Syngas, a CO and H2 mixture mostly generated from non-renewable fossil fuels, is an essential feedstock for production of liquid fuels. Electrochemical reduction of CO2 and H+/H2O is an alternative renewable route to produce syngas. Here we introduce the concept of coupling a hydrogen evolution reaction (HER) catalyst with a CDots/C3N4 composite (a CO2 reduction catalyst) to achieve a cheap, stable, selective and efficient route for tunable syngas production. Co3O4, MoS2, Au and Pt serve as the HER component. The Co3O4-CDots-C3N4 electrocatalyst is found to be the most efficient among the combinations studied. The H2/CO ratio of the produced syngas is tunable from 0.07:1 to 4:1 by controlling the potential. This catalyst is highly stable for syngas generation (over 100 h) with no other products besides CO and H2. Insight into the mechanisms balancing between CO2 reduction and H2 evolution when applying the HER-CDots-C3N4 catalyst concept is provided.
Syngas, a mixture of H2 and CO, is a critical feedstock for production of synthetic fuels and industrial chemicals via well-established industrial processes such as the Fischer–Tropsch process (commercialized by Sasol and Shell)1, 2. The H2/CO ratio in syngas is of a great significance for meeting the requirements for specific products: H2/CO = 2:1 for methanol and H2/CO = 1:1 for dimethyl ether for example. The conventional production approach of syngas is based on reforming non-renewable fossil fuels (e.g., coal, petroleum coke, and natural gas)3, which increases the consumption of fossil fuel and aggravates the energy crisis. Synthesizing syngas with a controlled H2/CO ratio by reduction of CO2 not only contributes to the solution of the energy crisis, but at the same time reduces the amount of greenhouse gases (CO2).
CO2 reduction to CO and hydrogen evolution reactions (HER) per se are two independent major and important fields. Electrochemical (EC) and photoelectrochemical (PEC) methods integrating CO2 reduction reaction and HER are key components of prospective technologies for renewable syngas4. Different types of semiconductors have been combined with an efficient catalyst for CO2 reduction to produce syngas by the PEC approach5,6,7,8. Cu-ZnO/GaN/n+−p Si was recently reported as a highly efficient PEC catalyst to produce syngas with a tunable H2/CO ratio (between 1:2 and 4:1)5. Metal or metal-based composites, including Ag9, Cu10, Ru(II) polypyridyl complex11, Ag/C3N412, and Re-functionalized graphene oxide13, have been investigated for EC reduction of CO2 and H+/H2O to syngas. The different methods used to tune the ratio of H2/CO include altering the CO2 flow rate14 and pressure15, the reaction temperature16, and the applied potential9. Different crystalline sites of Au catalyze different reaction channels (edge sites initiate CO generation and corner sites H2 generation)17, 18. A novel pulsed-bias technique using Cu as the catalyst was recently applied to tune the H2/CO ratio in syngas between ~32:1 and 9:16 by using different pulse times for the same working potential. The selectivity is however limited and CH4 and C2H4 by-products affect the purity of syngas10. Using Ag/C3N412, the H2/CO ratio in the produced syngas can be tuned from 100:1 to 2:1 by controlling the applied potential and the Ag loading on graphitic carbon nitride but the total current density is lower than 1 mA/cm2 at −0.6 V. The previous EC attempts to synthesize renewable syngas are still characterized by an unsatisfactory performance including some of the following disadvantages: a high onset overpotential necessary to initiate the CO2 reduction reaction, a low CO + H2 generation current density, a small selectivity of CO production, and a poor stability of the generation current density and the Faradaic efficiency (FE) of H2 and CO.
We hereby propose a design concept of a cheap composite EC catalyst for a tunable, stable, selective, and efficient production of syngas, made of three components: a HER catalyst, a CO2 reduction catalyst toward CO, and a catalyst which stabilizes the active hydrogen (H•) necessary to trigger both HER and the CO2 reduction reactions. For HER, we choose several known catalysts (Co3O4, Pt, MoS2, and Au). For CO2 reduction, we apply graphitic carbon nitride (C3N4) since carbon-bonded nitrogen groups including pyridinic N, pyrrolic N, and graphitic N have recently been proposed as active sites for CO2 reduction to CO19,20,21. C3N4 has a porous structure and is shown as a good substrate for dispersion of catalytic nanoparticles22,23,24,25. The selected catalyst for stabilization of active hydrogen (H•) are carbon dots (CDots)26, 27, which possess significant adsorption capabilities for H+28, 29 and CO230 and exhibit excellent ability of electron transfer31, 32 necessary for H• generation (H++e−→H•). CDots also improve the conductivity of C3N4.
