Charge Transfer
DOI: 10.1002/ange.201005377
Charge State of Gold Nanoparticles Supported on Titania under
Oxygen Pressure**
Soeren Porsgaard, Peng Jiang, Ferenc Borondics, Stefan Wendt, Zhi Liu, Hendrik Bluhm,
Flemming Besenbacher, and Miquel Salmeron*
Gold nanoparticles supported on TiO2 (Au/TiO2) are active
catalysts for a variety of reactions even below room temper-
ature.[1] This finding has initiated intensive research during
the last two decades aimed at explaining the mechanisms
responsible for this surprisingly high catalytic activity.[2]
Among the proposed explanations are quantum size
effects,[3] reactive low-coordinate Au atoms,[4] oxygen spill-
over effects to and from the support,[5] support-induced
strain,[4] and charge transfer between the support and the Au
nanoparticles.[6] Even though it is accepted that several of
these effects play a role,[7] there is still no general agreement
on the influence of these different mechanisms, particularly
the importance of the oxide support, activation of oxygen, and
the charge state[8] of the Au nanoparticles.
To study the changes in charge state of Au nanoparticles
on a TiO2 substrate under elevated-pressure reaction con-
ditions, ambient-pressure X-ray photoelectron spectroscopy
(APXPS)[9] is the technique of choice. By means of XPS, the
charge transfer can be revealed as a shift in the apparent
binding energy (BE), since a peak shift to a higher apparent
BE indicates donation of electrons, whereas a peak shift to a
lower apparent BE indicates acceptance of electrons.[10]
However, to correctly assign the origin of peak shifts, it is
necessary to use a proper reference that is unaltered during
the experiment. A particular challenge is the band-bending
effect, which has been observed in the case of TiO2.
[11] For
example, Kurtz et al.[11b] reported an irreversible band-bend-
ing effect on reduced TiO2 samples on O2 exposure caused by
oxidation of the reduced TiO2 surface, and band-bending
effects have also been observed for SnO2 by APXPS.
[12]
Here we report on a reversible band-bending effect in the
Au/TiO2 system and discuss its importance for the detection
of charge-transfer phenomena in general. Our results reveal
that reversible adsorption of O2 molecules on TiO2(110)
occurs through charge transfer from the oxide support to the
adsorbates, a finding which is important with regard to the
high catalytic activity observed on Au/TiO2 for numerous
reactions, including CO oxidation.
To establish a proper binding energy reference system in
our APXPS measurements, we designed a sample (sample 1)
with two separate Au areas on a rutile TiO2(110) single-crystal
surface: 1) an area 4 mm in diameter covered by Au and
connected to ground, and 2) an array of 200 mm-diameter
electrically floating Au islands (Figure 1a). This sample was
initially cleaned by sputtering/annealing cycles and subse-
quently oxidized by air, since the lithography process was
performed in a separate chamber (see Experimental Section).
A set of Au 4f spectra recorded in different gas atmospheres
Figure 1. a) Sketch of the fabricated sample of gold evaporated on top
of a rutile TiO2(110) single crystal (sample 1). This sample had two
distinct areas: 1) An area 4 mm in diameter covered by Au and
connected to ground, and 2) An array of 200 mm-diameter electrically
floating Au islands. b) XPS spectra of Au in the two sample areas for
different gas exposures. The total chamber pressure of O2, CO, and
the mixture (O2+CO) was 1 Torr. The spectra are shown offset for
clarity.
[*] P. Jiang, Prof. M. Salmeron
Material Sciences Division
Lawrence Berkeley National Laboratory
1 Cyclotron Rd, Berkeley, CA 94720 (USA)
E-mail: mbsalmeron@lbl.gov
S. Porsgaard, S. Wendt, Prof. F. Besenbacher
Interdisciplinary Nanoscience Center (iNANO)
Aarhus University, DK-8000 Aarhus C (Denmark)
F. Borondics, H. Bluhm
Chemical Sciences Division
Lawrence Berkeley National Laboratory (USA)
Z. Liu
Advanced Light Source
Lawrence Berkeley National Laboratory (USA)
[**] This work was supported by the Director, Office of Science, Office of
Basic Energy Sciences, Chemical Sciences, Geosciences, and
Biosciences Division, under the Department of Energy Contract No.
DE-AC02-05CH11231.
Zuschriften
2314  2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. 2011, 123, 2314 –2317

is shown in Figure 1b. To calibrate the energy scale, we also
recorded spectra from the grounded Au area of sample 1
(Area 1). In ultrahigh vacuum (UHV) we did not observe any
shift in the Au 4f peak positions for the electrically floating
Au (Area 2) relative to grounded Au. This indicates that the
Au islands in Area 2 do not accumulate charge during XPS
measurements, because the conductivity of the TiO2 substrate
is high enough to compensate the emission of photoelectrons.
