Title Page
 Table of Contents
 List of Tables
 List of Figures
 Statement of problem
 Surface science techniques
 Sample preparation
 XPS of prepared samples
 Depth profile of degraded...
 Discussion of results and...
 Biographical sketch

Title: Application of surface science techniques to degradation of superconducting YBa2Cu3O7-<delta, Greek letter>
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00090182/00001
 Material Information
Title: Application of surface science techniques to degradation of superconducting YBa2Cu3O7-<delta, Greek letter>
Physical Description: Book
Language: English
Creator: Büyüklimanli, Temel Hulusi
Publisher: Temel Hulusi Büyüklimanli
Publication Date: 1991
 Record Information
Bibliographic ID: UF00090182
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 - 001693332
oclc - 25221953

Table of Contents
    Title Page
        Page i
        Page ii
    Table of Contents
        Page iii
        Page iv
    List of Tables
        Page v
    List of Figures
        Page vi
        Page vii
        Page viii
        Page ix
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
    Statement of problem
        Page 11
        Page 12
    Surface science techniques
        Page 13
        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
        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
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
    Sample preparation
        Page 60
        Page 61
        Page 62
        Page 63
    XPS of prepared samples
        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
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
    Depth profile of degraded superconductors
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
    Discussion of results and conclusions
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
    Biographical sketch
        Page 125
        Page 126
        Page 127
Full Text







I am grateful to my supervisor, Dr. J.H. Simmons for his guidance throughout

this dissertation and for creating an atmosphere that makes working with him a

pleasure. I was lucky to be able to benefit from his broad knowledge of science. I

would like to thank my committee members, Dr. P.H. Holloway, Dr. D.E. Clark,

Dr. J.J. Mecholsky,Jr. and Dr. D.B. Tanner, for their help and constructive

criticism. Special thanks goes to Dr. Holloway who suggested me to work on this

subject and helped me learn surface science better. I also would like to extend my

gratitude to Greg Chandler, Iftikhar Ahmad, Carl Mueller and Kelly Truman for the

samples they provided, Eric Lambers for his help with surface measurements and for

their valuable discussions which enhanced this work.

Whatever I accomplished would have been almost impossible to do without my

family's support even though I couldn't be with them in their troubled times. And I

dedicate this work to my mother and to the memory of my father and my grandfather

who gave me the first spark to do academic work and continuing motivation when I

was faraway but couldn't live long enough to see me complete it.

Finally I am very much in debt to my wife Isn, who deserves most of the

credit, for her support, encouragement and understanding.


ACKNOWLEDGEMENTS .................................. ii

LIST OF TABLES .................. .................... v

LIST OF FIGURES ...................................... vi

ABSTRACT ........................................... ix


1 INTRODUCTION ................. ................. 1
1.1 Superconductivity ............................ 1
1.2 Corrosion of Ceramic Superconductors ............... 5
1.3 Application of X-ray Photoelectron Spectroscopy to
Superconductors ........................... 8

2 STATEMENT OF PROBLEM ........................... 11

3 SURFACE SCIENCE TECHNIQUES ...................... 13
3.1 X-ray Photoelectron Spectroscopy ................. 13
3.1.1 Theory .............................. 13
3.1.2 Qualitative analysis ..................... 16
3.1.3 Chemical Shifts ............... ...... 23
3.1.4 Quantitative Analysis .................... 25
3.1.5 Instrumentation ................ ...... 30
3.1.6 Data Collection ....................... 40 Instrument Settings ............... 40 Irradiation Effects ................ 43 Survey (Wide) Scan ............... 44 Standardization .................. 44 Data Manipulation ................ 44
3.2 Depth Profile ............................. .. 51
3.2.1 XPS with Ion Sputtering .................. 51
3.2.2 AES with Ion Sputtering .................. 53
3.2.3 Angle Resolved XPS .................... 56

4 SAMPLE PREPARATION ............................ 60
4.1 Pellets ................................... 60
4.2 Films ..................................... 61

5 XPS OF PREPARED SAMPLES ........................ 64
5.1 Experimental Set Up ........................... 64
5.2 Chemical Reactions on the Surface ................ 78
5.3 Concentration Changes ........................ 91


7.1 Phase Transformation ..................
7.2 Passivation of Degradation ...............
7.3 Summary and Conclusions ...............

REFERENCES .................................

BIOGRAPHICAL SKETCH ..........................

....... 106
....... 106
....... 113
....... 115


....... 126


Transition Temperature of Superconductors. ............... 5
Binding energy positions of the lines produced by YBa2Cu30
detected by XPS operated under Mg Ka X-rays. ........... 20
X-ray and spectroscopic notation. . . . . . . ..... ....... 23
Area ratio of the spin-orbit splitting doublets. . . . . . ... 24
Molybdenum XPS lines for Mg Ka X-rays. . . . . . ... 67
Changes in the relative atomic concentrations due to X-ray
exposure and Ar ion milling. ....................... 70
Components used to curve fit Ba 4d spectra of 123 compounds .84
Components used to curve fit O Is spectra of 123 compounds. .85
Heats of formation of barium compounds with oxygen,
hydrogen and carbon ............................. 86
Relative atomic concentrations of YBCO pellet and RF sputtered
film during the experiment. ........................ 94


Table I
Table II

Table III
Table IV
Table V
Table VI

Table VII
Table VIII
Table IX

Table X

Figure page
Figure 1 Low temperature resistivity of (a) a normal metal, and (b) a
superconductor (in zero magnetic field), containing nonmagnetic
impurities. ................................... 2
Figure 2 Magnetic field temperature phase diagram of Type-I
superconductors. ............................... 4
Figure 3 Photoemission process. ............................ 14
Figure 4 De-activation of a photoemission process by emitting a photon
or Auger electron and shake up transition. . . . . . .... 15
Figure 5 Wide XPS scan of a superconducting YBazCu307 pellet using
Mg Ka x-rays. .................. .............. 18
Figure 6 A schematic of a basic XPS instrument. . . . . . . ... 32
Figure 7 The use of Auger line shape to deconvolute F KLL spectrum in
measuring the concentration of Pb-F-Pb, Pb-F-Cd and
Cd-F-Cd type bridging fluorine. ................... 40
Figure 8 Surface sensitivity enhancement by variation of the electron
take-off angle. ...................... .......... 58
Figure 9 Angular dependence curves calculated for a clean flat thick
surface and a flat overlayer/substrate system. . . . . . ... 59
Figure 10 Comparison of Cu 2p (YBCO after ion sputtering) raw spectrum
with smoothed spectrum using 19 point Savitsky-Golay routine 8
times. .......................................... 66
Figure 11 The effects of 35 minute Ar+ ion sputter cleaning on (a) Cu
2P3/2 and (b) Cu LMM spectra of a YBCO pellet. ........... 69
Figure 12 Effects of Ar+ ion milling in Ba 4d of YBCO pellet while
removing surface adsorbates. . . . . . . . . . ... 71
Figure 13 C Is spectrum of YBCO after exposure to ambient air and 85%
relative humidity at 50 C. ......................... 73
Figure 14 Curve fitted C Is spectrum of YBCO pellet treated with 85%
relative humidity at 500C for 5 hours. .................. 74
Figure 15 Change due to X-ray radiation damage in the (a) Cu 2P3/2 and
(b) Cu LMM ................................. 76
Figure 16 The XPS wide spectrum of a superconducting YBa2Cu307-_
pellet after 5 point Savitsky-Golay smoothing, Cls referencing
(284.6 eV) and baseline and satellite subtraction. ............ 79
Figure 17 Flowchart of experimental procedure followed. . . . . ... 81

Figure 18 Second derivative of a Ba 4d spectrum from a YBCO treated 5
hours in the humidity chamber. 3 times 19 point smoothing and
satellite subtraction was carried out before the differentiation. ... 83
Figure 19 Curve fitting of the Ba 4d spectra acquired from a low density
pellet during the course of the experiment. . . . . . .. . 89
Figure 20 Curve fitting of the O Is spectra acquired from a low density
pellet during the course of the experiment. . .... . . . . 90
Figure 23 Relative concentration of barium and copper with respect
exposure to corroding environment. . . . . . . . ... . 95
Figure 24 Angle Resolved XPS spectrum of a YBCO. The distance from
surface is proportional to Sin 0 where 0 is the take-off angle. .98
Figure 25 Normalized atomic concentration qf Y, Ba, Cu calculated from
the spectra shown in Figure 24. . . . . . . . . . ... 99
Figure 26 Survey scan of a fresh pellet measured by SAM following a 20
min Ar ion sputtering (a), and after 7 point smoothing and 11
point differentiation (b). ........................ . 101
Figure 27 Depth profile of an as received 123 pellet measured by SAM
and its atomic concentration at end of the measurement. ....... 103
Figure 28 Ion assisted depth profile measured by SAM of the sintered
pellet (density = 3.0 g/cc) after 48 hours at 85% humidity and
500C. .... . ................... ..... .... 105
Figure 29 Ion assisted depth profile measured by SAM of the RF sputtered
film after 48 hours at 85% relative humidity and 500C. ...... 106
Figure 30 Depth profile of a laser ablated film after exposure to air and
humidity. ................... ... ............. 107
Figure 31 SEM photomicrograph of (a) the pellet and (b) the film before
exposure air and humidity ......... ......... ..... 111
Figure 32 Change in the Ba 4d512 XPS peak of (a) pellet and (b) film, due
to ambient atmosphere and humidity exposure calculated from
the peak areas of Figure 19 .................. ....... 113
Figure 33 Change in the O Is XPS peak of (a) pellet and (b) film, due to
ambient atmosphere and humidity exposure calculated from the
peak areas of Figure 20. ................. ... .... 114

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




August 1991

Chairman: Joseph H. Simmons
Major Department: Materials Science and Engineering

Surface modifications from exposure to ambient atmosphere and 85% relative

humidity at 500C were measured on YBa2Cu307_8 (YBCO) pellets with densities

varying from 3.0 g/cc to 5.5 g/cc and thin films fabricated by laser ablation and RF

sputtering of the 123 composition using x-ray photoelectron spectroscopy (XPS),

scanning Auger electron multiprobe (SAM) and ion-assisted depth profiling. The

corrosion reactions in air led to surface formation of BaCO3 and exposure to humid

atmosphere showed a predominance of BaCxOyHz and BaOHy and the Ba 4d52 and

O Is binding energies for these reaction products have been measured to be 89.6,

91.0, 92.1 and 530.2, 532.5, 534.0 eV's respectively. An O Is peak from water in

hydrated pellet was also detected at 535.6 eV.

No Cu and Y was detected in the surface layer of the sintered pellets by XPS and

SAM after exposure to 18 days of ambient atmosphere followed by 48 hours of moisture.

Accumulation of Ba and depletion of Cu was confirmed by ion-assisted Auger electron

spectroscopy (AES) depth profile.

The rate of accumulation of Ba and formation of reaction products (BaCxOyH)

increases with particle size (RF sputtered film to low density pellet). High negative heat

of formation values for BaOxH and BaCxOyHz seems to be the driving force for Ba to

diffuse to the surface provided that reaction paths are available. Grain boundary or

surface diffusion appear to be the most likely transport paths.



1.1 Superconductivity

Many metals and alloys, undergo a phase transition to a state of zero

resistivity at low temperatures. First observed by H. Kammerlingh Onnes' in Hg

near the liquid He vaporization temperature, this phenomenon has been extended

beyond 100 K in a set of ceramic perovskites discovered by Bednorz and Miller.

The temperature at which the transition takes place is called the critical temperature

(Te) and the zero resistivity (or infinite conductivity) state is called superconductivity.

Figure 1 shows the resistivity of a normal and a superconducting metal as a function

of temperature. The resistivity (p) behavior of a metal at low temperatures (also

followed by the superconductor above T,) is given by2

p(T) = po + AT5

where po is resistance caused by defects in the lattice such as impurities, grain

boundaries and twin boundaries and twin domains. The AT5 term arises from

electron-phonon interactions intrinsic to the material.

In the superconducting state the magnetic flux is expelled from the material

due to an ability to sustain shielding currents without energy loss. This is known as



Figure 1 Low temperature resistivity of (a) a normal metal, and (b) a
superconductor (in zero magnetic field), containing nonmagnetic

the Meissner effect. This property is often used for demonstrations of

superconductivity by levitating a magnet over a superconductor (or vice versa). The

magnetic induction inside the substance is known to be

B = po (H+ M) = go(1 + x)H

where H is the intensity of the external magnetic field, M the magnetization in the

medium, and X its magnetic susceptibility. Since B = 0 in the superconducting state,

it follows that M = H. The medium is therefore perfectly diamagnetic, and the

susceptibility is x = 1.

