Polyethylenimine promoted electrocatalytic reduction of CO₂ to CO in aqueous medium by graphene-supported amorphous molybdenum sulphide

Abstract

Efficiently and selectively converting CO₂ to value-added carbon compounds remains a major challenge in sustainable energy research. In this paper, we report the synthesis of a cost-effective catalyst, i.e. amorphous molybdenum sulphide on a polyethylenimine modified reduced graphene oxide substrate, for electrocatalytically reducing CO₂ into CO in CO₂ saturated aqueous NaHCO₃ medium with high efficiency and selectivity. The catalyst is capable of producing CO at overpotentials as low as 140 mV and reaches a maximum faradaic efficiency of 85.1% at an overpotential of 540 mV. At an overpotential of 290 mV with respect to the formation of CO, it catalyzes the formation of syngas with high stability. Detailed investigations reveal that PEI works as a co-catalyst by providing a synergetic effect with MoS_x.

---

Broader context

Electrochemical CO₂ reduction using renewable energy as a power source could potentially solve the energy and environmental problems arising from excessive use of non-renewable fossil fuels. Currently, electrocatalysts for this reaction generally suffer from high cost, low energetic efficiency and poor product selectivity. Nevertheless, molybdenum or tungsten containing formate dehydrogenases (FDH) can efficiently and selectively catalyze reduction of CO₂ to formate in Nature. Inspired by the active sites of FDH, amorphous molybdenum sulphide (MoS_x) on a polyethylenimine (PEI) modified reduced graphene oxide (rGO) substrate has been synthesized and used as a heterogeneous electrocatalyst for reduction of CO₂ to CO in an aqueous medium. The high efficiency and selectivity is accomplished by the synergetic effect of MoS_x and PEI while rGO improves electrical conductivity and stability of the catalyst.

Introduction

Energy and environmental problems are among the most challenging issues facing mankind in this century.^1–3 Despite tremendous efforts to develop renewable energy sources, the majority of energy used today is still from non-renewable, fossil fuel sources, which leads to an increase in greenhouse CO₂ levels in the atmosphere and may cause catastrophic climate changes. Electrocatalytic reduction of CO₂, which recycles CO₂ back to fuels and commodity chemicals by using renewable energy as a power source, could provide an attractive solution to this issue.⁴ However, CO₂ is very stable under environmental conditions and the hydrogen evolution reaction (HER) often prevails over CO₂ reduction in aqueous electrolytes under cathodic polarization,^5,6 making it essential to find a suitable catalyst to achieve cost-effective CO₂ reduction with high efficiency and selectivity.

The past several decades have witnessed reports of a number of homogeneous and heterogeneous catalysts for electrocatalytic CO₂ reduction; these mainly include metals,⁵ metal oxides,^7–9 metal complexes^10–13 and nitrogen-containing carbon materials.^14–17 Among them, noble metals such as Au, Ag and Pd have attracted particular interest for they can selectively convert CO₂ to CO at relatively low overpotentials.^5,18 Moreover, nanostructured noble metals have been developed recently to enhance their performance.^19–27 For example, oxide-derived Au nanoparticles exhibit highly selective CO₂ reduction to CO in aqueous solution at a lower overpotential than a polycrystalline bulk Au electrode.²⁰ However, precious metals are hampered by their low abundance and high cost, which restrict their practical, large-scale application. Thus, searching for abundant and efficient electrocatalysts with selectivity towards CO, or capable of producing syngas that could be readily utilized in a Fischer–Tropsch process, is vital for commercializing CO₂ electroreduction technologies.²⁸ Although less expensive materials such as Cu, Sn, Bi and their oxides have demonstrated promise for reduction of CO₂ to hydrocarbons, formic acid and CO,^29–32 these systems generally show low activities and/or selectivity for CO production or need non-aqueous solvents which may limit their practical applications.

In Nature, formate dehydrogenases (FDH), which contain molybdenum or tungsten centers coordinated by a selenocysteine residue and amino acid residues, can catalyze the oxidation of formate to CO₂.^33,34 Hirst and co-workers^35,36 showed that they could efficiently and specifically catalyze the reversible transformation between CO₂ and formate under mild conditions. This inspires us to explore new catalysts resembling the FDH active site for electroreduction of CO₂. Molybdenum disulphide (MoS₂), a two-dimensional transition metal dichalcogenide, has sparked tremendous research interest as a substitute for Pt for effectively catalysing the HER in acidic media.^37–40 MoS₂ consists of a planar triangular lattice of Mo atoms sandwiched between two planes of S atoms and each Mo(IV) is coordinated to six S ligands, while each S center is connected to three Mo atoms. From a biomimic point of view, MoS₂ shares a similar structure with the active center of FDH, making it potentially an electrocatalyst for CO₂ reduction. Recently, Nørskov et al.⁴¹ demonstrated theoretically that MoS₂ could possibly be an electrocatalyst for CO₂ reduction and this was subsequently proven experimentally in ionic liquid (IL) medium with water as a proton source.⁴² In the present work we demonstrate that amorphous molybdenum sulphide (MoS_x) on a polyethylenimine (PEI) modified reduced graphene oxide (rGO) substrate can be a highly efficient and selective electrocatalyst for reduction of CO₂ to CO in simple aqueous media at low overpotentials. We show that the PEI layer considerably enhances the catalytic activity of the MoS_x, presumably by suppressing HER and stabilising the CO₂˙⁻ intermediate during CO₂ reduction, as has been suggested previously for PEI on carbon nanotube materials.⁴³

