Electrochemical Reduction of Protic Supercritical CO2 on Copper Electrodes

Article abstract of DOI:10.1002/cssc.201701205

The electrochemical reduction of carbon dioxide is usually studied in aqueous solutions under ambient conditions. However, the main disadvantages of this method are high hydrogen evolution and low faradaic efficiencies of carbon-based products. Supercritical CO₂ (scCO₂) can be used as a solvent itself to suppresses hydrogen evolution and tune the carbon-based product yield; however, it has received little attention for this purpose. Therefore, the focus of this study was on the electrochemical reduction of scCO₂. The conductivity of scCO₂ was increased through the addition of supporting electrolyte and a cosolvent (acetonitrile). Furthermore, the addition of protic solutions of different pH to scCO₂ was investigated. 1 m H₂SO₄, trifluoroethanol, H₂O, KOH, and CsHCO₃ solutions were used to determine the effect on current density, faradaic efficiency, and selectivity of the scCO₂ reduction. The reduction of scCO₂ to methanol and ethanol are reported for the first time. However, methane and ethylene were not observed. Additionally, corrosion of the Cu electrode was noticed.

Full text of DOI:10.1002/cssc.201701205

Accepted Article
Title: Electrochemical Reduction of Protic Supercritical CO2 on Copper
Electrodes
Authors: Olga Melchaeva, Patrick Voyame, Victor Costa Bassetto,
Michael Prokein, Manfred Renner, Eckhard Weidner, Marcus
Petermann, and Alberto Battistel
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: ChemSusChem 10.1002/cssc.201701205
Link to VoR: http://dx.doi.org/10.1002/cssc.201701205
A Journal of
www.chemsuschem.org

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
Electrochemical Reduction of Protic Supercritical CO2 on Copper
Electrodes
Olga Melchaeva,*[a,b,c] Patrick Voyame,[a] Victor Costa Bassetto,[a] Michael Prokein,[c] Manfred Renner,[c]
Eckhard Weidner,[b,c] Marcus Petermann,[b] and Alberto Battistel*[a]
Abstract: The electrochemical reduction of carbon dioxide is usually
studied in aqueous solutions under ambient conditions. However,
the main disadvantages of this method are high hydrogen evolution
and low faradaic efficiencies of carbon based products. Supercritical
CO2 (scCO2) can be used as a solvent itself to suppresses hydrogen
evolution and tune carbon based product yield, however, it received
low attention. Therefore, the focus of this study was on the
electrochemical reduction of supercritical CO2 (at 40 ˚C and 80 bar).
The conductivity of scCO2 was increased through addition of
supporting electrolyte and co-solvent (acetonitrile). Besides, the
addition of protic solutions with different pH to supercritical CO2 was
investigated. 1 M H2SO4, trifluoroethanol, H2O, KOH, and CsHCO3
solutions were used to determine the effect on current density,
faradaic efficiency, and selectivity of scCO2 reduction. Reduction of
supercritical CO2 to methanol and ethanol were reported for the first
time. However, methane and ethylene were not observed.
Additionally, corrosion of Cu was noticed.
the studies in this area are done by purging gaseous CO2 in
aqueous solutions under ambient conditions. Because of the
CO2 stability high overpotentials are required to reduce it. Thus,
in aqueous solutions hydrogen evolution is predominant and
competitive to CO2 reduction.[1,2] Additionally, the low solubility of
CO2 in aqueous solutions limits the rate of the reduction
reaction.[3] Thus, the process is usually associated with low
faradaic efficiency (FE) and low selectivity towards carbon-
based products.
Effect of High Pressure
However, this problem can be tackled by increasing the CO2
pressure.[2] Larger concentrations of CO2 at the electrode result
in higher reduction currents and yields. High-pressure CO2
(below 69 bar) suppresses hydrogen evolution and increases
selectivity towards formic acid[4], methanol[5], ethanol[6],
ethylene[7] and other hydrocarbons[6] even in aqueous solutions.
Besides, the decrease of hydrogen evolution and CO2 reduction
was proportional to increase of pressure.
At pressures and temperatures above 73.8 bar and 31 °C, CO2
reaches its supercritical conditions. A supercritical phase is a
fluid which takes up all the volume of its container, like a gas,
and like a gas has low viscosity, but it also conserves the high
density of a liquid. These properties make the supercritical CO2
(scCO2) an attractive solvent. Furthermore, any other gas is
readily solubilized in a supercritical phase.[3] On the opposite,
liquids have a maximum solubility and biphasic system can
form.[3]
In supercritical CO2, only little amounts of protic solutions (e.g.
water) are needed to facilitate the reaction to hydrocarbons,
organic acids, or alcohols. This can prevent significant hydrogen
evolution and reduce energy losses. Besides, scCO2 provides
higher mass transport of dissolved ions and molecules than in
liquids, ensuring its prompt delivery to the electrodes.[8,9] The
differences between conventional electrochemical reduction of
gaseous CO2 in aqueous solution at ambient conditions and in
supercritical CO2 at elevated pressure are provided in Figure 1.
In the left panel, CO2 is reduced in an aqueous phase, where at
Introduction
The depletion of fuel stocks and adverse impact on the
environment indicate the urgent need of alternative sources of
clean and renewable energy. Conversion of the waste materials
such as carbon dioxide (CO2) into valuable products as
chemicals and fuels has potential economic and environmental
benefits. Electrochemical CO2 reduction is considered as one of
the most attractive methods, due to the relative simplicity of the
equipment, low operating temperatures, and possibility to couple
with renewable energy sources.[1]
Carbon dioxide can be electrochemically converted into various
products as hydrocarbons, alcohols, and organic acids. Most of
[a]
Olga Melchaeva, Dr. Patrick Voyame, Victor Costa Bassetto, Dr.
Alberto Battistel
Laboratory of Physical and Analytical Electrochemistry (LEPA)
EPFL Valais Wallis
Rue de l'Industrie 17, 1950 Sion, Switzerland
E-mail: alberto.battistel@epfl.ch
ambient condition, it has
a
solubility of 0.059 mol.%
[b]
[c]
Olga Melchaeva, Eckhard Weidner, Marcus Petermann
Institute for Particle Technology
Ruhr University Bochum
Universitätsstraße 150, 44801 Bochum
Olga Melchaeva, Michael Prokein, Manfred Renner, Eckhard
Weidner
(0.033 mol L-1).[10] The reduction in supercritical phase is
depicted in the central panel, where water solubility is between
0.34 and 0.61 mol.% according to the conditions (50 °C at 80.8
and 141.1 bar[11]). In the right panel, the supercritical phase is
supersaturated with water, which forms an aerosol and
condensate on the electrode. The CO2 reaction in this case can
proceed in both phases. According to the nature and solubility of
the products, these can partition between the two phases.
Fraunhofer UMSICHT
Osterfelder Str. 3, 46047 Oberhausen
olga.melchaeva@umsicht.fraunhofer.de
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.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
Moreover, the water aerosol increases the total conductivity of
the electrolyte (see supporting information).
However, there is still relatively little literature on
electrochemistry in supercritical CO2. Up to now most of the
works have concentrated on the improvement of electrical
conductivity of scCO2[15,21,24], possibility of using scCO2 as a
medium for electrochemical synthesis on the example of
reduction and oxidation of ferrocene and DMFc[12,19–21,25], and on
[26]
photoreduction of scCO2. Only few works on electrochemical
reduction of supercritical CO2 have been reported.[3,23,27]
Effect of pH and Electrolyte
Effect of different protic solutions on supercritical CO2 reduction
was not yet investigated. However, there are number of
publications addressing aspects of pH and types of dissolved
ions in aqueous solutions. Protons are needed for CO2 reduction
to hydrocarbons, alcohols, and organic acids (equation 1-4). In
aprotic mediums hydrogen containing products can be neglected,
as formation of only CO and oxalate anions (COO–)2 is possible
(equation 5, 6).[3] Different protic solutions can provide a system
with a different amount of protons and increase current density,
efficiency, and product distribution.
