This is the pre-peer reviewed version of the Communications Chem. Int. Ed. 2019, 58,(19) Angewandte
following article: Angew.
5877 –
Chemie
5881, which has been published in final form at https://onlinelibrary.wiley.com/doi/abs/10.1002/anie.
201811669. This article may be used for non-commercial purposes in accordance with Wiley Terms and
International Edition: DOI: 10.1002/anie.201811669
Conditions forOxyhalogenation Hot Paper
Alkane Use of Self-Archived Versions.
German Edition:

DOI: 10.1002/ange.201811669

Halogen-Dependent Surface Confinement Governs Selective Alkane
Functionalization to Olefins
´
Guido Zichittella+, Matthias Scharfe+, BegoÇa Puÿrtolas, Vladimir Paunovic, Patrick Hemberger,
Andras Bodi, LµszlÛ SzentmiklÛsi, NÇria LÛpez, and Javier Pÿrez-RamÌrez*
Abstract: The product distribution in direct alkane functionalization by oxyhalogenation strongly depends on the halogen
of choice. We demonstrate that the superior selectivity to olefins
over an iron phosphate catalyst in oxychlorination is the
consequence of a surface-confined reaction. By contrast, in
oxybromination alkane activation follows a gas-phase radicalchain mechanism and yields a mixture of alkyl bromide,
cracking, and combustion products. Surface-coverage analysis
of the catalyst and identification of gas-phase radicals in
operando mode are correlated to the catalytic performance by
a multi-technique approach, which combines kinetic studies
with advanced characterization techniques such as promptgamma activation analysis and photoelectron photoion coincidence spectroscopy. Rationalization of gas-phase and surface
contributions by density functional theory reveals that the
molecular level effects of chlorine are pivotal in determining
the stark selectivity differences. These results provide strategies
for unraveling detailed mechanisms within complex reaction
networks.

The development of novel technologies for the selective
functionalization of light alkanes is a critical step to enable
the utilization of natural gas as an energy vector in the
transition between the oil and the renewables era.[1]
This pivotal advancement is limited by our understanding of
the mechanisms governing the heterogeneously catalyzed
´
[*] G. Zichittella,[+] M. Scharfe,[+] Dr. B. Puÿrtolas, Dr. V. Paunovic,
Prof. J. Pÿrez-RamÌrez
Institute for Chemical and Bioengineering, Department of Chemistry
and Applied Biosciences, ETH Zurich
Vladimir-Prelog-Weg 1, 8093 Zurich (Switzerland)
E-mail: jpr@chem.ethz.ch
Dr. P. Hemberger, Dr. A. Bodi
Laboratory of Femtochemistry and Synchrotron Radiation
Paul Scherrer Institute
5232 Villigen (Switzerland)
Dr. L. SzentmiklÛsi
Nuclear Analysis and Radiography Department, Centre for Energy
Research, Hungarian Academy of Sciences
Konkoly-Thege MiklÛsi Çt 29–33, 1121 Budapest (Hungary)
Prof. N. LÛpez
Institute of Chemical Research of Catalonia
The Barcelona Institute of Science and Technology
Av. PaÔsos Catalans 16, 43007 Tarragona (Spain)
[+] These authors contributed equally to this work.
Supporting information, including catalyst preparation, characterization, and evaluation, descriptions of the operando PGAA and
PEPICO techniques, DFT calculations, and the ORCID identification
number(s) for the author(s) of this article can be found under:
https://doi.org/10.1002/anie.201811669.
Angew. Chem. Int. Ed. 2019, 58, 5877 –5881

