- Email: [email protected]
sulfur-doping of graphene with cysteine as a heteroatom source for oxygen reduction electrocatalysis
Journal of Colloid and Interface Science 505 (2017) 32–37
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science journal homepage: www.elsevier.com/locate/jcis
Regular Article
Nitrogen/sulfur-doping of graphene with cysteine as a heteroatom source for oxygen reduction electrocatalysis Huanhuan Zhang, Yanli Niu, Weihua Hu ⇑ Institute for Clean Energy & Advanced Materials, Chongqing Key Laboratory for Advanced Materials and Technologies of Clean Energies, Faculty of Materials & Energy, Southwest University, Chongqing, China
g r a p h i c a l a b s t r a c t
a r t i c l e
i n f o
Article history: Received 8 March 2017 Revised 18 April 2017 Accepted 20 May 2017 Available online 23 May 2017 Keywords: N/S-doped graphene Oxygen reduction reaction Cysteine Electrocatalyst Graphene oxide
a b s t r a c t Heteroatom-doped graphene have encouraged intensive research as promising metal-free oxygen reduction reaction (ORR) electrocatalysts but the correlation between the precursor material and final ORR activity remains unclear. In this work a serial of nitrogen/sulfur (N/S)-doped graphene catalysts were synthesized by modifying graphene oxide (GO) with cysteine as a N/S source and sequential thermal annealing. It is disclosed that the cysteine-GO reaction time shows a significant influence on the ORR activity of N/S-doped graphene. A unique process of oxidation-induced in situ disulfide formation is further found to be involved in the synthesis of optimal N/S-doped graphene, which displays ORR activity superior to commercial Pt/C in alkaline media. This work suggests that the heteroatom source itself and careful optimization of reaction conditions are critical to obtain high performance doped-graphene electrocatalyst. Ó 2017 Elsevier Inc. All rights reserved.
1. Introduction Heteroatom-doped carbon has emerged as a type of promising oxygen reduction reaction (ORR) catalysts since the pioneer work by Dai and coworkers in 2009 [1–3]. The enhanced ORR activity
⇑ Corresponding author. E-mail address: [email protected] (W. Hu). http://dx.doi.org/10.1016/j.jcis.2017.05.069 0021-9797/Ó 2017 Elsevier Inc. All rights reserved.
is believed to originate from heteroatom-disturbed electroneutrality of sp2 carbon in the graphitic framework to favor the adsorption of molecular oxygen and subsequent electrochemical reduction [2,4–8]. Dually-doped carbons were further explored as this strategy offers more flexibility to modulate the electronic structure by optimizing doping sites, dopant densities, and dopant ratio for possible synergistic effect [9–17]. Accumulating evidence has unveiled that both the doping density and doping site of the
H. Zhang et al. / Journal of Colloid and Interface Science 505 (2017) 32–37
33
Fig. 1. Schematic showing the synthesis of N/S co-doped graphene.
heteroatom are significantly important to the final ORR activity [5,18,19]. In most cases, heteroatom-doping of carbon was achieved by thermal annealing of a precursor material composed of a carbon host with modified heteroatom source. It is not fully understood yet regarding how heteroatoms are conjugated into different sites of the host carbon lattice during the high temperature process, which restricts the rational design and synthesis of doped carbon for ORR catalysis. Cysteine is a well-known sulfur-containing redox amino acid implicated in various biological systems [20,21]. It has been used as a precursor to synthesize N/S-doped carbon quantum dots [22]. Due to the presence of reductive thiol, cysteine possesses considerable reducing ability and is able to be oxidized by dissolved oxygen at neutral pH to form a cystine dimer [23]. At the same time, cysteine is able to bind to graphene oxide (GO) surface via electrostatic attraction and/or hydrophobic interaction [24] or via the ring-opening of epoxides on GO sheets [25]. Considering the abundant chemistry of cysteine and its various ways binding to GO, in this work we used it as a green N/S source to functionalize GO under different reaction condition, followed by annealing in an inert Argon atmosphere to synthesize N/S-doped graphene catalysts. Systematic characterizations and electrochemical comparisons unveiled that although N/S-doping is successfully achieved in all synthesized catalysts, their ORR activity varies significantly upon the change of the cysteine-GO reaction time. The optimal catalyst demonstrates ORR performance superior to commercial Pt/C in alkaline media. A unique process of oxidationinduced in situ disulfide formation with optimal grafting density is further disclosed to be critical in the synthesis of the optimal catalyst, as shown in Fig. 1. This work suggests that the heteroatom source itself and careful optimization of reaction conditions are critical to obtain high performance doped-graphene electrocatalyst, and thus may offer a valuable clue on the rational design of metal-free ORR electrocatalysts.
