Susceptibility of a heterogeneous catalyst, Rh/γ-alumina, to rapid structural change by exposure to NO

Article abstract of DOI:10.1039/b106846f

Metal particles in a Rh/γ-Al₂O₃ catalyst of differing particle size are oxidised by NO/He within 5 seconds at 313 K; rapid, highly exothermic dissociative chemisorption of NO is the initial step.

Full text of DOI:10.1039/b106846f

Susceptibility of a heterogeneous catalyst, Rh/g-alumina, to rapid
structural change by exposure to NO†
Tom Campbell,a‡ Andrew J. Dent,b Sofia Diaz-Moreno,c John Evans,*a Steven G. Fiddy,a Mark A.
Newton*a and Sandra Turina
a Department of Chemistry, University of Southampton, Southampton, UK SO17 1BJ.
E-mail: je@soton.ac.uk. E-mail: M.A.Newton@soton.ac.uk
b CLRC Daresbury, Warrington, Cheshire, UK WA4 4AD
c
The European Synchrotron Radiation Facility (ESRF), Grenoble F-38043, France
Received (in Cambridge, UK) 27th July 2001, Accepted 10th December 2001
First published as an Advance Article on the web 23rd January 2002
Metal particles in a Rh/g-Al2O3 catalyst of differing particle
size are oxidised by NO/He within 5 seconds at 313 K; rapid,
highly exothermic dissociative chemisorption of NO is the
initial step.
situ in a flow microreactor to 573 K under 4% H2/He (a), and (b)
that derived after exposure to 4% NO/He at 313 K for 5 seconds.
Considerable changes are evident even after a short exposure.
Data analysis indicates an initial state with a Rh site with ~ 8 Rh
neighbours (Rh–Rh 2.68 Å) and ~ 3 non-nearest neighbours
(Rh…Rh 3.75 Å). After exposure for 5 seconds, three shells are
apparent: ~ 1 N at 1.78 Å, ~ 2 O at 2.05 Å and ~ 2 Rh at 2.72
Å. There are parallels with the oxidised centres, which are also
produced after exposure to O2,3 the short Rh–N shell indicates
some structural differences. In addition regeneration of metal on
subsequent exposure to H2/He is considerably slower, only
The performance of rhodium-containing catalysts to facilitate
chemical processes such as CO oxidation and hydrogenation,
and also NO reduction using CO, H2, and hydrocarbons is very
dependent upon reaction conditions. We have previously
demonstrated that Energy Dispersive EXAFS (EDE) can be
used to monitor the formation of metal particles in situ, in this
case from reduction of Pt(acac)2/H1–SiO2.1 Used in combina-
tion with mass spectrometry, the available time resolution
allows a direct correlation between structural changes at the
metal centres and gas phase composition; in that example, the
reaction studied was between NO and RhCl(CO)2/g-Al2O3.2 In
this report, we apply these techniques to investigate the effects
of CO and NO, key elements of automotive exhaust catalysis,
on the local rhodium structure of a conventional supported
metal catalyst, Rh/g-Al2O3.
occurring above ca. 390 K (see supplementary Fig. S2 ESI†).3
Rh–Rh
Starting from either very small particles, N1
= 4 ( ~ 6
Rh–Rh
atoms), or larger ones, N1
= 8.1 ( 0.8) (ca. 40 < Rh atoms
< 200),4 affords a very similar product.
Fig. 2 shows the temporal variation of the white line intensity,
and the intensity of an EXAFS feature (the location of these
points is shown in the ESI†), observed to increase and diminish
respectively during the reaction. Data are shown for three cases:
Rh–Rh
Rh particles showing N1
of 4 (Tred = 313 K) during
The quartz microreactor system and the conditions of the X-
ray beamline (ID24 at the ESRF) and also the data analysis
techniques have been previously described.2 Typically acquisi-
tion of a spectrum took 300–400 ms, spectra being recorded
every 2–3 seconds. In Fig. 1 we show the Rh K-edge EDE-
derived EXAFS from a 4 wt% Rh/g-Al2O3 sample reduced in
exposure of 4% NO/He (black) and CO (blue) at 313 K; and for
Rh–Rh
Rh particles showing N1
= 8 (Tred = 573 K) during
exposure to 4% NO/He (red). It is well known that CO can
disperse very small rhodium particles (CN 5 4) over longer
timescales but it is clear that, under the conditions used here, CO
fails to elicit any change in the observed Rh EXAFS, even for
the most dispersed Rh sample.5 NO diluted in He, however,
effects rapid changes in the spectra derived from each supported
system accompanied by a transient but pronounced (12.5 ( 2.5)
K) exotherm (Fig. 3). Coincident with this temperature rise, and
Fig. 1 Rh K-edge EDE derived k3-weighted EXAFS data from a 4 wt% Rh/
Al2O3 sample: (a) after in situ reduction to 573 K in 4% H2/He and cooling
to 313 K (spectra taken under He flow of 6 ml min21), and (b) after exposure
to 4% NO/He for ca. 5 seconds. Black and red lines, respectively, indicate
experiment and theoretical fits derived from spherical wave analysis in
EXCURV98.11
Fig. 2 Temporal response of the Rh K-edge white line and an EXAFS
feature observed in the raw data 265 eV post edge (derived from 255
sequential EDE spectra (360 ms, 1 spectrum every ~ 2.5 seconds). Spectra
† Electronic supplementary information (ESI) available: data showing
monitoring positions in Fig. 2 and data showing reduction of the NO
oxidised Rh adduct on alumina. See http://www.rsc.org/suppdata/cc/b1/
b106846f/
obtained during switching of gas feed from He to 4% NO/He for pre-
Rh–Rh
reduced samples showing N1
= 4 (black) and 8 (red). The blue curves
‡ Present address: Johnson Matthey Plc, Royston, Hertfordshire, UK.
are those derived from gas switch from He to 4% CO/He (blue).
304
CHEM. COMMUN., 2002, 304–305
This journal is © The Royal Society of Chemistry 2002

