Front Cover
 Table of Contents
 Oxygen transport through high-purity,...
 A nondestructive depth profile...
 ESD and surface analysis studies...
 Appendix: Description of the ultrahigh...
 Biographical sketch
 Back Cover

Title: Surface studies related to the development of a hyperthermal oxygen atom beam generator
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00090190/00001
 Material Information
Title: Surface studies related to the development of a hyperthermal oxygen atom beam generator
Physical Description: vi, 89 leaves : ill. ; 29 cm.
Language: English
Creator: Davidson, Mark Rogers, 1962-
Publisher: Mark Davidson Rogers
Place of Publication: Gainesville, Fla.
Publication Date: 1990
Copyright Date: 1990
Subject: Surfaces technology -- Analysis   ( lcsh )
Oxygen   ( lcsh )
Diffusion   ( lcsh )
Silver   ( lcsh )
Chemical Engineering thesis Ph. D   ( lcsh )
Dissertations, Academic -- Chemical Engineering -- UF   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Statement of Responsibility: by Mark Rogers Davidson.
Thesis: Thesis (Ph. D.)--University of Florida, 1990.
Bibliography: Includes bibliographical references (leaves 85-89).
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00090190
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 001589732
oclc - 23067831
notis - AHL3707

Table of Contents
    Front Cover
        Page i
        Page ii
        Page iii
    Table of Contents
        Page iv
        Page v
        Page vi
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
    Oxygen transport through high-purity, large-grain silver
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
    A nondestructive depth profile study of oxygen-exposed large grain silver using ARAES and ISS
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
    ESD and surface analysis studies of silver / 0.5% zirconium alloy
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
    Appendix: Description of the ultrahigh vacuum surface facility
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
    Biographical sketch
        Page 90
        Page 91
        Page 92
    Back Cover
        Page 93
Full Text








The author wishes to thank his father and mother, who

instilled in him the desire to know how things work, and the

confidence to tackle the sometimes difficult and frustrating

task of learning how and why the things around us behave the

way they do. The author has been ready to give up on a

problem many times when he remembered that "F = ma, you can't

push on a rope, and the rest is just algebra," as his father

told him as he grew up. The author also wishes to thank his

wife Jan, who has put up with 6 years of being a graduate

student's wife and all of the many inconveniences which that

entails. Without her support, these years of education would

not have been possible. The author's appreciation goes to

both his freshman chemistry professor, Dr. Sam Colgate, and

Dr. H.A. Laitinen, in whose laboratory he worked as an

undergraduate. These two men, in particular, first showed him

how exciting research can be. The author wishes to thank his

advisor, Professor Gar B. Hoflund, for having a seemingly

infinite amount of patience and helpful guidance during the

learning process. Lastly, the author wishes to thank Dr. R.A.

Outlaw, who provided the motivation for this work and a great

deal of guidance in both the experimental aspects of this work


and in the interpretation of the data. He also thanks Dr.

Outlaw for performing the diffusivity and permeability

experiments on his permeability apparatus.



ACKNOWLEDGMENTS ...... ..... ........................... ii

ABSTRACT ............................................ iv

INTRODUCTION ........................................... 1

Motivation ........................................ 1
Past Research ..................................... 5

SILVER .......................................... 14

Introduction ...................................... 14
Permeation Analysis ............................. 15
Experimental ...................................... 18
Results and Discussion .......................... 24
Summary ........................................ 34


Introduction ...................................... 36
Experimental ...................................... 37
Results and Discussion .......................... 42
Conclusions ....................................... 54

ZIRCONIUM ALLOY ................................... 55

Introduction ...................................... 55
Experimental .................................... 58
Results and Discussion .......................... 67
Summary and Conclusions ......................... 79

FACILITY ........................................ 80

REFERENCES ............................................. 85

BIOGRAPHICAL SKETCH .................................. 90

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




May, 1990

Chairman: Gar B. Hoflund
Major Department: Chemical Engineering

Silver and silver alloys are important catalytic

materials for several commercially important reactions such

as ethylene epoxidation. In addition, silver and its alloys

have been used as oxygen permeation membranes in the

development of an hyperthermal neutral oxygen atom gun. The

permeability and diffusivity of oxygen in large-grain, high-

purity silver was measured. Both the diffusivity and the

permeability were found to be significantly lower than

previous investigators have reported. It is believed that

grain boundary transport played an important role in previous

studies. The silver used in this study had a very low grain

boundary density, so the transport rates were slower. This

hypothesis is supported by Auger electron spectroscopy (AES)

line scans of heated, oxygen-saturated silver which show the

oxygen accumulating near the grain boundaries. The apparent


activation energy for diffusion changed near 630 C, possibly

indicating some oxygen trapping is occurring. The surface

analysis techniques of angle resolved Auger electron

spectroscopy (ARAES), X-ray photoelectron spectroscopy (XPS),

ion scattering spectroscopy (ISS), and electron stimulated

desorption (ESD) have been used to study the chemical

interaction and mechanisms of diffusion of oxygen in pure

silver and 0.5% zirconium silver alloys. It has been shown

that a surface relaxation in the first several silver layers

occurs which strongly affects the oxygen concentration

profile. Oxygen and zirconium diffusion constants and neutral

and positive ion ESD cross sections from the Ag/Zr alloy have

also been obtained. A variation of ESD cross section with

electron beam incidence angle has been observed and, relative

changes with cross section have been measured for several

incidence angles.



This work was motivated by two seemingly unrelated

important industrial problems. The first involves the

degradation of materials in orbiting spacecraft due to

collision with neutral oxygen atoms. The main objective of

this research is to help in the development of a research tool

which could be used to study the degradation of proposed

spacecraft materials in ground-based laboratories. The second

application of the studies presented in this work has to do

with the optimization of the catalytic oxidation of ethylene

to form ethylene epoxide.

The atmosphere at typical orbital altitudes (200 to 1000

km) consists largely of neutral atomic oxygen atoms. This

combined with the large velocity of the orbiting vehicle (=8

km-s' ) results in a flux of hyperthermal neutral oxygen atoms

(E=5 eV, Flux = 1 x 1015 atoms-cm -s ) colliding with the

leading spacecraft surfaces. This highly energetic collision

of the atomic oxygen has been a major cause of materials

degradation during space shuttle missions [1]. A good example

of the problems which this can create was recently seen during

the retrieval of the long duration exposure facility satellite


(LDEFS). The planned duration of the satellite was 1 year,

after which it would be retrieved and the results of the on-

board experiments analyzed. Due to the space shuttle

Challenger accident, the retrieval was delayed 3 years. One

of the experimental panels on board consisted of a panel of

layered polymer films. When the satellite was grasped by the

retrieval arm, the panel turned to dust and the results of the

experiment were lost. It is believed that the disastrous

degradation of the polymer panel was caused by long-term

exposure to the hyperthermal oxygen atom flux. Possible

problems can also be expected with composite materials

proposed for use in the planned space station, exterior

coatings of the Hubble Space Telescope, and materials for use

in proposed orbiting laser communication systems. Therefore,

study of the proposed materials activity towards hyperthermal

oxygen in ground-based laboratories is required.

In order to perform these experiments, a source of

neutral oxygen atoms with an energy and flux approximating

that found in orbit is required. Current schemes for

producing neutral oxygen atom fluxes include radio frequency

(rf) discharge, thermal dissociation, and electron impact [2].

In order to accelerate these beams to appropriate energies,

techniques such as nozzle expansion or charge neutralization

are used. These methods generally have several problems which

include low flux density, low mean energy, wide energy

distribution, and production of excited state atoms. In


addition, these techniques are not generally compatible with

the pressure required for a detailed surface analysis. At a

typical background pressure for these methods of z1 x 10'6

Torr, the surface which is being studied is covered with

background gases in approximately 1 second (assuming a

sticking coefficient of nearly 1). Therefore, studies of

oxygen-atom-treated surfaces in such systems have been dosed

with as much of the background gas (typically CO, CO2, 02, and

Nz) as the dosing gas. Clearly, any device which is used to

produce hyperthermal oxygen atoms must be compatible with

vacuum levels in the 1011 Torr range.

Efforts toward producing such a hyperthermal oxygen atom

gun (HOAG) have relied on two phenomena. The first is the

unusually high permeability of oxygen through silver [1,3,4].

The second is the fact that electron stimulated desorption

(ESD) of adsorbates from metal surfaces yields predominantly

neutrals with energies in the 3 to 10 eV range [5-26]. The

proposed scheme for a continuous HOAG is to pressurize one

side of a hot silver membrane with oxygen while the other side

of the membrane is exposed to the UHV chamber. The

temperature of the membrane is kept below that at which the

oxygen will desorb on the vacuum side of the membrane. It

has been shown that oxygen diffuses atomically through the

silver [27,28]. The silver surface on the vacuum side is then

bombarded with electrons, which results in ESD of neutral

atomic oxygen with an energy of ;5 eV. A schematic diagram

100 Torr 02


o 0 (atomic)
0 55 eV


Ag membrane
T,500 t

Figure 1. A schematic of the proposed HOAG


of the HOAG is shown in figure 1. Preliminary transient

experiments have shown the scheme to be feasible [1].

The second major motivation for studying oxygen

interaction with silver and silver alloys is the use of silver

as an ethylene epoxidation catalyst. Ethylene epoxide is a

commodity chemical which is the precursor to many commercially

important chemicals including ethylene glycol (antifreeze).

The mechanisms of catalytic ethylene epoxidation by silver

have been extensively studied but are still poorly understood

[29-69]. Because the volume of ethylene epoxide produced is

very large, a small improvement in the catalytic activity or

decay rates of the catalyst results in relatively large

economic benefits. A great deal of traditional trial and

error catalyst optimization has been performed on these

catalysts and a fundamental understanding of the mechanisms

involved in the reaction and decay in activity of the catalyst

are now necessary to further improve the efficiency of the

catalyst. An understanding of the adsorption of oxygen on the

silver surface and its dissolution into the bulk is a first

step in this goal of an improved catalyst design based on an

understanding of the fundamentals of the process.

Past Research

As previously mentioned, the basic concept of the HOAG

has been tested in a transient experiment [1]. In this

experiment, a silver wire was saturated with 10 Torr of oxygen

at 500 C for 1 hr. The system was then pumped down to =0'11


Torr. The atomic oxygen flux was monitored by a quadrupole

mass spectrometer (QMS) which was operated in the appearance

potential mode in order to remove any contribution to the 16

AMU signal from fractionation products of CO, CO2, or other

oxygen-containing background gas species. The cross section

for ESD was found to be 7 x 10-19 cm2 and the flux for a

surface coverage of 0.5 was found to be 1.1 x 1012 cm-2-s1. In

addition, a diffusivity of 2.64 x 10-6 cm2-S-1 was measured at

a temperature of 5000C.

