Metal–nitrogen (Co-g-C₃N₄) doping of surface-modified single-walled carbon nanohorns for use as an oxygen reduction electrocatalyst

Abstract

A cobalt-doped graphitic carbon nitride (g-C₃N₄) polymer was supported on surface-modified single-walled carbon nanohorns (SWCNHs) to produce a new Co-g-C₃N₄ catalyst for the oxygen reduction reaction (ORR). X-ray photoelectron spectroscopy results show that Co–N bonds are formed and that strong electronic coupling occurs between the SWCNHs and Co-g-C₃N₄. The as-fabricated catalysts exhibit reasonable ORR activity, high selectivity, and outstanding electrochemical stability in alkaline media. These results are attributed to the large number of stable Co–N_x active sites, high percentage of pyridinic N and rapid mass transport on the “Dahlia flower”-like SWCNHs surface. The fabricated Co-g-C₃N₄/SWCNHs catalyst has potential for use as a non-precious metal cathodic catalyst in fuel cell applications.

---

1. Introduction

The slow kinetics of the oxygen reduction reaction (ORR) is the main factor preventing fuel cells from becoming an efficient, environmentally friendly power source. Commercial platinum-based electrocatalysts are currently considered to be the most effective cathodic catalysts, but the large-scale commercialization of these catalysts is hindered by their high cost and inferior durability.^1–3 Consequently, extensive efforts are underway to develop cheap electrocatalysts with high activity and durability as alternatives to precious metal catalysts for ORR.

Of the non-precious metal electrocatalysts employed in ORR, heteroatom-doped carbon materials, especially N-doped carbon materials, are a promising alternative.^4–6 These materials perform comparably to Pt/C in ORR in alkaline media because of their distinct electronic properties and structural features. Recently, graphitic carbon nitride (g-C₃N₄) has attracted considerable attention as a non-precious metal catalyst for ORR. In particular, its graphene-like sp² bonding structure and the abundant in-plane nitrogen dopants and defects makes its electronic structure attractive for ORR. Nevertheless, the widespread use of g-C₃N₄ is limited by two problems: the poor electroconductivity and low specific surface area of bulk g-C₃N₄ hinder electron transport during the ORR process. Therefore, researchers have combined carbon materials (carbon black, mesoporous carbon, and graphene)^7–9 to improve the electrical conductivity, which leads to greatly enhanced catalytic activity. These materials also have more active sites than other N–carbon materials. However, the specific surface areas of these different carbon matrices are still too small, and their surface smoothness leads to interactions between the active sites and carrier that are too weak; therefore, their mass activities are much lower than those of Pt-based catalysts.

Chen et al. reported the doping of g-C₃N₄ with transition metals. The rich pyridine-like nitrogen in g-C₃N₄ can trap many transition metal atoms to form potential active sites.^10–12 The formation of metal–nitrogen bonds (Me–N_x) is crucial for reducing the oxygen molecule, and metal doping is thus used to enhance the ORR catalytic activity of g-C₃N₄. Two key factors dominate the performance of non-precious metal catalysts: (1) the interactions between different transition metals and heteroatom compositions, which control the inherent nature of the active sites, and (2) the accessible surface area and porosity, which control the number of accessible active sites and the transport of ORR-relevant species.^13,14 Therefore, this work aims to synthesize catalysts with higher active sites (Me–N_x) concentrations using suitable nitrogen/non-precious metal precursors and carbon supports, thereby improving the catalyst activity and durability.

Single-walled carbon nanohorns (SWCNHs) are a new type of carbon material synthesized by the arc discharge method.¹⁵ These materials are desirable over other carbon morphologies because of their easily tunable surface area and microporosity. The most advantageous characteristic of this material is its distinct “Dahlia flower”-like morphology formed via aggregation of the angled structure. This morphology results in more defects and an open architecture, which are beneficial for supporting dispersed metal particles. More number of Me–N_x sites can be obtained by tuning the surface area and microporosity to achieve higher activity and durability without sacrificing the physical and chemical characteristics of the carbon matrix.^16,17 Hence, several previous studies demonstrated the excellent electrochemical performance of SWCNHs doped with noble metals (Pt, Ru, and Se).^18,19 To facilitate charge transfer/gas diffusion and increase the active site density, SWCNHs were used as a support for Co-doped g-C₃N₄, which was employed due to the doping effect of the transition metal Co and the good catalytic performance of g-C₃N₄ in ORR. Therefore, the SWCNHs and Co-g-C₃N₄ structures were combined to exploit their unique properties to obtain a composite material with the desired properties. The resulting Co-g-C₃N₄/SWCNHs material is shown to be an active, cost-effective, robust cathodic catalyst in alkaline fuel cells.