reactions for direct hydrocarbon upgrading, which can
provide design criteria for active, selective, and stable
catalysts. In the case of light-alkane functionalization—
generally requiring high temperatures and/or aggressive
reactants—the level of complexity is further amplified by
the possibility of gas-phase reaction pathways, in which highly
reactive radical species and/or radical initiators are liberated
from the surface to generate desired and undesired products.[2] A high level of selectivity control could be achieved if
alkane activation were confined on a catalyst surface. However, to unravel reaction pathways, as well as reactive
intermediates in gaseous and solid phases, and to model
complex and dynamic surfaces at the atomic level, the
combined use of strong experimental evidence with advanced
theoretical approaches is required. Catalytic oxychlorination,
which involves the reaction of an alkane with HCl and O2, has
recently demonstrated selective ( 95 %) generation of
ethylene from ethane over a wide range of catalyst families,[3]
while the use of HBr as a halide source results in a range of
products comprising alkyl bromide, CH4, and carbon oxides,
among others.[4] To understand the mechanistic origin of such
selectivity control, we combined kinetic studies with operando surface-coverage quantification by prompt-gamma
activation analysis (PGAA) and monitoring of gas-phase
radicals by photoelectron photoion coincidence spectroscopy
(PEPICO), ultimately rationalized at the molecular level by
density functional theory (DFT) calculations (Figure 1). An
iron phosphate catalyst was chosen because of its ability to
selectively ( 97 %) generate ethylene and propylene via
alkane oxychlorination.[3] Performance assessments in ethane
oxychlorination (EOC) and oxybromination (EOB), under
variable temperatures (573–853 K), revealed that the light-off
curve for oxybromination was shifted to about 150 K—
a lower temperature compared to that obtained in oxychlorination (Figure 2 a) and in agreement with previous
observations on other materials.[4b] Characterization of the
material before and after catalysis by means of N2 sorption, Xray diffraction (XRD), and Raman spectroscopy revealed
that the textural properties and the crystallographic structure
were preserved (Supporting Information, Figure S1,
Table S2). A comparison of the selectivity patterns obtained
at a similar alkane conversion level (ca. 20 %) showed that
C2H4 is the major product (selectivity ca. 95 %) when HCl is
used as a halide source, while oxybromination led to the
formation of C2H5Br, CH4, carbon oxides, and C2H4 (Figure 2 b; Figure S2). Both reactions are believed to follow
a consecutive mechanism, where the alkyl halide is the
intermediate to the olefin (Figure S2). Nevertheless, the
observed selectivity differences might be caused by the

