Immobilization of a Molecular Re Complex on MOF-derived Hierarchical Porous Carbon for CO2 Electroreduction in Water/Ionic Liquid Electrolyte

Full text of DOI:10.1002/cssc.202002014

Chemistry–Sustainability–Energy–Materials
Accepted Article
Title: Immobilization of a molecular Re complex on MOF-derived
hierarchical porous carbon for CO2 electroreduction in water/ionic
liquid electrolyte
Authors: Marc Fontecave, Domenico Grammatico, Huan Ngoc Tran,
Yun Li, Silvia Pugliese, Laurent Billon, and Bao-Lian Su
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.202002014
Link to VoR: https://doi.org/10.1002/cssc.202002014
01/2020

10.1002/cssc.202002014
ChemSusChem
1
2
3
Immobilization of a molecular Re complex on
MOF-derived hierarchical porous carbon for
CO2 electroreduction in water/ionic liquid
electrolyte
4
5
6
7
8
9
Domenico Grammatico, a c Dr. Huan Ngoc Tran, b Dr. Yun Li, b Silvia Pugliese, a b
Prof. Laurent Billon, c Prof. Bao-Lian Su a * and Prof. Marc Fontecave b *
10
11
a Laboratory of Inorganic Materials Chemistry (CMI), University of Namur, 61 rue de
Bruxelles, B‐5000 Namur, Belgium.
12
13
14
15
16
b Laboratoire de Chimie des Processus Biologiques, UMR CNRS 8229, Collège de France-
CNRS-Sorbonne Université, PSL Research University, 11 Place Marcelin Berthelot, 75005
Paris, France.
17
18
19
c Bio-inspired Materials Group: Functionality & Self-assembly, Université de Pau et des Pays
de l’Adour, E2S UPPA, CNRS, IPREM UMR 5254, 64000, PAU, France
20
21 * to whom correspondence should be addressed : marc.fontecave@college-de-france.fr;
22 bao-lian.su@unamur.be
23
24 KEYWORDS: CO2 electroreduction; catalysis; rhenium complex; heterogenization;
25 hierarchical porous carbon; ionic liquid
26
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
27 Abstract
28 The development of molecular catalysts for CO2 electroreduction within electrolyzers requests
29 their immobilization on the electrodes. While a variety of methods have been explored for the
30 heterogenization of homogeneous complexes, we here report a novel approach using a
31 hierarchical porous carbon material, derived from a Metal Organic Framework, as a support for
32 the well-known molecular catalyst [Re(bpy)(CO)3Cl] (bpy = 2,2’-bipyridine). This cathodic
33 hybrid material, named Re@HPC, has been tested for CO2 electroreduction using a mixture of
34 an ionic liquid (1-Ethyl-3-methylimidazolium tetrafluoroborate, EMIM) and water as the
35 electrolyte. Interestingly, it catalyzes the conversion of CO2 into a mixture of carbon monoxide
36 and formic acid, with a selectivity that depends on the applied potential. The present study thus
37 reveals that Re@HPC is a remarkable catalyst, enjoying excellent activity (turnover numbers
38 for CO2 reduction of 7835 after 2 h at -1.95 V vs Fc/Fc+ with a current density of 6 mA cm-2)
39 and good stability. These results emphasize the advantages of integrating molecular catalysts
40 onto such porous carbon materials for developing novel, stable and efficient, catalysts for CO2
41 reduction.
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
42 Introduction
43 Carbon dioxide electroreduction into energy-dense carbon-based liquid or gaseous products is
44 an attractive way to store renewable energies into chemical energy. However, because of the
45 high stability of CO2 and the requirement of multiple electron- and proton-transfers for its
46 transformation, catalysts are required to overcome the slow kinetics, minimize activation
47 energies and overpotentials and control product selectivity. While heterogeneous catalysts are
48 generally favored due to their stability as well as facile product and catalyst recovery,
49 homogeneous molecular metal complexes (coordination and organometallic complexes) have
50 also been developed for CO2 electroreduction since the 1980’s. In particular, they offer the
51 unique opportunity to tune the coordination environment of the metal center and its reactivity
52 via synthetic modifications of the ligands. The molecular strategy has been recently described
53 in review articles.[1,2]
54 To reconcile these two approaches, homogeneous catalysts have sometimes been immobilized
55 on heterogeneous conductive supports in order to provide original hybrid electrodes. However,
56 this strategy has been applied to few molecular catalysts with limited success so far in the
57 specific case of CO2 electroreduction.[3–5] To illustrate the various techniques developed for
58 providing access to such hybrid electrodes, the case of the prototypical catalyst
59 [Re(bpy)(CO)3Cl] (bpy = 2,2’-bipyridine), highly selective for CO2 to CO conversion, is
60 interesting, since its immobilization has been studied using a variety of approaches. The first
61 attempts to immobilize this complex were achieved by Meyer and coworkers[6] and Abruña
62 and coworkers [7] in 1985 and 1986 respectively, two years after the initial publication by Lehn
63 reporting the catalytic activity of this complex for CO2 reduction.[8] In both cases, the catalyst
64 was immobilized via electropolymerization of [Re(vbpy)(CO)3Cl] (vbpy = 4-vinyl-4'-methyl-
65 2,2'-bipyridine) on an electrode to afford catalytic poly-[Re(vpbpy)(CO)Cl] films, allowing
66 electroreduction of CO2 to CO with very high faradaic yields in CH3CN. However, these films
67 were shown to exhibit limited stability. A similar polymerization approach was recently
68 developed by Kubiak and co-workers using 5-ethynyl-bipyridine derivatives, instead of vinyl-
69 bipyridine, however in that case very fast deactivation of the catalytic films was observed.[9] A
70 different strategy for electropolymerization of a variant of [Re(bpy)(CO)3Cl] was reported by
71 Deronzier and co-workers.[10] Upon electrochemical oxidation of a solution of
72 [Re(pyrbpy)(CO)3Cl] (pyrbpy = 4-(4-Pyrrol-l-ylbuty1)4’-methyl-2,2’-bipyridine) using a Pt
73 electrode, a polypyrrole-based film was deposited and found to catalyze the electroreduction of
74 CO2 in CH3CN leading to CO as the only reduction product. However, the polymer was highly
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
75 unstable and substantial loss of current was observed during electrolysis. In 1993, Kaneko and
76 co-workers reported the immobilization, via simple adsorption, of [Re(bpy)(CO)3Br] within a
77 Nafion© membrane, which is stable in various solvents and easily interfaces with electrode
78 surfaces.[11] The polymer-confined catalyst was active for CO2 electroreduction in neutral
79 aqueous electrolyte, leading to mixtures of CO, H2 and formic acid, with a selectivity depending
80 on the applied potential.