Superconductivity is lost if a sufficiently large current is passed through the material

(critical current Je), or if a magnetic field above a critical value is applied to the

sample (critical magnetic field H,). It is found that He changes as a function of

temperature and the relation is given by3

H,(T) = H(O) 1- T (1)

This can be represented as a phase diagram of the superconductor (magnetic field vs

temperature) (Figure 2). These superconductors are know as Type I. In the Type II

superconductors however, the Meissner effect begins to breakdown, at least partially,

well before the critical field is reached. Type II materials are known as hard

superconductors, because they usually have high critical field.

H non-superconducting



Figure 2 Magnetic field temperature phase diagram of Type-I

The highest transition temperature recorded in metal alloys was 23 K measured

on Nb3Ge.4 Table I gives the transition temperatures of some metals and alloys.

Bednorz and Miller (1986) in their discovery of ceramic perovskite

superconductors (La2_ BaCuO4 system with a TK of -30 K)5 broke through the

previous long standing maximum of 23 K in Nb3Ge. Chu and his coworkers (1987)

raised the transition temperature above the liquid nitrogen boiling point = 77 K with

yttrium barium copper oxides YBa2Cu307_- (YBCO T, = 93 K).6 Intensive

exploration with Bi and Tl based copper oxides have pushed T, over 125 K. The

ceramic perovskite superconductors have received much attention since their

Table I


Transition Temperature of Superconductors.

Element Tc (K)t Compound Tc (K)*
Nb 9.26 Nb3Ge 23.2
Tc 7.77 Nb3Alo.8Ge0.2 20.1
Pb 7.19 Nb3Sn 18.1
La 3(fcc) 6.06 Nb3A1 17.5
V 5.30 V3Ga 16.5
La ca(hcp) 4.9 Nb3N 16.0
Ta 4.48 MoN 12.0
Hg c(rhomb) 4.15 Nb3Au 11.5

SN.W. Ashcroft and N.D. Mermin, Solid State Physics, Saunders
College, Philadelphia 729 (1976).

SM.A. Omar, Elementary Solid State Physics, Addison-Wesley,
Reading, MA 499 (1975).

discovery, and one of the major problems with almost all investigations has been the

lack of reproducibility of measurements and the appearance of aging effects. These

seem to result from environmental corrosion of the oxide components in various bulk

and thin film forms. The effect in YBCO is the subject of this dissertation.

1.2 Corrosion of Ceramic Superconductors

Several studies have reported environmental effects on 123 superconductors.

It is now well established that upon exposure to water YBCO loses its

superconductivity properties. While examining various aqueous mixtures, several

byproducts were measured using X-ray Diffraction (XRD). BaCO3, YBa2Cu305 and

CuO, were detected following such an experiment as explained by the set of reactions


3H20 + 2YBa2Cu307 Y2BaCuO5 + 3Ba(OH)2 + 5CuO + 0.502

Ba(OH)2 later forms BaCO3 in reaction with CO2 in the ambient atmosphere:

Ba(OH)2 + CO2 BaCO3 + H20

Effects of ambient atmosphere for several days were measured by neutron

diffraction.8 Changes in the diffraction pattern were attributed to absorption of

moisture in air. The degradation was concluded to occur at the grain boundaries

leading to crack growth. High resolution electron microscopy studies on YBa2Cu307,

at several temperatures between 20 and 4500C, and in three different atmospheres

(oxygen, air, and vacuum) have revealed that surfaces decompose rapidly at

temperatures above approximately 150C.9 Decomposition was also found to be

accelerated by oxygen and structural defects with emphasis on the size of free grain

surfaces. Exposure to air for one week resulted in the formation of planar defects.

Formation of additional CuO planes adjacent to existing CuO planes and collapse of

the structure at the surface were considered. It was observed that the diamagnetic

moment which measures the fractional amount of superconducting phase in the

material, decreases gradually with prolonged exposure of YBCO to air. Other TEM

investigations also report similar distortions of the lattice caused by planar

defects.10,11 Bulk samples treated in a humidity chamber to accelerate corrosion

have revealed hydroxide and carbonate formation along the grain boundaries observed

by SEM.12 X-ray analysis also showed the presence of Y2BaCuOx phase, CuO,

BaCO3, Cu(OH)2 and possibly Y(OH)3 and BaO.12,13 Extended analysis of fine

structure (EXAFS) and x-ray absorption near-edge structure (XANES) measurements

also proved that YBCO decomposes to BaCO3 and Ba(OH)2 following water

exposure. 14

YBa2Cu307_S samples subjected to low pressure moisture were examined in

x-ray photoelectron spectroscopy (XPS) and the formation of hydroxide OH- species

and molecular water (H20) were detected.15 These reactants were subsequently

removed by heating in ultra high vacuum (UHV). The changes in the oxygen and

barium spectra were speculated to be due to the formation of BaCO3 and Ba(OH)2.

Ion assisted XPS revealed depletion of Cu and enrichment of Ba following air

exposure for several days.16 The altered layer thickness depended on sample age.

On the other hand water treatment produced enrichment of Ba and Cu over Y keeping

Ba/Cu ratio almost constant and it was explained in subsequent reactions as follows:

I 4YBa2Cu307 + 6H20 -- 2Y2BaCuO5 + 10CuO + 6Ba(OH)2 + 02t

II 2Y2BaCuO5 -* 2BaCuO2 + 2Y203

III Ba(OH)2 + CO2 -- BaCO3 + H20

1.3 Application of X-ray Photoelectron Spectroscopy to Superconductors

Since the discovery of the new family of superconducting materials, many

analytical techniques have been attempted in order to determine the relation between

structure and mechanism of superconductivity. The valence states of Cu and O in

YBa2Cu307_3 have received much attention. This is considered to be important to

understand the Cu O interaction and thereby to lend support to the theoretical

models of superconductivity based on Cu 3d and O 2p. Since XPS has provided

precise information on the valence state of a variety of materials, this technique has

been employed extensively for studies of Cu 2p and O ls core levels in high Tc SC

YBa2Cu307_O It has also been used extensively to determine atomic coordination,

to detect impurities, to establish rough estimates of the composition, to calculate local

charge densities and potential energy curves, and to obtain bond structure

information. Probably the most frustrating fact about these compounds, while

performing these tests, was that they corroded rapidly when exposed to humid or even

ambient atmosphere. Since XPS is a surface sensitive technique, the extension of

results to the bulk structure requires a careful cleaning of surface adsorbates. There

is no generally accepted method for removing surface contamination. Many methods

have been suggested, but each has serious drawbacks. For example, vacuum fracture

commonly practiced in sample preparation for XPS studies has been criticized for

enhancement of grain boundaries at the fracture surface leading to a Ba-rich

layer17'18 except for single crystals; low energy ion milling in situ causes

differential sputtering of the elements, especially oxygen, and leads to incorrect

compound stoichiometries;17,19,20 surface scraping with a diamond or steel file is

the most often used method, yet suffers from the same grain boundary exposure

problem as sample fracture.17,19,21 Heating the sample until the contaminants are

evaporated seems to be the most reliable method for measurement of the bulk


The oxygen vacancy in the defect perovskite layered oxidic compounds,

YBa2Cu307-_ (YBCO), leads to charge anomaly. The analyses of Y 3d5/2 and

Ba 3d5/2 of YBCO samples by XPS have shown that their formal valencies were 3+

and 2+ respectively. If oxygen were assumed to have a 2- formal valency, then Cu

would have a non-integral valency (or mixed valence). For example, if 3 = 0, the

valence of Cu is 2.33 (or Cu2+ and Cu3+ in a 2:1 ratio) while it is 2+ for 6 = 0.5.

Alternatively', if the valence of Cu is fixed to 2+ then some of the oxygens should

have 1- valence. Moreover, the charge neutrality suggests (1-26) holes per unit

cell. This implies that a knowledge of the oxygen vacancy concentration, 6 can be

used for estimating the valence of constituent elements in YBCO compounds. Since 6

also governs the transition temperature, Te, a correlation between superconductivity

(Te) and the valence state of the constituents in YBa2Cu307_3 is desirable. Initially

most of the XPS work was concentrated on the valency of Cu as it was believed that

the extra charge (Cu3+) was the source of the "anomalous" superconductivity. But

the findings were often inconsistent and this debate still continues. X-ray irradiation

damage during the XPS measurement has been reported,22 resulting primarily in a

reduction of oxygen content leaving a more metallic surface. This further complicates

the determination of the valence states of Cu.

Photoemission of the valence states of YBa2Cu307_- was examined to

calculate the electronic structure. However, it appears that interpretations of results

by different research groups lead to different conclusions and controversy. The

available data on valence bands and theoretical calculations have not yet succeeded in

relating the mechanism of superconductivity with the electronic structure of the high

T, superconductors. The complication with the surface impurities affects the valence

band the most due to low intensity. It is therefore natural to turn to the XPS core

level studies of these superconductors.



Evidence of degradation prompted the measurements of surface changes upon

exposure to ambient atmosphere and water. X-ray photoelectron spectroscopy was

used to complement other techniques such as XRD7 to identify the products resulting

from reaction of these ceramics with different conditions of exposure. By purposely

corroding the 123 compounds separation of bulk components and contaminants is

possible upon comparison with the superconducting samples. This also helps to

determine the mechanism of corrosion. The objective of this study can be

summarized under 3 major parts:

1- Identification of surface species of YBCO compounds to help determine

the reaction of bulk components (123 structure) with atmospheric

components (surface reactants, carbonates, hydroxides, etc.),

2- Combining different surface analytical techniques (XPS, sam, ion

assisted depth profile) to study the variation in detected surface species

with depth into the bulk.

3- Understanding the degradation of YBCO by examining samples made

by different processes and exhibiting different structures.

In this study, the aim was to identify corrosion products in two types of

samples (pressed powder pellets and thin films) subjected to air and 85% relative

humidity at 500C. For this purpose XPS was employed extensively for surface and

near surface analysis and Scanning Auger Multiprobe (SAM) was performed to study

depth variations of surface components.

Pellet samples and film samples were prepared in pairs and one of each was

tested for superconductivity while the surface of the other was examined in the

corrosion study. All samples were of the 123 nominal composition and had T, in the

range 80 90K. Samples were analyzed immediately after preparation, following an

ambient air exposure of 18 days and after exposure to 85% relative humidity and

50C in a humidity chamber. The details of the sample preparation and experimental

conditions are given in chapters 4 and 5.



3.1 X-ray Photoelectron Spectroscopy

3.1.1 Theory

X-ray photoelectron spectroscopy or ESCA (Electron Spectroscopy for

Chemical Analysis) is a commonly used surface analysis technique for detecting

impurities, making rough estimates of the composition, calculating charge densities

and potential energy curves, and obtaining bond structure information. The principle

of XPS is very simple. When a sample is irradiated with x-rays, electrons may be

emitted from inner and outer shells of the constituent atoms (Figure 3). The kinetic

energy Ek of the emitted photoelectron is given by

Ek = hv-Eb-O (2)

where hv equals the energy of incident photons, Eb is the binding energy of electrons

(with reference to Fermi level) and 4I is the spectrometer work function. Equation (2)

only holds for compounds in electrical equilibrium with the spectrometer. Electrical

insulators build up positive charge on the surfaces, causing a shift in the kinetic



F.......... ................ .... erm i Level





Figure 3 Photoemission process.

energy of the emitted electrons. The resulting charging depends strongly on the x-ray

flux, the electrical resistivity of the sample, and electron back flux from the x-ray

source and other components in the system. For an electrical insulator the kinetic

energy of the photoelectron will be given as

Ek = hv-Eb-O-Q (3)

The additional Q term takes the sample charging into account. The relative positions

of the photoelectron lines will not be influenced by the surface charging and therefore

the choice of an appropriate reference line of known energy would allow an absolute

energy calibration. Apart from photoelectron lines, Auger lines will also be observed

in the spectrum. The excited ions formed in the photoelectron process can be

deactivated by the emission of a photon, or by the emission of an Auger electron

(Figure 4). In the latter an outer shell electron will make a transition to the vacant

Fermi level--------------------

Shake up
L 3 (a

L 2 T

Augej e


..............---............ ... Fermi level------...........


L 2



Figure 4 De-activation of a photoemission
electron and shake up transition.


prK mitin r--

process by emitting a photon or Auger


inner shell state with the simultaneous ejection of a second outer shell electron.