Experimental

Chemicals and apparatus

NaHCO₃ (ACS grade), NaOH and NaH₂PO₄ (AR grade) were purchased from Merck. Natural graphite (crystalline, 300 mesh) was from Alfa Aesar. Polyethylenimine (PEI, M_n = 10 000, branched), polyethylene glycol (PEG, M_n = 400), polydiallyldimethylammonium chloride (PDDA, 20 wt% in water, M_w = 100 000–200 000) and (NH₄)₂MoS₄ (99.97% trace metals basis) were purchased from Sigma-Aldrich. All the chemicals were used without further purification. Graphene oxide (GO) was synthesized from natural graphite using the method described by Hummers et al.⁴⁴ Phosphate buffer used for electrodeposition of MoS_x was prepared by adding 5.0 M NaOH solution to 0.1 M NaH₂PO₄ solution to adjust the pH to 6.81. All the aqueous solutions were prepared with Milli-Q water.

Transmission electron microscopic (TEM) images were collected on a FEI Tecnai G2 T20 TWIN TEM Instrument. Scanning electron microscopic (SEM) images and energy dispersive spectrum (EDS) were recorded on a FEI Nova NanoSEM 450 FEG SEM Instrument equipped with Bruker Quantax 400 X-ray analysis system. Raman spectra were obtained using a Renishaw inVia Microscope with a 532 nm laser source and IR spectra with a Digilab (7000) Stingray Imaging FT-IR Spectrometer. X-ray diffraction (XRD) data were collected with a Philip X-ray powder diffractometer (Cu K_α radiation). Gas chromatography (GC) was performed with an Agilent 7820 A gas chromatography system equipped with a HP-plot molesieve (5A) column and a thermal conductivity detector (TCD). The carrier gas was helium (99.99%) for CO analysis while nitrogen (99.99%) was used as carrier gas for H₂ analysis. The retention times were compared with known compounds. All the electrochemical experiments were conducted on a CHI 760E electrochemical workstation (CH Instruments, Austin, Texas, USA) at room temperature (22 ± 2 °C).

Fabrication of rGO–PEI–MoS_x electrodes

10 mg graphene oxide (GO) powder was dispersed in 10 mL water and pH was adjusted to around 9.0 by adding an appropriate amount of aqueous NaOH solution (5.0 M) dropwise. After sonication for at least 1 hour, 500 μL of a 5 wt% PEI aqueous solution was then added into the GO dispersion quickly and sonicated for another 30 minutes. Afterwards, it was stirred at 60 °C for 12 hours to form the GO–PEI solution. In order to obtain a well-dispersed solution, centrifugation at 3000 rpm was applied for 20 minutes to remove any large aggregates. The procedure for fabricating the rGO–PEI–MoS_x electrode is schematically described in Fig. 1. Firstly, 4 μL of GO–PEI solution was casted onto a glassy carbon electrode (GCE, 3 mm in diameter, CH Instruments, USA) to form the GO–PEI modified GCE and dried under an infrared lamp. The as-prepared electrode was then immersed into a solution composed of 2 mM (NH₄)₂MoS₄ and 0.1 M phosphate buffer (pH 6.81) for electrodeposition of MoS_x by cycling the potential from 0.3 V to −1.3 V (vs. Ag/AgCl (3 M KCl)) at a scan rate of 50 mV s⁻¹ for 10 times according to the procedure developed by Merki et al.⁴⁵ During this process, GO is electroreduced to reduced graphene oxide (rGO), while MoS_x is electrodeposited onto the rGO. MoS_x, rGO–PEI, rGO–MoS_x, rGO–PEG–MoS_x and rGO–PDDA–MoS_x electrodes were prepared by a similar procedure. To prepare TEM samples, catalyst materials were carefully scraped off from the electrode surface and dispersed in water by sonication for 30 minutes. Afterwards, 2 μL of the dispersion was drop cast on a copper grid and dried in air.

Fig. 1 Schematic illustration of the fabrication of rGO–PEI–MoS_x modified electrodes.

Electrochemical measurements

Cyclic voltammetric (CV) measurements were carried out in 0.5 M aqueous NaHCO₃ solution saturated with N₂ or CO₂ by bubbling N₂ or CO₂ for at least 30 minutes in a conventional three-electrode cell comprising a platinum wire counter electrode, an Ag/AgCl (3 M KCl) reference electrode and a modified or unmodified GCE as the working electrode. In some experiments, a piece of high surface area carbon cloth (1 cm × 0.5 cm) modified with rGO–PEI–MoS_x was used as working electrode. For hydrogen evolution reaction (HER) catalysis, CV measurements were conducted in 0.5 M H₂SO₄, 0.5 M NaHCO₃ and 0.5 M phosphate buffer (pH 7.2) solutions, respectively, at a scan rate of 5 mV s⁻¹ with iR compensation. Electrolysis at different applied potentials was performed in a gas-tight two-compartment electrochemical cell with a glass frit as the separator. Each compartment contained 12.5 mL electrolyte and approximately 20 mL headspace. A piece of carbon cloth (1 cm × 1.5 cm) was used as counter electrode, an Ag/AgCl (3 M KCl) as reference electrode and a modified glassy carbon plate (0.4 cm × 1.5 cm) as working electrode. Before electrolysis, the cell was degassed by bubbling CO₂ gas for at least 30 min. The solutions in both compartment were stirred during electrolysis. After the electrolysis, a small fraction of the cell's head space products (200 μL) was sampled by gas-tight syringe and analysed by GC. For each electrolysis, the products were sampled and analysed twice in a very short time interval and two electrolysis experiments were run at each applied potential. The durability test was conducted by chronoamperometry at a constant potential of −0.65 V using a rGO–PEI–MoS_x modified 3 mm diameter GCE. A slightly positive CO₂ pressure was applied to prevent air from entering the electrolysis cell and to maintain the concentration of CO₂.