Silver Oxygen Interaction

The majority of the previous research has been conducted

on single crystals [37-63] and most of the studies have been

performed on the (110) surface since the sticking coefficient

( 10-3) on that crystal face is about 1000 times greater than

on either the (100) or (111) faces [42]. It was found in

these studies that molecular oxygen exists on the surface at

temperatures below -1030C but that it dissociates and forms

adsorbed atomic oxygen above that temperature. Subsurface

atomic oxygen is believed to form above 1500C and is stable to

about 3270C. Based on low energy electron diffraction

measurements, the maximum coverage is believed to be half a

monolayer. In addition, a subsurface state has been proposed

for the (110) surface, which also desorbs above 3270C.

Polycrystalline studies of the oxygen/silver system are

much less common due to the complexity of the results [23,27,

28,64-69]. Polycrystalline samples may show characteristics


of all of the single crystal faces. In addition, defects and

grain boundaries can play an important role in the adsorption

and diffusion of gasses through solids. The polycrystalline

silver studies so far have shown four general oxygen species

present after varying exposures. Evidence for a molecular

oxygen species has been observed, but this species is not

stable at room temperature under vacuum and desorbs quickly.

Like the single crystal studies, atomic oxygen was found

adsorbed on the surface and dissolved in the bulk. In

addition, a strongly-bound oxygen species is formed and it is

postulated that this species is associated with grain

boundaries and surface imperfections [64].

It is interesting to note that in all of the studies, no

molecular oxygen was observed at temperatures above room

temperature. X-ray photoelectron spectroscopy (XPS) shows

that even though the oxygen is present as atomic oxygen, no

silver oxide is formed [27,28]. This indicates that the

oxygen forms a solid solution in the silver.

Two reported studies of the surface structure of silver

have been found. Ning and coworkers [70] have performed

calculations based on the embedded atom method to calculate

surface relaxations in a number of materials and crystal

planes including the (100),(110), and (111) planes of silver.

They show that the (110) and (111) planes undergo an

oscillatory relaxation in which the first two atomic layers

are contracted, the second two layers are expanded and the


third two layers are contracted. They also provide some

experimental results which confirm their findings. Kuk and

Feldman [71] have performed high-energy ion scattering/

channelling studies of a silver (110) surface which can give

a relatively direct measure of the surface relaxations. They

find, in agreement with Moghadam and Stevenson [69] that the

surface undergoes an oscillatory relaxation in which the first

two layers are contracted and the second two are expanded

relative to the bulk layer spacing.

Electron Stimulated Desorption

Electron stimulated desorption has been reviewed by Madey

and Yates[5] and more recently by Hoflund[6]. In addition,

a large body of research exists which involves studies of

systems of various adsorbates on several metal and insulator

surfaces [1-23]. Several mechanisms have been proposed for

ESD, two of which are predominantly accepted. The first is

known as the Menzel-Gomer-Redhead (MGR) model. The MGR

process is depicted in Figure 2. The adsorbate vibrates at

the ground vibrational state near the bottom of the potential

well. The ejection of an electron from a bonding orbital

results in the formation of an antibonding orbital which may

include either a neutral or ionized particle being ejected.

The second mechanism is known as the Knotek-Feibelman (KF)

mechanism and differs from the MGR mechanism in the way in

which the repulsive potential is created. This mechanism is

depicted in Figure 3 for

M*+ A


Figure 2. Potential energy diagram of the MGR model for ESD
of O ions from TiO2


desorption of 0+ from TiOz. A core hole in the metal atom

created which leaves the molecule in an excited state. One

mechanism for de-excitation is an interatomic Auger

transition, which removes two or three valence electrons from

the oxygen atom. This creates a positively charged oxygen

atom next to a positively charged metal atom. The oxygen ion

is then ejected by the coulombic force exerted on it. In both

the MGR and KF mechanisms, ions desorb and most are

neutralized as they leave the surface. Both theoretical and

experimental evidence exists which shows that in most cases

the ratio of neutrals to ions ranges from 100 to 1000.

Some recent studies of physisorbed adsorbates have shown

that nearly all of the desorbed particles are neutral. In

these studies, no ions could be detected [20-22,24-26].

Another mechanism has been proposed by Antoniewicz [26] to

account for these findings. In this mechanism, a valence

electron is ejected from the adsorbate. The potential well

for the ion in this case is deeper than that for the neutral

adsorbate due possibly to the attraction of the image charge

formed in the bulk metal. This attraction causes the ion to

move initially towards the surface. As the ion moves closer

to the surface, the probability for neutralization increases.

When the ion is neutralized, it is no longer in its

equilibrium position, and, depending on the position of the

ion when it is

e -
e e



T4+ 2-

Figure 3. Schematic potential energy diagram of the KF model
of the ESD mechanism


neutralized, the atom formed may be high enough on the

repulsive portion of the potential energy curve that it can

escape the surface and desorb as a neutral atom. In this

manner, nearly all of the atoms desorbed are neutral, since

desorption depends on neutralization. This mechanism is

depicted schematically in Figure 4.

M+ A+

M*+ A


Figure 4. Schematic potential diagram of the Atoniewicz model
for the mechanism for ESD of loosely bound adsorbates.



Past measurements of the diffusivity of oxygen in Ag

have, in all cases, ignored the effects of grain boundary

density and material purity [3,4,72-75]. Since both are

important concerns in determining the actual transport

mechanism, it is essential that the effects of each be

studied. If grain boundary diffusion is a significant part

of the transport mechanism for the diffusion of oxygen in Ag,

the grain boundary density would have a major effect.

Experiments with larger grain size would reduce the number of

transport paths and would provide smaller permeation rates

compared with past research (which most likely had small

grains and therefore higher grain boundary densities).

Furthermore, impurities (both species and amounts) could cause

oxygen trapping effects that would result in diffusivity

variations. Experiments with high-purity Ag would minimize

such effects.

This chapter presents results of permeation measurements

of oxygen through high-purity, large-grain Ag membranes


conducted under ultrahigh vacuum (UHV) conditions. Auger

electron spectroscopy (AES) results of oxygen charged Ag are

also presented.

Permeation Analysis

The permeation of planar membranes is generally described

in Crank[76]. If a diffusion coefficient D independent of

concentration C is assumed, then

C 2C(x,t)
--(x,t) = D ( --- ) (1)
t x2

where x is the variable distance through the membrane and t

is the time.

The boundary conditions applied to this equation can be

determined from the permeation process, which includes (a)

adsorption of the oxygen molecules on the high-pressure side;

(b) dissociation of the molecules to atoms; (c)incorporation

of the atoms into the lattice, grain boundaries, or defects;

(d) random walk migration through the membrane by means of

(c); (e) emergence from a solid solution at the vacuum

interface and entry into an adsorbed state; (f) reassociation

to molecules; and (g) desorption. Normally the rate

controlling step is that of (d).

Following Crank [76], Eq. (1) may be solved using the

following boundary conditions:

C(x,t=0) = g(x) (2)

C(x=0,t) = S1/2 = (3)

C(x=d,t) = SP/2 C (4)

where C is concentration, g(x) is and arbitrary function, d

is the membrane thickness, S is the solubility, Po is the

high-side pressure, and Pd is the pressure at the UHV

interface, which is usually taken to be zero at the beginning

of the experiment. Equations (3) and (4) describe the

concentration of the gaseous species in the metal assuming

that Sievert's law is obeyed. If the upstream side of the

membrane is pressurized, then the boundary conditions become

C1 = SP12, C2 = 0 and the gas flux J at steady state is

J(x=d,t->o) = DSP2/d = KP2/d = (5)

where the permeability K = DS

The steady-state flux Jss through the membrane is related

to the steady-state pressure Pss in the measurement chamber by

Jss = Pssf/AkT (6)
where f is the limiting conductance to the main ion pump, k

is Boltzmann's constant, T is the gas temperature, and A is

the membrane area. The permeability then is

K = Pssfd/P1/2AkT. (7)

In order to determine the diffusivity following the

establishment of a steady-state permeation rate, the upstream

of high-pressure side can be reduced to zero such that CI=C2=0,

and the flux becomes

J(x=d,t) = --- [-22 (-1)" exp( -Dn2it/d2)] (8)
d 1

The time varying flux AJ(t) and pressure P(t) within the

measurement system are related by

V----- = AkTJ(t) f[P(t)-Pp(t)] (9)

where V is the volume of the measurement chamber and Pp(t) is

the time varying pressure within the main pump. The

conductance f is chosen such that P(t) > P (t), and the flux


V dP(t) f
J(t) = -------- + -- P(t) (10)
AkT dt AkT

A comparison of the two terms in Eq. (10) indicates that a

worse case condition occurs at high temperatures when dp(t)/dt

is greatest, but even at the highest temperature encountered

the ratio of the first term to the second term is less than

z0.02. Therefore, the flux can be simplified to

J(t) (f/AkT)P(t). (11)

The ratio of Eq. (8) to Eq. (5) combined with Eq. (11) gives

P(t) 0
= -2Z (-1)" exp( -Dn272t/d2) (12)
Pss 1
where D is the diffusivity. Experimentally, the time for

P(t)/P, to decay to some predetermined value can be measured,

and then D can be found by numerical iteration of (12).


However, because of the rapid decay of the exponential

function, no significant error is introduced by considering

only the first terms of the summation.


Membrane Preparation

The permeation membranes used in this work (99.999 + %

vacuum-melted Ag) were spark machined into a disk geometry

that fits between two mini-flanges to separate the UHV

detection volume from the oxygen supply volume. The membrane

surfaces were prepared by polishing to a 30im grit finish,

followed by degreasing with laboratory detergent in an

ultrasonic cleaner, rinsing in deionized H20 in an ultrasonic

cleaner, chemically cleaning in 4 parts HNO3 to 1 part

deionized water, rinsing in deionized water, and drying with

oil-free and filtered N2. The AES analysis of the surface

conducted after the cleaning procedure closely resembles the

spectrum of a sputter cleaned Ag surface. Small contaminant

levels of S, Cl, C, and O were observed but the S, Cl, and C

peaks were removed during the first few permeation runs of

oxygen through the membrane (54], thus providing a clean Ag

sample for experiments.

The grain growth of the Ag membranes was stabilized

during the prerun vacuum degassing before oxygen was admitted

to insure maximum growth of the grains. The grain size

distribution varies from approximate diameters of 1 to 5 mm


with a mean diameter of about 1.1 mm. Grain densities

determined by the linear intercept method were found to be

1.25 grains/mm2. Some twinning also was observed. Details of

these membranes have been reported previously [77].