Three energy levels are involved: the level of the inner shell vacancy (with an energy

El ), the level of the electron which fills up the initial vacancy (E ), and the level of

the ejected Auger electron (E3). The kinetic energy EA(123) of the Auger electron is

approximately given as

EA = E1-E2-E3 (4)

For example in Figure 4,

EA = E-E -EL

In contrast to the photoelectrons the energy of the Auger electrons does not depend on

the energy of the x-ray source. Owing to surface charging, the Auger electrons

emitted by an electrical insulator will be changed in exactly the same waj as the

photoelectron energy. The binding energy and therefore the measured kinetic energy

of an emitted photoelectron will be characteristic of the atom and the activated

electronic orbital which itself is unique to the element in the specimen to which it


3.1.2 Qualitative analysis

Equation (3) can also be written as

Eb = hv-Ek-,-Q


so that the photoelectron spectrum can be plotted as a function of binding energy Eb

expressed in electron volts (eV). Figure 4 is an illustration of an XPS spectrum

obtained using a magnesium anode (hv = 1253.6 eV) for a superconducting

YBa2Cu3O7_3 pellet. The photoemission process is inelastic if the photoelectron

suffers an energy change (usually an energy loss) between photoemission from an

atom in the solid and detection in the spectrometer. The broad background increasing

to high binding energy (or low kinetic energy) is due to inelastic photoemission

(energy loss within the solid). Photoemission by Bremmsstrahlung (white) radiation

dominates this general background in the low binding energy region of the spectrum.

Secondary electrons resulting from inelastic photoemission increasingly dominate the

background at higher binding energies. The peaks observed in Figure 4 can be

grouped into three basic types: peaks due to photoemission from core levels, from

valence levels, and peaks due to x-ray excited Auger emission (or Auger series).

Binding energy line positions of these levels are given in numerical order in Table II.

Other features include plasmons, shake-up peaks, satellites and ghosts. Energy

loss peaks, also called plasmons, result from discrete energy losses associated with a

collective excitation or oscillation of free electrons inside the solid; they are present

on the high binding energy side of every peak. Corresponding to another type of

discrete energy loss, the shake-up peaks are characteristic of only certain elements.

They are typically observed for 2p levels of transition metals with unpaired d

electrons paramagneticc species). For example, for a fraction of 2p electrons ejected

from Cu2+, photoemission will be accompanied by the simultaneous promotion of a

Binding energy positions of the lines produced by
detected by XPS operated under Mg Ka X-rays.


Energy (eV)*



O 2s, Ba 5s
Y 4p
Y 4s
Cu 3p3/2
Cu 3P1/2
Ba 4d5/2
Ba 4d3/2
Cu 3s
Y 3dS/2
Y 3d3/2
Ba 4P3/2
Ba 4P1/2
Ba 4s
C Is
Y 3P3/2
Y 3P1/2
Y 3s

Energy (eV)*


C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder and (ed.) G.E.
Muilenberg, Handbook of X-ray Photoelectron Spectroscopy, Perkin Elmer
Corporation, Eden Prairie, MN (1979).

Ss, p, d levels represent photoelectron and KLL, LMM, MNN levels represent
the Auger lines)

Table II


O ls
Ba 3d5/2
Ba 313/2
Cu 2P3/2
Cu 2P1/2
Ba 3P3/2
Cu 2s
Ba 3P1/2

C -
2 V

S' I I I

1000 800 600 400 200 0
Binding Energy (eV)

Figure 5 Wide XPS scan of a superconducting YBa2Cu307.6 pellet using Mg Ka x-rays.

3d electron to an upper level. The energy required for this process is taken from the

kinetic energy of the outgoing 2p photoelectron and an additional peak will appear on

the high binding energy side of the parent 2p peak. This peculiar type of satellite

peak has practical applications; it is indeed very useful as a fingerprint of the

oxidation and/or spin state of those elements susceptible to giving it. This feature is

clearly evident in CuO,23 which was explored intensely to examine the electronic

density of states,24'25 the oxygen coordination variation and in search for Cu3+ in

superconducting ceramics.26,27 For example the effects of excessive heating19

and Ar etching20 for purposes of surface cleaning were observed as a loss of oxygen

and thereby a reduction in Cu valence, detected by a reduced shake up satellite

intensity and a narrowing of the 2P3/2 line. On the other hand scraping the surface

also reveals the shake up satellite.28 X-ray irradiation damage was detected as

reported before,29 mostly as a reduction of oxygen content leaving a more metallic

surface after 50 min of irradiation by comparing measured spectra to those of Cu,

Cu20, CuO and NaCuO2. In the same work the area under the shake up satellite of

YBCO was compared to that of NaCuO2 to determine the presence of 3+ valency.

The shake up satellite was also identified with the presence of the superconducting

Cu-123 phase3,031 and the absence of this feature was observed in corroded


Standard x-ray sources are not monochromatic. Besides the Bremmsstrahlung

radiation mentioned above and the principal Kol,2 line, magnesium and aluminum

targets also produce a series of lower intensity lines, referred to as x-ray satellites.

The transitions giving rise to Kal,2 radiation (an unresolved doublet) are

2P3/2,112 is for both the Mg and Al targets. Satellites arise from less probable

transitions (e.g. Kfl; valence band -, Is) or transitions in a multiply ionized atom (e.g.


X-ray ghosts are due to excitations arising from impurity elements in the x-ray

source. The most common ghosts are Al Kol,2 from a Mg Ka source. This arises

from secondary electrons produced inside the source hitting the aluminum window

(present to prevent same electrons hitting the sample). This radiation will therefore

produce weak ghost peaks 233.0 eV (1486.6 Al Kao 1253.6 Mg Kao) lower binding

energy than those excited by the dominant Mg Kal,2 radiation.

The core level structure in Figure 4 is a direct reflection of the complex

electron structures in these ceramic superconductors. Magnesium Kao radiation is

only energetic enough to probe the core levels of for example Ba up to the 3p shell.

It is immediately clear that the core levels have variable intensities and widths and

that non-s levels are doublets.

Any orbital that does not possess spherical symmetry (i.e., for any orbital

quantum number I different from 0) will appear as a doublet. Because the coupling

between 1 and the spin magnetic moment s equal to + /2 or V2 lifts the degeneracy.

Any subshell with I ; 0 therefore splits into two levels characterized by the quantum

number j = 1 Is\, where \sl = 2h which gives rise to doublets in the XPS

spectrum33 such as Ba 3d5/2 and Ba 3d3/2. These notations are summarized in

Table III.

Table HI X-ray and spectroscopic notation.

Quantum numbers

n I j

1 0 'A
2 0 h
2 1 2
2 1 %3
3 0 I2
3 1 1/2
3 1 %3
3 2 %
3 2 %5
i "

X-ray suffix

X-ray level





The magnitude of the spin-orbit splitting depends on the element itself and on

the subshell, due to parallel or anti-parallel nature of the spin and orbital angular

momentum vectors of the remaining electron. The difference in energy of the two

states, AE is proportional to the spin-orbit coupling constant t which depends on

the expectation value (1/r3) for the particular orbital. The separation can be many

electron volts. Therefore AE, is expected to increase as the atomic number (Z)

increases for a given subshell (constant n, 1) or to increase as I decreases for constant

n (e.g. Table II shows for barium that the splitting of 3p(75 eV)> 3d(15 eV) ).

The relative intensities of the doublet peaks are given by the ratio of their

respective degeneracies (2j + 1). Thus the area ratios and designations (nl ) of spin-

orbit doublets are given in Table IV.

The peak width, defined as the Full Width at Half-Maximum (FWHM) AE, is

a convolution of several contributions

AE = (AE2 + AE +AE )12

where AEn is the natural or inherent width of the core level, AEp is the width of the

photon source (x-ray line) and AEa the analyzer resolution, all expressed as FWHM,

with the assumption that all components have a Gaussian line shape.

The valence band is readily observable close to the zero binding energy. The

core levels of all the elements present in the sample, are measurable if their binding

energy is less than the energy of the incident x-rays.

3.1.3 Chemical Shifts

Peak assignment is easily done with the help of a binding energy table.

Indeed, the chemical process of valence orbital identification manifests itself only as a

first-order perturbation on the binding energy of the core levels. In different chemical

compounds the binding energy of inner shell electrons may vary due to differences in

nearest neighbor chemical or structural configuration. This perturbation is the

chemical shift which has an order of magnitude of 1 10 eV (to be compared with a

total binding energy of 50 1000 eV). It can be detected and used for structural

Table IV Area ratio of the spin-orbit splitting doublets.

Subshell j Area (2j + 1) ratio*

s 1/2
p 1/2, 3/ 1:2
d %,/2 2:3
f /2, /2 3:4

D. Briggs and J.C. Rivibre, Practical Surface Analysis by Auger and
X-ray Photoelectron Spectroscopy, ed. D. Briggs and M.P Seah, John
Wiley & Sons Ltd, New York 113 (1983).

analysis. Immediate qualitative elemental analysis is therefore rather straightforward

with XPS; cases of peak overlap from different elements are relatively rare and easily

overcome because they rarely affect more than one core level of an element at a time.

Common sense in the examination of XPS spectra for the presence and absence of the

other core levels of the suspected element with their proper intensity ratios helps solve

virtually any potential ambiguity. A simple model that has had some success in

interpreting chemical shifts is the charge potential model.34 The binding energy of

a core level in atom i surrounded by other atoms j is given in the model by

Ei = E+ kq+E q/r,

where E ? is a reference energy, k is a constant, qi is the valence charge on atom i,

and ri. is the radius of the sphere centered on i and passing through thej atoms. The

last term can also be written as Vi, since for an ionic solid the summation extends

over the whole lattice and therefore Vi in that case is the Madelung potential. A

change in Ei due to a change in the valence charge qi can thus be written as

AEi = kAqi + A V

The chemical shift AEi is therefore related to both the change in the valence charge

and the change in the potential due to the surrounding shells of atoms, and since the

two terms on the right-hand side do not necessarily have the same sign, it is possible

for changes in chemical state to produce negligible chemical shifts. Most oxidized

metals, however, are subjected to distinct chemical shifts compared to pure metals.

This can lead to useful structural and chemical information about the specimen.

3.1.4 Quantitative Analysis

The number of photoelectrons with a given energy reflects the concentration of

a given species with a certain type of bonding. The current of photoelectrons of

kinetic energy Ei excited from the core level X of the i th atomic species in the

surface of a sample may be written as35

I(Ei,) = KJ(hv)Ni oi (hv, 0,X)X(Ei )R(Ei,Ea )T(Ea )D (Ea )G

where K is the proportionality constant, J(hv) is the x-ray flux incident on the sample

at energy hv, Ni is the density of atoms of the i th species, ai (hv,0,X) is the cross-

section for photoionization of level X at the x-ray energy hP and at the ejection angle

0, X(Ei) is the inelastic mean free path for electrons of energy Ei, R(Ei,Ea) is the

electron-optical factors for retardation of electrons from energy Ei to the analyzer

pass energy Ea, T(Ea) is the analyzer transmission function for energy Ea D(Ea) is

the efficiency of detection for energy Ea and G is a geometric factor. Under normal

experimental conditions the x-ray flux would be kept constant, as would the detection

efficiency and the geometry, in which case the above equation reduces to


The photoionisation cross-section is given by35

ai (hv,0,X) = oi (hv,)F(O,X,hv)di

where oa (hv,X) is the total photoionization cross-section, F(O,X,hv) is the angular

asymmetry factor, and dQ is the solid angle of acceptance of the analyzer. The

photoionisation cross-sections for different conditions have been calculated by

Scofield.36 Angular asymmetry factors however vary between instruments

F(O,X,hv) = 1 + P ((X,hp)( sin2m -1)

where f is the asymmetry parameter. Values of f have been calculated by Reilman et


There is considerable debate in the literature regarding experimental values for

inelastic mean free paths and the theoretical development of such values. In an

earlier work done by Penn, for kinetic energy Ei > 200 eV it was derived to be38

X(E,) = Ei [a(lnEi + b)]-1

where a and b are parameters.

The expressions derived by Seah and Dench,39 based on the compilation of

many experimental measurements from the literature, are, for elements

X(Ei) = 538Ei-2 + 0.41(aEi)12

and for organic compounds

X(Ei) = 2170Ei-2 + 0.72(aEi)12

where X is measured in monolayers and a is the monolayer thickness.

Wagner et al. on the other hand, have shown that most of the available

experimental data can be interpreted as the power function40

X(E,) = CEim

where C and m are functions of the solid material. They have shown that for kinetic

energies greater than 300 eV, m is approximately 0.75 for inorganic solids and is in

the range of 0.7 to 1.0 for organic.

