Mechanistic Insights into Catalytic Ethanol Steam Reforming Using Isotope-Labeled Reactants

Article abstract of DOI:10.1002/anie.201604388

The low-temperature ethanol steam reforming (ESR) reaction mechanism over a supported Rh/Pt catalyst has been investigated using isotope-labeled EtOH and H₂O. Through strategic isotope labeling, all nonhydrogen atoms were distinct from one another, and allowed an unprecedented level of understanding of the dominant reaction pathways. All combinations of isotope- and non-isotope-labeled atoms were detected in the products, thus there are multiple pathways involved in H₂, CO, CO₂, CH₄, C₂H₄, and C₂H₆product formation. Both the recombination of C species on the surface of the catalyst and preservation of the C?C bond within ethanol are responsible for C₂product formation. Ethylene is not detected until conversion drops below 100 % at t=1.25 h. Also, quantitatively, 57 % of the observed ethylene is formed directly through ethanol dehydration. Finally there is clear evidence to show that oxygen in the SiO₂-ZrO₂support constitutes 10 % of the CO formed during the reaction.

Full text of DOI:10.1002/anie.201604388

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201604388
German Edition: DOI: 10.1002/ange.201604388
Reaction Mechanisms
Mechanistic Insights into Catalytic Ethanol Steam Reforming Using
Isotope-Labeled Reactants
Stephen Crowley and Marco J. Castaldi*
Abstract: The low-temperature ethanol steam reforming
(ESR) reaction mechanism over a supported Rh/Pt catalyst
has been investigated using isotope-labeled EtOH and H2O.
Through strategic isotope labeling, all nonhydrogen atoms
were distinct from one another, and allowed an unprecedented
level of understanding of the dominant reaction pathways. All
combinations of isotope- and non-isotope-labeled atoms were
detected in the products, thus there are multiple pathways
involved in H2, CO, CO2, CH4, C2H4, and C2H6 product
formation. Both the recombination of C species on the surface
ꢀ
of the catalyst and preservation of the C C bond within ethanol
are responsible for C2 product formation. Ethylene is not
detected until conversion drops below 100% at t = 1.25 h. Also,
quantitatively, 57% of the observed ethylene is formed directly
through ethanol dehydration. Finally there is clear evidence to
Figure 1. Possible reaction pathways during ethanol steam reforming.
show that oxygen in the SiO2-ZrO2 support constitutes 10% of
Circled numbers correlate to numbered Equations (1)–(12). Arrows
the CO formed during the reaction.
represent forward reaction for simplicity.
P
recious metal nanoparticles dispersed on high-surface-area
carriers have been demonstrated to exhibit superior capa-
bilities in catalyzing chemical reactions.[1,2] Currently there is
an interest in employing precious metal catalysts in oxy-
genated fuel reforming to produce hydrogen.[3–5] It is pro-
jected that the use of oxygenated feedstocks to produce
chemicals will grow nearly fourfold from $3.6 billion in 2011
to $12.2 billion by 2021.[6] Ethanol is an excellent example of
a platform feedstock targeted for reasons of current cost and
availability.[7,8] However, the role of specific metals and
supports in the reforming process is only well understood for
oxygen-free hydrocarbon fuels. Base metals such as nickel
Equations (1)–(12) and Figure 1 represent a fraction of what
have been determined to be the dominant contributors.
Steam Reforming : C2H5OH þ H2O Ð 2 CO þ 4 H2
Dehydrogenation : C2H5OH Ð C2H4O þ H2
Decomposition : C2H5OH Ð CH4 þ CO þ H2
Dehydration : C2H5OH Ð C2H4 þ H2O
ð1Þ
ð2Þ
ð3Þ
ð4Þ
Methane Cracking : CH4 Ð C þ 2 H2
ð5Þ
ꢀ
and copper are ineffective at breaking the C C bond within
Ethylene Cracking : C2H4 Ð 2 C þ 2 H2
ð6Þ
oxygenates and deactivate readily through coke forma-
tion.[9,10] Thus, precious metals are under investigation for
the steam reforming process with bimetallic rhodium-based
formulations showing increased activity and resistance to
deactivation.[11–13] Despite their superior performance, there is
still no consensus on how precious metals interact with
reactant species to give rise to the products.
