"E22 :.' .2".

‘d: 21
.y. -

um. uv «2.7. ”-1.”.

4g- 2"! pt? .2

 

.. 2 22‘..2_

$1;

« ~42

 

im b‘fl-WL‘.‘
Mgr; 122.22. c u : . w .. "mg?” ., :: W?—
V “ " ‘6 3“;- ’3; ., ‘7 ‘ 2215:“ :

J‘fé.‘
‘ 2"” 2. .. 2.“

‘L':

”'3.“ u ' “I:- '

£5?» ‘C‘ k '3 \ «a. uni-Jr: \- in“: 22:3?“1‘54312'

A 33:;:!‘ 01‘ ..r ~ . « 2 2 .2 m-

‘i' , ME: ~ ~- - 2 “3.25- “2;; ‘ ”~22 new.
.2... “.2 2:... fi, 3 “(.419

2.23.5” “'“.. 2.2.13»;
9,»? ,. .... ”ft.

. 3‘ ‘vé‘~ ‘
‘ "“f‘uiz‘bfir
‘ ‘ .‘ ‘LI‘

.222,“
nihiltfi-EA

‘2T 2 «7 3 2,. “k 7 1 , 2 .12.
" ’1'?” “A x . - n. -: 2 Ju- , .
‘ ' 2 .2 2,27 ~.. . m- "fi‘fip‘fifi? “21:39“:

‘L ‘1'“ “1"2‘1' " 1 WV "axlil‘gytru‘ fig}

2 ;: q » .2 _‘
”‘7’ “~ 13-2 .2. 2E ’ '7 4.2! ‘ $51.:
‘ ‘ d l”;:;:
. . . -

2». f2

‘ ‘2‘; L23;

. ~22‘r

v.33? “:2. 3
22-6: 2\ «(J-2.1."
ft $32.31,} «2:14:45
\:
‘r .3:

.45‘ «H
"7;! 2
'5 s.

.2 -.
.' .
'I r‘
"2‘2 ,.
32g.
, 2

' 2 I” _.
- :51; :
1‘“ w: «A
"‘ £322": ‘34,.

.22-
: M5":.',~:' {’1‘

e
JC 1.2-
2 ‘5"

: 5W.
. ..1:r~.-'.».: €({:’:t;:-"-
" 5:221:21: 3" :,3:?'{_ 1’3
J x:.

”r I
I {A 3‘
I» x .3
5"2‘ 0 \rlifi'r
, .
2 v-
'51,“ " 3":

.,
3:54

um.
I"

. ”waging.
11:21:... firmhzm .2.”

H1vn+1‘:-W""
W5": 2: .mra...
..... 2.

«-
¢
.0" 'vw’t 15 "L Lyn—7'
.43.... 1:72;}? 5‘“ rv.:,’!‘.:.»

“a #flb

Illllllllllllllllll ‘

This is to certify that the

thesis entitled

AN EVALUATION OF HIGH ENERGY ION
IMPLANTATION FOR NEAR STUDIES

presented by
Todd L. Rachel

has been accepted towards fulfillment
of the requirements for

M. 5. degree in Mechanical
Engineering

Major professor

Date W

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

I I—'_'_—_"_. 7, 7 _ _.

—....r

r *‘W
LIBRARY
Michigan State

I University

———-_

 

 

*—

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES mm on or before date due.

DATE DUE DATE DUE DATE DUE l
T 'r
{46% i 7‘32” L n ,
APR 2 2 2003 D

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU In An Affirmative Action/Equal Opportunity lnetltution
emails-9.1

 

AN EVALUATION OF HIGH ENERGY ION IMPLANT ATION
FOR WEAR STUDIES

By

Todd L Rachel

A THESIS

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

MASTER OF SCIENCE

Department of Mechanical Engineering

1991

55%4/27

ABSTRACT

AN EVALUATION OF HIGH ENERGY ION IMPLANT ATION
FOR WEAR STUDIES

By

Todd L Rachel

The use of radiotracers in wear studies has shown to have a clear advantage over
other wear measuring techniques. One type of radiotracer measurement technique is
that of Ion Implantation in which an accelerated ion beam penetrates the surface of an
engine part to be worn and implants radioactive beam nuclei. As the part is subjected
to mechanical friction, the ions are collected in the engine oil and detected as the im-
planted material is wom away. This technique has many advantages, but one possi-
ble disadvantage is the change in nature of the target material, which has the possibil-
ity of changing the wear rate of the part. To determine if material properties are af—

fected, a 20Ne , 2.5 MeV / nucleon non-radioactive beam was used to implant varying

amounts of ions (2x1012, 2x10”, and 21:1014 ions/cmz) into ceramic and stainless
steel specimens. The specimens were then tested for a change in hardness and elas-
tic modulus due the implantation of ions using a Nano Indentation Machine. The ce-
ramic material, silicon nitride, and the stainless steel 17-4PH both showed a de-
crease in hardness and elastic modulus due to ion implantation at the ion dosage of

2x1014 ions/cmz. The 2x1013 and 2x1012 ions/cm2 dosages showed no change in
hardness or elastic modulus. Finally, an ion implantation detection system required
to detect gamma radiation for engine wear studies was designed.

ACKNOWLEDGMENTS

I would like to give thanks to several important people.

To Dr. Schock, my major professor, who through his knowledge and guidance made
this project successful.

To my good friend Tom Stuecken for the machining of associated hardware and who
gave me a life time of knowledge and invaluable assistance.

To Dr. Ronningen for the tedious instruction on the use of the gamma ray detector, pa-
tience and late nights especially during the cyclotron run.

To Richard Shalek for the stainless steel preparation, fabrication of the gold foil de-
graders, helpful knowledge and companionship at Oak Ridge National Laboratory.

To Dr. Grummon for his instruction and gold foil fabrication.
To Dr. McHarris for his helpful comments.

To my advisory committee Dr. Schock, Dr. Grummon and Dr. Sommerton for their
guidance.

To my wife for being so understanding and supportive throughout my masters pro-
gram.

And to my family, who with their scholastic persistence in my younger years, their
constant support and substantial financial contributions, made my education possible.

iii

TABLE OF CONTENTS

Page

LIST OF TABLES ....................................................................................................... vi

LIST OF FIGURES ..................................................................................................... vii
CHAPTER 1 - INTRODUCTION ............................................................................... 1
1.1 Problem Statement ....................................................................... 1
1.2 Literature Survey ......................................................................... 3
CHAPTER 2 - FACILITIES AND SYSTEM COMPONENTS ................................. 5
2.1 Wear Specimens ........................................................................... 5
2.2 Buehler Polisher/ Grinder Machine ............................................ 8
2.3 Cyclotron ..................................................................................... 9

2.4 Nuclear Radioisotopes ................................................................. 11

2.5 Ion Implanting Depth Model (TRIM) .......................................... 15

2.6 Gamma Ray Analysis System ...................................................... 17

2.6.1 Gamma ray Detection System ............................................ 18

2.6.1.1 Gamma ray Detector and Accessories ................... 18

2.6.1.2 Preamplifier ............................................................ 19

2.6.1.3 Crysotat/Dewar System ......................................... 19

2.6.2 Pulse Modification System ................................................. 20

2.6.2.1 Amplifier ................................................................ 20

2.6.2.2 Analog to Digital Converter (ADC) ...................... 22

2.6.2.3 Digital Stabilizer .................................................... 24

2.6.3 Liquid Nitrogen System ...................................................... 25

2.6.4 Multichannel Analyzing System ......................................... 27

2.6.4.1 System 100 ............................................................. 27

2.6.4.2 MicroSampo ........................................................... 28

2.7 N ano— Indentation Machine ......................................................... 29

iv

Page

CHAPTER 3 - PREPARATION AND PROCEDURES ............................................. 37

3.1 Specimen Preparation ................................................................. 37

3.2 TRIM .......................................................................................... 41

3.3 Cyclotron Ion Implantation ......................................................... 45

3.4 Nano Indentation ......................................................................... 52

CHAPTER 4 - RESULTS AND DISCUSSION ......................................................... 61

4.1 Detection of Specimen’s Radioactivity ....................................... 61

4.2 Nano Indentation .......................................................................... 64

4.3 TRIM ........................................................................................... 91

CHAPTER 5 - SUMMARY AND CONCLUSIONS .................................................. 95

CHAPTER 6 - RECOMMENDATIONS .................................................................... 97

LIST OF REFERENCES ............................................................................................. 98
APPENDIX - DESIGN OF ION IMPLANTATION WEAR SYSTEM ..................... 100

LIST OF TABLES

Page
Table 1 Silicon Nitride Components ...................................................................... 5
Table 2 Silicon Nitride Properties ........................................................................... 6
Table 3 l7-4PH Stainless Steel Components .......................................................... 7
Table 4 l7-4PH Stainless Steel Properties ............................................................... 8
Table 5 Surface Preparation Procedure for Silicon Nitride ..................................... 39
Table 6 Surface Preparation Procedure for Stainless Steel ...................................... 40
Table 7 ’Doubled Specimens’ Test Matrix .............................................................. 48
Table 8 Actual Number of Ions Implanted in Specimens ........................................ 50
Table 9 Cyclotron Beam Current Variation ............................................................. 50
Table 10 Beam Degrader Foil Thickness ................................................................... 51
Table 11 ’Single Specimen’ Test Matrix ................................................................... 51
Table 12 Activity of Silicon Nitride no. 6 ................................................................. 61
Table 13 Spectrum Energy Calibration ...................................................................... 63
Table 14 Varying Hardness Values ........................................................................... 67
Table 15 Exhaust Activity Verses Blower Flow Rate ............................................... 105

LIST OF FIGURES

Page

Figure 1 Target Pod Set Up ........................................................................................ 10
Figure 2 Stable Nuclei ................................................................................................ 12
Figure 3 Nucleon Binding Energy .............................................................................. 14
Figure 4 Gamma Ray Detection System ..................................................................... 17
Figure 5 Schematic of Nano Indenter Machine .......................................................... 31
Figure 6 Displacement Sensing System ...................................................................... 33
Figure 7 Typical Load Displacement Curve ............................................................... 34
Figure 8 Berkovich Indenter ....................................................................................... 35
Figure 9 Dimensions of the Specimens ...................................................................... 37
Figure 10 Ion Implantation Depth from a Non Smooth Surface ................................ 38
Figure 11 Ideal Ion Implantation Profile .................................................................... 42
Figure 12 TRIM’s Calculated Degraded Ion Implanted

Concentration Depth .................................................................................. 43
Figure 13 Dose Depth Relation for a Non Degraded Beam

into Silicon Nitride ..................................................................................... 44
Figure 14 Dose Depth Relation for a Non Degraded Beam

into Stainless Steel ..................................................................................... 44
Figure 15 Cyclotron Beam ’Wings’ on Scintillator Plate ........................................... 47
Figure 16 Nano Indenter Specimen Holder ................................................................ 53
Figure 17 Master Menu of Nano Indenter Computer ................................................. 54
Figure 18 Runs A, B and c on Silicon Nitride no 6 (1014 ions) ................................ 56
Figure 19 Run D on Silicon Nitride no. 7 (1013 ions) ............................................... 56
Figure 20 Run E on Silicon Nitride no. 10 (control specimen) .................................. 57
Figure 21 Runs F, G and H on stainless steel no.4 (1014 ions) .................................. 58
Figure 22 Run Ion Silicon Nitride no.7 (1013 ions) .................................................. 58
Figure 23 Run K, L, and M on stainless steel no.2 (1013 ions) .................................. 59
Figure 24 Run P on stainless steel no.1 (1012 ions) .................................................. 60
Figure 25 Run Q, R, and s on Silicon Nitride no.8 (1012 ions) ................................. 6O

vii

Page

Figure 26 Silicon Nitride no. 6 Detected Spectrum ................................................... 62
Figure 27 Averaged Control Hardness of Silicon Nitride .......................................... 64
Figure 28 Averaged Control Elastic Modulus of Silicon Nitride ............................... 65
Figure 29 Hardness of Silicon Nitride no.6 (Run B) .................................................. 68
Figure 30 Elastic Modulus of Silicon Nitride no.6 (Run B) ....................................... 69
Figure 31 Hardness of Silicon Nitride no.6 (Run C) .................................................. 71
Figure 32 Elastic Modulus of Silicon Nitride no.6 (Run C) ....................................... 72
Figure 33 Hardness of Silicon Nitride no.7 (Run D and I) ......................................... 73
Figure 34 Elastic Modulus of Silicon Nitride no.7, Run D and I ............................... 75
Figure 35 Hardness of Silicon Nitride no.8, Runs Q and R ........................................ 77
Figure 36 Elastic Modulus of Silicon Nitride no.8 (Runs Q and R) ........................... 78
Figure 37 Hardness for Stainless Steel control Runs H and M .................................. 80
Figure 38 Elastic Modulus for Stainless Steel Control Runs H and M ..................... 81
Figure 39 Hardness of Stainless Steel no.4, Run F and G ......................................... 82
Figure 40 Elastic Modulus for Stainless Steel no.4 Run F and G ............................. 84
Figure 41 Hardness of Run K and L of Stainless Steel no.2 ....................................... 86
Figure 42 Elastic Modulus of Run K and L of Stainless Steel no.2 ........................... 87
Figure 43 Hardness of Run P on Stainless Steel no.1 ................................................. 89
Figure 44 Elastic Modulus of Run P on Stainless Steel no.1 ..................................... 90
Figure 45 Beam Incident on Magnified Specimen Contact Surfaces ......................... 92
Figure 46 Varying Ion Implantation Depth with Jagged Surface ............................... 93
Figure 47 Indentation of Deeper Implanted Regions .................................................. 93
Figure 48 Oil Detection System .................................................................................. 103
Figure 49 Exhaust Dilution System ............................................................................ 106

viii

CHAPTER 1
INTRODUCTION
1.1 Problem Statement

Inherent to any internal combustion engine is the matter of mechanical wearing of
components subjected to high contact areas such as bearings and piston rings. Many
techniques and much research has gone into the study of trying to reduce the wear of
the engine components. Different materials affect the wear rates and patterns of a
given mechanical system. Materials other than metals, like ceramics, are being stud-
ied to possibly help reduce the wear problem. Another area that is being focused up-
on is lubrication. In creating a lubricant which reduces the surface friction, the wear
rate could be lowered. And of course, changes in the overall design to reduce the fric-
tional surface area and still maintain dependability of the part would reduce the
amount of wearing. Several factors external to the engine also have an effect on the
wear rate. Owners maintenance, the driving techniques and frequency of use can
change the rate of wear of engine parts.

There are several techniques available to measure the wear of a part in an en-
gine. One technique to measure wear is to disassemble the engine and physically
measure the dimensions of the part of interest. This tear down technique changes the
subsequent wear pattern of the engine since exact reassembly is impossible. Anoth-
er method is ’Part Dimensional Change’ where a sensor is connected to the part
which senses a change in dimension of the part. The ’Part Weight Change’ method,
is a technique used in which the part must be weighed on a sensitive scale and then
put back into the engine. All of the above mentioned techniques depend on access to
the engine and accuracy of measurement. The above also have the possibility of poor
repeatability and reproducibility of data.

Another method of measuring mechanical wear is the Radioactive Tracer Method.
This technique uses a in situ method of measuring the wear of a engine part while the
engine is running. This method is very repeatable, able to make reproducible tests,
and quick once the system is set up. The technique is able to evaluate transient
wearing behavior, and can measure specific areas of a component. Wear measure-
ments as small as one micron can be detected.

One type of radiotracer measurement technique is that of Surface Layer Activation
(SLA) in which a heavy ion beam hits the individual engine part to be worn and caus-

es the nucleus of the atoms which make up the material to become unstable and thus
release energy. One form of released energy is in the form of gamma rays. The now
unstable atom is a radioactive ion. The detectable radioactive ions are worn away
with the part’s material and are detected in the engine’s oil in a quantitative manner.
The amount of material worn from the part is related to the amount of detected radio-
activity.

SLA activates the surface of the part to be studied, so it depends on the nuclear
properties of the elements that make up the part. Thus SLA is limited to those ele-
ments that can be transmuted into radioactive species with reasonably long half lives
and easily identifiable radiation, usually elements having the atomic number greater
than or equal to 24. The material limitation is currently quite extensive, e.g. chromium
and iron, where p, d or He ion bombardment can produce an isotope with a detectable
gamma ray. Hence material such as plastics, rubber and ceramics which are part of
the wear scene and Z < 24 can not be used in SLA technology. Furthermore this
method has the capability of changing the wear patterns of the material since the ma-
terial has undergone a change itself.

Ion Implantation is another radiotracer technique which uses the detection of radio—
active ions. But instead of bombarding the target with a high energy, heavy ion cyclo-
tron beam to change the atomic nature of the material, the beam merely implants ions
into the surface, thus retaining the original atomic structure. This method can be
used on any material, including ceramics and plastics, not just materials with the
atomic number above 24.

Even though damage done by Ion Implantation is assumed to be considerably less
than SLA (proposed by R.Ronningen [1]), the question still arises of does this meth-
od of testing change the material properties in any way? Our goal is to determine if
the implantation of ions into the experimental specimens causes a material change.
The implanting of radiotracer ions into a material is a physical one which takes a lot of
energy and may create heat from the collisions. The amount of radiation damage in-
curred is primarily dependent on the ion energy, the ion dose, the target temperature,
and the mass and chemical bonding of the target atoms. This overall method has the
potential of causing damage to the implanted part. If sufficient damage occurs, the tri-
bological properties of the target may change thus effecting the wear rate. If the im-
plantation did effect the wear results then the experimental specimens would be con-
sidered to have a different rnicrostructure than the original specimen. By implanting
experimental specimens with varying amounts of ions and then comparing the speci-
men’s wear rate and hardness against a control (non-implanted) specimen, we can

determine if there exists a possible threshold dose of ions implanted where the tribo-
logical properties are unaffected.

1.2 Literature Survey

Much research has been done on the effects of irradiation on ceramic materials.
Most of this research has been performed by the materials community, while the auto-
mobile engineer merely uses the technique with out knowing the potential problems
associated with the method. Most of the following work was taken from the En gi-
neered Materials Journals, where the research was plentiful.

Bhattacharya et al. [2] used optical microscopy, surface prof'rlometry, Auger elec-
tron spectroscopy and a transmission electron microscopy to study the high energy
ion irradiation effects on the surfaces of silicon carbide and hot pressed silicon nitride.

Optical microscopy and surface profilometry revealed that 1 MeV Ni+ ion irradiation

016

at a dose of 1x1 cm’2 produced swelling and roughness on both SiC and Si3N4.

Auger electron spectroscopy in combination with sputtering revealed that the SiC sur-
face became carbon rich, whereas the Si3N4 surface became Si rich due to ion irradia-

tion. Cross section TEM analysis of the surface layer of irradiated SiC showed that

1pm of material at the surface became amorphous due to irradiation with 1 MeV Ni+

016 2

at a dose of l x 1 cm' . Amorphization has also been observed in Si3N4.