Instrument transmission functions have been considered by Seah with reference

to XPS quantitation.41 The transmission function for a particular instrument is

usually available from the manufacturer also. Thus using the Scofield cross-sections,

the asymmetry parameter P, and the instrument transmission function, one can

calculate an overall sensitivity factor for each element and orbital, for each instrument


Given the appropriate values for a and for the instrument transmission

function, coupled with reasonable estimates of free path dependencies on kinetic

energy for the sample of interest one can perform a rigorous elemental analysis for

samples that are, or assumed to be homogenous to depths much greater than the

intrinsic mean free path. This can be done in two ways.42 Firstly, a comparison of

the intensities of photoelectron currents from the same core level of a given element in

different materials can be used. Since the same photoelectron energy is being

measured each time, the cross-section a will remain the same, and so will the

instrumental parameters that depend on kinetic energy, so that one can write

I(E) N1X1(E,)
,2(E) N22(Ed)

where the subscripts 1 and 2 refer to two different materials. Thus if the inelastic

mean free paths are known for the energy Ei in each material, the measured

intensities can be used to obtain relative elemental atomic concentrations.

Secondly, a comparison of the intensities of core level emission from different

atoms in the same materials can be done, which is the situation normally studied.

Since different kinetic energies are involved, the expression is more complex, and the

ratio of intensities of emission from two different atoms can be written

11(E1) Noa ,RT,
12(E2) N2o2X2R2T2

If the Penn expression for X is used the expression simplifies to

I,(Ej) Na,(lnE,+b)RT,
12(E2) N2a2(lnE +b)R2T2

In the latter case it is not necessary to know absolute values of the inelastic

mean free paths, but the cross-sections and the instrumental parameters need to be


In practice quantification is often attempted by simply measuring the intensities

of chosen photoelectron peaks from pure elements, and assuming that the intensity of

the same peak for the same element in an alloy or a compound is related to the atomic

concentration of that element in proportion to the 'pure' intensity. This is the method

of relative elemental sensitivity factors so widely used. In other words, if I~" and

N1" are the intensity and density appropriate to the pure element, then in any other


N," I,
N1 I/

so that in a multi-component situation the assumption is that the atomic fraction of

element 1 in a sample is given by

X1 -

j 1

For a binary system of elements 1 and 2, this leads to

I I =N, IN
12/12 N2/N:

Whereas without the simplification, the intensity ratios would be

11/f NIN I 1/2 I, 2 N1N; 172
12-12 Nz/N2 22X 1 or 2 ,,
1211: N2/N: )12/X:j If NfN2 X712

However, the assumptions based on the use of pure elemental sensitivity

factors can lead to serious errors.

3.1.5 Instrumentation

A typical XPS instrument consists of an excitation source, sample mounting

assembly, an analyzer, a detector, and a vacuum enclosure for these items. Also

included are an electronic data processing, display and output units (Figure 6). As

discussed before an x-ray source is needed to excite electrons from the sample

examined. These electrons are then collected, sorted, amplified and transmitted to

data processing unit by an electron detection system. The whole system runs under

vacuum conditions ranging from 10-7 to 10-11 Torr. In addition to these basic parts

other tools can be installed to improve analysis. These might include

1. Dual or quad anode x-ray source with targets other than Mg Ka. Al Kct and

Zr Ka are the most common additions. This makes it possible to measure low

intensity or impurity peaks and core levels with binding energies greater than

the Mg Ka photon energy, and to separate Auger from photoelectron lines if

they happen to overlap in Mg Ko radiation.


Figure 6

A schematic of a basic XPS instrument.

2. Ion gun to remove surface adsorbates and to perform depth profile studies.

This is achieved by bombardment of a beam of positive noble gas ions of

energies between 500 eV and 5 keV, typically, directed at the surface.43 As

a result of the exchange of energy in the surface and subsurface regions, some

atoms or clusters of atoms at the surface are given enough kinetic energy to

leave the surface. In other words, the surface is eroded. The gas commonly

used is argon, chosen as a compromise between efficiency of removal of

material, which increases with atomic weight and vapor pressure at the

temperature of liquid nitrogen trap, if in use; the heavier noble gases tend to

condense at low temperatures and be difficult to pump away. The process of

erosion is called sputtering or etching and the source of ions is called an ion

gun. Although ion bombardment is used universally for surface cleaning,

either by itself or in combination with heating, the potential user should be

aware of the artifacts that can be introduced by it. These take the form of

either or both chemical or topographical changes induced in the surface.44

Chemical changes here represent the alterations in the elemental composition

or chemical state, or both together, of the major constituents that can occur at

the surfaces of alloys and compounds during ion bombardment due to different

sputtering yields. Topographical changes are usually in the direction of the

increased surface roughness, or, occasionally, of statistically induced surface

roughness. Sputtering is also used as a method for obtaining compositional

information as a function of depth below the surface. When used with a

technique, such as XPS, the information is built up stepwise, i.e. by alternate

cycles of bombardment and analysis. This technique is called depth profiling

(section 3.2). All the artifacts that may be introduced by cleaning the surface

by ion bombardment are of course present and are equally important during

continued bombardment to produce a depth profile; indeed their effect may

increase with depth.

3. Temperature controlled sample holder capable of heating or cooling. This is

particularly important for surface cleaning. Several elemental materials,

including silicon, and the refractory materials, can be cleaned by heating them

to sufficiently high temperatures, generally a few hundred degrees Celsius

below their melting points. It can also be valuable if measurement of

temperature effects is desired.

4. Scraper is usually used to remove the top surface layer by mechaftical means.

It is generally a razor blade, diamond tip or stainless steel file mounted on a

rod extended to the vacuum assembly. Soft metals such as lead, indium and

tin, and most superconducting ceramics (YBCO) have been successfully

cleaned by scraping.28

5. Sample preparation chamber where mechanical (surface scraping), ion sputter

or heated cleaning can be performed to prevent contaminating the analysis

chamber. This chamber can also be used for sample fabrication or surface

preparation such as gold decoration for charge calibration which will be

discussed later.

6. Multiple sample holder carousel which allows unattended data collections for

more than one specimen (as many as 10). This increases the efficiency of data

collections considerably, since no additional pump down period is needed

between samples and the automatic acquisitions of data allow the use of the

instrument more frequently, expanding the operational time and generating

more data for a given period of time.

7. Manipulator for changing the position of the sample. Since the sizes and

shapes of samples can vary the optimum position in terms of sensitivity and

energy resolution for the surface analysis of one sample will not necessarily be

optimum for the next. Manipulator also makes it possible to move the sample

laterally in a reproducible way to allow different areas to be analyzed. It is

often desirable to be able to rotate the sample about an axis through its

surface, again in reproducible way, either to alter the angle of incidence of

primary ions during ion bombardment or to alter the angle of take-off of

electrons accepted by the analyzer, for purposes of profiling by variation of

escape depth (angle-resolved-XPS, section 3.2.1). It may also be necessary to

adjust the sample position for other surface science techniques available inside

the same chamber. These manipulators can be manual or computer controlled,

or a combination of both. A computer controlled manipulator makes angle

dependent analysis very practical, where the data are collected automatically.

8. Complimentary surface techniques, such Ion Scattering Spectroscopy (ISS),

Secondary Ion Mass Spectroscopy (SIMS), Auger Electron Spectroscopy

(AES), Low Energy Electron Diffraction (LEED), etc., are possible in the

same analysis chamber. This permits comparison of data without having to

expose the sample to air.

9. Networked computers make data processing possible from remote terminals.

10. Flood gun directs a beam of low energy electrons to the surface of an insulator

to compensate the positive charge built up. This is particularly important in

monochromatic systems, since white radiation is lost and the lack of

Bremmsstrahlung radiation can induce substantial charge shifts. As the sample

begins to charge, electrons produced by the flood gun are directed and

attracted towards the sample surface, flooding the surface with low-energy

electrons which provide charge compensation. The electron flux is usually

adjusted to drive the shifted peak to expected binding energy position.

Occasionally, because of heterogeneous charging and other considerations, it

has been found to be more optimum to overdrive, that is to place the surface at

a net negative surface potential, generally 6 to 7 eV negative. This results in

sharper, more well-defined peaks and better separation of chemical bonding

features.45 This probably works because the process results in a surface

potential, minimizing heterogeneous surface charge effects.

11. X-ray monochromator can be used to improve energy resolution of XPS lines.

It eliminates satellite interference, improves signal-to-background ratio by

eliminating the Bremmsstrahlung continuum and the selection of an individual

line from the unresolved principal line doublet can be achieved. Typical

resolution with a conventional x-ray source (Mg Ka line width 0.75 eV) on

Ag 3d5/2 peak is 0.8 1.1 eV, with a monochromatic Al Ka (line

width = 0.4 eV) this becomes -0.3 eV. The use of monochromator with an

insulator however emphasizes the charging problem due to reduced intensity of

the photon flux and the loss of bremmsstrahlung radiation, which in turn

induces peak broadening. Thus the use of a flood gun becomes crucial to

recover resolution.46

12. Small area analyzer on the other hand increases the surface spatial resolution

so that smaller regions of the surface can be examined. This is achieved by

focusing the x-ray source and using smaller apertures in the analyzer input

slits. This apparatus which has recently been introduced to the market, may

enable XPS to achieve spatial discrimination in surface analysis (as in scanning


13. Fracture stage on the sample holder makes it possible to probe the bulk

structure. This is especially useful for surfaces reactive with ambient air.

Some materials such as alkali halides, silicon, germanium, etc., will cleave

easily along well-defined cleavage planes, so that all that is required to

produce a clean surface with a composition close to that of the bulk is some

mechanical ingenuity in the method of holding the material, cleaving it and

offering it to the energy analyzer for analysis. In other materials such as

polycrystalline metals and alloys, which are too ductile at room temperature to

cleave, fracture may still be possible by lowering the temperature to the point


at which they become brittle. The fracture path may then either pass through

the individual grains (trans-granular) or follow the grain boundaries (inter-

granular); in either case a surface free from contaminants is produced,

although the inter-granular surface may contain certain impurities segregated

from the bulk.

14. Multichannel electron detector can speed up the data collection several fold

compared to single channel detectors due to higher count rates.

In this work both of the XPS instruments (KRATOS XSAM-800 and Perkin Elmer

PHI ESCA 5000) in the Department of Materials Science and Engineering, at the

University of Florida, were utilized.

The Kratos XSAM-800 is equipped with, in addition to basic parts, a dual

anode (Mg and Al) x-ray source, temperature controlled (both cooling add heating)

sample holder, sample preparation chamber with a manual linear motion rod that can

be modified as a scraper, multiple sample holder carousel capable of holding ten

specimens 1.5 cm diameter in size, a manual XYZ and rotator manipulator

(Huntington) in the sample analysis chamber, a flood gun. Although it is capable of

performing SIMS, it is not operational at this time.

The Perkin Elmer PHI ESCA 5000 has the following options, a dual anode

(Mg and Zr) x-ray source, an Ar ion gun for cleaning purposes (not powerful enough

for depth profiling), a sample holder for heating only to clean surface adsorbates (no

temperature measurement available), a manual Z-direction manipulator, and a

computer controlled rotator (for angle resolve XPS analysis-section 3.2.3), ISS


The detection performance, i.e. the count rates and energy resolution of both

instruments differs to the extent of their maintenance rather than the specifications of

their components. They are more suitable for performing different tasks. Thus,

depending on the information sought it may be more appropriate to chose one over

the other. For example, PHI is better for angle resolved XPS and ion sputter cleaned

surfaces, and Kratos can be more useful if numerous samples are needed to be

analyzed in a short period of time and temperature effects are needed to be

investigated. Data analysis differences also exist. The advantages of each can be

listed as follows; to get structural and chemical information, x-ray induced Auger line

shapes can also be investigated.47 This can be complicated if more than one shape

i.e. species are present. To separate and measure the abundance of each, one has to

curve fit the envelope of the peak using the line shapes of the individual components.

However Auger components usually do not have a Gaussian or Lorentzian shape that

is available in common curve fitting routines. To overcome this difficulty, DS800 in

Kratos will let the user save a real spectrum (e.g. a single component) and use that

line shape while curve fitting a multi-component spectrum. This procedure was used

recently to study the fluorine bridging statistics in CdPbF4 glass systems

(Figure 7).48 PHI however, has a very fast curve fitting routine for photoelectron

lines, which makes this usually tedious job, as discussed in section 3.1.7, easier.

Comparing the spectra of different samples is often necessary to inspect changes.

. Binding Energy (eU)

8 59 (
Binding Enersg (eU)


602 600 598 596 594
Binding Energy

Figure 7 The use of Auger line shape to deconvolute F KLL spectrum in
measuring the concentration of Pb-F-Pb, Pb-F-Cd and Cd-F-Cd
type bridging fluorine.