The UHV permeation system (ultimate pressure less than 2

x 10-11 Torr) employed in this work was equipped with a main

ion pump of 500 liters s'1, a titanium sublimation pump, a

quadrupole mass spectrometer (QMS), and ion gauges for signal

detection. A small calibrated conductance f was used to

separate the measurement chamber from the main ion pump in

order to provide a known conductance limitation and therefore

a known flow rate out of the measurement chamber. Measurement

of the pressure, P(t), then allowed the determination of the

permeation flux J(t). The high-pressure side of the membrane

was also ion pumped to permit appropriate cleanup of the high-

pressure side of the membrane surface and of the oxygen gas

supply line.

The room temperature background gas spectra observed in

the measurement chamber before and after introducing 100 Torr

oxygen (into the high-pressure side of the system) were

identical, thus indicating no leakage through the membrane.

Eventually, however, the 02 peak was observed and became the

dominant signal as the Ag membrane was elevated in



A separate apparatus was used for an AES study of the

large-grain Ag. The apparatus is described in detail in

appendix 1. The AES data were taken using the CMA operated in

the nonretarding mode with the coaxial electron gun operated

at an energy of 3 keV. The CMA was also used to perform ion

scattering spectroscopy (ISS). In ISS a primary beam of 1 keV

He' ions at a flux of < 10 gA/cm2 and a scattering angle of

about 1470 were used. The Ag sample was heated by electron

bombardment using a tungsten filament place behind the sample.


Permeation measurements

After vacuum degassing the membrane at 800 C, oxygen was

admitted to the high-pressure side, equilibrium flow was

established, and the upstream pressure was removed to

determine the pressure decay in the measurement chamber. From

the equilibrium pressure obtained for each run the

permeability is computed from Eq. (7). After the high-

pressure side is evacuated, a linear decay on a In P(t) versus

time plot is predicted by Eq. (12), from which the diffusivity

can be determined. Figure 5 shows representative data for

oxygen breakthrough to steady state (TAg = 550 C, Po = 150 Torr

02) and then pressure decay (TAg = 550 C, Po -> 0 Torr 0)).

After several cycles of this procedure to clean the membrane

PO =150torr02 P 0 -102 lo02


fr 10 8



1 '10 11 12 13 14 15

t. mmin

Figure 5. The 02 pressure rise to steady state and subsequent
pressure decay following pump down of upstream side. Dashed
curve represents normal pump down without evolution of oxygen
from the membrane.





on both faces, data were taken at 250C increments over the

temperature range of 400-8000C and over the high-side pressure

range of 5-250 Torr by monitoring the signals of both the ion

gauge and the QMS. The emission current of each instrument

was maintained at low values (0.4 and 0.5 mA, respectively) to

minimize instrument pumping.

The curvature that appears in Fig. 5 after the linear

decay in the signal is due to a weak incorporation of oxygen

into the stainless steel walls of the vacuum system. When the

high-pressure side is evacuated, a linear decay occurs as

predicted by Eq. (12); but ultimately, as the pressure

descends toward the background level, the data begin to curve

because the charged walls now see an opposite concentration

gradient that causes oxygen to desorb back into the gas phase.

The increase in oxygen gas flux results in an increased

background pressure, but is substantially less than the gas

flux evolving from the membrane immediately after the high-

pressure side is evacuated, so it does not alter the linear

decay necessary to determine the diffusivity. This effect has

also been substantiated by separate experiment. Note the

dashed line in Fig. 5 that represents the decay when the

membrane is uncharged with oxygen and the system has been

backfilled to the same 02 pressure that occurs during


At the higher temperatures studied, the vapor pressure of


the Ag is sufficiently large to cause some material loss from

the membrane. For example, at 8000C, the vapor pressure is

approximately 5 x 10-5 Torr, which results in a sublimation

rate of 0.01 mm/h. This represents a significant mass loss

and required corrections for each experimental run.

The procedure described above was used to study five

different Ag membranes from which consistent and reproducible

results were obtained.

AES experiments

For the AES study an Ag membrane was prepared as

previously described and then charged with oxygen at 100 Torr

for 2 h at 5000C. The sample was then mounted in the UHV

analytical system and sputtered until AES and ISS showed that

the surface was clean. The sample was then heated by electron

bombardment and the temperature was measured using a

thermocouple spot welded to the edge of the Ag sample. Since

the Ag membranes were thin and Ag has a high vapor pressure,

it was difficult to heat the sample evenly above 5000C. Above

this temperature much of the energy input to the sample caused

vaporization of the Ag. Because the energy could not be

evenly distributed in the thin membranes, a large temperature

gradient (>3500C) across the sample was formed. The

temperature at the edge of the sample was established at 600C,

and the canter was near the melting point (9600C). The sample

was then cooled to room temperature and studied with AES.


Results and Discussion

Permeation Measurements

The permeability data are presented in Fig. 6. The

linearity is quite good and was repeated over six separate

runs. Several initial runs were required, however, before the

data became repeatable, presumably because of contaminant

removal by the oxygen. An important feature of the permeation

technique used in this work is that bulk cleanup of the

material studied occurs more readily because very thin

membranes (z0.254 mm) were used. Segregation of the

impurities to grain boundaries and then to the free surface

can occur much faster, thus providing a more rapid cleanup by

oxidation and desorption of reaction products. The mass

spectra observed in the measurement chamber indicated a

nominal fragmentation pattern for 02 and showed even less

contamination than the UHV work of Beavis [78]. The low

levels of CO and CO2 and the absence of other contaminants

reflect a very clean membrane. A comparison with the apparent

permeability determined by Eichenauer and Muller [3] (fig. 6)

shows that the data reported here are lower in magnitude by

a factor of 3.2 but similar in slope. One possible

explanation for the lower magnitude is that the grain boundary

diffusion component was minimized in this work due to the

large grains. Silver is known to be sensitive to impurities

T, C
1000 800700 600 500


I I I 1 I I I I I I


Eichenauer and Miller (Ref. 1)

KEM-5.2xl10 exp [-22 ]

- K=1.01 x1018exp [-2

- 10 torr
- 100 torr
- 150 tor
- 200 torr


75 1.00

1 25
103/T. K1




Figure 6. Arrhenius plot of oxygen permeability of the large-
grain Ag compared to past research.



1 01 3



101 1



restricting grain growth since the segregation of the

impurities to the grain boundaries tend to slow down or stop

intergranular transfer of Ag atoms from smaller to larger

grains. This suggests that more grain boundaries and

therefore many more transport paths are present in less pure

Ag. Unfortunately, there is virtually no information on the

microstructure of the Ag samples used in the earlier studies,

but they are probably small grain samples since the Ag was not

high purity. The purest Ag studied by past researchers was

reported by Eichenauer and Muller (99.99 % vacuum melted at an

ultimate pressure of =10-3 Torr).

The permeability equation determined from a least-squares

fit of the data presented here is

K = 1.01 x 1018 exp(-21870/RT)cm'i's" (13)

The activation energy associated with these data compares well

to that of Eichenauer and Muller (Ea = 22 860 cal/mol). It is

important to note that all permeation data with upstream

pressures of Po = 10, 100, 150, and 200 Torr, plotted on the

same straight line (fig. 6). This agrees with Sievert's Law

and the assumption that the 0 atoms move through the Ag in the

atomic state.

The diffusivity data showed some variation in the first

few runs as the membrane became more and more contaminant

free, but ultimately became very repeatable. As shown in Fig.

7, a break in linearity occurs in the diffusivity at =6300C.

1000 800 700 600 500

I I I 1 I I I I I I I



D- Z.96 x 103 'exp- 11-50

5 x 10 5






D= 3.197 0xp[- 0]


1 00

103 T. K.

.1 50



Figure 7. Arrhenius plot of oxygen diffusivity in Ag compared
to past research. Note the diffusivity break in linearity at
T = 630 C.

Eichenauer and Miller (Ref. 1)

D 3.66 x 10-3 Qxp 11000




A break is commonly seen in systems where there is a

transition from grain boundary diffusion to lattice diffusion

with increasing temperature, but in all such cases the

activation energy for lattice diffusion is greater. In this

work the higher temperature regime has a lower activation

energy. The equations determined from the diffusivity plots


D = 2.96 x 10-3 exp(-ll 050/RT)cm2s ,

6300C < T < 8000C, (14a)

D = 3.2 x 10-2 exp(-15 330/RT)cm2s ,

4000C < T < 6300C. (14b)

In the higher temperature range, the value of the activation

energy (Ea = 11 050 cal/mol) is almost exactly that obtained

by Eichenauer and Muller (Ea = 11 000 cal/mol). At

temperatures below 6300C, however, the value of the activation

energy (Ea = 15 330 cal/mol) is considerably higher.

In order to examine what occurs in the higher temperature

regime, the desorption of oxygen from the membrane was

studied. Figure 8 shows the desorption spectrum of 02 from

the membrane (vacuum on both sides) observed during the

increase in membrane temperature to 8000C. Note that at 630C

sufficient thermal energy exists so that the onset of vacuum

desorption of oxygen from the membrane occurs. This indicates

that the oxygen atoms are probably migrating with an increased

mobility above this temperature. This increased mobility can



-- -4 deg/min



L 6




f Onset of oxygen desorption

450 500 550 600 650 700 750 800 850

T, C

Figure 8. The 02 evolution (slow temperature increase) from
Ag membrane detected downstream when the upstream pressure is
at p = 10'8 Torr.


be correlated with the diffusivity data represented by Eq.

(14a). In terms of diffusivity, whatever differences that

exist between the microstructure and the purity of our Ag

compared to that of Eichenauer and Muller must have been

overcome by the thermal energy, since the diffusivity values

are the same. In the low-temperature regime, the larger

activation energy and the lower mobility indicate a different

control of oxygen atom transport, possibly trapping. Mitchell

et al.[79] have examined the trapping effects of hydrogen in

copper due to the addition of substitutional impurities Er,

Zr, and Ti. They found no change in the permeability of to

4 at. % of Zr, but they did observe a significant decrease in

the magnitude of the effective diffusivity (Deff represents the

effect of the trap in contrast to the D in pure Cu).

Furthermore, the slope of the Deff (for all impurities ) in an

Arrhenius plot was also significantly greater than for the D

in the pure Cu and represents an increased activation energy

because of the traps. According to the analysis of Mitchell

et al., the observed activation energy E is the sum of the

trap energy ET and the activation energy Ea. The trap energy

ET for this work is 4280 cal/mol. Johnson and Lin have

observed a diffusivity break similar to that observed in this

work for the system H/Fe [80]. They attribute the higher

activation energy in the lower temperature range to the higher

efficiency of traps, which, in their case, were vacancies.


