1

v *5: -
a“ 3'*;3!::.£w
~ wads?
. ' V ‘

Iv
m

.
1‘ 4:; mg" “"51"
33:35?" ‘ 9?.

W 4.1“
n: ' ’.-r
A- .... _,,
33m WEE;
""235 “yr "
1. 43% r .
7r

_ 1.
I] "Eé;
S. ai-fiua‘ii' ,

4' - xAc";

n. w:
. ‘
IR'
. “

.AA.
a.
u...
‘

q,
rm:-

«3&5:

5-.
.fik-t‘

c“.

15619;!
' “Jim
"a‘g‘fifi
{3'3“ - ,

,c

 

Date

0-7639

This is to certify that the

thesis entitled

SURFACE MODIFICATION OF ZINC SULFIDE BY
ION IMPLANTATION WITH 200 KEV NICKEL IONS

presented by

STEPHEN CURTIS WHITE

has been accepted towards fulfillment
of the requirements for

MM

Major professor

 

 

IZ/zz/ZS

MS U i: an Affirmative Action/Equal Opportunity Institution

 

 

Illllllllllllllll

 

LIBRARY
Mlchigan State
University

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SURFACE MODIFICATION OF ZINC SULFIDE BY
ION IMPLANTATION WITH 200 KEV NICKEL IONS

BY

Stephen Curtis White

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Metallurgy, Mechanics, and Materials Science

1993

ABSTRACT

SURFACE MODIFICATION OF ZINC SULFIDE BY
ION IMPLANTATION WITH 200 KEV NICKEL IONS

BY

Stephen Curtis White

Zinc sulfide is an important infrared transmitting optical
window material. Because of it's brittleness and low strength it
is vulnerable to erosion and impact damage in severe environments.
It has been suggested by Kriven, et.al., that the compound nickel
sulfide (NiS) could act as a metastable transformation toughener
due to the positive volume expansion associated with the
displacive transformation to the low temperature allotrope. ZnS
was irradiated with l x 1017 ions/cm2 of 200 keV Ni+ at
approximately 400 K, aimed at producing metastable NiS
precipitates capable of improving the strength and toughness of
the surface layer of ZnS. The effect of ion implanation and post-
irradiation anneals upon the mechanical properties of ZnS were
investigated through hardness evaluations, and evidence of
precipitates were observed with transmission electron microscopy.
Significant improvement in the hardness was found for post-

irradiated annealed zinc sulfide.

Copyright by

STEPHEN CURTIS WHITE

1994

ACKNOWLEDGEMENTS

I would like to thank Dr. David Grummon of the Department of
Metallurgy, Mechanics, and Materials Science at Michigan State
University for his technical guidance, encouragement, and patience
throughout the duration of completing this work.

I would also like to thank Dr. Eldon Case of the Department
of Metallurgy, Mechanics, and Materials Science at Michigan State
University for his guidance and council with many technical issues
within this project.

I thank Dr. A. J. Armini of Implant Sciences in Danvers,
Massachusetts for contributing the implantation processing and
computer modeling.

And, I thank my wife, Kathy, for her endless encouragement,

motivating spirit, and sacrifices she made for this work, and
mostly for her love.

in

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

CHAPTER 1. INTRODQCTION

1.1 BACKROUND

1.2 ION IMPLANTATION

1.3 ION IMPLANTATION IN CERAMICS
1.4 ZINC SULFIDE

1.5 TRANSFORMATION TOUGHENING

1.6 NICKEL SULFIDE AS A TRANSFORMATION TOUGHENER

CHAPTER 2. EXPERIMENTAL PROCEDURE

2.1 SPECIMEN PREPARATION

2.1.1 Cutting

2.1.2 Polishing

2.1.3 Mbunting

2.1.4 Ion Implantation

2.1.5 Dimple Grinding

2.1.6 Chemical Thinning

2 . 1 .7 Ion Milling

2.2 X-RAY DIFFRACTION

2.3 TRANSMISSION ELECTRON’MICROSCOPY
2.4 INDEXING TEN DIFFRACTION PATTERNS

2.5 HARDNESS

iv

mi

14
3O
33

36

41
41
43
44
46
47
48
SO
55
64
67
69

71

2.5.1 Nickel-implanted ZnS
2.5.2 Annealed Nickel-implanted ZnS
2.6 SCANNING ELECTRON MICROSCOPY

2.7 X-RAY PHOTOELECTRON SPECTROSCOPY

HAPTER RE T DI SSIO

3.1 ION MILLING EFFECTS

3.2 X-RAY DIFFRACTION

3.3 TEM RESULTS

3.4 INDEXING DIFFRACTION PATTERNS
3.4.1 Zinc Sulfide

3.4.2 Nickel-implanted Zinc Sulfide
3.5 HARDNESS

3.5.1 Nickel-implanted Zinc Sulfide
3.5.2 Annealed Nickel-implanted Zinc Sulfide
3.6 SCANNING ELECTRON MICROSCOPY

3.7 X-RAY PHOTOELECTRON SPECTROSCOPY

3.8 DIFFUSION

HAPTER 4 L I
APPENDICES
APPENDIX A. VICKERS.INDENT HARDNESS DATA
APPENDIX B. BASIC PROGRAM - HARDNESS IN GIGAPASCALS

LIST OF REFERENCES

71

73

76

77

78

78

89

92

96

96

103

115

116

118

125

133

139

147

151

156

159

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

10.

11.

LIST OF TABLES

Chemical thinning durations and observations.

XRD diffraction data for B—ZnS.

Indexing data for SAD pattern of unmodified ZnS
using camera constant of 2.035.

Indexing data for SAD pattern of unmodified ZnS
using camera constant of 2.707.

Indexing data for irradiated ZnS using calibrated
camera constant of 2.035.

Indexing data for "reflections" in irradiated ZnS
with a camera constant of 2.097.

Indexing data for ring pattern of Figure 24 with
camera constant of 2.097.

Indexing data for ring patterns from enlargement of
Figure 24 with calibrated camera constant of 2.097.

Comparison of Ni—irradiated ZnS d—spacings to various
phases of nickel, sulfur, and zinc.

Vickers hardness data for nickel-irradiated and
non-irradiated ZnS pre— and post-anneal.
and sulfur.

Binding energies of nickel, zinc,

Figure
Figure
Figure
Figure

Figure

Figure

Figure

Figure
Figure
Figure
Figure
Figure
Figure

Figure

Figure
Figure

Figure

10.

11.

12.

13.

14.

15.

16.

17.

LIST OF FIGURES

Infra-red transmission wavelengths [2].
Defect generation by ion implantation.
Effect of varying ion dose in ceramic [13].
Relative hardness of Cr—implanted alumina [18].

Flexural strength of implanted alumina
[l9].

Flexural strength of alumina
with ion species [20].

Nickel concentration in alumina
[l9].

Knoop microhardness for YSZ [23].

Atomic structure of zinc sulfide [29].
Nickel—sulfur phase diagram [27].
Slicing ZnS for TEM specimens.
ZnS specimen arrangement for irradiation.
Gatan hand-held disc grinder[35].

Gatan Model 656 Dimple Grinder
(top view)[35].

Dimple grinder specimen mounts[35].
Ion mill specimen holder cross-section.

Anneal profiles of Ni-irradiated ZnS.

vfi

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

TEM micrograph of thinned edge
of 200keV Ni-irradiated ZnS.
(Mag:100k)

"Saw tooth" features observed
in TEM micrograph of 200keV
Ni-irradiated ZnS. (Mag:100k)

Bright-field TEM micrograph
of 200keV Ni—irradiated ZnS.
(Mag:100k)

"Reflections" observed in dark
-field TEM micrograph of 200keV
Ni—irradiated ZnS. (Mag:100k)

Bright—field TEM micrograph
in precipitant-rich region of
Ni-irradiated ZnS. (Mag:100k)

Dark—field image of region in
Figure 22 showing precipitant
reflections in Ni—irradiated ZnS.
(Mag:100k)

Selected-area diffraction
pattern of 200keV nickel-
irradiated ZnS. (Mag:100k)

X—ray diffraction pattern of ZnS.
Selected-area diffraction
pattern of zinc sulfide.

(Mag: 100k)

Illustration of ZnS selected-area diffraction
pattern of Figure 26.

Monograph matching ZnS in Figure 26.[39]

an

Figure
Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

Illustration of Ni:ZnS pattern of Figure 24.
Monograph matching Ni:ZnS in Figure 24.[39]

Increase in Vickers hardness for
200keV Ni—irradiated ZnS and
control samples with annealing.

Scanning electron micrograph
of zinc sulfide. (Mag:1100)

Scanning electron micrograph
of zinc sulfide surface "voids."
(Magz3300)

Scanning electron micrograph
of 200keV Ni—irradiated ZnS.
(Mag:1100)

Scanning electron micrograph of
Ni-irradiated ZnS annealed to 750
degrees celsius. (Mag:1100)

Scanning electron micrograph of
Ni—irradiated ZnS annealed to 850
degrees celsius. (Mag:1100)

X-Ray photoelectron spectrum of 200keV
nickel—irradiated zinc sulfide.

X—Ray photoelectron spectrum of 200keV
nickel-irradiated zinc sulfide in region
of nickel peak.

X—Ray photoelectron spectrum of 200keV
nickel-irradiated zinc sulfide in region
of sulfur peak.

X—Ray photoelectron spectrum of 200keV

nickel-irradiated zinc sulfide annealed to
800 degrees celsius.

ix

Figure 41. Distribution of 200keV nickel
ions implanted into zinc sulfide.

CHAPTER 1. INTRQDQCTION

1.1 BACKROUND

Thermal imaging systems detect energy in the form of
radiation emitted by warm bodies which exhibit spectra in the
wavelength range defined as infrared (IR). For operation in the
earth's atmosphere, the application of IR optics is limited by
selective transmission to three wavelength bands: 0.8-1.3 microns,
3-5 microns, and approximately 8—14 microns as shown in Figure 1
[1,2]. Use of these systems on aircraft expose the infrared
optical material, or "window," to hostile environments. (This use
of the term "window" refers to the actual optical material, such
as ZnS or ZnSe, used in an IR detector, rather then the
conventional reference to a band of wavelengths in the IR
transmission spectra.) The purpose of the IR window is to protect
the delicate imaging system from such environments while
permitting reception and processing of input received via IR
signals that pass through the window material. Output data is
then generated by the input, and is used for telemetry and
guidance.

Improvements in material selection and processing continue
to be explored to enhance the effectiveness of IR windows in
fulfilling these objectives. However, difficulties may lie in
altering the window material properties and performance while
maintaining the optical characteristics necessary for system

1

2

operation. Wien's displacement law [3] gives the wavelength of
maximum emission, km, for a blackbody radiating at a given
absolute temperature T:

XmT = 2898 microns-Kelvin (1)
For objects at an approximate temperature of 20 degrees celsius
(298 Kelvin) the IR wavelength hand between 8 and 14 microns is
suitable for their detection. Zinc sulfide (ZnS) is IR
transmitting in this range.

It has been observed that ZnS experiences mechanical damage
and transmission losses as a result of impact with rain droplets
[4,5,6]. Ceramics are most susceptible to failure in tension,
thus it is probable that pre-existing flaws in the near—surface
region of this ceramic experience growth initiated by radial
tensile stresses. Rayleigh surface waves generated by the impacts
with water droplets were reported to be the source of these
stresses and considered to ultimately be responsible for the
growth of surface cracks. Donald and McMillan [7] conclude that
"any improvement in the mechanical properties of existing infra—
red materials should ideally be achieved by increasing the strain
to failure of the material" and that the techniques which are the
most suitable for achieving this are those involving surface
modification.

Ion implantation has been studied considerably in the 1980's
as a method to modify surface and near-surface regions of metals,
ceramics, and polymers to improve properties such as wear
resistance, hardness, and fracture toughness. Significant

widespread interest in the implantation process for affecting such

3
a diverse range of materials attests to the unique advantages
offered by ion implantation.

Fundamentally, ion implantation is the physical acceleration,
bombardment, and injection of a particular ion species into the
surface of a substrate. With typical energies ranging from
100,000 electron-volts (100 keV) to as high as 1 million electron-
volts, a single ion may enter the target and come to rest in
approximately 10'12 seconds (1 picosecond). Essentially the
effect is that of an extremely rapid cooling rate, or "quench."
Thus, the thermodynamic constraints present during conventional
surface—altering or alloying processes do not apply with this
technique. Although thermal diffusion of atomic species is
minimal, other mechanisms of atomic mobility may occur both during
and shortly after the ion implantation process is complete. This
will be explored in greater detail in the next section.

Ion implantation also offers no distinct interface between
modified and unmodified regions. This is achieved by the intimate
mixing of the projectile and target species as well as a gradual
decrease in ion concentration with depth towards the bulk region.
Therefore, interface delamination or debonding are not experienced
by ion implanted materials.

The objective of the investigation described herein was to
introduce a transformation toughener into the subsurface region of
zinc sulfide. This was attempted by the ion implantation of
nickel—beam ions into the subsurface region of a prepared surface
of ZnS, followed by thermal annealing. Changes in hardness were

observed as well as physical evidence of the atomic species

4
proposed to be responsible for improvements in the surface of this

infra—red transmitting ceramic.

.HNH mcumcmam>mz coammfifimcmuu ownimuucH

31.500230;
0.3.0.3 0n. 0.: 0.0. 0.0 0.0 0t. 0.0 0.0 .0.

v o.m

.H musmflm

0.N -0;

 

A

 

_

—

a

_

_

 

_

_

 

_

 

 

0m
00
00

Om

 

 

00—

(0/0) uogssgwsuml

1.2 ION IMPLANTATION

Ion implantation is gaining momentum as a method to alter the
surface and sub-surface properties of ceramics used in various
applications. This process is relatively advanced and can be
considerably expensive, however, its advantages over conventional
surface modification techniques have resulted in experimental
investigations with respect to metals, ceramics, and polymers.
These advantages are l) a clean process, 2) lack of sharp
interface thus no interface delamination, 3) potential to implant
any material, 4) with any ion species, 5) at a wide range of
accelerating voltages, 6) achieving a specified dose, and 7)
without the thermodynamic equilibrium constraints normally
associated with conventional processing.

The process of ion implantation is one of rapidly
accelerating a selected ion species into a solid target. The
atomic composition and structure of the near-surface region of the
target are altered by the presence of the new species and
resulting defects. Typically, the equilibrium solubility of the
implanted species in the target material is exceeded. This occurs
because of two characteristics of ion implantation.

First, is the relatively shallow depth to which a typically
high dose of implanted ions penetrate into the target material in
a very short time, resulting in a high concentration per unit
volume of the implanted species in the sub—surface region. Second
is due to the lack of thermal energy available for driving

6

7
temperature-dependent diffusion of this high concentration of
implanted ions. The event marked by the ions impact and
subsequent decceleration and dissipation of kinetic energy is a
process which occurs in approximately 10'12 seconds [8]. The
effects of such a "quench rate" coupled with the ability to
maintain the implanted substrate at low temperatures releases the
thermodynamic constraints associated with equilbrium dependent
processes.

This implanted region may be considered to be in a metastable
energy state. That is, the entire layer affected by ion
implantation is in a higher Gibbs free energy state relative to
the bulk material. However, each implanted ion comes to rest at a
position that represents a local minimum energy state. The ions,
having dissipated their kinetic energies through atomic and
nuclear collisions, now possess insufficient energy to mobilize
further towards an even lower Gibb's free energy position. An
additional driving force such as thermal energy would be required
to initiate this motion. The subsurface region, in effect, is one
that is super-saturated with the implanted species, and metastable
in terms of its energy state; a condition which lends itself to a

precipitate transformation reaction of the type:

(1'--> a +B (2)
where, a' = a metastable supersaturated solid solution,
a = a structure identical to a' but closer to an

equilibrium concentration, and

a stable or metastable precipitate [9].

'5)
ll

An accelerating ion incident upon a substrate loses its
kinetic energy by interacting with the target material's atoms
("nuclear stopping") and electrons ("electronic stopping").
Elastic scattering by the atomic nuclei which are shielded by
their electron configuration, and the ionizaiton of these
electrons, constitutes the means by which most of the energy
dissipation occurs [10]. Figure 2 depicts a binary atomic
structure in which ions are being implanted. The path labeled
"1" illustrates a non—displacing collision. In this case, as the
incident ion passes by a stationary lattice atom, a portion of the
ion's kinetic energy is imparted to the atom, temporarily
increasing it's vibrational energy. The path of the ion is
altered by this elastic scattering event and the lattice atom
remains in its original position. (Many sources are available
which discuss in detail the dynamics of particle collisions and
the phenomena of vacancy and interstitial defect generation,
radiation—enhanced diffusion, and ion channeling [8,10,11,121.)

A significant number of additional energy—loss processes
also occur which result in the generation of a large number of
defects by atomic rearrangement. One mechanism for this is the
case in which the amount of energy imparted to a lattice atom
exceeds its threshold displacement energy and the atom is
displaced (primary displacement), leaving a vacant atomic
position, or "vacancy" in the lattice structure. In Figure 2,
this is depicted by path 2. The displaced lattice atoms with
their newly absorbed vibrational energies, themselves, become

"incident" ions. Each of these ions in turn lose their energy

9
by displacing additional atoms in the target ("secondary
displacement"). This is shown by path 4 in which a lattice atom,
having been displaced by an incident ion's second collision event
(path 3), leaves its site and impacts and displaces another
lattice atom of the same type.

Multiple occurrences of this type of event along an incident
ion's path result in a "collision cascade" region. Many of the
atomic positions closest to the ion‘s path become vacancies. The
increased number of vacancies, which are point defects,
significantly enhances the diffusivity of the various species, for
vacancies are central to the mechanism by which substitutional
atoms diffuse.

At the end of the displacement phase, many of the displaced
atoms do not regain a position as in the original structure.
Depending on the incident and target species characteristics,
there can be a high concentration of these atoms at interstitial

positions within the lattice. These are referred to as "self—

interstitial atoms" (SIA's). Many of these SIA's end up adjacent
to vacancies. Such a vacancy—interstitial pair, referred to as a
"Frenkel pair," is shown in the atomic model of Figure 2.

The incident ions form defects not only by the energy they
impart but by their presence. Once they come to rest, an
interstitial or substitutional defect is formed (labeled "i" or
"s" in Figure 2). The ion species have sizes and energy
characteristics that differ from the target material atoms and the

substitutional ions may bond with the cations or anions of

10
the target material. As mentioned previously, their concentration
usually exceeds equilibrium solubility limits, and their
redistribution is dependent upon the various processes which occur
in ion irradiated materials.

Atoms present in a region of high concentration in a material
typically diffuse to regions of lower concentration, for this
reduces the overall Gibb's Free Energy. "Radiation—enhanced
diffusion" refers to the increased potential for diffusing species
in irradiated materials due to the excess number of point defects
and defect clusters that are introduced by the implantation
process. As is true for the implanted species, the number of
other defects generated by ion implantation typically far exceed
their equilibrium concentrations. The diffusion of these defects,
whether they are vacancy, substitutional, or interstitial, can be
activated thermally either during ion implantation or with
subsequent annealing. The result is an increased flux of defect
migration due to the increased quantity of diffusing species.

This redistribution is significant because it determines the
change in properties of the modified material. Radiation-enhanced
diffusion is a thermally activated process, therefore the
microstructure is driven towards a state of equilibrium.

"Radiation—induced segregation" occurs by the influence of
irradiation-generated defect fluxes upon the implanted species.
The implanted ions "couple" to the defect fluxes and result in
nonequilibrium concentration gradients. For example, such a

gradient will exist between regions of high and low

11
concentrations of vacancies generated during implantation. As
vacancies in the high region move across the gradient to the low
region, the implanted atomic species will move accordingly, in the
opposite direction of the vacancy flux. This type of vacancy
motion can occur during the implantation process and is driven by
the abrupt atomic changes taking place within the bombarded atomic
structure. Activated independent of thermal energy, this activity
typically drives towards a nonequilibrium state. Since this
results in local changes in composition, precipitation of a second
phase may be induced.

Both radiation—induced segregation and radiation-enhanced
diffusion play significant roles in irradiated materials which
undergo subsequent heat treament. Due to the nature of their
mechanisms, radiation-induced segregation would be expected to be
prevalent during ion implantation in which the target material is
kept at a low temperature throughout the process. This results in
a supersaturated region that has been rapidly generated.
Therefore, it would be metastable and more conducive to
precipitation of a second phase, possibly assisted by the
segregation of newly bonded species to second-phase embryo sites.

It follows that radiation—enhanced diffusion would most
likely be the dominant effect during annealing of the implanted
material. Second—phase precipitants which may have formed prior
to annealing would grow in volume, thus moving the atomic
structure closer to equilibrium.

The final results on morphology and mechanical properties

12
are ultimately governed by the characteristics of the constituents
that are combined, and the parameters of the ion implantation
process. Variations in both of these are reviewed in Section 1.3

on ion implantation in ceramic materials.

13

 

Figure 2. Defect generation by ion implantation.

1.3 ION IMPLANTATION IN CERAMICS

Various studies have shown that ion implantation into
particular ceramic materials improves the mechanical properties of
the material. It is known that when a load is applied to a
ceramic, the tendency towards failure is due to the tensile
components of the load. The corresponding stresses induced in the
surface initiate and propagate cracks until failure occurs. Pre-
existing cracks typically occur in ceramic materials and therefore
are even more susceptible to cause damage or failure. These
surface phenomena are amenable to modification by ion
implantation.

The controllable factors which determine the characteristics
of implantation are the ion dose (typically 1013 to 1018 ions per
square centimeter for ceramics), the energy (typically 20 to 200
keV), incident ion mass, and target material. Of these, most
commonly the ion dose, or "fluence," is varied.

McColm and Clark [13] illustrate the effect on the near-
surface structure of a ceramic with varying ion dose (Figure 3).
Three regions considered unique are described by the authors whom
speculate on mechanisms which may be responsible for mechanical
property improvements. A summary of these descriptions follow.

