Facile Synthesis of Fe@C Loaded on g-C3N4 for CO2 Electrochemical Reduction to CO with Low Overpotential.

Electrochemical CO2 reduction has been acknowledged as a hopeful tactic to alleviate environmental and global energy crises. Herein, we designed an Fe@C/g-C3N4 heterogeneous nanocomposite material by a simple one-pot method, which we applied to the electrocatalytic CO2 reduction reaction (ECR). Our optimized 20 mg-Fe@C/g-C3N4-1100 catalyst displays excellent performance for the ECR and a maximum Faradaic efficiency (FE) of 88% with a low overpotential of -0.38 V vs. RHE. The Tafel slope reveals that the first electron transfer, which involves a surface-adsorbed *COOH intermediate, is the rate-determining step for 20 mg-Fe@C/C3N4-1100 during the ECR. More precisely, the coordinating capability of the g-C3N4 framework and Fe@C species as a highly active site promote the intermediate product transmission. These results indicate that the combination of temperature adjustment and precursor optimization is key to facilitating the ECR of an iron-based catalyst.

Introduction

The rising carbon dioxide (CO2) concentration in the atmosphere, which causes a variety of problems such as global warming, aggravation of desertification, and a decline in biodiversity, is gradually threatening the sustainable development of human beings.[1−4] Electrochemical reduction of CO2, which can be powered by electricity from renewable sources, is considered to be a potential efficient way to transform CO2 into value-added fuels and chemical feedstocks.[5−7] However, the electrochemical CO2 reduction reaction (ECR) is impeded by the high overpotential and poor selectivity due to the high energy barriers of CO2 and the hydrogen evolution reaction (HER).[8,9] For these reasons, the exploration of efficient, original, and useful electrode materials for the ECR is significant at this current stage.

Recently, carbon-based catalysts, as abundant and cheap materials, have been considered to be hopeful electrocatalytic carbon dioxide candidates due to their high specific surface area, remarkable electrical conductivity, and outstanding chemical stability.[10−12] However, pure carbon catalysts display no activity for CO2 valorization which therefore hampers their roles in the electrocatalytic CO2 process. On the grounds of experimental research, we know that a transition metal (e.g., Fe,[13] Co,[14] Ni,[15] and Mn[16]) anchored on a carbon-based material can enable an apparent catalytic activity to improve the efficiency and selectivity of ECR. Especially, Li et al.[17] verified that, a superb proton activation capability, the Fe dopant could reduce the reaction barriers and decrease the overpotential. Hence, it is feasible to develop an Fe@C material to assist in electrochemical catalysis reactions.

Furthermore, loading an Fe@C material onto support materials is a useful strategy to avoid reunion. An ideal support material not only offers a large surface area during the electrocatalytic reactions but also accelerates abundant interface links between the carrier and the metal nanoparticles. For example, Zhang et al.[18] investigated detailed mechanisms of Fe/g-C3N4, Co/g-C3N4, and Ni/g-C3N4 catalysts for electrochemical CO2 reduction. Fe/g-C3N4 showed the highest electrocatalytic activity in comparison to Co/g-C3N4 and Ni/g-C3N4 in terms of the activity and stability. On the basis of the above research, g-C3N4, as a promising support material, has a strong affinity to CO2 and high oxophilicity for adsorption of an adsorption intermediate.[19] Consequently, integrating g-C3N4 with Fe@C generates a stable product that constitutes part of a potential class of active catalysts for the ECR.