⌫ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

5877

Communications

Angewandte

Chemie

operando conditions (for the principle of PGAA
see Scheme S1), for which FePO4 is a suitable
material because of its moderate neutron-capture
cross-section.[6] The experimental setup comprises
a continuous-flow tubular reactor, where the
catalyst is non-destructively irradiated with cold
neutrons[7] that are captured by the nuclei of all
atoms in the sampling volume during the oxyhalogenation reaction. The de-excitation of the
thus formed compound nuclei results in emission
of prompt g-rays with element-specific energies
and with intensities that are proportional to the
number of emitting atoms (Figure 1). The quantified halogen uptake with respect to iron, and at
variable temperatures and inlet HX concentrations, is correlated with the yield of ethylene in
Figure 3. In particular, at the onset of oxychlorination,
FePO4
exhibited
approximately
4.1 mmolCl molFeˇ1 chlorine uptake, which corresponds to 76 % of surface iron sites occupied by
Figure 1. A representation of the multi-technique strategy used in this study that
chlorine. A rise in reaction temperature led
enabled unraveling of the surface-driven and gas-phase radical-chain pathways
through which alkane activation occurs in catalytic ethane oxyhalogenation. This
to an increase of the yield of ethylene while the
approach concertedly combined 1) kinetic analysis for assessment of catalytic
chlorine uptake decreased to approximately
activity and selectivity patterns with operando 2) PGAA and 3) PEPICO spectro3 mmolCl molFeˇ1 (equivalent to 55 % surface covscopies, enabling quantification of the surface composition and detection of
erage) and remained constant at higher tempergaseous intermediates, respectively. The experimental findings were ultimately
atures. On the other hand, the bromine uptake at
rationalized by 4) DFT calculations. Each of these techniques results in spectra, or
the onset of oxybromination at 573 K reached full
energy levels (inner graphs), that can be subsequently evaluated and/or correlated
coverage at approximately 5.4 mmolBr molFeˇ1,
to observable parameters (outer graphs), revealing a detailed mechanistic picture
for a complex chemical process.
which drastically decreased to 2.5 mmolBr molFeˇ1
at 633 K. Further increase in temperature to 723 K
did not allow bromine uptake quantification
more kinetically favorable elimination of HCl compared to
because of statistically negligible halogen content, as corroHBr over a catalyst surface, as shown in catalytic methyl
borated by the absence of bromine in the used catalyst by
halide coupling.[5] To probe this hypothesis, we conducted
elemental analysis (Table S2). Interestingly, an increment of
ethyl halide dehydrohalogenation to C2H4 over FePO4 (Figthe halide feed resulted in an increased olefin yield that
correlated with a higher halogen coverage in oxychlorination,
ure S3) and observed that the light-off curve for C2H5Cl
dehydrochlorination was shifted by about 40 K, to lower
while the opposite trend was observed in oxybromination. To
translate these observations into causality correlations, dentemperature, compared to C2H5Br. When O2 and HX were
sity functional theory calculations were performed on the
co-fed with ethyl halide over FePO4 (Figure 2 c), dehydrochlorination led mainly to C2H4, while a competition between
experimentally relevant (102) surface of trigonal FePO4
polybromination and dehydrobromination was evident for
(Figure S1). Cl2 and ClC desorption are endothermic by
ethyl bromide. Additionally, catalytic hydrogen halide oxida2.70 eV and 2.66 eV, respectively, indicating that chlorine
tion and gas-phase ethane halogenation were investigated so
atoms are persistent on the surface. In contrast, bromine can
as to measure the ability of the catalyst to produce molecular
desorb as Br2 (DE = 2.10 eV), and even more favorably as BrC
halogen and the intrinsic reactivity of the halogen with the
(DE = 1.52 eV, Figure 5; Figure S5). Therefore, the evolution
alkane at comparable conditions to oxyhalogenation (Figof bromine species is much more energetically favored
ure 2 a; Figure S4). While chlorine-based reactions occurred
compared to chlorine species, which explains the lower
in distinct temperature regions, the bromine-based analogues
surface bromination under comparable reaction conditions
overlapped, which indicates that adsorbed HCl cannot evolve
(Figures 2 a and 3; Figure S5). These results suggest that gasas Cl2 from the catalyst under typical oxychlorination
phase pathways can be initiated by the formation of bromine
radicals, which were evidenced by the detection of Br2, BrC,
conditions, in contrast to Br2 that can be generated during
and C2H5C radicals by operando PEPICO (Figure 4; Figoxybromination. These findings suggest that ethane oxychlorination could occur via a surface-driven mechanism in
ures S6–S11). This spectroscopic technique, whose principle is
which the adsorbed chlorine species might be important for
schematized in Scheme S2, has recently revealed evidence of
the catalyst performance. On the contrary, gas-phase bromithe gas-phase radical-chain routes in methane oxybrominanation with in-situ-generated bromine species is likely the
tion.[2e] Therein, the gaseous species exiting the PEPICO
preferred pathway for ethane activation in oxybromination.
reactor form a molecular beam that is skimmed and ionized
To address the impact of the halogen coverage on the
by monochromatic vacuum ultraviolet (VUV) radiation to
performance, we conducted PGAA spectroscopy under
yield photoelectrons and photoions that are detected in

5878

www.angewandte.org

⌫ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angew. Chem. Int. Ed. 2019, 58, 5877 –5881

Communications

Angewandte

Chemie

Figure 3. Yield of ethylene as a function of halogen uptake in ethane
oxyhalogenation at different temperatures and HX (X = Cl, Br) inlet
concentration as determined in operando PGAA. More details on the
working principles of this technique are summarized in Scheme S1.
Conditions: C2H6 :HX:O2 :Ar:He = 6:6:3:4.5:80.5, FT = 100 cm3 minˇ1,
Wcat = 1 g, P = 1 bar. Dotted lines represent the calculated halogen
uptake corresponding to full surface coverage. Error bars are presented
for each catalytic point of the measured halogen uptake.

