Recommended values of clean metal surface work functions

Gregory N. Derry, Megan E. Kern, and Eli H. Worth

Citation: J. Vac. Sci. Technol. A 33, 060801 (2015); doi: 10.1116/1.4934685

View online: http://dx.doi.org/10.1116/1.4934685
View Table of Contents: http://avs.scitation.org/toc/jva/33/6
Published by the American Vacuum Society
REVIEW ARTICLE

Recommended values of clean metal surface work functions

Gregory N. Derry,a) Megan E. Kern, and Eli H. Worth
Department of Physics, Loyola University Maryland, 4501 N. Charles St., Baltimore, Maryland 21210
(Received 13 August 2015; accepted 14 October 2015; published 30 October 2015)
A critical review of the experimental literature for measurements of the work functions of clean
metal surfaces of single-crystals is presented. The tables presented include all results found for
low-index crystal faces except cases that were known to be contaminated surfaces. These results
are used to construct a recommended value of the work function for each surface examined, along
with an uncertainty estimate for that value. The uncertainties are based in part on the error distribu-
tion for all measured work functions in the literature, which is included here. The metals included
in this review are silver (Ag), aluminum (Al), gold (Au), copper (Cu), iron (Fe), iridium (Ir), mo-
lybdenum (Mo), niobium (Nb), nickel (Ni), palladium (Pd), platinum (Pt), rhodium (Rh), ruthenium
(Ru), tantalum (Ta), and tungsten (W). VC 2015 American Vacuum Society.

[http://dx.doi.org/10.1116/1.4934685]

I. INTRODUCTION In order to address these needs, the present work includes

The work function of a solid has long been considered a critical review of all existing values found by the authors
one of the most fundamental surface properties of a mate- for work functions of clean, low-index faces of single-crystal
rial.1–8 Defined essentially as the amount of energy required metal surfaces. Polycrystalline samples and surfaces meas-
to remove an electron from the solid, the work function ured under poor vacuum conditions are not included in these
depends on the electronic structure of the bulk as well as the tables. In addition to the complete tables of all reasonably
charge density relaxation due to the presence of the surface. reliable literature values, we also include a separate table of
In addition to this fundamental significance, the work func- recommended values for each surface, based on the aggre-
tion is also a crucial parameter in a number of technologi- gated results from all of the individual measurements. These
cally important processes such as the behavior of the metal/ recommended values also include uncertainty estimates that
semiconductor interface, the electrode/electrolyte interac- have been made as realistic as possible given the limitations
tion, and the formation of electron beams. of the information available. Protocols and assumptions used
Despite the acknowledged importance of the work func- to calculate the recommended work functions and their
tion, both fundamental and practical, presently existing com- uncertainties are detailed below.
pendia of measured values for the work functions of clean,
low-index surfaces of single-crystal pure metals are defi- II. METHODS
cient. Several influential tables were published decades The precision and reliability of a work function measure-
ago,3–6 and these tables are still the major sources of infor- ment depend on both the experimental technique employed
mation in standard reference works.9 A more recent review to make the measurement and the state of the sample surface.
of the work function literature by Kawano8 has made an The state of the surface used to make the measurements, in
extremely valuable contribution to improving the situation turn, depends on the preparation and cleaning methods,
by updating the published measurements, but this does not while knowledge of this state requires the characterization of
entirely solve the problem, because no critical review of the the surface. Typical characterization methods include a
reported values is included, and the tables that are provided means of assessing the cleanliness (such as Auger spectros-
mix reliable measurements with unreliable measurements copy) and a means of observing the structural order (such as
and also with values that are not measured at all but rather low energy electron diffraction). The measurement techni-
computed from models. As a result of the present situation,
ques that have been used to measure work functions all have
investigators who require a value for the work function of a
potential weaknesses that must be addressed in order for the
particular surface sometimes resort to literature values of
techniques to produce valid results, and some of these tech-
poor accuracy that do not offer reliable information with
niques are inherently less precise and/or reliable than others.
which to vet theoretical calculations or provide quality input
In the present review, work function measurements for
for practical model systems of technological importance.
which there is evidence that the surfaces were contaminated
or for which there is evidence that the measurement tech-
a)
Author to whom correspondence should be addressed; electronic mail: nique used is invalid are not included in the tables provided.
[email protected] Of those measurements that are included in the tables, some

060801-1 J. Vac. Sci. Technol. A 33(6), Nov/Dec 2015 0734-2101/2015/33(6)/060801/9/$30.00 C 2015 American Vacuum Society
V 060801-1
060801-2 Derry, Kern, and Worth: Recommended values of clean metal surface work functions 060801-2

are of extremely high quality (using techniques that are well

documented to be valid and precise, with good descriptions
of the precautions taken to ensure that necessary conditions
were met, and also including a full accounting of the clean-
ing procedures and characterization methodology employed)
while others are somewhat less so (using basically valid
experimental techniques but not documenting special care in
making certain that all necessary conditions have been met
to ensure high precision, or using incomplete characteriza-
tion methods). However, in order to avoid the possibility of
subjective judgments affecting the results, all of these meas-
urements are treated on an equal footing and the recom-
mended values are found by taking simple unweighted
averages of all the individual measurements deemed reliable
enough to include. In addition to the recommend values, all
of the original work is referenced here so that investigators
who wish to examine it may exercise that option.
Estimating the uncertainties of these final average results
entails several difficulties. The number of available work FIG. 1. Histogram of the relative error distribution of all measured work
function measurements varies greatly from one surface to functions of surfaces for which at least four measurements exist. The solid
another. In cases where the number of measured values is curve is a best-fit Lorentzian function for this distribution; the broken curve
large, the scatter in these values is frequently quite large, is a best-fit Gaussian function.
despite the small uncertainties reported by the investigators.
fit curves for a Gaussian function and a Lorentzian function.
One of the underlying causes of the difficulty in estimating
The Gaussian distribution is clearly a poor fit by inspection.
uncertainties is the empirical fact that the uncertainties
In addition to further justifying the conclusion, as stated
reported in the literature must surely be unrealistic, given the
above, that the errors are non-normally distributed, these
fact that the spread in the measured values (in cases where a
large number of such values exist) are routinely about an results suggest a methodology for estimating the uncertainties
order of magnitude greater than the uncertainties reported in the recommended values for cases where only a small num-
for each individual value. Three important implications fol- ber of measurements exist. In these cases, we can still con-
low from this fact: First, the uncertainties reported in the lit- struct reasonable uncertainty estimates by assuming that the
erature, although they are included for completeness in the distribution of errors for any particular surface is well
tables presented here, are not really useful in making realis- approximated by the distribution for the entire literature.
tic uncertainty estimates of the recommended values. Although the Lorentzian curve fits the scatter better than
Second, the real errors do not appear to be normally distrib- the Gaussian curve, we do not suggest that the scatter
uted, given the contradiction between the real scatter and the actually follows a Cauchy distribution.10 The data are insuf-
nominal error estimates, and thus many of the standard ficient to determine the actual distribution of the errors.
methods of estimating uncertainty are not valid in this However, since the errors are reasonably approximated by a
instance. Thirdly, many surfaces have only one or two meas- Lorentzian over the limited range we are interested in, we
ured values of the work function available in the literature, can use it for numerical computations when needed. In prac-
and for these cases it becomes especially problematic to tice, this turns out to be the case for surfaces where only one
know how reliable the final result is and to formulate an or two measurements exist, because in these cases we have
appropriate uncertainty estimate. no information whatsoever on the scatter, so we need to
To address these problems, we have developed an approxi- depend entirely on the properties of the entire distribution.
mation to the actual error distribution of the work function For cases where at least three reliable measurements exist,
values found in the literature. Although no individual surface we can use the information represented by the scatter in
offers enough measurements to ascertain this error distribu- these empirical values instead of the properties of the distri-
tion, the data from all extant measurements can be aggregated. bution. Combining the results for these two regimes (N  2
We do this by computing the differences between the meas- and N > 2), we are able to find uncertainty estimates for all
ured values and the average values for all of the surfaces hav- of the surfaces for which work function measurements exist.
ing at least four useable measurements. These differences are Although we do not know the probability distribution of
then normalized by dividing them by the average value itself the errors, we are still able to estimate the uncertainties by
for that surface, resulting in a relative difference. Since these using Chebyshev’s inequality, which is independent of this
relative differences, referenced to the average value for each distribution. For some confidence probability a, we can use
surface, are all comparable, they can then be aggregated to the expression11
obtain a distribution for the entire data set. This procedure  
generates enough information to yield meaningful statistics. A a ¼ P j/   /j  k prffiffiffiffi  1  4 1 ; (1)
histogram of these results is shown in Fig. 1, along with best- N 9 k2