Townsend [3] looked at the optical effects of ion implantation. Townsend stated
that by the selection of ion energy and ion dose, one can inject trace impurities that
control luminescence, generate absorption bands or complex defect aggregates, stimu-
late production of new crystalline phases or destroy crystallinity. Such property
changes are valuable in diagnostic studies and indeed the high rates of energy deposi-
tion along an ion track allow production of defect clusters or excited states which are
otherwise unattainable. Implantation invariably stimulated luminescence which pro-
vided a sensitive means of analysis to measure purity and near surface defect concen-
trations. Post implantation measurements revealed changes in many physical and
chemical properties of the materials.

Burggraaf et al. [4] studied ion implantation and monolayer techniques for modify-
ing ceramic surfaces. The authors suggested that thin surface layers with a thickness

up to lum can be prepared within a ceramic matrix or on top of it. Concentration pro-
files of the dopant atoms were produced with different shapes while the structure was

crystalline or amorphous. Thin layers or complex layer structures were produced or
thermodynamically stable solid solutions as well, from metastable compounds or com-
posites. A variety of properties were changed, e.g., optical, electrical, mechanical,
catalytic, and corrosive. Mono layer techniques were found to be especially interest-
ing for modifying internal surfaces of porous materials and for changing chemical reac-
tivity.

Cochran et al. [5] investigated the mechanical and structural properties of ion im-
planted yttria stabilized zirconia ceramics. The authors implanted yttria stabilized

zirconia with 100er Ca+, Al"’ and 02 + ions at fluences of 1015 to 6 x 1016

ions/cmz. Knoop microhardness with a shallow indenter penetration depth peaked at
a dose of 1016 ions/cm2 for both Al+ and 02 + but Ca+produced no effect on micro-

hardness. Vicker’s microhardness with a much greater indenter penetration depth did
not show detectable change by implantation but fracture toughness measurements

from the same Vicker’s indentations exhibited 1o23% increases at the highest 02 +

doses and 20-25% increases at high Al+ doses. Reflection electron diffraction

showed that the surface had not been made amorphous by the 6 x 1016 Al+ dose as
a well crystallized diffraction pattern was obtained.

Oliver et al. [6] studied the hardness of ion implanted ceramics. The authors state
that it has been established that the wear behavior of ceramic materials can be modi-
fied through ion implantation. Studies have been done to characterize the effect of im-
plantation on the structure and composition of ceramic surfaces. To determine the ef-
fects of an ion implanted A1203, a commercially available ultra low load hardness

tester was used to characterize A1203 with different implanted species and doses.

The hardness of the base material is compared with the highly damaged crystalline
state and the amorphous material.

Even though much research has been done on the implantation damage of materi-
als, the critical dosage required to produce a hardness change has not been accurately
determined for automotive frictional materials. The current research finds the thresh-
old ion dosage which causes material hardness elastic modulus changes. These prop—
erty changes may cause the materail to wear differently than the orginal material.

CHAPTER 2

FACILITIES AND SYSTEM COMPONENTS

2.1 Wear Specimens

Ceramic and a metal materials were used as test specimens which are applicable
to wear applications. The ceramic was a Silicon Nitride (Si3 N 4) NC-l32 supplied by

Norton Corporation. This ceramic is often used in advanced engine wear research due
to its hardness and stability under high temperatures. This ceramic is known to un-
dergo amorphization during irradiation which could be detectable in changing material
hardness or wear patterns [7]. The Si3 Ni 4 components shown in Table 1.

Table l Si3 Ni 4 Components

Element Max. wt % Typical wt %
Mg 1.0 0.4-0.6

A1 0.5 0.18-0.3
Fe 0.75 0.16-0.35
Ca 0.05 0.006-0.03
Mn - 0.05

B - 0.003

W 3.0 1.5-2.0

Silicon and Nitride components make up the remaining weight percentage. NC-
132 is a high strength, hot pressed ceramic that is considered to be the standard for

5

advanced ceramic materials. NC-132 retains its strength and oxidation resistance
above useful temperature ranges of nickel-cobalt super alloys. N C-132 is character-
ized by high density, excellent thermal shock resistance, chemical inertness and high
hardness. The Silicon Nitride is > 98% dense. NC-132 has the material properties
shown in Table 2.

Table 2 Silicon Nitride Material Properties

English SI
Bulk Density 199/2001bs/it3 3.2/3.3 g/cc
Hardness, KHNIOO 2000
Strength 3 pt. 20 c 140/160 x 103 psi 990/1100 MPa
Fracture Toughness 4.9/6.0 x 103 Lb-in'3/2 5.4/6.6 MNnt'3/2
Creep Rate %/HR 0.1 /0.007
@ 1350 c, 20 ksi
Young’s Modulus, E 45 x 106 psi 310 GPa
Thermal Expansion 1.9 x 10'6 /F 3.5 x 10'6 / C
Thermal Conductivity 220 BTU-in/hr-ft2 -F 32 W/m K
Critical QuenchT 1110/1300F 600/700C

Max. Use Temperature 2460 F 1350 C

The metallic specimen selected was Stainless Steel 17—4PH which originally con-
tained mostly austenite. Grummon and Shalek [8] suggested that the metallic alloy
be processed (heat treated and quenched) to produce a metastable structure (of 98%
martensite) which could be a sensitive gage of radiation induced structural change in-
duced by low dose irradiation. Since the phase was metastable, it could revert back
to the austinite phase if sufficient kinetic energy was applied to the Specimen. The
typical properties of the 17-4PH are shown in Table 3.

Table 3 17-4PH Stainless Steel Components

Element Wt. %
C 0.014
MN 0.61
SI 0.37
P 0.025
S 0.027
CR 15.15
NI 4.41
CU 3.46
CB+TA 0.19
N 0.019
MO 0.18

The element iron makes up the remaining weight percentage. The solution was
treated by Shalek [8] at 1900 F (1040 C), held at temperature for a time commensu-
rate with section thickness, and rapidly cooled as required to below 90 F (32 C).

Table 4 17 -4PH Stainless Steel Properties

Yield Strength, (.20 PCT.) KSI 192.0

Tensile Strength, KSI 199.0
Elongation in 1.4 ins.,PCT 14.0
Reduction in area, PCT 48.0
Hardness, BHN 401
Hardness, HB 321

A review of literature indicates a threshold dose of 1 x 1016 ion/cm2 for observ-
able material property changes in irradiated metals.

2.2 Buehler Polisher I Grinder Machine

The Buehler Polisher/Grinder machine is used for planar grinding and polishing
of test specimens. The system is electronically controlled from two front panels. The
machine consists of a base containing an abrasive platen and a head which holds test
specimens on a shaft. Three abrasive platens are used, two ceramic platens of vary-
ing hardness and a nylon platen for polishing. The shaft on to which a five specimen
holding wheel is attached, rotates and is lowered onto the base platen. Both the head
shaft and the platen rotate, in counter rotation to each other for best results suggest-
ed by Buehler. The rubbing of the specimen surfaces on the platen, mixed with abra-
sive solutions, causes the specimen material to be worn away. The system is con-
trolled from a upper and lower control panel, in which specimen holder force on platen,
rotation direction, duration, water lubrication and platen speed can be chosen. The
platens are Metlap #6, Metlap #10 and a nylon platen. The higher the Metlap num-
ber, the harder the plate material. The abrasive solutions are a diamond suspension
liquid whose varying diameter diamond particles are suspended in either a oil or wa-
ter base for lubrication. For the final polishing stages, a MasterMet solution which
contains no diamonds is used. The diamond size varies from a 45m to 0.25pm dia-
mond diameter. The larger diamond suspension is for removing specimen material
while the smaller diamond suspension is for polishing. Diamond suspension is used
for ceramic specimen preparation since it is harder than the ceramic specimen material

which we are using.

The specimen holder is a five piece holding wheel which holds 1.25 inch diame-
ter specimens contained in mounting material. Normally the individual specimen
mounts are made using an epoxy mix from Buehler called EPO-KWICK epoxy. The
mixed hardener is poured on top of the specimen in a plastic cup and allowed to dry.
While drying, the Specimens are placed in a vacuum cell to remove any air trapped in
the viscous epoxy.

Once hardened and at room temperature, the specimens are placed in the spec-
imen holding wheel by use of a specimen mounting rig. This rig holds the specimen
holding wheel firmly, approximately 0.125 inches from its base plate. When the speci-
men mounts are placed in the specimen holding wheel, the mounts fall to the base
plate and are secured, thus resulting in five specimen mounts that are all equidistant
from the holding wheel.

Once all the specimens are mounted into the holding wheel, the wheel is then
placed on the polishing head where it ’floats’, meaning that it can wobble thus con-
forming to any non flat platen surface. The appropriate settings are chosen for platen
speed, force of the specimens on the platen, duration of run, etc. Once the machine is
running, it must be manned and diamond suspension sprayed onto the surface approx-
imately every 15 seconds. The machine will run until the time duration runs out or the
session is cancelled.

2.3 Cyclotron

In order to implant ions into a target material, the National Super Conducting
Cyclotron Laboratory (NSCL) at Michigan State University is used. The cyclotron im-
plants ions into the specimen by use of a ion beam which is generated at its core and
is transported down a beam line. The beam line is made of nuclei of atoms of virtually
any single element. The nuclei are stripped of some of their electrons before entering
the beam line. The collision of the beam with a target creates useful particles which
are then implanted into the specimen. A group at Michigan State University which in-
cludes researchers from the College of Engineering and the NSCL are experimenting
with possible useful beams concurrent with Ion Implantation Research [9].

In order to make use of the cyclotron accelerated beam, the target specimens
must be placed at the end of the beam line and organized as needed. The components
of interest lie at the end of the beam line in a vacuum sealed stainless steel pot. The
experimental pot has a bolted flange for beam line connection, a windowed flange for

10

television monitoring, a top flange for access to the components and a bottom flange
which held several movable shafts. All the flanges were lip type which use copper
gaskets to seal off the vacuum. Pressure solenoids that moved the shafts were con-
nected to the external portion of the bottom flange. Objects that were to be moved in
and out of the beam were connected to the movable shafts, see Figure l. The pres-
sure solenoids were remotely operated from the cyclotron control room outside the ra-
diation vault. The pot contained four items. The item that was that in the beam line
was a stationary 0.125 x 0.625 inch sloted scintillator plate. The side facing the beam
was painted with a material that scintillates when exposed to the cyclotron beam a1-
lowing the beam shape to be viewed. The slot only allowed a controlled portion of the
larger beam to pass through.

 

 

/

Passage
Slot

. . F . . .
Scrntrllator 839W Scrntlllator and Specrmens
Plate “1’ inHolding Rig
\ Scintillator
Beam Line
9 __§
’4 \ Set Screws
\\
\

 

 

 

 

 

 

 

 

 

 

 

Stationary t
Movable Shafts

Figure 1 Target Pod Set Up

The second item in line is a Farady cup used to measure the beam current.
This piece could be moved in and out of the beam line. The Faraday cup was connect-
ed to wires leading out of the pot and connected to a count meter in the control room.
The third item was a movable piece that contained foil degraders of varying thick-

11

ness. The degraders are used for degrading the beam intensity before hitting the
specimens. The last item was a rig that held six specimen mounts and one scintilla-
tor piece in order to view beam location.

2.4 Nuclear Radioisotopes

The nucleus of an atom consist of N neutrons (no charge) and Z protons (positive
charge), both of which are considered nucleons. The mass number (A) of an element
is the number of nucleons. The atomic number of an element is the number of protons
(Z). An isotope is a nuclei with the same number of protons and a different number of
neutrons than the definition of the element. The stability of an isotope depends on the
number of protons and neutrons in the nucleus.

12

 

1151) I

 

 

 

 

 

Iii!)

 

'1 lilrrtrlc'lrlrcici
"‘I,I- '1 ll" llrlrittrrrl

 

 

 

o It" curritlltrr
0 l‘. r or electron capture emitters
I”T a «emitters ‘ifll l'. r 0' EC.

 

 

 

allI‘GIIIETUIIS 1t~rtalrlc 1,! II!!!‘

 

 

lllll

 

 

 

ml

 

 

§

 

 

 

 

J
I

 

 

 

L
a
1
rec

 

lierrtrrrrt Irthnrlrer (It!)
8

 

 

 

 

 

 

 

 

. ..§‘
.:r:~
‘°' 593‘
.°:.
0 :g:
,5(: :zi::
gig;

“K: :ii

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o‘roroaoaosocovoaoaoloouo
l’rrrttrrr lllllllllllr (£2!) '

Figure 2 Stable Nuclei

13

Figure 2 shows the plot of the neutron number N verses the atomic number Z
for all known nuclei. Each small square represents an isotope and the solid black
squares represent stable isotopes. Two smooth curves, one through the stable iso—
topes and the other for N=Z, have been drawn. It is evident that for low values of N
and A, the stable isotopes have N=Z. For heavier elements, the stability curve grad-
ually departs from the N=Z line, reaching a slope of 1.6 for A=238. The isotopes on
both sides of the curve are radioactive and eventually decay in such a way as to form
the final stable isotopes.

The forces between the nuclei must be attractive, or there would be no stable
nuclei because of the repulsive Coulomb force between the protons. Also, these at-
tractive forces must be much stronger than the Coulomb repulsive forces, or there
would be no stable heavy nuclei, because the Coulomb repulsion between the large
number of protons would disrupt the heavy nuclei. The binding energy of the nucleus
is decreased by the Coulomb repulsion of the protons in the nucleus. Because Cou-
lomb forces are long range forces, each proton effects every other proton, not just its
immediate neighbors. The repulsive force, therefore, goes on increasing with increas-
ing 2 and hence, with increasing A. This Coulomb repulsive force has two conse-
quences, (1) The average binding energy per nucleon drops as A increases as seen
in Figure 3 or (2) the locus of the stable nuclei should depart from the N=Z towards
the direction of a higher number of neutrons, shown in Figure 2.

14

Binding energy (Mevl

 

onwunrmtnmmmmmwm
A-r'

Figure 3 Binding Energy per Nucleon.

When an unstable nucleus decays, it emits an alpha particle, a beta particle, or
a gamma ray. In the process, the original particle decays into another particle, and
the element is called radioactive. Alpha particles have very weak penetrating power
and can be stopped by a sheet of paper, but they can cause intense ionization in the
gas through which they travel. Beta particles have less ionizing power but more pen-
etrating power than alpha particles. Beta particles can easily pass through thin foils
of metals a few millimeters thick. Gamma rays cause even less ionization but can
penetrate through slabs of different materials many centimeters thick, including any
engine block. Gamma rays carry no charge, thus they are not deflected by electric or
magnetic field and they exhibit all the characteristics of electromagnetic waves. Gam-
ma rays are used for the wear study in question.

When a nucleus emits a particle, it has essentially decayed, thus the radioac—
tivity decreases. Experimental evidence shows that radioactive decay follows an ex-

At

ponential law. An expression used for the exponential decay is N = No e' where

N is the number of radioactive nuclei present at time ’t’. N0 is the original number of

radioactive nuclei present. The probability of decay per unit time = A. = 1n 2 / (half life
of isotope). If an element has a half life of one day, then the element will contain one
half of the number of radioactive nuclei in one day then it did to begin with. It is im-

15

possible to predict when a particular unstable nucleus will decay.
The activity of a radioactive specimen is defined as the number of disintegra-
tions per second (dps) that occur in the specimen. A unit of activity used often is the

curie which is equal to 3.7 x 1010

dps. Radioactive decay is a law of nature and con-
forms strictly to it. A given isotope always has the same half life and set of associat-
ed gamma ray energies, independent of how, when or where it was made. Each iso-
tope gives off one or several gamma rays with specific energies. These gamma ray

energies for all known isotopes are documented in such articles such as the ’Chart of
N uclides’ [10]. As an example, the isotope 5A’Mn has only one gamma ray at the en-

ergy of 834.83 KeV. It also shows that 5A’Mn has a half life of 312.7 days and a
branch ratio of 99.98.

2.5 Ion Implanting Depth Model (TRIM)

In order to estimate the penetration depth of energetic 20Ne ions into solid
target specimens, an ion implantation depth model was used, known as TRIM
(Transport of Ions in Matter) [11]. TRIM is primarily concerned with the quantitative
evaluation of how the ions lose energy to the solid and the final distribution of these
ions after they stop with in the solid. Also considered are the first order effects on
the atoms of the solid, particularly the electronic excitation of the atoms, the displace-
ment of lattice atoms by energetic collisions and the production of plasmons and pho-
nons in the solids by the passing ions. The ions may have energies from 10 eV to 2
GeV/amu. The program will accept complex targets made up of compound materials
with up to three layers made up of different materials. It will calculate both the final
distribution of the ions and also all kinetic phenomena associated with the ion’s ener-
gy loss: target damage, ionization, and photon production. All target atom cascades
in the target are followed in detail, and the redistribution of these target atoms is de-
terrrrined. TRIM only calculates the beam implantation for a monoenergetic beam.
Computer simulations of the polyenergetic interacting beam are achieved by running
several different monoenergetic beams of varying energies and nature and combining
them into one file. Computer simulations of the polyenergetic beam are necessary to
understand the overall interaction process. The TRIM simulation consists of follow-
ing a large number of individual ion or particle histories in a target. Each history be-
gins with a given energy, position, and direction. The particle is assumed to change
direction as a result of binary nuclear collisions and move in straight free-fight paths

16

between collisions. The energy is reduced as a result of nuclear and electronic energy
losses, and a history is terminated either when the energy drops below a pre-speci-
fied value or when the particle’s position is outside the target. The target is consid-
ered amorphous with atoms at random locations, and thus the directional properties of
the crystal lattice are ignored. The nuclear and electronic energy losses or st0pping
powers are assumed to be independent. Therefore, particles lose energy in discrete
amounts in nuclear collision and lose energy continuously from electronic interac-
tions.

TRIM helps to understand the physical results of ion implantation and to aid
in the preparation of the properties test. The deposit of ions depends on the size and
energy of the ions and the material of the target. Some combinations give sharp con-
centration peaks at a specific depth, where as others are more apt to be largely dis-
tributed. In any case a direct beam into a specimen will not give a uniform concentra-
tion of ions through some depth.

17

2.6 Gamma Ray Analysis System

The Gamma Ray Analysis System is used for detecting radioactive ions that
emit gamma rays. The gamma my analysis system contains four components: a gam-
ma ray detection system, a pulse modification system, a liquid nitrogen system and a
multichannel analyzing system. The purpose of the detection system is to detect
gamma rays of various energies.

 

 

 

ADC Digital
I I
. O O
W->— N
Gamma Ray >——l .
Detector f: ............... : . ‘

 

 

 

 

 

 
    
   

''''''''''''''''''''''''

 

- Electronic Valve

i
I

 

 

 

 

 

 

 

 

 

 

I -
2500 V _.x
Power Supply Auto Fill Multi Channel
Controller Analyzer
50 Liter Liquid
Nitrogen Tank

Figure 4 Gamma Ray Detection System

18
2.6.1 Gamma Ray Detection System

The gamma detector contains four components; a semi-conductor gamma ray de-
tector, a preamplifier, a cryostat/dewar system which maintains the detector element
in a vacuum at cryostat temperatures and a liquid nitrogen dewar. The purpose of this
system is to detect gamma rays and to produce a voltage peak proportional to the
gamma ray’s energy.