PHI will allow up to four different spectra, whereas DS800 will allow only two.

DS800 is able of doing quantitative analysis (including curvefits) automatically to 25

separate acquisitions, provided they don't have more than 1 2 eV chemical or

charging shift compared to PHI's only one acquisition. DS800's command line

interface makes it easier to write macros to speed up the spectrum massage process.

3.1.6 Data Collection

In this section emphasis will be given to the initial measurements of a new

substance before starting a study of surface effects using XPS. The steps reviewed

here will help reassure the reliability of the data collected. The results of a

measurement can easily be misinterpreted if the necessary precautions are not taken. Instrument Settings

The following tests are a good starting point for finding the optimum running

conditions for the best resolution.

1. Resolution (Pass energy): Number of channels used per energy step

to collect data. Controlled through the software while setting up a run. "Kratos has 3

choices (HI, ME, LO) HI usually gives the best result for chemical shift

measurements. PHI also has 3 levels of setting (survey, utility, high), however the

pass energy can be adjusted for a particular resolution. The pass energies varies from

189.95 eV to 4.45 eV.

2. Magnification: Size of the region to be examined (spacial

resolution). This option is only available on Kratos. It is particularly important if the

sample used is smaller than the holder, to expose only the region of interest. HI

magnification (smaller region) improves the FWHM slightly.

3. Mode: Electron detection method. There are two modes; 1. FRR

(Fixed Retarded Ratio) in which electrons are retarded by the same amount

throughout the energy spectrum. This is used to obtain more accurate concentration

calculations. 2. FAT (Fixed Analyzer Transmission) where electron transmission is

kept constant by changing the ratio with respect to Kinetic Energy thereby increasing

sensitivity for low KE electrons. These are set in the acquire set up menu of the

Kratos. PHI on the other hand always acquires in FAT mode, but in the data

massage one can edit the spectrum to FRR by normalized to E option.

4. X-ray source: Mg Ka and Al Ka (Kratos) or Zr La (PHI).

Aluminum and Zr are more intense but have a broader line width than Mg.

Magnesium is the most often used x-ray source mainly because of its narrower line.

Higher peak resolutions usually requires more intense Al x-ray targets, especially

when monochroinators are used. Aluminum Ka (hp = 1486.6 eV) and Zr La

(hp = 2042.4 eV) are used to detect electrons with binding energies greater than

Mg Ka radiation (hp = 1253.6 eV). The separation between the x-ray source and

sample should be optimum for obtaining maximum count rates. This can be achieved

by adjusting the height of the sample while continuously measuring count rates at a

fixed energy (usually the energy of the orbital to be examined). 300 watts (15

kV x 20 amps) is the most common setting to power the x-rays used on both


5. Metal Standard: X-ray photoelectron spectroscopy analysis of metal

foils such as silver and gold should be performed frequently to check the work

function 4P and peak resolution. The work function can be measured by first setting

the work function to zero and subsequently acquiring a known spectrum (e.g. Ag 3d).

Then the difference between its measured and known position is the work function. It

is advisable to inspect the work function before using conductors and it is not as

crucial for insulators, since charging of the surface will shift the spectrum even if the

work function is accurate. Copper and gold combination is also used to inspect the

relative positions of the peaks.

6. Slit: A mechanical slit setting is available on Kratos. It is used to

improve the peak resolution by limiting the peak intensity. Four different slit

openings were available which are set manually on the instrument. All of them were

tested with various combination of the above mentioned mode of operation and was

concluded that the widest opening (default) was the most reasonable since any other

setting reduces the intensity drastically.

7. Step size: Energy increment to be used while scanning an energy

window. This determines the density of data points. Both instruments allow the user

to choose any size desired. 0.05 eV is most common for chemical shift analysis and

0.3 0.5 eV is for qualitative (survey) study.

8. Flood gun: Low energy electron gun used for surface charge neutra-

lization. Although some literature and my past experiments have indicated that the

use of a flood gun can improve the resolution, it should be noted that there is a debate

in the literature about whether or not the effect of this beam of electrons is known and

whether or not the gun should be used. Users need to test the flood gun before using


it in order to find the optimum setting, because the charging varies from sample to


9. Electron detector saturation voltage: The potential applied to the

detector to amplify the counts must be increased periodically to its saturation point to

achieve maximum possible intensities. Irradiation Effects

Almost every surface analysis technique has shortcomings. XPS has always

been known to yield a high sensitivity to the upper layers of the surface (top

0.5 3 nm or -2 to 10 atomic layers) and to be relatively non-destructive.

However, some compounds can be sensitive to x-rays and irradiation damage can

cause structural and stoichiometric changes. Thus this step is important especially

with materials for which no consistent data are available in the literature. The

software packages which are used in depth profile analysis display data collected

between layers removed by energetic ions thus allowing an examination of the depth

induced. This procedure can be used to inspect the effect of radiation. A run can be

set up so that data are again collected periodically but the power of the ion gun is

turned off. Recording the x-ray exposure time, one can find the safe limits for XPS

 Survey (Wide) Scan

Samples are examined for any kind of impurity and oxidation state (hydration

or carbonation) using a low resolution and maximum sensitivity spectrum (e.g.

Figure 4). The fingerprint spectrum of every element is available in the

handbooks.49 A comparison, starting from the most intense line in the data,

unambiguously helps determine the elements present in the sample of interest. Standardization

Before studying mixed species, one has to determine the standard. The

response of a certain element at a specific settings can be used to deconvolute the

mixed species. The standard consists of binding energy position with respect to an

internal (e.g. metal core level in a metal oxide) or external (e.g. C Is) calibration

peak, FWHM, the separation and area ratio of a doublet, and line shape." The

standardization test must be repeated if a different instrument is used or any condition

on the instrument is changed. Data Manipulation

1. Charge calibration: There are several procedures for compensating

surface charging in insulating samples. Most common is measuring and correcting

the energy of the C Is peak (known to be hydrocarbon or surface adventitious

carbon), which is present in every spectrum unless it is purposely removed by

cleaning techniques. In the event that there are various carbon species, then a curve

fit with the lowest energy assigned to CHx can be employed. The energy of CHx has

been taken as 285.0 eV or 284.6 eV and occasionally 284.0 eV. Since the rest of the

energy measurements or assignments are relative, any of the three can be used as long

as it is reported with the data. For cleaned samples where this carbon is removed, an

internal peak can be used for calibration and metal core levels lines sometimes can be

used for this purpose, given that they are not shifted relative to the peak of interest.

A gold thin film deposited on the sample (gold decoration) can be used by assigning

the Au 4f7/2 to 83.8 or 84 eV.50 Some problems with the use of gold as a standard

have been reported,51,52 thus caution should be taken while applying this method.

A flood gun can also be used for charge neutralization purposes. It is essential in a

monochromator, provided that it is modified properly.

2. Data Smoothing: Type of smoothing routine. The most often used

technique is a 2nd degree least squares central point Savitsky-Golay53 smoothing

routine. Smoothing raw data may produce unrealistic results if the number of points

for the smoothing routine is not chosen appropriately. This number depends on the

FWHM of a single species peak and the step size (energy increment). Displaying the

smooth data over the raw data for different numbers may help determine the best

choice. One method suggested calculates the number by 0.7 x number'of points

(channels) across the FWHM (eV) or 14 x FWHM (eV) if 0.05 eV is used for step


3. Deconvolution: X-ray line shape subtraction used to remove the

peak broadening caused by the source. The several combinations of parameters

(number of iteration, number of points) should be tested for a suitable selection before

applying this technique. It usually does not require drastic changes to improve the

investigation of peak shape.

4. Background subtraction: Elimination of background to get improved

peak shape measurements. Separate methods are available with the Kratos (linear,

Shirley, Tougaard) and the PHI (linear, integral) software. It is generally

indicated55,56 that linear subtraction is the best choice for high resolution spectra.

Besides choosing the type of subtraction algorithm it is necessary to insure that the

derived background has a slope representative of the data. This can be achieved often

by fixing the end points where the tails of the peak meet the background. For low

resolution survey scans acquired in FAT mode integral subtraction should be used. A

new formulation has been reported, which was based on the effect of electron

scattering.57 More recent work by Tougaard has shown that the peak background

can also be used for an in-depth concentration profile.58

5. 2n Derivative: A method for estimating positions of overlapping

peaks.59 When a spectrum with adequate signal-to-noise intensity is produced it is

generally possible to differentiate it numerically. The second derivative is calculated,

which is often helpful in determining the number of components in a peak, the

relative intensities and the binding energy position for each component.60 It is clear

that differentiation of a multi-component spectrum can lead one to deduce the

appropriate subcomponents of the spectrum. The data must generally be smoothed

between each differentiation in order to eliminate the development of random noise

into distinct peaks. Given a second or even fourth derivative spectrum, one can often

pick out the binding energy of appropriate subcomponents. This binding energy can

then be specified in the curve resolution routine. The FWHM can also be specified,

given the data for reference compounds run on the same instrument. Derivative

spectra cannot give accurate peak positions since one is examining the envelope of

overlapping peaks so that the observed maximum of each peak is always shifted by

the presence of other peaks. Nevertheless derivative spectra provide useful

information. In general the most accurate peak position is observed when

differentiating convoluting intervals similar to the FWHM of the component peaks,

though there is a need to balance convoluting interval with resolution. In addition to

slight errors in peak position, the relative intensity of the peaks is only approximate.

This is due to the cancellation of effects which occur when the positive lobe of one

component peak overlaps with the negative lobe of the other peak. In general this

effect will depend upon the number of component peaks, their widths and intensities.

6. Curve Fitting: This is the most common method for XPS analysis.

It is used to interpret and quantify measurements by deconvoluting the overlapping

separate structure peaks and by changing the line shape, position (B.E.), intensity,

and FWHM. Theoretical and experimental work on the binding energy measurement

for elements in different chemical states, peak shapes, background and satellite

subtraction has allowed a realistic analysis of complex electron spectroscopy data.

Employing this information, users have deconvoluted XPS spectra with

Gaussian-Lorentzian peak shapes.61,62 This technique has proven useful mainly

because it is straightforward. Its main disadvantage is that the result is not

necessarily unique. Because of the uncertainty of each guess, this process can take a

long time to reach convergence as it involves determining the number of peaks in a

spectrum, width of each component and combination there of. Attempts have been

made to speed up this routine by applying "Damped Non-linear Least Squares"

technique.63 Recent considerations of asymmetrical line shapes have proven to be

an improvement for realistic curve fitting.64

Curve fitting routines supplied by the manufacturers are usually adequate for

basic tasks. The quality of the fits can be judged by the accuracy of the fits and by

comparison of the envelope of the summed peaks to real data profiles. Manual or

Auto fit can be done. When autofit is chosen, one has to make sure that B.E.

positions and FWHM of the peaks change the same amount towards a better fit. This

can be done in the setup screen. The only problem with autofit is that the final fit

depends strongly on the initial set of peaks. Therefore setting up expected intensities

before the routine begins helps speed up the process and the result is more realistic.

7. Factor Analysis: Factor analysis and other pattern recognition

techniques were originally employed in psychometric research and during the last

decade their use became widespread in many scientific areas. Chemical applications

have usually involved spectroscopic data which generate a large number of peaks.65

In the early stages it was applied to Nuclear Magnetic Resonance, Infrared and Mass

spectroscopies with large sample sets involving multi-element analysis.66 Target

Factor Analysis (TFA) was first introduced to surface analysis by Gaarenstroom who

applied the technique to Auger Electron Spectroscopy to interpret chemical state

information, depth profiles and interfaces.67'68'69 It was later applied to XPS to

measure, x-ray exposure effects70 and through angle-resolved XPS oxidation of

tungsten by different chemical treatments.71 In the latest work this technique was

applied to time dependent analysis of oxidation of chromium using both XPS and

AES.72 Factor analysis is a mathematical method of decomposing a data matrix

into a product of two smaller matrices, one matrix consisting of a set of linearly

independent basis vectors (components or factors) and the other matrix containing the

relative magnitudes of each basis vector present in each data vector (factor loadings).

Thus the original data can be regenerated by the linear sum of products of basis

vector points. Rather than reviewing the details of the mathematical procedure that

leads to Principal Component Analysis, a criterion for this technique will be given.

The original data are represented by a matrix D with i rows and j columns. i is the

number of spectra acquired on a particular sample during treatment (chemical, x-ray

exposure, ion-sputter depth profile, etc.), and j is the number of data points

(channels, binding energy, kinetic energy, etc.) per spectra. The first constraint for

this method is that the number of independent chemical components (m) which

contribute to the set of data that is embodied by D should be smaller than both i and j

or D is a matrix of rank m. The next step is to diagonalize the matrix A = DDT

The number of non-zero (+e) eigenvalues will then denote the number of independent

components, if the spectrum is noise free or e is less than some limit. The upper

bound of e is the standard deviation of the data.68 By trying different ways of

plotting the eigenvectors the original spectrum is reproduced, leading to quantitative

information of the components.

