J. Appl. Phys. 106, 093510 (2009); https://doi.org/10.1063/1.3253329 106, 093510
? 2009 American Institute of Physics.
Low energy photoelectron diffraction
analysis at high angular resolution of Cu and
Mn/Cu surfaces
Cite as: J. Appl. Phys. 106, 093510 (2009); https://doi.org/10.1063/1.3253329
Submitted: 09 July 2009 . Accepted: 21 September 2009 . Published Online: 09 November 2009
G. P. Cousland, A. E. Smith, J. D. Riley, and A. P. J. Stampfl
ARTICLES YOU MAY BE INTERESTED IN
New display-type analyzer for the energy and the angular distribution of charged particles
Review of Scientific Instruments 59, 545 (1988); https://doi.org/10.1063/1.1139884
Wide-angle display-type retarding field analyzer with high energy and angular resolutions
Review of Scientific Instruments 88, 123106 (2017); https://doi.org/10.1063/1.4990769
Time- and angle-resolved photoemission spectroscopy of solids in the extreme ultraviolet at
500 kHz repetition rate
Review of Scientific Instruments 90, 023104 (2019); https://doi.org/10.1063/1.5081938
Low energy photoelectron diffraction analysis at high angular resolution
of Cu and Mn/Cu surfaces
G. P. Cousland,
1
A. E. Smith,
1,aH20850
J. D. Riley,
2
and A. P. J. Stampfl
3
1
School of Physics, Monash University, Clayton, Victoria 3800, Australia
2
Department of Physics, La Trobe University, Bundoora, Victoria 3086, Australia
3
Bragg Institute, Australian Nuclear Science and Technology Organisation, Lucas Heights, New South
Wales 2234, Australia and School of Chemistry, The University of Sydney, New South Wales 2006, Australia
H20849Received 9 July 2009; accepted 21 September 2009; published online 9 November 2009H20850
X-ray photoelectron diffraction simulations using a real-space approach are shown to accurately
produce the extraordinarily detailed photoelectron diffraction pattern from CuH20853111H20854 at an electron
kinetic energy of 523.5 eV. These same simulations show that most sensitivity is obtained when
using low energy electrons at high angular resolution. Structural differences are observed to be
greatest around a kinetic energy of H11011100 eV and many of the features observed in the
photoelectron diffraction patterns may be directly related to phenomena observed in low energy
electron diffraction patterns from the same surface. For CuH20853100H20854, simulations of buckled surfaces
with a Mn overlayer predict that low energy photoelectron diffraction can easily discriminate
chemical and structural differences. Even the effects of the relaxed surface of CuH20853100H20854 is indeed
observable along azimuthal scans around a kinetic energy of 100 eV. Our results show that low
energy photoelectron diffraction is extremely sensitive to changes in surface structure if high
resolution patterns are acquired. ? 2009 American Institute of Physics. H20851doi:10.1063/1.3253329H20852
I. INTRODUCTION
X-ray photoelectron diffraction H20849XPDH20850 is an established
experimental technique used for the determination of surface
structure.
1,2
One of the strengths of XPD is that it uses spe-
cific core levels and is therefore atom specific enabling dif-
fraction patterns from adsorbates and particular components
of alloys.
3,4
If chemical shifts are large enough atoms with
particular bonding can be selected as electron sources.
Typical XPD energies in the range of 200?1000 eV are
used in which regime multiple scattering significantly con-
tributes to the electron intensity distribution as does forward
scattering. In its earliest incarnation, much structural infor-
mation was obtained with diffraction patterns at energies
above 300 eV from forward focusing and Kikuchi-like
effects,
5,6
for data along the lines of atoms.
7
More recently
simulations and R factor analysis have provided a means of
comparing data with complex structures and enabling atomic
positions to be determined with reasonable accuracy. Single
scattering models were widely used but as multiple scatter-
ing analyses appeared it became clear that single scattering
models did not yield correct interatomic spacing.
Two developments have offered an improvement for
XPD determination of surface structures. Electron diffraction
in atomic clusters H20849EDACH20850 software
8
has enabled the calcula-
tion of multiple scattering processes for large clusters and
can simulate XPD patterns with a higher degree of reso-
lution. Electron analyzers have improved resolution enabling
data to be obtained with more detailed angular structures
being resolved and data taken over the full 2H9266 steradians. In
this paper we compare the EDAC calculation of the diffraction
patterns from a CuH20849111H20850 surface and the experimental data
set to show detailed comparison which is now possible.
