Production of formate by CO₂ electrochemical reduction and its application in energy storage

Abstract

Production of liquid fuels by electrochemical CO₂ reduction (eCO₂R) is an attractive option for energy storage in the form of renewable energy. This study focuses on efficient formate production using an eCO₂R system and its application in generating power using a direct formate fuel cell (DFFC). A carbon black supported SnO₂ catalyst was used for the eCO₂R in a gas diffusion reactor using a 1.0 M KOH electrolyte. An average faradaic efficiency of 80% for formate production was achieved over a wide electrode potential range (−0.63 to −1.43 V vs. RHE). Since the overall current density varied linearly with the overpotential, the rate of formate production could be easily controlled by varying the applied potential. At a current density of 251 mA cm⁻² (−1.43 V), a high formate production rate was achieved at 3 mg min⁻¹ cm_WE⁻² resulting in 0.5 M formate being produced within 1 hour. This formate solution was directly used as the fuel for a DFFC, without pre-treatment. The fuel cell consisted of a Pd–CeO₂/C anode and FeCo/C cathode and produced a peak power density of 92 mW cm⁻². A closed loop of “electricity–formate–electricity” has been realized in this study, signifying the promising future of sustainable CO₂ conversion to liquid fuels for CO₂ fixation as well as for energy storage.

---

Introduction

Electrochemical CO₂ reduction (eCO₂R) in aqueous media has aroused attention in the CO₂ utilization field. Converting CO₂ and water with renewable electricity to produce carbonaceous chemicals and fuels provides a new means for achieving a low-carbon economy. The carbonaceous products from eCO₂R can be wide-ranging, both gases, such as CO, CH₄, and C₂H₄, which can be formed at current efficiencies greater than 70%,^1–3 and liquid products, such as formate, methanol and ethanol,⁴ produced with faradaic efficiencies greater than 90%,^5,6 40%,⁷ and 40%,⁸ respectively. Liquid products have advantages in terms of transportation, storage, and handling. In terms of electric energy generation, formate is superior to alcohols as it is non-flammable and non-toxic. The thermodynamic cell voltage of a formate fuel cell is 1.48 V,⁹ which is higher than that of methanol (1.21 V), ethanol (1.14 V) and hydrogen (1.23 V).

Vo et al.¹⁰ demonstrated the concept of electricity storage and release using formate produced by CO₂ electro-reduction and investigated the possibility of this closed-loop energy conversion for the first time. However, electric energy generation from formate produced from CO₂ has not been realized due to low concentration of the products. In a normal direct formate fuel cell (DFFC), a high concentration of formate (≥1 M) is used.¹¹ Table S1† summarizes related studies^12–27 of selective formate production by eCO₂R published in the past 5 years, with their corresponding formate yields and its concentration in the residual electrolyte – which were rarely mentioned in the original publications. The formate concentrations are characteristically below 0.1 M after 1 hour of electrolysis, far less than the level required for energy generation and storage. The reasons for the low formate concentration could be poor CO₂ mass transfer efficiency in an aqueous medium and low current density by using an electrolyte with low alkalinity. Lee et al.²⁸ developed a catholyte-free electrolyzer using a vapor stream on the cathode, with a 25 cm² electrode area. A formate concentration of 0.9 M was produced at 70 °C but it decreased to 0.5 M at 30 °C. However, supplying a vapour stream would increase the energy and cost of such a process.

In this study, we developed a cost-effective and high-efficiency gas diffusion system to achieve controllable formate production by eCO₂R in an aqueous medium. 0.5 M formate could be produced directly in alkaline media within 1 hour under ambient conditions. This formate product was then directly utilized as the fuel for a DFFC without pre-treatment.

Experimental section

SnO₂/C-GDE cathode fabrication

An SnO₂/C catalyst was prepared by physically mixing SnO₂ (nanopowder, ≤ 100 nm, Sigma-Aldrich) and carbon black (Vulcan XC 72R, Cabot) particles in the desired mass ratios of 0.5, 1, and 3.5. Individual SnO₂ and C were also used for comparison. To apply the catalyst on a tailored gas diffusion electrode (GDE), the catalyst ink was prepared by dispersing 15 mg catalyst powder in 200 μL isopropanol (>99.8%, VWR Chemicals) and 33 μL Nafion binder (10 wt% in total catalyst) (Nafion suspension, 5 wt%, Sigma-Aldrich). The ink was sonicated for 10 hours before coating onto a 2 cm² surface of a gas diffusion layer (GDL) (H2315 I2 C6, Freudenberg) using an air brush (Evolution AL plus, Harder & Steenbeck) and drying at 45 °C for several minutes. Painting and drying were repeated until the desired total catalyst loading of ∼4 mg cm⁻² was achieved.