81 Besides this immobilization strategy involving polymers and membranes, covalent grafting to
82 the electrode has also been explored. As o-quinone moieties found on the edge planes of
83 graphite can condense with substituted o-phenylenediamines, [Re(5,6-diamino-
84 phenanthroline)(CO)3Cl] was used to modify graphite electrodes.[12] The increased conjugation
85 between the catalyst and the electrode was proposed to overcome poor conductivity issues in
86 some polymeric films deposited onto electrodes. This system was very selective for CO
87 formation (faradaic yield of 96%) in CH3CN with high turnover numbers (TONs). A conjugated
88 polymer material incorporating [Re(bpy)(CO)3Cl] motifs and covalently attached to a glassy
89 carbon electrode was obtained by electropolymerization of a Re complex containing 2,2’-
90 bipyridine-5,5’-bis-(diazonium) ligands.[13] These polymer films proved highly stable, allowing
91 thousands of turnovers for selective production of CO in CH3CN. As a last illustration of this
92 class of grafting methods, a [Re(4-(4-aminophenetyl)-4’-methyl =-2,2’-bipyridine)(CO)3Cl]
93 complex has been attached to a high surface area porous carbon electrode via the
94 electrooxidation of the amino group.[14] While efficient for selective CO2 to CO conversion,
95 very fast leaching of the complex into the bulk solution was observed after few minutes of
96 electrolysis.
97 Finally, non-covalent immobilization was achieved, exploiting π-π interactions between
98 pyrolytic graphite and [Re(pyrene-bpy)(CO)3Cl], containing a pyrene-substituted bpy.[15] This
99 system was catalytically active for CO production but was rapidly deactivated, likely because
100 of reduction of the pyrenyl moiety. The best result with such a non-covalent immobilization
101 strategy was obtained by Kubiak and co-workers who developed a hybrid electrode resulting
102 from the incorporation of [Re(tBu-bpy)(CO)3Cl], with tBu-bpy = 4,4’-di-tert-butyl-2,2’-
103 bipyridine, into multi-walled carbon nanotubes (MWCNTs). This material proved stable and
104 active in an aqueous electrolyte for highly selective CO production, at relatively low
105 overpotential and with a current density of 4 mA cm-2.[16]
106 One of the drawbacks of all these immobilization strategies, with the exception of the MWCNT/
107 [Re(tBu-bpy)(CO)3Cl] system, resides in the need to introduce functionalities to the bipyridine
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
108 ligand. This not only generates synthetic issues but also results in significant modifications of
109 the electronic properties of the catalyst as well as of its reactivity. We thus thought of
110 developing methods that allow the immobilization of the [Re(bpy)(CO)3Cl] itself without
111 synthetic modifications. Here we report the preparation and characterization of an original
112 conductive carbon porous material and its utilization as a support for the [Re(bpy)(CO)3Cl]
113 catalyst, which can be easily fixed within the pores of the solid material. The novel hybrid
114 material proved stable and active as a catalyst for CO2 electroreduction, with high current
115 densities and high turnover numbers. Interestingly, this catalyst converts CO2 into a mixture of
116 CO and HCOOH, whose ratio depends on the applied potential.
117
118
119
120 Scheme 1: Synthesis of HPC and Re@HPC.
121
122 RESULTS AND DISCUSSION
123 Synthesis and characterization of the hierarchical porous ZIF8-derived carbon-HPC
124 The synthesis of hierarchical porous carbon (named HPC in the following) derived from the
125 Metal Organic Framework ZIF8 was synthesized following a previous report,[17] however with
126 significant modifications (Scheme 1). Briefly, the ZIF8 precursor was first infiltrated into the
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
127 polystyrene spheres (PS) template and crystallized using a solution of CH3OH/NH3·H2O.
128 Second, a treatment at 450 °C under argon atmosphere, instead of washing with organic
129 solvents used in the previous report,[17] was carried out in order to remove the PS template and
130 thus obtain the hierarchical porous ZIF8 (HPZIF8). HPZIF8 shows slightly lower BET surface
131 area (Figure S1 and Table S1) as compared to the previous report,[17] likely due to remaining
132 PS beads in the sample. Scanning electronic microscope (SEM) and powder x-ray diffraction
133 (XRD) characterizations of HPZIF8 show expected well-defined ZIF8 crystals (Figures S2-S3).