2. Experimental GO was prepared using the modified Hummers method [26,27]. Graphite powder (0.5 g) and sodium nitrate (NaNO3, 0.5 g) were mixed with concentrated sulfuric acid (H2SO4, 98%, 23 mL) and stirred for 30 min in an ice bath. Potassium permanganate (KMnO4, 3.0 g) was slowly added into the mixture with vigorous stirring. The mixture was further stirred for 1 h in a 35 °C water bath before water (40 mL) was slowly added under vigorous stirring. The mixture was stirred for 30 min at 90 °C. Finally, 100 mL of water was added to the mixture, followed by hydrogen peroxide (H2O2, 30%, 3.0 mL). After cooling down to room temperature, the products were collected by using centrifugation, and washed with 5 wt%
HCl and deionized water for several times. The yellow-brown suspension of GO was obtained by dispersing the products in water with the assistance of ultrasonication. To synthesize N/S-doped graphene, cysteine (2.0 mg mL1) was added to GO suspension (1.0 mg mL1) and the pH of the final solution was around 5.0–6.0, which was further adjusted to 8.0 with 0.1 M KOH solution. After stirring the mixture at ambient atmosphere for 7 d, the product was collected by centrifugation, freezing-dried, and annealed at 900 °C (Ar atmosphere, heating rate of 5 °C min1, maintaining at 900 °C for 3 h), as shown in Fig. 1. This catalyst was denoted as N/S-G-7d. For comparison, N/ S-G-1d, N/S-G-4d and N/S-G-14d were also obtained by adjusting the cysteine-GO reaction time to 1, 4 and 14 d, respectively. GO was also processed with the same annealing procedure to obtained thermally reduced GO (RGO). An Autolab PGSTAT302N potentiostat system was used for electrochemical measurements in a three-electrode cell with Pt foil as the counter electrode. Hg/HgO/1M KOH electrode was used as the reference electrode, to which all potentials reported in present work are referring. Also the potential could be easily converted to the one versus Reversible Hydrogen Electrode (ERHE) by using the Nernst equation:
ERHE ¼ EHg=HgO þ 0:059 pH þ E0Hg=HgO where the pH is the electrolyte pH (13 in this case) and the E0Hg=HgO is 0.098 V versus Normal Hydrogen Electrode (NHE). The working electrode was prepared by casting catalyst slurry (dispersing 2.0 mg of catalyst and 50 lL of 5 wt% Nafion solution in 1.0 mL of ethanol) onto glassy carbon electrode (2.0 mm diameter, 5 lL applied for cyclic voltammetry CV measurements, and 5.0 mm diameter, 25 lL applied for linear scanning voltammetry, LSV measurements) and the solvent was allow to naturally evaporate at room temperature. Before each experiment, the freshly prepared electrolyte (0.1 M KOH) was bubbled with pure N2 or O2 for 30 min. Fourier transform infrared (FTIR) spectra were measured on Nicolet FTIR 6700 spectrophotometer. X-ray photoelectron spectrum (XPS) was collected on an ESCALAB 250Xi system from Thermo Fisher. Transmission electron microscopy (TEM) and Scanning electron microscopy (SEM) images were obtained on a JEM2100 system and a JSM-7800F system from JOEL, respectively. Raman spectra were collected on a RENISHAW inVia Raman Microscope with 532 nm excitation. 3. Results and discussion The annealing temperature was first optimized and 900 °C was chosen as the optimal temperature for thermal treatment of the precursor, as shown in Fig. S1. The ORR activity of as-prepared N/
34
H. Zhang et al. / Journal of Colloid and Interface Science 505 (2017) 32–37
Fig. 2. Comparison of ORR activity of various catalysts in O2-saturated 0.1 M KOH. (a) LSV curves of N/S-G-1d, N/S-G-4d, N/S-G-7d, and N/S-G-14d; (b) LSV curves of RGO, N/SG-7d and Pt/C, scan rate of 5 mV s1 at 1600 rpm.
S-doped graphene with different cysteine-GO reaction time was evaluated with LSV measurement. As shown in Fig. 2a, the N/SG-7d sample exhibits a very positive onset potential of 0.09 V, which is more positive than that of other N/S-doped graphene samples tested (N/S-G-1d, N/S-G-4d, and N/S-G-14d). The limiting current density of N/S-G-7d sample is also highest among all four catalysts. The cysteine-GO reaction time shows a significant influence on the ORR activity of as-prepared catalyst, which may be tightly associated with the different bonding way between cys-
teine and GO and will be discussed in details below. When compared to commercial Pt/C (20 wt% Pt on Vulcan carbon XC-72) and thermally reduced GO (RGO), N/S-G-7d also demonstrates more positive onset potential and half-wave potential than commercial Pt/C with comparable limiting current density, indicating high ORR activity (Fig. 2b). LSV measurements at different rotating rates were carried out to reveal the ORR selectivity on N/S-G-7d sample. All electrocatalytic measurements are highly reproducible. As shown in
Fig. 3. ORR activity of N/S-G-7d in O2-saturated 0.1 M KOH. (a) rotating-disk voltammograms of N/S-G-7d at different rotating rates at a scan rate of 5 mV s1 and (b) corresponding K-L plots at different potentials; (c) RRDE voltammograms and (d) thereof calculated peroxide yield and electron-transfer-number of N/S-G-7d at a scan rate of 5 mV s1 at 1600 rpm.
35
H. Zhang et al. / Journal of Colloid and Interface Science 505 (2017) 32–37
Fig. 4. (a) FTIR spectra of GO, cystine and GO after reacting with cysteine for different time at pH 8.0; (b–d) high-resolution XPS spectra of C1s (b), N1s (c) and S2p (d) of N/SG-7d.
Table 1 Comparison of doping densities as determined by XPR and ORR performance of various N/S-doped graphene catalysts synthesized in present work.
*
No.
Sample
Condition
N at.%
S at.%
S/N ratio
1 2 3 4 5 6 7 8
N/S-G-7d N/S-G-1d N/S-G-4d N/S-G-14d
GO + cysteine@pH 8.0, 7 d GO + cysteine@pH 8.0, 1 d GO + cysteine@pH 8.0, 4 d GO + cysteine@pH 8, 14 d GO + cysteine@pH 11, 7 d GO + cysteine@pH 6, 7 d GO + cystine, pH 12.5–8.0, 1 d GO + cystine, pH 1.5–8.0, 1 d
1.02 0.80
1.32 0.52
1.29 0.65
2.69 2.38
1.96 1.65
0.73 0.69
1.95
1.39
0.71
J @-0.2 V (mA cm2)
E1/2* (V)
3.361 1.242 2.937 2.398 2.497 2.078 2.545 2.546
0.093 0.224 0.122 0.127 0.146 0.164 0.154 0.166
Half wave potential (E1/2) is defined as the potential at which the current density reaches half of the current density at 0.6 V for convenience.
