A Precious-Metal-Free Hybrid Electrolyzer for Alcohol Oxidation Coupled to CO2-to-Syngas Conversion

Article abstract of DOI:10.1002/anie.202002680

Electrolyzers combining CO₂ reduction (CO₂R) with organic substrate oxidation can produce fuel and chemical feedstocks with a relatively low energy requirement when compared to systems that source electrons from water oxidation. Here, we report an anodic hybrid assembly based on a (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO) electrocatalyst modified with a silatrane-anchor (STEMPO), which is covalently immobilized on a mesoporous indium tin oxide (mesoITO) scaffold for efficient alcohol oxidation (AlcOx). This molecular anode was subsequently combined with a cathode consisting of a polymeric cobalt phthalocyanine on carbon nanotubes to construct a hybrid, precious-metal-free coupled AlcOx–CO₂R electrolyzer. After three-hour electrolysis, glycerol is selectively oxidized to glyceraldehyde with a turnover number (TON) of ≈1000 and Faradaic efficiency (FE) of 83 %. The cathode generated a stoichiometric amount of syngas with a CO:H₂ ratio of 1.25±0.25 and an overall cobalt-based TON of 894 with a FE of 82 %. This prototype device inspires the design and implementation of nonconventional strategies for coupling CO₂R to less energy demanding, and value-added, oxidative chemistry.

Full text of DOI:10.1002/anie.202002680

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: A Precious-metal-free Hybrid Electrolyzer for Alcohol Oxidation
Coupled to CO2-to-syngas Conversion
Authors: Mark Bajada, Souvik Roy, Julien Warnan, Kaltum Abdiaziz,
Andrewas Wagner, Maxie Roessler, and Erwin Reisner
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.202002680
Link to VoR: https://doi.org/10.1002/anie.202002680

10.1002/anie.202002680
Angewandte Chemie International Edition
A Precious-metal-free Hybrid Electrolyzer for Alcohol
Oxidation Coupled to CO2-to-syngas Conversion
†
†
†
Mark A. Bajada1, , Souvik Roy1, , Julien Warnan1, , Kaltum Abdiaziz2,3, Andreas
Wagner1, Maxie M. Roessler2, and Erwin Reisner1*
1 Department of Chemistry, University of Cambridge, Cambridge CB2 1EW, UK
2 Department of Chemistry, Imperial College London, Molecular Sciences Research
Hub, White City Campus, London W12 0BZ, UK
3 School of Biological and Chemical Sciences and Materials Research Institute, Queen
Mary University of London, London E1 4NS, UK
*e-mail: reisner@ch.cam.ac.uk
† These authors contributed equally
1
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
Abstract
Electrolyzers combining CO2 reduction (CO2R) with organic substrate oxidation can
produce fuel and chemical feedstocks with a relatively low energy requirement when
compared to systems that source electrons from water oxidation. Here, we report an
anodic hybrid assembly based on a (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO)
electrocatalyst modified with a silatrane-anchor (STEMPO), which is covalently
immobilized on a mesoporous indium tin oxide (mesoITO) scaffold for efficient alcohol
oxidation (AlcOx). This molecular anode was subsequently combined with a cathode
consisting of a polymeric cobalt phthalocyanine on carbon-nanotubes to construct a
hybrid, precious-metal-free coupled AlcOx–CO2R electrolyzer. After three hour
electrolysis, glycerol is selectively oxidized to glyceraldehyde with a turnover number
(TON) of 1000 and Faradaic efficiency (FE) of 83% and the cathode generated a
stoichiometric amount of syngas with a CO:H2 ratio of 1.25±0.25 and an overall cobalt-
based TON of 894 with a FE of 82%. This prototype device inspires the design and
implementation of nonconventional strategies for coupling CO2R to less energy
demanding, and value-added, oxidative chemistry.
2
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
Introduction
The electrosynthesis of fuels is being pursued as a potential solution to intermittent
electricity production via renewable wind and solar technologies.[1] Conventional
electrolyzers couple the oxygen evolution reaction (OER) at the anode to a hydrogen
evolution reaction (HER) or CO2R at the cathode.[2,3] However, the kinetic hurdles of
the anodic four-electron process and consequently large overpotential for the OER,
tied to the limited commercial value of O2, are spurring interest in employing more
synthetically useful and facile organic oxidative reactions.[4–11]
Recent technoeconomic studies have shown that 90% of the overall energy
requirements for current commercial approaches in CO2 electrolysis stem from the
OER, and that lower cell potentials for fuel-generating reductive chemistry can be
achieved by substituting the OER for AlcOx.[12] In particular, by combining theory and
experiment, it was shown that glycerol, a biomass-derived platform chemical and a by-
product from the production of biodiesel and soap,[13,14] is an attractive candidate
which can greatly improve the economics of the overall redox process.
Reports on tandem AlcOx–CO2R electrolyzers suffer from two main drawbacks to
date.[7,9] Firstly, precious metal-containing components are employed in the
electrolysis cells. Secondly, homogeneous catalysts and mediators are required in
excess in the solution, which complicates post-reaction processing of the liquid
products. The use of catalysts in solution presents an additional challenge during
electrosynthesis in that only a tiny fraction of the catalyst (positioned in the Helmholtz
layer) can turnover, while the rest remains inactive in the bulk.
Here, we consider glycerol as a commercially viable resource to develop a robust and
precious-metal-free anodic assembly for coupling with the cathodic CO2R reaction.
3
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
We modified TEMPO, an efficient, non-toxic, molecular (electro)catalyst that can
oxidize a wide range of alcohol substrates under mild conditions,[15,16] with a silatrane
anchor (giving STEMPO), for robust immobilization on a mesoITO scaffold (Figure 1).