Figure 1. Comparison of CO2 reduction in different conditions. Left: reduction
in aqueous solution, where CO2 has low solubility. Center: reduction in
supercritical CO2 where water has low solubility. Right: reduction in
supercritical CO2 supersaturated with water. In this case supercritical and
liquid phase coexist. Water condenses on the electrode and forms an aerosol.
CO2 + H+ + 2 e–  HCOO–
(1)
(2)
(3)
(4)
(5)
(6)
CO2 + 2 H+ + 2 e–  CO + H2O
CO2 + 6 H+ + 6 e–  CH3OH + H2O
CO2 + 8 H+ + 8 e–  CH4 + 2 H2O
The first studies of electrochemistry in supercritical CO2 were
done by the group of Wightman in the late 1980s.[12–14] They
could observe well-defined redox wave of ferrocene
(Fc/Fc+).[12,13] Bartlett and Branch et al. provided extensive work
on conductivity of scCO2[15]. Low dielectric constant of CO2 (ε <
1.8) requires the application of bulky hydrophobic electrolytes.
Electrolytes based upon tetrabutylammonium cations and
weakly coordinated anions were considered as the most
2 CO2 + 2 e–  (COO–)2
–
2 CO2 + e–  CO + CO3
In conventional electrochemistry in aqueous solutions, a vast
amount of water is available for the reaction, which is also often
associated with predominant hydrogen evolution. Multiple
studies have shown an effect of the cation size of the supporting
electrolyte and pH change on the efficiency and product
distribution in aqueous electrolytes.[28–30] However, no study was
made on the effect of protic solutions for scCO2 reduction.
Therefore, for better understanding of the process, a comparison
to conventional electrochemistry in aqueous solutions is
provided.
promising. Use of tetrabutylammonium cations [NBun ]+
4
associated with [BF4]–, [PF6]– anions have been reported in
several publications and can give reasonable conductivity[15,16,17]
.
Increase of anion size and degree of fluorination also have an
effect on increase of solubility and conductivity.[15]
As an alternative method, polar co-solvents, as methanol[15,18]
,
It was generally noticed that strong basic solutions increase
selectivity towards CO and C2O4 2–; slightly basic solutions to
hydrocarbons (CH4, C2H4) and alcohols (CH3OH, C2H5OH); and
acidic and neutral towards HCOOH and methanol.[1,29,31,32]
According to Hori et al. CO2 reduction is associated with the OH–
accumulation on the surface of the electrode which leads to the
increases of the local pH at the electrode in comparison to the
bulk solution.[33] High pH is related to low amounts of molecular
CO2 and corresponding rise of HCO3– and CO32– at the electrode.
The decrease in CO2 concentration near the electrode results in
a reduction of faradaic efficiency for complex hydrocarbon and
alcohols and in increased CH4 and H2 production.[30] However,
the application of a low pH is also associated with challenges as
higher hydrogen evolution reaction is prevalent.[34] Therefore,
neutral or slightly acidic medias are generally used for the CO2
reduction.[31,34]
dimethylformamide[19], acetone, or acetonitrile[15,20] can be used
to increase electrical conductivity of scCO2. These co-solvents
play the role of intermediate polarity mixtures, increasing
dielectric constant, solubility, and dissolution of electrolyte salts
and water in CO2. According to Bartlett et al., an addition of 10 to
18 mol.% of CH3CN is crucial to achieve a dielectric constant of
7 and therefore a higher current density.[17] Besides, solubility of
supporting electrolytes was noticed to be 5 times higher in
acetonitrile rather than in methanol.[21]
The idea of electrochemical scCO2 reduction was introduced by
Bandi et al., where its application was foreseen in their
patent.[8,9] Abbott and Harper were the first to conduct
investigations on the electrochemical reduction of supercritical
CO2 in 1998.[22] No hydrogen production was detected; however
the overall faradaic efficiency also did not exceed 28.8 % for
HCOOH on Pt and 15.6 % for H2C2O4 on Pb.[22] In the
consecutive study of Abbott et al., in a system with no added
proton donor the only found products were CO and (COOH)2.[3]
Zhao et al. used CO2/H2O system in combination with ionic
liquids as an electrolyte. Products as H2, CO, and CH2O2 were
reported with 50.6, 38.1 and 1.5 % FE, respectively.[23]
Singh et al. reported the possibility of offsetting the effects of
electrolyte polarization occurring near the cathode surface. They
have found a dependence of local pH change on the cation size
of aqueous electrolyte.[28] Hydrated cations showed an ability to
promote proton transfer and soften the local pH change at the
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
cathode surface, acting as a buffer. It was noted that the pKa
Copper was used as one of the most attractive catalysts in CO2
value decreases with increasing size of cations near the cathode. conversion to hydrocarbons, such as CH4, C2H4; CO; alcohols
Thus, it can help to decrease and buffer the local pH change,
assuring higher amount of CO2 present on the cathode surface.
As a result of this study, an increase of the alkali metal cation
size from Li to Cs lead to twice higher current density and
significant shift of Cu and Ag catalysts selectivity towards
C2H5OH and C2H4. For example, faradaic efficiency of CO2
reduction on copper in 0.1 M LiHCO3 solution resulted in 78.1 %
H2, 11.1 % CH4, 6.7 % C2H4 and 0.6 % C2H5OH, while in 0.1 M
CsHCO3 37.8 % H2, 12.9 % CH4, 39.6 % C2H4 and 9.7 %
C2H5OH. Schizodimou and Kyriacou also reported an increase
in CO2 reduction rates on copper catalysts in acidic solutions
with increase in surface charge of the cations as follows: Na+ <
Mg2+ < Ca2+ < Ba2+ < Al3+ < Zr4+ < Nd3+ < La3+.[29] This effect
was even more pronounced in acidic solutions at pH > 4. In 2 M
HCl solution the main products were CH3OH, HCOOH, and
CO.[29] Studies of Kaneco et al. and Thorson et al. have also
shown the effect of hydrated alkali metal solution on Cu[35] and
Sn[36] electrodes, respectively. They have attributed the
selectivity of the process to the specific adsorption of metal
cations. It was suggested, that the small cations (Li+, Na+) might
not easily adsorb on the electrode surface due to their strong
hydration, while bigger size cations (Cs+) might have a better
adsorption and decrease the electrode hydration.[1,35,36] This
observation, however, applies best not to strongly acidic or
alkaline solutions.[28] Theoretical study of Singh et al. have also
suggested that the optimal bulk pH for CO2 reduction shall be
around 7.[37]
as methanol, ethanol and propanol; organic acids like HCOOH,
(COOH)2, as well as esters, paraffins, and olefins.[31,38] There is
a number of excellent studies covering several aspects, such as
mechanism reactions of CO2 reduction on copper[7,39] catalytic
activity and binding CO energies[40]
,
effects of surface
morphology[41], electrolyte[7], pH[7,28], pressure[7], etc. Therefore,
as the most studied catalyst in CO2 reduction, plane copper was
chosen for reference and comparison to literature at ambient
conditions.
Results and Discussion
One of the major challenges when working with scCO2 medium
is associated with its very low conductivity and poor solvation of
polar compounds. Therefore, our experiments have started with
optimization of scCO2 conductivity by varying several
parameters as concentration of supporting electrolyte, co-
solvent and water amount. When reasonable conductivity was
found, effect of various protic solutions on product distribution
has been tested.
Variation in Supporting Electrolyte Concentration
In order to achieve sufficient conductivity of CO2 various
amounts of supporting electrolyte tetrabutylammonium
hexafluorophosphate (TBAPF6) were tested. For the dissolution
of TBAPF6 salt, constant amount of acetonitrile was used. The
dependence between resistance of the solution and TBAPF6
concentration had an exponential decaying function (supporting
information, Figure S2). A significant drop of the solution
resistance (Rs) from 294 to 68 Ohm cm2 was observed for
electrolyte concentrations of 5 and 35 mM, respectively. Further
increase of TBAPF6 concentration from 35 to 80 mM resulted in
only little decrease of Rs to 48 Ohm cm2 indicating that
scCO2/acetonitrile system was close to saturation. Therefore,
35 mM concentration of the electrolyte was used in most trials,
as a balance between electrolyte costs and optimal conductivity
of the electrolyte solution and to avoid the supersaturation of the
acetonitrile which would then not dissolve in the supercritical
phase.
CO2 and H2O reduction result in the release of OH– and thus
high local pH at the surface of the electrode.[7] However,
increase of CO2 pressure leads to the carboxylation of water
–
(CO32–) and formation of bicarbonates (HCO3 ) in case of alkali
metal hydroxide solutions. Besides, use of high CO2 pressure
contributes to the buffering effect of the solution, due to the
reaction with OH– and bicarbonate formation. These leads to
lowering and buffering of local pH with pressure.[7] Thus, for
example, lowering of local pH by the CO2 pressure increase
from 1 to 9 bar in 0.5 M KHCO3 electrolyte resulted in significant
increase of ethylene selectivity.[7]
In this paper, we used supercritical CO2 as a solvent and
reactant itself for CO2 reduction on copper at 80 bar and 40 ˚C.
Our goal was to investigate the behaviour of electrochemical
reduction of scCO2. To increase the electrochemical conductivity
of CO2 we studied the effects of supporting electrolyte, co-
solvent (acetonitrile), and water. Besides, we compared the
effect of different protic solutions dissolved in scCO2 on current
density, efficiency, and selectivity of the products according to
their pH and metal cation size. Water and acidic solutions as
1 M H2SO4 and trifluoroethanol (TFE) were used as proton
donors. Basic aqueous solutions, as 1 M KOH and CsHCO3
were used due to their reported activity to buffer changes in the
local pH and increase selectivity towards ethanol and ethylene,
which is reported to be linked with the increase of the size of
alkali metal cation.[28] The effects of pH on selectivity are further
discussed.
Molar conductivity (훬푚) as a function of the salt concentration
was obtained from the Kohlrausch’s law for small concentrations
(see supporting information). Linear approximation was used for
the Debye-Hückel law for 1:1 electrolytes:
1
훬0푚
퐾
=
−
√퐶
푅푚푒푠 ∙ 퐶 푘푐푒푙푙 푘푐푒푙푙
where Rmes is the measure of the solution resistance, C the
electrolyte concentration, 훬0푚 the limiting molar conductivity, K
an empirical constant and 푘푐푒푙푙 the cell constant. The equation
allows to calculate the cell constant 푘푐푒푙푙, which is a function of
the geometrical ratio of the cell and current line distribution
between the electrodes. To the best of our knowledge, no
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
limiting molar conductivity of TBAPF6 in supercritical mixture
CO2/acetonitrile system was reported. Using the value reported
result in the “liquid-like” behavior of the solution and therefore
low mass transport.[15,17] However, for further experiments, a
concentration of 29.3 vol.% was employed to achieve higher
conductivity.
for TBAPF6 in acetonitrile at ambient conditions 훬0푚,푇퐵퐴푃퐹6
=
167 Ω–1 cm2 mol–1, a cell constant kcell equal to 1.05 cm–1 was
obtained. The value can vary as the influence of high pressure
CO2 was not taken in account. For a system with low
conductivity as scCO2, kcell values below 1 are advisable. To
increase the conductivity, a small gap between the electrodes
and large electrodes are advisable.
Effect of Co-Solvent
Low dielectric constant of scCO2 is linked with poor solubility and
dissolution of ionic species and low conductivity.[17] The dielectric
constant is influenced by interatomic and intermolecular
attractions and determines the dissolution of the electrolyte.[42]
Dielectric constant of CO2 can be increased by the use of polar
miscible co-solvents as methanol, acetone, and acetonitrile. This
would increase dissolution of electrolyte salts and solubility of
water in scCO2. Acetonitrile is commonly used as aprotic solvent
for CO2 reduction.[43,44] Solubility of CO2 in acetonitrile is 8 times
higher than in water.[45] Besides, it has a relatively high stability
(large potential window)[44] and high dissolution ability for
tetrabutylammonium electrolyte salts (5 times higher than that of
methanol).[15] Additionally, it cannot be mistaken for one of the
products of CO2 reduction and it can be employed as media for
CO2 products analysis.
Various concentrations of acetonitrile with 0.029 vol.% H2O were
added to scCO2/TBAPF6 system. The experiments were
performed in the reactor with the sapphire window for phase
change observation. Acetonitrile in the concentrations above
15.1 vol. % with 0.029 vol.% H2O and 35 mM TBAPF6 are fully
miscible and form one phase with scCO2 at 40 ˚C and 80 bar.
According to literature and as observable in Figure 2 a, CO2
reduction can be seen in the range from approximately –0.6 until
–1.7 V vs. Ag/Ag+.[1] At more negative potentials, acetonitrile
Figure 2. Cyclic voltammograms (CV) recorded on a Cu electrode in a 35 mM
TBAPF6 scCO2 solution containing varying amounts of acetonitrile and
0.029 vol.% of water at 80 bar and 40 °C at 100 mV s–1. The arrows represent
the directions of cathodic and anodic scans. (a). Electrolyte resistance as
function of acetonitrile concentration (b).
reduces
(see
supporting
information).
The
cyclic
voltammograms (CV) show an increase in current density at
higher CH3CN concentrations. The conductivity of the system
without acetonitrile is too low and very small current can be
driven (in the picoampere range). In the case with 7.6 vol.% of
acetonitrile the system showed two phases mostly because the
amount of electrolyte was too high and saturated the acetonitrile.
Therefore the amount of co-solvent and electrolyte in the
supercritical phase was very low compared with the other
experiments. From 22.7 vol.% the voltammograms show a
plateau in the current which moves toward more negative
currents increasing the concentration of acetonitrile. After
34.0 vol.% of acetonitrile, no significant enhancement of CO2
reduction current was observed. Figure 2 b shows the resistance
of the electrolyte measured by impedance spectroscopy. The
resistance halved between 22.7 and 29.3 vol.% and still
decreased another ten percent at 34.0 vol.%. This is connected
with the higher polarity of the electrolyte and therefore with the
decreasing of ion pairs and improved mobility of ionic species.
According to Bartlett et al., addition of 10–18 mol.% of
In the cyclic voltammetries, a plateau in the current can be
observed over a potential range of approximately 600 mV. Most
of the reactions pathways for CO2 reduction involve protons (see
eq. 1-4) and they are a one proton per electron reaction (eq. 2-4).
Therefore, given the low amount of water and the fact that the
pH was not too acidic a limiting diffusion current for protons can
be expected.