2.6.1.1 Gamma ray Detector and Accessories

The gamma ray detector is a Hi gh-Purity Germanium coaxial photon detector,
manufactured by EG&G Ortec, model number Gem 90220-P-Plus. The detector con-
tains a single crystal of germanium 75.8mm in diameter and 92.9mm in length and is
located in the center of the cylinder 3mm from the end cap of the detector. The crystal
is located in the detector chamber which contains a vacuum and a 2500 volt electric
field across it. This chamber is also kept at a temperature of -198.5 C to reduce
’noise’ of radio frequency signals to the sensitive detection element. The detector is
maintained at these low temperatures by having liquid nitrogen flow around it. The
detection chamber needs approximately 24 hours to cool fiom ambient temperature in
order to operate correctly. The GEM series detector uses a p-type material with a
lithium diffused outer contact in the detection chamber. The closed end coaxial detec-
tor configuration blends the closed outer contact end into the outside cylinder surface,
resulting in small absorptions cross sections. The low capacitance, closed end detec-
tors have lower noise and better resolution. The purpose of the detector is to produce
a voltage pulse proportional to the sensed gamma ray energy. Electron hole pairs are
produced by the absorption of a gamma ray into the single crystal of germanium locat-
ed in the detector chamber. The electron hole pairs are then swept across a 2500 volt
electric field to oppositely charged contacts in the detection chamber. The electrons
moving though the electric field produces an induced current. This current is then sent
to the preamplifier.

A radioactive specimen holder was fabricated to hold one inch diameter thin
calibrated specimen wafers. The holder is made of clear plastic and has 0.5 inch dis-
tance increments positions from the detector. The plastic holder slips over the end of
the detector, open box end facing up ward. The solid calibrated specimen is placed in
a 2.5 x 3.2 inch cardboard rectangle with a one inch hole in the middle for the Speci-
men. When detecting a source, it is important that the source is in the same position

19

for every count test.

Solid Angle is a topic that is of concern when counting a specimen. Radioac-
tive specimens give off gamma rays in every direction. If the detector were able to
surround the entire specimen, as a sphere, then it could intercept every gamma ray
being emitted. But the detector only sees a fraction of the gamma rays emitted and
the geometric name for the cone of interception is called the solid angle. This angle is
a fraction of the entire solid angle which is a sphere.

2.6.1.2 Preamplifier

The preamplifier is a consealed sheath approximately the same diameter as
the detector’s outer sheath. The absorption of the gamma ray by the detector produc-
es a current pulse at the preamplifier input. Each current pulse is integrated by the
charge sensitive loop which is essentially a FET input operational amplifier with ca-
pacitive feedback. A feedback resistor R in parallel with the feedback capacitor C re-
moves the charge stored by the input capacitance. This results in the exponential de-
cay with time constant 1: = R*C, of each charge loop pulse. The pole zero cancella-
tion network differentiates the charge loop output to give a shorter decay time
constant and is adjusted to eliminate pulse undershoot following differentiation. The
preamplifier requires power of 24 volts which is supplied by the amplifier in the Pulse
Modification System discussed later. The detector cryostat contains a temperature
sensing element attached to the cooling path of the detector. The sensing element
connects to a hybrid monitoring circuit which is incorporated into the preamplifier. An
output cable from the preamplifier is connected to the remote shutdown input on the
rear panel of the bias supply in the power box. The bias supply will reduce the detec-
tor’s bias voltage to zero if the temperature resistance is greater than approximately
135 ohms. The monitoring circuit in the preamp provides this condition if the detector
temperature becomes too high. Although no alarm is provided, the bias supply meter
will indicate zero voltage and system noise will greatly increase after shutdown oc-
curs. The automatic shutdown circuin should be placed in operation before applying
any bias to the detector. This is a critical safety device to prevent damage to the ex-
pensive detector.

2.6.l.3 Crysotat/Dewar System

The cryostat/dewar system maintains the detector element in a high vacuum

20

at close to liquid nitrogen temperatures. The three liter dewar which is connected to
the detector serves as a reservoir of liquid nitrogen while the cryostat provides a path
for heat transfer from the detector element to the liquid nitrogen reservoir. Both the
dewar and the cryostat rely upon a vacuum to insulate cold inner parts from the outer
surfaces. The cryostat provides an outer envelope which can be maintained internally
at high vacuum. The cryostat also contains a mount for the detector element and elec-
trical components. Vacuum feed through in the cryostat allow electrical connections
to be made from the detector to the preamp and high voltage filter. The cryostat con-
tains a cryogenic vacuum pump that consists of one or more reservoirs of molecular
sieve - a pelletized material with an extremely large surface to volume ratio which ab-
sorbs large volumes of gas when cooled to liquid nitrogen temperatures. This method
of cryosorption pumping provides a good cryostat vacuum for continuous operation
over a period of approximately five years. The three liter reservoir must be filled ev-
ery 24 hours due to vaporization of the liquid nitrogen. The dewar is filled from a con-
nection in the back of the container. This dewar is part of the liquid nitrogen system
which has the sole purpose of providing liquid nitrogen to the cryostat to keep the de-
tector element cool.

2.6.2. Pulse Modification System

The Pulse Modification System’s purpose is to take a rather crude input volt-
age pulse of given magnitude and convert it to a digital number which is proportional
to the magnitude of the voltage pulse. To achieve this, the system requires three
components: an amplifier, an analog to digital converter and a digital stabilizer.

2.6.2.1 Amplifier

The purpose of the amplifier is to create a well defined pulse from a poorly de-
fined pulse from the preamplifier. The model 2020 amplifier’s objective is to shape
pulses, received from the detector preamplifier, into a near Gaussian shaped pulse so
that there is consistency of shape entering the analog to digital controller.

Found on the front panel, the course and fine gain controls adjust the peak am-
plitude. By changing the amplitude, the relative energy associated with the incoming
gamma ray will change, thus changing the energy at which the gamma ray appears on
the resulting software screen. A super fine gain 22 turn adjustment screw is used for
resolution better than 0.0063%. The polarity switch sets the model 2020 for the polar-

21

ity of the incoming preamplifier signal. The input polarity switch is always set to posi-
tive for this setup.

In the shaping section of the panel, the P/Z 22 turn adjustment screw optimiz-
es the pole / zero setting which optimizes amplifier baseline recovery and overload
performance for the preamplifier fall time constant. A correct pole / zero setting
makes for a smooth pulse restoration to the baseline. The multiplier button multiplies
the shaping time setting by l or 2. The shaping time knob selects the shaping time
constant. This constant is somewhat proportional to the amount of time the amplifier
spends on shaping the incoming pulse. The longer the shaping constant, the better re-
solved the peak but the slower the system, and vice-versa holds true.

The restorer rate in the restorer section is a two position toggle switch used
to set the baseline restorer rate. In AUTO mode the baseline restorer rate is auto-
matically optimized by internal circuitry as a function of unipolar output signal duty cy-
cle and count rate. In HIGH mode the baseline restorer is fixed into a high rate. The
high restorer rate can be used where a higher than normal restoration rate is re-
quired. The HIGH restorer rate is normally not used since there is a loss of resolu-
tion due to the increases correlated noise inherent in do restoration. The restorer
mode is a two position toggle switch that selects SYMmetrical or ASYMmetrical
baseline restorer modes. In symmetrical mode, the restoration currents are identical
for above and below the baseline. In asymmetrical mode, the restorer current above
the baseline is much less than that below the baseline. The asymmetrical restorer
mode offers superior high count rate performance for high resolution Ge spectroscopy
and is used on this system. The symmetrical mode is used on Ge systems with low
quality preamplifier, scintillation and proportional counting systems. The threshold
auto/var switch is a two position toggle switch which is used to set the baseline re-
storer threshold. In auto mode, the baseline restorer threshold is automatically opti-
nrized by internal circuitry as a function of the unipolar output signal noise level. In
variable mode, it provides a manually variable baseline threshold adjustment range of
0 to 200mV dc.

The pile up reject (PUR) switch is a two position switch that enables or dis-
ables the PUR / LTC (live time correction). Nuclear spectroscopy with high count
rates can be made more accurate by testing for pulse pileup and rejecting contaminat-
ed pulses. Pulse pileup produces distortions in the higher ranges of the spectrum be-
cause sucessive shaped amplifier pulses tend to merge, and cause aberrations in the
precise amplitude of any single pulse as measured by the analog to digital controller.
Since many spectra have higher count rates at lower energies, the pileup of these

22

pulses strongly affects the less intense higher energy counts and can cause signifi-
cant errors. A pileup reject circuit provides substantial improvement to the spectrum.
The PUR DISC is for adjusting the PUR discriminator threshold level. The LED is a
visual aid for setting the DISC threshold at an optimum level.

The BNC connectors at the bottom of the panel are for input and output con-
nectors. The input (on left) accepts positive or negative preamplifier pulses and
should be connected to the detector’s preamplifier. The unipolar output is used in this
setup and connects to the ADC. The unipolar output provides a positive, linear ac—
tively filtered near-Gaussian shaped pulses, amplitude linear to +10 volts.

On the back of the amplifier, there are three input/output connectors which are
connected in parallel to the front panel connectors, meaning they are used for the
same thing. The current setup uses the front panel. Also on the back there are DT
(dead time) and PUR connections which should be connected to the corresponding
connections on the ADC. There is a power outlet of 24v which should be connected to
the power inlet of the detector’s preamplifier.

2.6.2.2 Analog to Digital Converter (ADC)

The analog to digital converter translates an incoming voltage to a digital num-
ber. The incoming voltage is proportional to the energy of the detected ion from the
gamma detector. The ADC model 8077 incorporates a Wilkinson type, 450 MHz
clock rate , counter/ramp converter optimized for high count rates, high resolution and
low differential non-linearity performance.

The GAIN knob is a six position rotary switch used to select full scale resolu-
tion of the input signal. A selection of 512, 1024, 2048, 4096, 8196, or 16384 can be di-
vided into the incoming 10v signal. The GAIN controls the ABC’s resolution. A di-
rect coupled input of 0 to +10v is digitized according to the front panel GAIN control
into a binary code which corresponds to 512 through 16384 channels. Thus the incom-
ing voltage is divided into a choose number of channels set by GAIN. The larger the
number of channels selected, the higher the resolution but the longer it takes to pro-
cess the data, thus increasing the dead time. As an example, with GAIN set to 8192
and no offset, the maximum conversion time is 18.7ttsec. With the GAIN set to 512
and no offset, the maximum conversion time is 1.84ttsec. The RANGE rotary switch
is set equal to the size of the memory group assignment used in the multichannel ana-

lyzer’s software. For instance, a memory assignment of 4096 channels in the soft-
ware would require a range of 4096 channels on the ADC.

23

A LED is displayed on the front panel which shows the amount of dead time
the system is encountering. Dead time occurs when there is an overload of incoming
data. The ADC cannot keep up and the data is backlogged, thus resulting in a live
counting time less than real time. The live counting time is stretched until all the data
that was inputted could be processed. For example a system that has been counting
for 60 seconds (real time) and has a high dead time may indicate a live count time of
only 10 seconds. The PUR and LTC account for this overload and it is they who slow
the live time down so that all the data can be processed. The LED displays the
ADC’S dead time in increments of 10%.

The front panel six offset toggle switches provide the suppression of digital ze-
ro. The offset is used to shift the given toggle switch channel number to zero in the
multichannel analyzer’s memory. With no offset (all switches down), the ADC’s
channel numbers are the same as the software’s memory channel numbers. For ex-
ample, if the GAIN is set to 8192 and the software’s memory assignment is only
4096, an offset of zero will allow only the lower half of the full scale conversions to be
stored, that is pulses up to five volts out of the possible 10v. If in this example, the
offset were set to 4096 (the 4096 switch up), channel zero of the software memory
would be shifted to correspond to channel 4096 of the ADC. This offset would allow
the upper half of the full scale conversion, those above five volts, to be stored in the
assigned software’s memory.

The lower level discriminator (LLD) and the upper level discriminator (ULD)
controls set the limits for the input signals to be accepted by the ADC for conversion.
If an input falls within the selected window of channels (higher than the LLD but low-
er than the ULD setting) the input will be converted. If the input does not fall within
the window, the input will not be converted. The window check is made at the conclu-
sion of the linear gate time, an input pulse not within the window will add to the
ADC’s dead time, since it takes the time to check.

The zero control varies the zero intercept of the ADC conversion function so
that the zero energy is stored in channel zero of memory. This zero setting is adjust-
able with a small screw driver and a pulser. A guide to setting this zero is found in
the Canberra’s "Fast Analog to Digital Converter Model 8077 Operator’s Manual."

Peaks that are not well defined may be from noise or system interruptions.
The peak detect function allows for some control over the chosen peaks. Peak Detect
is a toggle switch used to select either automatic or delayed peak detection to start
the ADC conversion. In auto mode, the linear gate, which allows a valid input pulse
to be acquired, opens when the input rises above the input threshold. The input

24

threshold level tracks the LLD setting up to 0.1 volts max, therefore the input Sig-
nal’s baseline must be less than 0.1 volts for proper ADC operation. A high amplifier
gain setting will raise the entire baseline over 0.1 and cause the system to overload
thus resulting in a 100% dead time for the system.

The linear gate closes when the pulse falls below 90% of its peak amplitude.
Wide pulses may make the automatic detection of the 90% point less certain. To
overcome this uncertainty, the DELAYED peak feature may be used. This feature al-

lows input pulses up to 100 usec wide to be converted. In this mode the linear gate
closes at the end of the selected delay time. The delay time can be selected with the
ADJ control while monitoring the INSP test point next to the ADJ control. The IN-
SPect test point is provided to monitor the linear gate time.

The BN C connector on the front panel is used as the input from the amplifier.
This accepts a positive unipolar or bipolar pulses for pulse height analysis (PHA) or
specimen voltage analysis (SVA). In the current set up, the switch is set on PHA.
The SVA is only for system testing. On the back panel, the PUR and DT connections
should be connected to the corresponding connections on the back of the amplifier.
The STAB connection is connected to the Digital Stabilizer and the DATA is connect-
ed to the IBM computer.

2.6.2.3 Digital Stabilizer

The digital stabilizer stabilizes the ADC. When converting the analog signal
to a digital number the ADC might not convert a 2.5V peak to the same number as it
did 15 minutes earlier. The ADC wanders a little because it has no relative zero
point. This wandering ADC causes the peaks on the computer screen to wander also
since the peak’s screen position is determined by the ADC. This wandering makes
for repeatability very difficult The Digital Stabilizer is designed to take care of the
wandering ADC, it stabilizes the ADC by working in parallel with it. On the front
panel, references may be entered to keep the output stable. A low energy peak chan-
nel number is entered for the ’zero’ setting and a high energy peak channel number
entered for the ’gain’ setting. These two entered points make a reference for the
ADC resulting in a stable peak. The two point stabilization can be turned to on, hold
and off by toggle switches on the front panel of the digital stabilizer. When the
switches are turned off, stabilization is disabled and the correction voltage is held to
zero volts. When the switches are in the hold position, the correction voltage is not
updated, but the last correction voltage is maintained at the output.

25

The % CORRECTION indicator is a ten segment LED bar graph, which indi-
cates the percentage of full scale correction being applied in increments of twenty per—
cent. The five digit display shows one of the four parameters for the Zero or Gain por-
tion of stabilization. The four parameters are PEAK, WINDOW WIDTH, STAB ILI-
ZATION RATE and ANALOG RANGE.

The Peak display indicates the peak channel which the stabilizer is correcting
on. The Window width indicated the number of channels used by the stabilizer as its
sampling range, with the window is centered about the peak. The Stabilization rate
indicates the count rate of the input data which gives rise to a change in correction
voltage. This display is in counts per second (CPS). The Analog Range indicates the
value of the Analog Attenuation Factor. The attenuation is applied to the 5v range of
the correction voltage from the Digital Stabilizer to the ADC.

2.6.3 Liquid Nitrogen System

The purpose of the liquid nitrogen system is to keep the detector element at
close to liquid nitrogen temperatures. The liquid nitrogen (LN2 ) system has several

components to achieve this task. A 50 liter self pressurizin g liquid nitrogen tank is
used as a large reservoir of LN2 . The 50 liter tank is connected to a three liter dewar

which is fixed to the detector which was mentioned earlier. Every 24 hours the 50 li-
ter tank automatically fills the dewar by the use of an auto fill controller. The dewar
provides LN2 to the cryostat which is used as a path for heat transfer from the detec-

tor element to the dewar.

The 50 liter tank is manufactured by Cryofab and is insulated by a vacuum.
Once filled, the tank must relief the nitrogen gas build up due to vaporization. A relief
valve is located on the top of the tank and has been modified to be adjustable by sim-
ply screwing in the end piece for a higher tank pressure or screwing out to lower the
pressure. The tank pressure should be kept at 3 to 5 psi. The relief valve will make a
continuous hissing sound as it relieves the pressure build up. During normal opera-
tion the tank should be filled once every week at the cyclotron. The tank also has a
special self pressurizing feature which creates large pressures form its own vaporiza-
tion virtually instantaneously. This self pressurizing system is regulated by a black
knobbed valve at the top of the tank. When high pressure is not needed, the valve
should be closed, as is the case in the current set up. The tank also contains several
safety features which should not be tampered with.

26

The dewar can be filled manually via a modified funnel or by the 50 liter tank
which is the current set up. The dewar comes with a plug for the fill connection mak-
ing the detector portable. The cryostat is an internal system which cannot be seen
and needs no attention.

The auto fill controller is located in the power supply bin with the other elec-
tronic equipment. The purpose of the auto fill control is to fill the dewar every 24
hours from the tank of liquid nitrogen. The controller is a time based unit which
counts down from 24 hours. When the timer runs out, a solenoid valve opens on the
line between the dewar and the tank, permitting the tank to fill the dewar. A sensor
located on the dewar’s vapor vent line senses the presence of liquid nitrogen in the
vapor line and causes the solenoid to close, thus filling stops. This sensor, consisting
of a resistor in the flow of the vapor vent, is very sensitive. When the dewar is filling,
the liquid nitrogen replaces the nitrogen vapor which exits out the vent line. When
the dewar is full, liquid nitrogen will overflow the dewar and exit out the vent line,
causing the resistor to cool and reach a certain resistance which triggers the auto fill
controller to turn off the fill line solenoid. On the side of the controller, there is a re-
sistance threshold adjustment screw which changes at what temperature the control-
ler turns the filler solenoid off. If the sensor is not properly set, the controller will not
fill the dewar all the way up, thus not keeping the detector cool and the detector might
not power up (safety device). Also if the sensor is not set properly, the fill solenoid
will never turn off and liquid nitrogen will flow out the vent line until the 50 liter tank
is empty, so proper adjustment is required.

The time to next fill is displayed on the front panel of the Auto Fill Controller
and can be changed by a row of dip switches accessible through the side panel of the
controller. The system can be filled at any time by using the manual button on the
front panel. When the system is filling the green light will be lunrinated. The system
filling can be prevented by setting the NORM / INHIBIT toggle switch to inhibit.
Place the switch in NORM mode for normal operation. A maximum fill time can be
set by toggle switches on the side panel to prevent the system from overfilling.
When this preset max fill time is reached, an alarm will sound and the fill time exceed
light will lunrinate. The alarm can be reset by pressing the ’error’ and ’reset’ button
simultaneously and the pressing the alarm ’reset’ button. The solenoid toggle switch
is for abiling and disabling the electronic fill solenoid. When set to off, the solenoid
sill not open to let the passage of liquid nitrogen flow to the dewar. The power to the
will solenoid is from a connection on the back cover of the auto fill controller. The va-
por vent sensor is connected to the ’sense’ connection.

