Water Oxidation at Neutral pH using a Highly Active Copper-Based Electrocatalyst

Article abstract of DOI:10.1002/cssc.202001444

The sluggish kinetics of the oxygen evolution reaction (OER) at the anode severely limit hydrogen production at the cathode in water splitting systems. Although electrocatalytic systems based on cheap and earth-abundant copper catalysts have shown promise

Full text of DOI:10.1002/cssc.202001444

Chemistry–Sustainability–Energy–Materials
Accepted Article
Title: Highly active copper-based electrocatalyst for water oxidation at
neutral pH
Authors: Shiguo Zhang, Hussein A. Younus, Yan Zhang, Matthias
Vandichel, Nazir Ahmad, Kari Laasonen, Francis Verpoort,
and Ce Zhang
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.202001444
Link to VoR: https://doi.org/10.1002/cssc.202001444
01/2020

10.1002/cssc.202001444
ChemSusChem
Highly active copper-based electrocatalyst for water oxidation at neutral pH
Hussein A. Younus,ab Yan Zhang,*a Matthias Vandichel,ce Nazir Ahmad,g Kari Laasonen,e
Francis Verpoort,*d Ce Zhang,f and Shiguo Zhang*a
a
College of Materials Science and Engineering, Hunan Province Key Laboratory for Advanced
Carbon Materials and Applied Technology, Hunan University, Changsha 410082, China.
b Chemistry Department, Faculty of Science, Fayoum University, Fayoum 63514, Egypt.
c
Department of Chemical Sciences and Bernal Institute, University of Limerick, Limerick V94
T9PX, Republic of Ireland.
d
Laboratory of Organometallics, Catalysis and Ordered Materials, State Key Laboratory of
Advanced Technology for Materials Synthesis and Processing, Wuhan University of Technology,
Wuhan 430070, China.
e School of Chemical Engineering, Aalto University, 02150 Espoo, Finland.
f
Nanophotonics and Optoelectronics Research Center, Qian Xuesen Laboratory of Space
Technology, China Academy of Space Technology, Beijing 100094, China.
g Department of Chemistry, GC University, Lahore, 54000, Pakistan.
Emails:
Yan Zhang: zyan1980@hnu.edu.cn,
Francis Verpoort: francis.verpoort@ku.ac.ae,
Shiguo Zhang: zhangsg@hun.edu.cn.
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
Abstract
The sluggish kinetics of the oxygen evolution reaction at the anode severely limits the hydrogen
production at the cathode in water spitting systems. While electrocatalytic systems based on cheap
and earth-abundant metal copper catalysts have been promising for water oxidation under basic
conditions, only very few examples with high overpotential can be operated under acidic or neutral
conditions, even though hydrogen evolution in the latter case is much easier. This work presents
an efficient and robust Cu-based molecular catalyst, which self-assembles as a periodic film from
its precursors under aqueous conditions on the surface of glassy carbon electrodes (GCE). This
film catalyzes the oxygen evolution reaction (OER) under neutral conditions with impressively
low overpotential. In controlled potential electrolysis, a stable catalytic current of 1.0 mA/cm2 can
be achieved at only 2.0 V (vs. RHE) and no remarkable decrease in the catalytic current is observed
even after prolonged bulk electrolysis. The catalyst displays first-order kinetics and a single site
mechanism for water oxidation with a TOF (kcat) of 0.6 s-1. DFT calculations are performed to
study the OER behavior of the periodic Cu(TCA)2 (HTCA = 1-mesityl-1H-1,2,3-triazole-4-
carboxylic acid) film and reveal that TCA defects within the film create Cu(I) active sites which
can provide a low overpotential route for OER. This route involves Cu(I), Cu(II)-OH, Cu(III)=O
and Cu(II)-OOH intermediates and is enabled at a potential of 1.54 V (vs. RHE), requiring an
overpotential of 0.31 V. This corresponds well with an overpotential of ~ 0.29 V obtained
experimentally for the grown catalytic film after 100 CV cycles at pH=6. However, to reach a
higher current density of 1 mA cm−2, an overpotential of 0.72 V is required.
Keywords: Electrochemical water oxidation; Molecular catalyst; Cu metal-organic-material;
Neutral conditions; DFT calculations.
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
1.
