kV with a magnification of
1000x. SEM micrographs for sulfide prepared by combustion synthesis were obtained at
5 kV with a magnification of 5000x view. An Horiba K-700 particle analyzer was used
to perform light-scattering experiments for determination of particle sizes in the range of
40
0
50 p.m; samples were dispersed in a mixture of 50% H20(1)/50% glycerol. X-ray
diffraction data were recorded on a Siemens D5000 X-ray powder diffractometer by
using Cu Ka radiation. Data were collected over the 20 range 10-150° by counting for 2
s at each step of 0.01°. Crystallite size of the oxide precursor was determined by the
Warren-Averbach method [13] (win-crysize software, Siemens 1995). Crystallite sizes
and strains of the sulfide products were derived from Williamson-Hall plots [14].
Photoluminescence measurements were made at room temperature by following
procedures that have been described elsewhere [15].
Cathodoluminescence were
measured at Phosphor Technology Center of Excellence located at Georgia Tech.
Research Institute. Chromaticities were determined from the emission spectrum by using
the principle described in chapter 1. Excitation spectra were corrected by using sodium
salicylate as a quantum counter.
Results and Discussion
In the high-temperature preparation of MgS, CaS, or solid solutions between the
two components, the critical synthesis step is generally the sulfurization of the oxide
MO(s) + H2S(g) (or CS2(g)) > MS(s) + H20(g) (or CO2(0)
AG > 0,
where M = alkaline earth metal. Because the free energy of the reaction is positive, the
reactants are naturally favored under equilibrium conditions. To shift the equilibrium in
favor of products, a large excess or flow of H2S (or CS2) is generally used [9]. In prior
work on the solid-solution system Mgi_xCaS, an inmiscibility gap was reported to cover
the range 0.15 5_ x 5_ 0.89.
Kasano, Megumi, and Yamamoto [3], however, have
prepared samples across the entire range 0
x 5 1 by using PC13 as a promoter.
To further develop the sulfurizaton process of the oxide, we have examined the
reaction by using an oxide precursor prepared by the combustion method [10-12], which
results in the production of a very fine, unagglomerated ceramic powder. The basic
principle of the process is the decomposition of an oxidizer, such as a metal nitrate, in the
presence of a nitrogen-based fuel; common fuels are urea [10], glycine [11], and azides
[12]. Two key features are important to the success of the combustion method: (1) the
energy required to drive the chemical reaction is provided by the reaction itself and not
41
by an external heat source, and (2) the large volume of gas generated during the reaction
will prevent agglomeration of oxide powder, resulting in a very fine particle size on the
submicron or even nanometer scale. This gas also prevents sintering by dissipating much
of the heat generated by the reaction. In this study, we have examined the reaction of a
nitrate/glycine mixture. Glycine will complex both lanthanide and alkaline-earth cations
through ammine and carboxyl ligation modes, thereby increasing cation solubility and
limiting selective precipitation as water is evaporated.
Characterization of Oxide
An X-ray diffraction pattern of the oxide product (nominal formula Mgo.9Cao.iO)
prepared by this method is shown in Figure 3.1. The diffraction peaks of the product shift
to lower 20 angles in comparison with those of pure MgO. This phenomenon indicates
the formation of a solid solution. This is a little amazing because Mgi,Cax0 can only be
formed at about 1873 K and x can just get up to 0.02 [16]. But the reported maximum
temperature for a glycine combustion reaction can get is below 1673 K [10-12].
The very broad diffraction peaks in Figure 3.1 also indicate the crystallite size of
the product is very small. The results are shown in Figure 3.2. The median crystallite
size of Mgo.9Cao.iO is 13nm.
42
1600
1400
1200
1000
800
600
400
200
0
30
110
50
150
130
Two Theta
Figure 3.1: X-ray powder pattern of Mgo,9Cao,iO prepared by combustion
synthesis. The line is diffraction pattern of MgO.
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
(a)
0.1
0
0
2
4
8
10 12 14 16 18 20
Length (nm)
43
1
02
0.8
0.7
0.6
0.5
0.4
0.3
(b)
02
0.1
10
12
14
16
18
20
Length (nm)
Figure 3.2: Plot of the Fourier size coefficient against (a). Crystallite
size distribution and (b). Median crystallite size of Mga9Ca0.10.
Characterization of Sulfide
By treating the oxide precursors under flowing H2S(g) at elevated temperatures,
the corresponding sulfides are readily formed.
Scanning electron microscopy (SEM)
images of the material Mg0.96Ca0.04S:0.1% Eu from the oxide precursor and that obtained
by the conventional H2S-treatment of a sulfate mixture (MgSO4 + CaSO4) are shown in
Figure 3.3. Significant differences are evident in particle size and shape between the two
samples. Particles prepared by the combustion method have well-defined edges and a
rather homogeneous size distribution around 2 pm.
Samples prepared by the
conventional method, however, have poorly defined shapes and a rather inhomogeneous
size distribution with several larger particles around 10 pm.
Because size of the
precursor particles formed from the combustion reaction is sufficiently small, it should be
possible to generate varying sulfide particle sizes by controlling the reaction time and
temperature.
44
Figure 3.3. SEM micrographs of Mgo.96Cao, 04S: 0.1%Eu2+ prepared
by sulfate method (upper) and combustion method (lower)
45
Results from the light-scattering measurements on the sulfide powder are
summarized in Figure 3.4. Here, the median particle size is determined to be 4-5 [tm.
This result is similar to the visual estimates from the electron micrographs. The powder
pattern of Mg0.8Ca0.2S was shown in Figure 3.5. The arrow indicated the highest peak
that can be associated with MgO. The ratios of maximum peak intensity of MgO to that
of MgS are given in Table 3.1.
14
12
10
t
8
6
4
2
0
10-9
9-8
9-7
7-6
6-5
5-4
4-3
3-2
2-1
Size Range
Figure 3.4: Particle size distribution of Mg0.9Caa1S determined by light
scattering.
1-0.9
46
Table 3.1: Characterization of Mgi_CaxS
IMg0/1MgS
Crystallite Size
a (A)
Strain
(Pm)
MgS
1.9%
5.203±0.0003
0.132
5.33E-5
Mgo.96Ca0.04S
0.61%
5.226±0.0003
0.131
8.54E-5
Mgo.93Cao.o7S
0.75%
5.241±0.0003
0.159
2.16E-4
Mgo.9Cao.iS
0.79%
5.257±0.0002
0.154
3.55E-4
Mgo.s6Cao,14S
0.58%
5.279+0.0001
0.145
2.05E-4
Mgo.sCa0.2S
1.4%
5.310+0.0003
0.183
6.36E-4
Mgo.9Cao.10
-
-
0.0129
3E-3
9000
8000
7000
6000
5000
4000
3000
2000
MgO
1000
0
10
30
50
70
90
110
130
Two Theta
Figure 3.5: X-ray powder diffraction pattern of MgasCaa2S:0.1%Eu
150
47
The production of MgS from MgO is much more difficult than the production of
the corresponding sulfide from CaO, SrO, or BaO. Indeed, the production of a material
having a high MgS/MgO ratio has apparently been achieved only by including PC13(g) in
the reaction [3]. A representative X-ray powder diffraction pattern of Mg0.8Ca0.2S:0.1%
Eu is shown in Figure 3.5. The narrow peak width and high peak intensity indicates the
sulfide is in very good crystalline form. However, a peak attributable to the MgO has
been identified; the ratio of maximum peak intensity of MgO to that of MgS is given in
Table 3.1.
The lattice constant as a function of composition in the series Mgi.
xCaxS:0.1at% Eu (0 5 x < 0.2) is shown in Figure 3.6. Because Ca (r=1.14 A) is larger
than Mg (r=0.86 A) [17], the lattice parameter increases with increasing Ca content. This
indicates a solid solution is formed.
5.34
5.32
5.3
5.28
5.26
5.24
5.22
5.2
5.18
0
0.05
0.15
0.1
x
Figure 3.6: Lattice constants of Mg 1_,cCax-S.
0.2
0.25
48
For the solid solution Mgi_xCa.S, we just synthesized compositions to x = 0.2.
There are two reasons for this. First, from the following results, Mgo.gCao.2S:Eu shows
the true red chromaticity (matches the chromaticity of Y203:Eu) needed by red phosphor,
and second, it is difficult to prepare the solid solution with higher Ca contents by using
the combustion synthesis with glycine. However x = 0.2 already overcomes the limit of
solid solution (x=0.15) prepared by conventional sulfate/H2S method [16]. The lattice
constant of MgS:0.1% Eu (a = 5.203±0.0003 A), is a little larger than MgS:0.04% Eu (a
= 5.2015 A) [3] and undoped MgS (a = 5.200 A). The slight increase of lattice constant
is thought to be caused by Eu doping, because Eu
A) is larger than Mg (r----0.86
A) [17].
Particle size of the sulfide solid solution has been determined by SEM and light
scattering method as mentioned in the previous section. However, there are several
domains (crystallites) defined by the grain boundaries in one particle.
It is known the
electron-hole pair can recombine on such defects, so the efficiency will be affected. To
better define these effects, the crystallite size of the sulfide was determined from a x-ray
method, Williamson-Hall plots [14]. The principle of Williamson-Hall plot is that the
broadening of diffraction peak is caused by both small crystallite size and strain:
FWHM = 4(, /2)xtan0 + X./(txcos0)
where FWHM is the full width at half maximum of diffraction peak (expressed in
radian), /2 is the tensile or compressive strain,
is the wavelength of X-ray, t is the
crystallite size, and 20 is the diffraction angle. If we plot sine on the abscissa and
FWHMxcos0 on the ordinate, the strain and crystallite size can be deduced from the
slope and intercept, respectively. The results are shown in Figure 3.7 and Table 3.1. The
crystallite size determined from x-ray is one order of magnitude smaller than the particle
size determined by SEM and light scattering. The strain also increases as the content of
Ca increases.
49
0
0.5
1
1.5
2
2.5
3
3.5
4
4*sin(theta)
Figure 3. 7: Representative Williamson-Hall plot for Mg0.9Ca0,1S: 0.1%Eu2+
50
Photoluminescence
Room-temperature photoluminescence spectra for Mgi_xCaS:0.1% Eu are shown
in Figure 3.8.
1
x=0
x=0.04
x=0.07
x=0.1
x=0.14
x=0.2
0.8
7,4
.c7)
F-, 0.6-
0.4
z0
0.2
0
500
550
600
650
700
750
Wavelength (nm)
Figure 3.8: Emission spectra of Mgi_,CaiS:0.1%Eu. 2,=315nm.
As expected for Eu2±, broad-band emission associated with the transition
4f65d1>4f7 is observed. As seen from Figure 3.8, incorporation of Ca into MgS:Eu leads
to a broadening and a red shift of the luminescence band. As noted from the X-ray
diffraction results, a solid solution of CaS in MgS is formed at these compositions. So
there is not a separate phase of CaS:Eu (,(peak emission) = 650 nm) that is giving rise to
the boadening and shifting of the bands. The broadening likely arises both from the
relative concentration and inhomogeneous distribution of Ca in the matrix. At lower Ca
51
concentrations, e.g., x = 0.04, there is only a 25% probability that a Eu dopant atom will
occupy a site with Ca as a next-nearest neighbor. As a result, two rather distinct types of
sites may be expected
those without the next-nearest neighbor Ca atoms (a site much
like that in MgS) and those with the next-nearest neighbor Ca atoms. These types of sites
and their inhomogeneous distribution through the host likely give rise to the rather broad
spectral peak for the composition x = 0.04. As the Ca concentration increases, e.g., x =
0.07, the probability for location of a Eu atom in a site having a next-nearest neighbor Ca
atom increases, the distribution begins to favor occupation of such a site, and the
emission band becomes more asymmetric and skewed toward longer wavelengths. At
sufficiently high Ca concentrations, e.g., x > 0.1, the probability for association of Eu
with a next-nearest neighbor Ca atom is sufficiently high that the emission band becomes
more symmetrical and largely broadened by the inhomogeneous distribution of Ca.
A
MgS
Mg.93Ca.07
Mg.86Ca.14
0
230
260
290
320
350
380
410
440
Wavelength (nm)
Figure 3.9: Excitation spectra of Mgi_xCa,S:0.1%Eu.
470
500
530
52
By monitoring the emission maximum for each sample, excitation spectra have
been obtained (Figure 3.9).
In agreement with Asano and Nakao [18], three main
excitation bands labeled A, B, and C are observed.
On the basis of computational
studies, bands A and C have been assigned to transitions of the type 4f7- *4f65d', and
band C is assigned to the transition 4f7--->4f66s1. Hence, band A corresponds to transitions
to the t2g, d set, band B to the 6s level, and band C to the eg, d set.
According to the configuration coordinate diagram, the distance AR between the
lowest energy points of two adiabatic potential curves, which represent the ground and
excited states, is a qualitative measurement of the interaction between electrons and
vibrations of the optical center under consideration. AR is anticipated to be smaller in the
case of MgS:Eu in comparison to the cases of CaS:Eu and Mgi.CaS:Eu, because of the
difference of alkaline-earth radii (r(Mg) = 0.86 A, r(Ca) = 1.14 A.) This effect also
contributes to the more pronounced phonon structure in the excitation spectrum of MgS
in comparison to the (Mg,Ca)S materials and the larger Stokes shifts observed for the
higher Ca concentrations.
