Visible-light-driven oxidation of organic contaminants in air with gold nanoparticle catalysts on oxide supports

Article abstract of DOI:10.1002/anie.200800602

(Graph Presented) Staying cool in the sun: When illuminated with visible light, gold nanoparticles dispersed on various metal oxide supports show significant activity in the oxidation of formaldehyde and methanol in air at room temperature. HCHO conversion under illumination of blue light (see picture, blue bars) and under red light (red bars) are compared.

Full text of DOI:10.1002/anie.200800602

Angewandte
Chemie
DOI: 10.1002/anie.200800602
Nanostructured Catalysts
Visible-Light-Driven Oxidation of Organic Contaminants in Air with
Gold Nanoparticle Catalysts on Oxide Supports**
Xi Chen, Huai-Yong Zhu,* Jin-Cai Zhao, Zhan-Feng Zheng, and Xue-Ping Gao*
One of the great challenges in catalysis is to devise new
catalysts that have high activity when illuminated by visible
light.[1] Solving this challenge will allow us to use sunlight, an
abundant and clean low-cost energy source, to drive chemical
reactions. Visible light (wavelength l > 400 nm) constitutes
around 43% of solar energy,[2] and the energy of sunlight to
the Earth is about 10000 times more than the current energy
consumption of the world.[3] Many approaches have been
proposed to develop visible light photocatalysts, including
doping TiO2 with metal ions or metal atom clusters,[4,5]
incorporating nitrogen[6] and carbon[7] into TiO2, and employ-
ing other metal oxides or polymetallates as catalyst materi-
als.[2,4,5,8] Research has been mainly concentrated on semi-
conductor oxides. Sulfides have also been studied, but they
are not suitable catalysts because of their poor chemical
stability.[1,4] However, searching for catalysts that can work
under visible light should not be limited to semiconductor
materials with band-gap structure, but can be extended to
other materials, such as gold nanoparticles.
It can be said that glaziers in medieval forges were the first
nanotechnologists who produced colors with gold nanopar-
ticles of different sizes,[9] although they had little under-
standing of the modern day principles which have become a
hot topic in the last two decades. In recent years there have
been numerous studies on the optical properties of gold
nanoparticles.[10] Gold nanoparticles absorb visible light
intensely because of the surface plasmon resonance (SPR)
effect.[10,11] The electromagnetic field of incident light couples
with the oscillations of conduction electrons in gold particles,
resulting in strong-field enhancement of the local electro-
magnetic fields near the rough surface of gold nanoparti-
cles.[12] The enhanced local field strength can be over
500 times larger than the applied field for structures with
sharp apices, edges, or concave curvature (e.g. nanowires,
cubes, triangular plates, and nanoparticle junctions).[13] The
SPR absorption may cause rapid heating of the nanoparti-
cles.[14,15]
Gold nanoparticles supported on metal oxides are effi-
cient catalysts for important oxidation process, including
selective oxidation of hydrocarbons and oxidation of various
volatile organic compounds (VOCs), such as CO, CH3OH,
and HCHO at moderately elevated temperatures.[16,17] There-
fore, the combination of the SPR absorption and the catalytic
activity of gold nanoparticles presents an important oppor-
tunity: if the heated gold nanoparticles could activate the
organic molecules on them to induce oxidation of the organic
compounds, then oxidation on gold catalysts can be driven by
visible light at ambient temperature. Moreover, the SPR is a
local effect, limited to the noble metal particles, so that the
light only heats gold nanoparticles, which generally account
for a few percent of the overall catalyst mass.[16] This leads to
significant saving in energy consumption for catalyzing
organic compound oxidation.
To verify the possibility of driving the VOC oxidation with
visible light at room temperature, we prepared gold particles
supported on various oxide powders. ZrO2 and SiO2 powders
were first chosen as supports, because their band gaps are
circa 5.0eV [18] and circa 9.0eV, [19] respectively, which are
much larger than the energies of the photons of visible light
(less than 3.0eV). Thus, the light cannot excite electrons from
the valence band to the conduction band. It is also impossible
for the gold nanoparticles on ZrO2 to reduce the band gaps of
ZrO2 enough for visible light photons to be absorbed and
excite electrons in ZrO2. Thus, the catalytic activity is not
caused by the same mechanism as occurs in semiconductor
photocatalysts, but is due to the SPR effect of gold nano-
particles. The changes in the concentrations of the reactant
(HCHO, 100 ppm) and product (CO2), when gold supported
on ZrO2 was used as the catalyst, are depicted in Figure 1a.