Introduction
Economical hydrogen production to address the increased demand for developing a sustainable
energy system is one of the most challenging aspects. Electrolysis of water into hydrogen and
oxygen would be the most efficient process, wherein the water oxidation half reaction or oxygen
evolution reaction (OER) produces protons and electrons necessary for the hydrogen evolution at
the cathode. This step, however, is the bottle neck of water electrolysis and determines the overall
performance because of slow kinetics, large potential, and instability of electrocatalysts. Therefore,
the development of innovative water oxidation electrocatalysts with improved catalytic
performance is essential for practical applications. Although the library of heterogeneous catalysts
for OER is growing rapidly,[1] molecular catalysts are particularly attractive since their catalytic
properties can be relatively tuned by systematic structural and electronic modifications, and it can
be incorporated into chromophore–catalyst assemblies for their applications in
photoelectrochemical cells. Furthermore, the development of more active/stable OER catalysts can
be stimulated and guided via mechanistic computational studies.[2] Various molecular complexes
have been identified as catalysts in solution or immobilized on metal-oxide surfaces, or as
precursors to catalytically active metal oxides or films.[3] On the other hand, research on water
oxidation should focus on two key points. The first one is lowering the overpotential, which in
turn might improve the catalyst stability and robustness, since in most cases, catalyst robustness
and activity towards water oxidation go hand in hand. The smoother the flow of electrons go
through the system, the more robust the catalyst is. On the contrary, if the charge transfer resistance
is too high and the charge stays too long in the system, it will ultimately destroy molecular building
blocks. The second factor is reducing the catalyst cost with using earth abundant elements, so that
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
enabling cost effective hydrogen production. In this context, cobalt, copper, nickel, iron, and
manganese based-molecular catalysts have been presented.[4]
Recently, Cu-based water oxidation catalysts (WOCs) have been widely developed because of
their high abundance, low cost, rich redox properties with a variety of oxidation states, and
biologically relevance.[5] After the first report by Mayer and co-workers on a Cu(II) 2,2′-
bipyridine catalyst for water oxidation,[6] several copper catalysts were presented with simple Cu(II)
complexes or salts.[7] Several motifs were incorporated in the design of copper molecular catalysts
ranging from monodentate to pentadentate ligands as well as coordination polymer.[8] A slight
modification in the ligand structure in copper-catalyzed water oxidation could substantially reduce
the overpotential by more than 200 mV under alkaline conditions, emphasizing the critical role of
the ligand and the possible tuning of the catalyst activities via manipulating the electronic
environment around the metal center.[9] Recently, a water-soluble cationic mononuclear CuII
porphyrin was reported to catalyze water oxidation under neutral conditions with a low
overpotential of 310 mV (estimated at j = 0.1 mA cm−2). Rotating ring-disc electrode experiments
and pH-dependent electrochemical studies proposed that during the catalytic process, a H2O2-
related species, instead of high-valent CuIII–O•/CuIV=O, was generated to form the O–O bond
which might be the reason for the observed low overpotential.[10] Copper metal could be an ideal
catalyst for water oxidation, however its dissolution in aqueous conditions under operating
potentials is another challenge for its utilization. Anodization of a copper electrode in a carbonate
buffer induces the formation of very small copper-oxide particulate materials which have shown
high stability and activity with reasonable overpotential.[11] Though this noticeable growth in
copper based catalysts for water oxidation, one can only find few examples that operate under
neutral or acidic conditions where hydrogen evolution is much easier, which might be correlated
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
to the high overpotential required by copper catalysts in such conditions.[12] Even some excellent
systems revealing high activity in basic conditions (pH 9.0), they didn’t show any catalytic activity
at pH-levels below 7.5.[8a, 8d] Thus, to develop copper-based molecular OER catalysts with high
efficiency and low onset potential in neutral or acidic conditions is highly desired.
In this context, we report a mononuclear copper (II) complex [Cu(TCA)2] (complex 1) as a
molecular catalyst for electrochemical water oxidation, which is generated in-situ from copper salt
and an organic ligand, 1-mesityl-1H-1,2,3-triazole-4-carboxylic acid (HTCA) under aqueous
conditions. This molecular catalyst is able to catalyze water oxidation in neutral conditions at very
low overpotential and display a very stable oxygen evolution current density of about 1 mA/cm2
at 2.0 V (versus RHE) for more than 15 h, with its activity superior to most reported Cu-based
catalysts under neutral conditions.[12a]
2. Experimental
2.1. Materials and physical measurements
All solvents and reagents were purchased from Sigma Aldrich or Aladdin and used as received.
All solutions were prepared in ultrapure water (Millipore MilliQ® A10 gradient, 18.25 MΩ cm,
1
13
2–4 ppb total organic content). H and C NMR spectra were recorded at 400 MHz and at 100
MHz, respectively, if not otherwise stated using Bruker 500 MHz NMR spectrometer. UV/Vis
spectra were recorded on a Shimadzu UV-1800 spectrophotometer at room temperature. Field
Emission Scanning Electron Microscope (FE-SEM) images and SEM-EDX data were taken on a
ZEISS ULTRA PLUS-43-13 in connection to an Energy Dispersive Spectrometer (OXFORD X-
Max 50). The acceleration voltage was 15 kV. High-resolution mass spectra were collected on a
Bruker Daltonics Solarix 7.0 Tesla Fourier Transform Mass Spectrometer.
2.2. Electrochemistry
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
Electrochemical measurements were collected in a three-electrode glass cell with a 3 mm glassy
carbon working electrode in aqueous acetate solution. Electrolyte pHs were adjusted using 1M
CH3COOH or 1M NaOH. The counter electrode was platinum wire, and the reference was
Ag/AgCl (E = 0.197 V vs. NHE). The overpotential was calculated as follows η= (ENHE
+
0.059*pH) – 1.23. The glassy carbon disk was polished mechanically with an aqueous slurry of
0.3, 0.1 and 0.05 μm alumina successively, on a Microcloth polishing fabric, until a mirror finish
was achieved. After polishing, the electrode was ultrasonically cleaned in Milli-Q (Millipore)
water then acetone each for 1–2 minutes after each polishing step and rinsed thoroughly with pure
water. The platinum counter electrode was flame annealed and washed with pure water before
placing it into the cell. The ITO slides were cleaned in succession with detergents and acetone and
ultimately washed with Millipore water. Electrochemical measurements were conducted using a
CHI660E electrochemical workstation (CH Instruments, Shanghai, China).
2.3. Synthesis
Aryl azide[13]
A round bottom flask was charged with Aryl amine (0.05 mol) and water (60 mL). Concentrated
HCl (8.4 mL, 0.105 mol) was added to the vigorously stirred reaction mixture in an ice-water bath.