This paper explores the sensitivity of full hemisphere
photoelectron diffraction to variations in the bulk terminated
structure of CuH20849100H20850 surfaces and to a bulk terminated struc-
ture with a Mn overlayer. It is noted that most low energy
electron diffraction H20849LEEDH20850 determination of surface struc-
tures is performed at low electron energies so the examina-
tion of sensitivities ranges from electron energies in the usual
range, 300 eV down to typical LEED energies of 50 eV.
Comparison is made between the bulk terminated diffraction
pattern and the varied surfaces using an R factor whose val-
ues give an objective determination of the changes in the
diffraction data and the possibility of structure determination
from full hemispherical photoelectron diffraction H20849FHPEDH20850.
Comparison is shown between the experimental diffrac-
tion pattern from a CuH20849111H20850 surface for 3p electrons with an
energy of 523 eV and the calculated diffraction pattern using
EDAC code. The remarkable agreement suggests that using
EDAC to explore the sensitivity of FHPED as a function of
energy for different surface structures will guide and demon-
strate its value for surface determination.
The present work has been prompted by an investigation
of the exchange properties at the interface of antiferromag-
netic and ferromagnetic materials with spin glass surfaces.
Cu
3
Mn has been chosen as the initial material, with azimuth-
scanned synchrotron XPD patterns to be used to determine
structure. In the present work, owing to the growth properties
of the CuMn alloy,
9
XPD diffraction patterns for Cu H20853100H20854
surfaces are considered in comparison with Mn/CuH20853100H20854 us-
ing low photoelectron energies. Due to the enhanced ability
of current electron spectrometers, XPD can now be obtained
at higher angular resolution than has previously been the
case. While only preliminary experimental results are
available,
10
computer simulations are carried out using EDAC
aH20850
Electronic mail: andrew.smith@sci.monash.edu.au.
JOURNAL OF APPLIED PHYSICS 106, 093510 H208492009H20850
0021-8979/2009/106H208499H20850/093510/6/$25.00 ? 2009 American Institute of Physics106, 093510-1
to determine how well low energy photoelectron diffraction
analysis can find structure. These simulations are thus pre-
dictions of high resolution XPD results in the energy range
below H11011300 eV.
As the variation with energy of the reflected intensity is
an integral part of LEED surface analysis,
11
XPD patterns
may depend strongly on low kinetic energies. Accordingly
CuH20853100H20854 XPD patterns at values in the range 73.5?303.5 eV
for bulk truncated, relaxed,
12
buckled, and Mn incorporated
surfaces
9
are investigated using EDAC.AnR-factor analysis
is used to compare data from the bulk truncated CuH20853100H20854
simulation with the other Cu surfaces simulated, for all en-
ergies investigated.
In order to check the accuracy of the simulations for
copper surfaces, EDAC calculations are carried out for the
CuH20853111H20854 surface as experimental results for this surface are
now available from a second generation toroidal electron
spectrometer, which is able to achieve high angular
resolution.
13
While comparisons of EDAC simulations with
experimental results for CuH20853111H20854 have been made
previously,
14
the present work shows how performing a high
resolution analysis enables the observation of an extraordi-
nary amount of detail in a XPD pattern of CuH20853111H20854.
II. SIMULATION REQUIREMENTS
EDAC simulates XPD and LEED by summing in real
space over scattering paths, with atoms represented by muf-
fin tin potentials,
15
arranged in clusters. Results were calcu-
lated with the modified recursion method, which is an exten-
sion of Haydock?s method.
16
Diffraction patterns were
determined for hemispherical 2H9266 scans over all polar and
azimuthal angles at fixed energies. Simulations were per-
formed for a parabolic cluster, with the emitting atom posi-
tioned at the focus of the solid paraboloid.
The simulation time scales with the product nN
2
H20849l
max
+1H20850
3
where n is the number of iterations, N the number of
atoms in the cluster, and l
max
the maximum angular momen-
tum quantum number for the scattering event, estimated from
the photoelectron energy.
15
N depends on energy through the
corresponding volume, proportional to the cube of the inelas-
tic mean free path, which is energy dependent.
17
For conver-
gence to be achieved at high resolution, the number of itera-
tions required was typically greater than 40. Together with
required cluster sizes of approximately 1300 atoms, this was
substantially greater than the literature values.