The porous electrode enables high catalyst loading, while the immobilization
procedure permits direct electronic communication between the electrode surface and
the molecular species, leading to constant catalytic turnover, easier product isolation
and catalyst recyclability. The performance of the mesoITO|STEMPO assembly was
assessed under different conditions with a range of alcohols.
We then coupled the anode to a hybrid CO2R cathode, featuring a cobalt
phthalocyanine (CoPc) based electrocatalyst. CoPc was polymerized onto carbon
nanotubes (CNT) to form a CNT-polymeric CoPc composite (CNT-CoPPc; where
CoPPc denotes polymeric cobalt phthalocyanine), and then deposited onto porous
carbon paper (CP, cathode assembly henceforth denoted as CP|CNT-CoPPc)[17]
(Figure 1). The AlcOx–CO2R electrolyzer oxidizes glycerol to glyceraldehyde with
good catalytic performance and high selectivity, and generates a stoichiometric
amount of syngas at the cathode.
4
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
Results and Discussion
Metal oxide (MOx) electrodes present an interesting platform for catalyst
immobilization and offer a number of advantages such as affinity for a variety of
anchoring units, and the possibility to morphologically tune the surface to enhance the
loading of molecular components.[18,19] MOxs can exhibit different electronic properties,
as demonstrated by the metallic behavior of ITO and the semiconductive properties of
TiO2, thus offering a versatile electroactive platform to combine with surface-anchored
molecular catalysts.[20,21] Several mechanistic studies have highlighted the effect of pH
on the TEMPO catalytic cycle, with enhanced oxidation rates recorded under more
basic conditions.[22–24] This stringent criterion implies that some of the more commonly
used anchoring groups compatible with MOx scaffolds, such as carboxylic acids and
phosphonic acids (pH stability < 4 and 7, respectively),[25] may not be suitable for
TEMPO immobilization on an ITO electrode. We therefore designed STEMPO, where
a silatrane unit offers increased linker stability under more alkaline conditions,[26,27]
a
result of the strong siloxane bonds being formed upon hydrolysis of the caged moiety
on the MOx surface (Figure 1).
STEMPO was synthesized in good yield by reacting the acyl chloride of 4-carboxy-
TEMPO with 3-aminopropylsilatrane. Full synthetic and characterization details (high-
resolution mass spectrometry, infrared spectroscopy (Figure S1) and elemental
analysis) are provided in the Supporting Information.
The mesoITO|STEMPO anode was assembled by incubating the mesoITO electrode
(film thickness 4.5 μm, Figure S2) in a solvent bath mixture, and heating the solution
to 70 °C under a N2 atmosphere for a duration of 6 h. Multi-scan cyclic voltammetry
(CV) measurements were used to deduce the optimal mixture, in which the surface
5
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
loading of STEMPO (STEMPO) on the mesoITO scaffold was both maximal and stable,
with STEMPO being determined by integrating the charge passed in the oxidation wave
of the consecutive cyclic voltammograms (see Supporting Information, equation (S1)).
The best mixture consisted of a STEMPO solution (2 mM) with 2%(v/v) acetic acid
(AcOH) and 1%(v/v) H2O in acetonitrile (MeCN). With regards to stability of the
immobilized STEMPO compound, MeCN was the most suitable solvent from those
attempted (Figure S3). The combination of AcOH and H2O facilitated the hydrolysis of
the silatrane cage on the mesoITO surface,[28] and was deemed necessary for
instigating the anchoring process (Figure S4 and S5). Optimal STEMPO was thus
typically found to be 20–25 nmol cm−2, which is in the expected range for
nanostructured ITO surfaces.[29,30]
X-ray photoelectron spectroscopy (XPS) showed binding signals in the Si2p and N1s
regions (Figure 2a and Figure S6, respectively), where the Si2p signal agrees with XPS
reference spectra for the siloxane-bearing group.[31]
Multiple CV scans of the mesoITO|STEMPO electrode reveals a reversible redox
wave at E1/2 = 0.83 V vs. NHE (Figure 2b), which corresponds to the
nitroxide/oxoammonium species, and is only slightly more positive than that of TEMPO
(E1/2 = 0.74 V vs. NHE, Figure S7a). At low scan rate (10 mV s−1), the peak-to-peak
separation is below 20 mV and is thus in good agreement with the ideal value of 0 mV
for a reversible response of a surface-adsorbed species (Figure S7b). The full width
at half-maximum is 116 mV (Figure S7b), only slightly broader than the theoretically
predicted value of around 91 mV for a 1 e− process (at 25 °C).[32] This slight deviation
from ideal behavior can be attributed to multilayer formation,[33,34] stemming from the
6
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
cross-polymerization of Si-O-Si bonds between adjacent anchoring units in the
mesoporous scaffold and film resistance of the mesoITO electrode.
A deeper analysis of the electron-transfer dynamics of the mesoITO|STEMPO system
was inferred using the Laviron method,[35] which relies on the change in the peak
potential (Ep) with scan rate (). The resulting trumpet plot for the mesoITO|STEMPO
assembly is portrayed in Figure 2c. The intercepts of the linear regions of the plot can
be used to deduce the critical scan rate (c) and the apparent electron transfer rate
constant (kapp) for the system (see Supporting Information for further details). Values
for c and kapp were determined to be equal to 72 ± 2 mV s−1 and 0.68 ± 0.02 s−1,
respectively. The rate of electron transfer appears to be slow (hence the low value for
c), but is comparable with other covalently linked redox species in the literature.[36]
Figure S8 depicts the CV scans measured over a range of scan rates to highlight the
change in the peak-to-peak separation for the STEMPO redox wave as the applied
scan rate exceeds c. The linear relationship between the peak current (ip) and , for
 < c (Figure 2d) is characteristic for a surface-immobilized redox entity.[32] The pH
stability of the mesoITO|STEMPO assembly was investigated using a multi-scan CV
approach, whereby the electrode was subjected to several redox cycles in solutions
of differing pH (Figure 2e and Figure S9). Very good stability was obtained after 200
scans at pH 7 and 8 (decrease in signal intensity of 34% and 39%, respectively,
relative to scan 1), and the decay curve only began to be more severe at pH 10. These
results support that the assembly is suited to operate under the basic conditions
required for enhanced TEMPO catalysis.