After stirring and cooling to 0 °C for 20–30 min, a freshly prepared, ice cold solution of NaNO2
(3.25 g, 0.05 mol) in water (10 mL) was added dropwise to the reaction mixture while keeping the
internal temperature between 0–5 °C. After addition of NaNO2 (aq.), the reaction mixture was
stirred for an additional 10 min. A freshly prepared solution of sodium azide (3.25 g, 0.05 mol) in
water (30 mL) was added dropwise to the reaction mixture via additional funnel while maintaining
the internal temperature of the reaction mixture below 5 °C. Upon complete addition of the sodium
azide solution, the reaction mixture was stirred for an additional 20–30 min at 0 °C, followed by
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
stirring at room temperature for another 3 hr. The reaction mixture was extracted with ethyl acetate
EtOAc (2×100 mL). The combined organic layers were dried over anhydrous Na2SO4 (s), filtered
and concentrated under reduced pressure. The crude product was further purified by flash column
chromatography using petroleum ether as eluent.
Mesityl azide
1
Yellow liquid, yield: 60.5%; H NMR (500 MHz, CDCl3) δ 6.84 (s, 2H), 2.33 (s, 6H), 2.26 (s,
3H); 13C NMR (126 MHz, CDCl3) δ 135.3, 134.3, 131.8, 129.5, 20.7, 18.0.
Phenyl azide:
Dark red liquid, yield: 80.5%; 1H NMR (500 MHz, CDCl3) δ 7.43 (t, J = 7.9 Hz, 2H), 7.22 (t, J =
13
7.4 Hz, 1H), 7.11 (d, J = 8.1 Hz, 2H); C NMR (126 MHz, CDCl3) δ 139.64, 129.72, 124.51,
118.55.
3-Nitro phenyl azide:
Orange solid, yield: 75.0%; 1H NMR (500 MHz, CDCl3) δ 9.62 (s, 1H), 8.82 (t, J = 2.1 Hz, 1H),
8.46 (dd, J = 1.4, 1H) 8.40-8.33 (m, J = 8.2 Hz, 1H), 7.92 (t, J = 8.2, 1H).
1-Mesityl-1H-1,2,3-triazole-4-carboxylic acid (HTCA)
The compound was synthesized according to a published method, with some modifications.[14]
Under Ar atmosphere in 10 mL vial CuSO4.5H2O (40 mg, 5 mol %, 0.16 mmol), sodium ascorbate
(64 mg, 10 mol %, 0.32 mmol) was dissolved in water (1.0 mL). Mesityl azide (525 mg, 1.0 equiv,
3.26 mmol) and propiolic acid (254 μL, 1.2 equiv, 3.92 mmol) were dissolved independently each
in 0.5 mL t-BuOH and sequentially injected the previously prepared click catalyst. The mixture
was stirred at room temperature for 24 h under Ar atmosphere. A yellow precipitate appears then
the reaction was quenched with saturated NaHCO3 (15 mL), washed twice with ether (2×40 mL),
acidified with 0.5 M aq. solution of H2SO4 (~100 mL), and extracted with ethyl acetate (3×50 mL).
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
The combined organic solutions were dried over Na2SO4, filtered, and concentrated under reduced
pressure to afford acid the product as a white solid (450 mg, 72.5%): 1H NMR (500 MHz, CDCl3)
13
δ 8.24 (s, 1H), 7.04 (s, 2H), 2.39 (s, 3H), 2.00 (s, 6H).; C NMR ((126 MHz, CDCl3) δ 163.74,
140.81, 139.35, 134.82, 132.60, 130.24, 129.33, 21.15, 17.31; FTIR (neat): νmax=3147, 2958, 2921,
2646, 2569, 1687, 1539, 1496, 1411, 1353, 1292, 1190, 1033, 935, 850, 771, 582, 547 cm-1.
1-Phenyl-1H-1,2,3-triazole-4-carboxylic acid (PTCA):
White solid, yield: 70.2%; 1H NMR (500 MHz, CDCl3) δ 8.63 (s, 1H), 7.80 (d, J = 7.4 Hz, 2H),
7.57 (dt, J = 30.3, 7.3 Hz, 3H), 7.28 (s, 1H), 13C NMR (126 MHz, CDCl3) δ 163.41, 139.81, 129.83,
129.78, 126.03, 120.92. FTIR (neat): νmax = 3467, 3136, 2927, 2856, 1724, 1691, 1595, 1550, 1498,
1400, 1234, 1039, 983, 850, 756, 682, 563, 509, and 432 cm-1.
