Synergistic Interaction between Oxides of Copper and Iron for Production of Fatty Alcohols from Fatty Acids

Article abstract of DOI:10.1021/acscatal.5b01664

The selective hydrogenation of fatty acids to fatty alcohols can be achieved under moderate conditions (180 °C, 30 bar H₂) by simultaneously supporting copper and iron oxides on mesoporous silica nanoparticles. The activity of the cosupported oxides is significantly higher than that of each supported metal oxide and of a physical mixture of both individually supported metal oxides. A strong interaction between both metal oxides is evident from dispersion, XRD, TPR, and acetic acid TPD measurements, which is likely responsible for the synergistic behavior of the catalyst. Copper oxide is reduced in situ to its metallic form and thereby activates hydrogen. It is proposed that hydrogen spills over to iron oxide where fatty acids bind and are selectively reduced to the alcohol.

Full text of DOI:10.1021/acscatal.5b01664

Letter
pubs.acs.org/acscatalysis
Synergistic Interaction between Oxides of Copper and Iron for
Production of Fatty Alcohols from Fatty Acids
‡
‡
‡
Kapil Kandel, Umesh Chaudhary, Nicholas C. Nelson, and Igor I. Slowing*
U.S. Department of Energy, Ames Laboratory, and Department of Chemistry, Iowa State University, Ames, Iowa 50011, United States
*
S Supporting Information
ABSTRACT: The selective hydrogenation of fatty acids to fatty
alcohols can be achieved under moderate conditions (180 °C,
3
0 bar H ) by simultaneously supporting copper and iron oxides
2
on mesoporous silica nanoparticles. The activity of the
cosupported oxides is significantly higher than that of each
supported metal oxide and of a physical mixture of both
individually supported metal oxides. A strong interaction
between both metal oxides is evident from dispersion, XRD,
TPR, and acetic acid TPD measurements, which is likely
responsible for the synergistic behavior of the catalyst. Copper
oxide is reduced in situ to its metallic form and thereby activates
hydrogen. It is proposed that hydrogen spills over to iron oxide
where fatty acids bind and are selectively reduced to the alcohol.
KEYWORDS: fatty alcohols, fatty acids, biorenewables, mesoporous silica, iron oxide, copper oxide
he instability in the supply of fossil fuel resources and the
increase in consumer demand are leading to a growth in
catalysts still capable of working at low H pressures and
2
moderate reaction temperatures.
T
1
,2
the market of renewable chemicals. Biorenewable feedstocks
rich in fatty acids such as microalgae, vegetable oils, or waste oil
have been recently targeted for producing commodity
chemicals such as solvents, lubricants, flavors, pharmaceuticals,
Many research groups including ours have reported the
activity of inexpensive Ni nanoparticles for the hydrogenation
of fatty acids to hydrocarbons between 250 and 290 °C and 30
15−17
to 40 bar H
pressure.
More recently, we discovered that
2
3
,4
as well as alternative fuels. Fatty alcohols are surfactant range
molecules that can be used in lubricants, plasticizers, cosmetics,
among others, with a world market valued at around $1.9
Fe nanoparticles catalyze the conversion of fatty acids into
hydrocarbons with less cracking and decarbonylation than Ni
18
under similarly moderate conditions. Fatty alcohols were
identified as the major reaction intermediates, becoming
hydrocarbons upon dehydration and further olefin hydro-
genation. Thus, we considered that it could be possible to stop
the reaction at the alcohol stage if the conditions were mild
enough to prevent dehydration of the intermediate. To test this
hypothesis, we prepared an iron oxide catalyst by impregnating
mesoporous silica nanoparticles (MSN) with aqueous Fe-
5
billion and projected growth rate of 7.3% per year.
Commercially, the production of fatty alcohols involves
methanolysis of triglycerides or fatty acids, followed by
hydrogenolysis using copper chromite catalysts at temperatures
ranging from 200 to 400 °C and H pressures of 200 to 300
2
6
−9
bar.