8
Separate simulations were run for emitters on different
atomic planes and the output added together in order to con-
struct a diffraction pattern. These simulations could thus be
run concurrently on separate processors in a high perfor-
mance computing environment. The composite data were
then displayed as a stereographic projection of an XPD
pattern.
III. R-FACTOR ANALYSIS
To examine simulation data quantitatively, a modified
R-factor formulation is used to compare the bulk truncated
CuH20853100H20854 with the relaxed, buckled, and Mn incorporated sur-
faces. The R-factor is found by first calculating a H9273 value that
is the H20849normalizedH20850 difference between the intensity at a point
IH20849H9258,H9278H20850 and the averaged azimuthal intensity over the corre-
sponding polar angle IH20849H9258H20850, for each data set,
18
H9273H20849H9277,H9278H20850 = H20851IH20849H9277,H9278H20850 ? I
0
H20849H9277H20850H20852/I
0
H20849H9277H20850. H208491H20850
Following Zheng et al.,
19
the R-factor was then determined
as the sum of squares of the difference of the H9273 values be-
tween two data sets employing an appropriate normalization,
R =
H20858
i
H20849H9273
ci
?H9273
ei
H20850
2
H20858
i
H20849H9273
ci
2
+H9273
ei
2
H20850
, H208492H20850
where H9273
ci
are the simulated and H9273
ei
the experimental H20849includ-
ing background subtractionH20850 H9273 values. For comparison pur-
poses in the present work the bulk truncated CuH20853100H20854 simu-
lation was taken as the ?experimental? H9273 values.
IV. CU?111? AT ?3P? ELECTRON KINETIC ENERGY
OF 523.5 EV
An experimental diffraction pattern for CuH20853111H20854 was ac-
quired at an electron kinetic energy of 523.5 eV using a
toroidal electron spectrometer designed and built by the Cen-
tre for Materials and Surface Science H20849CMSSH20850 at the La
Trobe University, Department of Physics, and now installed
and operational at the Berlin synchrotron facility BESSY.
13
The spectrometer allows the simultaneous analysis of a range
of kinetic energies of emitted electrons at all polar angles.
This approach dramatically reduces measurement time in
comparison to a conventional angle resolved photoelectron
spectrometer, which measures photoelectrons for one combi-
nation of polar and azimuthal angles at a time. The CMSS
spectrometer allows for the rotation of the sample for obtain-
ing data at different azimuth angles. By combining a set of
polar scans taken at different azimuth angles, a full hemi-
sphere scan can be constructed. Importantly the toroidal de-
sign and parallel nature of acquisition allows the full angular
resolution of the analyzer to be used without increasing data
acquisition times unduly. Typically three dimensional dif-
fraction patterns taken from single crystal Cu surfaces take
less than 30 min to acquire.
Figure 1 shows symmetrized experimental and EDAC
simulated diffraction patterns for CuH20853111H20854 at an electron ki-
netic energy of 523.5 eV from a Cu 3p core level. All dif-
fraction patterns shown are contrast images, where low and
high intensity features are shown as dark and light regions,
respectively. As the resolution of the toroidal analyzer is 0.5?
and 1? in the polar and azimuthal angles, respectively, the
angular resolution for both the azimuthal and polar angles in
the EDAC simulations was set to 0.5? and accordingly allows
the testing of features found in the experiment at this
resolution.
The results shown in Fig. 1, at a photoelectron kinetic
energy of 523.5 eV, may be compared with lower resolution
results obtained for CuH20853111H20854 using the 2p core level excited
with Mg KH9251 and Si KH9251 at electron kinetic energies of E
kin
=320.8 eV and E
kin
=807.2 eV, respectively.
14
These ex-
perimental results were successfully simulated by EDAC, with
the main features attributed to forward focusing and
Kikuchi-like bands.
093510-2 Cousland et al. J. Appl. Phys. 106, 093510 H208492009H20850
It should be emphasized that in order to reproduce the
high resolution of the present calculation compared with the
earlier work, larger clusters and a higher number of iterations
are required for the calculations to converge. In particular the
present calculation uses 44 iterations and cluster sizes of
H110112700 atoms, which is approximately ten times greater than
that of the earlier work. These findings for CuH20853111H20854 show that
EDAC clearly replicates the XPD experimental result at high
resolution, including forward focusing H20849labeled ?A? in fig-
ureH20850 and Kikuchi bands H20849BH20850.