The initial concentration of HCHO in the glass vessel was
100 ppm. HCHO content decreased by 64% in two hours
under the irradiation of six light tubes of blue light (with
wavelength between 400 and 500 nm and the irradiation
energy determined to be 0.17 WcmÀ2 at the position of glass
slides coated with the gold catalysts), and the CO2 content in
the vessel increased accordingly. These results confirm that
the oxidation of formaldehyde to carbon dioxide proceeds to
a large extent at ambient temperature. The turnover fre-
quency was calculated as being about 1.2 ” 10À3 molecules of
[*] Dr. X. Chen, Prof. H.-Y. Zhu, Z.-F. Zheng
School of Physical and Chemical Sciences
Queensland University of Technology
Brisbane, Qld 4001 (Australia)
Fax: (+61)7-3864-1804
E-mail: hy.zhu@qut.edu.au
Prof. X.-P. Gao
Institute of New Energy Material Chemistry
Nankai University
Tianjin 300071 (China)
Fax: (+86)22-2356-2604
E-mail: xpgao@nankai.edu.cn
Prof. J.-C. Zhao
Institute of Chemistry
The Chinese Academy of Science
Beijing 100080 (China)
[**] Financial supports from the Australian Research Council (ARC) and
NCET (040219) of China are gratefully acknowledged. Thanks are
also due to Dr. Hongwei Liu and Dr. Yong Yuan for conducting the
TEM experiments and to Younan Zhu for his careful revision.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200800602.
Angew. Chem. Int. Ed. 2008, 47, 5353 –5356
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5353

Communications
Figure 2. Influence of the oxide support on the activity of HCHO
oxidation reaction. Blue bars: HCHO conversion (%) under illumina-
tion of blue light; red bars: conversion under red light.
and 16% of methanol was oxidized under irradiation of six
blue and red light tubes in two hours, respectively, whereas no
oxidation was observed for cyclohexane. The polarity of a
molecule is in proportional to the value of its dipole moment.
The dipole moments of formaldehyde, methanol and cyclo-
hexane are 2.3, 1.7 and 0Debye, respectively. The results
demonstrate that for the reactant molecules of similar
structural features, the molecule with higher polarity is
much easier to activate and thus be oxidized on gold
nanoparticles under visible light.
Figure 1. Oxidation of 100 ppm HCHO on gold nanoparticles sup-
ported on ZrO2 (Au/ZrO2) under blue light or sunlight at 258C.
a) Changes in reactant HCHO (blue lines) and product CO2 (red lines)
~
&
*
under different light intensities ( 0.17, 0.13, 0.08, and
!
0.02 W cmÀ2) as a function of irradiation time. b) The relationship
The function of the oxide supports in the catalysts used in
the present study may be similar to that of the oxide supports
of the gold catalysts used for the thermal oxidation.[16,21] If an
oxide of a transition metal or rare earth element is used as
support, the gold catalysts exhibited marked activity
(Figure 2), whereas a gold catalyst on SiO2 exhibited moder-
ate activity under red light only. This situation is similar to
that of the thermal oxidation of CO with gold catalysts, in
which the oxide supports in the gold catalysts have been
classified into inert and active supports.[21] Silica is an inert
support, which has the function of stabilizing the small gold
nanoparticles but does not adsorb oxygen, whereas the active
supports, such as transition metal oxides and rare earth
oxides, can adsorb oxygen molecules. To understand the
function of the oxide supports in the catalytic oxidation driven
by light, we conducted the catalytic reaction with the Au/ZrO2
and Au/SiO2 catalysts, respectively, in pure nitrogen
(99.99%), instead of dry air while maintaining the other
experimental conditions unchanged. About 5% of HCHO
was degraded on the Au/ZrO2 catalysts under blue light, but
no HCHO oxidation was observed on the Au/SiO2 catalysts.
The results indicate that oxygen, the oxidation agent for the
reaction, mainly comes from air. If we replaced air with pure
nitrogen, there was no oxygen for the oxidation with Au/SiO2
catalyst, therefore no oxidation was observed. However, for
Au/ZrO2 catalyst, the strongly adsorbed oxygen remained on
ZrO2 support after air was replaced with nitrogen, and was
thus able to take part in the oxidation. The adsorbed oxygen
on the active supports also explains the high activity of the
catalysts on these supports when the reaction proceeds in dry
air.
between HCHO conversion and light intensity.