134 HPZIF8 was finally pyrolyzed under an inert atmosphere at 800 °C to obtain the final
135 hierarchical porous carbon (HPC) material. HPC was characterized by various techniques. XRD
136 pattern confirmed that the material was not crystalline anymore (Figure S2). SEM images
137 (Figure 1a and Figure S3c) showed well-defined and well-distributed carbon particles with a
138 three-dimensional tetrakaidecahedron morphology and with a uniform monodisperse diameter
139 of 2 µm. This specific morphology is due to the use of PS as a template which controls the
140 shape of the material.[17] The highly ordered porous carbon material showed a hexagonal close-
141 packed arrangement of pores (inset Figure 1a) with interconnected windows and porosity
142 uniformly distributed.[18,19] N2 physisorption analyses (Figure 1b and Figure S1) allowed to
143 establish a BET surface area of 495 m2 g-1 and total pore volume of 1 cm3 g-1 with around 25%
144 of the total pore volume assigned, by the Horwart-Kawazoe method, to the micropores. The
145 increase of the surface area due to the pyrolysis step leading to HPC (495 m2 g-1) from HPZIF8
146 (145 m2 g-1) is likely due to a complete removal of PS. N2 adsorption isotherms are of typical
147 type I, indicating the microporous character of the material with a pore size of 0.5 nm. The
148 presence of mesopores and macropores has been revealed by two N2 adsorption uptakes at p/po
149 of 0.5 and 0.95-1.0, respectively, as shown in Figure 1b. The BJH analysis method gave a very
150 sharp peak centered at 5 nm, corresponding to the hysteresis loop at p/po of 0.5, [20] while the
151 sharp N2 adsorption uptake at p/po of 0.95-1.0 is generated by the macropores with a diameter
152 of 380 nm observed by SEM (Figure 1a). The above analysis suggests HPC material contains
153 a hierarchical macro-meso-microporous structure.
154
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
155
156 Figure 1: a) SEM image of HPC with enlargement (inset); b) N2 adsorption-desorption
157 isotherms of HPC and Re@HPC and pore size distribution for HPC (inset) made by BJH
158 method.
159
160 Functionalization of HPC material with a molecular complex
161 HPC was then used as a support for heterogenizing the molecular [Re(bpy)(CO)3Cl] complex,
162 named Re in the following. Re was immobilized without any further functionalization taking
163 advantage of the interconnected porosity of the support which improves the hosting properties.
164 The heterogenization of Re on HPC has been carried out as follows. Briefly, Re was first
165 dissolved in CH2Cl2 and the solution added dropwise to a sonicated dispersion of HPC in the
166 same solvent, with an initial Re:HPC weight ratio of 1:9 (Scheme 1). After stirring the solution
167 for 1 h and centrifugation the as-obtained solid hybrid system, named Re@HPC, was dried
168 overnight at room temperature. Then, Re@HPC was characterized not only to check for the
169 presence of the molecular complex within the solid but also to define morphology and structure
170 of the final material.
171 SEM images of Re@HPC showed that the morphology of the carbon particles was retained
172 (Figure S4). The presence of Re within the solid has been verified by Energy Dispersive X-Ray
173 mapping (EDX) (Figure S4) showing a uniform dispersion of the complex, as well as the
174 presence of residual Zn from ZIF8, even after pyrolysis. N2 physisorption analysis after Re
175 loading showed a decrease of the BET surface area in agreement with the incorporation of the
176 molecular complex within the pores of the support (Figure 1b). The integrity of Re within the
177 solid material was proved by 1H NMR spectroscopy after releasing the complex from Re@HPC
178 by washing with CDCl3. The spectrum of the solution proved identical to that of the pure freshly
179 prepared homogeneous complex (Figure S5). Finally, the amount of extractible complex from
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
180 30 mg of Re@HPC using an organic solvent (CH2Cl2) was determined by UV-Visible
181 spectroscopy from the intensity of the absorption band at 387 nm, characteristic of the complex
182 (Figure S6). We thus determined that the sample contained 4.8 µmol of complex leading to 0.16
183 µmol mg-1 of complex within Re@HPC.
184
185 Preparation and characterization of the working Re@HPC/GDL electrode
186 An Re@HPC/GDL electrode (1 cm2) was prepared by deposition of 5 mg Re@HPC on a
187 commercial gas diffusion layer (GDL), as described in the experimental section. The total
188 amount of Re in that electrode was determined after extraction by UV-Visible spectroscopy as
189 described above for Re@HPC (Figure S6). The electrode was shown to contain 0.6 mol cm-2.
190 The CV of Re@HPC/GDL electrode in 0.1 M TBAPF6 in CH3CN at low scan rate of 10 mV s-
191 1 (Figure S7) served to calculate the density of electroactive species from the charge integration
192 of the peak at about -1.6 V vs Fc/Fc+, corresponding to the reoxidation of the complex.[21] The
193 data led to a surface density of 20 nmol cm-2, corresponding to approximately 3% of the total
194 amount of complex present in the solid material. These numbers compare well with those
195 determined for the best hybrid [Re(tBu-bpy)(CO)3Cl]/MWCNT material (13 nmol cm-2, 1-8%
196 of the total catalyst loaded) reported by Kubiak et al.[16] Finally, the electrochemical active
197 surface area of the electrode was determined as 2.18 cm2 per 1 cm2 of geometric surface area
198 by a standard method (Figure S8).