There are a myriad competing reactions contributing to
the overall mechanistic understanding of ethanol steam
reforming (ESR)[14] and they number in the hundreds.
Acetaldehyde Decomposition : C2H4O Ð CH4 þ CO
Boudouard Reaction : 2 CO Ð C þ CO2
ð7Þ
ð8Þ
Water Gas Shift ðWGSÞ : CO þ H2O Ð CO2 þ H2
Reverse Carbon Gasification : CO þ H2 Ð C þ H2O
Methanation : CO þ 3 H2 Ð CH4 þ H2O
ð9Þ
ð10Þ
ð11Þ
ð12Þ
Ethylene Hydrogenation : C2H4 þ H2 Ð C2H6
[*] S. Crowley, Prof. Dr. M. J. Castaldi
Department of Chemical Engineering
City College of the City University of New York
New York, NY 10031 (USA)
Classic reaction-model development focuses on measur-
ing products over a range of test conditions combined with
a proposed set of equilibrium reactions to match observed
experimental data.[14–17] In a recent review on catalytic ESR
by Hou et al., the authors highlight that there is no agreed
upon reaction pathway for the overall ESR process.[14] Many
researchers have contributed extensive experimental and
E-mail: mcastaldi@ccny.cuny.edu
Supporting information and the ORCID identification numbers for
the authors of this article can be found under:
http://dx.doi.org/10.1002/anie.201604388.
Angew. Chem. Int. Ed. 2016, 55, 1 – 7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

Angewandte
Communications
Chemie
modeling efforts to elucidate the precise reaction sequence.
Vesselli et al. utilize X-ray photoelectron spectroscopy (XPS)
of adsorbed ethanol on a Rh(111) surface and UHV
desorption experiments, while Resta et al. use density-func-
tional theory (DFT) to determine major species formed
during ethanol decomposition, thus showing experimentally
contribution of each metal to the overall product distribution.
Bare SiO2-ZrO2 was found to be inactive in the process. All
experiments were performed in a packed bed reactor at
3508C and 1 atm. Other test parameters are available in the
Supporting Information. The product distribution profile for
the 3 wt% Rh/1 wt% Pt catalyst is shown in Figure 3 for the
steam reforming of the isotope-labeled reactants. Product
distribution profiles for the single-metal catalysts are avail-
able in the Supporting Information.
ꢀ
and computationally that C C bond cleavage is preferential
[18–21]
ꢀ
to C O bond scission.
In addition, the dehydrogenation
reaction [Eq. (2)] on Rh/CeO2 was studied by the group of
Chen, thus showing that
an oxametallacycle is
formed with subsequent
ꢀ
C C bond cleavage and
desorption to yield CH4,
H2, and CO.[22] How-
ever, there is currently
no consensus on the
origin of the atoms con-
stituting
the
final
observed products for
the steam reforming of
higher-order hydrocar-
bons.
Isotope labeling is
a
longstanding tech-
nique used to gain
insight into the likeli-
hood of particular reac-
tion pathways.[23–25] Song
et al. used deuterated
Figure 3. Product distribution profile for the reforming of [1-13C]ethanol with [18O]water over 3 wt% Rh/1 wt% Pt
on SiO2-ZrO2 support. T=3508C, S/C=1.5, GHSV=44000 hꢀ1
.
ethanol and water to determine the adsorption/desorption
behavior of reactants,[26] as well as to investigate the reaction
pathway of ethanol and water over CeO2- and ZrO2-
supported cobalt.[27] Here we seek to provide new insights
into a rhodium-based catalyst which dynamically changes
oxidation states during reforming.[28] Through employing both
isotope-labeled ethanol and water, it is now possible to track
the atomic partitioning of the reactants into the products.
Chemical formulas of the reactants, [1-13C]ethanol and
[18O]water are shown in Figure 2.
The product distribution changed dynamically within the
first four hours on stream, and is consistent with earlier
work.[28] Evidenced in Figure 3, initially there was a steady
increase in the amount of H2, CO, and CH4 produced, but CH4
reached a maximum after 10 minutes then steadily decreased.