In this regard, the development of catalysts to produce
fatty alcohols under milder conditions has been a long-standing
industrial target. Additionally, chromium-free catalysts are
desired in order to reduce environmental pollution. Doping
Cu with promoters such as boron has allowed eliminating Cr
(
NO ) ·9H O and calcining in air at 350 °C. We tested the
3 3 2
hydrogenation of stearic acid using the iron-oxide-based catalyst
5.5 wt % Fe, based on ICP-OES measurement) at different
(
temperatures for 3 h under 30 bar H . The initial screening
and performing the reaction at lower H pressures (on the
2
2
gave poor to moderate selectivity for octadecanol over all the
temperatures tested (Figure 1a). Although full conversion was
obtained in only 3 h at 250 °C, the yield of alcohol was merely
order of 100 bar) while keeping moderate reaction temper-
atures (240 °C). Much more important reductions in reaction
1
0
pressures (down to 20 bar H ) have been achieved with high
2
2
9%, and the main product was the fully reduced alkane.
Since Cu is known to catalyze the hydrogenation of esters
conversion and selectivities by replacing Cu with noble metals
11−14
such as Ru, Rh, Re, Pd, or Pt.
Perhaps the most
19,20
and carboxylic acids to alcohols,
we prepared a second
outstanding result may be a recent report by Hardacre et al.
in which carboxylic acids were hydrogenated to alcohols using
Pt/TiO catalysts at only 130 °C and 20 bar H (over 80%
catalyst by impregnating MSN with Cu(NO ) ·2.5H O and
3 2
2
2
2
14
conversion and 90% selectivity after 12 h). Although these
results are remarkable, it would be desirable to replace the
expensive noble metals with more abundant and economic
Received: July 31, 2015
Revised: October 7, 2015
©
XXXX American Chemical Society
6719
DOI: 10.1021/acscatal.5b01664
ACS Catal. 2015, 5, 6719−6723

ACS Catalysis
Letter
calcination in air at 350 °C. The amounts of metals used for
this new material were selected to keep the total metal load
comparable to those of the single metal-MSN catalysts. Metal
loading of the resulting catalyst was 2.8 wt % Cu and 3.5 wt %
Fe based on ICP-OES measurement. Both elements appeared
to be well dispersed and similarly distributed in the material
according to TEM/STEM imaging and EDS mapping (Figure
S2). The composite catalyst displayed a significantly higher
activity than the single metal oxide catalysts, giving almost
quantitative conversion after 3 h at 180 °C and 30 bar H2
(
Figure 2). For comparison, the single metal oxide catalysts
Figure 1. Effect of temperature on the conversion of stearic acid and
the yields of 1-octadecanol and alkanes using (a) iron oxide and (b)
copper oxide supported on MSN. Reaction conditions: stearic acid 1
mM in 10 mL of hexanes, 10 mg of catalyst, 30 bar H , 3 h.
2
Figure 2. Effect of temperature on the conversion of stearic acid and
the yields of 1-octadecanol and alkanes using copper and iron oxides
supported on MSN, (2.8 wt % Cu, 3.5 w% Fe). Reaction conditions:
calcining at 350 °C (7.1 wt % Cu based on ICP-OES) to
compare with our iron-oxide-based catalyst. The catalyst gave
an identical temperature dependence profile for the conversion
of stearic acid as the original iron oxide-based MSN catalyst,
showing almost full conversion after 3 h at 250 °C and 30 bar
stearic acid 1 mM in 10 mL of hexanes, 10 mg of catalyst, 30 bar H , 3
2
h.
gave less than 3% conversion under the same reaction
conditions. Furthermore, the catalyst provided high selectivity
for the alcohol under these conditions, giving up to 95% yield
of 1-octadecanol. However, the alcohol selectivity dropped
significantly at higher temperatures with increased dehydra-
tion−hydrogenation to produce octadecane.