199
200 Electrocatalytic CO2RR in water/1-Ethyl-3-methylimidazolium tetrafluoroborate
201 (EMIM) electrolyte
202 Since the electroreduction of CO2 using the Re@HPC/GDL in water yielded mainly hydrogen
203 (data not shown), the reaction was studied in a CO2-saturated solution of an ionic liquid in the
204 presence of a small amount of water as a source of protons. Indeed, ionic liquids have been
205 shown to facilitate CO2 electroreduction catalyzed by solid catalysts due to their ability to
206 increase CO2 solubility and activate CO2.[22] A previous report showed that using an ionic liquid
207 as both the solvent and electrolyte resulted in decreased overpotential and increased second-
208 order rate constant for CO2 electroreduction to CO catalyzed by [Re(bpy)(CO)3Cl].[23]
209 However, with such homogeneous systems, the viscosity of the medium results in strong
210 limitations in mass transport (small diffusion coefficients) of the molecular catalyst and low
211 current densities. This issue does not apply to an immobilized molecular complex. Thus, a
212 mixture of an ionic liquid (1-Ethyl-3-methylimidazolium tetrafluoroborate) and water (named
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
213 H2O/EMIM in the following) has been chosen as the electrolyte for further CO2
214 electroreduction studies using the Re@HPC/GDL electrode. The Linear Sweep
215 Voltammograms (LSVs) of the Re@HPC/GDL electrode in 5% v/v H2O/EMIM saturated with
216 either Ar or CO2 are presented in Figure 2a. A catalytic wave at an onset potential of -1.55 V
217 vs Fc/Fc+ and developing further up to an applied potential of -2.05 V was observed only in the
218 presence of CO2 and was thus assigned to CO2 reduction (Figure 2a). At potentials more
219 negative than -2.1 V, a second wave was observed however also present for the Re@HPC/GDL
220 electrode under argon, thus likely corresponding to proton reduction to H2. Unfortunately, the
221 evaluation of overpotential values is not trivial here as the equilibrium potentials of the
222 CO2/HCOOH and CO2/CO couples in H2O/EMIM are unknown.
223 Controlled Potential Electrolysis (CPE) at different potentials between -1.75 V and -2.05 V vs
224 Fc/Fc+ coupled with quantification of CO2 reduction products has been carried out for 15 min
225 in order to characterize the CO2 reduction reaction. As expected the total current density
226 increased with increased driving force up to 11 mA cm-2 at -2.05 V and was found to be stable
227 during electrolysis at all potentials (Figure 2b). The high current densities translate into
228 remarkably high turnover numbers (TONs) and high Turnover Frequencies (TOFs). For
229 example, at -2.05 V, 2026 TONs were reached after 15 min, which corresponds to a TOF value
230 of 2.3 s-1. As shown below, under these conditions, formate was the major product.
231 The only detected products in all experiments were H2 and CO in the gaseous phase and formic
232 acid in the liquid phase. Surprisingly, the selectivity of the reaction was found to be greatly
233 dependent on the applied potential, with CO being the major product at the most anodic
234 potentials and formic acid becoming the major product at potentials more cathodic than -1.9 V,
235 while H2 was a minor product at all potentials. This is clearly shown in Figure 2c which displays
236 the Faradaic yields (FY) of the various products at different potentials. At -1.75 V vs Fc/Fc+,
237 FY(CO) and FY(H2) were 79% (TONCO 260 and TOFCO 0.3 s-1) and 9% respectively, and no
238 formate could be found in the liquid phase. Screening the reaction at more negative potentials,
239 CO formation decreased dramatically while FY(HCOOH) and FY(H2), the latter to a smaller
240 extent, increased. At -2.05 V, FY(CO), FY(H2) and FY(HCOOH) were 3%, 20% and 76%
241 (TONHCOOH 1950 and TOFHCOOH 2.2 s-1) respectively (Figure 2c). Thus, the selectivity between
242 CO vs HCOOH formation during electroreduction of CO2 catalyzed by Re@HPC can be finely
243 tuned by varying the applied potential.
244 As expected, the selectivity could also be tuned with the water content. When the 15 min CPE
245 was carried out at -1.75 V vs Fc/Fc+ in 10% v/v H2O/EMIM, the current was slightly larger but
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
246 the system was less selective for CO2 reduction with a FY(H2) of 33 % (Figure S9). When
247 instead the water content was decreased (1% v/v H2O/EMIM), lower current densities were
248 obtained, with comparable selectivity. The 5% v/v H2O/EMIM electrolyte thus proved a good
249 compromise for achieving high current densities and high selectivity for CO2 reduction (Figure
250 S9).
251
252
253 Figure 2: Controlled potential electrolysis of CO2 using the Re@HPC/GDL electrode: a) LSV
254 in 5% v/v H2O/EMIM saturated with CO2 (black) or with Ar (red), scan rate 20 mV s-1; b) total
255 current density at various applied potentials as a function of time; c) Faradaic Yields for CO,
256 H2 and formate after 15 min electrolysis at different potentials in 5% v/v H2O/EMIM saturated
257 with CO2.
