- Email: [email protected]
Efficiency enhancement in dye sensitized solar cells using dual function mesoporous silica as scatterer and back recombination inhibitor
Chemical Physics Letters 658 (2016) 276–281
Contents lists available at ScienceDirect
Chemical Physics Letters journal homepage: www.elsevier.com/locate/cplett
Research paper
Efficiency enhancement in dye sensitized solar cells using dual function mesoporous silica as scatterer and back recombination inhibitor Tanvi a, Aman Mahajan a,⇑, R.K. Bedi a, Subodh Kumar b, Vibha Saxena c,⇑, D.K. Aswal c a
Material Science Laboratory, Department of Physics, Guru Nanak Dev University, Amritsar 143005, India Department of Chemistry, Guru Nanak Dev University, Amritsar 143005, India c Thin Film Devices Section, Technical Physics Division, Bhabha Atomic Research Centre, Mumbai 400 085, India b
a r t i c l e
i n f o
Article history: Received 11 April 2016 In final form 24 June 2016 Available online 25 June 2016 Keywords: Mesoporous silica Light harvesting Recombination Dye sensitized solar cells
a b s t r a c t In the present work, we report the usage of mesoporous silica for improving light harvesting as well as for suppression of back recombination without affecting the extent of dye loading on TiO2 films. Synthesized mesoporous SiO2 was characterized by X-ray photoelectron spectroscopy, X-ray diffraction, Brunauer Emmett and Teller measurement, Scanning electron microscopy and Transmission electron microscopy. DSSCs were fabricated by incorporating different wt% of mesoporous SiO2 in TiO2 paste. An improvement of 50% was observed for devices fabricated using 0.75 wt% of mesoporous SiO2. The mechanism behind the improvement was investigated using electrochemical impedance spectroscopy and UV–Vis spectroscopy. Ó 2016 Published by Elsevier B.V.
1. Introduction Owing to their eco-friendly attributes and cost effectiveness, dye sensitized solar cells (DSSCs) have emerged as promising candidates among different generation solar cells [1]. So far the maximum reported efficiency for DSSCs is approximately 13%, which is below the theoretical estimate by approximately 20% [2]. Efforts are therefore underway to enhance DSSC efficiency through different approaches. So far improvement in efficiency has been achieved either via light harvesting or by reduced recombination. In order to improve DSSC efficiency via light harvesting, various approaches like co-sensitization of TiO2 electrode with different dyes [3], plasmonic enhancement [4,5] synthesis of new sensitizers with higher molar extinction coefficients [6] and incorporation of scattering particles have been implemented [7]. On the other hand, reduced back electron transfer (or recombination) was achieved using post and pre-treatment of mesoporous TiO2, blocking layer on FTO, insulating layer on mesoporous TiO2 or co-adsorbents in the electrolyte [8–12]. Of late the usage of large sized nanoparticles as well as core–shell structures as scatterers in TiO2 have attracted attention since the incorporation of scatterers within TiO2 layer results in enhancement of interaction of photons with dye molecules. This enhancement is attributable to the scattering of light by large sized ⇑ Corresponding authors. E-mail addresses: [email protected] (A. Mahajan), [email protected]. in (V. Saxena). http://dx.doi.org/10.1016/j.cplett.2016.06.071 0009-2614/Ó 2016 Published by Elsevier B.V.
particles, which in turn results in enhancing the device efficiency [13,14]. This approach has generated considerable interest due to the availability of a variety of nanoparticles. It has been reported that silica acts as an efficient scattering center for improving light harvesting owing to higher difference in refractive index from surrounding medium, (the refractive index of TiO2 is 2.56 while that of silica is 1.46) [15]. Rho et al. had observed 22% enhancement in efficiency with silica beads as scatterers [16]. On the other hand, 20% improvement in photovoltaic performance was reported when SiO2/TiO2 core/shell nanoparticles were used [14]. Park et al. had observed 46% improvement with hybrid silica conjugated TiO2 nanostructures [17]. However, these approaches also resulted in reduced dye loading and in turn reduced light harvesting. To address this drawback, a two layered approach, each having different sized nanoparticles have been explored. In this approach, the first layer consists of small sized mesoporous TiO2 nanoparticles (10–15 nm) while second layer contains large sized scattering particles. Wang et al. had reported 18% improvement in efficiency by using multilayer approach as compared to single layer approach [18]. Niaki et al. had observed 7% enhancement in photovoltaic performance [19] using double layer film doped with Zn ions [20]. However, single reflection of light limits the improvement that is attainable in photovoltaic current density and therefore restricts the device efficiency [18]. We have earlier proposed that a significant improvement in efficiency can be achieved by improving both light harvesting and reduced recombination through a novel strategy of co-sensitization [3,21].
Tanvi et al. / Chemical Physics Letters 658 (2016) 276–281
In the present case, mesoporous SiO2 was chosen since it not only acts as good scatterer but also provides efficient barrier between TiO2 and electrolyte. In our approach, mesoporous SiO2 particles with controlled pore size were synthesized so that TiO2 sites remain available for dye loading (Fig. 1). We have thus demonstrated that an improvement of 50% in device efficiency can be achieved by proper choice of mesoporous SiO2 owing to cumulative effect of enhanced light harvesting and reduced back recombination.
277
0.75 wt% and 1 wt%, of SiO2 are referred as 0.5 S, 0.75 S and 1 S, respectively, the reference electrode (i.e. TiO2) is referred as T hereafter. These electrodes were sensitized with N719 dye by dipping electrodes in 0.3 mM dye solution in ethanol for 24 h. The thickness of photoanodes was estimated to be 2.5 lm from their cross-sectional view in SEM image. Photocathodes were prepared by drop casting platinum catalyst precursor onto the cleaned FTO followed by annealing the same at 450°C. Finally, the cells were assembled by using parafilm and electrolyte containing I/I 3 as redox couple in 3-methoxypropionitrile.