The highest CO2 production was observed at t = 0 and
immediately decreased. C2H6 was produced at a near-constant
rate during the first 1.25 hours on stream while no C2H4 was
detected. At 1.25 hours, maxima were observed in H2 and CO
production followed by a steady decrease, thus indicating
catalyst deactivation. This deactivation was further supported
by ethanol conversion simultaneously falling below 100%. A
selectivity change from C2H6 to C2H4 was also observed and it
stabilized after 2.25 hours on stream with C2H4 concentrations
of 0.007 mol% and no C2H6 detected. Identical features in the
reforming product distribution profile are observed after
In this study, isotope enrichment was 99% 13C for
[1-13C]ethanol and 97.40% 18O for [18O]water (confirmed by
GCMS). All non-H-atoms can be distinguished from one
another, thus allowing for atom origin differentiation.
ESR was performed over 3 wt% Rh/1 wt% Pt, 4 wt%
Rh, and 4 wt% Pt on SiO2-ZrO2 catalysts to determine the
a
1 hour regeneration in 5% O2 in N2. Only CO2
(0.12 mole%) was detected during this time, thus suggesting
oxidation of carbon deposited during reformation.
Interestingly, the single-metal catalyst formulations
showed signs of deactivation after 0.5 hours on stream as
opposed to 1.25 hours for the bimetallic catalyst (see Fig-
ures S1 and S2 in the Supporting Information). Overall,
similar reforming behavior was observed between the 4 wt%
Rh and 3 wt% Rh/1 wt% Pt formulations, and is not
surprising given that they are both primarily rhodium-
containing catalysts, with the bimetallic formulation provid-
ing the highest level of H2 production at its maximum.
However, the 4 wt% Pt catalyst exhibited a much lower
Figure 2. Isotope-labeled reactants [1-13C]ethanol and [18O]water.
2
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 7
These are not the final page numbers!

Angewandte
Communications
Chemie
selectivity for H2 and increased selec-
tivity for the C2 species, thus suggesting
platinum is less capable of breaking the
ꢀ
C C bond within ethanol.
Thus far we have shown what is
commonly reported in the literature:
the amount of each product detected as
a function of time on stream. However,
the origins of the atoms within each
product are still unknown. Through
examining the isotopic breakdown of
individual species, we can now deter-
mine how the reactants partition into
the products, thus giving us quantitative
mechanistic insight.
The simplest isotope-labeled prod-
uct differentiation occurs in CH4. Two
forms are possible, 12CH4 and 13CH4,
and both were detected. Total mol% of
CH4 as a function of time on stream is
presented in Figure 4A, and a reproduc-
tion of the data presented in Figure 3 is
shown with shaded areas representing
each type of CH4 detected. For the
bimetallic catalyst, a maximum amount
of 13CH4 was detected initially at
0.18 mol% (33% of total CH4), which
decreased to 0.01 mol% after
2.25 hours. This outcome is surprising
since 13CH4 formation requires both the
12
13
13C and the
C
16O bonds in the
ꢀ
C
ꢀ
ethanol to break. Similar behavior was
observed for the 4% Rh catalyst for-
mulation, though maximum 13CH4 was
detected at 10% of total CH4 rather
than 33%. 4 wt% Pt provided lower,
near constant 13CH4 at 2% of total
methane.
From this data set, rhodium seems
ꢀ
to be more capable of breaking the C C
bond in ethanol with either the 12
13C
ꢀ
C
bond breakage rate decreasing with
time on stream or remaining constant
while selectivity toward other 13C-con-
taining products increases, thus
accounting for the decrease in observed
13CH4. However the monotonic decline
in the CH4 for all catalyst formulations
ꢀ
Figure 4. Atomic partitioning of 12C,13C, 16O, and 18O into A) methane, B) ethane, C) ethylene,
D) carbon monoxide, and E) carbon dioxide over 3 wt% Rh/1 wt% Pt on SiO2-ZrO2. Bold colored
text indicates isotope labeling.
suggests the 12
C
13C bond scission becomes less favored since
[1-13C]ethanol and recombination with another 12C species,
and is consistent with the formation of 13CH4. Similar
behavior is observed for the rhodium-only formulation.
ꢀ
the concentration profiles change more abruptly for other 13
C
products. This observation will be discussed further with
respect to the C2 species. Similar tracking was performed for
the C2Hx and COy products (Figures 4B–E).
