EP4405519B1

EP4405519B1 - Electrochemical carbon dioxide reduction catalyst for formate production

Info

Publication number: EP4405519B1
Application number: EP22793137.5A
Authority: EP
Inventors: Charles CREISSEN; Marc Fontecave; Moritz Wilhelm SCHREIBER
Original assignee: College de France; TotalEnergies Onetech SAS
Current assignee: College de France; TotalEnergies Onetech SAS
Priority date: 2021-09-24
Filing date: 2022-09-21
Publication date: 2025-07-09
Anticipated expiration: 2042-09-21
Also published as: US20240328007A1; ES3039636T3; EP4405519A1; WO2023046714A1; PT4405519T; FI4405519T3; DK4405519T3; CA3231505A1

Description

Technological field

The present disclosure relates to uses of catalysts and catalytic methods for electrochemical carbon dioxide reduction, for example to upgrade greenhouse gases such as carbon dioxide to valuable fuels and feedstocks. In particular, the present disclosure is about a gas diffusion electrode (GDE) suitable for carbon dioxide electrolysis, a gas-fed flow cell comprising such GDE as well as a method for producing such GDE and to a process using such GDE.

Technological background

Electrochemical carbon dioxide reduction (CO₂R) offers an attractive route to upgrade greenhouse gases such as CO₂ to valuable fuels and feedstocks. However, today it is curtailed at least in part by the limits of having a high selectivity into specific products. Achieving a narrow product distribution with cheap CO₂R catalysts is challenging and conventional material modifications offer limited control. In particular, in the case of formate (HCOO^-) while a high selectivity has been reported by using a gold-based CO₂R catalyst, it could be interesting to achieve such high selectivity with cheaper CO₂R catalysts. Indeed, formate is a high energy density and cost-beneficial product and can be used as a hydrogen storage material and as an energy source in direct formate fuel cells. Formate also has wide applications in the textile, leather and pharmaceutical industries.
The study of Fang Y. et al., entitled "Electrochemical Reduction of CO2 at Functionalized Au Electrodes" (J. Am. Chem. Soc., 2017, 139, 3399-3405) has evaluated Au electrodes functionalized with thiol-tethered ligands for their ability to alter the selectivity of CO₂R reaction. It was found that upon the use of 4-pyridylethylmercaptan-modified Au CO₂R catalyst, the Faradaic efficiency (FE) amounts to 21% at -1.00 V vs RHE, while on Au surfaces, the formate FE was only 11% at -1.01 V vs RHE. The partial current density for formate, which is an indication of the formate production rate, has been determined to be as high as -4.1 mA/cm² when the pyridylethylmercaptan-modified Au CO₂R catalyst has been used, while on the untreated gold surface, the partial current density was only -1.37 mA/cm². The good selectivity for formate is explained by the protonation of the pyridine moiety of the ligand which is anchored on the gold surface to effectively promote the activation and conversion of CO₂ to HCOOH - see for example the study of Li. F et al., entitled "Understanding the Role of Functional Groups of Thiolate Ligand on Electrochemical CO2 Reduction over Au(111) from First-Principles" (J. Mater. Chem. A, 2019, 7, 19872-19880) and to the adsorption of a hydrogen atom on the gold surface which favour the electrophilic attack of the CO₂ to yield activated HCO₂*.
The study of Tao Z. et al., entitled "Copper-Gold Interactions Enhancing Formate Production from Electrochemical CO2 Reduction" (ACS Catal., 2019, 9, 10894-10898) describes the electrochemical CO₂R to formate on a Cu/Au bimetallic system. The authors have shown that interactions with gold can turn copper, which by itself is neither selective nor active for the electrochemical CO₂ reduction to formate, into an improved catalyst for the same reaction. The Cu/Au bimetallic catalyst described allows for the reduction reaction to achieve a partial current density of 10.4 and a Faradaic efficiency of 81% at -0.6 V vs RHE.
The study of Li J. et al., entitled "Electroreduction of CO2 to Formate on a Copper-Based Electrocatalyst at High Pressures with High Energy Conversion Efficiency" (J. Am. Chem. Soc., 2020, 142, 7276-7282) describes a square-wave electrochemical redox cycling treatment of ultrapure copper foil to produce submicron-thick films rich in (111)-oriented Cu₂O nanoparticles anchored on Cu for the CO₂R reaction in a high-pressure electrolyser under CO₂-partial pressure of 0.1 to 6 MPa. It was observed that at 0.1 MPa, maximum Faradaic efficiency of 27.1% at -0.64 V vs RHE in formate was obtained while at high pressure of about 4.5 MPa, the Faradaic efficiency in formate increase up to 97.7% at -0.64 V vs RHE. However, at this potential, the partial current density attains only -5.2 and for attaining a higher partial current density (of about -33 mA/cm²), it is required to work at -1.24 V vs RHE which as for consequence to decrease the Faradaic efficiency in formate to about 53%. Such designs, although offering a good selectivity, are hampered due to the requirement of working under high pressure.
The design described in the study of Puring K. J. et al., entitled "Assessing the influence of supercritical carbon dioxide on the electrochemical reduction to formic acid using carbonsupported copper catalysts" (ACS Catal., 2020, 10, 12783-12789) further requires work with supercritical carbon dioxide to control the selectivity into formate.
In the study of Fang Y. et al., entitled "Carbon dioxide electrochemical reduction at thiolatemodified bulk Au electrodes" (Catal. Sci. Technol., 2019, 9, 2689), it is shown that thiol ligands containing no nitrogen-based heterocycle and introduced by treating polycrystalline Au with an ethanolic solution of thiols, can alter the activity and selectivity of Au for the electrochemical carbon dioxide reduction, suppressing the formation of hydrogen or in other cases the formation of carbon monoxide.
In the study of Garcia-Muelas R. et al, entitled "Origin of the selective electroreduction of carbon dioxide to formate by chalcogen modified copper" (J. Phys. Chem. Lett., 2018, 9, 24, 7153-7153), it is revealed that on sulfur-, selenium- or tellurium-modified copper, the chalcogen adatoms are present on the surface and actively participate in the reaction, either by transferring a hydride or by tethering carbon dioxide thus suppressing the formation of carbon monoxide. Thus, a Faradaic efficiency for formate amounting to about 75% at -0.6 V vs RHE has been obtained. At this potential, the partial current density was about -10 mA/cm². Similar results have been described at -0.9 V vs RHE in the study of Deng Y. et al, entitled "On the role of sulfur for the selective electrochemical reduction of carbon dioxide to formate on CuSx catalysts" (ACS Appl. Mater. Interfaces, 2018, 10, 28572-28581) or in the study of Huang Y. et al, entitled "Rational design of sulfur-doped copper catalyst for the selective electroreduction of carbon dioxide to formate" (Chem. Sus. Chem., 2018, 11, 320-326).
KR 10 2018 0129307 discloses copper electrocatalysts for carbon dioxide reduction. Pandey et al., Journal of Materials Chemistry B, 2019, 7, 6630 discloses copper-cysteamine nanoparticles. Guo et al., Journal of Molecular Structure, 2011, 911, 103 discloses 4-mercaptopyridine adsorbed to gold, silver and copper oxide films. Weng et al., Phys. Chem. Chem. Phys., 2018, 20, 16973; Kopfjar et al., Chemical Engineering Technology, 2016, 39, 2042; Kopfjar et al., J. Appl. Electrochem., 2014, 44, 1107; and Higgins et al., ACS Energy Letters, 2019, 4, 317 disclose electrochemical CO₂ reduction using gas diffusion electrodes.
Although the selectivity and production rate of formate could be considered as being important, there is still room for improvement.
There is a still need for an improvement of catalyst materials for efficient electrochemical carbon dioxide reduction (CO₂R) reactions; in particular, there is still a need for a catalyst that shows improved selectivity for formate; for a system comprising such a catalyst and for process using such catalysts.