HCHO per gold atom per second. This frequency is com-
parable to those for the CO oxidation on gold catalysts by
heating the reaction system to 808C or above, which is
between 10À2 and 10À3 molecules of CO per gold atom per
second.[20]
The catalytic activity was dependent on the intensity of
light irradiation. When the experiment was conducted with-
out light irradiation, no changes in HCHO content were
observed and no CO2 was detected. When the light intensity
was reduced by turning off two and four of the six light tubes,
the conversion of HCHO decreased correspondingly (Fig-
ure 1a). Sunlight was also used as the light source for the
HCHO oxidation, and 8% of HCHO was converted into CO2
in two hours. The intensity of the sunlight is much lower
(0.02 WcmÀ2) than those of the light tubes. The HCHO
conversion rate is plotted against the light intensity (Fig-
ure 1b), which clearly demonstrates the dependence of
conversion on light intensity. When a red light with wave-
lengths between 600 and 700 nm was used, the catalytic
activity of Au/ZrO2 catalyst is slightly lower compared with
that of the blue light; 50% of HCHO was oxidized in two
hours (Figure 2). These facts indicate that the reaction is
undoubtedly driven by visible light.
As the SPR effect involves enhanced local electromag-
netic fields, the oscillating fields should interact more
intensively with polar than nonpolar molecules and chemical
bonds. This conclusion was verified experimentally by con-
ducting the oxidation of methanol and cyclohexane using Au/
ZrO2 catalyst under the same experimental conditions. 18%
As seen in transmission electron microscopy (TEM)
images (Figure 3), the size of the gold nanoparticles on
5354
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 5353 –5356

Angewandte
Chemie
particles in a catalyst can be obtained from the difference
between the spectra of the catalyst and oxide support (see the
Supporting Information, Figure 2). The irradiation energy
absorbed by the gold nanoparticles is then derived from the
overlap area of the absorption spectrum of the gold nano-
particles and the spectra of the irradiation tube sources (see
the Supporting Information, Figure 3) as well as the irradi-
ation energy (for instance 0.17 WcmÀ2 for blue light). The
turnover frequency per unit of the irradiation energy absor-
bed by gold nanoparticles of the catalysts, the normalized
turnover frequency, was calculated from the data of gold
content and HCHO conversion of the catalyst. The gold
content, turnover frequency derived from HCHO conversion,
and gold content, and normalized turnover frequency, are
listed in Table 1. The absorbed energy by the gold nano-
particles on SiO2 is lower than those particles on other oxides.
This could be an important reason for the relative low activity
of Au/SiO2 sample. The activities of the catalysts on various
oxides in the normalized turnover frequency are given in the
Supporting Information, Figure 4.
Based on these facts, we propose a tentative reaction
mechanism for the light-driven catalytic oxidation. The
irradiation of incident light with a wavelength in the range
of the SPR band may result in two consequences. The first is
that light absorption by the gold nanoparticles[9–11] could heat
these nanoparticles up quickly.[14] We calculated the energy of
the light absorbed by gold nanoparticles in the experiment
with Au/ZrO2 catalysts (gold particle size 20–30 nm) under
blue light (0.17 W cmÀ2), assuming the particles are spherical,
and estimated how fast such energy could heat the particles
up, namely at a rate of 3–58C per second. This means that gold
particles of this catalyst could be heated to 1008C, at which
the HCHO oxidation proceeds significantly,[22] in 2–3 min,
even if the energy conversion efficiency is 10% or below.[15]
The second consequence is that the interaction between the
oscillating local electromagnetic fields and polar molecules
also assists in activating the molecules. The activated polar
organic molecules react with oxygen in close proximity.
The fact that it is difficult to activate nonpolar molecules
under the same conditions indicates that the interaction is
crucial to the catalytic reaction. The oxygen adsorbed on the
active supports can migrate to the gold nanoparticles, and the
adsorption of oxygen on the oxide support increases the
concentration of oxygen around the organic molecules on the
gold nanoparticles and thus the opportunity to react with the
molecules, accelerating the oxidation. The proposed reaction
mechanism is distinctly different from that for the reaction
catalyzed by semiconductor photocatalysts.
Figure 3. TEM images of the gold catalysts on different oxide supports.
a) Au/ZrO2, b) Au/CeO2, c) Au/Fe2O3, d) Au/SiO2. Arrows indicate gold
nanoparticles.
different oxides can be different. The size distributions of the
gold nanoparticles on four supports were calculated from the
TEM images (see the Supporting Information, Figure 1). The
distributions on ZrO2 and SiO2 are broad, and have peak
values at 27 and 53 nm, respectively. Most of the gold
nanoparticles on CeO2 are below 10nm, and most of the
particles on Fe2O3 supports are between 10and 30nm. The
mean sizes of gold particles are in the order Au/CeO2 < Au/
Fe2O3 < Au/ZrO2, and the performance of the catalysts on the
three oxides are in the order Au/ZrO2 > Au/CeO2 > Au/
Fe2O3. It appears that gold particle size is not the determining
factor, and the synergistic effect of the active metal oxide has
as a much stronger influence on the catalytic activity than
particle size.