V. CU?100? IN ?3P? ELECTRON KINETIC ENERGY
RANGE OF 73.5?303.5 EV
Figure 2 shows the results of XPD simulations from bulk
truncated CuH20853100H20854 at 300 K performed for electron kinetic
energies of 73.5, 103.5, 203.5, and 303.5 eV. The results are
displayed in Fig. 2 as contrast images of CuH20853100H20854 diffraction
patterns at 300 K, where photoelectrons are emitted from 3p
core levels. In these simulations the sample is moved
through polar and azimuthal angles, with an angle of 55?
being maintained between source and detector that are in the
same plane. This configuration corresponds to XPD synchro-
tron experiments performed using a SPECS 150 MCD elec-
tron spectrometer at the National Synchrotron Radiation Re-
search Centre, Taiwan. In this series of measurements the
detector was set to an angular resolution of 1? in both polar
and azimuthal directions, which were defined by a five axis
manipulator from OmniVac.
Forward focusing can be expected at energies above
H11011300 eV for XPD patterns,
5
but Fig. 2 predicts this at lower
energies. While not evident at 73.5 eV, the 103.5 eV XPD
pattern shows forward focusing along the H20851111H20852 principal
axes H20849label ?C? in Fig. 2H20850 as does the 203.5 eV pattern along
the H20851011H20852 axes H20849label DH20850. At 103.5 eV there is forward focus-
ing in the center of the pattern along the H20851001H20852 principal axis
H20849label EH20850 but this does not occur at 73.5 eV. As in LEED
experiments for CuH20853100H20854,
11
these real-space simulations pre-
FIG. 3. Stereographic projection of H20849aH20850 CuH20853100H20854 with surface relaxation of
H110021.10% for first, +1.70% second, +1.0% third interlayer spacing, H20849bH20850
CuH20853100H20854 with an 0.5 ML Mn overlayer that is buckled, and H20849cH20850 CuH20853100H20854 with
the same buckle as in H20849bH20850 for electron kinetic energy of 73.5 eV and at a
temperature of 300 K. Bulk truncated CuH20853100H20854 is displayed in Fig. 2. Fea-
tures in H20849bH20850 include Kikuchi-like bands H20849fH20850 that make the central white cross
H20849gH20850 appear smaller. The parameters for surface relaxation in H20849aH20850 come from
Davis and Noonan H20849Ref. 12H20850 and the buckle in H20849bH20850 and H20849cH20850 comes from
Wuttig et al. H20849Ref. 9H20850. R factors comparing data from these results with Fig.
2H20849aH20850 are shown in Table I.
FIG. 1. Comparison between experimental XPD patterns for CuH20853111H20854 with a
photoelectron kinetic energy of 523.5 eV and stereographically projected
EDAC simulations. H20849aH20850 Experimental H20849bH20850 simulation with photoelectron emit-
ted from the 3p shell at a kinetic energy of 523.5 eV at 300 K; H20849cH20850 the planar
and axial projections for CuH20853111H20854. LabelAmarks forward focusing along the
H20851001H20852 principal axes and label B Kikuchi bands.
FIG. 2. H20849Color onlineH20850 Stereographic projection of EDAC output at tempera-
ture of 300 K for electron kinetic energies H20849aH20850 73.5 eV, H20849bH20850 103.5 eV, H20849cH20850
203.5 eV, and H20849dH20850 303.5 eV. H20849eH20850 is the stereographic projection for bulk
truncated CuH20853100H20854. Label C marks the H20851111H20852 principal axis, D H20851011H20852,EH20851001H20852.
The images demonstrate forward focusing and Kikuchi-like bands at photo-
electron energies of 103.5 eV and above. Note the minimum in the H20851001H20852
center H20849eH20850 of the diffraction pattern at a photoelectron energy of 73.5 eV.
093510-3 Cousland et al. J. Appl. Phys. 106, 093510 H208492009H20850
dict that XPD patterns are energy dependent. For instance
XPD simulations show an overall drop in intensity at higher
energies as seen in LEED. It is also possible to identify XPD
features with LEED intensity variations. For example, the
central minimum for the H20851001H20852 axis at 73.5 eV corresponds to
the minimum of intensity in the 00 beam for LEED at
H1101173 eV.
