An enhanced electrochemical CO₂ reduction reaction on the SnO_x–PdO surface of SnPd nanoparticles decorated on N-doped carbon fibers

Abstract

In the electrocatalytic CO₂ reduction reaction (CO₂RR), tin-based catalysts are known for their high formate faradaic yield. However higher overpotentials are required to attain a high faradaic yield with high partial current density for formate. Here, we describe the increase in the electrocatalytic CO₂RR activity of Sn nanoparticles decorated on nitrogen-doped carbon fibers (NCFs) by adding a small amount of Pd. Nitrogen-doped carbon fibers decorated with SnPd nanoparticles (Sn_100−yPd_y–NCF) of different Sn : Pd ratios were synthesized using the electrospinning method and their electrocatalytic CO₂RR activity was studied. The Sn_100−yPd_y–NCF catalyst with 3 wt% (y = 3) Pd displayed superior activity for the CO₂RR and attained a faradaic efficiency of 85%, whereas the NCF with Sn nanoparticles (Sn₁₀₀–NCF) attained only 57% efficiency at the same potential. The surface electronic configuration, Tafel slope (79 mV dec⁻¹) and bicarbonate reduction activity of the catalyst reveal that the combination of SnO_x–PdO on the catalyst surface is responsible for the superior CO₂RR activity.

---

1. Introduction

Recent analyses have confirmed that the atmospheric CO₂ concentration reached 412 ppm, and its contribution to global warming is alarming.^1–4 In this context, reducing the global CO₂ emissions and converting CO₂ to useful products are the best solutions to control the CO₂-related climate change issues.^5–8 The electrochemical conversion of CO₂ to liquid fuels, such as formic acidt/formate, methanol, and ethanol, or chemical feedstocks such as syngas, CO + H₂ (ref. 9–15), is the most promising method for the effective utilization of CO₂. Recently, extensive investigations have been conducted across the world for the effective electrochemical reduction of CO₂. As a result, the electrocatalytic CO₂RR activity of a large number of materials was tested, which include metals,^16–18 metal nanoparticles,^19–23 bimetallic nanoparticles, metal nanoalloys,^24–27 and many types of carbon materials.^28–32 These materials have different capabilities and product selectivity for the CO₂RR. The product selectivity and efficiency vary with metals,¹⁸ the size of the metal particles,³³ the roughness of the material surface,^34,35 the dopant in the carbon structure^28–30 and even the cations present in the electrolytes.³⁶ Therefore, tailor-made materials are important for the selective, efficient, and stable conversion of CO₂ to a product.

A variety of products have been produced through the electrochemical CO₂RR;³⁷ however, techno-economic studies claim that formic acid/formate from CO₂ is one of the most economically viable products.³⁸ In the renewable energy industries, formic acid with a volumetric capacity of 53 g L⁻¹ of hydrogen³⁹ and liquid formate-based fuel cells (DFFCs), which have a higher theoretical cell voltage (1.45 V) and low poisoning effect,⁴⁰ confirm the importance of formic acid/formate as an energy carrier. Recently, researchers have focused on the development of highly efficient electrocatalysts and membranes for the formate oxidation reaction, which may facilitate the commercialization of direct formate fuel cells.^41,42 Even though DFFCs are considered as a source of green energy, the industrial production of formic acid/formate by the hydrolysis of methyl formate or formamide is not an eco-friendly method.^43,44 In this context, the efficient electrochemical reduction of CO₂ to formate using renewable energy sources is a promising method for utilizing CO₂ and producing clean and sustainable fuels.

Among the metal catalysts, Sn-based catalysts have an efficient electrochemical CO₂ reduction activity. Tin has a high hydrogen overpotential and excellent selectivity for CO₂ to formate conversion;^19,45–50 tin is also a non-toxic, inexpensive, and eco-friendly element. Furthermore, the native oxide layer of the Sn surface increases the catalytic activity by stabilizing the CO₂ radical (CO₂˙⁻).⁵⁰ Although tin displays high selectivity and efficiency for the CO₂ reduction reaction (CO₂RR), the requirement of high overpotential is the main drawback. A variety of modifications have been introduced to reduce the overpotential and improve the catalytic activity of Sn,^19,45–49 but the ability to achieve high faradaic efficiency and high partial current density remains a substantial challenge. In the present study, we prepared Sn nanoparticles with a low amount of Pd that were supported on N-doped carbon fibers (NCFs) using a simple electrospinning method and investigated the electrocatalytic CO₂ reduction activity of these materials. Compared to the usual carbon support for nanoparticles, the N-doped carbon surface is more suitable since pyridinic nitrogen in N-doped CNFs exhibits catalytic CO₂ reduction reaction activity.^28,29 We chose palladium as an additive to tin to increase the CO₂RR activity because the Pd surface is capable of stabilizing and reducing bicarbonate ions to formate at a near thermodynamic potential.^51,52 Since the bicarbonate ions are the proton source for the CO₂ reduction reaction in bicarbonate solution,^19,21 the catalyst surface stabilized bicarbonate ions may enhance the quick protonation of CO₂˙⁻. Based on these aspects, we prepared Sn_100−yPd_y–NCF catalysts with different amounts of Pd loading. As a result of this rational design of the electrocatalyst, we could achieve a maximum faradaic efficiency of 85% and a partial current density value of 12.75 mA cm⁻² during the electrochemical reduction of CO₂ on the Sn_100−yPd_y–NCF catalyst containing 3 wt% (y = 3) Pd. The material characterization and analysis of the electrochemical activity of these catalysts confirm that the combination of SnO_x–PdO (where SnO_x is the sum of the SnO and SnO₂) on the catalyst surface is responsible for this superior activity.

