Atomic XAFS as a tool to probe the electronic properties of supported noble metal nanoclusters

Article abstract of DOI:10.1021/ja043107l

Atomic XAFS is a very attractive technique for probing electronic properties of supported metal nanoclusters. For platinum nanoparticles on different supports, the technique is found to be in good agreement with infrared CO adsorption measurements. The advantages of AXAFS, however, are that no probe molecule is required and that real-time measurements under reaction conditions are possible. Copyright

Full text of DOI:10.1021/ja043107l

Published on Web 02/17/2005
Atomic XAFS as a Tool to Probe the Electronic Properties of Supported
Noble Metal Nanoclusters
†
†
†
‡
‡
Ad M. J. van der Eerden, Tom Visser, T. Alexander Nijhuis, Yasuo Ikeda, Muriel Lepage,
†
,†
Diek C. Koningsberger, and Bert M. Weckhuysen*
Department of Inorganic Chemistry and Catalysis, Debye Institute, Utrecht UniVersity, Sorbonnelaan 16,
3
584 CA Utrecht, The Netherlands, and Material Engineering DiVision, Toyota Motor Engineering &
Manufacturing Europe, Technical Centre, Hoge Wei 33B, B-1930 ZaVentem, Belgium
Received November 16, 2004; E-mail: B.M.Weckhuysen@chem.uu.nl
Supported noble metal nanoclusters find widespread applications
in heterogeneous catalysis as their catalytic activity can be altered
by changing the support composition and architecture or by adding
Supported Pt particles (1 wt %) were prepared via a dry
impregnation step of the support oxide materials with the appropri-
ate aqueous solutions of Pt(NH ) (NO ) . After impregnation and
3
4
3 2
1
the appropriate promoting elements. These effects are related to
drying at 353 K in N for 12 h, calcination was carried out by
2
changes in the electronic properties of the noble metal, as measured
with IR spectroscopy after, for example, CO adsorption. It is
generally accepted that the ratio of linear-to-bridged metal-
coordinated CtO reflects the electronic properties of the adsorbing
noble metal nanocluster, and this ratio increases with increasing
ionization potential of the metal particle.2
drying in a high air flow for 12 h at 393 K followed by increasing
the temperature to 573 K. Reduction was performed in pure H at
2
573 K for 2 h. After reduction and flushing with N at room
2
temperature, the samples were passivated by admitting a small
amount of air. The obtained materials were characterized in detail
with XRF, EXAFS, and TEM, and well-defined supported Pt
nanoclusters were observed on each of the support oxides (Sup-
porting Information). It was found that the microporous supports
contain mainly 1 nm Pt clusters, whereas the main fraction of Pt
clusters in the mesoporous and macroporous supports has dimen-
sions around 1.5-2 nm.
Another attractive, but almost unexplored, technique for probing
the electronic structure of supported noble metal nanoclusters is
atomic X-ray absorption fine structure spectroscopy (AXAFS).
3
Holland et al. first recognized this feature, and the groups of Rehr,
O’Grady, Baberschke, Ramaker, and Koningsberger did its further
4
development. Whereas extended X-ray absorption fine structure
The CtO adsorption IR measurements were performed on self-
spectroscopy (EXAFS) is known to originate from the scattering
of the outgoing electron against the potential of neighboring atoms,
AXAFS represents the scattering against the potential of the electron
cloud of the absorber atom itself. The embedded potential of probed
atoms is dependent on the chemical and electronic environment of
the atoms and can be influenced by the support characteristics. The
intensity and position of the AXAFS peak is a function of the
bonding of the absorbing atom with its environment. Therefore,
any change in the support oxide altering the embedded potential
of the absorbing noble atom will be reflected in its AXAFS
spectrum.
The introduction of AXAFS as a powerful new tool for studying
heterogeneous catalysts is hampered by the lack of sufficiently broad
experimental data to support the relation between the AXAFS
intensity of catalytic systems and the corresponding changes in the
electronic properties. Here, we show for a wide set of different
support oxides that AXAFS accurately probes the electronic
properties of supported Pt nanoclusters. A new tool with the
potential to probe the electronic changes in metal catalysts under
reaction conditions is explored.