Fig. 3a, ORR current density is improved with the increase of rotating rate from 400 to 2500 rpm. The Koutecky–Levich (K-L) plots of N/S-G-7d catalyst exhibit good linearity at different potentials, as shown in Fig. 3b. The electron-transfer number (n) was calculated to vary from 3.47 to 3.72 in the range of 0.30 to 0.70 V according to the K-L plots [7,28], indicating that the detrimental twoelectron ORR process is efficiently inhibited on N/S-G-7d. The ORR pathway on N/S-G-7d was further investigated by rotating ring-disk electrode (RRDE) measurements at 1600 rpm. Fig. 3c exhibits high disk current (Id) while low ring current (Ir),
suggesting negligible hydrogen peroxide intermediate generated and excellent four-electron selectivity on N/S-G-7d [29–31]. The hydrogen peroxide yield and overall electron-transfer-number were calculated and shown in Fig. 3d by using Ir, Id and collecting efficiency of 37% [3,32]. The yield of peroxide species is from 3.7% to 16.8% when the potential was swept from 0.2 to 0.75 V. The electron-transfer-number ranges from 3.66 to 3.92 depending on the potential, which is in good agreement with the values calculated from K-L plots, indicating excellent selectivity for favorable four-electron pathway of N/S-G-7d. According to accelerated
36
H. Zhang et al. / Journal of Colloid and Interface Science 505 (2017) 32–37
Fig. 5. Comparison of ORR activity of different N/S-doped graphene catalysts. (a) LSV curves of N/S-doped graphene prepared by annealing cystine-grafted GO at 900 °C. The cystine, instead of cysteine, was grafted by slowly adjusting solution pH with 0.1 M KOH or HCl from 12.5 or 1.5 to 8.0 in 24 h. (b) LSV curves of N/S-doped graphene prepared with the same procedure as for N/S-G-7d except for that the pH value of cysteine-GO reaction solution was 6, or 11, but not optimal 8.0. Scan rate: 5 mV s1 at 1600 rpm in O2-saturated 0.1 M KOH.
durability test (ADT) and chronoamperometric curve of methanol addition, N/S-G-7d also exhibits excellent stability and good methanol tolerance, both of which are highly desirable for fuel cell applications, as shown in Fig. S2 [33,34]. To disclose the underlying mechanism under the excellent ORR activity of N/S-G-7d, the bonding between GO and cysteine for different reaction time was investigated by FTIR. As shown in Fig. 4a, pristine GO shows its characteristic absorption peaks originating from C@O, C@C, OAH, and CAO stretching vibrations (1732, 1629, 1228, 1071 cm1) [35,36]. After reacting with cysteine for 1 d, the characteristic peak of carboxyl group at 1732 cm1 disappears, indicating partial reduction of GO sheets by cysteine [24]. Longer reaction results in newly emerged bands at 1586, 1488, 1409, 1383, and 1084 cm1, all of which are in good agreement with that of cystine, the disulfide dimer of cysteine after being oxidized. This suggests that the cysteine was grafted onto GO in the form of disulfide dimer [23]. It has been reported that in the presence of dissolved oxygen, cysteine is able to be spontaneously oxidized at neutral pH to form cystine dimer [23], whose solubility is quite low in neutral pH [37]. Therefore, it is speculated that when the reaction proceeds for short time, e.g., 1 d, cysteine binds to GO surface via electrostatic attraction and/or hydrophobic interaction [24] or the ring-opening of epoxides [25]. When prolonged to, e.g., 4 d or 7 d, these pre-bound cysteine act as seeds to initiate the formation of cystine dimer on the nanosheets [24]. Together with the LSV curves in Fig. 2a, these FTIR spectra imply that the cystine dimer is more efficient than cysteine as N/S source for graphene doping. The doping nature of N/S-G-7d was analyzed by XPS and Raman. XPS spectra confirm the presence of N and S (atomic abundance of 1.02% and 1.32%, respectively, as shown in survey spectrum in Fig. S4). The high-resolution C1s, N1s and S2p spectra were shown in Fig. 4b–d, suggesting successful doping in graphene framework (details about XPS analysis were provided in Supplementary material). Remarkably, in other two samples tested (N/S-G-1d, and N/S-G-14d, respectively), successful doping is also achieved, as unveiled by their XPS spectra in Figs. S5 and S6. In N/S-G-7d the S atomic percentage is higher than that of N but it is not the case for other catalysts, as summarized in Table 1. More experimental evidences however, are needed to directly relate the ORR activity to the S/N doping ratio in this class of catalysts. On Raman spectra (Fig. S7), the G band at 1594 cm1 is induced by sp2 carbon atoms and the D band at 1348 cm1 is assigned to defects in graphene sheets. The intensity ratio (ID/IG) of the D and G bands for RGO is 0.96, and it
increases to 1.06 for N/S-G-7d, suggesting successful heteroatom doping of graphene. The unique dimer-mediated grafting for N/S-G-7d inspires us to directly graft disulfide cystine on GO to synthesize N/S-doped graphene. As the solubility of cystine is much higher in acidic or alkaline solution than that in neutral pH [37], we successfully achieved cystine grafting via adjusting solution pH, as confirmed by FTIR in Fig. S8. XPS confirms the successful N/S-doping and the doping density is even higher than that of optimal catalyst (Fig. S9 and Table 1). Electrochemical evaluation, however, suggests that the resultant N/S catalyst is still inferior to the optimal N/S-G-7d catalyst, as shown in Fig. 5a. Together with the comparison in Fig. 2a, the results may suggest that there is an optimal grafting density for cystine on GO, which is easily achieved via in situ cystine dimer formation process rather than via pH-controlled dimer grafting method. The same reason may also lie behind the experimental observation that higher or lower pH value (Fig. 5b) results in inferior ORR activity because the formation rate of disulfide dimer is sensitive to pH and thus induces deviation from the optimal grafting density. Detailed optimization is still ongoing in the authors’ lab. 4. Conclusions In summary, we have demonstrated a simple yet efficient approach to synthesize N/S-doped graphene electrocatalyst by using cysteine as an environmental-friendly N/S source. Interestingly, the cysteine-GO reaction time shows a significant influence on the ORR activity of as-prepared catalyst. The optimal N/S-G7d catalyst displays superior-to-Pt/C ORR performance in terms of half-wave potential, methanol tolerance and durability in alkaline electrolyte. Its limiting current density and four-electron reduction selectivity are close to that of Pt/C, thus exhibiting great promise as a highly efficient ORR electrocatalyst. It is further unveiled that the cysteine is grafted on GO via a unique process of oxidation-induced in situ disulfide dimer formation. This work suggests that the heteroatom source itself and careful optimization of reaction conditions are critical to obtain high performance doped-graphene electrocatalyst, and thus may offer a valuable clue on the rational design of metal-free ORR electrocatalysts. Acknowledgements We would like to gratefully acknowledge the financial support from National Natural Science Foundation of China (No.