Immobilization and direct wiring of STEMPO to the mesoITO electrode was confirmed
by film electrochemical electron paramagnetic resonance (FE-EPR) spectroscopy
7
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
(Supporting Information).[37] The combined FE-EPR spectroelectrochemical technique
allows for the appearance and disappearance of paramagnetic species to be
monitored as a function of the applied potential in the absence of any mediators. The
high electrical conductivity combined with the surface-modification properties of ITO
make it a suitable platform for carrying out FE-EPR spectroscopy. Carbon-based
electrodes tend to give rise to large radical signals and are thus unsuitable for such
studies.[37]
For FE-EPR spectroscopy, cylindrical mesoITO (C-mesoITO) electrodes were
employed for use in the EPR spectroelectrochemical cell. The unpaired electron in the
TEMPO moiety is delocalized around the N and O atoms with nuclear spins (I) of 1
and 0, respectively, and thus only couples with N nuclei. This interaction gives rise to
a triplet pattern in which the peaks, for the case of a diffusional species tumbling rapidly
in solution at room temperature, are all the same intensity (EPR spectrum for
diffusional TEMPO presented in Figure S10a, black trace). A triplet pattern is also
discernible for the C-mesoITO|STEMPO assembly, but the peak intensities are
distorted in this case (Figure S10a, red trace). This change in line-shape of the EPR
spectrum relative to the diffusional case arises from a slower tumbling rate which can
be a consequence of the impaired mobility of the TEMPO moiety upon STEMPO
immobilization.[38]
Figure 2f highlights the results from the FE-EPR investigation. C-mesoITO|STEMPO
samples were poised at a particular potential, using a three-electrode setup, and then
flash-frozen to allow for low-temperature EPR characterization. Examples of EPR
spectra, at three different potentials, are presented in Figure 2f (inset) (full range in
Figure S10b), where an increase in the applied bias is accompanied by a drop-in signal
intensity, that eventually vanishes due to the oxidation of the radical to EPR-silent
8
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
STEMPO+. The shape of the EPR spectra for E < 1 V vs. NHE are typical of nitroxide
radicals measured at low temperatures (100 K).[39] The normalized signal area of each
individual EPR spectrum was plotted as a function of the potential, and is a close fit to
the anticipated 1 e− Nernst equation (solid line, Figure 2f).
Next, we investigated the catalytic performance of the mesoITO|STEMPO assembly.
Figure 3a depicts the catalytic behavior of the system as a function of the solution pH,
where 4-methylbenzylalcohol (MBA) was chosen as a model substrate. The current
density increases with increasing pH, accompanied by a lower onset potential for
catalysis (from 0.75 V at pH 7.3, to 0.68 V at pH 10, vs. NHE), which is comparable to
previous reports for immobilized TEMPO on carbon-based electrodes.[40,41] This
observation is in-line with the established TEMPO-mediated oxidation mechanism,
whereby alcohol deprotonation leads to formation of a pre-oxidation complex via
nucleophilic attack of the alkoxide on the electrophilic nitrogen of the oxidized TEMPO
moiety (the oxoammonium cation), prior to aldehyde formation.[41–44] However, the
enhancement starts to plateau between pH 9 and pH 10, contrary to what is observed
for TEMPO, and related nitroxyl derivatives, in solution.[44] The plateau shown in Figure
3a for the mesoITO|STEMPO system could be due to a combination of factors, and
we rationalize this behavior to stem from the relatively slow electron transfer between
the ITO electrode and immobilized STEMPO, as well as from mass transport
limitations of the substrate in the mesoporous network.
Controlled potential electrolysis (CPE) was then conducted at an applied potential
(Eapp) of 1 V vs. NHE at room temperature, to further probe the effect of pH on the
mesoITO|STEMPO system. From Figure 2e, we observed that the stability of the
anodic, molecular assembly is high at pH 7 and 8, but less so at pH 10. However, the
TEMPO-mediated catalysis, and hence reaction kinetics, are favored under more
9
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
alkaline conditions (Figure 3a). To compare the overall mesoITO|STEMPO
performance as a function of pH, the TON and FE (Supporting Information, equations
(S4) and (S5), respectively) were calculated after a 3 h CPE experiment with MBA (30
mM) as the substrate, at the four different pH values (Figure 3b, Figure S11). The
moles of product, 4-methylbenzylaldehyde (nMBAd), originating from selective MBA
oxidation, was quantified by high performance liquid chromatography (HPLC)
(Supporting Information). The TON metric for STEMPO experiences a clear maximum
at pH 8, reaching a value close to 3000, highlighting the fine balance between
immobilization stability and catalytic reactivity for long-term electrolysis experiments.
On either side of the maximum, there is a corresponding decrease in the TON. At
lower pH, this can be attributed to a slower rate of substrate oxidation thereby resulting
in less nMBAd, whereas higher pH adversely affects the stability of the
mesoITO|STEMPO assembly, likely leading to a loss of the catalytic sites on the
electrode over reaction time. Post-CPE (at pH 8) XPS conducted on the
mesoITO|STEMPO electrode reveals peaks in the Si2p and N1s regions (Figure S12),
similar to those observed on a freshly assembled electrode (Figure 2a and Figure S6),
and thus indicates that the gradual drop in activity could be primarily due to hydrolysis
of the amide bond (and subsequent loss of the TEMPO moiety from the assembly).
On the other hand, the FE metric is invariant with the pH (average of 86 ± 3% as
calculated across the pH range, Figure 3b), implying that the charge passed at the
electrode|catalyst interface is utilized in the same, selective manner (being directed
towards substrate oxidation) throughout the pH range.