258
259 Production of formate is unexpected since the Re complex is known to be a very selective
260 catalyst for CO2 electroreduction to CO in organic solvents. Interestingly, when, as a control
261 experiment, CPE was carried out using the soluble Re complex in the same CO2-saturated 5%
262 v/v H2O/EMIM electrolyte, the reaction was shown to yield mainly H2 with minor amounts of
263 CO and almost no HCOOH, at all visited potentials (Figure S10). Furthermore, the current
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
264 densities were much lower since at -1.85 V the current density reached a value of 0.3 mA cm-2
265 (Figure S10) for 12 mol of Re catalyst (1 mM in 12 ml) as compared to 3 mA cm-2 for 0.6
266 mol Re catalyst present in Re@HPC/GDL (Figure 2). This thus demonstrates the great impact
267 of the solid support on the reaction outcome, both in terms of the activity and the selectivity.
268 The effect on the activity is likely to be partly due to the hydrophobic nature of the environment
269 provided by the pores in which the catalyst is bound, favouring CO2 uptake and increasing CO2
270 local concentration and to the absence of mass transport limitations which challenge the
271 homogeneous catalyst.
272 In contrast, the effect of the solid support on the selectivity is difficult to explain. Formation of
273 formate cannot be explained by the catalytic properties of the support itself since, as a control
274 experiment, CPE of CO2 using an HPC/GDL electrode, not loaded with the Re complex, under
275 identical conditions, gave very different results (Figure S11). For example, at -1.85 V, with a
276 current density of 1 mA cm-2, no formate could be detected, as compared to 3 mA cm-2 and a
277 FY(HCOOH) of 20% for the Re@HPC/GDL electrode, and, at -1.95 V, with a current density
278 of 2 mA cm-2, a FY(HCOOH) of 15% was obtained, as compared to 6 mA cm-2 and a
279 FY(HCOOH) of 50% for the Re@HPC/GDL electrode. These results confirm that the specific
280 catalytic properties of the Re@HPC/GDL electrode are essentially due to the immobilized Re
281 complex, with limited or no contribution from the residual Zn. Formation of formate also
282 excludes Zn as an active species since this metal is well-established for selective formation of
283 CO. [24]
284 There is, to our knowledge, one study reporting formate production during electroreduction of
285 CO2 catalyzed by the Re complex, with a HCOOH/CO selectivity depending on the applied
286 potential as well.[11] Interestingly, in that case, the catalyst was incorporated into a Nafion
287 membrane and electrolysis was carried out in a phosphate aqueous electrolyte, explaining why
288 the major product at almost all potentials applied was H2. The activity, in terms of current
289 density, of that system was furthermore much lower than that reported here. While formate
290 production using the Re@HPC catalyst is likely to proceed via a Re-H hydride species reacting
291 with CO2,[25,26] in contrast to CO production which derives from a Re-CO2 intermediate, it is so
292 far unclear why the HPC support favours such a mechanism at very cathodic potentials and how
293 the support tunes the HCOOH/CO selectivity. These mechanistic aspects deserve further
294 investigation.
295
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
296 Finally, CPE was carried out for 2 h at -1.75 V (FY(CO)=80% and FY(H2)=10%) and -1.95 V
297 (FY(CO)=20%, FY(H2)=20% and FY(HCOOH)=55%) vs Fc/Fc+ (Figure 3). The system
298 proved very stable during electrolysis in terms of both current density and selectivity. After 2 h
299 at -1.75 V, 3155 TONCO were achieved corresponding to a TOFCO of 0.4 s-1. Furthermore, post-
300 electrolysis characterization of the electrode by SEM showed that the porosity of the system
301 was retained even if slight agglomeration of the particles was observed (Figure S12). After 2 h
302 at -1.95 V, 7835 TONs (TONCO + TONHCOOH) were achieved corresponding to a TOF (TOFCO
303 + TOFHCOOH) of 1.1 s-1. A longer experiment, during 5 hours, confirmed the stability of the
304 catalyst (Figure S13).
305
306
307 Figure 3: Two-hour electrolysis: current density and FY as a function of time during CPE at -
308 1.75 V vs Fc/Fc+ (a) and at -1.95 V vs Fc/Fc+ (b) under CO2 in 5% v/v H2O/EMIM using the
309 Re@HPC/GDL electrode. Faradaic yields for CO (red dots), H2 (black squares) and formate
310 (blue diamonds).
311
312 Conclusion
313
314 We have developed a novel hybrid solid catalyst, Re@HPC, for CO2 electroreduction to CO
315 and HCOOH. Thanks to the hierarchical porosity of the porous and conductive carbon material
316 HPC, large amounts of the molecular catalyst [Re(bpy)(CO)3Cl] can be easily immobilized
317 without any functionalization. We anticipate that HPC can be further used for heterogeneization
318 of other complexes. In order to optimize catalysis, CO2 electroreduction has been studied in an
319 ionic liquid electrolyte in the presence of water, known to favour CO2 solubility and activation.
320 To the best of our knowledge, this is the first example of a heterogeneized molecular complex
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
321 characterized for CO2 reduction catalysis in an ionic liquid. Re@HPC/GDL electrodes thus
322 provide high current densities (up to 11 mA cm-2), high TOFCO+HCOOH (2.3 s-1) and display good
323 stability. When compared to other hybrid solid electrodes in which [Re(bpy)(CO)3Cl] has been
324 immobilized,[6–12,14–16] the Re@HPC/GDL electrode clearly shows significant improvements
325 in terms of catalyst loading, TOF values and, most remarkably, in terms of stability: with the
326 exception of the MWCNT/ [Re(tBu-bpy)(CO)3Cl] material developed by Kubiak and
327 collaborators[16], all previously Re-based hybrid catalysts greatly suffer from low stability
328 (Table S3). It is also interesting to note that the Re@HPC/GDL electrode compares well, in
329 terms of current density at comparable potentials, with a standard CO-selective Ag electrode in
330 a CH3CN/EMIM electrolyte.[27] In addition, this novel electrode is unique in providing not only
331 CO but also HCOOH as CO2 reduction products, with FY(H2) not exceeding 20%, in marked
332 contrast with the other comparable systems which are selective for CO production and in
333 contrast with the homogeneous catalyst which favours proton reduction over CO2 reduction and
334 is quite inefficient in the same water/ionic liquid system. The CO/HCOOH ratio can be tuned
335 with the applied potential since it decreased as the potential is shifted to more cathodic values.