Chromaticities for each of the compounds are given in Table 3.2. Results of low-
voltage cathodoluminescence measurements are summarized in Table 3.3 and Figure
3.10.
Table 3.2: Emission properties of illgi_,CaS:Eu and Y203:Eu
Chromaticity
Max. Emission A, (nm)
*
x
y
MgS
588
0.562
0.438
Mg0.96Ca0.04S
607
0.605
0.395
Mg0.93Ca0.07S
619
0.610
0.390
Mg0.9Cao. i S
624
0.625
0.375
Mgo.86Cao.14S
626
0.632
0.368
Mg0.8C43.2S
633
0.645
0.355
Y203:Eu
611*
0.64
0.36
line emission
53
Table 3.3: Luminous Efficiency (lm/w) at Different Voltage.
Voltage
Y203: 4%Eu
(v)
Mg0.9Cao.IS: 0.1%Eu2+
Mgo.9Cao.IS: 0.1%Eu2+
(combustion method)
(Sulfate/H2S method)
500
1.801
0.880
0.245
1000
4.084
2.34
0.660
2000
7.739
8.07
2.90
3000
8.92
4.36
1.52
Combustion
G *Y203: Eu
Sulfate/H2S
0
500
1000
1500
2000
2500
3000
Voltage (v)
Figure 3.10: Comparison of himinous efficiency of Y203:4%Eu and
klgo.9Cao.1S:0.1%Eu prepared by combustion and sulfate/H2S method
3500
54
The phosphor prepared by using the combustion-derived precursor is not only
about three times as efficient and bright as that prepared by the conventional method, but
it also exhibits an efficiency comparable to that of the commercial red phosphor Y203:Eu
[19].
Summary
Highly sulfurized Mgi.,,CaxS:Eu luminescent materials have been prepared by
treating with H2S(g) a Mgi,Cax0 precursor obtained from a combustion reaction. The
process yields well-crystallized particles in a narrow size-distribution range.
High
cathodoluminescent brightness and efficiency are observed.
Acknowledgments
This work was supported by DARPA through the Phosphor Technology Center of
Excellence.
References:
1.
W. Lehmann and F. M. Ryan, J. Electrochem. Soc., Vol 118, 477 (1971).
2.
Sey-Shing Sun, E. Dickey, J. Kane, P.N. Yocom, Proc. 17th Int. Display Research
Conf , Morreale, J., Ed.; Society for Information Display: Toronto, p201, 1997.
3.
H. Kasano, K. Megumi, and H. Yamamoto, J. Electrochem. Soc., Vol 131, 1953-1960
(1984).
4. B. T. Collins and M. Ling, J. Electrochem. Soc., Vol 140, 1752-1755 (1993).
5.
D. Poelman, R.Vercaemst, R. L. Van Meirhaeghe, W. H. Laflere, F. Cardon, J.
Lumin., Vol 65 7-10 (1995).
6. 0. Missous, F. Loup, J. Fesquet. H. Prevost, Eur. J. Solid State Inorg. Chem., Vol 28
(supplement) 163-166 (1991).
7. K. Chakrabarti, V. K. Mathur, J. F. Rhodes, and R.J. Abbundi, J. Appl. Phys., Vol 64,
1363-1366 (1988).
55
8.
Y. Tamura, Jpn. J. Appl. Phys., Vol. 33, 4640-4646 (1994).
9.
N. Singh, O.R. VIJ, J. Mater. Sci.,Vol. 29, 4941-4945 (1994).
10. D. A. Fumo, M. R. Morelli, and A. M. Segadas, Mater. Res. Bull., Vol. 31, 12431255(1996).
11. L. A. Chick, L. R. Pederson, G. D. Maupin, J. L. Bates, L. E. Thomas, and G. J.
Exarhos, Mater. Lett., Vol. 10, 6-11 (1990).
12. M. kottaisamy, D. Jeyakumar, R. Jagannathan, and M. M. Rao, Mater. Res. Bull.,
Vol. 31, 1013-1020 (1996).
13. P. Klug and L. E. Alexander, X-ray Diffraction Procedures, 2nd ed., Chap. 9, New
York, Wiley, (1974).
14. G. K. Williamson and W. H. Hall, Acta. Metall. Vol. 1, 22 (1953).
15. A. Akella and D. A. Keszler, Mater. Res. Bull., Vol. 30, 105-111 (1995).
16. Gmelins handbuch der Anorgnischen Chemie, 28(a) Hb/B3 13463/4(1961).
17. Shannon and Prewitt, Acta Crystallogr. B25 (1969) 925. Shannon, ibid., A32 (1976)
751.
18. S. Asano and Y. Nakao, J. Phys. C: (solid State Phys.), Vol. 12, 4095-4108 (1979).
19. Database of FED phosphors from PTCOE.
56
CHAPTER 4
NIOBATE PHOSPHORS PREPARED BY SOL-GEL METHOD
Dong Li and Douglas A. Keszler*
Department of Chemistry and Center for Advanced Materials Research
Gilbert Hall 153, Oregon State University, Corvallis, OR 97331-4003
57
Abstract
GdNbO4:Bi and YNbO4:Bi phosphors have been prepared by the sol-gel method;
photoluminescence spectra and cathodoluminescence efficiencies were determined.
Compared to Y2SiO5:Ce, GdNbO4:Bi has a similar luminous efficiency, while YNbO4: Bi
exhibits a higher efficiency. The relatively high efficiencies of these niobates are partly
associated with their long wavelength emission tails that extend into the green portion of
the spectrum. Hence, the efficiency under true-blue filtering needs to be determined to
assess their efficacy as blue FED phosphors.
Introduction
Steady improvements in the area of ceramic spacer, cathode development have
been made in FED industry. Based on these developments, commercial product will be
due to the market this year.
However there are still no suitable phosphors that are
particularly designed for FED.
Some companies, such as Candescent, still use
conventional CRT phosphors P22 in their FED products. The shortage of FED phosphors
was most severe for blue. The sulfide phosphors are not stable enough and the oxide
phosphors do not show high enough brightness. The emission spectrum of YNb04: Bi is
similar with that of classical ZnS: Ag, although it is broader. We expect the highly
refractory oxides to be stable.
Niobate phosphors have been studied from 1960's [1]. The majority compounds
have nearly octahedral Nb06 with sharing of corners, edges, or faces with nearest
neighbor Nb06. There are also materials such as rare earth niobates LnNbO4 ( LnY, La,
Gd, Lu), which have isolated nearly tetrahedral Nb04 complexes with no sharing of
oxygen with other complex. However the crystal structure refinement based on valance
calculation has shown the coordination of Nb5+ ion is really 4+2 with two oxygens at
relatively long distance [2]. The distorted Nb06 polyhedra are linked by edges and form
chains in the [001] direction. The LnNbO4 niobates show reasonable efficient deep blue
emission with a high quenching temperature [3].
58
The luminescence of pure YNbO4 and GdNbO4 have been investigated previously
[2, 4]. Blasse and Bril also studied YNbO4:Bi. Further work by Grisafe and Fritsch [5]
extended investigation to GdNbO4:Bi.
The low voltage cathode ray phosphor
characteristics of YNbO4:Bi were studied by Vecht and Co-workers [6]. They found
YNbO4: Bi showed higher luminance and efficiency compared to commercially available
P47 (Y2SiO5: Ce).
In view of the high preparation temperature (1300°C-1500°C) and the marked
volatility of Bi203, it is difficult to prepare Bi3+-activated phosphors in a controlled way.
In this study YNbO4: Bi and GdNbO4: Bi were prepared by sol-gel method. We hope in
this way the Bi3+ concentration can be controlled. The phosphors were also prepared by
flux method for comparison.
Experimental
Sol-gel process provides a new approach to the preparation of ceramics [7,8,9].
Starting from molecular precursors, an oxide network is obtained via inorganic
polymerization reactions based on hydroxylation and condensation. Two different routes
are usually described in the literature depending on whether the precursor is an inorganic
salt (cheap, less versatile) or a metal organic compound (versatile, expensive).
The major steps involved in sol-gel process includes [10]:
1.
Hydrolysis: Acid or base catalysts are usually employed to speed up the reaction.
2. Polymerization: Condensation reactions occur between adjacent molecules in which
H2O and /or ROH are eliminated and metal oxide linkages are formed. The colloidal
dispersion form connected by polymer networks is termed a sol.
3.
Gelation: The polymer networks link up to form a 3-D network throughout the liquid.
The system becomes rigid. Solvent remains in the pores of the gel.
4. Drying: Water and alcohol are removed at low temperature (<100°C)
5.
Dehydration: Moderate temperatures, 400-800°C, are required to drive off the
residual organic and chemically bound water.
6.
Densification: Temperature typically in excess of 1000°C cause elimination of
porosity and formation of a dense metal oxide.
59
The key steps are hydrolysis and polymerization:
h=0
0<h<2N
h=2N
M- (OH) -M
Tolation
[M(OH)h(OH2)N-hr.")+ hydrooxo-aquo-ion
[M(0112)N]A+
[M(OH)Nr-A)- hydro-ion
aquo-ion
[M0x(OH)N.x] (N+x-A)- oxo-hydro-ion
[mois](2N-A)
4,oxolation
M-O-M
MONH2N.h is the key specie that must present for condensation. Because it has
both nucleophilities due to M<co) and good leaving group H2O. The only difference for
metal alkoxide is some coordination groups are ROH instead of H2O.
C. Alquier and co-worker found there are four ways to synthesis Nb205 gels [11].
The best process is the chloroalkoxide route. This is a good compromise between the
flexibility of the alkoxides and the lower cost of the mineral precursor NbC15. In this
study, this method was adopted and adjusted to prepare YNbO4:Bi and GdNb04:Bi.
9.466g of NbC15 (AESAR 99.9%) was dissolved into 250m1 of ethanol
(McCormik distilling Reagent Quality Co.) to form chloroalkoxides solution according to
the following equation:
NbC15 + 3CH3CH2OH > 3HC1 + NbC12(OCH2CH3)3
The mixing needs to be fast in order to avoid moisture. It appears to be very exothermic
and HC1 gas is evolved. The solution is stored in a dry atmosphere for at least 24 hours
which is named stock solution
1.
Stock solution 2 was prepared by dissolving
stoichiometric amount of Gd203, Bi203 in nitric acid solution or Y(NO3)3.6H20 (AESAR
99.9 %(REO)) in H2O. The hydrolysis is performed on 10m1 of stock solution 1. 7 ml of
stock solution 2 was added dropwise under stirring for 15minutes. The solution mixture
was put at 50°C. White opaque gel is formed after 30 minutes. The mixture was put
there overnight to complete gel transition. The gel was heated at 110°C for 6 hours to
evaporate excessive water and alcohol. The sample is then moved to alumina crucible
and heated at 200°C/hour to 700°C. Then the sample is taken out and grounded. The
60
final firing is at 1050°C for 20 hours. The YNbO4 and YNb04:Bi phosphors was also
prepared by solid state reaction at 1050°C by adding 5 weight percent KF.
Results and Discussion
X-ray diffraction shows YNbO4 and GdNbO4 were formed by sol-gel method.
The peak high is higher than that of the sample prepared by solid state reaction. This
indicates the crystallinity is higher. YNbO4 shows different emission peaks under 254nm
and 304nm excitation. This is shown in Figure 4.1. When Bi is doped, the emission peak
shifted to the red.
The corrected emission and excitation spectra of YNbO4 and
YNbO4:Bi are shown in Figure 4.2.
1
y'
0.9
254
0.8
\\,
.2 0.7
nm
304 nm
0.6
..017)
0.5
.N
71 0.4
g
0.30.2
0.1
0
340
380
420
460
500
540
Wavelength (nm)
Figure 4.1: Emission of YNbO4 under 254nm and 304nm excitation.
580
61
240
270
300
330
360
390
420
450
480
510
540
570
600
Wavelength (nm)
Figure 4.2: Corrected emission and excitation spectra of YNbO4 and YNbO4:Bi.
GdNbO4 also has different emission peaks under 254nm and 307nm excitation,
although not as pronounced as that of YNbO4. This is shown in Figure 4.3. This result is
in contrary to the observation of Grisafe and co-worker who claimed GdNbO4 is
essentially inert under 254nm excitation [5] and in agreement with the results of Dyer and
White who state all rare earth niobates show cathodoluminescence [12]. The absence of
luminescnece in GdNb04 were explained by Blasse and Bril as energy transfer from
niobate to gadolinium, but the latter shows no luminescence due to concentration
quenching [3]. However it is not much overlap from our determination. The Bi doping
also causes continuous red shifting up to 5% (the range we studied). While Grisafe and
co-worker found the Bi peak emission is independent of the activator concentration. The
corrected excitation and emission spectra of GdNb04 and GdNbO4:Bi are shown in
Figure 4.4.
62
1
0.9
307nm Ex
254nm Ex
0.8
0.7
0.6
0.5
0.4
E
O 0.3
Z
0.2
0.1
0
360
390
420
450
510
480
540
570
600
Wavelength (nm)
Figure 4.3: Emission of GdNbO4 under 254nm and 307nm excitation.
1
0.9
0.8
0.7
E-1
0.6
0.5
N
-7; 0.4
2 0.3
0.2
0.1
-
0
240
290
340
I
390
440
490
540
Wavelength (nm)
Figure 4.4: Excitation and emission of GdNBo4 and GdNb04:Bi.