In this study of high-purity Ag, vacancies and interstitial

impurities are more likely to be trapping centers than are

substitutional impurities. The vacancy concentration in Ag (

Ef z 25.3 kcal/mol) is about 4 x 10i1 cm-3 at 8000C, but these

vacancies could be filled by oxygen atoms at elevated

temperature that would then stabilize them in the low-

temperature regime.

Lattice diffusion is the dominant mechanism in this work

because of the very large grains, but grain-boundary self-

diffusion is well known to occur in Ag and predominates over

lattice diffusion below 7500C for small grain material [81].

Therefore, the diffusion of oxygen in the grain boundaries

might also be quite significant since oxygen is a smaller atom

than Ag and has less chemical affinity for Ag than Ag has for

itself. For these reasons, grain boundary diffusion of oxygen

was likely to have been a significant component in past

research where the Ag grain size was small.

AES experiments

Following the heat treatment of the Ag sample, AES line

scans of the 0 and Ag peaks were taken at room temperature and

found to be quite repeatable (see Fig. 9). These scans were

obtained by slowly moving the 1.0 gm primary electron beam

from the center of the sample to the edge of the sample.

Several interesting features are revealed by this reproducible

line scan. First, essentially no oxygen is present near the

Ag (MNN) Peak

0 (KLL) Peak

T = 600'C

0 2 4 6 8 10

Distance, mm

Figure 9. The AES line scans from the sample edge to the
center of the Ag membrane. The insert in the center is a
trace of the Ag membrane grain boundaries.

T 960C


region where the sample was near the melting point (10 mm)

from the edge of the disk). This implies that the oxygen

either desorbed or moved to a cooler portion of the sample.

The region between 7 and 9 mm from the edge of the sample

contains the largest amount of surface oxygen. This region

was hot enough for the oxyegn to be driven to the surface from

the bulk Ag but not so hot that it desorbed or migrated away

from the region. An AES sepctrum taken from the region where

the oxygen signal is a maximum shows only a very small amount

of oxygen (as indicated by the O/Ag ratio of less than 6%).

Moving the electron beam from 7 mm to the edge of the sample

results in a decreasing oxygen signal. This is due to a

decreasing temperature across the sample which results in less

oxygen migrating to the surface. The structure that is

particularly apparent in the Ag peak height line scan is due

to the presence of grains and grain boundaries.

Microstructure measurements show an average grain size of

about 1.25 mm, which corresponds well with the average

distance between the observed Ag minima and 0 maxima. The

insert in Fig. 9 is an outline of the grain boundaries from a

micrograph of the Ag. Note that the Ag peaks and 0 valleys

are about the same distance apart as the average grain size.

Past research in the surface properties for the O/Ag

system shows a very small sticking coefficient and virtually

no adsorption on the (100) and (111) planes and only about 5%

on the (110) plane [39,41]. This indicates little oxygen

affinity for adsorption on single crystal Ag and suggests a

barrier to oxygen dissolution into the bulk. In this work we

have observed a very small oxygen AES signal in the middle of

the grains of oxygen-charged Ag but have observed a much

larger signal at the grain boundaries. Further, as the sample

was heated to above 6000C, intragranular oxygen signals became

more detectable, but the signals at the grain boundaries

dramatically increased, suggesting rapid segregation to the

grain boundaries and ultimately to the free surface.


We have determined the permeability and diffusivity of

oxygen through high-purity large-grain Ag. The magnitude of

the permeability found in this work compared with past

research is lower by a factor of 3.2. The diffusivity curve

shows two linear regimes. The high-temperature regime agrees

very well with past research, but the low-temperature regime

has a higher activation energy, probably due to trapping. The

AES experiments indicate there is a much smaller intragranular

oxygen signal intensity compared to that observed at the grain

boundaries and that this difference is further accentuated

with increasing temperature. Although lattice diffusion is

the dominant mechanism in this work, the permeability

comparisons, the AES results, and the implications of past

research strongly indicate the importance of grain boundaries


in the O/Ag system. The study of oxygen through fine-grained

Ag (Ag 0.5% Zr) and Ag single crystals will ultimately provide

more definitive answers to this question.



In the previous chapter, evidence that the transport

properties of oxygen through polycrystalline silver are

influenced heavily by defects and grain boundaries is given.

In addition, some knowledge of the position of the oxygen in

relation to the surface is desired in order understand any

possible mechanisms for transport and ESD of oxygen.

Several studies of single crystal Ag(llO) have indicated

that a subsurface oxygen species might be formed under certain

conditions. Segeth, Wijngaard, and Sawatzky [82] recently

found evidence using angle resolved ultraviolet photoelectron

spectroscopy (UPS) that a surface relaxation can occur on

oxygen exposed Ag(llO). In addition, Kuk and Feldman [83]

have found using high energy ion scattering spectroscopy

(HEISS) that Ag(llO) undergoes a multilayer surface relaxation

the first two layers are contracted and the second two layers

are expanded relative to the bulk. Backx et. al. [49] have

suggested that a subsurface oxygen species is formed by oxygen

exposure of Ag(llO) which can be transferred to the surface


by heating to 2000C. It is likely that the subsurface oxygen

which has been observed is located in the subsurface region in

which the lattice spacing is expanded relative to the bulk.

These studies indicate that at least some of the surfaces

exposed in a polycrystalline sample might undergo relaxation

and that their interaction with oxygen may not be a

straightforward surface adsorption and bulk diffusion. In

this study large-grain, high-purity silver was exposed to

oxygen and analyzed using ion scattering spectroscopy (ISS)

and angle resolved Auger electron spectroscopy (ARAES) [84].

By combining these techniques, a non-destructive depth profile

of the near surface region can be obtained, and that

information used to elucidate a schematic oxygen potential

distribution as a function of depth for the first few atomic



The sample used in this study was vacuum melted high

purity (99.999%) silver with an average grain size of about 1

mm. The sample was solvent cleaned in a manner similar to

that discussed previously. Residual contaminants were removed

in the analysis chamber by repeated heating to 5000C and Ar

sputtering. Oxygen doses were performed in a preparation

chamber attached to the main analysis chamber. Doses were

performed at 75, 150, 200, and 2500C in 75 Torr of oxygen.

The sample heater was designed with the hot filament protected

from oxygen exposure so that the high pressure doses could be


Energy analysis was performed using a Perkin Elmer PHI

model 25-270AR double pass cylindrical mirror analyzer

equipped with an angle-resolving aperture. ISS and ARAES data

were taken in the non-retarding mode. ISS data were generated

by using 1 keV He* ions and a defocused ion beam in order to

minimize damage. Scattering angles of 1440 and 640 were used

for ISS. Standard backscattered (1440) ISS data were obtained

for oxygen quantification, but backscattered ISS proved to be

insensitive to trace quantities of low molecular weight

impurities so forward scattered (640) was used to determine

sample cleanliness (Fig. 10). From the data presented in

figure 10, it is estimated that the sensitivity of forward

scattered ISS to surface oxygen is about 1000 times higher

than that for backscattered ISS. It should be noted that

although a silver surface can appear clean by backscattered

ISS, the forward scattered ISS reveals several surface

contaminants. Although these contaminants are present at low

levels, the behavior of silver towards oxygen is drastically

changed by very small levels of contaminants since most

contaminants bind oxygen much more strongly than does the


AES data were obtained by using the coaxial electron gun

operated at 3 keV. Detection of the Auger electrons was

at X200

Forward Scattered Ag



I Illll ifll ll1 l nlll i ili111111 1 il I i f l llli ll tl
0 0.2 0.4 0.6 0.8 1

Figure 10. Comparison of forward scattered and backscattered
ISS data for the same sample. Note the contaminants which are
evident in the forward scattered spectrum are not detectable
in the backscattered ISS spectrum.


performed using a lock-in amplifier with a peak-to-peak

modulation of 0.5V in the outer cylinder. AES data were

obtained with the sample surface normal to the electron beam

and the electrons emitted in a cone approximately 420 about

the normal were collected. In ARAES, the sample is tilted

so that the electron beam strikes the sample at 200 from the

surface and the a 900 angle resolving aperture is positioned

such that only electrons emitted at an angle of about 200 from

the surface are collected. The technique has been described

in detail elsewhere [84].

The information obtained in the techniques of ISS, AES,

and ARAES can be combined to give a semiquantitative depth

profile of the surface region. It is known that ISS is

sensitive to only the first monolayer since ions which

penetrate beneath the surface have a nearly 100% chance of

being neutralized or undergoing multiple scatterings and

simply contributing to the "inelastic background" and not to

the elastic peaks which are the basis for quantification

[85,86]. AES, however, is sensitive to about 30 atomic layers

as shown in fig. lla. The data presented in Fig. 11 is

calculated based on the mean free paths of the incoming and

the emitted electrons. The mean free path of both the

incident and emmitted electrons was assumed to be 10 Angstroms

[88] and the lattice constant used was that determined by Kuk

and Feldman for the Ag(ll0) surface [83]. The mean free path


Percentage of total signal by layer

7 -
Normal Auger

I- S



S7 8 Q7

y "Numer

Figure 11. Results of the calculation of depth sensitivity
of AES and ARAES as a function of the silver layer number.

o g

Figure 11. Results of the calculation of depth sensitivity
of AES and ARAES as a function of the silver layer number.


data measured thus far show a scatter of several Angstroms

in both the primary electron and Auger electron energy ranges

so the results of the calculations presented, while correct

in relative magnitude of AES and ARAES signals, should be

considered approximate. It should also be noted that the

calculation considers only the angular dependence of the mean

free path and not the energy dependence. In addition, a

surface relaxation was assumed to take place and was also

modeled after the findings of Kuk and Feldman [83] for the

Ag(llO) surface. The assumption of a surface relaxation in

the polycrystalline sample is justified based on the findings

of this work. The calculated depth sensitivity of the ARAES

data is given in figure llb. Notice that the sensitivity of

the ARAES is much greater near the surface than AES and that

ARAES is sensitive only to about 15 atomic layers.

Results and Discussion

Determination of Relative ISS Cross Section

In order to quantify ISS data it is generally necessary

to experimentally measure the relative ISS sensitivities of

the species involved. Theoretical calculations are unreliable

since they depend strongly on the model chosen for the

interaction potential between the ion and the target atom

[89]. In addition, ion neutralization probabilities can

differ between target elements and this aspect of the cross

section is also very difficult to model.