supported catalyst wafers in a transmission cell. After heating in
vacuum and reduction in H at 573 K, the catalyst was exposed to
2
a static pressure of 20 mbar CO, followed by evacuation at room
temperature. All IR measurements were duplicated in order to
ensure the reproducibility of the obtained results. Figure 1 shows
some representative Pt CtO IR spectra for the samples Pt/K-Y,
Pt/Ca-Y, Pt/SiO
2
, and Pt/MCM-41. The IR spectra are characterized
-1
by a strong IR absorption band at around 2050 cm and a weaker,
broad band at around 1800 cm-1. The former band is assigned to
the stretching vibration of a linearly (L) Pt-coordinated CtO,
whereas the latter band is due to bridged (B) Pt-CtO stretching
5
vibrations. It is clear that the IR L:B intensity ratio obtained by
integrating the spectra decreases in the order of Pt/MCM-41 (14)
>
2
Pt/SiO (10) > Pt/Ca-Y (4.5) > Pt/K-Y (2.7), reflecting an
increasing electron density on the supported Pt nanoclusters. The
assumption made is that the extinction coefficients for adsorbed
CtO are influenced to the same extent for the linearly and bridged
Pt-coordinated CtO stretching vibrations, and the error in the ratio
is estimated to be around 10%.
Pt L
wafers in a transmission cell, identical to that used for the IR
measurements with the exception that the CaF windows have been
3
edge XAFS data were collected on self-supported catalyst
For this purpose, the AXAFS intensities of 14 different supported
Pt catalysts [ranging from microporous (H-USY and zeolite Y
+
+
+
+
2+
2+
2+
2+
2
exchanged with H , Na , K , Rb , Mg , Ca , Sr , and Ba
)
replaced by Be windows. After measurement, a precise background
subtraction was performed, one which optimizes the AXAFS and
EXAFS contributions in the XAFS data, while leaving the double-
electron excitations mostly in the background (Supporting Informa-
over mesoporous (SBA-15 and MCM-41) to macroporous (SiO
and SiO loaded with Cs and Ba ) support oxides] are compared
2
2
+
2+
with the corresponding linear-to-bridged Pt-coordinated CtO ratios
as obtained with IR spectroscopy on the same set of samples after
adsorption of CtO at room temperature in the same spectroscopic
in situ cell.
6
tion). Details on this procedure can be found elsewhere. The
AXAFS contribution is then isolated from the total XAFS data by
subtracting the EXAFS contributions from the experimental XAFS
data. The error for the isolation of the AXAFS contribution is
†
Utrecht University.
Toyota Motor Engineering & Manufacturing Europe, Technical Centre.
‡
7
estimated to be around 10%. The AXAFS spectra of Pt/K-Y, Pt/
3272
9
J. AM. CHEM. SOC. 2005, 127, 3272-3273
10.1021/ja043107l CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
Figure 3. AXAFS peak intensity of supported Pt nanoclusters as a function
of the corresponding IR L:B intensity ratio (for error discussion, see text).
Figure 1. IR spectra of CtO adsorbed on (a) Pt/K-Y, (b) Pt/Ca-Y, (c)
Pt/SiO2, and (d) Pt/MCM-41.
in principle, of clusters of any atom amenable to the XAFS
technique. In addition, the obtained information is fully consistent
with CO IR measurements. More importantly, however, is that
AXAFS does not need any probe molecule. As a consequence,
AXAFS can be used in the future to probe the electronic properties
of supported noble metal nanoparticles under reaction conditions
in real time, delivering mechanistic insight on the working catalyst.
Acknowledgment. This work has been funded by Toyota Motor
Europe and a NWO-CW VICI grant. We kindly thank fruitful
discussions with people from Toyota Motor Corporation and Toyota
Central R&D Labs, Japan. The authors also acknowledge beamtime
grants from the DUBBLE Grenoble and X1.1 Hasylab beamline
stations.
Supporting Information Available: Experimental details, an
EXAFS analysis sample, and full analysis results (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
1
Figure 2. Fourier transforms (k , ∆k ) 2.5-8 Å-1) of the AXAFS spectra
of (a) Pt/MCM-41, (b) Pt/SiO2, (c) Pt/Ca-Y, and (d) Pt/K-Y.
References
(
1) (a) Linic, S.; Barteau, M. A. J. Am. Chem. Soc. 2004, 126, 8086-8087.
(b) Bell, A. T. Science 2003, 299, 1688-1691. (c) Wei, J. M.; Iglesia, E.
Angew. Chem., Int. Ed. 2004, 43, 3685-3688. (d) Carrettin, S.; Concep-
cion, P.; Corma, A.; Nieto, J. M. L.; Puntes, V. F. Angew. Chem., Int.