H. Zhang et al. / Journal of Colloid and Interface Science 505 (2017) 32–37
21273173), Fundamental Research Funds for the Central Universities (XDJK2015B014), and Natural Science Foundation Project of CQ CSTC (cstc2016jcyjA0493). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcis.2017.05.069. References [1] L. Wang, H. Dong, Z. Guo, L. Zhang, T. Hou, Y. Li, J. Phys. Chem. C 120 (2016) 17427–17434. [2] L. Zhang, J. Niu, M. Li, Z. Xia, J. Phys. Chem. C 118 (2014) 3545–3553. [3] K. Gong, F. Du, Z. Xia, M. Durstock, L. Dai, Science 323 (2009) 760–764. [4] Y. Jiao, Y. Zheng, M. Jaroniec, S.Z. Qiao, J. Am. Chem. Soc. 136 (2014) 4394– 4403. [5] D. Guo, R. Shibuya, C. Akiba, S. Saji, T. Kondo, J. Nakamura, Science 351 (2016) 361–365. [6] X. Bai, Y. Shi, J. Guo, L. Gao, K. Wang, Y. Du, T. Ma, J. Power Sources 306 (2016) 85–91. [7] X. Chen, Phys. Chem. Chem. Phys. 17 (2015) 29340–29343. [8] V. Thi Huong, D. Shin, J. Jeong, J. Lee, J. Phys. Chem. C 120 (2016) 22342–22348. [9] A.Z. Yazdi, H. Fei, R. Ye, G. Wang, J. Tour, U. Sundararaj, A.C.S. Appl, Mater. Interfaces 7 (2015) 7786–7794. [10] J. Liang, Y. Jiao, M. Jaroniec, S.Z. Qiao, Angew. Chem. Int. Ed. 51 (2012) 11496– 11500. [11] J. Zhang, L. Qu, G. Shi, J. Liu, J. Chen, L. Dai, Angew. Chem. Int. Ed. 55 (2016) 2230–2234. [12] Y. Zheng, Y. Jiao, L. Ge, M. Jaroniec, S.Z. Qiao, Angew. Chem. Int. Ed. 52 (2013) 3110–3116. [13] Y. Hua, T. Jiang, K. Wang, M. Wu, S. Song, Y. Wang, P. Tsiakaras, Appl. Catal., B 194 (2016) 202–208.
37
[14] J.-S. Han, D.Y. Chung, D.-G. Ha, J.-H. Kim, K. Choi, Y.-E. Sung, S.-H. Kang, Carbon 105 (2016) 1–7. [15] W. Yan, X. Cao, J. Tian, C. Jin, K. Ke, R. Yang, Carbon 99 (2016) 195–202. [16] Y. Zhou, R. Ma, S.L. Candelaria, J. Wang, Q. Liu, E. Uchaker, P. Li, Y. Chen, G. Cao, J. Power Sources 314 (2016) 39–48. [17] E. Gracia-Espino, J. Phys. Chem. C 120 (2016) 27849–27857. [18] T. Sharifi, G. Hu, X. Jia, T. Wagberg, Acs Nano 6 (2012) 8904–8912. [19] L. Chen, Z. Chen, Z. Huang, Z. Huang, Y. Wang, H. Li, H. Zhou, Y. Kuang, J. Phys. Chem. C 119 (2015) 28757–28765. [20] K.G. Reddie, K.S. Carroll, Curr. Opin. Chem. Biol. 12 (2008) 746–754. [21] S. Akter, J. Huang, C. Waszczak, S. Jacques, K. Gevaert, F. Van Breusegem, J. Messens, J. Exp. Bot. 66 (2015) 2935–2944. [22] Y. Dong, H. Pang, H.B. Yang, C. Guo, J. Shao, Y. Chi, C.M. Li, T. Yu, Angew. Chem. Int. Ed. 52 (2013) 7800–7804. [23] D.L. Dewey, J. Beecher, Nature 206 (1965) 1369–1370. [24] J. Liu, S. Fu, B. Yuan, Y. Li, Z. Deng, J. Am. Chem. Soc. 132 (2010) 7279–7281. [25] H.R. Thomas, A.J. Marsden, M. Walker, N.R. Wilson, J.P. Rourke, Angew. Chem. Int. Ed. 53 (2014) 7613–7618. [26] X. Liu, W. Hu, Rsc Adv. 6 (2016) 29848–29854. [27] H. Zhang, X. Liu, G. He, X. Zhang, S. Bao, W. Hu, J. Power Sources 279 (2015) 252–258. [28] X. Hou, P. Zhang, S. Li, W. Liu, Phys. Chem. Chem. Phys. 18 (2016) 22799– 22804. [29] S. Zhang, H. Zhang, Q. Liu, S. Chen, J. Mater. Chem. A 1 (2013) 3302–3308. [30] S. Zhang, B. Liu, S. Chen, Phys. Chem. Chem. Phys. 15 (2013) 18482–18490. [31] Y. Niu, X. Huang, W. Hu, J. Power Sources 332 (2016) 305–311. [32] T. Liu, P. Zhao, X. Hua, W. Luo, S. Chen, G. Cheng, J. Mater. Chem. A 4 (2016) 11357–11364. [33] Y. Li, Y. Li, E. Zhu, T. McLouth, C.-Y. Chiu, X. Huang, Y. Huang, J. Am. Chem. Soc. 134 (2012) 12326–12329. [34] H. Wang, X. Bo, A. Wang, L. Guo, Electrochem. Commun. 36 (2013) 75–79. [35] H.-P. Cong, P. Wang, M. Gong, S.-H. Yu, Nano Energy 3 (2014) 55–63. [36] W. Hu, G. He, H. Zhang, X. Wu, J. Li, Z. Zhao, Y. Qiao, Z. Lu, Y. Liu, C.M. Li, Anal. Chem. 86 (2014) 4488–4493. [37] J. Wu, L. Zhao, X. Xu, N. Bertrand, W.I.I. Choi, B. Yameen, J. Shi, V. Shah, M. Mulvale, J.L. MacLean, O.C. Farokhzad, Angew. Chem. Int. Ed. 54 (2015) 9218– 9223.