The detected ISS Ni signal can be approximated by [85]

N. = K Si C,


K = a factor which depends on the primary flux,

instrument transmission, and other factors which don't

vary between target elements

Si = the elemental sensitivity

Ci = the concentration of species i at the surface

If we assume that the total density of atoms on the surface

doesn't change with coverage, then for a binary system (0

adsorbed on Ag) it can be shown that for any two different


(N01 N02) / (NAg2 NAg) = S / SA9

In order to determine the relative sensitivity, a silver

sample was cleaned and dosed with oxygen. Several ISS spectra

were then obtained during sputtering. Since the oxygen is

sputtered preferentially, ISS spectra of varying surface

oxygen coverages were obtained. The differences in oxygen

signal were then plotted vs. differences in silver signal for

all possible combinations of spectra obtained (fig. 12). The

slope of the best fit line was determined to be 0.045 and

represents the relative ISS sensitivity of 0 to Ag.



' 0.22




D 0.14

* 0.12

S 0.1

O 0.08




0 2 4 6

Silver Signal Change (arb. units)

Figure 12. Plot of data used to
sensitivities for a binary system. The
is given by the slope.

find relative ISS
relative sensitivity

Oxygen Dosing Results

The ISS data as a function of dosing temperature for a

pressure of 75 Torr are shown in fig. 13. The "clean" Ag

surface shows no ISS oxygen signal. The rest of the ISS data

show a rise in the oxygen surface concentration with

increasing dosing temperature up to a maximum of about 50% of

a monolayer. The normal AES data are shown in fig. 14. Like

the ISS data, the "clean" silver shows no detectable oxygen

signal greater than the noise level. However, as the dosing

temperature increases, the normal AES signal saturates at an

oxygen concentration of about 14 at. %. The ARAES data are

given in fig. 15. Unlike either the ISS or AES data for the

"clean" sample, the ARAES data show a detectable oxygen signal

corresponding to 3 at. % oxygen. Much like the AES data the

ARAES data also show that the silver saturates at an oxygen

concentration of about 14 at. %. The oxygen concentrations

calculated for all of the dosing temperatures and the "clean"

sample are summarized in fig. 16 and are discussed in detail


The results for the clean sample seem quite unusual. The

ISS data show clearly that the surface is free of oxygen, and

the AES data show the surface region is apparently free of

oxygen, but the ARAES data show that there is a small amount

of oxygen present. ARAES is normally sensitive to the

0 0.2 0.4 0.6 0.8

Figure 13. ISS spectra for the various
a) "clean"; b) 750C; c) 1500C; d) 2000C;

dosing temperatures.
e) 2500C

Kinetic Energy

Figure 14. Normal (bulk sensitive) AES data for the various
dosing temperatures. a) "clean"; b) 750C; c) 1500C; d) 200C;
e) 2500C


a) b) c) d) e)
Ag Ag Ag Ag Ag


0 0 0 00

Kinetic Energy

Figure 15. ARAES (near surface sensitive) data as a function
of dosing temperature. a) "clean"; b) 750C; c) 1500C; d) 2000C;
e) 2500C






Clean 75

150 200 250



Dosing Temperature ( C)

Figure 16. Summary of the atomic concentration calculated
from the ISS, AES, and ARAES data for all of the dosing



near surface region, however, in this case, ISS data show

conclusively that there is no surface oxygen. The results can

be explained by considering the relative sensitivity as a

function of depth between normal AES and ARAES (fig. 11). The

percentage of total signal for the first 4 layers in ARAES is

about double that for normal AES in the same region. If the

oxygen detected in the ARAES spectrum resided between the

first and fourth Ag layers in small quantities (<5%) then it

is possible that the oxygen signal would be below the normal

limit of detectability of the normal AES of about 1%. It is

reasonable to expect that the oxygen might be found

exclusively in a subsurface region if the surface undergoes

a relaxation in which one of the silver layers expanded

relative to the bulk. If this were the case, the oxygen would

have a lower energy in this region and would preferentially

fill the expanded region.

After a 75 OC oxygen dose, the surface concentration of

oxygen increases to 14% as indicated by ISS and the ARAES data

show about 7% oxygen. The fact that the surface oxygen

concentration increases faster than the near subsurface oxygen

concentration indicates that the potential barrier to surface

adsorption is lower than that for the oxygen to diffuse into

the near surface region. The normal AES data indicate an

oxygen concentration lower than that given by the ARAES data.

This is due to the oxygen not yet diffusing a great distance


into the bulk. Since a large portion of the normal AES signal

comes from depths beyond 5 atomic layers where the oxygen has

not yet diffused, the oxygen concentration given by the total

normal AES oxygen signal is lower than that given by ARAES.

After the 150 oC dose the surface concentration increases

to 40% oxygen and the AES data show an oxygen concentration

increase to 9% while the ARAES data indicate that the oxygen

content remains the same at 7%. Given that there is a

subsurface region in which the oxygen preferentially resides,

the ARAES data seem to be inconsistent. Since the surface

concentration increases it would be expected that the ARAES

data would show a larger oxygen concentration than the normal

AES data. The only way to explain this apparent discrepancy

is by postulating that part of the surface relaxation involves

the first two Ag layers which contract relative to the bulk

and, therefore, increases the potential energy of the oxygen

in that region. This suggests that the first layer below the

surface is comparatively depleted of oxygen.

Dosing temperatures of 200 oC and 250 oC saturate the near

surface region at about 14 at.%. Dosing at temperature up to

6000C and 100 Torr resulted in the same oxygen concentration

as measured by AES. Both the normal AES and ARAES data

indicate this concentration while the ISS data show an oxygen

concentration of 50 at. %. The fact that the ISS give a high

oxygen concentration while ARAES and normal AES give the same


lower oxygen concentration supports the hypothesis that a

layer near the surface is oxygen deficient.

A schematic model of the potential variation and final

oxygen concentration distribution as a function of depth which

is consistent with the data is given in fig. 17. The fact

that surface oxygen binds in relatively large quantities at

low temperature doses indicates a low activation energy for

surface adsorption but the relatively slow rate of movement

of oxygen into the subsurface and bulk regions indicates an

activation energy barrier between the surface and the bulk

which is significantly higher than that for surface

adsorption. The ARAES data for the "clean" sample indicate

that there is a subsurface region in which the oxygen

potential barrier to diffusion is significantly lower than in

the bulk. In addition, comparison of the data at higher

temperatures indicates that there is a layer in which the

oxygen potential is increased relative to the bulk. Analysis

of the relative depth sensitivities of ARAES and normal AES

indicates that these layers are in the first four atomic

layers of the Ag. Beyond the first few layers the potential

can be expected to be some uniform bulk value between layers

with some activation energy to go between layers. The model

proposed is consistent with a surface relaxation consisting

of a contraction relative to the bulk between the first two



x 3




0 d


Silver Layer Number

Figure 17. A model schematic 0 potential energy and final
concentration profile which is suggested by the data

Subsurface bulk
Surface [ Subsurface


silver layers and an expansion relative to the bulk between

the second and third silver layers.


ISS, AES and ARAES are useful together as a semi-

quantitative depth profile of the near surface region. For

the oxygen exposed large-grain polycrystalline silver,

evidence has been found for two subsurface regions with

differing oxygen affinities. In one region, oxygen adsorption

is enhanced relative to the bulk and, in the second, oxygen

adsorption is reduced relative to the bulk. In addition, the

oxygen adsorbs relatively easily on the surface, but diffuses

more slowly into the bulk, indicating that the activation

energy for adsorption is lower than the activation energy for

bulk diffusion.



In order for a successful HOAG to be constructed, a

relatively large increase in oxygen atom flux over that found

in the transient experiments is required. The two factors

limiting the flux are transport rate and ESD cross section.

The data presented in chapter 2 suggest that grain boundary

density can have an effect on oxygen transport rates through

silver metal. These results suggest that increasing grain

boundary density can increase the transport rates through the

metal. In addition, a higher ESD yield is desired, either

through an increased ESD cross section or an increased oxygen

concentration on the surface.

A search of the literature provided no references to

solid metals which had a higher permeability for oxygen at

practical temperatures than that which was measured for the

pure silver. However, it is known that small amounts of

impurities in the silver can decrease the grain size, thereby

increasing the grain boundary density. If the preferred

transport mechanism for the oxygen through the silver is


through the grain boundaries, then the increased grain

boundary density would increase the transport rates.

In addition to an increased oxygen transport rate through

the membrane, an increased ESD cross section is desired. In

his review of ESD phenomena[6], Hoflund stated that the

desorbing particle flux can be written as

ip = ip (EeQeIe ,e, Ep, pp, ,T,t,f(r))


Ee is the electron energy

Qe is the electron flux

0ene are the polar and azimuthal angles of incidence

Ep is the desorbing particle energy

0ene are the polar and azimuthal angles of desorption

is the surface coverage

T is the temperature

t is the time of exposure to the electron beam

f(r) is the defect density function

For the polycrytalline case the azimuthal angles are not

expected to play a role if the grain orientation is

sufficiently random but all of the other factors can be

important. Due to the difficulty of performing repetitive ESD

experiments, very little is known about the effects which

variation of these parameters on the ESD yield. In

particular, very little is known about the effect of electron

incidence angle on the ESD cross section. In order to


maximize the ESD cross section, the effect of the primary beam

angle is examined in this study.

The Ag/0.5% Zr system was chosen for several reasons.

The first, as mentioned before, is that grain growth in silver

is inhibited by additions of small amounts of impurities such

as zirconium. The addition of the zirconium tends to stop

the intergranular transfer of silver from smaller grains to

larger, thereby increasing the grain boundary density. The

second reason for choosing zirconium is that it is readily

oxidized to ZrO2. If the Zr in the alloy segregates to the

surface and oxidizes, it would cause an increase in surface

oxygen concentration which would cause an increase in the

oxygen ESD flux from the surface. The last reason for using

Zr as the additive is that it forms a maximal valency oxide.

It is theorized that ESD cross sections for ESD which occurs

by the KF mechanism are maximized for maximal valency oxides.

The KF mechanism depends on an interatomic Auger transition

which involves valence electrons from the oxygen atom relaxing

into a core hole on the metal atom. A relaxation involving

intra-atomic valence electrons is much more probable than the

interatomic transition if the valence electrons are available.

Since the metal atoms of maximal valency oxides have their

valence electrons used in the bonding orbitals, an interatomic

transition is more likely than if the metal atom has one or

more available valence electrons.


Another parameter which is of interest in the design of

the HOAG is the ratio of neutral atoms to ions produced. As

discussed in the introduction, the ratio predicted can vary

greatly depending on the mechanism. Experimentally measured

values reported in the literature range from 10 Neutrals/Ion

to greater than 4 x 106 Neutrals/Ion [5,22]. The total ESD

cross section was measured by monitoring the surface

concentration as a function of time, and the relative number

of neutrals to ions is not known. For a HOAG to be useful in

the study of orbiting spacecraft materials, it must produce

almost exclusively neutrals.