336 The combination of a water/ionic liquid electrolyte system, the hierarchical porosity of the
337 support facilitating mass exchange and transfer,[28] and the hydrophobic environment provided
338 by the pores in which the molecular complex is fixed is likely to explain the remarkable and
339 unique performances of Re@HPC as a CO2 reduction catalyst. However, further studies are
340 required to understand how the support drives and controls the reactivity of the immobilized
341 complex.
342
343
344
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
345 Experimental section
346 General methods
347 All chemicals were received from commercial sources and used as received. Zinc nitrate
348 hexahydrate (Zn(NO3)·6H2O, 99%), 2-methyilimidazole (98%), tetrabutylammonium
349 hexafluorophosphate (TBAPF6, 98%), Nafion© perfluorinated resin (10 µL of a 5 wt% solution
350 in mixture of lower aliphatic alcohols containing 5% water), absolute ethanol (CH3CH2OH),
351 methanol (CH3OH, 99%), dichloromethane (CH2Cl2, 99 %), acetonitrile (CH3CN, 99.8%),
352 chloroform-d (CDCl3, 99.8%) and ammonia solution (NH3·H2O, 32% ) were purchased from
353 Sigma-Aldrich. 1-Ethyl-3-methylimidazolium tetrafluoroborate (>98%) was purchased from
354 IOLITEC Ionic liquid technologies GmbH. [Re(bpy)(CO)3Cl] was synthesized as previously
355 reported by Kubiak et al.[29]
1
356 UV−vis spectra were recorded using a Cary 100 UV−vis spectrophotometer (Agilent). H
357 spectra were recorded on a Bruker Avance-III 300 NMR spectrometer (300 MHz) at room
358 temperature. The Rhenium concentrations in the electrolyte were assessed using an Agilent
359 7900 quadrupole ICP-MS. Liquid samples were sprayed through a micro-nebulizer in a Scott
360 spray chamber prior to ionization. An indium internal standard was injected after inline mixing
361 with the samples to correct for signal drift. Calibration solutions with Re concentrations
362 encompassing the full range of sample concentrations were used to convert measured counts to
363 concentrations. Reported uncertainties were calculated using algebraic propagation of blank
364 substraction and sample count standard deviations (n=3).
365 Material characterization
366 Scanning electron microscope (SEM) images were performed by using a JEOL 7500F
367 microscope operating at 15kV with EDX detector incorporated and a Hitachi S-4800 operating
368 at 5kV. X-ray diffraction (XRD) characterization was carried out by Panalytical X’Pert PRO
369 diffractometer (Cu Kα radiation, Bragg-Brentano geometry, sealed tube operated at 45 mA 30
370 kV X’Celerator linear detector). Nitrogen physisorption analyses were performed with ASAP
371 2420 using a platinum resistance device and liquid nitrogen as adsorbed molecule.
372 Synthesis of HPZIF8
373 Polystyrene spheres (PS) (the size of 400 nm) were synthesized as previously reported[17] and
374 used as a template for the preparation of a hierarchical porous ZIF8 (HPZIF8). ZIF8 with
375 polystyrene (ZIF8@PS) was also prepared using a protocol from Chen et al.[17], via infiltration
376 of the ZIF8 precursor into the PS spheres template. The following steps were modified from
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
377 the previous protocol to obtain the HPZIF8. This, was prepared from ZIF8@PS treating it at
378 450 °C during 4 h under Argon atmosphere for removing the PS beads.
379 Synthesis of HPC
380 The hierarchical porous carbon (HPC) derived from ZIF8 was obtained by heating the HPZIF8
381 at 800 °C during 3 h under Argon atmosphere. The HPC was used without further treatment.
382 Preparation of Re@HPC
383 The preparation of Re complex loading on HPC material was carried out as follows: the solution
384 of [Re(bpy)(CO)3Cl] (6 mg) in CH2Cl2 (10 mL) was added dropwise to a sonicated dispersion
385 of HPC (50 mg) in CH2Cl2 (20 mL). The suspension was then centrifuged for removing the
386 supernatant. The obtained material was then dried overnight under vacuum condition at room
387 temperature and used without further treatments.
388 These materials were characterized by: N2 physisorption analyses, SEM images, EDX mapping
389 and XRD (Supporting Information).
390 The amount of Re complex loaded was quantified by UV-Visible spectroscopy (Figure S6):
391 after extraction of 30 mg of Re@HPC with 20 mL of CH2Cl2 the solution was analysed for its
392 absorption at 387 nm, characteristic of [Re(bpy)(CO)3Cl] and the latter was compared to
393 calibration curve prepared with pure [Re(bpy)(CO)3Cl]. Re@HPC was also treated with CDCl3
394 for extraction of the complex and characterization by 1H NMR (Figure S5), in order to confirm
395 the integrity of the complex.