The versatility of the hybrid electrode was demonstrated by extending the substrate
scope to glycerol, cellulose-derived hydroxymethylfurfural (HMF), and the lignin model
compound 2-phenoxy-1-phenylethanol (PP-ol; Table S2).[45] A turnover frequency
10
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
(TOF) analysis based on the sigmoidal catalytic response of the CV trace was
performed for the STEMPO system in the presence of the different substrates
(Supporting Information).[46] Figure 3c and 3d depict concentration profiles obtained
for glycerol and HMF, respectively, and the concentration profile for MBA is shown in
Figure S13a (corresponding ‘maximum current density vs. concentration’ plots for
these three substrates are presented in Figure S13b-d). PP-ol was poorly soluble in
pure aqueous electrolyte, and thus a CV trace for this compound was recorded in a
MeCN-water mixture (Figure S14). The estimated TOFs for the four compounds, and
the relevant experimental conditions, are summarized in Table S2. The results show
that the mesoITO|STEMPO system can be used to oxidize a variety of alcohol-based
substrates, with the primary benzylic alcohols MBA and HMF showing highest activity
(TOF = 0.677 and 0.680 s−1, respectively), followed by the aliphatic triol glycerol
(0.557 s−1). The results from this analysis therefore signify the use of low-cost and
abundant alcohols like glycerol for electrocatalytic applications with the
mesoITO|STEMPO electrode. PP-ol gave the lowest TOF (0.268 s−1), which agrees
with the expected trend that primary alcohols are oxidized more rapidly than secondary
alcohols by TEMPO in basic solution.[42]
Having characterized the anodic assembly and demonstrated the electrocatalytic
compatibility of mesoITO|STEMPO with a variety of biomass representative alcohols,
we turned towards applying this system within a coupled electrolyzer. Conversion of
CO2-to-syngas as the cathodic half-reaction in the electrolyzer presents an attractive
strategy to utilize the electrons from alcohol oxidation by mesoITO|STEMPO. To
facilitate a cost-efficient redox cycle, use of robust, earth-abundant catalysts for
selective CO2R is essential. While many 3d transition metal-based molecular catalysts
have been developed over the years,[47] CoPc has emerged as one of the most
11
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
promising catalysts for CO2-to-CO reduction because of its enhanced performance
upon immobilization onto polymers and carbon-based electrodes.[48–50] In the coupled
electrolyzer, we employed a CNT-CoPPc hybrid electrode, fabricated by in situ
polymerization, and subsequently deposited on CP.[17] The CP|CNT-CoPPc cathode
catalyzes the electrochemical reduction of CO2 to syngas, with a CO:H2 ratio
dependent on the applied potential.[17,51]
The CV trace recorded for CP|CNT-CoPPc under N2 displays a broad quasi-reversible
redox process (Figure 4a, E1/2  –0.71 V vs. NHE), which corresponds to the metal-
centered CoII/CoI reduction of CoPPc. The surface concentration of electroactive
cobalt centers was estimated to be 18.3 ± 1.6 nmol cm–2 from integration of the CoI/CoII
oxidation wave (Figure S15). This corresponds to 5.6 ± 0.5% cobalt sites being
electrochemically accessible, whereby the total amount of Co was determined using
inductively coupled plasma optical emission spectroscopy (ICP-OES) measurements
(Supporting Information, equation (S6)).
A catalytic onset from the CP|CNT-CoPPc electrode was observed in a CO2 saturated
solution at a potential close to –0.84 V vs. NHE (Figure 4a). Electrocatalytic
performance of the cathode was probed by stepped constant potential
chronoamperometry in the range of –0.70 V to –1.00 V vs. NHE, with 50 mV
increments and 30 min steps (Figure S16). Product formation was monitored via a
continuous flow gas chromatography (GC) method (Supporting Information). H2 was
the only product until –0.80 V and CO evolution started at more negative potentials
( –0.85 V). The selectivity of the electrode towards CO increases sharply at more
negative potentials, reaching 76% at –1.00 V vs. NHE (overpotential, η = 0.46 V, where
12
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
E(CO2/CO) = −0.54 V vs. NHE at pH 7.3)[52]. Within the same potential range, the blank
CNT electrode did not generate any H2 or CO (Figure S16a, purple trace).
To elucidate the working principle of the coupled mesoITO|STEMPO–CP|CNT-CoPPc
electrolyzer, initial experiments were conducted using MBA. A catalytic wave for the
mesoITO|STEMPO assembly in the presence of MBA (30 mM) was observed, which
appeared to plateau at around 3 mA cm–2, at an applied potential just above 1 V vs.
NHE (Figure 4b). The mesoITO|STEMPO electrode displayed slightly lower current
densities than CP|CNT-CoPPc and was therefore selected as the working electrode
(WE) in the coupled electrolyzer assembly, while the cathode assumed the role of the
counter electrode (CE). A two-compartment electrochemical cell was employed with a
Selemion-AMV anion-exchange membrane to separate the compartments. A Ag/AgCl
reference electrode (RE) was placed in the working compartment and the three-
electrode configuration was adopted prior to testing in a two-electrode fashion to be
able to precisely control the Eapp at the WE versus a known reference (Supporting
Information). This also allowed us to record the exact potential at the CE (ECE) during
electrolysis against the same reference, thus providing a more detailed description of
the cell parameters over reaction time.