590
63
Figure 4.5 shows the relative brightness as the bismuth concentration in the
gadolinium orthoniobates under 254nm excitation. The optimum bismuth concentration
is 1 mol%, which is a little lower than 2 mol% observed by Grisafe and co-worker [5].
110
100
90
80
70
GdNBO4
-s- YNbO4:2 %Bi
60
50
0
0.01
0.02
0.03
0.04
0.05
0.06
Mole% Bi3+
Figure 4.5: Relative brightness as a function of the bismuth concentration.
Three samples were sent to Georgia Technology Institute for FED determination.
Their results are shown in Figure 4.6 and Table 4.1. The efficiency of GdNbO4: Bi
prepared by sol-gel method is lower than the efficiency of YNb04: Bi prepared by flux
method. The sample prepared by sol-gel method is very hard after sintering and need
more grinding.
We think the excessive grinding process caused the decrease of
efficiency. The chromaticities are also shown in Table 4.2. The difference of the value is
due to the different emission range.
64
Table 4.1: Luminous efficiencies of some phosphors.
500V
700V
1000V
2000V
3000V
Y2SiO5:Ce
0.22
0.35
0.52
1.02
1.14
Y1.75Sc0.25Si05:2%Ce
0.1265
0.1536
0.2657
0.5378
0.7001
Sr2P207: Sn2+
0.1265
0.1446
0.2531
0.4998
0.6327
YNb04 (1(F)
0.091
0.122
0.213
0.582
0.84
YNb04,2%Bi3+ (KF)
0.087
0.161
0.33
1.167
1.754
GdNb04:1%Bi3+ (sol-
0.061
0.085
0.165
0.44
0.632
0.013
0.025
0.081
0.44
0.632
(LiF)
gel)
KNbOP2O7
Table 4.2: Chromaticities of some phosphors
400-700nm(PTCOE)
360-600nm(Our Lab)
x
y
x
y
Y2SiO5:Ce
-
-
0.151
0.039
YNb04 (x.F)
0.185
0.188
0.154
0.128
YNb04:2%Bi3+ (1(F)
0.223
0.315
0.160
0.172
GdNb04:1%Bi3+ (sol-gel)
0.198
0.259
0.153
0.119
KNbOP2O7
0.219
0.307
-
-
65
2
Sr2P2O7: Sn2+
1.8
1.6
-*-- Y1.75Sc0.25Si05:2%Ce (LiF)
L4
YNb04:2%Bi3+ (KF)
6-11.2
GdNbO4:1 %Bi3+ (sol-gel)
.r
1
oo
0.8
§ 0.6
a
0.4
0.2
0
0
500
1000
1500
2000
2500
3000
3500
Voltage (V)
Figure 4.6: Dependence of luminous efficiency on voltage.
The optical transition of the complex ion Nb04 can be classified as a charge
transfer type [1, 4]. The highest ground state level is non-bonding 7c-orbital t16 and
contained mainly the 2p-orbitals of oxygen.
The two lowest excited states were
antibonding ti5e and ti5t2, where the e and t2 orbitals were mainly derived from the metal
3d orbitals.
The two excited levels will be further split by crystal field due to site
symmetry. The 6P>8S transition of Gd is about 310nm. There is some overlap with
niobate emission band. Therefore, energy transfer exist between them.
Bi3+ is a 6S2 ion. The ground state of S2 yields only one level 'S0. The excited
state configuration, sp, yields 3P0, 3P1, 3P2, and 'P1 [1, 13]. The energy diagram is shown
in Figure 4.7. The spin selection rule is relaxed by spin-orbit coupling. Due to the
selection rule iJ =0, ±1 (0>0 is forbidden), the transitions 1S0-43P0 and 1S0>3P2 should
66
remain forbidden. The crystal field effect will make transitions requiring AJ=2 become
allowed. The emission is due to the 3130,1>1So transition. Whether 3P0 or 3P1 is the initial
level depends on their energy difference and the temperature. Excited 3P state also
undergoes spin-orbit interaction and John-Teller interaction.
IP1
3D
r2
3
P1
3P0
'so
Figure 4.7: Schematic representation of the energy level scheme of a free s2 ion.
Metal-Metal charge transfer from s2 ions to d° ions makes this case even more
complicated [14].
In some cases, the charge transfer state changes the quenching
temperature drastically.
The peaks in excitation spectra can be assigned as follows from the above
discussion. The very broad band before 290nm is mainly due to Nb4--0 charge transfer
and 'S0>3P1 direct excitation of the Bi3+ center. So when Bi concentration is increased,
this band intensity also increased as in Figure 4.4. Another contribution is possibly from
Bi*Nb metal-metal charge transfer.
The sharp peak around 304nm is due to Gd3+
intraion transition and Bi +-O charge transfer. Datta mentions that the 320nm absorption
67
in Y203: Bi may also be due to a charge transfer mechanism involving bismuth and
oxygen [15]. However there is also a sharp peak around 304nm in pure YNbO4. We
think this may be due to some impurity Bi3+ in it. Because we prepared YNbO4 and
YNbO4: Bi samples side by side.
The emission is ascribed to both transition from 3P-->1 So of Bi3+ and niobate
emission. The niobate emission is at shorter wavelength. So when the sample is excited
by 254nm light, both Bi and niobate are excited. When the sample is excited by 310 nm
light, only Bi is excited. Therefore, the emission is at longer wavelength.
In survey of the FED properties of niobate system, we found some compounds
with a short Nb-O bond [16]. The oxygen is unshared by other atoms. Blasse and coworker found the niobate contains this niobyl (NbO) groups shows strong luminescence
and the niobate has regular Nb06 octahedra is almost non-luminescent [17]. We prepared
a
sample with niobyl group
cathodoluminescence.
KNbOP2O7 and determined its low voltage
The results are shown in Table 4.1.
It has about the same
efficiency as YNbO4 and GdNbO4: Bi above 2000v. However the efficiencies at low
voltage are much lower.
Conclusion
GdNbO4:Bi was prepared by sol-gel method at relatively lower temperature. This
phosphor has lower cathodoluminescence efficiency at low voltage than that of YNbO4:
Bi prepared by flux method, although it shows high photoluminescence efficiency under
254nm excitation.
The low FED efficiency may be due to excessive grinding.
KNbOP2O7, which has a niobyl (NbO) group, does not shown efficient low voltage
cathodoluminescence efficiency.
Compared to Y2SiO5: Ce and Y1.75Sco.25Si05: Ce,
YNbO4 and GdNbO4: Bi shows similar efficiency and YNbO4: Bi shows higher
efficiency. However the efficiency of filtered blue needs to be determined to examine
their property as blue FED phosphors.
68
References:
1.
K. H. Bulter, Fluorescent Lamp Phosphors: Technology and Theory, Pennsylvania
State University Press (1980).
2. L.N. Kinzhibalo, V. K. Trunov, A. A. Evdokimov, and V. G. Krongauz, Sov. Phys.
Crystallogr. 27(1), 22-25 (1982).
3.
G. Blasse and A. Bril, J. Luminescence Vol. 3, 109(1970).
4. R. C. Ropp, Luminescence and the solid state, Elsevier (1991).
5.
D. A. Grisafe and C. W. Fritsch, J. Solid Statae Chem. Vol. 17, 313-318(1976).
6.
A. Vecht, D. Charlesworth, D. W. Smith, SID 97 Digest, 588-590(1997).
7.
J. Livage, M. Henry and C. Sanchez, Prog. Solid State Chem. Vol. 18, 259341(1998).
8. J.
D. Ramsay, "Sol-gel Processing" in"Cotrolled particles, droplet and bubble
formation", Ed. D. J. Wedlock, Butterworth-Heinemann, 1994.
9.
R. C. Mehrotra Structure and Bonding, Vol. 77 2-35(1992).
10. C. P. Ballard and A. J. Fanelli, "Sol-gel chemistry" in "Chemistry of Advanced
Materials", Ed. by C. N. R. Rao, Blackwell Sci. Pub. (1993).
11. C. Alquier, M. T. Vndenborre and M. Henry, J. Non-Crystalline Solids, Vol 79, 383395(1986).
12. D. White, Trans. Brit. Ceram. Soc. Vol. 63, 301(1964).
13. G. Blasse, Topics in current chemistry, Vol. 17, 2-24.
14. G. Blasse, Structure and bonding, Vol. 76, 154-185(1991).
15. Datta, J. Electrochem. Soc., Vol. 114, 1137(1967).
16. S. Oyetola, A. Verbaere, Y. Piffard and M. Tournoux, Eur. J. Solid State Inorg.
Chem., t.25, 259(1988).
17. G. Blasse, G. J. Dirksen and M. F. Hazenkamp, Eur. J. Solid State Inorg. Chem., t.26,
497(1989).
69
CHAPTER 5
LUMINESCENCE OF SrS:Cu
Dong Li and Douglas A. Keszler*
Department of Chemistry and Center for Advanced Materials Research
Gilbert Hall 153, Oregon State University, Corvallis, OR 97331-4003
70
Abstract
Results of steady-state photoemission and excitation studies, decay-time
measurements, and thermal-quenching observations are presented for SrS:Cu powder
phosphors containing combinations of 1+ and 3+ codopants. By adjusting the relative
concentrations of these ions, selected emission colors can be produced. The shapes and
positions of the emission bands depend primarily on the surroundings of the individual
Cu atoms. As a result, the luminescence results are explained on the basis of isolated Cu
coordination environments and the charge-compensation process.
Introduction
Luminescence in alkaline-earth sulfides has been an active area of study for more
than 70 years [1]. They are of considerable interest for application in electroluminescent
(EL) displays, where their performance characteristics have been largely unsurpassed by
other materials. Among these binary hosts, SrS is the most widely studied example.
SrS:Ce was first investigated for use in EL displays in 1984 [2], and it is still of interest
for use in color by white devices. A highly efficient, blue-emitting EL phosphor, SrS:Cu,
was reported in 1997 [3]. This was a very important breakthrough in the development of
EL devices, as no high-efficiency blue EL phosphor was previously known.
To date, investigations on SrS:Cu have led to considerable uncertainty about the
nature of the luminescent centers [4-10]. Three different emission bands have been
observed: strong bands near 470 and 520 nm and a weaker feature near 545 nm. The
appearance of a specific band depends on many factors, such as concentration of Cu, the
presence of codopants, preparative conditions, and temperature. To achieve improved
phosphor performance, it is important to understand and to control the emission in this
host. In this contribution, we present results on SrS:Cu codoped samples that provide
new information on the luminescence process as well as a means to control the
luminescence color.
71
Fundamental properties, such as band gap and defects, have considerable effects
on the performance of alkaline-earth sulfides in devices. These sulfides adopt the face-
centered cubic NaCl structure, and they are characterized by a high degree of ionicity.
The experimental band gaps, as determined by optical reflection, increase in the order
BaS < SrS < CaS < MgS with the value of SrS at about 4.6 eV. On the basis of bandstructure calculations, CaS, SrS and BaS are all predicted to be direct band-gap materials,
while the situation for MgS is still unresolved [13].
Computational studies by Mishra
and co-workers [11] reveal that the upper levels of the valance band are S 3p in character,
while the lower portion of the broad conduction band has S 3d character. The phonon
frequency of SrS is approximately 200 cm-I.
Intrinsic point defects in these sulfides have been studied by using electron-spin
resonance. Two defect centers are generated by S vacancies: the F center, comprising a
sulfur vacancy with two electrons, and the F+ center, comprising a S vacancy with one
electron [12, 13]. Rennie and co-workers suggest an equilibrium exists between the F
and F+ centers; where the concentration of r centers dominates at low temperatures, and
F centers dominate at high temperatures [12].
As the temperature changes, the
equilibrium between the types of defect shifts, causing wavelength shifts in the
absorption bands and producing a difference between absorption and excitation bands.
The S vacancies also contribute to the production of color in the samples. The colors are
red for CaS, and pink to black, depending on the defect density, in the case of SrS.
Intrinsic emission has been observed in the undoped sulfides; the details of a
given spectrum depend on preparative conditions. Two peaks are observed in SrS, a
dominant feature at 380nm and a weaker broad band at 460 nm [12]. As Ca is introduced
(Sri_CaS, x>0), the intermediate compositions show the longer wavelength emission
shift between 380 to 540 nm [14].
Brightwell and co-workers have examined the
excitation spectra for these materials [14]. For SrS, two peaks are observed at 270 and
370 nm.
The excitation energies are lower than the band-gap energy, indicating
excitation occurs from the valance band to a localized defect-energy level associated with
anion vacancies. As Ca is substituted for Sr, additional excitation peaks appear below the
band-edge feature. Much of the work on these aspects of the defect chemistry has been
summarized in a review by Pandey and Sivaraman [13].
Additionally, they review the
72
spectroscopic studies for the many different activators that have been doped into these
hosts.
Experimental
Th Cu-doped SrS phosphors were prepared from a mixture of SrCO3 (AESAR,
99.99%) and prescribed quantities of activator, co-dopant, or both. Samples were heated
in a graphite boat at 1223 K for 1 hour under a flowing stream of H2S (g). The H2S (g)
flow was changed to a static 4%H2/96%Ar atmosphere, and the furnace is cooled to room
temperature at approximately 300 K/h.
