Reaction-driven surface restructuring and selectivity control in allylic alcohol catalytic aerobic oxidation over Pd

Article abstract of DOI:10.1021/ja200684f

Synchronous, time-resolved DRIFTS/MS/XAS cycling studies of the vapor-phase selective aerobic oxidation of crotyl alcohol over nanoparticulate Pd have revealed surface oxide as the desired catalytically active phase, with dynamic, reaction-induced Pd redox processes controlling selective versus combustion pathways.

Full text of DOI:10.1021/ja200684f

COMMUNICATION
pubs.acs.org/JACS
Reaction-Driven Surface Restructuring and Selectivity Control in
Allylic Alcohol Catalytic Aerobic Oxidation over Pd
,
†
†
‡
§
†
Adam F. Lee,* Christine V. Ellis, James N. Naughton, Mark A. Newton, Christopher M. A. Parlett, and
†
Karen Wilson
†
Cardiff Catalysis Institute, School of Chemistry, Cardiff University, Cardiff CF10 3AT, U.K.
‡
Department of Physics, University of York, York YO10 5DD, U.K.
§
European Synchrotron Radiation Facility, 6 Rue Jules Horowitz, BP-220, F-38043 Grenoble, France
S Supporting Information
b
ABSTRACT: Synchronous, time-resolved DRIFTS/MS/
XAS cycling studies of the vapor-phase selective aerobic
oxidation of crotyl alcohol over nanoparticulate Pd have
revealed surface oxide as the desired catalytically active
phase, with dynamic, reaction-induced Pd redox processes
controlling selective versus combustion pathways.
here has been much recent interest in the development of
heterogeneously catalyzed, aerobic selective oxidation
T
Figure 1. Synchronous DRIFTS/MS/XAS measurements were recorded
(
selox) routes to transform diverse primary alcohols to
in a stainless steel reactor with vitreous carbon windows that permit high
21
1
ꢀ3
X-ray transmission with minimum interference to the XAS signal. Heated
inlet and outlet gas lines prevent condensation of less-volatile components.
aldehydes.
Allylic aldehydes are valuable intermediates for
the fine chemical, pharmaceutical, and agrochemical sectors. For
example, crotyl alcohol (CrOH) is an important agrochemical and
precursor to the food preservative sorbic acid, while cinnamaldehyde
confers a cinnamon aroma. The commercial synthesis of allylic
aldehydes often proceeds via oxidation of their alcohol analogues by
hazardous stoichiometric oxidants. In addition to safety considera-
tions, such current technologies are also atom-inefficient because of
poor selectivity and/or additional separation and waste treatment
accurate control over the oxygen versus alcohol concentrations during
the reaction. Similar provisos apply to complementary ATR-IR
17
measurements, wherein reactively formed CO and benzoate accu-
mulated at the surface of Pd catalysts under reducing environments.
15
Pre- and in situ reduction protocols both suppress crotyl and
cinnamyl alcohol selox over PGMs. Time-resolved X-ray photoelec-
18
19
tron spectroscopy (XPS) and thermal desorption studies over
model Pd catalysts showed that metallic (111) facets promote
crotonaldehyde decarbonylation and implied an important role for
coadsorbed oxygen in ameliorating such self-poisoning. Indeed, we
steps to isolate the product and therefore economically disad-
4
vantageous. Solid catalysts able to utilize O can circumvent these
2
5
limitations to afford alternative continuous aerobic selox processes.
The most promising of such heterogeneous catalysts derive from
2þ
6
7,8
9
recently postulated a direct relationship between Pd surface (and
nanocrystalline Au and Pd and bimetallic formulations thereof,
9,11
10
bulk) concentrations determined ex situ and the associated turnover
typically dispersed on high-area carbon or oxide supports.
8
frequency toward crotyl, benzyl, and cinnamyl alcohol selox.
Despite these advances, future catalyst design and optimization is
hampered by our limited understanding of the active site responsible
Here we utilized the powerful combination of energy-disper-
sive X-ray absorption spectroscopy (XAS), diffuse-reflectance
Fourier transform IR spectroscopy (DRIFTS), and online mass
12,13
for the (rate-limiting) oxidative dehydrogenation step
and its
associated response to reaction conditions.