J. Vac. Sci. Technol. A, Vol. 33, No. 6, Nov/Dec 2015

060801-3 Derry, Kern, and Worth: Recommended values of clean metal surface work functions 060801-3

where r is the standard deviation of a set of work function The identification in Table II of the experimental
data for some surface and N is the number of measured val- techniques employed to make the work function measure-
ues for that surface. The choice of the confidence probability ments have been coded by letters using the following
sets the value of the parameter k, and hence the confidence correspondences:
interval of the uncertainty. In the present work, we have cho-
sen a ¼ 0.8, yielding k ¼ 1.49. This method was used for all Technique Letter code
surfaces with N > 2, the p uncertainties
ffiffiffiffi of the recommended Photoemission valence band width A
values quoted being kr= N for these cases with an 80% threshold photoemission B
confidence interval. For the cases in which N  2, computa- field emission C
tion of the standard deviation would have no meaning, so for thermal emission D
these cases we use the properties of the entire error distribu- target current spectroscopy E
tion as described above. The 80% confidence interval for ionization desorption F
this distribution is found to be at D/=/  ffi 0:06 based on contact potential difference, diode method G
field emission retarding potential H
integration of the best-fit Lorentzian curve. Thus, the
pffiffiffiuncer-
ffi contact potential difference, Kelvin probe I
tainty for these work functions is taken to be 0:06= N mul- retarding field, secondary emission J
tiplied by the recommended value. Though not completely private communication, no detail provided K
rigorous, this methodology should be robust in a practical scanning tunneling microscopy L
sense and result in realistic error estimates. metastable induced electron spectroscopy M
This analysis does not identify the various sources of
error involved nor does it distinguish between ransom sour-
ces of error and systematic errors in the measurements, The vast majority of the measurements were made using
which will vary in different experimental methods. In addi- the first four techniques listed, each of which has advantages
tion to the sources of error inherent in the different method- and disadvantages.3 Thermionic emission introduces system-
ologies, the state of the sample surface is itself a source of atic differences due to the temperature dependence of the
error. Defects in the surface crystal structure and residual work function, and both thermionic emission and threshold
contamination below the detection level both can affect the photoemission are susceptible to systematic errors due to the
work function, resulting in errors additional to the random Schottky effect (influence of electric fields used to bias the
experimental errors. sample). Both of these methods are also unreliable if patch
effects are present, but this should not be problematic in any
III. RESULTS of the work reported here since we have limited the focus to
The main results of this review are presented in Tables I single crystals. Most of the work employing field emission is
and II. Because many investigators will be primarily inter- unable to directly relate the needed field strength to the meas-
ested in the recommended value of the work function of some ured voltage, and thus must depend on knowledge of the aver-
particular surface, along with its estimated uncertainty, these age tip work function to measure the work function of
recommended values are summarized first in Table I. The individual crystal faces. Also, it is difficult to explicitly mea-
number of measurements N that the recommended values are sure the surface composition to demonstrate freedom from
based on is not given in Table I because it is easily found by contamination. Photoemission valence band width measure-
consulting Table II, which is a detailed listing of all the litera- ments are difficult to do with high precision because the sec-
ture results that are used to construct the recommended values ondary cut-off (and frequently also the Fermi level) are not
and their uncertainties given in Table I, computed using the extremely sharp due to temperature effects and spectrometer
methodology outlined above. If an investigator is mainly resolution, and further systematic errors are possible if the ge-
interested in knowing an estimate of the numerical value of ometry is not optimized properly. Of the remaining techniques
the work function for some surface based on existing meas- on the list, only the contact potential difference methods have
urements (and an estimate of the uncertainty of this value), been used extensively. The major disadvantage of these meth-
then Table I is sufficient for that purpose. ods is that they only provide relative information that must be
More extensive information is found in Table II. Work compared to some standard surface with a known work func-
function results for every paper that was used in the construc- tion, and thus the results can be no better than the accuracy
tion of Table I are included in Table II, organized by metal. with which this standard work function has been measured.
Also included are the reference numbers to the original While the information provided in Table II is self-
papers,12–171 the nominal uncertainty of the result if it was explanatory and sufficient for many purposes, we also provide
reported, and the experimental technique used to make the here a brief narrative description of the results in order to con-
measurement. Three measurements12,42,118 were excluded textualize this information and summarize the experimental
from the averages due to combinations of insufficient charac- situation. Silver and copper are both comparatively well
terization, measurement technique issues, and outlier status. measured, with several reliable values (found using different
In Table II, the nominal uncertainties are only provided for methods) that are in good agreement with each other.
completeness, as it is clear that the real uncertainties of the However, in both of these cases, the number and consistency
measurements are much larger in many of these cases. of the values are not as good for the (110) surface as they are

JVST A - Vacuum, Surfaces, and Films

060801-4 Derry, Kern, and Worth: Recommended values of clean metal surface work functions 060801-4

TABLE I. Recommended values of the work functions of clean, low-index, single-crystal metal surfaces and their estimated uncertainties, based on the experi-
mental literature. Uncertainties found using Fig. 1 results are indicated by an asterisk *.

