Chemical Behaviour of CaAg2 under Ethylene Epoxidation Conditions

Article abstract of DOI:10.1002/ejic.201800710

The binary compound CaAg₂ is examined as a catalyst for the ethylene epoxidation reaction. During the induction phase, conversion and selectivity increase and then remain stable for several hundred hours. The presence of ethyl chloride as a promoter is crucial. The pristine CaAg₂ reacts with the gaseous reactants and forms a porous microstructure of calcium-containing oxidation products on the surface, in which particles of elemental silver are embedded. The microstructure is remarkably stable, and in particular, prevents further sintering of the silver particles.

Full text of DOI:10.1002/ejic.201800710

A Journal of
Accepted Article
Title: Chemical behaviour of CaAg2 under ethylene epoxidation
conditions
Authors: Iryna Antonyshyn, Olga Sichevych, Karsten Rasim, Alim
Ormeci, Ulrich Burkhardt, Sven Titlbach, Stephan Andreas
Schunk, Marc Armbrüster, and Yuri Grin
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Inorg. Chem. 10.1002/ejic.201800710
Link to VoR: http://dx.doi.org/10.1002/ejic.201800710

European Journal of Inorganic Chemistry
10.1002/ejic.201800710
FULL PAPER
Chemical behaviour of CaAg
conditions
2
under ethylene epoxidation
Iryna Antonyshyn,*[a] Olga Sichevych,[a] Karsten Rasim,[a] Alim Ormeci,[a] Ulrich Burkhardt,[a] Sven
Titlbach,[b] Stephan Andreas Schunk,[b] Marc Armbrüster,[c] and Yuri Grin*[a]
Abstract: The binary compound CaAg
2
was examined as catalyst for
development of suitable support and the search of new synthesis
routes of catalyst.[11–16]
the ethylene epoxidation reaction. During the induction phase,
conversion and selectivity increase and then remain stable for several
hundred hours. The presence of ethyl chloride as promoter is crucial.
Based on a variety of studies of the epoxidation reaction, it is
evident that the electronic structure of the catalyst surface layer
has a tremendous influence on the adsorption properties of
reactants and the occurrence of specific reaction intermediates.
Depending on the adsorption strength of reactants and
intermediates as well as desorption rates of reaction products, the
rate-determining step can vary and largely govern reaction
pathways and therefore overall selectivity. Bimetallic catalysts
were assumed to be useful systems for tuning of the catalytic
activity by modifying the electronic structure. DFT calculations on
simplified models where individual Ag atoms on the surface are
replaced by atoms of another metal predict catalytic performance
2
The pristine CaAg reacts with the gaseous reactants and forms a
porous microstructure of Ca-containing oxidation products on the
surface, in which particles of elemental silver are embedded. The
microstructure is remarkably stable and in particular prevents further
sintering of the silver particles.
Introduction
Silver is
a
unique catalyst for two industrially relevant
[17]
for Ag-Cu, Ag-Au and Ag-Pd catalysts. The enhancement of the
processes: ethylene epoxidation and partial oxidation of methanol
to formaldehyde.[1] Nevertheless, the selectivity towards the target
product ethylene oxide (EO) in the epoxidation reaction amounts
to only 40–50 % for unpromoted silver catalysts[2] and can be
selectivity towards ethylene oxide for supported Ag-Cu and Ag-
Pd bimetallic catalysts compared to elemental Ag was also
experimentally confirmed.[17, 18] However, advanced investiga-
tions of Ag-Cu catalysts under reaction conditions reveal no
metallic or alloyed copper on the surface of these catalysts.
Instead, thin oxide-like “CuO” layers are detected and the
improved to about 85–90 % by using a variety of promoters, which
are typically used in technical processes.[
1, 3]
Aiming for a rational
understanding and/or design of alternative catalysts, a deep
insight into the underlying reactions and the dynamic changes of
the catalyst material are decisive. Consequently, the last decades
saw enormous efforts in studies using both experimental
techniques and computational methods. The “hot” topics are
adsorption of oxygen on Ag,[1, 4] the nature of the active
mechanism of ethylene epoxidation turns out as strongly surface-
[19]
structure dependent.
There seems to be no unique surface
structure which is responsible for the catalytic activity. The
chemistry of the Ag-Cu catalyst system is highly dynamic and not
trivial. The complexity of such catalytic systems led sometimes
also to discrepancies between theoretical calculations and
experimental evidences. For example, the Ag-Cd catalyst is
species[
1, 4c]
and reaction intermediates, the mechanism of
[5]
[
1, 6]
the role of the support,[1, 7] promoters[1, 8]
ethylene epoxidation,
and size effects.[
unselective towards ethylene oxide according to density
functional theory (DFT) calculations,[20] whereas experimental
1, 9]
Still debated in detail, there are two fundamentally different
approaches for the discussion of the ethylene epoxidation
[21]
studies show enhanced selectivity compared to elemental Ag.
This can be explained by oversimplified assumptions in the model
for the calculation: the oxidizing conditions of ethylene
epoxidation dramatically change the oxidation states of surface
atoms and thus their adsorption properties. In the same direction
points the finding that zinc oxide ZnO segregates to the surface
reaction: (i) the consideration of different types of oxygen species
and their role in the ethylene epoxidation[
“
1, 4, 10]
and (ii) the
oxametallacycle” model[1, 6] which largely relies on suitable
oxygen species from model (i). Nowadays, intense research is
focused on the nature of electrophilic oxygen species, theoretical
predictions and experimental proofs of reaction intermediates (e.g.
oxametallacycle OMC), optimization of catalyst particle size, the
[22]
during kinetic studies of supported Ag-Zn catalysts.
The need for a further understanding the oxygen incorporation
into Ag-based catalysts was the reason to investigate CaAg , one
2
of the binary compounds in the system Ca–Ag with ordered
crystal structure. The crystal structure of intermetallic compounds
(
IMCs) as the consequence of the chemical bonding makes them
[
a]
Dr. I. Antonyshyn, Dr. O. Sichevych, Dr. K. Rasim, Dr. A. Ormeci,
Dr. U. Burkhardt, Prof. Yu. Grin
Max-Planck-Institut für Chemische Physik fester Stoffe
Nöthnitzer Str. 40, 01187 Dresden (Germany)
E-mail: Iryna.Antonyshyn@cpfs.mpg.de; Juri.Grin@cpfs.mpg.de
Dr. S. Titlbach, Dr. S.A. Schunk
useful model systems for understanding the catalysts’ behaviour
under reaction conditions. This was well documented for
hydrogenation reactions. Here, a combined experimental and
computational study on the CaAg
epoxidation conditions is presented.