H . The selectivity, however, heavily favored octadecanol over
2
the hydrocarbon (82% yield at 250 °C and 92% at 230 °C,
Figure 1b). One possible reason for the difference in
selectivities between the catalysts may be their different
2
1
acidities. The less acidic copper oxide may be unable to
promote the dehydration of the alcohol, a conversion that the
more acidic iron oxide is known to catalyze under these
Increasing the relative loading of Cu in the composite (6.1 wt
% Cu, 0.6 wt % Fe) led to a catalyst with a slightly lower activity
1
7,22
reaction conditions.
(85% conversion after 3 h at 180 °C and 30 bar H ) but 100%
2
Previous work has shown that the reduction of fatty acids
catalyzed by metal oxides involves two different sites: a
hydrogen activation site corresponding to in situ formed metal
alcohol selectivity (Figure S3a). This high-loading Cu
composite preserved the alcohol selectivity at higher temper-
atures, which was consistent with the behavior of the copper
oxide MSN catalyst (Figure 1b). In contrast, a catalyst with a
much larger amount of Fe than Cu (0.9 wt % Cu, 5.5 wt % Fe)
had a lower activity under the same reaction conditions (36%
conversion) (Figure S3b) and displayed high temperature
selectivity similar to the iron oxide MSN (Figure 1a).
Comparison between different catalyst compositions suggests
that the highest activity should be found at Cu:Fe mole ratios
close to 1:1 (Figure 3). However, Cu seems to have a larger
effect on the activity of the composites, as inferred from the
significantly lower activity for the low-loading Cu catalyst than
for the low-loading Fe catalyst.
18,23,24
and a metal oxide site that binds the carboxylic groups.
Ponec and co-workers suggested the strength of the metal−
oxygen bonds was a key factor in determining the capacity of
the oxides to bind and activate acids for their reduction and
used the heat of formation per oxygen atom in the highest
23
oxide as an indicator of bond strength. They identified iron
oxide as having an optimal metal−oxygen bond strength, which
was sufficiently strong to bind carboxylates, yet weak enough to
allow the bound oxygen to react and escape as water to
regenerate the binding site. Because of the higher affinity of
25−27
carboxylic acids for the surface of iron oxide
and the easier
reduction of copper oxide to its metallic form, we considered
producing a hybrid catalyst containing both metal oxides. It was
expected that copper could be reduced to its metallic form
Temperature-programmed reduction (TPR) analysis of the
2.8 wt % Cu and 3.5 wt % Fe catalyst showed a low
temperature reduction peak with maximum at 149 °C (range
80 to 170 °C, Figure 4a, top), which was at a slightly lower
temperature than the reduction peak observed for the copper
oxide-MSN catalyst (168 °C, Figure 4a, middle). The amount
during the reaction and act as site for H binding and activation,
2
whereas iron oxide could bind and activate the carboxylic acid.
Although H chemisorption on copper has an activation barrier,
2
oxides of more electropositive metals, including iron, are known
of H consumed at this peak was 0.62 mmol/g, which is almost
2
28−32
to act as promoters for H activation on Cu.
twice as much as would be needed to reduce all CuO in the
2
0
The copper and iron oxides catalyst was prepared by
coimpregnation of the metal nitrates on MSN followed by
sample to Cu (0.33 mmol/g). XPS analysis of the material
reduced at 180 °C showed that while Cu had become metallic,
6
720
DOI: 10.1021/acscatal.5b01664
ACS Catal. 2015, 5, 6719−6723

ACS Catalysis
Letter
dispersion of the oxides and/or lack of crystallinity, which is
usual for small loadings of metals in mesoporous silicas calcined
18,35
at low temperatures.
Samples with higher loadings of the
metals on MSNs were prepared by the same procedures to
better understand the possible structure of the materials. A 27.2
wt % iron oxide-MSN sample exhibited an XRD pattern
corresponding to α-Fe O , (JCPDS No. 33-0664, Figure S5b).