1-(3-Nitrophenyl)-1H-1,2,3-triazole-4-carboxylic acid (NTCA):
Off-white solid, yield: 65.4%;1H NMR (500 MHz, DMSO-d6) δ 13.40 (s, 1H), 9.62 (s, 1H), 8.82
(t, J = 2.2 Hz, 1H), 8.46 (ddd, J = 8.2, 2.2, 0.9 Hz, 1H), 8.36 (ddd, J = 8.3, 2.3, 0.9 Hz, 1H), 7.92
13
(t, J = 8.2 Hz, 1H). C NMR (126 MHz, DMSO-d6) δ 161.78, 148.96, 141.42, 137.28, 132.00,
128.21, 127.10, 124.11, 115.90. FTIR (neat) vmax=3576, 3104, 1700, 1533, 1350, 1297, 1254, 1175,
1036, 774, 739, 669, 660, 582 cm-1
3. Results and discussion
3.1. In situ catalyst generation and characterizations
The catalyst is generated in-situ by dissolving 2.0 mM of ligand, 1-mesityl-1H-1,2,3-triazole-4-
carboxylic acid (HTCA) and copper salt (Cu(CH3OO)2) with 2:1 molar ratio in 0.1 M sodium
acetate (Scheme 1). The resulted pale green solution suggested the formation of complex 1, as
established by both Ultraviolet-visible (UV-vis) spectroscopy and high-resolution electrospray
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
ionization mass spectrometry (HR-ESI-MS). UV-Vis spectra of an aqueous solution of HTCA and
copper acetate in 2:1 ratio (in 0.1 M sodium acetate, pH 7.0) showed a broad peak at λmax = 737
nm (Figure S3), a characteristic of a copper(II) dzx,dyz → dx2−y2 transition in a tetragonal ligand
field, in which the copper(II) has
a
distorted square-pyramidal coordination
environment.[15] Beyond a Cu(II): TCA ratio of 1:2, there are barely discernible changes in the
spectra consistent with the formation of a bis complex, [CuII(TCA)2] as the dominant form in
solution (Figure S3), while excess ligand will remain uncoordinated. We have also checked the
catalyst stability in the acetate solution using UV-Vis spectroscopy, the spectra did not show any
noticeable change within 10 days, which suggests the high aqueous stability of the catalyst under
neutral conditions (Figure S4). Furthermore, HR-ESI-MS analysis of the in-situ generated complex
solution in 0.1 M sodium acetate revealed a molecular ion peak at m/z = 276.0732, likely
corresponding with ([Cu(TCA)2+Na++H++5H], z = 2, calc. 276.07611) (Figure 1). Increasing the
molar ratio of ligand: metal to 3:1 did not cause any change in the spectrum. However, this peak
was absent in the spectrum of copper acetate solution when tested under the same conditions
(Figure S5). Since aqua ligand is rather labile and it can easily de-coordinate under operating
conditions of ESI-MS, we have tried the softer matrix-assisted laser desorption and ionization
time-of-flight mass spectrometry (MALDI-TOF MS) to analyze the complex solution. The latter
did not show any peak that corresponds to the diaqua structure of [Cu(TCA)2(H2O)2] at m/z =
561.1522, but instead displayed a peak at m/z = 529.1087, which matches with [Cu(TCA)2+4H]
(Figure 1). The crystalline product was also characterized by FTIR (Figure S6). The IR spectrum
of the free ligand has several bands appearing at 1695, 1546, 1492, and 1415ꢀcm−1 due to
carboxylic C=O, and imine C=N stretching vibrations in the solid state. The two strong absorption
peaks at 1404 and 1610 cm–1 refer to the symmetric and antisymmetric stretching vibrations of the
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
carboxylate groups from the coordinated ligand in the complex structure, respectively, which seem
to shift to different frequencies with respect to the free ligand, further indicating their coordination
with copper.[16] Notably, a broad peak centered in the range of 3200-3650 cm-1 is attributed to the
hydroxyl group from the water molecules of crystallization.
Scheme 1. Synthesis of Cu(TCA)2
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
Figure 1. High Resolution Electro Spray Ionization Mass Spectra (HR ESI-MS) (A, B), and
Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectra (MALDI-TOF-MS) (C,
D) of the complex solutions generated in-situ by direct mixing the ligand (HTCA) and copper
acetate in 0.1 M sodium acetate of pH 7.0 at different ratios of ligand: metal; 2:1 (A, C) and 3:1
(B, D), respectively. Copper concentration was maintained at 1.0 mM in the all cases.
The molecular structure of the Cu(TCA)2 was unambiguously confirmed by single crystal X-ray
diffraction (Figure 2). Crystals suitable for X-ray crystallography were isolated from the aqueous
solution of the complex (1 mM) in acetate electrolyte (0.1 M) at pH 6.9 within 3 months as greenish
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
blue crystals. This supports the notion that the complex is highly stable in buffer solution for a
long time. Interestingly, the crystals isolated from complex solutions of different ratios of ligand
to metal ions (2:1 and 3:1 ligand to metal ion) had the same structure, which agrees with UV-vis
results and confirms that extra ligand remains uncoordinated to the copper ions. The selected bond
lengths and angles of the complex as well as crystallographic data are given in Tables S3 and S4
in supporting information. The complex has a distorted square pyramidal geometry configuration
with two TCA ligands in the basal square plane are bidentatively coordinated to copper through
both nitrogen and carboxylate-oxygen donor atoms. In accordance with spectroscopic
characterizations, there was no coordinated water in the complex structure and instead an oxygen
atom from a carboxylate group of another molecule occupies the apical site. In this case the
carboxylate group of triazole moiety serves as a bridging group to finally make 1D-coordination
chains, where one molecule is nearly perpendicular on the other one (Figure 8b and Figure S7).
This is different from the diaqua cobalt complex of similar structure [Co(TCA)2(H2O)2] which
might be due to the hydrophobicity of Cu(II) as compared to Co(II), also discussed for molecular
Co(TCA)2 and Cu(TCA)2 complexes within the Supporting Information.[2a] This is further
supported by the longer Cu−O bond length of 2.197 Å as compared to Co−O of 2.104 Å in its
cobalt analogues structure. However, the Cu−O bond length in Cu(TCA)2 remains close to several
copper complexes reported for water oxidation, and among the shortest. The TCA ligand
approaches the metal more closely with shorter Cu−N distances ∼1.968−1.998 Å, as compared to
Cu−N distances in some other copper(II) complexes based on pyridine ligands.[17]
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
Figure 2. X-ray crystal structure of Cu(TCA)2.