Catalyst characterization

X-ray diffraction (XRD) was used to evaluate the crystal structure of the catalyst (Philips X-ray diffractometer PW 1730 equipped with a Cu X-ray tube (Cu-Kα; λ = 0.154 nm)) operated at 40 kV and 40 mA. A scanning electron microscope (SEM, Hitachi SU-70) equipped with an energy dispersive X-ray detector (EDX, Bruker Quantax 400) was used to analyse the catalyst surface morphology and composition.

The eCO₂R system and electrochemical analysis

A GDE reactor fabricated using 3D printing was used to carry out the mass transfer enhanced eCO₂R, as illustrated in a previous study.²⁹ Fig. S1† shows the 3D drawing of the reactor set-up. The prepared SnO₂/C was applied to the cathode with a geometric surface area 2 cm² and the anode was a coiled platinum wire. The RE-61AP reference electrode (0.117 V vs. SHE, ALS CO., Ltd) was used for the alkaline solution. The applied potentials (vs. RE-61AP) in the three-electrode system were all converted to the reversible hydrogen electrode (RHE) according to eqn (1).

E (vs. RHE) = E (vs. RE-61AP) + 0.117 V + 0.0591 V × pH (1)

All the electrochemical reactions and measurements were carried out at ambient temperature and pressure using a potentiostat (Metrohm Autolab PGSTAT128N). The flow rate of CO₂ (BOC 99.99%) was maintained at 15.8 ml min⁻¹ using a flow meter (Cole-Parmer TMR1-010462). 1 M KOH (Emsure®, 85%) and 5 M KOH solutions were used as the catholyte and anolyte, respectively, separated by a cation exchange membrane (CEM) (F-950, Fumapem). In the catholyte flow mode, a peristaltic pump (120U/DM2, Watson Marlow) was used to supply fresh catholyte at a flow rate of 1.0 ml min⁻¹ to maintain the local pH and to remove liquid products to attain reaction equilibrium. In the catholyte fixed mode, 13 ml catholyte was maintained without circulating and flowing.

Cyclic voltammetry (CV) was carried out in triplicate between −1.46 to 0.84 V (vs. RHE) with a scan rate of 10 mV s⁻¹ to initially explore the cathode electrochemical behaviour. eCO₂R was measured chronoamperometrically (CA) by recording the current at a particular applied potential over reaction time. The current density (j) was calculated based on the geometric surface area of 2 cm² of the working electrode.

Product analysis of eCO₂R

A gas chromatograph (Shimadzu Tracera GC-2010) equipped with a Barrier Ionization Discharge (BID) detector was used to analyze gas and alcoholic liquid products. A ShinCarbon ST micropacked column 80/100 (Restek) was used to quantitatively analyze permanent gases and light hydrocarbons, while a Zebron ZB-WAXplus capillary column (Phenomenex) was used for alcoholic liquids. An ion chromatograph (Eco IC, Metrohm) equipped with a “METROHM 6.1005.200” column was used for quantifying volatile fatty acids (VFA) including formic acid. Inductively coupled plasma optical emission spectrometry (ICP-OES) for the determination of inorganic anion concentration was carried out on a PerkinElmer Optima 8000 ICP-OES instrument. The faradaic efficiency (FE) and half-cell energy efficiency (EE) for each product were calculated as reported by Zhu et al.³⁰

Analysis of the direct formate fuel cell (DFFC)

The fuel cell was based on a membrane electrode assembly (MEA) as previously reported, using nickel foam supported Pd/C–CeO₂ as the anode, a Tokuyama A201 anion exchange membrane, and carbon cloth coated with FeCo/C as the cathode.³¹ A model fuel consisting of 0.5 M HCOOK (Merck, Germany) in a 1.0 M KOH (Merck, Germany) solution was used as the control for comparison. The fuel solution was supplied to the anodic chamber at a flow rate of 1 ml min⁻¹ and a temperature of either 25 °C or 60 °C, while the oxygen flow on the cathodic side was held at 0.1 L min⁻¹. Fuel cell polarization curves were obtained at a scan rate of 10 mV s⁻¹ from the open circuit voltage (OCV) to a 0.5 V cell voltage and monitored using an ARBIN BT-2000 5A 4 channel instrument (Arbin Instruments, College Station, TX, USA). Galvanostatic experiments were undertaken at a constant current density of 25 mA cm⁻² while circulating 25 ml of the fuel solution. The cell potential was monitored, and the tests were stopped when the potential reached 0 V. The delivered energy for a single cell load, the faradaic efficiency for fuel utilization and energy efficiency for the complete oxidation of formate to carbonate were calculated according to a related study.³²

Results and discussion

Catalyst characterization

SEM was employed to characterize the morphology of the SnO₂/C catalyst with a mass ratio of SnO₂/C = 3.5. The fresh SnO₂/C catalyst (Fig. 1a) shows a uniform mixture of SnO₂ (∼100 nm irregular particles) and Vulcan XC-72 carbon black (∼50 nm spherical particles) and can clearly be distinguished from carbon black (Fig. S2†). Clusters (>300 nm) of irregular SnO₂ particles are also present. The catalyst morphology changed after eCO₂R as shown in Fig. 1b, with the size of Sn-based crystal now 200–500 nm. However, the shape and size of carbon particles remained unchanged.