Summary of the disclosure

One or more of the above needs can be fulfilled by the gas diffusion electrode according to the present disclosure comprising copper nanoparticles functionalized with one or more pyridine-containing ligands being pyridine-containing ligands.
According to a first aspect, the disclosure provides a gas diffusion electrode suitable for carbon dioxide electrolysis, said gas diffusion electrode having a gas diffusion membrane, the gas diffusion electrode further comprising an ink deposited on the gas diffusion membrane; wherein the ink comprises an ion-conducting polymer, said gas diffusion electrode is remarkable in that the ink further comprises a catalyst comprising copper nanoparticles functionalized with one or more pyridine-containing ligands, wherein the one or more pyridine-containing ligands have an anchoring group comprising a sulphur atom tethered to the copper nanoparticles.
It has been found that copper nanoparticles functionalized with one or more pyridine-containing ligands show good results in the electrolysis of carbon dioxide when implemented into a gas-diffusion electrode of a gas-fed flow cell, notably in the selectivity and/or production rate in formate. Indeed, surprisingly, a selectivity into formate equivalent to a faradaic efficiency of at least 75% at -100 mA/cm^-2 can be reached.
With preference, the pyridine-containing ligands are selected from 4-pyridylethylmercaptan, 4-mercaptopyridine, 2,6-dimethyl-4-mercaptopyridine, 2-mercaptopyridine and any mixture thereof; more preferably, the one or more pyridine-containing ligands are or comprise 4-mercaptopyridine.
For example, the one or more pyridine-containing ligands have a thiol group.
For example, the one or more pyridine-containing ligands are grafted onto the copper nanoparticles. For example, the one or more pyridine-containing ligands are present in a surface concentration ranging from 5 nmol cm^-2 to 40 nmol cm^-2 as determined by reductive desorption and UV-visible spectroscopy as set out in the description; preferably, ranging from 8 nmol cm^-2 to 20 nmol cm^-2.
Advantageously, the one or more pyridine-containing ligands are grafted onto the copper nanoparticles and form a monolayer.
In an embodiment, the copper nanoparticles have an average diameter ranging from 5 nm to 200 nm as measured by transmission-electron microscopy, preferably from 10 nm to 150 nm, more preferably from 20 nm to 100 nm.
In an embodiment, the copper nanoparticles comprise facets selected from Cu(200) facets, Cu(111) facets and any mixture thereof.
Advantageously, the catalyst is suitable for electrochemical carbon dioxide reduction (CO₂R) reactions to produce formate.
One or more of the following features advantageously define the gas diffusion membrane of the gas diffusion electrode of the disclosure:

The gas diffusion membrane is a hydrophobic porous support. With preference, said hydrophobic porous support shows a pore size ranging from 400 nm to 500 nm as determined by scanning electron microscopy preferably, from 420 nm to 580 nm or from 440 nm to 560 nm.
The gas diffusion membrane is a hydrophobic, porous and chemically inert support; with preference, the gas diffusion membrane is not soluble in KOH.
The gas diffusion membrane is or comprises polytetrafluoroethylene.
The gas diffusion membrane has a thickness ranging from 50 µm to 150 µm as measured by scanning electron microscopy, preferably from 60 µm to 120 µm more preferably from 70 µm to 100 µm.

One or more of the following features advantageously define the ion-conducting polymer of the gas diffusion electrode of the disclosure:

The ion-conducting polymer comprises an ionomer.
The ion-conducting polymer is or comprises an ionomer with a tetrafluoroethylene backbone group (-CF₂-CF₂-).
The ion-conducting polymer is or comprises a perfluorinated sulfonic acid, such as Nafion^®.
The ion-conducting polymer is or comprises tetrafluoroethylene-perfluoro(3-hydrophobioxa-4-pentenesulfonic acid) copolymer, such as Aquivion^®.

In an embodiment, the ink layer has a thickness ranging from 2 µm to 20 µm as measured by scanning electron microscopy, preferably from 5 µm to 15 µm, more preferably from 8 µm to 12 µm.
For example, the ink has a weight ratio of the copper nanoparticles functionalized with one or more pyridine-containing ligands over the ion-conducting polymer ranging from 10 to 40, preferably from 15 to 35, or from 20 to 30. With preference, the ink has a weight ratio of the copper nanoparticles functionalized with thiol-tethered ligands over the ion-conducting polymer ranging from 10 to 35, preferably from 15 to 30, or from 25 to 35.
For example, the gas diffusion electrode has a mass loading of the ink onto said gas diffusion membrane ranging from 1.50 mg/cm² to 2.00 mg/cm², preferably from 1.60 mg/cm² to 1.90 mg/cm².
For example, the copper nanoparticles have an average diameter ranging from 5 nm to 200 nm as measured by transmission electron microscopy, preferably from 10 nm to 150 nm, more preferably from 20 nm to 100 nm.
According to a second aspect, the disclosure provides a method for producing the gas diffusion electrode suitable for carbon dioxide electrolysis as defined according to the first aspect, said method is remarkable in that it comprises the following steps:

a) providing copper nanoparticles;
b) obtaining a first dispersion by dispersing the copper nanoparticles into a solvent selected from water and/or a first organic solvent;
c) adding one or more pyridine-containing ligands to the first dispersion to obtain a suspension with copper nanoparticles functionalized with one or more pyridine-containing ligands, wherein the one or more pyridine-containing ligands have an anchoring group comprising at least one sulphur atom;
d) obtaining a second dispersion by dispersing the copper nanoparticles functionalized with one or more pyridine-containing ligands into a second organic solvent;
e) adding an ion-conducting polymer to the second dispersion to obtain an ink;
f) providing a gas-diffusion membrane and depositing said ink onto said gas-diffusion membrane to obtain the gas diffusion electrode.

In an embodiment, the solvent is water and step (b) of dispersing the copper nanoparticles to obtain a first dispersion in an aqueous solution.
With preference, the first organic solvent used in step (b) is a polar solvent selected from dichloromethane, ethyl acetate, acetone, dimethylformamide, acetonitrile, n-butanol, n-propanol, methanol, ethanol and any mixture thereof; with preference, the first organic solvent is or comprises dimethylformamide and/or methanol.
With preference, the second organic solvent used in step (d) is selected from dichloromethane, ethyl acetate, acetone, dimethylformamide, acetonitrile, n-butanol, n-propanol, methanol, ethanol and any mixture thereof; with preference, the second organic solvent is or comprises dimethylformamide and/or methanol; more preferably, the second organic solvent is or comprises methanol.
The first organic solvent can be the same or different from the second organic solvent.
For example, step (b) further comprises a sub-step of sonicating the first dispersion. With preference, the sub-step sonicating the first dispersion is performed at room temperature; for example, at a temperature ranging from 20°C to 30°C. With preference yet, the sub-step of sonicating the first dispersion is performed for at least 10 minutes, preferably for at least 15 minutes and/or for at most 60 minutes, preferably at most 45 minutes.
For example, step (b) and step (c) are performed simultaneously and step (c) of adding one or more pyridine-containing compounds comprises adding a solution of one or more pyridine-containing compounds into the organic solvent used in step (b). With preference, the concentration of the solution of one or more pyridine-containing ligands ranges from 30 nmol per mg of Cu to 200 nmol per mg of Cu; preferably from 50 to 180 nmol per mg of Cu; more preferably from 70 to 160 nmol per mg of Cu; even more preferably from 90 to 150 nmol per mg of Cu and most preferably from 110 to 140 nmol per mg of Cu.
For example, steps of washing and drying are performed after step (c) and/or before step (d). With preference, the step of washing is performed with an organic solvent and the step of drying lasts at least 12 hours, preferably at least 24 hours and/or lasts no more than 48 hours, preferably no more than 36 hours. The organic solvent used in the step of washing is selected from dichloromethane, ethyl acetate, acetone, dimethylformamide, acetonitrile, n-butanol, n-propanol, methanol, ethanol and any mixture thereof; with preference, the organic solvent used in the step of washing is or comprises dimethylformamide and/or methanol.
For example, the step of depositing the ink onto the gas-diffusion membrane is performed by spray-deposition.
For example, said method further comprises the step (g) of drying the gas diffusion electrode under reduced pressure, for example at a pressure ranging from 0.09 MPa to 10^-4 MPa. According to a third aspect, the disclosure provides a gas diffusion electrode obtained by the method according to the second aspect. The third aspect is not claimed.
According to a fourth aspect, the disclosure provides a gas-fed flow cell suitable for carbon dioxide electrolysis, said gas-fed flow cell comprising a gas chamber, a catholyte chamber and an anolyte chamber, wherein said gas chamber is separated from the catholyte chamber by a gas diffusion electrode which is attached to the catholyte chamber by an electrically conductive connection, said gas diffusion electrode having a gas diffusion membrane being comprised within said gas chamber, wherein said catholyte chamber and said anolyte chamber are separated by an anion exchange membrane, and wherein said catholyte chamber and said anolyte chamber comprise respectively a cathode and an anode, said gas-fed flow cell is remarkable in that the gas diffusion electrode is as defined according to the first aspect and/or with the third aspect.
The cathode and the anode are different from the gas diffusion electrode.
Surprisingly, it has been demonstrated that the gas-fed flow cell of the present disclosure, namely comprising the gas diffusion electrode of the first aspect and/or with the third aspect, allows for obtaining good results in the electrolysis of carbon dioxide, notably in the selectivity and/or in the production rate of formate. Thus, it is demonstrated in the present disclosure that selectivity into formate equivalent to a faradaic efficiency of at least 75% at -100 can be reached, preferably at least 80%, and of at least 65% at -300 mA/cm², preferably at least 70%.
It is also demonstrated in the present disclosure that the formate production rate is reflected by the partial current density, is of at least 200 mA/cm², preferably of at least 210 mA/cm². These results, in the selectivity and/or in the production rate of formate have never been obtained before, especially the results in terms of formate production rate.
With preference, said reference electrode is an Ag/AgCl electrode filled with 3.4 M of KCI. With preference, said anode is a Ni foam anode.
Advantageously, the electrically conductive connection from the gas diffusion electrode and the catholyte chamber is achieved by applying copper tape on said gas diffusion electrode, the copper tape being electrically connected to a metallic rod in contact with the catholyte chamber. For example, the metallic rod is a steel rod, preferably a stainless-steel rod.
According to a fifth aspect, the disclosure provides a process for electrolysing carbon dioxide, said process comprising the following steps:

