Role of ethylene on surface oxidation of TiO2(110)
Y. Murata, V. Petrova, I. Petrov, C. V. Ciobanu, and S. Kodambaka
Citation: Appl. Phys. Lett. 101, 211601 (2012); doi: 10.1063/1.4767954
View online: http://dx.doi.org/10.1063/1.4767954
View Table of Contents: http://apl.aip.org/resource/1/APPLAB/v101/i21
Published by the AIP Publishing LLC.
Additional information on Appl. Phys. Lett.
Journal Homepage: http://apl.aip.org/
Journal Information: http://apl.aip.org/about/about_the_journal
Top downloads: http://apl.aip.org/features/most_downloaded
Information for Authors: http://apl.aip.org/authors
Downloaded 16 Jul 2013 to 138.67.128.86. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions
APPLIED PHYSICS LETTERS 101, 211601 (2012)
Role of ethylene on surface oxidation of TiO2(110)
Y. Murata,1 V. Petrova,2 I. Petrov,2 C. V. Ciobanu,3 and S. Kodambaka1,a)
1
Depatment of Materials Science Engineering, University of California Los Angeles, Los Angeles,
California 90095, USA
2
Frederick Seitz Materials Research Laboratory, University of Illinois, Urbana, Illinois 61801, USA
3
Department of Mechanical Engineering and Materials Science Program, Colorado School of Mines, Golden,
Colorado 80401, USA
(Received 16 October 2012; accepted 2 November 2012; published online 21 November 2012)
Using in situ high-temperature (700-1000 K) scanning tunneling microscopy (STM), we studied the
influence of ethylene on the surface dynamics of oxygen-deficient, rutile-structured TiO2(110). STM
images were acquired during annealing the sample as a function of time, oxygen and ethylene
pressures, and temperature. With increasing oxygen pressure and/or decreasing temperature,
TiO2(110) surface mass increased, consistent with previous results. Interestingly, annealing the
sample in ethylene with traces of oxygen also results in the growth of TiO2 at higher rates than those
observed during annealing in pure oxygen. Our results indicate that ethylene promotes oxidation of
C 2012 American Institute of Physics. [http://dx.doi.org/10.1063/1.4767954]
TiO2(110). V
Transition-metal (TM) oxides,1 notably titanium dioxide
(TiO2),2–4 have a wide variety of applications in photocatalysis, for splitting of water5,6 and photo-assisted purification of
volatile organic compounds (VOCs),7 as a support for metal
catalysts,8 in dye-sensitized solar cells,9 as electrodes in batteries,10 and even as pigments in cosmetics. Among all of
the above-mentioned applications, TiO2 is well-known for
its photo-assisted oxidation of VOCs into CO2 and H2O.
Catalytic performance of TiO2 is sensitive to surface
atomic structure, composition, and roughness. Surface defects
such as oxygen (O) vacancies can alter its electronic band
structure11–13 and are essential to promote adsorption and
reaction of gas molecules.14 As O atoms desorb from the surface of TiO2, Ti4þ reduces to Ti3þ. And, the Ti3þ site is energetically favorable for the adsorption of oxygen while Ti4þ is
the preferred adsorption site for ethylene and water.15 Therefore, a fundamental understanding of the factors controlling
the surface stoichiometry is critical for rational design of high
efficiency catalysts with superior lifetime performance.
Considerable research efforts, experimental as well as
theoretical,16–20 have investigated the structure of rutile
TiO2(110),3,21–24 energetically the most stable surface.
While majority studies focused on understanding the photochemical reaction kinetics, relatively little is known concerning the thermally activated catalytic reactions. Here, we
study rutile TiO2(110) surface compositional and structural
evolution at high temperatures in presence of a relatively
simple VOC, ethylene (C2H4).15,25–28
One of the distinct characteristics of TiO2(110) is that
the surface exhibits structural phase transition whose kinetics
depend on the temperature and oxygen pressure.29–36 For
example, annealing TiO2(110) in vacuum leads to an
O-deficient surface with a (2 1) surface reconstruction,
while annealing in oxygen results in stoichiometric (1 1)
surface. Therefore, the surface structure and reactivity
depend on the temperature and oxygen pressure.37–40
a)
Author to whom correspondence should be addressed. Electronic mail:
kodambaka@ucla.edu.