Contents lists available at ScienceDirect
Journal of Colloid and Interface Science journal homepage: www.elsevier.com/locate/jcis
Regular Article
Nitrogen/sulfur-doping of graphene with cysteine as a heteroatom source for oxygen reduction electrocatalysis Huanhuan Zhang, Yanli Niu, Weihua Hu ⇑ Institute for Clean Energy & Advanced Materials, Chongqing Key Laboratory for Advanced Materials and Technologies of Clean Energies, Faculty of Materials & Energy, Southwest University, Chongqing, China
g r a p h i c a l a b s t r a c t
a r t i c l e
i n f o
Article history: Received 8 March 2017 Revised 18 April 2017 Accepted 20 May 2017 Available online 23 May 2017 Keywords: N/S-doped graphene Oxygen reduction reaction Cysteine Electrocatalyst Graphene oxide
a b s t r a c t Heteroatom-doped graphene have encouraged intensive research as promising metal-free oxygen reduction reaction (ORR) electrocatalysts but the correlation between the precursor material and final ORR activity remains unclear. In this work a serial of nitrogen/sulfur (N/S)-doped graphene catalysts were synthesized by modifying graphene oxide (GO) with cysteine as a N/S source and sequential thermal annealing. It is disclosed that the cysteine-GO reaction time shows a significant influence on the ORR activity of N/S-doped graphene. A unique process of oxidation-induced in situ disulfide formation is further found to be involved in the synthesis of optimal N/S-doped graphene, which displays ORR activity superior to commercial Pt/C in alkaline media. This work suggests that the heteroatom source itself and careful optimization of reaction conditions are critical to obtain high performance doped-graphene electrocatalyst. Ó 2017 Elsevier Inc. All rights reserved.
1. Introduction Heteroatom-doped carbon has emerged as a type of promising oxygen reduction reaction (ORR) catalysts since the pioneer work by Dai and coworkers in 2009 [1–3]. The enhanced ORR activity
⇑ Corresponding author. E-mail address: [email protected] (W. Hu). http://dx.doi.org/10.1016/j.jcis.2017.05.069 0021-9797/Ó 2017 Elsevier Inc. All rights reserved.
is believed to originate from heteroatom-disturbed electroneutrality of sp2 carbon in the graphitic framework to favor the adsorption of molecular oxygen and subsequent electrochemical reduction [2,4–8]. Dually-doped carbons were further explored as this strategy offers more flexibility to modulate the electronic structure by optimizing doping sites, dopant densities, and dopant ratio for possible synergistic effect [9–17]. Accumulating evidence has unveiled that both the doping density and doping site of the
H. Zhang et al. / Journal of Colloid and Interface Science 505 (2017) 32–37
33
Fig. 1. Schematic showing the synthesis of N/S co-doped graphene.
heteroatom are significantly important to the final ORR activity [5,18,19]. In most cases, heteroatom-doping of carbon was achieved by thermal annealing of a precursor material composed of a carbon host with modified heteroatom source. It is not fully understood yet regarding how heteroatoms are conjugated into different sites of the host carbon lattice during the high temperature process, which restricts the rational design and synthesis of doped carbon for ORR catalysis. Cysteine is a well-known sulfur-containing redox amino acid implicated in various biological systems [20,21]. It has been used as a precursor to synthesize N/S-doped carbon quantum dots [22]. Due to the presence of reductive thiol, cysteine possesses considerable reducing ability and is able to be oxidized by dissolved oxygen at neutral pH to form a cystine dimer [23]. At the same time, cysteine is able to bind to graphene oxide (GO) surface via electrostatic attraction and/or hydrophobic interaction [24] or via the ring-opening of epoxides on GO sheets [25]. Considering the abundant chemistry of cysteine and its various ways binding to GO, in this work we used it as a green N/S source to functionalize GO under different reaction condition, followed by annealing in an inert Argon atmosphere to synthesize N/S-doped graphene catalysts. Systematic characterizations and electrochemical comparisons unveiled that although N/S-doping is successfully achieved in all synthesized catalysts, their ORR activity varies significantly upon the change of the cysteine-GO reaction time. The optimal catalyst demonstrates ORR performance superior to commercial Pt/C in alkaline media. A unique process of oxidationinduced in situ disulfide formation with optimal grafting density is further disclosed to be critical in the synthesis of the optimal catalyst, as shown in Fig. 1. This work suggests that the heteroatom source itself and careful optimization of reaction conditions are critical to obtain high performance doped-graphene electrocatalyst, and thus may offer a valuable clue on the rational design of metal-free ORR electrocatalysts.