The lower dose of 1016 ions per square centimeter is labeled
as Region I. The damage resulting from this fluence is
low enough so that the solid maintains crystallinity, i.e. both
short- and long-range order is maintained in the lattice

l4

15

structure. Damage is present and substitutional and/or
interstitial solid solutions contribute to surface hardening. By
selecting a large ion relative to the substrate atomic size,
implantation may result in a compressive stress which can inhibit
crack initiation and propagation, thereby improving the surface
toughness.

An amorphous layer just below the surface is marked by Region
II in the figure from McColm and Clark. This layer is attributed
to ion doses around 1017 ions per square centimeter. Surface
softening, as observed by Burnett and Page in silicon [I4], is
expected due to the amorphous region. An increase in surface
toughness may be realized due to the plasticity of this layer.

Region III corresponds to ion doses exceeding 1018 ions per
square centimeter and possesses a thicker amorphous layer than
Region II. Additionally, Region III begins at the material
surface. It is likely that this is due to removal of the original
surface by sputtering, thereby exposing the amorphous region of
material. Flaw geometries in this region are altered and
contribute to both the strength and the surface toughness. Many
of the experimental investigations performed on ceramics subjected
to ion implantation have led to observations which may be
supported by the strengthening mechanisms suggested by McColm and
Clark.

McHargue et. a1. [15] implanted polycrystalline Ti82
(titanium diboride) with nickel ions at an energy of 1 MeV
(1 MeV = 106 electron—volts). One use of TiB2 is as a thermal

barrier coating for erosive and wear environments. Powder was

16
vacuum hot—pressed to produce specimens with a theoretical density
of 98.4 percent and grain sizes ranging from 50 to 100 microns
(10'6 meters). A fluence of 1x1017 ions per square centimeter
yielded TiB2 with a nickel-titanium ratio of about 0.12.
Specimens were prepared to permit transmission electron microscopy
(TEM) of the cross—section of the implanted region and for
hardness testing using a KnOOp microhardness technique.

Bright—field and weak-beam dark—field micrographs of the
implanted specimen revealed dislocation tangles in the implanted
layer with a thickness of about 550 nanometers (1 nanometer = 10‘9
meters). The dislocation density was moderate and relatively
uniform. The local regions between these dislocation structures
was viewed to be free of defects. Defects ranging from 5 to 10
nanometers were found to occupy the layered region extending from
550 nanometers to 750 nanometers. These defects were present in a
greater density than that of the dislocation tangles. After 750
nanometers no damage was observed. As noted by the authors and
confirmed by the presence of these defect species, the implanted
layer remained crystalline.

Using a 15 gram load (0.147 Newtons) to measure Knoop
microhardness, it was determined that the depth of the
indentations (250-300 nanometers) would result in hardness data of
the composite consisting of the modified layer and the
unmodified substrate. Due to the relatively shallow depth of

penetration of accelerated ions, this is generally the nature of
the data for hardness of ion implanted ceramics. Thus, the

results are usually reported as relative hardness values which

17
indicate whether the effects provide improvement or degradation of
surface properties. Because of the hexagonal structure of TiB2
and the corresponding anisotropy, the hardness of individual
grains were measured.

Relative hardness values of 1.7 to 2.0 were found. This
represents a 70 to 100 percent increase over that of the
unimplanted sample. After partial recovery of irradiation damage
by annealing, a decrease in the hardness ratio from 1.7 to 1.45
resulted. An increase from 40 to 80 percent of indentation
fracture toughness was determined from 0.49 and 0.98 Newton loads
of Vickers indents. The methods of Evans [16] and Marion [17]
were used to calculate the apparent fracture toughness reported
above.

In a separate study, McHargue et. al. investigated the effect
of ion implantation of alumina [18]. In this work, single
crystals of aligned and random high purity A1203 were implanted
with various ion species (Cr, Zr, Nb, Al+O, Xe) at temperatures of
77, 300, or 640 degrees Kelvin. Energies between 40 and 300 keV
were used to implant fluences ranging from 1015 to 1017 ions per
square centimeter, with current densities of approximately 2
microamps per square centimeter. Characterization techniques used
were Rutherford backscattering-ion channeling (RBS-C),
transmission electron microscopy, and Knoop microhardness.

It was observed by RBS-C that the aluminum and oxygen
sublattices sustained significant damage at 300 degrees Kelvin,
and metastable interstitial solid solutions were formed. A

relative hardness increase of 50 percent was determined for a

18
specimen implanted with 1017 chromium ions per square centimeter
at 300 Kelvin (Figure 4). It was also noted that a hardness
increase only on the order of about 10 percent was found for a
solid solution of Cr203—A1203 prepared using a conventional
method, and possessing a cation concentration similar to that of
the implanted specimen, (approximately 10 percent Cr).

The response of fracture stress of alumina to the
implantation of ions was the subject of study conducted by T.
Hioki et. al. [19]. This work included a pre-anneal of single
crystal alpha-alumina specimens to eliminate mechanical damage
prior to the implantation step. Nitrogen, and in a separate
study, nickel ion doses from 1015 to approximately 1017 ions per
square centimeter were implanted at 400 keV and a current density
of approximately 3 microamps per square centimeter, while under
vacuum.

Three-point bending strength tests which placed the implanted
surfaces in tension were performed at room temperature on five
specimens from each of the conditions of interest. Figure 5 is a
plot of the flexural strength versus nitrogen ion dose. Increases
of 100 to 300 x 106 Pascals (106 Pascals = 1 Mega—Pasacal, or 1
MPa) of the average flexural strength were measured with respect
to an average initial value of 500 MPa for the unimplanted
material. These increases were accompanied by a decrease for the
ion dose greater than 1017 ions per square centimeter. Softening
of the surface layer due to amorphization at the higher fluence
has been suggested as being responsible for this decrease.

Further results of this work were summarized [20], and are

19

shown in Figure 6 as average relative flexural strengths versus
ion dose for argon, nickel, and nitrogen ion species. Figure 6
shows that an increase in the flexural strength of the single
crystal a-alumina of 37 percent was measured. Furthermore, it
appears that heavier ions have a greater effect if implanted at
lower fluences. It is suggested that the ion implantation
strength improvements are related to the damage incurred by the
substrate. This is consistent with the microstructural and
mechanical property changes observed by McHargue et. al. [15].

Volume expansion in the surface region of a-alumina was
reported by Krefft and EerNisse [21] to occur as a result of
implantation—induced damage. Hioki et. al. [19] proposed that "a
lateral surface compressive stress is induced" as a result of the
restraint imposed by the undamaged substrate material upon the
expanding implanted region. Additionally, each work suggests a
compressive stress, determined to be approximately 103 MPa in the
damaged surface, may be responsible for the improvement in surface
strength.

For the nickel—implanted sapphire (a-alumina) specimens in
the second study by Hioki, RBS-C data indicated extensive damage
to the surface region [19]. Nickel concentration profiles for
annealing temperatures of 800, 1000, and 1300 degrees celsius
were compared to the as-implanted concentration. Figure 7 shows
the significant change in the as-implanted profile to those of the
1000 and 1300 degrees celsius annealing results. A spinel
compound NiA1204 was detected by x—ray diffraction for specimens

that were annealed beyond 1000 degrees celsius. From optical and

20
scanning electron microscope inspections, a phase composition of
xNiAlZO4 + (1-x)A1203 was suspected of forming a surface layer
with a thickness of about 0.1 microns. Though recovery of the
implantation—induced lattice damage was considerable, it was
discovered that the flexural strength of the nickel—implanted
specimens (Figure 6) did not experience recovery (return to the
lower initial relative flexural strength of alumina of 1.0) upon
annealing.

One possible mechanism for this strengthening, as givin by
Hioki, is a residual surface compressive stress present after
cooling the composite implanted/non-implanted layers. This stress
results from the thermal expansion mismatch between the
polycrystalline NiA1203 and the substrate.

Yttria stabilized zirconia (YSZ) was subjected to ion
bombardment, heat treatment, and hardness and fracture toughness
testing by K. O. Legg et. al. [23]. Single crystals of skull
melted 9.4 mass percent YSZ were implanted with Al+ or Zr+ ions of
doses ranging from 1016 to 4x1017 ions per square centimeter at
190 keV. A 25 percent improvement in KnoOp microhardness values
(0.5 Newton indent load) was reported for aluminum implanted
samples annealed from 1300 to 1400 degrees celsius (Figure 8).
This result was anticipated, for alumina was found

to crystallize from the zirconia matrix, covering over 32 percent
of the surface, and alumina is harder than cubic zirconia.
However, thermal expansion mismatch was also considered as a
source of increased hardness by generation of a compressive stress

in the surface layer. It was reported that fracture toughness

21
increased only about 10 percent.

Significant improvements in the Al implanted specimens was
realized after annealing the higher dose (4x10l7 Al cm’2)
specimens at 1400 degrees celsius and a heating rate of 75 degrees
celsius per hour. What the authors referred to as a "nucleation
controlled microstructure" (NCM) was formed with this schedule.
Interlocking ribbons of the two phases possessed widths of
approximately 0.5 microns. It was suggested that their presence
was the mechanism responsible for a 40 percent increase in
hardness and an improvement in fracture toughness of 70 percent.
No significant changes were observed in the zirconia-implanted
YSZ.

In review, mechanical properties of ceramic surfaces can be
modified with the ion implantation of species that initiate the
mechanisms attributed to the modifications. Microstructural
changes and residual stresses induced by various methods have been
shown to result in these modifications [24]. In general, these
investigations found the changes in the atomic structure to be of
two types. One was extensive lattice damage in the form of
various defect types. The second was the presence of an
additional nucleated phase formed upon annealing.

Lattice damage is expected to be present in any irradiated
material. However, for work in which a second phase was not
found, the irradiation damage was recovered following heat
treatment, and a loss of the mechanical improvements measured
prior to annealing were diminished. This strongly suggests that a

second phase forming from the metastable region of implantation is

22

necessary in order to retain the improvements in mechanical
properties established by the surface compressive stress.

Improvements in relative hardness values and fracture
toughness for nickel-implanted TiB2 ranged from 70 to 100 percent
and 40 to 80 percent, respectively. McHargue et. al. [15] also
observed metastable interstitial solid solutions in alumina
implanted with chromium. The result was a 50 percent increase in
relative hardness. For both studies, thermal treatments reduced,
but did not eliminte the mechanical property improvements.

Increases of 100 to 300 MPa of the average flexural strength
of alumina containing implanted nitrogen ions were reported by
Hioki et. al. [19]. The modifications were attributed to a
compressive surface stress induced in the surface region by volume
expansion differences between the implanted and unimplanted
layers. For nitrogen—implanted alumina, the retention of flexural
strength in spite of rigorous annealing environments was
correlated to the thermal expansion mismatch of a dual phase
spinel compound which had formed during low temperature anneals.

Finally, yttria stabilized zirconia was found to develop a
"nucleation controlled microstructure" of "interlocking ribbons"
[23] following aluminum implantation and annealing. The presence
of this new microstructure was suspected to be responsible for a
40 percent increase in hardness and a fracture toughness increase
of 70 percent. Several mechanisms have been suggested for the
observations reviewed here. Though diverse, they share in common
the end result of establishing a residual compressive stress in

the subsurface region of a material. This condition is achieved

23

through volume expansion in the atomic structure of the original
sub-surface region and, to be sustained, may require the formation
of a second phase that will provide this increase in volume.

Solid solution-strengthening and defect interactions appear
to lead to improvements in hardness, fracture toughness, and
flexural strength in implanted crystalline layers [24]. Due to
the relatively shallow depth of implantation, only trends in
hardness measurements are reported for Vickers and Knoop hardness
values. These results coupled with transmission electron
micrographs of the altered material's structure provide insight
into further understanding of the mechanisms for changes in

mechanical properties of the surface.

24

.HmHH oHEmuoo ca omop cOH ocfl>um> mo poowmm .m musmflm

.NnEo . mco: mmoo

0. 20. :0. 20—
_ _

_ _

  

I. ‘ l I l I ‘ I l 8’ I l I.
I. l I. 1 II. ' I. [I I I I l— I
l
I I. l I l I I l l l I I l

t ' III II. I [ll

.. .5023084 I s i - ..

05:0320

l
l
l
l
l
l
l
t
l

t.\\\5\\vhkmkk\\\\hhk
_

H
ZO_Omm

I;

.. .- .. t .. 2.

EH " HH
_

_
20.051 20.0mm "

25

 

 

 

 

._mH. mcfiEDHm poucdeEAuHo mo mmmcpumc m>fiumamm .0 mucosa
Susa\§kz.wozw34u
”0.0.: O. m w v 0
_ 1 _ _ A 8
xnnnv _
l. I._ No
_ m...
l L. 0.0 w
u
_ m.
I. xoVwo Jr 2 H
\ w
0
. N
l. l n. 9
[x08 to)... l 0..
_ _ _ A t

26

_‘
c—A

 

 

A 400 keV N.
CE 10 ’- ~Al203 single-crystal
a Hood
9 9 T 0
Z .. 51. 5
-— 8 implanted
gt
8
:: 7 *
U)
‘9: 6r
3
3‘. s
E

L

or i l L_

 

 

3 . .
o 10" 10'5 (0'5 10'7 10'5
Dose (ions/cmz)

Figure 5. Flexural strength of implanted alumina
[l9].

27

 

 

 

 

l I l
0 U Arc.
o I N‘
E t
O 1.6
2
LU
(I
p.
U)
_l
<
(I
D 14 >-
X
W
.1
IL
LIJ
>
;
<
.J
u: 1.2 _
(I
POLYCRYSTALLINE ,"
l
/\O .I
’/
LO T ------ T '

 

1014 1015 1016 1017 101‘
FLUENCE (lONS/cm’)

Figure 6. Flexural strength of alumina
with ion species[20].

27

 

 

 

 

l l l
O 0 Aroo
o I N‘
E )—
15 —
3 SINGLE CRYSTALS
“4 e
E .'
U) I‘-
2’ ‘ l
C .I I‘
3 “ I . 7
El ' l
IL ,A .
g “‘~A" / ‘.
r- ' l
< .o.‘\; .
S L2 — . l —
a: e
POLYCRYSTALLINE ,"
/
/\O -"
’/
L0 3 T '

 

1014 1015 1016 1017 101‘
FLUENCE (IONS/cm?)

Figure 6. Flexural strength of alumina
with ion species[20].

 

12 .
Ni—>Al203(C-cut)
300 KeV, lxlO'7 crfiz . as implanted
:10” >- o 800 c
'5 53 2.... + moot
U P ‘ 1‘ .o o. A 13000C
Pb 8 ‘ ‘2 .° °.
r' A O
1 d o
u 5 A P 0
‘ ‘?
8 6p ‘ ‘ L ed 0
.. 1.0 ' , , '0
3 ‘ o .0: {O :'.: H ’92:..: 1:0 ‘ e
'6‘ Lb— . .0.9:09;‘ ’0 39......
A o O...’ O ‘ A .. .T,
LC) 0’ o A ‘ 9 ’.
O 1 o. J t 3 9.
U 2" ‘ " °fl ‘: .o 3
— A ' . ‘ Q. N
z ‘ : 6°. “5 cat. '.
L 0 q ‘ ’..
O ”‘3’: IQ” l “‘1AAA 1 JV ' n

 

 

 

OJ
DepthCum )

Figure 7.
[19].

0.2

Nickel concentration in alumina

Knoop Microhardnoss (Spa)

16

 

29

 

(D 50 mm
L’ o 100 gm
bl] NCM
~-- unhnpbntad z’T—‘T‘\
#- ‘___‘\

'§-l-0-I-i-I-O-O—.-I-0-0—.Q-O-I. .-

   

 

I I; I I I 1 I

 

 

RT 1200 1300 1400 1500

Annealing Temperature (’C)

Figure 8. Knoop microhardness for YSZ [23].

 

 

27

1.8

 

 

 

 

I l l
O 0 Are.
a I N‘

I
c L5 ~ ~
g SINGLE CRYSTALS
'“ o
m .
93 ,‘l
2* ' i
s 1. -’ 1
x I . T
LU I ‘
: ‘jAA.
g “u" I ‘.
t- ' l
< -°- I .
3 L2 — \' l a
‘6 o

POLYCRYSTALLINE ,"

/
/\O I'-
Lo T ------ T '

 

10” 10‘5 10'6 10" 10“
FLUENCE (IONS/cm?)

Figure 6. Flexural strength of alumina
with ion species[20].

 

 

 

 

)2 .
Ni-?Alzo3(C‘CU1)
3OO KeV, lxlO'7 crfi2 ° 85 Implanted

:10- 1,, o 800 c
'2 5" 3:. + iooot
U _ A ‘4 0° 0° ‘ IBOOOC
50 a ‘ .2 o o.
r‘ A 9 O
X o
u A A P 0

. °.~
g 6*.- T A 5 .0. O
“ ‘.° . 0 .0
E A o .0: .r':.0 '91.:...:‘0.o. .
E L h- : #1:: '0‘. 1“ O. 0....

' 9 o 9
g .' I A A 9 ..
O 1 9 g I 3 1
U 2“ f I' ° ‘1 To ;
i A ' °' 3 .3 T"
5‘ 9: Tb 5‘ 8‘0 .3.
O 111‘ z” lqh l “‘AAAA 1 J9" ’0.”
O 0.3

0.1
Depthfium )

Nickel concentrat
[19] .

Figure 7.

ion in alumina

Knooo Microhardnaaa (GPa)

 

16

14

 

 

 

29
O 50 gm
'- 0 100 om
_ D NCM
_.... unimplanted A
. ..-- \
+-

   

1 l l 1 l 1

 

Figure 8.

1200 1300 1400 1500

Annealing Temperature ('C)

Knoop microhardness for YSZ [23].

 

 

1.4 ZINC SULFIDE

ZnS is a II—VI type ceramic that transmits infra-red
wavelengths in the 8 to 14 micron wavelength range. The zinc—
sulfer bond has 82 percent covalent character and an open atomic
packing structure which is present due to the ratio of cation to
anion diameters. Atomic diameters for zinc and sulfur are 2.7 and
2.0 angstroms, respectively (1 Angstrom = 10'10 meters). The
principle structure types are face-centered cubic zincblende,
referred to as sphalerite, and the high-temperature hexagonal form
known as wurtzite. Their indices of refraction, according to one
source [25], are 2.356 and 2.378 for wurtzite, and 2.368 for
sphalerite, with no indication of wavelength of measure. A second
reference reports the refractive indices for chemically-vapor
deposited ZnS as 2.2520 at a wavelength of 4 microns and 2.2005 at
10 microns [26].

The melting point for sphalerite is estimated as 1830 degrees
celsius although, under atmospheric pressure, ZnS volatilizes
without decomposing at around 1000 degrees celsius. This is the
temperature at which sphalerite has been observed to transform to
wurtzite [27]. The Handbook of Chemistry and Physics [28] gives
the lattice constant of sphalerite as a = 5.4093 angstroms and of
wurtzite as a = 3.814 and c = 6.2576 angstroms. The structure of
sphalerite is illustrated in Figure 9 [29].

Commercial zinc sulfide material is generally made from
hot-pressed polycrystalline compacts which have been processed

3O

31
from compounds of pure inorganic materials. These materials are
placed in a mold which is then subjected to hydraulic pressure.
Induction heating of the compact while under pressure raises the
material temperature high enough for plastic deformation, without
melting, to obtain a relatively pore—free optically dense product.
Aggregates of small crystals form upon cooling and close packed
grains with minimal grain boundaries result. The scattering of
radiation by grain boundaries is present, but optical transmission

is prevalent [30].

32

 

 

 

 

 

 

 

 

Figure 9. Atomic structure of zinc sulfide [29].

1.5 TRANSFORMATION TOUGHENING

Transformation toughening is a method of improving the
ability for a material to resist fracturing due to the propagation
of cracks within the material structure. All ceramics contain
flaws within the subsurface region. The flaws usually originate
during the processing of the ceramic and may exist as
imperfections in the atomic structure, as mechanical abrasion at
the surface, and in some cases, resulting from the difference in
coefficient of thermal expansion between two phases of the
material. According to Griffith [31] local stresses are
concentrated at these flaws and the ceramic structure develops
openings in the atomic planes to relieve these localized stresses.
Such openings are referred to as "microcracks."

Fundamentally, a crack propagates through a material when the
stress at the crack tip exerts a force that exceeds the fracture
strength of the material. Transformation toughening, according to
Evans [32], is regarded as a process whereby the crack tip is
"shielded" from the stresses which induce this propagation;
"shielded" referring to any method by which the crack tip stresses
are reduced.

It is important to make the distinction between the stresses
which induce initial microcrack formation, and those stresses
which drive the pre-existing microcracks to propagate. The first
originates within the material at inherent flaws which
produce lattice strain and a corresponding stress concentrated

33

34
at these sites. In the mechanical equilibrium state, all stresses
existing in the material matrix are lower than that required for
crack propagation, thus the cracks are dormant. The second stress
results from the external application of forces, ie. by physical
impact with the surface or by loading. Since these sources are
external, stress-induced transformation toughening is a useful
method for improving the performance of materials which are more
susceptible to damage at the surface than in the bulk.

The stress above which a displacive phase transformation is
induced is the critical transformation stress, OCT. Once this
diffusionless transformation takes place, the volume of the new
structure may be greater than or less than the initial phase,
depending on the resulting structures atomic arrangement.
Diffusionless refers to atomic movement which is less than one
interatomic spacing. For transformation toughening to be imposed
upon a material, the phase transformation must result in an
increase in volume to the new phase. At the tip of a propagating
crack, this volume increase creates a constraint upon the material
matrix and compression of the atomic structure in the zone
surrounding the crack tip. This lowers the strain and the
corresponding stress to a level below that which is required for
further propagation of the crack.

From the previous section, it was seen that establishing a
residual surface compressive stress was the general principle for
improving the surface mechanical properties of
surface-sensitive ceramics. It was also seen that ion

implantation which resulted in formation of a second phase in the

35
sub-surface region experienced sustained surface enhancement.
These improvements were attributed to a compressive stress
established in the surface by the presence of the second phase.
It is likely, then, that the second phase structures were larger
in volume than the original matrix.