Although the procedure seems almost straightforward, the scientists who

applied this method have cautioned the interpretation of factors (eigenvalues) and PCA

(plot of eigenvectors).73 Kratos Analytical (makers of XSAM800 XPS) don't have

a software package for PCA. Perkin Elmer (makers of PHI-ESCA 5000 and PHI-

SAM 660) however is well equipped. Perkin Elmer scientists have taken advantage

of factor analysis and written algorithms on MATLAB (a programmable advanced

scientific mathematical software package for UNIX based Apollo and IBM-PC DOS

machines) to do PCA.74'75'76 The software can be used directly from PHI-

ACCESS menu which runs the instruments. Utilizing this package from PHI has

many advantages over creating an implementation based on the theory unless one is

experienced in this area. Most importantly the program can readily use the data

acquired by a PHI machine. Aside from PCA or TFA, it has other data massage

routines that are not available from the regular ACCESS software, MATLAB can also

be used independently on Apollo's which will allow other spectral data to be analyzed

and modelling that involves complex mathematical calculations can also be performed.

The difficulties in transferring data and the inexperience in this subject makes PHI-

MATLAB the best choice, provided that it is available.

3.2 Depth Profile

3.2.1 XPS with Ion Sputtering

X-ray photoelectron spectroscopy is a relatively nondestructive technique

compared with other methods of surface analysis. No surface species are removed

during XPS measurement, and the soft x-ray source used for excitation avoids many

of the problems associated with thermal degradation of sensitive materials. Sample

composition remains constant during analysis, allowing extensive characterization of a

surface layer 5 30 A thick without significant time-dependent effects. However, for

certain applications, a significantly broader scope of information (e.g. probing into the

bulk) can be obtained by coupling XPS with destructive tools, or techniques such as

SIMS, Rutherford Back Scattering (RBS). To obtain depth concentration profiles of

elements distributed in layers parallel to the surface, ion sputtering is used in most

surface analysis methods (e.g. XPS, AES) as an in situ tool for exposing the

underlying layers of the solid. Though AES has been used extensively for depth

profiling, this technique does not normally provide information about the chemical

state of the surface elements; in this respect therefore, XPS can provide more

information. Photoelectron peak intensities for different elements can be readily

normalized and correlated by using theoretically based and empirically verified

relative intensity factors (section 3.1.4), whereas Auger peak intensities are correlated

from a theoretical basis requiring the application of calculated correction factors to the

observed intensities. Furthermore, the measurement of XPS peak intensities is

straightforward since the number of electrons observed, N(E), is obtained directly

from the peak-to-peak amplitude of the derivative signal. However caution is

required since this method is subject to significant errors if the peak shapes changes

due to chemical reaction during the measurement.

As discussed before in section 3.1.5, caution should also be exercised in

chemical interpretation of XPS data following ion bombardment since the ion etching

process itself can cause changes to the surface being examined. For example, the

bombardment of an oxide surface by Ar+ ions can cause chemical reduction of the

surface to a lower oxide level or to the metal.44 This process can be used

advantageously to observe intermediate oxidation states as a higher oxide is reduced to

its metallic form. However, in cases where the true distribution of the oxidation

states is desired, care must be taken in using ion sputtering.

The successful use of ion sputtering as a complimentary tool to surface

analysis by XPS requires an understanding of the physico-chemical effects that are

associated with ion bombardment. These effects include differential sputtering of the

surface species and the incident ions, and gross heating of the surface due to the

transfer of energy from the incident ions to the sample.

3.2.2 AES with Ion Sputtering

Auger electron spectroscopy combined with ion sputtering is probably the most

common technique utilized to profile elemental concentrations with depth, especially

in thin films. Although it has disadvantages compared to XPS as discussed above, in

some other aspects it possesses unique characteristics. The most significant being the

ability to analyze a small area of the sample or providing better spatial resolution than

XPS. For example, this enables the analysis of craters formed by ion beam exposure.

In-depth profiling is accomplished by continuously sensing the elemental composition

of the crater bottom during sputter erosion. Simultaneous ion bombardment and

measurement are possible, since the number of secondary electrons created by the

electron beam is small compared with the secondary electron current.77 A static

pressure of 5 x 10-5 Torr Ar is required to operate the sputter ion gun"and it has a

negligible effect on both XPS and AES.

A mtjor advantage of the simultaneous ion etching and AES detection over a

sequential process in which the ion etching is stopped during the Auger measurement

is that surface contamination is greatly reduced. In typical profile measurements, the

surface is sputter eroded at a rate of several atomic layers per second. If the partial

pressure of active gases can be maintained at 10-7 Torr or below, the maximum

arrival rate for surface contaminants is 0.1 monolayers/sec. Under these conditions,

the concentration of surface contaminants cannot exceed a few percent of an atomic

layer. When active gases are removed from a static background of Ar atoms by a

liquid nitrogen cooled Ti sublimation pump, the surface concentration of residual gas

impurities can be maintained at a considerably lower level.

Several factors affect the depth resolution of the AES/ion sputtering profiling

technique. These include ion beam uniformity across the sampled area, the Auger

electron escape depth, and sample inhomogeneity. Loss of depth resolution from ion

beam non-uniformity is negligible if the ion beam is large compared with the electron

beam. The contribution from the Auger escape depth is only 5 20A and remains

constant with sputtered depth. Sample inhomogeneity is the most important factor

limiting depth resolution and can occur as a non-uniform thickness of thin films over

the sampled area, as polycrystalline grains which are comparable in dimension with

the thickness of the film, and as precipitation of constituent elements into particles. A

non-uniform film thickness obviously limits the depth resolution to a value less than

the percent uniformity of the film thickness over the sampled area. The depth

resolution is determined by relative sputtering rates of the precipitates and the matrix

in films containing precipitates, and by a variation in sputtering yield with grain

orientation in polycrystalline films. In general, the depth resolution of the AES/ion

etching technique is limited to about 10% of the sputtered depth, but can approach

3% in optimum cases. The best in-depth resolution that has been achieved occurs in

amorphous films.

Any quantitative evaluation of depth profiling data obtained by combination of

electron spectroscopy with ion sputtering requires a quantification of the signal

intensity and the depth scale together with a knowledge of the depth resolution

function, provided a composition change during the analysis is negligible or known.

The speed of analysis, background considerations and above all the small analysis spot

greatly favors high spatial resolution AES (or Scanning Auger Microscopy-SAM) over

XPS. Whereas the decisive parameters for quantitative AES and XPS (electron

escape depth and electron back-scattering in AES) are comparatively well defined, a

prediction of the large number of parameters (depending on the instrumental and ion

beam/sample interaction characteristics) appears highly speculative at present.

Alternate methods to conventional sputtering, like crater edge profiling,78 ball

cratering78 and effective escape depth variation, are useful approaches in special


After the first step in quantification (concentration and depth scales), a finite

difference to the true, original profile remains, which is determined by the

experimental depth resolution or, more precisely, the depth resolution fufiction. A

knowledge of this function allows the deconvolution of the normalized measured

profile to obtain a true profile by numerical methods.79 Alternatively, a

convolution of an assumed true profile and an iterative comparison with the measured

profile allows the determination of the true composition profile. In any case, the

parameter for the reliability and precision of the latter is the experimentally resolved

depth which depends on a careful optimization of the profiling conditions.

3.2.3 Angle Resolved XPS

A nondestructive but limited quantitative analysis as a function of depth from

the surface is also available. This technique is often called angle resolved XPS. As

the angle of electron exit relative to the sample surface is decreased, the effective

sample area is increased while effective sampling depth is proportionally decreased;

this has the effect of increasing the peak intensities of surface species. The practical

implication of this observation is that by tilting the sample at an angle to the

spectrometer, it is possible to enhance the surface sensitivity of ESCA and thus help

distinguish surface phenomena from features arising deeper into the sample. While

this has its greatest application in the area of polymer analysis, where mean free paths

tend to be long, it is also applicable to the study of very thin layers such as oxides on

metals and more recently to examine surface structures of ceramic


The reason for enhancement of surface sensitivity is shown diagrammatically

in Figure 8. If X is the inelastic mean free path (IMFP) of the emerging electron then

95% of the signal intensity is derived from a distance 3X within the solid. However,

the vertical depth sampled is clearly given by

d = 31sina

and this is maximum for a = 900. In the case of a substrate (s) with a uniform thin

overlayer (o) the angular variation of intensities is given by81

Id ,e -d/sinma
ds ^s




Figure 8 Surface sensitivity enhancement by variation of the electron take-off


Id = Io(1 e-a/ )

where X is the appropriate value for the observed photoelectron. In the ideal situation

these equations lead to curves of the type shown in Figure 9. However, the real case

is usually complicated by the fact that the system geometry imposes a response

function also dependent on a. This complication is avoided by measurements of

relative values of I/ols so that the instrument response function is canceled. Thus as







Figure 9 Angular dependence curves calculated for a clean flat thick surface and
a flat overlayer/substrate system.






I _

shown in Figure 9, at low values of a, Io/I increases significantly.

The major requirement for surface sensitivity enhancement is that the surface

is flat. Surface roughness leads to an averaging of electron exit angles and also

shadowing effects (both for the incident x-rays and emerging electrons) such that in

most cases the surface enhancement effect cannot be observed.82 In most

commercial spectrometers the angle between the incident x-rays and detected electrons

(180 a 6) is fixed. However, it has been shown that at very low 6 there is also a

surface sensitivity enhancement effect due to the rapid fall-off in x-ray penetration

depth as 6 tends to grazing incidence.83



The subject of this research was to study the formation of surface species

resulting from the exposure of YBCO superconductors to air and a wet atmosphere.

In order to detect the susceptibility of different structures, several types of samples

were tested. They can basically be divided into two categories: pressed pellets and

thin films. Two pellet samples from the same batch and two films from the same

deposition were prepared and one of each was tested for superconductivity while the

surface of the other was examined in the corrosion study. All samples were of the

123 nominal composition and had To in the range 84 93 K. Samples were analyzed

immediately after preparation, following an ambient air exposure of 18 days and after

exposure to 85% relative humidity and 500C in a humidity chamber.

4.1 Pellets

Pellets were obtained from Greg Chandler.84 They were prepared from

powders of Y203 (< 10Am diam.), BaCO3 (1/m diam.) and CuO (< 800/m

diam.).85 The powders had a nominal purity level of 99.999% for Y203 and CuO

and 99.9% for BaCO3. The molar ratio of Y:Ba:Cu in the mixture was 1:2:3. These


powders were mechanically mixed for 24 hours to ensure homogeneity. Mixed

powders were calcined in an A1203 crucible at 950C in air for one hour. The

powders were quenched in air and then ground with mortar and pestle. Calcination

and grinding steps were repeated twice. Pellets with an average height of 1.35mm

and average diameter of 4.85mm were pressed from 0.1 g of calcined powder at

430MPa in a laboratory press. The pellets were then sintered at 9500C for one hour

in an A1203 combustion boat. They were cooled slowly to 6000C and annealed for

three hours in flowing oxygen followed by slow cooling to room temperature in

flowing oxygen.85 Densities of the pellets varied from 3 to 5.5 g/cc

4.2 Films

Films several microns thick with different orientations deposited using laser

ablation and RF sputtering were tested. YBa2Cu307.B films obtained from Carl

Mueller86 were made by deposition onto unheated A1203 or SrTiO3 single crystal

substrates using pulsed CO2 laser radiation (10 Im, 250 nsec, 1 Hz, and energy

density of approximately 2000 J/cm2) on bulk superconductor target. The films were

grown with either 50 pulses or 5400 total pulses. One inch diameter YBa2Cu3O7-_

targets were continuously rotated during the deposition, and the substrate to target

distance was 3 cm. The base pressure was 5 x 10-7 Torr and rose to 1 x 10-5 Torr

during deposition. The films were post annealed in oxygen at 900C for 1 hour,

electrical resistance vs. temperature data were collected. In addition the films and

targets were characterized using SAM, scanning electron microscopy (SEM), XRD

and XPS. Only the results obtained from XPS will be reported. The details of other

measurements are given elsewhere.86

RF planar magnetron sputter deposition was also used to grow YBa2CU3078

films and films were obtained from Kelly Truman.87 Sputter targets were fabricated

by mixing milled and ground powders of Y203, CuO, and BaF2 and cold pressing

into disks 2 inches in diameter and 1/8 inch flat. They were fired in air at 500C to

prevent disassociation during exposure to the plasma. XRD measurements made

following the treatment have shown no reaction at this temperature. Si and SrTiO3

substrates were used, with the latter giving better results. The films deposited on

SrTiO3 were used for the surface study of degradation. The substrates were cleaned

with methanol prior to deposition and placed directly above the target (-6 cm).

