Mg clusters on MgO surfaces: Characterization with metastable impact
electron spectroscopy, ultraviolet photoelectron spectroscopy „HeI…,
and temperature programmed desorption
J. Günster, J. Stultz, S. Krischok, and D. W. Goodmana),b)
Department of Chemistry, Texas A&M University, College Station, Texas 77842-3012
P. Stracke and V. Kemptera)
Physikalisches Institut, Technischen Universität Clausthal Leibnizstrasse 4,
D-38678 Clausthal-Zellerfeld, Germany
~Received 14 October 1998; accepted 19 April 1999!
MgO films ~2 nm thick! were grown on Mo and W substrates while metastable impact electron
~MIES! and ultraviolet photoelectron spectroscopy ~UPS! ~HeI! spectra were collected in situ. Apart
from the valence band emission no additional spectral features have been detected with electron
spectroscopies. After exposing the oxide surface to Mg ~substrate temperature between 100 and 300
K! an additional peak not seen with UPS, located within the band gap, shows up in MIES. This
band, located at ;2 eV above the top of the valence band with a full width at half maximum of
;1 eV at the lowest exposures, can be detected in MIES until its intensity falls below a level of
1023 of that from the valence band. This additional emission is attributed to the formation of small,
nonmetallic Mg clusters. The energetic position of the cluster emission closely matches that of the
expected ionization of surface F 1
S /F S centers. It is proposed that in the initial phase of the Mg
exposure, F center-type defects are produced close to extended defects, such as steps, corners, etc.
These Mg-induced defects appear to play an important role as nucleation sites for cluster formation.
At sufficiently large exposures a densely packed Mg film forms. © 1999 American Vacuum
Society. @S0734-2101~99!20504-0#
I. INTRODUCTION
Several reasons exist why considerable attention is devoted to the addition of metal atoms to metal oxides surfaces.
Metal addition to oxides leads to an enhanced reactivity via
electron transfer to a variety of adsorbed molecules leading
to the formation of radical anion species.1,2 Moreover, the
interaction between the metal, in particular clusters, and
metal oxide supports plays a key role in a number of technologically important applications including catalysis,3,4
metal–ceramics composites, gas sensors, and microelectronics.5 The study of the interaction between metal atoms and oxide surfaces is of importance in understanding
phenomena like segregation of metal atoms at oxide
surfaces,6,7 diffusion of metal atoms on insulators,8 the formation of metal-induced point defects on oxides,2,9 and the
formation of nanoparticles in semiconductors and
insulators.10
Surprisingly little is known about the electronic structure
of metal/oxide interfaces, in particular about the type of the
metal adsorption sites, the electronic character of the firstlayer metal species, and regarding the nature of the interaction with nearest-neighbor metal atoms.5,11 Electronic structural studies of the early stage of metal growth on oxides are
complicated because nucleation occurs preferentially at defect sites, such as vacancies, or extended defect sites, such as
steps, kinks, etc. Unfortunately, the nature and extent of the
a!
Authors to whom correspondence should be addressed.
Electronic mail: [email protected]
b!
1657
J. Vac. Sci. Technol. A 17„4…, Jul/Aug 1999
defects are normally not well controlled in studies of metal
adsorption on metal oxides.
The present study sheds light on the early stage of Mg
cluster formation on MgO. Our main interest is the study of
the nucleation mechanism and the identification of the sites
at which the nucleation takes place. Often photoemission is
employed as a probe of electronic structure in studies of
metal aggregates on surfaces. When employed on insulator
surfaces, problems are encountered due to charging of the
surface. This problem is circumvented in the present work,
as suggested in Refs. 3 and 4, by using ultrathin oxide films
rather than bulk oxides. These films are thin enough to avoid
charging, but thick enough to exhibit the electronic structure
of the bulk oxide.3 However, in this approach spectral features from the density of states ~DOS! of the supporting substrate often obscure the electronic structure of very small
clusters.12 For this reason we have applied metastable impact
electron spectroscopy ~MIES! to measure changes of the
electronic structure during the early stage of cluster formation. MIES is only sensitive to the DOS of the film and the
charge density of the species adsorbed on the outermost
layer; it is not sensitive to the DOS of the underlying substrate supporting the oxide film. Moreover, MIES is very
sensitive to features resulting from s-valence electrons, such
as those encountered when studying the adsorption of alkali
and alkaline earth atoms.