A CO2 saturated carbonate buffer (0.5 M) was used in both compartments, which
yielded a solution pH close to 7.3 that remained relatively constant throughout the
experiment. Figure 4c depicts the results from the coupled electrolysis (three-electrode
configuration), with Eapp = 1 V vs. NHE at room temperature. Alcohol conversion to the
corresponding aldehyde, MBAd, was quantified by HPLC, whereas CO and H2 were
quantified via the continuous flow GC method (Supporting Information). Catalytic
metrics obtained for the respective anode and cathode highlight the effectiveness of
the combined system. MBA oxidation resulted in a TONSTEMPO of 1515 and FE close
13
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
to 90% after the 3 h CPE experiment. The TONSTEMPO was lower than expected from
the TOF analysis from CV scans (Table S2) due to the modest stability of the anodic
assembly, as demonstrated by the multiple CV scan measurements and prolonged
CPE (cf. Figure 2e and Figure 3b, respectively). A cobalt-based TON for syngas
generation of 1360 (TONCO = 599 and TONH = 761) and overall FEs for CO and H2
2
of 35% and 45%, respectively, were achieved for the CP|CNT-CoPPc cathode.
This performance encouraged the substitution of MBA for glycerol, on account of its
advantages as a potential substrate for coupling with CO2R in real-life applications. A
similar setup to that used for coupled MBA oxidation was employed, except in this
case, the anode compartment consisted of a carbonate buffer (0.5 M) at pH 8.3 (under
N2), whereas the catholyte was comprised of a CO2 saturated carbonate buffer (0.5 M)
at pH 7.3. This was deemed necessary for glycerol, as the STEMPO-mediated
catalysis involving this substrate was observed to be too sluggish at the quasi-neutral
pH of CO2 saturated carbonate buffer (i.e. pH 7.3), but increased in activity under more
alkaline conditions (as evidenced by the CVs recorded at pH 7.3 and pH 8.3, Figure
S17). Figure 5a illustrates the reaction time plot obtained with glycerol as the
substrate, with Eapp = 1 V vs. NHE. HPLC analysis revealed that glyceraldehyde
(GlyAd) was the primary anodic product from the coupled electrolysis experiment.
The two compartments maintained their individual pH values for the duration of the
electrolysis, and a TONSTEMPO and FE of 997 and 83%, respectively, were measured
for the anodic half-reaction. Although precautions were taken to minimize
overoxidation or further reaction of GlyAd, trace amounts of some side-product can
potentially form (not detected by HPLC), leading to the observed 7% drop in the FE
relative to the MBA electrolyzer. With regards to the cathode metrics, the cobalt-based
14
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
TON was determined to be equal to 894 (TONCO = 444 and TONH2 = 450), while similar
FEs for the gaseous products, relative to the MBA-based electrolyzer, were measured
(FE = 41% for CO, 41% for H2). A side-by-side comparison of the calculated FEs for
the liquid and gaseous products over reaction time, for the MBA- and glycerol-based
electrolyzers, is provided in the Supporting Information (Figure S18).
As mentioned, using a three-electrode configuration allowed for ECE (i.e. the potential
at the CP|CNT-CoPPc electrode) to be monitored throughout the course of the
electrolysis experiment. From the traces shown in Figure 5b, there is an alteration in
the CO:H2 ratio at the cathode over time, which seems to be a consequence of the
decrease in the absolute value of ECE. This decrease in the reducing potential at the
cathode is itself a result of the gradual decline in activity at the anode over time. The
change in the CO:H2 ratio as a function of the cathodic potential is in-line with the
stepped chronoamperometric experiments carried out for the CP|CNT-CoPPc
electrode (with Pt mesh as CE), as discussed previously (Figure S16). The time-lag
between the minima of the ECE trace and the maximum value of CO:H2 ratio on Figure
5b is likely caused by the slow diffusion of CO from the porous cathode.
We furthered our investigation into coupled glycerol oxidation and CO2R, and
performed a series of experiments in a two-electrode configuration, while varying the
applied cell potential (Vcell). Values for Vcell in the range of 1.8 to 2.1 V were chosen,
̅
̅
|
|
based on the rationale that: Ecathode − Eanode ≈ |ECE − Eapp| = 1.85 V, where ECE is
the average potential at the CE, over reaction time, as measured in the three-electrode
configuration (i.e. Figure 5b). Figure 5c depicts the combined FE at the cathode (for
CO and H2) and the CO:H2 ratio, over reaction time, for Vcell = 2.0 V. The trends agree
with those obtained for the three-electrode setup. The increase in the maximum of the
CO:H2 ratio for the two- versus three-electrode configuration (shown in Figure 5b)
15
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
could be a result of the increased driving force provided by the 2.0 V potential. This
bias most likely leads to more reductive potentials at the cathode, and, in accordance
with the stepped chronoamperometry data for CP|CNT-CoPPc (Figure S16), would
translate to a higher CO:H2 ratio.
Finally, we attempted to parametrize the cell energy efficiency (), as a function of Vcell.
The equation used to calculate this metric is defined in equation (1):[12]
+
|(FEH ∙EH /H + FECO∙ECO /CO) − FEGlyAd∙EGlyAd/glycerol
|
|
|
Ecell
Vcell
2
2
2
(1)
 =
=
Vcell
where EH /H , ECO /CO, and EGlyAd/glycerol denote the reduction potentials for H+, CO2,
+
2
2
and glyceraldehyde, respectively, under non-standard conditions (Table S3). A more
detailed breakdown regarding the thermodynamic analysis required to compute  is
provided in the Supporting Information. Figure 5d illustrates the FEs for the anodic and
cathodic processes, along with the corresponding  calculations, for different Vcell
values. There is a slight improvement in the CO selectivity upon increasing from 1.8
to 1.9 V (FECO = 36 and 46%, respectively), presumably a result of the higher driving
force at these applied voltages. This enhancement is met with an improvement in 
(from 16 to 18%), since the 100 mV additional bias is offset by the increase in FECO,
as governed by equation (1). However, for Vcell ≥ 2.0 V, the combined effects of a
largely unchanged CO:H2 ratio and anodic FE, causes a corresponding drop in the cell
efficiency to 16%, similar to that obtained for Vcell = 1.8 V. The cell energy efficiency
values measured for our hybrid electrolyzer are in accordance with those reported in
the literature, where for example an efficiency of 17%, at 1.8 V cell potential, was
measured for a tandem electrolyzer featuring benzyl alcohol oxidation coupled with
CO2 to CO and H2 conversion.[9] However, the previously reported system was
16
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
comprised of Ru-based molecular catalysts for the reductive and oxidative half-
reactions, and additionally, only one of the catalysts was immobilized. On the other
hand, we have incorporated immobilized cathodic and anodic catalysts in our
electrolyzer, both free of any precious-metals, and have also demonstrated the
applicability of the tandem AlcOx–CO2R device to couple the oxidation of more
commercially viable substrates, such as glycerol, with CO2-to-syngas conversion.