396 Electrode preparation
397 Re@HPC (5 mg) was sonicated 1 h in absolute ethanol (200 µL) and a solution of Nafion
398 perfluorinated resin (10 µL of a 5 wt% solution in mixture of lower aliphatic alcohols containing
399 5% water). The suspension was then, carefully, deposited by drop casting on a gas-diffusion
400 Layer, GDL (AVCarb GDS 3250; 1 cm2) in order to have a uniform deposition. The electrode
401 was then dried in air overnight at room temperature. For all the experiments the working
402 electrodes were prepared in the same way.
403 Electrode characterization
404 SEM images of the Re@HPC/GDL electrode were obtained before and after 2 h electrolysis.
405 The amount of Re complex present on the GDL was quantified by UV-Visible spectroscopy as
406 described above via extraction of the complex from the Re@HPC/GDL electrode immersed in
407 20 mL of CH2Cl2 (Figure S6) and quantification by comparison with the calibration curve.
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
408 Electrochemical characterization
409 All electrochemical characterization and electrolysis experiments were carried out using a Bio-
410 logic SP300 potentiostat with two-compartment cell with an Ag/AgCl/3M KCl reference
411 electrode, placed in the same compartment as the working Re@HPC/GDL electrode. A
412 platinum counter electrode was placed in a separate compartment. The two compartments were
413 separated by a membrane (Fumasep FBM-Bipolar Membrane). The electrolyte was CO2-
414 saturated 5% v/v H2O/EMIM (12.5 mL). The electrochemical cell was first purged with CO2 at
415 a flow rate of 20 mL min-1 for 1 h prior to catalytic tests using a mass flow controller (Bronkhorst
416 EL-FLOW model F-201CV). All potential values are given versus the potential of the Fc/Fc+
417 couple added as an internal standard to the solution after measurement. In 5% v/v H2O/EMIM:
418 E1/2(Fc/Fc+) = 0.35V vs Ag/AgCl (Figure S14).
419 Electrochemically active surface area of the electrode was estimated by probing the redox
420 reaction of the ferricyanide/ferrocyanide couple using cyclic voltammetry (CV). 0.1 M KCl
421 solution containing 25 mM ferrocyanide was initially degassed with Ar. Then the potential of
422 the working electrode was swept between 700 mV and –200 mV vs. Ag/AgCl (1 M KCl) at
423 different scan rates (mV s-1). Between each CV at different rates, the solution was bubbled with
424 Ar and shaken to quickly reach back to the initial conditions. Electrochemically active surface
425 areas (ECSA) were estimated from the Randles-Sevcik equation, as follows:
5
426
Ip = 2.69 × 10 n3/2 A D1/2 υ1/2
C
( )
427 with Ip: peak current, n: number of moles of electrons per mole of electroactive species, A: area
428 of electrode (cm2), D: diffusion coefficient (cm2 s-1), υ: scan rate (V s-1), C: concentration (mol
429 cm-3). The diffusion coefficient of ferricyanide is 6.7 × 10–6 cm2 s-1 and its concentration 25 ×
430 10–6 mol cm-3. The ECSA (A) is estimated from the slope of the plot of Ip versus υ 1/2.
431 CVs for the determination of surface density of electrochemically active sites were recorded in
432 CH3CN, 0.1 M TBAPF6 with a scan rate of 10 mV s-1.
433 The surface loading (Г[Re] as mol cm−2) of the catalyst was calculated through the integration
434 of the reoxidation wave in the CV scan (Figure S7) using the equation:
푞
435
훤[푅푒] =
푛퐹퐴
436 where q is the charge (C) obtained from integration of the oxidation wave, n the number of
437 electrons in the redox process per Re center (n = 1), F is the Faraday constant (96485 C mol-1),
438 and A is the geometrical electrode area (1 cm2).[30]
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
439 The homogeneous electrochemical experiments were carried out under the same conditions
440 described above using the same electrochemical cell. A glassy carbon electrode (1 cm2) was
441 used as working electrode in a solution 1 mM of [Re(bpy)(CO)3Cl] in 5% v/v H2O/EMIM.
442 H2 and CO were identified and quantified using a gas chromatograph (SRI 8610C) equipped
443 with a packed Molecular Sieve 5 Å column for permanent gases separation and a packed
444 Haysep-D column for light hydrocarbons separation. Argon (Linde 5.0) was used as carrier gas.
445 A flame ionization detector (FID) coupled to a methanizer was used to quantify CO while a
446 thermal conductivity detector (TCD) was used to quantify H2. The liquid-phase products were
447 quantified using an ionic exchange chromatography system (883 Basic IC plus; Metrohm).
448
449
450 Acknowledgements
451 D.G and S.P acknowledge financial support from the European School on Artificial Leaf:
452 Electrodes & Devices (eSCALED). This work is part of the eSCALED project which has
453 received funding from the European’s Union’s Horizon 2020 research and innovation
454 programme under the Marie Sklodowska-Curie grant agreement No 765376.
455 This research used resources of the Electron Microscopy Service located at the University of
456 Namur. This Service is member of the “Plateforme Technologique Morphologie –Imagerie”.
457
SEM images were also collected by F. Pillier at the Laboratoire Interfaces et Systèmes
458 Electrochimiques, Paris, France.
459
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
460 References
461 [1] R. Francke, B. Schille, M. Roemelt, Chem. Rev. 2018, 118, 4631–4701.