All potentials were converted to the reversible hydrogen electrode (RHE) reference scale using the formula: E(vs. RHE) = E(vs. Ag/AgCl) + 0.210 V + 0.059 V × pH. All the working electrodes were polished using a 0.3 μm aqueous alumina slurry on a polishing cloth, sonicated in water and acetone, rinsed with acetone, and then dried under a flow of nitrogen before use.

Results and discussion

Fourier transform infrared spectroscopic characterization of GO–PEI

Fourier transform infrared (FT-IR) spectroscopic characterization of GO and GO–PEI was undertaken to reveal the nature of the interaction between PEI and GO. As shown in Fig. 2a, these two materials exhibit distinctly different FT-IR spectra. Two additional peaks at 2841 and 2932 cm⁻¹ were found with GO–PEI, which were assigned to symmetric and asymmetric stretching modes of the CH₂ of the PEI chains.⁴⁶ Furthermore, a significant decrease of the peak at 1724 cm⁻¹ (attributed to C=O from the carboxyl group in pristine GO) and the appearance of a new band at 1632 cm⁻¹ (attributed to the C=O stretching of the primary amide) suggest that the carbonyl groups are converted to amides on reaction with the PEI.⁴⁷ This is further confirmed by the new bands observed at 1450 cm⁻¹ (C–N stretching vibration) and 1572 cm⁻¹ (N–H bending vibration) with GO–PEI.⁴⁸ This covalent interaction between PEI and GO is beneficial in maintaining the high stability of the modified electrode under long term electrolysis conditions. It is important to point out that the band at 1592 cm⁻¹ ascribed to the C=C vibration of the aromatic rings was observed in both GO and GO–PEI, implying that the sp² character in GO–PEI is preserved after the chemical conjugation,⁴⁹ which is crucial to preserve the conductivity of the fabricated electrode.

Fig. 2 (a) FT-IR spectra of GO (black line) and GO–PEI (red line). (b) Electrodeposition of rGO–PEI–MoS_x on GCE by repeating CVs for 10 cycles.

Fabrication and characterization of rGO–PEI–MoS_x

The rGO–PEI–MoS_x electrode was fabricated by electrodeposition of MoS_x onto a rGO–PEI modified electrode using the detailed procedure described in the Experimental section. The voltammogram (Fig. 2b) clearly demonstrates that a new reduction process at −0.1 V (vs. RHE, all potentials hereafter are with respect to this reference) and a new oxidation process at 0.4 V emerge and their current magnitude increases with cycling. The oxidation peak is attributed to the oxidation of [MoS₄]²⁻ to form MoS₃ and S₈ according to the following equation:^50,51

The reduction peak is assigned to the reduction of the electrogenerated S₈. In the potential region more negative than −0.25 V, reductive deposition of amorphous Mo^IVS₂ (i.e. MoS_x) takes place according to the previous electrochemical quartz crystal microbalance studies.⁵⁰

[MoS₄]²⁻ + 2H₂O + 2e⁻ → MoS₂ + 2HS⁻ + 2OH⁻

In the potential range of −0.4 V to −0.69 V, GO is electro-reduced to rGO.⁵² Thus, the deposition of MoS_x and reduction of GO occurred simultaneously in this potential region, which was confirmed by energy dispersive spectroscopy (EDS) and Raman spectroscopy (Fig. S1, ESI†). The presence of hydrophilic PEI on GO could increase the space between each GO layer and make it more easily accessible for [MoS₄]²⁻ to achieve efficient deposition of MoS_x.⁵³ Moreover, both rGO–PEI and rGO–PEI–MoS_x modified GCE show a large increase in capacitance compared to bare GCE (Fig. S2, ESI†) due to a significantly enlarged conductive surface area introduced by rGO.⁵⁴ The further enhanced capacitance of rGO–PEI–MoS_x arises from pseudo-capacitance processes associated with MoS_x.⁵⁵ To confirm the interaction between rGO and MoS_x, transmission electron microscopic (TEM) and scanning electron microscopic (SEM) images were used to characterize the morphology. As revealed in Fig. S3 (ESI†), rGO–PEI–MoS_x shows a translucent and wrinkled texture associated with flexible graphene sheets. Fig. 3a shows the magnified image and MoS_x nanoparticles of diameter 17 ± 3 nm (Fig. S4, ESI†) can be clearly seen. The fast Fourier transform (FFT) pattern indicates the amorphous nature of MoS_x nanoparticles (Fig. 3a inset), which is further confirmed by X-ray diffraction (XRD, Fig. S5, ESI†) with no characteristic peaks of any crystalline structures observed. Moreover, in the absence of GO–PEI, the MoS_x tended to aggregate into much larger particles (Fig. S6, ESI†). By contrast, as can be seen from Fig. 3b, the MoS_x nanoparticles grew uniformly on the surface of rGO–PEI allowing the exposure of more active sites of MoS_x to CO₂.