In this investigation detailed studies also have been carried out on small nonmetallic clusters at metal exposures far
below the critical coverage required for multilayer
formation.5,13 In order to gather information on the atomic
0734-2101/99/17„4…/1657/6/$15.00
©1999 American Vacuum Society
1657
1658
Günster et al.: Mg clusters on MgO surfaces: Characterization with MIES
1658
structure of the Mg clusters, MIES has been combined with
temperature programmed desorption ~TPD!. We have recently demonstrated the potential of combining these two
techniques.14
II. EXPERIMENT
The apparatus used in these studies has been described
previously.15,16 Briefly, it is equipped with a cold-cathode
gas discharge source for the production of metastable
He* ( 3 S/ 1 S) (E * 519.8/20.6 eV) atoms with thermal kinetic
energy and HeI photons (E * 521.2 eV) as a source for ultraviolet photoelectron spectroscopy ~UPS!. The intensity ratio 3 S/ 1 S is found to be 7:1. Additionally, the apparatus is
equipped with x-ray photoelectron spectroscopy ~XPS!, Auger electron spectroscopy ~AES!, and low energy electron
diffraction ~LEED!. Metastable and photon contributions
within the beam were separated by means of a time-of-flight
method using a mechanical chopper. MIES and UPS spectra
were acquired with incident photon/metastable beams 45°
with respect to the surface normal; electrons emitted in the
direction normal to the surface are analyzed. Collection of a
MIES/UPS spectrum requires approximately 120 s. The
measurements were performed using a hemispherical analyzer ~Leybold EA10/100! with an energy resolution of 250
meV for MIES/UPS. Since the metallic substrate and the
analyzer are in electrical contact, the Fermi energy of the
substrate appears at a constant energy and permits the work
function change of the surface to be measured directly from
the high energy cutoff of the electron spectra.17
A second apparatus, described elsewhere,18 equipped with
a MIES/UPS source of the same type as described above and
a setup for TPD was used to calibrate the Mg coverages.
Briefly, TPD spectra were collected with a differentially
pumped quadrupole mass filter in line-of-sight to the sample
while ramping the sample temperature linearly by 3 K/s. In
addition, this second apparatus is equipped with XPS, AES,
and LEED.
MgO layers with an approximate thickness of 2 nm were
prepared by evaporating Mg on a Mo~100! and W~110! substrate at room temperature, followed by a subsequent annealing to 800 K in oxygen ambient. As shown in previous MIES
and UPS studies,18,19 the as-prepared MgO films are very
similar to MgO~100! single crystals with respect to their
electronic properties.
III. RESULTS
A. Electron spectroscopy
Figure 1 shows MIES spectra acquired during Mg exposure to the MgO-covered Mo~100! surface at 100 K; similar
results are obtained at 300 K. As discussed previously,20
MIES spectra acquired from the MgO film, prior to Mg exposure ~bottom spectrum!, are essentially the MgO surface
density of states ~SDOS! as seen via an Auger deexcitation
~AD! process. Thus, the structure observed between 4 and 10
eV binding energy with respect to E F is due to ionization of
the MgO valence band states with O(2p) character. TypiJ. Vac. Sci. Technol. A, Vol. 17, No. 4, Jul/Aug 1999
FIG. 1. MIES spectra acquired from the MgO surface as a function of the
Mg exposure. The bottom spectrum shows the clean MgO surface; the topmost, the fully covered one. Inset: work function change vs exposure time.
The dashed spectra were acquired near the work function minimum.
cally for the MgO surface band gap, no intensity between E F
and 3.8 eV binding energy is apparent. Good agreement is
found between the SDOS derived from MIES and theoretical
results.19 The UPS data ~not shown here; see Refs. 19 and
20! provide essentially the same information as MIES. However, the valence band of the clean MgO surface reveals a
two-peak structure between 4 and 10 eV, which is discussed
in detail in Ref. 19.