Three Cu reagents were examined - Cu2O
(CERAC 99.9%); CuCl (Johnson Matthey, puratronic); and CuSO4.5H20 (Aldrich
99.999%). No discernable differences among the emission spectra were observed on the
basis of the Cu source.
CuSO4.5H20.
The results described herein were all obtained by using
CuSO4.5H20 and any co-activators [NaNO3 (Aldrich, 99.995%),
Y(NO3)3.6H20 (AESAR 99.9%) and La(NO3)3.6H20 (Johnson Matthey, 99.9%] were
dissolved in H20(1) and subsequently ground into the SrCO3.
Powder X-ray work was done with
a
Siemens D-5000 diffractometer.
Photoluminescence measurements were obtained with equipment that has been described
elsewhere [15].
Results
The room-temperature emission spectra of SrS doped with Cu; Cu, Na; and Cu,
Y; are shown in Figure 5.1. The sample containing 0.05% Cu exhibits a broad emission
band peaking near 465 nm. On increasing the Cu concentration to 0.2% and 1.0%, the
peak in the emission band shifts dramatically to 535 nm with a shoulder persisting in the
blue portion of the spectrum for the 0.2% sample. On codoping the 0.2% Cu sample with
0.2% Y, the emission peak shifts back to 465 nm. Several codopants - Sc3+, 1,a3+, Gd3+,
Ce3+, Bi3+, Zr4+, Hf4+, 1113+, A13±, Ga3+
- produce a similar effect, but the strongest blue
73
shift has been observed with Y. Emission spectra for samples at 77 K are given in Figure
5.2; excitation wavelengths have been set to 310 and 350 nm. When excited by 310nm,
all samples show emission shoulder in the blue region. The Y codoped sample shows the
bluest component. When excited by longer wavelengths, the emission peaks shift to the
green.
0.05%Cu
0.9
0.2%Cu
0.8
an
1%Cu
0.7
0.2 %Cu0.2 %Y
E. 0.6
%,,
0.2 %Cu0.2 %Na
0.5
0.4
P..
Z 0.3
0.2
0.1
0 -1
390
420
450
480
510
540
Wavelength(nm)
570
600
Figure 5.1: Emission of SrS doped samples, Aexe=310nm.
630
74
0.2%CuY, 310nm
0.2%Cu, 310nm
0.2%CuY, 350nm
74 0.8
0.2%Cu, 350nm
=
on
1%Cu, 310nm
a
0.6
1%Cu, 350nm
0.4
0
0.2
0
390
420
450
480
510
540
570
600
630
Wavelength(nm)
Figure 5.2: Effect of different excitation at 77 K.
The effect of temperature on the emission of SrS:0.2%Cu and SrS:0.2%Cu,Y is
demonstrated in Figures 5.3 and 5.4, respectively.
For the pure Cu sample, the peak
emission shifts to longer wavelengths with increasing temperature. The shoulder in the
blue becomes stronger with increasing temperature up to approximately 200 K; above
this temperature its relative contribution decreases. For the Y codoped sample (cf. Figure
5.4,) the peak emission shifts to shorter wavelengths with increasing temperature. A thin
film obtained from Planar American has also been studied to compare with our results
from powder samples; the spectrum is shown in Figure 5.5. The pure copper sample with
Ga2S3 as flux exhibits a temperature-dependent shift that is similar to that of the Y
codoped sample. For the codopants
Na+, and IC', the maximum emission is always
in the green portion of the spectrum, independent of Cu concentration; the results are
depicted in Figure 5.6. Chromaticities for selected samples are collected in Table 5.1.
75
1
0.9
0.8
0.7
v)
0.6
E-1
rt 0.5c.)
tg-1
04
0:1
I 0.3
0.2
0.1
0
390
420
450
480
510
540
570
600
630
Wavelength(nm)
Figure 5.3: Effect of temperature on emission of SrS:0.2%Cu, A.,=310nm
1
0.9
77 K
130 K
180 K
230 K
298 K
0.8
°"',=) 0.5
NI
NI
0.4
Z0 0 3
0.2
0.1
0
390
420
450
480
510
540
570
600
630
Wavelength(nm)
Figure 5.4: Effect of temperature on emission of SrS:0.2%Cu,Y, ile,=310nm.
76
Table 5.1: Chromaticities of selected samples
x
y
0.05%Cu
0.152
0.202
0.2%Cu
0.230
0.469
1%Cu
0.247
0.550
1%Cu, 1%La
0.171
0.233
0.2%Cu, 0.2%Y
0.146
0.173
0.2%Cu, 0.2%Na
0.306
0.587
1
0.9
80 K
130 K
0.8
180 K
0.7-
4
230 K
0.6
298 K
% 0.5
.N
7s.
0.4
E
2 0.30.2
0.1
or
390
420
450
480
510
540
570
600
630
Wavelength(nm)
Figure 5.5: Effect of temperature on emission of SrS:Cu(Ga2S3) film (N8-018),
/lexc =310nm.
77
1
0.2%Cu
0.4%Cu
0.6%Cu
0.9
0.8
T't
.,To 0.7
cn
H 0.6
("1"., 0.5
N
;7; 0.4
;@..
0 0.3
Z
0.2
0.1
0 --,
410
460
510
560
610
660
Wavelength(nm)
Figure 5.6: Emission of samples with x%Cu, 1%Na, A,=310nm.
Excitation
The temperature dependencies of the excitation spectra of SrS:0.2%Cu and
SrS:0.2%Cu, 0.2%Y are illustrated in Figures 5.7 and 5.8, respectively. In the sample
containing Cu only, four bands may be identified, centered near 260, 280, 310, and 350
nm. A similar set of bands is identifiable in the Y codoped sample, but the band near 355
nm is largely absent. The 310-nm band is much more pronunced for the Y codoped
sample. The longest wavelength band becomes much stronger on addition of Li (Figure
5.9) and at high Cu concentrations (Figure 5.10). Addition of a 3+ ion, e.g. La3+, greatly
suppresses this band, even for Cu concentrations as high as 5 at%.
78
1
77 K
0.9
130 K
0.8
180 K
4at 0.7
230 K
E. 0.6
298 K
ft 0.5
.N
74 0.4
E
0.3
0.2
0.1
0
260
280
300
320
340
360
380
Wavelength (nm)
Figure 5.7: Effect of temperature on excitation of SrS:0.2%Cu, A.,,=525nm
at both 298K and 230K, Aem=520nm at 180K, and )m= 515nm. at 78K.
400
79
1
77 K
0.9
130 K
0.8
180 K
a 0.7
230 K
E. 0.6
298 K
0.5
lV
"731 0.4
2 0.3
0.2
0.1
0
260
280
300
320
340
360
380
Wavelength (nn)
Figure 5.8: Effect of temperature on excitation of SrS:0.2%Cu, 0.2%Y;
2e,=-500nm at 77K; 495nm at 130K; 485nm at 180K; 475nm at 230K;
and 470nm at 298K.
400
80
250
270
290
310
330
350
370
390
410
Wavelength (nm)
Figure 5.9: Effect of codopants on excitation of SrS:0.2% Cu. A.,,n=525nm
for 0.2%Cu sample. Aem=475nm for 0.2%Cu,0.2%Y sample. Aeni=530nm
for 0.2%Cu, 0.2%Li sample.
430
81
1%Cu
5%Cu, 5%La
0
260
280
300
320
340
360
380
400
420
440
Wavelength (nm)
Figure 5.10: Effect of La3+ on the emission of SrS:Cu, 2,=310mn
Photoluminescence decay time
Decay curves for SrS:0.2%Cu and SrS:0.2%Cu, 0.2%Y are given in Figures 5.11
and 5.12, respectively. These data were obtained by using a 10-ns laser pulse at 266 nm
and measuring the light in the entire emission band by using an appropriate set of filters.
The decay curve for the Cu sample has been fit with a single exponential, corresponding
to a decary time of 7
The data for the Y-codoped sample has been modeled with two
separate exponential fits, corresponding to times of 0.1 and 2 p.s. Decay curves obtained
by monitoring only the green emission at 550 nm and only the blue emission at 450 nm
are shown in Figures 5.13 and 5.14, respectively. A consistent decay time near 6 1.1S is
observed for the green emission, while a time near 0.1 .ts is observed for the blue
emission. Decay times are summarized in Table 5.2.
82
0.0E+00
-2.5
5.0E-06
1.0E-05
2.0E-05
1.5E-05
2.5E-05
3.0E-05
-3
-3.5
-4
y = -136360x - 3.3154
R2= 0.9957
Decay Time: 7.3E-6s
Time(s)
Figure 5.11: Decay curve for SrS:0.2%C.'u, band emission.
0.0E+00
-4
5.0E-07
1.5E-06
1.0E -06
2.0E-06
1--
y = -8E+06x - 3.316
-4.5
R2= 0.9408
Decay Time=1.2E-7s
....
y = -563226x - 5.4882
R2 = 0.9681
decay Time = 1.8E-6s
-6
-6.5 -1
-7
Time(s)
Figure 5.12: Decay curve for S'rS:0.2%Cu, 0.2%Y, band emission.
2.5E-06
83
0.2%Na
1%Na
A 0.2%Na0.1%Y
x 0.1%Na0.1%Y
y = -156668x - 2.3798
R2 = 0.9795
Decay time = 6.4E-6s
Time(s)
Figure 5.13: Decay curves for SrS:0.2%Cu and with codopants, 2,, =550nm.
0.0E+00
-2
1.0E-07
2.0E-07
3.0E-07
4.0E-07
5.0E-07
A
X
4- Due to laser pulse
o 0.1%Y0.1%na
0.15%Y0.1%Na
A 0.3%Y0.1%Na
x 0.5%Y0.1%Na
0.2%Y
-2.5
0
0
a
a
0
0
0
a
0
O
0
0
a
0
a
0
0
0
A
-5
-5.5
y = -2E+06x - 4.2065
R2= 0.9534
Decay Time = 5E-7s
-6
Time(s)
Figure 5.14: Decay curves for SrS:0.2%Cu with codopants Ae, = 450nm.
0
84
Table 5.2: Decay times of SrS: 0.2%Cu, x%A(,us)
No Monochromator
At 450nm
At 550nm
-
-
0.1%Na0.1%Y
7.3, 3.6
6.4
6.4
6.4
6.4
0.1%Na0.15%Y
--6.4
0.2%Cu only
Pt
CD
CD
1%Na
0.2%Na
0.2%Na0.1%Y
to
CD
0.1%Na0.5%Y
0.2 %Y
-
-
6.4
6.4
6.4
6.4
--6.4
0.5
0.5
0.5
-
--0.5
-
1.8, 0.12
0.5
-
-
Thermal Quenching
Luminescence quenching has been measured by monitoring the photomultiplier
signal at a fixed emission wavelength as well as by monitoring the signal for the
integrated band. The curves generated by selecting emission wavelengths of 460 and 512
nm are shown in Figure 5.15 for the samples SrS:0.2%Cu and SrS:0.2%Cu,0.2%Y. The
blue emission (460 nm) increases in the range 80 to 120 K and then begins a gradual
descent up to room temperature. A gradual decrease in the green emission is observed
with increasing temperature for each sample, but the Y codoped sample exhibits a more
dramatic fall in emission. Quenching curves that have been obtained by monitoring the
integrated emission band are depicted in Figure 5.16.
The curves have been determined
by using both a fixed excitation wavelength (310 nm) as well as broad-band excitation.
With the exception of the Y codoped sample under 310-nm excitation, no decrease in
emission signal is observed with increasing temperature. The anomalies in these curves
near 200 K are unknown. Quenching curves for selected thin films provided by Planar
America and those for powders of SrS:0.2%Cu and SrS:0.2%Cu, 0.2%Y are given in
Figure 5.17. Quenching curves of both thin films are obtained by monitoring the
85
integrated emission band. The fixed excitation wavelengths are 310nm for SrS:Cu film
and 295nm for SrS:Cu, Ag film.
1.250
-4- 0.2%Cu(512nm)
--N-- 0.2%Cu(460nm)
1.000
0.2%CuY(512nm)
0.2%CuY(460nm)
s.
H 0.750
.--4-4
N
0.500
O
z
0.250
0.000
80
120
160
200
240
280
Temperature(K)
Figure 5.15. Quenching of SrS:0.2%Cu and SrS:0.2%Cu, 0.2%Y by
monitoring specific wavelengths noted in legend, itexc = 310 nm.
320
86
1.25
1
4,.
7's
E-1
- .-.-
0.75
4- ± + ± + + ± ± +
.1:so
.4'
,+-+
0.5
X 0.2%Cu, Ex=310nm
0
z
-U- 0.2%Cu,Y, Ex=310nm
0.25
+ 0.2%Cu, BEX
0.2%CU,Y, BEX
0
50
100
150
200
300
250
350
Temperature(K)
Figure 5.16: Quenching of SrS0.2%Cu and SrS:0.2%CuY by monitoring integrated
emission signal with excitation wavelength noted in legend (BEX = band excitation.)
1.25
E-4 0.75
r?.
1
0.5-
SrS: Cu(Ga2S3) Film
-s-- SrS: CuAg(Ga2S3) Film
N..
+-\\
-X- SrS:0.2%Cu Powder
0.25
SrS:0.2%CuY Powder
0
50
100
150
200
250
300
350
Temperature(K)
Figure 5.17: Comparison of thermal quenching of thin film and powder phosphors.