If a stress—induced transformation toughener could be formed
as a second phase in the surface of a ceramic by ion implantation,
the resistance to damage from external loads would be greatly
enhanced. Transformation to a phase with a larger volume upon
impact by a load could prevent fracture of the ceramic surface.
The following section introduces nickel sulfide as a

transformation toughener with potential for this application

1.6 NICKEL SULFIDE AS A TRANSFORMATION TOUGHENER

Millerite is the naturally occurring nickel sulfide at room
temperature and has a rhombohedral structure. It is also referred
to as B-NiS, whereas a—NiS is the hexagonal high—temperature form,
with the transition temperature reported to be at approximately
380 degrees celsius, and the melting point around 1000 degrees
celsius [27]. The composition by weight is approximately 65
percent nickel and 34 percent sulfur for stoichiometric nickel
sulfide. For a-NiS, the hexagonal lattice parameter ranges are a
= 3.438 to 3.426 angstroms and c = 5.348 to 5.316 angstroms for
the nickel-rich and sulfur rich compositions [27]. Figure 10
shows the nickel-sulfur equilibrium phase diagram.

Many materials have more than one crystallographic form in
which they can exist, their chemical and mechanical
stabilitydependent on factors such as temperature, pressure, and
stoichiometry. Pure zirconia, for example, transforms from the
high-temperature cubic phase to a tetragonal form at approximately
2400 degrees celsius, then to monoclinic at 950 degrees celsius.
A significant increase in volume occurs primarily for the second
of these transitions and results in the ceramic body crumbling
into powder, i.e., it self—destructs [34]. Typically, additives
to zirconia result in stability of this tetragonal phase. The
advantage this material then offers lies in the fact that the
phase transformation may be induced by an externally applied
stress. The stress at the tip of an advancing crack is

36

37
reduced by the positive change in volume that occurs upon
transformation, which results in a residual compressive stress.
The crack is "arrested" which results in improved toughness.

Similar volume changes for nickel sulfide were reported in
work by M.V. Swain [33] which involved the study of microcracking
in glass. The local regions of cracking observed in the glass
were identified to be in close proximity to inclusions of nickel
sulfide. A volume expansion of approximately 4 percent was
observed to accompany the microcracking. This change in volume
was attributed to the a -to-B phase transformation of a—NiS,
metastably retained in the quench-cooled glass. Transformation
was initiated during a subsequent heating of the glass as it
approached the transformation temperature of approximately 380
degrees celsius.

W. M. Kriven has explored several materials with potential as
transformation toughening alternatives to zirconia based upon
their crystallograpy and empirical results [34]. The criteria
recognized by Kriven for a transformation to be considered
"martensitic" (diffusionless) were that the particles undergo both
shear and volumetric strains. This translates to the distortion
of the material structure without breaking the atomic bonds. This
is a "displacive" transformation and is usually accompanied by a
change in volume due to the deformation of the lattice. Initiated
by Swain's work, research conducted by Kriven led to the
suggestion of NiS as a metastable stress-induced transformation
toughener.

Following Kriven's supposition, work was initiated to

38

evaluate the potential of introducing a-NiS as a stress-induced
transformation toughener in the surface of ZnS by ion implantation
and thermal treatment. The final system proposed in this work is
one in which nickel ions, implanted into zinc-sulfide, form bonds
with sulfur and precipitate as nickel—sulfide. The metastable
phase of a—NiS is desired as a fine distribution of particles in
the near—surface region of ZnS bulk material, a ceramic known to
be surface sensitive. When a stress is applied to the surface,
the associated shear and volumetric strains would then induce the
displacive transformation of a—NiS particles to B—NiS. This
transformation results in a positive change in volume, therefore,
the stresses at the microcracks which exist within the ZnS
subsurface region would be reduced and the cracks would not
propagate. Following one impact of the surface, the subsurface
region in which the transformations took place would retain the
residual compressive stress of the volume increase. Ceramics tend
to fail in tension, thus, a surface stress of a compressive nature
would further guard against mechanical failure due to crack
propagation in this region of the ZnS material.

Zinc sulfide has been ion implanted with nickel-beam ions of
a fluence of 1017 ions/cm2 at 200 keV. The goal was to form
metastable a-NiS precipitates in the ZnS surface region.
Annealing of the modified material was conducted towards this end.
Diffraction patterns of the specimens generated by
transmission electron microscopy were analyzed for evidence of
nickel sulfide precipitates. Additionally, hardness of the zinc

sulfide before and after implantation and annealing was determined

39
to investigate the effects on the mechanical properties.

Electron diffraction patterns did reveal the presence of a
second phase of very fine particles in the near—surface region of
the irradiated ZnS. Indexing of the pattern provided evidence
that a-Nis precipitants were present in the material. Increases
in hardness were also found for nickel-irradiated material that
was followed by heat treatment. Chapter 2 provides the detail of
the experimental procedures and Chapters 3 and 4 present the

results and conclusions, respectively.

4O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

WEIGHT PER CENT SULFUR
10 20 O 40 50' 60 70 no 90
1600 ' I ' 4 i 1 L J 1
1453. I REF. 1
“0° 1‘0 REF.“ '5 . .
\ '='=' ‘32.?-
z z
z_': #2
\ " 1 .. 1
.. x I l
:3 1000 N x\
u’ k \ / -._
“ t
a ”0 I fli— =1AT.
: ,,. . . ("fin—k
Id “A, ‘
u "f T 'l 14 l 1
'- 600 5350* (21.5)
r; r ' __._ 5 2, 44:13.00)
1 mass] 9 '3‘.
'°°—m~_-renfi--_--_-_. m—“m'nal’fi’fi
F”-—’ I orsu
_ coo. souo utsfl
mo j I
2" .3 :7: ~:'. "1’;
0 E '2‘ i"
O 10 20 30 40 so 60 70 IO 90 1”
Ni ATOMIC PER CENT SULFUR 3

Figure 10.

Nickel-sulfur phase diagram [27].

CHAPTER 2 EXPERIMENTAL PRCCEDURE

This Chapter describes the methods used for specimen
preparation of zinc sulfide for ion implantation. Specimens were
prepare for both TEM analyses and surface hardness testing. The
role of X—ray diffraction for improving the accuracy of an
internal standard is discussed as well as use of this standard in
the work completed by transmission electron microscopy. TEM
patterns were generated in the selected-area diffraction mode,
along with bright- and dark-field images of the sub-surface region
in the nickel-irradiated ZnS. Indexing of the diffraction
patterns is discussed briefly. Vicker's hardness testing was
conducted to measure the change in mechanical properties for the
irradiated samples and scanning electron microscopy provided a
close look at the surface grain morphology. X-ray photoelectron
spectrosc0py was conducted on experimental and control specimens

to investigate the presence of various constituents.

2.1 SPECIMEN PREPARATION

Two geometries of zinc sulfide ceramic material were prepared
for TEM analysis and hardness testing, respectively. A core saw
had been used prior to this work to cut 15 millimeter "plugs" from
a disc of zinc sulfide (Irtran II, Eastman Kodak Co., Rochester,
NY), approximately 25 centimeters in diameter. Four of these

were dedicated to ion implantation and four were retained as the

41

42
control samples for the surface hardness evaluation.

Once the plugs were removed, the remaining scrap material
shapes resembled a parallelpiped with dimensions of approximately
25 centimeters in length and 5 centimeters width and height. An
initial polishing test was conducted to determine the resistance
of the material to removal by a rough medium. It was found that
the ZnS could be easily shaped and formed by hand polishing with
400 grit polishing paper.

The method for obtaining the desired TEM specimen geometry
was developed for the ZnS material with the parallelpiped shape.
Though the cross-sectional geometry of these were not circular, as
is typical for metallic TEM specimens, it was concluded that a 3
millimeter diagonal of the parallelpiped cross-section would be
sufficient for securing a specimen in a TEM holder. After further
work it was found that the corners of this shape experienced
stress concentrations too high for reasonable handleability of the
ZnS ceramic material. These specimens fractured during transfer
operations using TEM tweezers.

A revised shape was obtained by carefully grinding the
material by hand on 600 grit polishing paper. The finer grit was
used to remove the corners and reduce the surface roughness to
achieve an approximately cylindrical shape, thereby decreasing the
possibility of damage to the material. Distilled water dispensed
from a plastic squeeze bottle was used for a coolant and to carry

particles away from the grinding paper.

43
2.1.1 Cutting

A diamond blade variable speed saw was used to cut specimens
to their initial TEM thickness. Figure 11 illustrates the
arrangement for slicing. Zinc sulfide was mounted to a clean
glass slide with thermoplast, a heat setting polymer-resin
commercially available (Beuhler) as a solid "stick" at room
temperature. Heating the slide with a Beuhler hot plate to
approximately 170 degrees celsius, and applying thermoplast to the
slide surface resulted in a small pool of thermoplast upon which a
specimen was located. Removing the heat source solidified the
resin, thus adhering the cylinder of zinc sulfide to the slide.
The glass slide was then mounted in the cutting saw clamp. This
arrangement, prevented distortion or damage to the specimen, and
made available the entire specimen bulk for sample preparation.
Approximately forty samples were cut from this arrangement,
ranging in thickness from 1000 to 700 microns. All transfer of
the specimens in this procedure was performed with nonmagnetic TEM
tweezers, size 3c (A. Dumont & Fils, Switzerland).

Following the removal of the specimens by reheating the glass
slide, the specimens were successively cleaned ultrasonically in
warm water with a mild surfactant, acetone, and ethyl alcohol to
remove contaminants such as cutting oil,thermoplast, and ZnS
residue. Care was exercised to expose the specimens to ultrasonic
energy transmitted only through the fluid medium (water) to
minimize exposure to the transducer vibrations. This was
accomplished by suspending a glass beaker of water and the samples

into the bath, making no contact with the reservoir surfaces.

44
2.1.2 Polishing

A polishing block was prepared to aid the polishing of the
ZnS ceramic specimens. A cylinder of aluminum alloy,
approximately 3 centimeters in diameter and 2 centimeters in
height, was polished on 320, 400, and 600 grit polishing paper,
successively. To ensure uniformity in the surface of the
polishing block, the polishing papers were attached to a
photographic plate approximately 30 centimeters square. A light
microscope at 50X was used to inspect the polishing progress.

TEM specimens were then mounted to the polishing block with
cryanoacrylate (Super Glue Corporation, Hollis, NY) in groups of
five. Care was exercised in arranging the specimens on the block
to promote an equal distribution of pressure throughout the hand
polishing steps. The glue was cured for one hour at room
temperature.

The polishing procedure included six different polishing
media. First, the specimens were coarse-polished on 600 grit
paper attached to a photographic plate. This was done until
"rocking" of the block-specimen arrangement was eliminated, to
assure uniform contact of each specimen. Preparation of the
surfaces continued with a fine, 1200 grit polishing paper. This,
too, was secured to a photographic plate.

The last four polishing steps were performed on motor driven
polishing wheels. The "laps" or cloths used on the wheels were
either nylon or "Micromet" polishing discs; both were backed with
adhesive. The media used on these wheels were 5 micron and 1

micron polishing alumina (A1203) on the nylon discs and 0.3 micron

45
alumina and 0.05 micron Buehler deagglomerated alumina on the
"Micromet" discs. Duration of polishing for each grade of
polishing media was approximately five minutes. After attaching
the zinc sulfide specimens to the aluminum block, and between each
polishing step, the samples were cleaned using the following

sequence:

* water rinse 2 mins
* distilled water rinse 30 secs
* distilled water ultrasonic 20 secs
* ethyl alcohol ultrasonic 20 secs

The appearance of the surfaces of the zinc sulfide specimens
changed with the polishing procedure from opaque, to translucent,
then to transparent with a slight hue of greenish—yellow,
attributed to normal scattering by inherent defects.

Upon completion of polishing one side, the specimens were
removed from the aluminum block by soaking in acetone for 2.5
hours, then remounted to permit a rough polish of the other
surfaces using 1200 grit polishing paper. These surfaces would be
prepared further with dimple grinding and ion milling, following
the implantation process, thus, the 1200 grit paper alone provided
the necessary finishes. This was followed by the procedure
described above for removal,then cleaning, and storage in a TEM
specimen holder. Once both sides of the specimens were polished
and approached TEM geometry requirements, they were divided into

two groups of twenty. One group was selected as the control,

 

. ..
”’EQOI

= u
1C3.

2.1.

2'4
“UV

‘3‘ v.

A-‘r.
.

1 c

A)!

ll)
4.

LJ.

'7"

‘A

‘¢

 

.1“

ll:
1 n

46
while the other specimens were prepared for exposure to the ion

implantation processing.

2.1.3 Mounting

An aluminum plate with thickness of approximately 0.5 inch
and measuring 6 inches by 6.5 inches, was designed and machined to
be used as the specimen's mount that would be attached to the
chill block in the ion implantation chamber. This plate and the
zinc sulfide sample arrangements on the plate are illustrated in
Figure 12. The plate was cleaned with a water—detergent mixture
followed by iso-propyl alcohol. To remove any corrosion products
remaining on the aluminum surface, the plate was soaked in
hydrofluoric acid for 3 minutes. Rubber gloves were used for
further handling and transfer of the plate following the acid
treatment to minimize contamination.

The twenty TEM specimens selected for ion bombardment were
attached to the aluminum plate with electrical and heat-conductive
silver paint in the arrangement shown in Figure 12. The four ZnS
discs 15 millimeters in diameter and 3 millimeters thick were also
attached. These were used for hardness and annealing experiments.
The discs and cluster of TEM specimens were contained within a 2
inch diameter centrally located on the aluminum plate. The plate
was heated in a low-temperature oven for 1 hour at approximately
90 degrees farenheit to aid in the drying of the silver paint.
IUEter cooling, the plate was placed in a single "Ziploc"
IPOlyethelene bag which was filled with liquid-nitrogen vapor and

Sealed. A second bag containing several dessicator capsules and

47
surrounding the first bag was similarly filled and sealed. A
corrugated cardboard structure was designed to trap the liquid-
nitrogen vapor in a "pocket" over the specimens. Additional
packaging materials were used to protect the aluminum plate and

ZnS specimens during shipping to the ion beam facility.

2.1.4 Ion Implantation

The specimens prepared for ion treatment were shipped to
Implant Sciences Corporation in Danvers, MA. Nickel ions were
implanted into the ZnS substrate at a dose of 1x1017 ions/cm2 with
200 keV of energy and a beam current of 165 x 10’6 amps. The
angle between the beam direction and horizontal was 90 degrees.
Freon cooling temperature was minus 45 degrees celsius and freon
cooling pressure was 9 pounds per square inch (psi). The beam
aperture was open 100 percent and the inert chamber gas used
during the implant was chlorine at 0.85 psi.

The surface of the implanted specimens as—received appeared
to be dark purple or scarlet in color. A similar but lighter
appearance was observed on the 2 inch diameter of the aluminum
plate. One of the TEM specimens was lighter in color and it
appeared as though partial masking of the beam occurred, for this
sample was located at the edge of the beam diameter. It could be
assumed that this specimen received a dose of nickel lower than
1017 ions/cmz.

Soaking in an acetone bath dissolved the silver paint and
permitted removal of the specimens from the aluminum base.

Ultrasonic cleaning in acetone, then absolute ethanol, preceded

48
storage of the TEM specimens in a TEM "grid—storage" container.
The 15 millimeter discs were stored in plastic laboratory
containers each filled with a layer of dessicant, then cotton, and

laboratory tissue paper.

2.1.5 Dimple Grinding

Further specimen preparation steps were necessary following
ion implantation to make TEM specimens of the 3 millimeter discs
of nickel—implanted zinc sulfide. Two separate methods typically
used for ceramic TEM specimen preparation were attempted. One
utilized dimple grinding followed by ion milling to physically
thin the material such that an electron transparent region exists
adjacent to a perforation edge. The other involved chemistry
following dimple grinding which resulted in a thinning effect by
'etching.

Dimple grinding is typically used to prepare specimens for
ion milling. The grinding provides material removal rates far
exeeding those achieved with ion milling, thus reducing the
duration necessary for the latter process which is both, more
precise in material removal (atomic effect), and more expensive.
The "dimple" or bowl-shaped contour of a previously flat specimen
surface is achieved by grinding material away from the specimen
with a fine diamond paste slurry carried on a brass
wheel. The specimen rests on a rotating stage during grinding,
thus, forming a dimple in the specimen with the lowest (thinnest)
point to be the target for ion milling.

The nickel-irradiated zinc sulfide specimens were

49
approximately 600 microns thick at this step of the preparation.
Typically, a specimen thickness conducive to dimple grinding is
between 50 to 100 microns. Therefore, prior to grinding, final
thinning and perforation, the bulk of the back side of each of the
irradiated ZnS specimens was reduced to yield thicknesses of
approximately 80 microns.

This was accomplished with a Gatan Disc Grinder 623, a hand-
held polishing unit consisting of a 0.375 inch diameter
cylindrical specimen mount centrally inserted into a 3 inch
diameter polishing base (see Figure 13). A rotating dial (control
knob in Figure 13) with a scale mounted on tOp of the base
indicates the thickness of the specimen attached (with melted wax)
to the specimen mount. The specimen surface is polished by
rotating the dial which lowers this surface just below the
polishing unit's face. This permits specimen contact to abrasive
paper affixed to a glass plate. The abrasive paper used was
Buehler 1200 grit polishing paper.

A top view of the Gatan Dimple Grinder-Model 656 is shown in
Figure 14 (Gatan, Inc. 6678 Owens Dr. Pleasanton, CA 94566). Once
the desired thickness is achieved, the specimen mount which was
used in the hand polisher is placed in a metal ring on the
specimen turntable platform of the dimple grinder. The grinding
wheel depth is adjusted with the micrometer drive, and the dial
indicator shows the thickness of the material removed in microns.
The speed of the turntable and the counterweight can be adjusted;
optimum settings must be determined empirically. The speed

control was set at 5 and the counterweight used was 1 gram for the

50
zinc sulfide specimens.

Initially, a standard specimen mount and a 15 millimeter
phosphor-bronze grinding wheel were used to attempt preparing TEM
specimens for ion milling. These are illustrated in Figure 15,
along with the smaller 10 millimeter wheel and modified mount.
The larger wheel removed too much of the specimen near the
perimeter for adequate strength to be retained. This resulted in
specimens fracturing during handling with either TEM tweezers or
with vacuum tweezers. The smaller wheel formed dimples with
diameters significantly smaller than the 3 millimeter diameter of
the specimens, thus retaining bulk material at the perimeter.
Vacuum tweezers were used successfully in the handling and

transfer of these fragile samples.

2.1.6 Chemical Thinning

According to P. Worthington, [36] zinc sulfide can be thinned
by a "specially developed chemical thinning method."
Worthington's success with thinning zinc sulfide with chemical
flow was made possible by the manufacture of a glass enclosure
that permitted the flow of thinning chemicals across the ZnS, the
observation of the material removal progress with a back-
illuminating microscope, and a distilled water wash system to
stop the process immediately upon observing a perforation.

For the present work, a South Bay Model 550C Single—Jet
Thinning Instrument (South Bay Technology inc , 5209 Tyler Avenue,
Temple City, CA) was used to introduce the thinning solution to

the zinc sulfide specimen surface. This instrument consists

51
of a chemically inert plastic platform shaped to form a round base
approximately 8 inches in diameter. Molded within this plastic
are housings for the pump and motor, inlet and outlet ports for
directing the chemical, and a specimen stage which included a
splash guard and manification lens for viewing the process. The
platform is grooved on the underside. A glass cylinder fits into
this gasket-lined groove and the thinning solution is poured into
and pumped from this reservoir.

The solution was prepared by pouring 900 milliliters of
hydrogen chloride into a 1500 milliliter Corning glass beaker and
adding 9 grams of chromium trioxide (CrO3). This was stirred with
a glass stirring rod for approximately one minute, then poured
into the jet thinner reservoir. To protect the metallic
components of the jet thinner, pressurized air from a chemical
hood source was run through a 1/4 inch flexible rubber hose to an
inlet in the thinning instrument's platform. The system beginning
at the inlet permits the pressurized air to form an envelope
between the chemically sensitive components in the jet thinner
platform and the chemical stored in the reservoir below.

Non—irradiated zinc sulfide specimens were used to test the
abilities of this technique. For irradiated specimens to be
thinned with this method, the irradiated surface must be protected
from the chemical. To test this, one surface of a non—irradiated
specimen was coated with Miccroshield lacquer using a fine oil
paint brush. The specimen to be thinned was then located on the

pedestal of a sample holder with a 2.3 millimeter hole to minimize

the exposure of the chemical to the specimen surface. A

 

52
polyethylene diaphragm with a centered hole was placed over the
specimen and held into a groove with a rubber ring.

The automatic shutoff feature of the Model 550C was not
utilized. Infrared optics are used for the detection of
perforation in thinned samples. Being transparent to infrared
radiation, specimen perforation was monitored visually with a hand
held 10x magnifier and an Olympus Transmission & Back-illuminating
optical microscope. The magnifier built into the splash guard,
which surrounds the pedestal and jet nozzle, was unable to provide
a view of the specimen due to the hydrogen chloride vapor.

A second chemical solution consisting of CrO3 and H20 was
used with the jet thinner to attempt making TEM specimens with the
zinc sulfide. A 3 x 3 disPO weigh boat was tared on a Mettler
Type H15 single pan balance (Mettler Instrument Corp. Hightstown,
New Jersery). CrO3 was placed in the boat with a stainless steel
spatula until 80 grams were measured. This was poured from the
boat into a polypropylene bottle with a screw top. A 250
milliliter glass beaker was filled to 100 milliliters with H20 and
stirred with a 2 inch stir bar and a Cole—Farmer Model 4812
Magnetic 6 x 6 inch Stirrer with Hot Plate. The chromium trioxide
was dissolved into the water by stirring the solution, heated to
85 degrees celsius, for 2 minutes. This solution was also used in
the jet thinner.