3.2. Electrochemical behavior and stability
The electrochemical properties of complex 1 are investigated using glassy carbon electrode (3 mm
diameter) as a working electrode and Ag/AgCl and Pt wire as reference and counter electrode,
respectively, in pH 7.5 sodium acetate solution. As shown in Figure 3, cyclic voltammograms
(CVs) of complex 1 displayed large, irreversible oxidative waves corresponding to catalytic water
oxidation. CVs of a solution containing only the ligand HTCA showed featureless voltammogram
demonstrating that the ligand is stable in this potential window and that the complex is the
responsible for the catalytic activity. The water oxidation occurs at an onset potential of 1.75 V vs.
RHE (all the potentials are presented versus reversible hydrogen electrode (RHE)) which
corresponds to a low overpotential (η) of 520 mV. The onset potential is defined as the potential
at a current density that is double to that of the background, approximately 0.2 mA cm-2. This
overpotential competes very well with most reported copper-based molecular catalysts for water
oxidation that operate under neutral conditions (See Table S6 in SI). To accurately compare the
catalytic activity of our catalytic system, the overpotential to attain a 1.0 mA/cm2 current density
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
was determined to be 680 mV, representing a cathodic shift by approximately 130 and 330 mV
with respect to that of the mononuclear [Cu(TMC)(H2O)](NO3)2 and dinuclear [Cu2(BPMAN)(µ-
OH)]3+ copper complexes that are recently reported as highly active catalysts under neutral
conditions.[12a, 12c] Furthermore, this overpotential at 1.0 mA/cm2 current density compares
favorably to bio-inspired Cu4O4 cubane catalysts, which displayed 620 - 760 mV to attain the same
current density, albeit these values were reported under strongly basic conditions (pH 12.0) and
the catalysts are tetranuclear.[18]
CVs of complex 1 at pH 7.5 under nitrogen atmosphere include CuII/I couple at E1/2 = 0.485 V
versus RHE (Figure 3a). This redox couple is different from that of Cu(II) salt, where the latter
was observed at slightly higher potential. The catalyst CVs at pH > 7 indicate an electrochemical-
chemical-electrochemical (ECE) mechanism for the Cu(II/I) couple, whereas at lower pH it looks
electrochemically reversible i.e. electrochemical-chemical (EC) mechanism, which may refer to a
change in the coordination mode of the ligand at different pHs (Figure 3b). In the ECE mechanism,
the oxidation of a Cu(I) to Cu(II) species is accompanied by chemically irreversible formation of
a new species at or near the electrode surface. This chemical irreversible process could be a partial
dissociation of one TCA ligand. In such case, two forms of the complex will exist at the electrode
interface, i.e. the starting complex Cu(TCA)2 and another species with partially dissociated TCA
ligand. This could explain the appearance of a second peak at 2.10 V after the main catalytic wave,
which can be assigned to water oxidation due to the starting complex, while the first one is due to
water oxidation by the defected structure. This is in a good agreement with DFT calculations,
where the defected structure has much lower overpotential as compared to the original complex.
In contrast, at lower pH values (pH<7.0), the Cu(II/I) couple is electrochemically reversible and
the complex maintains its coordination environment, thus its cyclic voltammogram has only one
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
catalytic peak. In EC mechanisms; at the slower scan rates, enough time passes between the
initiation of oxidation process and the reversed reduction of its oxidized form to allow complete
chemical conversion, and no return reduction is seen. For fast scan rates, the follow-up step can be
outrun, and all oxidized forms are reduced on the return scan as shown in Figure 4. Considering
the pH-dependent structure of the catalyst, it is thus expected to have different redox peaks for
Cu(II/I) redox couple, since the complex will have different coordination environment based on
the electrolyte pH. At pH>7 (pH 7.2-7.5), in the forward scan Cu(II/I) redox couple has the same
potential ~ 0.35 V (vs. NHE), while at pH <7 (i.e. pH 6.5), the peak was shifted to 0.198 with ~
150 mV difference, which can be only explained as results of changing the complex structure.
Furthermore, the variation of the onset potential with pH value has a slope of 58 mV/pH unit
(Figure 3B), which is mostly correlated to one electron oxidation of copper to higher oxidation
state, directly before water oxidation, that is coupled with one proton transfer. The peak currents
(ip) of the non-catalytic CuI/CuII waves at 0.61 V vary linearly with the square root of scan rates
(ν1/2) varying from 0.005 to 0.5 V s−1 (Figure 4), indicating that that the electrochemical process is
diffusion limited, and thus homogeneous.[19] A diffusion coefficient of 6.5×10-6 cm2 s-1 was
determined from the dependence of the ip on ν1/2 based on the Randles–Sevcik equation (eq. 1),
where n=1, is the number of electrons transferred, F is the Faraday constant, A is the area of the
working electrode (0.07 cm2), [cat] is the concentration of the catalyst (mol L-1), DCu is the catalyst
diffusion coefficient, R is the ideal gas constant, and T is the temperature. In addition, the catalytic
current (ic) (at 1.98 V) and the peak current (ip) of noncatalytic CuII reduction (at 0.61 V) both
displayed linear relations with the catalyst concentration from 0.25 to 1.0 mM (Figure S9). This
first-order kinetics also suggests a single site mechanism for water oxidation.[8b] Additionally, a
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
turn over frequency (TOF) (kcat) of 0.6 s-1 could be estimated from the catalyst dependence plots,[20]
which is comparable to data reported for other copper-based WOCs.[9, 12a, 21]
푖푝 = 0.446 푛퐹퐴[퐶푎푡] (푛퐹휈퐷퐶푢
)
(1)
1/2
푅푇
Figure 3. Electrochemical characterizations of complex 1. A) CV of solution containing 1.0 mM
Cu and 2.0 mM TCA in 0.1 M sodium acetate solution using glassy carbon electrode (GCE, 0.07
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
cm2), scan rate 100 mV s-1 at pH 7.5 in comparison to blank acetate solution (Blank AS) and TCA
ligand, the arrow indicates the initial sweep direction. B) CVs of the complex were recorded at
different pH values, the inset figure shows the plot of onset potential peak at different pH values.