Fig. 1 SEM images of the SnO₂/C(3.5) catalyst (a) before eCO₂R and (b) after 6 hours' eCO₂R at −1.43 V.

The XRD spectra in Fig. 2 show the crystal structure of the SnO₂/C(3.5) catalyst before and after eCO₂R, respectively, and that their compositions change from polycrystalline SnO₂ to polycrystalline Sn due to electrochemical reduction. This was confirmed by the CV measurements displayed in Fig. S3† showing that the reduction potential for SnO₂ to Sn was less negative than that for eCO₂R.¹⁹

Fig. 2 XRD spectra of the SnO₂/C(3.5) catalyst before eCO₂R (blue) and after 6 hours' eCO₂R at −1.43 V (orange).

eCO₂R activities

Effect of the carbon support. The eCO₂R catalytic performances of SnO₂/C mixtures (with mass ratios of 0.5, 1, and 3.5 in 4 mg cm⁻² catalyst) and individual SnO₂ and C were examined in a wide range of potentials from −0.63 to −1.43 V (vs. RHE) with a flowing catholyte. The normalized faradaic efficiency (FE) of all the products (H₂, CO, and formate) and overall current density (j) are displayed in Table S2† and summarized in Fig. 3. The partial current densities for H₂, CO, and formate as a function of the applied potential for each catalyst are presented in Fig. S4.†

Fig. 3 Current densities of eCO₂R catalysed by SnO₂/C mixtures and individual SnO₂ and C at a wide range of applied potentials.

As shown in Fig. 3, carbon black and SnO₂ both show catalytic activity for eCO₂R, producing H₂, CO, and formate as the major products. It is interesting to see that carbon black itself was active for eCO₂R with 50% FE for the C1 product (40% formate and 10% CO) at a low overpotential of −0.63 V, but the FE for C1 decreased at higher overpotentials. At the same applied potential, SnO₂ achieved much higher current densities and formate FE than carbon alone suggesting that SnO₂ possesses higher activity and faster kinetics. SnO₂ showed a similar trend to carbon in terms of the FE for C1 products, decreasing as the overpotential increases. The SnO₂/C composites show an enhancement in both formate FE and current density compared to carbon alone at any applied potential. However, in contrast to SnO₂ alone, the SnO₂/C composites generally need a high overpotential (i.e., an applied potential more negative than −0.87 V) to achieve higher formate FEs. Only SnO₂/C(3.5) shows a higher formate FE and current density than SnO₂ alone at any potential.

Sn and SnO₂ are known for their catalytic activity for formate production by eCO₂R,^22,23,27 which is also reflected in the results of this study. However, the formate FE showed a decreasing trend with the increasing overpotential when using SnO₂ alone, whereas it was maintained or even increased with the increasing overpotential when using SnO₂/C composites. One reason could be that the improved electrical conductivity of carbon facilitated the electron/charge transfer of the catalyst. For the catalytically optimum SnO₂/C(3.5), the molar ratio of SnO₂ : C on the reaction interface should be 1 : 3.6 (mol mol⁻¹). The carbon composition could provide efficient bridges for charge transfer between SnO₂ particles.

Another function of the carbon support could be adapting the SnO₂ catalyst pore structure to build a more effective reaction interface. Generally, inside the catalyst layer of a GDE, there are three pore conditions: flooded pores, wetted pores, and dry pores as illustrated by Weng et al.³³ and reproduced in Fig. S5.† The dry pores are gas transport channels, but they are inactive for eCO₂R due to the lack of an aqueous electrolyte and ion transfer. The flooded pores completely eliminate gas channels so they possess high mass-transport resistance for the gaseous CO₂ reactant. Thus, an ideal gas–liquid–solid three-phase interface that maximizes gas, liquid, and charge transfer was assumed to be facilitated by partially wetted pores.

Measuring the double-layer capacitances of the porous electrode in aqueous solution is normally used to evaluate the electrode–electrolyte phase boundary and to investigate the pore conditions.^34,35 The double-layer capacitances of 5 electrodes were measured in HClO₄ solution by a reported method;³⁶ smooth stainless-steel (SS) foil was also tested as a reference with the relative area of the electrode/electrolyte interface being 1, as presented in Fig. S6† and Table 1.