A nickel-irradiated specimen was polished by hand to a
thickness of 105 microns, then dimpled to yield a minimum
thickness of 20 microns. The polished side was coated with

protective lacquer and the specimen was loaded for jet thinning

53
with the CrO3/HC1 chemical solution. Using the jet thinner auto-
dispense timer, inspections were made at 10 second
intervals. After 4 minutes of thinning, it was determined that
perforation had occurred much earlier, but was obscured by the red
Miccrocoat lacquer. This perforation resulted in a specimen too
thick for TEM observations. Table 1 summarizes the chemical

thinning procedures and observations.

TABLE 1.

54

Chemical thinning durations and observations.

W

CfO3/H2O

CrO3/HCl

15

10

6O

10

17

3O

60

SEC

min

min

min

min

min

min

min

min

min

THINNING TIME

OBSERVATIONS

surface slightly
etched

no change on
lacquer

no change

preferential
material

removal

(appears "eroded")

little change

preferential
material

removal

approx 10 microns

slightly etched

dimple depression

deeper depression

further erosion

bowl shape evident

.NI

«3:

FL.

55

2.1.7 ION MILLING

Final perforation of the polished and dimpled nickel-
implanted zinc sulfide specimens was accomplished with ion
milling. This was done with a Maxmill Ion Gas Milling System,
Model 306/TM 170 (VCR Group Inc. San Francisco). The system
includes a liquid nitrogen cooled specimen stage and an automatic
perforation detector with variable sensitivity. The minimum beam
angle to the horizontal specimen plane that could be set for this
system was 18 degrees. An angle less than this would result in
detector—beam path interference from the cooling stage. Beam
angles of 18 and 30 degrees were used for ion milling the zinc
sulfide specimens. The argon back-filled chamber maintained a
vacuum of 2 x 10'5 torr. The liquid nitrogen stage temperature
average was approximatley minus 50 degrees celsius. An energy of
5 kilo—electron volts and gun current of 1.5 milliamps was typical
for the ion milling process. The speed of rotation of the stage
was fixed at approximately one-half revolution per minute.

Following ion milling the chamber remained undisturbed to
permit returning to room temperature without the aid of the
nitrogen gas warming mode. The objective was to minimize the
possibility for thermal shock to occur in the specimen. This
ambient warming averaged approximately 7 hours in duration.

Several arrangments were used to mount the specimens in the
ion milling chamber. The first held the specimen between two
washers of aluminum which were placed in the recess of a gear-

driven brass disc. A brass retainer which held the aluminum and

56
specimen in place was secured with three screws which threaded
into the bottom brass disc After ion milling, black residue was
apparent on the aluminum specimen washers. This may have been
deposited by the ionization of contaminants either existing in the
chamber, or carried into the chamber by the zinc sulfide or
aluminum materials, or both.

Platinum washers with central holes 2.3 millimeters in
diameter replaced the aluminum washers which are not as durable to
ion milling. The standard arrangement above was modified to
secure the specimens below the 2.3 millimeter hole in the absence
of stress by compression between two washers. This was
accomplished by using cyanoacrylate to adhere the specimen to the
bottom side of the washer with the non—irradiated specimen surface
below the hole and exposed to the ion beam for milling. This
arrangement is shown in Figure 16.

Attempts to use the auto—terminator were unsuccessful due to
instrumentation instability. Many attempts were made to stabilize
the detector with the ion beam at varying levels of sensitivity
and gun voltage. This may have been due to excessive use of the
mill with many different materials resulting in a high rate of gun
contamination.

Following the preparation of a potentially electron-
transparent specimen, the specimen was placed in a 50x50 copper-
mesh Gilder folding grid, (Ted Pella, Inc. PO Box 2318, Redding,
CA” Catalog No. GD 50/50) for TEM observation. This procedure was
observed with the aid of an Olympus Type JM back-reflection

rnicroscope (Olympus Optical Co, Ltd.) and Dupont Type 3c tem

57
tweezers. An electric vacuum-tweezers apparatus with a one
millimeter nozzle was used to transport the specimens from a glass
slide to the Olympus microscope stage for enclosure in the copper—
mesh TEM grid. Pella TEM specimen holders were used to store the

specimens, which were then stored in a dessicating cabinet.

58

.mcoEflooam Ema now mew OCAOAHm

pm<4102mm1H

 

.HH musoflm

(EmIi
QZOEflQ

 

m 9 4m fl
mm<i_o

59

.coaumflpmuufl you ucoEmmcmuum cmefloodm mam

\\51 l (1..

.\

a

 

.N; cHDOAA

 

mm102_N

 

.A-slir -;:(,;1¢;-v.

     

00000 00000

   

OOOOO

 
 

OOOOO

   

/

 

6O

Specimen
ejector

 

 

Hex set screw

 

. I“ 111““

‘/
Eli ILLIAI ’- .-

 
    

 

Retaining
ring

 

 

   

Ejector pin
#:0161010

 

 

 

  

 

Reversible ,
scale

 

vvvvv

 

 

Base—N ‘

 

 

...............

 

 

 

 

O-Rin Specimen Wave
#010g mount washer
623-0008

Figure 13. Gatan hand-held disc grinder[35].

 

61

counterweight

dial indicator

 

load scale

 

+y adjustment

 

F \

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

/ ..
/ \ return spring
. / adjustment
- \\ -:géE;:wg men /1
+x adjustment .,\
\ ABLE
©\ —-ry adjustment
raLse / lower __ N.‘ muminator
cam pivoted SPEED power socket
1 “‘“°"“ 0
me .
zero set ’ ‘ grinding
screw 1 j wheel
G” O
o
DIMPLE BINDER
k / \ \ J
adjustable micrometer bearing specimen plattorm
micrometer drive drive assembly over magnetic turntable
scale 656-0150

Figure 14.

Gatan Model 656 Dimple Grinder
(top view)[35].

 

@ 10mm

- :1me

L l

Specimen Mount 656-0116

 

eze-oooe -—~‘ L“ 10W

 

 

 

 

 

 

 

Standard Specimen Mount

Figure 15. Dimple grinder specimen mounts[35].

63

IBM:
m<mo

.COHuommlmmouo umpaon ccefioman Add: coH .mH ousmwm

2925QO

 

_ 1|.
—[ [— —[

 

Norm/E 02E
52:50 Ezzmm

 

§<mm ZO_
Hzmgoz.

’1

2.2 X-RAY DIFFRACTION

Zinc sulfide was used as an internal standard for indexing
the diffraction patterns generated by the zinc sulfide irradiated
with nickel ions. The reflection spots of a transmission electron

microscope (TEM) diffraction pattern are generated due to
Constructive interference of the waves of an electron beam which
pass through a small, selected area of a specimen, thus they are
C a lled "selected-area diffraction" (SAD) patterns. As they travel
1: hrough a TEM specimen, the beam is scattered by parallel atomic
p l anes in the specimen with interplanar spacing, d.
The known d-spacings for a standard material may be taken
f rom the Powder Diffraction File, or "PDF" [37] , or they may be
de termined by other experimental techniques, such as x—ray
d i ffraction (XRD) or x-ray microanalysis. The interplanar d-
Spacings given in the PDF for a selected material are those
Spvacings which were found using these and other methods for
particular specimens prepared under specific conditions, thus they
represent approximations. Further variations may occur due to the
p‘L‘esence of minor amounts of impurities.

X-ray diffraction was implemented to determine the actual d—
Spacings of the zinc sulfide ceramic used in the present work to
i"“Frrove the accuracy of the indexing calculations for the TEM

diffraction patterns. Diffraction peaks in the XRD pattern of a
material indicate the significant reflections from particular
plane orientations at specific 26—ang1es, the orientations being

64

 

65
an indicator of the material constituents.
A round specimen 15 millimeters in diameter and 7 millimeters
chick was placed in a holder made of corrugated plastic in which a
14 millimeter hole was cut. The surface of the ZnS protruded
approximately 1 millimeter above the holder surface. The surface
of the zinc sulfide was left in the "as received" condition, that
i s , the surface was macroscopically smooth but not optically
t: ransparent.
A General Electric X-ray diffractometer was used to project
X — rays into the zinc sulfide. The specimen surface relative to
C he x-ray beam was rotated between the 29-angles of 25 and 150
de grees at 2 degrees per minute. Detection sensitivity was 2000
C cunts per second and the time constant was set at 1.0. A second
Spe cimen was exposed to x—rays with identical instrument

parameters. This specimen had been implanted with nickel ions at

a dose of 1017 ions per square centimeter and 200 kilo—electron
VQ :Lts (accelerating voltage). The irradiated specimen diameter

a as approximately 13 millimeters with about 3 millimeters

thickness.

The x-ray diffractometer signal output graphs the reflection
intensities versus the diffraction angle, 29. Bragg's Law [38]
was used to calculate the d—spacings for the 15 largest
diffraction peaks from the diffraction angle and the wavelength of
the x-rays emitted from a copper source (I. = 1.542 angstroms).
These were then correlated to the d-spacings and their respective

diffracting planes for B—ZnS (sphalerite) listed in PDF Card No.
S-
0566 [37].

 

 

 

66
An additional resource for indexing diffraction patterns are
the standard spot pattern monographs of Rarey, Stringer, and
Edington [39]. These illustrations provide information on single
c:rystal patterns for various structure types including diffraction

ignlanes, radius ratios, and angles between reflections.

 

2.3 TRANSMISSION ELECTRON MICROSCOPY

A JEOL 100CX transmission electron microscope equipped with

STIEM (Scanning Transmission Electron Microscopy) and a LINK EDX

(Ianergy Dispersive X—Ray) analytical system was used for all TEM

Cplsservations of the nickel-irradiated and non-irradiated zinc

ssrLJlfide specimens. The imaging mode, in bright- and dark—field,

éaLJ:Ld the selected area diffraction (SAD) mode were used. A

t:.1_1aagsten filament was the source for the electron beam with an

.53.<::r<:elerating voltage of 100 kilovolts (100 x 103 volts).

Objective apertures 1 through 4 were 120, 60, 40, and 20

:1r1.:i_ <::3:ons in diameter, respectively. Objective apertures 1 and 2

‘R’NEE=-:2:-ea used for most of the observations in conjunction with a 10

FTTLZi. <::tz:on condenser aperture and a diffraction aperture of 20

"llJ_.<::::r?ons.

Preparation of the JEOL 100CX transmission electron

r1 ‘ . . . .
:1. <:::2:\oscope for speCimen observations was implemented according to

s t:: .
iEELJFJJdard TEM operation procedures. Vacuum tweezers were used to

ruc2>
T‘“’T<Ei specimens from a TEM specimen holder to the spring—loaded

AS:E;>
TEE=‘<:2:imen stage. This stage included a +/- 60 degree tilt

Capability.
Micrographs of images and diffraction patterns were taken
TUITTT 1::;:k3- Kodak 4489 electron microscope film. The majority were taken
::Tt:L ‘rrhagnifications of 72,000 or 100,000 in imaging mode, or with a
QTtTt‘era length of approximately 55 centimeters (55 x 10'2 meters) in
ESHEEE :JLI

‘53<:ted-area diffraction (SAD) mode.

67

‘

68
As discussed in Section 2.2, x-ray diffraction was
implemented to determine d-spacings for the zinc sulfide used in
this study. The ZnS structure, intrinsic to the effects under
study, acted as the internal standard for each of the electron
diffraction patterns of the ion implanted material. The thinned
region of the TEM specimens consisted primarily of zinc sulfide,
in spite of the nickel irradiation of the surface. Therefore, the
d i ffraction patterns generated by the irradiated material included
t: he reflections of the zinc sulfide structure and those
re flections generated by any new phase or species present due to
t: he introduction of nickel by ion implantation. An accurate
de termination of the d-spacings for new reflections was achieved
by calculating the calibrated camera constant for each individual
a l e ctron micrograph with the information provided in the pattern

by t. .he internal standard material.

 

 

2.4 INDEXING TEM DIFFRACTION PATTERNS

Transmission electron microscope diffraction patterns of a
"selected-area" of the specimen of nickel—irradiated and non-
irradiated zinc sulfide were indexed to obtain information
relevant to identifying the phases or species in the subsurface
region of the ZnS. Relative interplanar angles, pattern
orientation, and d-spacings between planes, which are determined
by radii measurements, all contribute to this analysis.
An Intertech hairspring divider, a compass-like precision
measuring device, and a transparent millimeter scale were used to
me asure the distances of diffraction spots from the central spot,
and the angles between them. These measurements were made on the
etT"I‘L13_sion side of the negative of the micrographs to eliminate
a rrors which may occur when measuring distances on a contact
Dr int. The radii were determined by measuring the distance
be 1:-V«reen spots symmetric about the center and dividing these values
by 2
The ratio of radii to focal length, R/L, in the electron
hi Qroscope has been shown to correlate to the ratio of
ac Celerating electron wavelength to d-spacing [40]. Using this
tionship, the interatomic spaCings for the non-irradiated Zinc
SR 1 fide pattern were calculated with A = 0.037 angstroms, the

”a
V- .
elength of electrons for an accelerating voltage of 100

“til

Qvolts [40], and a non-calibrated camera constant of L = 550
"i
3‘ limeters.

69

‘

 

 
 

rt. .1. rte u u l . A: Y1 Dc I he . P» «D «5
Hi. .2 C i .10 P0 .50 ”V rm. .«la 40 r; any at .‘ I n p ..
.1 . 0n ,1. Ft. al. .00 . . re. .u 1.. . . vi .n .. .n... 0.0
.q a 7.. 4.0 2a .5 .1 we . .C .. . «Ha I. 5 c . . we «\c n .

Y.

 

70
These d—spacings were then compared to those given by the x—
ray diffraction results and similarities were matched. A single
d-spacing from the XRD data was assumed as the most accurate based
on correlation to a significant reflection in the electron
diffraction pattern. Knowing the wavelength)», the spot radii, R,
and a d-spacing for the actual ZnS material used, gave a more
accurate calibrated camera constant, L. Revised values for the
additional d-spacings were then calculated. The reflection planes
g iven by PDF were correlated to these results, and data from the
appropriate monograph [39] were used to confirm the pattern
i nciexing.
For the electron diffraction pattern of nickel-irradiated

ZnS features identified as representing the internal standard ZnS

S f; ructure were indexed as above. Additional features resulting
f rom the ion implantation were then indexed in a similar manner.
The d-spacings were calculated with the calibrated camera constant
”11 1 Ch was determined by the measurements of the internal standard.
S‘L'2'r—1':Z>Sequently, various references for interplanar spacings were
DQJ37.-Sulted to identify the phases or species present in the

"‘Q -
@- lfied specimen which exhibit the reflections in the electron

Q i
:8 fraction pattern.

 

.u u . I. . a c I 7‘ F C P c -s
5 A. .. w” a. w. 5 we. Cc a. a
I 3.. c . Aime nu n x I ”be .. a “L. ..
2 . a" 0.. . . ”A 2 F .ql Ce :0

 

I
,‘C‘r
“

 

2.5 HARDNESS

To assess the influence of the surface modification upon
mechanical properties of zinc sulfide, the Vickers hardness of
nickel-irradiated ZnS was determined and compared with that of

non-irradiated ZnS. Further studies were conducted to examine
the effects of annealing upon the hardness of implanted and

unmodified zinc sulfide .

2 - 5 .1 Nickel-implanted ZnS

Discs of zinc sulfide 15 millimeters in diameter and
approximately 5 millimeters in thickness were prepared, as
d i S cussed in Section 2.1, to provide both irradiated and control
Sp e Cimens for evaluating the influence of the irradiation and
S1-1~h>sequent thermal treatment. Following the implantation
pro cedure, the irradiated and unmodified discs were each attached
: Q glass slides with a low-temperature thermoplast. This
at rangement was positioned above the saw blade with the same
Dbehtation to the blade as that for the TEM specimens (see Figure
L 3“ ) - Post—cutting sample dimensions approximated parallelpipeds
wi Ch top surface areas of about 30 square millimeters. This
Qu Q ting was completed with a diamond wheel saw blade on a Buehler

(in; e
I'TTQnd saw with standard cutting oil used as a lubricant and

Q
<:>.<::

Q ling agent. Cutting oil contamination was removed by soaking
11%

Q§
%
1Zone and alcohol ultrasounds followed. Specimens were stored

Specimens in warm water with a mild surfactant. Separate

71

‘

C’ A)
otb.
Afr:
vy‘ug

. 739
_‘ 1A..

A
N

‘l
A»
~3.g

4‘" '-
-»

fraCt

 

72
individually in S dram glass vials with screw tops which were then
stored in a non—evacuated dessicant chamber.

Due to the depth of the indents made by the Vickers hardness
trester, actual hardness values represented those of a composite
c:c3nsisting of the thin implanted layer and the unmodified bulk.

iIirle sensitivity of the indentor to hardnesses of the modified sub—
ss'Llrface region was maximized by using the lowest available load of
:3 .kilograms—force for most of the hardness testing. The load time
cEiJITl<j load speed were set at 15 seconds and 50 microns of indentor
t: :t:‘Eavel per second, respectively.
Spacings between respective indentations and the sample edges
“"<EE=LJ:?63 maintained at approximately 8 times the half-diagonal length
<:>§155- t:he indents for each of the specimens tested [41]. Specimens
“’"EE=‘JET‘53 soaked in acetone, then alcohol prior to hardness testing and
EE'Esi-‘ZZEJ:1 specimen was transferred to the specimen platform with Dupont
3 Q TEM tweezers.
Hardness was determined by entering Vickers indent data
( z\Zplpendix A) into a Basic program (Appendix B) which used the
E <::> :JL-ZJLcowing formula to calculate hardness in gigaPascals:
H = .464P/a2 (2)
thI‘e, H = hardness, GPa (1x1012 Pascals)
P = load, kilogram-force
a = average half—diagonal length of indent
.464 = material constant for ZnS [41].
=3; Toughness values were calculated for the non-irradiated
§§Qimens however, the decrease in contrast of the indented

%
Q .
(:H:ure crack-lengths to that of the surface fOllOW1ng

II‘IIIIIIIIIIIIIIIII-.__

‘I
;.

 

>—
E‘-

'1

.11
I I

‘l
LII

 

—
n
‘.

*-

'I‘J‘.

~

:3

't

A
v5r
‘ -

73
implantation and annealing was significant, and prevented

measuring the crack lengths for the nickel-irradiated ZnS.

2.5.2 Annealed Nickel-implanted ZnS

Annealing of nickel—irradiated zinc sulfide was implemented
to initiate thermal activity such as, radiation—enhanced
diffusion, and to enhance precipitation of nickel sulfide. To
test the effect of annealing upon the surface-modified ZnS with
respect to mechanical properties, a control specimen of ZnS was

included with an irradiated specimen for each anneal cycle, and

 

"(Al'zu .
4‘

pre-and post—anneal hardnesses were measured.

Annealing cycles were run in a Centorr Series M60 Front
Access vacuum furnace (Model No. M60-3x8—W—D-04S4-A-20, Centorr
Associates, Inc, Route 28, Suncook, New Hampshire 03275; 603—485—
9504). The furnace stage upon which the specimen crucibles were
placed was a tantalum plate which was located approximately 3
inches from the thermocouple. The initial chamber temperature was
15 degrees celsius and the furnace chamber was maintained under
constant argon pressure such that argon gas flow from the chamber
through a gas venting valve was approximately 10 cubic centimeters
per minute.

The specimens were ultrasonically cleaned with acetone and
200 proof alcohol, then placed in alumina crucibles (Coors

Ceramics Company, Golden, Colorado) with dimensions of 30 x 9 x 7
millimeters. Dupont Type 3c TEM tweezers cleaned in alcohol were
used for transport of the specimens. The crucibles were placed on

the tantalum stage in the furnace chamber with Teflon coated lab

a
-
"V

 

 
 

:. w.“ .9 ...fl . . .Hh .s; wL. .3 .3. C» . . L1.

74
tongs. To minimize transfer of volatiles or other species between
treated and untreated specimens, each specimen was placed in an
upright crucible with the surface for hardness testing facing
upward. This was then covered with another crucible of the same
size, inverted such that the edges of the "bowls" of the crucibles
met. Each such crucible arrangement was dedicated to irradiated
or non-irradiated specimen processing.

The chamber was evacuated to 50 millitorr, then backfilled
with argon. This procedure was repeated twice then followed by
initiation of the heating cycle. The three temperature profiles
which were implemented reached maximums of 750, 800, and 850
degrees celsius. An auto-controller unit was used to set the
profiles such that the ramps to maximum temperature were
approximately the same. The objective was to reach this maximum
as rapidly as the furnace/controller unit could accomodate while
still maintaining relative uniformity in the ramps to the peak
temperature between each profile. Upon reaching the target
temperature, the heat zone power was terminated and the argon flow
was maintained for approximately one hour. The three thermal
profiles used in the annealing of the specimens are shown in
Figure 1?. Specimens were removed six to eight hours later using
Teflon-coated tongs and TEM tweezers. Each was returned to their
respective glass vial and stored in a dessicant cabinet.

Vickers indent data and respective hardnesses for the
annealed specimens were obtained using the same procedure as
<iescribed for the nickel—irradiated zinc sulfide samples. This

data is also included in Appendix B.

 

Temperature (Degrees C)

900

800

700

600

500

400

300

200

100

75

 

 

I'l'l'l'l'l'l'lrl'l'Irl'l'l'l'lrl'

 

I 36 deg/min +750C
O 30 deg/min +800C
A 34 deg/min *BSOC

 

 

 

 

 

1 l l I l I l I l

 

O

10 20 3O 4O 50
Time (Minutes)

Figure 17. Anneal profiles of Ni-irradiated ZnS.

 

2.6 SCANNING ELECTRON MICROSCOPY

Nickel-irradiated zinc sulfide specimens underwent SEM
(scanning electron microscopy) and EDX (energy-dispersive x-ray)
analyses before, and after thermal annealing. Post-annealed non-
irradiated ZnS specimens were analyzed as well. SEM micrographs
of the zinc sulfide surfaces were obtained with a JEOL JSM-35C at
20 kilovolts. A Tracor Northern D500 X—ray system provided
information on the presence and relative concentration of nickel
in the ZnS utilizing a Standardless Quantitation computer program.