Herein, we report a Fe@C/g-C3N4 nanocomposite, where the Fe nanoparticles were uniformly deposited on the surface of g-C3N4 nanosheets. In addition, the composite catalyst was used in a traditional H-type cell assembly to complete the experiment. In order to evaluate the high activity, long stability, and selectivity of the catalysts, a series of Fe@C/g-C3N4 catalysts with different amounts of Fe-based materials (0, 10, 20, and 30 mg) as well as different sintering temperatures (900 and 1100 °C) were synthesized for the process of the ECR. Additionally, the characterizations revealed the magnetism of 20 mg-Fe@C/C3N4-1100 created a new platform to boost the performance of the catalyst in electrocatalytic CO2 reduction. Electrocatalytic performance tests confirmed that 20 mg-Fe@C/C3N4-1100 possessed the fastest charge transfer rate and the best selectivity. Potentially, the electrocatalyst material 20 mg-Fe@C/C3N4-1100 as a promising tool will open up a new pathway for the ECR.

Results and Discussion

As shown in Figure a, all of the X-ray diffraction (XRD) patterns of 10 mg-Fe@C/C3N4-1100, 20 mg-Fe@C/C3N4-1100, and 30 mg-Fe@C/C3N4-1100 displayed one pronounced diffraction peak at approximately 44.7°, which corresponds to the (110) plane of Fe0 (PDF #06-0696). A broad peak at close to 22° is derived from the (002) reflection of the graphitic carbon structure. Moreover, no other reflection peaks of iron nitride (PDF #49-1664 and #50-1087) were identified, suggesting the predominant formation of metallic Fe in the samples. By utilizing Scherrer’s equation relating the coherently scattering domains with the Bragg peak widths, L = κλ/B cos θ, in which κ = 0.89 for spherical particle and B is the full angular width at half-maximum of the peak in radians, the average crystal size was determined to be around 17.8 nm for 20 mg-Fe@C/C3N4-1100.[20] Moreover, we noted that improving the loading content of the iron precursor led to an increase in the particle size. Additionally, the particle size increased upon an increase in annealing temperature.

Figure 1

(a) XRD patterns of C/C3N4-1100, 10 mg-Fe@C/C3N4-1100, 20 mg-Fe@C/C3N4-1100, and 30 mg-Fe@C/C3N4-1100. (b) N 1s and (c) Fe 2p XPS spectra of 20 mg-Fe@C/C3N4-1100 and 20 mg-Fe@C/C3N4-900. (d) H-TPR spectra of 20 mg-Fe@C/C3N4-1100 and 20 mg-Fe@C/C3N4-900.

Subsequently, X-ray photoelectron spectroscopy (XPS) was employed to provide insight into the surface composition of the resulting 20 mg-Fe@C/C3N4-1100 and 20 mg-Fe@C/C3N4-900. A dominant N 1s peak at 400.7 eV was observed for both samples, which can be assigned to pyrrolic N.[21] An apparent peak at around 399.68 eV in 20 mg-Fe@C/C3N4-900, which disappeared in 20 mg-Fe@C/C3N4-1100, confirmed the presence of N similar to that of the Fe-N moiety.[22,23] The content of the pyridinic N improved the predominant formation of iron.[24] The N–Fe 2p3/2 binding energies (BEs) of 20 mg-Fe@C/C3N4-1100 were shifted to lower values in comparison with the BEs of 20 mg-Fe@C/C3N4-900 (Figure c), suggesting that 20 mg-Fe@C/C3N4-1100 composite material had more iron in comparison to the other samples.

To further characterize the redox properties of the prepared materials, H2-TPR was used to describe 20 mg-Fe@C/C3N4-1100 and 20 mg-Fe@C/C3N4-900, as shown in Figure d. Their H2 consumption peaks, centered at approximately 410 and 520 °C, corresponding to the reduction of iron trioxide to ferrous ion,[25,26] were detected for 20 mg-Fe@C/C3N4-900. The last H2 consumption peak (630 °C), corresponding to the reduction of Fe2+ to Fe0, was relatively weak.[27] However, there was no obvious H2 consumption peak for 20 mg-Fe@C/C3N4-1100, implying that the main component of this material was iron. Obviously, it has been confirmed that the H2-TPR results were consistent with the XRD and XPS results.