Here we found that the Co3O4-CDots-C3N4 is the best EC catalyst for syngas production studied within the present work. We thus first describe the structural characterization of this catalyst. We then present the design concept of the three-component catalyst for syngas production and detail its working mechanisms. We follow by showing that this concept is valid and that Co3O4-CDots-C3N4 is indeed an efficient, tunable, stable, selective, and cheap EC catalyst for syngas production. It initiates CO2 reduction to CO in aqueous solutions at a low overpotential (0.17 V vs. reversible hydrogen electrode (RHE)) and the total current density of H2 and CO generation may reach up to 15 mA/cm2 at a potential of −1.0 V vs. RHE. The H2/CO ratio of syngas generated by Co3O4-CDots-C3N4 is tunable from 0.07:1 to 4:1 by controlling the applied potential. Dedicated experiments highlight the different mechanisms by which syngas production applying Co3O4-CDots-C3N4 is controlled and manipulated. Finally, the generality of the catalyst design concept applying Pt, MoS2, and Au as the HER catalyst is demonstrated. The significance of this work is thus twofold. First, it presents Co3O4-CDots-C3N4 as an efficient, cheap, tunable, and stable catalyst for syngas production on one hand. Second, it provides an avenue to design catalysts for controlled and tuned production of syngas in particular and other chemicals of interest to the chemical industry in general.
Characterization of Co3O4-CDots-C3N4
Transmission electron microscopy (TEM) image (Fig. 1a) reveals that the Co3O4-CDots-C3N4 consist of nm-sized Co3O4 nanoparticles (NPs) and CDots dispersed on the C3N4 matrix. The two different types of nanocrystals shown in the inset of Fig. 1a are identified as Co3O4 NPs with a d-spacing of 0.24 nm consistent with Co3O4 (311)33 and CDots with an interplanar spacing of 0.33 nm (see Supplementary Fig. 1)34. Figure 1b displays a scanning transmission electron microscopy (STEM) micrograph and its corresponding chemical maps of C-K, N-K, O-K, Co-K, and Co-L for the Co3O4-CDots-C3N4. Co-K, Co-L, and O-K cover the entire C3N4 area monitored, indicating that Co3O4 NPs are evenly distributed on the C3N4 sheet. X-ray diffraction (XRD) shows (Supplementary Fig. 2) the typical diffraction lines of Co3O4 and C3N4. Besides the average size of Co3O4 NPs derived from the XRD spectrum of the Co3O4-CDots-C3N4 using the Debye−Scherrer equation (Supplementary Table 1), is about 10 nm, which is consistent with the size of Co3O4 NPs in the TEM image. Energy dispersive X-ray absorption (EDX) analysis of the Co3O4-CDots-C3N4 (Supplementary Fig. 3) reveals an elemental atom composition of C (46.65 at. %), N (48.74 at. %), O (3.01 at. %), and Co (1.60 at. %). X-ray photoelectron spectroscopy (XPS) of the composite (Supplementary Fig. 4) shows C 1s, N 1s, O 1s, and Co 2p peaks with a similar elemental composition to that deduced by EDX (C (39.9 at. %), N (52.5 at. %), O (5.8 at. %), and Co (1.8 at. %)) (Supplementary Table 2). The Co 2p3/2 peak was fitted (Supplementary Fig. 4e) by using two synthetic peaks positioned at binding energy (BE) = 780.0 and 781.3 eV (Co3+ and Co2+, respectively) and the Co 2p1/2 peak was fitted (Supplementary Fig. 4e) by using two synthetic peaks positioned at BE = 795.0 and 797.3 eV (Co3+ and Co2+, respectively). Figure 1c shows type IV N2 adsorption–desorption isotherms of the Co3O4-CDots-C3N4 with a H3-type hysteresis loop35, indicating the formation of a porous structure (the pore-size distribution was derived from the isotherms; for more details, see Supplementary Methods) is presented in Fig. 1c inset; specific surface area ~160 m2/g). The porous Co3O4-CDots-C3N4 was found to adsorb a large amount of CO2 (~0.33 mmol/g at 1.2 atm, Fig. 1d). This strong CO2 adsorption capability is mainly due to CO2 molecules adsorbed on the porous C3N4 surfaces, which constitute more than 90 wt.% of the composite catalyst (Supplementary Fig. 5a). It is nevertheless enhanced in a synergistic way by the incorporation of CDots (Supplementary Fig. 5b shows that the CO2 adsorption on Co3O4-CDots-C3N4 and on CDots-C3N4 is about the same, while that on Co3O4-C3N4 is much lower). The adsorption of H+ (Fig. 1e) is very rapid in the first 15 min, then it becomes slower with contact time, stopping additional absorption after 30 min. The amount of adsorbed H+ (for more details, see Supplementary Methods) on Co3O4-CDots-C3N4 and CDots-C3N4 is about 13.8 mg/g, while the amount of adsorbed H+ is only 5.11 mg/g for the Co3O4-C3N4 (Supplementary Fig. 6b). The incorporation of CDots thus significantly increases the adsorbed capacity of H+. All these results indicate that the present electrocatalyst has a strong adsorption capacity for both H+ and CO2, which is significantly important for both processes of CO2 reduction reaction and HER.