Table 1 Double layer capacitances and relative area of the electrode/electrolyte interface of SnO₂/C-GDEs with various SnO₂/C mass ratios, C-GDE, and SnO₂-GDE in contrast with smooth stainless-steel (SS) foil

Sample	SS foil	C	SnO2/C (0.5)	SnO2/C (1.0)	SnO2/C (3.5)	SnO2
Molar ratio of SnO2/C (mol/mol)	—	—	0.04	0.08	0.28	—
Double-layer capacitance (mF)	0.008	0.164	0.864	4.568	7.747	5.711
Relative area of the electrode/electrolyte interface	1	20.7	109.3	578.2	980.6	722.9

---

As shown in Table 1, the capacitance of carbon alone is 0.164 mF, nearly as small as that of the smooth SS foil (0.008 mF), although Vulcan carbon black is well-known for its large specific surface area > 200 m² g⁻¹.^37,38 This indicates that hydrophobic carbon alone results in a mostly “dry pores” situation. Ahn et al.³⁹ measured the contact angle of Nafion-bonded Vulcan XC72 to water to be ca. 150°, implying strong hydrophobicity. SnO₂ alone shows a much larger double-layer capacitance of 5.711 mF, indicating good electrolyte intrusion and a large active area at the electrode/electrolyte interface, whereas the contact angle of bare SnO₂ to water was 98°,⁴⁰ making it relatively more hydrophilic than carbon. When mixing SnO₂ and carbon in different ratios, the electrode capacitances were different. For SnO₂/C(0.5) with the smallest SnO₂/C molar ratio of 0.04 (mol/mol), the electrode/electrolyte interface area increased about 5-fold compared to that of carbon alone, showing the effect of the more hydrophilic SnO₂ in the SnO₂/C composite. With the increasing SnO₂ molar ratio, the electrolyte intrusion in the porous electrode increased. SnO₂/C(3.5) exhibited the largest electrode/electrolyte interface area, more than that of SnO₂ alone, and also shows optimum eCO₂R performance as shown in Fig. 3. This is probably because SnO₂/C(3.5) mostly contains partially wetted pores that favor eCO₂R. Moreover, since the size of Sn-based grains grew during eCO₂R as indicated in the SEM result (Fig. 1), an appropriate amount of carbon with the morphology unaffected by electrolysis should be beneficial to maintain the pore structure in SnO₂/C composites.

Since the overall current density almost linearly increased with the overpotential and the formate FE was maintained in the selected potential range by using SnO₂/C composites (especially the SnO₂/C(3.5)), the partial current density of formate production approximately followed a linear relationship with the applied potential as indicated in Fig. S4,† which means that formate production could be feasibly controlled by the energy input.

Controllable formate production. Fig. 4a shows the relationship of the overall CO₂ conversion rate and the formate production rate vs. applied potential when SnO₂/C(3.5) was used as the catalyst. The formate yield and CO₂ conversion increased approximately linearly with the increased overpotential, reaching about 3.03 mg min⁻¹ cm_WE⁻² and 23% respectively at −1.43 V. The formate yield per millilitre of CO₂ on each 1 cm² working electrode (WE) was calculated and compared to related studies published in recent 5 years (Fig. 4b). This work shows at least 10-fold improvement compared to the results in the literature.^{12–15,17,23,27}

Fig. 4 (a) The yield of formate and overall CO₂ conversion rate of eCO₂R catalysed by SnO₂/C(3.5) at specific applied potentials. (b) A comparison with related studies published in recent 5 years in terms of formate yield per millilitre of CO₂ feedstock on every 1 cm² working electrode (WE). The triangular symbols represent GDE-based results, while the round symbols represent the studies without using a GDE.

To produce formate with higher concentration, a batch reactor with a static catholyte was used. Fig. 5 shows the average FE and energy efficiency (EE) of formate production using the optimum SnO₂/C ratio (3.5) and operating the cell for one hour at various specific potentials. The formate FE at any potential is maintained around 73% which is lower than the results (∼80%) obtained in catholyte flow mode in Fig. 3, most probably due to the accumulated formate products affecting the reaction equilibrium and hindering the forward reaction. With the increasing overpotential, formate concentration in 1 hour electrolysis increased with decreasing EE. A concentration of 0.5 M formate could be produced at −1.43 V within 1 hour electrolysis with about 30% energy efficiency. The EE could be increased to higher than 50% when applying a lower overpotential of −0.63 V to produce a 0.5 M formate solution, but a much longer electrolysis time (9 h) would be needed as suggested in Fig. S7.†

Fig. 5 Formate concentration, average faradaic efficiency (FE) and energy efficiency (EE) with 1 hour electrolysis at specific applied potentials of −0.63 to −1.43 V.

Performance of a direct formate fuel cell (DFFC) using formate from eCO₂R

The 0.5 M formate solution produced by eCO₂R, denoted as “renewable fuel”, was then filtered using a PTFE membrane with a 0.2 μm pore size and directly utilized as the fuel for a DFFC. A model fuel of 0.5 M HCOOK dissolved in 1 M KOH prepared using commercial reagents was also tested in the DFFC for comparison. Table 2 summarizes the cation and anion concentrations of the fuel determined by IC and ICP measurement.

