Role of the Adsorbed Oxygen Species in the Selective Electrochemical Reduction of CO2 to Alcohols and Carbonyls on Copper Electrodes

Article abstract of DOI:10.1002/anie.201706463

The electrochemical reduction of CO₂ into fuels has gained significant attention recently as source of renewable carbon-based fuels. The unique high selectivity of copper in the electrochemical reduction of CO₂ to hydrocarbons has called much interest in discovering its mechanism. In order to provide significant information about the role of oxygen in the electrochemical reduction of CO₂ on Cu electrodes, the conditions of the surface structure and the composition of the Cu single crystal electrodes were controlled over time. This was achieved using pulsed voltammetry, since the pulse sequence can be programmed to guarantee reproducible initial conditions for the reaction at every fraction of time and at a given frequency. In contrast to the selectivity of CO₂ reduction using cyclic voltammetry and chronoamperometric methods, a large selection of oxygenated hydrocarbons was found under alternating voltage conditions. Product selectivity towards the formation of oxygenated hydrocarbon was associated to the coverage of oxygen species, which is surface-structure- and potential-dependent.

Full text of DOI:10.1002/anie.201706463

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Role of the adsorbed oxygen species in the selective
electrochemical reduction of CO2 to alcohols and carbonyls on
copper electrodes
Authors: Cécile S. Le Duff, Matthew J. Lawrence, and Paramaconi
Rodriguez
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: Angew. Chem. Int. Ed. 10.1002/anie.201706463
Angew. Chem. 10.1002/ange.201706463
Link to VoR: http://dx.doi.org/10.1002/anie.201706463
http://dx.doi.org/10.1002/ange.201706463

10.1002/anie.201706463
Angewandte Chemie International Edition
COMMUNICATION
Role of the adsorbed oxygen species in the selective
electrochemical reduction of CO2 to alcohols and
carbonyls on copper electrodes
Cécile S. Le Duff, Matthew J. Lawrence and Paramaconi Rodriguez[a]
Abstract: The electrochemical reduction of CO2 into fuels has
gained significant attention recently as source of renewable carbon-
based fuels. The unique high selectivity of copper in the
electrochemical reduction of CO2 to hydrocarbons has called much
interest in discovering its mechanism. In order to provide significant
information about the role of oxygen in the electrochemical reduction
of CO2 on Cu electrodes, the conditions of the surface structure and
the composition of the Cu single crystal electrodes were controlled
over time. This was achieved using pulsed voltammetry, since the
pulse sequence can be programmed to guarantee reproducible initial
conditions for the reaction at every fraction of time and at a given
frequency. In contrast to the selectivity of CO2 reduction using cyclic
voltammetry and chronoamperometric methods, a large selection of
oxygenated hydrocarbons was found under alternating voltage
conditions. Product selectivity towards the formation of oxygenated
hydrocarbon was associated to the coverage of oxygen species,
which is surface-structure-and-potential dependent.
distribution found on Cu nanocatalysts. A possible explanation
to the diverse selectivity is the dynamic nature of the structure of
copper surfaces under CO2 electrochemical reduction
conditions.[4]
Recently, it has been noted that the effect of the role of oxygen
content on the copper catalyst also adds complexity to the
1f, 5]
system.[1e,
Some reports have claimed the formation of
methanol during CO2 reduction on Cu2O electrodes with faradaic
efficiencies (FE) of 38 %.[2a, 6] On the other hand, in a similar
system – Cu2O films – others have reported the formation of CO
(FE ∼ 45 %), formic acid (FE 30 %), ethanol (FE∼ 4.7 %) and
ethylene (FE∼93.9 %), but no methanol.[1f] Kas et al. reported
the formation of formic acid as the only liquid product (FE 20 %)
during CO2 reduction on Cu2O prepared by electrochemical
oxidation of copper electrodes.[1e] The causes of the different
selectivity have been ascribed to a myriad of factors such as the
oxidation state of the Cu, oxygen content on the films and the
roughness of the electrode (surface structure). The origin of the
different selectivity remains a mystery and its elucidation would
impact significantly on the development of improved catalysts for
the CO2 reduction.
In order to provide significant information of the role of oxygen in
the electrochemical reduction of CO2 on Cu electrodes, one
turns to simple experiments which allow more control over the
experimental conditions and, therefore, allow the separation of
the parameters that influence the selectivity. In order for such an
enterprise to be achieved, we have utilised Cu single crystal
electrodes with crystal structure (100) and (111) and an
electrochemical strategy to control the oxygen content at the
surface. The surface structure and the composition of the single
crystal electrodes were controlled over time. This was achieved
using pulsed voltammetry,[7] since the pulse sequence can be
programmed to guarantee reproducible initial conditions for the
reaction (Figure 1A) at every fraction of time and at a given
frequency.