87
Concentration Quenching
Concentration quenching curves of SrS: x%Cu, x%La and SrS: x%Cu, 1%Na are
shown in Figure 5.18. Samples were excited with a low-pressure Hg lamp at 254 nm,
and the signal of the entire emission band was monitored at the PMT. The blue emission
of the La codoped sample attains a maximum near 0.2% Cu, 0.2% La, and the green
emission of the Na codoped sample have a maximum at about 0.1% Cu (results provided
by Ben Clark).
1.2
SrS: x%Cu, x%La
U -SrS: x%Cu, 1%Na
0.8
.% 0.6
0.4
0.2
0
0
0.2
0.4
0.6
0.8
x%
Figure 5.18: Concentration quenching of SrS:x%Cu,x%La and SrS:x%Cu,1%Na.
88
Discussion
The luminescence of Cu+ in ZnS-type phosphors is widely appreciated, studied,
and utilized. It is incorporated as an acceptor along with a donor impurity such as F or
A13+ to provide charge compensation. The emission is associated with an electron
transition from a level very close to the edge of conduction band (F- or Al3+ level) to a Cu
level that is situated above the edge of the valence band. Such a model, however, is not
applicable to Cu+ doped into alkaline-earth sulfides [7, 9].
Although ligand-metal charge transfer likely plays a role in the excitation and
emission processes of SrS:Cu [16], it is common to treat the problem in the context of
the crystal field-model with NaF:Cu+ [17, 18] serving as the model system. In this way,
absorption and emission are attributed to transitions of the type 3d94s1+33d10. The
energy levels of NaF:Cu+ are sketched in Figure 5.19 [17].
3Eg
emission
T2g
Err
26cm"
Tlg
is
1
Free Ion
Crystal Field
I
1 Alg
Jahn-Teller
Figure 5.19: Cu+ energy levels in alkali halide host NaF.
89
The two free ion levels 3D and ID are split by the octahedral field into Eg and T2g
components. The Eg states are further split as a result of the Jahn-Teller effect; the Jahn-
Teller energy is on the order of 2000 cm' [19]. These levels are further split by spinorbit coupling, where the Tig and T2g components arise from the 3Eg state. Emission
occurs from the S-0 component of the 3Eg state. Since the T2g (3Eg) can also mix with the
1T2g state by spin-orbit interaction, it acquires a singlet character (1-2%). Moreover,
there is only a 26 cm'' energy difference between T2g(3Eg) and Tig(3Eg). For temperatures
higher than 50 K, the radiative rate from the higher level T2g(3Eg) is much faster than that
from the level Tig(3Eg) [10, 17]. To assign the two bands in the emission spectrum of
NaF:Cu at room temperature, Moine and Pedrini propose they arise from a dynamic JahnTeller effect splitting the 3Eg state [18].
Differing explanations have been advanced to explain emission phenomena for
SrS:Cu.
Laud and Kulkarni [8] have assigned the 471-nm blue emission to Cu in
interstitial sites and the green emission to Cu in simple substitutional sites. They further
suggest that a transition from the conduction band to the Cu center gives rise to 548-nm
emission, while a transition from the Cu center to the valance band gives the 520-nm
emission. Yamashita and co-workers have proposed a different model [9]. They have
examined the photoluminescence spectra of SrS:Cu with different levels of dopant
concentration at 80 K. A band labeled a peaking at 513 nm is attributed to isolated Cu+
in substituional sites (monomer), a band labeled 13 peaking at 543 nm is attributed to
paired Cu+ ions (dimer), and all longer wavelength emission bands are assigned to highly
aggregated Cu+ centers. They do not, however, relate any of these emission centers to the
blue emission that is observed at room temperature. A clear structural view of the dimer
and aggregated centers has also not been presented.
When studying the activators and co-activators in CaS, Lehmann has proposed
that the luminescent center does not consist only of activator and coactivator ions
replacing ions of the host but involve interstitial and/or vacancies as well [21].
On the basis of the earlier luminescence data and the results presented here, we
propose a simplified model based on the defect chemistry to explain the results.
Monovalent Cu+ ions are assumed to substitute for the divalent Sr2+ ion in the SrS host.
90
There is a deficiency of positive charge that can be compensated by the formation of S
vacancies or the incorporation of other charge-compensating ions:
Cu+ + Na+ + Vs2+ > 2Sr2+
Cu+ +
Sr2+ + S2-
Cu+ + Y3+ -+ 2Sr2+
The latter two cases should lead to one Cu site retaining a coordination number of
the matrix, and the former should lead to one Cu site having a reduced coordination
number.
This difference in coordination number will certainly affect the optical
properties of the doped samples. On the basis of current results, we believe the blue
emission should be associated with emission from Cu+ occupying a Sr substutional site
with retention of an approximate octahedral envinronment, while the green emission
should be associated with the Cu+ ion on a site near a S vacancy, thus having a reduced
coordination number, perhaps five. Solid state NMR results provide some support for
this model. The results show the coordination surroundings are different for the two
emission centers [22]. Singh and co-workers have proposed an alternate mechanism
wherein both Cu+ and Na+ may substitute on a single Ca2+ site (Cu+ + Na+ > Ca2+) in
CaS [20]. Na+ may occupy an interstitial position or a substitutional position together
with Cu+. Considering the sizes of the atoms, the latter situation seems unlikely.
Excitation
The excitation spectra for SrS:0.2%Cu (green emission) and SrS:0.2%Cu,0.2%Y
(blue emission) are quite different. The two bands centering at 260 and 310 nm are
assigned to the transitions lAig >IT2g and lAg > lEg levels, respectively in the literature.
The 260nm band is overlapped with band excitation (-4.6eV, i.e. 270nm). The 310-nm
band is more pronounced for the Y codoped sample. A similar band centered near 289
nm (B band) was observed by Payne and co-workers for NaF: Cu (6 coordinated) [19]; it
is assigned as one of two Jahn-Teller split components.
The results in Figures 5.2 and 5.10 also indicate there is one blue emission center
and one green emission center, and the energy level of blue emission center is higher. It
91
seems the excitation from 260nm to 310 nm feeds both emission centers and 350 nm
excites more green emission.
Emission and Quenching
The transitions in the free Cu+ ion are parity forbidden. They become partially
allowed by electron-phonon interactions.
There is no inversion center for five
coordinated copper. The parity selection rule is completely relaxed. There is always
equilibrium between 5-coordinated and 6-coordinated copper. Co-doped with 3+ and 1+
ions will make one kind of copper center dominated, but could not eliminate another kind
of copper. The green emission dominates at low temperature, because there is little
electron-lattice interaction. When temperature is increased, the blue emission becomes
possible.
At even higher temperature, the barrier B can be passed thermally, green
emission dominant. This process is shown in Figure 5.20. There are few 5-coordinated
coppers in Y copped sample, elevated T can only relax the parity forbidden rule. So the
dominant emission is blue at higher temperature.
The explanation of the thermal quenching is complicated by the shifting of
emission peak at different temperature. The alternation of emission and excitation at
different temperature must also be considered. When samples are excited by 310 nm
light and monitored at either 512 nm (Figure 5.15) or the integrated emission band
(Figure 5.16), the dramatic fall of Y co-dopped sample is due to two factors.
<1>
Emission peak shift out of green region into blue (Figure 5.4). <2> Continued decrease
of efficiency at 310 nm excitation (Figure 5.8). The decreased efficiency of 310 nm
excitation may have relation with the alternation of band-gap (I guess smaller) due to
doping of acceptor (Cut) and donor (Y3+). When excited at 310 nm and monitored at 460
nm, the initial increase is mainly due to the shift of emission peak into the blue region
(Figures 5.3 and 5.4). The decrease after that is due to the emission peak shift back to
green for Cu sample and decrease of efficiency for 310 nm excitation for the Cu, Y
sample. When the samples are excited by the whole excitation band, which is similar to
the EL excitation with different impact energy, neither sample exhibits a significant
92
decrease in the emission signal with increasing temperature. This indicates the thermal
quenching for these phosphors is not so significant. The anomalies near 200 K maybe
due to the barrier between green and blue emission center and other unknown factor
which need further study.
Ground
State
Figure 5.20. Schematic configuration coordinate diagram for SrS:Cu with both green
and blue emission centers.
93
Decay Time (tentative explanation)
According to the relation T--X2, we expect the blue emission has faster decay time
than the green emission if they come from the similar emission center. The calculated
decay time for blue emission should be about 4 1.15 when the decay time for green
emission is 6.4 ps. However, the decay time we determined is one order faster than that.
In the energy diagram of NaF:Cu we discussed previously, the T2g (3Eg) can mix with the
'T2g state by spin-orbit interaction and acquires a singlet character (1-2%). If this is the
case, the decay time of blue should be much faster due to the relaxation of spin-forbidden
rule.
Concentration Quenching
The maximum intensity for the blue sample occurred at 0.2%. This result is in
line with study of other groups [3,4], although it is lower than 0.5at% found by Laud and
co-worker [9].
Na codoping changes most Cu-defect-Cu green emission centers to Cu-
defect-Na center, which prevents the energy transfer between two close Cu activators.
This provides advantage compared to the sample doped with pure Cu.
Based on above results, we designed one phosphor system: SrS:0.2%Cu,x%Na,
y%Y. By changing the amount of Na, Y, or both, the emission color can be tuned
between blue and green (Figure 5.21.) Some details on the characteristics of this system
and that of a green EL device based on SrS:0.2%Cu,1%K has been submitted for
publication [23].
94
0.2% Y
0.3% Y, 0.1% Na
0.15% Y, 0.1% Na
Cu
0.8
0.2% Na
=
to
. 0.6
Ern
a.
-ci
N
0.4
'') z
0.2
0
390
440
490
540
590
640
Wavelgenth (nm)
Figure 5.21. Photoluminescence of SrS:0.2%Cu,x%Na,y%Y.
Conclusion
A simple model is proposed to explain the luminescence colors that are observed
in SrS:Cu phosphors. The Cu+ cation occupying a substitutional site with approximate 6
coordination emits blue light, while the green emission is associated with the 5
coordinated Cu+ near a S vacancy.
Because the spectral-energy distribution in these
systems is known to be independent of excitation source, the results should be expected
to be reproduced in the function of electroluminescent devices.
95
References:
1.
P. Lenard, F. Schmidt, and R. Tomascheck, Handbook of Experimental Physics,
Vol. 23, Akademie Verlagsges, leipzig, 1928.
2.
W. A. Barrow, R. E. Covert and C. N. King, Digest of the 1984 SID Intl Display
Symposium, P249, (1984).
3.
S. S Sun, E. Dickey, J. Kane and N. Yocom, Proc. 17th Int Display Research
Conf,, Morreale, J., Ed.; Society for Information Display: Toronto, p201, 1997.
4.
M. Avinor, A. J. Carmi and Weinberger, Chem. Phys., Vol. 35, 1978(1967).
5.
D. Sharma and Amarsingh, Indian J. Pure Appl. Phys., Vol. 9, 810(1971).
6.
P. S. Kanari, Indian J. Pure Appl. Phys., Vol. 9, 1060(1971).
7.
W. Lehmann, J. Electrochem. Soc., Vol 117, 1389(1970).
8.
B. B. Laud and V. W. Kulkarni, J. Phys. Chem. Solid, Vol. 39, 555(1978).
9.
N. Yamashita, Jpn. J. Applied Phys. Vol. 30, 3335(1991).
10.
N. Yamashita, K. Ebisumori, K. Nakamura,
3846(1993).
11.
K. C. Mishra, K. H. Johnson, P. C. Schmidt, Mater. Sci. and Eng., B18,
Jpn. J. Applied Phys. Vol 32,
214(1993).
12.
J. Rennie, K. Takeuchi, Y. Kaneko and T. Koda, J. Lumin., Vol. 48&49,
787(1991).
13.
R. Pandey and S. Sivaraman, J. Phys. Chem. Solids, Vol. 52, 211(1991).
14.
J. W. Brightwell, B. Ray, P. Srphton and I. Viney, J. Cryst. Growth, Vol. 86,
634(1988).
15.
A. Akella and D. A. Keszler, Mater. Res.Bull., Vol. 30, 105-111(1995).
16.
G. Blasse, Structure and Bonding, Vol. 76, P17(1991).
17.
S. A. Payne, R. H. Austin, D. S. Mcglure, Phys. Rev. B, Vol. 29, 32(1984).
18.
B. Moine and C. Pedrini, Phys. Rev. B, Vol. 30, 992(1984).
96
19.
S. A. Payne, A. B. Goldberg, D. S. McClure, J. Chem. Phys., Vol.78, 3688(1983).
20.
N. Singh, G. L. Marwaha, V. K. Mathur, Phys. Stat. Sol. (a), Vol. 66, 761(1981).
21.
W. Lehmann, J. Lumin. Vol. 5, 87(1972).
22.
J. Janesky, W. Warren, D. Li, D. A. Keszler, unpublished results.
23.
D. Li, B.Clark, D. A. Keszler, P. Kier, and J. F. Wager, submitted for publish.
97
CHAPTER 6
PHOTOLUMINESCENCE OF (Zn,Ga)S:Mn
Dong Li and Douglas A. Keszler
Department of Chemistry and Center for Advanced Materials Research
Gilbert Hall 153, Oregon State University, Corvallis, OR 97331-4003
98
Abstract
(Zn,Ga)S:Mn
solid-solution phosphors have
been synthesized by
high-
temperature solid-state techniques. The emission colors range from yellow to deep red,
depending on the Ga concentration. Photoluminescence emission and excitation spectra
have been obtained at both liquid nitrogen and room temperatures; luminescent lifetimes
have also been obtained from laser-excitation measurements. Analysis of the excitation
spectra leads to Racah parameters B=0.0741 and C=0.410 eV, the crystal field parameter
Dq = 0.08 eV, and the covalency parameter s = 0.043.