2
3
Interestingly, a copper and iron oxides-MSN with a similar load
of Fe (26.8 wt %) and a small amount of Cu (3.1 wt %) did not
give well-defined patterns but broad peaks, indicating an
amorphous mixture of the oxides (Figure S5c). In contrast, a
copper and iron oxides-MSN containing 24.9 wt % Cu and 7.3
wt % Fe produced only a CuO pattern (Figure S 5d). These
results suggest a strong interaction between the oxides that
mainly limited the growth of iron oxide crystals. Copper has
been shown to promote dispersion of iron catalysts by inducing
the nucleation of a large number of small iron oxide
Figure 3. Stearic acid conversion (blue) and octadecanol yield (red) as
a function of the mole fraction of iron in copper−iron oxide catalysts
supported on MSN. Reaction conditions: stearic acid 1 mM in 10 mL
of hexanes, 10 mg of catalyst, 180 °C, 30 bar H , 3 h.
2
32,33,36
crystallites.
Treating the 24.9 wt % Cu and 7.3 wt %
Fe MSN composite with H at 180 °C gave a product with clear
2
0
Cu reflections (JCPDS No. 70-3038, Figure S5e), in good
agreement with the XPS, TPR, and dispersion results obtained
for the low load catalysts. Minor peaks were also observed that
could be assigned to small amounts of CuO and FeO. These
experiments suggest iron oxide interacts with copper oxide to
0
facilitate its reduction to Cu by H , and coexists with the metal
2
as a highly dispersed iron oxide.
Acetic acid TPD experiments of the 2.8 wt % Cu and 3.5 wt
%
Fe mixed oxides MSN material revealed a sharp decrease in
the strength of acid adsorption on the composite. The original
copper oxide-MSN and iron oxide-MSN catalysts had
desorption peaks centered approximately at 260 and 295 °C,
respectively (Figure S6), whereas the 2.8 wt % Cu and 3.5 wt %
Fe composite presented a bimodal distribution with centers
around 220 and 250 °C (Figure 4b). This TPD maxima shift
toward lower temperatures also supports the occurrence of a
significant interaction between both oxides in the material.
Treatment of the 2.8 wt % Cu and 3.5 wt % Fe composite with
Figure 4. (a) Temperature-programmed reduction of iron and copper
oxides (top), copper oxide (middle) and iron oxide (bottom)
supported on MSN. (b) Temperature-programmed desorption of
acetic acid from copper and iron oxides supported on MSN as
synthesized (blue) and after reduction at 180 °C (red).
Fe remained in the +3 oxidation state (Figure S4). In contrast
to the excess of H consumed by the copper iron oxides-MSN,
H at 180 °C changed the TPD profile to produce a single
2
2
the single metal CuO-MSN (7.1 wt %) consumed 1.02 mmol/
g, which is approximately the amount needed to reduce all the
metal in the sample to Cu (1.11 mmol/g). These results
suggest that the excess hydrogen consumed by the mixed metal
oxide-MSN in the sharp TPR peak should be adsorbed on the
material, either on the surface of the metallic Cu or spilled over
to the iron oxide. The observed decrease in reduction
temperature of Cu could be related to increased dispersion
event at approximately 300 °C (Figure 4b), which matched the
profile of iron oxide-MSN (Figure S6b). These findings suggest
that the composite consisted of an amorphous or microcrystal-
line, well-distributed mixture of the metal oxides, which upon
reduction at 180 °C segregated into metallic copper and iron
oxide. Such segregation has been previously observed during
0
37,38
0
the reduction of bulk copper−iron mixed oxides.
The Cu
and iron oxide containing material is believed to be the working
catalyst. Consistent with this hypothesis, performing the
caused by the interaction with iron oxide. N O chemisorption
2
experiments showed that the dispersion of Cu increased from
reaction with the prereduced catalyst (180 °C under H
2
8
(
% in the single oxide MSN to 67% in the mixed material
Table S1). This observation is consistent with previous reports
on the enhanced dispersion of copper oxide in the presence of
stream) gave a similar alcohol yield as the original material
(88% versus 95%).
The synergistic behavior observed in the catalytic process
was strongly dependent on the proximity between the two
species: in contrast to the 99% conversion obtained with the
3
3,34
iron oxide.