As discussed by Vesselli et al., it is likely that the 12
C
13C
bond within [1-13C]ethanol is cleaved prior to recombination
with a separate 12C species.[18] Following this logic, it seems
that the abundance of H313C12CH3 at levels of 35–45% can be
The C2H6 formed (Figure 4B) is initially observed to have
a composition of 65% H312C12CH3, and it decreases to 55% of
total C2H6 produced over the first 2 hours of ESR. Since the
two 12C atoms must come from separate [1-13C]ethanol
molecules, this shows that the C2H6 formation pathway
12
13C bond within
ꢀ
C
attributed to, first, a breaking of the
[1-13C]ethanol and subsequent breaking of the 13
16O bond,
ꢀ
C
thereby allowing the 12C and the 13C moieties to recombine
into ethane. This theory is supported by the presence of
12
13C bond within
ꢀ
occurs primarily by breaking the
C
Angew. Chem. Int. Ed. 2016, 55, 1 – 7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

Angewandte
Communications
Chemie
H313C13CH3. Even though this form of C2H6 is detected at low
levels, the only way for this molecule to form is through
clear evidence that the primary pathway for CO formation is
through direct reaction between the ethanol and water.
Roughly 25% 13C16O was detected throughout the test, thus
13
12
breaking the
C C
16O and the 13C bonds in ethanol, thus
ꢀ
ꢀ
allowing two 13C species to recombine into ethane. On the
contrary, the platinum-only catalyst formulation yields ethane
with H313C12CH3 making up 95% of the total, thus further
demonstrating that platinum is ineffective at breaking the
indicating that half as many
C
16O bonds within
13
ꢀ
[1-13C]ethanol are preserved in CO formation. It can be
seen that the 12C entity recombines with 18O from water to
form 12C18O, though only at 15% of total CO makeup.
However, no 12C16O is detected at any point during reforming
over the platinum-only catalyst. This result is expected since,
ꢀ
C C bond within ethanol. For all catalyst formulations,
however, much of the 13C goes to the COy species, thus
maintaining a low concentration of H313C13CH3.
upon first inspection, it would seem that the only source of 16
O
A similar explanation can be applied to C2H4 (Figure 4C)
with similar trends observed across the three catalyst
formulations. C2H4 and C2H6 formation have a strong corre-
lation since the formation of C2H6 (all isotopes) declines and
C2H4 concentration increases concurrently. This selectivity
change suggests that a portion of the C2H4 is formed directly
from the C2H6 by the gas-phase dehydrogenation path-
way.[29,30] However, this reaction is unfavored at the operating
temperature of 3508C,[31] thus implying that the C2H6
observed must be formed through ethylene hydrogenation
[Eq. (12)] on the catalyst surface, a commonly reported
feature of supported platinum-group metal catalysts.[32–34] We
hypothesize that the tradeoff between C2H6 and C2H4
production observed at t = 1.25 h implies that the ethylene
hydrogenation reaction is suppressed as the catalyst deacti-
vates, and it is consistent with the emergence of ethylene as
a stable product and the simultaneous decreases in H2 and
ethanol conversion.
is the carbon-bound oxygen atom in [1-13C]ethanol. It would
therefore be unlikely for the 13
16O bond to break simply for
ꢀ
C
the 16O moiety to combine with a 12C species on the catalyst,
thus making it surprising to find that 12C16O is present
throughout the reforming experiment for the rhodium-con-
taining formulations. Thus, we propose that the CO formation
pathway on rhodium-containing catalysts occurs not only
through recombination of the ethanol and water reactants
themselves, but also with the oxygen within the catalyst
support as the only other source for 16O.
Oxygen exchange between water and silica/zirconia sup-
ports has been well documented.[35–39] Tracking the composi-
tion of water in the effluent yielded a nearly constant isotopic
concentration of 55% H218O and 45% H216O. This outcome
proves that H216O is formed consistently throughout the test,
either through oxygen exchange with the support or through
the dehydration, reverse carbon gasification, and methana-
tion reactions [Eqs (4), (10), and (11) respectively]. In
addition, the CO2 concentration profile (Figure 4E) reveals
that the WGS reaction [Eq. (9)] initially occurs, but gradually
declines. Simultaneously, the CO concentration increases
from 0.45 mol% at t = 0 to 0.65 mol% at t = 1.25 h, that is,
more than the amount that can be provided by the decom-
position of CO2 (maximum observed concentration:
0.15 mole%) to CO. Consequently, the likely COy species
which initially forms on the catalyst is CO, and it undergoes
further oxidation to CO2, supplied with oxygen from either
the rhodium surface, the support itself, or water.