Fig. 3 TEM (a) and SEM (b) images of rGO–PEI–MoS_x. Inset of (a): Corresponding FFT pattern. Scale bar: (a) 100 nm, (b) 200 nm.

CO₂ reduction activities of rGO–PEI–MoS_x

To assess the possibility of using rGO–PEI–MoS_x for electrocatalytic reduction of CO₂, CV studies were firstly undertaken in N₂-saturated and CO₂-saturated 0.5 M aqueous NaHCO₃ solutions (pH values 8.5 and 7.2, respectively) at a scan rate of 50 mV s⁻¹ (Fig. 4a). Under both atmospheres, rapid increases of current were observed at around −0.22 V. In the more negative potential region, a broad process centred at −0.65 V appears under the CO₂ atmosphere. Similar processes were also reported in the previous studies with nitrogen-doped carbon nanofibers¹⁶ or Ag^23,56 as the catalysts and were assigned to the reduction of CO₂. Since CV cannot provide conclusive evidence about the nature of this reduction process, potentiostatic electrolysis was conducted by applying a potential in the range of −0.25 V to −0.95 V (corresponding i–t curve is depicted in Fig. S7, ESI†) and the faradaic efficiency (FE) is plotted as a function of the applied potential in Fig. 4b. The results reveal that H₂ and CO are the dominant products with a combined FE of around 100% over the whole potential range. Therefore, the liquid products, if any, were not analysed. The selectivity towards H₂ and CO is strongly dependent on the applied potential. At an applied potential of −0.25 V (corresponding to an overpotential of 140 mV for the formation of CO, since the equilibrium potential is −0.11 V vs. RHE;⁵ all the overpotentials hereafter are reported with respect to this equilibrium potential), CO with a FE of ∼5% was detected while H₂ is the dominant product. This overpotential for CO formation is significantly lower than that of polycrystalline Au or Ag⁵ and as low as that achieved in oxide-derived Au nanoparticles under the same conditions.²⁰ FE of CO formation overtakes H₂ evolution when the potential becomes more negative and it reaches a maximum of 85.1% at an applied potential of −0.65 V (overpotential of 540 mV), which exceeds the highest FE of the polycrystalline Ag electrode reported in the same electrolyte medium.⁵ H₂ regains dominance in the more negative potential region. This phenomenon was also observed in the case of Ag and Au catalysts in previous reports^5,27,56,57 and has been explained as follows: in the potential region of low CO₂/CO overpotentials, HER is dominant since it is thermodynamically and kinetically more favourable.^5,58 In the potential region of intermediate CO₂/CO overpotentials (i.e. −0.35 to −0.75 V vs. RHE), CO₂ reduction occurs to form CO, and HER is hindered due to the slow desorption of CO from the active sites.⁵⁹ In the potential region of high CO₂/CO overpotentials, kinetics of both CO₂ reduction and HER increase. However, HER regains dominance since H₂ evolution is the kinetically favoured process and H⁺ is much more readily available in the aqueous HCO₃⁻ medium. Therefore, CO₂ reduction process becomes relatively slower due to either kinetic or mass transport limitations.^21,42,56,57

Fig. 4 (a) CVs of rGO–PEI–MoS_x modified GCE in N₂-saturated (black curve) and CO₂-saturated (red curve) 0.5 M aqueous NaHCO₃ solution. Scan rate was 50 mV s⁻¹. Inset: Structure of PEI. (b) Faradaic efficiency for CO (red bars) and H₂ (blue bars) as a function of potential. Error bar associated with each value is also provided.

The source of CO is an important issue that needed to be confirmed. In our study, no CO was detected when the electrolysis was undertaken in a N₂ atmosphere, which confirms that the direct source of CO is not HCO₃⁻. Another possible source could be the catalyst materials, which include support electrode, rGO and PEI. ¹³C labelled CO₂ and NaHCO₃ could be used to provide direct evidence. However, the formal potential is 0.52 V vs. RHE for the reaction CO(g) + 2H⁺ + 2e⁻ ↔ C(s) + H₂O, which is much more positive than the applied potential (−0.25 to −0.95 V vs. RHE) used in our studies. Therefore, direct oxidation of C to generate CO is not a thermodynamically favorable process. Under the mildly reducing electrochemical conditions involved in our work, extensive oxidation of rGO and PEI to produce the amount of CO detected is also very unlikely. Moreover, we also measured the weight difference of the rGO–PEI–MoS_x modified glassy carbon plate electrode before and after electrolysis (see Note S1 in the ESI† for detail) and found that the weight loss is far less than the amount needed to supply the CO produced during electrolysis. Even carbon with the same mass as all of the modified catalyst materials cannot produce sufficient CO at 100% conversion yield. Therefore, neither rGO nor PEI can be the major source of CO.