Table 2 Fuel composition

	Ion concentration (mol L^−1)
	COOH^−	K^+	CO3^2−/HCO3^−	Sn^4+/Sn^2+	pH
Model fuel	0.50	1.50	—	—	13.8
Renewable fuel	0.50	2.82	0.62	1.66 × 10^−5	>14

---

The potentiodynamic (VI curves) results of these two fuels at 25 °C and 60 °C are given in Fig. 6a. Power densities were calculated based on the VI curves and plotted in Fig. 6b. Galvanostatic experiments for both fuels were carried out at 25 mA cm⁻² by circulating 25 ml fuel (60 °C), with the results shown in Fig. 6c. Fig. 6d shows the discharge energy, by integrating the galvanostatic curves over time until the cell voltage reaches 0 V. The parameters of DFFC performance by using these two fuels are summarized in Table 3.

Fig. 6 The performances of the DFFC using the model fuel and renewable fuel: (a) polarization curves at 25 °C and 60 °C, (b) power density curves at 25 °C and 60 °C, (c) galvanostatic curves at 25 mA cm⁻² (60 °C) by circulating 25 ml fuel, and (d) discharge energy curves at 25 mA cm⁻² (60 °C) by circulating 25 ml fuel.

Table 3 Parameters of DFFC performance using the model fuel and renewable fuel at 60 °C

	Model fuel	Renewable fuel
OCV (V)	0.98	0.99
Cell resistance (mΩ)	72	84
Peak power density (mW cm^−2)	118	92
Discharge current density (mA cm^−2)	25	25
Discharge energy (kJ)	1.15 (25 ml)	0.95 (25 ml)
Faradaic efficiency (%)	69	84

---

As shown in Table 3, the OCV of the DFFC using the renewable fuel reached almost 1 V and was relatively close to the theoretical potential 1.356 V of the formate fuel cell operated at 60 °C with a formate concentration of 0.5 M and pH = 14. This theoretical potential was calculated using the Gibbs–Helmholtz equation and Nernst equation as shown in Table S3.† The renewable fuel achieved a peak power density, discharge energy, and energy efficiency of 92 mW cm⁻², 0.95 kJ, and 30%, respectively. The energy efficiency of the cell is highly dependent on the current density discharge value. Discharging the cell at a very low current density will keep the cell potential high and the energy efficiency will be higher, but with a low power density. It is more practical to run the cell at a more useful and realistic current density, i.e., 25 mA cm⁻² as chosen here, which can be applied to charge a mobile phone battery⁴¹ within a reasonable amount of time.

The model fuel shows a slightly higher peak power density, discharge energy, and energy efficiency than the renewable fuel. From the composition analysis in Table 2, it can be seen that even though the formate concentrations in both fuels are the same, the renewable fuel from CO₂ reduction contained a high concentration of carbonate (0.62 M) and a small quantity of Sn (∼2 ppm) due to the reaction with dissolved CO₂ and etching of the Sn catalyst during eCO₂R. The existence of carbonate ions can shift the equilibrium of the DFFC anodic reaction (eqn (2)) and hinder the forward reaction to some extent. From the view of catalysis, the adsorption of COOH_ads is the rate determining step (RDS) for electrooxidation of formate over Pd (eqn (3)).⁴² Carbonate/bicarbonate groups are competitive adsorbates against COOH_ads, the chemisorption of which on the Pd surface was reported to be potential-independent,⁴³ and can be strengthened by high coverage of carbonate/bicarbonate ions. Thus, the lower cell voltage when using the renewable fuel should most probably be caused by the large concentration of CO₃²⁻/HCO₃⁻.

COOH⁻ + 3OH⁻ → CO₃²⁻ + 2H₂O + 2e⁻ (2)

However, the faradaic efficiency of the renewable fuel reached 84% which is 15% higher than that of the model fuel. This is probably attributed to the higher alkalinity of the renewable fuel, as the formate fuel with higher alkalinity enabled higher faradaic efficiency of the DFFC.^31,32 Although it was not possible to measure the pH value of the renewable fuel as it was beyond 14, the concentration of hydroxyl ions could be roughly calculated as presented in Table 4. The calculated concentration of OH⁻ is 1.22 M, showing higher alkalinity than that of the model fuel.

Table 4 Calculation of OH⁻ concentration in the residual catholyte after eCO₂R producing 0.5 M formate^a

	Reactions	Calculation	C OH ^− (M)
a C: molar concentration (M or mol L^−1), t: time (s), n: molar mass (mol), F: faradaic constant (96 485 C mol^−1), Q: quantity of electric charge (C), V: volume (L), I: current (A).
Original OH^−	—	—	1.0
Produced OH^−	CO2 + 2H2O + 2e^− → HCOOH + 2OH^− (4)		1.46
	CO2 + H2O + 2e^− → CO↑ + 2OH^− (5)
	2H2O + 2e^− → H2↑ + 2OH^− (6)
Consumed OH^−	CO2 + 2OH^− → CO3^2− + H2O (7)		−1.24
Net.			1.22