Surface Work function (eV) Uncertainty (eV) Surface Work function (eV) Uncertainty (eV)

Ag(100) 4.36 0.05 Ni(100) 5.17 0.11

Ag(111) 4.53 0.07 Ni(111) 5.24 0.07
Ag(110) 4.10 0.15 Ni(110) 4.72 0.13
Al(100) 4.31 0.18* Pd(100) 5.48 0.23*
Al(111) 4.32 0.06 Pd(111) 5.67 0.12
Al(110) 4.23 0.13 Pd(110) 5.07 0.20
Au(100) 5.22 0.31* Pt(100)[1X1] 5.75 0.13
Au(111) 5.33 0.06 Pt(100)[5X1] 5.67 0.14
Au(110) 5.16 0.22* Pt(111) 5.91 0.08
Cu(100) 4.73 0.10 Pt(110) 5.53 0.13
Cu(111) 4.90 0.02 Rh(100) 5.30 0.15
Cu(110) 4.56 0.10 Rh(111) 5.46 0.09
Fe(100) 4.60 0.33 Rh(110) 4.86 0.21*
Fe(111) 4.81 0.29* Ru(100) 5.40 0.11
Fe(110) 5.07 0.04 Ru(1010) 4.60 0.28*
Ir(100)[1X1] 5.97 0.23 Ta(100) 4.10 0.25*
Ir(100)[5X1] 5.95 0.25* Ta(111) 3.50 0.21*
Ir(111) 5.78 0.04 Ta(110) 4.74 0.09
Ir(110) 5.42 0.32* W(100) 4.70 0.06
Mo(100) 4.46 0.11 W(111) 4.44 0.03
Mo(111) 4.37 0.24 W(110) 5.44 0.14
Mo(110) 4.92 0.05 W(211) 4.84 0.07
Nb(100) 4.08 0.17*
Nb(111) 4.37 0.19*
Nb(110) 4.63 0.17

for the other low index faces. Two other metals for which a For the other low-index faces of these metals, only a small
large total number of reliable results exist are nickel and plati- number of work function measurements exist, with variable
num, but the majority of these are only for the (111) surface reliability and consistency, yielding larger uncertainties than
in both of these cases. The other low index faces have fewer the (111) surfaces. Two other metals that have N > 3 reliable
reliable values, though the uncertainties are still reasonable in work function measurements for a given crystal face are palla-
several cases due to the high degree of consistency among the dium and ruthenium. These two cases are a little different
measurements. Tungsten is an interesting special case, from those already discussed, because for the surfaces of pal-
because a very large number of work function measurements ladium and ruthenium that have these larger number of meas-
were made prior to the existence of techniques to determine urements [Pd(111) and Ru(100)], the scatter in the
surface cleanliness. Thus, the work functions of single-crystal measurements is considerable, so their uncertainties are larger
tungsten surfaces are not necessarily as well measured as the than the comparable surfaces of Rh, Ir, Al, and Fe. All the
very large number of measurements suggests. W(110), in par- other faces of Pd and Ru have fewer measurements and corre-
ticular, has a large uncertainty due to the scatter in the reliable spondingly greater uncertainties, similar to the other surfaces
values. In contrast, the work function of aluminum has been with N  3 good values. The remaining refractory transition
measured fewer times. Only Al(111) has multiple reliable metals (niobium, molybdenum, and tantalum) have compara-
measurements with a low degree of scatter, thus yielding a tively reliable work function values only for the (110) surfaces
small uncertainty. The other two low-index faces of aluminum of Mo and Ta, for which larger numbers of measurements
have fewer measurements, resulting in a higher uncertainty; exist. Of the rest of the low-index surfaces in this set of met-
this may be an overestimate of the error, but further measure- als, there are few measurements and/or a rather large amount
ments would be needed to confirm this. A similar situation of scatter. Finally, the work functions of the low-index surfa-
occurs with iron. A series of papers published in Japan pro- ces of gold are surprisingly poorly known, given the amount
vided reliable measurements of the work function for all of and quality of the data for the other group IB metals. The
the low-index faces of iron, but only Fe(110) was subse- Au(111) surface does have a good work function value,
quently measured by several more investigators, yielding four though, since the comparatively few measurements have only
reliable work function values with a low degree of scatter. a small degree of scatter.
Two other metals where this situation exists are rhodium and Although the primary purpose here is to provide a reliable
iridium. Both the Rh(111) and Ir(111) surfaces have multiple set of work function values for particular low-index surface
work function measurements, and in both of these cases, the structures, it is also of some interest to look at chemical
reliable values are all consistent, yielding a low uncertainty. trends of work functions across the Periodic Table, similar to

J. Vac. Sci. Technol. A, Vol. 33, No. 6, Nov/Dec 2015

060801-5 Derry, Kern, and Worth: Recommended values of clean metal surface work functions 060801-5