Conclusion
In conclusion, we have designed, fabricated and characterized an anode featuring a
silatrane-modified TEMPO molecule on a mesoITO scaffold, and demonstrated the
electrocatalytic ability of the hybrid MOx system to efficiently oxidize a variety of
biomass representative substrates. The siloxane anchor, formed upon hydrolysis of
the silatrane cage on the MOx surface, displays robust binding. The catalytically active
site (i.e. the oxoammonium cation) is both stable and readily regenerated under
electrocatalytic conditions,[53] and we believe that the long-term stability of the hybrid
electrode assembly is currently limited by the amide bond in STEMPO. Improvements
to the molecular design of the linker employed for STEMPO will provide a possibility
to enhance the stability and overall activity of the anodic assembly.
We further showed the advantage and versatility of our novel STEMPO anode by
coupling alcohol oxidation with an efficient CO2R cathode (CP|CNT-CoPPc) to
construct a fully molecular, precious-metal-free, AlcOx–CO2R electrolyzer. The
functionality and performance of the device was investigated using a three-electrode
configuration, first employing MBA as a model substrate, and later, using the
commercially applicable substrate, glycerol. It was found that in both cases,
17
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
stoichiometric amounts of a selective oxidation product (the corresponding aldehyde)
and syngas were generated at the anode and cathode, respectively. FEs were typically
excellent for the hybrid system, exceeding 80% for both anode and cathode. TONs
were also high, approaching 1000 for mesoITO|STEMPO and 900 for CP|CNT-CoPPc
(with glycerol as substrate). The TON of the cathode in the electrolyzer is currently
limited by the prolonged stability issue of the anodic assembly during continuous CPE
experiments and the CoPPc-cathode on its own is known to maintain activity over a
longer time-period.[17] Further studies were then made using a demonstrator-type, two-
electrode setup for coupled glycerol oxidation at the anode and syngas generation at
the cathode, showing similarly promising metrics as per the three-electrode setup. Cell
energy efficiency calculations also revealed the advantages of operating at a lower
Vcell , with a maximum efficiency of 18% being measured at a cell potential of 1.9 V.
This molecular hybrid system is therefore a suitable model for the development of
future AlcOx–CO2R electrolyzers based on earth-abundant materials, which can
provide chemical feedstocks (aldehydes and syngas) from sustainable and abundant
resources, such as biomass-derived alcohols, CO2, and renewable electricity.
18
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
Figures
Figure 1: Coupled AlcOx–CO2R electrolyzer, featuring a mesoITO|STEMPO anode (right-
hand compartment) and a CP|CNT-CoPPc cathode. An SEM cross-section image of the
mesoITO electrode is shown on the right (film thickness 4.5 μm).
19
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
Figure 2: Characterization of mesoITO|STEMPO, assembled under optimized conditions. (a)
−
Si2p XPS spectrum of mesoITO|STEMPO. (b) Multi-scan CV, conditions: pH 8 aq. HCO3
/CO32− (0.2 M),  = 50 mV s−1, N2 (legend denotes scan number). (c) Trumpet plot deduced
−
from the variable scan rate CV measurements; conditions: pH 8 aq. HCO3 /CO32− (0.5 M), N2.
(d) ip vs.  plot, for  < c. (e) Stability curves as a function of pH (data fitted to a mono-
exponential decay), formulated by tracing the change in STEMPO (obtained through integration
of the oxidation wave in the CV) over scan number in the multi-scan CV experiment. (f) FE-
EPR potentiometric titration of C-mesoITO|STEMPO. Peak area of the STEMPO EPR signal
as a function of potential (colored dots), fitted to 1 e− Nernst equation (solid line). Inset: X-
band (9.5 GHz) EPR spectra of STEMPO at different applied potentials. Measurements
performed at 100 K, 2 mW microwave power, 100 KHz modulation frequency and 2 G
modulation amplitude.
20
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
Figure 3: (a) pH dependent CVs measured for the mesoITO|STEMPO system, in the
presence of 30 mM MBA. (b) TON and FE metrics compiled from the CPE experiment under
−
a range of pH values. Conditions. pH 7.3: CO2 saturated aq. HCO3 /CO32− (0.5 M); pH 8 – 10:
−
aq. HCO3 /CO32− (0.5 M) titrated under ambient conditions to the correct pH; for CVs:  = 20
mV s−1; for CPE: Eapp = 1 V vs. NHE, tCPE = 3 h, MBA (30 mM). Product quantification (via
HPLC) at the 3 h time-point was used for TON and FE calculation (Supporting Information,
equations (S4) and (S5)). Electrochemically determined concentration profiles for (c) glycerol,
and (d) HMF, using optimized pH (pH 8), and  = 20 mV s−1.
21
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
Figure 4: CVs for the (a) CP|CNT-CoPPc cathode, and (b) mesoITO|STEMPO anode,
recorded separately in a three-electrode setup, with Pt mesh as the CE and Ag/AgCl as RE.
