Efficient electrolyzer for CO₂ splitting in neutral water using earth-abundant materials

Significance

Electrochemical CO₂-to-CO conversion is one important option for storing intermittent, renewable electricity into chemical bonds so as to produce fuels and to use CO₂ as a feedstock for chemicals. The setup of an electrolyzer, associating cheap and abundant materials able to split CO₂ into CO and O₂, in environmentally friendly conditions (neutral pH, ambient temperature) with a high selectivity and stability, and a 50% energy conversion efficiency is reported. The results open the way to solar energy driving of the CO₂ /CO + 1/2 O₂ splitting by associating the electrochemical cell with a light-to-electricity conversion device, and more generally with surplus electricity from renewable intermittent sources.

Abstract

Low-cost, efficient CO₂-to-CO+O₂ electrochemical splitting is a key step for liquid-fuel production for renewable energy storage and use of CO₂ as a feedstock for chemicals. Heterogeneous catalysts for cathodic CO₂-to-CO associated with an O₂-evolving anodic reaction in high-energy-efficiency cells are not yet available. An iron porphyrin immobilized into a conductive Nafion/carbon powder layer is a stable cathode producing CO in pH neutral water with 90% faradaic efficiency. It is coupled with a water oxidation phosphate cobalt oxide anode in a home-made electrolyzer by means of a Nafion membrane. Current densities of approximately 1 mA/cm² over 30-h electrolysis are achieved at a 2.5-V cell voltage, splitting CO₂ and H₂O into CO and O₂ with a 50% energy efficiency. Remarkably, CO₂ reduction outweighs the concurrent water reduction. The setup does not prevent high-efficiency proton transport through the Nafion membrane separator: The ohmic drop loss is only 0.1 V and the pH remains stable. These results demonstrate the possibility to set up an efficient, low-voltage, electrochemical cell that converts CO₂ into CO and O₂ by associating a cathodic-supported molecular catalyst based on an abundant transition metal with a cheap, easy-to-prepare anodic catalyst oxidizing water into O₂.

The production of carbon-based fuels or chemicals using the most abundant carbon source (CO₂) requires designing efficient, cheap, selective, and sustainable processes able to convert CO₂ into useful products (1–7). Carbon monoxide production is an important step to fuels because it can be used as a feedstock in the Fischer–Tropsch process. Compared with water splitting, electrochemical reduction of CO₂ into CO is a greater challenge. This is particularly true when aiming to carry out this reaction selectively in friendly conditions, namely at neutral pHs, ambient temperature, and with abundant and cheap materials as catalysts as opposed to solid-state high-temperature electrolyzers (8).

We recently discovered that substitution of the four paraphenyl hydrogens of iron tetraphenylporphyrin by trimethylammonio groups provides a water-soluble iron porphyrin (WSCAT) able to catalyze selectively the electrochemical conversion of CO₂ into CO in neutral water in homogeneous conditions (9). The next challenge was to efficiently immobilize this molecular catalyst onto the cathode and to set up an integrated electrochemical cell able to split CO₂ and H₂O into CO and O₂ according to

$${\text{CO}}_2+2\hspace{0.17em}{\text{H}}^{+}+2\hspace{0.17em}{\text{e}}^{-}\rightleftarrows \text{CO}+{\text{H}}_2\text{O}\hspace{1em}\left(E_{{\text{CO}}_2/\text{CO}}^{\text{0}}=-0\text{.}11\hspace{0.17em}\text{V}\right),$$

$${\text{H}}_2\text{O}\rightleftarrows \mathrm{½}\hspace{0.17em}{\text{O}}_2+2\hspace{0.17em}{\text{H}}^{+}+2\hspace{0.17em}{\text{e}}^{-}\hspace{1em}\left(E_{{\text{O}}_2/{\text{H}}_2\text{O}}^{\text{0}}=1\text{.}23\hspace{0.17em}\text{V}\right),$$

$${\text{CO}}_2\rightleftarrows \text{CO}+\mathrm{½}\hspace{0.17em}{\text{O}}_2\hspace{1em}\left(\mathrm{\Delta }{G}^{\text{0}}=2\text{.}68\hspace{0.5em}\text{eV}\right),$$

[potentials referred to the standard hydrogen electrode (SHE)].

Immobilization of the catalyst was achieved by preparation of a suspension containing Nafion, WSCAT, and carbon powder (Materials and Methods) (10). This solution was then sprayed onto a carbon support (glassy carbon electrode for cyclic voltammetry experiments and carbon felt or carbon Toray for electrolysis) and air-dried. Interactions between the positively charged catalyst and the negatively charged functionalities of the ionic polymer secure robust integration of the catalyst into the coated film. This was attested to by the absence of UV-vis signal corresponding to an iron porphyrin in a water solution in which the electrode was immersed for a few hours. The catalytic film consists of a thin coating of the electrode surface as confirmed by scanning electron microscopy (SEM) (Fig. 1).