TABLE II. Published values of work function measurements for low-index TABLE II. (Continued.)
surfaces of single-crystal metals, including references and experimental
measurement techniques (see text for technique codes). Work Uncertainty
Surface function (eV) (eV) Technique References
Work Uncertainty
Surface function (eV) (eV) Technique References Cu(111) 4.8 0.3 L 45
Cu(111) 4.88 — A 46
Ag, silver Cu(111) 4.88 0.05 A 47
Ag(100) 4.81 0.01 B 12 Cu(111) 4.93 — A 48
Ag(100) 4.22 0.04 B 13 Cu(111) 4.94 — A 49
Ag(100) 4.3 0.1 A 14 Cu(111) 4.946 0.01 A 50
Ag(100) 4.4 — K 15 Cu(110) 4.4 — J 37
Ag(100) 4.42 0.02 A 16 Cu(110) 4.48 0.03 B 38
Ag(100) 4.34 0.03 B 17 Cu(110) 4.92 0.019 I 42
Ag(100) 4.43 0.01 A 18 Cu(110) 4.5 — A 51
Ag(111) 4.75 0.01 B 12 Cu(110) 4.5 — A 52
Ag(111) 4.49 0.02 A 19 Cu(110) 4.59 — I 53
Ag(111) 4.56 — A 20 Cu(110) 4.87 — I 23
Ag(111) 4.46 0.02 B 13
Ag(111) 4.56 — A 21 Fe, iron
Ag(111) 4.45 0.03 B 17 Fe(100) 4.17 0.03 I 54
Ag(111) 4.50 0.1 A 22 Fe(100) 4.88 0.07 B 55
Ag(111) 4.69 — I 23 Fe(100) 4.75 0.03 B 56
Ag(110) 3.9 — A 24 Fe(111) 4.81 0.02 B 57
Ag(110) 4.14 0.04 B 13 Fe(110) 5.12 — B 58
Ag(110) 4.25 0.03 B 17 Fe(110) 5.0 — A 59
Fe(110) 5.05 — A 60
Al, aluminum Fe(110) 5.12 0.06 A 61
Al(100) 4.20 0.03 B 25
Ir, iridium
Al(100) 4.41 0.03 B 26
Ir(100)[1X1] 6.15 0.1 A 62
Al(111) 4.26 0.03 B 25
Ir(100)[1X1] 6.1 0.1 E 63
Al(111) 4.24 0.03 B 26
Ir(100)[1X1] 5.67 — C 64
Al(111) 4.25 0.05 A 27
Ir(100)[5X1] 6.0 0.1 A 62
Al(111) 4.29 0.02 B 28
Ir(100)[5X1] 5.9 0.15 E 63
Al(111) 4.3 0.1 A 29
Ir(111) 5.79 — C 64
Al(111) 4.4 — A 30
Ir(111) 5.8 — D 65
Al(111) 4.53 — A 31
Ir(111) 5.85 — G 66
Al(110) 4.06 0.03 B 25
Ir(111) 5.7 0.02 F 67
Al(110) 4.28 0.02 B 26
Ir(111) 5.76 0.05 H 43
Al(110) 4.35 0.05 A 27
Ir(110) 5.42 0.05 H 43
Au, gold
Mo, molybdenum
Au(100) 5.22 0.04 A 32
Mo(100) 4.45 — F 68
Au(111) 5.26 0.04 A 32
Mo(100) 4.25 — C 69
Au(111) 5.30 0.05 A 33
Mo(100) 4.53 0.02 B 70
Au(111) 5.3–5.6 0.1 B 34
Mo(100) 4.6 — A 71
Au(111) 5.3 — A 35
Mo(111) 4.05 — C 69
Au(110) 5.20 0.04 A 32
Mo(111) 4.55 0.02 B 70
Au(110) 5.12 0.07 I 33
Mo(111) 4.52 0.02 B 72
Cu, copper Mo(110) 4.80 — C 69
Cu(100) 4.76 0.05 A 36 Mo(110) 4.95 0.02 B 70
Cu(100) 4.58 — J 37 Mo(110) 4.95 — A 73
Cu(100) 4.59 0.03 B 38 Mo(110) 5.0 — D 74
Cu(100) 4.63 0.01 A 39 Mo(110) 4.92 0.02 I 75
Cu(100) 4.65 — A 40
Cu(100) 4.77 0.05 A 41 Nb, niobium
Cu(100) 5.155 0.054 I 42 Nb(100) 3.97 0.02 B 72
Cu(100) 5.10 0.05 H 43 Nb(100) 4.18 0.02 H 43
Cu(111) 4.85 — B 44 Nb(111) 4.08 0.02 B 72
Cu(111) 4.95 — J 37 Nb(111) 4.66 — B 76
Cu(111) 4.94 0.03 B 38 Nb(110) 4.51 0.02 B 72
Cu(111) 4.90 — A 40 Nb(110) 4.37 0.01 I 75
Cu(111) 5.54 0.012 I 42 Nb(110) 4.75 0.08 D 77
Cu(111) 4.88 — I 23 Nb(110) 4.87 — A 78

JVST A - Vacuum, Surfaces, and Films

060801-6 Derry, Kern, and Worth: Recommended values of clean metal surface work functions 060801-6

TABLE II. (Continued.) TABLE II. (Continued.)

Work Uncertainty Work Uncertainty

Surface function (eV) (eV) Technique References Surface function (eV) (eV) Technique References

Ni, nickel Pt(111) 5.91 0.07 F 67

Ni(100) 5.22 — B 79 Pt(111) 6.10 0.06 B 113
Ni(100) 5.17 0.05 B 80 Pt(111) 5.95 0.05 C 114
Ni(100) 4.89 0.03 B 81 Pt(111) 6.4 0.1 A 62
Ni(100) 4.95 0.05 G 82 Pt(110) 5.4 — C 101
Ni(100) 5.1 0.1 G 83 Pt(110) 5.7 0.1 A 115
Ni(100) 5.22 0.04 B 84 Pt(110) 5.49 0.05 C 114
Ni(100) 5.53 0.05 H 43
Rh, rhodium
Ni(100) 5.3 0.1 A 85
Rh(100) 5.11 — A 116
Ni(111) 5.0 0.2 M 86
Rh(100) 5.38 0.05 C 117
Ni(111) 5.42 0.04 B 87
Rh(100) 4.6 0.1 K 118
Ni(111) 5.15 0.1 A 88
Rh(100) 5.42 0.01 C 119
Ni(111) 5.3 — A 89
Rh(111) 5.6 0.04 A 120
Ni(111) 5.25 — A 46
Rh(111) 5.4 — A 121
Ni(111) 5.25 — A 61
Rh(111) 5.59 0.01 C 119
Ni(111) 5.36 — A 49
Rh(111) 5.3 — A 122
Ni(111) 5.2 0.07 B 90
Rh(111) 5.4 0.1 A 123
Ni(111) 4.9 0.1 A 29
Rh(110) 4.80 0.05 C 117
Ni(111) 5.3 — A 91
Rh(110) 4.91 0.01 C 119
Ni(111) 5.36 — A 92
Ni(111) 5.35 0.05 B 84 Ru, ruthenium
Ni(110) 4.7 0.1 B 93 Ru(001) 5.27 0.08 A 124
Ni(110) 4.7 0.1 G 83 Ru(001) 5.4 0.1 A 125
Ni(110) 5.04 0.02 B 84 Ru(001) 5.52 — A 126
Ni(110) 4.55 — A 94 Ru(001) 5.50 0.1 A 62
Ni(110) 4.62 0.05 B 95 Ru(10-10) 4.6 0.1 A 103