2. Experimental

2.1. Chemicals

N-Doped carbon fibers with Sn and Pd nanoparticles were prepared using SnCl₂·2H₂O (Wako Chemicals), Pd (CH₃COO)₂ (Sigma-Aldrich), polyacrylonitrile (PAN, MW = 150 000) (Sigma-Aldrich), and DMF (Wako Chemicals) as precursors. Electrolytes were prepared using KHCO₃ (Wako Chemicals), H₂SO₄ (Wako Chemicals), potassium formate (Wako Chemicals), disodium hydrogen phosphate, and monosodium hydrogen formate (Wako Chemicals). DMSO (Sigma-Aldrich) and D₂O (Wako Chemicals) were used for NMR analyses. All solutions were prepared in Milli-Q water (18.2 MΩ).

2.2. Synthesis and characterization of metal nanoparticles decorated on N-doped carbon fibers

Tin nanoparticles decorated on N-doped carbon fibers (Sn₁₀₀–NCF) were prepared by an electrospinning method using a NANON-01A (MECC Co., Ltd.) electrospinning system. A solution containing 1.2 g of polyacrylonitrile (PAN), 1 g of SnCl₂·2H₂O and 12.8 g of DMF was electrospun to produce Sn₁₀₀–NCF. The following electrospinning conditions were applied, an applied voltage of 25 kV, a distance of 150 mm that was maintained between the needle and collector, and a precursor solution fed with a feed rate of 0.8 mL per hour. After electrospinning, fibers were dried at 130 °C for 12 hours in a vacuum atmosphere, and the dried fibers were heated at 280 °C for 90 minutes in an air atmosphere at a heating rate of 5 °C per minute to stabilize the PAN fibers. The stabilized composite PAN fibers were then heated at 900 °C for 120 minutes at a ramping rate of 3 °C per minutes under a N₂/H₂ (100 sccm N₂ and 30 sccm H₂) atmosphere. Various Sn_100−yPd_y–NCF catalyst samples with different amounts of Pd were prepared using the method described above. For Sn_100−yPd_y–NCF the solution for electrospinning was prepared by adding the required amount of metal precursors to PAN (1.2 g) in a DMF (9.8 g) solution. Metal precursor solutions were prepared by separately dissolving Pd (Ac)₂ (25, 50, 75, and 100 mg) and SnCl₂·2H₂O (1 g) in 1.5 mL of DMF, and then the two solutions were mixed and sonicated for 30 minutes.

The ICP analysis was performed using a Shimadzu sequential plasma emission spectrometer ICP2-8100 to quantify the total metal loading and the weight percentage distribution of Sn and Pd on the NCF (details are provided in the ESI,† Table S1). Phase identification and crystallinity of the catalyst samples were determined by performing XRD analysis at a scan rate of 1° per minute using an Ultima IV system from Rigaku with a Cu Kα (λ = 1.5406 Å) and X-ray source operating at 40 kV and 40 mA. The surface compositions of the fibers were analyzed using an XPS Quantum 2000, ULVAC-PHI Inc., with a twin-anode X-ray source using Al Kα radiation (hυ = 1486.58 eV). A FESEM (S-4800, Hitachi High Technology) operating at an accelerating voltage of 5 kV and a HR-TEM (TOPCON EM-002BF-J) operating at an accelerating voltage of 200 kV with a twin EDS facility were used to visualize the morphology and to understand the elemental distribution of nanoparticle-decorated fibers.

2.3. Electrochemical characterization

The electrochemical activity of the catalysts was studied using an electrochemical measurement system (Hokuto Denko, HZ5000). Linear sweep voltammetry (LSV), cyclic voltammetry (CV) and constant potential electrolysis were performed in a two-compartment cell separated by a Nafion-117 membrane. An Ag/AgCl (KCl-saturated) reference electrode was used for all measurements, and Pt foil (area: 18 cm²) served as the counter electrode. Catalyst ink was prepared by mixing the required amount of catalyst in a solution containing 5 mL of isopropanol (25 wt%) and 20 μL of the Nafion® solution (5 wt%). The catalyst ink was sonicated for 4 hours in an ice bath to obtain a good dispersion. After achieving a well-dispersed solution, the catalyst was coated on the glassy carbon electrode using the drop-casting method, and the catalyst loading was maintained at 50 μg of metal per cm² for all the samples. The catalyst-coated glassy carbon electrode (Hokuto Denko) with an electrode area of 0.196 cm² was used for voltammetric and impedance analyses. The catalyst-coated, 1 cm² Sigracet 28 BA gas diffusion layer (GDL) electrode was used for electrolysis and stability studies. For the CO₂RR, a 0.5 M KHCO₃ solution was first saturated with CO₂ by bubbling it with CO₂ gas for 1 h. During electrolysis, a constant flow of CO₂ gas was maintained at the cathode compartment. Similarly, the bicarbonate reduction reaction was performed by passing N₂ instead of CO₂ in a 0.5 M KHCO₃ solution. All applied potential values reported here are iR corrected and with respect to a reversible hydrogen electrode (RHE). The conversion of the applied potential measured using the Ag/AgCl electrode to the RHE was calculated using a method described in the ESI† (section S4).

2.4. Product analysis

Liquid product quantification was performed using a Bruker 400 MHz NMR spectrometer with DMSO as the internal standard. The details of the analysis and quantification of the liquid products using NMR are provided in the ESI† (section S6.1).

Gas products were quantified using a gas chromatograph (Shimadzu GC-2010 Tracera) equipped with Micropacked ST in series with a Shincarbon ST column and barrier discharge ionization detector (BID). The details of the analysis and quantification of the gas products using gas chromatography are provided in the ESI† (S6.2).