The intrinsic catalytic activity toward CO₂ reduction for each MoS_x active site is indicated by turnover frequency (TOF), calculated using the following equation:

where I_total is the current (in A) in the CV measurement, FE is faradaic efficiency for CO formation, n is active sites number (in mol) and F is Faraday's constant (in C mol⁻¹). The factor 2 arises from the fact that two electrons are required to form one CO molecule from one CO₂ molecule. Taking I_total = 3.4 × 10⁻⁴ A, FE = 85.1%, n = 6.18 × 10⁻¹⁰ mol (estimated from voltammetric measurements, Fig. S8, ESI†), the TOF at the applied potential of −0.65 V is calculated to be 2.4 s⁻¹. Such a large TOF value implies that this catalyst could potentially be used for practical applications when it is immobilized on a porous support with a large surface area. Indeed, a current density as high as 55 mA cm⁻² at −0.65 V could be reached when the rGO–PEI–MoS_x composite catalyst was immobilized on a carbon cloth electrode (Fig. S9, ESI†).

It is also worth noting that the molar ratio of H₂ and CO is about 2 : 1, which is the composition of syngas, when the applied potential is −0.4 V (equivalent to 290 mV overpotential for CO formation) (Fig. 5). Thus, this produced gas mixture could readily serve as the source gas for the Fischer–Tropsch process, an important gas to liquids technology for generating methanol, synthetic petroleum and other liquid fuels.²⁸ More importantly, the FE of H₂ and CO remains constant at this potential over at least 3 hours of electrolysis (Fig. 5), showing great promise as a practical catalyst for this process.

Fig. 5 Amount and faradaic efficiency of H₂ (circles) and CO (squares). Potentiostatic electrolysis at −0.4 V in CO₂-saturated 0.5 M aqueous NaHCO₃ solution. Measurement interval: 0.5 h.

In addition, the durability of rGO–PEI–MoS_x was also studied by chronoamperometry at −0.65 V, which is the optimum potential with respect to maximum CO selectivity. As shown in Fig. S10 (ESI†), the current remains stable in the time scale of more than 16 hours (60 000 s). Initial decay of current, which cannot be completely attributed to the double layer charging process, is mainly due to the adsorption of gaseous products. This current was largely recovered when the electrode was left under open circuit condition for 500 seconds before applying the same potential for electrolysis, suggesting long-term durability of the rGO–PEI–MoS_x. SEM images and EDS of the electrode surface after 60 000 s are shown in Fig. S11 (ESI†). Compared with the SEM image taken before the electrolysis (Fig. 3b), little difference in morphology was observed after the long-term electrolysis experiment. The elemental analysis reveals the preservation of C, N, Mo and S elements, which also indicates the stability of the material. In HER catalysis, amorphous MoS_x often suffers from activity decay due to mechanical loss of catalyst materials or poisoning by surface adsorbates from the electrolyte or reference electrode rather than inherent property changes.^60,61 Chorkendorff et al. showed that catalyst support, such as carbon fibers, could play an important role in improving the mechanical stability of the amorphous MoS_x catalyst.⁶² Therefore, in our case, the presence of the rGO support enhances the mechanical stability of the electrodes under catalytic turnover conditions, as suggested in regard to other types of support.^55,62–64

Possible origins of the activity of rGO–PEI–MoS_x for CO₂ reduction

To reveal the possible origins of the activity of rGO–PEI–MoS_x for CO₂ reduction, the CV measurements were undertaken using rGO–PEI, rGO–MoS_x and rGO–PEI–MoS_x modified GCEs, respectively. As shown in Fig. 6a, in the absence of MoS_x, rGO–PEI modified GCE demonstrates a reduction onset potential that is about 500 mV more negative than the other two electrodes in both N₂-saturated and CO₂-saturated 0.5 M aqueous NaHCO₃ solutions which indicates that MoS_x is the effective catalyst towards the reduction reactions observed here. Electrolysis results (Fig. 6b) show only H₂ was detected with a FE of ∼93% at −0.85 V. The rGO–MoS_x modified GCE presents similar CV curves in N₂ and CO₂ atmospheres and potentiostatic electrolysis at −0.65 V reveals that H₂ is the only product with a FE of ∼95%. Interestingly, although the incorporation of PEI into rGO–MoS_x system results in a decrease in current density under either N₂ or CO₂ atmospheres, which will be discussed later, there is a significant difference in voltammetric characteristics and a considerably higher FE for CO at −0.65 V. Hence, a synergetic effect of PEI and MoS_x is responsible for the catalysis of CO₂ reduction.

Fig. 6 (a) CVs of rGO–PEI, rGO–MoS_x, rGO–PEI–MoS_x modified GCE in N₂-saturated and CO₂-saturated 0.5 M aqueous NaHCO₃ solution. Scan rate was 50 mV s⁻¹. (b) Faradaic efficiency for CO (red bar) and H₂ (blue bars) of rGO–PEI, rGO–MoS_x, rGO–PEI–MoS_x, potentiostatic electrolysis at −0.65 V in CO₂-saturated 0.5 M aqueous NaHCO₃ solution. CVs of rGO–PDDA–MoS_x (c) and rGO–PEG–MoS_x (d) modified GCE in N₂-saturated and CO₂-saturated 0.5 M aqueous NaHCO₃ solution. Scan rate was 50 mV s⁻¹. Inset of (c) and (d) shows the structure of PDDA and PEG, respectively.