0003-6951/2012/101(21)/211601/5/$30.00
In this Letter, we report in situ high-temperature (700–
1000 K) scanning tunneling microscopy (STM) studies of the
effect of ethylene-oxygen gas mixtures on structural evolution
of rutile-structured TiO2x(110). STM images were acquired
as a function of time, O2 and C2H4 partial pressures, and temperature. With increasing O2 pressure and/or decreasing temperature, we observe an increase in surface area due to the
formation of TiO2 via supply of Ti atoms transported from the
bulk and O atoms from the gas phase. These results are consistent with earlier high-temperature STM observations.29
Interestingly, annealing the sample in presence of C2H4 with
traces of O2 also leads to an increase in surface mass but the
growth rates are significantly higher than those observed
during annealing in pure O2. This surprising phenomenon
suggests that ethylene promotes surface oxidation TiO2(110).
All our experiments were carried out on a 9 2 0.5
mm3 rutile-TiO2(110) single crystal. In order to facilitate resistive heating of the sample, 2000 -Å-thick Ta layer was sputter deposited on the backside of the crystal. The TiO2(110)
sample was then air-transferred to an ultra high vacuum
(UHV, base pressure 1 1010 Torr) multi-chamber STM
system equipped with facilities for low-energy electron diffraction (LEED), residual gas analysis (RGA), Arþ sputtering,
and dosing O2 and C2H4 gases. The substrate was degassed in
UHV at 150 C for 14 h, cleaned by repeated cycles of sputtering with 1 keV Arþ for 30 min and annealing in vacuum at
1100 K for 5 min, and annealed in O2 (pressure, pO2 ¼ 9.8
108 Torr) for 30 min. This procedure resulted in TiO2x
(110) composed of 1 1 terraces and (2 1)-reconstructed
stripe structures oriented along [001]. Figure 1(a) shows a representative STM image and a LEED pattern acquired from the
as-prepared sample. A higher magnification STM image of
flat terrace in Fig. 1(b) shows a periodic structure with
3.3 6.3 Å2 spacing, characteristic of 1 1 surface reconstruction. This is consistent with the in-plane spacing of
3.0 6.5 Å2 measured from the LEED data.
In situ variable-temperature STM images were obtained
in the constant current mode using commercially available
Pt-Ir tips while heating the sample. O2 and C2H4 gases were
101, 211601-1
C 2012 American Institute of Physics
V
Downloaded 16 Jul 2013 to 138.67.128.86. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions
211601-2
Murata et al.
FIG. 1. (a) Typical room-temperature STM image (1000 1000 Å2) of
TiO2(110) composed of atomically-smooth 1 1 terraces and rectangular
2 1 stripes (brighter contrast). Inset shows a LEED pattern of TiO2(110)
acquired using incident electron energy of 200 eV. (b) Magnified image
(40 40 Å2) of the 1 1 region with the unit cell highlighted by a black
rectangle.
introduced to the desired value using UHV leak valves. RGA
data indicated that the partial pressures of H2O and CO2
gases, commonly found in UHV systems, were below
109 Torr during all of our annealing experiments, irrespective of the ambient (the amount of CO could not be quantified because it has the same molecular weight as that of
C2H4 and N2). During annealing in C2H4, however, we find
trace amounts of O2. From RGA spectra, we measured
pO2 ;C2 H4 of 1.2 109 Torr and 3.2 109 Torr while dosing
C2H4 at pC2 H4 of 9.8 109 Torr and 9.8 108 Torr,
respectively. Substrate temperatures were measured using
optical pyrometry and are accurate to within 50 K. Typical
tunnelling currents of 0.1 nA to 0.5 nA and bias voltages of
Appl. Phys. Lett. 101, 211601 (2012)
2.0 V to þ2.0 V were used. Point-mode scanning tunneling
spectroscopy (STS) measurements (I vs. V data) were
obtained over a range of bias voltages VT between –1 V and
þ1 V. During the measurements, tip-sample separation was
held constant by interrupting the feedback loop. Scan sizes
(50–200 nm), scan rates (50-100 s/frame), and tunneling parameters were varied to check for tip-induced effects. We
observed no such effects in the results presented here. STM
images were processed using WSXM software.41
We now focus on the effect of annealing ambient on the
TiO2(110) surface structure. Figure 2 shows a series of representative in situ STM images obtained from the same region
while annealing sequentially the sample at T ¼ 1000 K in (a)
C2H4 (pC2 H4 ¼ 9.8 108 Torr with pO2 ;C2 H4 ¼ 3.2 109
Torr) (b) UHV (3.8 1010 Torr), and (c) O2 (pO2 ¼ 3.8
109 Torr). The surface features in all of the images (a)-(c)
resemble the so-called cross-linked (1 2) structure, suggestive of a strongly O-deficient surface.22,39 Interestingly, we
observe the growth on the surface (for example, the isolated
two-dimensional island) during annealing the sample in ethylene (Fig. 2(a)). In contrast, we do not observe any significant changes in the island size or shape during annealing
the same sample in UHV (Fig. 2(b)) or in O2 (Fig. 2(c)).