The electrocatalyst design and reaction mechanism
We first present the design concept of the ternary electrocatalyst and highlight its operation mode to produce syngas in a controllable, tunable way. The proposed electrocatalyst consists of three parts: an electrocatalyst for HER (e.g., Co3O4), an electrocatalyst for CO2 reduction to CO (e.g., C3N4), and an electrocatalyst which triggers both reaction channels (e.g., CDots by trapping H+ and e− and generating H•). A schematic diagram of the Co3O4-CDots-C3N4 ternary electrocatalyst syngas generation is presented in Fig. 2. CDots first trap H+ from the solution and e− from the glassy carbon electrodes and combine them to form H•. The two other catalyst components compete for these H• species. They may diffuse to C3N4 (the CO2 reduction catalyst due to its different N active sites19,20,21) and reduce CO2 to CO. Alternatively they may diffuse to Co3O4 (the HER catalyst) and produce H2. We thus suggest that CDots are the generation site of H•, C3N4 is the generation site of CO, and Co3O4 is the generation site of H2. Syngas with tunable H2/CO ratio can be thus achieved by balancing the CO2 reduction channel and the HER channel. The different catalyst components have additional functions. Introduction of CDots to the composites enhances the adsorption of both CO2 and H+. Incorporation of CDots in the electrocatalyst improves its conductivity (Supplementary Fig. 7). C3N4 serves as the highly porous substrate in which the CDots and Co3O4 nanoparticles are incorporated. C3N4 thus provides a large surface area which enhances the reaction activity and also guarantees a proximity between the different generation sites (CDots for H• generation, C3N4 for CO generation, and Co3O4 for H2 generation). This proximity is essential for an efficient reaction. Experimental evidence substantiating this proposed mechanism will be given in the following sections.
The electrocatalytic performance of Co3O4-CDots-C3N4
We now show that the suggested design concept of the Co3O4-CDots-C3N4 indeed provides a tunable and stable production of syngas. The electrocatalytic performance of Co3O4-CDots-C3N4 for syngas production was tested in an airtight three electrodes electrochemical H-type cell. The gaseous reduction products monitored by a gas chromatography (GC) system were CO and H2 while no other reduction liquid products were found by H1 NMR. Figure 3a shows the current densities of CO (jCO, red trace) and H2 (jH2, black trace) vs. different potentials acquired in a CO2-saturated 0.5 M KHCO3 (pH = 7.2) solution. The curves indicate a significant generation of both CO and H2 but the jCO/jH2 ratio varies with the potential applied. The jCO curve shows that the CO2 electrocatalytic reduction starts at an initial potential of −0.28 V vs. RHE (all potentials reported throughout this paper are with respect to this reference). This operating voltage corresponds to a low overpotential of 0.17 V (the equilibrium potential of CO2/CO is at −0.11 V). Up to −0.45 V, both jCO and jH2 are very small and above −0.45 V both currents increase. Between −0.45 and −0.75 V, jCO is larger than jH2, indicating that the CO2 reduction reaction predominates. For potentials larger than −0.75 V, the HER channel becomes more significant than the CO2 reduction reaction. The increase of jH2 with potential is about constant (the jH2 curve is a straight line). In contrast, the slope of the jCO curve decreases with the potential applied. We attribute this behavior to the consumption of CO2 at high CO generation rates due to its low solubility or slow mass transfer. Indeed (inset of Fig. 3a), bubbling CO2 into the electrolyte increases jCO but the increase rate of jCO with potential is still lower than that of jH2. Figure 3b shows the FEs of CO and H2 generation vs. the applied cathodic potentials. The FE of CO generation increases with the applied potential and reaches a maximum of 89% at −0.6 V. For higher potential, it decreases to ~40% with increasing potential to −0.8 V, most probably due to the limited mass transport of CO2 in the electrolyte. In contrast, the FE of H2 generation is about 86% for a low applied potential, decreases with potential to about 5% at −0.6 V and then rises again reaching 55% at −0.8 V. The total FE of CO and H2 combined reaches up to 95%. Consequently, syngas with different H2/CO ratios can be obtained by altering the applied potential as shown in Fig. 3c. The volume ratio between H2 and CO can be tuned from 4:1 to 0.07:1 and syngas with H2/CO = 1:1 can directly be generated at the potentials of −0.45 V (total current density is 0.25 mA/cm2) and −0.75 V (total current density is 5.78 mA/cm2). Note that the catalyst mass activity for syngas generation reaches a high value of 10 A/gcatalyst at a potential of −0.6 V. The stability of the Co3O4-CDots-C3N4 catalyst for electrocatalytic CO2 reduction and HER (Fig. 3d) was studied as well. A negligible decay in current density was observed when operating the system for 100 h at the applied potential of both −0.6 V (Fig. 3dΙ, black trace, CO2 reduction reaction predominates) and −0.75 V (Fig. 3dΙ, red trace, HER and CO2 reduction are equal). It reveals that the Co3O4-CDots-C3N4 maintains high electrocatalytic stability during the 100-h test. An additional experiment validated the stability of both the current density and the FE for 30 h (it is very likely that the FE is stable for 100 h as well).