the initial adsorption of NO, the evolution of mass 28 is
observed followed by mass 44 (Fig. 3). Analysis of the
evolution of minor mass fragments (12, 14, 22, 26, 32, and 42)
indicates that these features are due to N2 and N2O respectively,
with minor contributions from species such as O2, CO, CN,
CNO and CO2. The approximate levels of NO uptake may be
calculated from these experiments, and expressed in terms of
NO/Rh atoms initially present. This yields an estimate of NO
uptake of ca. 1.5 ( 0.26) NO/Rh. The sensitivity of the mass
spectrometer toward N2O (44) and N2 (28) (relative to NO)
allows us to estimate that approximately 0.1 NO/Rh are reacted
to yield the N2 observed and ca. 0.14 NO/Rh are subsequently
reacted to form N2O. The remaining 1–1.2 NO/Rh are retained
on the catalyst.
Many previous studies on similar systems (utilising vibra-
tional spectroscopies)6 have indicated the predominant forma-
tion of a linear Rh–NO+ species, although contributions from
other nitrosyls, such as Rh(NO)2 are also observed. The EXAFS
obtained from our experiments is consistent with the pre-
dominant formation of an oxidic Rh phase with a central core
comprising of two (surface) oxygen bonds and, on average, a
single nitrosyl species surrounding the Rh. The analysable
datalength and the intrinsic noise associated with the data do not
permit absolute confirmation of the nitrosyl species formed.
The rapid, exothermic reaction of Rh/g-Al2O3 with NO must
involve dissociative chemisorption to afford N2 and N2O. Since
the sample bed is essentially solid Al2O3 (Cp ~ 79 J mol21 K21
6) the energy release may be estimated from the observed
exotherms. For a typical sample, this yields a value of 0.16–0.23
J (20 mg sample, exotherm ~ 10–15 K). If this energy release is
primarily due to dissociation of NO on the Rh particles, we can
estimate that the bed temperature rise equates to a potential
microscopic heating of the supported particles of ca. 640–960 K
(assuming a Cp of bulk Rh7). Further, the level of NO
dissociation can be calculated from the N2 desorption; from this,
the dissociative heat of adsorption (DEdiss) of NO may be
estimated, as 300–400 kJ mol21. Estimates of DEdiss NO for Rh
single crystals from both experiment8 and theory9 are also of the
order of 300–400 kJ mol21
.
The extent and rapidity of the NO induced fragmentation of
these supported Rh particles are in excess of those observed for
any disruptive process previously reported. The source of these
changes lies in the balance between the dynamics of molecular
dissociation on the supported particles, and the processes
available for the dissipation of that energy thus released from
the particle into its surrounding medium. The consequences for
mechanistic studies of highly dispersed oxide-supported hetero-
geneous catalysts are considerable. Clearly, structure reactivity