i. providing a gas-fed flow cell;
ii. providing at least one electrolyte flow into said gas-fed flow cell wherein the at least one electrolyte flow is a catholyte flow and an optional anolyte flow;
iii. activating said gas-fed flow cell;
iv. providing an input flow of carbon dioxide to produce an output flow of liquid component comprising at least formate;
v. recovering said output flow comprising at least formate;

^-1

For example, the electrolyte flow provided in step (ii) has a flow rate that is comprised between 2.5 mL min^-1 to 8.5 mL min^-1, preferably from 3.0 mL min^-1 to 8.0 mL min^-1, more preferably from 3.5 mL min^-1 to 7.5 mL min^-1.
For example, the electrolyte flow provided in step (ii) is a flow of an aqueous solution of one or more inorganic bases.
With preference, the aqueous solution of one or more inorganic bases has a concentration ranging from 1 M to 10 M; preferably from 3 to 7 M or from 5 M to 10 M. For example, the aqueous solution of one or more inorganic bases has a concentration that is at least 5 M.
With preference, the aqueous solution of one or more inorganic bases has a pH ranging from 7 to 15.
With preference, the one or more inorganic bases are alkali selected from NaOH, KOH, Ca(OH)₂, LiOH, Mg(OH)₂, RbOH, CsOH and any mixture thereof. With preference, the one or more inorganic bases are or comprise KOH and/or NaOH.
Advantageously, said step (iii) is performed by injecting carbon dioxide at a potential gradient starting at -0.3 V versus a reference electrode and ending at -2.0 V versus said reference electrode at a sweep rate ranging from 20 mV s^-1 to 30 mV s^-1, the reference electrode being preferably an Ag/AgCl electrode filled with 3.4 M of KCI. With preference, the potential gradient starts at -0.5 V versus a reference electrode and ends at -1.8 V versus said reference electrode.
With preference, said step (iv) lasts at least 1 hour, more preferably at least 2 hours, even more preferably at least 3 hours, most preferably at least 4 hours, even most preferably at least 5 hours or at least 6 hours.
For example, the input flow of carbon dioxide provided in step (iv) has a flow rate that is ranging from 10 mL min^-1 to 150 mL min^-1, preferably from 15 mL min^-1 to 100 mL min^-1, more preferably from 20 mL min^-1 to 80 mL min^-1, even more preferably from 25 mL min^-1 to 50 mL min^-1.
For example, the input flow of carbon dioxide provided in step (iv) comprises at least 95 mol% of carbon dioxide based on the total molar content of the input flow; preferably at least 98 mol%.
Advantageously, said step (iv) is performed at room temperature, for example at a temperature ranging from 20°C to 30°C.
Advantageously, said step (iv) is performed at atmospheric pressure, for example at a pressure ranging from 0.09 MPa to 0.11 MPa.
With preference, the process is operated with a cathodic voltage no lower than -0.5V vs. reversible hydrogen electrode (RHE).
According to a sixth aspect, the disclosure provides for the use of a catalyst for electrochemical carbon dioxide reduction (CO₂R) reactions to produce formate, the use being remarkable in that the catalyst is according to the first aspect. Thus, the disclosure provides for the use of a catalyst for electrochemical carbon dioxide reduction (CO₂R) reactions to produce formate, the use being remarkable in that the catalyst comprises copper nanoparticles functionalized with one or more pyridine-containing ligands, wherein the one or more pyridine-containing ligands have an anchoring group comprising a sulphur atom tethered to the copper nanoparticles.
With preference, the pyridine-containing ligands are selected from 4-pyridylethylmercaptan, 4-mercaptopyridine, 2,6-dimethyl-4-mercaptopyridine, 2-mercaptopyridine and any mixture thereof; more preferably, the one or more pyridine-containing ligands are or comprise 4-mercaptopyridine.
For example, the one or more pyridine-containing ligands have a thiol group.
For example, the one or more pyridine-containing ligands are grafted onto the copper nanoparticles. For example, the one or more pyridine-containing ligands are present in a surface concentration ranging from 5 nmol cm^-2 to 40 nmol cm^-2 as determined by reductive desorption and UV-visible spectroscopy; preferably, ranging from 8 nmol cm^-2 to 20 nmol cm^-2.
Advantageously, the one or more pyridine-containing ligands are grafted onto the copper nanoparticles and form a monolayer.
In an embodiment, the copper nanoparticles have an average diameter ranging from 5 nm to 200 nm as measured by transmission-electron microscopy, preferably from 10 nm to 150 nm, more preferably from 20 nm to 100 nm.
In an embodiment, the copper nanoparticles comprise facets selected from Cu(200) facets, Cu(111) facets and any mixture thereof.
The particular features, structures, characteristics or embodiments may be combined in any suitable manner, as would be apparent to a person skilled in the art from this disclosure, in one or more embodiments.

Description of the figures

Figure 1 illustrates the gas fed flow cell of the present disclosure. An insert in figure 1 represents a zoom of the gas chamber and the catholyte chamber, highlighting the gas diffusion electrode of the present disclosure.
Figure 2 shows Fourier-Transform Infrared spectra showing the successful anchoring of thiol ligands on copper nanoparticles.
Figure 3 is a transmission electron microscopy image showing an image of initial copper nanoparticles.
Figure 4 is a transmission electron microscopy image showing an image of copper nanoparticles with SPy.
Figure 5 is a transmission electron microscopy image showing an image of copper nanoparticles with SPy after 1 hour of electrolysis at -300 mA cm^-2.
Figure 6 is a scanning electron microscopy image showing images of the gas diffusion electrode of the invention.
Figure 7 illustrates the UV Visible spectrum of the desorbed ligand from the electrode.
Figure 8 is the zoom of the peak shown on figure 7 obtained after reductive desorption of the electrode under argon.
Figure 9 is a current-voltage response (iR-corrected) of a Cu-SPy GDE obtained from chronopotentiometric steps with a 3 min hold time.
Figure 10 is an iR-corrected current-voltage response obtained from chronopotentiometric steps comparing Cu and Cu-SPy electrodes.
Figure 11 shows the FE_HCOO- values obtained from controlled current electrolysis over 1 hour with varying current densities for Cu and Cu-SPy electrodes.
Figure 12 displays the faradaic efficiency for all the products obtained following the present disclosure.
Figure 13 displays the partial current densities for all the products obtained following the present disclosure
Figure 14 is a study of the selectivity as a function of the desorption of the nitrogencontaining ligands from the copper nanoparticles.
Figure 15 is a scanning electron microscopy image of the Cu-SPy electrode before electrolysis.
Figure 16 is a scanning electron microscopy image of the Cu-SPy electrode after electrolysis carried out at 300 mA cm^-2.
Figure 17 is a scanning electron microscopy image of the Cu-SPy electrode after electrolysis carried out at 500 mA cm^-2.
Figure 18 shows the single-pass conversion efficiency in function of the carbon dioxide flow rate. Conditions: 5 M KOH, 5.5 mL min^-1 electrolyte solution flow, anolyte and catholyte volume of 20 mL each, geometric electrode area of 1 cm².
Figure 19 represents a stability study of the gas-fed flow cell of the present disclosure.
Figure 20 is the ¹H-NMR spectrum of the catholyte for a Cu-SPy electrode following electrolysis.