Figure 4. (A) Cyclic voltammograms of complex 1 (1.0 mM) in 0.1 M acetate solution at pH 6.8
with scan rate varying from 10 to 500 mV s−1. (B) Plot of the anodic current density maximum of
the CuI/CuII couple as a function of the square root of scan rate.
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
We have further studied the effect of ligand structure and thus electronic environment around the
copper metal on complexes’ redox properties and catalytic activities towards water oxidation.
Three ligands were screened (Figure S10), where slight shifts in CuII/I redox peaks could be
observed between the screened catalysts. In addition, the non-substituted 1-phenyl-1H-1,2,3-
triazole-4-carboxylic acid ligand (PTCA) showed slightly lower onset potential (~ 60 mV) as
compared to the one with electron donating groups (TCA) or electron withdrawing group (1-(3-
nitrophenyl)-1H-1,2,3-triazole-4-carboxylic acid, NPTCA). However, one can observe that
introducing electron donating or withdrawing groups into phenyl ring has much less impact on the
onset potential as well as redox peaks as compared to some other copper catalysts,[21b] which might
be due to the indirect interaction of phenyl group with the metal ion, and thus substituents in the
triazole moiety might have more impact.
To check the catalyst stability, a water electrolysis experiment was performed, in a two-
compartment cell, followed by re-precipitation of the ligand via acidifying the complex solution
with acetic acid. NMR spectroscopy indicated that the ligand maintained its original structure
which supports the good stability of the complex under operating conditions. However, when
electrolysis is performed in a one compartment cell, a black precipitate was observed at the counter
platinum electrode which might be due to reduction of Cu(II) ions into Cu metal, and releasing
free ligand which also appeared as white precipitate in the electrolyte. To further check whether
this white precipitate is the original ligand or any oxidized forms. It was isolated and characterized
using 1H-NMR which confirms that the isolated ligand is the original ligand (Figure S11). These
findings demonstrate the nice stability of not only the complex but also the ligand under operating
conditions. Similar Cu(II) reduction was also observed for other copper systems on the counter
electrode.[6, 9]
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
The catalyst performance was evaluated via the long-term electrolysis. The experiment was
performed at a fixed potential (2.0 V versus RHE) in 0.1 M acetate electrolyte (pH 7.0) containing
1 mM of catalyst using high surface area ITO electrode (1 cm2). In the beginning a slight increase
in the catalytic current was observed which might be due to the formation of some electroactive
species near to the electrode surface, as already observed with some other molecular systems in
electrochemical water oxidations.[22] Afterwards, the system had a stable oxygen evolution current
density of about 1 mA/cm2 for more than 15 h, with a passed charge of 40 C/cm2 (Figure 5A).
During electrolysis, gas bubbles were clearly observed, and these bubbles were confirmed as
oxygen molecules by GC. To further investigate whether the observed catalytic current arises from
the complex or deposition products on ITO surface such as copper oxide/hydroxide, the as-used
ITO electrode was removed from catalyst solution after 15 h of electrolysis, thoroughly rinsed with
water and then tested in catalyst free acetate buffer under the same conditions. No apparent O2
evolution current was observed in this case, which suggests that the observed catalytic current
when using complex 1 is mainly from the complex and the complex is most likely homogenous in
nature. After long–term CPE, UV-vis spectra of the resulting solution also confirm the stability
and homogeneous nature of the catalytic system and no change or decrease was observed in the
absorbance of the Cu(II) d−d transitions peak (Figure S12).
Through the demonstrated catalyst stability during bulk electrolysis, we have observed film
deposition/adsorption on glassy carbon electrode during cyclic voltammetry experiments. The film
growth was observed in different pH ranges, however, to monitor this film growth using linear
sweep voltammetry (LSV), we work in slightly acidic conditions. For this purpose, consecutive
CV cycles of complex 1 in sodium acetate solution (pH=6.0) were swept, in the potential window
0.25 to 2.05 V using GCE (0.07 cm2) at 100 mV s-1 scan rate. Thereafter, the electrode was
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
thoroughly rinsed with water, but not polished, and then LSV was recorded in blank sodium acetate
solution (pH 6.0) after several numbers of CV cycles, also LSV of the catalyst solution was
recorded for comparison.[23] Firstly, LSV of complex 1 solution (pH 6.0) showed an oxidative
wave with an onset potential of about 1.91 V and a wave catalytic current of 0.5 mA/cm2 at 2.07
V. As the number of CV scans increases, the oxidation wave increases to six-fold in magnitude
and shifts to lower onset potentials within the screened 100 CV cycles (Figure 5B). Thus, the
grown film after 100 CV cycles demonstrated a low onset potential of 1.51 V at pH 6.0 which
corresponds to an overpotential of 285 mV (measured at current density j = 0.20 mA cm−2). In
addition, the overpotential required, for the deposited film, to reach a catalytic current density of
1 mA cm−2 is about 720 mV, which is smaller than that observed for most previously reported Cu-
based molecular WOCs especially those measured at near-neutral conditions.[24]
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
Figure 5. A) Controlled potential electrolysis (CPE) of 1.0 mM catalyst solution in acetate solution
of initial pH 7.0, as compared to as-used ITO under similar conditions, inset shows the amount of
charge passed during controlled potential electrolysis experiment in both cases, B) Linear sweep
voltammetry (LSV) collected using a freshly polished GCE in 1.0 mM solution of complex 1 in
0.1 M sodium acetate solution of pH 6.0 (blue), and a previously swept electrode for 10 (red), 30
(pale blue), 50 (pink), and 100 consecutive CV scans (green) in catalyst solution then tested in
blank acetate solution.