However, the amount of C2H4 observed cannot be solely
attributed to the prevention of the hydrogenation reaction.
Stable C2H4 production was observed at 0.007 mol%. Since
stable C2H6 production was observed at 0.003 mol%, this
leaves 0.004 mol% (57%) of C2H4 unaccounted for. By
examining the corresponding isotope-labeled components of
C2H6 and C2H4, insight can be gained into the most likely
12
ꢀ
ethylene production pathway. The amount of both
C
12C
13
and
C
13C ethane and ethylene observed before and after
the tradeoff in selectivity at t = 1.25 h indicates that the
ꢀ
12
As evidenced, 12C16O is formed when the
13C bond
ꢀ
C
ethylene hydrogenation reaction is blocked for ethylene
formed from identical carbon atoms. The 12
13C ethane and
within [1-13C]ethanol is broken, thus freeing a 12C species
which remains on the surface of the catalyst and allows
reaction with 16O on the support. This process is supported by
a feature of the CO isotope product distribution occurring
after 3.75 hours on stream. At this point, 12C16O is no longer
observed in the products and suggests one of two phenomena:
that the 16O in the support near the metal–support interface
has been completely depleted and replaced by 18O from the
labeled water, or that the bimetallic catalyst is no longer
a pure alloy with platinum segregating to the surface and
preventing recombination of the 12C and 16O species. Notably,
12C16O production resumes after the regeneration step and
suggests that the oxygen in the support is replenished with
16O, thus allowing the same phenomenon to occur in the
second ethanol reforming (t = 5.7 h and later).
ꢀ
C
ethylene, however, are a different matter. H313C12CH3 is
observed at 0.001 mol% during stable C2H6 production
whereas H213C12CH2 is observed at 0.005 mol% during
stable C2H4 production. Thus, the 0.004 mol% of unac-
counted-for C2H4 observed is entirely present as
H213C12CH2. Therefore, the decrease in 13CH4 observed, as
well as the minimal H213C13CH2 detected, provide support for
the ethanol dehydration pathway [Eq. (4)] as the most likely
source of ethylene production.
The products become more complex for the COy species
with two options for the carbon atom (12C and 13C) and two
options for the oxygen atom(s) (16O and 18O). For the
bimetallic and single-metal rhodium catalyst formulations,
all four possible forms of CO (Figure 4D) are detected, thus
proving that CO is not formed solely from the O-bound
carbon atom within ethanol. Interestingly, for the rhodium-
containing catalysts, 13C18O accounts for nearly 50% of the
total CO detected throughout the test, and increases to 65%
during the last 15 minutes of reforming. This occurrence is
Interestingly, all six possible species of CO2 are detected
throughout the reforming experiments for the three catalyst
formulations. While the amount of CO2 decreased after an
initial maximum, CO2 showed the lowest variability in isotope
product breakdown with nearly constant makeup as detailed
4
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 7
These are not the final page numbers!

Angewandte
Communications
Chemie
Table 1: Isotope product distribution for CO2 during ethanol reforming.
CO2 Isotope: 16O12C16O 16O13C16O 18O12C16O 18O13C16O 18O12C18O 18O13C18O
Composition: 4.3% 10.3% 14.3% 32.5% 12.3% 26.3%
[1] G. Bond, Heterogeneous Catalysis: Principles and
Applications, Clarendon Press, Oxford, New York,
1987.
[2] S. Duan, S. Senkan, Ind. Eng. Chem. Res. 2005, 44,
6381 – 6386.
[3] Q. Ming, Catal. Today 2002, 77, 51 – 64.
in Table 1. Not surprisingly, the most abundant CO2 isotope is
[4] M. Li, K. Duraiswamy, M. Knobbe, Chem. Eng. Sci. 2012, 67,
26 – 33.
18O13C16O, in which the 13
16O bond within [1-13C]ethanol is
ꢀ
C
maintained and combines with the 18O from H218O. However,
[5] S. D. Angeli, G. Monteleone, A. Giaconia, A. A. Lemonidou,
Int. J. Hydrogen Energy 2014, 39, 1979 – 1997.