2. Materials and methods 2.3. Characterization 2.1. Materials Polyethylene block polymer, tetraethyl orthosilicate (TEOS) and di-tetrabutylammoniumcis-bis(isothiocyanato)bis(2,2-8bipyridyl4,4dicarboxylato) ruthenium(II) (N719) dye were purchased from Sigma Aldrich. Fluorine-doped tin oxide coated substrates (FTO), TiO2 paste (Ti-Nanoxide T) and platinum catalyst precursor (Platisol-T) were procured from Solarnix and electrolyte containing I/I 3 as redox couple in 3-methoxypropionitrile (EL-HSE) was received from Dyesol. 2.2. Synthesis of mesoporous silica and fabrication of DSSCs Mesoporous SiO2 was synthesized using Polyethylene block polymer and tetraethyl orthosilicate as precursors using the method reported earlier (S1) [22,23]. For fabrication of DSSCs, FTO substrates were first treated with 40 mM aqueous solution of TiCl4 at 70°C for 30 min and then annealed at 450°C for 30 min to form a compact layer of TiO2 (c-TiO2). Mesoporous SiO2 particles were added to the TiO2 paste in different weight% 0.5 wt%, 0.75 wt% and 1 wt% under continuous magnetic stirring for homogeneous mixing of these particles into TiO2 paste. The resultant paste was further used to prepare films using earlier procedure [3]. The electrodes thus prepared using 0.5 wt%,
Brunauer Emmett and Teller (BET) analysis was performed using Micromeritics porosity analyser ASAP 1020. The X-ray photoelectron (XPS) spectra was recorded using Mac-2 electron analyser with Ka as radiation source, Scanning electron microscopic images were taken with Carl Zeiss Supra 55 while Transmission electron microscopic images were obtained using JEOL JEM-2100. UV–Vis spectra were recorded with Shimazu, UV–VIS–NIR 3600. X-ray diffraction (XRD) pattern was obtained from D8 Focus, Bruker Ettlingen, Germany with Cu Ka as source. The photovoltaic properties of DSSCs were carried out using Sciencetech Solar Simulator equipped with 150 W Xenon lamp and Autolab ECO Chemie PGSTAT 30 Potentiostat/Galvanostat. The illumination intensity was 100 mW/cm2 at AM 1.5 G. Prior to the characterization, intensity of lamp was calibrated using standard silicon cell. 3. Results and discussions Synthesized mesoporous SiO2 was characterized using XPS and BET measurements. As shown in Fig. 2(a), the nitrogen adsorption–desorption isotherm exhibited a type IV shape with type H1 hysteresis, suggesting well-ordered mesoporous SiO2 with a narrow and uniformly distributed cylindrical pores [24]. The BET surface area of mesoporous structure of SiO2 was estimated to be
Fig. 1. Schematic representing absorption of photons (a–b) and recombination of charge carriers (c–d) before and after incorporating mesoporous SiO2 respectively.
278
Tanvi et al. / Chemical Physics Letters 658 (2016) 276–281
Fig. 2. (a) Nitrogen adsorption–desorption isotherms and pore size distribution (inset). XPS spectra of (b) Si 2p (c) O 1s.
501.23 m2/g. The pore size of synthesized mesoporous SiO2, as determined from the isotherm, revealed a narrow distribution in range 6–8 nm with average pore diameter of 7 nm (inset of Fig. 2 (a)). The pore size of synthesized SiO2 is beneficial for increasing dye loading since it can facilitate the transport of the dye. The XPS spectra of synthesized SiO2 exhibit peaks around binding energies 103.3 eV and 532.4 eV as shown in Fig. 2 (b) and (c). The peak at 103.3 eV is assigned to Si 2p peak and could be fitted with single Gaussian confirming the 4+ oxidation state of Si. On the other hand, O 1s peak at 532.4 eV is slightly asymmetric and is fitted with two Gaussian peaks corresponding to oxygen in SiO2 (532.4 eV) and physisorb oxygen (531.2 eV) [25]. The surface morphology and structure of synthesized mesoporous SiO2 was studied using scanning electron microscopy and Transmission electron microscopy. As shown in Fig. 3(a), the SEM images reveal the formation of uniformly distributed rod like structures without any aggregation. The dimensions of the rod were estimated to be around 370 nm (diameter) and 1.2 lm (length). The observation is consistent with the TEM images which also show rod like shapes (Fig. 3(a) inset). Further, rings or dotted pattern were observed to be absent in selected area electron diffraction (SAED) pattern of the synthesized mesoporous SiO2 thereby suggesting amorphous nature. This is corroborated with XRD pattern, which showed a broad peak indicating amorphous nature (S2) [23,26]. The HRTEM of single particle clearly reveals the presence of straight and parallel channels through-out the mesoporous SiO2 particle (Fig. 3(c)). The magnified view of HRTEM further confirms the mesoporous structure of SiO2 with pore size around 8.33 nm which is consistent with BET analysis (Fig. 3(d)). As discussed earlier, the pore size of mesoporous SiO2 (7 nm) is much bigger than that of N719 dye molecules (1 nm) [27] which facilitates the diffusion of dye molecules through proper channels and therefore does not affect dye loading when incorporated in mesoporous TiO2 [28]. Fig. 4(b) shows SEM images of mesoporous TiO2 incorporated with SiO2 revealing inter-connecting nanocrystallites having porous structure. The figure clearly shows that morphology of the mesoporous TiO2 (Fig. 4(a)) did not change when mesoporous SiO2 was incorporated in TiO2 and the rod like structure of mesoporous SiO2 is entirely covered by TiO2. The photovoltaic characteristics of the DSSCs fabricated using mesoporous SiO2 were studied by recording current density (J)-voltage (V) characteristics under 1 sun. The estimated photovoltaic parameters of the fabricated cells as a function of wt% of mesoporous SiO2 are presented in Fig. 5 [23]. It is to be noted that the efficiency of reference cell is 2.37% which is similar to the reported value in literature for 2.5 lm thick TiO2 film. A small deviation in DSSCs efficiency from its average value indicates high degree of reproducibility of fabricated devices. An optimized efficiency of 3.57% is obtained for 0.75 wt% of mesoporous SiO2. This improvement in efficiency is comparable with those reported