This phenomenon of film deposition/adsorption opposes the fact that a film deposition and current
enhancement are mostly indications of catalyst decomposition and the formation of metal oxides
on the electrode surface. However, taken into account the stability of the complex in bulk
electrolysis experiments and the absence of any electrochemical active deposits on ITO electrode
surface after 15 h of bulk electrolysis. In addition, glassy carbon electrode surface is hydrophobic
while that of ITO is hydrophilic,[25] and knowing that the ligand and thus the complex are also
hydrophobic[2a] suggest that a physisorption of either the starting complex or some
electrochemically generated species on the glassy carbon electrode might be responsible for the
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
current enhancement with consecutive CV scans. This is not unprecedented result, since cobalt
complex of the same ligand [Co(TCA)2(H2O)2] has assembled a catalyst film on electrode surface,
which proved not to be cobalt oxide/hydroxide as normally expected, but instead, it generated a
molecular cobalt specie that incorporated the organic ligand bound to cobalt ions.[2a] Further, we
recently reported a similar phenomenon in a ruthenium based catalytic system, where the partial
ligand dissociation enables surface generation of electrochemically active species on the surface
of GCE when an external potential was applied.[26] Likewise, we have reported the activity of two
similar Co(phenanthroline) catalysts which assemble into metal organic films at the electrode
surface from their aqueous solutions under external potential bias, where OH substituent groups
in Co(1,10-phenanthroline-4,7-diol)2 complex provided excellent electrode anchor points and can
thus lower the charge transfer resistance and lead to increased water oxidation activity.[27]
One extra indication pointing towards the molecular species deposition rather than CuOx/Cu(OH)2
is obtained by studying the deposition of the free ligand itself under similar conditions. To
investigate the possible attachment of TCA ligand onto GCE, the ligand solution was repeatedly
swept in the same potential window and under similar conditions to that used in case of the catalyst
deposition. In agreement with our expectations, a gradual increase in the current of cyclic
voltammogram on consecutive scans was observed suggesting adsorption of some species on the
surface of GCE. The film growth was monitored using electrochemical impedance spectroscopy
(EIS), since ligand adsorption at the GCE might alter the electrode/solution interface and thus it
changes transfer resistance through the electrode. The EIS measurements are performed at an
applied potential of 2.05 V vs. RHE in a frequency range of 100 000 to 1 Hz (Figure 6). For this
purpose, the film is grown via a number of CV cycles, then its EIS is recorded in a catalyst-free
sodium acetate solution (pH 6.0). As indication of film growth, the charge transfer resistance
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
decreases with the film growth and reaches a minimum after around 80 segments, which suggests
the complete coverage of the electrode with a single layer of the ligand at this stage. After 80
segments, the charge transfer resistance slightly increases which might be due to starting the
adsorption of multilayers of the ligand. From the Nyquist plot, it is clearly shown that a film of
ligand has a substantially lower charge transfer resistance as compared to bare GCE and thereby
faster electron transfer rate.
Figure 6. Nyquist plots of the films of free ligand deposited after number of CV segments in a
ligand free sodium acetate electrolyte (0.1 M) (pH = 6.0) at an applied potential of 2.05 V (vs.
RHE).
The deposited catalyst film on the GCE surface after 100 CV segments was analyzed after washing
and drying. Scanning electron microscope (SEM) revealed the homogenous coverage of the
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
electrode surface with the deposited materials (Figure S13). Further, Energy-Dispersive X-ray
spectroscopy (EDX) did not provide accurate results about the chemical composition of the film,
mostly due to the very low film thickness and the fact that this technique probes deeply into the
surface, which is mainly carbon and oxygen (from glassy carbon electrode). However, X-ray
photoelectron spectroscopy (XPS) probes only a few nanometers below the surface, and is more
sensitive as compared to EDX.[28] High resolution XPS demonstrates the presence of copper and
nitrogen in the deposited film with Cu:N ratios ranging from 1:1.9 to 1:2.26 based on several scans
in different points of the electrodeposited film (Table S5). The presence of nitrogen in the film
suggests the incorporation of the ligand into the film, which might exist complexed with copper
or as a free ligand adsorbed into CuOx/Cu(OH)y. This can be confirmed based on the binding
energies of both elements. The observed high carbon and oxygen content in the survey along with
the detection of some other elements arising from GCE background strongly recommend the
ultrathin nature of the deposited film (Figure S14), which also agrees with the results from EDX.