In this study, the surface composition of the Ag/0.5% Zr

alloy is measured using ISS. In addition, the ESD cross

section is determined from AES measurements and the neutral

to ion ratio is measured using a mass spectrometer. A

preliminary investigation of the effect of primary electron

beam angle on the ESD yield is also presented.


The alloy used in this study was prepared by vacuum

melting a mixture of Ag and 0.5 wt. % Zr in a manner similar

to the preparation of the high-purity large-grain silver. The

sample was solvent cleaned in an ultrasonic bath with toluene,

acetone, trichloroethylene,acetone, and ethanol in that order.

The alloy was cleaned in the UHV analysis system by repeated

Ar+ sputtering and annealing until no further contaminants


could be detected by AES and ISS. Zirconium is quite reactive

to many different compounds such as CO, CO2, and Cl so care

had to be taken to keep the background levels of the reactive

contaminants to a minimal level and the sample had to be

regularly sputter cleaned.

Oxygen dosing, ESD studies, and surface analysis were all

performed in the UHV vacuum system described the appendix.

For the ISS analysis, a 100 nA primary beam of 1 keV He+ was

defocussed over a large area to minimize possible sputter

damage. A scattering angle of 1440 was used and the energy of

the scattered ions was analyzed using the PHI model 25-270AR

CMA operated in the nonretarding mode. AES data were also

taken by using the CMA operated in the nonretarding mode with

a primary beam energy of 3 keV. Signal detection was

accomplished by applying a 5 kHz oscillation to the outer

cylinder and detecting the output of the multiplier with a

lock-in amplifier. In this manner, a dN(E)/dE spectrum is

obtained. Sample heating for the ISS studies was accomplished

by using both radiant heating and electron bombardment from a

tungsten filament placed behind the sample. The temperature

at which the ISS experiments could be performed was limited to

2500C since the vapor pressure of the silver was so high above

this temperature that components in the analysis chamber would

be coated with silver and malfunction during the relatively

long duration of these heating experiments.


ESD experiments were performed in the ESD vacuum chamber

attached to the main vacuum chamber. A schematic of the

layout is shown in figure 18. The electron source for both

the ion and neutral ESD studies is a PHI grazing incidence

angle electron gun whose energy can be varied from 100 eV to

5 keV. All of the ESD for these experiments was performed

using a primary beam energy of 1 keV. It was found that the

ESD yield increased dramatically in the range from 500 eV to

1 keV. Since the overall objective of these studies is to

increase the oxygen atom yield, a 1 keV beam was chosen. Two

different analyzers were used for the detection of desorbed

particles. For the detection of ions, a specially designed

SIMS/ESD analyzer was used. This analyzer consists of a

quadrupole mass spectrometer with a retarding-accelerating

type energy filter on the front of it. With the combined use

of the energy filter and the quadrupole mass analyzer it is

possible to determine the individual energy distributions for

each mass to charge ratio which is desorbed. The drawback of

this analyzer is its inability to detect neutral particles.

Since most of the particles desorbed in ESD are neutrals, this

results in the loss of much of the signal. In addition, an

important parameter in determining the mechanism of the ESD

is the ratio of ions to neutrals. For these measurements, an

EAI quadrupole mass spectrometer is mounted such that ESD

neutral particles can be ionized and detected.

Figure 18. A schematic of the ESD chamber including the QMS
and SIMS/ESD analyzer.


One problem with the detection of neutral particles is

the large background signal which is present in any vacuum

system (fig. 19a). This problem was overcome by utilizing the

fact that the desorbing particles have an energy in the 3 to

10 eV range before they enter the ionizer. The ionizer is an

electron impact ionizer which operates by passing a beam of

70 eV electrons through an ionization volume which, by

electron impact, ionize some fraction of the atoms in the

ionization volume. The ions which are formed are then

accelerated by a negative potential in the range of 3 to 10

volts. Any ions which have an energy below about 3 eV have

a nearly zero probability of getting through the quadrupole.

If the ion acceleration potential is reduced to zero, any ions

formed from the background gases will not be detected since

the background gas molecules have an energy in the thermal

range (=0.03 eV). Particles which are desorbed by ESD

however, typically have an energy in the range of 3 to 10 eV

and so can be detected (fig. 19b). From figure 19, it can be

seen that the relative magnitude of the signals is such that

the ESD signal could not be distinguished from the background

if the background signal were not removed.

In order to quantify the ion to neutral flux, it is

necessary to know the efficiency of the ionizer. This can be

calculated from the overall sensitivity of the analyzer to

background gases. The sensitivity of the mass spectrometer

110 -
100 -


4d -
70 -


0 4 0


LZ a

T .B-



0 2o 40 60 0o

Mass (amu)

Figure 19. Comparison of: a) the background signal with an
ion acceleration potential of 7 V, b) ESD signal with no ion
acceleration potential. Note the change in y scale.


to any given gas can be written as [88]

S (Amps/Torr) P (Torr) = Iout (1)


S is the sensitivity

P is the partial pressure of the gas

It is the signal detected.

The ideal gas law can be used to convert the pressure to a gas

density giving

S p R T = Iout (2)


p is the density

R is the universal gas constant

T is the temperature

For a flux of particles such as the ESD particles the density

can be given by

P = Jin / Vin (3)


J. is the flux of particles entering the ionizer

vin is the velocity of the particles entering the ionizer

In addition, from the definition of current,

Iout = Jut A (4)


A is the area of the aperture into the ionizer(0.503cm2)

Substitution of (3) and (4) into (2) gives

= (5)
Jin v A

Since S is known to be 5 x 10"4 Amps/Torr for a 1mA ionizer

electron emission of and the energy of the incoming particles

are near 5 eV, eqn. (5) can be used to calculate an ionizer

efficiency for the desorbing ion flux of 3.2 x 107 ions/atom.

The results presented here were obtained for a 0.2mA ionizer

emission so the efficiency for these experiments is 1/5 that

for 1mA emission or 6.4 x 10.8 ions/ atom.

Preliminary investigation of the dependence of the ESD

yield and energy distribution on the electron incidence angle

was performed by using the ESD/SIMS head. The angle of the

detector with the surface has to be held constant in order to

study the dependence of the ESD on the electron incidence

angle, since the detected ESD yield is known to be a function

of both incidence and detection angle. This was accomplished

in the ESD chamber by comparison of ESD ion energy

distribution (ESDIED) measurements in which the sample angle

was varied 200 to either side of the normal of the analyzer.

This is shown in figure 20. In this manner, two spectra were

obtained with the same angle to the detector (sample normal

200 from analyzer axis) but with two different electron

incidence angles (200 and 600 from the surface normal).




Figure 20. A schematic of the procedure used to examine the
dependence of the ESDIED on electron incidence angle while
holding the collection angle constant.


Results and Discussion

ISS Data

In order to determine what the composition of the surface

might be at the desired operating temperature range of near

500C, an ISS study was performed in which the sample was

sputter cleaned until free from any contaminants and dosed

with oxygen at 5000C and 100 Torr for 1 hour. Based on the

results for the pure silver, it is believed that this

treatment would saturate the sample with oxygen. The sample

was then transferred back to the main chamber where it was

sputtered until no further changes in the ISS spectrum were

noted. This treatment produced a non-equilibrium surface due

to the preferential sputtering of the O and to a lesser extent

the Zr. The sample was then heated to 2500C and its surface

composition was monitored with ISS. As noted previously, the

temperature could not be held at 5000C for this experiment due

to evaporation of silver. The ISS relative sensitivities

could not be calculated for the three component spectra as

they could for the O/Ag system because the assumption of

constant surface atom density was not valid. In order to

quantify the ISS data, the same O/Ag sensitivity of 0.045 as

the pure O/Ag case was used. For the relative sensitivity of

Zr/Ag the relative sensitivity was calculated using the method

of Parilis [86] to be 0.98. Although it was shown that the


O/Ag theoretical relative sensitivity was not accurate, the

Zr and Ag are more similar in mass and chemical character so

it can be expected that the theoretical relative sensitivity

nearly accurate. The surface concentration vs. time for the

three components are shown in figure 21. The Zr and O tend

to move towards the surface and partially cover the silver.

The silver concentration tends to stabilize near 30 at. %.

It can be seen from figure 22 that the O to Zr ratio tends

toward ZrO2. The ISS data can be interpreted as either Ag

islands on a layer of ZrO2 or, more likely, a layer of ZrO2

which does not completely cover the Ag (ie. islands of ZrO2 on


It is known from the previous studies that the silver

saturates with oxygen at about 14%. If islands of ZrO2 do

form over the silver then there would be two separate

mechanisms for ESD and, therefore, two peaks in an ESDIED

spectrum. ESDIED spectra from oxygen saturated Ag and Ag/0.5%

Zr are given in figure 23a and b respectively. Note that the

ESDIED from the pure silver shows only one peak at an ion

energy of about 5.5 eV but the ESDIED from the alloy shows a

peak at about 5 eV and another at about 4 eV. This is

confirmation that there are two separate composition regimes,

probably ZrO2 islands on Ag. It should also be noted that the

overall yield of O atoms from the Ag/Zr alloy surface is about

a factor of 8 larger than that for the pure silver.





.o 40

S 35
2 30



O 4 8 12 16 20 24 28
Time (rmino

Figure 21. Surface concentration as a function of time at
2500C from ISS.


S 1.3

0 0.9

0 .3 I I I I I I I I I I I I
0 4 8 12 16 20 24 28

Time (min.)

Figure 22. Ratio of 0 to Zr on the surface as measured by ISS
as a function of time at 2500C.

I I\II l I I I I l l 'i"

0 2 4 6 8 10 12 14 16 16 20
Ion Kinetic Energy (eV)

Figure 23. A comparison of the ESDIED from a) Ag,
b) Ag/0.5% Zr

Neutral to Ion Ratio Measurement

Some difficulty was encountered in the attempt to measure

the neutral to ion cross section. In order to maximize the

signal at the QMS the ionizer was placed as close to the

sample as possible. Unfortunately, the ionizer was close

enough that some of the ionizer electrons struck the sample

and produced a background ESD signal which was present whether

the main electron beam was on or off. This background signal

had to be quantified in order to determine the ratio of

neutrals to ions. The background signal was determined by

performing several experiments with different combinations of

ionizer electrons and primary electrons. The results of these

experiments are summarized as follows.

Case # Primary beam Ionizer Mass 16 signal
(arb. units)

1 on off 100
2 on on 314
3 off on 143

For case one, the signal detected includes only ions from the

primary beam since the ionizer is off. In case 2, the signal

detected includes both neutrals and ions from both the primary

and background ESD signals. Case three includes only ions and

neutrals generated by the background electrons. These

relationships can be summarized as follows.