20
spectrometry (MS) in a dynamic in situ study of the vapor-
Early mechanistic studies aimed at elucidating the active site in
alcohol selox by platinum-group metals (PGMs) were compro-
phase selox of CrOH (CH CHdCHꢀCH OH) over diverse
3
2
1
nanoparticulate Pd catalysts that are active toward this reaction in
mised by the use of corrosive electrolytes and/or strongly adsor-
8
the liquid phase. Crucially, by working in the vapor phase, we
bing surfactants, which are now known to influence the oxidation
14
were able to minimize artifacts arising from diffusion limitations
or solvent effects common in liquid-phase selox and interrogate
the nature of adsorbates present on the surface of the catalyst
state and morphology. Subsequent in situ, liquid-phase extended
X-ray absorption fine structure (EXAFS) spectroscopy investiga-
tions identified important on-stream changes in the oxidation state
of supported Pd clusters during alcohol oxidations. In the first
dynamic quick EXAFS study of a solid catalyst during a liquid-
(
DRIFTS) and any associated restructuring (XAS) while mon-
itoring the global reactivity via the gaseous exhaust stream
composition with 1 Hz time resolution. Adoption of a transient
approach enabled us to delineate quantitative structureꢀfunction
1
5
phase reaction, we reported a correlation between PdO content
and crotyl and cinnamyl alcohol selox rates, while the converse was
1
6
noted for benzyl alcohol selox. Unfortunately, the complexities
of these three-phase studies make it impossible to eliminate mass-
transport limitations or competitive solvent adsorption and to achieve
Received: January 24, 2011
Published: March 25, 2011
r 2011 American Chemical Society
5724
dx.doi.org/10.1021/ja200684f
|J. Am. Chem. Soc. 2011, 133, 5724–5727

Journal of the American Chemical Society
COMMUNICATION
2 2 3
Figure 2. DRIFTS/MS/XANES intensities as a function of alternating CrOH/O cycles over (a) 2.37 wt % Pd/mesoporous Al O , (b) 2.5 wt % Pd/
1
ꢀ
SBA-15, and (c) Pd nanoicosahedra/mesoporous Al O at 80 °C: gray b, 1712 cm νCdO; blue line, H O (m/z 18); red line, O (m/z 32); O,
2
3
2
2
2
þ
crotonaldehyde (m/z 70); blue b, fitted XANES Pd concentration.
relations on a dispersed catalyst with a precision never previously
achieved via conventional steady-state liquid- or vapor-phase
measurements for any heterogeneously catalyzed organic trans-
formation. We have unequivocally demonstrated that low tem-
peratures stabilize partially oxidized Pd clusters that are selective
for the production and release of reactively formed crotonalde-
hyde, while >150 °C catalyst reduction promotes combustion.
Figure 1 illustrates the in situ reaction cell developed for
and propene (omitted for clarity). The temporal evolution of all four
products closely mirrored that of the oxygen reactant for each
catalyst, with a return to gas-phase CrOH almost instantaneously
suppressing their liberation from the catalyst surface.
The synchronous DRIFTS spectra show that exposure of the
impregnated and sol-immobilized catalysts to gas-phase CrOH
immediately induced surface vibrational modes characteristic
of both the alcohol reactant and crotonaldehyde (Figure S13).
The concentration of reactively formed crotonaldehyde on the
21
simultaneous energy-dispersed DRIFTS/MS/XAS measurements.