As a consequence of the Mg dosing, an additional peak,
located at ;2 eV above the top of the valence band ~2 eV
binding energy!, develops within the band gap. Since the
work function of the MgO film ~2.7 eV! is considerably
lower than the He 2s binding energy, no unoccupied states
are available at the surface into which resonant transfer of
the 2s electron can take place, even when taking into account an eventual shift of the 2s level during the He* interaction with the surface. Consequently, the band gap feature
also arises from an AD process ~involving electrons in Mginduced occupied states below the Fermi level!. At larger Mg
exposures the MgO valence band emission between 4 and 10
eV weakens considerably; the disappearance of the O(2 p)
structure in the topmost MIES spectrum in Fig. 1 indicates
that the entire surface is covered by Mg. The shape of the
resulting structure is very similar to that seen for Mg films.
Also shown in Fig. 1 ~see inset! is the work function
change during the Mg exposure. When Mg atoms are dosed
to the oxide, the work function decreases by ;0.5 eV while
the valence band intensity decreases by ;10% ~dotted MIES
spectra!. Simultaneously, the top of the valence band shifts
towards larger binding energy by approximately the same
amount. A decrease of the work function upon exposure to
metal atoms is generally attributed to charge transfer from
1659
Günster et al.: Mg clusters on MgO surfaces: Characterization with MIES
1659
FIG. 3. Intensity, energetic position, and FWHM of the band gap feature vs
exposure time.
FIG. 2. MIES spectra acquired from the MgO surface as a function of the
Mg exposure. The bottom spectrum shows the clean MgO surface. The
topmost spectrum corresponds to spectrum 4 of Fig. 1.
the adatoms to the substrate.5 This, in turn, indicates that the
adsorbed atoms exhibit some degree of ionicity. At larger
exposures the work function plateaus at a level typical for
metallic Mg films ~3.6 eV!, consistent with a model in which
Mg islands grow with lateral bonding similar to bulk Mg.
In order to gain more detailed information about the
changes in the low binding energy region during Mg dosage
to the MgO surface, MIES spectra were acquired using a
lower Mg evaporation rate and a higher energy resolution.
Data for MgO on W~110! at 300 K are presented in Fig. 2.
The top spectrum of Fig. 2 corresponds to spectrum 4 of Fig.
1. The bottom spectrum shows ~as does Fig. 1! the clean
MgO surface. No Mg-induced density of states is evident at
E F over the entire exposure range in Fig. 2. Thus, the species
responsible for the band gap feature exhibits nonmetallic behavior. The band gap feature can be detected until its intensity falls below a level of 1023 from that of the valence band
O(2p) emission of the clean MgO surface. As shown in Fig.
3, both the energy position and the peak width depend
~weakly! on the exposure time. A Gaussian-shaped DOS of
the band gap feature is observed where the full width at half
maximum ~FWHM! increases from ;1 to 1.9 eV over the
exposure range. Additional measurements show that the
band gap feature is stable up to 500 K and has virtually
disappeared upon heating to 600 K. UPS data have also been
collected simultaneously to the MIES spectra presented in
Fig. 2 ~not shown here!. These measurements show no
changes within the range of exposures; in particular, the Mginduced band gap feature is not apparent.
A similar band gap feature was obtained when dosing Na
atoms; however, it is less stable thermally and disappears
between 350 and 400 K upon annealing. Li and Cs additives
on MgO have been shown to produce similar band gap
features.20,21
JVST A - Vacuum, Surfaces, and Films
B. Temperature programmed desorption
TPD is an excellent technique for determining surface
coverages and the lateral interaction between adsorbates.
Figure 4 compares TPD and MIES spectra acquired from a
Mg-covered MgO surface as a function of the Mg exposure.