---

When circulating 25 ml renewable fuel to achieve a more adequate fuel utilization, the overall chemical to electrical energy efficiency of the single cell reached 30%, which is comparable to those using higher concentrations of the formate fuel (C_COOH⁻ ≥ 1 M) in stronger alkaline media (C_OH⁻ ≥ 2 M).^31,32 Using the DFFC, a closed loop of “electricity–formate–electricity” has been achieved. With a high overpotential of −1.27 V, the overall EE was 30% leading to an efficiency of 9% for “electricity–formate–electricity”. However, it proves the concept of converting CO₂ to fuels as energy storage materials for renewable energy. Further overall efficiency can be improved by using lower overpotentials (shown in Fig. 5) and enhancing the anodic oxygen evolution reaction (OER), which is currently not optimised.

Conclusion

This work utilized a carbon black supported SnO₂ catalyst to efficiently convert CO₂ to formate in a gas diffusion cell with an alkaline catholyte (1 M KOH) at high current densities. An optimum SnO₂/C mass ratio was obtained suggesting that the interaction of SnO₂ and carbon at this ratio built an appropriate three-phase reaction interface with efficient charge and mass transfer leading to higher selectivity of formate production. The faradaic efficiency of formate under the optimized conditions could be maintained at around 80% in the potential range of −0.63 to −1.43 V (vs. RHE) and with overall current densities from 30–250 mA cm⁻². This enabled controllable formate production by tuning the applied potential; the yield of formate reached as high as 3 mg min⁻¹ cm_WE⁻² when applying −1.43 V cathodic potential. 0.5 M formate which was produced within 1 hour with an energy efficiency of 30% was directly used as the fuel in a direct formate fuel cell (DFFC). The fuel cell with a Pd–CeO₂/C anode and FeCo/C cathode achieved a peak power density of 92 mW cm⁻² and an energy efficiency of 30%. A closed loop of “electricity–formate–electricity”, which also closed the carbon recycling loop, was achieved signifying the promising future for electrochemical CO₂ reduction to fuels for energy storage.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge the EPSRC LifesCO₂R project (EP/N009746/1), EPSRC NECEM (EP/R021503/1) and NERC MeteoRR (NE/L014246/1). Hang Xiang acknowledges the Doctoral Training Awards (SAgE DTA, 2015 cohort) from the Faculty of Science, Agriculture and Engineering, Newcastle University for supporting the PhD study. The authors also acknowledge Ente Cassa di Risparmio di Firenze for funding (project EnergyLab) and the PRIN 2017 Project funded by the Italian Ministry MUIR Italy (No 2017YH9MRK). Dr Da Li is also appreciated for engaging in scientific discussion of this work. Data supporting this publication is openly available under an ‘Open Data Commons Open Database License’. Additional metadata are available at: https://doi.org/10.25405/data.ncl.9907877.v1.