A second, much broader peak extending from
about 180 to 500 °C in the TPR of the composite (Figure 4a,
top) suggested a progressive reduction of iron oxide in the
material. This peak spanned a temperature range very similar to
the iron oxide-MSN catalyst (Figure 4a, bottom), which
suggests formation of Cu does not significantly affect the
reduction behavior of iron oxide.
Powder XRD analysis of the samples did not show any
distinct peaks, the only exception being copper oxide-MSN,
whose pattern matched that of CuO (JCPDS No. 71-3645,
Figure S5a). The lack of XRD reflections in iron oxide-MSN
and the mixed oxide materials was likely due to the high
composite catalyst after 3 h at 180 °C and 30 bar H , a physical
2
mixture of copper oxide-MSN and iron oxide-MSN gave only
8% conversion under the same reaction conditions. It must be
noticed that the conversion obtained with the physical mixture
matches well with the sum of the conversions of the individual
metal oxides-MSN under the same reaction conditions.
Therefore, the conversion using the physical mixture should
take place by a mechanism similar to the individual catalysts. To
the contrary, the catalytic process observed in the composite
material may involve cooperativity due to the proximity
0
6
721
DOI: 10.1021/acscatal.5b01664
ACS Catal. 2015, 5, 6719−6723

ACS Catalysis
Letter
between the two species. This proximity-dependent behavior
can be rationalized considering that hydrogen chemisorbed on
Cu and then spilled over to iron oxide. Hydrogen spillover can
spillover from copper to iron oxide sites is likely required for
the conversion and takes place when both sites are close to each
other. We consider that this composite catalyst can be a viable
low-pressure, chromium-free, and non-noble metal alternative
for the production of fatty alcohols from biorenewable
feedstocks. Further work aimed at exploring the structure and
role of the interface between the copper and iron oxide species
with a simplified model is currently in progress.
0
take place only on reducible oxides like iron oxide, but not over
3
9,40
silica.
Thus, when copper and iron oxide are not in contact
with each other, the transfer of activated hydrogen cannot take
place and they behave as individual species.
On the basis of the above results and the Mars−Van
Krevelen mechanism previously proposed for the reduction of
carboxylic acids on iron oxides, we propose the following
mechanism for the synergistic hydrogenation of stearic acid on
the composite catalyst (Scheme 1). In the proposed scheme,
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.5b01664.
■
*
S
Scheme 1. Proposed Mechanism of Fatty Acid Reduction on
Experimental methods, additional table and figures
a
Copper−Iron Oxides Supported on MSN
(PDF)
AUTHOR INFORMATION
Corresponding Author
E-mail: islowing@iastate.edu.
■
*
Author Contributions
‡
These authors contributed equally to the work (K.K., U.C.,
and N.C.N.).
a
Notes
(
a) Hydrogen is activated on copper and spills over to iron oxide
The authors declare no competing financial interest.
creating an oxygen vacancy, (b) carboxylic oxygen binds at the vacancy
and is reduced by a second equivalent of hydrogen, and (c) reduced
alcohol desorbs from the surface leaving behind an oxygen atom.
ACKNOWLEDGMENTS
■
The authors thank James W. Anderegg for the acquisition of
XPS spectra of the materials and BASF for the generous
donation of Pluronic P104 surfactant. This research is
supported by the U.S. Department of Energy, Office of Basic
Energy Sciences. The Ames Laboratory is operated for the U.S.
Department of Energy by Iowa State University under Contract
No. DE-AC02-07CH11358.
0
hydrogen chemisorbed on Cu spills over to closely located
iron oxide. There it reacts with lattice oxygen to produce water,
creating a vacancy upon water desorption (Scheme 1a). The
vacancy is then occupied via dissociative chemisorption of the
fatty acid (Scheme 1b). Because of the strength of the Fe−O
bond, the C−O bond can be broken by reaction of the surface
0
species with additional H activated at Cu . This leads to
REFERENCES
■
formation of the fatty alcohol, which desorbs regenerating the
lattice oxygen at the iron oxide site (Scheme 1c).