Investigation of the catalytic mechanisms
A series of controlled experiments were carried out to validate and further understand the catalytic mechanism operating with the Co3O4-CDots-C3N4 proposed in a previous section (Fig. 2). Issues investigated included: the generation sites of H•, CO, and H2; the role of H+; the roles of CDots, C3N4, and Co3O4; the role of the proximity between the CDots, C3N4, and Co3O4; ways in which the relative significance of the different reaction channels and the resulting CO/H2 volume ratio can be manipulated; other HER-CDots-C3N4 systems with different HER materials. We first studied the linear sweep voltammetry (LSV) curves of different combinations of the three components comprising the ternary Co3O4-CDots-C3N4. The comparison of the different LSV curves (Fig. 4a) and the complementary study of the CO and H2 composition generated during the electrocatalytic processes catalyzed by the different components (Fig. 4b) enabled a clear determination of both the generation sites of the CO and H2 and the role of the three basic components used. We first discuss the LSVs showing the apparent current density (current density per geometrical area), which is also equivalent to the mass activity of the catalysts (current per gram catalyst) as shown in Fig. 4a. Further on, we also give the corresponding turn on frequencies (TOFs) and the real current densities derived from the real surface areas of the components of the different combinations of catalysts in Fig. 4a. LSVs of C3N4 (Fig. 4a, curve 1, black trace, morphology of C3N4 is shown in Supplementary Fig. 8a), CDots (Fig. 4a, curve 2, cyan trace, morphology of CDots is shown in Supplementary Fig. 1), Co3O4 (curve 3, gray trace, morphology of Co3O4 is shown in Supplementary Fig. 8b), CDots-C3N4 (curve 4, blue trace, morphology of CDots-C3N4 is shown in Supplementary Fig. 8c), Co3O4-C3N4 (curve 5, green trace, morphology of Co3O4-C3N4 is shown in Supplementary Fig. 8d), Co3O4-CDots (curve 6, purple trace, morphology of Co3O4-CDots is shown in Supplementary Fig. 8e), and Co3O4-CDots-C3N4 (curve 7, red trace) were performed using a CO2-saturated 0.5 M KHCO3 (pH = 7.2) solution (Fig. 4a). The LSVs of C3N4 (curve 1, black trace), CDots (curve 2, cyan trace), Co3O4 (curve 3, gray trace), CDots-C3N4 (curve 4, blue trace), and Co3O4-C3N4 (curve 5, green trace) show poor electrocatalytic performances evident by their low current densities and high onset potentials (Fig. 4a). Only the Co3O4-CDots (curve 6) and the Co3O4-CDots-C3N4 (curve 7) exhibit high current densities and low onset potentials (Fig. 4a). CDots generate H2 only, but not CO (Fig. 4b). We explain it by their trapping and stabilizing H•, which may generate a small amount of H2 even in the absence of a HER catalyst. The LSV of Co3O4 (curve 3, gray trace) shows a relatively small activity producing only H2 (Fig. 4b). In comparison, Co3O4-CDots (curve 6, purple trace) produces a much larger amount of H2 compared to pure CDots or pure Co3O4. We attribute it to the effect of the CDots, which greatly enhance the electrocatalytic performance of Co3O4 by providing H•, which is necessary for generation of H2. We thus conclude that Co3O4 is the H2 generation site while CDots are the generation site of H• and both are needed for a large generation rate of H2. Now, we prove that the generation site of CO is C3N4. Figures 3a, b and 4a, b indicate that significant amounts of CO are generated by either the CDots-C3N4 or the Co3O4-CDots-C3N4. Co3O4 was shown to generate H2 only, which leaves CDots-C3N4 as the producer of CO. CDots per se produce only H2 as previously discussed. They however are essential for CO generation by CDots-C3N4 since they significantly enhance the CO2 adsorption, adsorb H+, and stabilize H•. C3N4 per se catalyzes only a very small (negligible) current (Fig. 4a, curve 1) so that the very little amount of gas produced by pure C3N4 (Fig. 4b) contains more H2 than CO. The supply of H• by the CDots is necessary to promote the generation of CO in C3N4. Figure 4a shows that each catalyst component adds to the activity by enhancing one of the three reactions (H• generation, H2 generation, and CO generation) but the complete three components composite Co3O4-CDots-C3N4 is necessary for intense generation of syngas. We further studied the effect of the type of mixing of the different catalyst components on the catalytic activity. The LSVs of physical mixtures of catalysts (Co3O4 + C3N4, CDots + C3N4, and Co3O4 + CDots + C3N4) were compared to those of composites of the same components chemically blended (Co3O4-C3N4, CDots-C3N4, and Co3O4-CDots-C3N4). The chemically prepared composites had much larger activities (current densities) than their corresponding physical mixtures (Supplementary Fig. 9). We attribute the large activity of the chemically prepared composites to the close proximity between the different catalysts (active sites) in the composite materials. In contrast, physical mixing does not provide such a proximity so that the large distance between the active sites hinders the activity of the physically mixed catalysts.