Starting at the uppermost MIES spectrum acquired for the
clean MgO surface, the Mg coverage increases monotonically towards the bottom spectrum. In the MIES spectrum
for the highest Mg coverage ~bottom spectrum!, the O(2p)
band has essentially disappeared, indicating complete coverage of the MgO surface by Mg. The relative peak area of the
corresponding TPD Mg feature (m/e524) is 264 times
larger than the Mg feature in the uppermost TPD spectrum,
which corresponds to the detection threshold of our mass
FIG. 4. Comparison between TPD and MIES data. The TPD spectra were
taken from the surface characterized with MIES. The topmost MIES spectrum shows the clean MgO surface; the Mg coverage increases monotonically towards the bottom spectrum.
1660
Günster et al.: Mg clusters on MgO surfaces: Characterization with MIES
FIG. 5. Mg-TPD and MIES O(2 p) peak intensities from Fig. 4 vs the deposition time on MgO~100! at 295 K.
filter. However, even at this very low coverage, the Mginduced band gap feature appears fully developed in MIES.
Analysis of the Mg TPD peaks shows an exponential increase in intensity toward lower temperature with coverage,
indicating that the desorption follows zeroth-order kinetics.
Since this behavior is typically observed for desorption with
scission of adsorbate–adsorbate bonds, this result suggests
that even at the lowest Mg concentrations in Fig. 4, formation of three-dimensional ~3D! islands occurs. This conclusion is supported by the nonlinear Mg uptake versus exposure time in Fig. 5 and is inconsistent with a simple layerby-layer growth mode and a constant sticking probability, as
postulated for the adsorption of potassium on MgO~100! at
150 and 295 K.22
The TPD data have been analyzed using a complete
analysis23 in which an Arrhenius plot of the natural logarithm of the desorption rates yields a straight line for a particular Mg coverage ~see Fig. 6!. The slope of each line
corresponds to the activation desorption energy (E des) for
that particular coverage. The inset in Fig. 6 shows that the
desorption energy increases initially from 100 kJ/mol at very
low coverages up to 138 kJ/mol. This latter value is close to
140.5 kJ/mol, the tabulated value for the heat of sublimation
for bulk Mg.24 A comparison of Figs. 4 and 6 indicates the
formation of Mg clusters with bulk metal properties with
respect to the Mg–Mg bond strength. The results of Figs. 1
and 2 also suggest the formation of bulk Mg clusters with
respect to their electronic structure. Such clustering occurs in
a Mg coverage regime where the band gap feature is fully
developed and the work function has traversed through its
minimum.
IV. DISCUSSION
Basically the uppermost spectrum in Fig. 1 shows a metallic Mg layer on the MgO substrate. This interpretation is
supported by the onset of the spectrum at E F and the corresponding TPD data. The clean MgO surface ~bottom specJ. Vac. Sci. Technol. A, Vol. 17, No. 4, Jul/Aug 1999
1660
FIG. 6. Arrhenius plot of the desorption rates obtained in the complete
analysis of the TPD data in Fig. 4. Inset: calculated desorption activation
energy vs relative Mg coverage.
trum in Figs. 1 and 2! shows no intensity near E F . However
immediately after the onset of Mg exposure, a peak develops
within the surface band gap. This peak is near E F and, due to
the limited resolution of our electron spectrometer, attenuates towards E F . Since it is well separated from the MgO
valence band emission, we believe this feature corresponds
to a surface defect state rather than metallic behavior of the
adsorbed species. At higher Mg coverages in Fig. 1 this peak
develops into a continuum near E F indicating the metallic
identity of the surface. Figure 2 shows data for a MgOcovered W~100! substrate exposed to a low flux of Mg. The
MIES spectra were collected with enhanced energy resolution. The maximum Mg exposure in Fig. 2 corresponds approximately to the fourth spectrum from the bottom in Fig. 1.
For that reason, no metallization of the adsorbed Mg species
in Fig. 2 occurs since there is no intensity near E f . However
the band gap feature appears fully developed.
Due to the Mg-induced feature located within the band
gap, dramatic changes of the intensity near E F are only observed at the onset of the Mg dosage. Metallization of the
surface is manifested in the change in the spectra from a
distinct feature in the band gap to a continuum.