Ip = 100 (6)


s (Np + Nb) + Ip + Ib = 314 (7)

s Nb + Ib = 143 (8)


Ip is the flux of ions due to the primary beam

N is the flux of neutrals due to the primary beam

Ib is the background flux of ions

Nb is the background flux of neutrals

s is the efficiency of the ionizer (6.4 x 10'8)

By solving these equations it can be shown that the neutral

to ion ratio is 1.1 x 107 neutrals/ion.

The result that nearly all of the particles desorbed from

this surface are neutrals has two major implications. The

first is that this surface is quite good for a HOAG with

respect to production of neutral oxygen atoms. The second has

to do with the mechanism for ESD from this surface. The KF

and MGR models generally predict neutral to ion ratios near

1000. The only mechanism proposed in the literature which

predicts neutral ratios of greater than 107 is the Antoniewicz

model which relies on the equilibrium position of the ionized

adsorbate state being closer to the surface than the ground

state. This mechanism, however, is thought to only occur for

adsorbates which are weakly bound to the surface, such as

condensed inert gases [21,22]. For the ZrO2 this is clearly

not the case. As was discussed previously, ZrO2 is a maximal

valency oxide, and the KF mechanism is generally favored on


one of the desorption states is due to the ZrO2. The

mechanism for oxygen desorption from the Ag regions is

probably the MGR mechanism and, the mechanism for desorption

from the ZrO2 is most likely the KF mechanism.

One possible explanation for the unusually high neutral

yield lies in the fact that a 1 keV primary electron beam was

used. For a primary beam of electrons of such a high energy,

many secondary electrons are produced, which form an electron

cloud around the spot where the electron beam strikes the

sample. The ions which are desorbed from the surface would

have to pass through this electron cloud and would, therefore,

have a relatively high probability of being neutralized by

electron attachment.

C. Measurement of Total ESD Cross Section

The total cross section for ESD is defined by [5]

d(0)/dt = n Q 0, (9)


n is the flux of electrons (cm-2s-1)

8 is the surface coverage of the species which is in the

binding state affected by ESD. (atoms/cm2)

Q is the total cross section for desorption (cm2)

This equation may be integrated to give

8(t)/80 = exp (-JOt/e}, (10)


J is the electron current density (A/cm2)


e is the electronic charge (C)

00 is the initial coverage.

The overall cross section can be determined from a plot of log

O(t)/00 vs. time. The surface coverage can be monitored by

AES but, a portion of the peak to peak height of the oxygen

peak in an Auger spectrum of the oxygen saturated Ag/0.5% Zr

alloy is proportional to the surface coverage and part of the

signal is due to the oxygen beneath the surface. In order to

separate these two signals the sample was kept under the

electron beam for several hours until no further change in the

oxygen signal was seen. The signal at that point was assumed

to be that portion of the AES signal which was due to oxygen

beneath the surface and this residual signal was subtracted

from the total signal in order to obtain a quantity which was

proportional to surface coverage. Substituting the corrected

AES signal into eqn. 10 gives

ln( (N(t)-N,)/(tJo-N, ) = -(JQ/e)t (11)

A plot of this quantity vs. time is given in figure 24. From

the slope of a linear least squares fit to this data, a total

cross section for ESD was found to be 3.9 x 10'19 cm2.

This cross section is slightly smaller than the O ESD

cross section for pure silver which was determined to be 7 x

10"19 cm2 [1]. This is consistent with the observed total

yield for ESD from the alloy which is about 8 times higher

than for the pure silver since the oxygen concentration is

0.00 -

.-0.20 -

Z -0.40




* 4.e
4. 4.c

4. 4. 4. 4



1000.00 1500.00


Figure 24. Plot of the AES data used to calculate the total
cross section for ESD. Slope of the best fit line is -



higher for the fully oxidized alloy surface than that for the

pure silver.

Effect of Electron Incidence Angle

The ESDIED for the two electron incidence angles which

were obtained for this study are shown in figure 25. Both

of the spectra were obtained with the same collection angle

but differing incidence angles. The overall ion signal for

the more glancing angle of incidence (Fig. 25b) is about three

times larger in magnitude than the overall ion yield for the

more normal incidence. In addition, the energy maximum of the

more glancing incidence spectrum is about 1 eV higher than

that for the more normal angle of incidence. The difference

in magnitude could be explained by a change in the shape of

the electron cloud which would change the neutralization

probability, but this would not explain the shift in energy.

The differences could be explained by considering the

mechanisms for desorption. All of the mechanisms proposed

rely on excitation of a bonding orbital by an incoming

electron. The bonding orbitals are generally located between

the surface atom and the adsorbate. An electron which was

approaching the adsorbate bond from above would tend to be

partially screened from the bonding orbital by the adsorbate

atom while an electron which approaches from a glancing angle

would less screened. The glancing electrons would have more

access to the bonding orbitals and, perhaps, access to some

Ion Kinetic Energy (eV)

Figure 25. ESDIED for two different incidence angles. a) 20
from surface normal, b) 600 from surface normal.


orbitals which are nearly blocked from the surface normal.

The results for this experiment as well as the hypothesis to

explain them are preliminary and further investigation into

this phenomenon is being planned.

Summary and Conclusions

All of the studies presented in this chapter indicate

that the Ag/0.5% Zr alloy will make a surface which is

superior to pure Ag with regard to generation of a

hyperthermal neutral oxygen atom beam by ESD. The ratio of

neutrals to ions for ESD of oxygen by 1 keV electrons from

this surface is more than 1 x 107. The energy distribution of

desorbing oxygen from the alloy is broader than that for the

pure silver but it is still centered near 5 eV and shows a

total ESD oxygen yield about 8 times higher than pure silver.

The cross section which was measured for the alloy is about

half that measured of O ESD from silver but the higher

concentrations of oxygen in the alloy cause the oxygen ESD

yield to be higher. A more grazing electron incidence angle

was shown to give an increased oxygen ion yield and slightly

higher energy maximum but the reasons for this are not yet

well understood.



A top view of the system is shown in Figure 26. This

system has many unique capabilities for the study of surfaces.

A system has been developed which consists of a sample

preparation chamber, a pretreatment chamber, and the UHV

analysis chamber. A sample manipulator system has been

developed which makes it possible to move the sample between

the various chambers and heat the sample in the preparation

or analysis chamber. It is also possible to prepare model

catalysts in a highly controlled atmosphere with intermediate

surface characterizations. Then, the sample can be moved from

the vacuum system and inserted into a high pressure reactor

and back into the UHV system for analysis without exposure to


It is possible to prepare numerous types of catalytic

surfaces in this type of system. Many of our studies have

involved electrochemically prepared catalysts such as tin

oxide- or titania- supported platinum. With this system oxide

surfaces can be prepared and characterized in the UHV system.

Facilities for metallization are also contained in the


preparation chamber so that mixed oxide catalysts and

supported metal catalysts can be prepared totally within the

system. Furthermore, the same facilities which are used for

catalyst preparation can be used for treatment of materials

with high pressures of gases at temperatures up to =800 C.

The system utilizes ion pumping and Ti sublimation

pumping to reach a base pressure in the 10"11 Torr range. A

turbomolecular pump is used to rough the system, pump inert

gas away after sputtering or performing ion scattering

spectroscopy (ISS), and differentially pump other portion of

the system. A PHI Model 25-270AR double-pass CMA charged-

particle energy analyzer is suspended vertically from port 6.

Port 5 contains a manipulator with capabilities for moving,

heating, and cooling the sample. The sample is placed on this

manipulator by a long stroke manipulator after sample

preparation and pretreatment.

Many surface techniques are included in this UHV system

(see Table I). The CMA is used both as a charged particle

energy analyzer and as a time-of-flight mass spectrometer.

Numerous sources are mounted in the ports focusing at the CMA

focal point including an ion gun, an ultraviolet photon

generator, an X-ray source, electron guns and a molecular beam

doser. This allows us to perform many techniques using the

CMA including X-ray photoelectron spectroscopy (XPS), Auger

electron spectroscopy (AES) and scanning Auger microscopy


(SAM), electron energy loss spectroscopy (ELS), Auger electron

appearance potential spectroscopy (AEAPS), electron stimulated

desorption with angular resolution and energy analysis (ESD,

ESDIAD, ESDIED), ultraviolet photoemission spectroscopy (UPS),

ion scattering spectroscopy (ISS), and secondary ion mass

spectrometry (SIMS). This chamber also contains a mass

spectrometer and a secondary electron detector so that we can

perform temperature programmed desorption (TPD), isotope

exchange experiments, and work function measurements. It is

possible to perform many of these techniques in an angularly

resolved manner over a very broad temperature range from that

of liquid nitrogen to above 1200 C.


Table I

Available Techniques

I. Cylindrical Mirror Analyzer

*A. X-ray Photoelectron Spectroscopy
*B. Ultraviolet Photoemission Spectroscopy (UPS)
*C. Auger Electron Spectroscopy (AES)
D. Scanning Auger Microscopy (SAM)
*E. Electron Energy Loss Spectroscopy (ELS)
F. Auger Electron Appearance Potential Spectroscopy
*G. Electron Stimulated Desorption (ESD, ESDIED, ESDIAD)
H. Ion Scattering Spectroscopy (ISS)
I. Depth Profiling

II. Others
A. Work Function Measurements
B. Temperature programmed Desorption
C. Isotope Exchange Experiments
D. Secondary Ion Mass Spectrometry (SIMS and Tandem
E. Gas Dosing
F. Heating > 1500 C
G. Cooling to 10 K
H. Inverse Photoelectron Spectroscopy
I. Metal Deposition
J. High Pressure Pretreatment

*Can be performed in the angle-resolved mode





Gas Exposures
Sample Heating
Metal Deposition


Figure 26. Schematic diagram of the vacuum system used in
these studies.