The morphological properties of 20 mg-Fe@C/C3N4-1100 were characterized by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Figure a,b showed that the Fe nanoparticles were successfully deposited on the surface of g-C3N4 while the nanosheet structure remained intact. In addition, the size distribution histogram of Fe nanoparticles in 20 mg-Fe@C/C3N4-1100 was analyzed.[28] As can be seen from the inset of Figure c, the representative diameters are from 5 to 50 nm, and the average diameter is 15 nm. Moreover, the d spacing of 0.2014 nm in Figure d, which agreed well with the lattice space of the Fe (110) plane, confirmed the generation of Fe nanoparticles.[29,30] As shown in Figure e, the (110) plane of Fe was also confirmed by fast Fourier transform (FFT) of the region, as shown in Figure e. Therefore, TEM and HR-TEM further testified to the formation of the 20 mg-Fe@C/C3N4-1100 composite in this work.

Figure 2

(a, b) SEM images of 20 mg-Fe@C/C3N4-1100. (c, d) TEM images of 20 mg-Fe@C/C3N4-1100. Inset in (c): size distribution of Fe nanoparticles. (e) Fast Fourier transform (FFT) of the region shown in (d).

The magnetic characteristics of Fe@C/C3N4 with different iron contents have been confirmed by VSM, and the magnetization curves are shown in Figure a. It was noted that 30 mg-Fe@C/C3N4-1100, 20 mg-Fe@C/C3N4-1100, and 20 mg-Fe@C/C3N4-900 had magnetic saturations of ∼7.5, 16, and 12.5 emu/g at an ∼3000 kOe field strength, respectively. As expected, 20 mg-Fe@C/C3N4-1100 had higher magnetic properties in comparison to the other samples.[31,32] More importantly, the prepared catalyst had certain magnetism, which provided a novel train of thought for the development of further different catalysts.[33] Additionally, the electron paramagnetic resonance (EPR) signal intensity of 20 mg-Fe@C/C3N4-1100 also increased in comparison to that of 20 mg-Fe@C/C3N4-900, implying that the Fe magnetic states had increased (Figure b).[34] To sum up, the 20 mg-Fe@C/C3N4-1100 precatalyst displayed a transformation of Fe microstructure and exhibited excellent magnetism.[29]

Figure 3

(a) VSM curves illustrating the saturation magnetization of 30 mg-Fe@C/C3N4-1100, 20 mg-Fe@C/C3N4-1100 and 20 mg-Fe@C/C3N4-900. (b) EPR spectra of 20 mg-Fe@C/C3N4-1100 and 20 mg-Fe@C/C3N4-900.

To evaluate the CO2 reduction reaction activity of 20 mg-Fe@C/C3N4-1100, linear sweep voltammetry (LSV) measurements of a well-prepared catalyst/carbon paper was conducted with N2 -or CO2-saturated 0.1 M KHCO3 electrolytes from −0.1 to −1.3 V vs the RHE at a scan rate of 0.1 V s–1. As shown in Figure a, the as-prepared electrode displayed a potential of −0.6 VRHE at 5 mA cm–2 under CO2-saturated conditions in comparison with −1.1 VRHE in an N2-saturated solution. Thus, it can be seen expressly that the prepared electrode was active for CO2 reduction.[35] When a solution was saturated with CO2, there was a shift in the onset potential toward lower negative potentials along with an increased current density due to the electroreduction of CO2.

Figure 4

(a) LSV of a 20 mg-Fe@C/C3N4-1100 electrode in N2- and CO2-saturated 0.1 M KHCO3 solutions. (b) Faradaic efficiency (FE) of CO at different potentials for 0 mg-Fe@C/C3N4-1100, 10 mg-Fe@C/C3N4-1100, 20 mg-Fe@C/C3N4-1100, and 30 mg-Fe@C/C3N4-1100. (c) CO current densities at different potentials and (d) stability tests of a 20 mg-Fe@C/C3N4-1100 electrode at −0.68 V vs the RHE for 5 h.