3. Results and discussion

3.1. Synthesis and characterization of the catalyst

In the stabilization process of the as-spun fibers at 280 °C in air, polyacrylonitrile undergoes crosslinking to form a stable fibrous structure.⁵³ Pyrolysis at 900 °C under a N₂/H₂ atmosphere after stabilizing the fibers with metal precursors produces metal nanoparticle-decorated graphitized N-doped carbon fibers. Based on the amount (wt%) of Sn and Pd present in the NCF, the catalyst samples are designated as Sn₁₀₀–NCF, Sn₉₇Pd₃–NCF, Sn₉₅Pd₅–NCF, Sn₉₄Pd₆–NCF, and Sn₉₁Pd₉–NCF, here the numbers represent the percentage of Sn and Pd in the total metal loaded on the NCF in all other cases.

The X-ray diffraction (XRD) patterns of all the samples are presented in Fig. 1a. The XRD pattern of the Sn₁₀₀–NCF sample indicated that the metallic tin was a major phase, along with a small amount of SnO₂. The diffraction patterns correspond to 2θ values of 30.67, 32.06, 43.92, and 44.98° represent the metallic tin (200), (101), (220) and (211) planes, respectively (JCPDS No: 83-7910), and 2θ values of 26.61, 33.89 and 51.79° correspond to the SnO₂ (110), (101), and (211) planes, respectively, (JCPDS No: 41-1445). The XRD patterns of Sn₉₇Pd₃–NCF and Sn₉₅Pd₅–NCF also showed the presence of Sn and SnO₂, similar to the Sn₁₀₀–NCF sample. In addition to Sn and SnO₂, the presence of PdO was also observed at 2θ values of 34.8 and 42.4°, representing the PdO (101) and (110) planes, respectively (JCPDS No 43-1024). For the catalysts with comparatively high Pd concentration, Sn₉₄Pd₆–NCF and Sn₉₁Pd₉–NCF, the XRD peaks correspond to SnO₂ and PdO were not observed; however, a slight shift in the diffraction peaks of Sn to higher 2θ values, 0.05 and 0.13 degrees, respectively, was noticed (Fig. 1b) compared to Sn₁₀₀–NCF, Sn₉₇Pd₃–NCF and Sn₉₅Pd₅–NCF, indicating the possible formation of an SnPd alloy.⁵⁴ Although the total amount of Pd present in Sn₉₇Pd₃–NCF and Sn₉₅Pd₅–NCF is low, the appearance of the PdO peak indicated that most of the Pd is located on the surface and exists as PdO.

Fig. 1 XRD patterns of (a) Sn₁₀₀–NCF and Sn_100−yPd_y–NCF catalysts, (b) highlighted, 2θ values between 30 and 31 degrees, and (c–e) the deconvoluted XPS spectra of N, Sn and Pd present in Sn₉₇Pd₃–NCF.

XPS analysis of the catalysts was performed to identify the surface composition (Fig. 1c–e, S1, and S2†). The deconvoluted high-resolution XPS N-1s spectra of all catalyst samples indicated the presence of graphitic (BE = 400.76 eV), pyrrolic (BE = 399.60 eV), and pyridinic (BE = 398.30 eV) nitrogen.³² The deconvoluted Sn-3d spectra of all catalysts showed the presence of Sn, SnO, and SnO₂. The binding energy values of Sn (3d_5/2 – 485.1 eV), SnO (3d_5/2 – 486.2 eV), and SnO₂ (3d_5/2 – 487.2 eV)⁵⁵ were approximately the same for all catalysts. The atomic percentages (at.%) of Sn, SnO, and SnO₂ were calculated from the corresponding peak area of the deconvoluted Sn-3d_5/2 spectrum (Table S2a†). For all catalysts, the amount (at.%) of SnO_x was greater than 95%, suggesting that the catalyst surface consists of an oxide layer of Sn. The deconvoluted Pd-3d spectrum was different for different Sn_100−yPd_y–NCF samples (Fig. 1e and S2†); the spectrum of Sn₉₇Pd₃–NCF contained only the PdO peak (3d_5/2 – 336.9 eV), but peaks for both PdO and Pd (3d_5/2 – 335.5 eV) were observed in the spectrum of Sn₉₅Pd₅–NCF. For Sn₉₄Pd₆–NCF and Sn₉₁Pd₉–NCF samples, both Pd and PdO peaks were slightly shifted to higher binding energy.⁵⁶ For Sn₉₄Pd₆–NCF,Pd and PdO peaks were observed at 3d_5/2 – 336.2 eV and 3d_5/2 – 337.3 eV, respectively, and at 3d_5/2 – 336.4 eV and 337.5 eV for Sn₉₁Pd₉–NCF, indicating a change in the binding energy values of Pd and PdO probably due to the formation of SnPd alloy.^57,58 The ratio of Pd/PdO of the Sn_100−yPd_y–NCF systems increased with the increase in the Pd concentration (Table S2b†). The XRD and XPS analyses revealed that the core of metal nanoparticles in Sn₁₀₀–NCF and Sn_100−yPd_y–NCF consisted of metallic Sn and the surface in the oxidized form. The surface of Sn₉₇Pd₃–NCF solely consists of SnO_x and PdO, whereas the surface of the Sn₉₅Pd₅–NCF system is mainly composed of SnO_x, PdO, and a low amount of Pd. For the Sn₉₄Pd₆–NCF and Sn₉₁Pd₉–NCF samples, the surface concentration of Pd is higher than the PdO concentration. In general, the XPS results corroborate the XRD results.