ꢀ
1
Alternating cycles of CrOH (in He) or O (in He) were passed over
catalyst surfaces was monitored by its associated 1712 cm band
(νCdO). A striking antiphase pattern between crotonaldehyde
production and surface confinement during CrOH cycles and
subsequent liberation of this desired selox product into the gas
2
three distinct as-prepared catalysts while maintaining an isothermal
bed temperature: 2.37 wt % Pd/mesoporous Al O and 2.2 wt % Pd/
2
3
SBA-15 prepared via incipient wetness impregnation and and a 5 wt
poly(vinylpyrrolidone) (PVP)-stabilized Pd nanoicosahedra sol
immobilized on mesoporous Al O . These experiments were re-
%
phase upon switching to an O stream is apparent. This observation
2
is precisely in accordance with our recent single-crystal investiga-
tions of CrOH selox, wherein coadsorbed oxygen adatoms pro-
moted crotonaldehyde desorption from Pd surfaces, thereby
2
3
peated between 80 and 250 °C (CrOH could not be maintained in
the vapor phase at lower temperatures).
The resulting low-temperature catalyst behavior is shown in
Figure 2, which compares X-ray absorption near-edge structure
19
suppressing secondary reactions. There was no evidence of maleic
anhydride or crotonic acid under these conditions. We should also
ꢀ
1
(
XANES) analysis of the time-dependent Pd K-edge XAS spectra
note the presence of weak features at 1915 and 2035 cm for the
2.37 wt % Pd/mesoporous Al O sample, which we assign to
with the principal surface and evolved gas-phase products detected
2
3
by DRIFTS and MS, respectively. Only the integrated O reactant
MS signal (m/z 32) is shown; the alternating CrOH reactant (m/z
molecular CO bound at metallic Pd bridge and atop sites, respec-
tively. The extinction coefficient of linearly bound CO is orders of
magnitude greater than the carbonyl stretch in propionaldehyde or
νCdC of allyl alcohol (4.95 vs 0.4 vs 0.026 mL mg mm ), so
this CO coverage represents only a few percent of the adsorbed
2
7
2) has been omitted for clarity. Rapid and reproducible switching
ꢀ
1
ꢀ1 23
of the reactant feedstream was achieved, with adsorbate and
effluent concentrations equilibrating in under 12 s.
The first important observation is that the palladium oxidation
state remained unperturbed throughout the cycling experiments
irrespective of the Pd catalyst, fitting to partially oxidized phases
oxygenate. Surface PdꢀH could not be detected, but requires H
2
24
pressures >9 Torr, exceeding that generated in situ from CrOH. It
is important to note the qualitatively similar behavior of our three
catalysts, independent of the support or mode of Pd nanoparticle
genesis, evidencing a common reaction mechanism and active site.
The system behavior changed dramatically above 180 °C, with
both the catalyst structure and reaction products evidencing
different surfaceꢀadsorbate chemistry. XANES showed the Pd
oxidation state for all three catalysts to exhibit a dynamic response
(
containing ∼20ꢀ40% total PdO) characteristic of the fresh
catalysts. There was also good agreement between the correspond-
ing fitted EXAFS spectra (Figure S8 in the Supporting Information),
evidencing high stability under the cycled redox environments at
8
0 °C. It is important to note that while XAS is an averaging
22
technique, Pd oxidation is a surface reaction rate-limited process,
2
þ
so oxide features are dominated by the selvage of our nanoparticles.
MS data show that the only reactively formed products desorbed
from all three catalyst surfaces at any temperature were crotonalde-
to the external reactant environment, falling to ∼10ꢀ27% Pd
under the reducing CrOH/He stream and rapidly reoxidizing to
2
þ
∼28ꢀ50% Pd upon switching to the O /He stream (Figure 3a
2
hyde and water, CO/propene, and CO , indicative of oxidative
and Figure S5a,b). Normalized spectra are shown in Figure S6 for
the 2.37 wt % Pd/mesoporous Al O catalyst, with the associated
2
dehydrogenation, decarbonylation, and combustion pathways, re-
spectively. Product desorption was negligible under CrOH cycles at
2
3
EXAFS fits in Table S1 and Figure S8. Although the magnitude of
these oxidation-state changes depended on the Pd catalyst, the
time/reactant dependence was invariant, highlighting a phenom-
enon general to Pd. This reactant-induced surface restructuring was
8
0 °C, but the switchover to gas-phase O triggered the rapid release
2
of water and crotonaldehyde (Figure 2, m/z 18 and 70), precisely as
expected from oxidative dehydrogenation, accompanied by trace CO
5
725
dx.doi.org/10.1021/ja200684f |J. Am. Chem. Soc. 2011, 133, 5724–5727

Journal of the American Chemical Society
COMMUNICATION
Figure 3. (a) MS/XANES intensities during alternating CrOH/O cycles over the 2.37 wt % Pd/mesoporous Al O catalyst at 250 °C: red line, O (m/z
2
2
3
2
2þ
3
2); 2, CO
2
(m/z 44); blueb, fitted XANES Pd concentration. (b) Evolution of CO
2
yieldandPdoxidation state duringCrOH/HefO
2
/He switchover
2 3
of the reactant stream over the 2.37 wt % Pd/mesoporous Al O catalyst at 250 °C. Combustion occurs before reoxidation of metal nanoparticles.