The annealed and non-annealed nickel-irradiated zinc sulfide
specimens were prepared for SEM by attachement to aluminum SEM
specimen mounts with a small drop of Conducting Graphite Paint
[Ladd Research Industries, Inc., Burlington, VT]. These were then
placed in a low-temperature oven set at approximately 60 degrees
celsius for approximately seven minutes followed by ten minutes of
cooling.

The areas scanned by EDX for each specimen was approximately
10,000 square microns. Tilting angles of the detector were 27 and
0 degrees with a detector distance of 55 millimeters and a working

distance of approximately 39 millimeters.

76

“h.
..\v

Q

va;

7‘7-

k3;

».

and

n-5

 

2.7 X-RAY PHOTOELECTRON SPECTROMETRY

Surface analysis by ESCA, or XPS (X-ray Photoelectron
Spectrometry), involves irradiating the surface of a specimen with
monoenergetic soft x—rays and sorting by energy the resulting
emitted electrons. Each element has a set of electron binding
energies that are unique. The kinetic energies of the inner—shell
electrons ejected from the surface are equal to the incident
photon energy minus the electron's binding energy at its
particular level in the electron shells of the solid.

A plot of the number of emitted electrons per energy interval
versus their respective kinetic energies yields the unique
elemental spectrum. The peaks in the spectrum from a compound are
approximately the sum of the elemental peaks from the individaul
constituents. The mean free paths of the electrons are very
small, therefore only the top few atomic layers actually emit
electrons from the specimen surface which are registered by the
detector. Typical mean escape depths range from 0.3 to 3
nanometers for a kinetic energy of 100 eV.

XPS was performed to measure the presence of nickel or nickel
sulfide in the subsurface region of the nickel-irradiated ZnS. A
Perkin-Elmer PHl 5400 Small Spot X-ray Photoelectron Spectrometer
was used for this analyses. Specimens were the round nickel-
irradiated ZnS samples, 15 millimeters in diameter and 3
millimeters thick, and the annealing specimens described in
Section 2.5.

77

 

CHAPTER 3. RESULTS AND DISCUS§ION

The results of the analytical work which lead to the
conclusions in Chapter 4 are reviewed and discussed here. First
is a preview of the role ion milling effects may have had in the
images of the microstructures. Quantitative results from x-ray
diffraction lead into the main work conducted with transmission
electron microscopy. Hardness comparisons between irradiated
specimens and specimens annealed following irradiation is
presented with a review of possible mechanisms. The mechanisms
proposed are discussed relative to results from scanning electron
microscopy and x-ray photoelectron spectrometry. This is followed

by a section on diffusion.

3.1 ION MILLING EFFECTS

Zinc sulfide specimens irradiated with nickel ions were
thinned by dimple grinder perforation followed by ion milling.
The procedure of ion milling involves the conversion of inert gas
atoms to a plasma of ions which are then accelerated towards the
target material at a glancing angle to physically bombard and
sputter the target atoms by transfer of energy. The glancing
angle of the incident ions results in removal of the atoms and a
"thinning" of the target. Eventually a perforation in the target
material forms and the area immediately adjacent to the
Perforation is transparent to incident electrons. The use of

78

79
high energies and the physical nature of this phenomenon result in
various effects which may be exhibited through transmission
electron microscope observations.

Figure 18 is a micrograph of an ion-implanted ZnS specimen
prepared by ion milling and observed in the imaging mode with a
JEOL 100CX Transmission Electron Microscope. Exposure time was
1.4 seconds. The view given by this micrograph is from a point
perpendicular to the specimens plane; this is the reference for
all of the following micrographs. At a magnificaton of 100,000x
(100k), the thinning effect ion milling had upon the specimen is
evident. The extreme edge adjacent to the perforation is the
thinnest part of the material. Moving towards the thicker region,
detail of the material structure is less apparent due to the
decrease in electron transparency.

In addition to thinning a material to electron transparent
dimensions, ion milling may also contribute undesired effects
which may alter the structural detail of an irradiated material,
thus prohibiting the analyses of a ceramic which has only
experienced the effects of ion implantation. Ion milling
physically removes atoms, therefore, an atomically rough surface
results. Variations in thickness may be present such that maximum
and minimum variations occur over very small distances, i.e.
nanometers. Transmission electron micrographs containing
artifacts which may be attributed to ion milling effects are shown
in the following figures.

A "saw tooth" appearance can be seen in the dark-field image

of Figure 19 which may have resulted from variations in ion

 

3..

‘H

 

.r»

.n.»

«\u

rs

2‘

wk:

80
beam energy and intensity. Another observation, illustrated in
Figure 20, is the presence of light regions which appear as
"trails" in the material with an orientation corresponding to the
direction of the ion milling beam. Sensitivity of the material's
removal rate to beam—energy oscillations may have caused these
variations. (The ZnS thinning rate was determined empirically to
be approximately 150 nanometers per minute at 5 keV and 68 degrees
to specimen normal). The rotation of the thinned edge through the
beam by stage rotation may add further variability to the milling
conditions. This, however, is only one speculation as to the
cause of these artifacts. The ZnS also underwent dimple grinding,
in which an imperfect grinding wheel may have influenced grain
orientation in the surface of the material.

Figure 21 is a TEM dark-field image of the nickel irradiated
ZnS, and corresponds to the location of the bright-field image
shown in Figure 20. This micrograph was taken at 100k
magnification with an exposure time of 32 seconds. The striations
are less apparent in this image, however, small bright "spots" can
be seen along the edge. This is evidence that precipitants of one
or more phases may be present in the near-surface region of the
bulk material.

A bright—field image of the specimen edge adjacent to that
shown in Figure 21 can be seen in Figure 22. This image, also
magnified at 100k, and exposed for 0.7 seconds, shows dark
regions, larger than the spots in the previous micrograph, located
at the extreme edge of the specimen. Originally suspected to be

precipitants, these spots were estimated to be ranging in size

81
from approximately 10 to 50 nanometers. Their wide spacing and
non-uniform distribution are characteristics that were neither
anticipated for the nickel ions implanted in the ZnS, nor for any
initial precipitations which may have occurred by the combining of
nickel with either zinc or sulfer. Therefore, these regions may
also be artifacts of ion milling, or they may be indicative of
structural detail at the edge. However, viewing this same region
in dark—field mode (Figure 23) revealed reflections with
relatively small size and spacing, as well as uniformity in
distribution. These features appear to be similar to those
observed in Figure 21.

The selected—area—diffraction pattern generated by the
specimen image of Figure 23 is shown in Figure 24. The large
parallel reflections are those of zinc sulfide. Concentric rings
of much smaller reflections also appear in this pattern, and were
suspected of resulting from diffraction by the small reflections
observed in the imaging mode. The ZnS x-ray diffraction results
are presented in the next section, followed by the discussion on

the imaging and indexing of reflections and rings in Figure 24.

82

 

Figure 18. TEM micrograph of thinned edge
of 200keV Ni—irradiated ZnS.
(Mag:100k)

83

 

Figure 19. "Saw tooth" features observed
in TEM micrograph of 200keV
Ni—irradiated ZnS. (Mag:100k)

84

 

Figure 20. Bright-field TEM micrograph
of 200keV Ni-irradiated ZnS.
(Mag:100k)

85

 

Figure 21. "Reflections" observed in dark
—field TEM micrograph of 200keV
Ni—irradiated ZnS. (Mag:100k)

 

86

 

Figure 22. Bright—field TEM micrograph
in precipitant-rich region of
Ni—irradiated ZnS. (Mag:100k)

 

Figure 23.

87

 

Dark-field image of region in
Figure 22 showing precipitant
reflections in Ni-irradiated ZnS.
(Mag:100k)

 

88

 

Figure 24. Selected—area diffraction
pattern of 200keV nickel-
irradiated ZnS. (Mag:100k)

-“ "W

.‘Fm .. *4

 

 

3.2 X-RAY DIFFRACTION

Figure 25 shows the diffraction pattern generated by the
detection of x-rays deflected at various angles from the atomic
planes of zinc sulfide. The 26-angles and the reflecting plane
indices are indicated. Table 2 lists the 26—angles measured from
the diffraction pattern, an intensity rating for each peak (1 =
strongest peak), and the corresponding plane's d-spacing as
calculated from Braggs Law. These are also listed for sphalerite
(B-ZnS) from PDF card no. 5-0566. A rating was assigned to the
relative intensities given for ZnS in the Powder Diffraction File
to show the intensity correlations between the PDF and the XRD
peaks. The d—spacings determined by the XRD analysis were used
for the internal standard method of electron diffraction pattern

calibration.

89

 

 

90

Table 2. XRD Diffraction data for B-Zns.

 

 

Peak
Intensity d (A) d (A)
Rating 2 de XRD PIR PDF

7 26.00 3.427

1 28.75 3.106 1 3.123

6 33.50 2.675 4 2.705
13 43.00 2.104 -

2 47.75 1.905 2 1.912
15 50.80 1.798 -

3 56.70 1.624 3 1.633
12 59.50 1.554 9 1.561
10 69.70 1.349 6 1.351

5 76.80 1.241 5 1.240
14 79.25 1.209 9 1.209

4 88.60 1.1039 5 1.1034

9 95.60 1.0408 7 1.0403
11 107.50 0.9560 8 0.9557

8 115.00 0.9142 7 0.9138

(111)

(200)

(220)

(311)

(222)

(400)

(331)

(420)

(422)

(511)

(440)

(S31)

 

91

 

.mcw mo :uwuumd coflpumummfip amuux .mm musmwm

“a...
'Q I. 'oll. .--.'I4.

:7.
.
n
.2 O '- . - "io. , . O I I ...O
. _ 1.. .
- .- ~ - 0 Del .

.ld...llv.ollooluolluuuon. I I.I'I' '-

a‘...... I

 

 

3.3 TEM RESULTS

The bright- and dark-field transmission electron micrographs
of the nickel—irradiated zinc sulfide are shown in Figures 22 and
23. These images show illuminated spots which are small in size
and display a relatively uniform distribution. These features
indicate the presence of precipitation activity within the
implanted subsurface region of the ZnS. A greater concentration
appears to be present at the surface of the specimen. However, a
decrease in the electron transparency of the specimen away from
the edge would reduce the number of spots it could illuminate, in
spite of concentration uniformity. Therefore, the concentration
of the "precipitants" may in fact be uniform. (This may be
investigated by observing a cross-section type sample using the
procedure described in work by McHargue et. al. [18].)

The illuminations were estimated to have an approximate size
of 5 to 10 nanometers. This observation is consistent with the
fact that the specimen was not annealed, hence, minimal growth of
precipitants would be expected. The only apparant sources
available for thermally motivated effects were the ion milling
procedure and specimen heating from the transmission electron
beam. Ion milling may elevate specimen temperatures several
hundred degrees celsius if a method of cooling is not implemented.
The irradiated ZnS specimens, however, were thinned while in
contact with a cooling stage which maintained the specimens at or
below room temperature. Relative to typical annealing

92

 

 

93
temperatures, the effects would be negligible, particularly with
respect to the bulk material. However, heating from the TEM
electron beam may have been sufficient to result in modifications
to the very thin implanted layer in the specimen. This applies,
as well, to the potential for electron beam heating of the
specimen surface.

As was discussed in Section 1.2, radiation-enhanced diffusion
(RED) is dependent upon thermal activation and moves an atomic
system in the direction of equilibrium. The enhancement is
realized as the large number of radiation-induced vacancy,
substitutional, and interstitial defects move through the atomic
structure towards positions with lower free energies. Typically,
their concentrations far exceed their equilibrium solubility
limits. For this reason, the thermal energy required for their
activation to diffuse is less than that required for a stable,
equilibrium condition. In the event that a low but sufficient
amount of thermal energy was available to initiate diffusion and
precipitation of a nickel species, the growth of the precipitants
would be expected to have sizes consistent with limited growth
kinetics.

Their sizes may also be limited by the competing phenomenon
of radiation-induced segregation (RIS). This process enhances the
mobility of species due to defect fluxes which are localized, and
independent of thermal energy. These fluxes exist due to the
local variations in concentration. Solutes (and other species)
preferentially couple to these fluxes and are transported to

various defect sinks, such as vacancies, grain boundaries, or the

S

 

94
surface. Though the microstructure is moved in a direction away
from equilibrium arrangement by this effect, the local
compositional changes induced by R18 may be sufficient to
precipitate a phase with composition that is normally unstable.
It is likely that the diffusion and segregation mechanisms were
both operative during the implantation process. Also, it is
possible that each contributed to the formation of the
precipitants observed.

It is suggested that the constituents of the precipitate
phase are nickel atoms bonded with sulfur atoms. This infers that
implanted nickel ions predominately displaced zinc atoms and
formed bonds with sulfur atoms. The displacement energy for zinc
is between 7 and 9 electron-volts, and for sulfur is between 15
and 20 electron-volts[42]. ZnS may, therefore, have become
supersaturated with NiS in a metastable state. This condition in
the presence of a high concentration of vacancies is one in which
the precipitate transformation reaction may occur. Using the
general equation from Section 1.2, the reaction for the irradiated
ZnS may be described as:

(1‘--> (1 +6 (4)

where, (x' = the metastable supersaturated solid solution of
nickel, zinc, and sulfur;

a = ZnS structure with nickel retained closer to an
equilibrium concentration;

0 = stable and metastable precipitants of B—NiS and
a—Nis, respectively.

 

95

The selected—area diffraction pattern, formed in the back
focal plane of the electron microscope, of the specimen of Figure
23 was exposed for 64 seconds, and is shown in Figure 24. This
micrograph includes two significant features. First is the
pattern of reflections forming three rows, with approximately
equal spacing between each other, and between the reflections in
each row. Second is the group of concentric rings which encircle
the center spot. The rings consist of very fine reflections,
indicative of a very fine crystalline structure. This was
determined by inspection of the rings in an enlargement of Figure
24. Also in this figure, diffuse rings, typical of a diffraction
pattern for an amorphous structure, were not observed. From this
it can be concluded that no amorphization of the ZnS structure

occurred as a result of the implantation.

3.4 INDEXING DIFFRACTION PATTERNS

3.4.1 Zinc Sulfide

Figure 26 shows the selected-area diffraction pattern of the
unmodified zinc sulfide ceramic. Exposure time was 110 seconds at
an approximate camera length of 55 centimeters. The reflections
are illustrated in Figure 27 and are labeled A through E.
Interplanar angles are also shown. A single—crystal diffraction
pattern monograph from reference [39] matches the pattern of
Figure 26, and is shown in Figure 28 with the reflection indices
and relevant angles indicated.

Table 3 lists the data and calculations determined by
indexing for the diffraction pattern of Figure 26. The second and
third columns list the radii for reflections A through E and the
corresponding d-spacings. The spacings were calculated from
Bragg's Law, using the approximate camera length of 550
millimeters. D-spacings from the ZnS x-ray diffraction data,
which were determined to be the closest in value to the calculated
spacings, are also listed along with the plane indices. Next are
the indices for each of the reflective planes found to be the
sources for the reflection positions by comparison of these
positions with those in the reference monograph. Finally,
agreement between the plane families indicated by the monograph
and those determined through d-spacing correlations are noted.

A single agreement exists by comparing planes in Table 3 for

96

 

(l.

{/1

 

97

reflection B which has a d—spacing of 2.675 angstroms and indices

of (200). Substituting this d-spacing into the camera constant
equation, ?(L = Rd, results in a camera constant of 2.207 (square
millimeters). Using this calibrated camera constant, re-

calculated d-spacings for reflections A through B were found and
are listed in Table 4. Comparing the XRD planes with those from
the monographs resulted in agreement between the plane families
for each of the reflections from the diffraction pattern for zinc
sulfide. The matching angles shown in the monograph and the
diffraction pattern add further confirmation that this monograph
is representative of the ZnS selected—area diffraction pattern.
The pattern of Figure 24 was produced by electron beam
transmission through ZnS at a particular crystallographic
orientation. Patterns from different orientations will result in
different reflection arrangements, interplanar angles, and
reflection intensities. However, the presence of at least one
reflection that was common between the unmodified and modified ZnS
SAD patterns allowed the use of the internal standard method of
calibration, regardless of crystallographic orientation. This was
observed for the pattern generated by electron transmission

through the modified ZnS.

 

98

Table 3. Indexing data for SAD pattern of unmodified ZnS
using camera constant of 2.035.

 

Reflection Closest Monograph
Label Radius (mm) d-spacinq (A) XRD Spacing Plane qur
A 7.50 2.713 2.675
(200) (111) no
B 8.25 2.467 2.675
(200) (200) yes
C 12.25 1.661 1.624
(311) (022) no
D 14.25 1.428 1.349
(400) (311) no
E 18.68 1.089 1.041

(511) (133) no

99

Table 4. Indexing data for SAD pattern of unmodified ZnS
using camera constant of 2.707.

 

 

 

Reflection Closest Monograph

Label Radius (mm) d—spacinq (A) XRD Spacing Plane Aqqr

A 7.50 2.943 3.132
(111) (111) yes

B 8.25 2.675 2.675
(200) (200) yes

C 12.25 1.802 1.905
(220) (022) yes

D 14.25 1.549 1.624
(311) (311) yes

E 18.68 1.181 1.241

(331) (133) yes

100

 

Figure 26. Selected-area diffraction
pattern of zinc sulfide.
(Mag:100k)

 

101

 

.mm musoflm mo cuwuumd
COADUMHMLHU mmumlpmuomamm mcm mo coflumuumzaaH

 

O
O
o a.
O
O
O
C .................. . ............
.o ..<\\
oNM \\\ m
.x.
. H
mm. 1

.am magmas

 

 

102

.mm1.©m mesons as mam sedansms camumocoz

.am maaaaa

 

 

fin
o a o
O
O
IIIIIIIIIIIIIIIIII .IIIuIIIIIIII
5 bk
-. .
m (X <
0 mm on Q. a
\\ 8m
x O O

«No

 

 

 

3.4.2 Nickel-Implanted Zinc Sulfide

The selected-area diffraction pattern obtained with the
nickel-irradiated zinc sulfide specimen is shown in the micrograph
of Figure 24, and is illustrated with labels in Figure 29.
Reflection spots have been assigned the letter designations of A
through H, and numbers 1 through 7 identify the individual rings.
The indexing data for the irradiated ZnS pattern are given in
tables similar to those for the unmodified ZnS SAD pattern.

Table 5 lists the radii for the reflection spots, the initial
approximated d—spacing calculations (camera constant = 2.035), the
closest d—spacings from the XRD data, the plane of reflections
indicated by the reference monograph, and any correlation
ascertained by comparison. The reference monograph pattern for

the irradiated ZnS diffraction image is illustrated in Figure 30.

103

104

Table 5. Indexing data for irradiated ZnS using calibrated
camera constant of 2.035.

Reflection Closest Monograph
Label Radius§mm2 d—spacinq (A) XRD spacing Plane Aggr

 

A 6.50 3.131 3.106

(111) (111)* yes*
B 6.75 3.105 3.106

(111) (111) yes
C 9.25 2.200 2.104
D 17.00 1.197 1.209

(420) (242) no
E 17.50 1.163 1.209/1.241

(420)/(331) (331) yes

F 17.75 1.146 1.104

(422) (420) no
G 20.50 0.993 0.956(

(440) (511) no
H 23.50 0.866 0.855

(620) (153) no

(* - by comparison to Figure 28)

105

The data in Table 5 shows that agreement exists between the
XRD plane indices and those indicated by the reference monograph
for three reflections. The first is of reflection spot A. The
illustration of Figure 29 indicates this reflection is atypical.
It is not included in any of the three rows of reflections which
represent the bulk ZnS. Additionally, its location is clearly
interior to the inner ring, thus it is not believed to be
associated with the ring reflections. Yet, this spot does appear
to be part of a regular pattern. Three additional spots are
present within ring 1 which form, with spot A, an arrangement
which is symmetrical about the central axis of the diffraction
pattern. Table 5 indicates they have plane indices of {111}.
However, reflection B also appears to be (111) related. The
symmetry of these four reflections, the angles between them, and
their d-spacing all indicate they are the (111) reflections
indexed for the non-irradiated ZnS pattern, as shown in Figure 27.

The SAD patterns of Figures 26 and 24, excluding the rings,
have both been presented as those of ZnS. The reflection
arrangements appear different as a result of variation in the
angle of the incident electron beam with respect to the crystal
orientation. Double diffraction, the phenomenon whereby a
secondary diffracted beam acts as a primary incident beam, may
have been the cause for the atypical reflections in the irradiated
specimen. This may be the reason that a pattern similar to that
observed in Figure 26, for the unmodified ZnS is present in the
irradiated ZnS reflection arrangement. Spot A in Figure 24 may be

the same reflection as that represented as A in Figure 26.

 

106

The second agreement noted in Table 5 exists between the
plane indices indicated by the XRD, and monograph data for
reflection B. The spot is located in a main row of reflections,
therefore it is not suspected as resulting from double
diffraction. Additionally, the calculated d-spacing of 3.105
angstroms (lodfimeters) is a close match to 3.106 angstroms, the d—
spacing from the (111) XRD peak. Furthermore, the (111) peak was
that which exhibited the greatest x—ray diffraction intensity,
which lends further confidence to the d-spacing correlations.
With E as the internal standard reflection with a d—spacing of
3.106 angstroms, the calibrated camera constant was calculated as
2.097.

Table 6 lists the indexing data, revised by the calibrated
constant, for the pattern of reflections in Figure 24.
Correlation was found for five of the eight reflections which
strongly suggests that the main pattern for the irradiated
specimen does represent zinc sulfide. The angles between
reflections which are illustrated in Figure 29 and the monograph
pattern of Figure 30, further indicate correlation for this
pattern.

The reflection labeled C in the diffraction pattern is shown
in both tables as having correlation to the XRD d-spacing of 2.104
angstroms. This d—spacing was calculated for the XRD peak of
43.00 degrees 20 which had a Peak Intensity Rating of 13, i.e., an
extremely low intensity. No extraneous reflections were observed
in the unmodified ZnS SAD pattern, and the d-spacing of 2.267

angstroms calculated from the modified ZnS data was not a value in

 

107
the unmodified indexing. Further assessment of this reflection is
discussed with the results of indexing the rings in the modified
ZnS pattern.