2. Experimental GO was prepared using the modified Hummers method [26,27]. Graphite powder (0.5 g) and sodium nitrate (NaNO3, 0.5 g) were mixed with concentrated sulfuric acid (H2SO4, 98%, 23 mL) and stirred for 30 min in an ice bath. Potassium permanganate (KMnO4, 3.0 g) was slowly added into the mixture with vigorous stirring. The mixture was further stirred for 1 h in a 35 °C water bath before water (40 mL) was slowly added under vigorous stirring. The mixture was stirred for 30 min at 90 °C. Finally, 100 mL of water was added to the mixture, followed by hydrogen peroxide (H2O2, 30%, 3.0 mL). After cooling down to room temperature, the products were collected by using centrifugation, and washed with 5 wt%
HCl and deionized water for several times. The yellow-brown suspension of GO was obtained by dispersing the products in water with the assistance of ultrasonication. To synthesize N/S-doped graphene, cysteine (2.0 mg mL1) was added to GO suspension (1.0 mg mL1) and the pH of the final solution was around 5.0–6.0, which was further adjusted to 8.0 with 0.1 M KOH solution. After stirring the mixture at ambient atmosphere for 7 d, the product was collected by centrifugation, freezing-dried, and annealed at 900 °C (Ar atmosphere, heating rate of 5 °C min1, maintaining at 900 °C for 3 h), as shown in Fig. 1. This catalyst was denoted as N/S-G-7d. For comparison, N/ S-G-1d, N/S-G-4d and N/S-G-14d were also obtained by adjusting the cysteine-GO reaction time to 1, 4 and 14 d, respectively. GO was also processed with the same annealing procedure to obtained thermally reduced GO (RGO). An Autolab PGSTAT302N potentiostat system was used for electrochemical measurements in a three-electrode cell with Pt foil as the counter electrode. Hg/HgO/1M KOH electrode was used as the reference electrode, to which all potentials reported in present work are referring. Also the potential could be easily converted to the one versus Reversible Hydrogen Electrode (ERHE) by using the Nernst equation:
ERHE ¼ EHg=HgO þ 0:059 pH þ E0Hg=HgO where the pH is the electrolyte pH (13 in this case) and the E0Hg=HgO is 0.098 V versus Normal Hydrogen Electrode (NHE). The working electrode was prepared by casting catalyst slurry (dispersing 2.0 mg of catalyst and 50 lL of 5 wt% Nafion solution in 1.0 mL of ethanol) onto glassy carbon electrode (2.0 mm diameter, 5 lL applied for cyclic voltammetry CV measurements, and 5.0 mm diameter, 25 lL applied for linear scanning voltammetry, LSV measurements) and the solvent was allow to naturally evaporate at room temperature. Before each experiment, the freshly prepared electrolyte (0.1 M KOH) was bubbled with pure N2 or O2 for 30 min. Fourier transform infrared (FTIR) spectra were measured on Nicolet FTIR 6700 spectrophotometer. X-ray photoelectron spectrum (XPS) was collected on an ESCALAB 250Xi system from Thermo Fisher. Transmission electron microscopy (TEM) and Scanning electron microscopy (SEM) images were obtained on a JEM2100 system and a JSM-7800F system from JOEL, respectively. Raman spectra were collected on a RENISHAW inVia Raman Microscope with 532 nm excitation. 3. Results and discussion The annealing temperature was first optimized and 900 °C was chosen as the optimal temperature for thermal treatment of the precursor, as shown in Fig. S1. The ORR activity of as-prepared N/
34
H. Zhang et al. / Journal of Colloid and Interface Science 505 (2017) 32–37
Fig. 2. Comparison of ORR activity of various catalysts in O2-saturated 0.1 M KOH. (a) LSV curves of N/S-G-1d, N/S-G-4d, N/S-G-7d, and N/S-G-14d; (b) LSV curves of RGO, N/SG-7d and Pt/C, scan rate of 5 mV s1 at 1600 rpm.
S-doped graphene with different cysteine-GO reaction time was evaluated with LSV measurement. As shown in Fig. 2a, the N/SG-7d sample exhibits a very positive onset potential of 0.09 V, which is more positive than that of other N/S-doped graphene samples tested (N/S-G-1d, N/S-G-4d, and N/S-G-14d). The limiting current density of N/S-G-7d sample is also highest among all four catalysts. The cysteine-GO reaction time shows a significant influence on the ORR activity of as-prepared catalyst, which may be tightly associated with the different bonding way between cys-
teine and GO and will be discussed in details below. When compared to commercial Pt/C (20 wt% Pt on Vulcan carbon XC-72) and thermally reduced GO (RGO), N/S-G-7d also demonstrates more positive onset potential and half-wave potential than commercial Pt/C with comparable limiting current density, indicating high ORR activity (Fig. 2b). LSV measurements at different rotating rates were carried out to reveal the ORR selectivity on N/S-G-7d sample. All electrocatalytic measurements are highly reproducible. As shown in
Fig. 3. ORR activity of N/S-G-7d in O2-saturated 0.1 M KOH. (a) rotating-disk voltammograms of N/S-G-7d at different rotating rates at a scan rate of 5 mV s1 and (b) corresponding K-L plots at different potentials; (c) RRDE voltammograms and (d) thereof calculated peroxide yield and electron-transfer-number of N/S-G-7d at a scan rate of 5 mV s1 at 1600 rpm.
35
H. Zhang et al. / Journal of Colloid and Interface Science 505 (2017) 32–37
Fig. 4. (a) FTIR spectra of GO, cystine and GO after reacting with cysteine for different time at pH 8.0; (b–d) high-resolution XPS spectra of C1s (b), N1s (c) and S2p (d) of N/SG-7d.
Table 1 Comparison of doping densities as determined by XPR and ORR performance of various N/S-doped graphene catalysts synthesized in present work.
*
No.
Sample
Condition
N at.%
S at.%
S/N ratio
1 2 3 4 5 6 7 8
N/S-G-7d N/S-G-1d N/S-G-4d N/S-G-14d
GO + cysteine@pH 8.0, 7 d GO + cysteine@pH 8.0, 1 d GO + cysteine@pH 8.0, 4 d GO + cysteine@pH 8, 14 d GO + cysteine@pH 11, 7 d GO + cysteine@pH 6, 7 d GO + cystine, pH 12.5–8.0, 1 d GO + cystine, pH 1.5–8.0, 1 d
1.02 0.80
1.32 0.52
1.29 0.65
2.69 2.38
1.96 1.65
0.73 0.69
1.95
1.39
0.71
J @-0.2 V (mA cm2)
E1/2* (V)
3.361 1.242 2.937 2.398 2.497 2.078 2.545 2.546
0.093 0.224 0.122 0.127 0.146 0.164 0.154 0.166
Half wave potential (E1/2) is defined as the potential at which the current density reaches half of the current density at 0.6 V for convenience.