References

1. S. Kaneco, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical reduction of CO₂ to methane at the Cu electrode in methanol with sodium supporting salts and its comparison with other alkaline salts, Energy Fuels, 2006, 20(1), 409–414.
2. C.-T. Dinh, T. Burdyny, M. G. Kibria, A. Seifitokaldani, C. M. Gabardo, F. P. G. de Arquer, A. Kiani, J. P. Edwards, P. De Luna and O. S. Bushuyev, CO₂ electroreduction to ethylene via hydroxide-mediated copper catalysis at an abrupt interface, Science, 2018, 360(6390), 783–787.
3. S. Rasul, D. H. Anjum, A. Jedidi, Y. Minenkov, L. Cavallo and K. Takanabe, A highly selective copper–indium bimetallic electrocatalyst for the electrochemical reduction of aqueous CO₂ to CO, Angew. Chem., 2015, 127(7), 2174–2178.
4. A. K. Singh, S. Singh and A. Kumar, Hydrogen energy future with formic acid: a renewable chemical hydrogen storage system, Catal. Sci. Technol., 2016, 6(1), 12–40.
5. S. Rasul, A. Pugnant, H. Xiang, J.-M. Fontmorin and H. Y. Eileen, Low cost and efficient alloy electrocatalysts for CO₂ reduction to formate, J. CO₂ Util., 2019, 32, 1–10.
6. S. Rasul, A. Pugnant and E. Yu, Electrochemical Reduction of CO₂ at Multi-Metallic Interfaces, ECS Trans., 2018, 85(10), 57–66.
7. J. Albo and A. Irabien, Cu₂O-loaded gas diffusion electrodes for the continuous electrochemical reduction of CO₂ to methanol, J. Catal., 2016, 343, 232–239.
8. N. Theaker, J. M. Strain, B. Kumar, J. P. Brian, S. Kumari and J. M. Spurgeon, Heterogeneously catalyzed two-step cascade electrochemical reduction of CO₂ to ethanol, Electrochim. Acta, 2018, 274, 1–8.
9. S. Fukuzumi, Production of liquid solar fuels and their use in fuel cells, Joule, 2017, 1(4), 689–738.
10. T. Vo, K. Purohit, C. Nguyen, B. Biggs, S. Mayoral and J. L. Haan, Formate: an energy storage and transport bridge between carbon dioxide and a formate fuel cell in a single device, ChemSusChem, 2015, 8(22), 3853–3858.
11. L. An and R. Chen, Direct formate fuel cells: a review, J. Power Sources, 2016, 320, 127–139.
12. Q. Wang, H. Dong and H. Yu, Development of rolling tin gas diffusion electrode for carbon dioxide electrochemical reduction to produce formate in aqueous electrolyte, J. Power Sources, 2014, 271, 278–284.
13. J. Wu, F. G. Risalvato, S. Ma and X.-D. Zhou, Electrochemical reduction of carbon dioxide III. The role of oxide layer thickness on the performance of Sn electrode in a full electrochemical cell, J. Mater. Chem. A, 2014, 2(6), 1647–1651.
14. X. Lu, D. Y. Leung, H. Wang and J. Xuan, A high performance dual electrolyte microfluidic reactor for the utilization of CO₂, Appl. Energy, 2017, 194, 549–559.
15. E. Irtem, T. Andreu, A. Parra, M. Hernández-Alonso, S. García-Rodríguez, J. Riesco-García, G. Penelas-Pérez and J. Morante, Low-energy formate production from CO₂ electroreduction using electrodeposited tin on GDE, J. Mater. Chem. A, 2016, 4(35), 13582–13588.
16. Z. Bitar, A. Fecant, E. Trela-Baudot, S. Chardon-Noblat and D. Pasquier, Electrocatalytic reduction of carbon dioxide on indium coated gas diffusion electrodes—Comparison with indium foil, Appl. Catal., B, 2016, 189, 172–180.
17. J. Chung, J. Koh, E.-H. Kim and S. I. Woo, Hierarchical Cu pillar electrodes for electrochemical CO₂ reduction to formic acid with low overpotential, Phys. Chem. Chem. Phys., 2016, 18(8), 6252–6258.
18. Y. Li, J. Qiao, X. Zhang, T. Lei, A. Girma, Y. Liu and J. Zhang, Rational design and synthesis of SnO_x electrocatalysts with coralline structure for highly improved aqueous CO₂ reduction to formate, ChemElectroChem, 2016, 3(10), 1618–1628.
19. Y. Fu, Y. Li, X. Zhang, Y. Liu, J. Qiao, J. Zhang and D. P. Wilkinson, Novel hierarchical SnO₂ microsphere catalyst coated on gas diffusion electrode for enhancing energy efficiency of CO₂ reduction to formate fuel, Appl. Energy, 2016, 175, 536–544.
20. T. Lei, X. Zhang, J. Jung, Y. Cai, X. Hou, Q. Zhang and J. Qiao, Continuous electroreduction of carbon dioxide to formate on Tin nanoelectrode using alkaline membrane cell configuration in aqueous medium, Catal. Today, 2018, 318, 32–38.
21. Q. Zhang, Y. Zhang, J. Mao, J. Liu, Y. Zhou, D. Guay and J. Qiao, Electrochemical Reduction of CO₂ by SnO_x Nanosheets Anchored on Multiwalled Carbon Nanotubes with Tunable Functional Groups, ChemSusChem, 2019, 12(7), 1443–1450.
22. F. Li, L. Chen, G. P. Knowles, D. R. MacFarlane and J. Zhang, Hierarchical mesoporous SnO₂ nanosheets on carbon cloth: a robust and flexible electrocatalyst for CO₂ reduction with high efficiency and selectivity, Angew. Chem., Int. Ed., 2017, 56(2), 505–509.
