Journal of the American Chemical Society
ARTICLE
the “indirect route”. It is suggested that the process proceeds via
halohydrin route where, e.g., 2-chloroethanol is formed by
oxidation of ethene with unspecified electrochemically formed
chloro-species near the electrode surface.8 2-Chloroethanol
subsequently undergoes homogeneous base-promoted dehydro-
genhalogenation to form the oxirane.17 As a result, the indirect
epoxidation in alkaline electrolytes gives high yields of epoxides
and the electro-oxidation in acidic solutions leads to accumula-
tion of halohydrins.17ꢁ19 The evidence to support the proposed
reaction mechanisms presented in the latter papers, however, is
rather inconclusive. In particular, the role of the halide ions in the
electrochemical oxidation of alkenes remains unclear.
The present effort explores the possibility of electrocatalytic
oxidation of ethene to produce oxirane in acid media, most
compatible with fuel cell technology.20 The role of Clꢁ in the
reaction process is evaluated. It is suggested here that Clꢁ may
display a dual function as catalytic epoxidation promoter and
reactant to form Cl2 and 2-chloroetanol. The aim is to investigate
to what extent the two functions of Clꢁ can be decoupled by
choice of catalyst, thus isolating a promoting effect of Clꢁ. RuO2
has been long known as an active and stable electrode material for
water and Clꢁ oxidation in acid media.21 In addition, these
electrocatalytic properties of the Ru based oxides can be tuned by
controlling the particle size22 or by doping with, e.g., Ni23 or
Co.24 In the present paper, the efficiency of an analogous
electrocatalysts, i.e., nanocrystalline RuO2 and Co doped
RuO2, toward ethene oxidation in acid media is investigated by
online mass spectrometry. To clarify the role of chloride ions in
the reaction mechanisms, DFT calculations are employed. The
comparison between results of theory and experiment is
discussed.
The particle shape analysis showed predominance of 101 (∼55%) and
110 (∼35%) oriented planes for both materials.22 The Co surface
concentration in Ru0.8Co0.2O2 determined by XPS was higher compared
to bulk and was around ∼30%.24 The analysis of particle shape and size
distribution of Ru0.8Co0.2O2 showed similar results as RuO2; the mean
diameter was 56 ( 14 nm and (101) was also the predominant surface
orientation followed by (110).24
Electrodes. The electrodes for electrochemical experiments were
prepared from RuO2-400, RuO2-900, and Ru0.8Co0.2O2 materials by
sedimentation of nanocrystalline powder from a water-based suspension
on Ti mesh. The duration of the deposition was adjusted to obtain the
surface coverage of about 1ꢁ2 mg/cm2 of active oxide layer. The
deposited layers were stabilized by annealing the electrodes for 20 min at
400 °C in air. The initial oxide suspensions were prepared in an
ultrasound bath and contained approximately 5 g/L of ruthenium-based
oxide in water.
Elecrochemical Setup. The electrochemical behavior of the
prepared materials was studied by cyclic voltammetry combined with
differential electrochemical mass spectrometry (DEMS). All experi-
ments were performed in a homemade Kel-F single compartment
cell27 in a three electrode arrangement controlled by a PAR 263A
potentiostat. Pt and saturated calomel electrode (SCE) were used as an
auxiliary and a reference electrode, respectively. For clarity, measured
potentials were recalculated and all reported data in the paper are
referenced versus the normal hydrogen electrode (NHE; þ0.244 V vs
SCE). The measurements were carried out in 0.1 M HClO4 aqueous
solution, both ethene-free and saturated with ethene. The effect of
chloride ions was studied in 0.1 M HClO4 containing 0.001, 0.01, and
0.3 M NaCl, respectively. The potential sweep rate was 15 mV sꢁ1 in all
experiments, which provided optimal compromise between obtaining
near steady state behavior and reasonable measurement times with
respect to the lifetime of the mass detector. The cell was sealed against
the ambient atmosphere, and saturation of the solution with ethene
(5.2 mM) was maintained by ethene circulation loop at the ambient
pressure and temperature (25 °C). The electrochemical compartment
was separated from the mass spectrometer vacuum inlet by PTFE
membrane (Gore, 100 nm pore size). The investigated electrode was
pressed against the membrane allowing immediate detection of any
volatile products formed at the electrode.
