Cobalt sulfide nanoparticles impregnated nitrogen and sulfur co-doped graphene as bifunctional catalyst for rechargeable Zn–air batteries

Abstract

We report cobalt sulfide nanoparticles decorated nitrogen and sulfur co-doped graphene nanosheets as an effective bifunctional oxygen catalyst. Its oxygen evolution reaction (OER) catalytic activity outperforms Pt/C, giving a current density of 38.19 mA cm⁻² at 1.0 V. Zn–air batteries using this material on air-cathode show high stability and good rechargeability.

---

The ever-growing awareness of sustainable development has been a key drive for environmentally friendly electric vehicles (EVs). Zn–air rechargeable batteries using aqueous electrolytes have emerged as promising candidates thanks to their high specific energy density (1084 Wh kg⁻¹), low cost and good safety^1,2 as compared to the prevailing lithium-ion batteries that offer a typical energy density of 200–250 Wh kg⁻¹ at relatively high cost with safety concerns. However, the commercialization viability of rechargeable Zn–air batteries is hindered by certain issues, such as the poor rechargeability, low power density and low energy efficiency, which are associated with the low columbic efficiency, sluggish oxygen reactions and high overpotential at the air-cathode of the batteries. A solution that helps circumvent such issues is an efficient bifunctional oxygen catalyst on air-cathode that facilitates oxygen reduction reaction (ORR) during discharge and oxygen evolution reaction (OER) during charge in battery operations.^3,4

Traditionally, bifunctional oxygen catalysts contain precious metals, such as Pt, Ru, and Ir in high content, which are far too costly. For applications using large battery packs, e.g., in electric vehicles, the bifunctional oxygen catalysts employed need to be inexpensive, environment-friendly and catalytically efficient.^5–7 Herein, we report such a catalyst prepared by integrating nitrogen and sulfur co-doped two-dimensional graphene nanosheets (GNs) of good ORR catalytic activity^8–10 with transition metal chalcogenide nanoparticles (NPs) of proven OER activity^11–15 via a facile one-pot hydrothermal reaction. This composite catalyst, N and S co-doped GNs decorated with CoS_x NPs, is found to be an effective bifunctional oxygen catalyst, giving good cycling stability in rechargeable Zn–air batteries.

Co-doping graphene with two different elements (particularly N and S) is a promising pathway to produce an ORR catalyst because it helps to create non-neutral electron sites that enable efficient four-electron transfer ORR activity.¹⁰ We use thiourea as the single precursor to react with graphene oxide (GO) to form N and S co-doped graphene nanosheets (NS-GNs), referring to a reported method;^16,17 however, the solvent in our reaction was changed from ethylene glycol (EG) to water, and the reaction temperature was reduced from 200 °C to 180 °C. Despite the lower reaction temperature, the autogenic pressure inside the enclosed vessels for our water-based system is ∼7 times higher than that of the EG-based system (10 vs. 1.25 bar). As a consequence, instead of somewhat flat flakes, the resulted NS-GNs are heavily corrugated with interlinked nature of wrinkles, as shown in the SEM (JEOL JSM-6700F) and TEM (Philips CM300-FEG instrument with operating voltage at 300 kV) images (Fig. 1a and b). As thiourea reacts with GO in the presence of Co²⁺, the nucleation of CoS_x nanoparticles starts from the surface of NS-GNs, the reductive co-doping product of graphene oxide, and these nuclei further grow into CoS_x nanoparticles of considerably larger sizes. A representative TEM image is shown in Fig. 1c. One can see that spherical CoS_x nanoparticles with an average diameter of 50 nm are decorated on NS-GNs. The corrugation of the underlying NS-GNs in this case is not as significant as it was earlier in the absence of Co²⁺. A representative high resolution TEM (HR-TEM) image (Fig. 1d) reveals that underneath the highly crystalline CoS_x nanoparticles, the NS-GNs are distorted with an average thickness of 7 layers. Such re-stacking of NS-GNs is likely caused by the high autogenic pressure of water at 180 °C.