The morphology of the materials was analyzed using FE-SEM and HR-TEM. The average diameter of N-doped carbon fibers was 650 nm (Fig. 2a), and the nanoparticles were spherical with a size of less than 10 nm uniformly decorating the carbon fibers (Fig. 2b and c). Lattice fringes with a d-spacing of 0.29 nm (inset of Fig. 2d) correspond to the (200) plane of Sn (JCPDS No: 83-7910). The energy-dispersive X-ray spectroscopy (EDS) map revealed the uniform distribution of C, N, O, Sn, and Pd on the Sn_100−yPd_y–NCF surface (Fig. 2f–j and S3†).

Fig. 2 (a) General morphology of Sn₉₇Pd₃–NCF (SEM image). HR-TEM images of (b) a single fiber of Sn₉₇Pd₃–NCF and (c) SnPd nanoparticles dispersed on the carbon fiber (d) showing the lattice fringes corresponding to the Sn (200) plane. (e–i) STEM-EDS mapping images of C, N, O, Sn and Pd, respectively.

3.2. Electrochemical reduction of CO₂

The electrocatalytic CO₂ reduction activity of all materials was primarily evaluated using linear sweep voltammetry (LSV). Fig. 3a and S5† show the voltammograms of Sn₁₀₀–NCF, Sn_100−yPd_y–NCF, and NCF, recorded between −0.1 V and −1 V at a scan rate of 10 mV s⁻¹ in CO₂ saturated 0.5 M KHCO₃. Based on a previous report, the potential at which the catalyst attains a current density value of −0.15 mA cm⁻² is considered as the onset potential.²⁹ The onset potential values of the NCF, Sn₁₀₀–NCF, Sn₉₁Pd₉–NCF, Sn₉₄Pd₆–NCF, Sn₉₅Pd₅–NCF, and Sn₉₇Pd₃–NCF were −0.36, −0.21, −0.17, −0.16, −0.14 and −0.12 V, respectively. At −1.0 V the NCF alone attained a current density value of 0.77 mA cm⁻², whereas the Sn₁₀₀–NCF catalyst attained a current density value of 8.23 mA cm⁻² at the same potential. Sn₉₇Pd₃–NCF, Sn₉₅Pd₅–NCF, Sn₉₄Pd₆–NCF, and Sn₉₁Pd₉–NCF attained a current density value of 19.75, 18.88, 14.12, and 12.07 mA cm⁻², respectively, at −1.0 V. However, the current density alone does not ensure the CO₂RR activity of the catalysts. So the product analysis was performed after 30 minutes of electrolysis at different potentials and the faradaic efficiency of each product was calculated. The liquid and gaseous products of the CO₂RR were qualitatively and quantitatively analyzed using NMR and gas chromatography, respectively. For all the catalysts, formate was the initial CO₂RR product obtained. Formate production was observed for Sn_100−yPd_y–NCF samples at −0.37 V, whereas for Sn₁₀₀–NCF and the NCF, the potential was −0.57 V, which is 0.2 V higher than the Sn_100−yPd_y–NCF catalysts. N-Doped carbon can produce CO and formate from CO₂,^28,29 however in this work the NCF produced only formate. For all other catalysts, the products of the CO₂RR were formate and CO, of which formate was the major product. The faradaic efficiency (FE_CO2RR) and partial current density (j_CO2RR) of CO₂RR products of all the samples were measured and the results are presented in Fig. 3b and c, S8–S10† and Table 1. The NCF attained a FE_CO2RR of 56% and j_CO2RR of 0.41 mA cm⁻² at −0.97 V, whereas at the same potential Sn₁₀₀–NCF attained a faradaic efficiency of 65% for the CO₂RR at −0.97 V with j_CO2RR of 4.60 mA cm⁻², the formate faradaic efficiency (FE_formate) was 57% and the corresponding partial current density of formate (j_formate) was 4.03 mA cm⁻².

Fig. 3 Electrocatalytic CO₂ reduction in 0.5 M KHCO₃ saturated with CO₂ on Sn₁₀₀–NCF and Sn_100−yPd_y–NCF catalysts. (a) iR-Corrected linear sweep voltammogram measured at a scan rate of 10 mV s⁻¹, (b) formate faradaic efficiency, (c) CO faradaic efficiency and (d) Nyquist plots measured at −0.37 V and −0.87 V.

Table 1 The potential at which the maximum FE_CO2RR is obtained, and corresponding FE_CO2RR, FE_formate, FE_CO, FE_H2, j_CO2RR and j_formate, of the catalysts

Catalyst	E at maximum FECO2RR (V)	FECO2RR (%)	FEformate (%)	FECO (%)	FEH2 (%)	j CO2RR (mA cm^−2)	j formate (mA cm^−2)	j CO (mA cm^−2)	j H2 (mA cm^−2)
Sn91Pd9–NCF	−0.87	67	58	9	33	3.79	3.28	0.51	1.87
Sn94Pd6–NCF	−0.87	70	60	10	30	4.20	3.60	0.60	1.80
Sn95Pd5–NCF	−0.92	81	70	11	19	10.84	9.36	1.48	2.54
Sn97Pd3–NCF	−0.92	85	77	8	15	12.75	11.55	1.20	2.25
Sn100–NCF	−0.97	65	57	8	35	4.60	4.03	0.57	2.4
NCF	−0.97	56	56	0	44	0.41	0.41	—	0.32