Figure 2. a) Ethane conversion as a function of temperature in oxyhalogenation over FePO4 and in non-catalytic ethane halogenation over
inert quartz particles, and hydrogen halide conversion as a function of
temperature in HX (X = Cl, Br) oxidation over FePO4. These reactions
occur in distinct temperature regions where chlorine chemistry is
concerned, whereas they overlap for bromine chemistry. b) Selectivity
patterns obtained in catalytic ethane oxyhalogenation and non-catalytic
ethane halogenation at ca. 20 % ethane conversion, illustrating that
only ethane oxychlorination leads to selective ethylene generation.
c) Selectivity to products as a function of ethyl halide conversion in the
dehydrohalogenation of ethyl halide in the presence of HX and O2 over
FePO4, showing that ethylene generation is favored with an ethyl
chloride intermediate over its brominated counterpart. The dotted gray
lines denote the yield of product j and the different symbols refer to
the reaction temperature. Conditions in (a) and (b): C2H6 :HX(X2):O2 :
Ar:He = 6:6(3):3(0):4.5:80.5(86.5), FT = 100 cm3 minˇ1, Wcat = 1 g,
P = 1 bar. Conditions in (c): C2H5X:HX:O2 :Ar:He = F1:6:3:4.5:85.5,
FT = 100 cm3 minˇ1, Wcat = 1 g, P = 1 bar.

delayed coincidence. In this way, isomer-selective identification of reactants and products is permitted, including radical
intermediates in both ethane oxychlorination and oxybromination (Figures 1 and 4; Figures S6–S11).[8] In particular, the
evolution of signals corresponding to reaction intermediates,
Angew. Chem. Int. Ed. 2019, 58, 5877 –5881

as a function of temperature, was correlated with that of the
reaction products; while the Br2 signal decreased sharply,
signals associated with C2H5C and BrC radicals increased along
with C2H5Br, until 600 K. At higher temperatures, the signal
corresponding to C2H5Br dropped to zero, while that of the
radicals and C2H4 increased, which is consistent with a higher
rate of dehydrobromination compared to ethyl bromide
generation (Figure 4; Figure S8). On the other hand, neither
Cl2 nor ClC radicals were observed under oxychlorination
conditions in the investigated temperature range (Figure 4;
Figure S6), which is in line with the DFT results. Furthermore,
C2H5Cl decreased sharply with temperature while C2H4
increased, corroborating the occurrence of a consecutive
mechanism in which ethane is transformed into ethyl halide
that is consequently dehydrohalogenated to C2H4, as supported by steady-state experiments (Figure S2). Finally,
measurements conducted using an empty reactor showed no
production of radicals or products (Figures S10 and S11).
Computational investigations at the molecular level
revealed that all the elementary steps in the gas-phase
pathways of oxybromination were barrierless. Additionally,
ethane or ethyl bromide activation with a BrC radical, to form
HBr, is endothermic by only 0.5 eV, while recombination of
the produced radicals with BrC to form C2H5Br or C2H4Br2,
which is occurring with a comparable collision frequency as
activation, is strongly exothermic (> 2.5 eV, Figure S13). This
implies that, once a bromine radical evolves from the surface,
it will directly activate ethane or ethyl bromide, or consume
any hydrocarbon radicals by recombination, thereby allowing
inordinate polybromination instead of ethylene generation.
Additionally, this mechanism of ethane activation can explain
the apparent lower activation energy for ethane conversion in
oxybromination (Figure 2 a), which consistently opposes the
DFT results for surface-catalyzed alkane activation in this
reaction (Figure S14). Nonetheless, C2H5C radicals were
observed by operando PEPICO in oxychlorination as well,
and its signal increased with the reaction temperature
(Figure 4). Its occurrence in oxychlorination, in which the