The comparison between the LSVs of the different catalyst components and their composites (Fig. 4a) should be done carefully. The catalyst areal density was kept the same (0.127 mg/cm2) for all (3 mm in diameter) electrodes. Since the composition of the Co3O4-CDots-C3N4 was 6 wt% Co3O4, 1 wt% CDots, and 93 wt% C3N4, it follows that the amount of CDots or Co3O4 in the CDots, Co3O4, or Co3O4-CDots electrodes was much larger than in the C3N4 containing composite electrodes. This means that the reaction activity per catalyst mass of curves 2, 3, and 6 is very low compared to that of the Co3O4-CDots-C3N4 electrode (Fig. 4a, curve 7). Since we deal with three different catalysts with different functions, we should calculate the mass activity (current per gram catalyst) for each component separately. This was done in Supplementary Fig. 10. The reaction activity per mass of a single catalyst component (Co3O4 or CDots) of Co3O4-CDots (Fig. 4a, curve 6), e.g., is actually lower by more than an order of magnitude than the activity per mass of a single catalyst component of that of the Co3O4-CDots-C3N4 catalyst (Fig. 4a, curve 7) though curve 6 appears to indicate (Fig. 4a) a larger activity than curve 7 (compare Fig. 4a to Supplementary Fig. 10). Supplementary Fig. 10 clearly shows that starting from a single catalyst component, the addition of each of the two other components increases the activity of syngas production, i.e., all components are necessary for an optimized syngas generation. Similarly, we calculate the turn on frequencies (TOFs) of the catalysts compositions of Fig. 4a. TOF = (number of reacted electrons per time/number of catalyst active sites). We approximate the number of active sites by the number of catalyst atoms. Since we have three different catalyst components with three different functions, we calculate the TOFs per each catalyst component (CDots, Co3O4, and C3N4). Supplementary Fig. 11 shows (similar to Supplementary Fig. 10) that the addition of each single component increases the syngas production activity and all components are necessary for the optimal performance. We further measure the BET and the electrochemical real surface areas of the catalyst components of each combination shown in curves 1–7 (Supplementary Tables 3–4). We use the electrochemical surface areas (ECSAs) to calculate the real current densities related to the specific catalyst components that constitute the seven combinations shown in curves 1–7. Supplementary Fig. 12 shows that similar to the mass activities and the TOFs curves, the real current densities of one component increase with the addition of a second component and are the largest when all three catalyst components combine to a three-component catalyst composite. Figure 3 shows that the ternary concept design indeed yields an optimal performance of its different components balancing between CO2 reduction and HER. This explains the high value of 10 A/gcatalyst obtained at −0.6 V. The CO2 and H+ adsorption measurements (Supplementary Figs. 5 and 6) of the individual different catalyst components (CDots, Co3O4, and C3N4) and their composites (CDots-C3N4, Co3O4-C3N4, CDots-Co3O4, and CO3O4-CDots-C3N4) reveal another effect which improves the ternary composite catalyst performance. The synergism of the three components acting simultaneously enhances the adsorption of the ternary composite by a factor of 2–3 with respect to the adsorption of the individual components. Real surface area measurements (BET and electrochemical) of the catalyst compositions (Supplementary Table 3) indicate that the inclusion of the C3N4 component results in a high surface area ((Sreal/Sgeometrical) = 200 for BET and 30 for electrochemical) while it is lower by an order of magnitude for the nanoparticle catalysts (CDots, Co3O4, or CDots-Co3O4). The dispersion of the catalysts nanoparticles on the C3N4 surface seems to explain the synergistic adsorption behavior.