---

The Sn_100−yPd_y–NCF catalysts have displayed an interesting trend in FE_CO2RR. Sn₉₄Pd₆–NCF and Sn₉₁Pd₉–NCF catalysts exhibited a high FE_CO2RR at the low overpotential region compared to Sn₉₇Pd₃–NCF and Sn₉₅Pd₅–NCF. However, at slightly higher overpotential, Sn₉₇Pd₃–NCF and Sn₉₅Pd₅–NCF attained very high FE_CO2RR and j_CO2RR values compared to the other two Sn_100−yPd_y–NCF catalysts. The faradaic efficiency values of formate and CO of Sn_100−yPd_y–NCF revealed that formate is the major CO₂RR product for all the catalysts. Among the Sn_100−yPd_y–NCF catalysts Sn₉₇Pd₃–NCF has attained high FE_formate and j_formate values, when evaluating the catalysts in terms of the faradaic efficiency of CO (FE_CO), the Sn₉₁Pd₉–NCF has attained high FE_CO, 30% at −0.77 V (Fig. 3c). Comparing the activity of efficient Sn_100−yPd_y–NCF catalysts with Sn₁₀₀–NCF in terms FE_formate, the highly efficient Sn₉₇Pd₃–NCF has a 27% higher FE_formate and 8 mA cm⁻² higher j_formate than Sn₁₀₀–NCF at −0.92 V (Fig. 3b and S9b†). Similarly, comparing the activity in terms of FE_CO, Sn₉₁Pd₉–NCF exhibited a 26% higher FE_CO and 0.96 mA cm⁻²j_CO than Sn₁₀₀–NCF at −0.77 V (Fig. 3c, S9c†). These results suggest that a small amount of Pd addition on Sn–NCF, either as Pd or PdO can enhance the CO₂RR activity. Based on the material characterization and faradaic efficiency of formate and CO results of the most efficient Sn_100−yPd_y–NCF catalysts we can say that SnO_x–Pd enhances CO production and SnO_x–PdO enhances formate production. The CO₂RR activity of pure Pd is well studied and it is reported that at low overpotentials Pd produces formate and CO and at higher overpotentials, Pd preferably produces CO,^59,60 but the activity of PdO is not well understood. We performed CO₂ reduction over the PdO surface (section S8, Fig. S11a†). During the initial stages of electrolysis, the FE_formate reached 64% but it quickly reduced and stabilized at 34%, the j_formate was <1 mA cm⁻². The XRD analysis at different stages of electrolysis confirmed that PdO reduced to Pd during the electrolysis (Fig. S11b†).

Since PdO was quickly reduced to Pd during the electrolysis, we tested the catalytic stability of Sn₉₇Pd₃–NCF (Fig. S12†) with periodic analysis of faradaic efficiency, followed by XPS and HR-TEM of the catalyst (Fig. S13†). Interestingly, a stable FE_CO2RR was retained for Sn₉₇Pd₃–NCF for up to 9 hours of electrolysis, and then a decrease in FE_CO2RR was observed. After 15 hours, the FE_CO2RR decreased to 70% from 85%. FE_formate and FE_CO were 68% and 2%, respectively. XPS analysis was performed after 9 and 15 hours of electrolysis. After 9 hours of electrolysis the PdO at.% was only reduced to 60 from 100%, the at.% of SnO_x remained almost the same, 96.4 from 96.7, (Fig. S13a and b†). The XPS analysis after 15 hours of electrolysis indicated that the at.% of Sn at the surface was increased from 3.26 to 18 and the at.% of SnO_x reduced to 82 from 96.74 (Fig. S13c†). A very low-intensity broad Pd-3d peak was obtained. The TEM image of the catalyst after 15 hours of electrolysis shows agglomeration of the particles on the surface (Fig. S13d†), whereas the particles inside the fibers have not changed. The stability test and followed material characterization confirms that the PdO reduction is slower in Sn₉₇Pd₃–NCF and the FE_CO2RR decreases with the reduction in PdO.

Impedance and Tafel slope analyses were performed to understand the charge transfer kinetics and the reaction mechanism. The electrochemical impedance was measured at a lower overpotential, −0.37 V, and a higher overpotential, −0.87 V in a CO₂-saturated KHCO₃ solution. At −0.37 V, Sn₉₁Pd₉–NCF and Sn₉₄Pd₆–NCF displayed a lower charge transfer resistance (R_CT) than Sn₉₇Pd₃–NCF and Sn₉₅Pd₅–NCF. The trend changed when the R_CT value was measured at −0.87 V, the Sn₉₇Pd₃–NCF and Sn₉₅Pd₅–NCF samples displayed a lower R_CT value than the Sn₉₁Pd₉–NCF and Sn₉₄Pd₆–NCF catalysts (Fig. 3d and Table S3†). Furthermore, we performed the EIS analysis before and after 30 minutes of electrolysis for the Sn₉₇Pd₃–NCF and Sn₉₁Pd₉–NCF catalysts (Fig. S14†). The R_CT values of Sn₉₁Pd₉–NCF increased with the time of electrolysis, whereas the change in R_CT values was not significant for Sn₉₇Pd₃–NCF. The increase in the R_CT values of Sn₉₁Pd₉–NCF with the electrolysis time is probably because of the adsorption of intermediates or products on the catalytic surface and hinders the electron transfer. To verify this possibility the Sn₉₁Pd₉–NCF electrode was kept at 0.5 V for 30 s after electrolysis and then EIS was performed; the R_CT was similar to the initial R_CT, suggesting the adsorption of an intermediate or reaction products on the catalyst surface. Since the Pd surface is susceptible to CO poisoning,^59–61 the increased R_CT values with respect to electrolysis time are possibly due to CO poisoning. The low or insignificant change in the R_CT values of Sn₉₇Pd₃–NCF after electrolysis revealed that the PdO surface is not susceptible to CO poisoning.⁶² The increase in R_CT values of Sn₉₁Pd₉–NCF and Sn₉₄Pd₆–NCF at high overpotentials is due to surface poisoning and these results corroborate their low current density and faradaic efficiency at higher overpotentials.