⌫ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.angewandte.org

5879

Communications

Angewandte

Chemie

strips a hydrogen and generates
the halide, thus closing the halogen cycle and highlighting the
importance of surface halogen
coverage for olefin formation.
A chlorinated site favors this
step by 0.18 eV over the brominated analogue (Figure S14)—
in line with the observed 40 K
light-off temperature shift in the
dehydrohalogenation of ethyl
halide (Figure S3). Furthermore, Cl stripping from the
C2H4X intermediate to form
C2H4 (Ea = 0.67 eV) is favored
over Br stripping (Ea = 1.50 eV,
Figure S14). However, considering the different extent of surface halogenation in oxychlorination and oxybromination
Figure 4. Extracted peak areas corresponding to molecular halogen, ethyl radical, ethyl halide, ethylene,
and halogen radical as a function of temperature in the oxychlorination and oxybromination of ethane
(as determined by operando
over FePO4 and using an empty reactor as determined in operando PEPICO (middle panel). More details
PGAA) and the possibility that
on the working principles of this technique are summarized in Scheme S2. The left and right panels
a halogenated site enables conillustrate the representative mass spectra of reactants, products, and intermediate radical species
certed H and Cl abstraction
detected in ethane oxychlorination and oxybromination over FePO4 at 723 K and 593 K, respectively.
3
ˇ1
ˇ2
from ethyl halide (Figure S14),
Conditions: C2H6 :HX:O2 :Ar = 2:2:1:17, FT = 22 cm min , Wcat = 0.05 or 0 g, P = 2 î 10 bar. The photon
which is favored over the indienergies at which the signals of each chemical species were recorded in all PEPICO experiments are
shown in Figures S6-S11.
vidual steps by 0.29 eV, the most
likely situation is that dehydrochlorination occurs in a congas-phase contributions are negligible, points to its role as an
certed pathway on a chlorinated Fe site, while dehydrobrointermediate in the surface-confined mechanism, while its
mination occurs on an oxidic site in two individual steps
spectroscopic detection can be rationalized owing to its easy
(Figure 5; Figures S16–S18). Notably, in addition to the
critical role that the halogen assumes in ethyl halide
desorption from the catalyst (endothermic by 0.47 eV,
activation, it is even indirectly involved in ethane activation
Figure 5; Figures S5 and S14–S18), and the low reaction
with oxygen, as evidenced by the complete absence of C2H5C
pressures (ca. 2 î 10ˇ2 bar). In general, surface oxyhalogenation reactions require an Fe site with a neighboring oxygen
radicals in the PEPICO studies with halide-free feeds
and its halogenated counterpart (the latter are more prom(Figures S7 and S9), and corroborated by the promoting
inent in oxychlorination, as determined by operando PGAA
effect of chlorine in oxidative dehydrogenation.[2a, 9] This can
(Figure 3)), which are responsible for ethane activation and
be explained by the change of the oxidation state of iron upon
for ethyl halide dehydrohalogenation, respectively (Figure S14, Table S3). In particular, the former site allows for
halogen-independent H stripping from ethane (Ea = 0.61 eV,
Figure 5), which is however
unlikely to catalyze H stripping
from ethyl halide (Ea > 2.10 eV,
Figure S14). Ethyl halide is
formed through barrierless halogen addition to a surfacebound C2H5C species (Figures
S14–S18), which entails that
only a small fraction evolves
to the gas-phase provided a halFigure 5. a) Energy barriers for halogen evolution, b) ethane activation, and c) ethyl halide dehydrohalogeogen atom is close. Further
nation in ethane oxyhalogenation over a halogenated FePO4(102) surface. The bottom panels illustrate
transformation of ethyl halide
the events corresponding to the respective bar plots that occur over the surface or in the gas-phase.
to ethylene requires a halogenColor code: Fe (light brown), P (yellow), O (red), Cl (green), Br (brown), C (gray), H (white). The
ated Fe site, where the halogen
complete reaction profile of the depicted most likely steps is provided in Figure S18.

