Reactivity Determinants in Electrodeposited Cu Foams for Electrochemical CO2 Reduction

Article abstract of DOI:10.1002/cssc.201801582

CO₂ reduction is of significant interest for the production of nonfossil fuels. The reactivity of eight Cu foams with substantially different morphologies was comprehensively investigated by analysis of the product spectrum and in situ electrochemical spectroscopies (X-ray absorption near edge structure, extended X-ray absorption fine structure, X-ray photoelectron spectroscopy, and Raman spectroscopy). The approach provided new insight into the reactivity determinants: The morphology, stable Cu oxide phases, and *CO poisoning of the H₂ formation reaction are not decisive; the electrochemically active surface area influences the reactivity trends; macroscopic diffusion limits the proton supply, resulting in pronounced alkalization at the CuCat surfaces (operando Raman spectroscopy). H₂ and CH₄ formation was suppressed by macroscopic buffer alkalization, whereas CO and C₂H₄ formation still proceeded through a largely pH-independent mechanism. C₂H₄ was formed from two CO precursor species, namely adsorbed *CO and dissolved CO present in the foam cavities.

Full text of DOI:10.1002/cssc.201801582

DOI: 10.1002/cssc.201801582
Full Papers
Reactivity Determinants in Electrodeposited Cu Foams for
Electrochemical CO2 Reduction
Katharina Klingan+,[a] Tintula Kottakkat+,[b] Zarko P. Jovanov,[c] Shan Jiang,[a] Chiara Pasquini,[a]
Fabian Scholten,[d] Paul Kubella,[a] Arno Bergmann,[c, e] Beatriz Roldan Cuenya,[d, e]
Christina Roth,[b] and Holger Dau*[a]
CO2 reduction is of significant interest for the production of
nonfossil fuels. The reactivity of eight Cu foams with substan-
tially different morphologies was comprehensively investigated
by analysis of the product spectrum and in situ electrochemical
spectroscopies (X-ray absorption near edge structure, extended
X-ray absorption fine structure, X-ray photoelectron spectros-
copy, and Raman spectroscopy). The approach provided new
insight into the reactivity determinants: The morphology,
stable Cu oxide phases, and *CO poisoning of the H2 formation
reaction are not decisive; the electrochemically active surface
area influences the reactivity trends; macroscopic diffusion
limits the proton supply, resulting in pronounced alkalization
at the CuCat surfaces (operando Raman spectroscopy). H2 and
CH4 formation was suppressed by macroscopic buffer alkaliza-
tion, whereas CO and C2H4 formation still proceeded through
a largely pH-independent mechanism. C2H4 was formed from
two CO precursor species, namely adsorbed *CO and dissolved
CO present in the foam cavities.
Introduction
Selective conversion of CO2 into high-quality chemical feed-
stock and fuels represents an emerging energy storage tech-
nology.[1] The electrochemical reduction of carbon dioxide is a
promising way to utilize intermittently available renewable en-
ergies to produce chemical feedstock and fuels for later use.
Copper has been a focus of recent research because of its abil-
ity to catalyze the electrochemical reduction of CO2 into
energy-dense hydrocarbons.[2] However, the electrochemical re-
duction of CO2 on copper results in the formation of many dif-
ferent compounds at a wide range of overpotentials. In addi-
tion to undesired H2, CO, C1 and C2 hydrocarbons and alco-
hols are formed in the gas headspace and in the liquid electro-
lyte. In recent years, there has been increasing effort to estab-
lish control over the product distribution by fine-tuning the
physical properties of the Cu electrode.[3] Different strategies
have been proposed for improving the activity and selectivi-
ty,[4] including variation of the thickness of the Cu metal over-
layers,[5] adjustment of the size,[6] shape,[7] and interparticle dis-
tance of the nanoparticles,[6b] selection of single crystal fac-
ets,[6c,8] utilization of electrolyte properties,[9] and grain boun-
dary strain effects.[10] The influence of the roughness or specific
nanostructures of the Cu electrode surface is of high impor-
tance.[11] Cu electrodes with high surface areas mainly produce
C2 products such as C2H4 and C2H6.[3b,12] However, high-surface-
area gas diffusion electrodes favor the C1 pathway.[13] The CO
yield at low potentials may or may not be related to the C2H4
yield at higher potentials.[6c,7c,11,14] In any event, adsorbed CO,
denoted as *CO, is a crucial reaction intermediate in the pro-
duction of CO, CH4, and C2H4.[15] Increasing the relative amount
of *CO sites favors increased production of C2H4.[10d,14c,16] The
formation of *CO intermediates may be favored by local alkali-
zation and mass transport limitations of protons.[9d,17] Local pH
changes probably lead to a change in reaction kinetics as the
[a] Dr. K. Klingan,+ S. Jiang, C. Pasquini, P. Kubella, Prof. Dr. H. Dau
Department of Physics
Freie Universitꢀt Berlin
Arnimallee 14, 14195 Berlin (Germany)
E-mail: holger.dau@fu-berlin.de
[b] Dr. T. Kottakkat,+ Prof. Dr. C. Roth
Department of Chemistry
Freie Universitꢀt Berlin
Takustr. 3, 14195 Berlin (Germany)
[c] Dr. Z. P. Jovanov, Dr. A. Bergmann
Department of Chemistry
Technische Universitꢀt Berlin
Straße des 17. Juni, 10623 Berlin (Germany)
[d] F. Scholten, Prof. Dr. B. Roldan Cuenya
Department of Physics
ꢀ
CO32ꢀ/HCO3 equilibrium is shifted, and CO2 may be depleted
Ruhr University Bochum
Universitꢀtstraße 150, 44801 Bochum (Germany)
from the electrode surface.[17b,18] The influence of surface area
and surface roughness on the product spectrum may be relat-
ed to the strength of local alkalization.[17b,19]
[e] Dr. A. Bergmann, Prof. Dr. B. Roldan Cuenya
Department of Interface Science
Fritz Haber Institute of the Max Planck Society
Faradayweg 4–6, 14195 Berlin (Germany)
[+] These authors contributed equally to this work.
The incorporation of Cu2O into the bulk or the surface of
the Cu electrodes can strongly enhance the selectivity towards
C2H4 and C2H6.[3b,7c,11] Ex situ XRD measurements resulted in
contradicting conclusions regarding the presence and stability
of Cu2O species in Cu foams, including no Cu2O,[20] the pres-
Supporting Information and the ORCID identification number(s) for the
author(s) of this article can be found under:
https://doi.org/10.1002/cssc.201801582.
ChemSusChem 2018, 11, 1 – 12
1
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
ence of Cu2O after deposition,[21] Cu2O reduction to Cu0 during
operation at catalytic potentials,[21b] and the presence of Cu2O
after operation at reductive potentials.[22] Huan et al. quantified
a 10% Cu2O mass contribution in their foams and 5% of this
phase was unaffected by cathodic electroreduction.[22a] In situ
deposition of a related dendritic Cu/Cu2O catalyst showed a
stable cuprite phase after electrolysis at negative potentials
during the production of ethylene.[22b] Oxygen species may be
formed from the exposure to air, as recently shown for oxide-
derived Cu catalysts,[23] so that reduction of copper oxide spe-
cies precedes the onset of CO2 reduction.[24]
To shine light on the role of oxide species, we studied Cu
foams by operando X-ray spectroscopy (XAS) and quasi in situ
X-ray photoelectron spectroscopy (XPS). The porous Cu films
(termed CuCats), were electrodeposited using the dynamic hy-
drogen-bubble template method.[21a] For the first time, local
pH values at the Cu cathode during CO2 electroreduction were
experimentally investigated through operando Raman spec-
2ꢀ
troscopy based on analysis of the shift in the HCO3ꢀ/CO3
equilibrium. For eight foams with substantially different mor-
phologies, we show how the area ratio of the inner-foam sur-
face (Aint) and macroscopic outer surface (Aout) determines the
partial current densities for the formation of H2, CH4, CO, and
C2H4 in distinctively different ways.