fully reversible and accompanied by a switchover from selective to
total oxidation with increasing temperature (see below). While the
latter is anticipated on simple thermodynamic grounds, the tem-
poral evolution of products revealed unexpected features of this
high-temperature chemistry. The transition from an oxidizing to a
reducing environment (and consequently a PdO to Pd catalyst)
x
occurred in less than 2 s as gas-phase O mixed with CrOH before
2
being swept from the system. Only weak, coincident CO pulses
2
were observed, indicating that PdO surfaces are almost inactive for
x
CrOH combustion. Lattice/surface oxygen extracted from the
catalysts during this PdO f Pd transformation was predominantly
x
consumed through oxidative dehydrogenation (see the next para-
graph). In contrast, subsequent exposure of reduced (hydrocarbon-
saturated) Pd surfaces to gas-phase O produced an instantaneous,
2
intense CO burst (Figure 3b), evidencing accumulation of
2
significant carbonaceous CH residues over metallic Pd nanopar-
x
2
Figure 4. DRIFTS/MS intensities during alternating CrOH/O cycles
ticles. Combustion was followed by slower Pd reoxidation, the
latter continuing throughout the oxidizing cycle. Nanoparticle sizes
were temperature-invariant for all three catalysts (Table S1 and
Figures S7 and S9), evidencing negligible sintering.
over the 2.37 wt % Pd/mesoporous Al O catalyst at 250 °C showing
2
3
higher crotonaldehyde (and crotonic acid) formation over PdO than over
x
Pd:blueline,H O(m/z 18);blackline,C H (m/z41);redline, CO(m/z
2
3 6
ꢀ1
28); O, crotonaldehyde (m/z 70); gray b, crotonic acid 1720 cm
.
The question of the fate of oxygen lost from the catalysts during the
O /He f CrOH/He switchover arises. Figure 4 shows correspond-
ing reaction profiles for the minor selective and partial oxidation
Transmission electron microscopy (TEM) and X-ray diffraction
provided no evidence of the former. If the latter scenario were true,
then the degree of oxidation should be proportional to the surface:
bulk Pd atom ratio (dispersion) of particles, with formation of a two-
2
products over the 2.37 wt % Pd/mesoporous Al O catalyst (Figure
2
3
S5b,c reveals similar behavior for the other catalysts). It is evident that
in the absence of gas-phase O , PdO is active for the oxidative
dimensional PdO capping layer passivating further oxidation of the
2
x
x
dehydrogenation of CrOH to crotonaldehyde via rapid abstraction of
lattice/surface oxygen adatoms. However, only a fraction of this
reactively formed aldehyde desorbs intact, since this selox step is
accompanied by the genesis of metallic Pd sites that in turn catalyze
decarbonylation and concomitant evolution of CO and C H into the
gas phase. The relative selox performance of Pd versus PdOx
nanoparticles can be directly estimated from the integrated crotonal-
dehyde desorption signals during their respective reducing/oxidizing
cycles; 66% more gas-phase crotonaldehyde is liberated from nano-
underlying Pd. We estimated the Pd dispersion of fresh nanoparticles
in our 2.37 wt % Pd/mesoporous Al O catalyst to be 47% by CO
2
3
chemisorption and TEM, in accordance with geometric predictions
for an unrelaxed 2.3 nm cuboctahedral particle (204 surface vs 429
bulk atoms). This value is indeed close to the 44% maximum PdO
observed in Figure 3a. In-situ XPS (Figures S10ꢀS12) confirmed
3
6
19
reduction of surface PdO upon CrOH adsorption at 250 °C.