27
2.6.4 Multichannel Analyzing System

The purpose of the Multichannel Analyzing System (MCA) is to take digital da-
ta from the Pulse Modification System and to display it in graphical form on the com-
puter screen. Once the data is in graphical form, it can be analyzed in different ways.
The Canberra System 100 MCA is the system that is used to acquire the data, put it
in graphical form and can be used to analyze the data. Another package called Can-
berra MicroSampo uses the data fiom the System 100 package to further analyze the
data. Both of these packages are on the IBM model 70 computer.

2.6.4.1 System 100

The System 100 is a software and hardware driven package. This system
must be run in Microsoft Windows and requires 2MB of RAM, an DOS of 3.0 or later,
EGA or VGA display and a hard disk. System 100 is the final stage of the detection
system. The current pulse which originated from the detector was amplified, shaped
and then converted into a digital number is then bused from the ADC to the System
100 board located on the mother board of the IBM model 70. The digital number corre-
sponding to the analog voltage peak of varying height enters the software and is
placed at a corresponding memory channel. The channel range is 4, 8 or 16 K which is
chosen while entering the System 100 software package, and is also chosen on the
ADC front panel discussed earlier. The larger the incoming digital number, the higher
the corresponding channel number. Channels are displayed on the horizontal axis of
the computer screen and the number of gamma rays detected are on the vertical axis.
When a digital voltage peak enters the software, a count is added to its correspond-
ing channel. A detected gamma ray is called a ’count’, so 500 counts on channel 1500
refers to 500 gamma rays detected in memory location 1500. This displayed graph is
the only screen of the System 100 software package. The main features used are dis-
cussed here.

The FILE function allows for saving and retrieving graphs that were made.

The ACQUIRE function that enables the computer to acquire the incoming data and
display it on the screen. Duration of acquire can also be set under this function.

The DISPLAY function selects the portion of the memory channels to be displayed.

28

The LEARN / EXECUTE function saves and executes previous keyboard sequences
for more automation of the program. This function saves of tedious repetition of keys.

The ENERGY CALIBRATION creates a mathematical relationship between the
channel numbers and the actual gamma ray energies. Once an energy calibration is
created and is turned on, the horizontal axis units are gamma ray energies instead of
memory channels.

The ANALYSIS function analysis specific peaks that have been marked with a
’region of interest’ (ROI). A ROI is placed on a peak by coloring the peak red using
the shift and arrow keys.

2.6.4.2 MicroSampo - Advanced Gamma Spectrum Analysis Software

MicroSampo is a software driven system with a software key required for op-
eration. The key is connected to the back serial port of the computer. This software
package does not require Microsoft Windows nor does it require as much memory as
System 100, but the software is not user friendly nor forgiving. MicroSampo includes
automated energy, efficiency and shape calibration features. It does the peak search
using generalized second differences and incorporates both linear and nonlinear peak
fitting using precalibrated peak shapes thus providing very accurate peak energy and
intensity information. It also does the nuclide identification and specific activity calcu-
lation using the least-squares solutions for overlapping peaks and nuclides. Micro-
Sampo is currently used for the analysis of specific isotope specimens for the calibra-
tion of the entire detection system. Once the detection system is operating correctly,
a detection calibration is performed so that an efficiency graph of the system is creat-
ed for the set up. Using this calibration, unknown isotopes can be identified.

MicroSampo can not acquire data thus it must retrieve a saved file created by
the System 100 software. MicroSampo has several functions which all must be used
in order to get results. The functions are as follows.

FILE - allows for the filing of System 100 and MicroSampo files.

SPECTRUM DISPLAY - displays the retrieved spectrum.

29
PEAK SEARCH - finds and displays the peaks of set significance.
PEAK FITTING - fits a mathematical curve to the shape of the peak.
PEAK SHAPING - fits an calculated smooth area under the peak curve.

DETECTOR EFFICIENCY CALIB RATION - uses above data to determine overall
detector efficiency.

Although of limited use in this study, the gamma ray detector will become an
integral part of the piston ring wear measurement system as described in the appen-
dix.

2.7 Nano- Indentation Machine

The nano-indentation machine is used to measure the hardness of materials
[12]. This machine can detect very slight changes in hardness, for instance due to ra-
diation damage. This machine is located in the High Temperature Materials Labora-
tory (HTML) of the Oak Ridge National Laboratory in Oak Ridge Tennessee. This
lab is part of the Department of Energy (DOE) and is currently operated by Martin
Marietta Energy Systems, Inc. It is designed to help solve material problems that
now limit the efficiency and reliability of advanced energy conversion systems. Coop-
eration with the US. industry is the particular focus of the HTML. The Mechanical
Properties Microprobe or the Nano Indenter is a special microhardness tester capable
of operating at loads in the microgram range. The instrument consists of three major
components: 1) the indenter, 2) an optical microscope, and 3) a precision table that
transports the specimen between the microscope and the indenter. Unlike conven-
tional hardness testers, it is not necessary to determine optically the area of an in-
dent in order to calculate the hardness. Instead, the position of the indenter relative
to the surface of the specimen is constantly monitored, thus allowing the depth of an
indent to be determined. The area of the indent is then calculated from a knowledge of
the geometry of the tip of the diamond indenter. The fact that the position of the in-
denter can be determined to with in 0.2nm means that the indenter can be used to
sample very small volumes of material. Typical applications include the characteriza-
tion of thin films, near surface properties, and the different phases in a multiphase ma-
terial. The instrument is also useful in studies of composite materials and in the de-

30

termination of interfacial strengths.

The Nano Indenter is essentially a load controlled system. However, a wide
variety of indentation parameters may be prescribed, and through the use of feedback
techniques, displacement controlled experiments may also be performed. Under both
load control and displacement control conditions, load and displacement are measured
continuously so that hardness can be determined as a function of indent depth on the
basis of a single indent.

The indenter consists of a rod attached at its center to the middle of a special
three-plate capacitive sensor. Movement of the capacitor plate is used to produce

the displacement of the indenter. The total movement possible is 100 pm. The depth
resolution of the system is approximately 0.2 nm. At the bottom of the indenter rod
is a three sided diamond pyramid ground so that the sides of the pyramid make an an-
gle of 65.3 degrees with the normal to the base. Thus the indents appear as equilater-
al triangles, and the length of a side of an indent is approximately 7.4 times its depth.
The depth of an indent is related to the diameter of the circle that circumscribes the
triangle the indent makes on a surface by the expression :

h =0.ll3d

where h is the depth of the indent and d is the diameter of the circumscribed circle.

A coil is attached to the top of the indenter rod and is held in a magnetic field,
thus passage of a current through the coil can be used to apply a force to the indenter.
There are two load ranges, 0 to 120 mN (0 to 12 grams) and 0 to 20 mN (0 to 2 gm).

The load resolution for the more sensitive range is approximately within 0.2 uN. The
gross vertical position of the entire indenter head is controlled by a motorized mi-
crometer, referred to as the 2 motor. The Z motor may be operated from the console
or by a manual control box.

The position of an indent on a specimen is selected using the optical micro-
scope that is built into the system. A TV camera mounted atop the microscope per-
mits the image of the specimen to be viewed remotely. The microscope has three tur-
ret-mounted objective lenses that in combination with the eye piece and additional
magnification produced by the TV camera provide magnifications of 130, 270 and 2000
X on the TV screen. Both turret rotation and focussing are done remotely from the
computer console.

The system used to apply the load to the indenter consists of a magnet and
coil in the indentation head and a high precision current source, Shown in Figure 5.

31

Programmable

Currcnt Source

Capacitive —m .
Displacement I
Gage Volunctcr

 

I INDENTER SUPPORT ARM]

 

.L ’
- V ‘ I I

y Mororizcd Table Vi Coriu‘zilccrs

UMENT B

 

 

 

Figure 5 Schematic of N ano Indenter Machine [11]

The current source is applied across a coil in a magnetic field, thus generating
a force. The current from the source, after passing through the coil, passes through a
precision resistor across which the voltage is measured. It is this voltage that is dis-

32

played on the displacement voltmeter. The current source is programmable with a re-
sulting range of displayed voltage of -2.0 to +2.0 volts in the high load range and from
+0.8 to +1.7 in the low range. The voltage is controlled through a 16 bit DAC, there-

fore the maximum resolution is 10uV.

The displacement sensing system is the most important part of the Nano In-
denter system. It is shown in Figure 6 and consists of a special three-plate capaci-
tive sensor. All three plates are circular disks approximately 1.5mm thick. The two
outer plates have a diameter of 50 mm, and the inner moving plate is half that size.
All three plates have a hole in their centers large enough to accommodate the indent—
er column, which is actually attached to the moving plate. This plate and indenter as-
sembly is supported by two leaf springs cut in such a fashion as to have very low stiff-
ness. At the top of the indenter column is a load coil that is used to raise and lower

the plate and indenter assembly through its 100nm of travel between the outer plates
of the capacitor. This same coil applies the force required to make an indent.

The x-y tables have a real resolution of 0.1 um. The motor controller can be
controlled by the computer. The screws have a pitch of 40 threads per inch. They
have a real accuracy of 20 millionths of an inch per inch and are mounted on bearings
with a maximum of 15 millionths of an inch horizontal play. Although the backlash in
the stage is virtually nonexistent due to special zero backlash nuts on the screws, it
is recommended that each position be approached from the same X and Y directions.
The stepping motors have a 50,000 steps per revolution. The motors must be de-en-
ergized between moves to prevent excessive heating during the experiment, which
leads to thermal drift in the displacement measurement.

33

5 a ‘— FORCE COIL

 

 

LEAF SPRING

 

/.

7 LOAD VOLTAGE

 

 

+1.8“ +1.9

OUTER PLATE 1” .3 “Hi
I
I
I
I
l
I

 

8

~

 

l i i
L movmo PLATE 1 E
i i

 

 

 

 

ouren PLATE ]"

 

+2.5 ~25
DISPLACEMENT VOLTAGE

 

LEAF SPRING

 

 

 

 

y/4— Z-MOTOR MOTION

 

 

Figure 6 Displacement Sensing System [11]

In the microprobe’s most sensitive mode, it operates with loads in the micro-
gram range, although a maximum load of 10 grams is possible. Indent dimensions are
not measured directly but are calculated from measurements of the position of the in-
denter relative to the surface of the specimen and a knowledge of the geometry of the
point of the diamond indenter. Sophisticated capacitance techniques are used to mea-
sure the position of the indenter relative to the surface with an accuracy of 0.2 nm.
Thus very small volumes of material can be studied, and the strength of a material can
be characterized on an extremely localized scale. Motion of the specimen stage in the
x-y plane is also precisely controlled. The indenter can be positioned with in 1.0 pm
of any chosen point on the specimen, and a series of indents, separated by steps as
small as 0.1 pm, may be made in any geometrical pattern. The entire operation of the
system is computer controlled, and one or several of indents may be specified and car-
ried out with out further operator intervention. The closest indent to the edge possi-

ble with this machine is 4 pm.

34

Because the indenter penetration is usually only a few nanometers deep, nano
indentation test are sensitive to the near surface features on the specimen. They are
also very sensitive to microstructural discontinuities such as grain boundaries, twin
boundaries, porosity and precipitates which have dimensions comparable to that of
the indenter tip’s incursion. With this in mind, localized microstructural effects, which
might be averaged out below a larger indentation, can have a major effect on hardness
number. Because nano indentations specimen only a shallow portion of the specimen
surface, the response is greatly influenced by the quality of the specimen preparation.
Note that it is possible to obtain a high polish on a surface, yet still retain the re-
menants of previous course grinding steps.

Load displacement (L-D) hardness testing is what is used in the nano indent-
er. L-D tests have several features which allow more information to be obtained than
from residual impression measurement methods alone. By monitoring the displace-
ment continuously, a curve is produced such as that shown in Figure 7.

 
 
  
     
   
  

 

hold at max. load —
- » down-
5 J load
5 up-lood curve:
a contains elastic and .
'3 plastic response 54* _.- (up/dz)
" / 91009
Qt-‘Cfét 1 used for
’ modulus
_ ‘ coic.
f—I'BCDVBI'BG depth—at

DEPTH (nm)

Figure 7 Typical Load Displacement Curve [11]

35

The curve consists of an upload portion, a hold at peak load, and an unloading
portion. The displacement information from the tip load portion contains both elastic
and plastic components so that the instantaneous hardness values computed from the
up load curve without elastic corrections would tend to appear somewhat lower than
values obtained from elastically recovered residual impressions made to the same
depth. The unloading curve indicates the elastic behavior of the specimen, and its
slope can be used to correct the instantaneous hardness number from the up load por-
tion of the L-D curve to compensate for elastic effects.

Not only elastic corrections can be obtained from the unload portion of the L-D
test results, the modulus of elasticity of the surface can also be obtained. This calcu-
lation involves assumptions about the indenter geometry, and involves the depth of
the indenter at maximum load and at the elastically relaxed condition.

The nano indenter uses a newer, ultra precision L-D instrument known as the
Berkovich indenter. This indenter is pyramidal Shaped with an equilateral triangular
base, shown in Figure 8.

TOP

   

angle A = 65 degrees

angle B (edge to face) =
141 deg 53 min

z= 0.1554L

A I, SIDE

Figure 8 Berkovich Indenter.

36

The major advantage of this indenter for precision work is that it is much easi-
er to get a sharp point with only three sides meeting than with four sides. This meth-
od can give hardness numbers just as Vickers and Knoop indenters can. The geome-
try of the Berkovich indenter is as follows:

Surface area S = 0.637 L2
Projected base area A = 0.5774 L2
Hardness Number HB = 1570 ( P/ L2)

where L is the altitude of the triangular base of the pyramid.

The elastic modulus is calculated from dz/dP shown in Figure 8.
(dz/d?) = on 2,912: / 24.5)0-5 (1 /1~:r)

where zp is the plastic depth, and Er is the composite elastic modulus of specimen
and indenter.

l/Er = (1-v2 )/E + (1-v02 )/1~:o

E is Young’s modulus , v is poissons ratio and the subscript ’0’ refers to the indenter.

CHAPTER 3

PREPARATION AND PROCEDURES

3.1 Specimen Preparation

Both of the ceramic and stainless steel specimen were made to the dimen-
sions indicated.

 

0.625 in ——>l

1.25 in. radius

   

0.125 in

Figure 9 Dimensions of the Specimens

The ceramic specimens were manufactured to size and a surface finish of 16
microinches on the radiused portion by Norton Corporation. Fifty specimens were
purchased at $110.00 dollars a piece. The stainless steel was machined at Michigan
State University from a two inch diameter hardened steel rod. The rod was cut into
wafers over 1/8 inch thick, then each wafer was cut into four approximately 1/2 x 5/8
inch rectangles. The piece was then surface ground to finish with the exception of the
soon to be radiused side. A special tool was built to put the radius on the pieces.
The tool was cylindrical with four slots cut into the outside of the cylinder. This tool
is set in a lathe with a specimen tightened down in each of its four slots. The diame-

37

38

ter was 2.50 inches which when exposed to a grinding wheel put a 1.25 inch radius on
the steel.

Now all of the specimens had to be prepared in an appropriated manner for the
cyclotron run. Two types of implanted specimens were desired for the cyclotron run.
One experiment was to implant the entire 5/8 inch long curved surface with ions,
meaning that the curved side must be exposed to the beam. Only the ceramic speci-
mens were to be used for this experiment and the manufactured l6 microinches
curved surface finish was adequate for this experiment. These specimens will be sub-
jected to a wear test not covered at this time. A well polished surface was essential
for good implantation results. The cyclotron would penetrate a set distance for a giv-
en material and beam. If the surface was not polished, peaks and crevices would be
apparent and the ions would deposit at a set distance into the material from the non
smooth surface, thus giving an implantation depth matching the surface profile. Note
also that the angle of incidence affects the implantation profile.

Beam

Unsmooth Surface

M

Ion Implantation
Depth Profile

Figure 10 Ion Implantation Depth into a Non Smooth Surface.

The other experiment was the implantation of ions into the large 1/2 x 5/8 sur-
face of the specimen. The specimens would be subjected to a nano-indenter to test
the hardness of the material where the beam hit. The cyclotron beam implants ions in

39

a peak at a set depth into the specimen, so if the specimens were bombarded, the
change in properties (if any) would occur not on the surface, but several microns into
the surface, resulting in being invisible to the surface testing nano—indenting ma-
chine. So it was decided to place two highly polished specimens together with the
large faces together and then sending the beam into the intersection of the two. This
’doubling up’ would cause the damage to occur on the inside surfaces of each speci-
men. The radiused portion of the specimens would not be used in this experiment.
This experiment would be run for both the silicon nitride and the stainless steel.

As stated before, the ceramic specimens which were to be implanted along the
full radiused portion, needed no further polishing, unlike the double type specimens.
The specimen mating surfaces should be as smooth as possible so that the double
piece seemed as a continuous material to the beam. In order to achieve this, the
specimens were first set with the large surface facing down in a Buehler fast curing
epoxy and then placed in the five specimen holding device for the planing and polish-
ing stage. Even though the specimens were manufactured to a tolerance less than
0.0005 inch, the specimens were not all perfectly planar so grinding must be done to
achieve planar specimens. After the abrasive planing stage was done, the specimens
were then polished. The surface preparation procedure for the Silicon Nitride is
shown in Table 5.

Table 5 Surface Preparation Procedure for Silicon Nitride

Surface Abrasive Time Force Speed
diamond suspension
Planar Grinding
Stage Metlap 10 45 pm 5 min 51bf/specimen 240 rpm
Specimen Integrity
Stage Nylon Texmet 6 um 7 min 5 lbf/specimen 50 rpm
Final Polishing

Stage Nylon Texmet l um 10 min 101bf/specimen 240 rpm

40

This procedure left a mirror like finish on the flat silicon nitride Specimens.
Since the Metlap 10 platen was approximately the same hardness as the specimens,
after about one hour of grinding and polishing, the platen became ’dished’. The platen
was then be reflatened by using the Buehler Platen Dressing Kit. This kit contained a
flat platen which connects to the Buehler head and several different sand paper grits
which removes all the high spots and eventually flattened the once dished Metlap
platen.

The procedure for grinding / polishing the stainless Steel is similar, but the
time and forces are reduced from the ceramic procedure since it is must softer than the
ceramic.