[23]
2
compound under ethylene
[
[
b]
c]
hte GmbH
Kurpfalzring 104, 69123 Heidelberg (Germany)
Prof. Dr. M. Armbrüster
Technische Universität Chemnitz
Fakultät für Naturwissenschaften
Institut für Chemie
Strasse der Nationen 62, 09111 Chemnitz (Germany)
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800710
FULL PAPER
Results and Discussion
hexagonal AlB
2
-type structure, the trigonal prisms around silver
atoms are strongly deformed and the coordination number of Ag
becomes 10 (6 Ca plus 4 Ag).
Characterization of as-prepared samples
2
Single-phase samples of CaAg were synthesized using the
heat treatment described in the Experimental Section. Obtained
samples were well crystallized (Figure 1, top) and brittle. They are
stable in argon atmosphere at room temperature and oxidize
slowly during exposure to air.
Figure 2. (top) Experimental (black), calculated (red), and difference (blue)
powder X-ray diffraction patterns of CaAg
2
(peak positions are shown in green);
along [010].
(
bottom) view of the crystal structure of CaAg
2
The single-phase nature of the synthesized samples was
Figure 1. Morphology of CaAg
2
particles (BSE detector, 8 kV, material contrast):
as-prepared (top) and after catalytic test (bottom).
additionally confirmed using differential thermal analysis (DTA),
Figure S1. One only sharp endothermic peak at 600 C on
o
heating corresponds to the decomposition of CaAg
2
following the
The binary compound CaAg
2
crystallizes with the KHg
2
-type
Figure 2, top presents
from
peritectic reaction CaAg  Ca Ag + L, which is in good
2
2
7
crystal structure (space group Imma).[
24]
agreement with literature data of 596
respectively.
o
C
[25a] and 597
oC,[25b–c]
the results of the crystal structure refinement of CaAg
2
powder X-ray diffraction data. Indexing of the reflections (from X-
ray powder diffraction pattern (PXRD) recorded using internal
2
These results reveal that CaAg samples can be synthesized
reproducibly with sufficient composition and structure control
using the described synthesis route.
standard LaB
6
)
and subsequent refinement yield lattice
parameters a = 4.691(1) Å, b = 7.295(2) Å, c = 8.135(1) Å, which
are in perfect agreement with literature data (a = 4.689(1) Å,
Catalytic measurements
b = 7.301(2) Å, c = 8.139(3) Å).[
24]
The summarized data are
-type of crystal structure.
= 2.731 Å, orange in Figure 2, bottom)
2
Ethylene epoxidation using CaAg as catalyst was performed
shown in Table S1 and confirm the KHg
In the “zig-zag” chains (d
2
varying temperature (T), gas hourly space velocity (GHSV), and
amount of ethyl chloride (EC) in the gas stream (Figure 3). The
selectivity towards ethylene oxide increases and reaches 60–
65 % in about 30 h, whereas the conversion level remains
relatively low during the whole experiment and does not exceed
1.5–2.5 %. The observed increase of selectivity at the beginning
indicates the presence of an induction period during which the
catalyst is transformed to its active state. Usually, the existence
1
along [100], the Ag atoms are interconnected forming planar
honeycomb-like nets. The interatomic distances Ag–Ag between
these chains are slightly longer d
bottom). The planar nets are interlinked to a three-dimensional
3D) framework by longer bonds of d = 2.900 Å along [010]. The
2
= 2.772 Å (green in Figure 2,
(
3
calcium atoms are located in the cavities of this framework.
Because of the orthorhombic deformation of the pristine
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800710
FULL PAPER
of an induction period reflects a noticeable modification, e.g. the
formation of oxygen-containing layer in oxidative catalysis.[ As
expected, a positive impact on conversion was observed at lower
26]
-
1
GHSV values (1000 h ) because of the longer contact time of the
reaction mixture with the catalyst (Figure 3).
An increase of conversion was achieved at elevated
o
temperatures (up to 280 C) but simultaneously a significant drop
of selectivity was observed. Even a slight increase of temperature
o
by 30 C reduced the selectivity from 64 to 53 %. The same trends
of conversion and selectivity depending on reaction temperature
[27]
are known for different types of silver catalysts: sponge-like,
powder,[ Ag(111) single crystals, Ag/-Al
28]
[29]
[30]
etc.
2
O
3
The reduction of EC concentration and finally its absence in the
reactive gas stream led to a significant drop of selectivity from 64
to 34% (Figure 3). Thus, the presence of the ethyl chloride
promoter is crucial for the catalytic performance.
Figure 4. CaAg
2
particle after catalytic test (BSE detector, 8 kV, material
contrast): (top) morphology of the surface; (bottom) elemental mapping (colour-
coded intensity distribution).
Figure 3. Ethylene epoxidation on CaAg
EO (violet) and selectivity towards EO (green) as a function of time on stream
TOS). Variation of experimental conditions (T, GHSV, amount of EC) are
2
: conversion of ethylene (red), yield of
SEM results are in agreement with powder X-ray diffraction data
Figure 5), which evidenced the presence of CaAg , Ca Ag and
2 2 7
elemental Ag under reaction conditions. Ca-related oxidation
products were not clearly identified via PXRD most probably due
to their small amount and low degree of crystallinity.
(
(
shown with solid lines.
Post-catalytic characterization of CaAg
2
The surface morphologies of the particles before and after
exposure to reaction conditions are completely different (Figure
1). Elemental mapping of as-synthesized sample shows a
homogeneous distribution of Ca and Ag within the particles, but
the presence of a tiny oxygen signal should also be mentioned.
o
-1
After the catalytic test (Tmax = 280 C, GHSV = 1000–2000 h ,
–4.5 ppm EC) the whole surface of the CaAg crumbs is covered
0
2
by small particles (Figure 1, bottom). Elemental mapping and
energy-dispersive X-ray spectroscopy (EDXS) at different points
and regions (Figure 4) shows the presence of elemental Ag
agglomerates (ca. 1–3 µm in size) on the surface (point 1), areas
with different Ca : Ag ratio (points 2–3) and enrichment in oxygen
of Ca-rich regions (point 3). The corresponding spectra are
presented in Figure S2. The more detailed determination of the
compositions for the Ca–Ag phases is hampered by the presence
of oxygen and the roughness of the surface.