A. Origin of the Mg-induced band gap feature
The origin of the band gap feature becomes clear from an
inspection of the results obtained during the exposure of ordered Al2O3 films to Pd.1 At low Pd exposures a very similar
Gaussian-shaped, band gap feature, roughly 2 eV wide and
centered ;2 eV below the Fermi level, appears. The Pdinduced feature for small exposures ~where no density of
states at E F is evident! is characteristic of clusters exhibiting
nonmetallic behavior. Consequently, we attribute the Mginduced band gap feature seen in Figs. 1 and 2 to molecular
Mg clusters. Figure 3 displays a linear rise of the clusterinduced emission up to a value that is about 30 times the
observed threshold value; a minimum in the work function
1661
Günster et al.: Mg clusters on MgO surfaces: Characterization with MIES
occurs at this point ~see Fig. 1!. At the highest exposure
studied in Figs. 1 and 4, the valence band emission of the
MgO film has essentially disappeared indicating that the
MgO surface is covered by at least a monolayer ~ML! of Mg.
Furthermore, from Fig. 4 it is evident that a fraction ~0.4%!
of this maximum coverage leads to a saturation of the intensity in the band gap. This gives an estimated detection
threshold of the Mg-induced feature in MIES of about 0.01%
of the maximum coverage in Fig. 4. In this context, a closer
inspection of the attenuation of the O(2 p) valence band
emission versus the Mg coverage yields a nearly parabolic
Mg uptake of the MgO surface while the O(2p) intensity
decreases linearly ~see also Fig. 5!. These data suggest 3D
island growth during Mg adsorption ~Volmer–Weber growth
mode!. For that reason it is difficult to estimate the absolute
Mg coverage near the detection threshold for the Mginduced features in MIES. However, an estimate of ;3 ML
(62 ML) Mg at maximum exposures in Figs. 1 and 4 appears realistic and yields a detection threshold of about
0.03%60.02% of a monolayer Mg. From this value we estimate that about 1012 sites/cm2 exist on the MgO surface at
which Mg nucleates at a density on the same order of magnitude found for a graphite surface.11
B. Identification of the nucleation sites
Clues for the identification of the Mg nucleation sites
come from EPR studies on MgO exposed to Mg and alkali
atoms,1,2,9 and from electron energy-loss spectroscopy
~EELS! studies during the growth of nonstoichiometric MgO
films.25 First, the EPR studies show that at low exposures F s
centers are generated on the MgO surface by the addition of
small amounts of Mg ~alkali! vapor. These results are corroborated by the numerical results obtained for the interaction of Rb atoms with MgO surfaces.26 Moreover, for larger
exposures of alkali, there is ample evidence that monomeric
and trimeric cationic centers are formed. For these cationic
centers the valence electron charge density is shared between
the metal aggregate and a suitable surface trap, e.g., an anion
vacancy at a step of the MgO film.2,9 Second, the EELS
studies of Ref. 25 indicate that F s centers are produced when
growing a nonstoichiometric MgO film with excess Mg at
the surface. Both findings suggest that in the nucleation
phase the Mg-induced formation of F s centers plays an important role. The valence charge of the Mg species can be
thought to be shared by the adatoms of the cluster and a
suitable surface trap. Our results ~Fig. 1! display a decrease
of the surface work function by about 0.5 eV for the Mg
exposure range studied; this suggests that charge transfer occurs during the formation of the Mg clusters.
The position of the Mg-induced peak with respect to the
top of the valence band of MgO is in close agreement with
that of F s centers on ideal and defective MgO surfaces.27,28
These calculations predict values between 2 and 2.7 eV depending on the location of the F s centers with respect to
eventual surface defects, as corners, steps, kinks, etc.
The shape of the cluster feature resembles that of an inhomogeneously broadened photoemission line. The width of
JVST A - Vacuum, Surfaces, and Films
1661
the line could reflect the probability for filling the vacancy
produced in the photoemission process at a particular cluster
atom in an electron transfer process from neighboring cluster
atoms or from the oxide film. We estimate that the lifetime
of the vacancy is of the order of 7310216 s.