Table 7 lists the radii and calculated d-spacings for each of
the rings observed in Figure 24. It was hypothesized that these
rings were generated by very fine crystalline precipitates of
nickel sulfide. Data supporting this hypothesis came from the d-
spacings of the rings. The data given in Table 7 are for d-
spacing calculations using the calibrated camera constant of
2.097, as determined by the internal standard based upon
reflection B.

An additional improvement in indexing the rings was obtained
by use of an 8 inch by 10 inch enlargement of the SAD pattern.
The rings in the original negative were spaced very close together
thus resulting in greater probability of error in measuring the
radii. Measurement of the radius for spot B on the original
negative was assumed to be the most accurate because of the
reflection intensity and uniformity. Therefore, the enlargement
factor was determined from the ratio of radii measurements for
reflection B on the negative and on the enlargement. This was
found to be 0.342. Multiplying this into the radii measured from
the enlargement resulted in more accurate ring radii from the
original SAD pattern negative. The results from this procedure,
and the corresponding d-spacings are given in Table 8.

Reflection C, discussed previously, is also listed in this
table with the corresponding corrected d—spacing as determined by

the enlargement measurement.

 

108

Table 9 shows the correlations made between the d—spacings
for the rings produced by the nickel-irradiated zinc sulfide
specimen and those for various phases of nickel sulfide. The
correlating phases are those which most closely match the ring
data out of all the available data for phases involving any two,
or all three of the nickel, zinc, and sulfur elements. The data
for (x- and.[3—NiS were taken from [42], the others from the Powder

Diffraction Files. [37]

109

Table 6. Indexing data for "reflections" in irradiated ZnS
with a camera constant of 2.097

 

 

 

Reflection Closest Monograph
Label Radius(mm) d-spacinq (A) XRD spacinq Plane Aqqr

A 6.50 3.226 3.427 -
B 6.75 3.106 3.106

(111) (111) yes
C 9.25 2.267 2.104
D 17.00 1.234 1.241

(331) (242) no
E 17.50 1.198 1.209/1.241

(420)/(33l) (331) yes

F 17.75 1.181 1.209

(420) (420) yes
G 20.50 1.023 1.041

(511) (511) yes
H 23.50 0.892 0.9142

(531) {153} yes

 

110

Table 7. Indexing data for ring pattern of Figure 24 with
camera constant of 2.097.

 

Ring Radii d—spacing
Number _(mm) (angstroms)
1 7.50 2.796
2 8.00 2.621
3 8.75 2.397
4 11.50 1.823
5 13.25 1.583
6 14.75 1.422

7 15.75 1.331

111

Table 8. Indexing data for ring patterns from enlargement of
Figure 24 with calibrated camera constant of 2.097.

 

 

Ring Enlargement Corrected d-spacing
Number Radii (mm) Radii (mm) (angstroms)
1 21.75 7.44 2.819
2 23.50 8.04 2.608
3 25.50 8.72 2.405
4 31.75 10.86 1.931
5 37.75 12.91 1.624
6 41.25 14.11 1.486
7 45.00 15.39 1.363

C 26.50 9.06 2.314

 

112

Table 9. Comparison of Ni-irradiated ZnS d-spacings to various
phases of nickel, sulfur, and zinc.

Ni-ZnS
d-spacings

(angstroms) a— NiS fl—NiS fi-ZnS Ni3__ _ NiS _ Ni3§4

2.819 2.830 2.850
(200) (311)
2.608 2.602
(101)
2.405 2.404
(220)
1.931 1.912
(220)
1.624 1.634 1.633 1.634
(321) (311) (311)
1.486 1.489
(200)
1.363 1.388 1.372 1.375 1.369
(600) (015) (410) (444)
(322)
2.314 2.227

(211)

113

.vm wusmflm mo cumuumq mcmuflz mo :oflpmuuwzaaH

.mm masons

 

 

 

 

 

114

 

_©m_.rm

 

muzofla :_ mzxuflz mcw50oma :Qoncoco: .cm ot:qfla
mm—
o O
:m 0 O
o Nam

Sm nmNm
- xx/a owl. o
P: .xx
NVN . m 1:11):
O m
m 0
m9 0 O
.
I m ovens

I|

‘
.8

‘Il

‘
|I
‘II

 

3.5 HARDNESS

The range of the nicke1~implanted layer prior to annealing
was approximated by the "Profile" software to be 1104 angstroms,
about a tenth of a micron. The depth of the Vickers indentor was
calculated from the data and simple geometry (Vicker's indentor
angle = 136 degrees) to be as high as 12 microns. Therefore, the
hardness results represent values for the composite consisting of
the nickel~irradiated ZnS and the untreated ZnS layer below. The
change in hardness from unmodified to modified, and modified—
annealed specimens represent trends in the hardness of the
irradiated surface alone. Though the depths of nickel or nickel
sulfide phases were examined analytically and theoretically, the
results were neither exhaustive nor conclusive. However, some
evidence suggests that following the annealing, diffusion of such
species away from the surface and into the bulk may have occurred.
If this were the case, the Vickers measurements would more closely
represent the hardness of the modified layer. Final hardness
results were determined from data generated using the lowest
available load on the Vickers indentor, thus, Optimizing the
results to best reflect the hardness attributable to conditions in

the modified layer alone.

115

116

3.5.1 Nickel-Implanted Zinc Sulfide

Following the irradiation of zinc sulfide with nickel, the
hardnesses of both modified and untreated specimens were measured
by Vickers indents. Appendix A contains the Basic program which
calculates hardness in GigaPascals (GPa), given the load used for
indenting and the indent measurements in microns.

The initial hardness values determined for irradiated
versus non-irradiated ZnS were 1.84 GPa and 1.77 GPa,
respectively. The indentor load used was 1.0 kilograms-force
and 14 indents were measured for each specimen. The standard
deviations were +/— 0.051 for irradiated ZnS and +/- 0.055
for unmodified ZnS, while the coefficients of variation for
each were 0.028 and 0.031, respectively. This small increase of
about 4 percent, and the standard deviations for both values,
appear to indicate that thermodynamic constraints typical of
conventional processing did not significantly affect the
"metastable" state of the implanted nickel. Cooling of the
substrate during implantation by use of a chill block and liquid
nitrogen flow may have been sufficient to permit only the action
of implanting nickel ions into the sub-surface region of the ZnS.
Once implanted, little thermally motivated activity occurred for,
and between, the nickel, zinc, and sulfur atoms. Further
confirmation of this can be seen from the hardnesses of irradiated
and non-irradiated pre—anneal specimens which were determined

using the decreased indentor load of 0.3 kilograms-force.

117

As discussed, the decrease in the indentor load resulted in
hardness data more representative of the nickel-implanted layer.
An average hardness of 1.77 GPa was calculated for three
irradiated ZnS specimens from 4 indents per specimen with a load
of 0.3 kilograms-force. The average of the standard deviations
and coefficients of variation were +/— 0.062 and 0.035,
respectively. The average hardness, standard deviation, and
coefficient of variation for the three non-irradiated specimens
were 1.75 GPa, +/— 0.067, and 0.038, respectively. These results
are included in Table 10 with the post-anneal hardness data. (The
raw data for all the hardness calculations are given in Appendix

B).

 

 

3.5.2 Annealed Nickel-implanted ZnS

The Vickers hardness of three specimens of nickel-irradiated
zinc sulfide were determined prior to, and following annealing to
their respective peak temperatures of 750, 800, or 850 degrees
celsius in a furnace chamber with a continuous flow of argon gas.
The change in hardness for the irradiated and the control
specimens of ZnS are shown in Table 10. All three of the control
specimens experienced an increase in hardness of approximately
6.5 percent. The irradiated specimens exhibited larger increases
in their Vickers hardness.

For the specimen annealed to 750 degrees celsius, the pre—
and post—anneal hardnesses were 1.77 and 2.13 GPa, respectively.
This represents a 20.3 percent improvement. A 28.6 percent
increase was observed for the 800 degree profile specimen which
exhibited 1.75 and 2.25 Gpa pre- and post—anneal hardnesses. The
850 degree celsius anneal schedule resulted in a 24.4 percent
increase, from 1.80 to 2.24 GPa. The indent load used was 0.3
kilograms-force for 4 indents pre-anneal and 12 indents post-
anneal testing of each specimen.

The objective in the selection of profiles for annealing
was to provide thermal energy to initiate the solid state of
the nickel-irradiated ZnS to move towards thermodynamic
equilibrium by assisting in overcoming diffusion barriers thus
resulting in precipitation of nickel sulfide. The melting point
of ZnS (sphalerite) has been estimated to be 1870 degrees celsius

118

 

 

119

although, it has been observed to transform to wurtzite at around
1000 degrees celsius under atmospheric pressure [26].

It would be advantageous to the performance of the modified
ZnS surface if the precipitated metastable transformation
toughener (NiS) was of minimal size, for this would maximize
uniformity in the distribution of the precipitant and, therefore,
in the surface modification effects. It was estimated that an
homologous annealing temperature (based on the melting point of
the ZnS bulk; 1830 degrees celsius) of 0.45 to 0.55 would be the
most conducive to this end [38]. The maximum annealing
temperatures were attained by heating as rapidly as possible, then
cooling with an even greater thermal descent to achieve, in
effect, a "quench" of the specimens undergoing thermal treatment.
This too, was implemented to maintain control of the precipitant
size.

Figure 17 illustrates the thermal profiles used in
annealing as temperature versus time, in degrees celsius and
minutes, respectively. It is apparant the slopes of these cycles
are similar, both during heating and cooling of the specimens.
The 750 celsius profile averaged a 36 degrees celsius per minute
rise in temperature between 15 and 750 degrees celsius. The
average heating rate for the 800 C profile was 34 degrees celsius
per minute, and the 850 C target temperature was met with a 30
degree celsius per minute average increase in heating. The times
to reach the target temperatures of 750, 800, and 850 degrees
celsius were 32.5, 39.0, and 39.5 minutes, successively.

Similar uniformity in thermal influences were attained

 

 

120
for the average cooling rates from peak temperature to just
over 100 degrees celsius. Each of these rates were within
+/- 10 percent of the others. Uniform cooling rates near the
beginning of the quench, where it is most critical, were
maintained within +/— 2.5 percent for the rates from the peak
temperatures to an average of about 360 degrees celsius.

The hardness values in Table 10 show that the irradiated
ZnS experienced an average increase in hardness of approximately
25 percent, compared to a 7 percent approximate increase in the
hardness of the non-irradiated zinc sulfide specimens. The
highest average hardness value of 2.25 GPa was measured for the
specimen annealed with the 800 degree profile. Figure 31
illustrates the changes in the hardness of both irradiated and
non-irradiated ZnS specimens with their annealing temperatures.
From this it can be seen that annealing to 800 degrees celsius
provided the maximum increase in hardness. Annealing beyond this
had little additional effect. The formation of N18 in the
subsurface region by the annealing may have been limited by
solubility constraints of nickel sulfide precipitants in the zinc
sulfide matrix..

Figure 31 clearly indicates that an increase in the annealing
temperature from ambient to 800 degrees celsius resulted in an
increase in the Vickers hardness of the nickel-irradiated ZnS.

If the diffusion-barrier model is considered, one may anticipate
no significant change in hardness with temperature followed by an
abrupt increase in hardness, signifying attaining the thermal

energy level necessary for the atomic "jump" across the barriers

 

121
to diffusion. This may be true if all the implanted nickel ions
were resting in the same metastable energy state. However, due to
the nature of the irradiation process, the atoms can be in various
states of mobility and rest, depending on the defect type with
which they were associated as well as their position with respect
to the surface.

Therefore, the increase in hardness observed for the anneal
at 750 degrees celsius is an indication that the proposed
mechanism may have been initiated. If this is accurate, then 750
degrees celsius must be sufficient thermal energy to induce the
diffusion of implanted nickel in zinc sulfide. It is likely that
a concentration of nickel was present at ZnS grain boundaries
prior to the annealing process, in the form of nickel sulfide and
as nickel atoms. Grain boundaries are sites for nucleation and
growth; processes which lower the free energy of the system. The
time required for diffusion of nickel atoms which are in the
grains themselves but immediately adjacent to the grain boundaries
would be less than for species further away, fully into the bulk
of the zinc sulfide grains. This would result in an increase in
hardness, as was observed for the 750 anneal, if the toughening
phase precipitated to a certain minimum concentration in the
subsurface region.

Annealing to a higher temperature or for a longer time
would then allow additional diffusion from the bulk and
precipitation to occur. However, it follows that those grains
which precipitated first would continue to grow at the expense of

other newly precipitating nucleation sites. Since a fine grain

 

122

size is desired for optimum uniformity and distribution of the
toughening phase,the time of anneal can be minimized while
increasing the utilization of precipitating species by increasing
the anneal temperature. Additionally, a finer grain size may be
obtained by a fast rate of heating which results in the formation
of more nucleation sites. The further increase in hardness
observed for the 800 degree celsius anneal may be due to an
increase in concentration of the toughening nickel-sulfide
precipitants, with grain sizes still kept at a minimum by
minimizing the time at temperature.

Through thermal elevation, nickel sulfide in the hexagonal
form (a—hi temp) enters the liquid state at approximately
995 degrees celsius and experiences a phase change to B-NiS
(millerite) upon cooling to around 380 degrees celsius.
A change to B—NiS is reported to exhibit a positive change
in volume of approximately 4 percent [26]. This was the apparant
increase in hardness determined for ZnS after the implantation

procedure.

Table 10.

Anneal
Tem C

pre

750

pre

800

pre

850

123

Vickers hardness data for nickel-irradiated and

non-irradiated ZnS pre— and post-anneal.

39!;IBBAQIAIEQ

 

Hardness
(GPa) +/-
1.75 .058
1.85 .046
1.77 .065
1.88 .084
1.74 .077
1.88 .079

.033

.025

.037

.045

.045

.042

CFV = coefficient of variation

IRRADIATED
Hardness
(GPa) +[—
1.77 .037
2.13 .073
1.75 .050
2.25 .095
1.80 .099
2.24 .095

.021

.034

.028

.042

.055

.042

Hardness (GPa)

2.4
2.3
2.2

I"
N...)

1.9
1.8
1.7
1.6

1.5

124

 

 

 

 

 

 

 

L
" 4
C.
).
L T
L —O— Irradiated
L -'9- Control
20 750 800 850

Annealing Temperature °C

Figure 31. Increase in Vickers hardness for
200keV Ni-irradiated ZnS and
control samples with annealing.

 

 

 

 

3.6 SCANNING ELECTRON MICROSCOPY

A non-irradiated specimen of zinc sulfide annealed with
the 750 degrees celsius schedule (see 2.5.2) was analyzed
using SEM and EDX. Percents, by weight, of the zinc and
sulfer were 78.58 and 21.30, respectively. Trace amounts of
various elements were attributed to contamination by the
conductive paint, the aluminum base, and handling. No trace
of nickel was found.

Figure 32 is a micrograph of the non—irradiated ZnS
surface at a magnification of 1100. Evidence of grain
"pullout" can by seen by the presence of voids in the surface
which appear to be greater than or equal to 2.0 microns.

Voids similar to those in Figure 32 appeared relatively
uniformly distributed with spacings of approximatley 20 to 30
microns. Figure 33 shows the voids at a higher magnification
(x3300). The surface of other grains canbe seen at the bottom of
these shallow voids.

A nickel concentration of 1.61 weight percent (1.24 atomic
percent) was found for a specimen of nickel-implanted zinc sulfide
prior to annealing. Small voids, or "micropores" were observed
in the surface before annealing. As shown in Figure 34, these
pores appear to be different from the voids in Figure 33 in that
no grain surfaces could be seen at the bottom, thus giving them
the appearance of a deep pore, or pit. Sizes averaged 1.0 micron
or greater, and the distribution was also relatively uniform, with
a pore spacing of about 15 to 25 microns. This may be evidence of
surface damage caused by irradiation.

125

126

Following annealing to 750 degrees celsius of the irradiated
ZnS, the concentration of nickel as determined by EDX decreased to
0.42 weight percent at the surface of a Vickers indentation, and
0.32 weight percent at the specimen surface. This may represent a
qualitative indication of nickel diffusion further into the bulk
of the ZnS. Figure 35 is a scanning electron micrograph of the
post-annealed irradiated ZnS.

Several observations can be made by commparison of Figure 35
with Figure 34. First, the "pitting" seen in the pre-anneal
specimen is not evident following the anneal. This may be due to
a thermal surface healing phenomenon, or may be less apparant
simply due to the increased contrast of the grain boundaries
following the thermal cycle. This leads to a second observation,
the appearance of an "etching" effect resulting in the enhanced
contrast seen in Figure 35. Grain-boundary definition appears to
be an artifact of the thermal exposure during annealing of the
specimen. Finally, the grain size of zinc sulfide for the various
specimens subjected to annealing did not change appreciably. The
average size regardless of processing was approximated to be 5
microns.

Figure 36 shows the nickel—irradiated ZnS following the
anneal to 850 degrees celsius. AGain, no appreciable change in
grain size is observed, and the "etching" effect is enhanced.
However, the weight percent of nickel decreased further, from the
levels for the non-annealed and 750-annealed specimens, to a
concentration too low for detection by the EDX system. This was

true regardless of variation in the detector angle. The annealing

127
temperature of 850 degrees celsius also represented the peak
anneal at which improvement in hardness of the modified zinc
sulfide appeared to level off. These two results, though
preliminary, indicate that the nickel ions implanted into the zinc
sulfide surface diffused away from the surface between the
temperatures of 750 and 850 degrees celsius.

The TEM results show evidence that various nickel sulfide
phases were present in the irradiated ZnS surface region. Theough
the elctron diffraction results came from irradiated ZnS not
subjected to annealing, the effects of ion milling and electron
beam heating may ahve been sufficient to initiate diffusion of the
implanted nickel and subsequent association and precipitation
of nickel sulfide particles.

Regarding the quantitative EDX data generated for nickel in
the ZnS, the relative weight percents determined for the nickel
were below the tolerance for EDX, which is generally approximated
as around plus or minus 3 to 5 percent, depending onthe system in
use. In spite of this, nickel was detected and measured where it
was known to be present, and the various concentrations followed a
qualitative trend consistent with observed annealing behavior for
various implanted species [8,18,23].

There may be a correlation between the improvement in
hardness of the irradiated ZnS and the extent of nickel
diffusion/dispersion away from the sub-surface region of the zinc
sulfide. Auger analysis may be used to further understand the

diffusion behavior of nickel in ZnS with annealing.

128

 

25KU X1188 8881 18.00 CEIIIB'E.‘

Figure 32. Scanning electron micrograph
of zinc sulfide. (Mag:1100)

 

19 EU EEO};

Figure 33. Scanning electron micrograph
of zinc sulfide surface "voids."
(Magz3300)

130

 

Figure 34. Scanning electron micrograph
of 200keV Ni—irradiated ZnS.
(Mag:1100)

 

Figure 35. Scanning electron micrograph of
Ni—irradiated ZnS annealed to 750
degrees celsius. (Mag:1100)

 

agru H1188

Figure 36. Scanning electron micrograph of
Ni—irradiated ZnS annealed to 850
degrees celsius. (Mag:1100)

3.7 X-RAY PHOTOELECTRON SPECTROMETRY

Figure 37 shows the XPS spectrum for ZnS irradiated with
nickel ions at 200 keV with a fluence of 1x103L7 ions/cmz.

Peaks were found which correlate to the inner-electron binding
energies for zinc and sulfur, respectively. These energies are
listed in Table 11 along with those that correlate to the
compounds of nickel-sulfide and zinc-sulfide. For those electron
energies indicating the presence of a compound, the element of
that compound to which the electron is associated is given in
parenthesis. Peaks which resulted from Auger, rather than inner
shell electron ejections are labeled in the spectrums, as are some
of the various electron levels. (Auger electrons are those
electrons which are ejected in response to the charge imbalance
created by the absence of an inner—shell electron). A small
nickel peak is also present in Figure 37 at a binding energy of
855.0 electron-volts. This is shown enlarged in Figure 38.

The carbon peak in Figure 37 was not present in the spectrum
after minimal sputtering of the specimen surface. Additional
sputtering of the specimen surface followed by analysis was
conducted, until approximately 50 nanometers of the top surface
were removed. The peak indicating the presence of nickel
increased very little in height and area, even up to 50 nanometers
in depth. Atomic concentration of the nickel at this depth was
measured to be 3.75 percent. However, the sensitivity of the
spectrometer for such a low concentration makes this value a rough

133

134

approximation. The computer generated data of the "Profile"
software calculated the concentration at this depth to be
approximately 7.0 atomic percent.

The binding energy for a 2p level electron of sulfur, if
the sulfur is associated with nickel, is 162.2 electron-volts.
The sulfur peak in Figure 39, corrected by 2.0 electron-volts (an
instrument calibration) had a binding energy of 162.0 electron-
volts; a reasonable correlation to 162.2 eV. This indicates that
nickel—sulfur bonds were present in the ZnS specimen prior to the
annealing operations.

The XPS spectrum of nickel-irradiated ZnS annealed to
800 degree celsius is given in Figure 40. There was neither
an indication of nickel in the spectrum nor an increase in
the nickel—sulfur bond peaks. Additionally, nickel peaks
were not found after sputtering the specimen surface up to
100 nanometers, the theoretical range for the pre-annealed
implanted nickel. The high level of diffusion associated
with nickel is considered to be the primary reason for this
result. Two oxygen peaks were removed following sputtering

while all other features in Figure 40 remained unchanged.