5880

www.angewandte.org

⌫ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angew. Chem. Int. Ed. 2019, 58, 5877 –5881

Angewandte

Communications
halogen adsorption and the resulting structural rearrangement of the surface (Figure S19), as indicated by a Bader
charge analysis (Table S4).[10] Thus, these changes of the
electronic structure allow easier ethane activation when
spectator halogen species are present—in contrast to a clean
surface where the ethane activation has to overcome
a 1.11 eV larger energy barrier.
In conclusion, we unraveled the mechanistic origin of the
distinct selectivity patterns between ethane oxychlorination
and oxybromination over iron phosphate by a multi-technique approach, which combined evidence from steady-state
kinetics and operando PGAA and PEPICO spectroscopies,
and was ultimately rationalized with theoretical calculations.
The sharp selectivity control in oxychlorination is achieved by
a purely surface-driven functionalization of ethane into ethyl
chloride, which is further dehydrochlorinated to ethylene
over a ClˇFe center. In contrast, alkane activation to ethyl
bromide in oxybromination occurs in the gas-phase with
evolved bromine and bromine radical species, thus leaving
a halogen-free surface that is more prone to additional
cracking and combustion pathways. These results demonstrate that chlorine-based processes hold great potential for
one-step olefin production in technical scale and provide
guidelines for catalyst design for direct alkane-to-olefins
transformation via oxyhalogenation. Furthermore, the findings provide a strategy for unraveling the mechanistic picture
in a complex reaction network, which is the typical scenario
encountered in virtually all hydrocarbon functionalization
processes.

Acknowledgements
This work was supported by an ETH Research Grant ETH-04
16-1. P.H. and A.B. acknowledge funding by the SFOE (SI/
501269-01). L.S. thanks the Jµnos Bolyai Research Fellowship
of the Hungarian Academy of Sciences, as well as the Project
No. 124068 of the National Research, Development and
Innovation Fund of Hungary, financed under the K_17
funding scheme. We thank the Budapest Neutron Centre s
transnational user access program for funding the PGAA
beamtime. The authors thank Prof. Ralph Spolenak for access
to Raman spectroscopy and Dr. Detre Teschner for providing
accessories for the PGAA experiments. We thank Ali Saadun,
Boglµrka MarÛti, and IldikÛ Harsµnyi for assistance with the
PGAA measurements, and Florian Goedicke for help with
the PEPICO experiments.

Conflict of interest
The authors declare no conflict of interest.