Since Ti was neither deposited nor removed during our
experiments, the observed growth on the surface can occur
via one or more of the following processes: deposition of
carbon from C2H4, carbothermal reduction of the TiO2 to
form carbide (TiC) or an oxycarbide (TiO1xCx) and/or mass
exchange between the bulk and the surface.29,30,35,37
FIG. 2. STM images (1000 1000 Å2)
acquired from a TiO2(110) sample during
annealing at 1000 K in: (a) a mixture of
C2H4 (9.8 108 Torr) þ O2 (3.2 109
Torr), (b) vacuum (3.8 1010 Torr), and
(c) O2 (3.8 109 Torr). The time t ¼ 0 in
the top most panel (a) is arbitrarily defined,
while t1 ¼ t þ 87 and t2 ¼ t þ 143 min.
Downloaded 16 Jul 2013 to 138.67.128.86. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions
211601-3
Murata et al.
Appl. Phys. Lett. 101, 211601 (2012)
FIG. 3. In situ high-temperature (700 K) STM images (700 700 Å2) acquired from the same TiO2(110) sample in: (a) ultra-high vacuum at t ¼ 0 and (b) after
annealing for 88 min in C2H4 (9.8 109 Torr) þ O2 (1.2 109 Torr) gas mixture. (c) Representative plots of (dI/dV)/(I/V) vs. V data obtained at 700 K using
point-mode scanning tunneling spectroscopy (STS) from the 1 1 regions. The black and red curves correspond to the black and red circles in (a) and (b),
respectively. For clarity, the curves have been vertically translated with the dashed lines indicating the zero values on the vertical axes.
As a means to identify the underlying mechanism leading to the observed growth, we first focused on the determination of surface composition. Figure 3 shows in situ hightemperature STS data along with STM images acquired from
nearly the same region on the TiO2(110) sample while
annealing at 700 K first in UHV (Fig. 3(a)) and after 88 min
in the presence of ethylene (Fig. 3(b)). Point-mode STS (I
vs. V data) were collected from the surface and used to
extract normalized conductance values [(dI/dV)/(I/V)], a
measure of local density of states (LDOS).42,43 The black
and red curves in Fig. 3(c) are typical plots of [(dI/dV)/(I/V)]
vs. V data obtained from black and red circled regions on
surfaces during annealing in UHV and in ethylene, respectively (the horizontal dotted lines in the plot indicate zero
conductance values for each of the curves). We find that
both the curves are nearly identical; the semiconducting
bandgap, as given by the range of voltages over which the
conductance values are zero in the plot, is ’1.2 eV for both
the surfaces. These results indicate that the surface electronic
structure and hence composition did not change due to the
introduction of ethylene. We note, however, that the measured bandgap value is considerably lower than the bandgaps
of bulk TiO2 (>3 eV), carbon-doped TiO2xCx (2.2 eV),44
and the newly discovered two-dimensional TiO2(110) surface phase (2.1 eV).45 Since the surface electronic structure
is unaffected by the introduction of ethylene, we suggest that
the significantly low bandgap in our samples is likely due to
a strongly O-deficient TiO2x. This is plausible since previous studies have noted drastic changes in the colors and
resistivities of rutile TiO2 crystals annealed in UHV at high
temperatures.40 Furthermore, we show below that the surface
growth mode is strikingly similar to that observed while
annealing in oxygen.
During annealing in ethylene (pC2 H4 ¼ 9.8 108 Torr
with pO2 ;C2 H4 ¼ 3.2 109 Torr), we observe the growth of
new terraces via the formation of alternating layers of 2 1
stripes and 1 1 terraces (see Fig. 4). Such oscillatory
growth has been previously observed during annealing of an
O-deficient TiO2(110) crystal in pure O2 atmosphere.30,31,33
From higher-resolution STM images, the measured in-plane
periodicities and the heights of the stripes are found to be consistent with the expected values for 1 1 terraces and 2 1
stripes on TiO2(110). This data provide further evidence that
the observed growth features are indeed TiO2x. Since
ethylene molecules alone cannot account for the additional
oxygen needed for the surface growth, we conclude that the
oxygen atoms must be supplied from the trace amounts of
oxygen gas present during dosing of ethylene. Based upon
these results, we infer that annealing TiO2(110) in C2H4 þ O2
gas mixtures is qualitatively similar to annealing in pure O2.