Tafel slopes (Fig. 4 and S15†) were plotted using the log of j_formate and overpotentials of each sample to understand the reaction mechanism. Tafel slope values for Sn₁₀₀–NCF, Sn₉₇Pd₃–NCF, Sn₉₅Pd₅–NCF, Sn₉₄Pd₆–NCF, and Sn₉₁Pd₉–NCF were 105, 79, 92, 105 and 117 mv dec⁻¹, respectively. The Tafel slope value of Sn₉₇Pd₃–NCF and Sn₉₅Pd₅–NCF was less than 120 mv dec⁻¹ and close to 59 mv dec⁻¹. The low Tafel slope value indicates that the first step of the reaction mechanism was the reversible one-electron transfer to CO₂ to form CO₂˙⁻ then the chemical rate-determining step, i.e. the protonation of C in CO₂˙⁻ to form formate.^19,48,49 Higher Tafel slope values closer to 120 mv dec⁻¹ suggest that the one-electron transfer to CO₂ to form CO₂˙⁻ is the rate-limiting step.^19,42,45 According to previous reports, the surface electronic structure plays a major role in the selectivity of the CO₂ reduction reaction by stabilizing CO₂˙⁻ and other intermediates.^21,33,34,48 Here, the low Tafel slope, high faradaic efficiency, and partial current density obtained for Sn₉₇Pd₃–NCF and Sn₉₅Pd₅–NCF confirm that the presence of SnO_x–PdO combination at the catalyst surface has an important role in their catalytic activity. A recent study by Zhang et al. on the electrochemical reduction of CO₂ using a Pd/SnO₂ system explained the role of the Pd–O–Sn combination in promoting the reaction through excellent CO₂ adsorption and decreasing CO poisoning on the Pd surface.⁶²

Fig. 4 Tafel slopes of (a) Sn₉₇Pd₃–NCF (b) Sn₉₅Pd₅–NCF (c) Sn₉₄Pd₆–NCF and (d) Sn₉₁Pd₉–NCF.

In the electrochemical CO₂ reduction reaction, bicarbonate ions in the electrolyte provide protons to protonate CO₂˙⁻ and produce formate.^19,21 A catalyst surface, which can stabilize the bicarbonate ions on its surface, can facilitate the quick protonation of CO₂˙⁻ to formate. We believe that the reason for the high current density and faradaic efficiency of the SnPd–NCF catalyst with the SnO_x–PdO surface is because the SnO_x–PdO surface effectively stabilizes CO₂˙⁻ and bicarbonate on its surface and protonates CO₂˙⁻ to produce formate. To ensure this possibility, we tested the bicarbonate reduction activity of all catalysts and analyzed the products. The NCF and Sn₁₀₀–NCF catalysts were unable to reduce bicarbonate to formate, whereas formate was produced when we performed bicarbonate reduction over Sn_100−yPd_y–NCF. Only a detectable amount of formate (FE_formate < 1%) was obtained from Sn₉₁Pd₉–NCF and Sn₉₄Pd₆–NCF. However, a quantifiable amount of formate (FE_formate > 3%) was obtained from the catalysts Sn₉₅Pd₅–NCF and Sn₉₇Pd₃–NCF (Fig. S16†). A low amount of formate formation suggests that the bicarbonate reduction is not kinetically feasible on the SnO_x–PdO surface, but the electrode stabilizes the bicarbonate ions on its surface. Compiling the results obtained from the material characterization, electrochemical CO₂RR, and bicarbonate reduction reaction, the catalyst surface with the SnO_x–PdO combination appeared to be more active than a catalyst surface with SnO_x and SnO_x–Pd. Further, we investigated the bicarbonate reduction activity of Pd and PdO by conducting bicarbonate electrolysis on bare Pd and the oxidized Pd electrode. The bicarbonate electrolysis performed on the Pd surface produced no formate, however, formate was obtained when the electrolysis was performed on PdO (section S-14 and Fig. S-17†), revealing that the PdO surface is capable of stabilizing and reducing the bicarbonate ion.

Based on these analyses and previously reported mechanisms,^19,62,63 we speculate the probable CO₂ reduction reaction mechanism on Sn₁₀₀–NCF with SnO_x (Scheme 1) and Sn_100−yPd_y–NCF with SnO_x–PdO (Scheme 2) as described below, however detailed in situ studies are warranted to confirm this mechanism. The CO₂ reduction mechanism on the Sn₁₀₀–NCF catalyst involves the formation and stabilization of CO₂˙⁻ at the catalyst surface. Stabilized CO₂˙⁻ further accepts a proton from the bicarbonate ions of the electrolyte to form formate (Scheme 1).^19,63 However, on the Sn_100−yPd_y–NCF surface with SnO_x–PdO, the bicarbonate ion adsorbed on the catalyst surface protonates CO₂˙⁻ to produce formate (Scheme 2).

Scheme 1 CO₂ reduction to formate on the Sn₁₀₀–NCF surface.

Scheme 2 CO₂ reduction to formate on the SnO_x–PdO surface.