Fig. 1.

SEM images of the Nafion/carbon powder/WSCAT (Left) and CoP_i (Right) films. (A and B) Electrode structures. (C and D) Catalytic films on the supporting material (polymer-wrapped carbon microfiber and stainless steel microwire, respectively). (E and F) High-magnification images of the films.

Conductive and catalytic properties for CO₂ reduction of the prepared electrode were characterized by cyclic voltammetry carried out in quasi-neutral water with no added buffer, thus preventing acid reduction (Fig. 2). Under argon, almost no capacitive current and no faradaic current were observed in the absence of carbon powder in the film, whereas the electrode exhibited a large capacitive current when carbon powder was incorporated in the coated film.

Fig. 2.

Cyclic voltammetry (pH 6.7). Nafion/WSCAT film deposited on a 3-mm-diameter glassy carbon electrode, v = 0.1 Vs⁻¹, pH = 6.7. Light gray, film without carbon powder under argon. Blue, film plus carbon powder under argon. Red, film plus carbon powder under 1 atm of CO₂.

This indicates that carbon powder renders the film conducting and thus makes addressable the molecular catalyst contained in the film. This is further confirmed by the observation of faradaic waves when started from the Fe^III complex. Although not being well defined, these waves may be assigned to the Fe^III/II, Fe^II/I redox couples. The slight increase of the current observed at more negative potentials (approximately −1.15 V vs. SHE) presumably reflects some catalysis of water reduction. In the presence of CO₂, a large increase of the current is seen, confirming the catalytic activity of WSCAT immobilized in the Nafion film toward CO₂ reduction. The onset of the catalytic wave is approximately −0.8 V vs. SHE corresponding to an overpotential of

$$\mathit{\eta }=E_{{\text{CO}}_2/\text{CO}}^0-\left(\text{RT}\ln 10/\text{F}\right)\times \text{p}\mathrm{H}=390\text{mV}.$$

An electrochemical membrane-separated cell was then assembled to couple the CO₂ reduction half-reaction to water oxidation. Although a significant overpotential is usually associated with the latter reaction, important progress has been recently made leading to a cheap and abundant material, easy to prepare, well characterized, and efficient at neutral pH, namely electrodeposited phosphate cobalt oxide (CoP_i), which could be used as water oxidation catalyst on the anodic side (6). A thin (50 mC/cm², i.e., approximately 330 nm) CoP_i film was thus deposited on a small stainless steel gauze (Fig. 1), with, as anodic electrolyte, a phosphate buffer (0.4 M, pH 7.3) ensuring a minimal anodic overpotential. The cell was constructed from flasks containing the electrolytes with a proton-permeable Nafion membrane and both electrodes clamped between the flasks (Fig. 3).

Fig. 3.

Proton exchange membrane electrolysis cell configuration. Cathodic half-reaction is CO₂ reduction to CO on a carbon electrode coated with a Nafion/carbon powder/WSCAT film and anodic half-reaction is oxidation of H₂O to O₂ on a stainless steel gauze with a CoP_i electrodeposited film.

A reference electrode was also introduced in the cathodic compartment so that both the cell voltage and the cathode potential electrode could be measured during electrolysis.

The electrochemical characteristics of the cell were obtained by applying a potential scan to the cathode while simultaneously recording cell voltage, hence leading to the potential of both the cathode and the anode including ohmic drop vs. the reference electrode and the current (Fig. 4A). At pH 7.3, the onset of the current (at 1 mA/cm²) corresponds to a cell voltage of 2.3 V with

$${\mathit{\eta }}_{cathodic}=\left\{E_{{\text{CO}}_2/\text{CO}}^0-\left(\text{RT}\ln 10/\text{F}\right)\times \text{p}\mathrm{H}\right\}-E_{cathodic}=320\text{mV}$$

and

$$\begin{array}{l}{\mathit{\eta }}_{anodic}+Ri=U_{\text{cell}}-E_{cathodic}-\left\{E_{{\text{O}}_2/{\text{H}}_2\text{O}}^0-\left(\text{RT}\ln 10/\text{F}\right)\times \text{p}\mathrm{H}\right\}\\ =630\text{mV}.\end{array}$$

An independent measurement indicated a 10-Ω cell resistance, amounting to an ohmic drop contribution of 50 mV to the overpotential.

Fig. 4.

Cell characterization and electrolysis at pH 7.3. (A) Current density as function of the electrode potential (vs. SHE) for both the cathode and the anode. (B) Entire cell overpotential as function of current density (dotted line corresponds to ohmic drop corrected data). (C) Charge transferred as function of time for an electrolysis at controlled cathodic potential, E_cathodic = −0.96 V vs. SHE (the dotted line corresponds to the partial charge for CO production). (D) Cell voltage as function of time for an electrolysis at controlled cathodic potential, E_cathodic = −0.96 V vs. SHE. (E) Product selectivity as function of time for an electrolysis at controlled cathodic potential, E_cathodic = −0.96 V vs. SHE. H₂ (o) and CO (red •). (F) Cell energy efficiency as function of time for an electrolysis at controlled cathodic potential, E_cathodic = −0.96 V vs. SHE.