Nonetheless, on the same metal oxide support, such as
ZrO2, the smaller the gold nanoparticle, the better the
catalytic activity. When the Au/ZrO2 catalysts was heated at
5008C, the gold particles in this catalyst increased from about
20–30 nm to over 100 nm, and the HCHO conversion activity
decreased substantially, from 64% to 8%.
A number of factors, such as particle size and morphology,
and the dielectric constant of the oxide medium, can affect the
SPR absorption of gold nanoparticles and thus the activity of
the gold catalyst. The absorption spectra for the gold nano-
Table 1: Gold content, absorption of irradiation energy, and catalytic activity of the gold nanoparticles on various oxides.
Au/ZrO2
Au/CeO2
Au/Fe2O3
Au/SiO2
Gold content [wt%]
2.44
0.168
2.62
0.159
3.10
0.124
2.24
0.108
0
Absorbed energy by gold nanoparticles under blue light [W cmÀ2
]
Turnover frequency under blue light [Au-atomÀ1 sÀ1
]
1.2”10À3
7.1”10À3
0.135
6.5”10À4
4.1”10À3
0.153
4.6”10À4
3.7”10À3
0.149
Normalized turnover frequency under blue light [cm2 JÀ1 Au-atomÀ1
Absorbed energy by gold nanoparticles under red light [W cmÀ2
Turnover frequency under red light [Au-atomÀ1 sÀ1
Normalized turnover frequency under red light [cm2 JÀ1 Au-atomÀ1
]
0
]
0.077
2.7”10À4
3.5”10À3
]
9.4”10À4
7.0”10À3
5.6”10À4
3.7”10À3
3.4”10À4
2.3”10À3
]
Angew. Chem. Int. Ed. 2008, 47, 5353 –5356
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5355

Communications
microgrid coated with a holey carbon film. The EDS experiment
was carried out on a FEI Quanta 200 Environmental SEM.
In summary, compared to the situation for conventional
oxidation with heating, the visible-light-driven process has a
significant advantage in that visible light interacts only with
the gold nanoparticles. Given that the gold content in this
study is 2–4 wt% of the supports, it requires much lesser
energy input to activate the reaction compared to the
conventional catalytic oxidation in which both the gold
particles and the support material are heated to high temper-
atures. The light-driven reaction can proceed at ambient
temperature at reaction rate similar to those of the catalytic
oxidation under heating as shown by the turnover frequency
of the reaction. Such a reaction at ambient temperature
provides great convenience, for example in indoor air
purification.
The finding that gold nanoparticles on ZrO2 and SiO2 can
catalyze the VOC oxidations under visible light is likely to
answer unsolved questions concerning the mechanism of
doping semiconductor photocatalysts with gold to narrow the
band gap of the semiconductor. The gold dopant itself may be
an important contributor to the visible-light absorption and
activity. Further possibilities to drive other chemical reactions
with sunlight at ambient temperature include, for instance,
reactions currently using gold catalysts.[17] This could open up
a new direction in catalysis and herald significant changes in
the economy and environmental impact of chemical produc-
tion.
Received: February 6, 2008
Revised: March 12, 2008
Published online: June 11, 2008
Keywords: gold · nanotechnology · oxidation · photochemistry ·
.
surface plasmon resonance
[1] N. S. Lewis, Nature 2001, 414, 589 – 590.
[2] Z. G. Zou, J. H. Ye, K. Sayama, H. Arakawa, Nature 2001, 414,
625 – 627.
[3] M. Grätzel, Nature 2001, 414, 338 – 344.
[4] M. I. Litter, Appl. Catal. B 1999, 23, 89 – 114.
[5] A. L. Linsebigler, G. Q. Lu, J. T. Yates, Chem. Rev. 1995, 95,
735 – 758.
[6] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science
2001, 293, 269 – 271.
[7] S. U. M. Khan, M. Al-Shahry, W. B. Ingler, Science 2002, 297,
2243 – 2245.
[8] S. Ikeda, M. Hara, J. N. Kondo, K. Domen, H. Takahashi, T.