(
1) Christensen, C. H.; Rass-Hansen, J.; Marsden, C. C.; Taarning,
E.; Egeblad, K. ChemSusChem 2008, 1, 283−289.
A dramatic decrease in activity was observed upon recycling
the catalyst, as conversion dropped to 23%. No significant
leaching of the metals was detected by ICP-OES analysis of the
reaction solution (less than 0.2 wt % of metals). XPS analysis of
(
2) Pfaltzgraff, L. A.; De bruyn, M.; Cooper, E. C.; Budarin, V.; Clark,
J. H. Green Chem. 2013, 15, 307−314.
3) Metzger, J.; Bornscheuer, U. Appl. Microbiol. Biotechnol. 2006, 71,
(
13−22.
0
the spent catalyst indicated Cu and Fe(III) (Figure S7), which
(4) Meier, M. A.; Metzger, J. O.; Schubert, U. S. Chem. Soc. Rev.
2007, 36, 1788−1802.
is consistent with the proposed reaction mechanism. Chem-
isorption measurements of the spent catalyst did not show
evidence of sintering, as Cu dispersion remained at 65%.
However, calcination of the spent catalyst at 350 °C led to
regeneration of the activity (84% conversion). The regenerated
catalyst is similar to the original coimpregnated material
according to XPS (Figure S7) and dispersion measurements
(
5) Holmberg, K. Curr. Opin. Colloid Interface Sci. 2001, 6, 148−159.
(6) Rieke, R. D.; Thakur, D. S.; Roberts, B. D.; White, G. T. J. Am.
Oil Chem. Soc. 1997, 74, 333−339.
(7) Kreutzer, U. R. J. Am. Oil Chem. Soc. 1984, 61, 343−348.
(8) Voeste, T.; Buchold, H. J. Am. Oil Chem. Soc. 1984, 61, 350−352.
(9) Rieke, R. D.; Thakur, D. S.; Roberts, B. D.; White, G. T. J. Am.
Oil Chem. Soc. 1997, 74, 341−345.
10) Chen, Y.; Chang, C. Catal. Lett. 1997, 48, 101−104.
́
(11) Toba, M.; Tanaka, S.-i.; Niwa, S.-i.; Mizukami, F.; Koppany, Z.;
(64%). The loss and recovery of activity upon redox cycling of
(
the catalyst suggests subtle changes in the interaction between
both metal oxides that may take place at their interface.
Investigation of these changes with a simplified model is
currently underway in our laboratory.
In conclusion, Cu and Fe oxides supported on MSN act
synergistically to catalyze the hydrogenation of stearic acid to 1-
octadecanol under milder conditions than the individual metal
oxides supported on the same material. The process involves in
situ-generated metallic copper, which likely acts as a binding
Guczi, L.; Cheah, K.-Y.; Tang, T.-S. Appl. Catal., A 1999, 189, 243−
250.
(
12) Mendes, M. J.; Santos, O. A. A.; Jordao
Catal., A 2001, 217, 253−262.
13) Rozmysłowicz, B.; Kirilin, A.; Aho, A.; Manyar, H.; Hardacre,
C.; Warna, J.; Salmi, T.; Murzin, D. Y. J. Catal. 2015, 328, 197−207.
14) Manyar, H. G.; Paun, C.; Pilus, R.; Rooney, D. W.; Thompson,
J. M.; Hardacre, C. Chem. Commun. 2010, 46, 6279−6281.
̃
, E.; Silva, A. M. Appl.
(
̈
̊
(
(
15) Zhao, C.; Bruck, T.; Lercher, J. A. Green Chem. 2013, 15, 1720−
and activating site for H , whereas iron oxide is thought to bind
2
1739.
and activate the carboxylic groups via a Mars−Van Krevelen
mechanism on in situ generated oxygen vacancies. Hydrogen
(16) Song, W.; Zhao, C.; Lercher, J. A. Chem. - Eur. J. 2013, 19,
9833−9842.