Table 6 Surface Preparation Procedure for Stainless Steel

Stage Surface Diamond Suspension Time Force Speed

Planar Grinding Metlap 10 30 um 3 min 5 lbf/specimen 240rpm
Specimen Integrity Nylon Texmet 6 um 3 rrrin 5 lbf/specimen 240 rpm
Polishing Nylon Texmet 1 pm 3 min 51bf/specimen 240 rpm

Nylon Texmet MasterMet 3 min 5 lbf/specimen 240 rpm

Once the large sides of the ceramic and stainless steels were polished, the
specimens were taken out of the epoxy mounts. The flat ends opposite the curve
needed to be polished. All the specimens were mounted in another specimen mount.
These mounts were to be used to hold the Specimens while being polished and also to
hold the specimens in the cyclotron holding pot while the specimens were being im-
planted with ions. The new mounts were made of copper and were 1.25 inches in di-
ameter, one inch tall with one inch flats on both sides. The one inch flats were for po-
sitioning the specimens in the cyclotron holder. A slit was milled in the center so that
the specimens would slide into the slit and be held by two set screws. The slot was
milled 0.03 inches shy of the 1/2 inch deep specimens so that the specimens would

41

protrude from the soft copper thus keeping the copper from being touched when the
Specimens were polished. There were two types of copper holders that were made,
one which held one specimen and another type which held two specimens. The single
Specimen holder exposed a single specimen radius which was to be hit by a multiply
degraded cyclotron beam for the wear test rig. The double specimen piece held two
highly polished surfaces together with the non-radiused end exposed. The cyclotron
beam was to penetrate a spot between the specimens and deposit ions on the inner
surface of the specimens. After implantation, the specimens would be separated and
the inner ion implanted face tested for a change in hardness.

Each copper specimen holder had two allen screws to hold the specimens in
place. The allen screws were machined flat so that a good flat surface contact was
made between the set screw and the specimens. A set screw with a burr could dam-
age the specimens or even break them. The specimen holders were made of copper
which has the best conductivity of any pure metal. Any heat generated by the beam
hitting the specimen would be absorbed into the copper away from the specimens.
The specimens with the radiused edge exposed were mounted in the single spaced
copper specimen holders and set aside since they were ready for implantation. The
highly polished sides were placed together and placed in the double spaced copper
holders with the flat ends exposed outward. These ends of these specimens were
then ground flat and then polished following the procedures listed above. The speci-
mens were polished so that the crack between the two specimens could not be seen
with the human eye, so that the two pieces seemed as one. All the Specimens were
now ready for implantation.

3.2 TRIM

A uniform dose depth profile was desired for the single Specimens that were to be im-
planted on the radius, shown in Figure 11.

42

Ion
Concentration
in Specimen

 

Distance into Specimen

Figure 11 Ideal Ion Implantation Profile

In order to achieve a uniform profile, degraders must be used which degrade
the beam energy before hitting the specimen. In degrading the beam, the ions will be
deposited at a shallower depth. The thicker the degrader, the shallower the implant.
If the degraders, of varying thicknesses are placed in succession in the beam line,
then the ion concentration profile will be a number of concentration peaks which are a
step in the direction of a uniform concentration profile. Since we are looking for a
change in hardness of the material at the implanted locations, we would like the ions
to be implanted at a uniform profile up to a set depth. This would show a consistent
wear rate through the implanted region, whether it was different from the non-im-
planted portion or not.

Since the beam deposits ions at a set depth ill a peak form, a combination of
varying thickness degraders were needed to give a somewhat uniform profile. Inex-
pensive, quick tests by TRIM enables us to determine the correct foil material and foil
thickness for our use. Gold and Nickel foils were chosen due to their dense material
and beam degrading ability. Gold seemed to be easiest to sputter into thin foils, so
gold was chosen. The preparation of the gold foils was done by Grumman and Shalek
[7]. The ion implantation depth desired was 4.5 um which was deep enough to pro-

43

vide some wear time but again shallow enough so that the hard Slow wearing ceramic
would not take too long. Due to size constraints and complexability, only six foils
were choose to be used, meaning that six varying thickness gold foils were to make
peaks for 0 to 4.5m deep. The TRIM operator, Y. Chung [12], set the beam energy
variable to 2.5 Mev / nulceon and TRIM gave a profile Similar to the one shown in Fig-
ure 12. The peaks corresponded to 9.0, 8.5, 8.0, 7.5, 7.0, 6.5 pm from left to right. The
deepest peak corresponding to the thinnest foil.

 

0.00045

0.0004 - D {

0.00035 -

I

0.0003 A
0
0.00025 '-

0.0002 '- i

RELATIVE ION DOSAGE

0.00015 '-

0.0001 - o

{L I 1
Q

 

 

 

0.00005

 

 

I
‘
w—
n
K—'

 

 

 

0 ‘ 1.
0 2

PENETRATDN DEPTH (MICRONS)

Figure 12 TRIM’s Calculated Degraded Ion Implanted Concentration Depth.

For the nano indenter run, we needed to know the distance into the specimen
where the maximum concentration of ions would occur. For a non degraded neon
beam of 2.5 Mev / nulceon into Silicon Nitride, TRIM gave the profile Shown in Figure

13. The peak ion concentration was found at 17.8 um into the silicon nitride.

 

0.00045

0.0004

0.00035

0.0003

0.00025

0.0002

RELATIVE ION DOSAGE

0.00015

0.0001

0.00005

 

 

 

 

14 to ._ 15 A i 20
emanation DEPTH (urcaous)

Figure 13 Dose Depth Relation for a Non Degraded Beam into Silicon Nitride.

 

0.00035

0.0003

0.00025

0.0002

0.00015

RELATIVE ION DOSAGE

0.0001

 

0.00005

 

 

 

. ... w 12
Pcurmmouomutmcrtous)

Figure 14 Dose Depth Relation for a Non Degraded Beam into Stainless Steel.

45

Likewise the same procedure was run for the stainless steel. TRIM gave the

profile shown in Figure 15. The peak ion concentration was found at 10.5 um into the
stainless Steel.

3.3 Cyclotron Ion Implantation

Ideally it was desired to achieve an ion implantation concentration of non-ra-
dioactive ions which would be similar to an implanted radioactive ion implantation.

An implantation using an equivalent of 2x1013 ions/cm2 of 22Na would yield the fol-
lowing radioactivity:

22Na : T1/2 = 950.4 days , branch ratio = 99.95

Probability of decay = A = ln2/T1/2 = 0.693 / 950.4 d = 0.0073 / d = 8.45 x 10'9 / sec.

Activity = 2. * (number of ions present)
= 8.45 x 109/ sec. * 1013 ions

= 8.45 x 104 disintegrations/ sec
= 2.28 uCi

Which is an acceptable activity value for radioactive tracer techniques. The 20Ne

beam was used for implantation due to its similarity to 22Na in atomic number and its
penetration into target materail profile.

The cyclotron irradiation procedure went as follows on 12/14/90. First the sili-
con nitride and stainless steel specimens were placed into the specimen holder rig
contained in the stainless steel pot. Starting from the top, the order of the specimens
were the scintillator, and then stainless steel number 1, 2, 4 and silicon nitride num-
ber 8, 7, 6. The foils put in the other foil holding rig were in the order of number of 18,
19, 20, 21, 22, and 23 which corresponded to the gold foil thickness of 9, 8.5, 8, 7.5, 7,

and 6.5 pm respectively as discussed earlier.

Television cameras were set up to monitor the scintillator plate and the scintil-
lator element on the specimen holding rig. The corresponding monitors were set up in
the cyclotron control room. The area was cleared, and the cyclotron vault was se-

46

cured in the proper manner. The cyclotron beam of 20Ne and 2.5 MeV was then actio
vated and sent to our test pot. The beam was tuned and shaped by the cyclotron op-
erator by viewing the beam on the scintillating plate. The Faraday cup was placed in
the beam line. The cyclotron operator condensed the beam to a size as small as pos-
sible and increased the current to as high as the 2.5 MeV energy limit allowed. The
scintillator plate had 0.5 inch markers drawn from the center of the plate, and the
beam was positioned in the middle of the target using these markers.

The beam current was read from the beam hitting the Faraday cup and creating
a current which was sent to a count integrator which summed up the number of counts
that it receives. The beam current could be calculated in the following manner. The
count integrator was set to 20 namps for 60 secs at 1000 counts I sec which was
60000 counts total range for 20 namps for the integrator. Using this data in a propor-
tion is used to find the real time current of the beam hitting the Faraday cup.

(count/min)/(60000cnt/min)=XneA/20 neA

Solving for the X gives the beam current on the Faraday cup. Using this calcu-
lated beam current and a beam time corresponding to the number of desired ions to be
implanted could be found.

A base reference was used for the time calculation of the 20 Ne 2+ beam. A
beam current of 30 enA has the energy flux of :

30 enA( 1.6 x 10'19 coul /e ) = 30 x 10'9 coal / see. which corresponds to
(30 x 10‘9 coul / sec )/ [(1.6 x 10'19 coul / charge)*( 2 charges / particle )]
= 9.38 x 1010 part/ sec implanted by the 30 neA beam.

Using the previous calculated beam current, and a method of proportions the
real particle implant rate could be calculated.

(beam current neA) / (30 neA) = (2 part /sec) / (9.38 x 1010 part/ see)

where Z is the real particle implantation rate. The desired number of ions implanted
divided by the calculated particle implantation rate gives the time required for the de-
sired number of ions to be implanted.

47

We had problems with collecting data from the Faraday cup so the scintillator
plate was used instead. The beam was moved so that it was positioned entirely on
the scintillator plate and beam counts were taken and a beam current was calculated
from this. Then the beam was moved to the aperture where only the top and bottom
fraction of the beam passing through the aperture, known as ’wings’, was measured
and a current was found.

Scintillator

Plate \

Aperture \

 

 

Aperture
er Wings

 

Figure 15 Cyclotron Beam ’Wings’ on Scintillator Plate.

The first count number reading from the count integrator from the total beam on
the scintillator plate was 9867 counts in one nrinute and the wings gave 759
counts/minute. The difference in these two numbers was considered to be the counts
passing through the aperture and thus reaching the specimens. Using the given pro-
portionality

48

(9108cnts/min)/(60000cnts/min)=XneA/(20neA)
and solving, a beam current of 2.9 neA was obtained. With this known beam current
and using the implantation rate proportionality

(2.9 neA)/(30neA)=Xpart/secl(9.38x 101°part/sec)

an implantation rate of 9.085 x 109 particles / second was calculated. If 1012 parti-

cles were to be implanted, a beam time of ( 1012 part ) / ( 9.085 x 109 part / sec ) =
110 seconds would be needed.

Using this equation with the known beam time gives a test time matrix for the
’doubled up specimens’ that were to be used for the nano-indentation experiment.
This experiment was run without the use of beam degraders.

Table 7 ’Doubled Specimens’ Test Matrix
SPECIMEN POSITION NUMBER OF IONS TIME

SCINTILLATOR 1 - -

SS # 1 2 1012 110 sec
SS # 2 3 1013 1110
SS # 4 4 1014 11000
SN # 8 5 1012 110
SN # 7 6 1013 1110
SN # 6 7 1014 11000

The Faraday cup was dropped out of the beam line and the beam traveled to
the specimens. The specimens were not implanted in the above order number but
from the shortest implant time to the longest.

The beam current fluctuated a lot but the implanting times were not changed.
The beam current was monitored and an implanted ion number was established from

49

that. Using the measured counts before and after a run, the average of the counts for
each run was used to determine the number of ions implanted.
The size of the beam was estimated by the cameras which gave the dimen-

sions to be approximately 1.00 x 0.5 inch which gives an total area of 0.5 inch2 = 3.23
cm2. This beam size is larger than the aperture size of 0.125 x 0.625 inch = 0.078

inch2 = 0.5 cm2. This is due to the refraction of the beam around the edges of the
scintillator plate. Since the beam intensity was measured at the scintillator plate of a
given dimension, the scintillator aperture total area will be used for the area implant-
ed on the specimen, since that is where the highest intensity of the beam would hit.

SS # 1
Pre-Aperture total - wing cnt = 9426 - 759 = 8667an min
Post-Aperture total - wing out = 9867 - 1150 = 8717 cnts/ min.
Average of intensity through aperture = 8692 cnts/min

using previous ratio 8692 / 60000 = x / 20 neA
x = beam current = 2.897 neA

then 2.897 neA / 30 neA = x /9.38 x 109 part/sec

x = particles second = 9.059 x 109 particles/second.

Total particles in 110 second run = 9.059 x 109 particles/second * 110 seconds

= 9.965 x 1011 particles

012 2

which gives 1.99 x 1 particles / cm .

Using this calculation for the rest of this experiments run, Table 8 was constructed.

50

Table 8 Actual Number of Ions Implanted in Specimen

SPECIMEN AVG. INTENSITY TIME PARTICLES PART/AREA
SS # 1 8692cnts/rnin 110 sec 9.97x1011 1.99x1012

SS # 2 6444 1100 6.72x1012 1.34x1012
SS # 8 14968 110 1.56x10 12 3.12x1012
SN # 7 17891 1100 1.87x10l3 3.73x1013
SS # 4 14943 10000 1.56x1014 3.11x1014
SN # 6 11031 10000 1.15x10l4 2.31x1014

Looking at the fluctuating intensities, Table 9 shows the percent change of the
beam intensity (counts / minute) which is related to the beam current.

Table 9 Cyclotron Beam Current Fluctuation

INTENSITY % VARIANCE FROM SMALLEST BEAM CURRENT
8692 cnts/min 34.9 2.89 neA
6444 0 2.15
14968 132 4.99
17891 178 5.96
14943 132 4.98

1 1031 71 3.67

Next the single ceramic specimens were implanted. This run required the use
of beam degraders to create a crude uniform concentration profile along the full length
of the radius portion of the specimen. At this time the intensity was 22716 clicks /

51

rrrinute which from a previous equation gives 7.57 neA. This current corresponds to

2.33x1010 particles/second. A time of 180 seconds was chosen for radiation time for
each of the six degraders, which gives a total implanting time of 1080 see which gives

013

2.36x1013 particles per specimens which is 5.11x1 ions/cmz. The Six gold foil de-

graders thicknesses are shown in Table 10.

Table 10 Beam Degrader Foil Thickness

Foil Degrader Thickness

A1 9.0um
A2 8.5um
A3 8.0um
A4 7.5um
A5 7.0um
A6 6.51m

Table 11 Test Matrix for the Single Specimens with Degraders

SPECIMEN POSITION NUMBER OF IONS TIME

SCINTILLATOR l - -

SN # 16 2 1013 1080 sec
SN # 15 3 1013 1080 sec
SN # 14 4 1013 1080 see
SN # 13 5 1013 1080 sec
SN # 12 6 1013 1080 see
SN # 11 7 1013 1080 see

So for each of the specimens, all six gold foils were placed in front of the speci-
mens one at a time for 180 seconds a piece.

52
3.4 Nano Indentation

The specimens were all kept in their copper mounts during transportation in or-
der to keep them in order. Our procedure was to first run a set of hardness tests on
the specimens with the greatest number of implanted ions in order to discover a
change in hardness. If the highest dosed specimens contained no damage, then the
lower dosed specimens were sure to contain no damage either. The specimens were
only removed from the mounts when needed. The first specimen used was Silicon N i-

tride no. 6 implanted with 1014 ions. Only one of the double mounted specimens were
used. The specimen was mounted on a metal stud by using a hot plate to heat the
stud and then melting a glue stick on the hot stud. The specimen was then put on the
stud and left to cool.

For mounting the mount holder was turned face down and propped up at its
ends originally with pieces of cardboard to permit the specimens to fall below the
specimen holder’s surface. The cardboard was quickly replaced with thin metal piec-
es which held the unit more Steady since mounting flatness could make the difference
between good or no data from the run. The Studs were then positioned so that the im-
planted edge was facing the front of the holder and that the edge was perpendicular to
the holding rig, see Figure 16.

53

 

Figure 16 Nano Indenter Specimen Holder

The Studs were then secured in the holder by a set screw which cocked the
studs when the torque of tightening made contact with the stud, which tilted the Spec-
imens non flat position. To get around this dilemma, the stud was first cocked counter
rotation and the set screw was not tightened very hard, but enough to hold the maxi-
mum 12 grams of indenter force. This routine of trail and error continued until a visibly
flat specimen surface was obtained. Once secure and flat, the specimens were
Sprayed with ’Dust Off’ and mounted in the Nano Indenter. The viewing lens was
then placed at the middle of the edge of the specimen with the computer controlled
system. The insulated cabinet surrounding the indenter was closed and the room
where the indenter sat was closed off.

The room temperature was monitored and was not allowed to fluctuate more
than 0.5 degrees centigrade. Since the sensitive machine accounts for thermal drift, a
large temperature drift would be considered too large for the machine to correct for.

The menu driven computer program was set up to run. The main computer
screen is shown in Figure 17.

54

«a. M"""7'"""" ‘~~'~W mvsvv-vmmev-vv‘h‘ a, h, t ._»«, . .3, ‘

I ygi. 'CINJB mist; Mil. .:\‘ 381mg...» .mm‘fln .x" " . “1.1" "“ “ ‘5‘
5N“

 

Master Menu lor System Control oi the Nanoindenter

 

>Take new Data
Manipulate Data
Manual Control oi the System
Re-calbrale

fry/.1
. ‘2}.
J.-
a".

91'
i
153
‘7.
T 34
. '- r
x a?
' a-
“ ‘ ‘1’.
eke}
i'éi
.. 3
Ni
1 (I '
“(‘34

You have selected choice 1

‘71., 4‘ 1‘ ,g \. ( A‘A.’ j, ~ l’r '
“4M0&&\\\\\'A . grout-i - rtuntt‘z a.

f." .9";
.- * ' 1‘ - b < .1 .
.fir’i‘nbfikw’bfixfi w

I f
xx
s ,

Useri Caps Command
Next Previous A ' ' All done

 

I r [as
\ th‘t‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 17 Master Menu of N ano Indenter Computer.

The indentation procedure had to be set up. The closest allowable distance to

the specimen’s edge was 4 urn. The specimen surface had to be flat since the indent-
er works on the basis to sensing the Specimen surface location. If the Specimen sur-
face was not flat, the diamond tip might touch and scratch the surface while moving to
the next indentation spot. On the other hand if the surface slopes down, the indenter
will move down past where it believes the surface Should be, thus aborting the in-
dent. When either of the two conditions arise, the Nano Indenter is said to ’Bomb
Out’, thus not picking up any data for that point. Since the surface can never be per-
fectly flat, a head repositioning step is performed at indent locations where the ma-
chine should reevaluate the surface plane. Head repositioning steps are automatical-
ly done at the beginning of each array of indents taken.

The operator called each indentation run, "material" which is confusing, so this

55

notation was changed from ’material’ to ’run’. The procedure that the indenter was
programmed to perform was as follows:

1- contact surface

2- load indenter to displacement of 50 nm deep
3- unload indenter by 15% of full rate to 80% retracted out of indent
4- load to 100 nm

5- unload at 15% rate to 80% retract

6- load to 150 nm

7- unload at 15% rate to 80% retract

8- load to 200 nm

9- hold indenter, take 20 data points

10- unload at 15% rate to 80% retract

11- hold indenter, take 60 data points

12- unload to 100% retracted

This procedure gave data to calculate the hardness, elastic modulus, plastic
deformation, stiffness and compliance of the material.