More notably, the Faradaic efficiencies toward CO (CO-FE) for the various kinds of Fe@C/C3N4-1100 catalysts as a function of the applied potential are given in Figure b. On comparison of the Faradaic efficiencies of CO produced by the four electrodes, it can be seen that the Fe@C/C3N4-1100 electrodes in 0.1 M KHCO3 with an increasing applied potential revealed that the FE CO increased in a potential window between −0.38 and −0.48 V (vs RHE), reaching a maximum, and then decreased between −0.48 and −0.78 V (vs RHE). Simultaneously, the 20 mg-Fe@C/C3N4-1100 electrode had a higher Faradaic efficiency, attaining a maximum of 88% at −0.48 V. However, CO and H2 are the only gaseous products and no other gaseous species could be detected by online gas chromatography (GC).[36] It is worth noting that a 20 mg-Fe@C/C3N4-1100 electrode had an excellent selectivity for CO (FE CO, 88%) with a low overpotential. Interestingly, Figure c provided the partial current densities of CO for 0 mg-Fe@C/C3N4-1100, 10 mg-Fe@C/C3N4-1100, 20 mg-Fe@C/C3N4-1100, and 30 mg-Fe@C/C3N4-1100 catalysts. One can see that the 20 mg-Fe@C/C3N4-1100 catalyst offered a higher CO current density in comparison to the other catalysts, implying a higher CO production capacity.[37]

As the stability of electrodes for the ECR is a vital element to measure the prospective practical applications,[38] the stability of the 20 mg-Fe@C/C3N4-1100 electrode was tested in 0.1 M KHCO3 at −0.68 V vs. RHE by using a proton exchange membrane pretreatment to segregate the anode and cathode chambers. Figure d showed that a current density of ∼5.5 mA cm–2 on the 20 mg-Fe@C/C3N4-1100 catalyst was almost completely maintained during 5 h of continuous electrochemical reduction, indicating that this electrode had a stable nature.

The charge transfer resistance at the electrolyte interface can be depicted on the grounds of the arc radius of a Nyquist curve acquired by EIS, where the charge transfer resistance decreases with decreasing arc radius. The Nyquist curves obtained for 20 mg-Fe@C/C3N4 with different sintering temperatures and 0 mg-Fe@C/C3N4-1100 are shown in Figure a. It was clearly demonstrated that the 20 mg-Fe@C/C3N4-1100 had the lowest charge transfer impedance (Rct) with the order being 20 mg-Fe@C/C3N4-1100 < 20 mg-Fe@C/C3N4-900 < 0 mg-Fe@C/C3N4-1100. This result was consistent with the experimental observation that the addition of Fe nanoparticles promoted the electron transfer and current density.[39] Prior research indicated that catalysts with lower Rct values had lower charge-transfer resistance between the reactant and the surface of the catalyst, providing a fast pathway for transferring electrons to CO2 in order to generate CO2• intermediates.

Figure 5

(a) EIS Nyquist plots of 0 mg-Fe@C/C3N4-1100, 20 mg-Fe@C/C3N4-900, and 20 mg-Fe@C/C3N4-1100. (b) The CO formation TOFs of 0 mg-Fe@C/C3N4-1100, 10 mg-Fe@C/C3N4-1100, 20 mg-Fe@C/C3N4-1100, and 30 mg-Fe@C/C3N4-1100 at different potentials. (c) Cdl values of 0 mg-Fe@C/C3N4-1100, 20 mg-Fe@C/C3N4-900, and 20 mg-Fe@C/C3N4-1100. (d) Tafel plots for the CO production on 20 mg-Fe@C/C3N4-1100 and 20 mg-Fe@C/C3N4-900, where jCO is the partial current density of CO.