C. Adsorption energetics of the Mg-induced species
It was noted in Sec. III that at large exposures the shape
and the energetic position of the band gap emission as well
as the work function approach the values of a metallic Mg
film. This is strong evidence that at this coverage large areas
of metallic Mg have formed. This interpretation is supported
by an increase of the activation desorption energies (E des) of
the Mg adatoms to a value close to that of bulk Mg. Thus far
the formation of Mg clusters on MgO has not been treated
theoretically; however, Cu clusters (n51 – 13) on an ideal
MgO have been modeled.29 While the adsorbate–substrate
interaction energy ~per Cu atom! decreases with increasing n,
the adsorbate–adsorbate interaction ~Cu–Cu cohesive energy! increases. Thus, the interaction between the Cu atoms
inside the cluster dominates the adsorption process. It remains to be verified that this interpretation also holds for
adsorption on defective MgO. However, the initially lower
E des for low Mg coverages in Figs. 4 and 6 suggests a significantly smaller Mg binding energy in small Mg clusters on
MgO~100! than in bulk Mg, i.e., large Mg clusters with metallic properties. As for the adsorption of alkali atoms on
metals, the work function exhibits a minimum before rising
toward the metallic Mg value. If we assume that MIES detects Mg dimers initially as the smallest stable cluster species, the clusters corresponding to the work function minimum are estimated to contain 50–100 Mg atoms. At this
cluster size metallization of the cluster takes place.11 The
phase transformation from molecular to metallic clusters appears to be associated with backdonation of charge from the
oxide film to the Mg cluster leading to the observed rise in
the work function.
In summary, the considerations in Secs. IV A–IV C suggest the following for the growth of Mg clusters. Initially,
Mg atoms adsorb near extended defects, such as steps, corners, etc. Stabilization takes place via the formation of
F s -type defects, whereby Mg valence electrons are shared
between the adsorbed Mg species and a suitable surface trap.
Such a surface trap can be either an intrinsic defect on the
MgO surface or one created during Mg exposure, with the
latter more likely. This follows since the MgO surface has
been oxidized thoroughly before Mg dosage and thus is expected to be free of oxygen vacancies. With increasing exposure these sites act as nucleation sites for the adsorption of
additional Mg atoms. Small two-dimensional Mg clusters are
formed whereby the valence charge density is shared between the cluster and a suitable surface trap. The identity of
the responsible surface traps remains to be determined.
V. SUMMARY
MgO films ~;2 nm thick! between 100 and 300 K were
exposed to Mg atoms. The in situ changes in the surface
1662
Günster et al.: Mg clusters on MgO surfaces: Characterization with MIES
electronic structure were monitored with MIES and
UPS~HeI! during the exposure process. The as-prepared
MgO films display emission from the ionization of the valence band states of essentially O(2 p) character. No additional features ~attributable to occupied point defects! can be
seen in the band gap. An additional peak, not seen with
UPS~HeI!, develops in the band gap during the Mg exposure
and is present in MIES until its intensity fell below a level of
1023 of that of the valence band. The band gap feature appears approximately 2 eV above the top of the valence band
with ;1.5 eV FWHM; both position and peak width depend
weakly on the Mg exposure. We attribute this additional
emission to the formation of small, nonmetallic Mg clusters.
Apparently, Mg atoms adsorb initially near surface defects,
such as steps, corners, etc. Mg valence electrons are shared
with a suitable surface trap. At larger exposure these sites act
as nucleation sites for the adsorption of additional Mg atoms
in the form of small molecular Mg clusters.
ACKNOWLEDGMENTS
The authors acknowledge with pleasure the support of
this work by the National Science Foundation ~Contract
No. DMR 9423707!, the European Union through the
HCM network ‘‘Charge Transfer Processes at Surfaces’’
~ERBCHRXCT 940571!, and the DAAD ~313-ARC-XI-97/
51!.