A first series of electrolysis was performed at controlled cathodic potential (E_cathodic = −0.86 V vs. SHE) for several hours, with an expansion vessel over the flasks to assess product selectivity and evaluate faradaic efficiency. The averaged current density was 0.7 mA/cm², the cell voltage 2.4 V (corresponding to 1.05-V overpotential), and gas chromatography analysis of gas in the headspace above solution leads to faradaic efficiencies of 90% for CO along with 10% for H₂ in the cathodic compartment and 99% for O₂ in the anodic compartment. It is worth noting that no hydrocarbons were detected in the cathodic compartment headspace. Ionic chromatography analysis of the cathodic electrolyte shows only traces of formate and oxalate. These results clearly indicate a high selectivity and a high faradaic efficiency of the cell for CO₂ and H₂O conversion to CO and O₂, respectively. A second series of electrolysis was performed at a controlled cathodic potential (E_cathodic = −0.96 V vs. SHE) for 30 h under CO₂ flow to evaluate cell stability. A linear increase of the charge passed vs. time was observed, corresponding to an averaged 1 mA/cm² current density (Fig. 4C); the cell voltage remained stable at a value of 2.5 V (Fig. 4D). Analysis of the gas mixture produced in the cathodic chamber over the course of the electrolysis showed an excellent stability of the cell selectivity (Fig. 4E). Note that the same experiment carried out with iron tetraphenylporphyrin instead of WSCAT led to a rapid decrease of the current to zero and a poor CO selectivity. The cell energy efficiency (EE) at 1 mA/cm² was evaluated to be 50% from the product of the faradaic yield for CO production (FY_CO = 0.9) and the ratio of the thermodynamics of the reaction over cell voltage: EE = FY_CO × {E⁰_O2/H2O − E⁰_CO2/CO}/U_cell (Fig. 4F). From the overpotential measured at the cathode (η_cathodic = 330 mV) and at the anode (η_anodic = 720 mV, ohmic drop being approximately 100 mV), the cathodic energy efficiency was evaluated as

$$\begin{array}{l}E{E}_{cathodic}=F{Y}_{\text{CO}}\times \\ \left\{E_{{\text{O}}_2/{\text{H}}_2\text{O}}^0-E_{{\text{CO}}_2/\text{CO}}^0\right\}/\left\{E_{{\text{O}}_2/{\text{H}}_2\text{O}}^0-E_{{\text{CO}}_2/\text{CO}}^0+{\mathit{\eta }}_{cathodic}\right\}=72\%,\end{array}$$

and the anodic energy efficiency as

$$\begin{array}{l}E{E}_{anodic}=F{Y}_{\text{CO}}\times \\ \left\{E_{{\text{O}}_2/{\text{H}}_2\text{O}}^0-E_{{\text{CO}}_2/\text{CO}}^0\right\}/\left\{E_{{\text{O}}_2/{\text{H}}_2\text{O}}^0-E_{{\text{CO}}_2/\text{CO}}^0+{\mathit{\eta }}_{anodic}\right\}=63\%.\end{array}$$

Further improvements of the overall cell energy efficiency concern both the anodic and cathodic catalytic films, and will allow running electrolysis at higher current densities. However, the results described here demonstrate that supported molecular catalyst based on the cheapest transition metal can be used to efficiently split CO₂ to CO and O₂ in neutral water and the performance of the low-cost, simply designed cell provides a robust basis to couple this electrolysis with a photovoltaic device producing electricity from solar energy (11–13), opening an avenue to CO₂-based solar fuels.

Materials and Methods

CO₂ splitting was performed in a two-compartment electrochemical cell where cathodic electrolyte was a CO₂-saturated 0.1 M KCl + 0.5 M KHCO₃ aqueous solution at pH 7.3 and anodic electrolyte was a 0.4 M potassium phosphate buffer at pH 7.3 degassed under argon. A Nafion NRE-212 membrane enabled proton transfer between aforesaid compartments. Cathodes were manufactured by spraying a suspension of WSCAT catalyst, conductive carbon powder, and Nafion solution in 2-propanol onto Toray carbon paper. CoP_i anodes were prepared by electrodeposition of a thin cobalt film on a stainless steel mesh in a 0.1 M potassium phosphate solution at pH 7 containing 0.5 mM Co²⁺. The headspace of the cathodic chamber was continuously purged with CO₂ and periodic manual injections in a gas chromatograph gave CO₂ reduction products selectivity. The anodic chamber was originally filled with electrolyte; gas evolution was gauged over time and oxidation products were investigated by gas chromatography after electrolysis had run to completion. Experimental details concerning the synthesis of WSCAT and full complementary experimental details can be found in the SI Appendix.