x
The temperature-dependent transition from selective CrOH
oxidation to combustion was quantified for the 2.37 wt % Pd/
mesoporous Al O catalyst (Figure 5). At low conversion (low T),
particulate PdO than from the metal.
x
2 3
There are two limiting scenarios to explain the partial catalyst
reduction/oxidation observed at high temperature: either the as-
prepared catalysts contained inhomogeneous distributions of discrete
small oxidic and large metallic Pd nanoparticles and the reactant
feedstock influenced their relative populations, or the observed redox
behavior reflects the reversible genesis of oxide shells encapsulating
metal cores uniformly across every nanoparticle in each catalyst.
crotonaldehyde was the dominant product via oxidative dehydro-
genation (the predicted oxygen consumption for this sole pathway
is indicated in green). Above 150 °C, alcohol and oxygen conver-
sion rose rapidly, the latter reflecting the emergence of total
oxidation pathways, which in turn dominated by 250 °C (the
predicted O consumption for combustion is indicated in red).
2
Together, these experiments provide direct evidence of a generic
5
726
dx.doi.org/10.1021/ja200684f |J. Am. Chem. Soc. 2011, 133, 5724–5727

Journal of the American Chemical Society
COMMUNICATION
combined DRIFTS/MS/XAS measurements for probing complex
reaction networks also paves the way for related vapor-phase fine-
chemicals catalysis studies within this milieu.
’
ASSOCIATED CONTENT
S
Supporting Information. Catalyst synthesis, characteri-
b
zation and reaction conditions. This material is available free of
charge via the Internet at http://pubs.acs.org.
’
AUTHOR INFORMATION
Corresponding Author
leeaf@cardiff.ac.uk
’
ACKNOWLEDGMENT
We thank the EPSRC (EP/E046754/1, EP/G007594/2) for finan-
cial support, a Leadership Fellowship (A.F.L.), and studentship
support (C.V.E., J.N.N., and C.M.A.P.) and the ESRF for
beamtime (CH2432). A.F.L. thanks Miss Carly J. Broderick for
invaluable support.
Figure 5. Switchover from selective to partial CrOH oxidation to combus-
tion with increasing temperature (conversion). Predicted relationships
2
between CrOH and O conversion for selective oxidative dehydrogenation
(green b) and total combustion (red b) are also shown.
’
REFERENCES
Scheme 1. Proposed Reaction Network Interrelating Nano-
particulate Catalyst Phase (Oxidation State) and Associated
Selectivity toward CrOH Oxidation upon the Surrounding
Reducing or Oxidizing Reactant Feedstream
(
(
1) Mallat, T.; Baiker, A. Chem. Rev. 2004, 104, 3037.
2) Meenakshisundaram, S.; Nowicka, E.; Miedziak, P. J.; Brett,
G. L.; Jenkins, R. L.; Dimitratos, N.; Taylor, S. H.; Knight, D. W.; Bethell,
D.; Hutchings, G. J. Faraday Discuss. 2010, 145, 341.
(3) Vinod, C. P.; Wilson, K.; Lee, A. F. J. Chem. Technol. Biotechnol.
2
011, 86, 161.
(
(
4) ten Brink, G. J.; Arends, I.; Sheldon, R. A. Science 2000, 287, 1636.
5) Sheldon, R. A.; Arends, I.; Hanefeld, U. In Green Chemistry and
Catalysis; Wiley-VCH: Weinheim, Germany, 2007.
6) Della Pina, C.; Falletta, E.; Prati, L.; Rossi, M. Chem. Soc. Rev.
008, 37, 2077.
7) Mori, K.; Hara, T.; Mizugaki, T.; Ebitani, K.; Kaneda, K. J. Am.