Figure 1. SEM of CuCats. Top-view of CuCats electrodeposited at
(a) 0.5 Acmꢀ2 for 10 s on Pt, (b) at 0.5 Acmꢀ2 for 10 s on Cu, (e) at 1.0 Acmꢀ2
for 20 s on Pt, and (f) at 1.0 Acmꢀ2 for 20 s on Cu. Cross-sections of CuCats
electrodeposited at (c) 0.5 Acmꢀ2 for 10 s on Pt, (d) at 0.5 Acmꢀ2 for 10 s on
Cu, (g) at 1.0 Acmꢀ2 for 20 s on Pt, and (h) at 1.0 Acmꢀ2 for 20 s on Cu.
Under the catalyst layer, the metal substrate and silicon glue are visible
above the copper tape used to fix the sample onto the sample holder. The
insets show a magnified region of the cross section. Complete cross-sections
of all the investigated CuCats are shown in Figure S1; complete SEM topog-
raphies are shown in Figures S2 and S3.
Results and Discussion
Physical properties
The material properties of the CuCats was tuned by the elec-
trodeposition process. Films of different thickness (Figure S1,
Supporting Information) with different micromorphological
structure (Figures S2 and S3) were realized by variation of the
deposition current, time of deposition, and substrates (Cu or
Pt). Figure 1 shows the top-view and the cross-sectional scan-
ning electron micrography (SEM) images of selected CuCats
electrodeposited on a Pt or Cu substrate either at low current
densities, short time (0.5 Acmꢀ2 and 10 s), or at double the cur-
rent density and time (1.0 Acmꢀ2 and 20 s). The thickness of
the catalyst film on Pt was significantly higher (80–360 mm)
than on Cu (40–90 mm). The influence of the substrate was re-
flected in pronounced differences in the morphology of the
CuCats on Pt metal (Figure S2) and Cu metal (Figure S3). Aside
from internal porosity and dendritic structures, there were ver-
tical macropores limited by dendritic structures (with pore di-
ameters of ꢁ35 mm), which formed a network of craters on
the catalyst surface (Figure 1 f). The significantly different mor-
phologies were related to the specific parameters of the used
“bubble-template” approach.[25] The surface pore sizes in-
creased with deposition time and higher deposition current;
larger pores and more pronounced fractal structures were cre-
ated on the Cu metal. Pore formation depends on a complex
process involving nucleation, growth, and detachment of H2
gas bubbles during Cu electrodeposition.[25] Small H2 bubbles
are quickly released on Pt, whereas H2 desorption is slower on
Cu, resulting in larger pores and deposition of less material.
A linear relation between the amount of Cu and the film
thickness (Figures 2, S4, and S5) indicated that the overall Cu
density of the investigated porous CuCats was approximately
the same (1=0.20ꢂ0.03 gcmꢀ3 versus 8.94 gcmꢀ3 for pore-
free metallic Cu metal foil). Therefore, the CuCats were low-
density foams (ꢁ45 times lower than Cu metal). Surprisingly,
the macroscopic foam density was roughly constant, regard-
less of the pronounced morphology differences at the nano-
meter and micrometer scale.
The electrochemically active surface areas (ECSAs) were de-
termined by cyclic voltammetry (CV) measurements at different
scan rates in the non-Faradaic current region (Table S1, Fig-
ures S6 and S7). For both substrate materials, the ECSA de-
pended, to a first approximation, linearly on the electrodepos-
ited Cu amount, but the electrochemical surface area per de-
posited Cu atom was 6 times higher for the films deposited on
Cu foil (1.90 cm2 mmolꢀ1 for the Cu substrate, 0.31 cm2 mmolꢀ1
for the Pt substrate). This difference was in qualitative agree-
ment with the SEM results shown in Figure 1, which also sug-
gested a larger surface area for CuCat deposition on a Cu foil.
&
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
2
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 2. (a) Relationship between the thickness (d in mm) of the CuCat films
and the electrodeposited amount of metal ions (nCu in mmol per geometrical
surface area in cm2). (b) Relationship between the electrochemically active
surface area (ECSA in cm2) of various CuCat films and the electrodeposited
amount of metal ions (nCu in mmol per geometrical surface area in cm2).
CuCats electrodeposited on a Cu substrate are shown in blue, and on the Pt
substrate in red. The thickness of the samples was estimated from the SEM
cross-sections (Figure S1). The amount of deposited Cu material was deter-
mined from TXRF (total reflection X-ray fluorescence) elemental analysis (Fig-
ure S4). The ECSA was calculated as ECSA=CDL/CCu metal. The double layer ca-
pacitance (CDL) was determined by scanning multiple CVs at different scan
rates in a non-Faradaic current range (Figures S6 and S7).
Figure 3. CVs in 0.1m CO2-saturated KHCO3 (pH 6.8, CO2 flow rate=20
mLminꢀ1, scan rate=100 mVsꢀ1): CuCats deposited at 0.5 Acmꢀ2 for 10 s
(cyan), at 0.5 Acmꢀ2 for 20 s (red), at 1.0 Acmꢀ2 for 10 s (green), at
1.0 Acmꢀ2 for 20 s (brown), Pt foil (black), and Cu foil (violet), on (a) a Pt sub-
strate or (b) a Cu substrate. The insets show the enlarged region of the re-
duction waves originating from *CO adsorption. ECSA normalized current
density during 1 h of chronoamperometric measurements as a function of
applied potential versus RHE for Cu foil (violet) and for CuCats deposited on
(c) Pt and (d) Cu substrate at 0.5 Acmꢀ2 for 10 s (cyan), at 0.5 Acmꢀ2 for 20 s
(red), at 1.0 Acmꢀ2 for 10 s (green), at 1.0 Acmꢀ2 for 20 s (brown). Tafel
slope for Cu foil (violet) and averaged Tafel slopes with the corresponding
standard deviations for CuCats/Pt (red) and CuCats/Cu (blue).
Total current densities and *CO current wave
Total current density
The electrochemical characterization of the catalysts was per-
formed in 0.1m KHCO3 solution saturated with CO2 at pH 6.8
in a custom-built H-cell (Figure S9). To ensure that the initially
present but unstable Cu oxide phases were reduced to metal-
lic Cu (see 1st cycle of CV in Figure S10), we swept the potential
slope may reflect a diffusion limitation that is common to all
investigated catalysts (even the pure metal foil).
between 0 and ꢀ1.2 VRHE with high scan rates (100 mVsꢀ1
)
*CO current wave
until reproducible currents were obtained (CVs in Figure 3),
then performed slow linear sweep voltammetry (LSV; 5 mVsꢀ1
;
A current plateau was observed for all the CuCats at approxi-
mately ꢀ0.57 VRHE in the LSVs (Figure S11, Figure S12) and from
approximately ꢀ0.65 to ꢀ0.70 VRHE in the CVs (Figure 3a–b).
This current wave was attributed mostly to CO2 reduction cou-
pled to the formation of adsorbed *CO,[14c,15a,26] but also to the
Figure S11). Very similar qualitative characteristics were ob-
served for LSV and CV: The onsets of the cathodic currents
were significantly shifted to more positive potentials for
CuCats of varying deposition currents and synthesis times
compared with those of metallic Cu foil. The overpotential nec-
essary to reach ꢀ5 mAcmꢀ2 was lowered to 350 mV by CuCats
in contrast to pure Cu. All CuCats reached significantly larger
current densities (up to 5 times) than the Cu metal foil. CuCats
had ꢁ8–65 times higher surface area than Cu metal foil
(Figure 2), which was most likely related to the larger current
densities. However, the ECSA-normalized current densities of
the CuCats were approximately one order of magnitude lower
than those of pure Cu foil (Figure 3c and d). Surprisingly, the
Tafel slopes (Figure 3c and d) were similar for all catalyst mate-
rials (ꢁ300 mV/decade). This unusually high value for the Tafel
ꢀ
reduction of HCO3 to HCOOꢀ.[27] Because the current-wave
magnitude did not correlate with jHCOO- (Figure S13), but was
proportional to jCO (as shown later in Figure 8), we favored the
assignment to adsorbed *CO. For a polycrystalline Cu metal
foil with a low amount of defect sites,[28] the *CO current wave
could not be detected. It was also not visible in the CVs col-
lected for the Cu metal foil, but clearly visible in the CVs of all
the Cu foams (Figure 3a, 3b, S14; see Figure S15 for the corre-
sponding redox charges obtained by integration of the current
waves). Extension of the potential window to more positive
potentials did not affect the size of this current plateau (Fig-
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
3
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
ure S16a). However, when the CV scan was started at more
negative potentials, the current plateau decreased (Fig-
ure S16b and c), suggesting that unbinding of *CO proceeded
Calculating the linear combinations of Cu0 metal foil and CuI2O
(with the weighting factors as fit parameters) leads to identical
results for XANES and EXAFS (extended X-ray absorption fine
structure) spectra, namely the presence of 15–45% Cu2O (Fig-
ure S20).
only at potentials more positive than approximately ꢀ0.4 VRHE
,
in line with recent findings.[29] Variation of the scan rate sug-
gested the involvement of a diffusion limitation in *CO forma-
tion (Figure S17).
The reason for the presence of the CuI2O detected by cryo-
XAS on dried CuCat foams could be either (i) the presence of
stable CuI2O,[30] or (ii) reoxidation originating from air expo-
sure[23] before freezing and the execution of the cryo-XAS ex-
periment. To investigate the role of air oxidation, we studied a
CuCat (Cu 0.5A-20s) by operando (or in situ) XAS. In Figure 4a–
c, the EXAFS and XANES of the films directly after electrodepo-
sition, and at ꢀ0.2 and ꢀ0.95 VRHE (applied for 30 min each
before collecting the data) are compared with reference sub-
stances (Cu metal, Cu2O, CuO, and [Cu(OH2)6]SO4). The as-de-
posited film contained CuꢀO distances of 1.92 ꢁ, which were
assigned to Cu2+ (Table S2), and three CuꢀCu distances as-
signed to metallic Cu (2.51, 3.58, 4.41 ꢁ). By applying a poten-
tial of either ꢀ0.2 or ꢀ0.95 VRHE for 30 min, the film was re-
duced to a metallic copper of comparably low order, similar to
observations reported for Cu nanocubes supported on
carbon.[31] To assess how fast these reductive changes occur at
the negative operating CO2 reduction potential (ꢀ0.95 VRHE),
we immersed the as-deposited CuCat into the CO2-reduction
buffer (0.1m KHCO3, pH 6.8) for various time periods (Fig-
Oxide content of Cu foams
To gain insight into the structure of the catalysts, ex situ X-ray
diffraction (XRD) patterns of different CuCats after a cycling
protocol (three CVs and one LSV) were recorded (Figure S18).
The main phase of the catalysts was identified as metallic Cu
with major contributions from crystallites with a diameter of
ꢁ100 nm. A minor and separated phase was identified as
Cu2O stemming from significantly smaller particles (<10 nm).
CuO crystallites were not detected.
For further structural information—specifically for XRD-invisi-
ble, noncrystalline CuCat fractions—Cu K-edge X-ray absorp-
tion spectroscopy (XAS) was performed on dry (electrolyte-
free) CuCat foams at 20 K (frozen after a cycling protocol con-
sisting of three CVs and one LSV). X-ray absorption near edge
structure (XANES) spectra of the CuCats resembled a mixture
of Cu metal contribution and Cu2O contribution (Figure S20).
Figure 4. Analysis of the Cu oxide content by operando X-ray absorption spectroscopy at the Cu K-edge. a) Fourier-transformed k3-weighted EXAFS spectra.
Solid lines represent the experimental data and shading represents the EXAFS simulation. The indicated reduced distances are approximately 0.4 ꢁ shorter
than the real distances determined by EXAFS simulations (Table S2). b) k3-weighted EXAFS spectra and simulations (colored lines) before Fourier transforma-
tion. c) XANES spectra at the Cu K-edge. Panels a, b, and c contain data for Cu0 (violet), CuCat at ꢀ0.95 VRHE (red), at ꢀ0.2 VRHE (grey), as deposited (orange),
Cu2O (green), [Cu(OH2)6]SO4 (light blue), the Cu2+ ion in aqueous solution (blue). d) XANES at the Cu K-edge: CuCat in 0.1m KHCO3 (pH 6.8) at open circuit po-
tential (OCP), and at ꢀ0.95 VRHE for the indicated time values, which indicate when the XANES scan was started (each scan was collected within 3 min).
&
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
4
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
ure 4d). Under open-circuit potential (OCP) conditions, a major
fraction of the Cu oxide becomes reduced. The XANES spec-
trum collected after 9 min at ꢀ0.95 VRHE indicated essentially
complete reduction to metallic copper; no further spectral
readily reduced to metallic copper. The presence of a minor
fraction of subsurface oxides that was below the detection
limit of our experimental equipment was not completely ex-
cluded. Complete reduction of the initially present Cu oxides
likely occurred at ꢀ0.2 VRHE (Figure S10). The transformation of
the Cu oxides into metallic copper resulted in a disordered
metallic phase, possibly with a metallic surface that offers a va-
riety of noncrystalline arrangements of copper atoms.
changes were observed for the following 18 min at ꢀ0.95 VRHE
.
Deconvolution of the final reduced state at ꢀ0.95 VRHE yielded
a 0–5wt% proportion of Cu2O (Figure S21). This indicated that
the complete absence of any Cu2O fraction was compatible
with the spectra. Comparison with the spectrum collected at
ꢀ0.2 VRHE revealed that at low negative potentials (ꢀ0.2 VRHE
)
the CuCat was already in its most reduced state, which could
be the oxide-free state.
Evidence for pronounced local alkalization
To demonstrate that CuCats experience diffusion-limited
proton supply and thus local alkalization, we investigated the
Additionally, the surface of a CuCat (0.5A-20s-Cu) was stud-
ied by quasi in situ X-ray photoelectron spectroscopy (XPS;
Figure 5). Details about the XPS measurements and evaluation
have been described elsewhere.[14b] The Auger spectra were
fitted with the line shape of CuI2O, CuIIO, and metallic Cu refer-
ence spectra (Figure 5b). Without any further electrochemical
treatment, the as-prepared CuCat contained 17% Cu0, 66%
CuI2O, and 27% CuIIO. After performing a cycling protocol, con-
sisting of three CVs (Figure S10, 3) and one LSV (Figure S11),
no more oxides were detected at the surface. XPS after catalyt-
ic operation for 20 min showed that the CuCat surface re-
mained metallic. Exposure of the as-deposited film to air seem-
ingly caused strong oxidation of the surface of the CuCat (mix-
ture of 66% CuI2O and 27% CuIIO).
2ꢀ
HCO3ꢀ/CO3 equilibrium through operando Raman spectros-
2ꢀ
copy. With increasing pH, the HCO3ꢀ/CO3 equilibrium gener-
ally shifted towards CO32ꢀ, resulting in a decrease in the 1015
and 1363 cmꢀ1 Raman bands (HCO3ꢀ) and an increase in the
1065 cmꢀ1 band (CO32ꢀ) (Figures 6a, S22–25). Figure 6 shows
the operando Raman spectra of a selected CuCat (0.5A-20s-Cu)
at OCP and comparably lower overpotentials (ꢀ0.6 and
ꢀ0.7 VRHE) in CO2-saturated 0.1m KHCO3 buffer.
In summary, the oxide species present in the as-deposited
CuCat film were not stable at negative potentials but were
Figure 5. Quasi in situ XPS measurements after ultrahigh vacuum transfer of
CuCat deposited at 0.5 Acmꢀ2 for 20 s on Cu substrate, obtained directly
after electrodeposition (as deposited), after performing three CVs and one
LSV (after cycling; Figure S10 and S11), and after 20 min of operation at
ꢀ0.95 VRHE in 0.1m KHCO3, saturated with CO2, pH 6.8. a) Binding energy of
the Cu2p region. b) Kinetic energy of the Cu LMM (LMM is the spectroscopic
notation of the atomic levels) Auger region.
Figure 6. (a) Reference Raman spectra of 0.1m KHCO3 at different pH values
(6.8, 8.8, 9.0, 9.5, 10.0, 10.5, 11.0, 11.5, 12.0, and 12.5). The pH was adjusted
by the addition of KOH. Operando Raman spectra of CuCat-0.5A-20s on Cu
at (b) OCP, (c) ꢀ0.6 VRHE, and (d) ꢀ0.7 VRHE in 0.1m KHCO3 (bulk pH 6.8) direct-
ly on the surface (0 mm; black line) and 200 mm away from the surface (blue
line).
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
5
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
ꢀ
At the OCP, mostly HCO3 was detected on the CuCat sur-
2ꢀ
face, as expected (Figure S24). However, a CO3 band was ob-
served at ꢀ0.6 VRHE directly on the surface. At ꢀ0.7 VRHE, the
1065 cmꢀ1 band was strongly enhanced compared with the
1015 cmꢀ1 band. Comparison to the reference spectra (Fig-
ure S24) showed that the local pH at ꢀ0.6 VRHE was 9.0 and
dropped to 7.8 at a 200 mm distance from the electrode. At
ꢀ0.7 VRHE the local pH was higher than 10, and alkalization
could still be detected at a distance of 200 mm from the sur-
face, at which the pH exceeded 9.5. Figure S25 shows that
local alkalization at low overpotentials was strongly reduced
on the Cu foil compared with the surface of Cu foam (at
ꢀ0.6 VRHE approximately pH 7.8, at ꢀ0.7 VRHE smaller than 9.5).
At low overpotentials, the CuCat was exposed to a drastic
local pH increase, which was likely related to the diffusion limi-
ꢀ
tations of HCO3 ions that act as proton carriers (in analogy to
ꢀ
the role of H2PO4 in water oxidation at neutral pH, see
Ref. [32]). The “local pH” is a macroscopic property because al-
kalization can be detected not only directly at the catalyst sur-
face but also at a distance of 200 mm. We think that the rela-
tion (ratio) between the proton-consuming catalytic processes
and macroscopic proton movements is decisive for the extent
of alkalization.
Correlation of ECSA and partial current densities
The major products detected from the electroreduction of CO2
on CuCats were CO, CH4, C2H4, HCOOꢀ, and the minor prod-
ucts were C2H5OH, C3H7OH, C2H6, and C3H6O (Figures S26–30),
with a stable product yield over the course of 8 h (Figure S31).
Clear differences were observed between the CO2-reduction
product spectrum of the various electrodeposited Cu foams
and Cu foil. The formation of minor products was generally en-
hanced on Cu foams compared with that on Cu foil, but no
distinct trends could be easily deduced for the minor products
from the complex data shown in Figures S26–30. However, in-
teresting (and often surprising) quantitative relationships were
observed between the ECSA and the partial current densities
for H2 (at ꢀ0.6 VRHE), CH4 (at ꢀ1.05 VRHE), CO (at ꢀ0.6 VRHE), and
C2H4 (at ꢀ0.85 VRHE). In the following, we focus on these infor-
mative relationships, which are summarized in Figure 7 (and
Figure S32 after normalization to ECSA).
Figure 7. The dependence of the geometrical partial current densities of
CuCats and Cu foil on the ECSA (in cm2). Partial current densities were deter-
mined at (a) ꢀ0.60 VRHE for H2 (black), (b) at ꢀ1.05 VRHE for CH4 (green), (c) at
ꢀ0.60 VRHE for CO (red), and (d) at ꢀ0.85 (blue), ꢀ0.95 (dark cyan), and at
ꢀ1.05 VRHE (cyan) for C2H4 in CO2-saturated 0.1m KHCO3 (pH 6.8). ECSA was
calculated from the experimental CDL (mF) (Figure S6, S7). The y-error bars
correspond to the average of three individual measurements, and x-error
bars indicate the accuracy of the determined ECSA. The data point in paren-
thesis is at the C2H4 detection limit; the real current density could be lower.
The slope values of the linear regressions were: (c) jCO 63.4ꢂ0.6 cm2 decꢀ1
and (d) jC 35.7ꢂ0.2 cm2 decꢀ1
.
2 H4
ceeding 35 mm, as observed in the 1.0A-20s-Cu foam, the effec-
tive macroscopic surface area may be larger than for the elec-
trode surface.
The outer macroscopic surface areas were approximately the
same for the Cu metal foils and all the porous Cu foams,
whereas the inner surfaces of the Cu foams were up to 65-fold
larger (as indicated by the ECSA determined in a noncatalytic
potential range). It has been suggested that rougher surfaces
suppress CH4 formation owing to a rise in the local pH, be-
cause CH4 formation involves rate-determining, pH-dependent
CO protonation steps.[2b,11,12b] Based on our Raman results,
which indicated pronounced alkalization, in line with earlier re-
ports,[9a,11,33] we attributed the suppression of CH4 production
to locally reduced proton concentrations resulting from macro-
scopic diffusion limitations in the supply of protons to the
CuCat electrode. Only the CuCat with the highest ECSA
Methane formation is independent of internal surface area
In contrast to the total current density and other partial cur-
rent densities, jCH was not enhanced by the increased ECSA of
4
the CuCat foams (Figure 7b). This surprising behavior was ex-
plained by the rate of CH4 synthesis, which was determined by
the area of the macroscopic outer surface of the catalyst mate-
rial (Aout), but largely insensitive to the area of the nanoscopic
inner surfaces of the CuCat foam (Aint). The outer, macroscopic
surface areas correspond to the macroscopic electrode area
covered by the CuCat foam. This area determines the macro-
scopic diffusion fluxes of educts and products from the bulk
electrolyte to the catalyst foam and vice versa. For the forma-
tion of the pronounced crater structures with diameters ex-
(65 cm2), had a higher jCH than the pure Cu foil. This could be
4
related to the characteristic dimensions of the sufficiently large
surface structures that enhance the macroscopic outer surface,
which is relevant for CH4 formation.
&
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
6
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
H2 formation is independent of internal surface area
reactions between (i) two adsorbed *CO species or (ii) between
one adsorbed *CO and dissolved carbon species. The solid
lines in Figure 7 represent a drastic approximation, but still
hint towards a quantitative trend. Their slope differs by a
factor of approximately 2, which is in line with a quadratic de-
pendence of jC H on the rate of CO production or number of
As shown in Figure 7a, at ꢀ0.6 VRHE, the H2-evolution partial
current was the same for all CuCat foams, irrespective of their
ECSA. This indicated that the H2 evolution current densities per
ECSA were strongly suppressed in the foam with a high surface
area compared with those with a low surface area (Figure S32).
The CH4 and H2 partial currents displayed very similar behavior,
which suggested a similar mechanistic explanation: A pro-
nounced local alkalization results from macroscopic diffusion
limitations in the supply of protons to the CuCat electrode,
which limits the H2-formation current density and suppresses
the relative yield of H2-formation when compared with CO for-
mation. An alternative explanation might be poisoning of the
inner CuCat surfaces by adsorbed *CO, which is discussed fur-
ther below (and considered to be less likely).
2
4
adsorbed *CO molecules, thus supporting the formation of
one C2H4 molecule from two precursor species in a bimolecular
reaction.
Role of *CO coverage
Interestingly, the CO partial current density was directly pro-
portional to the magnitude of the irreversible redox wave,
which was probably owing to the formation of adsorbed *CO,
as shown in Figure 8. The CuCat with the highest ECSA was an
CO formation is dependent on the internal surface area
In the CuCat foams, jCO significantly increased with increasing
ECSA (Figure 7c), in strong contrast to jH2 (Figure 7a) and jCH
4
(Figure 7b). The regression line in Figure 7c corresponded to
an exponential increase but did not provide strong evidence
for a truly exponential behavior. An alternative presentation of
the area-normalized partial current densities suggested that
the jCO was directly proportional to ECSA (roughly constant
value of jCO/ECSA in Figure S32). A proportional increase of the
partial current density and ECSA may be viewed as “normal”
behavior. Accordingly, the partial current densities of jH2 (Fig-
ure 7a) and jCH (Figure 7b) do not behave normally. After nor-
Figure 8. Geometric partial current density of CO in mAcmꢀ2 versus qred (or
*CO adsorption charge) in mCcmꢀ2. The charge was obtained by integrating
the plateau observed in the 2nd cycle of the CVs (Figure 3) of the CuCats.
Linear regression (red line) excludes 1.0A-20s-Cu with an ECSA of 65 cm2.
The x- and y-error bars correspond to the average of three individual
measurements.
4
malization to ECSA, they were strongly suppressed with in-
creasing ECSA (Figure S32). Again, diffusion limitations of pro-
tons may provide a possible explanation: Depletion of proton-
transporting carbonate ions results in pronounced alkaliza-
tion[17b] at the active sites of CO formation within the CuCat
foams, which inhibits CH4 formation and H2 formation at least
at low overpotentials, whereas CO formation proceeds through
a largely pH-insensitive pathway.[9a,d,19,34]
outlier, which could be owing to its large macropores and in-
crease in the effective outer surface compared with that of the
inner surfaces that dominate the ECSA. This clear correlation
can be simply explained as follows: The chemisorbed *CO
equilibrates with gaseous CO in solution, resulting in direct
proportionality of *CO coverage and CO release. The redox-
wave in the cyclic voltammogram becomes visible because
mostly complete *CO coverage of the CO2-reduction sites in
the foam interior is reached at approximately ꢀ0.7 VRHE, result-
ing in a drop of the net partial current density for CO forma-
tion at more negative potentials. The outlined events can also
explain why the rate of CO formation does not significantly in-
crease at higher overpotentials in most of the investigated
copper foams. A comparison of the *CO reduction charge to
the redox-active surface area revealed that a major fraction of
the surface sites could be occupied by chemisorbed *CO mole-
cules, as suggested by the following order-of-magnitude esti-
mate for the surface area per absorbed CO molecule ([Eq. 1];
Foam-enhanced C2H4 formation from two CO precursor
species
In the CuCat foams, jC H at ꢀ0.85 VRHE significantly increased
2
4
with increasing ECSA (Figure 7d). Considering that such en-
hancement was not observed at higher overpotentials (ꢀ0.95,
ꢀ1.05 VRHE; cyan and blue lines in Figure 7d), this correspond-
ed to a lowering of the onset potential for C2H4 formation
with increasing surface area in the foam interior. At ꢀ0.85 VRHE
,
the C2H4 partial current of the pure metal foil was below the
detection limit so that the cavities of the foam were conceived
as an essential prerequisite for low-potential C2H4 formation.
Interestingly, jC H (ꢀ0.85 VRHE) clearly increased more strongly
2
4
than jCO (ꢀ0.6 VRHE) with increasing ECSA. Although for jCO the
partial current densities per ECSA were constant, the “normal”
behavior, there was a strong increase in jC H per ECSA for the
2
4
low-surface-area foams compared with the high-surface-area
foams (Figure S32). This behavior could result from bimolecular
A*CO for CuCat/1A-10s-Cu; e: elementary charge):
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
7
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
filled by metallic copper, resulting in a total pore volume close
to 98%. Despite the significant variations in morphology, the
foam density remained approximately constant. In contrast,
the ECSA, which is a measure of the ratio between the foam-
internal surfaces (Ain) and the outer macroscopic surface (Aout),
depends on the film morphology and is strongly enhanced by
the electrodeposition on a Cu substrate, which can be ex-
plained by a predominantly dendritic structure of the internal
foam surfaces. The variation of the foam morphology promot-
ed the identification of properties that determined the catalyst
specificity and achievable partial current densities. The area of
the inner surfaces, not the total internal pore volume nor the
specific morphology, was decisive for the specificity and partial
current densities at low overpotentials.
A
A
A
¼ ECSA=ðqred=2eÞ
CO
CO
CO
*
*
*
¼ 37:7 cm2=ð9:1 mC=3:2 ꢃ 10ꢀ19 CÞ
¼ 13:2 ꢃ 10ꢀ20 m2 ¼ 13:2 ꢁ 3:6 ꢁ ꢃ 3:6 ꢁ
ð1Þ
The above surface area per adsorbed *CO, which corre-
sponds to approximately two copper surface atoms, is proba-
bly overestimated owing to concomitant CO release, but it
demonstrates that a near-saturating level of *CO coverage is
conceivable. Blockage of H2 by *CO “poisoning” within the
copper foam might contribute to suppression of H2 formation
in the copper foam interior. However, our data suggested that
the *CO coverage per ECSA was similar for low-surface-area
and high-surface-area copper foams (but nondetectable for
the metallic Cu foil). Therefore, it is unlikely that *CO poisoning
determines the activity trends shown in Figure 7.
The formation of CO and C2H4 at low overpotentials was
strongly favored by foams with large internal surface areas, in
contrast to H2 and CH4 formation. The correlation between the
partial current densities and Ain/Aout can be explained within
the following scenarios:
A bimolecular reaction of CO species present in the foam
cavities implied that for continuous catalytic operation, the ef-
fective residence time of chemisorbed *CO and gaseous CO is
a crucial determinant of the rate of C2H4 formation. Aside from
catalytic CO formation and conversion, the concentration of
chemisorbed *CO and gaseous CO molecules within the Cu
foams were determined by the *CO/CO equilibrium within the
cavities as well as the diffusion of gaseous CO molecules from
the inner-foam cavities to the bulk electrolyte. The resulting
*CO/CO concentrations may also be discussed in terms of an
effective *CO/CO residence time. This residence time was pre-
dicted to increase strongly with increasing internal surface area
of the foam in comparison to the outer macroscopic surface
area. The enhanced C2H4 formation efficiency for small average
distances between Cu nanoparticles has been explained
through a similar line of arguments.[16]
I. Macroscopic diffusion limited the proton supply and result-
ed in pronounced alkalization at the CuCat surface, as
shown herein by operando Raman spectroscopy. Because of
the low proton activity, H2 and CH4 formation was strongly
suppressed, whereas CO and C2H4 formation still proceeded
through a largely pH-independent mechanism.[2b,9a,19,34a]
II. Formation of chemisorbed *CO molecules could block
active surface sites and thereby might suppress formation
of H2.[6c,10d,16,17] The fraction of surface sites covered by *CO
could be significant. However, this fraction was not signifi-
cantly higher in high-surface-area foams compared with
that in low-surface-area foams. Consequently, *CO poison-
ing of the H2 formation reaction was unlikely to be the deci-
sive determinant of the reactivity trends reported in our
study.
Conclusions
III. Although jCO increased roughly in parallel with the ECSA of
the CuCat foams, a clearly stronger increase was observed
for jC H . Considering that jCO may be directly proportional to
Electrodeposited Cu foams initially consisted of both Cu oxide
(predominantly Cu+ but also Cu2+) and crystalline metallic do-
mains. The application of catalytic potentials that are sufficient-
ly negative for CO2 reduction resulted in massive reduction of
the Cu oxides such that a metallic copper phase prevailed.
However, this metallic copper differed from polycrystalline
copper metal owing to increased atomic disorder of the bulk
material. We could not exclude that this disordered metallic
copper contained dispersed (sub-surface) oxygen atoms, but
likely without the formation of any distinct Cu2O or CuO do-
mains, and not predominantly in the form of a surface oxide.
All the investigated copper foams initially contained a high
oxide content. Therefore, it is unlikely that the major reactivity
trends observed within the series of copper foams were deter-
mined by the initially present oxide fraction, which was trans-
formed into a disordered metallic phase at catalytic potentials.
A hydrogen-bubble template method for Cu electrodeposi-
tion produced foams with a thickness of 40–360 mm. A variety
of pore dimensions and morphologies were obtained depend-
ing on the deposition current, deposition time, and choice of
metal substrate (Pt or Cu) during Cu electrodeposition. For all
these foams, only approximately 2% of their total volume was
2
4
the surface-adsorbed *CO molecules, we propose that C2H4
was formed within the foam from two CO precursor species,
namely adsorbed *CO and dissolved CO present within the
foam cavities.
Experimental Section
Synthesis
The electrodeposition of Cu/Cu2O-based catalysts (CuCats) was per-
formed using a modified protocol described by Shin et al.[21a] Three
electrodes in a single compartment open electrolysis cell were
used without stirring or heating. An Ag/AgCl electrode (Methrom)
was used as a reference electrode, Pt foil was used as a counter
electrode, and cleaned and etched Cu or Pt foil (0.1 mm thickness,
99.99+% purity, Goodfellow) was used as the working electrode
and controlled by a Biologic SP300 potentiostat. Negative currents
of 0.5 or 1.0 A were applied for either 10 s or 20 s. The distance be-
tween the counter and working electrode was kept constant at ap-
proximately 3 mm. The electrolyte consisted of 0.2m CuSO4 and
&
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
8
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
1.5m H2SO4. Before use, the Cu or Pt foil was polished with dia-
mond paste (0.7 mm), cleaned with approximately 10% HNO3, and
then cleaned stepwise in an ultrasonic bath using milliQ H2O and
ethanol. Immediately before electrodeposition, the foils were
etched with approximately 10% HNO3 (in the case of Cu) or with
approximately 10% HCl (in the case of Pt). The backside of the
substrates was covered with silicon glue. Dark red films were ob-
tained, which were washed with milliQ water, followed by dry
blowing with N2 gas. Unless otherwise stated, prior to any further
ex situ analysis, the surface oxides were reduced with a set of CVs
(100 mVsꢀ1, from 0 to ꢀ1.05 VRHE) and LSV (5 mVsꢀ1, from ꢀ0.3 to
ꢀ1.1 VRHE).
B) Agilent Technologies 1200 Series HPLC was used for separation,
detection, and quantification of the carboxylic acids such as for-
mate and acetate. Separation was performed in an Organi-Acid
Resin column flushed with 0.005m H2SO4 at 1 mLminꢀ1 flow. De-
tection was achieved using a refraction index detector (RID). Cali-
bration was performed for formate and acetate.
Electrochemically active surface area
The electrochemical capacitance of the samples was determined
using CV[10a,35] at 8 different scan rates (10, 20, 30, 40, 50, 60, 80,
and 100 mVsꢀ1). Non-Faradaic responses were found in potential
ranges of 0.15 V around the OCP. All currents in this region were
assigned to double-layer charging.[35c] Prior to each CV, the starting
potential of the sample was held for 10 s. The charging current, i,
(as an average of the absolute values of the anodic and cathodic
sweep) of the double layer, was equal to the product of the scan
The Cu foil (0.1 mm thickness, 99.9999% purity, AlfaAeser) used as
a reference was polished with micropolish alumina suspension
(0.05 mm) on a microcloth, then cleaned stepwise in an ultrasonic
bath with milliQ H2O, ethanol, and milliQ H2O each for 10 min, fol-
lowed by etching in 10% HNO3 for 30 s.
rate, v, and the electrochemical double-layer capacitance, CDL
:
Electrochemical and chromatographic analysis of gaseous
products
i ¼ v ꢃ CDL
ð2Þ
The slope of the linear fit of the plot of i against the 8 different v is
the CDL. Calculation of the electrochemically active surface area
(ECSA) was performed by using Equation (3). For the specific ca-
pacitance Cs we used an experimental double layer capacitance
value of 38ꢂ3 mF for Cu foil.
The electrochemical CO2 reduction was performed in a custom
made gas tight two compartment electrochemical cell (Figure S9)
separated by a Nafion 211 membrane using a Gamry Reference
600 potentiostat. All the measured potentials were compensated
for iR drop and are represented with respect to reversible hydro-
gen electrode (RHE). Aqueous 0.1m KHCO3 solution was used as
the electrolyte after saturating with CO2 gas (99.995 vol.%, Air Liq-
uide) for 10 min. During electrolysis, the electrolyte in the working
electrode (WE) compartment was purged with CO2 from the
bottom of the cell at a flow rate of 20 mLminꢀ1 using a calibrated
mass flow controller. The electrolyte at the WE compartment was
stirred continuously to achieve maximum mass transport of CO2 to
the electrode surface. A leak-free Ag/AgCl reference electrode was
placed in front of the working electrode over a 1 cm2 area. A Pt
coil (Biologic) was used as the counter electrode and CO2 was
purged to the counter electrode compartment. The gaseous prod-
ucts formed from the CO2 reduction reaction were identified and
quantified using an online gas chromatograph (GC) (Shimadzu GC-
2014) coupled with a two-compartment electrochemical cell. The
product gases were analyzed after 20 min and 1 h of electrolysis at
a specified potential. No yield differences were observed between
20 min or 1 h. During sampling, the CO2 flow along with the prod-
uct gases in the loop were injected into the GC by switching a 10-
port valve. The GC was equipped with a thermal conductivity de-
tector (TCD) for the detection of H2 and other permanent gases. A
methanizer in series with a flame ionization detector (FID) was
used for detecting CO and hydrocarbons. Grade 5 Ar was used as
the carrier gas.
CDL
ECSA ¼
=
ð3Þ
Cs
X-ray spectroscopy
Cu K-edge XAS was performed at beamline KMC-3 at Helmholtz-
Zentrum-Berlin (BESSY II) at 20 K. Energy calibration was done by
collecting spectra of Cu metal foil (0.0075 mm, 99.99+%, Goodfel-
low) and shifting the energy values according to the first inflection
point of the first derivative of the Cu foil spectrum. Reference ma-
terial spectra of CuO, CuCl, CuSO4, and [Cu(OH2)6]2+ were collected
in absorption mode with ionization chambers in front and behind
the sample. Spectra of the Cu-based catalysts, Cu metal foil, and
Cu2O were all collected in fluorescence mode with a 13 element
windowless Germanium detector (Canberra) with a detuned X-ray
beam to avoid oversaturation and distortion owing to the high
metal content. The catalysts deposited on platinum were frozen in
liquid N2 after electrodeposition and three CVs and one LSV (0 to
ꢀ1 VRHE) were performed. In the case of the copper substrate, the
thin catalyst films could not be distinguished from the underlying
substrate. Therefore, the catalyst layers were carefully scratched off
the surface using plastic razors. To ensure that this method did not
affect the atomic structure of the sample, we compared the Cu K-
edge spectra of a thin film on Pt and one scratched off the sub-
strate surface. Figure S19 shows that no artefacts were introduced
by this method and the atomic structure remained identical re-
gardless if the sample was a thin film or a powder. Linear combina-
tions of Cu and Cu2O were selected by least-square fits (with the
weighing factors as fit parameters), which led to identical results in
the XANES and EXAFS spectra.
Analysis of liquid products
Quantitative analysis of the liquid products was performed with
two separate chromatographs (A and B). A) An offline gas chroma-
tograph Shimadzu GC-2010-Plus with AOC 150i (autosampling)
was used for detection and quantification of the volatile liquid
products such as acetaldehyde, propionaldehyde, ethanol, propa-
nol, allyl alcohol, and acetone. Microliter amounts of the liquid
sample were injected and converted to a gas phase at elevated
temperature in a quartz liner; some of the released gas was inject-
ed into the He carrier gas and to a SH-Stabilwax column for sepa-
ration. Aldehydes and alcohols were detected using a FID. Calibra-
tion was performed regularly for the six volatile liquid products.
Operando Cu K-edge measurements were performed at 293 K, as
described previously.[36] The CuCat was deposited on carbon paper
(FuelCellStore) at 0.5 A for 20 s. The deposited catalyst was placed
in the window of a home-made electrochemical cell (Teflon) by
connecting parts of pure carbon paper with Cu tape to connect
the electrode. The backside of the working electrode consisting of
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
9
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
only carbon paper (facing the X-ray beam) was covered with
Kapton tape. A small Ag/AgCl working electrode was mounted
close to the catalyst film (ꢁ0.2 mm). A platinum coil was used as a
counter electrode. The excited X-ray fluorescence passed through
the Kapton window and then through a 10 mm Ni metal foil shield-
ing against scattered light. The fluorescence was monitored per-
the Helmholtz-Zentrum Berlin (HZB) for beamtime allocation at
the KMC-3 synchrotron beamline of the BESSY synchrotron in
Berlin-Adlershof and Dr. Ivo Zizak for special support. We thank
Reza M. Mohammadi and Drs. Petko Chernev, Diego Gonzalez-
Flores, Vadim Sikolenko, Stefan Mebs, and Michael Haumann for
assisting with XAS data collection.
pendicular to the incident beam by
a scintillation detector
(19.6 cm2 active area, 51BMI/2E1-YAP-Neg, Scionix; shielded by ad-
ditional 2 mm Al foils against visible light). The detector consisted
of a scintillating crystal (YAP) converting X-ray photons into visible
light detected by a fast photomultiplier operated at 500 V connect-
ed to a Keithley multimeter for analog-to-digital data conversion.
Absorption spectra of Cu metal foil were measured for precise
energy calibration. At the start of each experiment, spectra of the
dry film (as deposited) were collected. Then, 0.1m KHCO3, saturat-
ed with CO2, pH 6.8, was filled into the cell, and a constant CO2
stream of 1 mLminꢀ1 was kept in front of the film surface during
operation.
Conflict of interest
The authors declare no conflict of interest.
Keywords: CO2 electroreduction · electrocatalysis · non-fossil
fuels · Raman spectroscopy · X-ray spectroscopy
Details about the quasi in situ XPS measurements and the corre-
sponding data evaluation have been described elsewhere.[14b]
[1] a) M. Gattrell, N. Gupta, A. Co, J. Electroanal. Chem. 2006, 594, 1–19;
b) D. D. Zhu, J. L. Liu, S. Z. Qiao, Adv. Mater. 2016, 28, 3423–3452;
c) K. D. Yang, C. W. Lee, K. Jin, S. W. Im, K. T. Nam, J. Phys. Chem. Lett.
2017, 8, 538–545; d) J. P. Jones, G. K. S. Prakash, G. A. Olah, Isr. J. Chem.
2014, 54, 1451–1466.
Operando Raman spectroscopy
[2] a) K. P. Kuhl, E. R. Cave, D. N. Abram, T. F. Jaramillo, Energy Environ. Sci.
2012, 5, 7050–7059; b) R. Kortlever, J. Shen, K. J. P. Schouten, F. Calle-
Vallejo, M. T. M. Koper, J. Phys. Chem. Lett. 2015, 6, 4073–4082.
[3] a) J. W. Vickers, D. Alfonso, D. R. Kauffman, Energy Technol. 2017, 5,
775–795; b) H. Mistry, A. S. Varela, S. Kꢃhl, P. Strasser, B. R. Cuenya, Nat.
Rev. Mater. 2016, 1, 16009.
[4] L. Zhang, Z.-J. Zhao, J. Gong, Angew. Chem. 2017, 129, 11482–11511.
[5] a) R. Reske, M. Duca, M. Oezaslan, K. J. P. Schouten, M. T. M. Koper, P.
Strasser, J. Phys. Chem. Lett. 2013, 4, 2410–2413; b) K. Manthiram, B. J.
Beberwyck, A. P. Alivisatos, J. Am. Chem. Soc. 2014, 136, 13319–13325.
[6] a) R. Reske, H. Mistry, F. Behafarid, B. Roldan Cuenya, P. Strasser, J. Am.
Chem. Soc. 2014, 136, 6978–6986; b) H. Mistry, F. Behafarid, R. Reske,
A. S. Varela, P. Strasser, B. Roldan Cuenya, ACS Catal. 2016, 6, 1075–
1080; c) Y. Huang, A. D. Handoko, P. Hirunsit, B. S. Yeo, ACS Catal. 2017,
7, 1749–1756.
[7] a) B. Roldan Cuenya, Acc. Chem. Res. 2013, 46, 1682–1691; b) H. S. Jeon,
S. Kunze, F. Scholten, B. Roldan Cuenya, ACS Catal. 2018, 8, 531–535;
c) D. Gao, I. Zegkinoglou, N. J. Divins, F. Scholten, I. Sinev, P. Grosse, B.
Roldan Cuenya, ACS Nano 2017, 11, 4825–4831.
[8] W. Luo, X. Nie, M. J. Janik, A. Asthagiri, ACS Catal. 2016, 6, 219–229.
[9] a) A. S. Varela, M. Kroschel, T. Reier, P. Strasser, Catal. Today 2016, 260,
8–13; b) G. Z. Kyriacou, A. K. Anagnostopoulos, J. Appl. Electrochem.
1993, 23, 483–486; c) S. Kaneco, K. Iiba, K. Ohta, T. Mizuno, J. Solid State
Electrochem. 1999, 3, 424–428; d) K. J. P. Schouten, E. Pꢄrez Gallent,
M. T. M. Koper, J. Electroanal. Chem. 2014, 716, 53–57; e) M. Akira, H.
Yoshio, Bull. Chem. Soc. Jpn. 1991, 64, 123–127.
[10] a) X. Feng, K. Jiang, S. Fan, M. W. Kanan, ACS Cent. Sci. 2016, 2, 169–
174; b) X. Feng, K. Jiang, S. Fan, M. W. Kanan, J. Am. Chem. Soc. 2015,
137, 4606–4609; c) C. W. Li, M. W. Kanan, J. Am. Chem. Soc. 2012, 134,
7231–7234; d) A. Verdaguer-Casadevall, C. W. Li, T. P. Johansson, S. B.
Scott, J. T. McKeown, M. Kumar, I. E. L. Stephens, M. W. Kanan, I. Chor-
kendorff, J. Am. Chem. Soc. 2015, 137, 9808–9811.
[11] H. Mistry, A. S. Varela, C. S. Bonifacio, I. Zegkinoglou, I. Sinev, Y.-W. Choi,
K. Kisslinger, E. A. Stach, J. C. Yang, P. Strasser, B. R. Cuenya, Nat.
Commun. 2016, 7, 12123.
[12] a) T. T. H. Hoang, S. Ma, J. I. Gold, P. J. A. Kenis, A. A. Gewirth, ACS Catal.
2017, 7, 3313–3321; b) S. Zhu, M. Shao, J. Solid State Electrochem. 2016,
20, 861–873; c) H. Xiao, W. A. Goddard, T. Cheng, Y. Liu, Proc. Natl. Acad.
Sci. USA 2017, 114, 6685–6688.
Raman spectra were collected with a Renishaw inVia Raman spec-
trometer coupled with a Leica microscope. Calibration was done
using a silicon wafer standard (521 cmꢀ1). A water immersion ob-
jective (Leica, 40ꢂ, numerical aperture=0.8) was used to focus
and collect the incident and scattered laser light during electro-
chemical measurements. A 0.013 mm thin optically transparent
Teflon film (DuPont) was used to protect the objective from alka-
line solution. The spectra were obtained in the range of 248–
1869 cmꢀ1 using a grating with 2400 linesmmꢀ1 and a slit size of
20 mm. A 473 nm laser (YAG laser, 100% power) focusing on a line
(ꢁ30ꢂ5 mm) served as an excitation source. The acquisition time
for a spectrum with 5 accumulations was 35 s. A home-built spec-
troelectrochemical cell made of polytetrafluoroethylene was inter-
faced with the Raman microscope for spectroscopic measure-
ments. The cell was equipped with an Ag/AgCl reference electrode
and a Platinum ring counter electrode and controlled with a SP-
200 Biologic potentiostat. To acquire spectra at different distances
from the catalyst surface, we performed depth scans at each po-
tential, employing the line-focus option of the inVia Raman spec-
trometer. The focusing point on the void structure of our catalyst
surface was set as zero position. Before measurement, we per-
formed three CVs and one LSV to reduce the surface oxides. The
depth scans were performed in 0.1m KHCO3 electrolyte saturated
with CO2, and the electrolyte was continuously purged with CO2
throughout the experiment. The desired potential was applied for
2 min before collecting the spectra to ensure steady-state condi-
tions. All spectra were smoothed and baseline corrected by Renish-
aw software. Peak area analysis was performed by in-house soft-
ware “gloFit”.
Further experimental details are provided in the Supporting Infor-
mation.
Acknowledgements
[13] R. Kas, K. K. Hummadi, R. Kortlever, P. de Wit, A. Milbrat, M. W. J. Luiten-
Olieman, N. E. Benes, M. T. M. Koper, G. Mul, Nat. Commun. 2016, 7,
10748.
[14] a) D. Kim, S. Lee, J. D. Ocon, B. Jeong, J. K. Lee, J. Lee, Phys. Chem.
Chem. Phys. 2015, 17, 824–830; b) D. Gao, F. Scholten, B. Roldan Cue-
nya, ACS Catal. 2017, 7, 5112–5120; c) C. S. Chen, A. D. Handoko, J. H.
Wan, L. Ma, D. Ren, B. S. Yeo, Catal. Sci. Technol. 2015, 5, 161–168.
This work was funded by the German Federal Ministry of Educa-
tion and Research (Bundesministerium fꢁr Bildung und Forschung
(BMBF)) under grants 03SF0523-A,B,C (CO2EKAT) as well as
033RCOO4D (eEthylene) and 05K16KE2 (IN-SITU-XAS), and the Eu-
ropean Research Council (OPERANDOCAT,ERC-725915). We thank
&
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
10
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
[15] a) Y. Hori, A. Murata, R. Takahashi, J. Chem. Soc. Faraday Trans. 1 1989,
85, 2309–2326; b) K. J. P. Schouten, Y. Kwon, C. J. M. van der Ham, Z.
Qin, M. T. M. Koper, Chem. Sci. 2011, 2, 1902–1909; c) A. A. Peterson, F.
Abild-Pedersen, F. Studt, J. Rossmeisl, J. K. Norskov, Energy Environ. Sci.
2010, 3, 1311–1315; d) F. Calle-Vallejo, M. T. M. Koper, Angew. Chem. Int.
Ed. 2013, 52, 7282–7285; Angew. Chem. 2013, 125, 7423–7426; e) A. J.
Garza, A. T. Bell, M. Head-Gordon, ACS Catal. 2018, 8, 1490–1499.
[16] X. Wang, A. S. Varela, A. Bergmann, S. Kꢃhl, P. Strasser, ChemSusChem
2017, 10, 4642–4649.
[17] a) H. Ooka, M. C. Figueiredo, M. T. M. Koper, Langmuir 2017, 33, 9307–
9313; b) D. Raciti, M. Mao, C. Wang, Nanotechnology 2018, 29, 044001;
c) C. Chen, B. Zhang, J. Zhong, Z. Cheng, J. Mater. Chem. A 2017, 5,
21955–21964.
[18] N. Gupta, M. Gattrell, B. MacDougall, J. Appl. Electrochem. 2006, 36,
161–172.
[19] M. Ma, K. Djanashvili, W. A. Smith, Angew. Chem. Int. Ed. 2016, 55, 6680–
6684; Angew. Chem. 2016, 128, 6792–6796.
[20] S. Sen, D. Liu, G. T. R. Palmore, ACS Catal. 2014, 4, 3091–3095.
[21] a) H. C. Shin, J. Dong, M. Liu, Adv. Mater. 2003, 15, 1610–1614; b) A.
Dutta, M. Rahaman, N. C. Luedi, M. Mohos, P. Broekmann, ACS Catal.
2016, 6, 3804–3814.
[22] a) T. N. Huan, P. Simon, G. Rousse, I. Genois, V. Artero, M. Fontecave,
Chem. Sci. 2017, 8, 742–747; b) C. Reller, R. Krause, E. Volkova, B.
Schmid, S. Neubauer, A. Rucki, M. Schuster, G. Schmid, Adv. Energy
Mater. 2017, 7, 1602114.
[26] Y. Hori, A. Murata, R. Takahashi, S. Suzuki, J. Chem. Soc. Chem. Commun.
1988, 17–19.
[27] R. Kortlever, K. H. Tan, Y. Kwon, M. T. M. Koper, J. Solid State Electrochem.
2013, 17, 1843–1849.
[28] 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.
[29] A. Wuttig, C. Liu, Q. Peng, M. Yaguchi, C. H. Hendon, K. Motobayashi, S.
Ye, M. Osawa, Y. Surendranath, ACS Cent. Sci. 2016, 2, 522–528.
[30] M. D. Merrill, R. C. Dougherty, J. Phys. Chem. C 2008, 112, 3655–3666.
[31] P. Grosse, D. Gao, F. Scholten, I. Sinev, H. Mistry, B. Roldꢅn Cuenya,
Angew. Chem. Int. Ed. 2018, 57, 6192–6197; Angew. Chem. 2018, 130,
6300–6305.
[32] K. Klingan, F. Ringleb, I. Zaharieva, J. Heidkamp, P. Chernev, D. Gonzalez-
Flores, M. Risch, A. Fischer, H. Dau, ChemSusChem 2014, 7, 1301–1310.
[33] J. H. Montoya, C. Shi, K. Chan, J. K. Nørskov, J. Phys. Chem. Lett. 2015, 6,
2032–2037.
[34] a) R. Kas, R. Kortlever, H. Yılmaz, M. T. M. Koper, G. Mul, ChemElectro-
Chem 2015, 2, 354–358; b) K. J. P. Schouten, Z. Qin, E. Pꢄrez Gallent,
M. T. M. Koper, J. Am. Chem. Soc. 2012, 134, 9864–9867; c) F. S. Roberts,
K. P. Kuhl, A. Nilsson, Angew. Chem. 2015, 127, 5268–5271.
[35] a) C. W. Li, J. Ciston, M. W. Kanan, Nature 2014, 508, 504–507; b) Y.
Song, R. Peng, D. K. Hensley, P. V. Bonnesen, L. Liang, Z. Wu, H. M.
Meyer, M. Chi, C. Ma, B. G. Sumpter, A. J. Rondinone, ChemistrySelect
2016, 1, 6055–6061; c) C. C. L. McCrory, S. Jung, J. C. Peters, T. F. Jaramil-
lo, J. Am. Chem. Soc. 2013, 135, 16977–16987.
[23] Y. Lum, W. Ager Joel, Angew. Chem. Int. Ed. 2018, 57, 551–554; Angew.
Chem. 2018, 130, 560–563.
[24] L. Mandal, K. R. Yang, M. R. Motapothula, D. Ren, P. Lobaccaro, A. Patra,
M. Sherburne, V. S. Batista, B. S. Yeo, J. W. Ager, J. Martin, T. Venkatesan,
ACS Appl. Mater. Interfaces 2018, 10, 8574–8584.
[36] D. Gonzalez-Flores, K. Klingan, P. Chernev, S. Loos, M. R. Mohammadi, C.
Pasquini, P. Kubella, I. Zaharieva, R. D. L. Smith, H. Dau, Sustainable
Energy Fuels 2018, 2, 1986–1944.
Manuscript received: July 13, 2018
Version of record online: && &&, 0000
[25] B. J. Plowman, L. A. Jones, S. K. Bhargava, Chem. Commun. 2015, 51,
4331–4346.
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
11
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

FULL PAPERS
K. Klingan, T. Kottakkat, Z. P. Jovanov,
S. Jiang, C. Pasquini, F. Scholten,
P. Kubella, A. Bergmann,
Inner values matter! Eight Cu foams
with substantially different morpholo-
gies were investigated by analyzing the
product spectrum. Electrochemical in
situ spectroscopies revealed that the re-
activity was controlled by internal sur-
face area.
B. Roldan Cuenya, C. Roth, H. Dau*
&& – &&
Reactivity Determinants in
Electrodeposited Cu Foams for
Electrochemical CO2 Reduction
&
ChemSusChem 2018, 11, 1 – 12
www.chemsuschem.org
12
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!