in literature e.g. the incorporation of nanocrystalline SiO2 in TiO2 resulted in efficiency improvement of 45% [29,30]. Similarly, Wang et al. reported that the caging of TiO2 nanoparticles within the curved microsheets of silica leads to 41% improvement in device efficiency as a result of light harvesting and suppression of recombination [31]. Other than SiO2, other nanoparticles such as GeO2 and CaF2 were also incorporated in DSSC devices, leading to efficiency improvement of about 30% [32,33]. Our results demonstrate the potential of mesoporous SiO2 in improving the device efficiency in an efficient manner. In order to investigate, the reasons for the device improvement, we first analysed estimated photovoltaic parameters. As shown in the figures, the open circuit voltage (Voc) remains almost unaffected while the short circuit current density (Jsc) increases gradually with increase in wt% of mesoporous SiO2. The improvement in JSC is attributable to enhanced light harvesting on account of incorporation of mesoporous SiO2. However, FF is maximum upon incorporation of 0.75 wt% of mesoporous SiO2 and this improvement is attributable to the blocking effect of SiO2. Incorporation of mesoporous SiO2 more than 0.75 wt% results in reduced fill factor, which may be due to the resistance posed by insulating SiO2. This is also evident from the dark J–V characteristics of DSSCs which shows increment in the value of onset potential upon incorporation of mesoporous SiO2 up to 0.75 wt% and decreases upon further incorporation of mesoporous SiO2 (Fig. 5(e)). Therefore, the overall highest efficiency (3.57%) is obtained for 0.75 wt% of mesoporous SiO2 which is an improvement of 50% over the reference cell. In order to investigate the mechanism behind the efficiency enhancement in photovoltaic parameters upon incorporation of mesoporous SiO2, UV–Vis and electrochemical impedance spectroscopy (EIS) were performed. Since Voc is defined by the difference between redox couple potential and quasi Fermi level of TiO2 [1], we have estimated the band gap of electrodes with different wt% of mesoporous SiO2 from Tauc plots. The band gap was found to lie in the range of 3.34–3.45 eV (S7, S8) [23] for TiO2 electrodes with different wt% of mesoporous SiO2. No significant change was observed in the band gap upon incorporation of SiO2, which clearly indicates that the conduction band edge of TiO2 remains almost unaffected by the incorporation of mesoporous SiO2 into TiO2. It is further confirmed by XPS as no peak corresponding to Ti–O–Si was observed in XPS spectra of all the samples incorporating mesoporous SiO2. In order to ascertain the reason for enhancement in JSC upon incorporation of mesoporous SiO2, we estimated dye loading on all the electrodes T, 0.5 S, 0.75 S and 1 S (S4, S5) [23] using methodology described earlier [3]. No significant change in absorbance is observed suggesting that incorporation of mesoporous SiO2 does not affect the dye loading. This was expected as the bigger pore size of mesoporous SiO2 than N719 dye molecules allows efficient dye loading and does not hinder sensitization of dye molecules to
Tanvi et al. / Chemical Physics Letters 658 (2016) 276–281
279
Fig. 3. (a) SEM of mesoporous SiO2, inset shows the TEM image of mesoporous SiO2 (b) SAED pattern (c) HRTEM of mesoporous SiO2 particle and (d) magnified view of (c).
Fig. 4. SEM of (a) mesoporous TiO2 and (b) mesoporous TiO2 incorporated with mesoporous SiO2 particles.
available TiO2 sites. Though the dye loading of photoanodes does not change significantly upon incorporation of mesoporous SiO2 particle in TiO2, an increase in the absorbance of photoanodes was observed, as shown in Fig. 6(a). The inset of Fig. 6(a) shows the percentage enhancement in absorbance of photoanodes prepared at different wt% of mesoporous SiO2. The enhancement of the absorbance with incorporation of mesoporous SiO2 is attributable to enhanced light harvesting owing to light scattering
properties rather than dye loading. This suggests that the synthesized mesoporous SiO2 can efficiently be utilized as scatterers in DSSC fabrication. Furthermore, the role of mesoporous SiO2 in reducing the recombination in DSSC was investigated using EIS, under 1 sun at open circuit potential and results are presented in Fig. 6(b). Typical impedance spectra (Nyquist plots) were found to consist two semi-circles: the first circle corresponds to charge transfer at the Pt/electrolyte interface (Rpt, Q1) while second one
280
Tanvi et al. / Chemical Physics Letters 658 (2016) 276–281
Fig. 5. Variation of (a) Jsc, (b) Voc, (c) FF and (d) with incorporation of mesoporous SiO2 in TiO2, error bars are also shown in the figures and, (e) J–V characteristics of DSSCs fabricated under dark conditions.
Fig. 6. (a) Absorption spectra of electrodes T, 0.5 S, 0.75 S and 1 S respectively after dye sensitization, Inset represents the variation of absorption in photoanodes with the wt% of mesoporous SiO2. (b) Impedance spectra of DSSCs assembled using T, 0.5 S, 0.75 S and 1 S under 1 sun at open circuit voltage.
is attributable to charge transport resistance in TiO2 (Rtr) as well as charge transfer resistance corresponding to recombination of electrons at the TiO2/electrolyte interface (Rrec, Q2) and chemical capacitance of TiO2 (Cl) (at Voc, Rtr in TiO2 network is negligible) [34]. The value of Rrec was estimated by fitting the EIS data using R(QR)(QR) model [21] (ZSimp Win 3.22 software). The electron life time (s) was calculated from Bode plots using the formula [35] s ¼ 2p1x, where x represents frequency peak in the low frequency range (S6) [23] and all EIS parameters for a typical DSSC are presented in Table 1. As evident from the table, the recombination resistance as well as electron life time increases with incorporation of mesoporous SiO2 up to 0.75 wt% and decreases thereafter. The increase in the recombination resistance upon incorporation of mesoporous SiO2 in TiO2 suggests hindrance of charge carriers towards back recombination. In addition, an increase in life time indicates that mesoporous SiO2 acts as energy barrier between TiO2 and electrolyte and therefore, reduced back electron transfer. This is also evident from the dark current measurements which shows that: (i) dark current is reduced substantially after incorporation of mesoporous SiO2 particles in TiO2 and (ii) the shift in dark Table 1 EIS parameters estimated for three devices each fabricated using T, 0.5 S, 0.75 S and 1 S. Electrodes
Rtr (Ohm cm2)
s (msec)
T 0.5 S 0.75 S 1S
7.04 11.78 13.44 10.94
11.47 16.48 23.66 16.48
current onset potential is larger for 0.75 S electrode as compared with 1 S electrode. Therefore the recombination between I 3 and CB electron of TiO2 is suppressed for 0.75 S electrode as compared with 1 S electrode leading to similar trend for FF as explained below: FF can be expressed in terms of Voc and b [34,36]
FF ¼
V r lnðV r þ 0:72Þ Vr þ 1
oc where V r ¼ bqV is the reduced voltage and b recombination Kb T
exponent. Since Voc remains almost unaffected therefore, FF is controlled by recombination exponent which is directly proportional to recombination resistance and hence follows the same trend as recombination resistance. Therefore, overall improvement in the device is observed for DSSCs fabricated with 0.75 wt% of mesoporous SiO2. 4. Conclusions In summary, we have synthesized mesoporous SiO2 with appropriate size which could be utilized as efficient scatterers and recombination inhibitors in DSSCs. The DSSCs fabricated exhibit power conversion efficiency as large as 3.57% using 2.5 lm thick TiO2 film incorporated with 0.75 wt% of mesoporous SiO2. The enhancement in photovoltaic performance was attributed to enhanced short circuit current density and FF, which is confirmed by absorption spectra, and electrochemical impedance spectroscopy.