Pd, palladium Ta, tantalum

Pd(100) 5.65 0.1 A 96 Ta(100) 4.1 — C 69
Pd(100) 5.3 — K 97 Ta(111) 3.5 — C 69
Pd(111) 5.6 — A 52 Ta(110) 4.47 0.02 I 75
Pd(111) 5.44 0.03 A 98 Ta(110) 4.73 — D 127
Pd(111) 5.6 — A 89 Ta(110) 4.74 0.02 B 128
Pd(111) 5.90 0.1 A 62 Ta(110) 4.75 0.06 D 129
Pd(111) 5.95 0.1 A 96 Ta(110) 4.94 0.05 B 130
Pd(111) 5.55 — A? 99 Ta(110) 4.8 — A 131
Pd(110) 5.2 — A 52 W, tungsten
Pd(110) 4.8 — A 100 W(100) 4.44 — D 132
Pd(110) 5.2 0.1 A 96 W(100) 4.52 — D 133
Pt, platinum W(100) 4.53 — D 134
Pt(100)[1X1] 5.5 0.2 E 63 W(100) 5.22 — D 135
Pt(100)[1X1] 5.8 — C 64 W(100) 4.55 — D 136
Pt(100)[1X1] 5.81 — C 101 W(100) 4.57 0.14 C 137
Pt(100)[1X1] 5.9 — E 102 W(100) 4.59 0.02 D 138
Pt(100)[5X1] 5.4 0.25 E 63 W(100) 4.6 0.1 D 139
Pt(100)[5X1] 5.7 0.1 A 103 W(100) 4.6 0.05 D 140
Pt(100)[5X1] 5.75 0.05 A 104 W(100) 4.63 — C 141
Pt(100)[5X1] 5.82 0.15 A 105 W(100) 4.63 0.02 H 43
Pt(111) 5.99 — C 101 W(100) 4.67 — C 142
Pt(111) 6.08 0.15 A 105 W(100) 4.65 — I 143
Pt(111) 5.6 0.1 A 106 W(100) 4.65 — C 144
Pt(111) 5.7 — A 89 W(100) 4.65 — C 145
Pt(111) 5.7 0.2 A 107 W(100) 4.65 — I 147
Pt(111) 5.84 0.05 A 108 W(100) 4.65 — ? 148
Pt(111) 5.8 — A 109 W(100) 4.7 — I 149
Pt(111) 5.85 0.05 A 110 W(100) 4.7 — C 150
Pt(111) 5.9 — A 111 W(100) 4.8 — C 152
Pt(111) 5.95 0.1 A 112 W(100) 4.8 0.05 C 153

J. Vac. Sci. Technol. A, Vol. 33, No. 6, Nov/Dec 2015

060801-7 Derry, Kern, and Worth: Recommended values of clean metal surface work functions 060801-7

TABLE II. (Continued.) TABLE II. (Continued.)

Work Uncertainty Work Uncertainty

Surface function (eV) (eV) Technique References Surface function (eV) (eV) Technique References

W(100) 4.82 — C 154 W(211) 4.66 — D 134

W(100) 4.9 — C 156 W(211) 4.89 0.15 C 137
W(100) 5.2 — C 157 W(211) 4.7 0.1 D 139
W(111) 4.3 — D 132 W(211) 4.71 0.05 D 140
W(111) 4.38 — D 133 W(211) 4.76 — C 144
W(111) 4.36 — D 134 W(211) 4.87 0.05 C 150
W(111) 4.54 — D 135 W(211) 4.9 — C 152
W(111) 4.4 0.1 D 139 W(211) 4.75 0.05 C 153
W(111) 4.47 0.02 H 43 W(211) 4.85 — C 154
W(111) 4.47 — C 144 W(211) 5.03 — C 155
W(111) 4.4 — C 145 W(211) 5.05 0.05 C 156
W(111) 4.4 — I 149 W(211) 5.25 0.05 F 158
W(111) 4.43 — C 150 W(211) 4.8 — C 170
W(111) 4.3 — C 152
W(111) 4.35 — C 153
W(111) 4.41 — C 154
W(111) 4.68 — C 155 the work done by Michaelson5 based on earlier data. To do
W(111) 4.78 — C 155 this, we suppress the structural information by averaging the
W(111) 4.60 0.05 C 156 work functions of all the crystal faces in order to obtain a
W(111) 4.4 — C 157 single materials parameter characteristic of that element.
W(111) 4.3 — C 160 These averaged work functions are plotted versus the atomic
W(111) 4.5 — C 162 number Z of each element in Fig. 2, showing the trends that
W(111) 4.32 — B&D 163 emerge along with the exceptions to these trends. For exam-
W(111) 4.45 — C 164 ple, there is a general tendency for the work functions to
W(111) 4.4 — C 165
increase with higher Z within a column of the Periodic
W(111) 4.43 — D 166
Table, but silver is clearly an exception to this rule. Note
W(111) 4.5 — G 169
W(111) 4.5 — C 170
that each different symbol corresponds to a different
W(110) 5.09 — D 132 Periodic Table column. Another interesting trend is the
W(110) 5.26 — D 133 increase in the work function with Z in each of the transition
W(110) 4.65 — D 134 metal rows, ending with a significant decrease when the
W(110) 4.54 — D 135 noble metal column is reached. This is seen mostly clearly in
W(110) 5.0 — D 139 the 4d metal row, where there is more data, but it can also be
W(110) 5.25 0.02 H 43 seen in the other rows.
W(110) 5.25 — C 144
W(110) 6.0 — C 145
W(110) 5.14 0.01 I 146
W(110) 5.11 0.02 I 147
W(110) 5.9 — I 149
W(110) 6.0 0.1 C 150
W(110) 5.96 — C 151
W(110) 5.8 — C 152
W(110) 5.07 — C 153
W(110) 5.85 — C 154
W(110) 5.83 — C 155
W(110) 5.76 — C 155
W(110) 5.8 — C 156
W(110) 5.25 0.05 F 158
W(110) >7.1 — C 159
W(110) 5.05 0.02 I 161
W(110) 5.15 0.01 I 75
W(110) 5.22 0.02 D 167
W(110) 5.25 0.02 A 168
W(110) 5.3 — G 169
W(110) 5.35 — C 170
W(110) 5.47 — C 171
W(211) 4.66 — D 132
W(211) 4.65 — D 133 FIG. 2. Work functions of the elements plotted against atomic number to
show periodic trends.

JVST A - Vacuum, Surfaces, and Films

060801-8 Derry, Kern, and Worth: Recommended values of clean metal surface work functions 060801-8