Angew. Chem. Int. Ed. 2019, 58, 5877 –5881

Chemie

Keywords: density functional theory · hydrogen halides ·
olefin selectivity · operando spectroscopies ·
reaction mechanisms
How to cite: Angew. Chem. Int. Ed. 2019, 58, 5877 – 5881
Angew. Chem. 2019, 131, 5935 – 5940
[1] a) R. Lin, A. P. Amrute, J. Pÿrez-RamÌrez, Chem. Rev. 2017, 117,
4182 – 4247; b) E. McFarland, Science 2012, 338, 340 – 342; c) H.
Schwarz, Angew. Chem. Int. Ed. 2011, 50, 10096 – 10115; Angew.
Chem. 2011, 123, 10276 – 10297; d) J. J. Sattler, J. Ruiz-Martinez,
E. Santillan-Jimenez, B. M. Weckhuysen, Chem. Rev. 2014, 114,
10613 – 10653; e) J. F. Hartwig, J. Am. Chem. Soc. 2016, 138, 2 –
24; f) R. Horn, R. Schlçgl, Catal. Lett. 2015, 145, 23 – 39; g) A. I.
Olivos-Suarez, ¿. Szÿcsÿnyi, E. J. M. Hensen, J. Ruiz-Martinez,
E. A. Pidko, J. Gascon, ACS Catal. 2016, 6, 2965 – 2981.
[2] a) C. A. G‰rtner, A. C. van Veen, J. A. Lercher, ChemCatChem
2013, 5, 3196 – 3217; b) F. Cavani, N. Ballarini, A. Cericola, Catal.
Today 2007, 127, 113 – 131; c) K. Kwapien, J. Paier, J. Sauer, M.
Geske, U. Zavyalova, R. Horn, P. Schwach, A. Trunschke, R.
Schlçgl, Angew. Chem. Int. Ed. 2014, 53, 8774 – 8778; Angew.
Chem. 2014, 126, 8919 – 8923; d) C. A. Carrero, R. Schlçgl, I. E.
Wachs, R. Schom‰cker, ACS Catal. 2014, 4, 3357 – 3380; e) V.
´
Paunovic, P. Hemberger, A. Bodi, N. LÛpez, J. Pÿrez-RamÌrez,
Nat. Catal. 2018, 1, 363 – 370.
´
[3] G. Zichittella, N. Aellen, V. Paunovic, A. P. Amrute, J. PÿrezRamÌrez, Angew. Chem. Int. Ed. 2017, 56, 13670 – 13674; Angew.
Chem. 2017, 129, 13858 – 13862.
´
[4] a) V. Paunovic, G. Zichittella, M. Moser, A. P. Amrute, J. PÿrezRamÌrez, Nat. Chem. 2016, 8, 803 – 809; b) G. Zichittella, B.
´
Puÿrtolas, V. Paunovic, T. Block, R. Pçttgen, J. Pÿrez-RamÌrez,
Catal. Sci. Technol. 2018, 8, 2231 – 2243; c) G. Zichittella, B.
´
Puÿrtolas, S. Siol, V. Paunovic, S. Mitchell, J. Pÿrez-RamÌrez,
ChemCatChem 2018, 10, 1282 – 1290.
[5] U. Olsbye, O. V. Saure, N. B. Muddada, S. Bordiga, C. Lamberti,
M. H. Nilsen, K. P. Lillerud, S. Svelle, Catal. Today 2011, 171,
211 – 220.
[6] G. L. Molnµr, Handbook of Prompt Gamma Activation Analysis
with Neutron Beams, Springer, Heidelberg, 2004.
[7] Z. Rÿvay, T. Belgya, L. SzentmiklÛsi, Z. Kis, A. Wootsch, D.
Teschner, M. Swoboda, R. Schlçgl, J. Borsodi, R. Zepernick,
Anal. Chem. 2008, 80, 6066 – 6071.
[8] a) D. L. Osborn, C. C. Hayden, P. Hemberger, A. Bodi, K.
Voronova, B. Sztaray, J. Chem. Phys. 2016, 145, 164202; b) X.
Tang, G. A. Garcia, L. Nahon, J. Phys. Chem. A 2015, 119, 5942 –
5950; c) P. Osswald, P. Hemberger, T. Bierkandt, E. Akyildiz, M.
Kohler, A. Bodi, T. Gerber, T. Kasper, Rev. Sci. Instrum. 2014,
85, 025101; d) P. Hemberger, V. B. F. Custodis, A. Bodi, T.
Gerber, J. A. van Bokhoven, Nat. Commun. 2017, 8, 15946; e) P.
Hemberger, A. J. Trevitt, T. Gerber, E. Ross, G. da Silva, J. Phys.
Chem. A 2014, 118, 3593 – 3604.
[9] C. P. Kumar, S. Gaab, T. E. M¸ller, J. A. Lercher, Top. Catal.
2008, 50, 156 – 167.
[10] A. Walsh, A. A. Sokol, J. Buckeridge, D. O. Scanlon, C. R. A.
Catlow, Nat. Mater. 2018, 17, 958 – 964.
Manuscript received: October 10, 2018
Accepted manuscript online: January 15, 2019
Version of record online: February 14, 2019

⌫ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.angewandte.org

5881