[6] N. Deschamps, The World Market for Bio-Based Chemicals,
2nd ed., SBI Energy: EL Analysts, Rockville, MD, 2012.
[7] P. B. Smith, G. F. Payne, Renewable and Sustainable Polymers,
American Chemical Society, Washington, DC, 2011.
[8] J. C. Serrano-Ruiz, R. Luque, A. Sepffllveda-Escribano, Chem.
Soc. Rev. 2011, 40, 5266 – 5281.
[9] A. Carrero, J. A. Calles, A. J. Vizcaíno, Chem. Eng. J. 2010, 163,
395 – 402.
[10] J. Sehested, Catal. Today 2006, 111, 103 – 110.
[11] A. Bshish, Z. Yaakob, B. Narayanan, R. Ramakrishnan, A.
Ebshish, Chem. Pap. 2011, 65, 251 – 266.
[12] M. A. Goula, S. K. Kontou, P. E. Tsiakaras, Appl. Catal. B 2004,
49, 135 – 144.
[13] F. Wang, W. Cai, C. Descorme, H. Provendier, W. Shen, C.
Mirodatos, Y. Schuurman, Int. J. Hydrogen Energy 2014, 39,
18005 – 18015.
[14] T. Hou, S. Zhang, Y. Chen, D. Wang, W. Cai, Renewable
Sustainable Energy Rev. 2015, 44, 132 – 148.
[15] M. Benito, J. L. Sanz, R. Isabel, R. Padilla, R. Arjona, L. Daza, J.
Power Sources 2005, 151, 11 – 17.
the next most abundant species is 18O13C18O where the
13
16O bond is broken and the 13C is bound to two 18O
ꢀ
C
centers. This step suggests that the second most likely process
for CO2 formation occurs by either the Boudouard reaction
[Eq. (8)], where two 13C18O molecules react, or by the WGS
reaction [Eq. (9)] where a 13C18O intermediate is formed and
reacts with [18O]water.
From our analysis, it is clear that the primary reaction
pathways are changing as a function of time on stream. A
figure has been included (Figure S5) and highlights the most
likely dominant pathways at various stages of ESR. By using
carefully selected isotope-labeled reactants, individual reac-
tion pathways have been examined with an unprecedented
level of clarity and trackability, thus allowing the unification
of theory and experimentally observed results. The selectivity
toward ethanol dehydration increased as the catalyst was
deactivated. The presence of platinum within the catalyst
makeup stabilizes rhodium performance and allows 100%
conversion of ethanol for a time period nearly threefold that
of the single-metal formulations. Rhodium is more effective
[16] S. M. de Lima, A. M. Silva, U. M. Graham, G. Jacobs, B. H.
Davis, L. V. Mattos, F. B. Noronha, Appl. Catal. A 2009, 352, 95 –
113.
ꢀ
at breaking the C C bond within ethanol and allows more
diverse species recombination than platinum. Support oxygen
was shown to play a crucial role in the formation of reforming
products, though platinum does not favor reaction between
this oxygen and carbon species. The presence of all C and
O isotopes within the products proves that bond preservation
is not the only pathway for product formation. A recombi-
nation of species on the catalyst surface is occurring
throughout the process for all catalyst formulations, thus
indicating that a classic depiction of possible reactions
[Eq. (1)–(12)] is incapable of completely describing species
formation.
[17] R. Trane-Restrup, S. Dahl, A. D. Jensen, Int. J. Hydrogen Energy
2013, 38, 15105 – 15118.
[18] E. Vesselli, A. Baraldi, G. Comelli, S. Lizzit, R. Rosei,
ChemPhysChem 2004, 5, 1133 – 1140.
[19] A. Resta, J. Blomquist, J. Gustafson, H. Karhu, A. Mikkelsen, E.
Lundgren, P. Uvdal, J. N. Andersen, Surf. Sci. 2006, 600, 5136 –
5141.
[20] E. Vesselli, G. Comelli, R. Rosei, S. Freni, F. Frusteri, S.
Cavallaro, Appl. Catal. A 2005, 281, 139 – 147.
[21] A. Resta, J. Gustafson, R. Westerstrçm, A. Mikkelsen, E.