4. Conclusions

In conclusion, the electrocatalytic CO₂ reduction reaction activity of SnPd nanoparticles decorated on N-doped carbon fibers with different Pd contents was analyzed. The catalysts were prepared using a simple one-step electrospinning method followed by pyrolysis. The Sn_100−yPd_y–NCF catalyst with a low Pd content, 3 wt% Pd, showed superior CO₂RR activity (total faradaic efficiency of 85% and a partial current density of 12.7 mA cm⁻² at −0.92 V compared with RHE) to the Sn_100−yPd_y–NCF catalysts with a high Pd content. The analysis of the surface composition revealed that Sn_100−yPd_y–NCF with 3 wt% Pd mainly contained PdO and SnO_x on its surface. The Tafel slope analysis and bicarbonate reduction activity of Sn₉₇Pd₃–NCF catalysts suggest a possibility that the SnO_x–PdO species at the catalyst surface effectively stabilize CO₂˙⁻ and bicarbonate on their surface. Simultaneous stabilization of bicarbonate anions and CO₂˙⁻ on the catalytic surface can promote the CO₂RR by quick protonation of CO₂˙⁻ to produce formate.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge the financial assistance from the Japan Science and Technology (JST) Core Research for Evolutionary Science and Technology, Japan Science and Technology Agency (JST-CREST, JPMJCR 1543), and this study was also supported in part by the “Five-Star Alliance”. Professor Ken Motokura, Tokyo Institute of Technology, is gratefully acknowledged for providing access to the gas chromatography facility.