Effect of Water
To increase the current density and promote CO2
electroreduction towards organic products, concentration of a
protic solution as water was optimized. Pure water was added to
the scCO2/CH3CN/TBAPF6 system in different concentrations
(Figure 3). The increase of the water content to 4.88 vol.%
acetonitrile is crucial, much higher concentrations (푥퐶퐻 퐶푁 > 0.5)
3
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
showed a decrease and even elimination of the plateau visible in
Figure 2. Moreover, higher H2O concentration shows a tenfold
increment in current density, from –0.36 to –4.31 mA cm–2 at –
1.5 V vs. Ag/AgPF6 with no H2O and 2.44 vol.%, respectively
(Figure 3). The saturation of scCO2/acetonitrile system occurred
with water concentration near 2.44 vol.%, which showed no
difference in the CV curves with an increase to 4.88 vol.%. The
test in the view cell with a sapphire window showed a one phase
system with water concentration near 2.44 vol.%, further
increase of water resulted in a two-phase system where water
formed an aerosol in the supercritical phase or condensed on
the window, at the walls and at the bottom of the reactor. Water
droplets had different size (between 0.3 and 3 mm) on the
sapphire glass and on electrodes (see supporting information).
After two hours at still conditions (no electrolysis), the aerosol
settled down and accumulated at the bottom of the reactor.
However, during the electrolysis more aerosol was produced by
the electrode reaction and filled the entire reactor. Within 4.5
hours from the end of the experiment, the view cell did not
become clear. The droplets formed on the surface of the
electrode stayed still. As shown in supporting information, a
water aerosol can increase the total conductivity of the system.
Figure 4. Schematic of sequestration of polar CO2 reduction products into the
water phase which is mainly condensed at the bottom of the reactor. A part of
the products is re-oxidized at the counter electrode and a part is partitionated
between the water phase and the supercritical phase according to their
dielectric constants.
For the electrochemical experiments, a reactor with a single
compartment for both working and counter electrodes was used.
Therefore, re-oxidation of CO2 reduction products on the counter
electrode was possible. To solve this problem, application of
slightly higher amount of water can be advisable. Taking into
account the polarity of the compounds in a mixture, higher
affinity of polar compounds, as organic acids and alcohols (CO2
reduction products), to polar water rather than to non-polar CO2
can be expected. Therefore, to predict the solubility of products
in scCO2/CH3CN/H2O/TBAPF6 at high pressure, we based our
prediction on the dielectric constants of all the constituents. For
example, dielectric constant (ε) of CO2 equals 1.8, in a mixture
with 30 mol.% acetonitrile it can increase to approximately 13.[17]
The dielectric constants of water, HCOOH, and CH3OH are 80.1,
55.9, and 32.7 at ambient conditions, respectively.[46] Increased
temperatures and pressures result in a slight decrease of ε for all
the components. This would mean that polar products formed on
the WE in low to medium polarity scCO2/CH3CN/H2O/TBAPF6
mixture, would have high affinity to dissolve in the excess of the
highly polar water, which is mostly placed on the bottom of the
reactor (Figure 4). The phase study of Yoon et al. in supercritical
CO2/CH3OH/H2O system (100 bar, 40 °C), also suggests higher
solubility of methanol in the systems with water excess (two
phase system).[47] Thus, slight excess of highly polar water can
offer an “escape way” for the polar products preventing their
oxidation at the counter electrode (Figure 4). For this purpose,
4.88 vol.% of water was used. This means that only
Figure 3. Cyclic voltammograms on Cu electrode in a 35 mM TBAPF6 scCO2
and 29.3 vol. % acetonitrile solution containing varying amounts of water at 80
bar and 40 °C, scan rate: 100 mV/s.
The plateau in the current shortened with increased water
content and almost disappeared at saturation. Some
experiments were performed in a stirred solution where no
plateau was visible (see supporting information). These facts
reinforce the assumption that the plateau is connected with the
limited availability of protons in the system. In fact, assuming a
diffusion coefficient in the order of 10–4 cm2 s–1, a concentration
of 20 μM H+, and a diffusion layer of 50 μm, a limiting diffusion
current of circa 4 mA cm–2 can be obtained, which is in
agreement with what observed. Furthermore, given the natural
turbulence of supercritical phase the current at the plateau was
wavier that in the rest of the voltammogram as already observed
in literature[21,25]. Besides, it was generally difficult to achieve
reliable impedance spectroscopies in this region.
approximately
half
of
it
was
dissolved
in
the
scCO2/acetonitrile/TBAPF6 mixture and participated in the
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
reaction, another half formed the aerosol and condensed on the
wall and on the electrodes and accumulated on the bottom of
the reactor. The presence of a water phase served as purpose
of polar products sequestration.
To reach higher and faster dissolution of products a few
parameters, like stirring, temperature, and amount of added
water should be optimized. These effects, however, are not
discussed in this work.
Effect of Protic Solutions in scCO2 Medium
As mentioned, with increase of pressure carboxylation of water
occurs. Formation of the bicarbonates at the electrode surface
can have a buffering effect at the electrode surface. However,
due to the small concentration of water used (4.88 vol.%) the pH
influence on the product distribution in scCO2 system might be
not as dramatic as just in aqueous medium at ambient pressure.
Reported bulk pH value in scCO2/water mixtures is
around 3.1.[48] Monitoring of the bulk pH change under high
pressure is difficult, while the one of the local pH (< 10 nm from
the electrode surface) change is even impossible; therefore they
were not performed in this study. The effect of the protic
solutions was evaluated from the product distribution, cyclic
volltamograms and chronoamperometric diagrams.
In the current study, neutral, acidic, and basic protic solutions
were tested for scCO2 reduction. Water was selected as neutral
source. Strong and weak acidic solution (1 M H2SO4 and
trifluoroethanol) were selected as proton reach components to
increase the current density and the availability of protons. Basic
protic solutions as 1 M CsHCO3 and KOH were used to buffer
the changes in local pH and shift selectivity towards ethylene or
longer chain products.
Figure 5. Typical cyclic voltammograms recorded on Cu electrode in a 35 mM
TBAPF6 solution of scCO2, 29.3 vol.% acetonitrile and 4.88 vol.% of various 1
M proton donors before (a) and after 2 h electrolysis (b) at ¬–1.55 V vs.
Ag/AgPF6 wire. Measurements conducted at 80 bar and 40 °C and scan rate
of 100 mV/s.
All the protic solutions were added in concentration of 4.88 vol.%
to
the
scCO2/acetonitrile/TBAPF6
mixtures.
solvents
Cyclic
in
voltammograms
of different protic
Chronoamperometry (CA) was used for the product synthesis
and stability measurements of the electrodes (Figure 6). At short
times, the current decayed because of capacitive mass transport
(Cottrell behavior) effects. In comparison to convectional
electrochemistry in aqueous solutions the current profiles were
fluctuating. Periodically decreasing and rising current can be
attributed to the formation of product layer on the surface of the
electrode and its periodical removal. Layer deposition on the
electrode would, therefore, hinder its catalytic activity and
restore it back when it is suddenly removed.
Besides, there is a lot of noise on the CA voltammograms.
Similar behaviour can be observed in aqueous solutions during
bubbles (gas) formation at the electrodes, causing its partial
coverage and therefore isolation from the electrolyte
solution.[49]The activity is then restored when a bubble leaves the
surface. In case of high-pressure electrolysis, this behaviour is
not to be expected, because of the total miscibility of scCO2 with
gases, unless a liquid layer is formed on the electrode. In this
case, small bubbles can develop in the water phase.
scCO2/acetonitrile/TBAPF6 mixture before and after two hours
electrolysis (at –1.55 V vs. Ag/AgPF6) are depicted in Figure 5.
One can see that the current density was increased by
application of H2SO4, CsHCO3, and KOH in comparison with
water at the potentials between –1.4 and –2.0 V vs. Ag/AgPF6.
Remarkably, the plateau region was also shorten due to the
higher supply of protons or buffering effect at the surface of the
electrode. After two hours electrolysis, the current density was
decreased for all the samples, which typically occurs due to the
altering of the electrode surface by the formed products.
However, the activity of the added acidic and basic protic
solutions still remained higher than the one of water. Application
of TFE yielded much lower current density than of other samples
and therefore was not considered for further electrochemical
investigation.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
(originating from the degradation of the supporting electrolyte)
were found. Small amount of silver was detected. Its source was
either an impurity in the copper material or came from the quasi-
reference electrode.
To understand whether carbon originated from CO2 reduction,
the electrode was also analysed after exposure to the same
conditions of the electrolysis experiments, but without applying
any current. In this case, C was present in a slightly lower
amount, while O amount increased. N was not detected
(Supporting information for details). This indicates that the
observed carbon and nitrites can be the product of acetonitrile
decomposition intensified by applied voltage and pressure.
According to literature, although acetonitrile is
compound, its polymerization could be affected by the extremely
high plasma or chemical
pressure[50] discharge[51]
a stable
,
,
environment[52], although all the conditions mentioned were
considerably more extreme than what experienced in our
experiments. As confirmed in this study, electrochemical
reduction of acetonitrile lies beyond –1.8 V vs. Ag/AgPF6 while
the electrolysis was performed at –1.55 V vs. Ag/AgPF6 (see
supporting information). Additionally, tetrabutylammonium salts
in the example of TBABF4 can be decomposed in acetonitrile
mixture at the potentials below –2.9 and above +2.0 V vs.
Ag/Ag+.[53] In this study potential of –1.55 V vs. Ag/AgPF6 were
applied, which is not supposed to lead to the salt decomposition.
The phenomenon of possible acetonitrile or electrolyte
decomposition was not investigated further in this work.
Moreover, as reported in several publication, graphitic carbon
might be a product of CO2 reduction and a cause of deactivation
of catalytic activity.[54] According, Kas et al., graphitic carbon can
be one of the products of CO2 reduction in the pathway of CH4
formation from CHO intermediate.[7] Additionally, we also noticed
that the deposition of black layer depended on the amount of
added water and did not occur in the absence of it. Copper oxide
formed without electrolysis was slightly higher than after
electrolysis. The natural copper oxide layer formed upon
exposure to air for more than a week was almost twice as low as
Figure 6. Typical chronoamperomogramms (CA) recorded for 2 h electrolysis
at -1.55 V vs. Ag/AgPF6 in a 35 mM TBAPF6 solution of scCO2, 29.3 vol.%
acetonitrile and 4.88 vol.% of various 1 M protic solutions. The measurements
conducted at 80 bar and 40 °C on Cu electrode.
Images of the electrodes taken before and after the electrolysis
illustrate the formation of a black layer on the copper electrode
surface (Figure 7 and supporting information). Other protic
solutions than water resulted in slightly higher amount of black
layer formed. This layer was predominantly formed at the places
where water droplets adsorbed. In case of the small size Cu
electrodes (0.2 cm2), agglomerated water droplets could cover a
significant part of the electrode, while in case of the bigger
electrode size (8 cm2) (Supporting information) one could see
distinctly black spots at the water/electrode interface. It is
noteworthy that the current distribution was more uneven on the
small electrode than on the larger one, which faced frontally a
flat counter electrode.
the
one
occurred
on
copper
exposed
to
scCO2/acetonitrile/H2O/TBAPF6 mixture, which proves that
copper oxidation is a result of chemical reaction in supercritical
fluid mixture. Furthermore pitting corrosion appeared on the
electrodes as visible from Figure 8 d after prolonged electrolysis
(16 hours).
Figure 7. Images of the copper electrodes before (a) and after
2 h
chronoamperometry at -1.55V vs. Ag/AgPF6 in the solutions with 4.88 vol.% of
1 M protic solutions: H2O (b), H2SO4 (c), KOH (d) and CsHCO3 (e). The
measurements conducted at 80 bar and 40 °C in 35 mM TBAPF6 solution of
scCO2, 29.3 vol.% acetonitrile.
During the experiments, a white colloidal dispersion appeared
around the electrode (supporting information, Figure S8) in one
phase mixture of scCO2/acetonitrile/water/TBAPF6 and it was
collected on the bottom of the reactor at the end of the
electrolysis. In order to have enough material for the elemental
analysis, the electrolysis was performed for 16 h. After
electrolysis, the system was depressurized and the white
powder in acetonitrile/water/TBAPF6 was dried on air. The color
of the powder changed from white to a bluish green tint after
drying. Elemental analysis showed residues of copper in the
analyte indicating the corrosion of Cu electrode (CuOx
formation) and its subsequent dissolution into the solution as
confirmed by the pitting corrosion visible in Figure 8 d. Copper
oxide formation was also observed in between CV when no
After electrolysis the large copper electrode was investigated
through scanning electronic microscopy and revealed multiple
features on the surface (Figure 8 a, b). The surface was quite
homogeneous, apart that on the bubbles spots. These spots
varied in size from 10 to 150 µm and were most likely formed in
the positions where water droplets adhered at the surface of the
electrode. ETD measurement indicated uneven coverage with C
and O elements on copper electrode surface. Besides in a few
spots nitrogen was detected. The bubble on Figure 8 a shows a
coffee ring effect enriched in O, while Figure 8 b show that the
bubble spot is mostly enriched in C and N. No P or F elements
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
reduction currents were applied. This layer was reduced during
conditions, where CH3OH is formed in larger amounts in slightly
acidic, slightly basic, or neutral solutions.
the first scan of
a subsequent CV or during electrode
pretreatment (supporting information). Copper oxidation is not to
be expected at the reduction potentials (in accordance to
Pourbaix diagram of copper), however, it could be the result of
chemical copper oxidation in scCO2/water/acetonitrile mixture.
After 8 h of chronoamperometry at –1.5 V vs. Ag/AgPF6, the
plain electrode became slightly porous, with the pore size of
approximately 1.0 μm. Dissolved Cu2+ ions could form CuCO3 as
an intermediate product. In the presence of water, CuCO3 is only
stable at CO2 pressure above 4.6 bar and pH between 4
and 8.[55] At ambient conditions, it is unstable and reacts with
water to form bluish green Cu2CO3(OH)2. Higher amounts of
water seemed to promote the electrodes corrosion and powder
formation. The solid particles formation was not found in the
systems without addition of H2O.
GC and HPLC analysis of gaseous and liquid products showed
formation of H2, CO, CH3OH, C2H5OH and HCOOH products,
which are summarized in Table 1. In contrast to literature, where
CH4 and C2H4 are common products on copper electrodes in
aqueous solutions at ambient conditions, they were not found in
scCO2 mixture. Besides, in comparison to the media with low
protic availability (like wet acetonitrile or ionic liquids), where CO
–
is the main product with possible formation of (COOH)2 and
– [3,23,56]
HCO3 .
In our case the synthesis in scCO2 have resulted
not only in CO but also in the organic products. Hydrogen was
present in all the samples in the range between 4.1 to
11.5 % FE, which is significantly lower than that at ambient
conditions. HCOOH formation and CO formation was observed
in all the samples.
Regarding the various protic solutions, a change in product
distribution was observed. In the solutions with 1 M KOH and
CsHCO3, no hydrocarbons were produced, ethanol was
observed only in the latter sample. HCOOH was formed for both
samples in scCO2, but was not reported at ambient conditions.
1 M KOH solution had a higher amount of hydrogen (average
10.3 % FE) and low amounts of CO (4.3 % FE). In literature,
slightly higher amounts of hydrogen were produced in KHCO3 in
comparison to CsHCO3 solution. The case of lower amount of
CO2 reduction products in hydrated K+ solution was attributed to
the increase amount of OH– and local pH, which results in slight
decrease in CO2 concentration at the surface of the electrode.[28]
However, in the case of scCO2, the amount of carbon based
liquid and gaseous products were smaller. Addition of 1 M
CsHCO3 aqueous solution to scCO2/acetonitrile/TBAPF6 mixture
resulted in lower amount of hydrogen formed (4.3 % FE). The
reduction was more selective towards ethanol (11.05 % FE),
which was noticed only in the presence of this cation. This could
occur due to the higher electrostatic interaction of Cs+ at the
electrode and its ability to locally decrease the pKa and buffer
the changes in pH. Application of neutral and acidic proton
donors as H2O and H2SO4 resulted in slightly higher amounts of
formic acid formation of 14.5 and 16.4 % FE, respectively. This
effect is similar to what is observed in aqueous solutions in
acidic medium.[32] Additionally, use of water yielded 7.5 % FE of
CH3OH. Similar behaviour was noticed in literature at ambient
Figure 8. SEM images of copper surface at different magnifications treated for
2 h at –1.55 V vs. Ag/AgPF6 in 35 mM TBAPF6, scCO2, 29.3 vol.% acetonitrile
and 4.88 vol.% H2O solution (a, b); pitting corrosion of copper electrode before
(c) and after (d) 16 h electrolysis at – 1.3 V V vs. Ag/AgPF6 in 35 mM TBAPF6,
scCO2, 29.3 vol.% acetonitrile and 1.48 vol.% H2O solution. Experiments were
performed, at 80 bar, 40 °C.
Difference in cyclic voltammograms and product distribution
demonstrate the successful transfer of protons and cations from
the aqueous solution to the surface of the electrode. However,
the effect of pH in scCO2 was different and not as dramatic as in
aqueous solution at ambient conditions, due to the predominant
amount of scCO2 and acetonitrile in the system and very little
amounts of aqueous solution dissolved in it. Besides, the bulk
pH of the initially neutral or basic solutions changed under high
pressure CO2 conditions to more acidic, due to the carboxylation
of water.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
Polymerisation of acetonitrile appeared to be more likely under
higher pressure, as some amounts of carbon and nitrogen were
found on the surface of the electrode.
Table 1. Range of faradaic efficiencies of gaseous and liquid products on
copper electrode at –1.55 V vs. Ag/AgPF6.
Sample
Charge
[C]
Faradaic efficiency [%]
Total faradaic efficiency did not exceed 34 %. Re-oxidation on
counter electrode, losses during sample transfer, or graphitic
carbon formation at the electrode were the main reasons. In
addition, copper corrosion was observed on the electrode
surface and copper carbonate was recovered in the cell.
Based on the performed experiments, several practical
strategies can be employed to increase electrocatalytic activity
and selectivity towards alcohols, by: i) use of bulky
tetrabutylammonium electrolytes and co-solvents like acetonitrile,
ii) application of aqueous solutions with large alkali metal cations
like CsHCO3, iii) separation of working and counter electrode
compartments, to save the products from re-oxidation.
H2
CO
CH3OH
C2H5OH
0
HCOOH
Sum
34
H2O
9.5–
6.4–
4.8–
6.8–
14.0–
10.4
7.4
5.6
8.1
14.9
KOH
7.4–
11.8
9.1–
11.5
3.2–
5.5
0
0
0
0
7.0–
12.4
24.4
23.3
28.7
CsHCO3
H2SO4
9.9–
11.4
4.1–
4.5
2.8–
3.4
8.2–
13.9
4.5–
5.1
5.7–
8.1–
3.6–
0
14.1–
9.5
8.7
4.2
18.6
Several questions remain open. The electrode surface
undergoes severe modification during the experiments and we
were not able to allocate a large amount of the charge injected
in the system. Furthermore, there were large variations in the
current transients during electrolysis which are probably linked
to the formation of a different phase on the electrode. Besides,
although the environment was reducing and the potentials
applied were clearly cathodic, copper underwent chemical
corrosion in the supercritical mixture.
Total faradaic efficiency was low in all cases; a maximum of
approximately 34 % FE was reached with water. The remaining
percentage of faradaic efficiency can account for losses during
depressurisation and gas and liquid sample transfer to the GC
and HPLC, product re-oxidation at the counter electrode and
black carbon formation on the electrode surface. Although most
of the products are water-soluble and should have partitioned in
the water phase at the bottom of the cell it cannot be neglected
that a part of these could have been re-oxidized at the counter
electrode.
Experimental Section
Electrochemical reduction of supercritical CO2 to methanol and
ethanol was reported for the first time. However, higher
efficiencies of the products are still desired.
Materials
High-purity CO2 (> 99.998 %, Carbagas) was used in all experiments.
1 M concentrations of aqueous solutions were prepared from KOH
(≥ 99.97 %, anhydrous, Sigma Aldrich), CsHCO3 (99.9%, Sigma Aldrich),
H2SO4 (98 %, Merck), 2,2,2-trifluoroethanol (TFE) (Reagent plus 99 %,
Sigma Aldrich) and water and used as proton donor solutions. Deionized
water was used throughout the experiments (Milli-Q, resistivity
18.2 MΩ·cm). Acetonitrile (> 99.9 %, Alfa Aesar, HPLC Grade) was used
as received. Tetrabutylammonium hexafluorophosphate (TBAPF6)
electrochemical analysis grade (≥ 99.0 %, Sigma Aldrich) was purified to
remove traces of methanol by dissolving it in acetonitrile and drying with
a vacuum pump at 65 °C.
Conclusions
In this paper, the electrochemical reduction of supercritical CO2
on copper catalysts was investigated. Supercritical CO2 was
used as a solvent and reagent itself. Several parameters such
as electrolyte, acetonitrile, and water concentration have been
tuned to increase the conductivity of the scCO2 media.
Two kinds of copper electrodes were used. The first electrode was made
from a copper rod (> 99.99 %) tightly sealed in PEEK material and had a
flat surface area of 0.2 cm2. The second electrode was a flat copper foil
(> 99.99 %) of 8.0 cm2 of area. Sandpaper (Buehler) with the grades 9, 3,
1, and 0.5 µm was used for polishing the electrodes.
The system was super-saturated with water which provided a
ten-fold increase in the current. The excess of water formed an
aerosol in the reactor, condensed on the walls and on the
electrodes, and concentrated at the bottom of the reactor. This
was used as expedient to separate the products of the
electrolysis.
Equipment
CO2 reduction under supercritical conditions has shown some
unique qualities over the conventional electrochemical reduction
by significant suppression of hydrogen evolution and possibility
to tune the selectivity towards certain products changing the
protic characteristic of the added water solution. The addition of
1 M H2SO4, KOH, and CsHCO3 to scCO2/acetonitrile/TBAPF6
systems allowed higher currents in comparison to pure water. In
case of H2O and 1 M CsHCO3, methanol and ethanol with
7.5 % and 11.1 % FE, respectively, were formed. Use of water
and H2SO4 solution had higher selectivity towards formic acid.
Unlike in literature, no formation of hydrocarbons was observed.
The experiments were conducted in two reactors. One reactor had a
volume of 41 ml and was mainly used for electrochemical experiments,
while the second reactor with a volume of 339 ml was equipped with
sapphire glass and was mainly used for optical observation of the phase
change. Both of the reactors were adapted for the electrochemistry in the
scCO2 by implementation of the three electrode setup. Copper disk
(0.2 cm2) or copper foil (8.0 cm2) were used as working electrodes, while
platinum mesh and silver wire as counter and quasi-reference electrodes,
respectively. The distance between working and counter electrodes was
5 mm, with quasi-reference electrode in the middle.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
The 41 ml reactor (stainless steel 304) was connected with a CO2 syringe
pump (Teledyne, ISCO Model 100DX) and a heating water pump
(Pharmacia Biotech, MultiTemp III). The reactor had a shape of cylinder
and was placed on a magnetic stirrer. Temperature was controlled by
circulating water through a stainless steel jacket wrapping the reactor.
More information on the experimental setup can be found in the
supporting information. The 339 ml reactor (view cell) (stainless steel
1.4542) was connected to a pressure pump (Lewa LDE3 K511) and
heated by four heating sticks equipped with temperature controller. The
reactor had a cylindrical shape and was stirred by the magnetic stirrer
(Premex 1.4980).
Keywords: copper
supercritical carbon dioxide • supercritical fluids
• electrochemistry • high pressure •
[1]
[2]
J. Qiao, Y. Liu, F. Hong, J. Zhang, Chem. Soc. Rev. 2014, 43, 631.
J. Qiao, Y. Liu, J. Zhang, Electrochemical reduction of carbon dioxide.
Fundamentals and technologies, CRC Press, Boca Raton, 2016.
A. P. Abbott, C. A. Eardley, J. Phys. Chem. B 2000, 104, 775.
a) S. Nakagawa, A. Kudo, M. Azuma, T. Sakata, Journal of
[3]
[4]
Electroanalytical Chemistry and Interfacial Electrochemistry 1991, 308,
339; b) K. Hara, A. Kudo, T. Sakata, Journal of Electroanalytical
Chemistry 1995, 391, 141; c) K. Hara, A. Kudo, T. Sakata, Journal of
Electroanalytical Chemistry 1995, 386, 257; d) M. Todoroki, K. Hara, A.
Kudo, T. Sakata, Journal of Electroanalytical Chemistry 1995, 394, 199;
e) F. Köleli, D. Balun, Applied Catalysis A: General 2004, 274, 237.
J. Li, G. Prentice, J. Electrochem. Soc. 1997, 144, 4284.
K. Hara, J. Electrochem. Soc. 1994, 141, 2097.
R. Kas, R. Kortlever, H. Yılmaz, M. T. M. Koper, G. Mul,
CHEMELECTROCHEM 2015, 2, 354.
Pressure was monitored using
a piezoelectric pressure sensor
(Swagelok) in both cases. Electrochemical measurements were
performed with a BioLogic potentiostat SP-300 operating in floating mode.
[5]
[6]
[7]
Method
[8]
[9]
A. Bandi, C. U. Maier, W. H. Bloss, DE4126349A1, 1993.
A. Bandi, M. Specht, T. Weimer, K. Schaber, Energy Conversion and
Management 1995, 36, 899.
The reactor was purged with CO2 for 10 minutes prior the measurement
to remove the air. Under constant CO2 stream, it was then filled with a
certain amount of TBAPF6 electrolyte, acetonitrile, and/or protic solution.
Afterwards, the reactor was sealed, heated to the temperature of 40 °C
and pressurized with CO2 up to 80 bar. The reaction mixture was
constantly stirred for about 20 min to achieve a homogeneous distribution
of all components. The electrochemical measurements were performed
10 minutes after stirring under quiescent conditions. Electrochemical
investigations on the effect of electrolyte were performed in scCO2
system with 29.3 vol.% (12 ml) acetonitrile and TBAPF6. Effect of
acetonitrile was tested in scCO2 mixture with 35 mM TBAPF6, acetonitrile,
and 0.029 vol.% of water. Effect of proton donors, was studied in scCO2
mixture with 4.88 vol.% of 1 M protic solution or water, 29.3 vol.% of
acetonitrile, and 35 mM TBAPF6. A silver wire (reference electrode) was
oxidized prior the measurements in the TBAPF6 acetonitrile electrolyte to
form Ag/AgPF6 layer for higher stability and reproducibility of the quasi-
reference electrode. Potential of Ag/AgPF6 in acetonitrile (without CO2) is
–200 mV vs. SHE.[57]
After the electrolysis, the reactor was cooled to 20 °C. All the gaseous
samples were transferred into a 30 ml stainless steel high-pressure
reactor for transportation to the analytical equipment. The gas samples
were analysed by gas chromatography equipped with TCD and FID
detectors (SRI instruments). Collection of liquid samples was followed by
alcohols quantification on gas chromatograph (Perkin–Elmer, Clarusgas
500) and organic acids on high-performance liquid chromatograph
(UltiMate 3000, Thermo Scientific).
[10] A. Kudo, J. Electrochem. Soc. 1993, 140, 1541.
[11] R. B. Gupta, J.-J. Shim, Solubility in supercritical carbon dioxide, CRC
Press, Boca Raton, 2007.
[12] M. E. Philips, M. R. Deakin, M. V. Novotny, R. M. Wightman, J. Phys.
Chem. 1987, 91, 3934.
[13] A. C. Michael, R. M. Wightman, Anal. Chem. 1989, 61, 270.
[14] D. Niehaus, M. Philips, A. Michael, R. M. Wightman, J. Phys. Chem.
1989, 93, 6232.
[15] P. N. Bartlett, D. C. Cook, M. W. George, J. Ke, W. Levason, G. Reid,
W. Su, W. Zhang, Phys. Chem. Chem. Phys. 2010, 12, 492.
[16] a) A. P. Abbott, C. A. Eardley, J. Phys. Chem. B 1999, 103, 2504; b) A.
P. Abbott, C. A. Eardley, J. C. Harper, E. G. Hope, Journal of
Electroanalytical Chemistry 1998, 457, 1; c) D. L. Goldfarb, H. R. Corti,
J. Phys. Chem. B 2004, 108, 3358.
[17] P. N. Bartlett, D. A. Cook, M. W. George, A. L. Hector, J. Ke, W.
Levason, G. Reid, D. C. Smith, W. Zhang, Phys. Chem. Chem. Phys.
2014, 16, 9202.
[18] J. Jun, P. S. Fedkiw, Journal of Electroanalytical Chemistry 2001, 515,
113.
[19] S. Chanfreau, P. Cognet, S. Camy, J.-S. Condoret, Journal of
Electroanalytical Chemistry 2007, 604, 33.
[20] K. E. Toghill, P. Voyame, D. Momotenko, A. J. Olaya, H. H. Girault,
Phys. Chem. Chem. Phys. 2012, 15, 972.
[21] J. A. Branch, P. N. Bartlett, Phil. Trans. R. Soc. A 2015, 373, 20150007.
[22] J. Harper, PhD, Leicester, 1998.
[23] G. Zhao, T. Jiang, B. Han, Z. Li, J. Zhang, Z. Liu, J. He, W. Wu, The
Journal of Supercritical Fluids 2004, 32, 287.
[24] a) A. P. Abbott, J. C. Harper, Phys. Chem. Chem. Phys. 1999, 1, 839;
b) A. P. Abbott, J. C. Harper, Faraday Trans. 1996, 92, 3895.
[25] J. A. Branch, D. A. Cook, P. N. Bartlett, Phys. Chem. Chem. Phys. 2015,
17, 261.
[26] a) M. A. Méndez, P. Voyame, H. H. Girault, Angew. Chem. Int. Ed. 2011,
50, 7391; b) P. Voyame, K. E. Toghill, M. A. Méndez, H. H. Girault,
Inorg. Chem. 2013, 52, 10949.
[27] A. Bandi, C. MAIER, W. BLOSS, DE4126349 C2, 1993.
[28] M. R. Singh, Y. Kwon, Y. Lum, J. W. 3. Ager, A. T. Bell, Journal of the
American Chemical Society 2016, 138, 13006.
[29] A. Schizodimou, G. Kyriacou, Electrochimica Acta 2012, 78, 171.
[30] H. Y. Murata A., Bull. Chem. Soc. Jpn. 1991, 64, 123.
[31] J. Albo, M. Alvarez-Guerra, P. Castaño, A. Irabien, Green Chem 2015,
17, 2304.
Electrodes were analysed by scanning electronic microscopy (SEM)
(Bruker Nano) coupled with Everhart-Thornley Detector (ETD) option
before and after the electrolysis. The solid particles formed after
electrolysis were analysed by inductively coupled optical emission
spectrometry (ICP-OES) (Varian 720-ES). In order to produce necessary
amount of particles, the electrolysis was performed for 16 h. The particles
were then dried from the solution on air at 50 ˚C.
[32] D. T. Whipple, E. C. Finke, P. J. A. Kenis, Electrochem. Solid-State Lett.
2010, 13, B109.
[33] C. G. Vayenas, R. E. White, M. E. Gamboa-Aldeco, Modern aspects of
electrochemistry, Springer Science+Business Media, New York, 2008.
[34] Y. Hori in Modern Aspects of Electrochemistry (Eds.: C. G. Vayenas, R.
E. White, M. E. Gamboa-Aldeco), Springer New York, New York, NY,
2008, pp. 89–189.
[35] S. Kaneco, K. Iiba, N. Hiei, K. Ohta, T. Mizuno, T. Suzuki, Electrochim.
Acta 1999, 4701.
[36] M. R. Thorson, K. I. Siil, P. J. A. Kenis, J. Electrochem. Soc. 2012, 160,
F69-F74.
[37] M. R. Singh, E. L. Clark, A. T. Bell, Physical chemistry chemical
physics : PCCP 2015, 17, 18924.
[38] M. Gattrell, N. Gupta, A. Co, Journal of Electroanalytical Chemistry
2006, 594, 1.
[39] a) K. J. P. Schouten, Y. Kwon, C. J. M. van der Ham, Z. Qin, M. T. M.
Koper, Chem. Sci. 2011, 2, 1902; b) R. Kortlever, J. Shen, K. J. P.
Acknowledgements
This work was supported by the Cluster of Excellence RESOLV
(EXC 1069) funded by the Deutsche Forschungsgemeinschaft.
The authors would like to acknowledge Prof. Hubert Girault for
the access to his laboratories and his support throughout the
project. We are also thankful to Prof. Raffaella Buonsanti and Dr.
Heron Vrubel for helping with gas and liquid chromatography.
Victor Costa Bassetto thanks the Swiss National Science
Foundation (SNSF project No. 154297).
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
Schouten, F. Calle-Vallejo, M. T. M. Koper, The journal of physical
chemistry letters 2015, 6, 4073.
Electrochimica Acta 2005, 50, 3726; c) R. A. Rincón, A. Battistel, E.
Ventosa, X. Chen, M. Nebel, W. Schuhmann, ChemSusChem 2015, 8,
560.
[40] K. P. Kuhl, T. Hatsukade, E. R. Cave, D. N. Abram, J. Kibsgaard, T. F.
Jaramillo, Journal of the American Chemical Society 2014, 136, 14107.
[41] a) W. Tang, A. A. Peterson, A. S. Varela, Z. P. Jovanov, L. Bech, W. J.
Durand, S. Dahl, J. K. Nørskov, I. Chorkendorff, Phys. Chem. Chem.
Phys. 2012, 14, 76; b) C. W. Li, M. W. Kanan, Journal of the American
Chemical Society 2012, 134, 7231.
[50] H. Zheng, K. Li, G. D. Cody, C. A. Tulk, X. Dong, G. Gao, J. J. Molaison,
Z. Liu, M. Feygenson, W. Yang et al., Angew. Chem. Int. Ed. 2016, 55,
12040.
[51] Q. Yu, M. Ye, L. Lu, J. Cheen, F. Wang, Chinese Journal of Polymer
Science 1988, 6.
[42] A. Jouyban, S. Soltanpour, H.-K. Chan, International Journal of
Pharmaceutics 2004, 269, 353.
[43] Y. Oh, H. Vrubel, S. Guidoux, X. Hu, Chemical communications
(Cambridge, England) 2014, 50, 3878.
[44] Á. Díaz-Duque, A. P. Sandoval-Rojas, A. F. Molina-Osorio, J. M. Feliu,
M. F. Suárez-Herrera, Electrochemistry Communications 2015, 61, 74.
[45] Y. Tomita, S. Teruya, O. Koga, Y. Hori, J. Electrochem. Soc. 2000, 147,
4164.
[52] E. Oikawa, K. Mori, G. Saito, BCSJ 1966, 39, 1182.
[53] J. K. Foley, C. Korzeniewski, S. Pons, Can. J. Chem. 1988, 66, 201.
[54] a) Y. Hori, H. Konishi, T. Futamura, A. Murata, O. Koga, H. Sakurai, K.
Oguma, Electrochimica Acta 2005, 50, 5354; b) D. W. DeWulf, J.
Electrochem. Soc. 1989, 136, 1686; c) J.-F. Xie, Y.-X. Huang, W.-W. Li,
X.-N. Song, L. Xiong, H.-Q. Yu, Electrochimica Acta 2014, 139, 137; d)
H. Yano, T. Tanaka, M. Nakayama, K. Ogura, Journal of
Electroanalytical Chemistry 2004, 565, 287.
[46] C. Wohlfarth, Numerical data and functional relationships in science and
technology. Group 4: @Macroscopic and technical properties of matter,
Springer, Berlin, Heidelberg, 1991.
[55] H. Gamsjager, W. Preis, J. Chem. Educ. 1999, 76, 1339.
[56] M. C. Figueiredo, I. Ledezma-Yanez, M. T. M. Koper, ACS Catal. 2016,
6, 2382.
[47] J. H. Yoon, M. K. Chun, W. H. Hong, H. Lee, Ind. Eng. Chem. Res.
1993, 32, 2881.
[48] C. Peng, J. P. Crawshaw, G. C. Maitland, J. P. Martin Trusler, D. Vega-
Maza, The Journal of Supercritical Fluids 2013, 82, 129.
[49] a) G. C. Barker, Journal of Electroanalytical Chemistry and Interfacial
Electrochemistry 1972, 39, 484; b) C. Gabrielli, F. Huet, R. P. Nogueira,
[57] R. A. Mantz, H. C. de Long, M. Hagiwara, P. C. Trulove, G. R. Stafford,
D. A. Costa, Molten salts XIV. Proceedings of the international
symposium, Electrochemical Society, Pennington, 2006.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201701205
ChemSusChem
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Olga Melchaeva*, Patrick Voyame,
Victor Costa Bassetto, Michael Prokein,
Manfred Renner, Eckhard Weidner,
Marcus Petermann, Alberto Battistel*
Page No. – Page No.
Electrochemical Reduction of Protic
Supercritical CO2 on Copper
Electrodes
Electrochemical reduction of CO2 in aqueous solutions is usually plagued with
massive hydrogen evolution. The reduction of CO2 in supercritical phase can be
tuned adding different protic aqueous solutions, like acidic, basic, or buffer
solutions. Different selectivities are achieved and, for the first time, it was possible
to convert supercritical CO2 to methanol and ethanol.
This article is protected by copyright. All rights reserved.