23. J. Yu, H. Liu, S. Song, Y. Wang and P. Tsiakaras, Electrochemical reduction of carbon dioxide at nanostructured SnO₂/carbon aerogels: The effect of tin oxide content on the catalytic activity and formate selectivity, Appl. Catal., A, 2017, 545, 159–166.
24. S. Y. Choi, S. K. Jeong, H. J. Kim, I.-H. Baek and K. T. Park, Electrochemical reduction of carbon dioxide to formate on tin–lead alloys, ACS Sustainable Chem. Eng., 2016, 4(3), 1311–1318.
25. S. Guo, S. Zhao, J. Gao, C. Zhu, X. Wu, Y. Fu, H. Huang, Y. Liu and Z. Kang, Cu-CDots nanocorals as electrocatalyst for highly efficient CO₂ reduction to formate, Nanoscale, 2017, 9(1), 298–304.
26. R. Kortlever, I. Peters, S. Koper and M. T. Koper, Electrochemical CO₂ reduction to formic acid at low overpotential and with high faradaic efficiency on carbon-supported bimetallic Pd–Pt nanoparticles, ACS Catal., 2015, 5(7), 3916–3923.
27. J. T. Feaster, C. Shi, E. R. Cave, T. Hatsukade, D. N. Abram, K. P. Kuhl, C. Hahn, J. K. Nørskov and T. F. Jaramillo, Understanding selectivity for the electrochemical reduction of carbon dioxide to formic acid and carbon monoxide on metal electrodes, ACS Catal., 2017, 7(7), 4822–4827.
28. W. Lee, Y. E. Kim, M. H. Youn, S. K. Jeong and K. T. Park, Catholyte-Free Electrocatalytic CO₂ Reduction to Formate, Angew. Chem., Int. Ed., 2018, 57(23), 6883–6887.
29. H. Xiang, S. Rasul, K. Scott, J. Portoles, P. Cumpson and H. Y. Eileen, Enhanced selectivity of carbonaceous products from electrochemical reduction of CO₂ in aqueous media, J. CO₂ Util., 2019, 30, 214–221.
30. D. D. Zhu, J. L. Liu and S. Z. Qiao, Recent advances in inorganic heterogeneous electrocatalysts for reduction of carbon dioxide, Adv. Mater., 2016, 28(18), 3423–3452.
31. H. Miller, J. Ruggeri, A. Marchionni, M. Bellini, M. Pagliaro, C. Bartoli, A. Pucci, E. Passaglia and F. Vizza, Improving the energy efficiency of direct formate fuel cells with a Pd/C-CeO₂ anode catalyst and anion exchange ionomer in the catalyst layer, Energies, 2018, 11(2), 369.
32. L. Wang, M. Bellini, J. Filippi, M. Folliero, A. Lavacchi, M. Innocenti, A. Marchionni, H. Miller and F. Vizza, Energy efficiency of platinum-free alkaline direct formate fuel cells, Appl. Energy, 2016, 175, 479–487.
33. L.-C. Weng, A. T. Bell and A. Z. Weber, Modeling gas-diffusion electrodes for CO₂ reduction, Phys. Chem. Chem. Phys., 2018, 20(25), 16973–16984.
34. A. Kamat, A. Huth, O. Klein and S. Scholl, Chronoamperometric investigations of the electrode–electrolyte interface of a commercial high temperature PEM fuel cell, Fuel Cells, 2010, 10(6), 983–992.
35. P. K. Sow, Z. Lu, H. Talebian, L. Damron and W. Mérida, Double layer capacitance measurements to characterize the water intrusion into porous materials, J. Phys. Chem. C, 2016, 120(43), 24794–24802.
36. C. W. Li and M. W. Kanan, CO₂ reduction at low overpotential on Cu electrodes resulting from the reduction of thick Cu₂O films, J. Am. Chem. Soc., 2012, 134(17), 7231–7234.
37. M. Lázaro, L. Calvillo, V. Celorrio, J. Pardo, S. Perathoner and R. Moliner, Study and application of carbon black Vulcan XC-72R in polymeric electrolyte fuel cells, Carbon black: production, properties and uses, 2011, p. 41.
38. M. Bevilacqua, C. Bianchini, A. Marchionni, J. Filippi, A. Lavacchi, H. Miller, W. Oberhauser, F. Vizza, G. Granozzi and L. Artiglia, Improvement in the efficiency of an OrganoMetallic Fuel Cell by tuning the molecular architecture of the anode electrocatalyst and the nature of the carbon support, Energy Environ. Sci., 2012, 5(9), 8608–8620.
39. M. Ahn, Y.-H. Cho, Y.-H. Cho, J. Kim, N. Jung and Y.-E. Sung, Influence of hydrophilicity in micro-porous layer for polymer electrolyte membrane fuel cells, Electrochim. Acta, 2011, 56(5), 2450–2457.
40. A. Kumar, A. Sanger, A. Kumar and R. Chandra, Highly sensitive and selective CO gas sensor based on a hydrophobic SnO₂/CuO bilayer, RSC Adv., 2016, 6(52), 47178–47184.
41. H. A. Attia, B. N. Getu, H. Ghadban and A. K. A. Mustafa, Portable solar charger with controlled charging current for mobile phone devices, Int. J. Energy Environ. Eng., 2014, 7(1), 17–24.
42. T. Gunji and F. Matsumoto, Electrocatalytic Activities towards the Electrochemical Oxidation of Formic Acid and Oxygen Reduction Reactions over Bimetallic, Trimetallic and Core–Shell-Structured Pd-Based Materials, Inorganics, 2019, 7(3), 36.
43. A. Berná, A. Rodes, J. M. Feliu, F. Illas, A. Gil, A. Clotet and J. M. Ricart, Structural and spectroelectrochemical study of carbonate and bicarbonate adsorbed on Pt (111) and Pd/Pt (111) electrodes, J. Phys. Chem. B, 2004, 108(46), 17928–17939.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9se00625g

---