After a background pressure of 1 x 10-6 Torr was reached, Ar gas was backfilled up

to 2 x 10-4 Torr, where it was kept during the deposition. A shutter between the

sample and the target was kept closed for 3 hours after the sputtering had begun. A

typical deposition took 6 hours at a rate of 50 A/min creating films of 1.8 1m

thickness. The deposited films were heat treated in a quartz annealing tube, where

the temperature was brought up to 850C exposed for one hour to a wet 02 gas

produced by passing 02 through H20 at a rate of 200 cc/sec. At this stage BaF2 has

reacted to form BaO and HF produced was removed by forming volatile SiF4:

BaF2 + H20 2HF + BaO

4HF + SiO2 -- 2H20 + SiF4 t

The samples were kept at 850C for one hour followed by a slow cool to room

temperature in 02 for 12 hours. XRD have shown strong c axis orientation. Use of

different substrates and techniques deposited can be found in an extensive review by

Kelly Truman87 and Leskeli et al.8




5.1 Experimental Set Up

X-ray photoelectron spectroscopy measurements were performed on Perkin

Elmer PHI ESCA 5000 and Kratos XSAM 800 at high resolution (pass energies at

17.9 and 8.95 eV) for quantitative and at low/survey resolutions (pass energy at

189.95 eV) for qualitative examinations. The spectra were collected with 0.05 eV

and 0.5 eV step sizes for high and low resolutions respectively. Magnesium Ka

x-rays were used extensively which were operated at 300 watts. The typical binding

energy resolution (FWHM) for both instruments was 1.2+0.2 eV on Ag 3d5/2 line of

a silver foil and 1.440.2 eV on Ba 4d5/2 of a YBCO sample. Angle resolved

measurements were made on the PHI ESCA instrument between 150 80*. Typical

vacua in both instruments were 10-8 Torr.

Data from direct measurements were enhanced by smoothing, and satellite

(except for the Auger peaks since they are independent of x-ray energy) and

background subtraction to improve interpretation of the spectra and enable comparison

of spectra from different samples. Since the peak width of the emission of a typical

species was measured to be 1.4 eV at 0.05 eV step size it was calculated that 19


points could be used several times for 2nd degree central point Savitsky-Golay

smoothing without any loss of spectral information (0.7 x 1.4/0.05 = 19.6 Section Figure 10 shows that this process can be applied several times without any

loss of spectral information provided that the correct number of points is used.

On the PHI ESCA the small pellet samples were first mounted on a standard

stainless steel holder clamped down by a molybdenum mask. However, the hole of

the mask exposing the sample to analysis was not large enough to allow detection of

Figure 10

945 940 935 930
Binding Energy (eV)

Comparison of Cu 2p (YBCO after ion sputtering) raw spectrum with
smoothed spectrum using 19 point Savitsky-Golay routine 8 times.

sample peaks by the electron detector. Although the Mo peaks (Table V) do not

interfere with the YBCO peaks (except for the Mo 3p3/2 and Y 3s, 396 and 395 eV

respectively), an oxygen peak from Mo203 could complicate the examination of the

O Is spectrum. To determine oxide level, a gold foil was covered with the mask was

inspected. Indeed it was discovered that a detectable oxygen level was present, high

enough to obscure the analysis of the YBCO samples. Thus a new approach was

used. Rather than clamping down the sample, it was placed in a circular boat made

out of Al foil without any retainers. No aluminVm was detected and this method also

Table V Molybdenum XPS lines for Mg Ka X-rays.

Energy (eV)* Linet

38 Mo 4p
65 Mo 4s
230 Mo 3d5/2
233 Mo 3d3/2
396 Mo 3P3/2
413 Mo 3P1/2
508 Mo 3s
1033 Mo MNN
1068 Mo MNN

C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder and (ed.) G.E.
Muilenberg, Handbook of X-ray Photoelectron Spectroscopy, Perkin Elmer
Corporation, Eden Prairie (1979).

t s, p, d levels represent photoelectron and MNN levels represent the Auger

allowed angle resolved XPS measurement. The films on the other hand were large

enough to be easily mounted using the Cu-Be foil springs attached with screws to the

holder. When using the Kratos XPS, pellets and films were affixed to the holders

without any special arrangement since it was a common practice to use double sided,

UHV safe Scotch 3M tape.

Analysis of the composition of the surface of superconductors is hampered by

the lack of generally accepted methods for removing surface contamination. Many

methods have been suggested, but each has serious drawbacks. For example, vacuum

fracture commonly practiced in sample preparation for XPS studies has been criticized

since it enhance grain boundaries at the fracture surface leading to a Ba-rich layer17,18

(except in single crystals); low energy ion milling in situ causes differential sputtering

of the elements, especially oxygen, and leads to incorrect compound

stoichiometries;17'19,20 surface scraping with a diamond or steel file is the most often

used method, yet it suffers from the same grain boundary exposure problem as sample

fracture.17,19,21 Heating the sample until the contaminants are evaporated seems to

be the most reliable method for measurements with reduced surface contaminants.19

Low energy Ar ion sputtering was tested for surface cleaning in this study.

The results confirmed the fact of metal reduction as indicated by a sharpening of the

Cu 2P3/2 and shake up peak loss (Figure 11-a) while removing the dominant carbon

contamination. A clear indication of Cu transforming from a possible combination of

valence states [1+ : Cu20-932.5 eV, 2+ : CuO,YBCO-933.7 eV, and conceivably

3+ : YBCO-934.5] to the 1+ states were observed. To illustrate this effect the

Initial spectrum I
After 35 min. low energy Ar on 2+




Binding Energy (eV)


i Initia spectmm

SAfter 35 min. low energy Ar on sputtring


355 350 345 340 335 330 325 320 315 310 305
Binding Energy (eV)


Figure 11 The effects of 35 minute Ar+ ion sputter cleaning on (a) Cu 2p3/2 and
(b) Cu LMM spectra of a YBCO pellet.

spectrum was smoothed 8 times with 19 points as discussed earlier, generally

followed by satellite and baseline subtraction. Satellite subtraction was used to

remove the contribution by Mg Kf photoemission of higher binding energy electrons

which might overlap with the Mg Ka photoemission of lower binding energy

electrons. The intensities were adjusted to have the same height at the 2P3/2 peak so

that the ratio of 2p3/2 to the shake up could be shown. This ration has been shown to

indicate the presence of 2+ valence state, although it cannot be used quantitatively

because of the presence of the 1 + valence also. The transformation could also be

observed as an energy shift and shape change in the major Cu x-ray induced Auger

transition line (L3M45M45-335 eV), and evolution of the L2Ms4M45 peak at 310 eV

(Figure 11-b). The spectral changes of the Ba 4d doublet showed similar

transformations (Figure 12). However the changes are not as evident in the Ba Auger

and Y 3d peaks. It was clear from the evolution of Ba and Cu together With changes

in the O Is line that ion etching for the purpose of cleaning causes chemical state

alterations in Ba and Cu oxidation. The analysis of relative elemental composition

variations reveals an enhancement of yttrium with respect to barium and copper

(Table VI), while the relative concentration of oxygen remains constant and as

expected, carbon decreases drastically. Therefore it was concluded that substantial

carbon content was due to a surface layer formed during the preparation and transfer

of the sample from the furnace to the XPS instrument while ion sputtering for

cleaning leads to chemical state and elemental composition changes. The study

presented here addresses the problems of surface contamination or degradation by


2' Initial spectrum
"E After 35 min. Ar ion sputtering



96 94 92 90 88 86 84 82
Binding Energy (eV)

Figure 12 Effects of Ar+ ion milling in Ba 4d of YBCO pellet while removing
surface adsorbates.

analyzing the surface reaction layer formed on 123 pellets and films under normal and

accelerated corrosion conditions. Therefore surface cleaning was not used.

Superconducting YBCO samples were conductive at room temperature and

thus no charge referencing was needed for the samples in contact with the

instrument's ground. Corroded samples however were insulators and a static positive

charge was formed during the XPS analysis. The C Is peak was used for charge

referencing. However a complex line shape is common in these superconductors

(Figure 13) and, in order to separate different components, curve fitting with usually

6 peaks (Figure 14) was carried out. The binding energies for these peaks were kept

Table VI Changes in the relative atomic concentrations due to X-ray exposure
and Ar ion milling.

Relative Atomic concentration Ratio
C O Y Ba Cu Y Ba Cu
Initial 29.1 57.5 3.0 7.5 2.9 1 2.5 0.97

120 min. 25.2 59.5 3.4 8.5 3.4 1 2.5 1.0

510 min. 26.7 57.9 3.4 8.2 3.8 1 2.4 1.1

Ion mill 7.1 59.9 11.4 12.5 9.1 1 1.1 0.8

the same throughout the experiment, however no attempt was made to identify them.

The curve fitting was only used to measure the position of the lowest binding energy

peak which was assigned to the adventitious hydrocarbon species. This peak was set

to 284.6eV and always used for charge correction.

In order to determine accurate peak positions, chemical shifts and peak widths,

standard samples were produced made from BaO, BaCO3, Y203 and CuO powders

cold pressed into discs. The metal core level and Auger transition data of these

samples helped identify the complex peaks obtained in YBCO samples. For the Ba

analysis, the 4d orbital was used since the chemical shifts of the higher electronic

state are greater. It was found that the 4d spin states (/2 and 3/2) to be separated by

2.5 eV as measured on BaO and BaCO3 samples after low energy ion sputter

cleaning. The binding energy of the 4d5/2 line was measured to be 89.6+0.4 eV for

both BaCO3 and BaO. The Ba 3d peak however was selected for quantitative analysis

since it is a more easily separated doublet (15 eV).

292 290 288 286 284 282 280

C Is spectrum of YBCO after exposure to ambient air and 85% relative
humidity at 500C.

Figure 13


294 292 290 288 286 284 282 280
Binding Energy (eV)

Figure 14 Curve fitted C Is spectrum of YBCO pellet treated with 85%
relative humidity at 500C for 5 hours.

Oxygen Is spectrum revealed the binding energy for BaO and BaCO3 to be

529.0+0.2 and 530.2+0.2 eV, respectively. Y203 and CuO O Is spectra was also

measured to be 529.0+0.2 eV. The Auger level spectra of these samples collected

were the KLL (for oxygen), LMM (for Cu) and MNN (for Ba) transitions. However

the complex peak shape of the x-ray induced Auger transitions lines makes it very

difficult to use them as a basis with which to separate the components of the spectrum

of mixed species. Special analysis tools such as factor analysis (or principal

component analysis, section are needed to be able to effectively study the

Auger transition changes.

X-ray irradiation damage was detected as reported before,22' 29 resulted in a

reduction of oxygen leaving a more metallic surface after prolonged exposures to

radiation. An energy shift and shape change was observed in the Cu 2P3/2 and

Cu LMM indicating a reduction towards lower valencies (Figure 15). Although a

possible loss of Cu3+ was strikingly evident in the core level spectrum as a decrease

in the shake up feature intensity and FWHM of the 2P3/2, the differences in the Auger

spectrum were not as clear. A reduction in Ba was also detected however it was not

as severe as the Cu reduction. No variations in the elemental composition were

detected in prolonged exposures to x-ray radiation (Table VI). In order to minimize

this effect during the measurements, fast data acquisitions were performed at the

expense of higher resolution. XPS spectra of Ba 4d and 3d5/2, O Is, Cu 2p3/2, Y 3d

and C Is orbitals, as well as x-ray induced Auger transitions of Ba, O and Cu were

collected. Valence band (0 40 eV) transitions were also examined, however low

intensity and lack of standard measurements hindered an unambiguous analysis of the

peaks detected in this region.

Wide scans of all of the YBCO samples were collected before a detailed study

of the effects. This step was taken to ensure that the data measured were free of

impurities that would alter the analysis. To accomplish this objective, the lowest

possible resolution (or largest pass energy-189.95 eV) was used for maximum






Figure 15

Binding Energy (eV)



Change due to X-ray radiation damage in the (a) Cu 2p3/2 and (b) Cu

sensitivity. The minimum atomic concentration detection of XPS is 1 3%. Since

binding energy resolution was not critical, large incremental (0.5 0.3 eV) energy

steps were used to acquire data more quickly. Integral baseline (background) and Mg

Kao satellite subtractions, charge referencing (to C Is) and smoothing with a smaller

number of points (5 9) were employed to improve the interpretation of the peaks

observed in the spectra. The survey spectrum of a superconducting YBCO pellet

after all the above steps were carried out is shown in Figure 16.