Crystalline MoS₂ has demonstrated its capability for effectively electroreducing CO₂ in an imidazolium based IL mixed with water as proton source.⁴² Similar observations have also been made previously by the same group in their study of electrocatalytic reduction of CO₂ at an Ag electrode in this medium.⁶⁵ They found that the imidazolium cation could act as a cocatalyst via the formation of an IL–CO₂ complex to reduce the activation energy barrier. Nørskov et al. recently predicted theoretically by DFT calculations that the edge sites of crystalline MoS₂ could be the catalysing sites for CO₂ reduction.⁴¹ They claimed that the bridging S atom at the edges of MoS₂ could selectively bind the intermediate HCOO over the product CO, which resulted in a deviation from the transition-metal scaling relationship of intermediates in the CO₂ reduction process and therefore significantly improved CO₂ reduction activities over transition-metal catalysts. Although there are no well-defined edges in amorphous MoS_x, there are many structurally and coordinatively unsaturated Mo and S atoms,^45,61,66 which are equivalent to the edge sites in the crystal from a catalysis point of view. On this basis, in principle, the amorphous MoS_x used in our study may be expected to be a better electrocatalyst for CO₂ reduction due to its abundance of such active sites.

However, in the absence of PEI, rGO–MoS_x modified GCE only showed activity towards HER, which infers the crucial role of PEI for CO₂ reduction. To investigate its role, PEI was replaced with other polymers, poly diallyldimethylammonium chloride (PDDA) or polyethylene glycol (PEG) for the fabrication of rGO–polymer–MoS_x modified electrodes. Fig. 6c and d present the CVs of rGO–PDDA–MoS_x and rGO–PEG–MoS_x modified GCE in N₂-saturated and CO₂-saturated 0.5 M aqueous NaHCO₃ solution at the scan rate of 50 mV s⁻¹, respectively. Again, both electrodes show similar CV curves in N₂ and CO₂ atmospheres with the same onset potential and only a slight decrease of current density in the present of CO₂ and further potentiostatic electrolysis demonstrated H₂ as the only product. Thus it is clear that the PEI possesses unique properties that produce synergetic effects with MoS_x for CO₂ reduction. Two factors may account for this effect:

(1) PEI can suppress HER and thus enhance the competing CO₂ reduction. To probe this further, the HER catalysis properties of rGO–MoS_x and rGO–PEI–MoS_x were investigated in aqueous media, containing 0.5 M H₂SO₄, 0.5 M bicarbonate (pH 8.5) or 0.5 M phosphate buffer (pH 7.2, at which catalytic CO₂ reduction was observed in our study). Results obtained are shown in Fig. S12 (ESI†). The kinetics of HER reaction were retarded in the presence of PEI in all media, as suggested by a larger overpotential. This effect of PEI on the HER is most likely due its amine groups. Rezaei et al. have shown that amine groups in ionic liquid electrolytes can inhibit hydrogen formation in aqueous H₂SO₄ solution.⁶⁷ Rosen et al. also found this effect when they used an ionic liquid as a co-catalyst for CO₂ reduction in mixed solvents containing water and [EMIM][BF₄].⁶⁸

(2) PEI may stabilize the intermediate species in the CO₂ reduction process. Tafel slopes of ∼120 mV dec⁻¹ have often been reported in the previous studies in aqueous media at metallic electrodes since the one-electron reduction of CO₂ to form CO₂˙⁻ is the rate determining step.⁵ By contrast, a much smaller value of 74 mV dec⁻¹ was obtained with our rGO–PEI–MoS_x modified electrode in the present work (Fig. S13, ESI†). Considering the presence of a competing HER, this value is close to the theoretical value of 59 mV dec⁻¹ predicted for the case where a reversible CO₂/CO₂˙⁻ process occurs prior to a rate-determining chemical step.^7,20,21 This enhanced kinetics of the CO₂/CO₂˙⁻ process was attributed in studies of N-doped carbon nanotubes, to the stabilization of CO₂˙⁻ by the amines on PEI, through hydrogen bonding and electrostatic interactions.⁴³

Conclusions

In summary, we have synthesized amorphous molybdenum sulphide (MoS_x) on polyethylenimine (PEI) modified reduced graphene oxide (rGO–PEI–MoS_x) and used it as a heterogeneous catalyst for electrocatalytic reduction of CO₂ to CO in CO₂ saturated aqueous NaHCO₃ solution with high efficiency and selectivity. The catalyst is capable of reducing CO₂ to CO at an overpotential as low as 140 mV. It reaches a maximum faradaic efficiency of 85.1% at an overpotential of 540 mV, under which this catalyst exhibits a high TOF value of 2.4 s⁻¹ for the formation of CO. Consequently, when the catalyst was deposited on the porous carbon cloth electrode with a large active surface area, a current density of about 50 mA cm⁻² can be reached at the same overpotential. At an overpotential of only 290 mV with respect to CO formation, this catalyst is also capable of producing syngas. The catalyst is highly stable in long-term electrolysis. The efficiency and selectivity towards CO₂ reduction rather than hydrogen evolution at the optimal applied potential are attributed to the synergetic combination of the properties of MoS_x and PEI. Our promising rGO–PEI–MoS_x composite catalyst may pave a new way for efficient CO₂ reduction catalysis based on inexpensive noble-metal-free materials. Catalytic performance could be further improved by fine tuning the properties of MoS₂ using the well-established doping methods.^39,51,69

Acknowledgements

This research was supported by the Australian Research Council (ARC) through the ARC Centre of Excellence for Electromaterials Science. DRM is also grateful to the ARC for his Australian Laureate Fellowship. The authors acknowledge the Monash Centre for Electron Microscopy for use of electron microscopes.