Tanvi et al. / Chemical Physics Letters 658 (2016) 276–281
Acknowledgements Authors are thankful to Dr. K.P. Muthe and Dr. Ajay Singh for their support and helpful discussions. One of the authors, Tanvi, is grateful to Department of Science and Technology, New Delhi (India) under INSPIRE program for fellowship. Authors A.M., R.K. B. and S.K. are thankful to UPE program under UGC, New Delhi (India). Financial assistance provided by BRNS, Mumbai to carry out the present research work is acknowledged. Appendix A. Supplementary material
[11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cplett.2016.06. 071. References [1] M. Gratzel, Acc. Chem. Res. 42 (2009) 1788. [2] H.J. Snaith, Adv. Funct. Mater. 20 (2010) 13. [3] V. Saxena, V. Putta, A.K. Chauhan, P. Jha, D.K. Aswal, S.K. Gupta, Appl. Phys. Lett. 100 (2012) 133303. [4] S.-J. Lin, K.-C. Lee, J.-L. Wu, J.-Y. Wu, Appl. Phys. Lett. 99 (2011) 043306. [5] Tanvi, A. Mahajan, R.K. Bedi, S. Kumar, V. Saxena, D.K. Aswal, J. Appl. Phys. 117 (2015) 083111. [6] Q. Yu, S. Liu, M. Zhang, N. Cai, Y. Wang, P. Wang, J. Phys. Chem. C 113 (2009) 14559. [7] T.G. Deepak, G.S. Anjusree, S. Thomas, T.A. Arun, S.V. Nair, A. Sreekumaran Nair, RSC Adv. 4 (2014) 17615. [8] Wu, X. Gao, M. Qin, J.-M. Liu, S. Hu, Appl. Phys. Lett. 99 (2011) 042106. [9] E. Palomares, J.N. Clifford, S.A. Haque, T. Lutz, J.R. Durrant, J. Am. Chem. Soc. 125 (2003) 475. [10] H.-J. Son, X. Wang, C. Prasittichai, N.C. Jeong, T. Aaltonen, R.G. Gordon, J.T. Hupp, J. Am. Chem. Soc. 125 (2003) 475.
[22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36]
281
Q. Hu, C. Wu, L. Cao, B. Chi, J. Pu, L. Jian, J. Power Sources 226 (2013) 8. W. Ma, Y. Jiao, S. Meng, Phys. Chem. Chem. Phys. 15 (2013) 17187. W.G. Yang, F.R. Wan, Q.W. Chen, J.J. Li, D.S. Xu, J. Mater. Chem. 20 (2010) 2870. S. Son, S.H. Hwang, C. Kim, J.Y. Yun, J. Jang, ACS Appl. Mater. Interfaces 5 (2013) 4815. Y. Wang, E. Chen, H. Lai, B. Lu, Z. Hu, X. Qin, W. Shi, G. Du, Ceram. Int. 39 (2013) 5407. Y. Rho, M. Wanit, J. Yeo, S. Hong, S. Han, J.-H. Choi, W.-H. Hong, D. Lee, S.H. Ko, J. Phys. D: Appl. Phys 46 (2013) 024006. K.-H. Park, H.-B. Gu, E.M. Jin, M. Dhayal, Electrochim. Acta 55 (2010) 5499. Z.-S. Wang, H. Kawauchi, T. Kashima, H. Arakawa, Coord. Chem. Rev. 248 (2004) 1381. A.G. Niaki, A. Bakhshayesh, M. Mohammadi, Sol. Energy 103 (2014) 210222. D. Sengupta, P. Das, U. Kasinadhuni, B. Mondal, K. Mukherjee, J. Renew. Sustain. Energy 6 (2014) 063114. V. Saxena, V. Putta, A. Gusain, P. Jha, J. Singh, S. Koiry, P. Varde, A. Chauhan, D. Aswal, S. Gupta, Org. Electron. 14 (2013) 3098. D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Science 279 (1998) 548. See supplemental material for XRD, dye loading, bandgap analysis, electron lifetime measurements and current–voltage characteristics. A. Sayari, B.H. Han, Y. Yang, J. Am. Chem. Soc. 126 (2004) 14348. Q. Lu, Z. Wang, J. Li, P. Wang, X. Ye, Nanoscale Res. Lett. 4 (2009) 646. Y. Xie, B. Yan, C. Tian, Y. Liu, Q. Liu, H. Zeng, J. Mater. Chem. A 2 (2014) 17730. L. Ellis-Gibbings, V. Johansson, R.B. Walsh, L. Kloo, J.S. Quinton, G.G. Andersson, Langmuir 28 (2012) 9431. A. Zrner, J. Kirstein, M. Dblinger, C. Bruchle, T. Bein, Nature 450 (2007) 705. S. Yuan, Q. Tang, B. He, L. Menand, H. Chen, Electrochim. Acta 125 (2014) 646. P. Xu, Q. Tang, B. He, Q. Li, H. Chen, Electrochim. Acta 134 (2014) 281. Z. Wang, Q. Tang, B. He, H. Chen, L. Yu, Electrochim. Acta 178 (2015) 18. Y. Duan, Q. Tang, Z. Chen, B. He, H. Chen, J. Mater. Chem. A 2 (2014) 12459. Z. Wang, Q. Tang, B. He, X. Chen, H. Chen, L. Yu, J. Power Sources 1 (2014) 175. F. Fabregat-Santiago, G. Garcia-Belmonte, I. Mora-Sero, J. Bisquert, Phys. Chem. Chem. Phys. 13 (2011) 9083. R. Kern, R. Sastrawan, J. Ferber, R. Stangl, J. Luther, Electrochim. Acta 47 (2002) 4213. Z. Huang, G. Natu, Z. Ji, M. He, M. Yu, Y. Wu, J. Phys. Chem. C 116 (2012) 26239.