11
Figures 3?6 show results of EDAC calculations for elec-
trons emitted from Cu 3p levels and various CuH20853100H20854 based
structures for kinetic energies of 73.5, 103.5, 203.5, and
303.5 eV. Each figure compares three surfaces: H20849aH20850 relaxed
CuH20853100H20854,
12
H20849bH20850 buckled CuH20853100H20854 with Mn overlayer,
9
and H20849cH20850
the same buckled CuH20853100H20854 with only Cu atoms. Bulk trun-
cated CuH20853100H20854 at the corresponding energies is displayed in
Fig. 2.
Results of R-factor calculations are displayed in Table I.
They show comparisons between the results displayed in Fig.
2 with those of Figs. 3?6 using a scheme based on Eqs. H208491H20850
and H208492H20850. Overall the diffraction patterns for CuH20853100H20854 with
surface relaxation closely resemble those for bulk truncated
CuH20853100H20854 at all energies shown. As seen in Table I, this can be
further verified by the low R factors.
Considering results at a kinetic energy of 73.5 eV, Fig.
3H20849bH20850 shows that CuH20853100H20854 with a buckle and Mn overlayer is
especially distinguished by dark lines corresponding to sec-
tions of H20853111H20854 Kikuchi-like bands H20849labeled ?F? in figureH20850 that
surround the central region. The R factor is 0.26, the highest
value obtained H20849see Table IH20850. These bands make the central
white cross H20849label GH20850 in Fig. 3H20849bH20850 appear smaller than in the
other XPD patterns of Fig. 3.At a kinetic energy of 103.5 eV
Fig. 4H20849bH20850 shows that the pattern for buckled CuH20853100H20854 with Mn
overlayer might appear most distinct, but the R-factor value
of 0.21 is in fact slightly smaller than for the buckled only
Cu atoms H20849R factor of 0.22H20850. In particular the H20853111H20854 Kikuchi
patterns H20849label ?H?H20850 together with four intensity minima that
can be associated with H20851112H20852 axes H20849label IH20850 appear more
clearly in Fig. 4H20849bH20850.
Finally, the visual differences between diffraction pat-
terns at 203.5 and 303.5 eV H20849Figs. 5 and 6, respectivelyH20850 for
relaxed and bulk truncated surfaces as well as for the Mn/Cu
systems investigated here are quite subtle and show that
higher energy patterns are less sensitive to fine changes in
surface structure. At 203.5 eV, the Mn/Cu pattern shows
some small differences, such as less distinct Kikuchi lines
H20849label ?J? in Fig. 5H20850 and generally lower intensity outside the
central region of the pattern. Any distinction at 303.5 eV is
difficult to discriminate, with little visible evidence of differ-
ence in the patterns for the various structures. The R factors
at 303.5 eV H20849see Table IH20850 also indicate this, with values of
0.0012 for the relaxed surface of Davis and Noonan
12
and
0.0031 for the surface due to Fowler and Barth.
20
At 303.5
eV the R factor of 0.24 for Cu with a Mn overlayer is close
to the value of 0.20 for the buckled Cu surface.
While overall the diffraction patterns for CuH20853100H20854 with
surface relaxation closely resemble those for bulk truncated
CuH20853100H20854 at all energies shown, and also evidenced by the R
factors given in Table I, interestingly, however, low energy
azimuthal diffraction scans show quite large differences be-
tween the currently different proposed relaxation models of
FIG. 5. Contrast images with electron kinetic energy of 203.5 eV. Other
parameters as in Fig. 3. Although variation between patterns is more subtle
than at lower energies, differences can still be observed. The comparison of
relaxed H20849aH20850 and buckled H20849cH20850 CuH20853100H20854 simulations indicates significant simi-
larity, and CuH20853100H20854 with an Mn overlayer H20849bH20850 exhibits some extra features,
including less distinct Kikuchi bands H20849jH20850 and a less intense H20849darkerH20850 outer
region.
FIG. 6. Contrast images at an electron kinetic energy of 303.5 eV. Other
parameters as in Fig. 3. No obvious differences are observed for the patterns
shown at this energy.