2. Experimental section

2.1. Materials

SWCNHs were purchased from XFNANO. Dicyandiamide (DCD), Co(OAc)₂·4H₂O, and carbon black (Vulcan XC-72R) were obtained from Cabot Corporation.

2.2. Catalyst preparation

The SWCNHs were thermally treated at 400 °C for 4 h in air to remove amorphous carbon impurities and to open the nanohorns.²⁰ The Co-g-C₃N₄/SWCNHs material was synthesized as follows. A preprocessed SWCNHs water suspension (100 mg, 10 mg mL⁻¹) was treated ultrasonically for 30 min. Subsequently, a solution of DCD (200 mg) and Co(OAc)₂·4H₂O (30 mg) was slowly added to the suspension. The resulting mixture was mechanically stirred at 80 °C for 10 h and then dried overnight to obtain a black solid. Finally, the sample was annealed at 600 °C under flowing Ar for 5 h, and the heat-treated sample was collected.

For comparison, g-C₃N₄/SWCNHs, g-C₃N₄/C, and Co-g-C₃N₄/C were prepared via the same process.

2.3. Catalyst characterization

Powder X-ray diffraction (XRD) data were collected using a D8 ADVANCE X-ray diffractometer (Bruker Biosciences Corporation, USA) with Cu Kα radiation (λ = 0.15406 nm) at 40 kV and 40 mA. Transmission electron microscopy (TEM) was performed using a Tecnai F30 field emission transmission electron microscope (Hillsboro, OR, USA). X-ray photoelectron spectroscopy (XPS) data were collected using an ESCA 3400 instrument (Kratos Analytical Ltd., Manchester, UK). Raman spectra analyses were performed using a Raman spectrometer (Renishaw) with a 514.3 nm Ar laser.

2.4. Electrocatalytic measurements

The electrochemical properties were evaluated using an electrochemical work station (CHI760D) and rotating disk electrode (RDE) instrument with a three-electrode test cell. Ag/AgCl (saturated KCl), modified glass carbon (GC) (3 mm in diameter), and a Pt pole were used as the reference, working, and counter electrodes, respectively. The catalyst-coated working electrode was prepared as follows. Approximately 5 mg of the sample were dissolved in 1 mL of ethanol, and 50 μL of 5% Nafion were added to give a homogeneous suspension after ultrasonication for 30 min. Approximately 20 μL of the Nafion catalyst ink were loaded onto a GC electrode surface and directly dried at room temperature.

Cyclic voltammetry (CV), linear sweep voltammetry (LSV), and stability experiments were performed in an N₂- or O₂-saturated 0.1 M KOH solution at different sweep rates.

3. Results and discussion

The TEM images of all of the SWCNHs, g-C₃N₄/SWCNHs, and Co-g-C₃N₄/SWCNHs catalysts are shown in Fig. 1. The distinct “Dahlia flower”-like morphology of the SWCNHs, which is not affected by the high-temperature treatment to produce the g-C₃N₄/SWCNHs and Co-g-C₃N₄/SWCNHs, is easily observed (Fig. 1a–d). Cobalt oxide nanoparticles are also clearly observed in Fig. 1c and d. This result conflicts with those in the literature²¹ and might be attributed the fact that only some of the Co atoms are into doped in g-C₃N₄, whereas the rest form CoO_x. In order to prove that Co can be doped into g-C₃N₄ with excellent distribution, Co and N element mapping images were obtained and are shown in Fig. 1e–g. It was found that the Co atoms are often adjacent or close to N atoms, suggesting that Co interacts with g-C₃N₄ by forming Co–N bonds.

Fig. 1 HR-TEM images of the (a) SWCNHs, (b) g-C₃N₄/SWCNHs, and (c and d) Co-g-C₃N₄/SWCNHs. (e and f) Corresponding Co and N elemental maps. (g) An overlaid Co and N elemental maps.