The CO formation turnover frequency (TOF) was measured for the presite activity of the catalyst material to generate CO (Figure b). Prominently, the TOF of 20 mg-Fe@C/C3N4-1100 was 0.63 s–1 versus 0.18 s–1 for 10 mg-Fe@C/C3N4-1100 and 0.35 s–1 for 20 mg-Fe@C/C3N4-1100 at −0.5 V vs RHE. Nevertheless, the CO formation TOF was only 0.09 s–1 for C/C3N4-1100 at −0.5 V vs RHE, implying that the pure C/C3N4-1100 catalyst produced little CO in the CO2 reduction.

The electrochemical active surface areas (ECSAs) of 0 mg-Fe@C/C3N4-1100 and 20 mg-Fe@C/C3N4 at different temperatures were investigated using the electrochemical double layer capacitance (Cdl), which were tested with cyclic voltammetry (Figure c).[40] The Cdl value of 20 mg-Fe@C/C3N4-1100 was 24.1 mF/cm2, higher than those of C/C3N4-1100 (12.6 mF/cm2) and 20 mg-Fe@C/C3N4-900 (14.5 mF/cm2), demonstrating large amounts of exposed active sites for 20 mg-Fe@C/C3N4-1100 during the ECR process.[41]

Furthermore, the Tafel slopes could be procured from LSV curves to explore the kinetics and possible mechanism for ECR on 20 mg-Fe@C/C3N4 catalysts at different temperatures. Slopes of 116 and 166 mV dec–1 were fitted for 20 mg-Fe@C/C3N4-1100 and 20 mg-Fe@C/C3N4-900 (Figure d), implying that 20 mg-Fe@C/C3N4-1100 had a faster increment of the CO2 reduction rate with increasing overpotential.[29,42] Simultaneously, this result reveals that the first electron transfer, which gives the surface-adsorbed *COOH intermediate, was the rate-determining step for 20 mg-Fe@C/C3N4-1100 during the ECR.

Results of the studies on the electrocatalytic reduction of CO2 to CO on different electrocatalysts are given in Table . It is worth noting that 20 mg-Fe@C/C3N4-1100 showed high selectivity at a low potential, thus proving its enormous potential for electrocatalytic CO2 performance.

Table 1

Comparison with Different Electrocatalysts for CO2 Reduction to COa

catalyst	potential vs RHE (V)	electrolyte	FE (%)	ref
20 mg-Fe@C/g-C3N4-1100	–0.38	0.1 M KHCO3	88	this work
Co1-N4	–0.8	0.1 M KHCO3	82	(43)
F-CPC	–1	0.5 M KHCO3	88.3	(44)
Fe–N–C	–0.46	0.5 M KHCO3	85	(45)
w-CCG/CoPc-A	–0.79	0.1 M KHCO3	91.5	(46)
NPCM-1000	–0.55	0.5 M KHCO3	92	(47)
ZIF-A-LD	–1.1	0.1 M KHCO3	90.57	(48)
Bi6Pd94-SAA NDs	–0.4	0.5 M KHCO3	90.5	(49)
Fe/NG-750	–0.57	0.1 M KHCO3	80	(50)

Abbreviations: fluorine-doped cagelike porous carbon, F-CPC; “washed” Co(II) phthalocyanine chemically converted graphene, w-CCG/CoPc-A; nitrogen–phosphorus codoped carbon materials, NPCM; coordination of Zn with N atoms in phenanthroline to form a ligand-doped product, ZIF-A-LD; Bi6–Pd94 single atom alloy (SAA) nanodendrites (NDs), Bi6Pd94-SAA NDs.

Conclusion

In this work, a useful ferromagnetic catalyst based on 20 mg-Fe@C/C3N4-1100 has been successfully synthesized. The electrocatalytic activity of the 20 mg-Fe@C/C3N4-1100 composite catalyst is visibly improved by the synergistic effect between g-C3N4 as the supporting substrate and iron particles with highly active sites heightening the transfer of charge, leading to a maximum FE of 88% for CO and a low overpotential of −0.38 V vs RHE. This is the best selectivity and activity for CO among many iron-based catalysts reported thus far. The Tafel slopes further display that the 20 mg-Fe@C/C3N4-1100 catalyst stabilizes the *OCOOH intermediate more effectively, thus increasing the potential for conversion of CO2 to CO. The present study offers a novel avenue to design potential precatalysts and gives new insights into the application of the ECR.