The first specimen, silicon nitride no.6 (1014 ions implanted), had three runs
on its surface. The runs were labeled "Run A, B and C" , with the run A in the middle
of the specimen, see Figure 18. The numbers in the parenthesis are the x and y coor-
dinates. The first run was to get familiar with the machine, so the large distance be-
tween indents could have stepped over the radiated area. Using the TRIM calculated
concentration peak profile, the full width at half maximum (FWHM) of the damaged
peak is 0.25pm, thus the ideal step distance between indents was 0.125 pm. But the
indents should be taken far enough apart so the area of indent is not effected by the
previous indent. An indent separation distance of 10 times the width of the indent is
a general rule of thumb. This would result in a indent width of 0.025 pm, which was
too small for the Nano Indentation machine. In order to get better resolution distance
perpendicular to the edge, we placed some of the indentation arrays at an angle. The
run C array was to focus around the area of damage at a slant for good resolution.

56

 

(42,483)
Run B
43 indents i. (0.34.87)
KKK Run A
‘1‘- , 6 indents (40.3.5).
\ l Run C
5‘ (0,1131) o-°°'

  

(0.15) ”“20 indents

    

(-12,14.3)

Specimen
Centerline

Material
/ Edge

    

Figure 18 Runs A, B and C on Silicon Nitride no 6 (1014 ions)

(-30,36.3)

Run D

43 indents
\.‘c

(0.6.3)

 

Specimen
Centerline

Figure 19 Run D on Silicon Nitride no. 7 (1013 ions)

57

Run E on Silicon Nitride no. 10 which is the control (non implanted) specimen.

(128,30)

‘
.
Q
.

R‘s. Run E
\‘20 indents

.
\
X

" (0,6)
I 9.76 degrees

Figure 20 Run E on Silicon Niu'ide no. 10 (control specimen)

 

The time required to take one indent was on the average of 15 minutes which made
for a long time to make a run. In order to make better use of time, several arrays
were done in one programmed setting. Runs F, G and H were done on stainless steel

no.4 (1014 ions). Run H was located off of the implanted region at the far left corner.
This was considered to be the control run for the stainless steel.

58

(-150, 31.3)

(6694.43) Run F

“3‘. Run H (4126.9) Run \18 mdents
\ 18 indents 13 indents
- (0.6.3)
‘x, (-606,3.53) 2..
(-6544,26.34)

_.r ldegrees |

 

Non-Implanted Region —>|<— Implanted Region —->

Figure 21 Runs F, G and H on Stainless Steel no.4 (1014 ions)
(-17.4,34)
3“... Run I
“-.._ 20 indents
,3...
(0.15)

 

Figure 22 Run I on Silicon Nitride no.7 (1013 ions)

59

Run J was done on a different material not concerning this subject. The runs
were not done in sequence of material number or number of implanted ions but rather
based on information obtained from previous runs.

Run K, L, and M were done on stainless steel no.2 (1013 ions)

 

(-6810,l94) (382,391) $334232)
“K... Run K
a... Run M Run L. 18 indents
\‘18 mdents 18 lndents
‘(-6669,169) (-240,14.7) (0,7)
Non Implanted
Region —> < Implanted Region ———->

 

Figure 23 Run K, L, and M on stainless steel no.2 (1013 ions)

Runs N and 0 were done on materials not part of this subject. Run P was on

stainless steel no.1 (1012 ions) which was the lowest ion implanted dose.

(-87.3,17)
\“-.,_ Run P
‘-...‘. 12 indents
“a. (0.6)

 

Figure 24 Run P on stainless steel no.1 (1012 ions)

 

 

-6395,82.2 -
g ) (33.23.32) .012’353)
Run S
. ‘-.,. Run R "-. R
a.“ 20 indents 20 indents 31332 indents
"-..(-6383,64.3)
(-6l.l,21) (0.17.9)
10 degrees “—-
Non Implanted
Region __._’ : “planted Region ———-’

 

Figure 25 Run Q, R, and S on Silicon Nitride no.8 (1012 ions)

CHAPTER 4

RESULTS

4.1 Detection of Specimen’s Radioactivity

The 2.5 MeV / nucleon cyclotron beam of 2ONe had a total energy of 50 MeV
which is above the Coulomb banier for both the ceramic and stainless steel speci-
mens, thus having the potential of causing radioactivity. After the implanted speci-
mens were removed from the pot, they were found to be radioactive. The specimens
exposed longest (the silicon nitride # 6 and stainless steel # 4) were found to be the
most active. The silicon nitride # 6 was the most active of the two and the gamma ray
was detected with a hand held safety detector, with the activity in counts / minute
which was read from a dial gauge on the detector. This detector was used instead of
the germanium detector at the Engine Research Laboratory since it was portable and
available. Table 12 shows the radioactivity of Silicon Nitride # 6

Table 12 Activity of Silicon Nitride # 6

DATE CNTS/MIN LN CNTS/MIN TIME REF. TIME

12/15/90 41864 10.64 5:22 0 hr

12/15/90 23000 10.04 6:35 1.22

12/15/90 15000 9.62 7:36 2.23

12/15/90 5000 8.52 11:24 6.03

12/16/90 300 5.70 15:24 10.03
M

Using the exponential decay law A = Aint exp'

300 cnts/min = 23000 cnts/min 4 exp -[(ln2 [half life) * 8.81 hr]
gives a half life of 1.4 hours. Using these numbers, an original activity of 41864.97

cuts/min was calculated which corresponds to 1.1311Ci.
Using the gamma detector at the Engine Research Laboratory, the energies at

61

62

which the gamma rays were being produced could be found. Five different detection
runs were taken on the following times and dates:

Date Time
12/18 10:15
12/18 16:15
12/19 9:00
12/19 23:00
12/20 21:15

The spectrum was taken on 4098 channels and the last three specu'um
showed no gamma peaks other than background. The first spectrum was stripped
with background shown in Figure 26. The x axis is the channel number and the y axis
is the number of gamma rays intercepted by the detector, known as counts.

 

l

”53 ' srlslrir '1

i

_l_ ..._..-

 

 

 

lrrrljlrlljulllllliiiinililI11:1::lli.:11

l M 2W 3m 4896

 

Figure 26 Silicon Nitride no. 6 Detected Spectrum

63
An energy calibration translates the spectrum channel number to the energy of

the gamma ray. The energy calibration shown in Table 13, sets the largest peak at
channel 1079.14 in the detected spectrum to a gamma ray energy of 271.07 KeV.

Table 13 Spectrum Energy Calibration

CHANNEL ENERGY
952.09 238.36 KeV

107 9. 14 271.07

1823.09 462.6

The second peak is to the right of the first and is located at slightly over chan-
nel 2000. This is the residue as a result of the subtraction of a large background peak
which had shifted slightly. The next spectrum showed the same 271.07 KeV gamma
ray but at a lesser intensity. There is no element in the Standard Table of Radioac-
tive Elements with this energy and a half life of 1.4 hours. The radioactivity had es-
sentially decayed too much by the time the germanium detector was used to make a
good assessment of the energy. Using the half life calculated from the first detection,

the activity was 3.136 x 10‘17 llCi by the time the first spectrum was read.

4.2 Nano Indentation

Control Silicon Nitride (Run E, S)

The Control (non-implanted) Runs of Silicon Nitride were examined first. The
hardness values of Run E on Silicon Nitride no. 10 and Run S on Silicon Nitride no.8
were smoothed and averaged giving the result is shown in Figure 27.

 

100 r 1
90 - Silicon Nitride Control Average .
so - a
70 e u
60 - d
Hardness
(GPa) 50 " d

40- -
30,. W .1

20 " ‘

 

 

 

10 1 l g l 1 l
5 10 15 20 25 30 35 40

DISTANCE FROM EDGE (pm)

Figure 27 Averaged Control Hardness of Silicon Nitride

This curve will be used as merely a light reference for all the silicon nitride
hardness plots. Each individual run has its own damaged and non damaged area sta-
tistics. which are all different from the others. The surface finish, exact beam location,
beam intensity and indent location are all t0pics of data variation. Since the variation
in data between runs is significant, the conclusions will be based most heavily on the
comparisons of the damaged against non-damaged areas for the given run. The sta-
tistical values for this run are as follows:

65

Averaged Control Run’s Hardness Statistics (Run E, S)
samples mean std. dev. sample std. dev.
14 30.78 1.28 1.327

All units in GPa. The sample standard deviation is always displayed and is
used when the number of samples is less than 50. The difference in formula between
standard and sample standard deviation is that sample standard has (n-l) replacing
a ’n’ in the denominator. Additional information about the implanted ion damage may
be obtained from the change in elastic modulus of the material. The average of elastic
modulus control runs is shown in Figure 28.

 

 

 

 

1000 u 1 fi

900 ~ Silicon Nitride Control Average 4

600 "' "'

700 " -4
Modulus .
(GPa) 6°” *

500 L W ‘

400 " "‘

300 " "

m . . . . . .

5 10 15 20 25 30 35 40
DISTANCE FROM EDGE (pm)

Figure 28 Averaged Control Elastic Modulus of Silicon Nitride.

This averaged curve will be used for control data representation of the elastic
modulus on silicon nitride for the same reasons as stated for the hardness data. The
statistical values for this data are as follows:

Averaged Control Entire Run Hardness Statistics (Run E, 8)
samples mean std. dev. sample std. dev.
14 462.3 22.1 22.95

66

2x1014ions/cm2 Dose Silicon Nitride (Run B, C)

Depths of 50, 100, 150 and 200 run where taken at each indent. The deepest in-
dent is the most accurate for representing the material hardness. The shallower in-
dents showed more surface and grain boundries effects. Also the percent error of
hardness value is lowered as more of the diamond tip is exposed to the material, thus
the deeper the indent, the less the error in measurement. For these reasons, only the
values for the 200nm will be used to predict change in hardness. The hardness values
shown may be considered high by some deep indent standards. There are several
methods to measure material hardness, none of which are directly related to the oth-
er. As explained by P. Blau [l3].

"Hardness numbers, even though expressed in the same units, need
not necessarily be the same if the same load is applied using indenters
of different shapes... Since hardness numbers depend on the method of
measurement, hardness is not a basic property, but rather a character-
istic of the material being tested under the specified indentation condi-
tions... Different hardness scales do not use the same area of the im-
pression in their calculations. For example, Brinell and Knoop hard-
ness are given as load per unit projected area while the Vickers
Hardness is expressed as a load per unit facet area. This difference is
one factor which can lead to different hardness numbers under the same
applied indenter load."

The main factor attributing to high hardness values was as the indents were tak-
en at deeper intervals (50, 100, 150 and 200nm), the hardness decreased, i.e. from
100 GPa for the 50 nm indent to 40 GPa for the 200nm indent. This phenomenon is
normal for nano indentation since only the surface is tested and not the bulk material
which represents the whole specimen. This hardness values did not reach a constant
number for the range tested, but they would decrease to a standard value as deeper
indents were taken. A further indication of high surface hardness values for nano indi-
cation is the high documented hardness values for materials tested by nano indenta-
tion methods as shown in Table 14.

67

Table 14 Varying Hardness Values

Material Standard Nano Indentation Measured
Hardness Hardness Hardness

NC 132 2 GPa [14] 22 GPa [15] 30 GPa

17-4PH 1 GPa [16] 8 GPa [13] 7 GPa

The ’standard hardness’ values are for the bulk material, measured by deep in-
dentation tests which result in lower hardness values than nano indentation tests.
The documented ’nano indentation hardness’ of 22 GPa [15] for the NC 132 was a re-
sult of a higher indentation load than what was used for the ’measured hardness’.
The measured hardness of 30 GPa used a light load resulting in a shallow indent. As
explained earlier, the hardness values decrease as the load is increased on the nano
indenter. Therefore, since the measured hardness tests for NC 132 used a lighter
load, the hardness values were greater than that of the documented nano indentation
hardness.

It is for this reason that most published nano indented hardness values are nor-
malized or used as reference. The values are left in original form for this research so
that a solid basis is maintained for future work. Thus the hardness values shown in
this work are the unique values for the shallow nano indentation tests.

Run B of Silicon Nitride no.6 (2x10l4 ions/cmz) showed a softening of the ce-
ramic in relation to its own average value which does not seem to correspond to the
Control Average Value. Figure 29 shows the hardness at a depth of 200 nm into the
material from the surface. The horizontal lines in Figure 29 indicate the 2 x standard
deviation window. This region is a 95.4% confidence region, meaning in this case
that the values located outside this region have a 95.4% probability of indicating a
change in hardness.

68

 

100 r ’r

9° ’ Silicon Nitride no. 6 (2X1014 i°“S/°m2)

 

so L 7
70 r J
so - .
Hardness
(GPa) 5. - 2 std. dev. region (20) y
n
.. . \A .

    

 

 

 

 

 

30 r l i
20 _ V Conn'ol Average 1
10 l l 1 1 l l

5 10 15 20 25 30 35 40

DISTANCE FROM EDGE (tun)

Figure 29 Hardness of Silicon Nitride no.6 (Run B)

The statistics on Run B are as follows, with the damaged portion being from 22 to
29pm:
2x1014 Dose (Run B) Silicon Nitride Non Damaged Portion Statistics

samples mean std. dev. sample std. dev.
34 39.48 3.91 3.97

2x1014 Dose (Run B) Silicon Nitride Damaged Portion Statistics
samples mean std. dev. sample std. dev.
7 29.81 4.76 5.144

The soft area is not a well determined peak with a full width at half max (FWHM) of
0.5rtm as TRIM suggested, but rather more of two soft spots approximately 3pm
apart, with the softest being at 25pm into the material. This deviation from the

69

TRIM’s calculated value will be discussed later. The graph shows a slight hardening
immediately before and after the soft portion with the values of 46.3 and 45.2 respec-
tively compared to the non damaged mean hardness of 39.48. The region of damage
has a 9.67 GPa average hardness drop which is a 24.5% average decrease in hard-
ness when compared to the outside damaged area (ODA) mean value If the softest
point is compared to the ODA mean, a 47.5% decrease in hardness results. A mean
hardness value of 39.48 GPa was found for the ODA, which is a 8.7 GPa difference
from the Control Average Run. This change could be attributed to the Nano Indenter
System Calibration, possibly from a noted 31% change in room humidity that occurred
between the B and Control Runs. Figure 30 shows the elastic modulus for Run B
compared with its Averaged Control Run.

 

1000 r u u I r

l

soo- Silicon Nitride no. 6 (2x1014 ions/cmz)

800 ' ‘

700 " (20) "

Modulus600 _ Run B .
(GPa)

 

     

 

 

 

 

 

500 r ‘JA\ ,t q
if \- \/
ANT? Air-”J“
400 ~ \ 4
Control Average
300 ~ 4
200 ‘ ‘ ‘ ‘ ‘ 1
5 10 15 20 25 30 35 40
DISTANCE FROM EDGE (um)

Figure 30 Elastic Modulus of Silicon Nitride no.6 (Run B)

As in the hardness plots, a decrease in elasticity is shown to occur at 25pm.
The double soft region is not as apparent in this case as in the hardness plot. Once
again the Control Average is much lower than that of Run B, which gives further evi-

70

dence of varying machine calibration. If the low extreme value is used along with the
ODA mean for this run, the Silicon Nitride decreased in elasticity by 46.5 %. This val-
ue is close to the 47.5% decrease in hardness. Even though the second soft peak is
not dominate, it still falls outside the ODA standard deviation indicating that the dam-
aged region for elasticity still exists for 22 to 29m.

The statistics on the Run are as follows, again with the damaged portion being
from 22 to 29pm:

2x10l4 Dose (Run B) Silicon Nitride Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
34 536.4 30.19 30.65

2x10l4 Dose (Run B) Silicon Nitride Damaged Portion Statistics
samples mean std. dev. sample std. dev.
7 494.6 98.32 106.2

In Run C, a window of the beginning of the damaged region can be seen in Fig-
ure 31.

71

 

 

 

 

 

 

 

100 I r
90 ~ . . . . 14 .
Silicon Nitride no. 6 (2x10 ions)

80 '- d

70 r -t

6. i- ..
“fag? so _ (26) Run c .

I

40 '- i

30 -—-—-——-—----- “NR-

20 _ Averaged Control ‘

10 l 1 l l l

14 16 18 20 22 24 26

DISTANCE FROM EDGE (um)

Figure 31 Hardness of Silicon Nitride no.6 (Run C)

This data is tough to decipher and does not show the signs of softening. This
was taken to the right of Run B and could have been outside the implanted area. An-
other possibility is that of miscalculated distance, so that the shown distance in from
the edge may not be correct. In any case, this run did not seem to traverse the dam-
aged region so it may be used as a Control Run.

2x1014 Dose (Run C) Silicon Nitride Entire Run Statistics
samples mean std. dev. sample std. dev.
20 40.55 3.89 3.99

Run C was performed on the same day as Run B. The mean here is 40.55 GPa
comparing it to the ODA Run B mean of 39.48 GPa shows that the indenter gave con-
sistent values for that day, but very different values for the other days. The entire run
standard deviation for Run C is 3.89, and 3.91 for the ODA region of Run B, so these
Statistics indicate that Run C is a ODA Run.

72

This same run gave the elastic modulus, shown in Figure 32..

 

 

 

 

 

 

1000 r . , , ,
,0, . Silicon Nitride no. 6 (2x1014 ions/cmz) ,
coo - _
700 ~ .

Modulus

(GPa) 600 - (20) Run C l
500 ~ I W rm. -
1 “-4 A x' ‘\./
z/}\/" v V ‘v/

400 h \ Averaged Control ‘
aoo - -

200 l 1 1 l 1 1
5 10 rs 20 25 30 35 4o

DISTANCE FROM EDGE (um)

Figure 32 Elastic Modulus of Silicon Nitride no.6 (Run C)

The statistics for this run are:

2x1014 Dose (Run C) Silicon Nitride Entire Run Statistics
samples mean std. dev. sample std. dev.
20 523.5 30.17 30.95

Notice that the range of distance of this plot is greater than the hardness plot,
giving the curve a more short and choppy appearance. The Standard deviation of
30.17 is greater than that of the Averaged Control Run, but all points are in an accept-
able confidence region. Thus this run can still be considered a Control Run.