Lundgren, J. N. Andersen, M. M. Yang, X. F. Ma, X. H. Bao,
W. X. Li, Surf. Sci. 2008, 602, 3057 – 3063.
[22] H.-L. Chen, S.-H. Liu, J.-J. Ho, J. Phys. Chem. B 2006, 110,
14816 – 14823.
[23] K. Otto, M. Shelef, J. T. Kummer, J. Phys. Chem. 1970, 74, 2690 –
2698.
Acknowledgments
[24] L. Shannon, J. G. Goodwin, Chem. Rev. 1995, 95, 677 – 695.
[25] R. M. Lum, J. K. Klingert, D. W. Kisker, S. M. Abys, F. A. Stevie,
J. Cryst. Growth 1988, 93, 120 – 126.
[26] H. Song, X. Bao, C. M. Hadad, U. S. Ozkan, Catal. Lett. 2011,
141, 43 – 54.
We thank Jeffrey LeBlanc and Prof. Dr. Ilona Kretzschmar
for their feedback on the manuscript and Megan Webster,
Robyn Smith, and the Combustion and Catalysis Laboratory
for their support. We thank the Department of Chemical
Engineering and the Earth Engineering Center at the City
College of New York for their financial support under project
number EEC j CCNY2012.
[27] H. Song, L. Zhang, U. S. Ozkan, Ind. Eng. Chem. Res. 2010, 49,
8984 – 8989.
[28] S. Crowley, Y. Li, A. I. Frenkel, M. J. Castaldi, in 249th Am.
Chem. Soc. Natl. Meet. Expo., Denver, CO, 2015.
[29] F. E. Frey, D. F. Smith, Ind. Eng. Chem. 1928, 20, 948 – 951.
[30] F. O. Rice, R. E. Varnerin, J. Am. Chem. Soc. 1954, 76, 324 – 327.
[31] H. J. Dar, S. U. Nanot, K. J. Jens, H. A. Jakobsen, E. Tangstad, D.
Chen, Ind. Eng. Chem. Res. 2012, 51, 10571 – 10585.
[32] Y. Okamoto, N. Ishida, T. Imanaka, S. Teranishi, J. Catal. 1979,
58, 82 – 94.
Keywords: heterogeneous catalysis · isotopic labeling ·
nanoparticles · reaction mechanisms · rhodium
Angew. Chem. Int. Ed. 2016, 55, 1 – 7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

Angewandte
Communications
Chemie
[33] S. Zhao, Y. Li, E. Stavitski, R. Tappero, S. Crowley, M. J.
Castaldi, D. N. Zakharov, R. G. Nuzzo, A. I. Frenkel, E. A.
Stach, ChemCatChem 2015, 7, 3683 – 3691.
[38] D. Martin, D. Duprez, J. Phys. Chem. 1996, 100, 9429 – 9438.
[39] J. Chelikowsky, D. Chadi, N. Binggeli, Phys. Rev. B 2000, 62,
R2251 – R2254.
[34] J. Grunes, J. Zhu, E. A. Anderson, G. A. Somorjai, J. Phys.
Chem. B 2002, 106, 11463 – 11468.
[35] D. A. Hutchison, J. Chem. Phys. 1954, 22, 758.
[36] W. G. Spitzer, J. R. Ligenza, J. Phys. Chem. Solids 1961, 17, 196 –
202.
Received: May 5, 2016
[37] B. W. Busch, W. H. Schulte, E. Garfunkel, T. Gustafsson, W. Qi,
R. Nieh, J. Lee, Phys. Rev. B 2000, 62, 290 – 293.
Revised: July 12, 2016
Published online: && &&, &&&&
6
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 1 – 7
These are not the final page numbers!

Angewandte
Communications
Chemie
Communications
Reaction Mechanisms
The tracking of atoms from reactants to
products during Rh/Pt catalytic ethanol
steam reforming using H312C13CH216OH
and H218O provided new insight into the
overall reaction mechanism. All combi-
nations of isotope- and non-isotope-
labeled atoms were detected in the
products, thus there are multiple path-
ways involved in H2, CO, CO2, CH4, C2H4,
and C2H6 product formation.
S. Crowley,
M. J. Castaldi*
&&&& — &&&&
Mechanistic Insights into Catalytic
Ethanol Steam Reforming Using Isotope-
Labeled Reactants
Angew. Chem. Int. Ed. 2016, 55, 1 – 7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7
These are not the final page numbers!