1. R.A. Outlaw, W.K. Peregoy, G.B. Hoflund, and G.R.
Corallo, NASA technical paper 2668, April 1987

2. R.A. Outlaw and F.J. Brock, J. Vac. Sci. Tech

3. W. Eichenauer and G. Muller, Z. Met. Kd., 53(1962)321

4. T.A. Ramanarayanan and R.A. Rapp, Metall. Trans.

5. T.E. Madey and J.T. Yates, Jr., J. Vac. Sci. Technol.,

6. G.B. Hoflund, Scanning Elect. Microsc., (1985)1391

7. R. Gomer in Desorption Induced by Electronic
Transitions DIET I, p. 80, Eds. N.H. Tolk, M.M. Traum,
J.C. Tully, and T.E. Madey, Springer-Verlag, New York,
NY, 1983

8. P. Feulner, S. Auer, T. Muller, A. Puschmann, and D.
Menzel in Desorption Induced by Electronic Transitions
DIET III, p. 58, Eds. R.H. Stulen and M.L. Knotek,
Springer-Verlag, New York, NY, 1988

9. G.E. Moore, J. Appl. Phys. 32(1961)1241

10. D.R. Sandstrom, J.H. Leck, and E.E. Donaldson, J. Appl.
Phys. 38(1967)2851

11. T.E. Madey and J.T. Yates, Jr., Surface Sci.,

12. D.A. Degras and J. Lecante, Nuovo Cimento Suppl.

13. P.A. Redhead, Nuovo Cimento Suppl., 5(1967)586

14. W. Ermrich, Nuovo Cimento Suppl., 5(1967)580

15. J.T. Yates Jr., T.E. Madey, and J.K.Payn, Nuovo Cimento
Suppl., 5(1967)558


16. P.A. Redhead, Can. J. Phys., 42(1964)386

17. J.W. Coburn, Surface Sci., 11(1968)61

18. S. Auer, P. Feulner, and D. Menzel, Phys. Rev. B,

19. P. Feulner and D. Menzel, Phys. Rev. Lett., 53(1984)671

20. E.R. Moog, J. Ungris, and M.B. Webb, Surface Sci.,

21. Q.J. Zhang and R. Gomer, Surface Sci. 109(1981)567

22. Q.J. Zhang, R. Gomer, and D.R. Bowman, Surface Sci.

23. R.A. Outlaw, G.B. Hoflund, and G.R. Corallo, App.
Surface Sci. 28(1987)235

24. D. Menzel in Desorption Induced by Electronic
Transitions DIET I, Eds. N.H. Tolk, M.M. Traum, J.C.
Tully, and T.E. Madey, Springer-Verlag, New York, NY,

25. P. Feulner, S. Auer, T. Muller, A Puschmann, and D.
Menzel in Desorption Induced by Electronic Transitions
DIET III, p. 61, Eds. R.H. Stulen, and M.L. Knotek,
Springer-Verlag, New York, NY 1988

26. P.R. Antoniewicz, Phys Rev. B 21(1980)3811

27. J.S. Hammond, S.W. Gaarenstroom, and N. Winograd, Anal.
Chem., 47(1975)2193

28. G. Schon, Acta Chemica Scandinavica, 27(1973)2623

29. R.E. Kenson and M. Lapkin, J. Phys. Chem.,74(1970)1493

30. P. Harriott, J. Catalysis, 21(1971)56

31. N.W. Cant and W.K. Hall, J. Catalysis, 52(1978)81

32. A.L. Larrabee and R.L. Kuczkowski, J. Catalysis.,

33. T. Inui and Y. Tanabe, J. Catalysis, 52(1978)375

34. H. Kanoh, T. Nishimura, and A. Ayame, J.
Catalysis, 57(1979)372


35. H.R. Dettweiler, A. Baiker, and W. Richarz, Helvetica
Chimica Acta, 62(1979)1689

36. S. Kagawa, M. Iwamoto, H. Mori, and
Tetsuro Selyama, J. Pys. Chem. 85(1981)434

37. P. Vishnu Kamath, K. Prabhakaran, and C.N.R. Rao,
Surface Sci. 146(1984)L551

38. P.J. Goddard and R.M. Lambert, Surface Sci.,

39. R.B. Grant and R.M. Lambert, Surface Sci. 146(1984)256

40. T.E. Felter, W.H. Weinberg, P.A. Zhdan, and G.K.
Boreskov, Surface Sci. 97(1980)L313

41. R.B. Grant and R.M. Lambert, Surface Sci., 146(1984)256

42. H.A. Engelhardt and D. Menzel, Surface Sci.,57(1976)591

43. J. Eickmans and A. Otto, Surface Sci., 149(1985)293

44. C.T. Campbell, Surface Sci., 157(1985)43

45. C.T. Campbell and M.T. Paffett, Surface
Sci. 143(1984)517

46. K. Bange, T.E. Madey, and J.K. Sass, Surface Sci.,

47. M. Bowker, M.A. Barteau, and R.J. Madix, Surface
Sci., 92(1980)528

48. C. Benndorf, M. Franck, and F. Thieme, Surface Sci.,

49. C. Backx, C.P.M. DeGroot, P. Biloen, and W.M.H.
Sachtler, Surface Sci. 128(1983)81

50. C. Backx, C.P.M. DeGroot, and P. Biloen, Surface
Sci., 104(1981)300

51. C.T. Au, S. Singh-Boparai, M.W. Roberts, and R.W.
Joyner, J. Chem. Soc., Faraday Trans. 1, 79(1983)1779

52. I.E. Wachs and S.R. Kelemen, J. Catalysis, 68(1981)213

53. R. Sporken, P.A. Thiry, J.J. Pireaux, R. Caudano, and
A. Adnot, Surface Sci. 160(1985)443

54. C.T. Campbell and M.T. Paffett, Surface Sci.

55. G. Rovida, F. Pratesi, M. Maglietta, and E. Ferroni, J.
Vac. Sci. Tech, 9(1971)796

56. S. Kagawa, M. Iwamoto, and S. Morita, 78(1982)143

57. G. Rovida, F. Pratesi, M. Maglietta and E. Ferroni,
Surface Sci., 43(1974)230

58. G. Rovida amd F. Pratesi, Surface Sci. 52(1975)542

59. W. Heiland, F. Iberl, E. Taglauer, and D. Menzel,
Surface Sci. 53(1975)383

60. M. Kitson and R.M. Lambert, Surface Sci., 109(1981)60

61. A. Puschmann and J. Haase, Surface Sci., 144(1984)559

62. K.C. Prince and A.M. Bradshaw, Surface Sci.,

63. W. Krakow, Surface Sci., 140(1984)137

64. L. Lefferts, J.G. van Ommen, and J.R.H. Ross, App.
Catalysis, 31(1987)291

65. M. Stoukides and C.G. Vayenas, J. Catalysis,

66. M. Peukert, Surface Sci., 146(1984)329

67. B. Reihl, R.R. Schlittler, and H. Neff, Surface
Sci., 152/153(1985)231

68. S. Kagawa, M. Iwamoto, and S. Morita, J. Chem. Soc.,
Faraday Trans. 1, 78(1982)143

69. F.K. Moghadam and D.A. Stevenson, J. Electrochem. Soc.,

70. N. Ting, Y. Qingliang, and Y. Yiying, Surface Sci.,

71. Y. Kuk and L.C. Feldman, Phys. Rev. B, 30(1984)5811

72. H. Rickert and R. Steiner, Z. Phys. Chem. 49(1966)127

73. E.H. Kolosov, H. Starkovscii, and B.M. Gryanznow, Zh.
Fiz. Khim., 48(1974)1861

74. B.M. Gryaznov, C.G. Gul'yanova, and C. Kanizius, Russ.
J. Phys. Chem. 47(1973)1517

75. R.E. Coles, Br. J. Appl. Phys., 14(1963)342

76. J. Crank, Mathematics of Diffusion, 2nd ed., p. 42
Oxford University Press, Oxford, 1967

77. R.A. Outlaw, W.K. Peregoy, and G.B. Hoflund, NASA Tech.
Pub. 2755(1987)

78. L.C. Beavis, Rev. Sci. Instrum., 43(1972)122

79. D.J. Mitchell, J.M. Harris, R.C. Patrick, E.P.
Boespflug, and L.C. Beavis, J. Appl. Phys., 53(1982)970

80. H.H. Johnson and R.W. Lin, in Proceedings of the 3rd
International Conference on Effect of Hydrogen on the
Behavior of Material, Eds. I.M. Bernstein and A.W.
Thompson, AIME, New York, NY, 1981

81. R.E. Hoffman and D.J. Turnbull, J. Appl. Phys.

82. W. Segeth, J.H. Wijngaard, and G.A. Savatzky, Surface
Sci., 194(1988)615

83. Y. Kuk and L.C. Feldman, Phys. Rev. B. 30(1984)5811

84. G.B. Hoflund, D.A. Asbury, C.F. Corallo, and G.R.
Corallo, J. Vac. Sci. Technol. A 6(1988)70

85. G.B. Hoflund and V. Young, Submitted to Surface

86. E.S. Parilis in Proceedings of the 7th International
Conference on Phenomena in Ionized Gases Vol. 1,
Belgrade, 1965, p. 129, Eds. B. Perovic and D. Tosic,
Gradevinska Knjiga Publishing House, Belgrade, 1966

87. A. Zangwill, Physics at Surfaces, Cambridge University
Press, New York, NY, 1988

88. QUAD 250 Residual Gas Analyzer Instruction Manual,
Electronics Associates, Inc., Palo Alto, California,


The author was born on July 10, 1962, in Chattanooga,

Tennessee, to A. Rogers and Mary C. Davidson. Soon after his

birth, his parents moved back to their native Florida. The

author was raised in Seminole, Florida, where he attended

Seminole High School and graduated in May, 1980. He received

a national merit scholarship to attend the University of

Florida in Gainesville. During his undergraduate education,

he worked as a lab assistant for Professor Herbert A. Laitinen

in the Department of Chemistry. While working for Professor

Laitinen, the author had some exposure to the work of

Professor Gar B. Hoflund with the Chemical Engineering

Department in the field of surface analysis. After graduating

with a B.S. in chemistry, the author chose to attend graduate

school in the Department of Chemical Engineering, in order to

learn the science of probing the outermost layers of solid

materials. In May 1986, the author was awarded a Master of

Engineering in chemical engineering. The author plans to

receive his Ph.D. in May, 1990.

I certify that I have read this study and that in my
opinion it conforms to acceptable standard of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.

Gar B. Hoflur Chairman
Professor of Chemical

I certify that I have read this study and that in my
opinion it conforms to acceptable standard of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.

'i, / I j/e ;' -, / ., < .
Herbert A. Laitinen
Graduate Research
Professor of Chemistry

I certify that I have read this study and that in my
opinion it conforms to acceptable standard of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.

Timothy J/ Anderson
Professor of Chemical

I certify that I have read this study and that in my
opinion it conforms to acceptable standard of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.

Gerasimos 7. Lyberatos
Associate Professor of
Chemical Engineering

I certify that I have read this study and that in my
opinion it conforms to acceptable standard of scholarly
presentation and is fully adequate, in scope and quality, as
a dissertation for the degree of Doctor of Philosophy.

Vaneica Y. Young,
Associate Professor of"

This dissertation was submitted to the Graduate Faculty
of the College of Engineering and to the Graduate School and
was accepted as partial fulfillment of the requirements for
the degree of Doctor of Philosophy.

May, 1990 L&Jl 0, 0,"
Winfred M. Phillips
Dean, College of

Madelyn M. Lockhart
Dean, Graduate School

/< (
/ ( i

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs