Sub-Second Time-Resolved Surface-Enhanced Raman Spectroscopy Reveals Dynamic CO Intermediates during Electrochemical CO2 Reduction on Copper

Article abstract of DOI:10.1002/anie.202104114

The electrocatalytic carbon dioxide (CO₂) reduction reaction (CO₂RR) into hydrocarbons is a promising approach for greenhouse gas mitigation, but many details of this dynamic reaction remain elusive. Here, time-resolved surface-enhanced Raman spectroscopy (TR-SERS) is employed to successfully monitor the dynamics of CO₂RR intermediates and Cu surfaces with sub-second time resolution. Anodic treatment at 1.55 V vs. RHE and subsequent surface oxide reduction (below ?0.4 V vs. RHE) induced roughening of the Cu electrode surface, which resulted in hotspots for TR-SERS, enhanced time resolution (down to ≈0.7 s) and fourfold improved CO₂RR efficiency toward ethylene. With TR-SERS, the initial restructuring of the Cu surface was followed (<7 s), after which a stable surface surrounded by increased local alkalinity was formed. Our measurements revealed that a highly dynamic CO intermediate, with a characteristic vibration below 2060 cm^?1, is related to C?C coupling and ethylene production (?0.9 V vs. RHE), whereas lower cathodic bias (?0.7 V vs. RHE) resulted in gaseous CO production from isolated and static CO surface species with a distinct vibration at 2092 cm^?1.

Full text of DOI:10.1002/anie.202104114

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
International Edition Chemie
www.angewandte.org
Accepted Article
Title: Sub-second Time-resolved Surface Enhanced Raman
Spectroscopy Reveals Dynamic CO Intermediates during
Electrochemical CO2 Reduction on Copper
Authors: Hongyu An, Longfei Wu, Laurens Mandemaker, Shuang
Yang, Jim De Ruiter, Jochem Wijten, Joris Janssens, Thomas
Hartman, Ward van der Stam, and Bert Marc Weckhuysen
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.202104114
Link to VoR: https://doi.org/10.1002/anie.202104114

Angewandte Chemie International Edition
10.1002/anie.202104114
RESEARCH ARTICLE
Sub-second Time-resolved Surface Enhanced Raman
Spectroscopy Reveals Dynamic CO Intermediates during
Electrochemical CO
2
Reduction on Copper
[
a]
[a]
[a]
[a]
[a]
Hongyu An , Longfei Wu , Laurens D. B. Mandemaker , Shuang Yang , Jim de Ruiter , Jochem H.
[
a]
[a]
[a]
[a]
J. Wijten , Joris C. L. Janssens , Thomas Hartman , Ward van der Stam * and Bert M.
[
a]
Weckhuysen *
[
a]
H. An, L. Wu, L. D. B. Mandemaker, S. Yang, J. de Ruiter, J. H. J. Wijten, J. C. L. Janssens, T. Hartman, W. van der Stam and B. M. Weckhuysen
Inorganic Chemistry and Catalysis,
Institute for Sustainable and Circular Chemistry,
Utrecht University, Universiteitsweg 99, 3584 CG Utrecht, The Netherlands.
b.m.weckhuysen@uu.nl and w.vanderstam@uu.nl
Supporting information for this article is given via a link at the end of the document
Abstract: Electrocatalytic reduction of carbon dioxide (CO
value-added products (e.g., ethylene) is a promising approach for
greenhouse gas mitigation, but many details of electrocatalytic CO
reduction reactions (CO RR) remain elusive. Raman spectroscopy is
suitable for in situ characterization of CO RR mechanisms, but the low
2
) into
deciding factors.[3] Recent research on oxide-derived Cu
electrocatalysts and pulsed-electrolysis discovered that oxide
species covering Cu surfaces are reduced under cathodic bias,
2
2
2
which led to improved CO RR activity ascribed to the formation of
an activated electrode through surface reconstructions.[4]
Adsorbed CO at the catalyst surface has distinct optical
signatures, and is usually considered to play a crucial role towards
hydrocarbon (e.g., ethylene and methane) formation during
2
signal intensity and resulting poor time resolution (often up to minutes)
hampers the application of conventional Raman spectroscopy for the
2
study of the dynamic CO reduction reaction, which requires sub-
second time resolution. By using Time-Resolved Surface Enhanced
Raman Spectroscopy (TR-SERS) we were able to successfully
CO
2
RR on Cu.[3d,4b,5] Through steady-state vibrational
spectroscopy studies, different adsorption modes of CO at the
electrode surface have been elucidated, such as CO adsorbed on
terrace- and defect-sites, and bridged CO.[6] However, many
details about the dynamic surface chemistry of CO intermediates
and C-C coupling, which are considered crucial for achieving C2+
products,[ remain unclear. Reconstruction of the catalyst surface,
as well as formation and coupling of CO intermediates, may
happen on a timescale near or below one second depending on
reaction conditions. Therefore, in situ time-resolved spectroscopic
techniques are necessary to follow the evolution of surface
reconstruction and adsorbed species in real-time, in order to gain
more insight into the reaction mechanisms. [6b,8]
monitor CO
Anodic treatment at 1.55 V vs. the reversible hydrogen electrode
RHE) and subsequent surface oxide reduction (below −0.4 V vs.
2
RR over Cu surfaces with sub-second time resolution.
(
RHE) induced roughening of the Cu electrode surface, which resulted
in hot-spots for TR-SERS, enhanced time resolution (down to ~ 0.7 s)
7]
2
and improved CO RR efficiency (i.e., four-fold increase in ethylene
faradaic efficiency). With TR-SERS, the initial formation of hot-spots
for SERS and CO2RR was followed (<7 s), after which a stable copper
surface surrounded by increased local alkalinity was formed. Our
measurements revealed that a highly dynamic CO intermediate, with
−
1
a characteristic vibration below 2060 cm , is related to C-C coupling
and ethylene production (−0.9 V vs. RHE), whereas lower cathodic
bias (−0.7 V vs. RHE) resulted in gaseous CO production from
isolated and static CO surface species with a distinct vibration at 2092
In situ vibrational spectroscopy is a paramount technique for
studying the steady-state chemistry of CO surface species and
intermediates on catalyst surfaces.[5d,5g,9] Due to the low scattering
and absorbing cross sections of water, in situ Raman
−
1
cm . Our results provide valuable time-resolved insights into the
dynamic nature of the electrode surface and adsorbed intermediates
spectroscopy has seen many successful applications in aqueous
environments.[
4c,5g,5f,10]
Compared to infrared spectroscopy (IR),
during CO
2
electrochemical reduction on copper and showcase the
potential of TR-SERS in copper-based electrocatalysis to follow
reaction dynamics.
Raman spectroscopy can also easily collect vibrational signal in
the low-wavenumber region, where valuable information about
catalyst structure is usually observed (e.g., surface oxide
species).[ However, extended acquisition times are typically
required to obtain good signal/noise ratios due to the low Raman
scattering probability (in the order of several minutes), which
results in a poor time resolution. Surface Enhanced Raman
Spectroscopy (SERS) can be exploited to boost the sensitivity
11]
Introduction
Copper (Cu) is a unique metal due to its outstanding ability to
produce ethylene and other C2+ products in the electrocatalytic
CO reduction reaction (CO
2 2
RR).[1] However, the exact reaction
towards adsorbed species through surface modifications at the
nanoscale, enabling shorter acquisition times.[12] In recent years,
mechanism for C2+ products is complex and still debated in
literature, since multiple electron- and proton-transfer steps are
involved that need to occur in a concerted manner.[2] Both the
structure of the electrode surface and the chemistry of surface
intermediates are considered to be important performance-
SERS has already seen many applications in catalysis
[11c,13]
research.
Interestingly, Cu is known to exhibit strong SERS
[13c, 13f]
activity in the near infrared excitation wavelength,
next to its
unique electrocatalytic abilities for C-C coupling and C2+ product
formation.
1
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202104114
RESEARCH ARTICLE
In this work, we take advantage of these two characteristics of Cu,
that TR-SERS may not be suitable for ideal, well-defined Cu
facets, which have poor intrinsic Raman signal enhancement.
which acts both as an active CO
active substrate in our experiments, to achieve sub-second in situ
time-resolved SERS under CO RR conditions. This unique
2
RR electrocatalyst and SERS-
External
signal
enhancement
through
Shell-Isolated
[
9b]
2
Nanoparticles is necessary to study flat Cu electrodes . After
performing an anodic treatment at 1.55 V for 120 s, the Raman
signal associated with surface oxide species increases in
signal/noise ratio, but again disappears within a second during
subsequent reduction at −0.4 V (Figure S4). In addition to the
combination enabled us to investigate the dynamic surface
reconstruction of Cu, as well as chemical processes of adsorbed
CO species on polycrystalline Cu electrodes during CO RR. Our
2
experiments reveal that anodic treatment (1.55 V, all biases in this
manuscript are vs. RHE) and subsequent reduction (<–0.4 V vs.
RHE) roughens the electrode surface, resulting in hot-spots for
enhanced SERS activity, as well as a four-fold increase in CO-
coupling efficiency towards C2+ products on Cu-based electrodes
x
disappearance of the CuO Raman signal, a clear peak at 1060
cm−1 is observed one second after the onset of a sufficient
cathodic bias (Figure 1d). This signal is assigned to adsorbed
2
−
[13d,18]
carbonate species (CO
3
) based on experiments
and
[
4b]
(
at −1.0 V vs. RHE). The time- and potential-dependent behavior
theoretical calculations on oxide-derived Cu electrodes . The
signal collected in the bulk electrolyte solution (Figure S5) is much
weaker than on anodized and reduced Cu-MP surface, and it
in our TR-SERS experiments reveals a dynamic CO surface
−
1
intermediate with a distinct vibration below 2060 cm , which only
appears for cathodic biases below −0.8 V vs. RHE. Selectivity
measurements suggest that this dynamic CO intermediate is
correlated to the production of ethylene at a cathodic bias of −0.9
V vs. RHE. At lower cathodic bias (−0.7 V vs. RHE), the Raman
spectra are dominated by a static vibrational signature at 2092
−
−1 [18]
mainly comes from dissolved HCO
3
(at 1012 cm ),
and no
CO 2
−
peak (at 1060 cm ) could be detected in the solution. This
−1
3
−
shows that the HCO
3
electrolyte ions are rapidly converted into
oxide. We attribute
CO 2
−
during the reduction of surface CuO
3
x
2
−
the formation of CO
3
in the initial stages after cathodic bias
onset to the deprotonation of HCO
−
1
−
cm , which is ascribed to CO adsorbed on undercoordinated
sites, resulting in gaseous CO production. Our results display that
3
due to the depletion of
protons during surface oxide/hydroxide reduction and hydrogen
evolution reaction (HER). The rapid appearance of the carbonate
band after the oxide stripping suggests that the reduced surface
of anodized Cu-MP is highly SERS-active.
The crucial role of anodization and subsequent reduction on the
strong Raman enhancement effect is further evidenced by the
Scanning Electronic Microscopy (SEM) and Atomic Force
Microscopy (AFM) images of pristine (Figure 1e) and anodized
Cu-MP (Figure 1f). The pristine Cu-MP electrode is rather smooth,
but also displays grooves as a result of mechanical polishing
2
CO RR and their intermediates are dynamic, and showcase the
need for improved time-resolved in situ spectroscopic
investigations down to milliseconds in order to investigate the
reaction kinetics in great detail.
Results and Discussion
In situ TR-SERS at −0.4 V: Anodization Leads to Formation of
(
Figure 1e). After anodic treatment and subsequent reduction,
nanoparticles can be observed on the electrode surface (Figure
f). Atomic Force Microscopy (AFM) results show that most of the
nanoparticles formed after anodic treatment are between 50 to
50 nm in size (Figure S6), comparable with previous reports
“
Hot-Spots”. Mechanically polished polycrystalline Cu (referred
to as Cu-MP hereafter) [14] is used as working electrode in our in
situ electrochemical Raman cell (three electrode cell, glassy
1
carbon counter electrode, Ag/AgCl reference electrode, CO
saturated 0.1 M KHCO aqueous electrolyte, pH = 6.8, see Figure
2
-
1
3
about SERS-active nanoparticles.[13c,13e,19] Electrochemical
capacitance tests (Figure S7) reveal that the double layer
electrochemical capacitance of Cu-MP increases by a factor of
eight after anodization and subsequent reduction, indicating an
increase in surface roughness. An increase in surface roughness
is also clearly observed in the AFM measurements, in which the
root mean square roughness (RMS) of the electrodes increases
from 13 nm to 41 nm after anodic treatment (Figure 1e,f and
Figure S6). We also recorded a live video of a Cu-MP electrode
during the entire potential cycling experiment with an optical
microscope, which also indicates surface roughening due to the
anodic treatment (video link and selected frames shown in Figure
S8).
S1). We have used both glassy carbon and Pt as counter
electrode, in order to exclude possible Pt contamination on the
working electrode.[15] No apparent differences were observed
(
see Supporting Discussion for comparison). The data with glassy
carbon as counter electrode is used in the main text of this
manuscript. The X-ray Diffraction (XRD) pattern (Figure S2)
shows that the Cu lattice is mainly (100) oriented without
indication of surface oxide species.[1c] We designed a potential
pulsing experiment (Figure 1a) to study the time-dependent
SERS intensity from Cu-MP and the effect of anodic treatment on
the surface structure. By continuously collecting spectra at a rate
of 717 ms per in situ Raman spectrum we obtained spectral
heatmaps, which allowed us to dynamically follow the time-
resolved behavior of the Raman signal. A weak Raman signal of
surface oxide can be observed with steady-state Raman
Besides the increased SERS effect that we have observed,
anodic treatment has also been applied to increase the Faradaic
efficiency (FE) and product selectivity toward C2+ products during
spectroscopy of pristine Cu-MP before reduction, evidenced by
two broad bands at 524 cm−1 and 614 cm−1 (Figure 1b).[13c, 16]
CO RR over Cu electrodes.[ Product analysis of pristine and
20]
2
anodized Cu-MP using an H-Cell (shown in Figure S9) is
displayed in Figure 1g. The reaction potential is set at −1.0 V,
which is a standard in literature for measuring ethylene formation
on Cu electrodes.[3f,16] For both samples, the major products
These bands, assigned to Cu oxide surface species, disappear
within one second after the onset of −0.4 V reducing potential, in
accordance with the cyclic voltammetry (CV) results (Figure S3),
evidencing the stripping of surface oxide species,[17] thereby
2
reduced from CO are methane, CO and ethylene. The FE of Cu-
exposing reduced and activated Cu surface for CO
2
RR. No
MP towards ethylene improved four-fold after anodization (23.0 %
compared to 4.8%). This reveals that C-C coupling of CO
intermediates are facilitated on the roughened anodized Cu-MP
obvious peaks can be observed in the carbonate region (900-
−
1
1200 cm , Figure 1c) for the pristine electrode, suggesting its
poor Raman enhancement after reduction. This also suggests
2
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202104114
RESEARCH ARTICLE
surface.[3a,3b,16] This is in line with the improved SERS
spots” for SERS enhancement and active sites for CO coupling
enhancement after anodization, suggesting that surface “hot-
are simultaneously created during nanoparticle formation.
Figure 1. (a) Potential pulsing procedure and related states of the Cu-MP electrode used in this figure. (b) Spectral heatmap from Time-Resolved Surface
Enhanced Raman Spectroscopy (TR-SERS) measurements of pristine Cu-MP during the first reduction step in the copper oxide region (Raman shift
between 400-700 cm–1), indicating rapid removal of surface oxide species, and (c) TR-SERS heatmap of pristine reduced Cu-MP in the carbonate region
(
Raman shift between 900-1200 cm–1), where no obvious bands are observed, indicating a poor SERS effect. (d) TR-SERS heatmap of anodized Cu-MP
in the carbonate region (between 900-1200 cm–1), displaying a strong carbonate vibration at 1060 cm–1, highlighting the strongly enhanced Raman signal
after anodic treatment and subsequent reduction. (e) Scanning Electronic Microscope (SEM, top) and Atomic Force Microscopy (AFM, bottom) images
of pristine Cu-MP and (f) anodized Cu-MP, showing the surface roughening and nanoparticle formation after anodic treatment. (g) Faradaic efficiency
(FE) of pristine and anodized Cu-MP during CO
2
RR at –1.0 V. The color bars of the heatmaps are based on photon counts of Raman spectra. Electrolyte:
flowing CO -saturated 0.1 M KHCO , pH = 6.8. Raman spectra interval: 717 ms.
2
3
In Situ TR-SERS at –0.7 V: CO Intermediates and Local
Alkalinity. It is generally accepted that adsorbed CO is a key
potentials (set at –0.7 V). The assignment of observed species is
depicted in Figure 2a.
intermediate in CO
2
RR towards hydrocarbon products on Cu
TR-SERS heatmaps of the stretching vibration of adsorbed CO
(in the 1900~2100 cm−1 region) on pristine and anodized Cu-MP
at −0.7 V are shown in Figure 2b and 2c, respectively. On pristine
Cu-MP, no SERS signal can be detected in the CO region (Figure
2b). For anodized Cu-MP, the CO Raman signal intensity
increases dramatically, which allows us to follow the dynamics of
the CO intermediates with sub-second time resolution (Figure 2c).
electrocatalyst surfaces.[3d,5] To investigate the reaction
mechanism, we performed in situ TR-SERS measurements to
investigate the dynamics of CO intermediates. The potential
pulsing procedure for the surface treatment and subsequent
reduction is the same as in Figure 1a, except for the reduction
3
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202104114
RESEARCH ARTICLE
By fitting the Raman spectra, we deconvolute the spectra of
anodized Cu-MP in the C=O stretching range into three peaks,
species from solution, which were created during the anodic
treatment.[22,23] Most of the re-deposition process is complete after
ca. 7 sec of reduction, which matches with the time scale of the
observed transition of the adsorbed CO species from LFB-
dominating to HFB-dominating (Figure 2b,c). TR-SERS results of
[
6b,6c,8a,20]
based on the previous work by Gunathunge et al.
(Figure
−
1
2
d): (1) bridged CO at ~2030 cm , (2) low-frequency band linear
−
1
CO (LFB-CO) at ~2060 cm , and (3) high-frequency band linear
−
1
−1
CO (HFB-CO) at ~2095 cm . The first species to be observed in
this region are two wide peaks centered at 2058 cm−1 (LFB-CO),
which appear within 2 s after the onset of cathodic bias. The LFB-
CO is typically associated with adsorbed CO on top of terrace-like
sites according to previous reports on well-defined facets.[6b,20]
This LFB-CO peak shifts towards lower Raman shifts over time
during the first ~7 s of cathodic bias onset. After this initial shift of
LFB-CO, a sharper peak at 2092 cm−1 becomes visible after 7 s,
which becomes the dominating species. The HFB-CO peak at
anodized Cu-MP in the oxide region (300-700 cm , Figure 2f)
show that after rapid oxide stripping within the first second, a new
−
1
peak at 520 cm emerges after ca. 7 sec of cathodic bias, which
is assigned to adsorbed Cu-OH at the electrode
surface.[11c,13c,16,24] This peak originates from the increase of local
alkalinity (and increase in the local hydroxide concentration),
which stems from the depletion of protons near the electrode
surface during HER.[16,18,25] These results suggest that during the
first 7 seconds after the onset of −0.7 V bias, the main reduction
2
092 cm−1 is ascribed to adsorbed CO on isolated defect-like sites, process is the re-deposition of Cu species leached into the
based on observations from previous in situ measurements on
well-defined systems.[21] The LFB-CO peak also shifts to 2072
cm−1 and remains as a weak shoulder next to the HFB-CO peak
after ca. 7 s. The peak positions and intensities of both HFB-CO
and LFB-CO remain stable after ~ 9 s of cathodic bias, up to 20
min of cathodic bias (Figure S10).
electrolyte, which results in the formation of nanostructures. After
7 sec, the HER and CO RR processes dominate, and the
2
increase of local alkalinity near the electrode surface results in the
accumulation of Cu-OH as major species. Furthermore, surface
hydroxide has also been reported to be a promoter for
CO
2
RR.[8b,26] These restructuring processes at the electrode
The chronoamperometry (CA) data of pristine and anodized Cu-
MP samples is shown in Figure 2e (complete 60 sec CA data in
Figure S11). In addition to the charging behavior of the
electrochemical double layer in both samples, an additional peak
is observed for anodized Cu-MP (Figure 2e), related to the
formation of nanostructures on the surface.[22] This additional
surface in turn changed the chemical state of the adsorbed CO
species, resulting in the observed dynamic nature of the bands
associated with adsorbed CO in the initial phase (first 7 s) of the
reaction (Figure 2c). Compared to IR, the ability of (TR-)SERS to
collect vibrational features at low wavenumbers (<800 cm )
allows us to correlate the re-deposition, local alkalinity and CO
intermediates in a time-resolved manner.
−
1
2
+
x y
peak is ascribed to re-deposition of dissolved Cu /CuO (OH)
4
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202104114
RESEARCH ARTICLE
Figure 2: (a) Surface species corresponding to the Raman signals observed in this figure. (b) TR-SERS heatmap of pristine Cu-MP during reduction at
0.7 V in the CO region (Raman shift between 1900-2150 cm−1). (c) TR-SERS heatmap of anodized Cu-MP during reduction at −0.7 V in the CO region
Raman shift between 1900-2150 cm−1), showing the dynamic behavior of adsorbed CO. (d) Fitted result of (c), allowing for deconvolution of the Raman
−
(
spectra into three bands: high-frequency band CO (HFB, blue bubbles), low-frequency band CO (LFB, orange bubbles) and bridged CO (yellow bubbles).
Bubble positions show time and Raman shift, and bubble sizes are proportional to peak intensities. The color bars of the heatmaps are based on the
photon counts of the Raman spectra. (e) Chronoamperometry (CA) curve of pristine and anodized Cu-MP during reduction at −0.7 V in the first 10 s (full
data is shown in Figure S11), showing the current associated with re-deposition of leached Cu after anodic treatment. (f) TR-SERS heatmap of anodized
Cu-MP during reduction at −0.7 V in the copper oxide and Cu-OH region (Raman shift between 300-700 cm−1), showing the formation of Cu-OH (peak at
5
20 cm−1) after 7 s of cathodic bias. Electrolyte: flowing CO
-saturated 0.1 M KHCO aqueous solution, pH = 6.8. Raman spectra interval: 717 ms.
2 3
In Situ TR-SERS beyond −0.8 V: Potential-Dependent
Dynamic CO and Product Selectivity. The product distribution
peak positions shown in Figure 3d) up to 20 min after cathodic
bias onset (Figure S13). Small variations in total signal intensity
are mostly caused by the rapid formation of bubbles due to
gaseous products at these overpotentials. However, the shift of
the LFB-CO peak position as a function of time suggests that it is
highly dynamic in nature and actively involved in surface reactions,
which is expected for CO intermediates that are involved in
complex reactions such as ethylene formation. The HFB-CO can
still be spotted in some observations at −0.9 V, but with weaker
intensity compared to the experiment at less cathodic biases, with
dynamic LFB-CO at ~2050 cm−1 being dominant (Figure S14).
of CO
2
RR on Cu is known to strongly depend on cathodic
potential.[
1a-1d, 3g]
To elucidate the relationship between adsorbed
RR selectivity, we performed TR-SERS on
CO species and CO
2
anodized Cu-MP at −0.8 V and −0.9 V (Figure 3). The potential
pulsing protocols are similar to Figure 1a, except for the more
cathodic reduction biases of −0.8 and −0.9 V. Steady-state in situ
Raman spectra (collected ca. 15 min after cathodic bias onset) at
all the three potentials (Figure 3a) show that the HFB-CO peak
positions are at slightly different positions for varying potentials
(
2093, 2090, and 2086 cm−1 at respectively −0.7, −0.8, and −0.9
2
The potential-dependent FE of anodized Cu-MP during CO RR is
V). The relative intensity of HFB-CO compared to LFB-CO
decreases at more cathodic potential, while an increase in relative
peak intensity of LFB-CO (compared to HFB-CO) is observed at
more cathodic potentials (Figure 3a). Meanwhile, the LFB-CO
peak also displays more drastic potential dependent shift to lower
Raman wavenumbers, from 2070 cm–1 at −0.7 V to ~2050 cm at
shown in Figure 3e. The clear trend that CO FE decreases (from
24.5% to 14.8%), while that of ethylene increases (from 2.2 to
16.0%) when the applied potential is varied from −0.7 V to −0.9 V
suggests that the dynamic LFB-CO species (at 2050 cm−1 Raman
shift at applied potential of −0.9 V) observed with our in situ TR-
SERS experiments are related to CO-CO coupling and ethylene
production, whereas the presence of the HFB-CO peak at 2092
cm−1 is more related to gaseous CO production.[3d,27] Since the
GC product analysis (injection every 4 min) is a lot slower than
our TR-SERS (sub-second collection of Raman spectra), we also
extended the TR-SERS time regime to 20 min (with a Raman
spectrum every 10 s) in order to cover the GC detection window,
and to analyze the stability or dynamics of the observed features
on those extended timescales (Figure S10 and S13). As
discussed above, we observe dynamic behavior of LFB-CO at
ethylene-dominating potentials of −0.9 V on the timescale of 20
min, whereas the HFB-CO SERS signal remains stable at −0.7 V
and −0.8 V over the course of the experiments. Furthermore, the
Cu-OH signal is also constant during 20 min of cathodic bias
(Figure S15), indicating that the surface of the electrode is stable
during our TR-SERS experiments. This further corroborates the
notion that the dynamic LFB-CO is related to electrocatalytic
reactions at the electrode surface at more cathodic bias, and not
to dynamics in the surface structure as a function of time.
–1
−
0.9 V. This is caused by the electrochemical Stark effect,[5e]
which results from the interaction between the applied electric
field and top-oriented adsorbed CO molecules, resulting in a
potential dependent shift.
During TR-SERS measurements at different applied potentials,
the formation and shift towards lower Raman shift of initial LFB-
CO at −0.8 V (Figure 3b and Figure S12) is similar to the scenario
at –0.7 V (see Figure 2c). Furthermore, the spectrum changes
from LFB-CO dominant to HFB-CO dominant after ca. 7 s, similar
to the behavior at −0.7 V. This further supports the notion that the
surface adsorbed CO is affected by local alkalinity and the re-
deposition of Cu nanoparticles in the first 7 s, also at −0.8 V. At
−0.9 V, LFB-CO exhibits a very dynamic peak shifting behavior,
as shown in Figure 3c, and the initially formed LFB-CO does not
transform into HFB-CO after ~ 7 s. Instead, it exhibits rapid peak
position shifting between 2040 and 2060 cm–1 (deconvoluted
5
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202104114
RESEARCH ARTICLE
Figure 3: (a) Comparison of steady-state Raman spectra (15 min after reduction) of anodized Cu-MP during reduction at −0.7 V, −0.8 V and −0.9 V.
Collection time is 5 s. (b) TR-SERS heatmap of anodized Cu-MP during reduction at −0.8 V in the CO region (Raman shift between 1900-2150 cm−1). After
7
(
s, an intense high-frequency band CO (HFB-CO) is observed. (c) TR-SERS heatmaps of anodized Cu-MP during reduction at −0.9 V in the CO region
Raman shift between 1900-2150 cm−1), showing that the low-frequency band CO (LFB-CO) is highly dynamic. (d) Fitted result of (c), showing the
increased contribution of LFB-CO to the spectrum at −0.9 V. Bubble position shows time and Raman shift, and bubble size is proportional to peak
intensity. (e) Faradaic efficiency (FE) of anodized Cu-MP during CO RR at −0.7 V, −0.8 V and −0.9 V. The colorbars of the heatmaps are based on photon
counts of Raman spectra. Electrolyte: CO -saturated 0.1 M KHCO with CO
2
2
3
2
bubbling, pH = 6.8. Raman spectra interval for TR-SERS: 717 ms.
The apparent differences in the time-dependent behavior of LFB-
CO at applied potentials between −0.7 V and −0.9 V imply that
the LFB-CO intermediates play a different, potential-dependent
gives rise to HFB-CO, and on step-edge sites to LFB-CO. Such
step-edge Cu sites shares more similarity to terrace sites than
isolated Cu sites, which explains the closer Raman peak position
of LFB-CO to terrace-CO (usually around 2050 cm ) [6b,20].
Meanwhile, CO on step-edge Cu sites has a higher chance of
coupling with a neighboring adsorbed CO intermediate compared
to isolated Cu sites, facilitating CO-CO coupling and C2+ product
formation.
-1
role in CO
2
RR. LFB-CO under ethylene-producing potentials of
−
0.9 V has a lower Raman shift than under CO-dominating
−
1
potentials of −0.7 V (2060 vs. 2070 cm , respectively). This lower
peak position suggests weakened C=O bonds, which would
facilitate subsequent CO-CO coupling steps. The significantly
higher tendency towards dynamic peak position shifting of the
LFB-CO at −0.9 V suggests more active chemical nature of LFB-
CO at ethylene-producing bias, revealing its relation to C-C
coupling. In contrast, the position of the HFB-CO Raman band
remains stable at all biases and over longer timescales, implying
it has a more static behavior. This leads to the conclusion that this
dynamic LFB-CO below 2060 cm−1 at −0.9 V is the active CO
We suggest a possible mechanism for improved Raman
enhancement and CO RR performance on anodized Cu-MP
2
based on our in situ TR-SERS measurements at different applied
potentials, as shown in Figure 4. Pristine Cu-MP is not very SERS
active due to the presence of surface oxide and its lack of
nanostructures (Figure 4a), but anodic treatment at 1.55 V vs.
RHE (Figure 4b) and subsequent reduction (Figure 4c) results in
intermediate for C
2
product formation, and cannot be assigned to
bridged CO, which is typically observed in an even lower Raman
-1
shift region (<2020 cm ). The improved FE towards ethylene after
anodization-reduction cycle suggests that CO-CO coupling
related by defect formation. We propose that two different defects
can be produced after the anodization-reduction cycle, namely
isolated defect sites and step-edge sites. CO on isolated sites
2
a highly active surface for both SERS and CO RR (two-fold
increase in ethylene FE at −0.9 V) due to nanoparticle formation.
Furthermore, our results show that the native surface oxide
species of anodized Cu-MP can be stripped within 1 s after
cathodic bias onset. In the first 7 s after cathodic bias onset, the
6
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202104114
RESEARCH ARTICLE
2
+
main process is the re-deposition of leached Cu /CuO
species, creating nanostructures and hot-spots for both Raman
enhancement and CO RR reaction. After 7 s, the local alkalinity
x
(OH)
y
reduction cycle simultaneously facilitates the formation of defect
sites, exhibits increased CO SERS signal intensity and promotes
ethylene production. Therefore, we assign this more active and
2
−
1
near the electrode surface starts to build up when the cathodic
bias is −0.7 V or beyond. At −0.7 V (Figure 4d), the dimerization
of the initial 2058 cm−1 terrace LFB-CO cannot be efficiently
triggered, and the LFB-CO band evolves into a stable HFB-CO on
isolated undercoordinated defect-like sites (Raman peak at ~2092
cm ). These isolated CO intermediates then desorb as gaseous
CO, which is the main CO reduction product at −0.7 V. This
suggests that an isolated defect Cu site stabilizing two adsorbed
CO molecules with enough proximity to induce CO-CO coupling
is unlikely. At high cathodic bias (−0.9 V), further reaction of
activated LFB-CO (Raman peak at 2058 cm ) can be triggered,
suppressing its conversion toward HFB-CO and subsequent CO
production (Figure 4e). It has been found that ethylene formation
is facilitated on high-index Cu facets rich in step-edge sites[3b,4g]
Compared to flat pristine Cu-MP electrode, our anodization-
dynamic species at or below 2060 cm to CO on step-edge Cu
sites with higher tendency towards CO-CO coupling, since the
coupling between terrace CO molecules would further weaken the
CO bond, resulting in the observed lower Raman shift compared
−1
to the LFB-CO at −0.7 V (at 2070 cm ). This assignment is further
corroborated by the more dynamic time-dependent shifting of this
CO band (even up to 20 min after cathodic bias onset), which
reveals that this intermediate is heavily involved in chemical
reactions and facilitates CO-CO coupling towards ethylene
formation (as evidenced by the activity measurements at −0.9 V).
This dynamic information of surface reconstruction and surface-
bound intermediates on the sub-second timescale is often
obscured in steady-state measurements, highlighting the
importance of time-resolved investigations under catalytically
relevant conditions.
−
1
2
−
1
.
2
Figure 4. Schematic representation of the observed processes on Cu-MP during CO RR using in situ TR-SERS. (a) The pristine electrode is rather smooth
and contains surface oxide species, but (b) anodic treatment (at 1.55 V) roughens the electrode surface and results in leached Cu-species in the
electrolyte. (c) Re-deposition of these dissolved Cu-species from the electrolyte at potentials <-0.7 V creates nanoparticles at the electrode surface,
which results in hot-spot formation and an increase in SERS activity. This is followed by the build-up of local alkalinity. Subsequent CO
results in CO adsorption on undercoordinated defect-like sites, and formation of gaseous CO as main product. (e) Reduction at a higher cathodic bias
-0.9 V) stabilizes CO at terrace-sites, thereby inducing C-C coupling and ethylene formation.
2
RR at (d) -0.7 V
(
edges) dominates at −0.7 V vs. RHE, characterized by a static
−
1
Conclusion
Raman band at ~ 2092 cm , and we correlate this more static CO
to gaseous CO product desorption. An activated CO intermediate,
−
1
We have successfully performed in situ Time-Resolved Surface
Enhanced Raman Spectroscopy (TR-SERS) during CO
characterized by a dynamic Raman band at < 2060 cm ,
dominates at −0.9 V vs. RHE, whose dynamic potential- and time-
dependent behavior suggests its tendency towards dimerization
and the formation of ethylene at more cathodic bias. Our results
demonstrate that in situ TR-SERS with sub-second time
resolution is a vital technique for achieving dynamic information
2
electrochemical reduction on Cu electrodes with sub-second time
resolution. Anodic pretreatment and subsequent reduction of the
Cu electrodes create nanostructures via re-deposition of
dissolved Cu-species, which act as “hot-spots” for SERS as well
as active sites for CO coupling at high overpotential (−0.9 V vs.
RHE). These nanostructures allow for TR-SERS measurements
with sub-second time resolution and a four-fold improvement in
2
of surface reactions during CO electrolysis.
ethylene production (up to 23% FE at −1.0 V vs. RHE). Our TR- Acknowledgements
SERS measurements reveal that surface reconstruction and
nanostructure formation happen within the first 7 s after cathodic
bias onset, after which an increase in local alkalinity creates a
stable Cu-OH electrode surface. CO adsorbed on isolated
undercoordinated defect sites (such as nanoparticle corners and
This work was supported by the Strategic Alliance between
Utrecht University, University Medical Center Utrecht and
Technical University Eindhoven. B.M.W. acknowledges funding
from the Netherlands Center for Multiscale Catalytic Energy
7
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202104114
RESEARCH ARTICLE
ACS Catal. 2018, 8, 7507–7516. c) Gunathunge, C. M.; Li, J.; Li, X.;
Hong, J. J.; Waegele, M. M. ACS Catal. 2020, 10, 6908–6923.
[7] a) Todorova, T. K.; Schreiber, M. W.; Fontecave, M. ACS Catal. 2020, 10,
Conversion (MCEC), an NWO Gravitation program funded by the
Ministry of Education, Culture and Science of the government of
the Netherlands.
1754–1768. b) Li, F.; Thevenon, A.; Rosas-Hernández, A.; Wang, Z.; Li,
Y.; Gabardo, C. M.; Ozden, A.; Dinh, C. T.; Li, J.; Wang, Y.; Edwards, J.
P.; Xu, Y.; McCallum, C.; Tao, L.; Liang, Z. Q.; Luo, M.; Wang, X.; Li, H.;
O’Brien, C. P.; Tan, C. S.; Nam, D. H.; Quintero-Bermudez, R.; Zhuang,
T. T.; Li, Y. C.; Han, Z.; Britt, R. D.; Sinton, D.; Agapie, T.; Peters, J. C.;
Sargent, E. H. Nature 2020, 577, 509–513. c) Huang, M. H.; Maljusch, A.;
Calle-Vallejo, F.; Henry, J. B.; Koper, M. T. M.; Schuhmann, W.;
Bandarenka, A. S. RSC Adv. 2013, 3, 21648–21654.
We would like to thank all the students and staff who have strived
to keep our lab running in this special period of our lives.
[
8] a) Gunathunge, C. M.; Ovalle, V. J.; Waegele, M. M. Phys. Chem. Chem.
Phys. 2017, 19, 30166–30172. b) Iijima, G.; Inomata, T.; Yamaguchi, H.;
Ito, M.; Masuda, H. ACS Catal. 2019, 9, 6305–6319. c) Zhu, S.; Jiang, B.;
Cai, W. B.; Shao, M. J. Am. Chem. Soc. 2017, 139, 15664–15667. d)
Sheng, H.; Oh, M. H.; Osowiecki, W. T.; Kim, W.; Alivisatos, A. P.; Frei, H.
J. Am. Chem. Soc. 2018, 140, 4363–4371.
2
Keywords: Electrocatalysis• Raman • in situ • CO reduction •
Spectroscopy
[
1] a) Zheng, Y.; Vasileff, A.; Zhou, X. L.; Jiao, Y.; Jaroniec, M.; Qiao, S. Z. J.
Am. Chem. Soc. 2019, 141, 7646–7659. b) Gao, D. F.; Aran-Ais, R. M.;
Jeon, H. S.; Roldan Cuenya, B. Nat. Catal. 2019, 2, 198–210. c) García
de Arquer, F. P.; Dinh, C.-T.; Ozden, A.; Wicks, J.; McCallum, C.; Kirmani,
A. R.; Nam, D.-H.; Gabardo, C.; Seifitokaldani, A.; Wang, X.; Li, Y. C.; Li,
F.; Edwards, J.; Richter, L. J.; Thorpe, S. J.; Sinton, D.; Sargent, E. H.
Science. 2020, 367, 661–666. d) Franco, F.; Rettenmaier, C.; Jeon, H. S.;
Roldan Cuenya, B. Chem. Soc. Rev. 2020, 49, 6884–6946. e) Kortlever,
R.; Shen, J.; Schouten, K. J. P.; Calle-Vallejo, F.; Koper, M. T. M. J. Phys.
Chem. Lett. 2015, 6, 4073–4082. f) Chen, C.; Li, Y.; Yu, S.; Louisia, S.;
Jin, J.; Li, M.; Ross, M. B.; Yang, P. Joule 2020, 4, 1688–1699. g)
Morales-Guio, C. G.; Cave, E. R.; Nitopi, S. A.; Feaster, J. T.; Wang, L.;
Kuhl, K. P.; Jackson, A.; Johnson, N. C.; Abram, D. N.; Hatsukade, T.;
Hahn, C.; Jaramillo, T. F. Nat. Catal. 2018, 1, 764–771. h) Smith, W. A.;
Burdyny, T.; Vermaas, D. A.; Geerlings, H. Joule. 2019, 3, 1822–1834.
2] a) Birdja, Y. Y.; Pérez-Gallent, E.; Figueiredo, M. C.; Göttle, A. J.; Calle-
Vallejo, F.; Koper, M. T. M. Nat. Energy. 2019, 4, 732–745. b) Wagner, A.;
Sahm, C. D.; Reisner, E. Nat. Catal. 2020, 3, 775–786.
3] a) Arán-Ais, R. M.; Rizo, R.; Grosse, P.; Algara-Siller, G.; Dembélé, K.;
Plodinec, M.; Lunkenbein, T.; Chee, S. W.; Roldan Cuenya, B. Nat.
Commun. 2020, 11, 3489. b) Simon, G. H.; Kley, C. S.; Roldan Cuenya,
B. Angew. Chemie Int. Ed. 2020, 59, 2-10. c) Huang, J.; Hörmann, N.;
Oveisi, E.; Loiudice, A.; De Gregorio, G. L.; Andreussi, O.; Marzari, N.;
Buonsanti, R. Nat. Commun. 2018, 9, 3117. d) Lum, Y.; Ager, J. W. Nat.
Catal. 2019, 2, 86–93. e) Wang, L.; Nitopi, S. A.; Bertheussen, E.;
Orazov, M.; Morales-Guio, C. G.; Liu, X. Y.; Higgins, D. C.; Chan, K. R.;
Norskov, J. K.; Hahn, C.; Jaramillo, T. F. ACS Catal. 2018, 8, 7445–7454.
f) Clark, E. L.; Resasco, J.; Landers, A.; Lin, J.; Chung, L.-T.; Walton, A.;
Hahn, C.; Jaramillo, T. F.; Bell, A. T. ACS Catal. 2018, 8, 6560–6570. g)
Zhang, Y.; Guo, S.-X.; Zhang, X.; Bond, A. M.; Zhang, J. Nano Today
[9] a) Hartman, T.; Wondergem, C. S.; Kumar, N.; van den Berg, A.;
Weckhuysen, B. M. J. Phys. Chem. Lett. 2016, 7, 1570–1584. b)
Hartman, T.; Geitenbeek, R. G.; Wondergem, C. S.; van der Stam, W.;
Weckhuysen, B. M. ACS Nano 2020, 14, 3725–3735. c) Kas, R.;
Ayemoba, O.; Firet, N. J.; Middelkoop, J.; Smith, W. A.; Cuesta, A.
ChemPhysChem 2019, 20, 2904–2925. d) Heidary, N.; Ly, K. H.;
Kornienko, N. Nano Lett. 2019, 19, 4817–4826. e) Gao, J.; Zhang, H.;
Guo, X.; Luo, J.; Zakeeruddin, S. M.; Ren, D.; Grätzel, M. J. Am. Chem.
Soc. 2019, 141, 18704–18714. f) Ren, D.; Ang, B. S. H.; Yeo, B. S. ACS
Catal. 2016, 6, 8239–8247.
[10] a) Yang, P.-P.; Zhang, X.-L.; Gao, F.-Y.; Zheng, Y.-R.; Niu, Z.-Z.; Yu, X.;
Liu, R.; Wu, Z.-Z.; Qin, S.; Chi, L.-P.; Duan, Y.; Ma, T.; Zheng, X.-S.; Zhu,
J.-F.; Wang, H.-J.; Gao, M.-R.; Yu, S.-H. J. Am. Chem. Soc. 2020, 142,
6400–6408. b) Jiang, S.; Klingan, K.; Pasquini, C.; Dau, H. J. Chem.
Phys. 2019, 150, 041718. c) Vasileff, A.; Zhi, X.; Xu, C.; Ge, L.; Jiao, Y.;
Zheng, Y.; Qiao, S. Z. ACS Catal. 2019, 9, 9411–9417.
[11] a) Stavitski, E.; Weckhuysen, B. M. Chem. Soc. Rev. 2010, 39, 4615–
4625. b) Weckhuysen, B. M. Angew. Chemie Int. Ed. 2009, 48, 4910–
4943. c) Mul, G.; Moradzaman, M. ChemElectroChem 2021,
doi:celc.202001598.
[12] a) Pilot, R.; Signorini, R.; Durante, C.; Orian, L.; Bhamidipati, M.; Fabris,
L. Biosensors. 2019, 9, 57.
[13] a) Hartman, T.; Geitenbeek, R. G.; Whiting, G. T.; Weckhuysen, B. M.
Nat. Catal. 2019, 2, 986–996. b) Zhang, H.; Duan, S.; Radjenovic, P. M.;
Tian, Z.-Q.; Li, J.-F. Acc. Chem. Res. 2020, 53, 729–739. c) Bodappa, N.;
Su, M.; Zhao, Y.; Le, J. B.; Yang, W. M.; Radjenovic, P.; Dong, J. C.;
Cheng, J.; Tian, Z. Q.; Li, J. F. J. Am. Chem. Soc. 2019, 141, 12192–
12196. d) Shan, W. Y.; Liu, R.; Zhao, H. C.; He, Z. L.; Lai, Y. J.; Li, S. S.;
He, G. Z.; Liu, J. F. ACS Nano 2020, 14, 11363–11372. e) Wang, Y.-H.
H.; Wei, J.; Radjenovic, P.; Tian, Z.-Q. Q.; Li, J.-F. Anal. Chem. 2019, 91,
1675–1685. f) Batista, E. A.; Temperini, M. L. A. J. Electroanal. Chem.
2009, 629, 158–163.
[
[
2020, 31, 100835. h) Vogt, C.; Monai, M.; Kramer, G. J.; Weckhuysen, B.
M. Nat. Catal. 2019, 2, 188–197. i) Wu, L.; Kolmeijer, K. E.; Zhang, Y.;
An. H.; Arnouts, S.; Bals, S.; Altantzis, T.; Hofmann, J. P.; Costa
Figueiredo, M.; Hensen, E. J. M.; Weckhuysen, B. M.; van der Stam, W.
Nanoscale 2021, 13, 4835–4844. j) Bagger, A.; Ju, W; Varela, A. S.
Strasser, P.; Rossmeisl, J. ACS Catal. 2019, 9, 7894−7899
[14] a) Engstfeld, A. K.; Maagaard, T.; Horch, S.; Chorkendorff, I.; Stephens, I.
E. L. Chem. Eur. J. 2018, 24, 17743–17755.
[15] Chen, R.; Yang, C.; Cai, W.; Wang, H.-Y.; Miao, J.; Zhang, L.; Chen, S.;
Liu, B. ACS Energy Lett. 2017, 2, 1070–1075.
[16] Zhao, Y. R.; Chang, X. Z.; Malkani, A. S.; Yang, X.; Thompson, L.; Jiao,
F.; Xu, B. J. J. Am. Chem. Soc. 2020, 142, 9735–9743.
[17] a) Velasco-Velez, J. J.; Velasco-Velez, J. J.; Mom, R. V.; Sandoval-Diaz,
L. E.; Falling, L. J.; Chuang, C. H.; Gao, D.; Jones, T. E.; Zhu, Q.; Arrigo,
R.; Roldan Cuenya, B.; Knop-Gericke, A.; Lunkenbein, T.; Schlögl, R.
ACS Energy Lett. 2020, 5, 2106–2111 b) Burke, L. D.; Collins, J. A.;
Murphy, M. A. J. Solid State Electrochem. 1999, 4, 34–41. c) Kortlever,
R.; Tan, K. H.; Kwon, Y.; Koper, M. T. M. J. Solid State Electrochem.
[
4] a) Arán-Ais, R. M.; Scholten, F.; Kunze, S.; Rizo, R.; Roldan Cuenya, B.
Nat. Energy 2020, 5, 317–325. b) Dattila, F.; Garclá-Muelas, R.; López, N.
ACS Energy Lett. 2020, 5, 3176–3184. c) Chou, T. C.; Chang, C. C.; Yu,
H. L.; Yu, W. Y.; Dong, C. L.; Velasco-Vélez, J. J.; Chuang, C. H.; Chen,
L. C.; Lee, J. F.; Chen, J. M.; Wu, H. L. J. Am. Chem. Soc. 2020, 142,
2857–2867. d) Ting, L. R. L.; García‐Muelas, R.; Martín, A. J.; Veenstra,
F. L. P.; Chen, S. T.; Peng, Y.; Per, E. Y. X.; Pablo‐García, S.; López,
N.; Pérez‐Ramírez, J.; Yeo, B. S. Angew. Chemie Int. Ed. 2020, 132,
21258–21265. e) Phan, T. H.; Banjac, K.; Cometto, F. P.; Dattila, F.;
2013, 17, 1843–1849.
García-Muelas, R.; Raaijman, S. J.; Ye, C.; Koper, M. T. M.; Lopez, N.;
Lingenfelder, M. Nano Lett. 2021, 21, 2059-2065. f) Kim, C.; Weng, L.-C.;
Bell, A. T. ACS Catal. 2020, 10, 12403–12413. g) Hori, Y.; Takahashi, I.;
Koga, O.; Hoshi, N. J. Mol. Catal. A: Chem. 2003, 199, 39−47. h) Bui, J.
C.; Kim, C.; Weber, A. Z.; Bell A. E. ACS Energy Lett. 2021, 6,
[
[
[
18] Zhang, Z. S.; Melo, L.; Jansonius, R. P.; Habibzadeh, F.; Grant, E. R.;
Berlinguette, C. P. ACS Energy Lett. 2020, 5, 3101–3107.
19] Jiang, K.; Huang, Y.; Zeng, G.; Toma, F. M.; Goddard III, W. A.; Bell, A. T.
ACS Energy Lett. 2020, 5, 1206–1214.
20] a) Gunathunge, C. M.; Li, X.; Li, J. Y.; Hicks, R. P.; Ovalle, V. J.;
Waegele, M. M. J. Phys. Chem. C 2017, 121, 12337–12344. b) Hollins,
P.; Pritchard, J. Surf. Sci. 1983, 134, 91–108.
1181−1188
[
5] a) Nitopi, S.; Bertheussen, E.; Scott, S. B.; Liu, X. Y.; Engstfeld, A. K.;
Horch, S.; Seger, B.; Stephens, I. E. L.; Chan, K.; Hahn, C.; Norskov, J.
K.; Jaramillo, T. F.; Chorkendorff, I. Chem Rev 2019, 119, 7610–7672. b)
Jeon, H. S.; Timoshenko, J.; Scholten, F.; Sinev, I.; Herzog, A.; Haase, F.
T.; Roldan Cuenya, B.; Roldan Cuenya, B. J. Am. Chem. Soc. 2019, 141,
[
[
21] Hollins, P. Surf. Sci. Rep. 1992, 16, 51–94.
22] Vvedenskii, A.; Grushevskaya, S.; Ganzha, S.; Eliseev, D.; Abakumova,
L. I. J. Solid State Electrochem. 2014, 18, 3437–3451.
[
[
[
[
23] Vavra, J.; Shen, T. H.; Stoian, D.; Tileli, V.; Buonsanti, R. Angew. Chem.,
19879–19887. c) Mandal, L.; Yang, K. R.; Motapothula, M. R.; Ren, D.;
2021, 60, 1347–1354.
Lobaccaro, P.; Patra, A.; Sherburne, M.; Batista, V. S.; Yeo, B. S.; Ager,
J. W.; Martin, J.; Venkatesan, T. ACS Appl. Mater. Interfaces 2018, 10 ,
24] Chen, X. Y.; Henckel, D. A.; Nwabara, U. O.; Li, Y. Y.; Frenkel, A. I.;
Fister, T. T.; Kenis, P. J. A.; Gewirth, A. A. ACS Catal. 2020, 10, 672–682.
25] Yang, K.; Kas, R.; Smith, W. A. J. Am. Chem. Soc. 2019, 141, 15891–
8574–8584. d) Malkani, A. S.; Dunwell, M.; Xuo, B. J. ACS Catal. 2019, 9,
474–478. e) Chernyshova, I. V; Somasundaran, P.; Ponnurangam, S.
15900.
Proc. Natl. Acad. Sci. 2018, 115, E9261–E9270. f) Perez-Gallent, E.;
Figueiredo, M. C.; Calle-Vallejo, F.; Koper, M. T. M. Angew. Chem. Int.
Ed. 2017, 56, 3621–3624. g) Li, Y. C.; Wang, Z.; Yuan, T.; Nam, D.-H.;
Luo, M.; Wicks, J.; Chen, B.; Li, J.; Li, F.; de Arquer, F. P. G.; Wang, Y.;
Dinh, C.-T.; Voznyy, O.; Sinton, D.; Sargent, E. H. J. Am. Chem. Soc.
26] Luo, M.; Wang, Z.; Li, Y. C.; Li, J.; Li, F.; Lum, Y.; Nam, D.-H.; Chen, B.;
Wicks, J.; Xu, A.; Zhuang, T.; Leow, W. R.; Wang, X.; Dinh, C.-T.; Wang,
Y.; Wang, Y.; Sinton, D.; Sargent, E. H. Nat. Commun. 2019, 10, 5814.
27] Kuhl, K. P.; Cave, E. R.; Abram, D. N.; Jaramillo, T. F. Energy Environ.
Sci. 2012, 5, 7050–7059.
[
2019, 141, 8584–8591.
[
6] a) Malkani, A. S.; Dunwell, M.; Xu, B. ACS Catal. 2018,9, 474–478. b)
Gunathunge, C. M.; Ovalle, V. J.; Li, Y. W.; Janik, M. J.; Waegele, M. M.
8
This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition
10.1002/anie.202104114
RESEARCH ARTICLE
Entry for the Table of Contents
Time-Resolved Surface Enhanced Raman Spectroscopy (TR-SERS) can monitor CO
2
RR over Cu surfaces with sub-second time
resolution of 717 ms per spectrum. Anodic treatment and subsequent reduction result in a highly active surface for both SERS and
ethylene production within 7 s. A dynamic CO intermediate below 2060 cm−1 is found to correlate with the CO-CO coupling and
ethylene formation, whereas a static CO intermediate with a distinct vibration at 2092 cm−1 is related to CO production.
9
This article is protected by copyright. All rights reserved.