On the basis of these
measurements, the emission centers are assigned as tetrahedrally coordinated Mn2+
cations.
Introduction
Although some progress has been achieved in the commercialization of full color
thin-film electroluminescnet (EL) displays, the amber monochrome display operating on
the basis of ZnS:Mn remains the most widely utilized EL device [1]. Currently the most
effective way to produce a full-color EL display is by the color-by-white method, which
comprises a yellow-emitting ZnS:Mn and blue-emitting SrS:Cu stack [2].
Improved
performance of such a stack could be achieved by utilizing a phosphor that exhibits a
luminescence color that is slightly red-shifted relative to that of ZnS:Mn. Of course, the
optimum performance of a full-color EL display would be achieved with the generation
of an efficient red, green, blue El-active phosphor set. In an earlier contribution [3], we
described the existence of the phosphor (Zn,Ga)S:Mn, which provides the opportunity to
realize both small red shifts in emission for application in the color-by-white method and
a true, chromatically red phosphor as one component of the tricolor set. In this report, we
describe some results on the photoluminescence behavior of this series of materials.
99
Experimental
Materials were prepared by high-temperature solid-state reactions. Stoichiometric
quantities of ZnS (CERAC, 99.99%), Ga2S3 (AESAR, 99.99%), and MnS (AESAR
99.9%) were mixed according to the following equations.
(0.99-1.5x) ZnS + 0.5x Ga2S3 +0. 01MnS
Zn(0.99-3 x/2)GaxMn0.13 IS
Five weight percent Li2CO3 (CERAC, 99.9%) was added as flux. The reagents were
ground together under hexane, placed in a graphite boat, and then inserted into a fused
silica tube. The sample was fired at 850°C in a CS2/N2 stream for 15 hours. ZnS:Mn and
ZnGa2S4:Mn were prepared under the same conditions without the Li2CO3 flux.
X-ray diffraction data were recorded on a Siemens D5000 X-ray powder
diffractometer by using Cu Ka radiation. Unit-cell parameters were determined from X-
ray measurements by collecting data at 0.02°/step and refining with the computer
program LAPOD [4].
Luminescence spectra were measured on a right-angle spectrometer. A Cary
model-15 prism monochromator was used to select excitation wavelengths, and the
sample emission was passed through an Oriel 22500 1/8-m monochromator followed by
detection with a Hamamatsu R636 photomultiplier tube (PMT.) Current from the PMT
was measured with a Keithley 486 picoammeter and transferred to a personal computer
for analysis. Emission spectra were corrected by using a tungsten lamp from Eppley
Laboratories, Inc. Excitation spectra were corrected by using either sodium salicylate or
rhodamine B as quantum counters.
Low-temperature spectra were obtained by using the equipment described above
in conjunction with a Cryo Industries flow cryostat equipped with a Conductus
temperature controller. Lifetime data were acquired by using a Nd:YAG laser having a
pulse width (FWHM) of 10 ns; a frequency mixing crystal was used to provide excitation
at 355 nm. The emission signal was detected with a PMT connected to a Tektronix digital
oscilloscope (model TDS 350) that was interfaced to a PC.
100
Results and Discussion
Lattice Constants
The two common forms of ZnS (Figure 6.1) are the cubic, low-temperature form
known as sphalerite and the hexagonal, high-temperature form known as wurtzite; the
phase transformation occurs at 1020°C [5].
The transition temperature can be
considerably affected by the addition of impurities such as Mn, Cu, Cl, and Dy [6].
Consistent with the cubic symmetry, the sphalerite form exhibits close packed layers of S
atoms that stack in the sequence ABCABC.... Zn atoms occupy tetrahedral sites within
this matrix, producing a tetrahedral coordination of S by Zn. For wurtzite, the stacking
sequence is ABABAB... with similar formation of tetrahedral environments for both S
and Zn. Many other S stacking sequences are known, giving rise to a number of ordered
polytypes such as 4H, 10H, 9R, and variations thereon. For our purposes, the mixing of
these sequences and their associated stacking faults generate nonequivalent sites for
dopant atoms, contributing to inhomogeneous broadening of excitation and emission
bands. The influence of these structural details on the optical properties of ions such as
Mn2+ are of much interest but remain largely unresolved [7, 8].
ZnGa2S4 crystallizes in the tetragonal system [9, 10] (Figure 6.1). The structure
may be viewed as a superstructure of sphalerite with an ordered array of S vacancies.
However Zn and Ga may be ordered [11], partially ordered [9] or disordered [12]. As in
sphalerite, the cations are coordinated by four S atoms, but the coordination number of
the S atoms drops from four to three.
ZnS:1%Mn crystallized in sphalerite structure at the preparing condition.
However the materials in the solid-solution series Zni-3,e2GaxS: 1 at% Mn (0 < x < 0.4)
formed in the wurtzite structure, even at 850°C. A similar effect is observed when ZnS is
doped with Dy3+ [6]. It appears that the creation of Zn vacancies facilitates the formation
of the wurtzite form.
101
ZnS-Cubic
ZnGa2S4
Zn
0
ZnS-Hexgonal
Figure 6.1. Structure of ZnS and ZnGa2S4.
102
The diffraction patterns show clearly that the solid solution was formed as
hexagonal structure until x approach 0.4. At x=0.4, ZnGa2S4 phase begins to appear.
Unit-cell parameters for selected members of the series Zn(l-3x/2)Ga,S: 1 at% Mn are
summarized in Table 6.1 and Figure 6.2. On the basis of the crystal radii for Ga3+ (r =
0.61 A) [13] and Zn2+ (r = 0.74 A) [13], a volume contraction is expected with increasing
Ga concentration. On the other hand, a slight volume expansion might be expected from
the associated production of Zn vacancies. By considering the S number densities in ZnS
(2.5 x 1022 cm 3) and ZnGa2S4 (2.8 x 1022 cm-3) (both calculations are based on the
volume of the unit cell), however, a volume decrease should be expected, and this is the
observed trend (Figure 1). The slope of the curve in the region 0 < x < 0.1, however, is
smaller than the slope in the region 0.1 < x < 0.35. This may arise from the incorporation
of u+ from the flux as a charge-compensating defect. The radius of Li+ (r = 0.73 A) [13]
is approximately the same as that of Zn2I-. It is possible that the solid solution formed in
the region 0 < x < 0.1 is Zn2.2xLi,,Ga.S, since 5-wt% Li2CO3 in the preparation
corresponds to x = 0.12. When the concentration of Ga corresponds to x > 0.1, there is
insufficient Li to compensate the formation of Zn vacanies, and the volume begins to
decrease more rapidly. A study on the samples prepared in evacuated silica tube and
prepared by using K2CO3 flux is being conducted.
Table 6.1: Lattice constants and lifetime of Zni_3x/2Ga,,S: 1%Mn.
x
a
c
V
Std error
lifetime
0
3.827
6.266
79.479
0.006
0.85ms
Zn0.85Ga0.1S
0.1
3.826
6.260
79.363
0.009
0.75ms
Zn0.7Ga0.2S
0.2
3.813
6.233
78.470
0.013
0.83ms
Zno.55Ga0.3S
0.3
3.796
6.206
77.437
0.011
0.75ms
Zno.4Ga0.4S
0.4
3.789
6.189
77.102
0.006
0.73ms
5.284
10.404
297.1/4
0.963
5.292
10.389
290.89/4
0.029
ZnS
ZnGa2S4
0.5
2.4ms
103
80
Li flux
79
0
-Na flux
K Flux
78
r.c.- 77
7; 76
75
74
73
72-,
0
0.2
0.1
0.3
0.4
0.5
0.6
X
Figure 6.2: Cell volume of Zn1-3x/2Ga,,S.
Photoluminescence
Emission and Thermal Quenching
Emission spectra (80 K and 300 K) for selected samples in the series Zni.
3,e2GaxS: I at% Mn (0
x
0.5) are given in Figure 6.3. As noted in our report on data
collected at room temperature [2], an increase in the Ga concentration leads to a
systematic red shift in the emission band of Mn.
Positions of emission peaks are
collected in Table 6.2 for both the low- and room-temperature measurements. Except for
the composition Zno.85Ga0AS, the emission peaks do not shift much with changing
temperature. As the temperature is lowered, the bandwidth is reduced.
104
Table 6.2: Emission peaks and chromaticity of ZniataGatS: 1%Mn
Peak Positions (nm)
Chromaticities
FWHM (eV)
300 K
78 K
300 K
78 K
X
Y
ZnS
584
586
0.200
0.16
0.543
0.455
Zn0.85Ga0.1S
594
602
0.238
0.18
0.575
0.425
Zn03Ga02S
618
618
0.274
0.21
0.620
0.379
Zno.55Gao.3S
629
629
0.279
0.22
0.640
0.360
ZnGa2S4
637
629
0.319
0.25
0.656
0.344
0.64
0.36
Y203: Eu3+
500
-
530
560
590
680
650
620
Wavelength (nm)
710
740
770
Figure 6.3a: Emission of Zn(l_3il2)Gd,S: 1%Mn (Room Temperature). Ae= 335nm.
105
1
0.9
0.8
.5, 0.7
cn
E-
0.6
0.5
1..)
0.4
E
z
0.3
0.2
0.1
0
1.6
1.7
1.8
1.9
2
2.1
2.2
2.3
2.4
Photon Energy(eV)
Figure 6.3b: Emission of Zn(1.3il2)GaxS: 1%Mn (78 K). A.,=335nm.
The chromaticity values were collected in Table 6.2 and figure 6.4. By forming
solid solution, the emission color was shifted from amber to red region. The chromaticity
of Zno.55Gaa3S:Mn is equal to that of Y203:Eu.
The thermal quenching was determined by monitoring the emission peak of each
phosphor (figure 6.5). Although there are many defects in (ZnGa)S solid solution, their
thermal quenching is better than that of ZnS.
106
0.8
0.6
y 0.4
0.2
0
0.2
0.4
0.6
0.8
x
Figure 6.4: C. I. E. diagram of Zn(l_3il2)GaAS:1%Mn phosphors.
107
120
170
220
270
320
Temperature (K)
Figure 6.5: Thermal quenching curve of (ZnGa)S:Mn by monitoring the emission peak.
Excitation Spectra
Excitation spectra for (Zn,Ga)S:Mn are given in Figures 6.6 and 6.7. As seen
from the band-gap region of Figure 6.6, addition of 10 at% Ga (x = 0.1) to ZnS:Mn
results in an approximte 0.1-eV red shift in the excitation band.
Incorporation of
additional Ga produces a blue shift in the band, culminating in a shift of approximately
0.05 eV at ZnGa2S4 (x = 0.5). It should be noted that these band-gap shifts do not
correlate with the shifts in the emission spectra that were considered in the previous
section of this report. The data in Figure 6.7 cover the range expected for d-d transitions
for the Mn2+ center. For ZnS:Mn and low Ga loadings, the Mn d-d transitions are well
resolved.
At higher Ga concentrations, these transitions appear to shrink in to the
background, presumably because of a greater inhomogeneous broadening that occurs at
high Ga concentrations. The total excitation spectra of selected Zn(1_3v2)GaxS:Mn are
shown in figure 6.8.
108
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
4
Photon Energy(eV)
Figure 6.6: Band excitation spectra of Zn(l_3,/2)GaxS: 1%Mn at 300 K by
monitoring the emission peak, respectively.
1
0.9
ZnS-78K
0.1Ga-78K
0.2Ga-78K
0.3Ga-78K
0.8
a 0.7
H 0.6
I
0.5
-a
4.)
N
73 0.4
g 0.3
0.2
0.1
0
2.3
2.5
2.9
Photon Energy(eV)
2.7
3.1
3.3
Figure 6.7: Excitation spectra of Zn(1_3,,,2)GaxS: 1%Mn at 78 K by monitoring
the emission peak, respectively
109
1
280
305
330
355
380
405
430
455
480
505
530
Wavelength (nm)
Figure 6.8: Excitation spectra include both band gap and Mn excitation.
Is it Mn emission?
Several possible centers may contribute to the red emission in (Zn,Ga)S:Mn:
ZnS:Ga,
octahedrally coordinated Mn2+,
donor-acceptor pairs, and tetrahedrally
coordinated Mn2+. While previous experience with ZnS:Mn provides a background for
favoring tetrahedrally coordinated Mn2+ as the emission center, it is worthwhile to
consider the other possibilities.
Davis and Nicholls [14] and Georgobiani and co-workers [15] have studied the
luminescence of ZnS:0.01at% Ga, separately. They assign the emission centers as selfactivated A-center. Davis Nicholls observed emission peaks at 1.32, 1.86, and 2.82 eV,
while Georgoiani and co-workers identified peaks at 2.47, 2.64, and 2.86 eV. Because
the emission peaks from the ZnS:Ga samples do not correspond to those observed in the
spectra of (Zn,Ga)S:Mn, we do not believe the emission derives from simple Ga doping.