Notes and references

1. Earth System Research Laboratory Global Monitoring Division, Mauna Loa observatory. https://www.esrl.noaa.gov/gmd/webdata/ccgg/trends/co2_data_mlo.png.
2. T. R. Karl and K. E. Trenberth, Science, 2003, 302, 1719–1723.
3. S. H. Schneider, Science, 1989, 243, 771–781.
4. D. Tong, Q. Zhang, Y. Zheng, K. Caldeira, C. Shearer, C. Hong, Y. Qin and S. J. Davis, Nature, 2019, 572, 373–377.
5. C. Song, Catal. Today, 2006, 115, 2–32.
6. E. Alper and O. Yuksel Orhan, Petroleum, 2017, 3, 109–126.
7. C. Ampelli, S. Perathoner and G. Centi, Philos. Trans. R. Soc., A, 2015, 373, 1–30.
8. M. Mikkelsen, M. Jørgensen and F. C. Krebs, Energy Environ. Sci., 2010, 3, 43–81.
9. E. E. Benson, C. P. Kubiak, A. J. Sathrum and J. M. Smieja, Chem. Soc. Rev., 2009, 38, 89–99.
10. S. Back, H. Kim and Y. Jung, ACS Catal., 2015, 5, 965–971.
11. G. Centi, C. Ampelli, C. Genovese, B. C. Marepally, G. Papanikolaou and S. Perathoner, Faraday Discuss., 2015, 183, 125–145.
12. S. Ma, M. Sadakiyo, R. Luo, M. Heima, M. Yamauchi and P. J. A. Kenis, J. Power Sources, 2016, 301, 219–228.
13. H. R. M. Jhong, S. Ma and P. J. Kenis, Curr. Opin. Chem. Eng., 2013, 2, 191–199.
14. Z. Chen, P. Kang, M.-T. Zhang and T. J. Meyer, Chem. Commun., 2014, 50, 335–337.
15. P. Kang, Z. Chen, A. Nayak and T. J. Meyer, Energy Environ. Sci., 2014, 07, 4007–4012.
16. K. P. Kuhl, T. Hatsukade, E. R. Cave, D. N. Abram, J. Kibsgaard and T. F. Jaramillo, J. Am. Chem. Soc., 2014, 136, 14107–14113.
17. R. Kortlever, J. Shen, K. J. P. Schouten, F. Calle-Vallejo and M. T. M. Koper, J. Phys. Chem. Lett., 2015, 6, 4073–4082.
18. Y. Hori, Mod. Aspects Electrochem., 2008, 42, 89–189.
19. S. Zhang, P. Kang and T. J. Meyer, J. Am. Chem. Soc., 2014, 136, 1734–1737.
20. D. Gao, H. Zhou, J. Wang, S. Miao, F. Yang, G. Wang, J. Wang, X. Bao, D. Gao, H. Zhou, J. Wang, S. Miao, F. Yang and G. Wang, J. Am. Chem. Soc., 2015, 137, 4288–4291.
21. Y. Chen, C. W. Li and M. W. Kanan, J. Am. Chem. Soc., 2012, 134, 19969–19972.
22. K. Manthiram, B. J. Beberwyck and A. P. Alivisatos, J. Am. Chem. Soc., 2014, 136, 13319–13325.
23. C. Kim, H. S. Jeon, T. Eom, M. S. Jee, H. Kim, C. M. Friend, B. K. Min and Y. J. Hwang, J. Am. Chem. Soc., 2015, 137, 13844–13850.
24. R. Kortlever, I. Peters, S. Koper and M. T. M. Koper, ACS Catal., 2015, 5, 3916–3923.
25. X. Zhao, B. Luo, R. Long, C. Wang and Y. Xiong, J. Mater. Chem. A, 2015, 3, 4134–4138.
26. Z. Xu, E. Lai, Y. Shao-Horn and K. Hamad-Schifferli, Chem. Commun., 2012, 48, 5626–5628.
27. S. Rasul, D. H. Anjum, A. Jedidi, Y. Minenkov, L. Cavallo and K. Takanabe, Angew. Chem., Int. Ed., 2015, 54, 2146–2150.
28. P. P. Sharma, J. Wu, R. M. Yadav, M. Liu, C. J. Wright, C. S. Tiwary, B. I. Yakobson, J. Lou, P. M. Ajayan and X. D. Zhou, Angew. Chem., Int. Ed., 2015, 54, 13701–13705.
29. H. Wang, Y. Chen, X. Hou, C. Ma and T. Tan, Green Chem., 2016, 18, 3250–3256.
30. N. Sreekanth, M. A. Nazrulla, T. V. Vineesh, K. Sailaja and K. L. Phani, Chem. Commun., 2015, 51, 16061–16064.
31. S. Pal, S. Narayanaru, B. Kundu, M. Sahoo, S. Bawari, D. K. Rao, S. K. Nayak, A. J. Pal and T. N. Narayanan, J. Phys. Chem. C, 2018, 122, 23385–23392.
32. H. R. M. Jhong, C. E. Tornow, B. Smid, A. A. Gewirth, S. M. Lyth and P. J. A. Kenis, ChemSusChem, 2017, 10, 1094–1099.
33. P. Iyengar, J. Huang, G. L. De Gregorio, C. Gadiyar and R. Buonsanti, Chem. Commun., 2019, 55, 8796–8799.
34. S. Narayanaru, J. Chinnaiah, K. L. Phani and F. Scholz, Electrochim. Acta, 2018, 264, 269–274.
35. S. Sen, D. Liu and G. T. R. Palmore, ACS Catal., 2014, 4, 3091–3095.
36. A. Murata and Y. Hori, Bull. Chem. Soc. Jpn., 1991, 64, 123–127.
37. K. P. Kuhl, E. R. Cave, D. N. Abram and T. F. Jaramillo, Energy Environ. Sci., 2012, 5, 7050–7059.
38. S. Verma, B. Kim, H. R. M. Jhong, S. Ma and P. J. A. Kenis, ChemSusChem, 2016, 9, 1972–1979.
39. J. Eppinger and K. W. Huang, ACS Energy Lett., 2017, 2, 188–195.
40. L. An and R. Chen, J. Power Sources, 2016, 320, 127–139.
41. S. Sankar, G. M. Anilkumar, T. Tamaki and T. Yamaguchi, ChemCatChem, 2019, 11, 4731–4737.
42. H. Kuroki, S. Miyanishi, A. Sakakibara, Y. Oshiba and T. Yamaguchi, J. Power Sources, 2019, 438, 226997.
43. A. Del Castillo, M. Alvarez-Guerra, J. Solla-Gullón, A. Sáez, V. Montiel and A. Irabien, Appl. Energy, 2015, 157, 165–173.
44. A. Álvarez, A. Bansode, A. Urakawa, A. V. Bavykina, T. A. Wezendonk, M. Makkee, J. Gascon and F. Kapteijn, Chem. Rev., 2017, 117, 9804–9838.
45. S. Lee, J. D. Ocon, Y. Il Son and J. Lee, J. Phys. Chem. C, 2015, 119, 4884–4890.
46. S. Zhao, S. Li, T. Guo, S. Zhang, J. Wang, Y. Wu and Y. Chen, Nano-Micro Lett., 2019, 11, 62.
47. X. Bai, W. Chen, C. Zhao, S. Li, Y. Song, R. Ge, W. Wei and Y. Sun, Angew. Chem., Int. Ed., 2017, 56, 12219–12223.
48. D. H. Won, C. H. Choi, J. Chung, M. W. Chung, E. H. Kim and S. I. Woo, ChemSusChem, 2015, 8, 3092–3098.
49. Y. Fu, Y. Li, X. Zhang, Y. Liu, J. Qiao, J. Zhang and D. P. Wilkinson, Appl. Energy, 2016, 175, 536–544.
50. Y. Chen and M. W. Kanan, J. Am. Chem. Soc., 2012, 134, 1986–1989.
51. M. Spichiger-Ulmann and J. Augustynski, J. Chem. Soc., Faraday Trans. 1, 1985, 81, 713–716.
52. N. Sreekanth and K. L. Phani, Chem. Commun., 2014, 50, 11143–11146.
53. S. Dalton, F. Heatley and P. M. Budd, Polymer, 1999, 40, 5531–5543.
54. W. Du, K. E. MacKenzie, D. F. Milano, N. A. Deskins, D. Su and X. Teng, ACS Catal., 2012, 2, 287–297.
55. Y. Li, J. Qiao, X. Zhang, T. Lei, A. Girma, Y. Liu and J. Zhang, ChemElectroChem, 2016, 3, 1618–1628.
56. B. Jiang, X. G. Zhang, K. Jiang, D. Y. Wu and W. Bin Cai, J. Am. Chem. Soc., 2018, 140, 2880–2889.
57. S. Sankar, G. M. Anilkumar, T. Tamaki and T. Yamaguchi, ACS Appl. Energy Mater., 2018, 1, 4140–4149.
58. Z. Li, Y. Chen, G. Fu, Y. Chen, D. Sun, J. M. Lee and Y. Tang, Nanoscale, 2019, 11, 2974–2980.
59. X. Min and M. W. Kanan, J. Am. Chem. Soc., 2015, 137, 4701–4708.
60. R. Kortlever, C. Balemans, Y. Kwon and M. T. M. Koper, Catal. Today, 2015, 244, 58–62.
61. S. Chatterjee, C. Griego, J. L. Hart, Y. Li, M. L. Taheri, J. Keith and J. D. Snyder, ACS Catal., 2019, 9, 5290–5301.
62. W. Zhang, Q. Qin, L. Dai, R. Qin, X. Zhao, X. Chen, D. Ou, J. Chen, T. Chuong and B. Wu, Angew. Chem., Int. Ed., 2018, 57, 9475–9479.
63. 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, ACS Catal., 2017, 7, 4822–4827.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d0cy01437k

---