1
E. Giamello, A. Ferrero, S. Colluccia, and A. Zecchina, J. Phys. Chem.
95, 9385 ~1991!.
2
D. Murphy and E. Giamello J. Phys. Chem. 99, 15172 ~1995!.
3
D. W. Goodman, Chem. Rev. 95, 523 ~1995!.
4
C. Xu and W. Goodman, Handbook of Heterogeneous Catalysis ~UCH,
Weinheim, 1997!.
5
C. T. Campbell, Surf. Sci. Rep. 27, 1 ~1997!.
J. Vac. Sci. Technol. A, Vol. 17, No. 4, Jul/Aug 1999
6
1662
R. Souda, Y. Hwang, T. Aizawa, W. Hayami, K. Oyoshi, and S. Hishita,
Surf. Sci. 387, 136 ~1997!.
7
T. Suzuki, S. Hishita, K. Oyoshi, and R. Souda, Surf. Sci. 391, L1243
~1997!.
8
T. Kizuka and N. Tanaka, Surf. Sci. 386, 249 ~1997!.
9
E. Giamello and D. Murphy, Mol. Eng. 4, 147 ~1994!.
10
Y. Kanzawa, T. Kageyama, S. Takeoka, M. Fujii, S. Hayashi, and K.
Yamamoto, Solid State Commun. 102, 533 ~1997!.
11
G. K. Wertheim, Z. Phys. D 12, 319 ~1989!.
12
Y. Q. Cai, A. M. Bradshaw, Q. Guo, and D. W. Goodman, Surf. Sci. 399,
L357 ~1998!.
13
A. P. Janssen, R. C. Schoonmaker, and A. Chambers, Surf. Sci. 49, 143
~1975!.
14
J. Günster, G. Liu, J. Stultz, and D. W. Goodman, J. Chem. Phys. 110,
2558 ~1999!.
15
W. Maus-Friedrichs, M. Wehrhahn, S. Dieckhoff, and V. Kempter, Surf.
Sci. 237, 257 ~1990!.
16
W. Maus-Friedrichs, S. Dieckhoff, and V. Kempter, Surf. Sci. 249, 149
~1991!.
17
Low Energy Electrons and Surface Chemistry, 2nd ed., edited by G. Ertl
and J. Küppers ~VCH, Weinheim, 1985!, p. 102.
18
J. Günster, G. Liu, V. Kempter, and D. W Goodman, J. Vac. Sci. Technol. A 16, 996 ~1998!.
19
D. Ochs, W. Maus-Friedrichs, M. Brause, J. Günster, V. Kempter, V.
Puchin, A. Shluger, and L. Kantorovich, Surf. Sci. 365, 557 ~1996!.
20
D. Ochs, M. Brause, W. Maus-Friedrichs, and V. Kempter, J. Electron
Spectrosc. Relat. Phenom. 88-91, 757 ~1998!.
21
M. Brause, D. Ochs, J. Günster, Th. Mayer, B. Braun, V. Puchin, W.
Maus-Friedrichs, and V. Kempter, Surf. Sci. 383, 216 ~1997!.
22
H. H. Huang, X. Jiang, Z. Zou, and G. Q. Xu, Surf. Sci. 376, 245 ~1997!.
23
D. A. King, Surf. Sci. 47, 384 ~1975!.
24
Handbook of Chemistry and Physics, 67th ed., edited by R. C. Weast
~Chemical Rubber, Boca Raton, 1986!.
25
K.-M. Schröder, D. Peterka, and H. Pfnür, DPG Frühjahrstagung Regensburg ~to be published!.
26
A. M. Ferrari and G. Pacchioni, J. Phys. Chem. 100, 9032 ~1996!.
27
L. N. Kantorovich, J. M. Holender, and K. Gillan, Surf. Sci. 343, 221
~1995!.
28
F. Illas and G. Pacchioni, J. Chem. Phys. 108, 7835 ~1998!.
29
V. Musolino, A. Selloni, and R. Car, Surf. Sci. 402-4, 413 ~1998!.