The sample crystal structures were characterized by XRD, and the results are shown in Fig. 2a. Two obvious diffraction peaks corresponding to the characteristic peaks of g-C₃N₄ are observed at 27.71° and 43.63°, suggesting that g-C₃N₄ is present in all of the synthesized catalysts.^7,22 The Raman results confirm that the Co is successfully doped into the g-C₃N₄ polymer. Fig. 2b shows that the I_D/I_G ratio for the SWCNHs, g-C₃N₄/SWCNHs and Co-g-C₃N₄/SWCNHs samples are 1.24, 1.49 and 1.75, respectively, indicating that the SWCNHs in the Co-g-C₃N₄/SWCHNs have the most defects. After the SWCNHs are coupled to the undoped g-C₃N₄, the G band hardly shifts. In contrast, when Co-g-C₃N₄ is supported on the SWCNHs, G band of the SWCNHs is obviously red-shifted, suggesting that electron transfer occurs between the Co species and g-C₃N₄ matrix.²³ This result indicates that the Co species is chemically coordinated to the g-C₃N₄ matrix through Co–N bonds. The explicit structure of Co-g-C₃N₄ was shown in Fig. 2c.

Fig. 2 (a) XRD patterns and (b) Raman spectra of the as-prepared SWCNHs, g-C₃N₄/SWCNHs and Co-g-C₃N₄/SWCNHs catalysts. (c) The structure of the Co-g-C₃N₄ complex.

XPS was employed to determine the catalyst chemical composition and elemental content as well as the states of the cobalt and nitrogen atoms in the prepared catalysts. In the XPS survey spectrum of the Co-g-C₃N₄/SWCNHs in Fig. 3a, C_1s, O_1s, N_1s, and Co_2p signals are observed, proving that C, O, N, and Co are present in the catalyst. Table 1 lists the elemental contents of the catalysts. After introducing Co into the SWCNHs, the N doping level increases from 7.17% for the g-C₃N₄/SWCNHs to 8.08% for the Co-g-C₃N₄/SWCNHs, indicating that Co and N interact. Moreover, the Co-g-C₃N₄/SWCNHs capture more N atoms to create more Co–N_x species, which might be beneficial for the ORR catalytic activity.

Fig. 3 (a) XPS survey scan and (b and c) high-resolution Co spectra of the Co-g-C₃N₄/C and Co-g-C₃N₄/SWCNHs catalysts.

Table 1 Elemental analysis (%) for the synthesized catalysts (XPS)

Catalyst	C	N	O	Co
g-C3N4/C	81.89	6.89	11.22
Co-g-C3N4/C	76.56	7.15	10.40	5.89
g-C3N4/SWCNHs	81.71	7.17	11.12
Co-g-C3N4/SWCNHs	76.21	8.08	9.52	6.19

---

High-resolution surveys of the metal compounds demonstrate the presence of various Co species (Fig. 3b and c). The high-resolution Co_2p spectra of the catalysts can be deconvoluted into two peaks with binding energies of 782 and 780.2 eV corresponding to Co–N_x and CoO_x, respectively.^24–26 As discussed previously, CoO_x nanoparticles are observed in the TEM images in Fig. 1. These results provide further evidence that not only does Co coordinate to the g-C₃N₄ matrix by Co–N bonding, but it is also present in the form of CoO_x in the synthesized catalysts.

The N_1s spectra of all of the N-coordinated metal catalysts have four component peaks (Fig. 4), namely the pyridinic N (398.2 eV), pyrrolic N (399.6 eV), graphitic N (401.2 eV), and N oxides (404.2 eV) peaks.^27–29 The percentages of the various nitrogen types are listed in Table 2. It was reported that pyridinic N, pyrrolic N and graphitic N all exhibit high ORR catalytic activities, and of these N species, pyridinic N has the highest activity.²⁶ As shown in Table 2, the Co-g-C₃N₄/SWCNHs catalyst has the highest percentage of pyridinic N, indicating that it might exhibit excellent ORR catalytic activity.

Fig. 4 High-resolution N_1s spectra for the (a) g-C₃N₄/C, (b) Co-g-C₃N₄/C, (c) g-C₃N₄/SWCNHs and (d) Co-g-C₃N₄/SWCNHs catalysts.