Recycling of CO2 into fuels is a promising way to address the
challenge of reducing the CO2 emissions that contribute to
global warming. The electrochemical conversion of CO2 into
fuels is one of the most promising paths towards the realisation
of a sustainable fuel economy because it can be powered by
renewable sources (e.g., solar, wind, tidal, etc.). To date, copper
and copper oxide-derived surfaces are the only catalysts that
can electrochemically convert CO2 to high value and energy-
dense products such as methane, ethylene, formic acid,
methanol and ethanol, among others;[1] however, the efficiency
and the selectivity of these catalysts are far from optimal and the
parameters controlling these factors are not-fully understood.
Such differences in reactivity and selectivity have been
attributed to the surface area, particle size, surface structure,
roughness of the electrode and even to the history of the
electrode.[1a, 1e, 2]
Some understanding of the surface structure reactivity and
selectivity relationship for the CO2 electrochemical reduction
reaction has been gained from theoretical and experimental data
on copper single-crystal electrodes.[1a, 1d, 3] It has been reported
that on Cu (111) facets, the formation of methane is favoured;
whereas, ethylene formation is predominant on Cu(100)
surfaces.[1a] Although these studies provide an important insight
into the mechanism of the CO2 reduction reaction and into some
of the parameters affecting the catalytic activity, they have not
been able to provide an explanation of the broad product
A short pulse at high potential (-0.2, -0.25, - 0.3 or -0.35 V vs
RHE) was superimposed onto a normal voltammetry cycle to
periodically recover the initial reaction conditions. These short
pulses resulted in the co-adsorption of anions and hydroxide
species onto the copper electrodes.[8] They may also favour the
desorption/oxidation of the accumulated adsorbed intermediates.
In addition, they may avoid possible structural changes
associated to H2 brightening at negative potentials.
Figure 2 shows the pulsed voltammetry of the Cu(111) and
Cu(100) electrodes in a phosphate buffer solution at pH=7.9±0.1,
in both the presence and absence of CO2 in the solution. As
indicated in the figure, each of the curves corresponds to
different upper potential pulses. The cyclic voltammetry under
the same conditions for each electrode was included for
comparison. As can be seen in the cyclic voltammetry on the
inset of Figure 2A, the selected potential for the positive step
occurred in the region where the co-adsorption of phosphates
[a]
Dr. Cecile S. Le Duff, Matthew J. Lawrence, Dr. Paramaconi
Rodriguez
School of Chemistry
University of Birmingham
Edgbaston, Birmingham, B15 2TT, United Kingdom
E-mail: p.b.rodriguez@bham.ac.uk
and OH species took place.[8b]
formation of Cu oxides and dissolution of copper resulted in
At higher potentials, the
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/anie.201706463
Angewandte Chemie International Edition
COMMUNICATION
changes of the surface structure and an increase of the step-
density.[8b, 9]
currents in the absence of CO2, even at low potentials (see the
comparison in Figure S2). This suggests that, under pulse
voltammetry conditions, Cu electrodes are capable of reducing
CO2 at potentials as low as -0.4 V vs RHE. The larger currents
of the pulsed voltammetry, in comparison with the cyclic
voltammetry, are observed in the potential region between 0 V
and 1.2 V. At higher potentials, the current densities of both
types of voltammetries are similar.
One objective of this work was to separate the different
parameters that contribute to the product selectivity during CO2
reduction on copper electrodes, as a function of potential. In
order to confirm the stability of the surface under pulse
voltammetry conditions, we have recorded the blank
voltammetry of the single crystal electrodes in NaOH before and
after each pulse voltammetry. It has been previously reported
that the cyclic voltammetry of Cu single crystal electrodes in
NaOH is sensitive to the presence of structural changes.[10] As
can be seen in Figures S3 and S4, the blank voltammetries of
Cu(111) and Cu(100) in NaOH are in agreement with previous
results. Our results do not show any significant structural
changes after the step potential protocol, neither in the absence
nor the presence of CO2, when the step potential is kept below -
0.25 V vs RHE. When the step potential on the pulse
voltammetry was as high as -0.2 V vs RHE, significant changes
in the voltammetries were observed. There are most likely
associated to changes in the microrugosity of the surface which
can be associated to the lattice expansion of the copper due to
the formation of copper oxides and to the copper dissolution.[7-8]
Therefore, in order to eliminate the effect of changes of the
structure, the product characterization is reported only for the
step potentials at -0.35 V, -0.3 V and -0.25 V vs RHE.