Once the initial tests were concluded the samples were studied with higher

energy resolution to examine the structural and chemical changes. However since

YBCO is sensitive to long x-rays exposures an optimum resolution needed to be

found. For this purpose several tests were performed by changing the combination of

step size and pass energy in order to obtain data rapidly and yet not increase the

width of the XPS line considerably so that an accurate analysis of the measured peaks

could be made. Step sizes of 0.025 0.05 eV were tested and it was observed that

0.025 eV raised the noise level without any significant improvement in the peak

shape. Furthermore, as a consequence of low signal-to-noise ratio this step size

required additional spectra to be collected which resulted in an increase in the x-ray

exposure. Thus the 0.05 eV step size was used without any loss of energy resolution.

The data were collected with the widest slit opening possible on the Kratos

XPS. Since there was a significant (- 1 eV) difference in the FWHM between

medium and high resolution, the latter was used for bond analysis. The PHI ESCA

on the other hand has several options for binding energy resolution. All of the





Figure 16

1000 900 800 700 600 500 400 300 200 100
Binding Energy (eV)

The XPS wide spectrum of a superconducting YBa2Cu307.. pellet after 5 point Savitsky-Golay smoothing, Cls
referencing (284.6 eV) and baseline and satellite subtraction.

provided pass energies between 4.45 35.75 eV were tested for optimum sensitivity

and resolution and it was found that the 17.9 eV pass energy could be used without a

major drawback.

5.2 Chemical Reactions on the Surface

The core, valence and Auger levels of the prepared samples were collected at

medium and high resolutions. Medium resolution data were used for atomic

concentration measurements while the high resolution data were used for the bond

analysis. As was indicated in CHAPTER 2 the main objective of this work was to

observe the surface changes of separately prepared YBCO samples when they were

subjected to the same environmental exposure. The measurement was always begun

with a freshly prepared sample (immediately out of the furnace), in order to minimize

ambient exposure effects. The first analyses of the samples were followed by

analyses of exposure to ambient atmosphere for 18 days and treatment for cumulative

periods of 5, 12 and 48 hours in a humidity chamber which was held at 85% relative

humidity and 500C. Samples exposed to humidity were normally analyzed

immediately. No changes were observed in Ba or O features when some samples

were reanalyzed after they had been kept in UHV (10-9 Torr) overnight. Figure 17

shows the complete method followed to study the changes on the surface. The search

for impurities and energy calibration (charge referencing) has already been addressed

in the previous section. The calculations for atomic concentration will be discussed in

the next section which will also include the analysis of depth measurements profiling

Preparation of 2
YBCO Samples

18 days in air and
5, 12, 48 hrs in
humidity chamber

Deconvolution of
Ba4d and 01s
XPS peaks

Ambient Atm. and
Humidity Treatment
(85% RH & 50C)

Flowchart of experimental procedure followed.

Ion Assisted Auger
Depth Profile

Figure 17

performed on superconducting and corroded YBCO after 48 hours of humidity

exposure. The technique used to examine photoelectron line changes in Ba 4d and

O Is will be discussed here.

X-ray photoelectron spectroscopy analysis of the surface can also yield

information about chemical changes occurring with various degrees of exposure by

means of chemical shift analysis of the deconvoluted peaks. To obtain consistent and

reproducible information from the peak synthesis (or curve fitting) the same approach

was used for every spectrum. These steps involved using C Is for charge

referencing, smoothing the spectrum and performing a second derivative to determine

the energy positions of the components more accurately. The curves that represent a

certain species were restricted to the same energy position within experimental error

(0.2 eV) after all of the spectra were fitted. They were also compared to

measurements reported in the literature. The width of components that represent the

same species was also kept constant (0.2 eV) with the assumption that the

instrumental broadening should not change. Curve smoothing enabled a faster fitting

process. The envelope of the fitted curves was then overlaid on the raw data for

visual inspection and to determine the accuracy of fit (x2).

The smooth complex peak was also used for 2nd derivative analysis of the

spectra. This technique was chosen to improve the first guess of the position and

height of the individual peak components, as required for computerized curve fitting

(Figure 18). The minima found represented the peaks in the raw data with the

location of the minima on the binding energy scale yielding the position and the depth



96 94 92 90 88 86 84
Binding Energy (eV)

Figure 18 Second derivative of a Ba 4d spectrum from a YBCO treated 5 hours in
the humidity chamber. 3 times 19 point smoothing and satellite
subtraction was carried out before the differentiation.

of the basin from the derivative of the background corresponding to the relative

intensity of the original peak. This step helped analyze the data faster since curve

fitting was one of the most time consuming processes in this study. As will be

discussed below, deconvolution of peaks can be misleading if not applied carefully

with some consistency. Although differentiated data decreased the ambiguity in

determining the components, it could often result from noise when there was low

signal-to-noise ratio. The number of points used while differentiating could lead to

incorrect interpretation as too small a number would enhance the noise while too large

a number could yield fewer components than present in the data. In light of these

considerations it was used in combination with curve fitting.

In the final step of the peak resolution process the binding energy of the

components was set according to the information gathered from the differentiated

spectrum, previous measurements on standard barium compounds and results reported

by other investigators. The O Is and Ba 4d peaks were deconvoluted into a series of

components, closely matching the binding energy positions reported in Ref. 80,

except for the superconducting cuprate species in Ba 4d. An additional component

which grew after exposure to humidity was identified as due to hydroxyl forms of Ba.

The binding energies and component specie assignments are shown in Table VII and

Table VIII for the Ba 4d5/2 and O Is XPS peaks. In the Ba and O XPS peak analysis

by Ford et al,19 the 531 eV O Is peak was interpreted as a combination of surface

and bulk species. Most studies16'80,89 have associated this peak with BaO2 or

BaCuO2 formed on the surface after heat-treatment19 and exposure to

atmosphere.15,16 It will be shown in the following section that humidity caused

depletion of copper from the surface. The observations reported here show that the

531 line is still present after Cu was no longer detected on the surface, therefore, the

assignment to BaO2 only is more appropriate as shown in Table VIII. The cuprate

species have lower binding energies (528 and 529 eV) as shown in Table VII. In

order to determine the identity of the higher binding energy lines we list heats of

formation (AHO ) of several Ba compounds with oxygen, hydrogen and carbon (since


Components used to curve fit Ba 4d spectra of 123 compounds


superconducting Ba (Y-Ba-Cu-O)

non-superconducting Ba (Y-Ba-Cu-O)

BaCO3, BaO



Ba 4d5/2






Table VIII Components used to curve fit O Is spectra of 123 compounds.


superconducting O (Y-Ba-Cu-O)

SC Y-Ba-Cu-O, CuO, Y203, BaO



COx, Ba(CxOyH)

Hydroxide I, Ba(OxHy)

Hydroxide II, H20

B.E.(eV)0.2 "








Table VII

the surface was rich in these elements) (Table IX). A high negative value indicates a

favored composition. Thus it can be concluded that the formation of compounds like

Ba(OH)2.8H20, BaO2.8H20 and BaCzO4.2HO2 was favored on the sample surface.

The hydroxide line was found to have a higher binding energy than predicted by

previous authors.21,80,90 The intensity of the 534 and 535.6 eV lines increased with

exposure to humidity and the lines most probably correspond to Ba(OH)2,

Ba(OH)2.8H20, BaC204.2H20 and/or molecular water.91 The O Is peaks often

detected in the 531 534 eV range have been attributed to a multiplicity of forms of

chemically and physically bound water on the surface. The chemisorbed and

physisorbed oxygen and hydroxyl are also found in this range making a clear

interpretation difficult.92'93 The high binding energy O Is peak in the

La2CuO4+5 system was also attributed to water.91 The possibility of the presence of

BaCxOyHz in this study arose from the fact that carbon had a component at 292.1 eV,

which is much higher than the reported BaCO3 peak at 288.60.3eV,30,32,94,95

also observed in this study (Figure 14). This carbon peak is most likely related to a

hydroxyl or a water compound since these usually have higher binding energies. In

making the assignments of Table VII and Table VIII, it was assured that no impurities

were present which would interfere with the interpretations. While curve fitting

different spectra, the binding energy positions were kept within +0.2 eV of the values

given in Table VII and Table VIII. While many of the peak assignments in XPS

studies were based on direct and indirect inference both in this paper and in most of

the literature, the assignments shown in Table VII and Table VIII agree well with


Table IX Heats of formation of barium compounds with oxygen, hydrogen and


Ba(OH)2.8H20 -799.5

BaO2.8H20 -719.3

Ba(C2H302)2.3H20 -567.3

BaC204.2H20 -470.1

BaC204. /2H20 -363.7

Ba(C2H302)2 -355.1

Ba(HCO2) -326.5

Ba(OH)2.H20 -299.0

BaCO3 -291.3

Ba(OH)2 -226.3

BaO2.H20 -223.5

BaO2 -150.5

BaO -133.4

BaH2 -40.9

AHf = The standard heat of formation of a given substance from its
elements at 250C kcal/g.mole.

S D.D. Wagman, W.H. Evans, V.B. Parker, R.H. Schumm, S.M. Bailey,
I. Halow, K.L. Chumey and R.L. Nuttall, National Bureau of Standards
Tech. Notes 270-3, 270-4, 270-5, 270-6, 270-7 and 270-8.

those of other authors and are consistent with a wide variety of

measurements.19'80,96 The Ba(OH)x species in this study had a larger binding

energy than predicted by the same reports. Instances where we differ from the

literature have been identified and justified by other measurements.

It was found that by first subtracting the most apparent features such as

pronounced shoulders or maximum in the peak to be analyzed, the speed of curve

fitting could be increased. For every Ba species it was necessary to use two peaks

with 2.5 eV separation and the intensity of the 4d5/2 peak was taken at 1.5 times the

4d3/2 peak intensity or integrated peak area as discussed before. These considerations

are in close agreement with earlier reports.19'80'96 Using these factors, peak

resolution of the Ba spectra was initiated from the lower binding energy side where

the superconducting YBCO component was well established in the early stages. After

fitting the right hand side of the contour (Ba 4d5/2) with 90% 100% Gaussian and

0 10% Lorentzian curves Ba 4d3/2 was fixed in place. This was followed by

placing the most intense peak at a binding energy position consistent with the previous

analysis. In other words the relative positions were kept constant according to the

tables given. The deconvolutions of the Ba 4d and O Is peaks for the bulk sample

are shown in Figure 19 and Figure 20 and for the Rf sputtered film are -shown in

Figure 21 and Figure 22, under all cases of exposure studied. The goodness of fit

varied between 0.02 to 2.0.

98 96 94 92 90 88

Figure 19

86 84

Curve fitting of the Ba 4d spectra acquired from a low density pellet
during the course of the experiment.

F -

Figure 20

Curve fitting of the 0 is spectra acquired from a low density pellet
during the course of the experiment.


5 hours (humidity)


Non a ,

Ba(O / \
... ...... .I \

94 93 92 91 90 89 88 87

S/ /-

S"5 hours (humidity)

1/8 days (air)

/. / \ \

SC 0 (Y-Ba-Cu-0)
SC Y-Ba-Cu-O, CuO, YO, BaO /

Hydroxide I. Ba(0y) / '\
Hydroxide II, HO -- ,

536 535 534 533 532 531 530 529 528 527 526
Binding Energy (eV)


5.3 Concentration Changes

In addition to analysis of peak contours which lead to examination of reactions

to corroding the environment, the area under the curves were calculated to investigate

the changes in elemental composition. The details of quantitative information

gathered from XPS have been given in section 3.1.4. The relative atomic

concentrations of Ba, Cu, Y, C and O were calculated for every sample after

treatments. Different background subtractions and smoothing were used to measure

experimental error. An estimated error of 1% was found. Various binding energy

values were tested to mark the background. It was concluded that smooth curves

generally improved the marking of the ends of the peaks yielding a better

representation of the background.

The changes in Cu with respect to Ba was the most striking. The Cu

concentration decreased with treatment indicating a depletion of copper and/or

enhancement of barium. After 48 hours in humidity however, it became clear that

both were occurring and Cu was completely lost from the surface for powder pellets.

Barium enhancement was not as significant while Cu was depleting in the films

especially when prepared by RF sputtering from a BaF2 target. A complete change in

the atomic concentration for the most severe cases (low density pellet and RF

sputtered film) are given in Table X. To illustrate the relative changes in barium and

copper these measurements can be plotted separately (Figure 23).

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