Notes and references

1. N. S. Lewis and D. G. Nocera, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 15729–15735.
2. H. A. Gasteiger and N. M. Markovic, Science, 2009, 324, 48–49.
3. T. A. Faunce, W. Lubitz, A. W. Rutherford, D. R. MacFarlane, G. F. Moore, P. Yang, D. G. Nocera, T. A. Moore, D. H. Gregory, S. Fukuzumi, K. B. Yoon, F. A. Armstrong, M. R. Wasielewski and S. Styring, Energy Environ. Sci., 2013, 6, 695–698.
4. J. Qiao, Y. Liu, F. Hong and J. Zhang, Chem. Soc. Rev., 2014, 43, 631–675.
5. Y. Hori, Modern aspects of electrochemistry, Springer, 2008, pp. 89–189.
6. J. Schneider, H. Jia, J. T. Muckerman and E. Fujita, Chem. Soc. Rev., 2012, 41, 2036–2051.
7. Y. Chen and M. W. Kanan, J. Am. Chem. Soc., 2012, 134, 1986–1989.
8. Y. Oh, H. Vrubel, S. Guidoux and X. Hu, Chem. Commun., 2014, 50, 3878–3881.
9. C. W. Li and M. W. Kanan, J. Am. Chem. Soc., 2012, 134, 7231–7234.
10. M. R. Dubois and D. L. Dubois, Acc. Chem. Res., 2009, 42, 1974–1982.
11. C. Costentin, M. Robert and J.-M. Savéant, Chem. Soc. Rev., 2013, 42, 2423–2436.
12. I. Bhugun, D. Lexa and J.-M. Savéant, J. Am. Chem. Soc., 1996, 118, 1769–1776.
13. S. Lin, C. S. Diercks, Y. B. Zhang, N. Kornienko, E. M. Nichols, Y. Zhao, A. R. Paris, D. Kim, P. Yang and O. M. Yaghi, Science, 2015, 349, 1208–1213.
14. Y. Oh and X. Hu, Chem. Soc. Rev., 2013, 42, 2253–2261.
15. E. Barton Cole, P. S. Lakkaraju, D. M. Rampulla, A. J. Morris, E. Abelev and A. B. Bocarsly, J. Am. Chem. Soc., 2010, 132, 11539–11551.
16. B. Kumar, M. Asadi, D. Pisasale, S. Sinha-Ray, B. A. Rosen, R. Haasch, J. Abiade, A. L. Yarin and A. Salehi-Khojin, Nat. Commun., 2013, 4, 2819–2826.
17. Y. Liu, S. Chen, X. Quan and H. Yu, J. Am. Chem. Soc., 2015, 137, 11631–11636.
18. 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.
19. X. Min and M. W. Kanan, J. Am. Chem. Soc., 2015, 137, 4701–4708.
20. Y. Chen, C. W. Li and M. W. Kanan, J. Am. Chem. Soc., 2012, 134, 19969–19972.
21. Q. Lu, J. Rosen, Y. Zhou, G. S. Hutchings, Y. C. Kimmel, J. G. Chen and F. Jiao, Nat. Commun., 2014, 5, 3242–3247.
22. Q. Lu, J. Rosen and F. Jiao, ChemCatChem, 2014, 7, 38–47.
23. S.-K. Amin, J. Huei-Ru, A. R. Brian, Z. Wei, M. Sichao, J. A. K. Paul and I. M. Richard, J. Phys. Chem. C, 2013, 117, 1627–1632.
24. D. Gao, H. Zhou, J. Wang, S. Miao, F. Yang, G. Wang, J. Wang and X. Bao, J. Am. Chem. Soc., 2015, 137, 4288–4291.
25. W. Zhu, R. Michalsky, O. Metin, H. Lv, S. Guo, C. J. Wright, X. Sun, A. A. Peterson and S. Sun, J. Am. Chem. Soc., 2013, 135, 16833–16836.
26. W. Zhu, Y. J. Zhang, H. Zhang, H. Lv, Q. Li, R. Michalsky, A. A. Peterson and S. Sun, J. Am. Chem. Soc., 2014, 136, 16132–16135.
27. 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.
28. Y. Zhang, G. Jacobs, D. E. Sparks, M. E. Dry and B. H. Davis, Catal. Today, 2002, 71, 411–418.
29. C. W. Li and M. W. Kanan, J. Am. Chem. Soc., 2012, 134, 7231–7234.
30. S. Zhang, P. Kang and T. J. Meyer, J. Am. Chem. Soc., 2014, 136, 1734–1737.
31. J. L. DiMeglio and J. Rosenthal, J. Am. Chem. Soc., 2013, 135, 8798–8801.
32. J. Medina-Ramos, J. L. DiMeglio and J. Rosenthal, J. Am. Chem. Soc., 2014, 136, 8361–8367.
33. J. C. Boyington, V. N. Gladyshev, S. V. Khangulov, T. C. Stadtman and P. D. Sun, Science, 1997, 275, 1305–1308.
34. K. Schuchmann and V. Müller, Science, 2013, 342, 1382–1385.
35. T. Reda, C. M. Plugge, N. J. Abram and J. Hirst, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 10654–10658.