6
722
DOI: 10.1021/acscatal.5b01664
ACS Catal. 2015, 5, 6719−6723

ACS Catalysis
Letter
(
17) Kandel, K.; Frederickson, C.; Smith, E. A.; Lee, Y. J.; Slowing, I.
I. ACS Catal. 2013, 3, 2750−2758.
18) Kandel, K.; Anderegg, J. W.; Nelson, N. C.; Chaudhary, U.;
Slowing, I. I. J. Catal. 2014, 314, 142−148.
19) Santiago, M. A. N.; Sanchez-Castillo, M. A.; Cortright, R. D.;
Dumesic, J. A. J. Catal. 2000, 193, 16−28.
20) Turek, T.; Trimm, D. L.; Cant, N. W. Catal. Rev.: Sci. Eng. 1994,
6, 645−683.
21) Jeong, N. C.; Lee, J. S.; Tae, E. L.; Lee, Y. J.; Yoon, K. B. Angew.
(
(
́
(
3
(
Chem., Int. Ed. 2008, 47, 10128−10132.
(
(
22) Zaki, T. J. Colloid Interface Sci. 2005, 284, 606−613.
23) Pestman, R.; Koster, R. M.; Pieterse, J. A. Z.; Ponec, V. J. Catal.
1
(
(
(
(
997, 168, 255−264.
24) Rachmady, W.; Vannice, M. A. J. Catal. 2002, 208, 158−169.
25) Rachmady, W.; Vannice, M. A. J. Catal. 2002, 208, 170−179.
26) Bowker, M.; Madix, R. J. Appl. Surf. Sci. 1981, 8, 299−317.
27) Chernyshova, I. V.; Ponnurangam, S.; Somasundaran, P.
Langmuir 2011, 27, 10007−10018.
28) Alexander, C. S.; Pritchard, J. J. Chem. Soc., Faraday Trans. 1
972, 68, 202−215.
29) Balooch, M.; Cardillo, M. J.; Miller, D. R.; Stickney, R. E. Surf.
Sci. 1974, 46, 358−392.
30) Okuhara, T.; Jin, T.; Zhou, Y.; White, J. M. J. Phys. Chem. 1988,
2, 4141−4147.
31) Claus, P.; Lucas, M.; Lu
A 1991, 79, 1−18.
32) Li, S.; Meitzner, G. D.; Iglesia, E. J. Phys. Chem. B 2001, 105,
743−5750.
33) Chen, C.-S.; Cheng, W.-H.; Lin, S.-S. Appl. Catal., A 2004, 257,
7−106.
34) Lendzion-Bielun, Z.; Bettahar, M. M.; Monteverdi, S. Catal.
Commun. 2010, 11, 1137−1142.
35) Tsoncheva, T.; Rosenholm, J.; Linden, M.; Ivanova, L.;
Minchev, C. Appl. Catal., A 2007, 318, 234−243.
36) Li, S.; Krishnamoorthy, S.; Li, A.; Meitzner, G. D.; Iglesia, E. J.
Catal. 2002, 206, 202−217.
37) Eguchi, K.; Shimoda, N.; Faungnawakij, K.; Matsui, T.; Kikuchi,
R.; Kawashima, S. Appl. Catal., B 2008, 80, 156−167.
38) Estrella, M.; Barrio, L.; Zhou, G.; Wang, X.; Wang, Q.; Wen, W.;
Hanson, J. C.; Frenkel, A. I.; Rodriguez, J. A. J. Phys. Chem. C 2009,
(
1
(
(
9
(
̈
cke, B.; Berndt, T.; Birke, P. Appl. Catal.,
(
5
(
9
(
(
(
(
(
1
(
(
13, 14411−14417.
39) Prins, R. Chem. Rev. 2012, 112, 2714−2738.
40) Lebedeva, O. E.; Chiou, W.-A.; Sachtler, W. M. H. J. Catal.
1
999, 188, 365−374.
6
723
DOI: 10.1021/acscatal.5b01664
ACS Catal. 2015, 5, 6719−6723