FIG. 7. H20849Color onlineH20850 Comparison of Cu 3p intensities for azimuthal scans
at polar angle of 30? for CuH20853100H20854 at 300 K and electron kinetic energy of
103.5 eV. Surface relaxation parameters from H20849+H20850 Davis and Noonan, H9004d
12
?1.10%, H9004d
23
+1.70%, H9004d
34
+1.0% H20849Ref. 12H20850, H20849H17034H20850 Fowler and Barth, H9004d
12
?2.0%, H9004d
23
+1.0% H20849Ref. 20H20850, H20849H17005H20850 without surface relaxation. The intensity
scale is linear. Note that at 0? and 90? the 3p intensities for relaxed surfaces
coalesce, whereas at 45? the intensity of the surface from Fowler and Barth
exhibits difference.
FIG. 4. Simulated diffraction patterns with electron kinetic energy of 103.5
eV. Other parameters as in Fig. 3. Extra features in H20849bH20850 include stronger
Kikuchi patterns H20849hH20850 and four intensity minima associated with H20851112H20852
axes H20849iH20850.
093510-4 Cousland et al. J. Appl. Phys. 106, 093510 H208492009H20850
CuH20853100H20854. Comparison between these
12,20
show differences as
great as 20%?30% of the total maximum intensity variation
within a particular azimuthal scan, see Fig. 7. This is further
evidenced by the R factors at 103.5 eV of 0.0015 for Davis
and Noonan
12
surface relaxed CuH20853100H20854
20
and 0.0054 for the
surface due to Fowler and Barth.
20
Although small, these
differences may be quantified by further R-factor
analysis
21,22
or compared visually for particular scans as in-
dicated by Fig. 7. These effects may be more evident at
particular energies and so this method may use photoelec-
trons at kinetic energies optimized according to the surface
being considered. Low energy photoelectron diffraction may
be more sensitive than LEED in this regard as azimuthal
distributions yield a plethora of information not attainable by
LEED angular patterns from well ordered single crystalline
surfaces.
VI. CONCLUSION
We have demonstrated how EDAC can be run at high
resolution to successfully simulate highly detailed XPD pat-
terns for the CuH20853111H20854 surface. In addition we have shown
how EDAC predicts visible changes in CuH20853100H20854 XPD patterns
as a function of energy and how working in a low kinetic
energy regime may be used to determine structure. Low en-
ergy XPD has features in common with LEED, including an
overall drop in intensity at higher energies and strong varia-
tions in intensity for the individual diffraction features. We
have predicted that surface relaxation has little visible effect
compared to bulk truncated CuH20853100H20854 for photoelectron ener-
gies of 73.5 and 103.5 eV as has been found in LEED. Un-
like conventional LEED, however, there are some conditions
in k-E space from low energy photoelectron diffraction that
do indeed show significant differences between competing
models in the low kinetic energy regime. This was evidenced
in CuH20853100H20854 surfaces with small relaxation parameter differ-
ences by both R-factor analysis of full data sets and visual
comparison of an azimuthal scan. Thus low energy photo-
electron diffraction may be more sensitive than LEED in
determining small changes to surface structure because of the
effect of multiple scattering from core-level electrons that
yields incredibly detailed diffraction patterns. R-factor analy-
sis used to compare bulk truncated CuH20853100H20854 with the other
cases indicates a noticeable effect caused by introducing
buckles and an Mn overlayer at all energies simulated. Al-
though visible similarity would indicate that there is little
difference in diffraction patterns at 303.5 eV, R-factor analy-
sis indicates differences in intensities comparable to those of
patterns at lower energies investigated. Finally we conclude
that high resolution XPD simulations for CuH20853100H20854 at 300 K
can visibly discriminate element types and buckles in sur-
faces, at kinetic energies below 303.5 eV, compared to the
bulk truncated case, and that having performed high reso-
lution experiments, a combination of data sets R factor
analysis and visual analysis will provide a method to accu-
rately distinguish structural models.
ACKNOWLEDGMENTS
G.C. would like to thank Dr. Steve Homolya for running
the original CuH20853111H20854 simulation and providing shell scripts
and many helpful ideas about operating EDAC. The simula-
tions were run at the Victorian Partnership for Advanced
Computing.
1
D. P. Woodruff, J. Electron Spectrosc. Relat. Phenom. 126,55H208492002H20850.
2
C. S. Fadley, in Synchrotron Radiation Research: Advances in Surface and
Interface Science, edited by R. Z. Bachrach H20849Plenum, New York, 1992H20850,
Vol. 1, p. 421.