Fig. 1 (a) A representative SEM image of N and S co-doped graphene nanosheets (NS-GNs), giving corrugated flake structure; (b) a representative TEM image of NS-GNs as ultrathin flexible flakes; (c) a representative TEM image of CoS_x@NS-GNs, where CoS_x nanoparticles are clearly visible on NS-GNs; (d) a representative HR-TEM image, indicating the existence of a few layers of NS-GNs in CoS_x@NS-GNs.

The co-existence of N and S on the surface of corrugated NS-GN flakes is evidenced by XPS (Thermo Scientific, Al K_α radiation) analysis (Fig. 2a), which was introduced via the reaction of the hydrolysis products of thiourea, ammonia and hydrogen sulfide with the oxygen-containing groups on GO under hydrothermal conditions. The contents of elemental N and S are found to be 2.32 and 1.88% respectively. The N1s spectrum is deconvoluted into two sets of signals (Fig. 2b), which are attributed to the pyridinic-N (398.8 eV) and the quaternary-N (401.2 eV). The direct proof of the S doping in graphene is provided by the high resolution S2p spectrum in Fig. 2c, which is resolved into 3 peaks at binding energies of 164.2, 165.7, and 168.5 eV. The first two match with the reported 2p3/2 and 2p1/2 positions of thiophene-S as a result of their spin–orbit coupling, while the 3rd one arises from the oxidized sulfur.^16,17 For CoS_x@NS-GNs, the XPS survey spectrum shows the presence of C, N, S, O, and Co (Fig. 2d). The ratio of Co to S is easily obtained from the EDS data (Fig. 2e) as 2 : 3 (39.6 : 60.4), which appears to indicate a composition of Co₂S₃ for the obtained cobalt sulfide. Nevertheless, with the possible contribution of sulfur atoms on the surface of NS-GNs, this ratio of coincidental value should not be over-interpreted. In the as-prepared composite catalyst, cobalt exists in a mixed valence state of Co(II) and Co(III), which will be further discussed later.

Fig. 2 (a) XPS survey spectrum of NS-GNs; (b) high resolution N1s spectra of NS-GNs; (c) high resolution S2p spectra of NS-GNs; (d) XPS survey spectrum of CoS_x@NS-GNs; (e) EDS spectrum of CoS_x@NS-GNs.

The crystalline phase of obtained NS-GNs and Co_xS/NS-GNs were determined by XRD (Bruker D8 General Area Detector Diffraction System using Cu K_α radiation) measurement, and the diffraction data are shown in Fig. 3a. One can see that the cobalt sulfide NPs grown on NS-GNs match the Co_1−xS phase (ICDD PDF #00-042-0826) with hexagonal structure. As both CoS_x and Co_1−xS are nonstoichiometric cobalt sulfides, which are essentially interchangeable, herein we will continue to use CoS_x@NS-GNs to denote the NS-GNs supported cobalt sulfides nanoparticles for consistency. To determine the content of CoS_x, the thermogravimetric analysis (TGA Q500, TA instruments) data of NS-GNs and CoS_x@NS-GNs were obtained from room temperature to 900 °C in air at a heating rate of 10 °C min⁻¹ (Fig. 3b). For NS-GNs (black line), the slow weight loss after 220 °C is likely to be due to the removal of sulfur and nitrogen from the graphitic lattices, and the major weight loss from 550 to 675 °C is attributed to the burn-off of graphitic carbons. In CoS_x@NS-GNs (red line), a similar major change takes place from 510 to 610 °C, but with a considerably smaller weight loss. This may be due to the weight gain by a complicated oxidation of CoS_x, which partially neutralizes the weight loss caused by the burn-off of graphitic carbons. The oxidized products of cobalt sulfide decompose further at temperatures above 700 °C to obtain stable cobalt oxides. Based on the TGA data, the composition of CoS_x@NS-GNs is estimated to be about 63 wt% of NS-GNs and 37 wt% of CoS_x.

Fig. 3 (a) XRD patterns of NS-GNs and CoS_x@NS-GNs, showing match of the cobalt sulfide with Co_1−xS; (b) TGA data of CoS_x@NS-GNs and NS-GNs in air from room temperature to 900 °C, revealing the composition.