Okubo, M. Kakihana, Chem. Mater. 1998, 10, 72 – 77.
[9] M. Ratner, D. Ratner, Nanotechnology: AGentle Introduction to
the Next Big Idea, Prentice Hall, New Jersey, 2003.
[10] a) P. Mulvaney, Langmuir 1996, 12, 788 – 800; b) X. H. Huang,
I. H. El-Sayed, W. Qian, M. A. El-Sayed, J. Am. Chem. Soc.
2006, 128, 2115 – 2120.
[11] a) D. L. Feldheim, C. A. J. Foss, Metal Nanoparticles: Synthesis,
Characterization, and Applications, Marcel Dekker, New York,
2002; b) H. Yuan, Chem. Mater. 2007, 19, 1592 – 1600; c) G. C.
Schatz, Acc. Chem. Res. 1984, 17, 370– 376.
[12] a) J. Jiang, K. Bosnick, M. Maillard, L. Brus, J. Phys. Chem. B
2003, 107, 9964 – 9972; b) L. M. Liz-Marzan, Langmuir 2006, 22,
32 – 41.
Experimental Section
[13] K. L. Kelly, E. Coronado, L. L. Zhao, G. C. Schatz, J. Phys.
Chem. B 2003, 107, 668 – 677.
[14] A. Takami, H. Kurita, S. Koda, J. Phys. Chem. B 1999, 103, 1226 –
1232.
[15] D. K. Roper, W. Ahn, M. Hoepfner, J. Phys. Chem. C 2007, 111,
3636 – 3641.
[16] a) M. Haruta, Gold Bull. 2004, 37, 27 – 36; b) M. D. Hughes, Y. J.
Xu, P. Jenkins, P. McMorn, P. Landon, D. I. Enache, A. F. Carley,
G. A. Attard, G. J. Hutchings, F. King, E. H. Stitt, P. Johnston, K.
Griffin, C. J. Kiely, Nature 2005, 437, 1132 – 1135; c) M. Haruta,
A. Ueda, S. Tsubota, R. M. T. Sanchez, Catal. Today 1996, 29,
443 – 447.
[17] a) M. Valden, X. Lai, D. W. Goodman, Science 1998, 281, 1647 –
1650; b) Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos,
Science 2003, 301, 935 – 938.
[18] A. Emeline, G. V. Kataeva, A. S. Rudakova, V. K. Ryabchuk, N.
Serpone, Langmuir 1998, 14, 5011 – 5022.
[19] J. Robertson, K. Xiong, S. J. Clark, Thin Solid Films 2006, 496,
1 – 7.
The gold nanoparticles supported on oxide powders were prepared by
the impregnation method: In general, 200 mg of support powder was
added into 50mL aqueous solution of HAuCl 4 and trisodium citrate
under stirring. The suspension was heated to and kept at 908C for
10min. The solid was then separated from the solution by centrifu-
gation and washed with distilled water and ethanol. The precipitate
was washed, and dried at 808C for 5 h. Gold nanoparticles on oxide
powders were suspended in ethanol to coat the solid onto glass slides
(2.5 cm wide and 7.5 cm long), which were heated at 808C.
The coated slides were placed in a glass vessel of about 6 liters.
The vessel is in a chamber with six light tubes (18 W/tube, Philips) as
the light source. Air conditioning was applied in the chamber to
maintain the temperature at 258C, as the light illumination could
otherwise cause an increase in the temperature of the vessel. The
vessel was filled with air and sealed after the coated glass slides were
mounted. Formaldehyde (HCHO) or methanol (CH3OH) was
injected into the vessel. The gaseous specimens were sampled
before the light was turned on and analyzed by a gas chromatograph
(Shimadzu GC-2014) used specially for this experiment. The gaseous
specimens were sampled during the reaction course with light on to
monitor the changes in the content of injected HCHO or CH3OH and
product CO2. TEM images were obtained using a Philips CM200
TEM or a FEI Tecnai 20TEM with an accelerating voltage of 200kV.
The specimens were sample powders deposited onto a copper
[20] V. Aguilar-Guerrero, B. C. Gates, Chem. Commun. 2007, 3210–
3212.
[21] M. M. Schubert, S. Hackenberg, C. Andre, M. Muhler, V. Plzak,
J. Catal. 2001, 197, 113 – 122.
[22] Y. N. Shen, X. Z. Yang, Y. Z. Wang, Y. B. Zhang, H. Y. Zhu, L.
Gao, M. L. Jia, Appl. Catal. B 2008, 79, 142 – 148.
5356
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 5353 –5356