9
IV. DISCUSSION CRC Handbook of Chemistry and Physics, 92nd ed., edited by W. M.
Haynes and D. R. Lide (CRC, Boca Raton, FL, 2011), pp. 12–124.
As stated previously, we believe that the assumptions and 10
B. P. Roe, Probability and Statistics in Experimental Physics, 2nd ed.
protocols employed in this work result in reliable estimates 11
(Springer-Verlag, New York, 2001).
of the work functions, based on the present literature, and S. G. Rabinovitch, Measurement Errors and Uncertainties, 2nd ed.
(Springer-Verlag, New York, 2000), pp. 102–105.
that the uncertainties given here are realistic and practical. 12
H. E. Farnsworth and R. P. Winch, Phys. Rev. 58, 812 (1940).
However, it should be emphasized that alternative assump- 13
M. Chelvayohan and C. H. B. Mee, J. Phys. C 15, 2305 (1982).
14
tions and protocols are also possible, which would change G. Ferrini, C. Giannetti, G. Galimberti, S. Pagliara, D. Fausti, F. Banfi,
and F. Parmigiani, Phys. Rev. Lett. 92, 256802 (2004).
these numerical results to some degree. For this reason, 15
D. R. Hamann, L. F. Mattheiss, and H. S. Greenside, Phys. Rev. B 24,
investigators should consult the primary literature if explor- 6151 (1981).
16
ing such alternatives is desired. K. Giesen, F. Hage, F. J. Himpsel, H. J. Riess, and W. Steinmann, Phys.
A few issues deserve further mention. One potential issue Rev. B 35, 971 (1987).
17
W. Y. Li, K. Goto, and R. Shimizu, Surf. Interface Anal. 37, 244 (2005).
to note is that, despite some apparent differences in the qual- 18
S. Schuppler, N. Fischer, T. Fauster, and W. Steinmann, Appl. Phys. A
ity and reliability of experimental measurements in the liter- 51, 322 (1990).
19
ature, no distinction has been made amongst them here in K. Giesen, F. Hage, F. J. Himpsel, H. J. Riess, and W. Steinmann, Phys.
order to avoid bias. Another obvious issue is the distinction Rev. Lett. 55, 300 (1985).
20
S. Ryu, J. Chang, and S. K. Kim, J. Phys. Chem. 123, 114710 (2005).
between cases where many measurements exist and cases 21
K. Giesen, F. Hage, F. J. Himpsel, H. J. Riess, W. Steinmann, and N. V.
where only one or two measurements exist. In the former Smith, Phys. Rev. B 35, 975 (1987).
22
cases, a good empirical basis exists upon which to estimate H. Li, X. Wang, S. He, K. Ibrahim, H. Qian, R. Su, J. Zhong, M. I. Abba,
the uncertainty in the recommended value, but in the latter and C. Hong, Surf. Sci. 586, 65 (2005).
23
D. Staub and F. J. Himpsel, Phys. Rev. B 33, 2256 (1986).
cases, we do not have much information to base these esti- 24
H. Q. Quian et al., Appl. Surf. Sci. 253, 2336 (2006).
mates on. A particular issue that arises due to the methodol- 25
R. M. Eastman and C. H. B. Mee, J. Phys. F: Met. Phys. 3, 1738 (1973).
26
ogy used here stems from the decision to use the entire error J. K. Grepstad, P. O. Gartland, and B. J. Slagsvold, Surf. Sci. 57, 348
(1976).
distribution for cases in which N  2, instead of using the 27
A. J. Maxwell, P. A. Bruhwiler, D. Arvantis, J. Hasselstrom, M.
consistency in the measurements. This protocol has the Johansson, and N. Martensson, Phys. Rev. B 57, 7312 (1998).
28
effect of making the uncertainty independent of whether two A. G. Borisov, H. Winter, G. Dierkes, and R. Zimny, Europhys. Lett. 33,
work function measurements are consistent with each other 229 (1996).
29
K. Gurtler and K. Jacobi, Surf. Sci. 134, 309 (1983).
or widely divergent. This seems undesirable, but treating 30
S. A. Sardar, J. A. Syed, K. Tanaka, F. P. Netzer, and M. G. Ramsey,
two measurements as a statistical ensemble also seems in- Surf. Sci. 519, 218 (2002).
31
valid. The root of the problem lies in the lack of a sufficient T. Chiang, G. Kaindl, and D. E. Eastman, Solid State Commun. 36, 25
number of existing measurements (for this reason, our labo- (1980).
32
G. V. Hansson and S. A. Flodstrom, Phys. Rev. B 18, 1572 (1978).
ratory is presently engaged in a program of precise work 33
J. Lecoeur, J. P. Bellier, and C. Koehler, Electrochim. Acta 35, 1383
function measurements to alleviate this situation), and the (1990).
34
point here is that investigators utilizing the information pre- D. Pescia and F. Meier, Surf. Sci. 117, 302 (1982).
35
C. Tzeng, W. Lo, J. Yuh, R. Chu, and K. Tsuei, Phys. Rev. B 61, 2263
sented in Table I should be aware of such issues.
(2000).
Despite these limitations, the tables presented here should 36
S. Kim, J. Kim, J. Han, J. Seo, C. Lee, and S. Hong, Surf. Sci. 453, 47
prove useful to the condensed matter community. There are, (2000).
37
in fact, a number of surfaces for which many reliable work G. A. Haas and R. E. Thomas, J. Appl. Phys. 48, 86 (1977).
38
P. O. Gartland, S. Berge, and B. J. Slagsvold, Phys. Rev. Lett. 28, 738
function measurements exist and are highly consistent. For (1972).
these cases, comparison with theoretical values can be a val- 39
T. Wegehaupt, D. Rieger, and W. Steinmann, Phys. Rev. B 37, 10086
uable tool to improve understanding of electrons at surfaces (1988).
40
and to develop computational methods for advanced materi- J. E. Rowe and N. V. Smith, Phys. Rev. B 10, 3207 (1974).
41
G. Tibbetts, J. Burkstrand, and J. C. Tracy, Phys. Rev. B 15, 3652 (1977).
als design. Conversely, the tables also show where work 42
T. A. Delchar, Surf. Sci. 27, 11 (1971).
function information is presently lacking and might spur fur- 43
R. W. Strayer, W. Mackie, and L. W. Swanson, Surf. Sci. 34, 225 (1973).
44
ther experimental effort in that direction, or at least discour- 45
P. O. Gartland and B. J. Slagsvold, Phys. Rev. B 12, 4047 (1975).
age comparison of theoretical values with unreliable J. F. Jia, K. Inoue, Y. Hasegawa, W. S. Yang, and T. Sakurai, Phys. Rev. B
58, 1193 (1998).
experimental numbers. 46
K. Giesen, F. Hage, F. J. Himpsel, H. J. Riess, and W. Steinmann, Phys.
Rev. B 33, 5241 (1986).
1 47
A. L. Hughes and L. A. DuBridge, Photoelectric Phenomena, 1st ed. G. D. Kubiak, Surf. Sci. 201, L475 (1988).
48
(McGraw-Hill, New York, 1932). N. Fischer, S. Schuppler, T. Fauster, and W. Steinmann, Surf. Sci. 314,
2
C. Herring and M. H. Nichols, Rev. Mod. Phys. 21, 185 (1949). 89 (1994).
3 49
J. C. Riviere, “Work function: Measurements and results,” in Solid State K. Tsuei, J. Yuh, C. Tzeng, R. Chu, S. Chung, and K. Tsang, Phys. Rev. B
Surface Science, edited by M. Green (Dekker, New York, 1969), Vol. 1, 56, 15412 (1997).
50
pp. 179–289. K. Takeuchi, A. Suda, and S. Ushioda, Surf. Sci. 489, 100 (2001).
4 51
V. S. Fomenko, Emissive Properties of Materials, 3rd ed. (Naukova C. Mariani, K. Horn, and A. M. Bradshaw, Phys. Rev. B 25, 7798 (1982).
52
Dumka, Kiev, 1970). W. Sesselmann, B. Woratschek, J. Kuppers, G. Ertl, and H. Haberland,
5
H. B. Michaelson, J. Appl. Phys. 48, 4729 (1977). Phys. Rev. B 35, 1547 (1987).
6 53
J. Holzl and F. K. Schulte, “Work functions of metals,” in Solid Surface D. Y. Li and W. Li, Appl. Phys. Lett. 79, 4337 (2001).
54
Physics, edited by G. Hohler (Springer-Verlag, Berlin, 1979), pp. 1–150. R. E. Simon, Phys. Rev. 116, 613 (1959).
7 55
S. Halas, Mater. Sci.-Pol. 24, 951 (2006). H. Kobayashi and S. Kato, Surf. Sci. 12, 398 (1968).
8 56
H. Kawano, Prog. Surf. Sci. 83, 1 (2008). K. Ueda and R. Shimizu, Jpn. J. Appl. Phys. 12, 1869 (1973).