Table 2 Relative contents of the nitrogen species in the catalysts (XPS)

Catalyst	Percentage of nitrogen type
	Pyridinic N	Pyrrolic N	Graphitic N	N oxides
g-C3N4/C	20.2	47.2	24.7	7.9
Co-g-C3N4/C	41.4	19.0	34.5	5.2
g-C3N4/SWCNHs	19.4	41.7	25.0	13.9
Co-g-C3N4/SWCNHs	43.7	17.0	33.2	6.1

---

The ORR catalytic activity of the Co-g-C₃N₄/SWCNHs was investigated by CV. As shown in Fig. 5a, when the electrolyte solution is saturated with N₂, the cathodic current in the CV has a featureless slope. In contrast, a well-defined cathodic peak appears in the CV when the electrolyte solution is saturated with O₂, suggesting that the Co-g-C₃N₄/SWCNHs exhibit significant catalytic activity. The electrocatalytic properties of the as-prepared catalysts were compared using the RDE curves obtained in an O₂-saturated 0.1 M KOH solution at a sweep rate of 5 mV s⁻¹, rotation rate of 1600 rpm and room temperature. The corresponding Koutecky–Levich (K–L) plots were obtained by collecting the ORR polarization curves at different rotation rates. According to K–L theory,³⁰ the electron transfer numbers (n) for a reaction can be calculated using the following equations:

where J and J_k are the measured and kinetic-limiting current densities, respectively; F is the Faraday constant; ω is the electrode rotation rate (rpm); n is the electron transfer number; C_O2 is the bulk O₂ concentration; ν is the kinematic viscosity of the KOH electrolyte; and D_O2 is the O₂ diffusion coefficient in 0.1 M KOH.

Fig. 5 (a) CV curves of the Co-g-C₃N₄/SWCNHs catalyst obtained at a scan rate of 50 mV s⁻¹ in O₂- (black) or N₂-saturated (red) 0.1 M KOH. (b) ORR polarization curves for the SWCNHs, g-C₃N₄/SWCNHs, acid-treated Co-g-C₃N₄/SWCNHs and Co-g-C₃N₄/SWCNHs catalysts obtained at a sweep rate of 5 mV s⁻¹ and rotation rate of 1600 rpm in O₂-saturated 0.1 M KOH. (c) Levich plots for the ORR on the prepared catalysts. The current density was measured at −0.4 V (vs. Ag/AgCl). (d) Corresponding RDE curves for the Co-g-C₃N₄/SWCNHs and Co-g-C₃N₄/C catalysts obtained at a sweep rate of 5 mV s⁻¹ and rotation rate of 1600 rpm in O₂-saturated 0.1 M KOH. (e) Levich plots for the ORR on the prepared Co-g-C₃N₄/SWCNHs and Co-g-C₃N₄/C catalysts. The current density was measured at −0.4 V (vs. Ag/AgCl).

As shown in Fig. 5b, the half-wave potentials (E_1/2) of the Co-g-C₃N₄/SWCNHs, g-C₃N₄/SWCNHs and SWCNHs are −0.157, −0.184, and −0.197 V, respectively, and their electron transfer numbers are 3.98, 3.42 and 2.78, respectively. The g-C₃N₄/SWCNHs exhibit higher catalytic activity than the SWCNHs, indicating that the g-C₃N₄ matrix has ORR active sites, which is consistent with literature results.⁷ However, it should be noted that both the E_1/2 and n values of the Co-g-C₃N₄/SWCNHs are higher than those of the g-C₃N₄/SWCNHs, showing that the presence of Co is important for achieving high catalytic activity. As shown by the TEM and XPS results, the synthesized Co-g-C₃N₄/SWCNHs catalyst contains both CoO_x and Co–N_x. It was reported that both of these species can act as effective ORR catalysts.^31,32 To elucidate the effect of Co doping on the Co-g-C₃N₄/SWCNHs catalytic activity, the synthesized catalyst was treated in 0.1 M H₂SO₄ at 60 °C for 6 h to remove the CoO_x species from the catalyst, and the ORR catalytic activity was subsequently evaluated under the same conditions. The acid-treated Co-g-C₃N₄/SWCNHs catalyst exhibits higher activity than the g-C₃N₄/SWCNHs and lower activity than the Co-g-C₃N₄/SWCNHs catalyst. This result indicates that doping Co into the g-C₃N₄ matrix can effectively enhance its ORR catalytic activity. Fig. 4d and 5b show that Co–N_x is present in the Co-g-C₃N₄/SWCNHs catalyst and that the percentage of pyridinic N in this catalyst is higher. Jiang et al. reported that cobalt species that form Co–N_x sites in the carbon shells are ORR active sites.³² Nonetheless, pyridinic N is more active for ORR than pyrrolic N and graphitic N. Therefore, it can be concluded that the improved catalytic activity of the Co-g-C₃N₄/SWCNHs catalyst is due to the presence of both Co–N bonds and pyridinic N.