To study the role of H+ in syngas generation, we investigated the electrocatalytic performance of Co3O4-CDots-C3N4 in aqueous solutions with the same pH value (HER in phosphate buffer solution, pH = 7.2; syngas reaction in CO2-saturated 0.5 M KHCO3 solution, pH = 7.2) and in an ionic liquid (without H+). The current density–time curves of pure HER and syngas reactions at the potential of −0.6 V (at −0.6 V the syngas reaction produces ~90% CO) are shown in Fig. 4c. A stable current density (~−1.21 mA/cm2) for syngas is observed at −0.6 V while −1.25 mA/cm2 for HER is reached at the same potential. Then, ionic liquid was used as the electrolyte solution to eliminate H+. The electrocatalytic performance of Co3O4-CDots-C3N4 was tested at the potential of −0.6 V in a CO2-saturated MeCN solution containing 0.5 M [BMIM]PF6. The current density is only about 0.61 mA/cm2 in a CO2-saturated ionic liquid (Fig. 4d), which is much lower than that obtained for CO2 reduction in CO2-saturated KHCO3 solution (−1.21 mA/cm2). Notably, no gaseous products of CO reduction could be detected suggesting that H+ plays an important role for CO2 reduction to CO in the present Co3O4-CDots-C3N4 catalyst system.
It was shown that the H2/CO volume ratio obtained using the Co3O4-CDots-C3N4 catalyst system is determined by the balance between the HER channel and the CO2 reduction channel to CO. This ratio was tuned by modifying the potential (Fig. 3). Another plausible tuning method is the decrease of the amount of the HER catalyst component (Co3O4) thus reducing the HER activity and increasing the CO2 reduction activity. We therefore measured the electrocatalytic activities of Co3O4-CDots-C3N4 produced using lower amounts of Co (0.04 and 0.02 M Co3O4 loadings with respect to the standard one of the present work (0.05 M)). Figures 3a and 4e, f show the jH2 and jCO obtained at different applied potentials for 0.05, 0.04, and 0.02 M Co3O4-CDots-C3N4, respectively. It is obvious that the CO/H2 ratio was strongly affected by the change of the Co3O4 HER catalyst amount.
Finally, the design concept of the HER-CDots-C3N4 ternary catalyst was applied for three additional HER active electrocatalysts: Pt, MoS2, and Au to form Pt-CDots-C3N4, MoS2-CDots-C3N4, and Au-CDots-C3N4. The structural characterization of these catalysts is given in Supplementary Fig. 13. The electrocatalytic activity of these three ternary composite catalysts for producing syngas were tested under the same conditions as for Co3O4-CDots-C3N4. Applying the Pt-CDots-C3N4 electrocatalyst (Fig. 5a), only hydrogen was detected in the gas phase and no reduction products from CO2 reduction were observed. Pt is considered as the most efficient electrocatalyst to facilitate HER36. The application of Pt-CDots-C3N4 as a catalyst shifts the balance between HER and CO2 reduction toward H2 generation, increases the intensity of the efficient HER channel and completely suppresses the CO2 reduction to CO. The MoS2-CDots-C3N4 electrocatalyst is still sufficiently HER active to produce close to 90% jH2, but a small amount of jCO (a few percent of the total current) is nevertheless observed. The FEs of CO are no more than 10% for the MoS2-CDots-C3N4 (inset of Fig. 5b). The Au-CDots-C3N4 electrocatalyst (Fig. 5c) still exhibits a higher HER activity than the CO2 reduction activity, but the amount of CO is much larger than for the MoS2-CDots-C3N4 catalyst (the FE of CO production reaches 25%, inset of Fig. 5c). It can be thus concluded that the concept of the ternary HER-CDots-C3N4 is general and valid for HER catalysts different than Co3O4. Achievement of a relatively high amount of CO/H2 however requires the application of a HER catalyst with only a medium activity pushing the balance between HER and CO2 reduction toward CO2 reduction. The smaller the HER activity, the larger relative amount of CO is obtainable. This conclusion was directly checked by reducing the amount of Au in the ternary Au-CDots-C3N4 electrocatalyst. The electrochemical tests (Fig. 5d) show that decreasing the amount of Au by a factor of two enabled the generation of syngas with a CO to H2 volume ratio larger than one. A striking property of the h-Au-CDots-C3N4 catalyst is its extremely high mass activity for producing syngas, i.e., >700 A/gAu for CO production and >700 A/gAu H2 production (~1500 A/gAu for the total current) at −0.7 V for the catalyst shown in Fig. 5d. This activity is two orders of magnitude larger than previously reported for efficient Au electrodes for CO production17, 18. This high mass activity allows a reduction of the electrode cost when precious catalysts (e.g., Au) are applied.