Detailed description

For the disclosure, the following definitions are given:
The terms "comprising", "comprises" and "comprised of" as used herein are synonymous with "including", "includes" or "containing", "contains", and are inclusive or open-ended and do not exclude additional, non-recited members, elements or method steps. The terms "comprising", "comprises" and "comprised of" also include the term "consisting of".
The recitation of numerical ranges by endpoints includes all integer numbers and, where appropriate, fractions subsumed within that range (e.g. 1 to 5 can include 1, 2, 3, 4, 5 when referring to, for example, a number of elements, and can also include 1.5, 2, 2.75 and 3.80, when referring to, for example, measurements). The recitation of endpoints also includes the recited endpoint values themselves (e.g. from 1.0 to 5.0 includes both 1.0 and 5.0). Any numerical range recited herein is intended to include all sub-ranges subsumed therein.
Figure 1 illustrates the gas-fed flow cell 1 of the present disclosure. Said gas-fed flow cell 1 comprises a gas diffusion electrode 7 suitable for carbon dioxide electrolysis. The following description first describes the gas diffusion electrode 7.

The gas diffusion electrode 7

The gas diffusion electrode 7 has a gas diffusion membrane 17, the gas diffusion electrode 7 further comprising an ink 19 deposited on the gas diffusion membrane 17; wherein the ink 19 comprises an ion-conducting polymer, said gas diffusion electrode 7 is remarkable in that the ink 19 further comprises a catalyst comprising copper nanoparticles functionalized with one or more pyridine-containing ligands, wherein the one or more pyridine-containing ligands have an anchoring group comprising one sulphur atom tethered to the copper nanoparticles. Advantageously, the one or more pyridine-containing ligands are grafted onto the copper nanoparticles and form a monolayer.
In an embodiment, the catalyst comprises copper nanoparticles wherein at least a part of the copper nanoparticles is functionalized with one or more pyridine-containing ligands, the one or more pyridine-containing ligands having an anchoring group comprising one sulphur atom tethered to the copper nanoparticles.
For example, the one or more pyridine-containing ligands have a thiol group before being grafted onto the copper nanoparticles, and thereby become thiol-tethered pyridine-containing ligands
With preference, the one or more pyridine-containing ligands are selected from 4-pyridylethylmercaptan, 4-mercaptopyridine 2,6-dimethyl-4-mercaptopyridine, 2-mercaptopyridine and any mixture thereof. For example, the one or more pyridine-containing ligands are or comprise 4-mercaptopyridine.
In an embodiment, the one or more pyridine-containing ligands are functionalized onto the copper nanoparticles and are present in a surface concentration ranging from 5 nmol cm^-2 to 40 nmol cm^-2 as determined by reductive desorption and UV-visible spectroscopy, preferably ranging from 6 nmol cm^-2 to 30 nmol cm^-2 or ranging from 8 nmol cm^-2 to 20 nmol cm^-2; or ranging from 10 nmol cm^-2 to 18 nmol cm^-2.
The gas diffusion membrane 17 allows for the diffusion of carbon dioxide as the main reactant of the electrolysis reaction into the electrochemical cell and is preferably a hydrophobic porous support. In a gas-fed flow cell 1, the gas diffusion membrane 17 is comprised within the gas chamber 3 of said gas-fed flow cell 1. With preference, said support shows a pore size ranging from 400 nm to 500 nm as determined by scanning electron microscopy, preferably from 420 nm to 580 nm or from 440 nm to 560 nm. The gas diffusion membrane is preferably selected from an ion-conducting polymer-based membrane, an ion-conducting inorganic material, a combination polymer/inorganic based membrane and the like.
It is preferred that the gas diffusion membrane is a hydrophobic, porous and chemically inert support; with preference, the gas diffusion membrane is not soluble in KOH. For example, the gas diffusion membrane 17 is or comprises polytetrafluoroethylene (PTFE). Examples of suitable membranes are commercially available from Fisher Scientific SAS under the commercial denomination Sartorius.
With preference, the gas diffusion membrane 17 has a circular shape and/or has a surface area of at least 1 cm², or of at least 2 cm². For example, the gas diffusion membrane 17 has a thickness ranging from 2 µm to 50 µm measured by scanning electron microscopy, preferably from 5 µm to 40 µm, more preferably from 8 µm to 30 µm.
An ink 19 is deposited on the gas diffusion membrane 17 and comprises an ion-conducting polymer. With preference, the ion-conducting polymer is or comprises an ionomer. For example, the ion-conducting polymer is or comprises an ionomer with a tetrafluoroethylene backbone group (-CF₂-CF₂-). Such ionomers are capable of creating strongly hydrophobic nanoporous networks. For example, said ion-conducting polymer is or comprises a perfluorinated sulfonic acid, such as Nafion^® (tetrafluoroethylene-perfluoro-3,6-dioxa-4-methyl-7-octenesulfonic acid copolymer); and/or the ion-conducting polymer is or comprises tetrafluoroethylene-perfluoro(3-hydrophobioxa-4-pentenesulfonic acid) copolymer, such as Aquivion^®. It can form a layer on the gas diffusion membrane, said layer having a thickness ranging from 2 nm and 50 nm measured by transmission electron microscopy, preferably from 5 nm and 40 nm, more preferably from 10 nm and 30 nm. For example, ink 19 has a ratio of the copper nanoparticles functionalized with thiol-tethered ligands over the ion-conducting polymer.
With preference, the ink has a weight ratio of the copper nanoparticles functionalized with thiol-tethered ligands over the ion-conducting polymer ranging from 10 to 35, preferably from 15 to 30, or from 25 to 35.
For example, the gas diffusion electrode 7 has a mass loading of the ink 19 onto said gas diffusion membrane 17 ranging from 1.50 mg/cm² to 2.00 mg/cm², preferably from 1.60 mg/cm² to 1.90 mg/cm². The mass loading can be determined by weighing before and after deposition and drying.
For example, the copper nanoparticles have an average diameter ranging from 5 nm to 200 nm as measured by transmission electron microscopy, preferably from 10 nm to 150 nm, more preferably from 20 nm to 100 nm.

The gas-fed flow cell 1

The gas-fed flow cell 1 suitable for carbon dioxide electrolysis of the present disclosure will then be described.
The gas-fed flow cell 1 comprises a gas chamber 3, a catholyte chamber 9 and an anolyte chamber 13. For example, the gas chamber 3 has a gas channel 5, through which a flow of CO₂ is circulating. The gas chamber 3 is separated from the catholyte chamber 9 by a gas diffusion electrode 7 which is attached to the catholyte chamber 9 by an electrically conductive connection. The catholyte chamber 9 and the anolyte chamber 13 are separated by an anion exchange membrane (AEM) 11. The catholyte chamber 9 and the anolyte chamber 13 comprise respectively a cathode (not represented) and an anode 15, for example a Ni foam anode. However, any oxygen evolution reaction (OER) catalyst and anode compartment design can be used. The gas-fed flow cell 1 of the present disclosure is remarkable in that the gas diffusion electrode 7 is as defined above and in that the gas diffusion membrane 17 of said gas diffusion electrode 7 is comprised within the gas chamber 3. The ink 19 comprising the ion-conducting polymer and the copper nanoparticles functionalized with one or more pyridine-containing ligands as described above is comprised within the catholyte chamber 9.
With preference, the cathode is a reference electrode. It is preferred that said reference electrode is an Ag/AgCl electrode filled with KCI at a concentration ranging from 3.0 to 3.8 M; preferably from 3.2 to 3.6 M; even more preferably with 3.4 M of KCI. In other implementation of the invention, the reference electrode could also be a reversible hydrogen electrode (RHE).
Advantageously, the electrically conductive connection from the gas diffusion electrode 7 and the catholyte chamber 9 is achieved by applying copper tape on said gas diffusion electrode 7, the copper tape being electrically connected to a metallic rod in contact with the catholyte chamber. For example, the metallic rod is a steel rod, preferably a stainless-steel rod.