Experimental Section

Materials

Ferric acetylacetonate (C15H21FeO6), Pluronic F-127, and KHCO3 were purchased from Aladdin Industrial Corporation. Melamine (C3H6N6) and concentrated hydrochloric acid (HCl) were purchased from Sinopharm Chemical Reagent Co. Ltd., China, and used without further purification.

Fabrication of g-C3N4

g-C3N4 were successfully synthesized by directly heating melamine. First, melamine was dried at 80 °C for 12 h in an oven. This dried melamine powder was then put into an alumina crucible with a cover and heated to 550 °C in a muffle furnace for 2 h at a heating rate of 5 °C/min. The resulting g-C3N4 was then subjected to grinding to increase the specific surface area.

Fabrication of 10 mg-Fe@C/g-C3N4-1100

In a typical synthesis of 10 mg-Fe@C/C3N4-1100, 3 g of g-C3N4, 6 g of Pluronic F-127 (the mass ratio is 1:2) and 10 mg of Fe(acac)3 were uniformly dissolved in 500 mL of deionized water by an ultrasonic treatment for 2 h. After this solution was vigorously stirring for another 6 h, the resultant products were separated by centrifugation and washed three times with ultrapure water and absolute ethanol. Subsequently, the black powder was carbonized in a tube furnace at 1100 °C for 1 h with a heating rate of 5 °C min–1 under a N2 atmosphere (40 mL/min). This powder was suspended in 2 M HCl solution for 24 h and then washed repeatedly with ethanol and ultrapure water to obtain the final sample.

Fabrication of 0 mg-Fe@C/g-C3N4-1100, 20 mg-Fe@C/g-C3N4-1100, and 30 mg-Fe@C/g-C3N4-1100

Typically, the preparation process was similar to that for 10 mg-Fe@C/g-C3N4-1100, except that 10 mg of Fe(acac)3 was replaced by 0, 20, or 30 mg Fe(acac)3 for 0 mg-Fe@C/g-C3N4-1100, 20 mg-Fe@C/g-C3N4-1100, and 30 mg-Fe@C/g-C3N4-1100, respectively.

Fabrication of 20 mg-Fe@C/g-C3N4-900

The preparation process was the same as that for 10 mg-Fe@C/g-C3N4-1100, except that the sintering temperature of 1100 °C was changed to 900 °C for 20 mg-Fe@C/g-C3N4-900.

Materials Characterization

The crystal structures, compositions, morphologies, and valence states of elements of the Fe@C/C3N4 composite materials were examined by X-ray powder diffraction (XRD), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), transmission electron microscopy (TEM), fast Fourier transform (FFT), and H2-temperature-programmed reduction (TPR). Their abilites to carry out the ECR were appraised under electrochemical conditions, and the reaction products were quantified by Weather Chromatograph, Model SP-2100 A, equipped with a thermal conductivity detector (TCD) and flame ionization detector (FID). These TCD and FID detectors were used to detect H2 and CO, respectively.

Electrochemical Tests

Electrochemical measurements were conducted with a traditional three-electrode system by means of a CHI 660E workstation (Shanghai Chenhua Instrumental Co., Ltd., China) using a sealed H-cell. An Ag/AgCl electrode was utilized as the reference electrode, and Pt foil served as the counter electrode. An Fe@C/C3N4 working electrode with an effective area of 1 × 1 cm2 was used for the ECR. In the cathode compartment, the catholyte utilized in our study was 0.1 M KHCO3. Before experiments, the catholyte was purged with N2 and CO2 for 30 min; the pHs of 0.1 M KHCO3 saturated with N2 and CO2 were 8.3 and 6.7, respectively. To transform all of the potentials to references to the reversible hydrogen electrode (RHE), the following formula was used