The design concept of the HER-CDots-C3N4 EC catalyst for syngas generation was introduced and its electrocatalytic performance for syngas production in aqueous solutions was studied, applying Co3O4, MoS2, Au, and Pt as the HER catalyst component. The Co3O4-CDots-C3N4 electrocatalyst was found the most efficient for syngas production among the composite combinations investigated. The Co3O4-CDots-C3N4 is capable of controlling the balance between the HER channel and CO2 reduction channel. The Co3O4-CDots-C3N4 initiates the reaction of CO2 reduction to CO in aqueous solutions at a low overpotential (0.17 V) while the total current density reaches up to 15 mA/cm2 at a potential of −1.0 V. The mass activity of the Co3O4-CDots-C3N4 is ~10 A/gcatalyst at −0.6 V when the total mass of the catalyst is considered and 1–2 orders of magnitude larger when the mass of the HER catalyst is considered (which is ~0.5–5% of that of the total catalyst weight). The Co3O4-CDots-C3N4 induces high FEs (95%) and is characterized by a stable production of syngas (over 100 h). Notably, the H2/CO ratio of syngas produced applying Co3O4-CDots-C3N4 is tunable from 0.07:1 to 4:1 by controlling the applied potential. The H2/CO may be also tuned by varying the amount of Co3O4 in the Co3O4-CDots-C3N4.
Dedicated experiments validated the catalyst design concept and provided additional insight to the syngas generation processes. C3N4 and Co3O4 are the activity sites for CO2 reduction reaction and HER, respectively. CDots are the generation site of H• needed to trigger both the reduction of CO2 to CO and the HER. The three-component catalyst concept is a general one and may be applied to a host of other materials. The versatility of the three components composite design may open a powerful pathway for the development of high-performance catalysts for syngas production as well as for other chemicals generation. Such an efficient and cost-effective electrocatalytic system has a high potential to be employed for the large-scaled production of syngas and controlled mixtures of other chemicals from CO2.
Transmission electron microscopy (TEM), high-resolution transmission electron microscopy (HRTEM), and scanning TEM (STEM) images were obtained using a FEI/Philips Tecnai G2 F20 TWIN transmission electron microscope. The energy dispersive X-ray spectroscopy (EDS) analyses were taken on a FEI-quanta 200 scanning electron microscope with an acceleration voltage of 20 kV. The crystal structure of the resultant products was characterized by X-ray diffraction (XRD) using an X’Pert-ProMPD (Holland) D/max-γAX-ray diffractometer with Cu Kα radiation (λ = 0.154178 nm). X-ray photoelectron spectroscopy (XPS) was obtained by using a KRATOS Axis ultra-DLD X-ray photoelectron spectrometer with a monochromatized Mg Kα X-ray source (hν = 1283.3 eV). The electrocatalysis reactions were tested by a Model CHI 660C workstation (CH Instruments, Chenhua, Shanghai, China). The electrochemical impedance spectroscopy (EIS) measurements were obtained applying a CHI 832 electrochemical instrument (CHI Inc., USA).
KHCO3 (99.7%) and Nafion perfluorinated resin solution (5 wt.%) were purchased from Sigma-Aldrich; hydrogen (99.999%), nitrogen (99.999%), and carbon dioxide (99.999%) were purchased from Airgas; Nafion®212 membrane was purchased from Dupont; Toray Carbon Paper (TGP-H-60) was purchased from Alfa Aesar; All chemicals were purchased from Sigma-Aldrich unless specifically stated. Milli-Q ultrapure water (Millipore, ≥18 MΩ/cm) was used throughout the work.
Fabrication of electrocatalysts
CDots were synthesized by our previously reported electrochemical etching method34. After 30-days reaction, a dark yellow solution containing CDots was formed in the reaction cell. It was then purified and concentrated to form a CDots solution of 3 mg/mL. For C3N4 fabrication, 10 g of melamine powder was put into an alumina crucible with a cover and then heated to 550 °C at a rate of 0.5 °C per min in a muffle furnace and maintained at this temperature for 3 h. The yellow powder (C3N4) was obtained after cooling down to room temperature. Co3O4 NPs was synthesized by hydrothermal method. Twenty mL Co(NO3)2 (0.01 M) solution was added into an alumina crucible with a cover and then heated to 550 °C at a rate of 0.5 °C per min in a muffle furnace and maintained at this temperature for 3 h. The black powder (Co3O4 NPs) was obtained after cooling down to room temperature. For preparation of Co3O4-C3N4, CDots-C3N4 or Co3O4-CDots-C3N4, 10 g of melamine powder was mixed with 10 mL solution containing 0.05 M Co(NO3)2, CDots, CDots and 0.05 M Co(NO3)2, respectively. Then, the mixture was put into an alumina crucible with a cover and heated to 550 °C at a rate of 0.5 °C per min in a muffle furnace and maintained at this temperature for 3 h. For preparation of Co3O4-CDots, 10 mL solution containing 0.05 M Co(NO3)2 and CDots (3 mg/mL) was put into an alumina crucible with a cover and heated to 550 °C at a rate of 0.5 °C per min in a muffle furnace and maintained at this temperature for 3 h.