To determine the effect of the SWCNHs support on the Co-g-C₃N₄ active species, the ORR catalytic activity of the Co-g-C₃N₄/SWCNHs is compared to that of the Co-g-C₃N₄/C catalyst in Fig. 5d. The Co-g-C₃N₄/SWCNHs exhibit higher ORR catalytic activity than the Co-g-C₃N₄/C catalyst as shown by their onset potentials (E₀), half-wave potentials (E_1/2), and electron transfer numbers (Fig. 5e). A comparison of these results to previous literature results shows that the onset potential (−0.066 V) of the Co-g-C₃N₄/SWCNHs is more positive than those of Co–N-GN (−0.098 V) and Co/PANI-PPYR (−0.13) catalysts.^33,34 These results suggest that the SWCNHs act as a synergistic, conductive support with a high surface area and microporosity, which allow the creation of more metal active sites and provide the reactants with sufficient access to the active sites to enhance the charge and mass transfer. Therefore, the catalyst and support exhibit beneficial chemical interactions that improve the catalyst efficiency and prevent the loss of active sites.

The catalyst stability was tested by cycling over the potential range from −0.8 V to 0.2 V in O₂-saturated 0.1 M KOH. The ORR current densities on the catalyst surface were recorded at −0.16 V after 1200 cycles (12 h). Compared to the Co-g-C₃N₄/C and commercial Pt/C catalysts, which exhibit a rapid current loss of approximately 20% after 4 h, the Co-g-C₃N₄/SWCNHs catalyst has a much more stable performance in alkaline media (Fig. 6), demonstrating that the Co-g-C₃N₄ and SWCNHs are strongly coupled in the Co-g-C₃N₄/SWCNHs catalyst.³³

Fig. 6 Effect of the scanning cycles on the ORR current density of the Co-g-C₃N₄/SWCNHs, Co-g-C₃N₄/C and Pt/C catalysts. The current density was measured at −0.15 V (vs. Ag/AgCl).

4. Conclusions

In summary, a Co-g-C₃N₄/SWCNHs material was synthesized via polycondensation and shown to exhibit reasonable ORR catalytic activity and outstanding electrochemical stability in alkaline media. The excellent ORR performance of this catalyst is attributed to its high surface area and microporosity, which provides the reactants with suitable access to the active sites to enhance the charge and mass transfer. Furthermore, more Co–N_x active sites are homogeneously distributed on the distinct flower-like SWCNHs structure, and the percentage of pyridinic N is higher. It is expected that Co-g-C₃N₄/SWCNHs can act as efficient non-precious metal catalysts and should be further developed in the future.

Acknowledgements

This work was supported by the National Natural Science Funds of China (NSFC, 21366027), the Doctor Foundation of Bingtuan (2013BB010), the Young Scientific and Technological Innovation Leaders of Bingtuan (2015BC001), and the Foundation of Young Scientists at Shihezi University (No. 2013ZRKXJQ03).