Table 11. Binding energies of nickel, zinc, and sulfur,
Ni NiS ZnS Zn
857.7 162.2 (8) 164.0 (S) 1021.8

855.1 (Ni) 1022 (Zn)

135

>mxoom Lo Esuoomdm cowoomHoODOSQ >mm x

o.omm o.oMm

.mpflwflsm ocfln Umumwpmuufi-fimxoflc
.bm muzmflm

>m .>uzuzu uz_=z_=
o.ovv =.=nm c.3og o.=- =.owa o.oaa 3.3:.—

 

 

 

 

 

 

:d__
I. _ .x 11.1-7). i i _ _ _ (liltii _ _ _ _
J\) “Hart,“ ()1;
k I. (N “n” p
m h an 7 .\
a=d 3 fl 3...
a. m ,N. ,1/
. ,. .. . 4..) .1... \\.r.\. 3!»... ~ p.17 . .). 4).... .
N .6“... a .: [VT/Qi/ —.. ... f \..PJ.T;.P~)L..\.:\ a 4) J \. K... .
a 3.... 1.. <.\‘.....-.~..
. 5.63.: 1...?
5.332.: . .....
—o.—U.—- )r#
_
surd~o_
. Axu—N .
sz .z~ ..
p _ n n p
- u q u d

‘F
‘-

 

uh

 

--

‘b

db
‘
‘-
q-
‘—
-
~-
—

z cam a: >0 amv.aa "aaaLzL mmaa ”\0 2 aka a .aaa.a_ u_emaas .as_a=a dream
_z ;._= eaz=a_ac_ o_=m2aa as_a_=m a=_~ ~m=~ “as”.

:_: mm.~um=~_ age was mv nunozc mw\m\m >u>zam cumu

:—

3/(3)N

13€5

.xmmd meofl: wo
cowmmu ca mpflwasm ocflu Umumflpmnufi-amxoflc

 

>mxoom mo Eduuomdm couuomeODOSQ >mm-x .mm ansmflm
>a .>uzuzu oz—=z~=
o.vvm o.mvw o.vmm o.mmm a.vom c.muw . a.vmw o.mmm o.vxm o.amw c.vaw
i. _ i i i i _ _ i i i _ + i i i _Illll|+l|):
_. --
.1. = --
rflfiee%_g g3 -
. _ I
1 s * -
_ 1... 5,).
_—__~ f. — .* --

-. >u0mm®

 

 

 

 

 

.1—
.1-
.1
-
Cl—
~-

2 com a: so sma.mm u>aaozo mac; m\o g mam.m .m~_.a
_z ;~_3 woucm_ac_ o_cmeoo o

c_: mm.muu:—_ age not mv nuaczc m Qua __zu4m

n—umdda .zopuce udcum
v_d_:m oc_N mch ”mam;
mmxwxm xmagmpgaz cumu

o

L”)

o—

3/(3)N

137

.xmmu usuasm mo
coflmon ca apfiuaam been paumflpmnnfl-fimxoflc

 

 

   

 

 

>mxoom mo Esnuomdm conuomHmODOQQ >mm-x .mm onsmflm
>6 .>ozuzu cza=z~=
o.wn_ o.ce_ o.mm_ c.vm~ o.mm— o.mw_ o.om_ o.~m_ c rm. o.mm_ o.mm_
.1)_))/_> S l 1):-.. 2-1. _ i i/ i _\. _ i i _ _ 1 TI).
x< . , _ 2 . .
. ./\/\, i /.>\ ./\-/ (>23: 2: 5% />/\.:( .x\/ \ 3
<<.
)0 0.7.: .
W .
i i i 1 i . i i i i i i i 1 i _ 7:11))-
5vnYNa_
Au.1.a;....:..a.._u wafl.c.- 2 com a: so oma.mm "aazeza mass M\0 x ava.a .mvm.~ u_imaaa .z=_sai Dream

:_= em.=uuz__ so: ass we ”L332: 4 as: .muso

am\m\m

_z 5:: tote—ac. 0:53:50 21.—:5“ o:_N mMZN n.._.=..._

x_L;—_L:= cum:

 

a

U?

o—

3/(3)N

138

ad

od__

.maflmfimo moanmmp oom
ob poammccm mpflmaam ocfln paumfipmuna-amxofia
>axoom mo Ennuomdm casuomHm0uond xmm-x .ov madman

>o .>o2mzH cz~=z~=

o.omm o.omm c.ovv o.omm c.ooo o.omm c.omw c.cam c.co__

 

 

.L . i i
-SHIFJ</XJ§
z.“

2.9 cm...—

or“

 

“ n _ u w n n w n u n “ n n u n .
on m .. 1
Sega -

pf.)
</ -
A<V(..M . .oq pie: l :11 > to; -
R¢gx<~( /.¢J/I2<l\\$/5JI .1s\¢$.$>.5433s\1.\ . »£..(.}<$4o£\».f;o.;r
f—U F‘VP-N Argo 9752/»

a . Q /

to), v I
.5 ._N < N.
p p n _
u a d -

 

 

 

‘-

2 com a: >0 omv.mm uroumzu mmc; mxo x Nm~.~

.Nmm.m_ "humane .zopucd mLcum

cucom ccc\egm x coc_ooam oo~.m:~ nwLng
c_c va.~umzmp cu: mac mv umLuzc mm\m\m >m>z3m cumm

c

If)

3/(3)N

3.8 DIFFUSION

Figure 41 is a plot of the implanted nickel as determined by
the software "Profile Code" developed by Implant Sciences
Corporation, Danvers, MA. The concentration profile was
calculated given the incident and target material data, as well as
the implantation specifications. Figure 41 indicates the range of
the nickel irradiated into ZnS to be 1105 angstroms, or 110
nanometers. The atomic concentration at this depth was 8.9x1021
nickel atoms/cm3. This maximum concentration is also given as
approximately 17 atomic percent. This result comes from dividing
the range concentration of nickel into the atomic concentration of
ZnS (5.23x1022 atoms/cm3). The maximum depth of the implanted
nickel was determined to be approximately 240 nanometers.

Also from Figure 41, the concentration of nickel at the
surface can be estimated as a few atomic percent. This is
consistent with the pre—anneal XPS concentration result of 3.75
atomic percent nickel. SEM/EDX measurements detected only 1.24
atomic percent. For such a low concentration, it is likely that
even a small difference in the XPS and EDX measurement volumes
could result in this difference.

While all of these values reflect the modeling of the
implantation results, and represent a useful qualitative
approximation, other mechanisms may have affected the true nickel
profile. Through the implantation process, radiation damage
results in a greater concentration of vacancies in the region of
irradiation. The maximum concentration of vacancies would most

139

140
likely lie between the surface and the region where most of the
nickel ions come to rest, ie. the range. For each nickel ion,
there exists a "collision cascade," which was initially discussed
in Section 1.2. This cascade is the corridor of damage remaining
in the path of each ion after it has traveled through the atomic
system dissipating its kinetic energy by collisions with other
atomic species.

As is true for the nickel atoms, radiation damage may also
result in a vacancy concentration which exceeds the equilibrium
solubility concentration in the atomic matrix. This condition
results in non-equilibrium concentration gradients. The driving
force for vacancies to move across such a gradient are the sudden
atomic changes occurring within the bombarded atomic structure.
Therefore, such a vacancy flux may be active during or shortly
following the irradiation process.

Diffusion of implanted nickel atoms may also occur, for the
vacancy flux is operative in proximity to the implanted species.
These atoms move in the direction opposite to the movement of a
vacant site and result in non-equilibrium concentration gradients.
The local changes in composition may induce precipitation of a
second phase at sites of nucleation. This phenomenom of
"radiation-induced segregation" is activated independent of
thermal energy,and may be the mechanism responsible for the
formation of the nickle-sulfide precipitants observed in Figure
23.

The movement of nickel atoms by the vacancy mechanism of

diffusion could result in zinc vacancies being filled by nickel

141
atoms, and subsequent bonding of the nickel to adjacent sulfur
atoms. The concentration of zinc vacancies is also believed to be
higher than that of sulfur due to their relative displacement
energies within a crystal of ZnS. During the ion implantation
process, sulfur atoms require 15 to 20 electron-volts of energy
transferred to them to be displaced from their atomic positions,
whereas zinc need only absorb 7 to 9 electron-volts to become
dislodged from their atomic positions [42].

Thermally assisted atomic motion may have commenced with the
minimal beam heating of the ion milling procedure and the TEM
studies. The result of this energy application on the conditions
described above may have been sufficient to result in additional
nickel-sulfide embryo formation leading to the precipitation
transformation reaction described in Section 1.2 (p.7).
Radiation—induced segregation moves a structure away from an
equilibrium state. Therefore, following irradiation, conditions
are highly-conducive to the diffusion of species due to the excess
number of defects in the metastable energy state. This radiation-
enhanced diffusion, being thermally activated, drives the energy
state of the matrix closer to a state of equilibrium.

Section 3.4 includes TEM observations of precipitants in the
implanted ZnS. The corresponding selected-area diffraction
pattern included very fine crystalline rings,indexed to be those
of one or more phase of nickel sulfide. Their presence in the
absence of the annealing procedures strongly suggest that one or
more of the mechanims discussed here may have been active prior to

heat treatment of the irradiated samples. It should be noted from

142
Table 9, the indications that the low-temperature stable phase of
B-NiS may have been present in greater concentration in the pre-
annealed samples than the precipitants of the meta-stable, high—
temperature phase, a-NiS. This outcome may be due to the relative
dominance of the two radiation-assisted diffusion mechanisms of
radiation-induced segregation and radiation—enhanced diffusion.
It may also be relevant to the 5 percent increase in hardness
measured for the pre-anneal samples.

During ion implanation, RIS may have been more active than
RED for several reasons. Due to the relatively high dose with
which nickel ions were implanted into zinc sulfide, a higher
concentration of vacancies would be generated and able to
contribute to the radiation—induced segregation mechanism.
Additionally, the zinc sulfide was maintained at a relatively low
temperature during implantation, therefore the temperature—
dependent RED activity was suspected of being mimimal at that
point.

The procedures and results of annealing nickel-irradiated ZnS
to temperature of 750, 800, and 850 degrees celsius in an argon
environment were presented in Sections 2.5 and 3.5, respectively.
The Vicker's hardness was shown to improve after annealing.
Complete analyses was not performed to determine the actual nickel
profile following the heat treatment cycles. However, attempts by
XPS to find nickel at the surface in the ZnS following the 800
degrees anneal were unsuccessful. The surface was then sputtered
to a depth of 100 nanometers, approximately the pre—anneal range

of the implanted nickel ions. Nickel was not detected by XPS even

143
at this depth. This result is consistent with the high
diffusivities which have been associated with nickel, even with a
moderate elevation in temperature. Although specific data was not
cited, Woodbury attested to the high diffusivity of nickel in zinc
sulfide in a commentary on work conducted by Pappalardo and Dietz
[43].

From [44], nickel and sulfur combined in a 1:1 ratio were
found to produce a-NiS in a temperature range of 400 to 900
degrees celsius. The corresponding range of time for this product
to form was between 4 and 168 hours. Assuming the minimum
temperature for the a—NiS reaction as 400 degrees celsius, the
thermal profiles in Figure 18 show the time duration above this
temperature for each anneal cycle. The time above 400 degrees for
the 750 degree anneal was 600 seconds, while the 800 and 850
degree anneals were above 400 degrees for 1,080 seconds. The heat
treatment durations for nickel—irradiated zinc sulfide were much
less than the time for the a-NiS annealing in [44]. However, the
highly metastable conditions within the sub—surface region of the
irradiated ZnS represented a state of enhanced diffusivity.

The self—diffusivity of nickel can be calculated from the

experimentally determined diffusion equation:

D = Doexp(—Q/RT) (7)

where, D = diffusion coefficient, of diffusivity, cmZ/sec

D0 = material dependent factor, cmZ/sec

144

Q 2 activation energy, kilo-Joules/mole

50
ll

8.314 Joules/mole—Kelvin

r—l
ll

temperature, Kelvin [44].

Do and Q for the self—diffusivity of nickel are 1.5 and 283.0
kilo-Joules/mole, respectively [44]. The following diffusivities
for nickel were calculated using the anneal temperatures for the

nickel-irradiated ZnS:

Dl(1023K) = 5.315x10‘ls cmZ/sec
D2(1073K) = 2.505x10'l4 cmZ/sec
D3(1123K) = 1.029x10'l3 cmz/sec

These values are approximations and serve more to highlight
the change from 1023K to 1123K. The increase from D1 to D2 is
greater than that from D2 to D3. This correlates to the increase
in hardness observed for the annealed specimens, illustrated in
Figure 31. The actual diffusivities may be significantly higher
than those calculated. Self—diffusivity is a measure of diffusion
of a species within itself. The melting point of nickel is 1453
degrees celsius, therefore, the bond-energy between nickel atoms
is significant. This bond-energy would be less for nickel atoms

surrounded by zinc and sulfur atoms. And again, a condition of

145
the highly metastable environment created by ion implantation is
one of enhanced mobility

Nickel sulfide formed in the subsurface region of the ZnS may
have diffused to depths much greater than the range during the
annealing procedure. Due to the depth of the indents made by the
Vicker's hardness tester, it was originally stated that the
resulting hardness values represented those of a composite
consisting of the thin implanted layer and the unmodified bulk.

If NiS did in fact diffuse to significant depths away from the
surface, as is eluded to by the failure of XPS to detect nickel at
the range depth, then the increase in Vicker's hardness observed
for the annealed samples more closely represents the hardness
improvements for the modified layer of the ZnS. This increases
the sensitivity of the Vicker's indentor to changes exhibited by
the presence of nickel sulfide, i.e. a positive change in volume
present from the a —> B phase transformation induced by the energy
from the indentor during the hardness test.

To estimate the nickel concentration after annealing at a
certain depth, several models may be develop after the diffusion
couple or thin-film solutions to Fick's Second Law for Diffusion
[45]. Auger analyses of annealed specimens would also provide
empirical data for understanding the diffusion behavior of nickel

in the ZnS structure.

Atoms/0M3

x1822

1.0

146

 

 

 

l !

 

 

sob

Figure 41.

1280 1808 2488

Distribution of 200keV nickel

ions implanted into zinc sulfide.

3m DEPTH

 

CHAPTER 4. CONCLQSION

Zinc sulfide, an infrared transmitting ceramic, was ion
implanted with a dose of 1017 nickel ions/cm2 at 200 keV under a
controlled atmosphere of argon, with the surface temperature
retained below 200 degrees celsius.

The specimens were then prepared for scanning electron and
transmission electron microscopic evaluations, surface hardness
and fracture toughness measurements, elemental analyses by x-ray
photoelectron spectroscopy, and x-ray diffraction.

Evaluation of the mechanical properties was conducted on
samples before and after heat treatments to assess the effect that
annealing may have on facilitating nickel diffusion and
precipitation of nickel sulfide, believed to be a metastable
transformation toughener.

The presence of nickel and any precipitants forming in the
implanted substrate was investigated by transmission electron
microscopy in selected-area diffraction mode. Zinc sulfide was
used as the internal standard by which the diffraction pattern
measurements were calibrated. This work revealed various nickel
sulfide phases in the subsurface region of the implanted ZnS.

Motivation for this work followed the supposition by w.M.
Kriven that nickel sulfide possesses the attributes of a stress-
induced transformation toughener. Implantation of nickel into
zinc sulfide, followed by thermal treatment, may induce the lower
volume metastable NiS phase to form in the subsurface region.

147

148
Upon the application of energy in the form of impact with an
accelerating force, this phase is believed to transform to a more
stable phase of N18 that requires a greater volume within the
matrix. This volume increase would then establish a compressive
stress in the surface region, thereby increasing the surface's
resistance to impact damage.

Transmission electron microscope diffraction patterns of
nickel-implanted zinc sulfide revealed both a and B nickel-sulfide
phases as precipitants, as well as N182. Precipitant sizes were
measured to be approximately 5 to 10 nanometers, thereby being
coherent in the matrix of zinc sulfide grains determined to be
approximately 5 microns. Beam heating may have induced initial
precipitant growth prior to annealing. Additional growth with
heat treatment may have produced precipitants semi-coherent with
respect to the matrix.

The average increase in hardness observed for annealed
nickel-implanted zinc sulfide as compared to unmodified zinc
sulfide was approximately 25 percent. This has been attributed to
the presence and action of metastable a-NiS in the subsurface
region of the ZnS. This phase of NiS undergoes a positive change
in volume when transformed to B-NiS. It is suggested that
transformation may have been induced by stresses associated with
the impact of the Vicker's hardness indentor. As the energy
propagated through the surface region, the atoms of a fine
distribution of metastable a-NiS phases absorbed sufficient energy
to displace their positions to the lower energy arrangement

of B-NiS. This represents an increased resistance to surface

149
damage which may then be sustained further in the region of impact
by the presence of a residual surface compressive stress due to
the volume increase. Additional "shielding" from crack
propagation is established by the semi-coherency of B-NiS
precipitants which constrain the ZnS matrix.

There was no significant change observed on the grain size
of the zinc sulfide following the annealing procedures, although
there was a small increase in hardness observed for control
samples of zinc sulfide. Certain measures were taken to minimize
possible contamination of the control by volatilized species in
the annealing chamber, however, this may be the cause for the
increase in hardness of the control samples. The ion implanted
nickel was determined by x-ray photoelectron spectroscopy to have
diffused further into the bulk away from the pre-anneal range of
1100 angstroms presumably by the vacancy mechanism of diffusion.
XPS also measured the presence of both nickel and sulfur, in pre-
anneal specimens, with binding energies indicative of nickel-
sulfer bonds.

A preliminary evaluation on the refractive index of zinc
sulfide by Fourier Transform Infra—Red spectroscopy indicated that
the ion implantation of nickel in the surface region of the ZnS
resulted in no observable degradation A change in refractive
index for a material is dependent upon changes in polarization,
absorption, and volume. Although the phase change from a to B
nickel sulfide results in a 4 percent increase in volume, the
relative increase for the entire bulk material of ZnS would be

negligible.

150

Primary sources of error in this work may potentially lie in
the camera constant calibration and selected-area diffraction beam
calibrations, as well as measurement errors on the SAD diffraction
patterns. The load used for Vickers hardness measurements was too
high to measure the hardness increase for the modified layer
alone. The post-anneal composite hardness did show an increase,
however, it would be difficult to extrapolate a hardness for the
modified layer from this data.

Additional work of interest would include investigating the
depth profile and diffusion behavior of nickel by XPS or Auger
Electron Spectroscopy in the post-annealed samples. Selected-area
diffraction patterns of irradiated and annealed specimens would
allow indexing the final precipitants and their sizes.
Additionally, microhardness measurements may be more
representative of the modified surface alone, rather than the
composite. Micro—micro diffraction of a single precipitant would
provide detail regarding the nature, size, and orientation of
precipitants in the subsurface region of an implanted target
material while in—situ annealing in a transmission electron
microscope would reveal information about the precipitation

process.

APPENDICES

APPENDIX A

VICKERS INDENT HARDNESS DATA

APPENDIX A "

VICKERS INDENT HARDNESS DATA

Po; speCimen ZnS spec.x non-irr. annealed test
-"dented on 7-24-89
ndent Load(kg) a1(um) a2(um) cl(um) :2;um)
1 0.30 53.00 55.10 C.00 C.DC
2 0.30 53.20 54.70 0.00 0.00
3 0.30 51.90 53.10 0.00 C.00
4 0.30 52.80 54.00 0.00 0.00
5 0.30 52.90 54.30 0.00 0.00
6 0.30 55.80 52.30 0.00 0.00
7 0.30 55.70 55.50 0.00 C.00
8 0.30 52.40 53.40 0.00 C.00
9 0.30 51.90 52.40 0.00 0.00
10 0.30 54.00 59.00 0.00 0.00
11 0.30 52.70 55.30 0.00 0.00
12 0.30 51.90 57.50 0.00 0.00
Break in 4600
0k
indent Amean(m) Cmean(m) Hard(GPa) Kic(MPam1/2)
1 27.0258-06 0.000E+00 0.187E+01 U.UUUUL+UU
2 26.975E-06 0.000E+00 0.187E+01 0.0000E+00
3 26.250E-06 0.000E+00 0.19BE+01 0.0000E+00
4 26.700E-06 0.000E+00 0.191E+01 0.0000E+00
5 26.800E-06 0.000E+00 0.19OE+01 0.0000E+00
6 27.0253-06 0.000E+00 0.187E+01 0.0000E+00
7 27.800E-06 0.000E+00 0.177E+01 0.0000E+00
8 26.450E-06 0.000E+00 0.195E+01 0.0000E+00
9 26.075E-06 0.000E+00 0.201E+01 0.0000E+00
10 28.250E-06 0.000E+00 0.171E+01 0.0000E+00
11 27.000E-06 0.000E+00 0.187E+01 0.0000E+00
12 27.350E-06 0.000E+00 0.182E+01 C.0000£+00
Mean hardness +/- std = 1.877393 8 4273:13-02

Coefficient of variation for hardness =

Mean Kic +/~ std = 0 0
Coefficient of variation for Kic = 0

1'tvtf1‘************************i‘f'l’i't*‘k‘k‘k

151

152

r.: specimen ZnS see: . no: 1:- c
incentec on 7-24-55
1‘:en: Loadtkg. al.3m; a: um
1 0.30 56.20 55.30
Z 0.30 53.30 55.40
3 0.30 55.80 57 70
4 0.30 55.00 55 LC
Break in 4600
3k
laden: Amean(m) Cmean(m)
1 27.875E-06 0.00DE+00
2 27.175E-06 0.0005+00
3 28.3755-06 0.0005+00
4 27.525E-06 0.0005+00
Ween hardness +/- std = 1 "744
Coefficient of variation for nardn
Mean 31c +/- Std = 0
Coefficient of variation for Ric =

I'Hrek'w

f'l'r"r*f**********************fiff'l'f'k'l’f?

Cimen

Load(kg)
0.30
0.30
0.30
0.30
in 4600

For specimen ZnS s
anented on 7-24-

a1(um)

p
8

ZnS spec.M irr.
7-24-89

55
56

~-

3 fl

.4

56

P
H

mtD

a2(um)
.40 56.50
.70 56.20
.90 55.50
.10 56.30
A irr pre-a

Indent Ameantm) Cmean(m)
1 27.9755-06 0.0005+00
2 28.225E-06 0.000E+00
3 27.350E-06 0.0003+00
4 28.100E-06 0.0005+00
Mean hardness +/- std = 1.7517
CoeffiCient of variation for hardn
Mean Kit +'- Std = 0
CoeffiCient of variation for Kic =

pre-anne

«firs-:2-
3... -“—
fl- 9‘"-
_._ mun

K. . v\.

F ".."