These observations are typical of all annealing experiments
carried out at temperatures T between 700 and 1000 K.
In order to better understand the observed phenomena
and the role of ethylene, we measured changes in surface
mass of TiO2 as a function of annealing time t, T, and O2
and C2H4 gas pressures. In each measurement sequence, the
increase/decrease in surface mass is determined from the difference in total area of 1 1 terraces and 2 1 stripes in
STM images acquired at each time t with respect to the area
measured in the same field of view in the image acquired at
t ¼ 0. (For areal calculations, we assumed that the 2 1
stripes are made up of 3.0 -Å-wide Ti3O6 structure.22 The
exact structure of the 2 1 stripes,24 however, has no effect
on our results since we are interested only in the relative
changes in areal coverages.) Figure 5(a) shows plots of
changes in TiO2 surface mass vs. t during annealing in UHV,
FIG. 4. Oscillatory growth of TiO2 stripes
and terraces during annealing in ethyleneoxygen gas mixtures. In situ high-temperature
STM images (1500 1500 Å2) of TiO2(110)
acquired during annealing at 900 K in C2H4
(9.8 108 Torr) þ O2 (3.2 109 Torr) gas
mixture. The time t ¼ 0 is defined arbitrarily
as the time at which the first image in this
measurement sequence was obtained. The
arrows highlight the growth of alternating
layers of stripes on top of a few islands.
Downloaded 16 Jul 2013 to 138.67.128.86. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions
211601-4
Murata et al.
Appl. Phys. Lett. 101, 211601 (2012)
FIG. 5. (a) Plot of surface mass (ML) of TiO2 as a function of annealing time t in: 3.8 1010 Torr vacuum (open circles), 3.8 109 Torr pure O2 (open
squares), and 9.8 108 Torr C2H4 with 3.2 109 Torr O2 (solid squares) at 1000 K. (b) Plot of dissolution rates of TiO2 during annealing in vacuum, as a
function of temperature T. (c) Growth rates of TiO2, normalized with respect to O2 pressure, plotted as a function of T during annealing TiO2(110) crystal in
C2H4 þ O2 (solid green triangles) and pure O2 (open circles).
pure O2, and C2H4 þ O2 at T ¼ 1000 K. For comparison, pO2
is set to be nearly same (3.8 109 Torr) as that of the
pO2 ;C2 H4 (3.2 109 Torr) in C2H4. During annealing in
UHV, we find that the surface mass decreases with time.
This result is expected and is a consequence of O desorption
from the surface and Ti diffusion into the bulk. In the presence of pure O2, we observe little changes in the surface
with time. This is likely due to the existence of a dynamic
equilibrium between desorption of oxygen from the surface
and the adsorption of oxygen from the gas phase at this
annealing temperature and pressure.33,34 These results are
consistent with previous reports,29,32–36 where it has been
shown that annealing at high temperatures in vacuum leads
to desorption of O atoms from the surface and diffusion of Ti
atoms into the bulk. And, annealing in oxygen reverses this
process, i.e., suppresses O atom desorption and promotes
surface growth via Ti diffusion from the bulk on to the surface. Note the dramatic increase in surface mass with the
addition of ethylene (Figs. 2 and 4), a clear indication that
the presence of ethylene promotes the growth of TiO2x.
The effect of temperature on the rates of dissolution/
growth of TiO2 is presented in Figs. 5(b) and 5(c). During
annealing in UHV, TiO2 surface mass decreases and the rate of
dissolution increases with increasing T as shown in Fig. 5(b).
This is because the partial pressure of O2 in equilibrium with
the bulk TiO2 crystal increases with increasing temperature and
leads to O desorption from the sample.33 With the introduction
of O2 gas, desorption of O atoms from the surface is counteracted by the incorporation of O atoms from the O2 gas and can
lead to growth as shown in Fig. 5(c). For a given crystal composition and oxygen pressure, there exists a threshold temperature33,34 below which growth occurs. Above this threshold
temperature, the O desorption rate exceeds the incorporation
rate leading to mass loss from the surface. In our experiments,
the threshold temperature is likely to be higher than 1000 K
since, from Fig. 5(a) data, we infer that annealing the sample at
all T 1000 K in pure O2 with pO2 3.8 109 Torr leads to
growth. And, from Fig. 5(c), the growth rate decreases with
increasing temperature. In contrast, annealing the sample in the
presence of ethylene results in growth and, the growth rates are,
within the measurements uncertainties, independent of temperature. While additional data is required to determine the exact
temperature dependence, the striking finding is that at all temperatures, the growth rates measured during annealing in
C2H4 þ O2 gas mixtures are higher than those obtained during
annealing in pure oxygen.