73

2x1013 ions/cm2 Dose Silicon Nitride (Run D, I)

Figure 33 shows Runs D and I and the Averaged Control Run on the next low-

013 2).

er implanted ion dose in Silicon Nitride, no.7 (2x1 ions/cm

 

100

so - Silicon Nitride no. 7 (2x1013 ions/cmz) *

 

 

  

 

 

 

 

 

 

 

 

 

do - ‘
70 - ‘
so - "
Hardness Run D
(GPa) 5" ’ . Control Average Run D‘
w Run 1(26) 7\ /\ I (29) 4* .
ll
so ~
20 -
v RunI
10 1 l 1 r l L
5 to 15 20 25 so as 40

DISTANCE FROM EDGE (1.1m)

Figure 33 Hardness of Silicon Nitride no.7 (Run D and I)

The data here is hard to decipher. Runs D and I show a resemblance to each

other but the resemblance to the Runs on 2x1014 dosage is somewhat lost. Run I
has a ODA mean hardness value of 26.79 GPa where Run D has a ODA value of 32.8
GPa. These runs were taken on different days. Run I shows a softening in two

places, one at 18pm and at 25pm. Run D has three soft spots at the depths of 18, 24
and 29pm. Unfortunately the resolutions of the points are not better, there is a possi-

bility of a softer point at 25pm for Run D according to the straddling of the point’s
slopes around it. The statistical data on these runs are as follows:

74

2x1013 Dose (Run D) Silicon Nitride Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
27 32.81 7.45 7.59

2x1013 Dose (Run D) Silicon Nitride Damaged Portion Statistics
samples mean std. dev. sample std. dev.
9 29.9 4.14 4.39

2x1013 Dose (Run 1) Silicon Nitride Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
14 26.79 6.82 7.08

2x1013 Dose (Run I) Silicon Nitride Damaged Portion Statistics
samples mean std. dev. sample Std. dev.
6 22.46 6.56 7.18

The damaged area portion is again taken to be 22 to 29pm where softening was

seen on the 2x1014 dose Silicon Nitride. The damaged area results are tough to con-
firm since the hardness from the non damaged areas is as low as the damaged area.
The standard deviations of the ODA runs are both higher than the damaged area’s
standard deviation. This data is much rougher than the previous runs, but the results
should be based on comparing the damaged region statistics to the ODA statistics for
same Specimen. Based on this data, this region of interest does not clearly Show ma-
terial damage.

75

The elastic modulus for Runs D and I are shown in Figure 34.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1000 r
m . Silicon Nitride no. 7 (2x1013 ions/cmz) .
‘ an. o A/Runl -
700 P \ Rl_rnI .
Modulus ‘ (20) Run D
l
(GPa) ‘°° ' i
500 '- T
400 r .4
300 *- ..
r (20)
200 . n 1 . n e
5 10 15 20 25 30 35 40

DISTANCE FROM EDGE (um)

Figure 34 Elastic Modulus of Silicon Nitride no.7 (Run D and I)

The statistics for these runs are as follows:

2x1013 Dose (Run D) Silicon Nitride Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
23 493.4 83.3 85.2

2x1013 Dose (Run D) Silicon Nitride Damaged Portion Statistics
samples mean std. dev. sample std. dev.
9 467.2 56.68 60.1 1

2x1013 Dose (Run 1) Silicon Nitride Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
14 475.0 136.5 141.6

76

2x1013 Dose (Run I) Silicon Nitride Damaged Portion Statistics
samples mean std. dev. sample std. dev.
6 385 42.4 46.44

As in the hardness plots, there exists regions of softening where for higher ion
dosages there were not any. As in the hardening plots, the apparent softening at 18
um clouds the 25 um softening. Notice that Run I has both the highest and lowest
values, thus a higher standard deviation of 136.5 for the ODA regions. This elastic
data is rougher than the hardness data, which led us to beleive that the material was
not as polished as the others. In any case, the softening is not noticeably apparent.

77

2x1012 ions/cm2 Dose Silicon Nitride (Run Q, R)

The lowest dose (2x1012) Silicon Nitride no. 8 is shown in Figure 35.

 

 

 

 

 

 

 

 

 

100 r r T
Silicon Nitride no. 8 (2x1012 ions/cmz)
80 " —1
60 *' ..
Hardness
(GPa) Run Q and R Run Q Run R
40 ~ ‘ * 4k _
Run Q g (20) °
20 - 20),
o . . . . .
10 15 20 25 30 35 40

DISTANCE FROM EDGE (um)

Figure 35 Hardness of Silicon Nitride no.8 (Runs Q and R)

The Statistical data from these runs are as follows:

2x1012 Dose (Run Q) Silicon Nitride Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
1 1 31.19 4.54 4.77

2x1012 Dose (Run Q) Silicon Nitride Damaged Portion Statistics
samples mean std. dev. sample std. dev.
6 32.7 6.55 7.18

78

2x1012 Dose (Run R) Silicon Nitride Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
12 29.1 9.10 7.41

2x1012 Dose (Run R) Silicon Nitride Damaged Portion Statistics
samples mean std. dev. sample std. dev.
5 29.64 2.57 2.87

Fortunately, this data is much better than the previous. The standard devia-
tions are much smaller, making it easier to see data deviation. This run Shows no
damaged area. Actually the mean values for the damaged region are greater than the
ODA regions.

The elastic modulus plot for this run is shown in Figure 36.

 

1000

900 - Silicon Nitride no. 8 (2x1012 ions/cmz)

 

 

 

 

 

 

 

 

 

 

000 t ‘
Run Q
no -1 Run Q (20')
Modulus
(GPa) 500 -
5°" "v RunQ k: A, , " .
. ........................ .. Wis? W
400 L a-
300 - ‘
zoo ‘ ‘ ‘ ‘ ‘
10 15 20 25 so as 40

DISTANCE FROM EDGE (um)

Figure 36 Elastic Modulus of Silicon Nitride no.8 (Runs Q and R)

79

The statistical data from these runs are as follows:

2x1012 Dose (Run Q) Silicon Nitride Non Damaged Portion Statistics
samples mean Std. dev. sample std. dev.
1 1 540.2 77.02 80.78

2x1012 Dose (Run Q) Silicon Nitride Damaged Portion Statistics
samples mean std. dev. sample std. dev.
5 502.2 63.53 71.03

2x1012 Dose (Run R) Silicon Nitride Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
13 496.5 84.96 88.4

2x1012 Dose (Run R) Silicon Nitride Damaged Portion Statistics
samples mean std. dev. sample Std. dev.
4 435.5 28.32 32.7

The data, even though very unsmooth, shows no change in the elasticity of the
material.

80
Control Stainless Steel (Run H, M)
The control (non-implanted) runs of stainless steel were again examined

first. Run H off to the side of Stainless Steel no.4 and Run S on Stainless Steel no.2
were averaged and are shown in Figure 37.

 

20 I T T I I

1. . Stainless Steel Control Specimen Average .
16 r -

r4~ -

Hardness ,2 _
(GPa)

10:- -l

o- W .

 

 

 

4 L L I l 1
5 10 15 20 25 30 35

DISTANCE FROM EDGE (tun)

Figure 37 Hardness for Stainless Steel (Control Runs H and M)

The statistical data for this Run is as follows:

Averaged Control Entire Run Statistics (Run H, M)
samples mean std. dev. sample std. dev.
14 7.75 0.305 0.316

The softer stainless steel made it easier to obtain a highly polished surface,
thus the stainless steel data was much smoother, resulting in smaller standard devi-

81

ations.

Figure 38 shows the elastic modulus for the Controlled Average Run.

 

 

 

 

800 r r r r r
Stainless Steel Control Specimen Average
700 t- -
600 '- ..
Modulus
(GPa) 500 P ~
400 - a
W
300 r ..
200 k 1 ‘ ‘ ‘
5 10 15 20 25 30 35

DISTANCE FROM EDGE (pm)

Figure 38 Elastic Modulus for Stainless Steel (Control Runs H and M)
The statistical data for this Run is as follows:
Averaged Control Entire Run Statistics (Run H, M)

samples mean std. dev. sample std. dev.
15 322.1 7.68 7.95

82

2x1014ionS/cm2 Dose Stainless Steel (Run F, G)

The highest implanted ion dose of Stainless Steel was no.4 (2x1014
ions/cmz). Runs G and F were done on Stainless Steel no. 4 and they Show good re-
peatability. Figure 42 shows Run G, F and the Averaged Control Run. A soft region
is clearly seen for both Run G and F, around the region of 12 to 21 pm, the softest re-
gion being 16 to 17pm.

 

20 I I 1 i

1

to ~ Stainless Steel no. 4 (2x1014 ions/cmz)
rs - ‘
14 - q

Hardness 12 - ‘

 

 

(GPa) Run G Run F
1° " M1: §_n_d G \ l (26) p

L Averaged Contro1\ \ I ll
6 , ................................. 11:: ----- ., .- q

..........

 
 

 

 

 

 

5 10 15 20 25 30 35
DISTANCE FROM EDGE (run)

Figure 39 Hardness of Stainless Steel no.4 (Run F and G)

The Statistical data on these Runs is as follows:

2x1014 Dose (Run F) Stainless Steel Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
10 7.98 0.306 0.322

83

2x1014 Dose (Run F) Stainless Steel Damaged Portion Statistics
samples mean std. dev. sample std. dev.
5 6.62 0.833 0.931

2x1014 Dose (Run G) Stainless Steel Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
7 7.91 0.239 0.258

2x1014 Dose (Run G) Stainless Steel Damaged Portion Statistics
samples mean std. dev. sample std. dev.
5 6.81 1.39 1.56

The damaged region was considered to be 12 to 21 um. Run G, which was
more focused on the damaged region, shows a leveling off to the same hardness as

Run F. The double soft peaks which were seen in the Silicon Nitride of 1014 dosage,
can be also seen here but in a less significant state. Comparing the control mean
hardness value of 7.75 GPa to that of ODA of Runs F and G, 7.98 and 7.91 respec-
tively shows a maximum deviation of 2.9%. The averaged softening of the material is
calculated to be 13.4%. But if the peak softness value is used, an overall maximum
softening of the material is 49.5%.

84

 

700 r r r r t

Stainless Steel no. 4 (2x1014 ions/cmz)

600 r 2
500 _ 4
Modulus 40°

(GPa)

 

300

 

200

 

 

 

100 14* Li

5 10 15 20 25 30 35

DISTANCE FROM EDGE (um)

 

Figure 40 Elastic Modulus for Stainless Steel no.4 (Run F and G)

There is a large similarity between this elastic modulus plot and the hardness
plot, both Show a drop at 25pm.

The statistical data for Runs G and F are as follows:

2x1014 Dose (Run F) Stainless Steel Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
10 307 21.95 23.13

2x1014 Dose (Run F) Stainless Steel Damaged Portion Statistics
samples mean std. dev. sample std. dev.
5 286 27.5 30.75

85

2x1014 Dose (Run G) Stainless Steel Non Damaged Portion Statistics
samples mean std. dev. sample std. dev.
7 317.1 16.08 17.36

2x1014 Dose (Run G) Stainless Steel Damaged Portion Statistics
samples mean std. dev. sample Std. dev.
5 281.6 46.75 52.27

This shows a 6.8 % decrease in the modulus of elasticity for Run F and a 11.4 % de-
crease for Run G, an average of 9.1 % using the mean values. Using the extreme val-
ues results in a 22.8 % and a 38.5 % decrease for Runs F and G respectively.

86

2x1013 ions/cm2 Dose Stainless Steel (Run K, L)

Run K and L were on Stainless Steel no.2 which is the next lower dose of ions

(1013 ions), the hardness graph is shown in Figure 41.

 

20 I f If

,, _ Stainless Steel no. 2 (2x1013 ions/cmz) 4

16 r a

14 r ‘

 

 

 

Hardness
(GPa) ‘2 ' ‘
m ’ Averaged Control "
a e
6 _ Run L Run K d

 

 

 

5 10 15 20 25 30 35
DISTANCE FROM EDGE (tun)

Figure 41 Hardness of Stainless Steel no.2 (Run K and L)

The softening of the material shown previously, but at a much lower significance.

2x1013 Dose (Run K) Stainless Steel Non Damaged Portion Statistics

samples mean std. dev. sample std. dev.
14 7.31 0.373 0.387

2x1013 Dose (Run K) Stainless Steel Damaged Portion Statistics
samples mean std. dev. sample std. dev.
4 7.22 0.199 0.229

87

2x1013 Dose (Run L) Stainless Steel Non Damaged Portion Statistics
samples mean Std. dev. sample std. dev.
1 1 7.29 0.379 0.397

2x1013 Dose (Run L) Stainless Steel Damaged Portion Statistics
samples mean std. dev. sample Std. dev.

4 7.09 0.453 0.523
The ODA mean harness values are within 8.5% of each other, making for good
repeatability. The average decrease in hardness here is 1.98% which is inside the

standard deviation, making the damaged region virtually undetectable according to
statistics. If the softest region is used, the decrease in hardness is 11%.

The elastic modulus of Runs L and K are shown in Figure 42.

700

 

Stainless Steel no. 2 (2x1013 ions/cmz)

 

    

600 ~ ~

500 ~ 1
Modulus 400 r 5
(GPa) Control ‘Average / Run K (20)

 

300 *

 

200 P -

 

 

5 10 15 20 25 30 35

DISTANCE FROM EDGE (pm)

100

 

Figure 42 Elastic Modulus of Stainless Steel no.2 (Run K and L)

88

The statistical data is as follows:

2x1013 Dose (Run K) Stainless Steel Non Damaged Portion Statistics
samples mean Std. dev. sample std. dev.
14 319.4 15.22 15.8

2x1013 Dose (Run K) Stainless Steel Damaged Portion Statistics

samples mean std. dev. sample std. dev.
4 322.5 12.36 14.27

2x1013 Dose (Run L) Stainless Steel Non Damaged Portion Statistics

samples mean std. dev. sample std. dev.
10 320 16.8 17.7

2x1013 Dose (Run L) Stainless Steel Damaged Portion Statistics

samples mean std. dev. sample std. dev.
5 307.8 20.9 23.37

Run K shows no apparent change at 25pm. Run L shows a 3.8% decrease in
damaged region but the decrease falls inside the large ODA standard deviation of
16.8 from the mean value of 320 which drops the allowable ODA deviation to 303.2
which is actually lower than the dip in the damaged region. Thus the stainless steel

can be assumed to have no change in elastic modulus at this dosage.

89

2x1012 ions/cm2 Dose Stainless Steel (Run P)

The lowest dose (2x1012 ions) was on Stainless Steel no.1 which is tested in
Run P. Only one Run was performed due to time constraints. Run P Shown in Figure
43 shows no distinct softening of material as would be expected from the trend of pre-
vious data.

 

20 r r

1

16 - Stainless Steel no. 1 (21't1012 ions/cmz)

16’ ‘

 

 

14 P -
Hardness

(GPa) 12 r -

10 *- ‘

o L IIUWII? (2‘1) ‘

e r .

 

 

I I l I l l

6 8 10 12 14 16 18

DISTANCE FROM EDGE (um)

 

Figure 43 Hardness of Stainless Steel no.1 (Run P)

The statistical data is as follows:

2x1012 Dose (Run P) Stainless Steel Entire Portion Statistics
samples mean std. dev. sample std. dev.
1 1 7.31 0.25 0.26

The Control Average has a mean of 7.75, a standard deviation of 0.305 and a
sample Standard deviation of 0.32 which is very similar to this run.

90

Figure 44 shows the elastic modulus of Run P.

 

 

 

 

 

 

 

700 r r r r r r
m _ Stainless Steel no. 1 (2x1012 ions/cmz) ‘
500 ~ ‘
Modulus
(GPa) 40° . Run P ‘
(26)
F n
300 r- \/\/\——— a
V
200 P j
100 ‘ L” l ‘ ‘ 1

6 8 10 12 14 16 18

DISTANCE FROM EDGE (um)
Figure 44 Elastic Modulus of Stainless Steel no.1 (Run P)

2x1012 Dose (Run P) Stainless Steel Entire Portion Statistics
samples mean std. dev. sample Std. dev.
11 317.45 36.39 38.17

The large point near 10pm, causes some problems, if it is eliminated the mean
is 307.4, the standard deviation is 18.56 and the sample standard deviation is 19.57
which drops the statistics dramatically but still does not drop them to the Control Run
values.

The smooth surface and low variation in data points outside the damaged por-
tion for the stainless steel indicated that the spacing between indents was adequate.
If the spacing had not been close enough, the data would be more erratic, much like
the silicon nitride data. Ironically with the stainless steel being a softer material, the
indenter penetrated deeper into the material thus requiring the indents further apart

91

than those in the hard silicon nitride. In the stainless steel, the indents were placed

approximately 8 pm apart, since the apparent width of the indent was 4 pm. Tra-
versing the damaged region at an angle, the resolution between indents from the edge

is at best 1.8 pm. The indents in the silicon nitride were less than 1 pm in width and

placing the run at an angle, a 0.8 pm resolution between indents from the edge was
achieved. Even with this better resolution, the points were too far apart for the sili-
con nitride’s rough surface. The data was too scattered for comfort, but fortunately
the damaged regions could still be seen.

4.3 TRIM

TRIM predicted a damaged depth of 17.8 um with a FWHM of 0.25 pm for
the silicon nitride, and a depth of 10.5 um and a FWHM of 0.5 pm for the stainless
steel. The results showed a damaged region in the silicon nitride from 22 to 29 pm
with the maximum damage at 25 pm. The stainless steel damaged region was from

12 to 21 pm with the most change at 16 pm. The silicon nitride shows a sort of dou-
ble soft region while the stainless steel shows more of a distinct softening at a point.
This error between the two is not distinctive of TRIM. One reason for the error may
be in that TRIM considers the target to be amorphous with atoms at random loca-
tions, and thus directional properties of the crystal lattice are ignored. Also the nucle-
ar reactions are not considered in TRIM. These may be subjects that might change
the results a little but certainly not enough to match the experimental results. The

2

plots of TRIM Shown were for an ion dosage of only 104 ions/cm due to the amount

of computer time required. If the desired 1014ions/cm2 dosage was run, the maximum

concentration peak would stay the same, but the FWHM would grow, resulting in a

damage region of greater than 0.5 pm. But the TRIM’s result would most likely be
far from the experimental result.

A more likely error would come from the physical set up of the specimens in
the beam line. The intersection of the flats of the two Specimens, where the beam
penetrated into the crack, could be the cause of conflicting implanted depths. Under
5000 X magnification, the intersection of the two specimens was quite obvious, the
crevice caused by the surfaces not being perfectly smooth. Under high magnification
the surfaces would be rough, with the specimen flats riding on each others ridges, see
Figure 45.

92

 

Specimen 1
———>
————>
Beam
—————->
———>
Specimen 2

Figure 45 Beam Incident on Magnified Specimen Contact Surfaces

As seen in TRIM, the beam penetrates a set amount of material and then de-
posits ions at that location. Using this scenario, the beam will enter the crack, travel
through several voids and material peaks until it has penetrated the set amount of
material. This will cause the ions to be implanted at a distance further in than if the
beam penetrated solid material. The beam that travels along the centerline of the
crack will encounter a less dense material and deposit further into the material, result-
ing in a distance implantation profile as shown in Figure 46.

93

Ion implantation depth profile Jagged Surface

 

Beam Direction

 

Specimen material

Figure 46 Varying Ion Implantation Depth with Jagged Surface.

The maximum depth of the indent was 200 nm which is 0.2 um. This depth
could be at the same magnitude as the surface finish of the silicon nitride, meaning
that the depth of the indent may not be deep enough to overcome the varying surface
implantation depths, see Figure 47.

Damaged Section
< = Jagged Surface

 

 

 

P

Beam Direction

 

Specimen material

Figure 47 Indentation of Deeper Implanted Regions

94

Using this scenario, the damaged regions would be located deeper than ex—
pected and would also be broader than expected. Also, the smoother the Specimen’s
surface, the closer the damaged depth would be to the calculated depth. With this in
mind, the error in damaged depth of the stainless steel is 5.5 pm where the error in

depth for the silicon nitride is 7.2 pm. According to the hardness data, the stainless
Steel had the smoother surface.