Figure 1. (A) Pulse program for the pulsed voltammetry and (B) alternate
voltage electrolysis program. The coloured lines indicate the different positive
step potentials. Both the pulse time and pulse interval used during the pulse
voltammetry were 0.1 s; alternate voltage electrolysis was performed with a
frequency of 10 Hz over 60 minutes.
The product selectivity for CO2 reduction on Cu(111) and
Cu(100), as determined by on-line electrochemical mass
spectrometry (OLEMS), during pulse voltammetry, is shown in
Figure 3.
In both cases, Cu(111) and Cu(100), the pulse voltammetry in
the absence of CO2 showed larger current contributions between
-0.5 V and -0.8 V, in comparison with the cyclic voltammetry. In
this potential window, the voltammetry of Cu(111) showed a
peak at -0.55 V, which shifted towards more positive potentials
when the pulse at the higher potentials became more positive. In
addition, the total change of the process in the given potential
region increased. The pulse voltammetries also showed a shift
of the onset of hydrogen evolution reaction towards more
negative potentials and showed a clear change on the reaction
rate.
The blank pulse voltammetries of Cu(100) showed similar
behaviour to that described for Cu(111); however, differences in
the peak position can be perceived. In the case of Cu(100), a
broader peak appears at a more negative potential of -0.70 V vs
RHE. The hydrogen evolution reaction is also shifted towards
more negative potentials.
In the presence of a CO2 saturated solution (Figures 2B and 2D),
the faraday currents associated to the reduction of CO2 are
Figure 2. Pulse voltammetries of Cu(100) (left column) and Cu(111) (right
column) electrodes in phosphate buffer solution. (A) and (C) show the blank
voltammetries in the absence of CO2; (C) and (D) show the voltammetries in
the presence of CO2. The black curves correspond to cyclic voltammetries,
taken at (=10 mV/s), in the absence and presence of CO2. The inset
correspond to the cyclic voltammetries of the Cu(111) and Cu(100) in the
absence of CO2 (=50 mV/s). The arrows indicate the positive potential steps
during the pulse voltammetries.
significantly larger than in the cyclic voltammetries.
A
representative set of transients, recorded for the Cu(111) and
Cu(100) electrodes during the pulsed voltammetry experiments,
are shown in Figure S3. The two processes observed in the
blank voltammetries can also be distinguished in the presence of
CO2; the current, in the presence of CO2, is larger than the
This article is protected by copyright. All rights reserved.

10.1002/anie.201706463
Angewandte Chemie International Edition
COMMUNICATION
The product distribution of the gaseous species from the cyclic
voltammetries are also included for comparison. As can be seen,
the formation of hydrogen (m/z=2) under the pulse potential
program is significantly compromised in both single-crystal
electrodes when compared with the hydrogen formation during
the cyclic voltammetry. In fact, as indicated by the arrows, the
hydrogen production decreases as the pulse potential becomes
less negative. Similar to the product distribution of CO2 reduction
using cyclic voltammetry,[1a] the formation of methane (m/z=15)
and ethylene (m/z=26) produced under pulse potential
voltammetry also depends on the surface structure and on the
applied potential. On Cu(111), the formation of CH4 and C2H4
under the pulse potential voltammetries program takes place at
higher overpotentials when compared with cyclic voltammetry. In
addition, the production of CH4 and C2H4 decreases as the pulse
potential becomes less negative. This is associated to the
decrease in the coverage of the protons at the negative step
potentials which occurs due to an increase in the coverage of
the oxide species at the upper step potential. As previously
reported, it was observed in the cyclic voltammetries that
Cu(100) is more selective toward the formation of C2H4, in
comparison to the formation of CH4, at low overpotentials. Under
step potential conditions, the formation of C2H4 and H2 on
Cu(100) decreases as the upper step potential becomes more
positive. At the upper step potential of -0.2 V, the formation of
C2H4 is almost negligible. This can be associated to the
suppression of protons adsorbed during the step potential
regime. Surprisingly, the electrochemical reduction of CO2 on
Cu(100) shows a significant increase in the production of CH4 at
potentials between -0.7 V and 1.2 V vs RHE, when the step
potential is applied. A maximum rate of production of CH4
species was observed between -0.8 V and -1.0 V vs RHE when
the upper potentials were stepped at -0.3 V and -0.35 V vs RHE.
It is important to note that the potential at which the maximum
formation rate of CH4 appears (-0.9 V vs RHE) coincides with
the potential of the crossover of the cyclic voltammetry curve
and the curve of the pulse voltammetries for CO2 reduction and
at the potential at which the formation of H2 increases. In
addition, at this potential, a change in the Tafel slope of the
pulse voltammetries can be distinguished. Currently, we
postulate that this might be associated with the coverage,
adsorbate geometries and available adsorption sites. In addition
to the well-known gaseous product reported by electrochemical
mass spectroscopy, m/z=31, presumably associated to the
formation of methanol, was also detected. Surprisingly, signals
of m/z=31 were observed in both electrodes, Cu(111) and
Cu(100), during pulse voltammetries at potentials as low as -0.6
V vs RHE. These signals were observed only under pulse
potential conditions when the pulse potentials were -0.3 and -
0.25. In both cases, the amount of the fragment m/z=31
increased when the pulse potential increased from -0.35 V to -
0.3 V vs RHE; however, the signals of m/z=31 were not
observed during the cyclic voltammetries, nor during the pulse
voltammetry with pulse potential -0.25. Although m/z=31 might
be associated to the fragment CH3OH+ of methanol, it might be
also attributed to fragments of other products reported for the
CO2 reduction on copper electrodes (e.g., ethanol).[4b] Further
insights into the mechanism will be presented below, after the
analysis of the liquid products by NMR.
Figure 3. Volatile products with m/z=2, m/z=15, m/z=26 and m/z=31, probed
by OLEMS during the reduction of CO2 on Cu(111) and Cu(100) electrodes.
The black line corresponds to the volatile products formed during a cathodic
scan of one cyclic voltammetry. The green line, blue line and red line
correspond to the volatile products detected during the pulse voltammetries at
-0.25 V, -0.3 V and -0.35 V, respectively. The orange arrows indicate, for the
different products, the trends observed, as functions of the upper limit potential.
In order to identify the liquid products during the electrochemical
reduction of CO2 under step potential conditions, we have
performed long term experiments in an electrochemical cell with
reduced volume, as described in the experimental section and in
Figure S1. The experiments were performed by applying a
square wave potential with frequency of 10 Hz between -0.8 V
vs RHE and the positive potentials of -0.25 V, -0.3 V and -0.35 V
vs RHE, as indicated in Figure 1B. The experiments were
performed during 1 h and 800 L samples were collected every
15 min and analysed by 1H NMR, as described in the
experimental section and supporting information. Figures S6-S9
show the 1H NMR characterisation of the products of the CO2
reduction after the square wave potential protocol for the step
potentials at -0.3 V and -0.25 V. Table 1 summarises the
products observed under these experimental conditions. Cyclic
voltammetries of both electrodes, before and after the
electrolysis, were recorded and confirm the stability of the
surface structure of the electrodes under these conditions
(Figure S10). The small changes observed in the blank
voltammetries of Cu(111) and Cu(100) in Figure S10 can be
associated to small changes in the surface structure (i.e., the
formation of steps and kinks) or to the presence of adsorbed
species generated during electrolysis.
This article is protected by copyright. All rights reserved.

10.1002/anie.201706463
Angewandte Chemie International Edition
COMMUNICATION
Previous works, using video scan rate in-situ STM, by
Matsushima et al. have shown that hydrogen induces
reconstruction of Cu(100) from a (1x1) unit cell to a c(p×8) unit
cell.[11] However, no reconstruction was found at pH 3, nor
higher. The authors have also reported that this reconstruction
process is reversible.
(COHads) or (CHOads) + 5e-+5H+  CH4 + H2O
(COHads) or (CHOads) + 3e-+3H+  CH3OH
(I)
(II)
The decrease in the formation of H2 on both electrodes and the
decrease in the formation of CH4 on Cu(111) (as observed in the
OLEMS under step potential conditions) can both be attributed
to the decrease of the H+ adsorption due to the presence of OH
adsorbed as induced at the most positive step potential. The
appearance of methanol or ethanol (m/z=32 in the OLEMS
experiments) is consistent with the decrease of hydrogen
coverage (as given by the stoichiometry in Eq. II vs Eq. I).
On the other hand, Soriaga’s group reported that, in alkaline
media,
a
copper polycrystalline electrode undergoes
reconstruction during hydrogen evolution conditions.[4] This slow,
sequential transformation showed first the formation of a
Cu(111) phase followed by the formation of Cu(100).
As can be seen on Table 1, a larger selection of oxygenated
hydrocarbons was found under alternate voltage conditions than
under DC chronoamperometric conditions.[2g, 4a] Interestingly, the
product selectivity towards C1, C2 and C3 oxygenated
hydrocarbons as a function of the surface structure was also
observed under the alternate voltage conditions.