The ORR catalytic activity of CoS_x@NS-GNs was evaluated on an Autolab Potentiostat (PGSTAT302N, Metrohm) using rotating ring–disc electrode technique. Linear sweep voltammograms (LSVs) on glassy carbon electrode (GCE), NS-GNs, CoS_x@NS-GNs, and Pt/C were performed at a rotating rate of 1600 rpm, and the results are presented together in Fig. 4a. Unsurprisingly, a typical LSV curve of non-noble metal ORR catalyst was observed for NS-GNs, which is in good agreement with the reported results.^9,10,16 Herein, the onset electrode potential for the ORR (E_ORR) is defined as the potential at which the additional cathodic current is kept at 10 μA cm⁻². In this case, with the decoration of CoS_x onto NS-GNs, the resultant composite catalyst CoS_x@NS-GNs shows improvement in ORR onset potential by 27 mV (−0.147 vs. −0.174 V) and ORR current by ∼10% of increase. Nevertheless, its ORR activity is inferior to commercial Pt/C (−0.045 V) because of a more negative onset potential and a slightly lower ORR current (2.818 vs. 2.900 mA cm⁻²) over the entire potential range. A summary of determinant parameters derived from Fig. 4a is given below in Table 1. One can clearly see the improved ORR activity from NS-GNs to CoS_x@NS-GNs, which can be attributed to a synergistic effect between CoS_x NPs and the NS-GNs support, as previously found in a similar system.¹⁸

Fig. 4 (a) Linear sweep voltammograms (LSVs) of GCE, NS-GNs, CoS_x@NS-GNs and Pt/C in O₂ saturated 0.1 M KOH aqueous solution at 1600 rpm with the sweep rate of 5 mV s⁻¹; (b) RRDE LSV of CoS_x@NS-GNs at a rotation rate of 1600 rpm; (c) the calculated number of electron transfer and H₂O₂% at varied potentials; (d) cyclic voltammograms of CoS_x@NS-GNs in O₂-saturated 0.1 M KOH aqueous solution after 50, 250, 500, 750, 1000 cycles of scans. The potential sweep rate is 50 mV s⁻¹.

Table 1 Summary of the test data from half-cell (ORR and OER) and full cell (battery) using air-cathode of CoS_x@NS-GNs or Pt/C

Catalyst	ORR^a		OER^b		Zn–air battery
	EORR (V)	j (mA cm^−2)	EOER (V)	j (mA cm^−2)	EOCV (V)	Ed^c (V)	Ec^c (V)
a j at E = −0.8 V.b E at j = 3 mA cm^−2 and j at 1.0 V.c E @ the 40th cycle. For the charging, the value was taken @ 50% of capacity.
CoSx@NS-GN	−0.147	2.818	0.674	38.19	1.394	1.243	2.118
Pt/C	−0.045	2.900	0.657	27.52	1.451	1.211	2.129

---

The ORR catalytic mechanism of CoS_x@NS-GNs was further investigated to understand the selectivity of the ORR process, i.e., whether it undergoes a two-electron (forming HO₂⁻ intermediate) or a desirable four-electron process. The RRDE LSVs results are shown in Fig. 4b, from which the number of electrons transferred (n) and the percentage of H₂O₂ released (%H₂O₂) during ORR are calculated using the following equations:

where I_d and I_r are the Faradaic current on the disk and the ring, respectively, with N being the collection efficiency of the ring, which was reported to be 0.37 by the manufacturer. From Fig. 4b, one can clearly see that I_r increases with I_d. Nevertheless, the current density on the ring electrode is notably smaller than that on the disk electrode, suggesting the formation of a small amount of hydrogen peroxide on the catalyst. The calculated number of electrons transferred during ORR is ∼3.2 with the H₂O₂% being ∼39% at potentials lower than −0.5 V (Fig. 4c). This indicates the involvement of two coexisting pathways, viz., two-electron and four-electron transfers (eqn (3) and (4)), in the entire process with the latter being the dominant one.¹⁹