CHAPTER 5

SUMMARY AND CONCLUSIONS

1) The implantation dosage of 2x1014 ions/cm2 caused a change in hardness

and elastic modulus in both the stainless steel and silicon nitride. The threshold local

dosages for both the stainless steel and silicon nitride is 2x1013 ions/cm2 , to be con-
servative. The stainless Steel having a better surface finish, enabled the data to be
smoother, thus easier to read. Both materials showed no radiation damage at the

012 ions/cmz, thus the threshold implantation dosage to create ma-

terial property change was achieved.

lower dose of 2x1

2) An implanted sample for wear studies would contain a series of these local
implants, the threshold for implant based on these studies is an order of magnitude
greater than that which caused local damage.

3) The measured hardness and elastic modulus values for both the silicon ni-
tride and the stainless steel were high when compared to tabulated values. The mea-
sured silicon nitride hardness was 30.78 GPa, compared to a typical documented val-
ue of 13 GPa. The measured modulus of elasticity was 462.3 GPa compared to 380
GPa typical. The measured stainless steel hardness was 7.9 GPa while the typical
value is 1.05 GPa. The measured modulus of elasticity was 322.1 GPa and 200 GPa
for the theoretical. The high values are typical of nano indentation tests. The data
was used in a relative sense, so exact hardness was not needed.

4) The implanted 2x1012, 2x1013 and 2x1014 ions/cm2 non radioactive ions
were carefully chosen to equal to the number of radioactive ions to produce a required

activity for wear analysis. Using these dosages, the resulting radioactivity of Na22

ions would yield 0.228, 2.28 and 22.8 llCi respectively. If Be7 ions were implanted at

these dosages, the resulting activity would be 4.06, 40.6 and 406 llCi respectively.

5) The nano indenter machine is set up so that no reading calibration is needed
other than horizontal plane calibration (x-y plane). Thus the same hardness values
should occur at different times, but this was not the case. Noting humidity change as
our only time variant, the variation in hardness was from 30.78 GPa for the Averaged
Control Run to 40.55 GPa for Run C, an unexplained variation of 24%. The same vari-
ation shows up in the modulus of elasticity values for that same day. The Nano In-

95

96

denter is an extremely sensitive machine, one which corrects for thermal drift. There
seems to be a need for other ambient corrections such as humidity change as well.

6) The detected gamma ray from the silicon nitride no. 6 specimen, had an
energy of 271.07 KeV. This energy could not be found in the Radiation Tables. This
measured energy value has an estimated error of at most 3 KeV, which still does not
match a known gamma ray to our created gamma ray. There is a possibility, in the
path of new research, to come across new results. Before introducing this possible
new gamma ray, more samples should be examined.

7) From these results, the Ion Implantation method is an excellent radiotrac-
er wear detection method as long as the threshold implanted ion dose is not reached.

8) A engine wear ion detection system was designed for internal combustion
engine studies, shown in the appendix.

Chapter 6

RECOMMENDATIONS

In order to continue the progress toward a better understanding of the damage
effects of Ion Implantation, these recommendations for future work seem appropriate.

1. A more highly polished specimen surface needs to be obtained. More non-dia-
mond compounds and chemical etching should be included in the final preparation pro-
cess.

2. A secure and easy specimen mounting system should be designed and fabricated
for the Nano Indentation Machine to ensure proper alignment and flatness of speci-
men mounted surfaces.

3. The cyclotron beam should be more controllable, resulting in less beam current
fluctuations. A higher beam current would most likely be more stable and also more
economical for implanting high ion doses.

4. The testing of implanted samples with polyenergetic ions will provide additional
verification of wear damage. Battelle Memorial Institute of Columbus, Ohio is plan-
ning to provide this testing. These implanted samples must be used in engine tests
to determine the factors that influence wear measurement under realistic conditions.
This work could begin with SLA activated parts.

97

REFERENCES

LIST OF REFERENCES

1. Personal communication.

2. Bhattacharya R.S., Rai A.K., Pronoko P.P.. "Journal of Applied Physics", 15
May 1987, 61, (10) 4791-4794.

3. Townsend P.D.,"Rep. Prog. Physics,", May 1987, 50, (5), 501-558.

4. Burggraaf A.J., Gelling P.J., Scholten D.. "High Tech Ceramics." (Part A)
[Processing Conference], Milan, Italy, 24-28 June 1986, Elsevier Science Publishers.

5. Cochran J.K., Legg K.O., Solnick H.F.. "Defect Properties and Processing of
High—Technology Nonmetallic Materials." [Processing Conference], Boston, USA,
14-17 Nov. 1983, Elsevier Science Publishers.

6. Oliver W.C., McHargue C.J., Farlow G.C., White C.W.. "Defect Properties
and Processing of High-Technology Non—Metallic Materials." [Processing Confer-
ence], Boston, Massachusetts, USA 2—4 Dec. 1985, Materials Research Society.

7. Burkhardt T., McHarris W., Chung Y., Lloyd 1., Rachel T., Schock H., Grum-
mon D., Shalek R., Ronningen R., Mallory M.. "Developments in High Energy Ion Im-
plantation for Wear Studies." Michigan State University, Status Report MSUERL—90-
5. 1990.

8. Personal communication.

9. Mallory M.. Schock H., McHarris W., Ronningen R., Sherill B.. "Production

and Implantation of 7Be For Wear Analysis", Michigan State University Report,
Nov. 17, 1988.

10. General Electric, ’Chart of the Nuclides’, Nuclear Energy Operations, 175
Curtner Avenue, M/C 684, San Jose, California 95125 USA.

11. Ziegler J.F. Biersack J.D., Littmark U.. ’The Stopping Power of Ions in Solids.’,

98

99
Pergamon Press Inc., Maxwell House, Fairview Park, Elmsford, NY. 10523 USA.
12. Cathcart J.V. "Operating Instructions for the Mechanical Properties Micro-
probe", High Temperature Materials Laboratory. Oak Ridge National Laboratory,
Oak Ridge, Tennessee. September 1988.

13. Personal communication.

14. Blau P., "The Lab Handbook of Microindentation Testing." Oak Ridge N ation-
al Laboratory, Oak Ridge, Tennessee. (1988).

15. NC 132 Material Documentation. Norton Company, Industrial Ceramics Divi-
sion, Worcester, Mass. 01606

16. Tanaka 1., Pezzotti G., Okamoto T., Miyamoto Y., Koizumi M.. "Hot Isostatic
Press Sintering and Properties of Silicon Nitride without Additives."Journal of Ameri-

can Ceramic Society., 72 [9] 1656-60 (1989).

17. Lindeburg M. Engineer in Training Review Manual, Sixth Edition. Profession-
al Publications, Inc. Belmont, CA 94002.

18. Mallory M. L. IEEE Trans. Nucl. Sci, 55-39, NQ,4 (1983) 2061.

19. Caskey G., Heilbronn L., Remington B., Galonshy A., Deak F., Kiss A., Seres
Z.,. MSUQL-QZQ, Michigan State University, Preprint (Nov. 1987).

20. Remington B. S., Caskey G., Galonshy A., Gelbke C.K., Heilbronn L., Heltsley
J., Tsang M.B., Deak F., Kiss A., Kasagi 1., Kolata J.J.. Phys. Rev. C, 34 (1986) 1985.

21. Lynch W. 6., Ann. Rev. Nucl. art. Sci., 111 (1987) 493.

22. Mallory M. L. ESE. 11m Int, Cont, on nglgmns and Their Applications, Ion-
ics, Tokyo (1987) 558.

23. Ajsenberg-Selove F. and Lauritsen T. , Nucl. Phys, A221(1974) 68.

APPENDIX

DESIGN OF A PISTON RING WEAR MEASUREMENT
SYSTEM FOR ION IMPLANTATION

Even though the main thesis is based on the question of material property
change due to ion implantation, a system to actually use the wear measurement tech-
nique has been designed and documented but has yet to be run. The system consists
of a one cylinder research engine, dynarnometer, oil and piston ring ion detection sys-
tems and an exhaust dilution system.

Research Engine

The engine choose for the study of radioactive piston ring wear was a one cyl-
inder 11 HP Briggs and Stratton Engine model number 254422, type 4015-01, with an
aluminum block, cast iron bore which was donated by Briggs and Stratton. The en-
gine is a compact, self sufficient unit that controls its engine speed via a carburetor
governor. The Briggs and Stratton engine is mechanically simple, reliable and parts
are readily available. The design of the engine makes it less sensitive to bolt torque
tolerances and leaves little room for mistakes in assembly, making it a good choice for
frequent tear down. The engine has several oil holes for easy connection to an exter-
nal oil detection system. The engine also has a low oil shut off system that shuts the
engine off when the oil level is low. This option is a good safe guard since the engine
oil will move through the engine into other components for oil detection. The engine is
turned off by grounding the ignition ground wire (not spark). The top ON/OFF switch
merely makes or breaks contact with the grounded engine cover. This switch was
disconnected and the grounding wire attached to a red cut out button on the dynamom-
eter’s control panel for quick engine cut out. The low oil cut out switch is operates us-
ing a float which rises up when the oil level is at the correct level, opening the switch
so that the ignition switch is not grounded. When the oil level drops to a certain lev-
el, the switch closes, grounding the ignition and thus stopping the engine. The
grounding of the oil switch also causes a red bulb to glow on the front cover of the en-
gine. The recommended oil capacity is a 1.25 quarts, but the engine is not full until 2
quarts are used.

100

101

Besides detecting the engine oil, a piston ring will be detected. The second
compression piston ring was chosen to be radiated in order to distribute more of the
radioactive ions in the oil. Placement of the detector in relation to the radioactive por-
tion of the ring is critical. In order to keep the second piston ring from rotating about
the piston, the piston ring was pinned. To pin the piston ring, notches were made in
both ends of the piston ring. Then a hole was drilled in the piston secondary ring
groove at an area of thick piston wall. A hardened pin was then pressed into the hole
and the piston ring installed so that the grooves matched up with the pin. This pinned
ring design makes it impossible for the pin to come out of the piston since the notch is
constantly pressing on the pin. The pinned ring was done early in the break in period
of the engine so that wear patterns were created in junction with the non rotating pis-
ton ring. The one gallon stock fuel tank were replaced by a five gallon fuel tank to ac-
commodate the fuel supply for the long wear data runs.

Dynamometer

The dynamometer is used to produce a load on the engine. With the dyna-
mometer, the engine can be run at a set Speed or a set torque can be placed on the re—
search engine. The dynamometer can also be used to motor or even start the engine.
The dynamometer is a EATON 50 HP eddie current dynamometer controlled by DY-
NAMATIC controllers, all of which was donated by NASA.

The dynamometer’s support bed is extended for mounting the engine on to the
bed. An engine stand was made to support the engine and to also match the engine’s
drive shaft with the dynamometer’s drive shaft. The engine stand was make of three
inch channel for the legs and half inch steel plate for the base plate. The half inch
plate was drilled and tapped to the engine’s bolt hole pattern. The dynamometer’s
bed was also drilled and tapped to secure the engine stand. A two inch tall Steel lip
was welded around the engine stand’s base plate as a safety measure to contain the
engine oil in case of a spill. The basin created by the lip allows for containment of all
the engine oil.

There must exist a flexible coupling between the dynamometer and the engine,
so that engine vibrations are not transmitted through the drive shafts. A Sure Flex
sleeve rated at 25 HP and fourteen degrees rotational flexibility was used as the cou-
pling, any less flexibility would shear under load.

Running the engine and dynamometer indoors caused noise and exhaust prob-
lems, so the dynamometer was made mobile by placing it on caster wheels. A rail

102

system was made to guide the wheels out a bay door to the outdoors. The track was
long enough as that the dynamometer could be completely removed from the building
and the bay door closed. To accommodate this mobile dynamometer, long water and
electrical lines were used and the control panel stand was also put on wheels. The
control panel controls the entire dynamometer / engine setup.

The dynamometer is always run in manual mode, and either in the drive or ab-
sorb mode. In absorb mode, the dynamometer absorbs power from the firin g engine
which can be controlled by the speed indicator. In drive mode, the dynamometer
drives the non firing engine. Drive mode is only used for starting the engine in the
present setup. The red button on the top of the control panel, when pulled out will en-
able the engine and dynamometer to operate. When pushed in, the engine’s ignition
and the absorbing power of the dynamometer is cut off. On the control panel, there is
a torque indicator which displays a voltage which is read from the load cell on the dy-
namometer. The gauge must be plugged into the load cell on the dynamometer. The
torque indicator always reads negative since the load cell is on the tension side of the
dynamometer. The load cell reads a voltage which must be multiplied by a calibrated
conversion factor, which is currently 60 ft*lbs/volt.

The dynamometer requires water as a cooling medium to flow through it. Safe-
ty features require water for the dynamometer to operate. Thermocouples are placed
in the inlet and outlet water lines which read the water temperature. A thermocouple
box on the bottom of the control panel stand must be plugged into the thermocouples
in order to view the inlet and outlet water temperatures. The dynamometer does not
automatically turn off when it gets too hot so the water temperature must be watched
closely.

Oil and Piston Ring Ion Detection System

The oil detection system transports the ion contaminated oil to the gamma de-
tector in a manner which makes the detector most susceptible to gamma ray intercep-
tion. The stock Briggs and Stratton engine contains no oil pumping system, the en-
gine components are oiled by the means of an arm connected to the crank which
splashes oil up on the components. An external oil pump is used to move the oil
around the oil detection system. The oil, which is heated in the engine, must be
cooled before surrounding the germanium detector. Warming of the detector would
first lead to an large increase in system noise resulting in increased dead time, then
the safety features would turn the detector off. The oil is cooled by a liquid to liquid

103

heat exchanger using water as the cooling medium. A Oil sheath was designed to slip
over the end Of the detector and cover the entire ion detecting chamber. The Oil enters
one side Of the Oil sheath through a mixing board and exits out the bottom Opposite
side. This design of totally encapsulating the detector with the contaminated Oil will
give the best rate of gamma rays interception by the detector. Flexible tubing was
used to transport the Oil between the components because some relative movement
is needed between the engine and other Oil components since the dynamometer rolls
in and out.

SINGLE CYLINDER ENGINE

 

 

 

 

 

 

  
 
 

 

SET AT 6 BAR\
WATER OUT
COILED TUBING AROUND "
DETECTING ELEMENT
TEMPERATURE _>
GAUGE
WATER 1” SHELL AND TUBE
3 HEAT EXCHANGER
PRES SURE
GAUGE

Figure 48 Oil Detection System

The total system Oil capacity is approximately 4.4 quarts. The recommended
time between Oil changes for the engine is 50 hours Of run time. Using a crude method
Of proportions, the Oil should be changed every 126 hours. The ideal case is not to
have to change the oil at all since an Oil change modifies the activity of the Oil.

TO detect ions exiting through the block, the germanium detector has to be
placed close to the block. An engine mounted stand was designed to hold the detec-
tor secure close to the block where the activated portion Of the piston ring was locat-

104

ed. The stand mounts to the head bolts and lower engine mount bolts. Gamma ray
detection at the engine block must be done while the engine has stopped and the pis-
ton is at top dead center. Positioning of the activated portion of the piston ring and
detector is critical.

Exhaust Dilution System

In order to meet safety regulations governing the amount Of radioactive ions
being exhausted to the atmosphere from the engine, an exhaust dilution system is re-
quired. The system has to dilute the contaminated exhaust and send it to a location
where winds would not have a tendency to blow the contaminant into the building.
The system had to dilute the exhaust so that the concentration Of radioactive ions
was well inside the limit of safety. The system had to also endure the weather condi-
tions since the system operated outdoors. The radioactive ion to be implanted in the

piston rings is 54Mn. 54Mn has a branch ratio of 99.98% , a half life Of 0.878 years

and a gamma ray energy Of 834.83 KeV. The engine has a displacement Of 392 cm3,

an assumed volumetric efficiency Of 70% and an operating speed of 3300 rpm. The ion
concentration in the exhaust was calculated for worst case as follows:

54Mn Probability of decay = 71. = 1n 2 / half life = 1.5125 x 10'6 /sec.

Volumetric flow rate Of engine = displacement/2 * rpm * pumping efficiency
= (392 cm3 /2) (3400 rev/min) (0.7) = 466480 cm3 /min.
= 7774.7 cm3 / sec.

Maximum expected activity of piston ring = 10 x 10'6 Ci (3.7 x 1010 dps / Ci)
= 3.7 x 105 dps.

Number Of ions of 54Mn associated with 3.7 x 105 dps = activity [probability Of decay.
= 3.7 x 105 dps/ 1.5125 x 10'6 lsec.
= 2.446 x 1011 ions.

Maximum ion wear rate Of engine = number Of ions / time to wear away all ions.

105

= 2.446 x 1011 ions / 200 hours (min time).
= 3.397 x 105 ions/sec.

Concentration of ions in exhaust = wear rate / mass flow rate Of engine.

= 3.397 x 105 ions/sec / 7774.7 cm3 /sec

= 43.7 ions /cm3 .

Activity of ion in exhaust = probability of decay * concentration Of ion in exhaust.
= (1.5125 x 10'6 / sec )*( 43.7 ions / cm3 )*(2.7 x 10'5 trCi / dps)
= 1.78 x10'9 llCi/cm3.

Hence the calculated undiluted ion concentration in the exhaust is 1.78 x 10'9

uCi I cm3. For 54Mn ion in an unrestricted area, the Office of Chemical and Biological
Safety (ORCBS) - "Standards for Protection Against Radiation" maximum permissi-
ble concentration form 10CFR20 states a maximum permissible concentration of 1 x
10'9 llCi / cc of air. SO a dilution system is needed in this worst case scenario. The
dilution is accomplished by adding more air to the exhaust. A positive displacement
blower was chosen since the flow rate is known for a given blower speed. Using vari-

ous blower flow rates in place of the engine’s volumetric flow rate, blower flow rates
affecting concentration can be found.

Table 15 Exhaust Activity Verses Blower Flow Rate.

ACTIVITY ( “Ci [cc of air) FACIUR OF SAFETY CFM REQUIRED

1.78 x 10'9 0.59 17.4
1.00 x 10'9 1.00 29.4
1.00 x 10'10 10.00 294.0
1.00 x 10'11 100.00 2940.0

1.00 x 10'12 1000.00 29400.0

106

The exhaust dilution system contains a blower, exhaust stack and ion detector
in the stack to monitor activity level. The diluted stack must be attached to the Side
Of the building extending 15 feet from the roof. This 15 feet was an estimated height
for a given building height and flow rate out of the stack from the MSU Building and
Planning Office. The exhaust stack is an aluminum four inch diameter tube, with all
joints sealed. The flexible connection between the movable dynamometer and the
permanent stack is made Of heat resistant rubber material. The blower and blower
motor assembly is mounted on wheels in order to be mobile.

A ion detection monitor is placed in the stack to show if ion concentrations
were getting too high, from possible mechanical failure. The monitor shows a count
per minute reading which can be related to the activity.

EXHAUST STACK

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ENGINE E :-:-= E FILTER AND DETECTOR
...__-L_.
I
DYNAMOMETER l’ ; I
l BLOWER] ‘

 

 

 

 

 

CONTROLLER

Figure 49 Exhaust Dilution System

The Operation Of the wearing ion detection in a running engine will occur at a later
date.