Fig. 3a, ORR current density is improved with the increase of rotating rate from 400 to 2500 rpm. The Koutecky–Levich (K-L) plots of N/S-G-7d catalyst exhibit good linearity at different potentials, as shown in Fig. 3b. The electron-transfer number (n) was calculated to vary from 3.47 to 3.72 in the range of 0.30 to 0.70 V according to the K-L plots [7,28], indicating that the detrimental twoelectron ORR process is efficiently inhibited on N/S-G-7d. The ORR pathway on N/S-G-7d was further investigated by rotating ring-disk electrode (RRDE) measurements at 1600 rpm. Fig. 3c exhibits high disk current (Id) while low ring current (Ir),
suggesting negligible hydrogen peroxide intermediate generated and excellent four-electron selectivity on N/S-G-7d [29–31]. The hydrogen peroxide yield and overall electron-transfer-number were calculated and shown in Fig. 3d by using Ir, Id and collecting efficiency of 37% [3,32]. The yield of peroxide species is from 3.7% to 16.8% when the potential was swept from 0.2 to 0.75 V. The electron-transfer-number ranges from 3.66 to 3.92 depending on the potential, which is in good agreement with the values calculated from K-L plots, indicating excellent selectivity for favorable four-electron pathway of N/S-G-7d. According to accelerated
36
H. Zhang et al. / Journal of Colloid and Interface Science 505 (2017) 32–37
Fig. 5. Comparison of ORR activity of different N/S-doped graphene catalysts. (a) LSV curves of N/S-doped graphene prepared by annealing cystine-grafted GO at 900 °C. The cystine, instead of cysteine, was grafted by slowly adjusting solution pH with 0.1 M KOH or HCl from 12.5 or 1.5 to 8.0 in 24 h. (b) LSV curves of N/S-doped graphene prepared with the same procedure as for N/S-G-7d except for that the pH value of cysteine-GO reaction solution was 6, or 11, but not optimal 8.0. Scan rate: 5 mV s1 at 1600 rpm in O2-saturated 0.1 M KOH.
durability test (ADT) and chronoamperometric curve of methanol addition, N/S-G-7d also exhibits excellent stability and good methanol tolerance, both of which are highly desirable for fuel cell applications, as shown in Fig. S2 [33,34]. To disclose the underlying mechanism under the excellent ORR activity of N/S-G-7d, the bonding between GO and cysteine for different reaction time was investigated by FTIR. As shown in Fig. 4a, pristine GO shows its characteristic absorption peaks originating from C@O, C@C, OAH, and CAO stretching vibrations (1732, 1629, 1228, 1071 cm1) [35,36]. After reacting with cysteine for 1 d, the characteristic peak of carboxyl group at 1732 cm1 disappears, indicating partial reduction of GO sheets by cysteine [24]. Longer reaction results in newly emerged bands at 1586, 1488, 1409, 1383, and 1084 cm1, all of which are in good agreement with that of cystine, the disulfide dimer of cysteine after being oxidized. This suggests that the cysteine was grafted onto GO in the form of disulfide dimer [23]. It has been reported that in the presence of dissolved oxygen, cysteine is able to be spontaneously oxidized at neutral pH to form cystine dimer [23], whose solubility is quite low in neutral pH [37]. Therefore, it is speculated that when the reaction proceeds for short time, e.g., 1 d, cysteine binds to GO surface via electrostatic attraction and/or hydrophobic interaction [24] or the ring-opening of epoxides [25]. When prolonged to, e.g., 4 d or 7 d, these pre-bound cysteine act as seeds to initiate the formation of cystine dimer on the nanosheets [24]. Together with the LSV curves in Fig. 2a, these FTIR spectra imply that the cystine dimer is more efficient than cysteine as N/S source for graphene doping. The doping nature of N/S-G-7d was analyzed by XPS and Raman. XPS spectra confirm the presence of N and S (atomic abundance of 1.02% and 1.32%, respectively, as shown in survey spectrum in Fig. S4). The high-resolution C1s, N1s and S2p spectra were shown in Fig. 4b–d, suggesting successful doping in graphene framework (details about XPS analysis were provided in Supplementary material). Remarkably, in other two samples tested (N/S-G-1d, and N/S-G-14d, respectively), successful doping is also achieved, as unveiled by their XPS spectra in Figs. S5 and S6. In N/S-G-7d the S atomic percentage is higher than that of N but it is not the case for other catalysts, as summarized in Table 1. More experimental evidences however, are needed to directly relate the ORR activity to the S/N doping ratio in this class of catalysts. On Raman spectra (Fig. S7), the G band at 1594 cm1 is induced by sp2 carbon atoms and the D band at 1348 cm1 is assigned to defects in graphene sheets. The intensity ratio (ID/IG) of the D and G bands for RGO is 0.96, and it
increases to 1.06 for N/S-G-7d, suggesting successful heteroatom doping of graphene. The unique dimer-mediated grafting for N/S-G-7d inspires us to directly graft disulfide cystine on GO to synthesize N/S-doped graphene. As the solubility of cystine is much higher in acidic or alkaline solution than that in neutral pH [37], we successfully achieved cystine grafting via adjusting solution pH, as confirmed by FTIR in Fig. S8. XPS confirms the successful N/S-doping and the doping density is even higher than that of optimal catalyst (Fig. S9 and Table 1). Electrochemical evaluation, however, suggests that the resultant N/S catalyst is still inferior to the optimal N/S-G-7d catalyst, as shown in Fig. 5a. Together with the comparison in Fig. 2a, the results may suggest that there is an optimal grafting density for cystine on GO, which is easily achieved via in situ cystine dimer formation process rather than via pH-controlled dimer grafting method. The same reason may also lie behind the experimental observation that higher or lower pH value (Fig. 5b) results in inferior ORR activity because the formation rate of disulfide dimer is sensitive to pH and thus induces deviation from the optimal grafting density. Detailed optimization is still ongoing in the authors’ lab. 4. Conclusions In summary, we have demonstrated a simple yet efficient approach to synthesize N/S-doped graphene electrocatalyst by using cysteine as an environmental-friendly N/S source. Interestingly, the cysteine-GO reaction time shows a significant influence on the ORR activity of as-prepared catalyst. The optimal N/S-G7d catalyst displays superior-to-Pt/C ORR performance in terms of half-wave potential, methanol tolerance and durability in alkaline electrolyte. Its limiting current density and four-electron reduction selectivity are close to that of Pt/C, thus exhibiting great promise as a highly efficient ORR electrocatalyst. It is further unveiled that the cysteine is grafted on GO via a unique process of oxidation-induced in situ disulfide dimer formation. This work suggests that the heteroatom source itself and careful optimization of reaction conditions are critical to obtain high performance doped-graphene electrocatalyst, and thus may offer a valuable clue on the rational design of metal-free ORR electrocatalysts. Acknowledgements We would like to gratefully acknowledge the financial support from National Natural Science Foundation of China (No.