O₂ + H₂O + 2e⁻ → HO₂⁻ + OH⁻ (3)

O₂ + 2H₂O + 4e⁻ → 4OH⁻ (4)

A durability test on CoS_x@NS-GNs was carried out using cyclic voltammetry scans in O₂-saturated 0.1 M KOH aqueous solution. As shown in Fig. 4d, little difference is visible after potential sweeps of 50, 250, 500, 750 and 1000 cycles, proving CoS_x@NS-GNs to have excellent electrochemical stability under alkaline conditions, which is important for the conventional rechargeable zinc–air battery used in this work.

To further understand the rechargeability, the OER activity (determining the charge behaviour of metal–air batteries) of NS-GNs, CoS_x@NS-GN, and Pt/C were studied. The OER polarization curves measured in O₂-saturated 0.1 M KOH aqueous solution at 1600 rpm are shown in Fig. 5. One can see that with the introduction of CoS_x onto NS-GNs, the onset potential at the current density of 3 mA cm⁻² is lowered to 0.674 V, which is comparable to Pt/C (0.657 V). More interestingly, the OER specific activity of CoS_x@NS-GNs at 1.0 V is measured to be 38.19 mA cm⁻², which is 1.4 times as high as that of Pt/C (Table 1). The higher OER activity is likely attributed to the CoS_x nanoparticles because NS-GNs are known to present poor OER activity. The main function of NS-GNs here, in addition to their ORR activity, is to provide a conductive network that facilitates efficient electron transfer. Nevertheless, some synergistic enhancement in the catalytic activity may arise from the interaction of CoS_x nanoparticles with the underlying NS-GNs, which is yet to be understood.^17,20

Fig. 5 OER curves of NS-GN, CoS_x@NS-GN, and Pt/C in O₂ saturated 0.1 M KOH at 1600 rpm with a sweep rate of 5 mV s⁻¹. The current density on CoS_x@NS-GNs at 1.0 V is 38.19 mA cm⁻², which is 1.4 times as high as that of Pt/C.

Zn–air batteries were built using CoS_x@NS-GNs or commercial Pt/C catalyst as air-cathode material to evaluate the rechargeability of the batteries in discharge–charge cycling experiments. The testing results are shown in Fig. 6a and also summarized in Table 1. One can see that when CoS_x@NS-GNs were used as cathode material, the battery could be easily discharged and charged for 50 cycles at 1.25 mA cm⁻² over a period of 50 h, with almost a constant discharge voltage of ∼1.23 V. In contrast, Pt/C suffers a loss of 11% in the discharge voltage (1.35 → 1.20 V). In the charge processes, a slight improvement in the charge voltage for CoS_x@NS-GNs upon cycling is observed; however, Pt/C encounters a clear increase in the charge voltage. Such difference in the change of charge and discharge voltages between both the samples during the repetitive charge–discharge process can be better viewed in Fig. 6b, in which the average discharge and charge potentials for the consecutive 50 cycles are plotted. A close-up view of the discharge–charge curves in the 40th cycle of both cells is given in Fig. 6c, in which the above-described difference is seen more clearly. It is also reflected by the larger difference between E_OCV and the discharge potential after 40 cycles of discharge–charge for Zn–air batteries using Pt/C based air-cathode (0.240 V) than that of its counterpart using the air-cathode of composite catalyst CoS_x@NS-GNs (0.151 V). With a comparable catalytic performance to Pt/C, CoS_x@NS-GNs will provide potential inexpensive alternatives for rechargeable Zn–air batteries.^21,22

Fig. 6 (a) Cycling performance of Zn–air battery at 1.25 mA cm⁻² and 1 h per cycle over 50 h using CoS_x@NS-GN and Pt/C as catalyst. (b) Average discharge and charge potentials for the consecutive 50 cycles. (c) Typical discharge and charge polarization curves of CoS_x@NS-GN and Pt/C for the fortieth cycle.