Contents lists available at ScienceDirect
Chemical Physics Letters journal homepage: www.elsevier.com/locate/cplett
Research paper
Efficiency enhancement in dye sensitized solar cells using dual function mesoporous silica as scatterer and back recombination inhibitor Tanvi a, Aman Mahajan a,⇑, R.K. Bedi a, Subodh Kumar b, Vibha Saxena c,⇑, D.K. Aswal c a
Material Science Laboratory, Department of Physics, Guru Nanak Dev University, Amritsar 143005, India Department of Chemistry, Guru Nanak Dev University, Amritsar 143005, India c Thin Film Devices Section, Technical Physics Division, Bhabha Atomic Research Centre, Mumbai 400 085, India b
a r t i c l e
i n f o
Article history: Received 11 April 2016 In final form 24 June 2016 Available online 25 June 2016 Keywords: Mesoporous silica Light harvesting Recombination Dye sensitized solar cells
a b s t r a c t In the present work, we report the usage of mesoporous silica for improving light harvesting as well as for suppression of back recombination without affecting the extent of dye loading on TiO2 films. Synthesized mesoporous SiO2 was characterized by X-ray photoelectron spectroscopy, X-ray diffraction, Brunauer Emmett and Teller measurement, Scanning electron microscopy and Transmission electron microscopy. DSSCs were fabricated by incorporating different wt% of mesoporous SiO2 in TiO2 paste. An improvement of 50% was observed for devices fabricated using 0.75 wt% of mesoporous SiO2. The mechanism behind the improvement was investigated using electrochemical impedance spectroscopy and UV–Vis spectroscopy. Ó 2016 Published by Elsevier B.V.
1. Introduction Owing to their eco-friendly attributes and cost effectiveness, dye sensitized solar cells (DSSCs) have emerged as promising candidates among different generation solar cells [1]. So far the maximum reported efficiency for DSSCs is approximately 13%, which is below the theoretical estimate by approximately 20% [2]. Efforts are therefore underway to enhance DSSC efficiency through different approaches. So far improvement in efficiency has been achieved either via light harvesting or by reduced recombination. In order to improve DSSC efficiency via light harvesting, various approaches like co-sensitization of TiO2 electrode with different dyes [3], plasmonic enhancement [4,5] synthesis of new sensitizers with higher molar extinction coefficients [6] and incorporation of scattering particles have been implemented [7]. On the other hand, reduced back electron transfer (or recombination) was achieved using post and pre-treatment of mesoporous TiO2, blocking layer on FTO, insulating layer on mesoporous TiO2 or co-adsorbents in the electrolyte [8–12]. Of late the usage of large sized nanoparticles as well as core–shell structures as scatterers in TiO2 have attracted attention since the incorporation of scatterers within TiO2 layer results in enhancement of interaction of photons with dye molecules. This enhancement is attributable to the scattering of light by large sized ⇑ Corresponding authors. E-mail addresses: [email protected] (A. Mahajan), [email protected]. in (V. Saxena). http://dx.doi.org/10.1016/j.cplett.2016.06.071 0009-2614/Ó 2016 Published by Elsevier B.V.
particles, which in turn results in enhancing the device efficiency [13,14]. This approach has generated considerable interest due to the availability of a variety of nanoparticles. It has been reported that silica acts as an efficient scattering center for improving light harvesting owing to higher difference in refractive index from surrounding medium, (the refractive index of TiO2 is 2.56 while that of silica is 1.46) [15]. Rho et al. had observed 22% enhancement in efficiency with silica beads as scatterers [16]. On the other hand, 20% improvement in photovoltaic performance was reported when SiO2/TiO2 core/shell nanoparticles were used [14]. Park et al. had observed 46% improvement with hybrid silica conjugated TiO2 nanostructures [17]. However, these approaches also resulted in reduced dye loading and in turn reduced light harvesting. To address this drawback, a two layered approach, each having different sized nanoparticles have been explored. In this approach, the first layer consists of small sized mesoporous TiO2 nanoparticles (10–15 nm) while second layer contains large sized scattering particles. Wang et al. had reported 18% improvement in efficiency by using multilayer approach as compared to single layer approach [18]. Niaki et al. had observed 7% enhancement in photovoltaic performance [19] using double layer film doped with Zn ions [20]. However, single reflection of light limits the improvement that is attainable in photovoltaic current density and therefore restricts the device efficiency [18]. We have earlier proposed that a significant improvement in efficiency can be achieved by improving both light harvesting and reduced recombination through a novel strategy of co-sensitization [3,21].
Tanvi et al. / Chemical Physics Letters 658 (2016) 276–281
In the present case, mesoporous SiO2 was chosen since it not only acts as good scatterer but also provides efficient barrier between TiO2 and electrolyte. In our approach, mesoporous SiO2 particles with controlled pore size were synthesized so that TiO2 sites remain available for dye loading (Fig. 1). We have thus demonstrated that an improvement of 50% in device efficiency can be achieved by proper choice of mesoporous SiO2 owing to cumulative effect of enhanced light harvesting and reduced back recombination.