J. Vac. Sci. Technol. A, Vol. 33, No. 6, Nov/Dec 2015

060801-9 Derry, Kern, and Worth: Recommended values of clean metal surface work functions 060801-9

57 110
H. Kobayashi and S. Kato, Surf. Sci. 18, 341 (1969). W. Ranke and W. Weiss, Surf. Sci. 414, 236 (1998).
58 111
S. Kato and H. Kobayashi, Surf. Sci. 27, 625 (1971). M. Kiskinova, G. Pirug, and H. P. Bonzel, Surf. Sci. 133, 321 (1983).
59 112
G. Broden and H. P. Bonzel, Surf. Sci. 84, 106 (1979). D. Fargues, J. J. Ehrhardt, M. Abon, and J. C. Bertolini, Surf. Sci. 194,
60
G. Pirug, G. Broden, and H. P. Bonzel, Surf. Sci. 94, 323 (1980). 149 (1988).
61 113
R. Fischer, N. Fischer, S. Schuppler, T. Fauster, and F. P. Himpsel, Phys. G. N. Derry and Z. J. Zhong, Phys. Rev. B 39, 1940 (1989).
114
Rev. B 46, 9691 (1992). J. Marien, Bull. Soc. R. Sci. Liege 45, 103 (1976).
62 115
J. Hulse, J. Kuppers, K. Wandelt, and G. Ertl, Appl. Surf. Sci. 6, 453 M. Alnot and J. J. Ehrhardt, Surf. Sci. 287/288, 325 (1993).
116
(1980). D. E. Peebles, H. C. Peebles, and J. M. White, Surf. Sci. 136, 463 (1984).
63 117
R. Drube, J. Noffke, R. Schneider, J. Rogozik, and V. Dose, Phys. Rev. B R. Vanselow and X. Q. D. Li, Surf. Sci. Lett. 264, L200 (1992).
118
45, 4390 (1992). J. G. Gay, J. R. Smith, and F. Arlinghaus, Phys. Rev. B 25, 643 (1982).
64 119
B. E. Nieuwenhuys, R. Bouwman, and W. M. H. Sachtler, Thin Solid R. S. Polizzotti and G. Ehrlich, J. Phys. Chem. 71, 259 (1979).
120
Films 21, 51 (1974). T. Janssens, G. Castro, H. Busse, U. Schneider, and K. Wandelt, Surf.
65
N. R. Gall’, S. N. Mikhailov, E. V. Rut’kov, and A. Y. Tontegode, Surf. Sci. 269/270, 664 (1992).
121
Sci. 226, 381 (1990). L. Q. Jiang and B. E. Koel, Phys. Rev. Lett. 72, 140 (1994).
66 122
R. E. Thomas, J. Appl. Phys. 41, 5330 (1970). P. Brault, H. Range, J. P. Toennies, and C. Woll, Z. Phys. Chem. 198, 1
67
M. Kaack and D. Fick, Surf. Sci. 342, 111 (1995). (1997).
68 123
N. R. Gall’, E. V. Rut’kov, and A. Y. Tontegode, Tech. Phys. 47, 484 A. Goldman, G. Rosina, E. Bertel, and F. P. Netzer, Z. Phys. B 73, 479
(2002). (1989).
69 124
M. Blaszczyszyn, Surf. Sci. 59, 533 (1976). C. Bromberger, H. Jansch, and D. Fick, Surf. Sci. 506, 129 (2002).
70 125
S. Berge, P. O. Gartland, and B. J. Slagsvold, Surf. Sci. 43, 275 (1974). J. Weimer, E. Umbach, and D. Menzel, Surf. Sci. 159, 83 (1985).
71 126
P. Soukiassian, R. Riwan, Y. Borensztein, and J. Lecante, J. Phys. C 17, K. Markert and K. Wandelt, Surf. Sci. 159, 24 (1985).
127
1761 (1984). D. L. Fehrs and R. E. Stickney, Surf. Sci. 8, 267 (1967).
72 128
I. Buribayev and N. A. Nurmatov, J. Electron. Spectrosc. 68, 547 (1994). J. E. Boggio and H. E. Farnsworth, Surf. Sci. 1, 399 (1964).
73 129
L. Q. Jiang and M. Strongin, Phys. Rev. B 42, 3282 (1990). W. T. Norris, J. Appl. Phys. 35, 467 (1964).
74 130
J. Desplat, J. Appl. Phys. 54, 5494 (1983). D. L. Fehrs and R. E. Stickney, Surf. Sci. 24, 309 (1971).
75 131
B. J. Hopkins and M. Ibrahim, Vacuum 23, 135 (1973). M. W. Ruckman, B. Xia, and S. L. Qiu, Phys. Rev. B 48, 15457 (1993).
76 132
R. M. Oman and J. A. Dillon, Surf. Sci. 2, 227 (1964). A. R. Hutson, Phys. Rev. 98, 889 (1955).
77 133
J. Jupille, B. Bigeard, J. Fusy, and A. Cassuto, Surf. Sci. 84, 190 (1979). G. F. Smith, Phys. Rev. 94, 295 (1954).
78 134
M. El-Batanouny, D. R. Hamann, S. R. Chubb, and J. W. Davenport, M. H. Nichols, Phys. Rev. 57, 297 (1940).
135
Phys. Rev. B 27, 2575 (1983). L. van Someren, Surf. Sci. 20, 221 (1970).
79 136
H. E. Farnsworth and H. H. Madden, J. Appl. Phys. 32, 1933 (1961). A. E. Abey, J. Appl. Phys. 39, 120 (1968).
80 137
W. Eib and S. F. Alvarado, Phys. Rev. Lett. 37, 444 (1976). T. V. Vorburger, D. Penn, and E. W. Plummer, Surf. Sci. 48, 417 (1975).
81 138
C. Weiser, Surf. Sci. 20, 143 (1970). A. A. Brown, L. J. Neelands, and H. E. Farnsworth, J. Appl. Phys. 21, 1
82
C. A. Papageorgopoulos and J. M. Chen, Surf. Sci. 52, 40 (1975). (1950).
83 139
G. E. Becker and H. D. Hagstrum, Surf. Sci. 30, 505 (1972). F. M. Gardner, F. E. Girouard, W. L. Boeck, and E. A. Coomes, Surf.
84
B. G. Baker, B. B. Johnson, and G. L. C. Maire, Surf. Sci. 24, 572 Sci. 26, 605 (1971).