36. A. Bassegoda, C. Madden, D. W. Wakerley, E. Reisner and J. Hirst, J. Am. Chem. Soc., 2014, 136, 15473–15476.
37. B. Hinnemann, P. G. Moses, J. Bonde, K. P. Jørgensen, J. H. Nielsen, S. Horch, I. Chorkendorff and J. K. Nørskov, J. Am. Chem. Soc., 2005, 127, 5308–5309.
38. T. F. Jaramillo, K. P. Jørgensen, J. Bonde, J. H. Nielsen, S. Horch and I. Chorkendorff, Science, 2007, 317, 100–102.
39. M. Chhowalla, H. S. Shin, G. Eda, L.-J. Li, K. P. Loh and H. Zhang, Nat. Chem., 2013, 5, 263–275.
40. J. D. Benck, T. R. Hellstern, J. Kibsgaard, P. Chakthranont and T. F. Jaramillo, ACS Catal., 2014, 4, 3957–3971.
41. K. Chan, C. Tsai, H. A. Hansen and J. K. Nørskov, ChemCatChem, 2014, 6, 1899–1905.
42. M. Asadi, B. Kumar, A. Behranginia, B. A. Rosen, A. Baskin, N. Repnin, D. Pisasale, P. Phillips, W. Zhu, R. Haasch, R. F. Klie, P. Kral, J. Abiade and A. Salehi-Khojin, Nat. Commun., 2014, 5, 4470–4477.
43. S. Zhang, P. Kang, S. Ubnoske, M. K. Brennaman, N. Song, R. L. House, J. T. Glass and T. J. Meyer, J. Am. Chem. Soc., 2014, 136, 7845–7848.
44. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
45. D. Merki, S. Fierro, H. Vrubel and X. Hu, Chem. Sci., 2011, 2, 1262–1267.
46. Y. Zhang, B. Chen, L. Zhang, J. Huang, F. Chen, Z. Yang, J. Yao and Z. Zhang, Nanoscale, 2011, 3, 1446–1450.
47. X. Zhou, Z. Chen, D. Yan and H. Lu, J. Mater. Chem., 2012, 22, 13506–13516.
48. Z. Y. Sui, Y. Cui, J. H. Zhu and B. H. Han, ACS Appl. Mater. Interfaces, 2013, 5, 9172–9179.
49. H. Kim, R. Namgung, K. Singha, I. K. Oh and W. J. Kim, Bioconjugate Chem., 2011, 22, 2558–2567.
50. H. Vrubel and X. Hu, ACS Catal., 2013, 3, 2002–2011.
51. C. G. Morales-Guio and X. Hu, Acc. Chem. Res., 2014, 47, 2671–2681.
52. S.-X. Guo, S.-F. Zhao, A. M. Bond and J. Zhang, Langmuir, 2012, 28, 5275–5285.
53. X. Yang, C. Cheng, Y. Wang, L. Qiu and D. Li, Science, 2013, 341, 534–537.
54. F. Li, J. Chen, X. Wang, M. Xue and G. Chen, Adv. Funct. Mater., 2015, 25, 4601–4606.
55. Z. Pu, Q. Liu, A. M. Asiri, A. Y. Obaid and X. Sun, J. Power Sources, 2014, 263, 181–185.
56. T. Hatsukade, K. P. Kuhl, E. R. Cave, D. N. Abram and T. F. Jaramillo, Phys. Chem. Chem. Phys., 2014, 16, 13814–13819.
57. Y. Hori, A. Murata, K. Kikuchi and S. Suzuki, J. Chem. Soc., Chem. Commun., 1987, 10, 728–729.
58. C. Delacourt, P. L. Ridgway and J. Newman, J. Electrochem. Soc., 2010, 157, B1902–B1910.
59. P. Kedzierzawski and J. Augustynski, J. Electrochem. Soc., 1994, 141, L58–L60.
60. D. Merki, H. Vrubel, L. Rovelli, S. Fierro and X. Hu, Chem. Sci., 2012, 3, 2515–2525.
61. J. D. Benck, Z. Chen, L. Y. Kuritzky, A. J. Forman and T. F. Jaramillo, ACS Catal., 2012, 2, 1916–1923.
62. A. B. Laursen, P. C. Vesborg and I. Chorkendorff, Chem. Commun., 2013, 49, 4965–4967.
63. D. J. Li, U. N. Maiti, J. Lim, D. S. Choi, W. J. Lee, Y. Oh, G. Y. Lee and S. O. Kim, Nano Lett., 2014, 14, 1228–1233.
64. S. Chen, J. Duan, Y. Tang, B. Jin and S. Z. Qiao, Nano Energy, 2015, 11, 11–18.
65. B. A. Rosen, A. Salehi-Khojin, M. R. Thorson, W. Zhu, D. T. Whipple, P. J. Kenis and R. I. Masel, Science, 2011, 334, 643–644.
66. H. Vrubel, D. Merki and X. Hu, Energy Environ. Sci., 2012, 5, 6136–6144.
67. B. Rezaei, S. Mallakpour and M. Taki, J. Power Sources, 2009, 187, 605–612.
68. B. A. Rosen, W. Zhu, G. Kaul, A. Salehi-Khojin and R. I. Masel, J. Electrochem. Soc., 2012, 160, H138–H141.
69. Y. Sun, S. Gao, F. Lei and Y. Xie, Chem. Soc. Rev., 2015, 44, 623–636.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ee02879e

---