462 [2] H. Takeda, C. Cometto, O. Ishitani, M. Robert, ACS Catal. 2017, 7, 70–88.
463 [3] N. Elgrishi, M. B. Chambers, X. Wang, M. Fontecave, Chem. Soc. Rev. 2017, 46, 761–
464
796.
465 [4] C. Sun, R. Gobetto, C. Nervi, New J. Chem. 2016, 40, 5656–5661.
466 [5] L. Sun, V. Reddu, A. C. Fisher, X. Wang, Energy Environ. Sci. 2020, 13, 374–403.
467 [6] T. R. O’Toole, L. D. Margerum, T. D. Westmoreland, W. J. Vining, R. W. Murray, T. J.
468
469 [7] C. R. Cabrera, H. D. Abruña, J. Electroanal. Chem. Interfacial Electrochem. 1986, 209,
470 101–107.
Meyer, J. Chem. Soc.{,} Chem. Commun. 1985, 1416–1417.
471 [8] J. Hawecker, J.-M. Lehn, R. Ziessel, J. Chem. Soc.{,} Chem. Commun. 1983, 536–538.
472 [9] A. Zhanaidarova, A. L. Ostericher, C. J. Miller, S. C. Jones, C. P. Kubiak,
473
474 [10] S. Cosnier, A. Deronzier, J.-C. Moutet, J. Electroanal. Chem. Interfacial Electrochem.
475 1986, 207, 315–321.
476 [11] T. Yoshida, K. Tsutsumida, S. Teratani, K. Yasufuku, M. Kaneko, J. Chem. Soc.{,}
477 Chem. Commun. 1993, 631–633.
478 [12] S. Oh, J. R. Gallagher, J. T. Miller, Y. Surendranath, J. Am. Chem. Soc. 2016, 138, 1820–
479 1823.
480 [13] N. M. Orchanian, L. E. Hong, J. A. Skrainka, J. A. Esterhuizen, D. A. Popov, S. C.
481 Marinescu, ACS Appl. Energy Mater. 2019, 2, 110–123.
482 [14] J. Willkomm, E. Bertin, M. Atwa, J.-B. Lin, V. Birss, W. E. Piers, ACS Appl. Energy
483 Mater. 2019, 2, 2414–2418.
484 [15] J. D. Blakemore, A. Gupta, J. J. Warren, B. S. Brunschwig, H. B. Gray, J. Am. Chem.
485 Soc. 2013, 135, 18288–18291.
486 [16] A. Zhanaidarova, S. C. Jones, E. Despagnet-Ayoub, B. R. Pimentel, C. P. Kubiak, J. Am.
487 Chem. Soc. 2019, 141, 17270–17277.
488 [17] K. Shen, L. Zhang, X. Chen, L. Liu, D. Zhang, Y. Han, J. Chen, J. Long, R. Luque, Y.
489 Li, B. Chen, Science 2018, 359, 206–210.
Organometallics 2019, 38, 1204–1207.
490 [18] M. Wu, Y. Li, Z. Deng, B. L. Su, ChemSusChem 2011, 4, 1481–1488.
491 [19] S. Suter, S. Haussener, Energy Environ. Sci. 2019, 12, 1668–1678.
492 [20] J.-P. Song, L. Wu, W.-D. Dong, C.-F. Li, L.-H. Chen, X. Dai, C. Li, H. Chen, W. Zou,
493
494
W.-B. Yu, Z.-Y. Hu, J. Liu, H.-E. Wang, Y. Li, B.-L. Su, Nanoscale 2019, 11, 6970–
6981.
495 [21] M. L. Clark, P. L. Cheung, M. Lessio, E. A. Carter, C. P. Kubiak, ACS Catal. 2018, 8,
496 2021–2029.
497 [22] B. A. Rosen, A. Salehi-Khojin, M. R. Thorson, W. Zhu, D. T. Whipple, P. J. A. Kenis,
498 R. I. Masel, Science 2011, 334, 643–644.
499 [23] D. C. Grills, Y. Matsubara, Y. Kuwahara, S. R. Golisz, D. A. Kurtz, B. A. Mello, J. Phys.
500 Chem. Lett. 2014, 5, 2033–2038.
501 [24] B. Qin, Y. Li, H. Fu, H. Wang, S. Chen, Z. Liu, F. Peng, ACS Appl. Mater. Interfaces
502 2018, 10, 20530–20539.
503 [25] J. Agarwal, R. P. Johnson, G. Li, J. Phys. Chem. A 2011, 115, 2877–2881.
504 [26] A. J. Morris, G. J. Meyer, E. Fujita, Acc. Chem. Res. 2009, 42, 1983–1994.
505 [27] G. P. S. Lau, M. Schreier, D. Vasilyev, R. Scopelliti, M. Grätzel, P. J. Dyson, J. Am.
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002014
ChemSusChem
506
507 [28] X. Zheng, G. Shen, C. Wang, Y. Li, D. Dunphy, T. Hasan, C. J. Brinker, B. L. Su, Nat.
508 Commun. 2017, 8, 1–9.
Chem. Soc. 2016, 138, 7820–7823.
509 [29] J. M. Smieja, C. P. Kubiak, Inorg. Chem. 2010, 49, 9283–9289.
510 [30] B. Reuillard, K. H. Ly, T. E. Rosser, M. F. Kuehnel, I. Zebger, E. Reisner, J. Am. Chem.
511
512
513
Soc. 2017, 139, 14425–14435.
This article is protected by copyright. All rights reserved.