Synthesis of Pt-CDots-C3N4
An aliquot of 0.3 g CDots-C3N4 (obtained by heating melamine at 550 °C for 3 h) was added into 10 mL H2PtCl6 (2 mM) aqueous solution and stirred for 12 h. After centrifuging, the precipitate was irradiated by UV light for 10 h. The resulting product was collected by centrifugation and dried in a vacuum at 60 °C for 12 h.
Synthesis of MoS2-CDots-C3N4
An aliquot of 0.3 g CDots-C3N4 (obtained by heating melamine at 550 °C for 3 h) was added into 15 mL aqueous solution containing Na2MoO4 (0.0625 g) and L-cysteine (0.1 g). The mixed solution was stirred 3 min. After that, the mixture was poured into a Teflon-lined stainless steel autoclave, and heated at 180 °C for 24 h. After the autoclave was cooled down to room temperature, the resulting black sediments were collected by centrifugation (10,000 rpm, 10 min) and washed with deionized water and ethanol for several times, and then dried in a vacuum oven at 80 °C for 12 h37.
Synthesis of Au-CDots-C3N4 and h-Au-CDots-C3N4
An aliquot of 0.3 g CDots-C3N4 (obtained by heating melamine at 550 °C for 3 h) was added into 10 mL HAuCl4 (2 mM or 1 mM) aqueous solution and stirred for 12 h. After centrifuging, the precipitate was irradiated by UV light for 2 h. The resulting product was collected by centrifugation and dried in vacuum at 60 °C for 12 h38.
Electrocatalysis activity test
Electrocatalysis activity test experiments were performed using a standard three-electrode configuration. A platinum wire was used as an auxiliary electrode and a saturated calomel electrode (SCE) was used as a reference electrode. The working electrode was either a catalyst-modified carbon fiber paper electrode (CFPE for short, 0.7 cm × 0.7 cm), or a catalyst-modified glassy carbon disk electrode (GCE for short, 3.0 mm diameter). For product analysis and constant-potential electrolysis experiment, the CFPE working electrode was a catalyst-modified carbon fiber paper electrode (0.7 cm × 0.7 cm). The preparation of the CFPE working electrode is as follows. An aliquot of 1.3 mg of electrocatalyst was ground with 0.1 mg polyvinylidene fluoride (PVDF) with a few drops of 1-methyl-2-pyrrolidone (MP) added to the produced mixture. The mixture was added into 10 mL 0.5% Nafion solution. After sonication, 1 mL dispersed solution was dropped directly onto the two sides of a 0.7 cm × 0.7 cm carbon fiber paper (the two sides of the carbon paper were modified by the catalyst). The bulk electrolysis was performed in an airtight electrochemical H-type cell with three electrodes. H-type cell consists of two compartments (volume of each part is 115 mL) separated by a Nafion®212 anion exchange membrane with 75 mL 0.5 M KHCO3 electrolyte in each chamber and. Besides, LSV experiments were used with the catalyst-modified GCE as the working electrode. The preparation of the GCE working electrode is as follows. 6 mg electrocatalyst was added into 2 mL 0.5% Nafion solution. After sonication, 3 μL dispersed solution was dropped on GCE. The mass density of catalyst was 0.127 mg/cm2. The electrochemical tests of different catalysts combinations were performed with full loading, i.e., the mass of the composite catalyst was 9 µg (Supplementary Table 4). The electrochemical surface area (ECSA) test of single catalyst components were performed using partial loading, i.e., 1% of 9 µg CDots and 6% of 9 µg Co3O4 (Supplementary Table 4). For LSVs experiments, initially, polarization curves for the modified electrode were carried out under an inert N2 (gas) atmosphere. After this, the solution was purged with CO2 (99.999%) for at least 30 min (CO2-saturated high purity aqueous 0.5 M KHCO3) and the electrocatalytic CO2 reduction was measured.
The data that support the findings of this study within the paper and its Supplementary Information file are available from the corresponding authors on request.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
A correction to this article is available online at https://doi.org/10.1038/s41467-018-02848-2.
This work was supported by the Collaborative Innovation Center of Suzhou Nano Science and Technology, the National Natural Science Foundation of China (51725204, 21771132, 51422207, 51572179, 21471106, and 21501126), the Natural Science Foundation of Jiangsu Province (BK20161216), and a project funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).