References

1. D. Wang, H. L. Xin, R. Hovden, H. Wang, Y. Yu, D. A. Muller, F. J. DiSalvo and H. D. Abruña, Nat. Mater., 2013, 12, 81–87.
2. S. M. Alia, K. O. Jensen, B. S. Pivovar and Y. Yan, ACS Catal., 2012, 2, 858–863.
3. J.-G. Oh, H.-S. Oh, W. H. Lee and H. Kim, J. Mater. Chem., 2012, 22, 15215–15220.
4. D. S. Su, S. Perathoner and G. Centi, Chem. Rev., 2013, 113, 5782–5816.
5. M. Liu, R. Zhang and W. Chen, Chem. Rev., 2014, 114, 5117–5160.
6. K. Gong, F. Du, Z. Xia, M. Durstock and L. Dai, Science, 2009, 323, 760–764.
7. J. Liang, Y. Zheng, J. Chen, J. Liu, D. Hulicova-Jurcakova, M. Jaroniec and S. Z. Qiao, Angew. Chem., Int. Ed., 2012, 51, 3892–3896.
8. K. Kwon, Y. J. Sa, J. Y. Cheon and S. H. Joo, Langmuir, 2012, 28, 991–996.
9. L. Xu, W.-Q. Huang, L.-L. Wang, Z.-A. Tian, W. Hu, Y. Ma, X. Wang, A. Pan and G.-F. Huang, Chem. Mater., 2015, 27, 1612–1621.
10. X. Wang, X. Chen, A. Thomas, X. Fu and M. Antonietti, Adv. Mater., 2009, 21, 1609–1612.
11. Y. Di, X. Wang, A. Thomas and M. Antonietti, ChemCatChem, 2010, 2, 834–838.
12. X. Chen, J. Zhang, X. Fu, M. Antonietti and X. Wang, J. Am. Chem. Soc., 2009, 131, 11658–11659.
13. F. Jaouen, E. Proietti, M. Lefèvre, R. Chenitz, J.-P. Dodelet, G. Wu, H. T. Chung, C. M. Johnston and P. Zelenay, Energy Environ. Sci., 2011, 4, 114–130.
14. M. Lefèvre and J.-P. Dodelet, ECS Trans., 2012, 45, 35–44.
15. N. Li, Z. Wang, K. Zhao, Z. Shi, Z. Gu and S. Xu, Carbon, 2010, 48, 1580–1585.
16. S. M. Unni, S. Ramadas, R. Illathvalappil, S. N. Bhange and S. Kurungot, J. Mater. Chem. A, 2015, 3, 4361–4367.
17. L. Brandão, C. Passeira, D. M. Gattia and A. Mendes, J. Mater. Sci., 2011, 46, 7198–7205.
18. T. Yoshitake, Y. Shimakawa, S. Kuroshima, H. Kimura, T. Ichihashi, Y. Kubo, D. Kasuya, K. Takahashi, F. Kokai and M. Yudasaka, Phys. B, 2002, 323, 124–126.
19. K. M. Eblagon and L. Brandao, J. Fuel Cell Sci. Technol., 2015, 12, 021006.
20. L. Zhang, N. Zheng, A. Gao, C. Zhu, Z. Wang, Y. Wang, Z. Shi and Y. Liu, J. Power Sources, 2012, 220, 449–454.
21. Q. Liu and J. Zhang, Langmuir, 2013, 29, 3821–3828.
22. S. Maldonado and K. J. Stevenson, J. Phys. Chem. B, 2005, 109, 4707–4716.
23. Q. Yu, J. Xu, C. Wu and L. Guan, RSC Adv., 2015, 5, 65303–65307.
24. H.-W. Liang, W. Wei, Z.-S. Wu, X. Feng and K. Müllen, J. Am. Chem. Soc., 2013, 135, 16002–16005.
25. G. Wu, Z. Chen, K. Artyushkova, F. H. Garzon and P. Zelenay, ECS Trans., 2008, 16, 159–170.
26. K. Niu, B. Yang, J. Cui, J. Jin, X. Fu, Q. Zhao and J. Zhang, J. Power Sources, 2013, 243, 65–71.
27. J. Casanovas, J. M. Ricart, J. Rubio, F. Illas and J. M. Jiménez-Mateos, J. Am. Chem. Soc., 1996, 118, 8071–8076.
28. M. Vikkisk, I. Kruusenberg, U. Joost, E. Shulga, I. Kink and K. Tammeveski, Appl. Catal., B, 2014, 147, 369–376.
29. Z. Lin, G. H. Waller, Y. Liu, M. Liu and C.-p. Wong, Nano Energy, 2013, 2, 241–248.
30. P. J. Kulesza, B. Grzybowska, M. A. Malik and M. T. Galkowski, J. Electrochem. Soc., 1997, 144, 1911–1917.
31. Y. Liang, Y. Li, H. Wang, J. Zhou, J. Wang, T. Regier and H. Dai, Nat. Mater., 2011, 10, 780–786.
32. Y. Jiang, Y. Lu, X. Wang, Y. Bao, W. Chen and L. Niu, Nanoscale, 2014, 6, 15066–15072.
33. S. Jiang, C. Zhu and S. Dong, J. Mater. Chem. A, 2013, 1, 3593–3599.
34. Q. Yi, Y. Zhang, X. Liu and Y. Yang, Sci. China: Chem., 2014, 57, 739–747.

---