On the other hand, ethylene is formed by dimerization of
intermediate species through
according to Equation (III):
a 6 electron-proton transfer
2(COHads) or 2(CHOads) + 6e-+ 6H+ C2H4 + 2H2O
(III)
If the reaction mechanism follows a Langmuir-Hinshelwood
mechanism, such coupling reaction requires the presence of two
adjacent C-containing adsorbates. The proximity of the reactants
will be given by the surface coverage and the surface structure
of the adsorbate species (e.g., surface oxides, phosphate anions
and reactive intermediates); therefore, the formation of C2H4
must be driven by an increase in the surface concentration of C-
containing adsorbate, to the detriment of proton adsorption.
The presence of alcohols and ketones in the C2 and C3
products indicates that the C–C coupling might occur before at
least one of the two carbon-oxygen bonds in CO2 is broken,
accompanied or followed by hydrogenation reactions. The
increase of COads and COHads species on the surface, together
with the decrease of the hydrogen availability on the surface, are
expected; the coupling of CHx ads and COads or COHads species
results in the formation of terminal alcohols and ketones.
After 60 minutes of electrolysis, it was observed that formic acid
was a major product. Previous studies proposed that the
mechanistic pathway toward formic acid formation is
independent from the formation of carbon monoxide, and that
formic acid cannot be further reduced to other products. It was
assumed, therefore, that the appearance of formic acid must be
associated to a change in the reaction mechanism over the
course of the reaction. It has also been shown that the formation
of formic acid and formate is favoured under alkaline conditions.
Therefore, our observation of the formation of formic acid is
consistent with a change of pH due to the consumption of
protons during the electrolysis. Despite the use of a buffer
solution, during the electrolysis, small changes of pH (from
pH=7.9 to pH=8.2 after 30 minutes and pH=8.6 after 1 hour)
were measured in the solution; these might result in small
changes in the selectivity.
Table 1. Liquid products probed by 1H NMR during CO2 reduction under the
alternate voltage electrolysis with a frequency 10 Hz. [compound] is present in
traces; compound is most intense peak.
Applied potential
Surface
Electrolysis
time
structure
-0.8 V to -0.3 V
-0.8 V to -0.25 V
C1: CH3OH
C2: CH3CH2OH
C3: (CH3)2CO
C1: CH3OH
C3: (CH3)2CO
30 min
60 min
Cu(100)
Cu(111)
C1: HCOO−
C2: CH3CH2OH
C3: (CH3)2CO
C1: HCOO−, CH3OH
C2: CH3CH2OH
C3: (CH3)2CO
C1: CH3OH
C3: [(CH3)2CO]
30 min
60 min
C1: [HCOO−], CH3OH
C1: HCOO−, CH3OH
C2: [CH3COO-]
C1: HCOO−, CH3OH
The C2 and C3 oxygenated products – ethanol, acetaldehyde
and acetone – were observed only on Cu(100); whereas, C1
oxygenated products – formic acid and methanol – were formed
on both surface structures. Other products, such as
acetaldehyde and acetate, were observed at step potentials
more positive than -0.2 V; however, significant changes in the
surface structure were observed after the electrolysis (see
Figure S11); therefore, changes in the selectivity could be
associated to uncontrolled factors during the electrolysis. Under
these experimental conditions, no other C3-C4 products
previously reported for the CO2 reduction on copper electrodes
were detected.[12]
It has been proposed, in the literature, that carbon monoxide is
the common intermediate in the formation, during the CO2
reduction, of a wide variety of products, such as CH4, CH3OH
and C2H4. During CO2 reduction, the adsorbed carbon monoxide
is formed first (2e-/2H+), followed by the formation of an
adsorbed carbonyl species, such as (CHOads) or (COHads).[13]
These intermediates later react to form methane and/or
methanol through an electron-proton transfer, as described in
Equation (I) or Equation (II):
Our results confirm that selectivity toward the formation of C1,
C2 and C3 products strongly depends on the geometry of the
copper surface; C2 and C3 products are more likely to form on
Cu(100) than on Cu(111). The formation of oxygenated species
– formate, methanol, ethanol, acetone and acetaldehyde – is not
necessarily linked to the surface structure, nor to the roughness
of the surface. The formation of oxygenated hydrocarbon may
be associated to the adsorption of OH species to the surface
when the potential is stepped to the upper potential. It has been
This article is protected by copyright. All rights reserved.

10.1002/anie.201706463
Angewandte Chemie International Edition
COMMUNICATION
proposed that the promotion of CO2 chemisorbed on oxygen-
modified Cu surfaces, and the subsequent hydrogenation, is a
key feature of the mechanism in the reduction of CO2 to
methanol.[14] Early work from Bowker reported the surface
structure sensitivity and the effect of the pre-dose oxygen
species in the formation of different intermediates in the gas
phase. Methoxy species (CH3O) are formed when dosing
methanol onto Cu(110). On the other hand, dosing methanol
onto a partially oxidised Cu(110) produced a formate species.[15]
A recent work has revealed the importance of copper suboxide
and the critical role of water in the first step of activation of CO2
on a Cu(111) surface in the gas phase.[16]
Product selectivity should be associated to OH coverage, which
is surface structure and potential dependent. Friebel et al.
reported that, at potentials close to the onset of anodic
dissolution/oxide growth of Cu(111), sulfate and water co-adsorb
to form a hexagonal superstructure with a (13×13) R13.9° unit
cell.[7a] Similarities in structure are expected on the phosphate,
given the nature of the anion. In the case of the Cu(100),
Cruickshank et al. reported the formation, in a diluted solution of
HClO4 and H2SO4, of a √2x√2 R45° structure associated to the
adsorption of either oxygen or OH on the (100) surface.[17]
However, it is important to mention that later results by Vogt et al.
suggested that such surface structure of the adlayer is
influenced by the presence of chloride contaminants in the
electrolyte.[18]
Experimental Section
Two sets of electrochemical cells were used in the experiments (see SI).
In all the experiments, a gold wire was used as a counter electrode and a
Hg/HgO electrode was used as a reference electrode. All the graphs are
presented in reversible hydrogen electrode (RHE) scale. Electrochemical
measurements were performed with an Autolab PGSTAT12. The
phosphate buffer solution (pH=7.9; 94 mL of 1 M K2HPO4 + 6 mL of 1 M
KH2PO4) was prepared from Sigma-Aldrich ACS reagent chemicals
(>98% and >99.0%, respectively), and ultra-pure water (Elga PureUltra,
18.2 MΩ cm, 1 ppb total organic carbon). Argon (Ar, N66) was used to
deoxygenate all solutions and CO2 (BOC) was used to saturate all
solutions. Cu(100) and Cu(111) single crystal electrodes from (icryst)
were used. Prior to each experiment, the electrodes were prepared by
electropolishing in 66% orthophosphoric acid at a voltage of +2 V and,
subsequently, rinsing with ultra-pure water. All the measurements have
been performed under steady-state conditions and all the
voltammograms correspond to the first cycle. The uncompensated
solution resistance was measured using the positive feedback mode of
the potentiostat and all voltammetries have been corrected by the
resistance.
Product characterization by On-line electrochemical mass spectrometry
and 1H NMR
On-line Electrochemical Mass Spectrometry (OLEMS) was used to
detect the gaseous products formed during the reaction. The reaction
products at the electrode interface were collected with a small tip
positioned close to the electrode.
Until now, the electrochemical approaches did not allow for
control over the surface coverage of an oxygen specie during an
experiment. Linear voltammetries or chronoamperometric
techniques are not appropriate techniques since the Cu-oxides
(O and OH species) depend on the electrode potential, the
direction of the scan and the previous history of the electrode. In
this report, we have demonstrated that conditions can be
controlled over time. This is achieved by using Cu single crystal
electrodes, pulsed voltammetry and the AC electrolysis because
the pulse potential sequence can be tuned to ensure
reproducible initial conditions without compromising the surface
structure (S4-5 and S10).
We propose here that the decrease of H2 availability for the
hydrogenation reaction results in the formation of C-C bonds
and of oxygenated products, and that the selectivity of the
oxygenated products is related to the surface coverage of OH on
the different surface structures. Further tuning of the product
selectivity can be achieved, not only by changing parameters in
the pulse program (such as potential window, frequency, duty
cycle and AC offset), but also by changing the surface coverage
of the OH specie. The different OH coverage can be tweaked
though anion adsorption, non-covalent interaction or pH.[19]
Even though it was not the main subject of this work, it could
also be concluded that the cooperativity effect between COads
and OHads could be playing a role in the reaction mechanism,
similar to that which has been reported for gold electrodes.[20]
In addition, this work opens the door to elucidate the effects of
oxide supports and promoters on the activity and selectivity of
supported Cu NPs and Cu thin-films in the conversion of CO2
into high added-value molecules.
NMR samples were prepared by taking 550 µl of the various solutions
resulting from the electrochemistry experiments and adding 50 µl of D2O
for lock purposes. We measured the samplesdata on a Bruker AVANCE
spectrometer console equipped with a 7T superconducting magnet,
operating at a 1H resonance frequency of 400 MHz, fitted with a 5 mm
BBO probe and situated in an air-conditioned room. NMR spectra were
acquired using a 1D 1H pulse sequence with excitation sculpting and
gradients to suppress the water (H2O) peak.[21] This was optimised on
each sample to account for possible small changes in room temperature
as well as sample conditions. A spectral width of 11 ppm with 16K
complex data points was measured and 1500 scans were acquired for
each sample. A delay (d1) of 5 s between scans was used to ensure full
signal recovery. The data were processed using TOPSPIN 3.1 software:
data were zero-filled to 32k to provide a final spectra resolution of 0.13
Hz per point and spectra were indirectly referenced to TMS at 0 ppm.
Further details can be found in the supporting information.
Acknowledgements
MJL acknowledges the University of Birmingham for the financial
support through a PhD scholarship at the School of Chemistry.
PR would also like to acknowledge the University of Birmingham
for financial support through the Birmingham fellowship program.
Keywords: CO2 reduction • pulse voltammetry • copper
electrodes • copper oxides • surface structure • selectivity
This article is protected by copyright. All rights reserved.

10.1002/anie.201706463
Angewandte Chemie International Edition
COMMUNICATION
1]
a) Y. Hori, I. Takahashi, O. Koga, N. Hoshi, J. Phys. Chem. B 2002,
106, 15-17; b) Y. Hori, in Modern Aspects of Electrochemistry, Vol. 42
(Eds.: C. Vayenas, R. White, M. Gamboa-Aldeco), Springer New York,
2008, pp. 89-189; c) A. A. Peterson, F. Abild-Pedersen, F. Studt, J.
Rossmeisl, J. K. Nørskov, Energy Environ. Sci. 2010, 3, 1311; d) J. H.
Montoya, A. A. Peterson, J. K. Nørskov, ChemCatChem 2013, 5, 737-
742; e) R. Kas, R. Kortlever, A. Milbrat, M. T. Koper, G. Mul, J.
Baltrusaitis, Phys. Chem. Chem. Phys. 2014, 16, 12194-12201; f) Y.
Lee, J. Suntivich, K. J. May, E. E. Perry, Y. Shao-Horn, J. Phys. Chem.
Lett. 2012, 3, 399-404.
[9]
Y. S. Chu, I. K. Robinson, A. A. Gewirth, J. Chem. Phys. 1999, 110,
5952-5959.
[10] K. J. P. Schouten, E. P. Gallent, M. T. M. Koper, J. Electroanal. Chem.
2013, 699, 6-9.
[11] a) H. Matsushima, C. Haak, A. Taranovskyy, Y. Grunder, O. M.
Magnussen, Phys. Chem. Chem. Phys. 2010, 12, 13992-13998; b) A. A.
Peterson, J. K. Nørskov, J. Phys. Chem. Lett. 2012, 3, 251-258.
[12] S. Lee, D. Kim, J. Lee, Angew. Chem., Int. Ed. Engl. 2015, 54, 14701-
14705.
[13] R. Kortlever, J. Shen, K. J. P. Schouten, F. Calle-Vallejo, M. T. M.
Koper, J. Phys. Chem. Lett. 2015, 6, 4073-4082.
[2]
a) M. Le, M. Ren, Z. Zhang, P. T. Sprunger, R. L. Kurtz, J. C. Flake, J.
Electrochem. Soc. 2011, 158, E45-E49; b) R. Kortlever, K. H. Tan, Y.
Kwon, M. T. M. Koper, J. Solid State Electrochem. 2013, 17, 1843-
1849; c) R. Kas, R. Kortlever, H. Yılmaz, M. T. M. Koper, G. Mul,
ChemElectroChem 2015, 2, 354-358; d) R. Reske, H. Mistry, F.
Behafarid, B. Roldan Cuenya, P. Strasser, J. Am. Chem. Soc. 2014,
136, 6978-6986; e) K. J. P. Schouten, E. Pérez Gallent, M. T. M. Koper,
J. Electroanal. Chem. 2014, 716, 53-57; f) W. Luo, X. Nie, M. J. Janik,
A. Asthagiri, ACS Catal. 2016, 6, 219-229; g) J. Monzó, Y. Malewski, R.
Kortlever, F. J. Vidal-Iglesias, J. Solla-Gullón, M. T. M. Koper, P.
Rodriguez, J. Mater. Chem. A 2015, 3, 23690-23698.
[14] a) G. C. Chinchen, K. C. Waugh, D. A. Whan, Appl. Catal. 1986, 25,
101-107; b) G. C. Chinchen, P. J. Denny, D. G. Parker, M. S. Spencer,
D. A. Whan, Appl. Catal. 1987, 30, 333-338.
[15] a) M. Bowker, R. J. Madix, Surface Science 1981, 102, 542-565; b) M.
Bowker, K. C. Waugh, Surface Science 2016, 650, 93-102.
[16] M. Favaro, H. Xiao, T. Cheng, W. A. Goddard, 3rd, J. Yano, E. J.
Crumlin, Proc. Natl. Acad. Sci. U. S. A. 2017, 114, 6706-6711.
[17] B. J. Cruickshank, D. D. Sneddon, A. A. Gewirth, Surface Science 1993,
281, L308-L314.
[18] M. R. Vogt, A. Lachenwitzer, O. M. Magnussen, R. J. Behm, Surface
Science 1998, 399, 49-69.
[3]
[4]
F. Calle-Vallejo, M. T. Koper, Angew. Chem., Int. Ed. Engl. 2013, 52,
7282-7285.
[19] a) M. R. Singh, Y. Kwon, Y. Lum, J. W. Ager, A. T. Bell, J. Am. Chem.
Soc. 2016, 138, 13006-13012; b) J. Resasco, L. D. Chen, E. Clark, C.
Tsai, C. Hahn, T. F. Jaramillo, K. Chan, A. T. Bell, J. Am. Chem. Soc.
2017, 139, 11277-11287; c) M. Liu, Y. Pang, B. Zhang, P. De Luna, O.
Voznyy, J. Xu, X. Zheng, C. T. Dinh, F. Fan, C. Cao, F. P. G. de Arquer,
T. S. Safaei, A. Mepham, A. Klinkova, E. Kumacheva, T. Filleter, D.
Sinton, S. O. Kelley, E. H. Sargent, Nature 2016, 537, 382-386; d) C.
Stoffelsma, P. Rodriguez, G. Garcia, N. Garcia-Araez, D. Strmcnik, N.
M. Marković, M. T. M. Koper, J. Am. Chem. Soc. 2010, 132, 16127-
16133.
a) Y. G. Kim, J. H. Baricuatro, A. Javier, J. M. Gregoire, M. P. Soriaga,
Langmuir 2014, 30, 15053-15056; b) Y.-G. Kim, A. Javier, J. H.
Baricuatro, M. P. Soriaga, Electrocatalysis 2016, 7, 391-399; c) Y.-G.
Kim, A. Javier, J. H. Baricuatro, D. Torelli, K. D. Cummins, C. F. Tsang,
J. C. Hemminger, M. P. Soriaga, J. Electroanal. Chem. 2016, 780, 290-
295.
[5]
a) A. D. Handoko, C. W. Ong, Y. Huang, Z. G. Lee, L. Lin, G. B. Panetti,
B. S. Yeo, J. Phys. Chem. C 2016, 120, 20058-20067; b) C. W. Li, M.
W. Kanan, J. Am. Chem. Soc. 2012, 134, 7231-7234; c) L. I. Bendavid,
E. A. Carter, J. Phys. Chem. C 2013, 117, 26048-26059.
[20] a) P. Rodriguez, N. Garcia-Araez, M. T. M. Koper, Phys. Chem. Chem.
Phys. 2010, 12, 9373-9380; b) P. Rodriguez, N. Garcia-Araez, A.
Koverga, S. Frank, M. T. M. Koper, Langmuir 2010, 26, 12425-12432;
c) P. Rodriguez, Y. Kwon, M. T. M. Koper, Nat. Chem. 2012, 4, 177-
182; d) P. Rodríguez, A. A. Koverga, M. T. M. Koper, Angew. Chem.,
Int. Ed. 2010, 49, 1241-1243.
[6]
[7]
K. W. Frese, J. Electrochem. Soc. 1991, 138, 3338.
V. Grozovski, V. Climent, E. Herrero, J. M. Feliu, Phys. Chem. Chem.
Phys. 2010, 12, 8822-8831.
[8]
a) D. Friebel, P. Broekmann, K. Wandelt, Phys. Status Solidi A 2004,
201, 861-869; b) C. Schlaup, S. Horch, Surface Science 2013, 608, 44-
54.
[21] T. L. Hwang, A. J. Shaka, J. Magn. Reson., Ser. A 1995, 112, 275-279.
This article is protected by copyright. All rights reserved.

10.1002/anie.201706463
Angewandte Chemie International Edition
COMMUNICATION
Layout 1:
COMMUNICATION
Pulse voltammetry and copper
single crystal electrodes were
implemented to determine the
influence of oxygen species
adsorbed on copper electrodes in
the electrochemical reduction of
CO2. Product selectivity towards the
formation of oxygenated
hydrocarbon was associated to the
coverage of oxygen species, which
is surface-structure-and-potential-
dependent.
Cécile S. Le Duff, Matthew J.
Lawrence, Paramaconi Rodriguez*
Page No. – Page No.
Role of the adsorbed oxygen
species in the selective
electrochemical reduction of CO2
on copper electrodes
This article is protected by copyright. All rights reserved.