In O2 or CO atmospheres at 1 Torr, the Au 4f7/2 peak position
in the floating Au also remained unchanged at 84.00
0.05 eV, that is, the Au 4f peak of the macroscopic Au
island is not altered on exposure to reducing and oxidizing
gases up to 1 Torr. These results show that calibration of the
binding-energy scale by using the Au 4f7/2 peak position of
grounded Au is reliable. Care was taken in these in situ
experiments to ensure that the reference material did not
undergo a chemical reaction that could produce chemical
shifts. As we have shown earlier,[13] gold can be oxidized by
prolonged exposure to X-ray photons at elevated oxygen
pressures.
We are now in a position to properly measure chemical
shifts of the core levels of the Ti and O atoms, adsorbates, gas
molecules, and Au in nanoparticles grown by thermal
evaporation. On a clean TiO2 sample (sample 2), the O 1s
and Ti 2p XPS spectra were found to be unchanged by
exposure to CO up to 1 Torr. However, exposure to O2
induces a shift of both the Ti 2p and the O 1s peaks towards
lower BE (Figure 2). Identical peak shifts were observed for
the Ti 3p peaks. The largest observed shift was 0.4 eV in an
atmosphere of 1 Torr O2. However, the absolute values of the
peak shifts may depend on the initial oxidation state of the
TiO2(110) crystal. After pumping the O2 gas out of the
vacuum chamber, all core-level peaks shifted back close to
their original position, that is, the peak-shift effect is
reversible. Figure 2b shows the peak shifts obtained from
the APXPS spectra by Gaussian fits to the peaks. A detailed
analysis reveals that the peaks do not shift completely back to
their original positions, and this suggests that two different
mechanisms are at play.
It could be speculated that the small irreversible peak shift
may be caused by O2 molecules, which are known to
dissociate on the TiO2 surface.
[14] However, dissociation of
O2 molecules cannot account for the observed reversible peak
shift, since recombination of O atoms on TiO2(110) at room
temperature can be ruled out.[14] Instead, we propose that the
reversible peak shifts are due to band bending induced by
charge transfer between adsorbed O2 molecules and the
substrate [Eq. (1)], which changes the surface potential.
O2ðgasÞ Ð O2ðadsÞ ð1aÞ
O2ðadsÞ þ eðTiO2Þ Ð O2ðadsÞ ð1bÞ
Molecularly adsorbed oxygen on TiO2 surfaces has been
previously observed by EPR spectroscopy.[15] Figure 3a shows
Figure 2. a) Series of O 1s and Ti 2p spectra measured on TiO2(110)
(sample 2) for different pressures of O2. The spectra acquired in UHV
(black lines) are the first in the series. The spectra are shown offset for
clarity. b) Shifts in the peak position relative to the positions in the
first spectrum.
Figure 3. a) Energy diagram of the sample and analyzer in UHV
(black) and in the presence of O2 (red). The left-hand side represents
the sample, and the right-hand side the analyzer, with work functions
FS and FA respectively. EB and EgasB are the binding energies under
UHV and under O2 gas pressure, respectively. Similarly, Ekin and E
gas
kin
are the measured kinetic energies, and EF is the Fermi level. b) Draw-
ing of the experimental setup, with X-rays incident at an angle of 178
from the surface normal. The take-off angle of the photoelectrons
relative to the sample normal is 448. c) Ar 2p gas-phase XPS spectra
for pure Ar and a mixture of Ar and O2, respectively.
Angewandte
Chemie
2315Angew. Chem. 2011, 123, 2314 –2317  2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.de

a schematic energy diagram for the TiO2 sample in UHVand
in an O2 atmosphere. The vacuum level changes gradually
between the sample and the nozzle of the analyzer. Since the
electronic levels of the gas-phase species are pinned to the
vacuum level,[16] their XPS peaks will shift following the
changes in contact potential (Figure 3b). To prove this we
performed a similar experiment as described above, but this
time the sample was also exposed to argon (0.15 Torr).
Figure 3c shows the gas-phase Ar 2p XPS spectrum at
0.15 Torr of argon. After addition of 1 Torr of O2 we observed
that the Ar 2p peak shifts by 0.3 eV towards lower BE, that is,
the vacuum level indeed shifts. Hence, we conclude that the
XPS peak shifts in Figure 2 are caused by a band-bending
effect.
The shift of the Ar 2p peak shown in Figure 3 is smaller
than that of 0.4 eV of the Ti 2p and O 1s peaks (cf. Figure 2).
This smaller peak shift can be attributed to the finite width of
the gas volume irradiated by the incident X-ray beam, with
gas molecules at different distances from the sample surface
and thus at different vacuum levels. Thus, the Ar 2p peak shift
represents a minimum value of the vacuum-level shift at the
sample surface. No peak shift was observed upon CO
exposure (not shown).