277
0.75 wt% and 1 wt%, of SiO2 are referred as 0.5 S, 0.75 S and 1 S, respectively, the reference electrode (i.e. TiO2) is referred as T hereafter. These electrodes were sensitized with N719 dye by dipping electrodes in 0.3 mM dye solution in ethanol for 24 h. The thickness of photoanodes was estimated to be 2.5 lm from their cross-sectional view in SEM image. Photocathodes were prepared by drop casting platinum catalyst precursor onto the cleaned FTO followed by annealing the same at 450°C. Finally, the cells were assembled by using parafilm and electrolyte containing I/I 3 as redox couple in 3-methoxypropionitrile.
2. Materials and methods 2.3. Characterization 2.1. Materials Polyethylene block polymer, tetraethyl orthosilicate (TEOS) and di-tetrabutylammoniumcis-bis(isothiocyanato)bis(2,2-8bipyridyl4,4dicarboxylato) ruthenium(II) (N719) dye were purchased from Sigma Aldrich. Fluorine-doped tin oxide coated substrates (FTO), TiO2 paste (Ti-Nanoxide T) and platinum catalyst precursor (Platisol-T) were procured from Solarnix and electrolyte containing I/I 3 as redox couple in 3-methoxypropionitrile (EL-HSE) was received from Dyesol. 2.2. Synthesis of mesoporous silica and fabrication of DSSCs Mesoporous SiO2 was synthesized using Polyethylene block polymer and tetraethyl orthosilicate as precursors using the method reported earlier (S1) [22,23]. For fabrication of DSSCs, FTO substrates were first treated with 40 mM aqueous solution of TiCl4 at 70°C for 30 min and then annealed at 450°C for 30 min to form a compact layer of TiO2 (c-TiO2). Mesoporous SiO2 particles were added to the TiO2 paste in different weight% 0.5 wt%, 0.75 wt% and 1 wt% under continuous magnetic stirring for homogeneous mixing of these particles into TiO2 paste. The resultant paste was further used to prepare films using earlier procedure [3]. The electrodes thus prepared using 0.5 wt%,
Brunauer Emmett and Teller (BET) analysis was performed using Micromeritics porosity analyser ASAP 1020. The X-ray photoelectron (XPS) spectra was recorded using Mac-2 electron analyser with Ka as radiation source, Scanning electron microscopic images were taken with Carl Zeiss Supra 55 while Transmission electron microscopic images were obtained using JEOL JEM-2100. UV–Vis spectra were recorded with Shimazu, UV–VIS–NIR 3600. X-ray diffraction (XRD) pattern was obtained from D8 Focus, Bruker Ettlingen, Germany with Cu Ka as source. The photovoltaic properties of DSSCs were carried out using Sciencetech Solar Simulator equipped with 150 W Xenon lamp and Autolab ECO Chemie PGSTAT 30 Potentiostat/Galvanostat. The illumination intensity was 100 mW/cm2 at AM 1.5 G. Prior to the characterization, intensity of lamp was calibrated using standard silicon cell. 3. Results and discussions Synthesized mesoporous SiO2 was characterized using XPS and BET measurements. As shown in Fig. 2(a), the nitrogen adsorption–desorption isotherm exhibited a type IV shape with type H1 hysteresis, suggesting well-ordered mesoporous SiO2 with a narrow and uniformly distributed cylindrical pores [24]. The BET surface area of mesoporous structure of SiO2 was estimated to be
Fig. 1. Schematic representing absorption of photons (a–b) and recombination of charge carriers (c–d) before and after incorporating mesoporous SiO2 respectively.
278
Tanvi et al. / Chemical Physics Letters 658 (2016) 276–281
Fig. 2. (a) Nitrogen adsorption–desorption isotherms and pore size distribution (inset). XPS spectra of (b) Si 2p (c) O 1s.
501.23 m2/g. The pore size of synthesized mesoporous SiO2, as determined from the isotherm, revealed a narrow distribution in range 6–8 nm with average pore diameter of 7 nm (inset of Fig. 2 (a)). The pore size of synthesized SiO2 is beneficial for increasing dye loading since it can facilitate the transport of the dye. The XPS spectra of synthesized SiO2 exhibit peaks around binding energies 103.3 eV and 532.4 eV as shown in Fig. 2 (b) and (c). The peak at 103.3 eV is assigned to Si 2p peak and could be fitted with single Gaussian confirming the 4+ oxidation state of Si. On the other hand, O 1s peak at 532.4 eV is slightly asymmetric and is fitted with two Gaussian peaks corresponding to oxygen in SiO2 (532.4 eV) and physisorb oxygen (531.2 eV) [25]. The surface morphology and structure of synthesized mesoporous SiO2 was studied using scanning electron microscopy and Transmission electron microscopy. As shown in Fig. 3(a), the SEM images reveal the formation of uniformly distributed rod like structures without any aggregation. The dimensions of the rod were estimated to be around 370 nm (diameter) and 1.2 lm (length). The observation is consistent with the TEM images which also show rod like shapes (Fig. 3(a) inset). Further, rings or dotted pattern were observed to be absent in selected area electron diffraction (SAED) pattern of the synthesized mesoporous SiO2 thereby suggesting amorphous nature. This is corroborated with XRD pattern, which showed a broad peak indicating amorphous nature (S2) [23,26]. The HRTEM of single particle clearly reveals the presence of straight and parallel channels through-out the mesoporous SiO2 particle (Fig. 3(c)). The magnified view of HRTEM further confirms the mesoporous structure of SiO2 with pore size around 8.33 nm which is consistent with BET analysis (Fig. 3(d)). As discussed earlier, the pore size of mesoporous SiO2 (7 nm) is much bigger than that of N719 dye molecules (1 nm) [27] which facilitates the diffusion of dye molecules through proper channels and therefore does not affect dye loading when incorporated in mesoporous TiO2 [28]. Fig. 4(b) shows SEM images of mesoporous TiO2 incorporated with SiO2 revealing inter-connecting nanocrystallites having porous structure. The figure clearly shows that morphology of the mesoporous TiO2 (Fig. 4(a)) did not change when mesoporous SiO2 was incorporated in TiO2 and the rod like structure of mesoporous SiO2 is entirely covered by TiO2. The photovoltaic characteristics of the DSSCs fabricated using mesoporous SiO2 were studied by recording current density (J)-voltage (V) characteristics under 1 sun. The estimated photovoltaic parameters of the fabricated cells as a function of wt% of mesoporous SiO2 are presented in Fig. 5 [23]. It is to be noted that the efficiency of reference cell is 2.37% which is similar to the reported value in literature for 2.5 lm thick TiO2 film. A small deviation in DSSCs efficiency from its average value indicates high degree of reproducibility of fabricated devices. An optimized efficiency of 3.57% is obtained for 0.75 wt% of mesoporous SiO2. This improvement in efficiency is comparable with those reported