relationships, like between facile (or structure insensitive) and
demanding (structure sensitive) reactions,10 based on the
surface area measurements and catalytic reactions carried out
under different conditions, may be problematical. With metal
nuclearities changing within a few seconds under mild condi-
tions, the active sites of high dispersion heterogeneous catalysts
cannot be viewed as being located on monolithic particles.
This work is dedicated to the memory of the late Dr Judith
Corker. This work was funded under the Catalysis and
Chemical Processes (CACP) initiative of the EPSRC. The
EPSRC is also thanked for postdoctoral funding to M. A. N. and
ICI for postdoctoral funding to S. G. F. We would like to thank
the ESRF for providing access to the synchrotron. The authors
are also extremely grateful to Bruce Hancock, John James,
Melanie Hill, Ralph Wiegel, and Sebastian Pasternak for their
technical expertise. Professor R Schlögl is also thanked for the
generous provision of the mass spectrometer facility available
on ID 24.
Notes and references
1 S. G. Fiddy, M. A. Newton, A. J. Dent, G. Salvini, J. M. Corker, S.
Turin, T. Campbell and J. Evans, Chem. Commun., 1999, 851; S. G.
Fiddy, M. A. Newton, T. Campbell, J. M. Corker, A. J. Dent, I. Harvey,
G. Salvini, S. Turin and J. Evans, Chem. Commun., 2001, 445.
2 M. A. Newton, D. G. Burnaby, A. J. Dent, S. Diaz-Moreno, J. Evans, S.
G. Fiddy, T. Neisius, S. Pascarelli and S. Turin, J. Phys. Chem. A, 2001,
105, 5965.
3 M. A. Newton, A. J. Dent, S. Diaz-Moreno, B. Jyoti, S. G. Fiddy and J.
Evans, manuscript in preparation.
4 A. Jentys, Phys. Chem. Chem. Phys., 1999, 1, 4059.
5 H. F. K. Van’t Blik, J. B. A. D. Van Zon, T. Huizinga, J. C. Vis, D. C.
Koningsberger and R. Prins, J. Phys. Chem., 1983, 87, 2264.
6 For instance see: T. Chafik, D. I. Kondaridis and X. E. Verykios, J.
Catal., 2000, 190, 446; D. I. Kondarides, T. Chafik and X. E. Verykios,
J. Catal., 2000, 191, 147.
7 Handbook of Chemistry and Physics, ed. D. R. Lide, CRC Press, Boca
Raton, FL, 72nd edn., 1991.
8 For instance see: Q. F. Ge, R. Kose and D. A. King, Adv. Catal., 2000,
45, 259, and references therein.
9 D. Loffreda, D. Simon and P. Sautet, J. Chem. Phys., 1998, 108,
Fig. 3 NO uptake and major gas (N2 and N2O) evolution during switch to
4% NO/He gas feed over the Rh/Al2O3 sample previously reduced in situ in
4% H2/He to 573 K along with the observed variation of sample bed
temperature (DT, inset). The switch to a NO/He feed occurs at 60 seconds,
and spectra are shown after subtraction of a null response to an identical gas
switch over the reacted sample bed. The mass and scaling factor for each
trace are indicated.
6447.
10 M. Che and C. O. Bennett, Adv. Catal., 1989, 36, 55; M. Boudart, Adv.
Catal., 1969, 20, 153.
11 EXCURV98, CCLRC Daresbury Laboratory Computer Program,
1998.
CHEM. COMMUN., 2002, 304–305
305