110
Octahedra lly coordinated Mn2+ is well known as a red emission center. The value
of Dq for Mn2+ in an octahedral environment is much different from that in tetrahedral
surroundings.
As presented in the previous section, the excitation spectra of
(Zn,Ga)S:Mn are consistent with a tetrahedral site for the Mn2+ cation. Soo and co-
workers have also examined by X-ray absorption fine structure analysis the local
environment of Mn in doped nanocrystals of ZnS [16]. The data are consistent with the
coordination of Mn by no more than four S atoms.
Because Li2CO3 flux was used in the preparation of samples, there is possible
donor-acceptor Ga-Li pairs could give rise to luminescence. To examine this possibility,
undoped Zno.7Ga0.2S was prepared by using Li2CO3. Although there is one emission peak
coincide with Mn emission in the emission spectra shown in Figure 6.9, the sample
exhibits no visible luminescence under the UV excitation. So the donor-acceptor
emission may be contribute a little to the luminescence, however it is definitely not
important in red emission we observed.
290
340
390
440
490
540
590
Wavelength (nm)
Figure 6.9: Spectra of pure and Mn doped Zno.7Ga0.2S.
640
690
111
To further confirm that the origin of the red emission is from tetrahedrally
coordinated Mn2+ cations, two additional experiments have been performed. Emission
spectra of Zno.7Gao,2S:1%Mn excited by light above the band gap and within the d
manifold of Mn2+ are given in Figure 6.10. Essentially the same emission spectrum is
observed in each case; hence the emission is independent of the excitation energy, a
result that is consistent with emission from a tetrahedrally coordinated Mn2+ center.
Luminescent lifetimes are summarized in Table 6.3 and Figure 6.11. The results are
entirely consistent with occupation of tetrahedral sites by Mn2+ in (Zn,Ga)S and the
origin of the emission from said centers.
1
0.9
430 nm
Band gap
0.8
to 0.7
7A
vp-
0.6
0.5
N
*-5.1
0.4
0 0.3
Z
0.2
0.1
0
540
560
580
600
620
Wavelength (nm)
640
660
680
Figure 6.10: Effect of band gap excitation (335nm) and Mn intraionic excitation
(430nm) at 78 K for Zn0 7Ga0.2S: 1%A/In.
112
0.0005
0.001
0.0015
0.002
0.0025
0.003
0.0035
. ZnS
0.85ms
A 0.2Ga
0.83ms
0.1 Ga
0.75ms
X 0.3Ga
0.75ms
0.004
AAAAAAAAAA
am
Time(s)
Figure 6.11: Decay time of Zno-3xi2)Gax,S: 1 ALVIn2+ at 300 K
Analysis of excitation band
Although the energetic position of the Mn 3d orbitals with respect to the valence
and conduction bands in ZnS has not been fully resolved, the consensus estimate places
them at about 3 eV below the upper valence band edge [17, 18]. As a result, the emission
arises completely from the d-d transitions on Mn2±. In the ZnS host, the degeneracy of
free-ion Mn2+, 3d5 configuration, is removed by the ligand-field effects and the intraionic electron-electron interactions. In the following discussion, we neglect the deviation
of the ligand field from Td symmetry in the case of the C3v site for the wurtzite structure.
The ground state is
A6 1(6)
N
corresponding to e2t23 configuration, and the quartet excited
states have configurations e3t22, e2t23, e't24. The calculated best fit energy level diagram
for Mn2' in ZnS is shown in Figure 6.12.
113
4T 1(4P)
Band Edge
3.5
S
4E(4D)
4T2(4D)
W
is
4E4A(4G
------ -41.
4G
LTT 2.5
4T2(4G)
4
Ti(4G)
1.5
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
Dq in eV
Figure 6.12. Energy level diagram of Mn2+ in ZnS with B=0.0741eV, C=0.410eV,
and e=0.043. The observed band energies (peak) are marked 0.
Various energy matrices have been used to analyze the absorption/excitation
spectra of Mn2+ [19-21]. Tanabe-Sugano's matrices [20] are the basis for all the other
matrices mentioned.
Because Stout's matrices [19] are developed to explain the
spectrum of Mn in crystal (MnF2) instead of aqueous solution or hydrated crystal, they
are adopted to evaluate the covalency parameter c and the Racah parameters B and C for
quartet excited states in the present work.
114
and 4E(4D) are
(.1
Among the excited quartet levels, the states 4A1(4G), 4E(4G),
independent of the crystal field parameter Dq. The energy of the 4A1(4G) state is given
by (1).
(1)
E[4A1(4G)] = (10B+5C)*(1-e)
The energies of 4E(4G) and 4E(4D) are obtained by solving the secular equation for the
following matrix (2):
13B+5C-s*(4B+2C)-E
2*31/2*B*(1..e)
(2)
E[4E(4G, 4D)]=
24,3132,Kbi 4y
(1-6)
14B+5C-s*(22B+7C)-E
The least square fitting of B, C, c solves B as a imaginary number.
So we choose a
different method that has been used for SrS:Mn [22]. The sum of 4E is given by (3).
E[4E(4G)] ±E[4E(4D.,
)]
27B+10C-s*(26B+9C)
(3)
By inserting the experimentally obtained energies into eqs. (1) and (3), we can obtain B
and C as functions of E. The value of c is determined by the best fit of the calculated
4E(4G) and 4E(4D) energies to the observed peak energies. From the analysis, we obtain c
= 0.0427, B = 0.0741eV, and C = 0.411eV. By using these values, we can solve the
secular equations similar to (2) for all the quartet levels by using Dq as an adjustable
parameter (see Ref. 19 for all the other matrices). Thus we obtain the energy diagram for
Mn2+ ion in ZnS as show in Fig. 12. A Dq value of 0.08 eV was chosen to fit the 4T1(4G)
level.
The observed excitation peak at 2.905 eV does not agree well with the energy
calculated for the 4T2(4D) level. This has also been observed in previous studies [17, 18],
although the peak has been assigned diffrently. Mehra assign the energy level of 4T1(4P)
lower than 4T2(4D) [17], while Gumlich and coworkers assigned 4T1(4P) as the highest
level which is the same as the current work. This misfit, however, is not the major
problem in the analysis. According to the simple point-charge model, Dq should scale
with the Mn(Zn)-S distance as dmr,_s-5. Therefore Dq should decrease for (Zn,Ga)S, if we
assume Zn(Mn)-S distance is proportional to the cell volume. The decrease of Dq should
shift the excitation peaks of 4T to lower energies, but this shift is difficult to discern from
the excitation spectra. Similar observations have been made for the series Zni_Mn,,S (0
115
< x <1) [23]. One explanation is the oversimplification associated with use of Ta site
symmetry rather than C3v.
The crystal-field calculation for octahedral example of
SrS:Mn has given much better agreement with experiment [22]. Covalency may also
play a role in affecting the results. In the (Zn,Ga)S solid solution, Zn vacancies are
formed and some S atoms are three coordinated as in ZnGa2S4. As a result of this
increased covalency, the Racah parameter B should decrease, shifting the d-d bands to
lower energies. If B is changes significantly, we can not apply the energy diagram for
ZnS:Mn to (Zn,Ga)S:Mn.
Because the 4E(4D) excitation peaks are not resolved in
(Zn,Ga)S:Mn, we could not apply the crystal-field calculations to them.
NDO-type quantum-mechanical calculations have also been applied in an attempt
to resolve some of these issues concerning the absorption features of ZnS:Mn [24]. Even
these efforts have provided a poor representation of the experiments.
The differences in
between calculated and observed energies for the two lowest lying 4T1 and 4T2 excited
states are significant. Clearly, the models applied to the analysis of Mn-doped ZnS and
(Zn,Ga)S are inadequate. Continued effort is required in this area to provide a more
detailed understanding.
Detailed results from single-crystal absorption studies of
(Zn,Ga)S:Mn may wll provide important information that can be used to guide the
computational effort.
Summary
Photoluminescence of Mn2+ in the host Zn1-3,e2GaxS has been examined.
The
emission can be conveniently controlled from the amber color of ZnS:Mn to the deep red
of ZnGa2S4:Mn by controlling the Ga concentration. Results of emission, excitation,
lifetime, and crystal-field calculations have been used to assign the emission process to
the Mn2+ centers. The important features leading to the red shift of the emission
wavelength remains an active area of investigation.
116
References:
1.
2.
C. N. King, R. E. Coovert, B. J. Dolinar, S. R. Hix, Advances in Display Technology
IV, SPIE Vol. 457 (1984) 74.
Sey-Shing Sun, E. Dickey, J. Kane, P.N. Yocom, Proc. 17th Int. Display Research
Conf., Morreale, J., Ed.; Society for Information Display: Toronto, p201, 1997.
3.
D. Li and D. A. Keszler, 4th Int. Conf. on Sci. and Tech. of Display Phosphors. Bend,
Oregon, P331 (1998).
4. J. Langford, J. Appl. Cryst., Vol. 4 (1971) 259 and Vol. 6 (1973) 190-196.
5. C-Y Yeh, Z. W. Lu, S. Froyen, A. Zunger, Phys. Rev. B, Vol 46, No.16 (1992)
P10086.
6.
P. R. Bote, P. K. Patil, J. K. Nandgave, P. D. Pawar, Solid State Communication Vol
79, pp5-7 (1991).
7.
H. E. Gumlich, J. Lumin, Vol 23 (1981) 73-99.
8. U.W. Pohl and H. E. Gumlich, Chap 3 in "Solid State luminescnece, theory, materials
and devices". Ed by A.H. Kitai, Chapman&Hall 1993.
9.
C. B. Carpenter, P. Wu, Y-M Gao, A. Wold, Mat. Res. Bull., Vol 24 P1077 (1989).
10. H-L Park, H-G Kim, W-T Kim, T-W Kim, Phys. Stat. Sol. (b) V180 K69 (1993).
11. X. Solans, A. M. Font etc., AFINA Vol. 45 (1988) 277.
12. C. Lowe-Ma, T. A. Vanderah, Acta Crystallographica C, Vol. 47 (1991) 919-924.
13. Shannon and Prewitt, Acta Crystallogr. B25 (1969) 925. Shannon, ibid., A32 (1976)
751.
14. J. J. Davis and J. E. Nicholls, J. Phys. C: Solid State Phys. Vol 12 (1979) P3329.
15. A. N. Georgobiani, Yu. V. Ozerov, E. E. Strumban, V. V. Anisimov, Phys. Stat. Sol.
A35 (1976) P251.
16. Y. L. Soo, Z. H. Ming. S. W. Huang, Y. H. Kao, R. N. Bhargava, D. Gallagher, Phys.
Rev. B, Vol. 50 (1994) P7602.
17. A. Mehra, J. Elecrrchem. Soc. Vol 31 (1971) 136.
117
18. H. E. Gumlich, R. L. Pfrogner, J. C. Shaffer, F. E. Williams, J. Chem. Phys. Vol 44
(1966) P3929.
19. J. W. Stout, J. Chem. Phys. Vol 31 (1959) P709.
20. Y. Tanabe and S. Sugano, J. Phys. Soc. Jpn., Vol. 9 (1954) 753 and 766.
21. L. E. Orgel, J. Chem. Phys., Vol. 23 (1955) 1004.
22. I. Matsuyama, N. Yamashita, K. Nakamura, Phys. Soc. Jpn., Vol 58 (1989) P741.
23. 0. Goede and W. Heimbrodt, Phys. Stat. Sol. (B)146, (1988) P11.
24. K.K. Stavrev, M.C. Zerner, International J. Quantum Chem., Vol 65 (1997) 877 and
papers cited in it.
118
CHAPTER 7
CONCLUSION
Selected phosphors of importance for continued development
of flat-panel
electroluminescent displays (ELDs) and field-emission displays (FEDs) have been
synthesized and studied.
By using ceramic-processing methods, solid-solution
techniques, and codopants, the luminous efficiencies and chromaticities of several
phosphors have been greatly improved. The major achievements of this works are:
1.
The maximum emission wavelength of Y2Si05:Ce3+ (the best rated blue FED
phosphor candidate) has been shifted by adding Sc to form the solid solution Y2xSCxS105 (0
2.
x
0.75.)
Eu2+-doped Mgi_xCaS luminescent powders were prepared by flowing H2S(g)
over nanoparticulate Mgi_xCa.0 that was prepared by the combustion method.
The phosphor powder is characterized by faceted grains with a median particle
size near 21Am and a high cathodoluminescent efficiency: 8.07 lm/w at 2000V,
which is comparable to that of Y203:Eu.
3.
The luminescence mechanism of SrS:Cu coactivated with different 1+ and 3+
ions is explained by considering the nature of the isolated Cu coordination
environments that result from the method of charge compensation. The emission
color can be tuned from blue to green.
4.
(Zn, Ga)S:Mn solid-solution phosphors were synthesized by using hightemperature solid-state reactions.
By adjusting the concentration of Ga, the
emission colors range from yellow to deep red.
This series of phosphors
provides the possibility for the development of a useful, true red EL phosphor.
On the basis of these results, a trichromatic phosphor system has been proposed for
upgrading ELD performance from monochrome to full-color capabilities.
119
BIBLIOGRAPHY
Akella , A., and Keszler, D. A., Mater. Res. Bull., Vol. 30(1995), pp.105-111.
Alquier, C., Vndenborre, M. T., and Henry, M., J. Non-Crystalline Solids, Vol.79 (1986),
pp.383-395.