Figure 5. X-ray diffraction patterns of CaAg
ethylene epoxidation experiment (600 h on stream, non-marked reflections
correspond to CaAg ).
2
before (black) and after (blue)
2
In general, the oxidation of CaAg
2
under ethylene epoxidation
conditions can be schematically described as
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800710
FULL PAPER
Ag catalyst. Nevertheless, the “stepwise” formed Ag catalyst
2 2 7
(CaAg  Ca Ag  Ag) is clearly more selective towards
ethylene oxide than Ag nanoparticles supported on CaO, which
were synthesized for comparison using classical impregnation
techniques (Figure 6). Parallel formation of Ag particles and Ca-
containing support yield a 3D microstructure in the subsurface
region. The latter ensures the enhanced selectivity towards
ethylene oxide and prevents the sintering of Ag particles which is
the main reason for the deactivation of industrially used Ag
catalysts.[30b, 32]
2
As Ca-containing product of the CaAg oxidation, CaO was
assumed in the above scheme, but due to the presence of carbon
dioxide and water vapour in the gas stream (products of the
undesired full oxidation of ethylene C
as well as ethyl chloride (as promoter), also the formation of
calcium hydroxide Ca(OH) , carbonate CaCO as well as chloride
CaCl cannot be excluded.
The amount of CaAg decreases significantly after prolonged
2 4 2 2 2
H + 3O  2CO + 2H O)
Electronic structure and chemical bonding
2
3
2
The crystal structure was fully optimized. The obtained lattice
parameters are a = 4.733 Å, b = 7.339 Å, c = 8.256 Å giving a 2.9
2
exposure to the gas stream, but even after 600 h it remains the
majority phase as inferred from PXRD data.
Catalytic tests at elevated temperatures (up to 350 C) led to full
%
larger unit cell volume compared to the experimental values (cf.
Characterization of as-prepared samples). The optimized atomic
coordinates, zCa = 0.5446, yAg = 0.0495, zAg = 0.1623, are in good
agreement with the experimental values (Table S1).
o
oxidation of CaAg
Ca(OH) and CaO, the corresponding reflections were identified
in PXRD patterns (Figure S3). This reveals that the formation of
Ca-related products as result of CaAg oxidation is dynamic and
2 3
to elemental Ag and a mixture of CaCO ,
2
2
depends on the amount and accessibility of water vapour and
carbon dioxide on the surface of the catalyst. The catalytic
performance of obtained catalyst resembles one after standard
o
catalytic test (Tmax = 280 C) over 600 h (S = 65%, conversion of
ethylene – 1-3%).
Figure 6. Ethylene epoxidation for different types of catalysts: conversion of
ethylene (total height of the bars) and yields of EO/CO
2
(coloured sections of the
bars) (T = 250 oC, 2.5 ppm EC) on A – CaAg
(TOS = 114 h), B1 – 5 wt. %
2
Ag@CaO (TOS = 192 h), B2 – 15 wt. % Ag@CaO (TOS = 192 h), B3 – 30 wt. %
Ag@CaO (TOS = 192 h). Selectivity values (S) are written on top of each bar
graph.
2
The short induction phase for CaAg in the ethylene epoxidation
experiment (Figure 3), as well as the appearance of small Ag
2
Figure 7. (top) Ca and Ag (4d) states for the optimized CaAg structure
compared to the Ag (4d) DOS in fcc Ag; (bottom) comparison of Ag (4d) DOS
particles on the surface reveal the in situ formation of Ag on a
in bulk, top and subsurface for the (100) surface.
2 3
CaO/Ca(OH) /CaCO support to which catalytic activity and
selectivity can be assigned. The formation of elemental Ag under
reaction conditions is also supported by the fact that the selectivity
is strongly dependent on the presence of ethyl chloride promoter
in the gas stream (Figure 3). The promotion effect of Cl is a well-
In the crystal structure of CaAg
near neighbours (distances 2 × d
the elemental fcc Ag with 12 nearest neighbours, this number of
Ag–Ag contacts in CaAg is rather small. Therefore, the implied
decrease in d – d hybridization is expected to reduce the band
width of the Ag (4d) states. Indeed, the Ag d band width of 2.5 eV
2
each silver atom has four Ag
1
, d and d ). In comparison to
2
3
described feature of elemental Ag catalysts in ethylene
2
epoxidation.[
8b, 31]
This means that under ethylene epoxidation
2
acts as precursor for the formation of an elemental
reaction CaAg
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800710
FULL PAPER
in CaAg
2
is smaller than 3.7 eV in fcc Ag (Figure 7, top). The top
very similar manner. The effective charges of the surface atoms
do not deviate significantly from the corresponding bulk values.
This finding suggests the scenario that when a species much
more electronegative than Ag is introduced to the pristine surface,
Ca atoms may prefer to form ionic bonds with that species
abandoning the Ag atoms. This may then result in the aggregation
of Ag atoms.
edge of the 4d band is located at –3.25 eV, and the center of
gravity at –4.49 eV. The corresponding values for the elemental
Ag are –4.32 and –2.7 eV, respectively.
The (100) surface has the lowest surface energy (see below
Surface energy calculations). This surface has the same
composition as the bulk, therefore the behaviour of the Ag d DOS
in the surface region is similar to that of the bulk. In a four-layer
slab model already the subsurface Ag closely resembles Ag in
2
bulk CaAg (Figure 7, bottom). The band width and center of
gravity values of subsurface Ag (4d) states are 2.5 and –4.53 eV,
essentially identical to the corresponding bulk values. The d
states of the top-layer Ag atoms shift to higher energies, the
center of gravity is at –4.21 eV, and since these Ag atoms are
missing one of their d
1
neighbours, the band width gets narrower,
2
.35 eV. However, the top edge of the d bands remains the same
in both bulk and surface.