La . vb

C th

.ab

A «A.

v . J)»
- R'—

H ._.'
r“ 4U-a)
f" h-
7czsu-
F“ q
Ecz+01
rh— A-
CD‘S-"NJ;
80E+'1

C
O
3.). (.95;

cltum)
0.00
0.00
0.00
0.00
nneal :es:

m

rd(GPa)
74E+Ol
7lE+01
.182E+01
1‘3—+01
s = Z

7?7Y7ff****f**‘l’f‘kf'k'k'f'f‘l’f‘l'i'f‘r'l'i'i'f‘l’f‘r‘r'r'r‘

11
w
(I)
1'

U) ‘0
(A) 01

} I

0‘.
Q

('1 (W ('3 (73

’I

()U()()H
(') (1) (D ()

(-. t“) t i (W L;

F-QA“

d--d~

(.7

mild
I‘l)
-Jm
him

0‘.
q 0

(3 C) or)

*1.

(J (300”

(_) O C.) (D -

mrnrwm I

x
3

C.) (_) (J C.)

O(D(DO

()L)()()
(\()I')()

‘v
-.

()O(J()}

.4

(JL)O(J%)

aml.
:~ac
E+OC
E+00
. ..
2+OC
aml

I)

153

5:: :tecimen ZnS Spec.2 annealed 750-0
-"certed or 8-10-89
-ndent Load(kg) a1(um) a2(um) :lium)
1 0.30 53.40 53.10 157 30
2 0.30 54.40 55.10 104 60
3 0.30 54.70 55.30 :16 0
4 0.30 55.40 55.00 122 7
5 0.30 53.00 54.90 10‘ 90
6 0.30 53.80 54.40 148 30
7 0.30 55.20 55.40 139.60
8 0.30 54.30 53.70 134.40
9 0.30 54.60 53.60 118.10
;: 0.30 54.10 55.90 109.00
11 0.30 54.00 53.30 11; 90
12 0.30 52.90 54.70 115 00
Break in 4600
3k
1 27.0255-06 0.0008+00 0.187E+01
2 26.9755-06 0.000E+00 0.187E+01
3 26.2505-06 0.000E+00 0.1985+01
4 26.7005-06 0.000E+00 0.19lE+01
5 26.8005-06 0.000E+00 C.1905+01
6 27.0255-06 0.000E+00 0.187E+01
7 27.8005-06 0.000E+00 0.177E+01
8 26.4505-06 0.0003+00 0.19SE+01
9 26.0755-06 0.000E+00 C.2015+01
10 28.2505-06 0.000E+00 0.1715+01
I; 27.0005-06 0.000E+00 C.187E+01
-2 27.3505~06 0.000E+00 c.1323+01
eat hardness +/- Std = 1.877393 5
:zefficien* of variation for hardness = 4
”ea: pic: +/- Std = O C
C:ef:icienr of variation for Kic = 0

v1171'"771*'r‘l’frff'r'l'f'l'f‘l‘f‘l'fi’tiffffi’ft1771

.OOOC
.OOOGE
.OOSOE+OO
.COOOE+OO

('1(Wh(fihhn(—Wnnnh

nlllUu U\ L.) '
DOC) (‘WOfl

\)(;)C)OUI
(700

~.) L)

0\ C7 (1‘

-a+oo
-00

Aflflhv—

A .

0 '~-:
.uuuu— J

.OOCQE+OO
.OOCQE+OO
.00003¢00
.COCQE+OO

A-—.~---

5:: sce:-men 715 Spec M 1:: annealed test Eat’;. C
;FQEEZEC on 7—31-89
:"aent Load(kg) a1(um) a2(um) 21(um) Sltum)
1 0.30 49.90 48.30 C.00 0.00
2 0.30 50.40 50.40 0.00 0.00
3 0.30 52.30 48.50 C.00 C.00
4 0.30 48.60 46.50 0.00 0.00
5 0.30 48.00 48.40 0.00 0.00
6 0.30 48.40 49.80 0.00 0.00
7 0.30 47.70 49.50 0.00 0.00
8 0.30 48.40 51.40 0.00 0.00
9 0.30 50.50 51.60 0.00 0.00
10 0.30 49.70 48.70 0.00 0.00
11 0.30 47.50 49.40 0.00 0.00
12 0.30 50.50 49.50 0.00 0.00
Break in 4600
Ok
ndent Amean(m) Cmean(m) Hard<GPa> Kic(MPam1{2}
l 24.550E-06 0.000E+00 0.226E+01 0.00005+00
2 25.2003-06 0.000E+00 0.215E+01 0.0000E+00
3 25.2005—06 0.000E+00 0.2158+01 0.0000E+00
4 23.7755-06 0.000E+00 0.24lE+01 0.00005+00
5 24.100E-06 0.000E+00 0.235E+01 0.0000E+00
6 24.550E-06 0.000E+00 0.226E+01 0.0000E+00
7 24.3008-06 0.000E+00 0.23lE+01 0.0000E+00
8 24.9505-06 0.000E+00 0.219E+01 0.00005+00
9 25.5255—06 0.000E+00 0.209E+01 0.00005400
10 24.6005-06 0.000E+00 0.22SE+01 0.00005+00
11 24.2255-06 0.000E+00 0.23ZE+01 0.00005+00
12 25.0005-06 0.000E+00 0.2188+01 0.00005+00
Mean hardness +/- std = 2.24517 9.4789855-02
Coefficient of variation for hardness = 4.2219478-02
Mean Kit +/- Std = 0 0
Coefficient of variation for Kic = 0

ftt‘t‘t‘rf'tf‘kf'k******************Y**7*Y‘l'i'i'

155

For specimen ZnS IRR. Spec.Q annealed 850 C
Indented on 8-10-89
Indent Load(kg) a1(um) a2(um) c1(um) c2(um)
l 0.30 47.80 51.30 0.00 0.00
2 0.30 48.60 50.50 0.00 0.00
3 0.30 48.90 50.30 0.00 0.00
4 0.30 45.90 47.90 0.00 0.00
5 0.30 49.40 51.30 0.00 0.00
6 0.30 47.60 51.30 0.00 0.00
7 0.30 48.80 50.00 0.00 0.00
8 0.30 51.30 50.00 0.00 0.00
9 0.30 48.90 48.20 0.00 0.00
10 0.30 47.70 49.80 0.00 0.00
11 0.30 48.50 49.70 0.00 0.00
12 0.30 48.50 52.60 0.00 0.00
Break in 4600
Ok
Indent Amean(m) Cmean(m) Hard(GPa) Kic(MPam1/2)
1 24.775E-06 0.000E+00 0.222E+01 0.0000E+00
2 24.775E-06 0.000E+00 0.222E+01 0.0000E+00
3 24.800E-06 0.000E+00 0.222E+01 0.0000E+00
4 23.450E—06 0.000E+00 0.24BE+01 0.0000E+00
5 25.175E-06 0.000E+00 O.215E+01 0.0000E+00
6 24.725E-06 0.000E+00 0.223E+01 0.0000E+00
7 24.700E-06 0.000E+00 0.224E+01 0.0000E+00
8 25.325E-06 0.000E+00 0.213E+01 0.0000E+00
9 24.275E-06 0.000E+00 0.231E+01 0.0000E+00
10 24.375E-06 0.000E+00 0.230E+01 0.0000E+00
11 24.550E-06 0.000E+00 0.226E+01 0.0000E+00
12 25.2755-06 0.000E+00 0.214E+01 0.0000E+00
Mean hardness +/- std = 2.241701 9.524485E-02
CoeffiCient of variation for hardness = 4.248777E-02
Mean Kic - Std = 0 0

/
+/
Coefficient of variation for Kic = 0

APPENDIX B

BASIC PROGRAM - HARDNESS IN GIGAPASCALS

1000
1010
1020
1040
1060
1080
1100
1200
1340
1360
1640
1720
1740
1760
1820
1860
1920
1940
2000
2020
2040
2045
2060
2080
2090
2120
th

2140
2160
2180
2200
2380
2400
2410
2420
2480
2680
2740
2750
2760

2780
2820
2880
2990
3000
3020
3040
3060
3080
3100
3180
3900
4000
4080
4090
4100
4120
4300
4320

APPENDIX B

BASIC PROGRAM - HARDNESS IN GIGAPASCALS

REM Hardness and toughness calculations

N = 30

DIM KG(N),A1(N),A2(N),C1(N),C2(N)

DIM H(N). KIC(N),AMEAN(N), CMEAN(N), HMEAN(N),KCMEAN(N)

READ NSPEC : REM read the number of specimens.npsec

CLS

FOR K = 1 TO NSPEC
GOSUB 3000 : REM Read input data for specimen
GOSUB 4000 : REM View input data on CRT

REM GOSUB 14000: REM Output initial data to lineprinter
GOSUB 2000 : REM Compute hardness
GOSUB 2400 : REM Calculate fracture toughness
GOSUB 6000 : REM Statiscics on hardness
GOSUB 7000 : REM Statistics on toughness
GOSUB 5000 : REM Output results to CRT

REM GOSUB 15000: REM Output results to lineprinter
NEXT K

END
REM Calculate hardness
FOR I = 1 TO NN

P(I) = KG(I)*9.8 :REM Multiply "mass" by acceleration due to gravity to
REM obtain the load P.
CONV = .000001 : REM convert microns (read from Buehler indentor) to meter
A1(I) = A1(I)*CONV : A2(I) = A2(I)*CONV
Cl(I) = C1(I)*CONV : C2(I) = C2(I)*CONV
AMEAN(I) = (A1(I) + A2(I))/4! : REM Amean is the average half—diagonal leng

ASQ = AMEAN(I)*AMEAN(I)

H(I) = (.464*P(I))/ASQ

H(I) = H(I)*9.999999E-10 : REM Convert hardness in Pa to GPa
NEXT I

RETURN

REM Calculate fracture toughness
FOR I = 1 TO NN

IF (Cl(I) = 0 AND C2(I) = 0) THEN FTEST = O

CMEAN(I) = (Cl(I) + C2(I))/2!

CAL = .016 : REM Cal is a material-independent calibration constant

FAC = SQR(E/H(I)) : REM Hard is the calculate hardness for the indent

REM Note: Both E and H are in GPa, so that FAC is unitless

DENOM = CMEAN(I)*(SQR(CMEAN(I))) : REM Kic depends on cmean to the 3/2 powe
KIC(I) = (CAL*FAC*P(I))/DENOM

KIC(I) = KIC(I)*.000001 : REM Convert Kit in Paml/2 to MPam1/2

NEXT I

RETURN

REM Input data for specimen
READ SPECS, DATS : REM Read the specimen label and date
READ NN,E: REM read the number of indents for this specimen
REM and the Young's modulus E, in GPa
FOR I = 1 TO NN
READ KG(I),A1(I),A2(I),C1(I),C2(I)

NEXT I

RETURN

REM View initial data

PRINT " " : PRINT " "

PRINT"For specimen ",SPECS : PRINT "Indented on ",DATS : PRINT " "
AS = " ### ##.## ###.## ###.## ###.## ###.##"
PRINT "Indent Load(kg) a1(um) a2(um) c1(um) c2(um)"

FOR J = 1 TO NN
PRINT USING A$;J,KG(J),A1(J).A2(J),C1(J),C2(J)

156

157

440C NEXT J

4600 STOP

499C RETURN

5000 REM Output hardness, toughness to CRT

5020 PRINT " " : PRINT " "

5040 PRINT"Por speCimen ".SPECS : PRINT "Indented on ",DATS : PRINT " "
5060 BS = "### ###.###““ #.###““ #.###““ #.####““ "
3080 PRINT"Indent Amean(m) Cmean(m) Hard(GPa) Kic(MPam1/2z "
5100 FOR I = 1 TO NN

5140 PRINT USING BS; I. AMEAN(I), CMEAN(I). H(I), KIC(I)

5400 NEXT I

5500 PRINT " "

5600 PRINT "Mean hardness +/— std = ",HMEAN,HSTD

5620 PRINT " CoeffiCient of variation for hardness = ",HCPV

5500 PRINT " "

5820 PRINT "Mean Kic +/- std = ",KMEAN,KSTD

5840 PRINT " Coefficient of variation for Kic = ",KCEV

5360 pRINT " H: pRINT"ttittttttiiitttiiittitti'ttittfitttitti":pRIIKIT""
5900 STOP

€000 REM This subroutine calculates the mean, variance, Standard deviation
5010 REM and coefficient of variation for vector H

5030 SUM = 0!

6040 SQSUM = 0!

6050 FOR I = 1 TO NN

6060 SUM = SUM + H(I)

6080 SQSUM = SQSUM + H(I)*H(I)

5100 NEXT I

6120 HMEAN = SUM/NN

6140 HVAR = (NN*SQSUM - SUM*SUM)/(NN*(NN - 11))

5160 HSTD = SQR(ABS(HVAR))

£180 HCFV = HSTD/HMEAN

€900 RETURN

"000 REM This subroutine calculates the mean, variance, standard deviation
“010 REM and coefficient of variation for vector KIC

“030 SUM = 0!
7040 SQSUM = 0!
"050 FOR I = 1

TO NN
"360 SUM = SUM + KIC(I)
"380 SQSUM = SQSUM + KIC(I)*KIC(I)
~100 NEXT I
"120 KMEAN = SUM/NN
"140 KVAR = (NN*SQSUM - SUM*SUM)/(NN*(NN - 13))
“160 KSTD = SQR(ABS(XVAR))
“180 KCFV = KSTD/KMEAN
“900 RETURN
14000 REM View initial data
14080 LPRINT " " : LPRINT " "
:4090 LPRINT"For specimen ",SPECS : LPRINT "Indented on ",DATS : PRINT " "
;4100 AS = " ### ##.## ###.## ###.## ###.## ###.##"
14120 LPRINT "Indent Load(kg) a1(um) a2(um) c1(um) c2(um)"
L4300 FOR J = 1 TO NN
14320 LPRINT USING A$;J,KG(J),A1(J),A2(J).C1(J),C2(J)

14400 NEXT J
14990 RETURN
;5000 REM Output hardness, toughness to lineprinter

15020 LPRINT " " : LPRINT " "

LSO40 LPRINT"Por specimen ",SPECS : LPRINT "Indented on ",DATS : PRINT " "
L5060 85 = "### ###.###‘““‘ #.###““ #.###“‘“ #.####““ "
5080 LPRINT"Indent Amean(m) Cmean(m) Hard(GPa) Kic(MPam1/2) "

15100 FOR I = 1 TO NN

LPRINT USING BS; 1. AMEAN(I), CMEANU). PHI). 1".

NEXT I
LPRINT
LPRINT
LPRINT
LPRINT
LPRINT
LPRINT
RETURN

158

"Mean hardness +/- std = ",HMEAN.HSTD
" Coefficient of variation for hardness = "
"Mean Kic +/- Std = ",KMEAN.KSTD

" Coefficient of variation for Kic = ".KCFV

H
H

ycpv
' O».

" ": LPRIN’I‘"tittifittittif'fitttttttittvttttit?tit?N :LPRINT'H'

LIST OF REFERENCES

Paul Kuttner, "Measuring The Critical Properties Of
Infrared Optical Materials," Laser Focus/Electro Optics,
April 1985, pp. 91-99.

D. F. Horne, Optical Production Technolpgy, 2 ed., Ch.8,
Adam Hilger Ltd., Bristol, 1983.

Kenneth Krane, Modern Physics, Ch.3, John Wiley and Sons,
Inc., 1983.

E. D. Case and A. G. Evans, "Length Of Maximal Impact
Damage Cracks As A Function Of Impact Velocity," Proc.
Int. Conf. on Erosion by Liquid and Solid Impact,
University of Cambridge, 1983.

E. D. Case, K. M. Louie, and A. G. Evans, "Statistical
Analysis Of Damage Induced By Water Drop Or Water Jet
Impact," Journal of Materials Science Letters, 3, 1984,
pp. 879—884.

J. E. Field, S. van der Zwaag, and D. Townsend, "Liquid
Impact Damage Assessment For A Range Of Infra-Red
Materials", Proc. 6th Int. Conf. on Erosion by Liquid and
Solid Impact, University of Cambridge, 1983.

I. W. Donald and P. W. McMillan, "Review: Infra-Red
Transmitting Materials: Part 1, Crystalline Materials,"
Journal of Materials Science, 13, 1978, pp. 1151-1176.

R.S. Averback and M.A. Kirk, "Atomic Displacement
Processes in Ion Irradiated Materials", Sprface Alloying
by Ion, ElectronI and Laser Beams, eds. L.E. Rehn, S.T.
Picraux, H. Wiedersich, American Society for Metals,
Metals Park, 1985.

David K. Brice, Ion Implantation Range & Energy

Deposition Distriputions Vol. 1, 1975, IFI/Plenum,
New York, p.l.

159

10.

11.

12.

160

J.K. Hirvonen, ed., Treatise On Materials Science And

Technology: Ion Implantation Vol. 18, 1980, Academic
Press, New York.

W. Hayes and A.M. Stoneham, Defects and Defect Processes
in Nonmetallic Solids, John Wiley and Sons, 1985.

I.J. McColm and N.J. Clark. Forming, Shaping and Working

of High-Performance Ceramics, 1986 Chapman and Hall,
New York, p.334.

13.

14.

15.

16.

17.

18.

19.

20.

21.

P.J. Burnett and T.F. Page, "Surface softening in silicon
by ion implantation", Journal of Materials Science, 19,
(1984), pp. 845-860.

C.J. McHargue, et. al. "Microstructure and Properties of .
TiB2 Implanted with 1 MeV Nickel", Nuclear Instruments 1
and Methods, B1 (1984) pp. 246-252.

A.G. Evans, (S.W. Freiman, ed.), Fracture Mechanics

Applied tp Brittle Materials, American Society for
Testing Materials, Philadelphia, PA, 1979, p.112.

R.H. Marion, Ref. 15, p.103.

C.J. McHargue, et. al., "The Structure and Properties of
Ion-Implanted A1203", Nuclear Instruments and Methods,
BIO/11, 1985, pp. 569-573.

T. Hioki, et. al., "Effect of Ion Implanation on Fracture
Stress of A1203", Nuclear Instruments and Methods, B7/8,
1985, pp. 521-525.

T. Hioki, et. al., Journal of Materials Science
Letters,(in press).

G.B. Krefft and E.P. EerNisse, Journal of Applied
Physics, Vol. 49, 1978, pp. 2725-2730.

Y.S. Touloukian, R.K. Kirby, R.E. Taylor and T.Y.R. Lee,
"Thermal Expansion, Nonmetallic Solids," Thermophysical
Properties of Matter Vol. 13, Plenum, New York, 1977.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

161

K.O. Legg. et. al., "Modification of Surface Properties
of Yttria Stabilized Zirconia by Ion Implantation,

Nuclear "Nuclear Instruments and Methods, B7/8, 1985, PP.
535-540.

Carl J. McHargue, "Structure and Mechanical Properties
of Ion Implanted Ceramics", Nuclear Instruments and
Methods, B10/11, 1987, pp. 1127-1133.

C.H. Mathewson, Zing: The Spience sng Technology of the
Metal, Its Alloys and Compounds, Plenum, New York, 1967.

B. Cooper, "IR Window Advances: FLIRS and Beyond,"
Photonic Spectra, Oct., 1988, pp. 149-156.

Max Hansen, Constitution of Binary Alloys, McGraw-Hill
Book Company, Inc. New York, 1958, p.1171.

R. C. Wheast, ed. Handbook of Chemistry and Physics
62nd ed., 1981-82, CRC Press, p.f-54.

W. D. Kingery, Bowen, Uhlmann, Introduption tp Ceramics
John Wiley and Sons, Inc., New York, 1976, p.63.

D. F. Horne, Optical Production Technology, 2 ed. 1983,
Adam Hilger Ltd., Bristol, p.249.

M. V. Swain, "Nickel sulphide inclusions in glass:

an example of microcracking induced by a volumetric
expanding phase change," Journal of Materials Science,
16, 1981, pp. 151-158.

W. M. Kriven, "Possible Alternative Transformation
Tougheners to Zirconia: Crystallographic Aspects,"

(in press) Dept. of Materials Science and Engineering.,
University of Illinois at Urbana-Champaign.

G. Thomas and M.J. Goringe, Transmissipn Eleptrpn

Microscopy of Materials, John Wiley and Sons, Inc.,
New York, 1979, p.13.

Powder Diffraction File, Joint Committee on Powder
Diffraction Standards, International Centre for
Diffraction Data, Swarthmore, PA.

B.D. Cullity, Elements pf X-rsy Diffraction, Addison
Wesley Pub. Co., Reading, MA., 1978, p.84.

35.

36.

37.

38.

39.

40.

41.

42.

162

Beeston, Horne, and Markham, Electron Diffraction and

Optical Diffraction Technigpes, American Elsevier
Pub. Co. Inc, New York, 1973, Ch.4.

C.R. Rarey, J. Stringer and J.W. Edington, Trans.
A.I.M.E., Vol. 236, p.811.

P. Worthington, "Observation of Polytypes and Stacking
Faults in Zinc Sulphide by Transmission Electron
Microscopy," Journal of Materials Science, Vol. 8, 1973,
pp. 1194-1200.

D.S. Grummon and E.D. Case, private communication.

Gatan Corp. Dimple Grinder Model 686 Operations Manual.
G. Kullerud and R.A. Yund, "The Ni—S System and Related
Minerals," Journal of Petrology, Vol. 3, Part 1, 1972,
pp. 126-175.

H.H. Woodbury, "Diffusion and Solubility Studies,"
Physic and Chemistry of II-VI Compounds, Aven and Prener
p.232.

Paul G. Shewmon, Diffusion in Solids, J. Williams Book
Co., Jenks, OK, 1983.

 

MICHIGAN smrE UNIV. LIBRRRIES
1|)WIN)Illlllml”IHIW)HI))NlllHlNl)“WWI
31293010432064