Method for producing the gas diffusion electrode 7

Preparation of the catalyst; i.e. of the copper nanoparticles

According to the disclosure, at least a part of the copper nanoparticles is functionalized with one or more pyridine-containing ligands. For example, at least 50 wt.% of the copper nanoparticles are functionalized, based on the total weight of the copper nanoparticles; preferably, at least 70 wt.%; more preferably at least 80 wt.%, and even more preferably at least 90 wt.%. In a preferred embodiment, 100 wt.% of the copper nanoparticles are functionalized with one or more pyridine-containing ligands.
The copper nanoparticles are firstly functionalized (i.e., grafted) with one or more pyridine-containing ligands presenting an anchoring group. To achieve that, the copper nanoparticles, provided in step (a), are in a step (b) dispersed in an organic solvent is selected from dichloromethane, ethyl acetate, acetone, dimethylformamide, acetonitrile, n-butanol, n-propanol, methanol, ethanol and any mixture thereof; with preference, the organic solvent is or comprises dimethylformamide and/or methanol.
The first dispersion that is obtained can be then sonicated in a sub-step. With preference, the sonicating sub-step is performed at room temperature, for example at a temperature ranging from 20°C to 30°C. In an embodiment, the sonicating sub-step is performed for at least 10 minutes, preferably for at least 15 minutes and/or for no more than 60 minutes, preferably no more than 45 minutes. The sub-step of sonicating the first dispersion is preferably performed before step (c).
Thereafter, in step (c), the one or more pyridine-containing ligands, preferably in a solution of the organic solvent, is added to the first dispersion to obtain a suspension. For example, the solution of the one or more pyridine-containing ligands in the organic solvent has a concentration ranging from 30 nmol per mg of Cu to 200 nmol per mg of Cu; preferably from 50 to180 nmol per mg of Cu; more preferably from 70 to 160 nmol per mg of Cu; even more preferably from 90 to 150 nmol per mg of Cu and most preferably from 110 to 140 nmol per mg of Cu.
Said suspension can be sonicated to form copper nanoparticles functionalized with one or more pyridine-containing ligands. For example, the sonication is performed for at least 30 minutes, preferably at least 45 minutes and/or for no more than 90 minutes, preferably no more than 75 minutes.
For example, the sonication is achieved at room temperature, i.e., at a temperature ranging from 20°C to 30°C.
For example, the copper nanoparticles functionalized with the one or more pyridine-containing ligands described above are washed and dried after step (c) and/or before the steps required to prepare the gas diffusion electrode 7. With preference, the step of washing is performed with an organic solvent and the step of drying lasts at least 12 hours, preferably at least 24 hours and/or last no more than 48 hours, preferably no more than 36 hours. With preference, the organic solvent is selected from dichloromethane, ethyl acetate, acetone, dimethylformamide, acetonitrile, n-butanol, n-propanol, methanol, ethanol and any mixture thereof; with preference, the organic solvent is or comprises dimethylformamide and/or methanol.

Preparation of the gas diffusion electrode 7 as such

The copper nanoparticles are, in step (d), dispersed into methanol to obtain a second dispersion. It is preferred that the second dispersion is sonicated. With preference, said sonicating is achieved at room temperature, for example at a temperature ranging from 20°C and 30°C. With preference yet, said sonicating is achieved for at least 10 minutes, preferably for at least 15 minutes and/or for no more than 60 minutes, preferably no more than 45 minutes.
Then, in step (e), an ion-conducting polymer, such as Nafion^®, is added to obtain the ink.
The ink is then deposited in step (f) on a gas-diffusion membrane, for example, a gas-diffusion membrane that is or comprises polytetrafluoroethylene. In a preferred embodiment, the ink is spray-deposited on the gas-diffusion membrane.
In a step (g), the gas diffusion electrode can be dried under reduced pressure, for example at a pressure ranging from 0.09 MPa and 10^-4 MPa.

Electrolysis of carbon dioxide and production of formate

Finally, the present disclosure is about a process for electrolysing carbon dioxide, said process comprising the following steps:

i. providing a gas-fed flow cell 1;
ii. providing at least one electrolyte flow 21 into said gas-fed flow cell 1 wherein the at least one electrolyte flow is a catholyte flow and an optional anolyte flow;
iii. activating said gas-fed flow cell 1;
iv. providing an input flow 23 of carbon dioxide to produce an output flow 25 of a liquid component comprising at least formate;
v. recovering said output flow 25 comprising at least formate;

flow cell

The process needs to have flowing catholyte but the anolyte can be flowing or static; with preference the anolyte is flowing as well.
For example, the current density is at least 100 mA, with preference, at least 200 mA.
The "electrolyte flow" hereafter refers to the catholyte flow but also apply to the anolyte flow in case of flowing anolyte.
For example, the electrolyte flow provided in step (ii) has a flow rate that is ranging from 2.5 mL min^-1 to 8.5 mL min^-1, preferably from 3.0 mL min^-1 to 8.0 mL min^-1, more preferably from 3.5 mL min^-1 to 7.5 mL min^-1.
For example, the electrolyte flow provided in step (ii) is a flow of an aqueous solution of one or more inorganic bases. With preference, the one or more inorganic bases are alkali selected from NaOH, KOH, Ca(OH)₂, LiOH, Mg(OH)₂, RbOH, CsOH and any mixture thereof. With preference, the one or more inorganic bases are or comprise KOH and/or NaOH
With preference, the aqueous solution of one or more inorganic bases has a pH ranging from 7 to 15.
With preference, the aqueous solution of one or more inorganic bases has a concentration of at least 1 M, of at least 3 M, or at least 4 M, or at least 5 M. With preference, the aqueous solution of one or more inorganic bases has a concentration of at most 10 M, of at most 8M, or at most 7 M, or most 6 M. For example, the concentration the aqueous solution of one or more inorganic bases is ranging from 1 M to 10 M; preferably from 3 M to 7 M, more preferably from 4 M to 6 M.
With preference, the aqueous solution of one or more inorganic bases is an aqueous solution of KOH at a concentration of at least 1 M, of at least 3 M, or at least 4 M, or at least 5 M. With preference, said at least one alkaline compound is an aqueous solution of KOH at a concentration of at most 10 M, of at most 8 M, or at most 7 M, or most 6 M. For example, the concentration of KOH in a solution of water is ranging from 1 to 10 M; preferably from 3 M to 7 M, more preferably from 4 M to 6 M.
Said step (iii) is performed by injecting carbon dioxide at a potential gradient starting at -0.3 V versus a reference electrode and ending at -2.0 V versus said reference electrode at a sweep rate ranging from 15 mV s^-1 to 35 mV s^-1 or from 20 mV s^-1 to 30 mV s^-1. With preference, the potential gradient starts at -0.5 V versus a reference electrode and ends at -1.8 V versus said reference electrode. For example, the reference electrode is an Ag/AgCl electrode filled with 3.4 M of KCI.
The output flow 25 of the liquid component, in addition to comprising formate, can also comprise one or more selected from acetate, ethanol and propanol.
In step (iv), it is possible to produce an additional output flow 27 of gaseous components which exits through the gas chamber 3 via the gas channel 5. The additional output flow 27 can comprise, for example, one or more ligands selected from hydrogen, carbon monoxide and ethylene. With preference, the additional output flow 27 is devoid of ethylene and but comprises hydrogen and/or carbon monoxide.
With preference, said step (iv) lasts at least 1 hour, more preferably at least 2 hours, even more preferably at least 3 hours, most preferably at least 4 hours, even most preferably at least 5 hours or at least 6 hours.
For example, the input flow of carbon dioxide provided in step (iv) has a flow rate that is ranging from 10 mL min^-1 to 150 mL min^-1, preferably from 15 mL min^-1 to 100 mL min^-1, more preferably from 20 mL min^-1 to 80 mL min^-1, even more preferably from 25 mL min^-1 to 50 mL min^-1. In case, the flow rate of carbon dioxide is too low, for example, in case the flow rate is below10 mL min^-1, the selectivity to formate can be lost.
Advantageously, said step (iv) is performed at room temperature, for example at a temperature ranging from 20°C to 30°C.
Advantageously, said step (iv) is performed at atmospheric pressure, for example at a pressure ranging from 0.09 MPa to 0.11 MPa.

Test Methods

Mass loading of the ink onto the gas diffusion membrane: The membrane was weighed using an analytical balance before deposition and after drying overnight in a vacuum desiccator.

Fourier Transform Infra-Red spectroscopy

Fourier-Transform Infrared (FT-IR) spectra were recorded using a Shimadzu Prestige 21 Spectrometer on solid powder samples in transmission mode.

Transmission Electron Microscopy

TEM analysis was conducted using a Jeol 2100F microscope equipped with Schottky Field Emission electron gun and an ultra-high resolution polar piece.

Scanning Electron Microscopy

SEM images were obtained using a SU-70 Hitachi FEG-SEM.

UV Visible spectroscopy

UV-vis absorption spectra were recorded on liquid samples using an Agilent Cary 100 spectrometer.

¹ H-NMR spectroscopy

Bruker Advance III 300 MHz spectrometer at 300 K has been used. D₂O was used as the lock solvent and an aqueous solution of terephthalic acid (TPA) was used as an internal standard for quantification.

Gas Product analysis

Gas products were detected online using SRI instruments 8610 GC with Ar as the carrier gas. The GC was fitted with a thermal conductivity detector for H₂ quantification, where the gas was separated using a HaySepD precolumn with a 3 m molecular sieve column. Carbon products were separated using either a 3 m molecular sieve column (CH₄) or a 5 m HaySepD column (CO, C₂H₄, C₂H₆) and detected using a flame-ionization detector fitted with a methanizer. Calibration was performed using a custom standard gas mixture in CO₂.
The FE for gas products was calculated using equation (1): $FE (%) = \frac{n_{product} \times n_{electrons} \times F}{(Q_{t = 0} - Q_{t = x})}$
Where n_product is the amount of product (mol), n_electrons is the number of electrons used to make the product, F is the Faraday constant (C mol^-1), Q _t=0 is the charge at the time of the injection, and Q_t=x is the charge at time x seconds before the injection, representing the time taken to fill the sample loop, with x depending on the combined flow rate of Ar and CO₂ as well as the loop size.
The full cell energy efficiency EE_full for formate was calculated using equation (2): ${EE}_{full} (%) = \sum \frac{{FE}_{product} \times E_{product}}{E_{cell}}$
Where E_product is the thermodynamic potential for formate (-0.02 V), E_cell is the measured cell potential, and FE_product is the faradaic efficiency (%). For EE_1/2 values, E_cell = E _1/2 + E _{H 2 O/} _O2, where E _1/2 is the iR-corrected potential measured in the cell (V vs. RHE) and E _{H 2 O/O 2} is 1.23 V.

Liquid Product Analysis

Liquid products were analysed using ¹H NMR with a Pre-SAT180 water suppression method. Formate values were confirmed with a standard calibration using sodium formate solutions (in 5 M KOH) to ensure the accuracy of the internal standard method. The crossover of formate through the anion exchange membrane was accounted for by also liquid sampling from the anode compartment.

Single-Pass Conversion Calculations

Calculations were based on the volumetric flow entering and leaving the cell as well as the consumed flow rates involved in both product generation and reactions with OH^-. The ideal gas law - see equation (3) - was used to relate the volumetric flow rate to the molar flow: ${PQ}_{f} = N_{f} RT$
wherein P is pressure, Q_f the volumetric flow, N_f the molar flow, R the gas constant, and T the temperature. The molar flow can be calculated using the molar values of each gas in the loop as long as the time is taken to fill the loop is known. In the present, the loop has a known size, so the time can be calculated given that the flow rate into the GC is known - see equation (4): $N_{f}^{product} = \frac{n_{product}}{time to fill loop}$
From here, the volumetric flow of products is calculated which represents the additional flow for gasses generated in the CO₂R reaction - see equation (5): $Q_{f}^{product} = N_{f}^{product} \times RT$
The difference between the outlet flow and the sum of the product volumetric flow rates is then simply the flow rate of unreacted CO₂ in the system. This contributes to the flow but is not going into the generation of products - see equation (6): $Q_{f}^{residual} = Q_{f}^{outlet} - \sum Q_{f}^{product}$
From here, the amount of CO₂ (in terms of volumetric flow) consumed by generating gas and liquid products can be calculated using equations (7) and (8): $Q_{gas}^{#} = \sum Q_{f}^{product} \times carbon atoms$
$Q_{liquid}^{#} = \sum \frac{n_{product} \times carbon atoms \times RT}{duration of experiment}$
This allows a total flow rate for CO₂ converted into carbonate to be determined using the inlet flow and the calculated product-based flow rates - see equation (9): $Q_{f}^{carbonate} = Q_{f}^{inlet} - Q_{f}^{residual} - Q_{gas}^{#} - Q_{liquid}^{#}$
A product conversion percentage can also be calculated - as in equation (10) - which represents the amount of consumed CO₂ that goes into product generation instead of consumption through reaction with OH^-: $Product conversion % = \frac{Q_{liq}^{#} + Q_{gas}^{#}}{Q_{f}^{carbonate}} \times 100$
The single-pass conversion efficiency - see equation (11) - is the best representation of the full system efficiency as it takes into account the CO₂ consumed and utilised as well as the products generated. $Single Pass Conversion % = \frac{Q_{liq}^{#} + Q_{gas}^{#}}{Q_{f}^{residual} + Q_{f}^{carbonate}} \times 100$

Examples

Material

Copper nanopowder (Sigma-Aldrich, 25 nm), 4-mercaptopyridine (ACROS organics, 96%), thiophenol (Sigma-Aldrich, ≥99%), N,N-dimethylformamide (Carlo Erba, 99.9%), and methanol (Carlo Erba, 99.9%) were used to form thiol-modified Cu nanoparticles. Polytetrafluoroethylene (PTFE) membranes (SartoriusTM, 0.45 µm pore size) and NafionTM (Sigma-Aldrich, 5 wt.% in lower aliphatic alcohols and water) were used for electrode preparation. Milli-Q H2O and KOH (Sigma-Aldrich, 99.99%) were used for electrochemical experiments. Terephthalic acid (Sigma-Aldrich, 98%), D₂O (99.9% D), and sodium formate (Sigma-Aldrich, ≥99%) were used for NMR experiments and calibration.

Synthesis of 2,6-dimethyl-4-mercaptopyridine

2,6-dimethyl-4-mercaptopyridine (DMSPy) was synthesised according to a previously reported procedure: Under inert conditions, 2,6-dimethyl-4-chloropyridine (1 g, 7.06 mmol) was dissolved in DMF and NaHS (0.99 g, 17.5 mmol) was added. The mixture was heated to 140 °C for 2 h then concentrated under vacuum. The product was purified using column chromatography (silica, DCM : MeOH = 10 : 1) and dried under vacuum. ¹H NMR (thione tautomer, d₆-DMSO): δ (ppm) = 6.90 (s, 2H), 2.20 (s, 6H), 12.22 (s, 1H).

Example 1 - Copper nanoparticles modification

Cu nanopowder (particle size of 25 nm as measured by transmission electron microscopy) (commercially available from Sigma-Aldrich, CAS number 7440-50-8) featuring a native oxide layer formed from ambient exposure were dispersed in N,N-dimethylformamide (DMF) (1 mL) and sonicated for 15 min at 25°C. A solution containing 4-mercaptopyridine (SPy, 10 mM, DMF) was added under inert conditions to obtain a mixture of 130 nmol mg_Cu ^-1. The suspension was sonicated for 1 h at 25°C then the particles were washed three times with DMF, twice with MeOH, and dried in vacuo for 24 h to form SPy-modified nanoparticles.
Thiophenol (SPh) modification was conducted using the same method with the same molar ratio of ligand to Cu nanoparticles.
Cu-SPy nanoparticles (25 nm) (i.e., Cu25-SPy nanoparticles) were formed in the same way but all treatments were carried out in a glovebox to avoid exposure of the particles to oxygen.
Successful anchoring of 4-mercaptopyridine (SPy) on Cu nanoparticles (NPs) was confirmed using Fourier transform infrared (FTIR) spectroscopy. Figure 2 shows indeed FTIR spectra of the bare Cu nanoparticles, the SPy compound and Cu-SPy nanoparticles.
X-ray photoelectron spectroscopy (XPS) has demonstrated that a mixture of thiol and thiolate environments were present.
No differences in morphology or particle size were observed between Cu and Cu-SPy particles. This can be observed in figures 3, 4 and 5 that are transmission electron microscopy images showing images of initial copper nanoparticles, the as-prepared Cu-SPy nanoparticles, and Cu-SPy nanoparticles after 1 hour of electrolysis at -300 mA cm^-2.

Example 2 - gas diffusion electrodes preparation

Gas diffusion electrodes (GDEs) were prepared by airbrushing a methanolic solution of ionomer and copper nanoparticles onto PTFE membranes (0.45 µm pore size) to form a porous network (approximately 10 µm thick) (see figure 6).
An ink containing a weight ratio of 4:3, Cu-SPy:Nafion (5%) was prepared in methanol and sonicated for 1 h at 25 °C. The ink was spray deposited onto a PTFE membrane (Sartorius, 0.45 µm pore size) confined to a circular diameter of 2 cm² to obtain a total mass loading of approximately 1.75 mg cm^-2 after drying under vacuum. The same mass loading was used for Cu, Cu-SPy, Cu-SPh, electrodes.

Example 3 - electrochemical Experiments

The GDEs of example 2 were electrically connected in a gas-fed flow cell for electrochemical testing.
All electrochemical experiments were conducted with a BioLogic VSP300 or VMP3 potentiostat with a 20 A current booster. Ohmic drop (iR) correction was conducted manually using resistance values obtained using electrochemical impedance spectroscopy.
Electrocatalysis was conducted in a custom-made gas-fed flow cell (Sphere Ltd). An anion exchange membrane (Sustanion, pre-treated in KOH), a Ni-foam anode, and a leak-free Ag/AgCl electrode filled with 3.4 M of KCI (reference electrode from Innovative Instruments Ltd.) were used. The PTFE-based GDEs were electrically contacted using Cu tape and confined to a geometric area of 1 cm². Pre-activation was required, which involves consecutive linear sweep voltammograms (LSVs) under CO₂ flow with a sweep rate of 25 mV s^-1 from -0.8 to -1.5 V vs. the reference electrode (Ag/AgCl electrode filled with 3.4 M of KCI) until stabilisation of the current response. A CO₂ inlet flow rate was maintained at 30 mL min^-1 using a mass flow controller (Bronkhurst) for initial studies and the electrolyte solution (5 M KOH) was circulated at a rate of 5.5 mL min^-1 using a peristaltic pump. The catholyte was constantly purged with Ar at a fixed flow rate of 35 mL min^-1 and the outlet was connected to the CO₂ outlet gas trap to carry any liquid saturated gas products to the GC. Additionally, calibrated flow meters (MesaLabs Defender 530+ and Ellutia 7000) were used to verify flow rates before and after the GC inlet to ensure the correct flow value was recorded and to establish the portion of CO₂ utilised to account for mass balance. The catholyte and anolyte volumes were 20 mL and the electrolysis time was 1 h for all experiments apart from the 6 h electrolysis, where the volumes were increased to 140 mL.
Gas products for CCE experiments were recorded at 20 and 45 minutes to ensure consistent selectivity and liquid products were taken after 1 h. For the 6 h electrolysis experiment, liquid samples were extracted using a syringe every 2 h and the gas products recorded every 30 minutes.
Potentials were converted to the reversible hydrogen electrode scale (RHE) using the Nernst equation: E_RHE = E_Ag/AgCl + 0.206 + 0.0591 × pH and were iR-corrected to account for the solution resistance, which was obtained from electrochemical impedance spectroscopy scans. Note that this does not account for any local pH changes at the electrode/solution interface, however, only small changes are expected for such highly alkaline systems.
A geometrical molecular loading of ≈15 nmol cm^-2 was determined by UV-Vis spectroscopy after reductive desorption of the thiol at highly cathodic potentials. The electrodes were exposed to potentials more negative than -2.2 V vs. Ag/AgCl/KCl3.4M under Ar flow for 1 h to ensure that all of the molecule was removed. Under Ar flow (see figures 7 and 8), the desorption of SPy forms 4,4'-dithiodipyridine with an absorption peak at 283 nm in KOH. The absorption peak for the desorbed molecule was correlated with a calibration curve for the complex to obtain a molar loading of SPy based on the geometrical area (1 cm²). The SPy loading value obtained was 14.6 ± 2.4 nmol cm^-2. Under CO₂ flow, the desorbed thiolate reacts to form a new species with an altered UV-Vis spectrum with a much higher absorption coefficient (peak at 258 nm used for analysis) - this likely corresponds to a thiocarbonate derivative. The low concentrations excluded conventional molecular characterisation, however through ligand stripping under CO₂ flow for a blank electrode, and comparison with the absorption peak at 258 nm for the experiments conducted under CO₂ flow, an approximate percentage of desorbed species could be obtained.
As shown by figure 9, it has been observed that in alkaline conditions (5 M KOH) these GDEs could reach -500 mA cm^-2 at applied potentials of approximately -0.5 V vs. RHE
There were no significant differences in the current-voltage responses for Cu and Cu-SPy samples confirming the negligible influence of the molecule on the physical properties of the catalyst (Figure 10). Also, no mass transport-limited current was observed.
Figure 11 shows the FE_HCOO- values obtained from controlled current electrolysis (CCE) over 1 hour with varying current densities for Cu and Cu-SPy electrodes. It was revealed that Cu-SPy GDEs displayed a high selectivity for HCOO^- whereas unmodified Cu showed a wide product distribution. Tables 1 and 2 display the faradaic efficiencies for the unmodified copper nanoparticles and the copper nanoparticles as prepared in the present disclosure respectively. FE_HCOO- values of 81 ± 4.3% at -100 mA cm^-2 and 72 ± 1.5% at -300 mA cm^-2 were obtained for Cu-SPy GDEs with a maximal partial current density for formate (j_HCOO-) of 217 ± 4.6 mA cm^-2.
Table 3 displays Cu-SPh, Cu25-SPy, and Cu-DMSPy faradaic efficiencies of main products from CO₂ reduction (1h) at different current densities. The results highlight the advantage of pyridine-containing ligands over thiophenol ligands.
At -500 mA cm^-2, the selectivity was lower (FE_HCOO- = 41 ± 3.5%) suggesting that ligand loss occurs at these potentials (≈ -0.5 V vs. RHE). This was verified using solution-phase UV-Vis spectroscopy of the electrolyte solutions following electrolysis.
Figure 12 displays the faradaic efficiency for all the products obtained following the present disclosure.
Figure 13 displays the partial current densities for all the products obtained following the present disclosure
It is also highlighted that at -300 mA cm^-2 (≈ -0.4 V vs. RHE) about 4% desorption was observed whereas at -500 mA cm^-2 about 7% of the total ligand loading was lost after 1h.
The loss of formate selectivity was correlated with an increase in ligand desorption using UV-Vis analysis throughout longer-term electrolysis at -500 mA cm^-2, as shown in figure 14 in which the correlation between FE_HCOO- and the percentage of SPy ligand lost to solution as determined by UV-Vis from an electrode held at -500 mA cm^-2 over the course of 2 hours is illustrated. The FE_HCOO- of a bare Cu GDE was recovered after 2 hours showing that the permanent effects of molecule desorption are minimal and highlighting the key role of the SPy compound in directing selectivity towards formate.
As shown in figures 15, 16 and 17 obtained by scanning electron microscopy, no clear morphological changes which could be responsible for ligand desorption were observed over the electrolysis.

The cathodic energy efficiency for HCOO^- (EE_½HCOO-) was 55 ± 3.2% at -300 mA cm^-2 with an optimal single-pass efficiency of 4.4% attained through alteration of the CO₂ flow rate as illustrated in figure 18. The optimum flow rate that does not affect FE_HCOO- is 15 mL min^-1. Each point was determined from 30-minute electrolysis at -300 mA cm^-2 with a fresh electrolyte solution.

Table 1 - Cu, Faradaic efficiencies of main products from CO₂ reduction (1h) at different current densities with corresponding average half-cell voltages and energy efficiencies for HCOO^-
j / mA cm^-2	E / V vs. RHE (-iR)	Faradaic Efficiency / %								EE_1/2 HCOO^- / %
j / mA cm^-2	E / V vs. RHE (-iR)	H₂	CO	HCOO^-	C₂H₄	Ethanol	Acetate	Propanol	Total
-100	-0.38 ± 0.01	12.4 ± 1.5	52.8 ± 2.6	26.0 ± 2.7	10.8 ± 1.7	1.4 ± 0.4	0.2 ± 0.2	1.7 ± 0.8	105.3 ± 0.3	20.3 ± 2.2
-300	-0.45 ± 0.04	8.2 ± 0.8	40.8 ± 3.1	21.8 ± 1.5	21.0 ± 0.8	2.7 ± 0.5	0.4 ± 0.1	2.9 ± 0.3	97.9 ± 2.3	16.3 ± 1.5
-500	-0.53 ± 0.08	9.6 ± 1.4	29.2 ± 4.0	16.8 ± 1.0	29.4 ± 2.8	3.9 ± 0.2	1.1 ± 0.4	3.6 ± 0.4	93.9 ± 1.9	12.0 ± 1.2

Table 2 - Cu-SPy, Faradaic efficiencies of main products from CO₂ reduction (1h) at different current densities with corresponding average half-cell voltages and energy efficiencies for HCOO^-
j / mA cm^-2	E / V vs. RHE (-iR)	Faradaic Efficiency / %								EE_1/2 HCOO^- / %
j / mA cm^-2	E / V vs. RHE (-iR)	H₂	CO	HCOO^-	C₂H₄	Ethanol	Acetate	Propanol	Total
-100	-0.33 ± 0.01	11.6 ± 4.2	8.3 ± 6.7	81.4 ± 4.3	-	-	-	-	101.3 ± 0.5	64.9 ± 3.7
-300	-0.41 ± 0.09	6.0 ± 1.9	21.1 ± 3.1	72.4 ± 1.5	0.8 ± 0.3	0.2 ± 0.3	-	-	102.6 ± 1.2	55.2 ± 3.2
-500	-0.46 ± 0.06	8.1 ± 2.4	35.2 ± 4.5	41.3 ± 3.5	11.8 ± 4.3	1.5 ± 1.0	0.3 ± 0.3	1.4 ± 0.6	99.7 ± 0.6	30.7 ± 3.4