We now turn our attention to shifts in the Au 4f peak
positions of supported Au nanoparticles. After Au evapora-
tion onto a clean TiO2(110) substrate (sample 3), the 4f core-
level peak of Au in the nanoparticles was found at a slightly
higher BE than that of 84.0 eV for the Au foil. This is usually
explained by a combination of initial- and final-state
effects.[2b,e, 17] It is unclear whether this effect, measured
under UHV conditions, is relevant at the higher pressures
of real catalysts. To study the influence of the reactants for
CO oxidation, we sequentially exposed the Au/TiO2(110)
sample to 200 mTorr of O2 and 200 mTorr of CO (Figure 4).
No shift of the Au 4f peak was observed upon exposure to
either O2 or CO. However, the O 1s and Ti 3p peaks shift on
O2 exposure. If the Ti 3p peak were used as a BE reference,
however, an apparent shift of Au 4f towards higher BE would
be obtained and erroneously interpreted as electron donation
from the Au nanoparticles. The use of a grounded piece of Au
foil as external reference instead allowed us to reveal that the
Au 4f peak actually does not shift (within the 0.05 eV
accuracy of the experiment) on exposure of the Au/TiO2
sample to O2. We estimated the order of magnitude limit on
the charging of Au nanoparticles by modeling them as
spherical capacitors. For a sphere of 2 nm radius a voltage
of 0.05 V would correspond to a charge of roughly 0.07
electrons or approximately 105 electrons per atom.
In conclusion, our results show that all TiO2-related peaks,
that is, Ti 2p, Ti 3p, and O 1s, shift together by the same
amount under exposure to O2, while those of supported Au
nanoparticles do not. The binding energies from the core
levels of the TiO2 substrate cannot be used as energy
reference for calibration of the XPS spectra. We have
shown that the BE shifts of these peaks are due to band-
bending effects caused by molecular O2, adsorbed on the TiO2
support at an O2 pressure of 1 Torr. The band-bending also
affects the vacuum level of TiO2, causing the XPS peaks of
nearby gas species to shift by similar amounts. The present
results clarify the origin of peak shifts of Au nanoparticles and
improve our fundamental understanding of catalytic reaction
mechanisms, for example, in CO oxidation, because adsorbed
O2 and O2
 may play a key role in supplying reactive oxygen
for surface-catalyzed reactions. Our findings demonstrate that
Au nanoparticles supported on TiO2 do not donate or accept
additional charge when the sample is exposed to either CO or
O2, at least at levels that would produce measurable shifts in
the XPS peaks.
Experimental Section
The experiments were carried out at beamline 11.0.2 and beamline
9.3.2 at the Advanced Light Source, Berkeley.[9] The samples were
fresh 5 10 mm rutile TiO2(110) single crystals that were cleaned by
sputtering/annealing cycles until no contamination was detected by
XPS. The Au/TiO2 sample (sample 3) was produced by exposing the
TiO2 crystal to Au vapor after exposure to O2.
To test the validity of the Au reference we fabricated a sample
(sample 1) with a gold pattern consisting of a circular area of 4 mm
diameter and an array of 200 mm-diameter islands (Figure 1). This
reference sample was fabricated in a separate chamber and trans-
ferred through air to the experimental chamber. After air exposure,
no contamination was detected. To minimize shifts of the photon
energy due to physical movement of the monochromator, all XPS
measurements with Au nanoparticles were carried out with a constant
photon energy of 830 eV. In all other experiments, the photon energy
was 690 eV. For reference purposes XPS spectra were recorded
frequently on a grounded Au foil in UHV to obtain a Au 4f7/2
spectrum, which is calibrated to 84.0 eV. This procedure was followed
throughout all the experiments to check for changes in the photon
energy. All measurements were done with a typical acquisition time
of 1 min per spectrum, and multiple spots were measured to minimize
beam-damage effects. This careful procedure excludes effects such as
beam-induced oxidation, as observed in previous studies.[13] The
values for the peak shifts were obtained by subtracting the Shirley
background and by Gaussian fitting to the peaks in the APXPS
spectra. All experiments were carried out at room temperature, and
gases were dosed by backfilling the chamber.
Received: August 27, 2010
Revised: December 1, 2010
Published online: February 3, 2011
Figure 4. Au 4f and O 1s XPS spectra for different gas and pressure
conditions on Au/TiO2(110). The spectra acquired in UHV are the first
in the series. The spectra are shown offset for clarity.