H. Zhang et al. / Journal of Colloid and Interface Science 505 (2017) 32–37
21273173), Fundamental Research Funds for the Central Universities (XDJK2015B014), and Natural Science Foundation Project of CQ CSTC (cstc2016jcyjA0493). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcis.2017.05.069. References [1] L. Wang, H. Dong, Z. Guo, L. Zhang, T. Hou, Y. Li, J. Phys. Chem. C 120 (2016) 17427–17434. [2] L. Zhang, J. Niu, M. Li, Z. Xia, J. Phys. Chem. C 118 (2014) 3545–3553. [3] K. Gong, F. Du, Z. Xia, M. Durstock, L. Dai, Science 323 (2009) 760–764. [4] Y. Jiao, Y. Zheng, M. Jaroniec, S.Z. Qiao, J. Am. Chem. Soc. 136 (2014) 4394– 4403. [5] D. Guo, R. Shibuya, C. Akiba, S. Saji, T. Kondo, J. Nakamura, Science 351 (2016) 361–365. [6] X. Bai, Y. Shi, J. Guo, L. Gao, K. Wang, Y. Du, T. Ma, J. Power Sources 306 (2016) 85–91. [7] X. Chen, Phys. Chem. Chem. Phys. 17 (2015) 29340–29343. [8] V. Thi Huong, D. Shin, J. Jeong, J. Lee, J. Phys. Chem. C 120 (2016) 22342–22348. [9] A.Z. Yazdi, H. Fei, R. Ye, G. Wang, J. Tour, U. Sundararaj, A.C.S. Appl, Mater. Interfaces 7 (2015) 7786–7794. [10] J. Liang, Y. Jiao, M. Jaroniec, S.Z. Qiao, Angew. Chem. Int. Ed. 51 (2012) 11496– 11500. [11] J. Zhang, L. Qu, G. Shi, J. Liu, J. Chen, L. Dai, Angew. Chem. Int. Ed. 55 (2016) 2230–2234. [12] Y. Zheng, Y. Jiao, L. Ge, M. Jaroniec, S.Z. Qiao, Angew. Chem. Int. Ed. 52 (2013) 3110–3116. [13] Y. Hua, T. Jiang, K. Wang, M. Wu, S. Song, Y. Wang, P. Tsiakaras, Appl. Catal., B 194 (2016) 202–208.
37
[14] J.-S. Han, D.Y. Chung, D.-G. Ha, J.-H. Kim, K. Choi, Y.-E. Sung, S.-H. Kang, Carbon 105 (2016) 1–7. [15] W. Yan, X. Cao, J. Tian, C. Jin, K. Ke, R. Yang, Carbon 99 (2016) 195–202. [16] Y. Zhou, R. Ma, S.L. Candelaria, J. Wang, Q. Liu, E. Uchaker, P. Li, Y. Chen, G. Cao, J. Power Sources 314 (2016) 39–48. [17] E. Gracia-Espino, J. Phys. Chem. C 120 (2016) 27849–27857. [18] T. Sharifi, G. Hu, X. Jia, T. Wagberg, Acs Nano 6 (2012) 8904–8912. [19] L. Chen, Z. Chen, Z. Huang, Z. Huang, Y. Wang, H. Li, H. Zhou, Y. Kuang, J. Phys. Chem. C 119 (2015) 28757–28765. [20] K.G. Reddie, K.S. Carroll, Curr. Opin. Chem. Biol. 12 (2008) 746–754. [21] S. Akter, J. Huang, C. Waszczak, S. Jacques, K. Gevaert, F. Van Breusegem, J. Messens, J. Exp. Bot. 66 (2015) 2935–2944. [22] Y. Dong, H. Pang, H.B. Yang, C. Guo, J. Shao, Y. Chi, C.M. Li, T. Yu, Angew. Chem. Int. Ed. 52 (2013) 7800–7804. [23] D.L. Dewey, J. Beecher, Nature 206 (1965) 1369–1370. [24] J. Liu, S. Fu, B. Yuan, Y. Li, Z. Deng, J. Am. Chem. Soc. 132 (2010) 7279–7281. [25] H.R. Thomas, A.J. Marsden, M. Walker, N.R. Wilson, J.P. Rourke, Angew. Chem. Int. Ed. 53 (2014) 7613–7618. [26] X. Liu, W. Hu, Rsc Adv. 6 (2016) 29848–29854. [27] H. Zhang, X. Liu, G. He, X. Zhang, S. Bao, W. Hu, J. Power Sources 279 (2015) 252–258. [28] X. Hou, P. Zhang, S. Li, W. Liu, Phys. Chem. Chem. Phys. 18 (2016) 22799– 22804. [29] S. Zhang, H. Zhang, Q. Liu, S. Chen, J. Mater. Chem. A 1 (2013) 3302–3308. [30] S. Zhang, B. Liu, S. Chen, Phys. Chem. Chem. Phys. 15 (2013) 18482–18490. [31] Y. Niu, X. Huang, W. Hu, J. Power Sources 332 (2016) 305–311. [32] T. Liu, P. Zhao, X. Hua, W. Luo, S. Chen, G. Cheng, J. Mater. Chem. A 4 (2016) 11357–11364. [33] Y. Li, Y. Li, E. Zhu, T. McLouth, C.-Y. Chiu, X. Huang, Y. Huang, J. Am. Chem. Soc. 134 (2012) 12326–12329. [34] H. Wang, X. Bo, A. Wang, L. Guo, Electrochem. Commun. 36 (2013) 75–79. [35] H.-P. Cong, P. Wang, M. Gong, S.-H. Yu, Nano Energy 3 (2014) 55–63. [36] W. Hu, G. He, H. Zhang, X. Wu, J. Li, Z. Zhao, Y. Qiao, Z. Lu, Y. Liu, C.M. Li, Anal. Chem. 86 (2014) 4488–4493. [37] J. Wu, L. Zhao, X. Xu, N. Bertrand, W.I.I. Choi, B. Yameen, J. Shi, V. Shah, M. Mulvale, J.L. MacLean, O.C. Farokhzad, Angew. Chem. Int. Ed. 54 (2015) 9218– 9223.