Ed. 2004, 43, 2538-2540. (e) Chen, M. S.; Goodman, D. W. Science
2
Mg-Y, Pt/SiO , and Pt/MCM-41 in the R-range from 0 to 1.5 Å
are given in Figure 2. One can notice that the AXAFS data of the
different supported Pt nanoclusters differ in their AXAFS intensity,
as well in the peak centroid. A decrease in the AXAFS intensity
results in a shift of the peak centroid to higher R values, and the
AXAFS intensity decrease follows the same order as that of the
2
004, 306, 252-255. (f) Guzman, J.; Gates, B. C. J. Am. Chem. Soc.
004, 126, 2672-2673. (g) Derrouiche, S.; Grevejat, P.; Bianchi, D. J.
2
Am. Chem. Soc. 2004, 126, 13010-13015.
(2) (a) Mojet, B. L.; Miller, J. T.; Koningsberger, D. C. J. Phys. Chem. B
1
999, 103, 2724-2734. (b) van Santen, R. A. J. Chem. Soc., Faraday
-
2
-2
IR L:B intensity ratios, that is, Pt/MCM-41 (2.2 × 10
Å
) >
Trans. 1 1987, 83, 1915-1934. (c) Crowell, J. E.; Garfunkel, E. L.;
Somorjai, G. A. Surf. Sci. 1982, 121, 303-320. (d) Sheu, L. L.; Karpinski,
Z.; Sachtler, W. M. H. J. Phys. Chem. 1989, 93, 4890-4894.
-
2
-2
-2 -2
Pt/SiO
0.95 × 10
The above-described procedure has been applied to the 14
2
(1.6 × 10
Å
) > Pt/Ca-Y (1.25 × 10
Å ) > Pt/K-Y
-
2
-2
(
Å
).
(3) Holland, B. W.; Pendry, J. B.; Pettifer, R. F.; Bordas, J. J. Phys. C 1978,
11, 633-642.
(4) (a) Rehr, J. J.; Booth, C. H.; Bridges, F.; Zabinsky, S. I. Phys. ReV. B
different supported Pt catalysts under study. Figure 3 plots the
AXAFS peak intensities of the different samples versus their
corresponding IR L:B ratios. The IR L:B ratios span a range of
more than 12 units, whereas the AXAFS intensities change between
1
994, 49, 12347-12350. (b) Wende, H.; Baberschke, K. J. Electron.
Spectrosc. Relat. Phenom. 1999, 101-103, 821-826. (c) Ramaker, D.
E.; Mojet, B. L.; O’Grady, W. E.; Koningsberger, D. C. J. Phys Condens.
Mater. 1998, 10, 8753-8770.
(5) (a) Hippe, C.; Lamber, R.; Schulz-Ekloff, G.; Schubert, U. Catal. Lett.
1
997, 43, 195-199. (b) Sheppard, N.; Nguyen, T. T. AdV. Infrared Raman
-
2
-2 -2
0
.8 × 10 and 2.6 × 10 Å . It is clear that the AXAFS intensity
Spectrosc. 1978, 5, 67-148. (c) Bourane, A.; Dulaurent, O.; Bianchi, D.
J. Catal. 2000, 196, 115-125.
decreases with decreasing IR L:B ratio, and both the AXAFS peak
intensity and the IR L:B ratio can be considered as complementary
measures for the electron charge of the Pt nanoclusters. In other
words, the AXAFS peak intensity and the IR L:B decrease with
decreasing ionization potential of the Pt atoms in the nanoclusters.
In conclusion, AXAFS is a very attractive technique for probing
the electronic properties of supported noble metal nanoclusters and,
(
6) (a) Ramaker, D. E.; van Dorssen, G. E.; Mojet, B. L.; Koningsberger, D.
C. Top. Catal. 2000, 10, 157-165. (b) Koningsberger, D. C.; de Graaf,
J.; Mojet, B. L.; Ramaker, D. E.; Miller, J. T. Appl. Catal. A: General
2000, 191, 205-220.
(7) Tromp, M.; van Bokhoven, J. A.; Slagt, M. Q.; Klein Gebbink, J. M.;
van Koten, G.; Ramaker, D. E.; Koningsberger, D. C. J. Am. Chem. Soc.
2004, 126, 4090-4091.
JA043107L
J. AM. CHEM. SOC.
9
VOL. 127, NO. 10, 2005 3273