3
C. Westphal, Surf. Sci. Rep. 50,1H208492003H20850.
4
J. Osterwalder, P. Aebi, R. Fasel, D. Naumovic, P. Schwaller, T. Kreutz, L.
Schlapbach, T. Abukawa, and S. Kono, Surf. Sci. 331?333, 1002 H208491995H20850.
5
W. F. Egelhoff, Jr., Crit. Rev. Solid State Mater. Sci. 16,213H208491990H20850.
6
D. Naumovic, A. Stuck, T. Greber, J. Osterwalder, and L. Schlapbach,
Phys. Rev. B 47, 7462 H208491993H20850.
7
M.-L. Xu, J. J. Barton, and M. A. Van Hove, Phys. Rev. B 39,8275
H208491989H20850.
8
F. J. Garcia deAbajo, M. A. Van Hove, and C. S. Fadley, Phys. Rev. B 63,
075404 H208492001H20850.
9
M. Wuttig, C. C. Knight, T. Flores, and Y. Gauthier, Surf. Sci. 292,189
H208491993H20850.
10
G. P. Cousland, A. E. Smith, J. Riley, S. Homolya, A. Stampfl, and J.
King-Lacroix, Proceedings of the 32nd Condensed Matter and Materials
Meeting, 29 JanuaryH110021 February 2008, Wagga Wagga. Available from:
http://www.aip.org.au/wagga2008/.
11
J. B. Pendry, J. Phys. C 4, 2514 H208491971H20850.
12
H. L. Davis and J. R. Noonan, J. Vac. Sci. Technol. 20, 842 H208491982H20850.
13
L. Broekman, A. Tadich, E. Huwald, J. Riley, R. Leckey, T. Seyller, K.
Emtsev, and L. Ley, J. Electron Spectrosc. Relat. Phenom. 144?147, 1001
H208492005H20850.
14
L. Despont, D. Naumovic, F. Clerc, C. Koitzsch, M. G. Garnier, F. J.
Garcia deAbajo, M. A. Van Hove, and P. Aebi, Surf. Sci. 600,380H208492006H20850.
15
J. B. Pendry, Low Energy Electron Diffraction H20849Academic, London, 1974H20850.
16
R. Haydock, Solid State Phys. 35, 215 H208491980H20850.
17
C. J. Powell and A. Jablonski, J. Phys. Chem. Ref. Data 28,19H208491999H20850.
TABLE I. R-factor comparisons between simulated values of bulk truncated CuH20853100H20854 with the simulated data
for H20849aH20850 CuH20853100H20854 with surface relaxation parameters from Davis and Noonan H20849Ref. 12H20850 and H20849bH20850 with surface
relaxation of H110022.0% first, 1.0% second interlayer spacing from Fowler and Barth H20849Ref. 20H20850, H20849cH20850 CuH20853100H20854 with a
Mn overlayer that is buckled, and H20849dH20850 CuH20853100H20854 with the same buckle as in H20849cH20850. See Fig. 3 for details of structural
parameters.
Kinetic energy
H20849eVH20850
Relaxed CuH20853100H20854,
Davis and Noonan
Relaxed CuH20853100H20854,
Fowler and Barth
CuH20853100H20854 with buckle
and Mn overlayer
CuH20853100H20854
with buckle
73.5 0.0041 0.0061 0.26 0.16
103.5 0.0015 0.0054 0.21 0.22
203.5 0.0042 0.0068 0.17 0.14
303.5 0.0012 0.0031 0.24 0.20
093510-5 Cousland et al. J. Appl. Phys. 106, 093510 H208492009H20850
18
R. Fasel, P. Aebi, J. Osterwalder, L. Schlapbach, R. G. Agostino, and G.
Chiarello, Phys. Rev. B 50, 14516 H208491994H20850.
19
J.-C. Zheng, C. H. A. Huan, A. T. S. Wee, M. A. Van Hove, C. S. Fadley,
F. J. Shi, E. Rotenberg, S. R. Barman, J. J. Paggel, K. Horn, Ph. Ebert, and
K. Urban, Phys. Rev. B 69, 134107 H208492004H20850.
20
D. E. Fowler and J. V. Barth, Phys. Rev. B 52,2117H208491995H20850.
21
E. Zanazzi and F. Jona, Surf. Sci. 62,61H208491977H20850.
22
J. B. Pendry, J. Phys. C 13, 937 H208491980H20850.
093510-6 Cousland et al. J. Appl. Phys. 106, 093510 H208492009H20850