Conclusion

In conclusion, we synthesized an inexpensive composite catalyst, CoS_x@NS-GNs, from a water-based reaction that shows comparable performance to Pt/C catalyst. The zinc–air rechargeable battery built using CoS_x@NS-GNs as air cathode material exhibits excellent cycling stability, proving the feasibility and significance of integrating N/S co-doped graphene nanosheets with nanosized metal sulfides in the design of ORR and OER bifunctional catalysts. Such catalysts as sustainable electrode material are likely to take a more central stage in the current surge of interest in metal–air battery research.

Acknowledgements

This work was supported by the project IMRE/12-2P0504 under the SERC Advanced Energy Storage Research Programme, and Institute of Material Research and Engineering (IMRE), A*STAR, Singapore. The authors thank Dr Zheng Zhang for his help with the XPS measurement.

Notes and references

1. J.-S. Lee, S. T. Kim, R. Cao, N.-S. Choi, M. Liu, K. T. Lee and J. Cho, Adv. Energy Mater., 2011, 1, 34–50.
2. P. Sapkota and H. Kim, J. Ind. Eng. Chem., 2009, 15, 445–450.
3. J.-G. Zhang, P. G. Bruce and X. G. Zhang, in Handbook of Battery Materials, Wiley-VCH Verlag GmbH & Co. KGaA, 2011, pp. 757–795.
4. V. Neburchilov, H. Wang, J. J. Martin and W. Qu, J. Power Sources, 2010, 195, 1271–1291.
5. F. Cheng and J. Chen, Chem. Soc. Rev., 2012, 41, 2172–2192.
6. R. Cao, J.-S. Lee, M. Liu and J. Cho, Adv. Energy Mater., 2012, 2, 816–829.
7. L. Jörissen, J. Power Sources, 2006, 155, 23–32.
8. L. Dai, Acc. Chem. Res., 2012, 46, 31–42.
9. S. Yang, L. Zhi, K. Tang, X. Feng, J. Maier and K. Müllen, Adv. Funct. Mater., 2012, 22, 3634–3640.
10. J. Liang, Y. Jiao, M. Jaroniec and S. Z. Qiao, Angew. Chem., Int. Ed., 2012, 51, 11496–11500.
11. N. Alonso-Vante, in Electrocatalysis in Fuel Cells, ed. M. Shao, Springer, London, 2013, vol. 9, pp. 417–436.
12. M.-R. Gao, J. Jiang and S.-H. Yu, Small, 2012, 8, 13–27.
13. H. Wang, Y. Liang, Y. Li and H. Dai, Angew. Chem., Int. Ed., 2011, 50, 10969–10972.
14. N. Mahmood, C. Zhang, J. Jiang, F. Liu and Y. Hou, Chem.–Eur. J., 2013, 19, 5183–5190.
15. N. M. Dimitrijevic, S. Li and M. Grätzel, J. Am. Chem. Soc., 1984, 106, 6565–6569.
16. Y. Su, Y. Zhang, X. Zhuang, S. Li, D. Wu, F. Zhang and X. Feng, Carbon, 2013, 62, 296–301.
17. Q. Liu, J. Jin and J. Zhang, ACS Appl. Mater. Interfaces, 2013, 5, 5002–5008.
18. Y. Liang, Y. Li, H. Wang, J. Zhou, J. Wang, T. Regier and H. Dai, Nat. Mater., 2011, 10, 780–786.
19. M. Seredych, J.-C. Idrobo and T. J. Bandosz, J. Mater. Chem. A, 2013, 1, 7059–7067.
20. W. G. Hardin, D. A. Slanac, X. Wang, S. Dai, K. P. Johnston and K. J. Stevenson, J. Phys. Chem. Lett., 2013, 4, 1254–1259.
21. Z. Chen, A. Yu, D. Higgins, H. Li, H. Wang and Z. Chen, Nano Lett., 2012, 12, 1946–1952.
22. Y. Li, M. Gong, Y. Liang, J. Feng, J.-E. Kim, H. Wang, G. Hong, B. Zhang and H. Dai, Nat. Commun., 2013, 4, 1805.

---