The chemical bonding in an intermetallic compound is an
essential component of the overall assessment of that
compound’s suitability for heterogeneous catalysis.[
23c]
The
behaviour of IMC in the presence of various species under
reaction conditions can be understood by analysing the atomic
interactions.[23c] To have a better understanding of the chemical
2
behaviour of CaAg , the electronic structure calculations were
combined with a real-space study of the chemical bonding. The
topological analysis of the electron density (ED) within QTAIM[
reveals that a Ca atom loses 1.3 electrons so that each Ag atom
gains 0.65 electrons. Hence, Ag becomes more nucleophilic than
in the elemental solid. The shapes of the Ca and Ag QTAIM
basins are very different (Figure 8, top). The Ca QTAIM basin is
close to spherical implying cationic character (ionic bonding). The
boundaries between two Ag QTAIM basins are flat indicating
covalent character of interaction, while those between the Ag and
Ca ones are round as if cut from a sphere, characteristic for polar
33]
(
ionic) interactions.
This finding already emphasizes large ionic contribution to the
Ca – Ag bonding. Further insight was obtained from an analysis
of the electron localizability indicator (ELI-D). The spatial
distribution of ELI-D in CaAg
in the vicinity of the Ag–Ag contacts d
The population of the bond basin b (0.5 electrons) is formed by
two silver atoms separated by d (90 %) and by two neighbouring
Ca atoms (10 %). The bond basin b contains 1.8 electrons: 1.2
of these originate from the silver atoms separated by d (0.6
electrons each), another 0.44 are contributed by two Ag atoms
with the d distance to the latter two, and the remaining 10 % is
coming from two Ca atoms. Thus, both b and b reflect mainly
2
reveals only two types of attractors
1
and d (Figure 8, middle).
2
1
1
2
2
1
1
2
the covalent interactions between two and four Ag atoms,
respectively. These attractors reveal a formation of covalently
bonded puckered anionic silver layers perpendicular to [010]. The
calcium cations are embedded in-between.
The ELI distribution was computed for the four-layer slab model
of the (100) surface (Figure 8, bottom). The Ag atoms in the top
and bottom layer have one of their d neighbours missing,
1
therefore the according dangling bonds protruding into the
vacuum region can be observed on both sides of the slab. The
bonding features inside the slab correspond to bulk features in a
Figure 8. Chemical bonding in CaAg
their effective charges; (middle) The ELI-D distribution in CaAg
the three shortest Ag–Ag distances d , d , d : isosurface value is 0.93, b
stand for the bonds, four Ag atoms contributing to the b bond are identified;
bottom) The ELI-D distribution for the four-layer slab model of the (100) surface
for CaAg : the orange isosurface value is 1.045, it represents the dangling
bonds; the yellow isosurface value is 0.954 showing the localization domain of
2
: (top) The Ag and Ca QTAIM basins with
unit cell with
and
2
1
2
3
1
b
(
2
2
2
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800710
FULL PAPER
the combined dangling, b
showing the b - and b -like bonds in the inner region of the slab; the red
isosurface value is 0.931 which depicts the b -type bond formed by the surface
and subsurface Ag atoms. The black lines show the unit cell (top and middle)
and calculated part of the structure (bottom).
2 1
and the b bonds; the green isosurface value is 0.938
In two latter cases the crystal structure allows for different surface
terminations. In order to obtain the surface energy of one surface
structure only, the upper and lower sides of the simulation slabs
have to be symmetric. However, this requirement leads to non-
stoichiometric (with respect to the 2:1 Ag:Ca ratio) slabs for the
cases (010)-I and (001)-III. The individual surface energies of
these non-stoichiometric surfaces (Figure 10) thus depend on the
chemical potential of the two components which is limited by the
heat of formation of the bulk compound. In this sense, a low-lying
non-stoichiometric surface energy, such as the Ca-rich one for
1
2
2
Surface energy calculations
The three principal directions [100], [010] and [001] were
chosen for construction of possible surface termination in CaAg
Figure 9).
2
(
(010)-I, indicates a tendency to transform Ag-rich surfaces after
cleaving into the Ca-terminated variant by Ag-precipitation – e.g.
upon annealing. The most favourable cleavage energies occur for
flat surface terminations (variants I). In general, one observes
almost identical values of the most favourable surface energies
for different terminations, implying the equivalent ability to cleave
in all directions. Furthermore, a relaxation of the pristine surfaces
changes the surface energies only to a negligibly small degree.
2
Figure 10. Surface energies of the terminating surfaces of CaAg (unrelaxed).
In view of the epoxidation reaction, the character of the surfaces
under oxidizing conditions is of particular importance. A number
of fully relaxed oxygen adsorption configurations was calculated.
All results are related to di-oxygen at the GGA-PBE level of theory.
Given the electropositive character of the calcium, extremely
large adsorption energies were observed for all configurations.
Depending on the particular local environment, the energy gain
varies between 2 and 4.6 eV per oxygen atom. Adsorption on Ag
(111) in comparison yields merely a gain of 0.45 eV. A favourable
local adsorption energy correlates with the number of locally
available Ca atoms. Oxygen atoms occupying the centers of Ca-
tetrahedra or Ca-triangles show the highest local adsorption
energy. Such large adsorption energies will therefore provide
a sizable driving force for the transformation of the intermetallic
compound to silver and CaO.
A more realistic view on this matter is given by an “ab initio
thermodynamics approach”.[34] It relates adsorption energies
Figure 9. Cleavage of the crystal structure of CaAg
2
and formation of possible
terminating surfaces in different crystallographic directions.
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800710
FULL PAPER
computed by DFT to the partial pressure p of a gas species (O
this case) via the expression:
2
in
regions. The configurations of the lowest energy (Figure 11,
bottom) already contain oxygen atoms beneath the surface. The
indicated dashed lines correspond to the “last” surface structure
that accommodates oxygen atoms exclusively at the surface by
which the region of the first atomic layer in a geometrical sense is
meant. Concomitantly with oxygen in-migration one observes the
expulsion of silver out of the quite small surface unit cell. Any
given Ca atom tends to optimize its oxygen coordination. A prime
example is shown for the (100) termination in which 3-atomic
silver clusters form.
where γads is the adsorption energy per surface area, Nox the
number of oxygen atoms adsorbed onto the surface area A of the
slab and Eads the adsorption energy per oxygen atom with respect
to the oxygen molecule. μ is tabulated and accounts for the
difference of entropy between the O
2
molecules in the gas phase
The experimental and theoretical results above clearly indicate
and the immobile oxygen atoms at the surface. The values are
taken from the NIST-JANAF thermochemical tables.[ Each line
in Figure 11, top corresponds to a particular calculated adsorption
state characterized by the number of adsorbed oxygen atoms Nox
and their local configuration.
Configurations with the highest density of oxygen atoms by
definition have the steepest negative slope with increasing
oxygen partial pressure. The different absolute values thus
correspond to the particular geometrical arrangements that can
be more or less favourable. Representative high-coverage
surface configurations are shown in Figure 11, bottom.
a pronounced instability of CaAg surfaces against reaction with
2
35]
oxygen under epoxidation conditions which is governed by the
extremely strong driving force towards the formation of calcium
oxide. Potentially passivating layers that would qualify as
energetically stable and geometrically dense were not found. On
the one hand, this instability is consequently coupled to a
segregation of silver atoms. On the other hand, the similar
oxidation in different crystallographic directions may produce a 3D
microstructure of oxidation products on the surface, which may
limit the size of the Ag particles and thus ensures stability of the
catalytic system over a long time at reaction conditions (Figure
12). Therefore, one may consider the intermetallic compound
CaAg under epoxidation conditions as model precursor for Ag-
2
based catalysts.
Figure 12. Microstructure of CaAg
experiment with a porous 3D layer of CaO/Ca(OH)
elemental Ag. Inset: schematic representation of the particle cross section.
2
particle after ethylene epoxidation
/CaCO with particles of
Figure 11. (top) Adsorption energies per surface area at 600 K of oxygen atoms
2
3
(
2 2
wrt. O ) on the most stable pristine surfaces (c. f. Figure 10) of CaAg .
Lines with increasing slope correspond to an increasing oxygen coverage (Nox).
Dashed lines indicate structures with full surface coverage and no subsurface
oxygen atoms. (bottom) Structural illustrations for typical oxygenated surface
configurations for (100), (010) and (001) facets (from left to right).
Conclusions
2
CaAg was tested as a catalyst for ethylene epoxidation. The
The comparable degree of stabilization for each surface
termination allows to conclude that CaAg particles react in a
2
presence of more electronegative oxygen in the gas stream and
high affinity of calcium to oxygen lead to the formation of calcium
oxide and a substantial segregation of silver to the surface. In
addition to oxygen, carbon dioxide and water vapour from the
undesired total combustion are in contact with the surface of
similar manner in all directions in agreement with experimental
data (Figure 1, bottom). The corrosion process can proceed
unhindered as the resulting structures are quite open and allow
an easy access of the gas phase to the subsurface regions to
oxidize deeper-lying calcium atoms: the simple addition of more
and more oxygen atoms above the already oxygen-covered
surface slabs leads to an oxygen migration (during the DFT
relaxation, hence without a kinetic barrier) towards the subsurface
CaAg
2
leading to a variety of possible Ca-related products of
oxidation, e.g. CaO, Ca(OH) and CaCO . PXRD and SEM
particles after the catalytic test as well as
calculations of adsorption energies of oxygen atoms at different
CaAg
2
2
3
analysis of CaAg
2
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800710
FULL PAPER
partial pressures confirmed the expulsion of Ag towards the
surface and formation of 3D microstructure, which hinders further
switch from nitrogen to reaction mixture (35 vol. % ethylene, 7 vol. %
oxygen, 5 vol. % argon (internal standard) and 53 vol. % nitrogen as carrier
gas). As promoter ethyl chloride (EC) was added in a controlled way to the
bulk oxidation of CaAg
temperatures (up to 350 C). Continuous dynamics of Ag particles
2
. The latter can be only achieved at higher
o
gas mixture. Different experiments varying the temperature (230–350 C),
o
-
1
gas hourly space velocity (GHSV = 1000–2000 h ) and amount of ethyl
chloride (0–4.5 ppm) were performed. All tests were carried out at ambient
pressure. Gas products analysis was done via gas chromatography (GC)
and mass spectrometry (MS) in analogy to the procedures published
elsewhere.[38]
which are homogeneously distributed in CaO/Ca(OH)
D-support prevents their sintering and provide uniform values of
selectivity towards ethylene oxide during almost 600 h of
experiment. The comparison of CaO/Ca(OH) /CaCO supported
Ag catalyst (the product of CaAg oxidation) with an Ag/CaO
2 3
/CaCO
3
2
3
The Fritz-Haber Institute ab initio molecular simulations package (FHI-
aims)[39] was employed for investigating the electronic structure and
chemical bonding features, geometry optimizations and surface
calculations. The FHI-aims is an all electron, full-potential electronic
structure code that employs atom-centered numerical basis functions.
Exchange-correlation effects were taken into account within the
generalized gradient approximation (GGA) to the density functional theory
as parametrized by Perdew-Burke-Ernzerhof (PBE).[40] The scalar
relativistic effects are included by the zero order regular approximation
2
catalyst (prepared using a classical impregnation method) reveals
the unique catalytic behaviour of the Ag-based catalyst obtained
using CaAg as a precursor.
2
Experimental Section
(
ZORA) approach.[41]
Taking into account the high vapour pressure and chemical reactivity of
calcium, all preparation steps were carried out in argon-filled glove boxes.
Silver shot (1–6 mm, Alfa Aesar, 99.999 %) and dendritic calcium pieces
Chemical bonding analysis in position space was performed within the
electron localizability approach combining topological analysis of electron
density (ED) based on the Quantum Theory of Atoms in Molecules
(Alfa Aesar, 99.98 %) were weighted in molar ratio 2:1 and placed into
[42]
(
QTAIM)[33] and electron localizability indicator (ELI). ELI was calculated
tantalum containers which were sealed under argon. The latter were
placed into quartz glass ampoules and sealed under vacuum. The initial
in the ELI-D representation[ by an own code using an interface to the
FHI-aims.[44] Topological analysis of the ED and the ELI-D was carried out
by the program DGrid.[45]
43]
o
reaction was carried out in a resistance furnace by heating up to 1000 C,
o
dwelling at this temperature for 1 h, cooling down to 570 C and
All surface energy and adsorption calculations were carried out using
the slab approach within FHI-aims. The slabs contained at least five atomic
layers with a vacuum region of approximately 20 Å. 1×1 surface unit cells
subsequent annealing at this temperature for 340 h. After above-
mentioned heat treatment the samples were quenched in ice water by
breaking of quartz glass.
Supported Ag nanoparticles were prepared by incipient wetness
impregnation of CaO support (Sigma Aldrich) with an aqueous solution of
of the (100), (010) and (001) orientations (8.26 × 7.33, 4.73 × 8.26, or 4.73
2
×
7.34 Å ) were used throughout. These dimensions correspond to the
) followed by drying (T = 80 oC) and calcination (T =
80 C) steps. The loadings of silver were 5, 15 and 30 wt. %.
optimized bulk lattice parameters (GGA-PBE). In the case of one-sided,
thus asymmetrical adsorption configurations, a dipole-field correction as
implemented in FHI-aims was employed. Surface energies and all
adsorption energies were calculated by relaxing the adsorbate and all
atoms of the top three surface layers.
silver nitrate (AgNO
2
3
o
Powder X-ray diffraction (PXRD) patterns were recorded using a Huber
1
Imaging Plate Guinier Camera G670 (Cu Kα radiation, λ = 1.54059 Å) and
lanthanum hexaboride LaB (a = 4.1569 Å) as internal standard. The
6
phase analysis via comparison of experimental and calculated PXRD
patterns was done using WinXPow software.[36] For peak positions
determination, indexing of powder X-ray diffraction patterns and Rietveld
refinement of the crystal structure, WinCSD package[37] was employed.
Acknowledgements
Differential
thermal
analysis / thermogravimetry
(DTA/TG)
The authors are thankful to P. Scheppan and M. Eckert for SEM
measurements, S. Scharsach and M. Schmidt for thermal
analysis, S. Hückmann for X-ray diffraction experiments.
measurements under Ar flow were done using Netzsch STA 449 C Jupiter
F1 apparatus, situated in an Ar-filled box. The heating up to 650 oC and
subsequent cooling to room temperature with a rate of 10 oC/min were
carried out using tantalum crucibles without cap.
Scanning electron microscopy (SEM) using JEOL 7800 with an attached
EDX/EBSD system (Quantax 400, Bruker, Silicon-Drift-Detector (SDD),
acceleration voltages 0.1 kV – 30 kV) was utilized for several purposes: (i)
checking of initial sample quality; (ii) determination of the composition of
synthesized compound; (iii) elemental mapping of the sample before and
after the catalytic test; (iv) morphology imaging of the sample surface after
its exposure to reaction conditions. All preparation steps for SEM
investigation were carried out in Ar atmosphere and a dedicated shuttle-
transfer system was used for sample insertion into the microscope. The
samples were cold fixed in silver-containing epoxy resin or directly
mounted on graphite tabs without embedding (the latter was used in case
of morphology studies). SiC papers and diamond powder with grain sizes
of 3 µm or smaller were used for surface polishing.
Keywords: intermetallic compound • heterogeneous catalysis •
epoxidation • chemical bonding • adsorption energy
[1] V.I. Bukhtiyarov, A. Knop-Gericke in Nanostructured catalysts. Selective
oxidations, Chapter 9: Ethylene epoxidation over silver catalysts (Eds.: C.
Hess, R. Schlögl), RSC Publishing, 2011, pp. 214–247, and references
therein.
[2] R.A. van Santen, H.P.C.E. Kuipers, Adv. Catal. 1987, 35, 265–321.
[3] a) J.G. Serafin, A.C. Liu, S.R. Seyedmonir, J. Mol. Catal. A – Chem. 1998,
131, 157–168; b) W.E. Evans, P.I. Chipman (Shell Oil Company), US
Patent 6717001 B2, 2004.
[4] a) T.C.R. Rocha, A. Oestereich, D.V. Demidov, M. Hävecker, S. Zafeiratos,
G. Weinberg, V.I. Bukhtiyarov, A. Knop-Gericke, R. Schlögl, Phys. Chem.
Chem. Phys. 2012, 14, 4554–4564; b) T.E. Jones, T.C.R. Rocha, A. Knop-
Gericke, C. Stampfl, R. Schlögl, S. Piccinin, Phys. Chem. Chem. Phys.
2015, 17, 9288–9312; c) T.E. Jones, T.C.R. Rocha, A. Knop-Gericke, C.
Stampfl, R. Schlögl, S. Piccinin, ACS Catal. 2015, 5, 5846–5850; d) T.E.
Jones, R. Wyrwich, S. Böcklein, E.A. Carbonio, M,T. Greiner, A.Yu.
Catalytic tests were performed using crushed bulk samples (ca. 1 cm3)
sieved to the fraction of 500–1000 µm in particle size. The catalyst was
filled between two layers of α-Al O (Alodur WSK F60, 250–315 µm) inside
2 3
the tubular stainless steel reactor (R0139). Pre-conditioning was carried
out via heating up to 160 oC (rate 1 C/min) under nitrogen flow and
dwelling for 15 h. Further heating up to 230 oC was accompanied with a
o
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800710
FULL PAPER
Klyushin, W. Moritz, A. Locatelli, T.O. Mentes, M.A. Niño, A. Knop-Gericke,
M. Friedrich, D. Teschner, G. Wowsnick, M. Hahne, P. Gille, L.
Szentmiklósi, M. Feuerbacher, M. Heggen, F. Girgsdies, D. Rosenthal, R.
Schlögl, Yu. Grin, Nature Mater. 2012, 11, 690–693; c) M. Armbrüster, R.
Schlögl, Yu. Grin, Sci. Technol. Adv. Mater. 2014, 15, 034803–1–17.
[24] M. Pani, M.L. Fornasini, F. Merlo, Z. Kristallogr. 2007, 222, 218–225.
[25] a) N. Baar, Z. Anorg. Allg. Chem. 1911, 70, 352–394; b) W.A. Alexander,
L.D. Calvert, A. Desaulniers, H.S. Dunsmore, D.F. Sargent, Can. J. Chem.
1969, 47, 611–614; c) M.R. Baren, Bull. Alloys Phase Diagr. 1988, 9, 228–
229.
R. Schlögl, S. Güntner, J. Wintterlin, S. Piccinin, ACS Catal. 2018, 8, 3844–
3
852; e) E.A. Carbonio, T.C.R. Rocha, A.Yu. Klyushin, I. Piš, E. Magnano,
S. Nappini, S. Piccinin, A. Knop-Gericke, R. Schlögl, T.E. Jones, Chem. Sci,
018, 9, 990–998.
2
[
5] a) M. Egashira, R.L. Kuczkowski, N.W. Cant, J. Catal. 1980, 65, 297–310;
b) S. Linic, M.A. Barteau, J. Am. Chem. Soc. 2002, 124, 310–317; c) S.
Linic, M.A. Barteau, J. Catal. 2003, 214, 200–212; d) S. Linic, M.A. Barteau,
J. Am. Chem. Soc. 2003, 125, 4034–4035.
[
[
6] a) M.O. Ozbek, I. Onal, R.A. van Santen, J. Catal. 2011, 284, 230–235; b)
M.O. Ӧzbek, R.A. van Santen, Catal. Lett. 2013, 143, 131–141.
7] a) A. Chongterdtoonskul, J.W. Schwank, S. Chavadej, J. Mol. Catal. A –
Chem. 2012, 358, 58–66; b) A. Chongterdtoonskul, J.W. Schwank, S.
Chavadej, Catal. Lett. 2012, 142, 991–1002.
[26] a) C. Stampfl, M.V. Ganduglia-Pirovano, K. Reuter, M. Scheffler, Surf. Sci.
2002, 500, 368–394; b) K. Reuter, M. Scheffler, Appl. Phys. A 2004, 78,
793–798; c) J.K. Nørskov, M. Scheffler, H. Toulhoat, MSR Bull. 2006, 31,
669–674.
[27] N.W. Cant, W.K. Hall, J. Catal. 1978, 52, 81–94.
[
8] a) M.O. Ӧzbek, I. Ӧnal, R.A. van Santen, Chem. Cat. Chem. 2013, 5, 443–
[28] R.A. van Santen, J. Moolhuysen, W.M.H. Sachtler, J. Catal. 1980, 65, 478–
480.
4
51; b) T.C.R. Rocha, M. Hävecker, A. Knop-Gericke, R. Schlögl, J. Catal.
014, 312, 12–16; c) M.T. Greiner, T.C.R. Rocha, B. Johnson, A. Klyushin,
2
[29] R.B. Grant, R.M. Lambert, J. Catal. 1985, 92, 364–375.
[30] a) A. Ayame, Y. Uchida, H. Ono, M. Miyamoto, T. Sato, H. Hayasaka, Appl.
Catal. A – Gen. 2003, 244, 59–70; b) G. Boskovic, N. Dropka, D. Wolf, A.
Brückner, M. Baerns, J. Catal. 2004, 226, 334–342; c) J. Jiang, T. Xu, Y.
Li, X. Lei, H. Zhang, D.G. Evans, X. Sun, X. Duan, Bull. Korean Chem. Soc.
2014, 35, 1832–1836.
A. Knop-Gericke, R. Schlögl, Z. Phys. Chem. 2014, 228, 521–541; d) W.
Diao, C.D. DiGiulio, M.T. Schaal, S. Ma, J.R. Monnier, J. Catal. 2015, 322,
1
4–23.
9] J. Lu, J.J. Bravo-Suárez, A. Takahashi, M. Haruta, S.T. Oyama, J. Catal.
005, 232, 85–95.
[
[
[
2
10] T.E. Jones, T.C.R. Rocha, A. Knop-Gericke, C. Stampfl, R. Schlögl, S.
Piccinin, Phys. Chem. Chem. Phys. 2014, 16, 9002–9014.
[31] a) C.T. Campbell, M.T. Paffett, Appl. Surf. Sci. 1984, 19, 28–42; b) C.T.
Campbell, J. Catal. 1986, 99, 28–38; c) S.A. Tan, R.B. Grant, R.M. Lambert,
J. Catal. 1987, 106, 54–64; d) K.L. Yeung, A. Gavriilidis, A. Varma, M.M.
Bhasin, J. Catal. 1998, 174, 1–12.
11] T.E. Jones, R. Wyrwich, S. Böcklein S., T.C.R. Rocha, E.A. Carbonio, A.
Knop-Gericke, R. Schlögl, S. Günther, J. Wintterlin, S. Piccinin, J. Phys.
Chem. C 2016, 120, 28630–28638.
[32] a) A.E.B. Presland, G.L. Price, D.L. Trimm, J. Catal. 1972, 26, 313–317; b)
G.L. Montrasi, G.R. Tauszik, M. Solari, G. Leofanti, Appl. Catal. 1983, 5,
359–369; c) E. Ruckenstein, S.H. Lee, J. Catal. 1988, 109, 100–119; d)
X.G. Zhou, W.K. Yuan, Chem. Eng. Sci. 2004, 59, 1723–1731; e) G.
Boskovic, D. Wolf, A. Brückner, M. Baerns, J. Catal. 2004, 224, 187–196.
[33] R.F.W. Bader, Atoms in Molecules, A Quantum Theory, Clarendon Press,
Oxford, 1995.
[
[
[
[
12] T. Suttikul, T. Sreethawong, H. Sekiguchi, S. Chavadej, Plasma Chem.
Plasma Process 2011, 31, 273–290.
13] J.E. van den Reijen, S. Kanungo, T.A.J. Welling, M. Versluijs-Helder M.,
T.A. Nijhuis, K.P. de Jong, P.E. de Jongh, J. Catal. 2017, 356, 65–74.
14] A. Indra, M. Greiner, A. Knop-Gericke, R. Schlögl, D. Avnir, M. Driess,
ChemCatChem 2014, 6, 1935–1939.
15] a) D. Sun, H. Wang, G. Zhang, J. Huang, Q. Li, RSC Adv. 2013, 3, 20732–
[34] K. Reuter, M. Scheffler, Phys. Rev. Lett. 2003, 90, 046103–1–4.
[35] M. W. Chase, NIST-JANAF Thermochemical Tables, National Institute of
Standards and Technology, Washington (U.S.), 1998.
2
0737; b) X. Jing, H. Wang, H. Chen, J. Huang, Q. Li, D. Sun, RSC Adv.
014, 4, 27597–27603; c) X. Jing, J. Huang, H. Wang, M. Du, D. Sun, Q.
2
Li, Chem. Eng. J. 2016, 284, 149–157.
[36] WinXPow (Version 2.25), STOE and Cie GmbH, Darmstadt (Germany),
2009.
[
16] a) S.S. Sangaru, H. Zhu, D.C. Rosenfeld, A.K. Samal, D. Anjum, J.-M.
Basset, ACS Appl. Mater. Interfaces 2015, 7, 28576–28584; b) A. Ramirez,
J.L. Hueso, H. Suarez, R. Mallada, A. Ibarra, S. Irusta, J. Santamaria,
Angew. Chem. Int. Ed. 2016, 55, 11158–11161; Angew. Chem. 2016, 128,
[37] L. Akselrud, Yu. Grin, J. Appl. Crystallogr. 2014, 47, 803–805.
[38] S.A. Schunk, D. Demuth, A. Cross, O. Gerlach, A. Haas, J. Klein, J.M.
Newsam, A. Sundermann, W. Stichert, W. Strehlau, U. Vietze, T. Zech in
High-throughput screening in heterogeneous catalysis, Chapter 2:
Mastering the challenges of catalyst screening in high-throughput
experimentation for heterogeneously catalyzed gas-phase reactions (Eds.:
A. Hagemeyer, P. Strasser, A.F. Volpe), Wiley-VCH, 2004, pp. 19–62.
[39] V. Blum, R. Gehrke, F. Hanke, P. Havu, V. Havu, X. Ren, K. Reuter, M.
Scheffler, Comput. Phys. Commun. 2009, 180, 2175–2196.
[40] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865–3868.
[41] E. van Lenthe, E.J. Baerends, J.G. Snijders, J. Chem. Phys. 1994, 101,
9783–9792.
1
1324–11327; c) D. Kim, Y. Lee, Y. Kim, K. Mingle, J. Lauterbach, D.A.
Blom, T. Vogt, Y. Lee, Chem. Eur. J. 2018, 24, 1041–1045.
[
[
17] S. Linic, J. Jankowiak, M.A. Barteau, J. Catal. 2004, 224, 489–493.
18] a) J.T. Jankowiak, M.A. Barteau, J. Catal. 2005, 236, 366–378; b) J.T.
Jankowiak, M.A. Barteau, J. Catal. 2005, 236, 379–386; c) J.C. Dellamorte,
J. Lauterbach, M.A. Barteau, Catal. Today 2007, 120, 182–185; d) J.C.
Dellamorte, J. Lauterbach, M.A. Barteau, Appl. Catal. A – Gen. 2011, 391,
2
81–288.
[
19] a) S. Piccinin, S. Zafeiratos, C. Stampfl, T.W. Hansen, M. Hävecker, D.
Teschner, V.I. Bukhtiyarov, F. Girgsdies, A. Knop-Gericke, R. Schlögl, M.
Scheffler, Phys. Rev. Lett. 2010, 104, 035503–1–4; b) S. Piccinin, N.L.
Nguyen, C. Stampfl, M. Scheffler, J. Mater. Chem. 2010, 20, 10521–10527;
c) N.L. Nguyen, S. Piccinin, S. de Gironcoli, J. Phys. Chem. C 2011, 115,
[42] M. Kohout, Int. J. Quantum Chem. 2004, 97, 651–658.
[43] a) M. Kohout, F.R. Wagner, Yu. Grin, Int. J. Quantum Chem. 2006, 106,
1499–1507; b) M. Kohout, Faraday Discuss. 2007, 135, 43–54.
[44] S.A. Villaseca, A. Ormeci, S.V. Levchenko, R. Schlögl, Yu. Grin, M.
Armbrüster, Chem. Phys. Chem. 2017, 18, 334–337.
1
0073–10079; d) N.L. Nguyen, S. de Gironcoli, S. Piccinin, J. Chem. Phys.
013, 138, 184707–1–8; e) M.T. Greiner, J. Cao, T.E. Jones, S. Beeg, K.
2
[45] M. Kohout, Program DGRID, version 4.6, Radebeul (Germany), 2011.
Skopurska, E.A. Carbonio, H. Sezen, M. Amati, L. Gregoratti, M.-G.
Willinger, A. Knop-Gericke, R. Schlögl, ACS Catal. 2018, 8, 2286–2295.
20] J.C. Dellamorte, M.A. Barteau, J. Lauterbach, Surf. Sci. 2009, 603, 1770–
[
1
775.
[
[
21] B.K. Bonin, X.E. Verykios, J. Catal. 1985, 91, 36–43.
22] N. Toreis, X.E. Verykios, S.M. Khalid, G. Bunker, Z.R. Korszun, J. Catal.
1
988, 109, 143–155.
[
23] a) M. Armbrüster, K. Kovnir, M. Behrens, D. Teschner, Yu. Grin, R. Schlögl,
J. Am. Chem. Soc. 2010, 132, 14745–14747; b) M. Armbrüster, K. Kovnir,
This article is protected by copyright. All rights reserved.

European Journal of Inorganic Chemistry
10.1002/ejic.201800710
FULL PAPER
FULL PAPER
Oxidation of CaAg
under ethylene epoxidation conditions
leads to the formation of 3D
2
intermetallic
Dr. Iryna Antonyshyn,* Dr. Olga
Sichevych, Dr. Karsten Rasim, Dr.
Alim Ormeci, Dr. Ulrich Burkhardt, Dr.
Sven Titlbach, Dr. Stephan Andreas
Schunk, Prof. Dr. Marc Armbrüster and
Prof. Yuri Grin*
microstructure
composed
of
elemental Ag particles embedded in
the less-crystallized support of
2 3
CaO/Ca(OH) /CaCO . This ensures
Page No. – Page No.
the enhanced selectivity towards
ethylene oxide and prevents the
sintering of Ag particles which is the
main reason of deactivation of
industrially used Ag catalysts.
Chemical behaviour of CaAg under
ethylene epoxidation conditions
2
This article is protected by copyright. All rights reserved.