How does ethylene promote the growth of TiO2? Previous studies have shown that TiO2, in presence of UV radiation, oxidizes ethylene to form CO2 and water.15 While our
experiments are conducted in the absence of UV light, the
high substrate temperatures (700–1000 K) employed are
likely to promote the oxidation of ethylene to produce water
molecules. Given that our crystals are strongly reduced, surface reactivity is expected to be higher,37,39 which can lead
to dissociative chemisorption of H2O molecules46,47 and
hence relatively faster growth of TiO2 than in the presence
of O2. Clearly, more detailed experiments are necessary to
better understand the observed phenomena.
In conclusion, we used in situ high-temperature scanning tunneling microscopy and investigated the effect of ethylene/oxygen gas mixtures on the surface dynamics of rutilestructured TiO2(110). We found that the presence of ethylene
leads to growth of TiO2 on the surface, a phenomenon that is
qualitatively similar to that observed during annealing in
pure oxygen. Our observations provide valuable insights into
the mechanisms controlling the surface composition and
structure during annealing in the presence of reactive gases.
We expect that our results may help improve the catalytic
performance, lifetime operation, and regeneration capability
of transition-metal oxides.
We gratefully acknowledge support from the AFOSR
(Dr. Ali Sayir) FA9550-10-1-0496 and from the NSF
through the Grant No. CMMI-0846858 and CMMI-1200547.
This work has benefited from the use of the facilities at the
Frederick Seitz Materials Research Laboratory Center for
Microanalysis of Materials.
1
C. T. Campbell, Surf. Sci. Rep. 27(1–3), 1–111 (1997).
F. A. Grant, Rev. Mod. Phys. 31(3), 646 (1959).
U. Diebold, Surf. Sci. Rep. 48(5-8), 53–229 (2003).
4
O. Carp, C. L. Huisman, and A. Reller, Prog. Solid State Chem. 32(1-2),
33–177 (2004).
5
A. Fujishima and K. Honda, Nature (London) 238(5358), 37–38 (1972).
6
S. U. M. Khan, M. Al-Shahry, and W. B. Ingler, Science 297(5590),
2243–2245 (2002).
7
P.-C. Maness, S. Smolinski, D. M. Blake, Z. Huang, E. J. Wolfrum, and
W. A. Jacoby, Appl. Environ. Microbiol. 65(9), 4094–4098 (1999).
8
S. J. Tauster, S. C. Fung, and R. L. Garten, J. Am. Chem. Soc. 100(1),
170–175 (1978).
9
A. Hagfeldt and M. Gratzel, Acc. Chem. Res. 33(5), 269–277 (2000).
10
L. Kavan, M. Gratzel, J. Rathousky, and A. Zukal, J. Electrochem. Soc.
143(2), 394–400 (1996).
11
M. V. Ganduglia-Pirovano, A. Hofmann, and J. Sauer, Surf. Sci. Rep.
62(6), 219–270 (2007).
12
N. Serpone, J. Phys. Chem. B 110(48), 24287–24293 (2006).
2
3
Downloaded 16 Jul 2013 to 138.67.128.86. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions
211601-5
13
Murata et al.
C. Di Valentin, G. Pacchioni, and A. Selloni, Phys. Rev. Lett. 97(16),
166803 (2006).
14
W. G€opel, G. Rocker, and R. Feierabend, Phys. Rev. B 28(6), 3427
(1983).
15
S. Yamazaki, S. Tanaka, and H. Tsukamoto, J. Photochem. Photobiol. A:
Chem. 121(1), 55–61 (1999).
16
C. L. Pang, R. Lindsay, and G. Thornton, Chem. Soc. Rev. 37(10), 2328–
2353 (2008).
17
H. Michael, Surf. Sci. 400(1-3), 203–219 (1998).
18
I. M. Brookes, C. A. Muryn, and G. Thornton, Phys. Rev. Lett. 87(26),
266103 (2001).
19
H. Idriss, V. S. Lusvardi, and M. A. Barteau, Surf. Sci. 348(1-2), 39–48
(1996).