in literature e.g. the incorporation of nanocrystalline SiO2 in TiO2 resulted in efficiency improvement of 45% [29,30]. Similarly, Wang et al. reported that the caging of TiO2 nanoparticles within the curved microsheets of silica leads to 41% improvement in device efficiency as a result of light harvesting and suppression of recombination [31]. Other than SiO2, other nanoparticles such as GeO2 and CaF2 were also incorporated in DSSC devices, leading to efficiency improvement of about 30% [32,33]. Our results demonstrate the potential of mesoporous SiO2 in improving the device efficiency in an efficient manner. In order to investigate, the reasons for the device improvement, we first analysed estimated photovoltaic parameters. As shown in the figures, the open circuit voltage (Voc) remains almost unaffected while the short circuit current density (Jsc) increases gradually with increase in wt% of mesoporous SiO2. The improvement in JSC is attributable to enhanced light harvesting on account of incorporation of mesoporous SiO2. However, FF is maximum upon incorporation of 0.75 wt% of mesoporous SiO2 and this improvement is attributable to the blocking effect of SiO2. Incorporation of mesoporous SiO2 more than 0.75 wt% results in reduced fill factor, which may be due to the resistance posed by insulating SiO2. This is also evident from the dark J–V characteristics of DSSCs which shows increment in the value of onset potential upon incorporation of mesoporous SiO2 up to 0.75 wt% and decreases upon further incorporation of mesoporous SiO2 (Fig. 5(e)). Therefore, the overall highest efficiency (3.57%) is obtained for 0.75 wt% of mesoporous SiO2 which is an improvement of 50% over the reference cell. In order to investigate the mechanism behind the efficiency enhancement in photovoltaic parameters upon incorporation of mesoporous SiO2, UV–Vis and electrochemical impedance spectroscopy (EIS) were performed. Since Voc is defined by the difference between redox couple potential and quasi Fermi level of TiO2 [1], we have estimated the band gap of electrodes with different wt% of mesoporous SiO2 from Tauc plots. The band gap was found to lie in the range of 3.34–3.45 eV (S7, S8) [23] for TiO2 electrodes with different wt% of mesoporous SiO2. No significant change was observed in the band gap upon incorporation of SiO2, which clearly indicates that the conduction band edge of TiO2 remains almost unaffected by the incorporation of mesoporous SiO2 into TiO2. It is further confirmed by XPS as no peak corresponding to Ti–O–Si was observed in XPS spectra of all the samples incorporating mesoporous SiO2. In order to ascertain the reason for enhancement in JSC upon incorporation of mesoporous SiO2, we estimated dye loading on all the electrodes T, 0.5 S, 0.75 S and 1 S (S4, S5) [23] using methodology described earlier [3]. No significant change in absorbance is observed suggesting that incorporation of mesoporous SiO2 does not affect the dye loading. This was expected as the bigger pore size of mesoporous SiO2 than N719 dye molecules allows efficient dye loading and does not hinder sensitization of dye molecules to
Tanvi et al. / Chemical Physics Letters 658 (2016) 276–281
279
Fig. 3. (a) SEM of mesoporous SiO2, inset shows the TEM image of mesoporous SiO2 (b) SAED pattern (c) HRTEM of mesoporous SiO2 particle and (d) magnified view of (c).
Fig. 4. SEM of (a) mesoporous TiO2 and (b) mesoporous TiO2 incorporated with mesoporous SiO2 particles.
available TiO2 sites. Though the dye loading of photoanodes does not change significantly upon incorporation of mesoporous SiO2 particle in TiO2, an increase in the absorbance of photoanodes was observed, as shown in Fig. 6(a). The inset of Fig. 6(a) shows the percentage enhancement in absorbance of photoanodes prepared at different wt% of mesoporous SiO2. The enhancement of the absorbance with incorporation of mesoporous SiO2 is attributable to enhanced light harvesting owing to light scattering
properties rather than dye loading. This suggests that the synthesized mesoporous SiO2 can efficiently be utilized as scatterers in DSSC fabrication. Furthermore, the role of mesoporous SiO2 in reducing the recombination in DSSC was investigated using EIS, under 1 sun at open circuit potential and results are presented in Fig. 6(b). Typical impedance spectra (Nyquist plots) were found to consist two semi-circles: the first circle corresponds to charge transfer at the Pt/electrolyte interface (Rpt, Q1) while second one
280
Tanvi et al. / Chemical Physics Letters 658 (2016) 276–281
Fig. 5. Variation of (a) Jsc, (b) Voc, (c) FF and (d) with incorporation of mesoporous SiO2 in TiO2, error bars are also shown in the figures and, (e) J–V characteristics of DSSCs fabricated under dark conditions.
Fig. 6. (a) Absorption spectra of electrodes T, 0.5 S, 0.75 S and 1 S respectively after dye sensitization, Inset represents the variation of absorption in photoanodes with the wt% of mesoporous SiO2. (b) Impedance spectra of DSSCs assembled using T, 0.5 S, 0.75 S and 1 S under 1 sun at open circuit voltage.