The low N:Cu ratio might be due to the deposition/adsorption of a molecular catalyst with 1:1 ratio
of copper and ligand as a result of the steric hindrance of mesityl group (which agree with the
structural defects proposed by DFT calculation, as will be discussed in computational studies
section) or it might be due to deposition of some other species such as CuOx or Cu(OH)2 along
with the starting complex. The low sensitivity of the technique to low atomic weight elements such
as N might be another reason. High resolution XPS of copper revealed the presence of two main
peaks at 935.0 and 954.9 eV for Cu 2p3/2 and Cu 2p1/2, respectively. These binding energies are
∼0.5−1.5 eV higher than those typically observed for Cu+ and Cu2+ oxidation states in Cu2O and
CuO, however, they are consistent with the values that have been observed for Cu2+ complexes.[29]
In addition, the broad shakeup satellite peaks at 939.8 and 344.1 eV for 2p3/2 peak, and the ratio of
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
the satellite peaks to the main peak are characteristic for Cu(II) coordination compounds with NO
coordination sites.[30] Further, we can observe a shoulder peak at 932.4 eV, which might be
indication for the incorporation of copper oxide species into the film.[31] The N 1s peak was
deconvoluted into three components at 399.4, 400.7, and 402.2 eV, where the two peaks at high
binding energy (400.7 and 402.2 eV) can be assigned to free nitrogen atoms from the triazole
structure.[32] A shift in the binding energy of metal bound nitrogen has been reported in literature
to a lower value of about 399.2 eV. Thus, the third nitrogen atom from the triazole moiety most
likely is coordinated to copper ions in the deposited film (Figure 7).[33] The O1s peaks of CuxO
and Cu(OH)2 appear in the range of 529.6-531.2 eV.[34] Thus, the high binding energy of O1s
observed at 533.6 eV mainly arises from carboxylate group, while, the other contribution appearing
at 531.9 eV might originate from the water of crystallization impeded into the film. The C1s core-
level spectrum displayed several peaks owing to different functionalized carbon atoms. The
spectrum suggests the presence of Csp2, Csp3, C=N, and COO in the deposited film, as observed
in the peaks at 284.8, 285.3, 286.8, and 288.8 eV, respectively, along with π-π* shake-up satellite
peak at 292.5 eV, which is characteristic for conjugated π-systems. Although XPS analysis could
establish that the thin film deposited on the surface of GCE has a metal-organic nature and combine
both the organic ligand and Cu2+ ions bound together, we can’t exclude the deposition of metal
oxide/hydroxide. In such case, CuOx/Cu(OH)2 might be generated from the decomposition of the
copper complex which is then re-dissolved into the electrolyte, leaving the metal-organic film of
very low thickness.
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
Figure 7. High resolution XPS of Cu 2p (A), N1s (B), C1s (C), and O1s (D) of the deposited film
on GCE. The film was deposited via 100 consecutive CV cycles in 1.0 mM solution of complex 1
in 0.1 M sodium acetate solution of pH 6.0.
3.3. Computational studies
To elucidate the potential water oxidation mechanism on complex 1, DFT calculations were
performed on the homogenous Cu(TCA)2 complexes, with according to the same methodology
which had been applied to study Co(TCA)2 complexes with similar morphology (see electronic
supporting information for more details).[2a] It is noteworthy to mention that for the Cu(TCA)2
complexes in gas phase, much higher computational overpotential was observed than for the
Co(TCA)2 complexes, we refer to SI for an elaborate discussion. Strikingly, the experimental OER
behavior shows onset overpotential of even 0.29 V. The reason for this computational/experimental
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
mismatch could be that we are considering the wrong type of active site. Within the field of
coordination polymers and metal-organic frameworks, the catalytic activity arises mostly from
defects within the structures.[35] Therefore, we decided to study defects within the periodic
Cu(TCA)2 films (Figure 8b-c). The respective intermediates (M-*, M-OH, M=O, M-OOH) of the
films are shown in Figure 9. When a TCAdefect is created, a Cu+ site appears within the framework
and at this active site, the OER reaction can proceed with much lower overpotential (Figure 8a).
The first step (M-*  M-OH) and third step (M=O  M-OOH) costs about 1.54 eV while the
second step (M-OH  M=O) requires only 1.3 eV. Therefore, the rate limiting step requires an
overpotential of 0.31 V, which forms the lower limit. Furthermore, this potential lies very close
proximity of the measured onset overpotential of 0.29 eV (see Figure 5B). Therefore, TCAdefects
might thus explain the low overpotential behavior of the Cu(TCA)2 film.
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
Figure 8. The impact of TCA defects on the oxygen evolution reaction (OER) comparing the
defective Cu(TCA)2-x with Cu+ active site with the defect-free Cu(TCA)2 with Cu2+ active sites
and the graphical representation of the (b) Cu(TCA)2-catalyst crystal without defect and (c)
Cu(TCA)2-x with one TCA defect in a unit cell containing 8 Cu(TCA)2 units. Color code: Cu
(orange), O (red), N (blue), C (gray), H (white).
This article is protected by copyright. All rights reserved.

10.1002/cssc.202001444
ChemSusChem
Figure 9. Atomic representation on the optimized species within the periodic lattice. Color code:
Cu (orange), O (red), N (blue), C (gray), H (white).
4. Conclusions
In summary, we reported a triazole-based copper molecular catalyst for electrochemical water
oxidation at neutral conditions. The catalyst was simply generated in-situ from its precursors in
acetate solution, and it showed very low overpotential as compared to reported copper-based
WOCs under similar conditions. The catalyst achieves a stable oxygen evolution current density
of about 1 mA/cm2 at only 1.59 V (versus NHE) under neutral conditions, and durable stability
during bulk electrolysis for at least for 15 hours without remarkable decrease in the catalytic
current, which represents remarkable progress in water oxidation by earth-abundant metals. In
contrast, negligible current was observed in the absence of the catalyst, even when using the ITO
electrode that was previously used for bulk electrolysis without polishing. This result also reveals
that the observed catalytic current when using complex 1 is mainly from the complex and the
complex is homogenous in nature. The catalyst displayed first-order kinetics suggesting a single
site mechanism for water oxidation with a TOF (kcat) of 0.6 s-1. On sweeping the complex solution
This article is protected by copyright. All rights reserved.