−
Conditions: CO2 saturated pH 7.3 aq. HCO3 /CO32− (0.5 M),  = 20 mV s−1 (pH 8 for cathode
N2 trace). (c) Coupled electrolyzer, showing moles of product (liquid and gaseous) and e− (as
recorded on the potentiostat) over reaction time. Conditions: two-compartment cell fitted with
anion-exchange membrane; three-electrode configuration with mesoITO|STEMPO as WE,
−
2−
CP|CNT-CoPPc as CE and Ag/AgCl as RE; CO2 saturated pH 7.3 aq. HCO3 /CO3 (0.5 M)
in both compartments; MBA substrate (30 mM) present in anodic compartment, Eapp (anode)
= 1 V vs. NHE, tCPE = 3 h; MBAd quantified by HPLC, CO/H2 via continuous flow GC.
22
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
Figure 5: (a) Similar profile to Figure 4c, but with glycerol as the substrate (Eapp (anode) = 1
V vs. NHE). (b) Trend in the CO:H2 ratio and CP|CNT-CoPPc CE potential (ECE) over reaction
time. (c) Combined FE for CO and H2, and CO:H2 ratio for the two-electrode configuration
incorporating glycerol as the substrate (applied cell potential = 2.0 V). (d) Cell energy
efficiency and FE plotted as a function of Vcell in the same two-electrode setup. Conditions: for
both the three-electrode (plots (a) and (b)), and two-electrode (plots (c) and (d)) configurations,
a two-compartment cell (fitted with anion-exchange membrane) was used; anode
−
compartment: N2 saturated pH 8.3 aq. HCO3 /CO32− (0.5 M); cathode compartment: CO2
−
saturated pH 7.3 aq. HCO3 /CO32− (0.5 M); glycerol substrate (50 mM) present in anodic
compartment, tCPE = 3 h. Glycerol oxidation and gaseous products quantified via HPLC and
continuous flow GC analysis, respectively.
23
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
Acknowledgments
This work was supported by the Endeavour Scholarship (M.B.), the Christian
Doppler Research Association (Austrian Federal Ministry for Digital and
Economic Affairs and the National Foundation for Research, Technology and
Development), OMV (M.A.B., J.W., A.W., E.R.), and a Marie Skłodowska-Curie
Fellowship from European Commission H2020 (S.R., GAN745604, CO2RED),
the Leverhulme Trust (M.M.R & E.R, RPG-2018-183), and the EPRSC
(EP/L011972/1 to CAESR). The authors thank Dr. Christian Pichler for
assistance with HPLC measurements, Mr. Adam Brown for XPS
measurements, Dr. Bertrand Reuillard and Dr. Sam Cobb for helpful
discussions, and the Centre for Advanced ESR (University of Oxford) for EPR
measurement time.
Conflict of interest
The authors declare no conflict of interest.
Keywords: alcohols · carbon dioxide · electrocatalysis · catalyst immobilization
· energy conversion
24
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
References
[1]
[2]
[3]
P. De Luna, C. Hahn, D. Higgins, S. A. Jaffer, T. F. Jaramillo, E. H. Sargent,
Science (80-. ). 2019, 364, eaav3506.
P. Kang, Z. Chen, A. Nayak, S. Zhang, T. J. Meyer, Energy Environ. Sci. 2014,
7, 4007–4012.
C. C. L. McCrory, S. Jung, I. M. Ferrer, S. M. Chatman, J. C. Peters, T. F.
Jaramillo, J. Am. Chem. Soc. 2015, 137, 4347–4357.
[4]
[5]
H. G. Cha, K.-S. Choi, Nat. Chem. 2015, 7, 328–333.
E. J. Horn, B. R. Rosen, Y. Chen, J. Tang, K. Chen, M. D. Eastgate, P. S. Baran,
Nature 2016, 533, 77–81.
[6]
[7]
[8]
B. You, X. Liu, N. Jiang, Y. Sun, J. Am. Chem. Soc. 2016, 138, 13639–13646.
T. Li, Y. Cao, J. He, C. P. Berlinguette, ACS Cent. Sci. 2017, 3, 778–783.
T. Li, T. Kasahara, J. He, K. E. Dettelbach, G. M. Sammis, C. P. Berlinguette,
Nat. Commun. 2017, 8, 390.
[9]
Y. Wang, S. Gonell, U. R. Mathiyazhagan, Y. Liu, D. Wang, A. J. M. Miller, T. J.
Meyer, ACS Appl. Energy Mater. 2019, 2, 97–101.
[10] C. Huang, Y. Huang, C. Liu, Y. Yu, B. Zhang, Angew. Chem. Int. Ed. 2019, 58,
12014–12017.
[11] E. Reisner, Angew. Chemie - Int. Ed. 2019, 58, 3656–3657.
[12] S. Verma, S. Lu, P. J. A. Kenis, Nat. Energy 2019, 4, 466–474.
[13] S. S. Yazdani, R. Gonzalez, Curr. Opin. Biotechnol. 2007, 18, 213–219.
[14] C. Dai, L. Sun, H. Liao, B. Khezri, R. D. Webster, A. C. Fisher, Z. J. Xu, J. Catal.
2017, 356, 14–21.
[15] M. F. Semmelhack, C. R. Schmid, J. Am. Chem. Soc. 1983, 105, 6732–6734.
[16] T. Miyazawa, T. Endo, S. Shiihashi, M. Okawara, J. Org. Chem. 1985, 50, 1332–
1334.
[17] N. Han, Y. Wang, L. Ma, J. Wen, J. Li, H. Zheng, K. Nie, X. Wang, F. Zhao, Y.
Li, et al., Chem 2017, 3, 652–664.
[18] P. G. Hoertz, Z. Chen, C. A. Kent, T. J. Meyer, Inorg. Chem. 2010, 49, 8179–
8181.
[19] D. Mersch, C. Y. Lee, J. Z. Zhang, K. Brinkert, J. C. Fontecilla-Camps, A. W.