20
R. A. Bennett, P. Stone, R. D. Smith, and M. Bowker, Surf. Sci. 454–
456(0), 390–395 (2000).
21
U. Diebold, J. F. Anderson, K. O. Ng, and D. Vanderbilt, Phys. Rev. Lett.
77(7), 1322–1325 (1996).
22
M. Li, W. Hebenstreit, and U. Diebold, Phys. Rev. B 61(7), 4926–4933
(2000).
23
T. J. Beck, A. Klust, M. Batzill, U. Diebold, C. Di Valentin, and A. Selloni, Phys. Rev. Lett. 93(3), 036104 (2004).
24
X.-Q. Gong, N. Khorshidi, A. Stierle, V. Vonk, C. Ellinger, H. Dosch, H.
Cheng, A. Selloni, Y. He, O. Dulub, and U. Diebold, Surf. Sci. 603(1),
138–144 (2009).
25
X. Fu, L. A. Clark, W. A. Zeltner, and M. A. Anderson, J. Photochem.
Photobiol. A: Chem. 97(3), 181–186 (1996).
26
T. N. Obee and S. O. Hay, Environ. Sci. Technol. 31(7), 2034–2038
(1997).
27
S. Koeppen and W. Langel, Phys. Chem. Chem. Phys. 10(14), 1907–1915
(2008).
28
M. E. Zorn, D. T. Tompkins, W. A. Zeltner, and M. A. Anderson, Environ.
Sci. Technol. 34(24), 5206–5210 (2000).
Appl. Phys. Lett. 101, 211601 (2012)
29
H. Onishi and Y. Iwasawa, Phys. Rev. Lett. 76(5), 791–794 (1996).
R. A. Bennett, P. Stone, and M. Bowker, Faraday Discuss. 114, 267–277
(1999).
31
R. D. Smith, R. A. Bennett, and M. Bowker, Phys. Rev. B 66(3), 035409
(2002).
32
K. F. McCarty and N. C. Bartelt, Surf. Sci. 540(2-3), 157–171 (2003).
33
K. F. McCarty, Surf. Sci. 543(1-3), 185–206 (2003).
34
K. F. McCarty and N. C. Bartelt, Surf. Sci. 527(1-3), L203–L212 (2003).
35
K. F. McCarty and N. C. Bartelt, Phys. Rev. Lett. 90(4), 046104 (2003).
36
K. F. McCarty and N. C. Bartelt, J. Cryst. Growth 270(3-4), 691–698
(2004).
37
M. Li, W. Hebenstreit, L. Gross, U. Diebold, M. A. Henderson, D. R. Jennison, P. A. Schultz, and M. P. Sears, Surf. Sci. 437(1–2), 173–190 (1999).
38
R. A. Bennett, P. Stone, N. J. Price, and M. Bowker, Phys. Rev. Lett.
82(19), 3831–3834 (1999).
39
M. Bowker and R. A. Bennett, J. Phys.: Condens. Matter 21(47), 474224
(2009).
40
M. Li, W. Hebenstreit, U. Diebold, A. M. Tyryshkin, M. K. Bowman, G.
G. Dunham, and M. A. Henderson, J. Phys. Chem. B 104(20), 4944–4950
(2000).
41
I. Horcas, R. Fernandez, J. M. Gomez-Rodriguez, J. Colchero, J. GomezHerrero, and A. M. Baro, Rev. Sci. Instrum. 78(1), 013705 (2007).
42
J. A. Stroscio, R. M. Feenstra, and A. P. Fein, Phys. Rev. Lett. 57(20),
2579–2582 (1986).
43
N. D. Lang, Phys. Rev. B 34(10), 5947–5952 (1986).
44
K. Nagaveni, M. S. Hegde, N. Ravishankar, G. N. Subbanna, and G. Madras, Langmuir 20(7), 2900–2907 (2004).
45
J. Tao, T. Luttrell, and M. Batzill, Nat. Chem. 3(4), 296–300 (2011).
46
O. Bikondoa, C. L. Pang, R. Ithnin, C. A. Muryn, H. Onishi, and G. Thornton, Nat. Mater. 5(3), 189–192 (2006).
47
U. Martinez, L. B. Vilhelmsen, H. H. Kristoffersen, J. Stausholm-Mïller,
and B. Hammer, Phys. Rev. B 84(20), 205434 (2011).
30
Downloaded 16 Jul 2013 to 138.67.128.86. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions