This	document	is	the	Accepted	Manuscript	version	of	a	Published	Work	that	appeared	in	final	form	in	Journal	of	Catalysis	353	(2017)	
171e180	hGp://dx.doi.org/10.1016/j.jcat.2017.07.013
Journal of Catalysis 353 (2017) 171–180
Contents lists available at ScienceDirect

Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat

Mechanism of ethylene oxychlorination over ruthenium oxide
M.D. Higham a, M. Scharfe b, M. Capdevila-Cortada a, J. Pérez-Ramírez b,⇑, N. López a,⇑
a
b

Institute of Chemical Research of Catalonia, ICIQ, The Barcelona Institute of Science and Technology, Av. Països Catalans 16, 43007 Tarragona, Spain
Institute for Chemical and Bioengineering, Department of Chemistry and Applied Biosciences, ETH Zurich, Vladimir-Prelog-Weg 1, 8093 Zurich, Switzerland

a r t i c l e

i n f o

a b s t r a c t

Article history:
Received 21 May 2017
Revised 16 July 2017
Accepted 17 July 2017
Available online 4 August 2017

The oxychlorination of ethylene is an industrially relevant process within the manufacture of polyvinyl
chloride (PVC). Although RuO2 is the best performing catalyst for the Deacon process (4HCl + O2 ?
2H2O + 2Cl2), experiments demonstrate a modest activity in the selective oxychlorination to vinyl chloride, favouring oxidation and polychlorinated saturated products. From the computational modelling
three main contributions are found to control the performance: (i) coverage effects that alter the configuration of intermediates; (ii) the monodimensional arrangement of the active sites, in which the reaction
of coadsorbed species works on a ‘‘first-come, first-served” basis; and (iii) the high reactivity of the oxygen species. Competition between oxidation and chlorination processes results in variable selectivity,
depending on the reaction conditions (particularly temperature and reactant partial pressures), which
influence the surface composition. From the analysis of the complex reaction network, the essential
requirements for a good oxychlorination catalyst are formulated.
Ó 2017 Elsevier Inc. All rights reserved.

Keywords:
Ruthenium dioxide
Ethylene oxychlorination
DFT
HCl oxidation
Vinyl chloride
Coverage effects

1. Introduction
RuO2 is an active catalytic material in a variety of processes,
including catalytic CO [1–3], NH3 [4], alcohol [5], and Hg oxidation
[6], as well as in electrochemical phenolic wastewater oxidation
[7], or as an anode in water splitting cells [8]. In addition to its electrochemical applications [9,10], one of the most significant industrial uses of RuO2 is in HCl oxidation (Deacon process), where it is
the best performing catalyst to produce molecular chlorine at low
temperature [11–14]. Therefore, the chemistries of oxygen and
chlorine on the RuO2 surfaces are intertwined, as evidenced by
the linear scaling relationships (Cl and O energies scale one with
the other) [15,16] and thus the material is prone to exhibit a complex selectivity behaviour. In fact, when seawater is employed in a
RuO2-based electrochemical water splitting cell, the selectivity
towards the O2 evolution reaction is compromised as Cl2 evolution
emerges as a competitive path [10,15,17,18].
Given this ability to activate adsorbates such as HCl, O2, and
hydrocarbons [19], RuO2 could potentially be a suitable catalyst
for chlorination or oxychlorination reactions. This is especially relevant as polyvinyl chloride (PVC) production represented a market
of 53 billion USD in 2015, still expected to grow at an annual rate of
5% in the coming years [20]. Therefore, vinyl chloride (VCM), the

⇑ Corresponding authors.

E-mail addresses:
(N. López).

jpr@chem.ethz.ch

(J.

Pérez-Ramírez),

http://dx.doi.org/10.1016/j.jcat.2017.07.013
0021-9517/Ó 2017 Elsevier Inc. All rights reserved.

nlopez@iciq.es

monomer of PVC or its intermediate 1,2-dichloroethane (commonly known as ethylene dichloride, EDC) are target synthetic
platform molecules. To date, VCM is either produced through
hydrochlorination of acetylene [21,22], or derived from EDC
through energy-intensive thermal cracking [23]. The latter process
releases HCl in equimolar amounts to VCM, making HCl recovery
crucial for sustainable large-scale PVC production. One way to
recycle this HCl is by using it as a reactant together with O2 in ethylene oxychlorination in order to produce EDC (or directly VCM)
[13,23,24].
The current industrial catalyst for ethylene oxychlorination
comprises cupric chloride (CuCl2) as the active phase, impregnated
on a porous support such as alumina, and promoted by numerous
additives to reduce the metal loss and/or improve the activity and
selectivity [25–28]. However, even though these cupric chloride
catalysts are widely applied in industry due to their remarkable
selectivity towards EDC (99%), they still suffer from active phase
volatilisation or particle stickiness [29]. Previous efforts to describe
the mechanism of ethylene oxychlorination on classical CuCl2based catalysts usually included in situ and operando characterisation by XAFS, XANES, and IR or EPR spectroscopies [30–35]. Briefly,
the process follows a redox mechanism in three steps: chlorination
of ethylene by reduction of CuCl2 to CuCl, that then forms a cupric
oxychloride, and finally CuCl2 is recovered by re-chlorination of the
oxychloride with HCl [24,26,27,30,31,33–43].
More recently, metal oxides like RuO2, IrO2, and CeO2 were
reported as active but very diverse in selectivity [44,45]. Whilst

172

M.D. Higham et al. / Journal of Catalysis 353 (2017) 171–180

IrO2 leads only to combustion of ethylene to CO and CO2, RuO2 and
CeO2 exhibited formation of EDC and VCM (the combined selectivity at same conversion level of 10% is 10% and 98%, respectively)
along with COx [44]. Alternatively, some catalysts based on lanthanide oxychlorides were found active for one-step VCM synthesis, of which EuOCl emerged as an excellent oxychlorination
catalyst, achieving a VCM selectivity up to 100%, but at comparatively lower conversions [45]. Nevertheless, the mechanism of oxychlorination on these materials remains unknown, which hinders
the catalyst design to overcome their specific drawbacks. A recent
review highlights the challenges presented by catalytic oxyhalogenation processes [46].
In this study, we intend to shed light on the complex ethylene
oxychlorination reaction network by combining state-of-the-art
Density Functional Theory (DFT) calculations and a selected set
of catalytic experiments to gain insight on the behaviour of RuO2
in a broad range of operating conditions. Thus far, there is no
mechanistic study of oxides in ethylene oxychlorination and
RuO2 presents a relatively uncomplicated electronic (conductive)
and geometric (rutile) structure, allowing the extended investigation of the complete reaction network, which comprises (de-)
chlorination, (de-)hydrogenation, and oxidation reactions (encompassing a total of 34 elementary steps). The main network characteristics could in the future be simplified when investigating other
materials like CeO2 or EuOCl, which are more appealing but present a more complex electronic structure. Furthermore, RuO2 exhibits both oxychlorination and combustion activity, making it an
ideal system to investigate the interplay between these two competing processes. Whilst recent theoretical studies suggest that
surface uptake of Cl by RuO2 is limited [47], previous combined
experimental and theoretical investigations found that the Cl
uptake is significant and does play some role in affecting HCl oxidation performance. Specifically, operando prompt gamma activation analysis (PGAA) studies show that under typical reaction
conditions, Cl is present in considerable quantities in the catalyst
sample [48,49]. Moreover, more recent theoretical ab initio thermodynamics studies suggest that the experimentally observed Cl
uptake is limited to surface replacement of bridging oxygen atoms
and the undercoordinated Ru positions and is present under typical
reaction conditions [50]. Hence, whilst the kinetics of RuO2 in oxychlorination are as yet unknown, it is reasonable to assume that
the degree of chlorination may very well play an important role
in oxychlorination activity, as much as it is believed to affect HCl
oxidation performance [48]. Moreover, oxychlorination has profited once before from the inheritance of the original Deacon catalyst [25–28] and bears the potential to profit again from the recent
major leap in HCl oxidation, i.e. the implementation of outstanding
low-temperature catalysts RuO2/TiO2-rutile and RuO2/SnO2cassiterite [14,51]. As such, this progress inspires the present work
to develop a full understanding of ethylene oxychlorination on
RuO2 and derive intrinsic requirements for a good VCM production
catalyst.

2. Materials and methods
2.1. Computational details
Density Functional Theory (DFT) as implemented in the Vienna
Ab initio Simulation Package (VASP), version 5.3.3, was applied to
bulk RuO2 [52,53]. The ionic positions and cell volume were optimised for a cell containing two formula units, using a planewave cut-off of 600 eV and a k-point sampling mesh of
9 Â 9 Â 13, employing the Monkhorst and Pack scheme [54]. The
optimised bulk lattice parameters obtained (a = b = 4.520 Å,
c = 3.118 Å) are in good agreement with experimental values

obtained from single-crystal X-ray diffraction; the difference
between calculated lattice parameters and the experimental values
is less than 0.7% [55]. From the optimised bulk structure, slabs representing the lowest energy rutile (1 1 0) facet of RuO2 were constructed. Previous works have established that the (1 1 0) facet is
the most stable, and hence it forms the largest contribution to
the surface of polycrystalline RuO2, therefore being the most representative of the main component of surface activity [56]. A p(4 Â 1)
supercell was employed, with a k-point sampling of 2 Â 4 Â 1. A
slab consisting of five layers, interleaved by 15 Å of vacuum, was
used. The top two layers of the slab were optimised, whilst the
remaining bottom three were fixed. Two distinct surface environments were identified for both Ru and O (Fig. S1). Oxygen atoms
are present as rows of 2-coordinate ‘‘bridging” oxygen atoms (Ob)
on the surface, as well as 3-coordinate atoms lying within the
plane of the surface. For Ru, 6-coordinate centres are present
located beneath the Ob atoms, whilst 5-coordinate unsatured sites
(Rucus) provide the main adsorption sites for catalytic activity.
Owing to their coordinatively unsaturated nature, the Ob and Rucus
sites play the most important roles in catalysis. Forces were converged to within 0.015 eV ÅÀ1. In order to eliminate the spurious
electrostatic interactions associated with asymmetric relaxation
of the slab, a dipole correction was applied to the vacuum. The
Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional
was employed throughout [57]. Inner electrons were replaced with
projector augmented-wave (PAW) [58], and the valance states
were expanded in plane-waves with a cut-off energy of 450 eV.
The chosen functionals provide a reasonable reproduction of
experimental results [48,59]. Transition states were identified
using the climbing image nudged elastic band (CI-NEB) method
[60], and vibrational analyses performed numerically with a step
of 0.015 Å. Structures are available through the ioChem-BD repository [61,62].
To analyse the complete reaction network, a simple microkinetic (MK) model was set up based on our previous models for
the Deacon reaction [49]. Rate coefficients of the elementary steps
were determined using the thermodynamic and kinetic parameters
obtained from the DFT calculations. The gas-phase entropies were
taken from the NIST database [63], the attempt frequency for the
rate calculations was set to 1013 for simplicity, and the Knudsen
equation was taken for adsorption. Then MK simulations were performed using a batch-model reactor; further details can be found
in Ref. [49]. In the MK simulations, the initial relative pressures
and temperatures correspond to those employed experimentally,
and the initial conditions are described in the text. The reaction
order of HCl at the experimental 623 K conditions was derived
from the MK models.
2.2. Experimental details
Commercial RuO2 (Sigma-Aldrich, 99.9%) was calcined at 723 K
in static air using a heating rate of 5 K minÀ1 and an isothermal
step of 5 h prior to its catalytic evaluation. The gas-phase HCl oxidation and ethylene oxychlorination were investigated at ambient
pressure in a continuous-flow fixed-bed reactor. The set-up consists of (i) mass flow controllers to feed C2H4 (PanGas, 20.15% in
He), HCl (Air Liquide, purity 2.8, anhydrous), O2 (Messer, 10.06%
in He), He (PanGas, purity 5.0) as a carrier gas, and Ar (PanGas, purity 5.0) as an internal standard; (ii) an electrically heated oven
hosting a quartz micro-reactor equipped with a K-type thermocouple whose tip reaches the centre of the catalyst bed; (iii) downstream heat tracing to avoid any condensation of the reactants
and products; and (iv) a gas chromatograph coupled to a mass
spectrometer (GC–MS) for on-line analysis. The effluent stream
was neutralised by passing it through an impinging bottle containing an aqueous NaOH solution (1 M). The catalyst (Wcat = 0.5 g,

173

M.D. Higham et al. / Journal of Catalysis 353 (2017) 171–180

3. Results
As shown in Table 1, the reactions above differ in the stoichiometry for O2 and HCl. Among them, the most exothermic one is the
irreversible oxidation of ethylene, thus this process represents a
thermodynamic sink, whilst the formation of EDC, VCM, and 1,2DCE are exothermic by a similar amount, much higher than HCl
oxidation. In the following, we will analyse the reaction schemes
that drive the selectivity of the reaction network, first by describing the experimental data for RuO2, making comparison with other
well-known Deacon catalysts, and then by the detailed evaluation
of the reaction profile through first-principles simulations.
In order to link HCl oxidation and ethylene oxychlorination
behaviour, both processes were compared in a continuous-flow
fixed-bed reactor at a temperature T20, where the catalyst reaches
20% (±1%) conversion in the respective reactions (Fig. 1a). The T20
of ethylene oxychlorination over RuO2 is 623 K, substantially
higher than for the Deacon process (462 K). A stability test confirmed the catalyst stability of the discussed ethylene oxychlorination conditions at 623 K, exhibiting constant conversion and
selectivities during 50 h (Fig. S2a). In addition, previous work
showed that bulk RuO2 continues to perform well as a HCl oxidation catalyst up to 673 K [56]. Comparatively, RuO2 is more active
than IrO2 or CeO2 for both processes, especially for HCl oxidation

a)

b)
Conversion, Selectivity / %

T20 / K

673

573

473

100
80
60
40
20

X(C2H4)
S(EDC)
S(VCM)
S(CO)
S(CO2)

0

RuO2

473

573

673

Temperature / K

c)
Conversion, Selectivity / %

particle size, dp = 0.4–0.6 mm), diluted with quartz particles
(Wquartz = 2.5 g, dp = 0.2–0.3 mm, calcined at 1173 K) was loaded
in the micro-reactor (10 mm inner diameter) and pretreated in
He at 473 K for 30 min. Thereafter, a total flow (FT) of 100 cm3
STP minÀ1 containing 3 vol% C2H4, 1–9 vol% HCl, 0–6 vol% O2, and
3 vol% Ar, balanced in He was fed to the reactor at a bed temperature (T) of 460–773 K and pressure (P) of 1 bar. Note that relatively
low feed concentrations were selected to prevent corrosion, enable
safe handling, and minimise the formation of hot spots in the catalyst bed due to the high reaction exothermicity. The standard conditions of 3 vol% C2H4, 3 vol% HCl, and 1.5 vol% O2 were chosen to
reflect the reaction stoichiometry to form VCM (see Eq. (5), Table 1).
Prior to the analysis of the reaction mixtures, the catalysts were
equilibrated for at least 1 h under each condition. The gas composition at the reactor outlet containing reactants (C2H4, O2, HCl) and
products (EDC, VCM, CO, CO2, ethyl chloride (EC), 1,2dichloroethene (1,2-DCE)) was analysed online using a gas chromatograph equipped with a GS-CarbonPLOT column coupled to a
mass spectrometer (Agilent GC 7890B, Agilent MSD 5977A) with
a triple-axis detector and an electron multiplier. In the oxidation
tests, HCl (3 vol%) and O2 were fed in a 1:2 volumetric ratio. The
Cl2 production was quantified by offline iodometric titration (using
a Mettler Toledo G20 Compact Titrator) of triiodide, formed by
purging the Cl2 containing reactor outlet through an aqueous KI
(Sigma-Aldrich, 99.5%) solution (0.1 M), with 0.01 M sodium thiosulfate solution (Sigma-Aldrich, 99.99%). Further details on the calculation of conversions and selectivities are given in the
Supplementary Material.

100
80
60
40
20
0
1

3

5

7

9

1

3

5

7

HCl feed content / vol.% O2 feed content / vol.%
Fig. 1. (a) T20 of HCl oxidation (white bar) and ethylene oxychlorination (grey bar)
over RuO2. Conditions: Wcat = 0.5 g; FT = 100 cm3 STP minÀ1 containing 3 vol% HCl
and 6 vol% O2 for HCl oxidation, 3 vol% C2H4, 3 vol% HCl, and 1.5 vol% O2 for
oxychlorination at P = 1 bar, each point was stabilised for 2 h on stream. Conversion
and selectivity versus (b) temperature with 3 vol% C2H4, 3 vol% HCl, and 1.5 vol% O2,
and (c) volumetric HCl and O2 feed contents at T = 623 K with 3 vol% C2H4, 1–9 vol%
HCl, 1.5 vol% O2 and 3 vol% C2H4, 3 vol% HCl, 0–6 vol% O2. Other conditions as in (a).

[44,45,64]. In terms of selectivity, RuO2 exhibits mostly combustion to CO2, but also shows some a limited degree of chlorinated
hydrocarbon formation (Fig. 1b). By comparison, IrO2, the least
active, exhibits purely combustion of C2H4 to CO, and CO2 (the
major product). CeO2, like RuO2, shows formation of both combustion products and chlorinated hydrocarbons, with selectivity
towards EDC prevailing [44]. None of the three catalysts exhibit
Cl2 formation under oxychlorination conditions, but most likely
for different reasons. On IrO2 and RuO2, consumption of O2 for
combustion is strongly favoured over HCl oxidation, whilst on

Table 1
List of reactions that take place competitively on the surface of the catalyst and their entalpies as obtained from the NIST database [63].
No.

Reaction

Equation

1
2
3
4
5
6
7

HCl oxidation
Ethylene oxidation
EC synthesis
EDC synthesis
VCM synthesis
1,2-DCE (Z/E) synthesis
EDC dehydrochlorination

4HCl + O2 ? 2Cl2 + 2H2O
C2H4 + 3O2 ? 2CO2 + 2H2O
C2H4 + HCl ? C2H5Cl
C2H4 + 1/2O2 + 2 HCl ? C2H4Cl2 + H2O
C2H4 + 1/2O2 + HCl ? C2H3Cl + H2O
C2H4 + O2 + 2 HCl ? C2H2Cl2 + 2H2O
C2H4Cl2 ? C2H3Cl + HCl

DH/kJ molÀ1
À114
À1323
À69
À242
À180
À353
+62

174

M.D. Higham et al. / Journal of Catalysis 353 (2017) 171–180

a first order behaviour as well, although with a lower slope with
respect to EDC. The formation of CO is not observed. In HCl oxidation the conversion was found to exhibit a 0.2 order on HCl [49].
In order to rationalise the experimentally observed temperature
and input feed composition dependence of the catalyst selectivity,
DFT calculations were performed simulating the surface under low
Cl coverage conditions, probing the energetics of the relevant elementary processes. A large reaction network was studied, considering processes involving both chlorination and oxidation of
C2H4, separately and in conjunction. Deacon-type processes take
place under the reaction conditions, with HCl and O2 dissociative
adsorption being prerequisites for further reaction with C2H4, as
well as HCl being oxidised by O2 to Cl2 and H2O as in the standard
Deacon process. A summary of the complete reaction network is
presented in Fig. 2.
Firstly, the processes that are common to both the Deacon and
oxychlorination reactions are summarised in Table 2 and denoted
by Roman numerals. Adsorption of O2 is exothermic by À1.20 eV
(process I, Table 2), with subsequent dissociation to 2O⁄⁄ also being
exothermic by an extra À0.74 eV with an activation barrier of 0.28
eV (process II, Table 2). Dissociative adsorption of HCl is exothermic by À1.79 eV (process III, Table 2), with H being abstracted
by bridging O atoms (Ob) present on the catalyst surface. In the
Deacon process, Cl recombination yields the product Cl2 which
subsequently desorbs from the surface (process IV, Table 2). Surface H arising from HCl dissociative adsorption leaves the surface
in the form of water, with H transferring from bridging O sites

CeO2 most of the Cl is consumed in the formation of EDC and VCM
under these conditions (HCl is substoichiometric with respect to
EDC). This result, in addition to previous works [19,44,45,64], suggests that contrary to expectations, a good Deacon catalyst may not
necessarily be a good oxychlorination catalyst, but the reasons
behind this were not investigated. Thus, we picked the best Deacon
catalyst, RuO2, which exhibits only moderate oxychlorination performance, for further investigation in order to identify the
limitations.
A variation of temperature from 693 K to 473 K (Fig. 1b)
revealed that the conversion stays constant at about 20% until
623 K before it significantly drops to 7% and less. At high temperatures, combustion products dominate, with CO only being formed
at the highest investigated temperature, whilst at low temperature
until 523 K only EDC is observed. Even though VCM formation is
favoured at higher temperatures in ethylene oxychlorination
(Fig. S2), it is formed only in the temperature window of 523–
653 K, exhibiting a maximum at 573 K. Considering the temperature with still significant VCM formation and already about 20%
conversion (623 K), the HCl feed content was varied from 1 to
9 vol%, in order to achieve a similar coverage history of the sample
from low to high chlorine coverage as in the temperature variation.
Across the range of feed HCl concentrations, the conversion
stays approximately constant, exhibiting a small minimum at
3 vol% HCl (Fig. 1c). With rising HCl content, the selectivity to CO2
is decreased linearly (order À1 with respect to HCl) and the selectivity to EDC shows a first order dependence to HCl. VCM exhibits

H H

(5)

H

Ru

C2H4

Ru

H

H

H
Cl

(7)

(8)

Ru

O

Ru

(6)

(2)
+H

H

Ru

H

H

O

H

Cl

Ru

(9)

H H

H
H

O

Ru

O

Ru

(1)

H
H

O

H

Ru

O
Ru

(10)

O
Ru

H

O

O

Ru

Ru

EC
(4)

(3)

H

(11)

H

VCM
H

Cl
H
H

H
Cl

H H

(13)

Cl

Cl

+H

Ru

H

(12)

H

H

H

H

Ru

(16)

(19)

Cl

H H
Cl

O
Ru

Ru

O C O

H

Ru

EDC
(14)
H
Cl

H (15)
Cl

H

H

Cl

(Z)-1,2-DCE

H
Cl

H Cl

H
H

(21)

Cl
Ru

Ru

CO2

(20)
H Cl

H
H

H

O
Ru

Ru

Ru

O

H

H
Cl

(E)-1,2-DCE
(17) +H (18)
Chlorination
Oxidation
+/– H

H

H

H

Cl
Cl

H

1,1-EDC

(22)

H
H

H

Cl

Cl

H

O

H
H

Cl

1,1-DCE

Fig. 2. Reaction network for the processes investigated. Deacon-type processes, which are common to all of the reaction pathways, are omitted for clarity. Adsorption and
desorption processes are omitted for similar reasons. Numbers refer to processes in Table 2. Possible end products are color-coded. The starting point (C2H* ) and the observed
4
products are also indicated with bold boxes.

175

M.D. Higham et al. / Journal of Catalysis 353 (2017) 171–180
Table 2
Reaction and activation energies (DE and Ea, in eV) for various elementary steps and imaginary frequencies (cmÀ1) of the corresponding transition states.
No.

DE

Ea

ti

I
II
III
IV
V
VI
VII

HCl oxidation
O2 + 2* ? O**
2
O** ? 2O*
2
HCl + Ob + * ? Cl* + ObH
Cl* + Cl* ? Cl2 + 2*
O* + ObH ? OH* + Ob
OH* + ObH ? H2O* + Ob
H2O* ? H2O + *

À1.20
À0.74
À1.79
+2.23
À0.02
À0.11
+1.12

–
0.28
–
–
0.35
0.18
–

–
538
–
–
1073
845
–

A1
1
2
A2
3
A3
4a
4by
A4

C2H4 chlorination
C2H4 + * ? C2H*
4
C2H* + Cl* ? C2H4Cl* + *
4
*
C2H4Cl + ObH ? C2H5Cl* + Ob
C2H5Cl* ? C2H5Cl + *
C2H4Cl* + Ob ? VCM* + ObH
VCM* ? VCM + *
C2H4Cl* + Cl* ? EDC* + *
C2H4Cl* + Cl* ? EDC* + *
EDC* ? EDC + *

À0.66
+0.14
+0.23
+0.61
À0.21
+0.88
+0.17
À0.87
+0.71

–
0.26
0.96
–
1.07
–
1.85
0.00
–

–
201
1345
–
1555
–
193
–
–

5
6
A5
7
8
9
10
11a
11by
A6

C2H4 oxidation
C2H* + O* ? OMME**
4
OMME** ? CH3CHO* + *
CH3CHO* ? CH3CHO + *
OMME** ? CH* + H2CO*
2
CH* + O* ? H2CO* + *
2
*
*
H2CO + O ? OCH2O**
OCH2O** + Ob ? OCHO* + ObH + *
OCHO* + Ob ? CO* + ObH
2
OCHO* + Ob ? CO* + ObH
2
CO* ? CO2 + *
2

À0.55
À0.88
+0.90
+0.96
À1.30
À0.89
À2.30
À0.04
À2.09
0.20

0.35
1.22
–
1.06
0.02
0.00
0.54
1.70
0.00
–

261
990
–
71
71
–
1358
1169
–
–

12
13
14
A7
15
A8
16
17
A9
18
A10

VCM chlorination
VCM* + Cl* ? ClCHCH2Cl* + *
ClCHCH2Cl* + ObH ? EDC* + Ob
ClCHCH2Cl* + Ob ? Z-C2H2Cl* + ObH
2
Z-C2H2Cl* ? Z-C2H2Cl2 + *
2
*
ClCHCH2Cl + Ob ? E-C2H2Cl* + ObH
2
E-C2H4Cl* ? E-C2H2Cl2 + *
2
VCM* + Cl* ? CH2CHCl* + *
2
CH2CHCl* + ObH ? CH3CHCl* + Ob
2
2
*
CH3CHCl2 ? CH3CHCl2 + *
*
*
CH2CHCl2 + Ob ? CH2CCl2 + ObH
CH2CCl* ? CH2CCl2 + *
2

+0.56
+0.47
À0.16
+0.32
À0.15
+0.53
+0.72
+0.25
+0.53
À0.25
+0.54

1.09
1.18
1.37
–
0.97
–
0.99
0.99
–
1.39
–

248
1307
892
–
1429
–
207
1314
–
1407
–

19
20
A11
21
22
A12
y

Reaction

VCM oxidation
VCM* + O* ? OCH2CHCl**
OCH2CHCl** ? CH2ClCHO* + *
CH2ClCHO* ? CH2ClCHO + *
VCM*+O* ? CH2CHClO**
CH2CHClO** ? CH3COCl* + *
CH3COCl* ? CH3COCl + *

À0.04
À1.05
+1.13
À0.27
À1.23
+0.72

0.98
1.10
–
0.73
1.19
–

323
946
–
272
1118
–

High coverage. Structures are available through the ioChem-BD repository [62].

(Ob) to coadsorbed O (Ocus) to generate OcusH (process V, Table 2),
followed by a second H transfer to yield H2O (process VI, Table 2)
which then desorbs from the surface (process VII, Table 2). Both H
transfer processes have mildly exothermic reaction energies
(À0.02 eV and À0.11 eV for processes V and VI, respectively) and
comparable low activation energies (0.35 eV and 0.18 eV for processes V and VI, respectively). Desorption of H2O is endothermic
by 1.12 eV (process VII, Table 2). Our results are thus in line with
previous calculations [2,56,65–68]. A recent theoretical work identifies that water evolution is a key process in HCl oxidation, acting
as an energetic bottleneck for the overall reaction [47]. However,
with regards to ethylene oxychlorination, eventual evolution of
water is a common process to all of the possible reactions, and
hence will not have any appreciable impact on selectivity of the
chlorinated products.
The addition of ethylene further expands the list of competitive
reactions. Physisorption of C2H4 on the RuO2(1 1 0) slab was
determined to be exothermic by À0.66 eV (process A1, Table 2).

Subsequently, addition across the C@C bond can occur with either
coadsorbed Cl or O occupying adjacent adsorption sites to C2H4,
yielding intermediates bound to the surface. The chlorination route
(process 1, Table 2) is endothermic whilst the oxidation route
(process 5, Table 2) is exothermic (DECl = 0.14 eV, compared to
DEO = À0.55 eV). On the other hand, chlorination has a slightly
lower activation barrier than oxidation (Ea,Cl = 0.26 eV compared
to Ea,O = 0.35 eV). Hence, competition exists between the two
processes, and the product distribution will vary depending on
whether the system is under kinetic or thermodynamic control.
The chlorination pathway begins with the formation of a ClCH2CH⁄ intermediate (process 1, Table 2). From here, a number of pro2
cesses can take place. Firstly, direct protonation of the ClCH2CH⁄
2
intermediate from ObH can occur (process 2, Table 2), yielding
ethyl chloride (C2H5Cl); this process was found to possess a feasible activation barrier (Ea = 0.96 eV) but was also calculated to be
slightly endothermic (DE = 0.26 eV). Desorption of C2H5Cl is
endothermic by 0.61 eV (process A2, Table 2). A second alternative

176

M.D. Higham et al. / Journal of Catalysis 353 (2017) 171–180

is VCM generation by H abstraction from the ClCH2CH⁄ intermedi2
ate (process 3, Table 2); this process is both exothermic and kinetically feasible (DE = À0.21 eV, Ea = 1.07 eV). Desorption of VCM
from the surface is endothermic by 0.88 eV (process A3, Table 2).
Finally, direct chlorination of the ClCH2CH2⁄ intermediate yields
the product EDC; the process was found to be slightly endothermic
(DE = 0.17 eV) but with a high activation barrier (Ea = 1.85 eV)
(process 4a, Table 2).
Coverage effects are crucial in HCl oxidation [48] and thus are
likely to be important also under oxychlorination conditions.
Under high Cl coverage conditions, EDC formation from direct
chlorination of ClCH2CH⁄ (process 4by, Table 2) yielded results
2
more consistent with experimental observation; no activation barrier was observed for this process (thus it would be controlled by
the Cl diffusion on the surface), and an exothermic reaction energy
was determined (DE = À0.87 eV). Under these conditions the
ClCH2CH⁄ intermediate is destabilised by 1.42 eV, which is offset
2
by the dissociative adsorption of HCl (À1.79 eV). In other words,
the high coverage responsible for the induced destabilisation is a
direct consequence of a high degree of HCl dissociative adsorption.
Hence, HCl dissociation and the induced distortion are directly
coupled with a net exothermic energy balance. Desorption of
EDC is endothermic by 0.71 eV (process 15, Table 2).
The competing oxidation route, the oxametallacycle intermediate (OMME) formed by the addition of coadsorbed O to C2H4 (process 5, Table 2), can undergo a variety of processes. Formation of
acetaldehyde (by 1,2-H shift, process 6, Table 2) is a possibility,
as discussed in previous works [69]. Acetaldehyde formation is
exothermic by À0.88 eV and has an activation barrier of Ea = 1.22
eV; desorption of acetaldehyde is endothermic by 0.90 eV (process
A5, Table 2). Under thermodynamic control, it is likely that this
species will ultimately undergo combustion to CO2 and H2O. It is
found that cleavage of the CAC bond in OMME, resulting in fragmentation, has a lower activation barrier of 1.06 eV, making this
the more feasible route (process 7, Table 2). The fragmentation
yields formaldehyde and a CH⁄ species bound to the surface. The
2
CH⁄ fragment can then react with coadsorbed O to generate a sec2
ond equivalent of formaldehyde (process 8, Table 2); this step is
highly exothermic (DE = À1.30 eV) and has a negligible activation
barrier (Ea = 0.02 eV) owing to the instability of the CH⁄ fragment.
2
Subsequently, formaldehyde can re-adsorb to the surface, forming
an intermediate bridging the coordinatively unsaturated Ru site
(Rucus) and another Ocus, adding across the C@O carbonyl bond in
the process (process 9, Table 2). The formation of the bridging
intermediate is exothermic (DE = À0.89 eV) and was determined
to have a negligible activation barrier. CO2 is then generated by
two consecutive H abstractions from the intermediate. The first
H abstraction (process 10, Table 2) is highly exothermic
(DE = À2.30 eV) and has a low activation barrier (Ea = 0.54 eV).
The second H abstraction (process 11a, Table 2), on the other hand,
was determined to have a high activation barrier of 1.70 eV and
was found to be thermoneutral (DE = À0.04 eV) under the low coverage model applied. Further calculations using a high Cl coverage
model (process 11 by, Table 2) yielded a much more exothermic
reaction energy (DE = À2.09 eV), and a negligible activation barrier. The intermediate was calculated to have been destabilised
by 1.66 eV, again this energy can be compensated for by the
exothermicity of HCl dissociative adsorption. Desorption of CO2 is
endothermic by 0.20 eV (process A6, Table 2). Finally, the H
abstracted from the intermediates by vicinal Ob will ultimately
be transferred to coadsorbed Ocus, thus evolving water and preserving the surface Ob atoms, as in the case of the Deacon reaction
(processes V, VI and VII, Table 2).
Since VCM, like C2H4, possesses a C@C bond, the same processes
(both chlorination and oxidation) that apply to C2H4 may also
apply to VCM, at least in theory, resulting potentially in products

containing both O and Cl. Several pathways were investigated.
Addition of either Cl or O across the C@C results in two possible
intermediates, since there are two approach trajectories (as
opposed to only one for C2H4, owing to the fact that in the latter
molecule the two C atoms are equivalent). Considering first VCM
chlorination, it is noted that there are two possibilities arising from
Cl addition to VCM; either a vicinal dichloride (ClCH2ClCH⁄) or a
geminal dichloride (Cl2CHCH⁄ ). Formation of the geminal dichlo2
ride (process 16, Table 2) is endothermic by 0.72 eV and has an
activation barrier of 0.99 eV. The Cl2CHCH⁄ intermediate can then
2
either gain H from ObH, originating from HCl dissociative adsorption, yielding CH3CHCl2 (DE = 0.25 eV, Ea = 0.99 eV, process 17,
Table 2), or alternatively Ob can abstract H from the intermediate
to give an unsaturated product, CH2CCl2 (DE = À0.25 eV, Ea = 1.39
eV, process 18, Table 2). Both CH3CHCl2 and CH2CCl2 have similar
adsorption energies (À0.53 eV and À0.54 eV, processes A9 and
A10 respectively, Table 2). Formation of the vicinal dichloride is
endothermic by 0.56 eV and has an activation barrier of 1.09 eV
(process 12, Table 2). As with the geminal dichloride intermediate,
potential products can be evolved by either gaining H from Ob, or
by H abstraction by Ob. Addition of H leads to an alternative route
to obtain EDC (DE = 0.47 eV, Ea = 1.18 eV, process 13, Table 2). H
abstraction can yield two unsaturated products, Z-C2H2Cl2 and EC2H2Cl2, which form a pair of geometrical isomers depending on
the conformation of the intermediate. The reaction energies for
the two processes are very similar (DE = À0.16 eV, DE = À0.15
eV, for Z-C2H2Cl2 and E-C2H2Cl2 respectively, processes 14 and
15, Table 2). However the activation barrier for the formation of
Z-C2H2Cl2 is significantly higher than that of E-C2H2Cl2 (Ea = 1.37
eV, c.f. Ea=0.97 eV), likely due to steric repulsion between Cl on
the same side of the molecule in the case of Z-C2H2Cl2. Additionally, the E-C2H2Cl2 desorption energy is more endothermic at
0.53 eV (process A8, Table 2) compared to 0.32 eV for Z-C2H2Cl2
(process A7, Table 2). Indeed, it is possible that the unsaturated
products from VCM chlorination could in turn undergo further
reaction in an analogous manner. However, since all the reaction
energies associated with the formation of VCM chlorination products are endothermic, and all of these processes have high activation barriers, further reactions beyond this point have not been
examined.
Turning to VCM oxidation, similarly two distinct intermediates
can be formed, CH2CHClO⁄⁄ (DE = À0.27 eV, Ea = 0.73 eV, process
21, Table 2), and OCH2CHCl⁄⁄ (DE = À0.04 eV, Ea=0.98 eV, process
19, Table 2), depending on the side of the VCM molecule from
which O approaches. As with OMME, H transfer can yield an aldehyde, CH2ClCHO, from OCH2CHCl⁄⁄ (DE = À1.05 eV, Ea = 1.10 eV,
process 20, Table 2), and in the case of CH2CHCl⁄⁄ an acid chloride,
CH3COCl, is formed (DE = À1.23 eV, Ea = 1.19 eV, process 22,
Table 2). Desorption of the aldehyde is considerably more
endothermic at 1.13 eV (process A11, Table 2), compared to 0.72
eV for the acid chloride (process A12, Table 2).

4. Discussion
4.1. Origin of selectivity
Selectivity between oxidation and chlorination of C2H4 is governed by the interplay between thermodynamic and kinetic factors. Thermodynamic selectivity can originate from differing
adsorption energies between species. Products with highly
exothermic adsorption energies are more likely to undergo further
reaction than to be desorbed as an observed product, thus dictating
selectivity [70].
Fig. 3 depicts the adsorption energies for a variety of chlorination products as well as C2H4 as a function of surface coverage

M.D. Higham et al. / Journal of Catalysis 353 (2017) 171–180

Fig. 3. Adsorption energies of ethylene, EC, EDC, VCM, and Z-DCE for a variety of
surface compositions. (a) Adsorption energy as a function of Cl coverage on the
RuO2 surface where half of the bridging oxygen atoms were replaced with Cl. (b)
Adsorption energy as a function of Cl coverage. (c) Adsorption energy as a function
of O coverage.

(by O or Cl). In general, higher coverages lead to destabilised
adsorbed species. It is notable that VCM is more strongly adsorbed
than other species at all but the highest of coverages, owing to its
compact size and its ability to interact with Ru active sites through
either Cl or the C@C bond, or indeed both (occupying two active
sites) at lower coverages. Therefore, if formed, VCM is likely to stay
on the surface, as it is the most stable product at low coverages. It
is also important to note that the C2H4 adsorption energy remains
relatively unchanged at high O coverage, compared to the adsorption energy for a clean surface. Since C2H4 is the precursor for combustion products, the stronger C2H4 adsorption further supports
oxidation to CO2 and H2O. However, at high coverages, VCM shows
a lower adsorption energy than C2H4, thus favouring the thermodynamic selectivity. This phenomena has been reported for the
gold-catalysed selective hydrogenation of CAC triple bonds in

177

alkynes, to the exclusion of hydrogenation of C@C bonds in the
resulting alkenes, since competitive adsorption between the reactants and products ensure that the products do not undergo further
reaction, being displaced in favour of the adsorption of the reactant
[69]. The lower adsorption energy for VCM compared to C2H4
means that under high coverage conditions, the two species compete for the remaining adsorption sites, with C2H4 being preferentially adsorbed. Consequently, VCM desorbs from the catalyst
surface once it is generated in favour of adsorption of C2H4, rather
than undergoing further reaction and thus leading to an enhanced
selectivity.
As for the kinetic aspects of the reaction, the first step in the
reaction is crucial, i.e. whether chlorination (process 1) or oxidation (process 5) of C2H4 takes place. The activation barrier for addition of Cl to C2H4 is lower than for the addition of O, however the
chlorination process is endothermic compared to the exothermic
oxidation route. Whilst it is very likely that the initial and final
states for Cl addition to C2H4 exist in equilibrium, the higher activation barrier for oxidation cannot be breached. Thus, at lower
temperatures, the system is more prone to kinetic control and only
chlorinated products are observed, as the intermediate C2H4Cl⁄ is
consumed to form chlorinated products over time. On the other
hand, at higher temperatures, thermodynamic control prevails,
and since the oxidation process is much more exothermic, it is
more likely that the formation of OMME is irreversible, and hence
at higher temperatures more combustion products are observed to
the exclusion of chlorination products.
Fig. 4 depicts the reaction profile for the oxidation and chlorination of C2H4, with chlorination progressing ultimately to EDC formation, at high and low Cl coverage. At low coverage, the
activation barrier towards EDC is prohibitively high, whereas at
high Cl coverage (process 4b, Table 2), the same process yielded
results more consistent with experimental observation. It was calculated that the ClCH2CH⁄ intermediate is destabilised by 1.42 eV,
2
which can be easily offset by the dissociative adsorption of HCl
(À1.79 eV). At high Cl coverage, however, adsorbed species are
somewhat destabilised by the coadsorbed chlorine atoms, thus
diminishing the activation barrier of the processes as the energetic
gap is closed between the initial and transition states of each

Fig. 4. (a) Reaction profiles for oxidation and chlorination of C2H4 at low (LC) and high (HC) chlorine coverage. The chlorination route depicts the eventual formation of EDC.
The white background region indicates where the oxidation and chlorination processes compete at the specific coverage, whilst the green background region indicates that
additional HCl has been dissociatively adsorbed to allow the reaction to continue to the formation of EDC, at high and low Cl coverages. (b) Diagrams depicting the considered
configurations, where green circles indicate chlorine atoms, green/white circles indicate either a chlorine or no atom, and green/red circles indicate either chlorine or oxygen
atoms.

178

M.D. Higham et al. / Journal of Catalysis 353 (2017) 171–180

process. This is reflected in differences in the atomic distances at
high Cl coverage compared to low Cl coverage. At low Cl coverage,
the distance between a Cl atom and the C atom of a coadsorbed C2H5Cl⁄ intermediate (Table 2, process 4a Fig. 5b) is 3.53 Å. However,
at high coverage (Table 2, process 4by), the corresponding distance
is only 3.08 Å, reflecting the crowding on the surface responsible
for the destabilisation of the C2H5Cl⁄ intermediate. It is also notable
that whilst the activation barrier for C2H4 chlorination is lower
than that of oxidation under low coverage conditions, at high coverages the pattern is reversed and the activation barriers are comparable. This implies that at high coverage, the competition
between oxidation and chlorination of C2H4 is enhanced, leading
to reduced selectivity.
Under kinetic control, it is important to consider why EDC, and
not C2H5Cl or VCM, is produced. The key to this again is the surface
coverage, which can confine and destabilise intermediates. For EDC
generation under high Cl coverage, the additional coadsorbed Cl
destabilises the C2H4Cl⁄ intermediate, with the destabilisation
being compensated for by the exothermicity of HCl dissociative
adsorption. Hence, the energetic barrier between the intermediate
and the transition state is eliminated by the destabilisation of the
intermediate. Formation of EDC takes place via the intermediate
reacting with one of the adjacent coadsorbed Cl, thus relieving
the species of its confinement, and as such, the process is highly
exothermic. On the other hand, gaining (losing) H from (to) ObH
(Ob) to generate C2H5Cl (VCM) does not involve the coadsorbed
Cl responsible for the destabilisation and hence, the confinement
is not relieved, therefore rendering it very likely that the thermodynamics and kinetics of these processes at high coverage would
be largely unchanged from their low coverage values. As such,
the experimental observation that the main chlorination product
is EDC, rather than VCM, is justified theoretically.
Previous studies [44] suggest that EDC dehydrochlorination,
generating VCM and HCl, requires sufficiently acidic sites on the
catalyst surface, which RuO2(1 1 0) lacks, hence in situ dehydrochlorination does not occur and VCM is not a major product.
Furthermore, it was suggested that at high partial HCl pressures,
a significant proportion of Ob atoms are replaced by Cl, thus limiting the number of Ob sites able to abstract hydrogen from intermediates, or indeed provide hydrogen to intermediates having
deprotonated HCl during dissociative adsorption [37]. This would
also hinder any processes that require such sites, thus further
favouring the generation of EDC which does not depend on H
transfers to or from Ob. Under thermodynamic control, it is evident
that oxidation dominates, and it is clear why combustion to CO2 is
preferred over intramolecular formation of CH3CHO. The activation
barrier for fragmentation of OMME to CH⁄ and H2CO is consider2

ably lower than that of the alternative H transfer processes. Additionally, all of the subsequent steps under low coverage (except for
the final H abstraction) are strongly exothermic and possess either
small or negligible activation barriers.
4.2. Requirements for a good VCM catalyst
From the description above, we have identified three main
requirements needed for a successful ethylene-to-VCM catalyst.
Firstly, we consider the thermodynamic term. The adsorption
energy for ethylene must fall in the same range as at least one of
the other reactants (fulfilled by O2 in this case). Moreover, it has
to be more strongly bound than the desired product. For RuO2, this
is only achieved for high coverages where the bonding energy of
VCM is lower than that of ethylene. Hence, competition for active
sites on the catalyst surface between adsorbates means that VCM
is displaced in preference of C2H4, limiting the extent to which
VCM undergoes subsequent reactions. These persistent active oxygen atoms are a severe threat to the production of VCM. The CAO
bond is stronger than the CACl bond (3.71 vs. 3.39 eV). This is an
indication for the calculated effect that chlorination is reversible
on the surface, whilst CAO formation is exothermic and thus irreversible. Energetically, this restriction implies that oxygen cannot
diffuse across Rucus positions close to adsorbed ethylene. Secondly,
the geometry of the (1 1 0) surface poses also the additional problem of dimensional confinement, shared by the other lowest
energy surfaces: (1 0 1) and (0 0 1) [71]. Transport of the active
Cl and O atoms takes place only along the [0 0 1] Rucus direction.
This implies that once ethylene is adsorbed, the first atom in the
neighbouring sites is the first one to be incorporated on the surface. As chlorination is reversible, the presence of a single O atom
in the neighbouring site precipitates the production of CO2. Therefore, confined transport along a single direction is detrimental to
the formation of VCM, with the constrained diffusion of ethylene
across the surface limiting the number of coadsorbed species it
can react with, making it more difficult for the kinetically preferred
chlorination to occur.
The final point concerns the effective abstraction of the H atoms
from the chlorinated fragments or even EDC. Basic sites are needed
for stripping H atoms from the intermediates leading to VCM. As
the presence of Ocus limits selectivity by favouring combustion,
the basicity of the geometrically reachable Ob needs to be strong
enough to allow effective H abstraction. On the RuO2 surface,
HCl-O2 mixtures produce surfaces where Ob are partially replaced
by Cl atoms ($50% replacement) [56]. Whilst this limit is crucial to
ensure the stability of the catalyst under Deacon conditions, it
impedes an effective dehydrogenation of the intermediates, most

Fig. 5. (a) Side view of the optimised geometry of C2H4Cl* at high coverage. (b) C2H4Cl* under low Cl coverage. (c) CHO* under high Cl coverage. (d) CHO* under low Cl
2
2
coverage. Colour legend: blue (Ru), pale red (surface O), bright red (adsorbed O), green (Cl), dark grey (C), and light grey (H).

M.D. Higham et al. / Journal of Catalysis 353 (2017) 171–180

likely leading to enhanced EDC production. Therefore, a successful
VCM catalyst needs to maintain effective O centres able to strip
hydrogen from the intermediates.
Additionally, in order to further validate our approach, we set
up a microkinetic model. The model is built on our previous results
on Deacon reaction [49] and thus is only briefly commented on
here. Notice that since the surface presents strong anisotropy and
monodimensional character [59], mean-field approaches can be
biased and thus they can only be taken as qualitative. Still, the
reaction order for HCl in the Deacon is found to be 0.3, in reasonable agreement with the experimental values. In contrast, the reaction orders of HCl for EDC and VCM production over RuO2 are 0.89
and À0.14, respectively. Thus, ethylene oxychlorination to EDC
shows a remarkable difference compared to the Deacon reaction,
in terms of the dependence on HCl. The kinetic similarities
between reaction orders close to zero for Cl2 and VCM can also represent a challenge in improving VCM production on RuO2. Finally,
it is noted that the production of Cl2 was negligible under the conditions at which EDC was observed, in excellent agreement with
experiments.
5. Conclusions
Under the conditions required for catalytic oxychlorination of
ethylene over RuO2(1 1 0), a variety of chemical reactions are possible, with competing oxidation and chlorination processes. Whilst
oxidation, and ultimately combustion to CO2 and H2O, is favoured
thermodynamically, conversion to the desired oxychlorination
products is possible at lower temperatures through kinetic control.
The coverage and composition of the surface is crucial to the behaviour of the system. Under high Cl coverages, competitive adsorption between C2H4 and VCM ensures that VCM is displaced by
C2H4, owing to its less exothermic adsorption energy, preventing
the excessive formation of products derived from VCM. The effect
of dimensional confinement of C2H4 means that coadsorbed species react on a ‘‘first-come, first-served” basis, thus resulting in a
potentially considerable degree of combustion where coadsorbed
O is involved. The effect of substitution of bridging O sites by Cl,
is key to understanding the prevalence of EDC over VCM under
kinetic control, since the Ob sites required to abstract H from intermediates are essential to VCM formation. Therefore, the surface
composition of the catalyst has a major impact on activity and
selectivity and is an important consideration in the development
of new catalysts. It is intended that this study will inform future
work in developing highly selective, robust and active catalysts
for ethylene oxychlorination, defining the basic requirements for
a good catalyst by offering mechanistic insights at the molecular
level.
Acknowledgments
Financial support from the Swiss National Science Foundation
(Project No. 200021-156107), the ICIQ Foundation, and the Spanish
MINECO (CTQ2015-68770-R) is acknowledged. M.C.-C. also
acknowledges MINECO for a ‘‘Juan de la CiervaÀFormación” fellowship (FJCI-2014-20568).
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2017.07.013.
References
[1] J. Wang, C.Y. Fan, K. Jacobi, G. Ertl, The kinetics of CO oxidation on RuO2(1 1 0):
bridging the pressure gap, J. Phys. Chem. B 106 (2002) 3422–3427.

179

[2] H. Over, Y.D. Kim, A.P. Seitsonen, S. Wendt, E. Lundgren, M. Schmid, P. Varga, A.
Morgante, G. Ertl, Atomic-scale structure and catalytic reactivity of the
RuO2(1 1 0) surface, Science 287 (2000) 1474–1476.
[3] H. Over, A.P. Seitsonen, E. Lundgren, M. Schmid, P. Varga, Direct imaging of
catalytically important processes in the oxidation of CO over RuO2(1 0 0), J. Am.
Chem. Soc. 123 (2001) 11807–11808.
[4] Y. Wang, K. Jacobi, W.-D. Schöne, G. Ertl, Catalytic oxidation of ammonia on
RuO2(1 1 0) surfaces: mechanism and selectivity, J. Phys. Chem. B 109 (2005)
7883–7893.
[5] B.-Z. Zhan, M.A. White, T.-K. Sham, J.A. Pincock, R.J. Doucet, K.V.R. Rao, K.N.
Robertson, T.S. Cameron, Zeolite-confined nano-RuO2: a green, selective, and
efficient catalyst for aerobic alcohol oxidation, J. Am. Chem. Soc. 125 (2003)
2195–2199.
[6] Z. Liu, V. Sriram, J.-Y. Lee, Heterogeneous oxidation of elemental mercury vapor
over RuO2/rutile TiO2 catalyst for mercury emissions control, Appl. Catal. B
Environ. 207 (2017) 143–152.
[7] A.S. Fajardo, H.F. Seca, R.C. Martins, V.N. Corceiro, I.F. Freitas, M.E. QuintaFerreira, R.M. Quinta-Ferreira, Electrochemical oxidation of phenolic
wastewaters using a batch-stirred reactor with NaCl electrolyte and Ti/RuO2
anodes, J. Electroanal. Chem. 785 (2017) 180–189.
[8] Y.-H. Fang, Z.-P. Liu, Mechanism and Tafel lines of electro-oxidation of water to
oxygen on RuO2(1 1 0), J. Am. Chem. Soc. 132 (2010) 18214–18222.
[9] H. Over, Surface chemistry of ruthenium dioxide in heterogeneous catalysis
and electrocatalysis: from fundamental to applied research, Chem. Rev. 112
(2012) 3356–3426.
[10] K.S. Exner, J. Anton, T. Jacob, H. Over, Controlling selectivity in the chlorine
evolution reaction over RuO2-based catalysts, Angew. Chem. Int. Ed. 53 (2014)
11032–11035.
[11] K. Iwanaga, K. Seki, T. Hibi, K. Issoh, T. Suzuta, M. Nakada, Y. Mori, T. Abe, The
development of improved hydrogen chloride oxidation process, Sumitomo
Kagaku. (2004) 1–11.
[12] K. Seki, Development of RuO2/Rutile-TiO2 catalyst for industrial HCl oxidation
process, Catal. Surv. Asia. 14 (2010) 168–175.
[13] J. Pérez-Ramírez, C. Mondelli, T. Schmidt, O.F.-K. Schlüter, A. Wolf, L. Mleczko,
T. Dreier, Sustainable chlorine recycling via catalysed HCl oxidation: from
fundamentals to implementation, Energy Environ. Sci. 4 (2011) 4786–4799.
[14] A. Wolf, L. Mleczko, O.F.-K. Schlüter, S. Schubert, Method for producing
chlorine by gas phase oxidation, EP2026905 A1, 2007.
[15] H.A. Hansen, I.C. Man, F. Studt, F. Abild-Pedersen, T. Bligaard, J. Rossmeisl,
Electrochemical chlorine evolution at rutile oxide (1 1 0) surfaces, Phys. Chem.
Chem. Phys. 12 (2010) 283–290.
[16] F. Studt, F. Abild-Pedersen, H.A. Hansen, I.C. Man, J. Rossmeisl, T. Bligaard,
Volcano relation for the Deacon process over transition-metal oxides,
ChemCatChem. 2 (2010) 98–102.
[17] S. Trasatti, Physical electrochemistry of ceramic oxides, Electrochim. Acta 36
(1991) 225–241.
[18] S. Trasatti, Electrocatalysis: understanding the success of DSAÒ, Electrochim.
Acta. 45 (2000) 2377–2385.
[19] N. López, G. Novell-Leruth, Rules for selectivity in oxidation processes on
RuO2(1 1 0), Phys. Chem. Chem. Phys. 12 (2010) 12217–12222.
[20] Businesswire, Global polyvinyl chloride (PVC) market worth USD 72.330
billion by 2021 – Analysis, Technologies & Forecasts Report 2016–2021 Vendors: INOVYN, Axiall Corp, Westlake Chemical Corp - Research and
Markets, 2016.
[21] G.J. Hutchings, Vapor phase hydrochlorination of acetylene: correlation of
catalytic activity of supported metal chloride catalysts, J. Catal. 96 (1985) 292–
295.
[22] P. Johnston, N. Carthey, G.J. Hutchings, Discovery, development, and
commercialization of gold catalysts for acetylene hydrochlorination, J. Am.
Chem. Soc. 137 (2015) 14548–14557.
[23] K. Weissermel, H.-J. Arpe, Vinyl-halogen and Vinyl-oxygen Compounds, in:
Ind. Org. Chem, Wiley-VCH, Weinheim, 2008, pp. 217–238.
[24] J.S. Naworski, E.S. Velez, Oxychlorination of ethylene, in: B. Leach (Ed.), Appl.
Ind. Catal., Academic Press, 1983, pp. 239–273.
[25] M. Malentacchi, C. Rubini, Oxychlorierungsatalysatorzusammensetzung zur
Regelung von exothermischen Reaktionen in einem Festbett, EP1020222 B1,
2004.
[26] N.B. Muddada, T. Fuglerud, C. Lamberti, U. Olsbye, Tuning the activity and
selectivity of CuCl2/c-Al2O3 ethene oxychlorination catalyst by selective
promotion, Top. Catal. 57 (2014) 741–756.
[27] N.B. Muddada, U. Olsbye, L. Caccialupi, F. Cavani, G. Leofanti, D. Gianolio, S.
Bordiga, C. Lamberti, Influence of additives in defining the active phase of the
ethylene oxychlorination catalyst, Phys. Chem. Chem. Phys. 12 (2010) 5605–
5618.
[28] J.P. Henley, M.E. Jones, D.A. Hickman, K.A. Marshall, D.J. Reed, W.D. Clarke, M.
M. Olken, L.E. Walko, Process for vinyl chloride manufacture from ethane and
ethylene with partial HCl recovery from reactor effluent, US 6933417 B1, 2005.
[29] F.E. Van Rooijen, J. Johan, A. De Bruijn, Catalytic oxychlorination, US 0054708
A1, 2009.
[30] M. Loos, J. Goulon, M. Bertucci, R. Bachelard, Structural studies using X.A.S. of
the active species catalysing the oxychlorination of ethylene – II, Phys. B
Condens. Matter. 158 (1989) 188–190.
[31] C. Lamberti, C. Prestipino, F. Bonino, L. Capello, S. Bordiga, G. Spoto, A.
Zecchina, S. Diaz Moreno, B. Cremaschi, M. Garilli, A. Marsella, D. Carmello, S.
Vidotto, G. Leofanti, The chemistry of the oxychlorination catalyst: an In Situ,
time-resolved XANES study, Angew. Chem. Int. Ed. 41 (2002) 2341–2344.

180

M.D. Higham et al. / Journal of Catalysis 353 (2017) 171–180

[32] D. Carmello, E. Finocchio, A. Marsella, B. Cremaschi, G. Leofanti, M. Padovan, G.
Busca, An FT-IR and reactor study of the dehydrochlorination activity of CuCl2/
c-Al2O3-based oxychlorination catalysts, J. Catal. 191 (2000) 354–363.
[33] G. Leofanti, A. Marsella, B. Cremaschi, M. Garilli, A. Zecchina, G. Spoto, S.
Bordiga, P. Fisicaro, G. Berlier, C. Prestipino, G. Casali, C. Lamberti, Aluminasupported copper chloride 3. Effect of exposure to ethylene, J. Catal. 202 (2001)
279–295.
[34] G. Leofanti, A. Marsella, B. Cremaschi, M. Garilli, A. Zecchina, G. Spoto, S.
Bordiga, P. Fisicaro, C. Prestipino, F. Villain, C. Lamberti, Alumina-supported
copper chloride: 4. Effect of exposure to O2 and HCl, J. Catal. 205 (2002) 375–
381.
[35] N.B. Muddada, U. Olsbye, T. Fuglerud, S. Vidotto, A. Marsella, S. Bordiga, D.
Gianolio, G. Leofanti, C. Lamberti, The role of chlorine and additives on the
density and strength of Lewis and Brønsted acidic sites of c-Al2O3 support
used in oxychlorination catalysis: a FTIR study, J. Catal. 284 (2011) 236–246.
[36] R.V. Carrubba, J.L. Spencer, Kinetics of the oxychlorination of ethylene, Ind.
Eng. Chem. Proc. Des. Dev. 9 (1970) 414–419.
[37] E.I. Gelperin, Y.M. Bakshi, A.G. Zyskin, Y.S. Snagovsky, A.K. Avetisov, Kinetics
and mechanism of ethylene oxychlorination, Russ. Chem. Ind. 28 (1996) 356–
363.
[38] D. Gianolio, N.B. Muddada, U. Olsbye, C. Lamberti, Doped-CuCl2/Al2O3 catalysts
for ethylene oxychlorination: influence of additives on the nature of active
phase and reducibility, Nucl. Instr. Meth. Phys. Res. B. 284 (2012) 53–57.
[39] N.B. Muddada, U. Olsbye, G. Leofanti, D. Gianolio, F. Bonino, S. Bordiga, T.
Fuglerud, S. Vidotto, A. Marsella, C. Lamberti, Quantification of copper phases,
their reducibility and dispersion in doped-CuCl2/Al2O3 catalysts for ethylene
oxychlorination, Dalt. Trans. 39 (2010) 8437–8449.
[40] A.J. Magistro, J.A. Cowfer, Oxychlorination of ethylene, J. Chem. Educ. 63 (1986)
1056–1058.
[41] D. Shi, R. Hu, Q. Zhou, C. Li, Effect of Cr-doping on CuCl2-KCl-CeO2/c-Al2O3
catalysts for ethane oxychlorination, Appl. Catal. A Gen. 506 (2015) 91–99.
[42] S.A. Kurta, Investigating active centers of industrial catalysts for the oxidative
chlorination of ethylene on a c-Al2O3 surface, Catal. Ind. 3 (2011) 136–143.
[43] S.A. Kurta, Catalysis of ethylene oxychlorination into 1,2-dichlorethane in the
presence of CuCl2/CuCl active centres on the surface of c-Al2O3, Chem. Chem.
Technol. 6 (2012) 1–8.
[44] M. Scharfe, P.A. Lira-Parada, V. Paunovic, M. Moser, A.P. Amrute, J. PérezRamírez, Oxychlorination-dehydrochlorination chemistry on bifunctional
ceria catalysts for intensified vinyl chloride production, Angew. Chem. Int.
Ed. 55 (2016) 3068–3072.
[45] M. Scharfe, P.A. Lira-Parada, A.P. Amrute, S. Mitchell, J. Pérez-Ramírez,
Lanthanide compounds as catalysts for the one-step synthesis of vinyl
chloride from ethylene, J. Catal. 344 (2016) 524–534.
[46] R. Lin, A.P. Amrute, J. Pérez-Ramírez, Halogen-mediated conversion of
hydrocarbons to commodities, Chem. Rev. 117 (2017) 4182–4247.
[47] F. Hess, H. Over, Rate-determining step or rate-determining configuration?
The Deacon reaction over RuO2(110) studied by DFT-based KMC simulations,
ACS Catal. 7 (2017) 128–138.
[48] D. Teschner, G. Novell-Leruth, R. Farra, A. Knop-Gericke, R. Schlögl, L.
Szentmiklósi, M. González Hevia, H. Soerijanto, R. Schomäcker, J. PérezRamírez, N. López, In situ surface coverage analysis of RuO2-catalysed HCl
oxidation reveals the entropic origin of compensation in heterogeneous
catalysis, Nat. Chem. 4 (2012) 739–745.
[49] D. Teschner, R. Farra, L. Yao, R. Schlögl, H. Soerijanto, R. Schomäcker, T.
Schmidt, L. Szentmiklósi, A.P. Amrute, C. Mondelli, J. Pérez-Ramírez, G. NovellLeruth, N. López, An integrated approach to Deacon chemistry on RuO2-based
catalysts, J. Catal. 285 (2012) 273–284.

´
[50] M. Moser, V. Paunovic, Z. Guo, L. Szentmiklósi, M.G. Hevia, M. Higham, N.
López, D. Teschner, J. Pérez-Ramírez, Interplay between surface chemistry and
performance of rutile-type catalysts for halogen production, Chem. Sci. 7
(2016) 2996–3005.
[51] T. Hibi, H. Nishida, H. Abekawa, Process for producing chlorine, US 5871707,
1999.
[52] G. Kresse, J. Furthmüller, Efficiency of ab-initio total energy calculations for
metals and semiconductors using a plane-wave basis set, Comput. Mater. Sci. 6
(1996) 15–50.
[53] G. Kresse, J. Furthmüller, Efficient iterative schemes for ab initio total-energy
calculations using a plane-wave basis set, Phys. Rev. B. 54 (1996) 11169–
11186.
[54] H.J. Monkhorst, J.D. Pack, Special points for Brillouin-zone integrations, Phys.
Rev. B. 13 (1976) 5188–5192.
[55] C.-E. Boman, Refinement of the crystal structure of ruthenium dioxide, Acta
Chem. Scand. 24 (1970) 116–122.
[56] N. López, J. Gómez-Segura, R.P. Marín, J. Pérez-Ramírez, Mechanism of HCl
oxidation (Deacon process) over RuO2, J. Catal. 255 (2008) 29–39.
[57] J.P. Perdew, K. Burke, M. Ernzerhof, Generalized gradient approximation made
simple, Phys. Rev. Lett. 77 (1996) 3865–3868.
[58] P.E. Blöchl, Projector augmented-wave method, Phys. Rev. B. 50 (1994) 17953–
17979.
[59] S. Pogodin, N. López, A more accurate kinetic Monte Carlo approach to a
monodimensional surface reaction: The interaction of oxygen with the
RuO2(1 1 0) surface, ACS Catal. 4 (2014) 2328–2332.
[60] G. Henkelman, B.P. Uberuaga, H. Jónsson, A climbing image nudged elastic
band method for finding saddle points and minimum energy paths, J. Chem.
Phys. 113 (2000) 9901–9904.
[61] M. Álvarez-Moreno, C. de Graaf, N. López, F. Maseras, J.M. Poblet, C. Bo,
Managing the computational chemistry big data problem: the ioChem-BD
platform, J. Chem. Inf. Model. 55 (2015) 95–103.
[62] https://dx.doi.org/10.19061/iochem-bd-1-38.
[63] P.J. Linstrom, W.G. Mallard (Eds.), NIST Chemistry WebBook, NIST Standard
Reference Database Number 69, National Institute of Standards and
Technology, Gaithersburg, MD.
[64] M. Moser, C. Mondelli, A.P. Amrute, A. Tazawa, D. Teschner, M.E. Schuster, A.
Klein-Hoffman, N. López, T. Schmidt, J. Pérez-Ramírez, HCl oxidation on IrO2based catalysts: from fundamentals to scale-up, ACS Catal. 3 (2013) 2813–
2822.
[65] H. Over, Ruthenium dioxide, a fascinating material for atomic scale surface
chemistry, Appl. Phys. A Mater. Sci. Process. 75 (2002) 37–44.
[66] S. Zweidinger, D. Crihan, M. Knapp, J.P. Hofmann, A.P. Seitsonen, C.J.
Weststrate, E. Lundgren, J.N. Andersen, H. Over, Reaction mechanism of the
oxidation of HCl over RuO2(1 1 0), J. Phys. Chem. C 112 (2008) 9966–9969.
[67] A.P. Seitsonen, H. Over, Oxidation of HCl over TiO2-supported RuO2: a density
functional theory study, J. Phys. Chem. C 114 (2010) 22624–22629.
[68] H. Over, Atomic-Scale Understanding of the HCl oxidation over RuO2, a novel
Deacon process, J. Phys. Chem. C 116 (2012) 6779–6792.
[69] Y. Segura, N. López, J. Pérez-Ramírez, Origin of the superior hydrogenation
selectivity of gold nanoparticles in alkyne + alkene mixtures: Triple- versus
double-bond activation, J. Catal. 247 (2007) 383–386.
[70] D. Crihan, M. Knapp, S. Zweidinger, E. Lundgren, C.J. Weststrate, J.N. Andersen,
A.P. Seitsonen, H. Over, Stable Deacon process for HCl oxidation over RuO2,
Angew. Chem. Int. Ed. 47 (2008) 2131–2134.
[71] G. Novell-Leruth, G. Carchini, N. López, On the properties of binary rutile MO2
compounds, M = Ir, Ru, Sn, and Ti: a DFT study, J. Chem. Phys. 138 (2013)
194706.