Chem. Soc. 2004, 126, 10657.
8) Hackett, S. E. J.; Brydson, R. M.; Gass, M. H.; Harvey, I.; Newman,
A. D.; Wilson, K.; Lee, A. F. Angew. Chem., Int. Ed. 2007, 46, 8593.
9) Enache, D. I.; Edwards, J. K.; Landon, P.; Solsona-Espriu, B.;
(
2
(
(
(
Carley, A. F.; Herzing, A. A.; Watanabe, M.; Kiely, C. J.; Knight, D. W.;
Hutchings, G. J. Science 2006, 311, 362.
(
10) Bianchi, C. L.; Biella, S.; Gervasini, A.; Prati, L.; Rossi, M. Catal.
Lett. 2003, 85, 91.
11) Li, C. L.; Zhang, Q. H.; Wang, Y.; Wan, H. L. Catal. Lett. 2008,
20, 126.
12) Lee, A. F.; Hackett, S. F. J.; Hargreaves, J. S. J.; Wilson, K. Green
Chem. 2006, 8, 549.
(13) Grunwaldt, J. D.; Caravati, M.; Baiker, A. J. Phys. Chem. B 2006,
110, 25586.
(
1
(
mechanism for Pd-catalyzed selox wherein metallic Pd nanopar-
ticles favor CrOH decarbonylation and combustion pathways
2þ
while Pd favors oxidative dehydrogenation to crotonaldehyde
at low temperature, with oxygen insertion to form surface-bound
crotonic acid competing at high temperature. The interdepen-
dence of the reactant environment, catalyst oxidation state, and
surface chemistry above 150 °C is shown in Scheme 1.
We have demonstrated the first application of time-resolved
synchrotron X-ray absorption spectroscopy to simultaneously
follow dynamic nanoparticle surface restructuring and the evolu-
tion of surface and gas-phase products during an organic reaction.
(14) Lee, A. F.; Wilson, K.; Lambert, R. M.; Hubbard, C. P.; Hurley,
R. G.; McCabe, R. W.; Gandhi, H. S. J. Catal. 1999, 184, 491.
(15) Lee, A. F.; Wilson, K. Green Chem. 2004, 6, 37.
(16) Keresszegi, C.; Ferri, D.; Mallat, T.; Baiker, A. J. Phys. Chem. B
2
005, 109, 958.
17) Ferri, D.; Mondelli, C.; Krumeich, F.; Baiker, A. J. Phys. Chem. B
006, 110, 22982.
18) Lee, A. F.; Chang, Z.; Ellis, P.; Hackett, S. F. J.; Wilson, K.
(
2
(
J. Phys. Chem. C 2007, 111, 18844.
Surface PdO , not Pd, has been identified as the catalytic species
(19) Naughton, J.; Lee, A. F.; Thompson, S.; Vinod, C. P.; Wilson, K.
Phys. Chem. Chem. Phys. 2010, 12, 2670.
x
responsible for the selective oxidation of CrOH to croto-
naldehyde. Elevated reaction temperatures facilitate reversible
nanoparticle redox processes and concomitant catalytic selectivity
loss in response to reaction conditions. These discoveries highlight
(20) Newton, M. A.;Belver-Coldeira, C.; Martinez-Arias, A.;Fernandez-
Garcia, M. Nat. Mater. 2007, 6, 528.
(
(
(
(
21) Newton, M. A. Top. Catal. 2009, 52, 1410.
22) Salmeron, M.; Schlogl, R. Surf. Sci. Rep. 2008, 63, 169.
23) Chong, T. S.; Tan, S. T.; Fan, W. Y. Chem.—Eur. J. 2006, 12, 5128.
24) Ratajczykowa, I. Surf. Sci. 1975, 48, 549.
the importance of stabilizing surface PdO and minimizing catalyst
x
reducibility in order to achieve high selox yields and will hopefully
aid the future design of Pd-derived selox catalysts. The utility of
5
727
dx.doi.org/10.1021/ja200684f |J. Am. Chem. Soc. 2011, 133, 5724–5727