Table 3 - Cu-SPh, Cu25-SPy, and Cu-DMSPy faradaic efficiencies of main products from CO₂ reduction (1h) at different current densities
j / mA cm^-2	Faradaic Efficiency / %
j / mA cm^-2	H₂	CO	HCOO^-	C₂H₄	Ethanol	Acetate	Propanol	Total
Cu-SPh
-100	16.0	39.5	28.1	14.7	1.3	0.6	0.5	100.7
-300	17.0 ± 1.1	35.0 ± 1.3	25.4 ± 1.5	20.6 ± 5.1	1.7 ± 0.5	0.6 ± 0.1	1.9 ± 0.4	98.2 ± 2.4
-500	23.0	13.0	17.7	34.4	3.0	2.0	3.2	96.4
Cu25-SPy
-100	15.9	4.5	79.9	0	0	0	0	100.4
-300	11.3	15.0	71.9	2.4	0.8	0.3	0.6	102.4
-500	31.1	17.2	26.4	14.7	2.3	1.1	1.0	93.7
Cu-DMSPy
-300	13.1	15.8	67.9	2.1	0.4	0.3	0.3	99.9

The stability of the gas-fed flow cell at a potential required to drive CO₂ reduction coupled with oxygen evolution using a Ni foam anode, by conducting CCE at -300 mA cm^-2 for 6 h (see figure 19) was studied. A stable potential of -2.7 V was observed and an average FE_HCOO-value of approximately 67% was maintained giving a full-cell energy efficiency for HCOO^- of 30.5% and single-pass conversion efficiency of 4.5%. After 6 hours, a fresh solution of electrolyte was added, which is depicted in figure 19 by the vertical line at 6 hours. Retention of the molecule after electrolysis was confirmed with UV-Vis spectroscopy, which showed no additional signal from desorbed species after 6 h. The molecular stability in a gas-fed flow cell, therefore, permits the use of a broad range of organic thiols for CO₂R.
Figure 20 is the ¹H NMR spectrum of the catholyte solution taken after carbon dioxide electrolysis and that shows the presence of formate (main product obtained from the electrolysis), terephthalic acid (as reference) and water (coming notably from the aqueous electrolyte solution).

Claims

Gas diffusion electrode (7) suitable for carbon dioxide electrolysis, said gas diffusion electrode (7) having a gas diffusion membrane (17), the gas diffusion electrode (7) further comprising an ink (19) deposited on the gas diffusion membrane (17); wherein the ink (19) comprises an ion-conducting polymer, said gas diffusion electrode (7) is characterized in that the ink (19) further comprises a catalyst comprising copper nanoparticles functionalized with one or more pyridine -containing ligands, wherein the one or more pyridine-containing ligands have an anchoring group comprising a sulphur atom tethered to the copper nanoparticles.
The gas-diffusion electrode according to claim 1 is characterized in that the one or more pyridine-containing ligands are selected from 4-pyridylethylmercaptan, 4-mercaptopyridine, 2,6-dimethyl-4-mercaptopyridine, 2-mercaptopyridine and any mixture thereof.
The gas-diffusion electrode according to any one of claims 1 or 2 is characterized in that the one or more pyridine-containing ligands are present in a surface concentration ranging from 5 nmol cm^-2 to 40 nmol cm^-2 as determined by reductive desorption and UV-visible spectroscopy as set out in the description; preferably ranging from 8 nmol cm^-2 to 20 nmol cm^-2.
The gas-diffusion electrode according to any one of claims 1 to 3 is characterized in that, the copper nanoparticles comprise facets selected from Cu(200) facets, Cu(111) facets and any mixture thereof.
The gas-diffusion electrode according to any one of claims 1 to 4 is characterized in that the copper nanoparticles have an average diameter ranging from 5 nm to 200 nm as measured by transmission electron microscopy, preferably from 20 nm to 100 nm.
The gas-diffusion electrode according to any one of claims 1 to 5 is characterized in that the ink (19) has a weight ratio of the copper nanoparticles functionalized with one or more pyridine-containing ligands over the ion-conducting polymer ranging from 10 to 40.
Gas-fed flow cell suitable for carbon dioxide electrolysis, said gas-fed flow cell (1) comprising a gas chamber (3), a catholyte chamber (9) and an anolyte chamber (13), wherein said gas chamber (3) is separated from the catholyte chamber (9) by a gas diffusion electrode (7) which is attached to the catholyte chamber (9) by an electrically conductive connection, said gas diffusion electrode (7) having a gas diffusion membrane (17) being comprised within said gas chamber (3),
wherein said catholyte chamber (9) and said anolyte chamber (13) are separated by an anion exchange membrane (11), and

wherein said catholyte chamber (9) and said anolyte chamber (13) comprise respectively a cathode and an anode (15),

said gas-fed flow cell (1) is characterized in that the gas diffusion electrode (7) is according to any one of claims 1 to 6.
Method for producing the gas diffusion electrode (7) suitable for carbon dioxide electrolysis according to claims 1 to 6, said method is characterized in that it comprises the following steps:
a) providing copper nanoparticles;

b) obtaining a first dispersion by dispersing the copper nanoparticles into a solvent selected from water and/or a first organic solvent;

c) adding pyridine-containing ligands to the first dispersion to obtain a suspension with copper nanoparticles functionalized with one or more pyridine-containing ligands, wherein the one or more pyridine-containing ligands have an anchoring group comprising at least one sulphur atom;

d) obtaining a second dispersion by dispersing the copper nanoparticles functionalized with pyridine-containing ligands into a second organic solvent;

e) adding an ion-conducting polymer to the second dispersion to obtain an ink (19);

f) providing a gas-diffusion membrane (17) and depositing said ink (19) onto said gas-diffusion membrane (17) to obtain the gas diffusion electrode.
Process for electrolysing carbon dioxide, said process comprising the following steps:
i) providing a gas-fed flow cell (1);

ii) providing at least one electrolyte flow (21) into said gas-fed flow cell (1) wherein the at least one electrolyte flow is a catholyte flow and an optional anolyte flow;

iii) activating said gas-fed flow cell (1);

iv) providing an input flow (23) of carbon dioxide to produce an output flow (25) of liquid components comprising at least formate;

v) recovering said output flow (25) comprising at least formate;
said process is characterized in that the gas-fed flow cell (1) provided at step (i) is as defined in claim 7, and in that said step (iii) is performed by injecting carbon dioxide at a potential gradient starting at -0.3 V versus a reference electrode and ending at -2.0 V versus said reference electrode at a sweep rate ranging from 15 mV s^-1 to 35 mV s^-1, the reference electrode being preferably an Ag/AgCl electrode filled with 3.4 M of KCI.
The process according to claim 9, is characterized in that the electrolyte flow (21) provided in step (ii) is a flow of an aqueous solution of one or more inorganic bases; with preference, the aqueous solution of one or more inorganic bases has a concentration ranging from 1 M to 10 M; preferably from 3 M to 7 M or from 5 M to 10 M.
The process according to claim 9 or 10, is characterized in that the one or more inorganic bases are alkali selected from NaOH, KOH, Ca(OH)₂, LiOH, Mg(OH)₂, RbOH, CsOH and any mixture thereof; and/or in that it is operated with a cathodic voltage no lower than -0.5V vs. reversible hydrogen electrode.
Use of a catalyst for electrochemical carbon dioxide reduction reactions to produce formate, the use is characterized in that the catalyst is comprising copper nanoparticles functionalized with one or more pyridine-containing ligands, wherein the one or more pyridine-containing ligands have an anchoring group comprising a sulphur atom tethered to the copper nanoparticles; preferably, the one or more pyridine-containing ligands are selected from 4-pyridylethylmercaptan, 4-mercaptopyridine, 2,6-dimethyl-4-mercaptopyridine, 2-mercaptopyridine and any mixture thereof.
The use according to claim 12 is characterized in that the one or more pyridine-containing ligands are present in a surface concentration ranging from 5 nmol cm^-2 to 40 nmol cm^-2 as determined by reductive desorption and UV-visible spectroscopy as et out in the description; preferably ranging from 8 nmol cm^-2 to 20 nmol cm^-2.
The use according to any one of claims 12 to 13 is characterized in that, the copper nanoparticles comprise facets selected from Cu(200) facets, Cu(111) facets and any mixture thereof.
The use according to any one of claims 12 to 14 is characterized in that the copper nanoparticles have an average diameter ranging from 5 nm to 200 nm as measured by transmission electron microscopy, preferably from 20 nm to 100 nm.