Altomare, A., Cascarano, G., Giacovazzo, C., and Guagliardi, A., J. Appl. Cryst. Vol.26
(1993), pp.343.
Asano, S., and Nakao, Y., J. Phys. C: (solid State Phys.), Vol. 12(1979), pp.4095-4108.
Avinor, M., Carmi, A. J., and Weinberger, Chem. Phys., Vol. 35(1967), pp.1978.
Ballard, C. P., and Fanelli, A. J., "Sol-gel chemistry" in "Chemistry of Advanced
Materials", Ed. by C. N. R. Rao, Blackwell Sci. Pub. (1993).
Barrow, W. A., Covert, R. E., and King, C. N., Digest of the 1984 SID Int'l Display
Symposium, P249, (1984).
Binh, V. T., Garcia, N., and Prucell, S. T., Advances in Electronics and Electron Physics,
Vol. 95 (1996), pp. 63-150.
Blasse, G., Dirksen, G. J., and Hazenkamp, M. F., Eur. J. Solid State Inorg. Chem.,
t.26(1989), pp.497.
Blasse, G., and Grabmaier, B. C., Luminescent Materials. Springer-Verlag 1994.
Blasse, G., Structure and bonding, Vol. 76(1991), pp.154-185.
Blasse, G., and., Bril, A., J. Luminescence Vol. 3(1970), pp.109.
Blasse, G., Structure and Bonding, Vol. 76(1991), pp.17.
Blasse, G., Topics in current chemistry, Vol. 17, pp.2-24.
Bote, P. R., Patil, P. K., Nandgave, J. K., and Pawar, P. D., Solid State Communication
Vol. 79(1991), pp.5-7.
Brightwell, J. W., Ray, B., Srphton, P., and Viney, I., J. Cryst. Growth, Vol. 86(1988),
pp.634.
Brodie, I., and Spindt, C. A., Advances in Electronics and Electron Physics, Vol. 83
(1992), pp. 2-108.
120
BIBLOGRAPHY (Continued)
Bulter, K. H., Fluorescent Lamp Phosphors: Technology and Theory, Pennsylvania State
University Press (1980).
Carpenter, C. B., Wu, P., Gao, Y-M, Wold, A., Mat. Res. Bull., Vol. 24(1989), pp.1077.
Chakrabarti, K. V., Mathur, K., Rhodes, J. F., and Abbundi, R. J., J. Appl. Phys., Vol.
64(1988), pp.1363-1366
Chalamala, B. R., Wei, Y., and Gnade, B. E., IEEE Spectrum, April (1998), pp. 42-51.
Chick, L. A., Pederson, L. R., Maupin, G. D., Bates, J. L., Thomas, L. E., and Exarhos,
G. J., Mater. Lett., Vol. 10(1990), pp.6-11.
Collins, B. T., and Ling, M., J. Electrochem. Soc., Vol. 140(1993), pp.1752-1755.
Database of FED phosphors from PTCOE.
Datta, J. Electrochem. Soc., Vol. 114(1967), pp.1137.
Davis, J. J., and Nicholls, J. E., J. Phys. C: Solid State Phys. Vol. 12 (1979), pp.3329.
Ellis, R. K., SID Seminar lecture notes, M-8/3 (1997).
Felsche., J., Structure and Bonding Vol. 13 (1973), pp.99.
Fumo, D. A., Morelli, M. R., and Segadas, A. M., Mater. Res. Bull., Vol. 31(1996),
pp.1243-1255.
Georgobiani, A. N., Ozerov, Yu. V., Strumban, E. E., Anisimov, V. V., Phys. Stat. Sol.
A35 (1976), pp.251.
Gmelins handbuch der Anorgnischen Chemie, 28(a) Hb/B3 13463/4(1961).
Goede, 0., and Heimbrodt, W., Phys. Stat. Sol. (B)146, (1988), pp.11.
Grisafe, D. A., and Fritsch, C. W., J. Solid Statae Chem. Vol. 17(1976), pp.313-318.
Gumlich, H. E., Pfrogner, R. L., Shaffer, J. C., Williams, F. E., J. Chem. Phys. Vol. 44
(1966), pp.3929.
Gumlich, H. E., J. Lumin, Vol. 23 (1981), pp.73-99.
Henderson, B., and Imbusch, G. F., Optical Spectroscopy of Inorganic Solids, Oxford
Scientific Publications, 1989.
121
BIBLIOGRAPHY (Continued)
Janesky, J., Warren, W., Li, D., and Keszler, D. A., unpublished results.
Kahn, Frederic J., SID Seminar Lecture Notes, M-1/3 (1997).
Kanari, P. S., Indian J. Pure Appl. Phys., Vol. 9(1971), pp.1060.
Kasano, H., Megumi, K., and Yamamoto, H., Electrochem. J., Soc., Vol. 131(1984),
pp.1953-1960.
King, C. N., 1988 SID Seminar lecture notes, Vol. 1 (1988) 2A/1.
King, C. N., Coovert, R. E., Dolinar, B. J., Hix, S. R., Advances in Display Technology
IV, SPIE Vol. 457 (1984), pp.74.
Kinzhibalo, L.N., Trunov, V. K., Evdokimov, A. A., and Krongauz, V. G., Sov. Phys.
Crystallogr. 27(1), pp.22-25 (1982).
Klug, P., and Alexander, L. E., X-ray Diffraction Procedures, 2nd ed., Chap. 9, New
York, Wiley, (1974).
Kottaisamy, M., Jeyakumar, D., Jagannathan, R., and Rao, M. M., Mater. Res. Bull., Vol.
31(1996), pp.1013-1020.
Lammers, M. J., and Blasse, G., J. Electrochem. Soc., Vol. 34 (1987), pp.2068.
Langford, J., J. Appl. Cryst., Vol. 4 (1971), pp. 259 and Vol. 6 (1973), pp.190-196.
Larson A. C., and Von Dreele, R. B., General Structure Analysis System, University of
California (1985-1990).
Laud, B. B., and Kulkarni, V. W., J. Phys. Chem. Solid, Vol. 39(1978), pp.555.
Lehmann, W., and Ryan, F. M., J. Electrochem. Soc., Vol.118 (1971), pp.477.
Lehmann, W., J. Lumin. Vol. 5(1972), pp.87.
Lehmann, W., J. Electrochem. Soc., Vol. 117(1970), pp.1389.
Lenard, P., Schmidt, F., and Tomascheck, R., Handbook of Experimental Physics, Vol.
23, Akademie Verlagsges, leipzig, 1928.
Li, D., and Keszler, D. A., 4th Int. Conf on Sci. and Tech. of Display Phosphors. Bend,
Oregon, pp.331 (1998)
122
BIBLIOGRAPHY (Continued)
Lin, J., Su, Q., Zhang, H. J., and Wang, S. B., Mater. Res. Bull., Vol. 31 (1996), pp. 189.
Livage, J., Henry, M., and Sanchez, C., Prog. Solid State Chem. Vol.18(1998), pp.259341.
Lowe-Ma, C., Vanderah, T. A., Acta Crystallographica C, Vol. 47 (1991), pp.919-924.
Maksimov, B. A., Iljukhin, V. V., Khantanov, Ju. A., and Be lov, N. V., Soviet PhysicsCrystallography, Vol. 15 (1970), pp.806.
Matsuyama, I., Yamashita, N., Nakamura, K., Phys. Soc. Jpn., Vol. 58 (1989), pp.741.
Mehra, A., J. Elecrrchem. Soc. Vol 31 (1971), pp.136.
Mehrotra, R. C., Structure and Bonding, Vol. 77(1992), pp.2-35.
Meijerink, A., Shipper, W. J., and Blasse, G., J. Phys. D-Applied Physics, Vol. 24 (1991)
pp. 997.
Mishra, K. C., Johnson, K. H., Schmidt, P. C., Mater. Sci. and Eng., B18(1993), pp.214.
Missous, 0., Loup, F., Fesquet. J., and Prevost, H., Eur. J. Solid State Inorg. Chem., Vol.
28(1991) (supplement), pp.163-166.
Moine, B., and Pedrini, C., Phys. Rev. B, Vol. 30(1984), pp.992.
Ono, Y. A., Electroluminescent Display, World Scientific Publishing Co., 1995.
Orgel, L. E., J. Chem. Phys., Vol. 23 (1955), pp.1004.
Oyetola, S., Verbaere, A., Piffard Y., and Tournoux, M., Eur. J. Solid State Inorg. Chem.,
t. 25(1988), pp.259.
Pandey, R and Sivaraman, S., J. Phys Chem. Solids, Vol. 52(1991), pp.211.
Park, H-L, Kim, H-G, Kim, W-T, Kim, T-W, Phys. Stat. Sol. (b) V180 K69 (1993).
Payne, S. A., Goldberg, A. B., McClure, D. S., J. Chem. Phys., Vol. 78(1983), pp.3688.
Payne, S. A., Austin, R. H., Mcglure, D. S., Phys. Rev. B, Vol. 29(1984), pp.32.
Poelman, D., Vercaemst, R., Van Meirhaeghe, R. L., Laflere, W. H., andn Cardon, F., J.
Lumin., Vol. 65(1995), pp.7-10.
123
BIBLIOGRAPHY (Continued)
Pohl, U.W., and Gumlich, H. E., Chap 3 in Solid State Luminescnece, Theory, Materials
and Devices. Ed by A.H. Kitai, Chapman&Hall 1993.
Powder Diffraction File Card No. 36-1476.
Powder Diffraction File Card No. 41-4.
Rabinovich, E. M., Shmulovich, J., Fratello, V. J., and Kopylov, N. J., Am. Ceram. Soc.
Bull., Vol. 66 (1987), pp.1505.
Ramsay, J. D., "Sol-gel Processing" in"Cotrolled Particles, Droplet and Bubble
Formation", Ed. D. J. Wedlock, Butterworth-Heinemann, 1994.
Rennie, J., Takeuchi, K., Kaneko, Y., and Koda, T., J. Lumin., Vol.48&49(1991), pp.787.
Ropp, R. C., Luminescence and the Solid State, Elsevier (1991).
Shannon and Prewitt, Acta Crystallogr. B25 (1969), pp. 925. Shannon, ibid., A32 (1976),
pp.75 L
Sharma, D., and Amarsingh, Indian J. Pure Appl. Phys., Vol. 9(1971), pp.810.
Shumulovich, J., Berkstresser, G. W., Brand le, C. D., and Valentino, A., J. Electrochem.
Soc., Vol. 135 (1988), pp. 3141.
Singh, N., Marwaha, G. L., Mathur, V. K., Phys. Stat. Sol. (a), Vol. 66(1981), pp.761.
Singh, N., VIJ, O.R., J. Mater. Sci., Vol. 29(1994), pp.4941-4945.
Smith, R. T., Information Display Vol. 14 (1998), No.2, pp. 12-16.
Solans, X., Font, A. M., etc., AFINA Vol. 45 (1988), pp.277.
Soo, Y. L., Ming, Z. H., Huang, S. W., Kao, Y. H., Bhargava, R. N., Gallagher, D., Phys.
Rev. B, Vol. 50 (1994), pp.7602.
Stavrev, K.K., Zerner, M.C., International J. Quantum Chem., Vol. 65 (1997), pp. 877
and papers cited in it.
Stout, J. W., J. Chem. Phys. Vol. 31 (1959), pp.709.
Summers, C. J., SID Seminar lecture notes, M-5/1 (1997).
Sun, S. S, Dickey, E., Kane, J., and Yocom, N., Proc. 17th Int. Display Research Conf.,
Morreale, J., Ed.; Society for Information Display: Toronto, pp.201, 1997.
124
Bibliography (Continued)
Tamura, Y., Jpn. J. Appl. Phys., Vol. 33(1994), pp.4640-4646.
Tanabe, Y., and Sugano, S., J. Phys. Soc. Jpn., Vol. 9 (1954), pp.753 & 766.
Tannas, Lawrence E. jr., SID Seminar Lecture Notes, M-1/3 (1995).
Vecht, A., Charlesworth, D., and Smith, D. W., SID 97 Digest, pp.588-590(1997).
Vecht, Aron, SID Seminar lecture notes, F-2/3 (1996).
Vij, D. R., Luminescence of Solids, Plenum, 1998.
Wager, J. F., and Keir, P. D., Ann. Rev., Mater. Sci., Vol. 27 (1997), pp. 223-248.
Walker,N., and Stuart, Acta Cryst. A39 (1983), pp.158-166.
Whitaker, J. C. Electronic Display, Technology, Designs, and Applications. McGrawHill Inc., 1994.
White, D., Trans. Brit. Ceram. Soc. Vol. 63(1964), pp.301.
Williamson, G. K., and Hall, W. H., Acta. Metall. Vol. 1(1953), pp.22.
Williamson, S. J., and Cummins, H. Z., Light and Color in Nature and Art, John Wiley &
Sons Inc., 1983.
Yamashita, N., Ebisumori, K., Nakamura, K., Jpn. J. Applied Phys. Vol. 32(1993),
pp.3846.
Yamashita, N., Jpn. J. Applied Phys. Vol. 30(1991), pp.3335.
Yeh, C-Y, Lu, Z. W., Froyen, S., Zunger, A., Phys. Rev. B, Vol. 46(1992), No.16
pp.10086
Young, R. A., "The Rietveld Method", International Union of Crystallography, Oxford
Science Publications, 1993.