DEMS. The DEMS apparatus consisted of Prisma TM QMS200
quadrupole mass spectrometer (Balzers) connected to TSU071E tur-
bomolecular drag pumping station (Balzers). The mass spectrometer
was set up to monitor all mass to charge ratio (m/z) channels that could
be expected from fragmentation of likely products of ethene oxidation.
Ethylene glycol, acetic acid, acetaldehyde, oxirane, and 2-chloroethanol
were considered. Fragmentation patterns for all monitored species were
obtained from online NIST spectral library28 and adjusted for the
employed spectrometer. The calibration of the mass detector was carried
out by measuring its response after injection of acetaldehyde and ethene
to the electrochemical cell attached to the mass spectrometer. In the
same way injected ethylene glycol did not penetrate the PTFE mem-
brane and cannot be detected by the employed equipment. The
fragmentation patterns for the rest of considered species were corrected
according to the calibration assuming the same trend of the detector
sensitivity increase toward fragments with lower mass as for ethene and
acetaldehyde. The calibrated fragmentation patterns of all monitored
species are summarized in Table 1.
’ METHODS
Chemicals. Cobalt(II) nitrate (ACS grade), 30% H2O2 solution
(semiconductor grade), acetaldehyde (ACS grade), 70% redistlilled
HClO4 (99.999%), and NaCl (p.a.) were purchased from Sigma-
Aldrich. Ruthenium(III) nitrosylnitrate and 25% tetramethylammo-
nium hydroxide (electronic grade) were supplied by Alpha Aesar.
Titanium mesh (20% open area) was from Goodfellow (U.K.). Ethene
(98%) gas was obtained from Linde Gas (Czech Republic). Milli-Q
(Millipore Inc.) water was used in all experiments.
24,26
Materials Preparation. The RuO222,25 and the Ru0.8Co0.2O2
samples were prepared using a solꢁgel approach described previously.
Starting solutions of either ruthenium(III) nitrosylnitrate or of
ruthenium(III) nitrosylnitrate with cobalt(II) nitrate in a mixture of
ethanol and propane-2-ol (1:1) were precipitated with aqueous solution
of tetramethylammonium hydroxide. In the case of starting solution
containing both Ru and Co the actual Ru/Co ratio was 4:1. Both
precipitation procedures led to a formation of amorphous precursors
that were filtered and annealed. The RuO2 precursor was treated with
H2O2 prior to the annealing. The precursor containing both Ru and Co
was aged in a PTFE lined stainless steel autoclave at 100 °C for 40 h to
facilitate the filtration process. The RuO2 precursor was annealed at
400 °C (RuO2-400) and 900 °C (RuO2-900) and the Ru0.8Co0.2O2
precursor at 800 °C in air for 4 h.
Materials Characterization. The crystallinity and phase purity of
the prepared samples were checked using Bruker D8 Advance powder
X-ray diffractometer with Vantec-1 detector and CuKR radiation source.
The X-ray diffraction analysis confirmed that all materials were of a
single-phase character with rutile type structure (P42/mnm).22,26 Both
RuO2-400 and RuO2-900 materials showed good particle size mono-
dispersity with mean diameters of 15 ( 9 and 35 ( 9 nm, respectively.22
Computational Details. The initial crystal structure was of rutile
type (P42/mnm) with lattice parameters a = b = 4.559 Å and c = 3.166 Å
as resulted from geometry optimization (DMOL329). Oxygen termi-
nated two-layer as well as three-layer slabs with (101) crystallographic
orientation were obtained by cleavage of the initial structure and were of
the size of 2 ꢂ 2 unit cell, with respect to the surface plane (see Figure S1
in Supporting Information). In both cases, the bottom layers were frozen
5883
dx.doi.org/10.1021/ja109955w |J. Am. Chem. Soc. 2011, 133, 5882–5892