Zuschriften
2316 www.angewandte.de  2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. 2011, 123, 2314 –2317

.Keywords: charge transfer · gold · nanoparticles ·
photoelectron spectroscopy · supported catalysts
[1] M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, M. J. Genet,
B. Delmon, J. Catal. 1993, 144, 175 – 192.
[2] a) M.-C. Daniel, D. Astruc, Chem. Rev. 2004, 104, 293 – 346;
b) Q. Fu, T. Wagner, Surf. Sci. Rep. 2007, 62, 431 – 498; c) M.
Haruta, Catal. Today 1997, 36, 153 – 166; d) A. S. K. Hashmi,
G. J. Hutchings, Angew. Chem. 2006, 118, 8064 – 8105; Angew.
Chem. Int. Ed. 2006, 45, 7896 – 7936; e) R. Meyer, C. Lemire,
Sh. K. Shaikhutdinov, H.-J. Freund,Gold Bull. 2004, 37, 72 – 124.
[3] M. Valden, X. Lai, D. W. Goodman, Science 1998, 281, 1647 –
1650.
[4] M. Mavrikakis, P. Stoltze, J. K. Nørskov, Catal. Lett. 2000, 64,
101 – 106.
[5] Z. P. Liu, P. Hu, A. Alavi, J. Am. Chem. Soc. 2002, 124, 14770 –
14779.
[6] a) D. Ricci, A. Bongiorno, G. Pacchioni, U. Landman, Phys. Rev.
Lett. 2006, 97, 3106; b) A. Sanchez, S. Abbet, U. Heiz, W.-D.
Schneider, H. Hkkinen, R. N. Barnett, U. Landman, J. Phys.
Chem. A 1999, 103, 9573 – 9578.
[7] B. Hvolbæk, T. V. W. Janssens, B. S. Clausen, H. Falsig, C. H.
Christensen, J. K. Nørskov, Nano Today 2007, 2, 14 – 18.
[8] a) J. G. Wang, B. Hammer, Phys. Rev. Lett. 2006, 97, 136107;
b) B. Yoon, H. Hkkinen, U. Landman, A. S. Wrz, J.-M.
Antonietti, S. Abbet, K. Judai, U. Heiz, Science 2005, 307, 403 –
407.
[9] D. F. Ogletree, H. Bluhm, E. D. Hebenstreit, M. Salmeron,Nucl.
Instrum. Methods Phys. Res. Sect. A 2009, 601, 151 – 160.
[10] K. Siegbahn, ESCA—Atomic, Molecular and Solid State Struc-
ture Studied by Means of Electron Spectroscopy, Vetenskaps-
societeten i Uppsala, Uppsala, 1967.
[11] a) E. N. Figurovskaya, V. F. Kiselev, F. F. Volkenshtei˘n, D. A. N.
SSSR 1965, 161, 1142 – 1145; b) R. L. Kurtz, R. Stockbauer, T. E.
Madey, E. Romn, J. L. De Segovia, Surf. Sci. 1989, 218, 178 –
200; c) J. Nowotny, T. Bak, L. R. Sheppard, M. K. Nowotny, J.
Am. Chem. Soc. 2008, 130, 9984 – 9993.
[12] a) Y. Gassenbauer, R. Schafranek, A. Klein, S. Zafeiratos, M.
Hvecker, A. Knop-Gericke, R. Schlgl, Phys. Rev. B 2006, 73,
245312; b) Y. Gassenbauer, R. Schafranek, A. Klein, S. Zafeir-
atos, M. Hvecker, A. Knop-Gericke, R. Schlgl, Solid State
Ionics 2006, 177, 3123 – 3127.
[13] P. Jiang, S. Porsgaard, F. Borondics, M. Kber, A. Caballero, H.
Bluhm, F. Besenbacher, M. Salmeron, J. Am. Chem. Soc. 2010,
132, 2858 – 2859.
[14] a) M. A. Henderson, W. S. Epling, C. L. Perkins, C. H. F. Peden,
U. Diebold, J. Phys. Chem. B 1999, 103, 5328 – 5337; b) E. Lira,
J. O. Hansen, P. Huo, R. Bechstein, P. Galliker, E. Laegsgaard, B.
Hammer, S. Wendt, F. Besenbacher, Surf. Sci. 2010, 604, 1945 –
1960.
[15] H. Liu, A. I. Kozlov, A. P. Kozlova, T. Shido, K. Asakura, Y.
Iwasawa, J. Catal. 1999, 185, 252 – 264.
[16] H. Bluhm, J. Electron Spectrosc. Relat. Phenom. 2010, 177, 71 –
84.
[17] A. Howard, D. N. S. Clark, C. E. J. Mitchell, R. G. Egdell, V. R.
Dhanak, Surf. Sci. 2002, 518, 210 – 224.
Angewandte
Chemie
2317Angew. Chem. 2011, 123, 2314 –2317  2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.de