140
(1971). G. A. Haas and R. E. Thomas, J. Appl. Phys. 40, 3919 (1969).
85 141
K. Christmann and J. E. Demuth, Surf. Sci. 120, 291 (1982). R. L. Billington and T. N. Rhodin, Phys. Rev. Lett. 41, 1602 (1978).
86 142
B. Lescop, A. Galtayries, and G. Fanjoux, J. Phys. Chem. B 108, 13711 Y. Gong and H. Zeng, Surf. Sci. 246, 169 (1991).
143
(2004). B. J. Hopkins and G. R. Shah, Vacuum 22, 267 (1972).
87 144
T. A. Callcott and A. U. MacRae, Phys. Rev. 178, 966 (1969). Y. Yamamoto and T. Miyokawa, J. Vac. Sci. Technol. B 16, 2871
88
F. J. Himpsel, J. A. Knapp, and D. E. Eastman, Phys. Rev. B 19, 2872 (1998).
145
(1979). A. van Oostrom, J. Phys. Chem. 47, 761 (1967).
89 146
J. E. Demuth, Surf. Sci. 65, 369 (1977). T. J. Lee, B. H. Blott, and B. J. Hopkins, Appl. Phys. Lett. 11, 361
90
E. Kisker, W. Gudat, M. Compagna, E. Kuhlmann, H. Hopster, and I. D. (1967).
147
Moore, Phys. Rev. Lett. 43, 966 (1979). B. J. Hopkins and K. R. Pender, Surf. Sci. 5, 155 (1966).
91 148
P. Dolle, M. Tommasini, and J. Jupille, Surf. Sci. 211/212, 904 (1989). C. A. Papageorgopoulos and J. M. Chen, Surf. Sci. 39, 313 (1973).
92 149
J. A. Barnard, J. J. Ehehardt, H. Azzouzi, and M. Alnot, Surf. Sci. T. A. Delchar and G. Ehrlich, J. Phys. Chem. 42, 2686 (1965).
150
211/212, 740 (1989). L. D. Schmidt, J. Phys. Chem. 46, 3830 (1967).
93 151
W. Gudat, E. Kisker, E. Kuhlmann, and M. Campagna, Phys. Rev. B 22, R. D. Young and E. W. Muller, J. Appl. Phys. 33, 91 (1962).
152
3282 (1980). R. Blaszczyszyn, M. Blaszczyszyn, and R. Meclewski, Surf. Sci. 51, 396
94
K. Jacobi, Y. Hsu, and H. Rotermund, Surf. Sci. 114, 683 (1982). (1975).
95 153
R. Clauberg, W. Gudat, E. Kisker, and E. Kuhlmann, Z. Phys. B 43, 47 A. D. Adsool, R. Pande, R. B. Sharma, M. A. More, and D. S. Joag,
(1981). Appl. Surf. Sci. 87/88, 37 (1995).
96 154
K. Wandelt and J. E. Hulse, J. Chem. Phys. 80, 1340 (1984). L. Schmidt and R. Gomer, J. Phys. Chem. 45, 1605 (1966).
97 155
J. G. Gay, J. R. Smith, F. Arlinghaus, and T. Capehart, Phys. Rev. B 23, C. Lea and R. Gomer, J. Phys. Chem. 54, 3349 (1971).
156
1559 (1981). T. Engel and R. Gomer, J. Phys. Chem. 50, 2428 (1969).
98 157
R. Fischer, S. Schuppler, N. Fischer, Th. Fauster, and W. Steinmann, A. A. Holscher, J. Phys. Chem. 41, 579 (1964).
158
Phys. Rev. Lett. 70, 654 (1993). F. L. Hughes and H. Levinstein, Phys. Rev. 113, 1029 (1959).
99 159
G. D. Kubiak, J. Vac. Sci. Technol. A 5, 731 (1987). R. D. Young and H. E. Clark, Appl. Phys. Lett. 9, 265 (1966).
100 160
J. Kuppers, F. Nitschke, K. Wandelt, and G. Ertl, Surf. Sci. 87, 295 S. A. Shakirova and E. V. Serova, Surf. Sci. 422, 24 (1999).
161
(1979). B. J. Hopkins and K. R. Pender, Br. J. Appl. Phys. 17, 281 (1966).
101 162
B. E. Nieuwenhuys, Surf. Sci. 59, 430 (1976). S. A. Shakirova and E. V. Serova, Appl. Surf. Sci. 219, 363 (2003).
102 163
R. Drube, V. Dose, and A. Goldman, Surf. Sci. 197, 317 (1988). D. F. Stafford and A. H. Weber, J. Appl. Phys. 34, 2667 (1963).
103 164
H. P. Bonzel and T. E. Fischer, Surf. Sci. 51, 213 (1975). T. Biernat and R. Blaszczyszyn, Appl. Surf. Sci. 230, 81 (2004).
104 165
B. Pennemann, K. Oster, and K. Wandelt, Surf. Sci. 249, 35 (1991). A. van Oostrom, Phys. Lett. 4, 34 (1963).
105 166
M. Salmeron, S. Ferrer, M. Jazzar, and G. A. Somorjai, Phys. Rev. B 28, J. L. Coggins and R. E. Stickney, Surf. Sci. 11, 355 (1968).
167
6758 (1983). B. Krahl-Urban, E. A. Niekisch, and H. Wagner, Surf. Sci. 64, 52 (1977).
106 168
D. M. Collins and W. E. Spicer, Surf. Sci. 69, 114 (1977). U. Thomann, C. Reub, Th. Fauster, F. Passek, and M. Donath, Phys. Rev. B
107
G. B. Fisher, Chem. Phys. Lett. 79, 452 (1981). 61, 16163 (2000).
108 169
M. Alnot, J. J. Ehrhardt, and J. A. Barnard, Surf. Sci. 208, 285 (1989). T. M. Gardiner, H. M. Kramer, and E. Bauer, Surf. Sci. 112, 181 (1981).
109 170
J. Cousty, C. A. Papageorgopoulos, and R. Riwan, Surf. Sci. 223, 479 C. J. Todd and T. N. Rhodin, Surf. Sci. 42, 109 (1974).
171
(1989). R. S. Polizzotti and G. Ehrlich, Surf. Sci. 91, 24 (1980).

JVST A - Vacuum, Surfaces, and Films