Rutherford, E. Reisner, J. Am. Chem. Soc. 2015, 137, 8541–8549.
[20] T. E. Rosser, C. D. Windle, E. Reisner, Angew. Chem. Int. Ed. 2016, 55, 7388–
7392.
[21] T. E. Rosser, E. Reisner, ACS Catal. 2017, 7, 3131–3141.
[22] J. M. Bobbitt, C. Brückner, N. Merbouh, Org. React. 2010, 103–424.
25
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
[23] M. Rafiee, B. Karimi, S. Alizadeh, ChemElectroChem 2014, 1, 455–462.
[24] D. P. Hickey, D. A. Schiedler, I. Matanovic, P. V. Doan, P. Atanassov, S. D.
Minteer, M. S. Sigman, J. Am. Chem. Soc. 2015, 137, 16179–16186.
[25] K. L. Materna, R. H. Crabtree, G. W. Brudvig, Chem. Soc. Rev. 2017, 46, 6099–
6110.
[26] B. J. Brennan, M. J. Llansola Portolés, P. A. Liddell, T. A. Moore, A. L. Moore,
D. Gust, Phys. Chem. Chem. Phys. 2013, 15, 16605–16614.
[27] K. L. Materna, B. J. Brennan, G. W. Brudvig, Dalt. Trans. 2015, 44, 20312–
20315.
[28] C.-J. Huang, Y.-Y. Zheng, Langmuir 2019, 35, 1662–1671.
[29] N. M. Muresan, J. Willkomm, D. Mersch, Y. Vaynzof, E. Reisner, Angew. Chem.
Int. Ed. 2012, 51, 12749–12753.
[30] K. L. Materna, B. Rudshteyn, B. J. Brennan, M. H. Kane, A. J. Bloomfield, D. L.
Huang, D. Y. Shopov, V. S. Batista, R. H. Crabtree, G. W. Brudvig, ACS Catal.
2016, 6, 5371–5377.
[31] P. Louette, F. Bodino, J.-J. Pireaux, Surf. Sci. Spectra 2007, 12, 133–138.
[32] A. J. Bard, L. R. Faulkner, Electrochemical Methods: Fundamentals and
Applications, Wiley, New York, 2001.
[33] A. F. Loftus, K. P. Reighard, S. A. Kapourales, M. C. Leopold, J. Am. Chem.
Soc. 2008, 130, 1649–1661.
[34] J. Zhao, X. Zhu, T. Li, G. Li, Analyst 2008, 133, 1242.
[35] E. Laviron, J. Electroanal. Chem. 1979, 101, 19–28.
[36] C. M. Hanna, C. D. Sanborn, S. Ardo, J. Y. Yang, ACS Appl. Mater. Interfaces
2018, 10, 13211–13217.
[37] K. Abdiaziz, E. Salvadori, K. P. Sokol, E. Reisner, M. M. Roessler, Chem.
Commun. 2019, 55, 8840–8843.
[38] M. M. Haugland, E. A. Anderson, J. E. Lovett, in Electron Paramagn. Reson.,
2017, pp. 1–34.
[39] C. Tansakul, E. Lilie, E. D. Walter, F. Rivera, A. Wolcott, J. Z. Zhang, G. L.
Millhauser, R. Braslau, J. Phys. Chem. C 2010, 114, 7793–7805.
[40] D. P. Hickey, R. D. Milton, D. Chen, M. S. Sigman, S. D. Minteer, ACS Catal.
2015, 5, 5519–5524.
[41] A. Das, S. S. Stahl, Angew. Chem. Int. Ed. 2017, 56, 8892–8897.
[42] W. F. Bailey, J. M. Bobbitt, K. B. Wiberg, J. Org. Chem. 2007, 72, 4504–4509.
[43] D. P. Hickey, D. A. Schiedler, I. Matanovic, P. V. Doan, P. Atanassov, S. D.
Minteer, M. S. Sigman, J. Am. Chem. Soc. 2015, 137, 16179–16186.
[44] M. Rafiee, K. C. Miles, S. S. Stahl, J. Am. Chem. Soc. 2015, 137, 14751–14757.
26
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
[45] X. Wu, X. Fan, S. Xie, J. Lin, J. Cheng, Q. Zhang, L. Chen, Y. Wang, Nat. Catal.
2018, 1, 772–780.
[46] C. Costentin, S. Drouet, M. Robert, J. M. Savéant, J. Am. Chem. Soc. 2012,
134, 11235–11242.
[47] K. E. Dalle, J. Warnan, J. J. Leung, B. Reuillard, I. S. Karmel, E. Reisner, Chem.
Rev. 2019, 119, 2752–2875.
[48] X. Zhang, Z. Wu, X. Zhang, L. Li, Y. Li, H. Xu, X. Li, X. Yu, Z. Zhang, Y. Liang,
et al., Nat. Commun. 2017, 8, 14675.
[49] M. Wang, K. Torbensen, D. Salvatore, S. Ren, D. Joulié, F. Dumoulin, D.
Mendoza, B. Lassalle-Kaiser, U. Işci, C. P. Berlinguette, et al., Nat. Commun.
2019, 10, 3602.
[50] Y. Liu, C. C. L. McCrory, Nat. Commun. 2019, 10, 1–10.
[51] S. Roy, E. Reisner, Angew. Chem. Int. Ed. 2019, 58, 12180–12184.
[52] A. J. Bard, R. Parsons, J. Jordan, Eds. , Standard Potentials in Aqueous
Solution, New York, 1985.
[53] J. E. Nutting, M. Rafiee, S. S. Stahl, Chem. Rev. 2018, 118, 4834–4885.
27
This article is protected by copyright. All rights reserved.

10.1002/anie.202002680
Angewandte Chemie International Edition
TOC artwork
28
This article is protected by copyright. All rights reserved.