is attributable to charge transport resistance in TiO2 (Rtr) as well as charge transfer resistance corresponding to recombination of electrons at the TiO2/electrolyte interface (Rrec, Q2) and chemical capacitance of TiO2 (Cl) (at Voc, Rtr in TiO2 network is negligible) [34]. The value of Rrec was estimated by fitting the EIS data using R(QR)(QR) model [21] (ZSimp Win 3.22 software). The electron life time (s) was calculated from Bode plots using the formula [35] s ¼ 2p1x, where x represents frequency peak in the low frequency range (S6) [23] and all EIS parameters for a typical DSSC are presented in Table 1. As evident from the table, the recombination resistance as well as electron life time increases with incorporation of mesoporous SiO2 up to 0.75 wt% and decreases thereafter. The increase in the recombination resistance upon incorporation of mesoporous SiO2 in TiO2 suggests hindrance of charge carriers towards back recombination. In addition, an increase in life time indicates that mesoporous SiO2 acts as energy barrier between TiO2 and electrolyte and therefore, reduced back electron transfer. This is also evident from the dark current measurements which shows that: (i) dark current is reduced substantially after incorporation of mesoporous SiO2 particles in TiO2 and (ii) the shift in dark Table 1 EIS parameters estimated for three devices each fabricated using T, 0.5 S, 0.75 S and 1 S. Electrodes
Rtr (Ohm cm2)
s (msec)
T 0.5 S 0.75 S 1S
7.04 11.78 13.44 10.94
11.47 16.48 23.66 16.48
current onset potential is larger for 0.75 S electrode as compared with 1 S electrode. Therefore the recombination between I 3 and CB electron of TiO2 is suppressed for 0.75 S electrode as compared with 1 S electrode leading to similar trend for FF as explained below: FF can be expressed in terms of Voc and b [34,36]
FF ¼
V r lnðV r þ 0:72Þ Vr þ 1
oc where V r ¼ bqV is the reduced voltage and b recombination Kb T
exponent. Since Voc remains almost unaffected therefore, FF is controlled by recombination exponent which is directly proportional to recombination resistance and hence follows the same trend as recombination resistance. Therefore, overall improvement in the device is observed for DSSCs fabricated with 0.75 wt% of mesoporous SiO2. 4. Conclusions In summary, we have synthesized mesoporous SiO2 with appropriate size which could be utilized as efficient scatterers and recombination inhibitors in DSSCs. The DSSCs fabricated exhibit power conversion efficiency as large as 3.57% using 2.5 lm thick TiO2 film incorporated with 0.75 wt% of mesoporous SiO2. The enhancement in photovoltaic performance was attributed to enhanced short circuit current density and FF, which is confirmed by absorption spectra, and electrochemical impedance spectroscopy.
Tanvi et al. / Chemical Physics Letters 658 (2016) 276–281
Acknowledgements Authors are thankful to Dr. K.P. Muthe and Dr. Ajay Singh for their support and helpful discussions. One of the authors, Tanvi, is grateful to Department of Science and Technology, New Delhi (India) under INSPIRE program for fellowship. Authors A.M., R.K. B. and S.K. are thankful to UPE program under UGC, New Delhi (India). Financial assistance provided by BRNS, Mumbai to carry out the present research work is acknowledged. Appendix A. Supplementary material
[11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cplett.2016.06. 071. References [1] M. Gratzel, Acc. Chem. Res. 42 (2009) 1788. [2] H.J. Snaith, Adv. Funct. Mater. 20 (2010) 13. [3] V. Saxena, V. Putta, A.K. Chauhan, P. Jha, D.K. Aswal, S.K. Gupta, Appl. Phys. Lett. 100 (2012) 133303. [4] S.-J. Lin, K.-C. Lee, J.-L. Wu, J.-Y. Wu, Appl. Phys. Lett. 99 (2011) 043306. [5] Tanvi, A. Mahajan, R.K. Bedi, S. Kumar, V. Saxena, D.K. Aswal, J. Appl. Phys. 117 (2015) 083111. [6] Q. Yu, S. Liu, M. Zhang, N. Cai, Y. Wang, P. Wang, J. Phys. Chem. C 113 (2009) 14559. [7] T.G. Deepak, G.S. Anjusree, S. Thomas, T.A. Arun, S.V. Nair, A. Sreekumaran Nair, RSC Adv. 4 (2014) 17615. [8] Wu, X. Gao, M. Qin, J.-M. Liu, S. Hu, Appl. Phys. Lett. 99 (2011) 042106. [9] E. Palomares, J.N. Clifford, S.A. Haque, T. Lutz, J.R. Durrant, J. Am. Chem. Soc. 125 (2003) 475. [10] H.-J. Son, X. Wang, C. Prasittichai, N.C. Jeong, T. Aaltonen, R.G. Gordon, J.T. Hupp, J. Am. Chem. Soc. 125 (2003) 475.
[22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36]
281
Q. Hu, C. Wu, L. Cao, B. Chi, J. Pu, L. Jian, J. Power Sources 226 (2013) 8. W. Ma, Y. Jiao, S. Meng, Phys. Chem. Chem. Phys. 15 (2013) 17187. W.G. Yang, F.R. Wan, Q.W. Chen, J.J. Li, D.S. Xu, J. Mater. Chem. 20 (2010) 2870. S. Son, S.H. Hwang, C. Kim, J.Y. Yun, J. Jang, ACS Appl. Mater. Interfaces 5 (2013) 4815. Y. Wang, E. Chen, H. Lai, B. Lu, Z. Hu, X. Qin, W. Shi, G. Du, Ceram. Int. 39 (2013) 5407. Y. Rho, M. Wanit, J. Yeo, S. Hong, S. Han, J.-H. Choi, W.-H. Hong, D. Lee, S.H. Ko, J. Phys. D: Appl. Phys 46 (2013) 024006. K.-H. Park, H.-B. Gu, E.M. Jin, M. Dhayal, Electrochim. Acta 55 (2010) 5499. Z.-S. Wang, H. Kawauchi, T. Kashima, H. Arakawa, Coord. Chem. Rev. 248 (2004) 1381. A.G. Niaki, A. Bakhshayesh, M. Mohammadi, Sol. Energy 103 (2014) 210222. D. Sengupta, P. Das, U. Kasinadhuni, B. Mondal, K. Mukherjee, J. Renew. Sustain. Energy 6 (2014) 063114. V. Saxena, V. Putta, A. Gusain, P. Jha, J. Singh, S. Koiry, P. Varde, A. Chauhan, D. Aswal, S. Gupta, Org. Electron. 14 (2013) 3098. D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Science 279 (1998) 548. See supplemental material for XRD, dye loading, bandgap analysis, electron lifetime measurements and current–voltage characteristics. A. Sayari, B.H. Han, Y. Yang, J. Am. Chem. Soc. 126 (2004) 14348. Q. Lu, Z. Wang, J. Li, P. Wang, X. Ye, Nanoscale Res. Lett. 4 (2009) 646. Y. Xie, B. Yan, C. Tian, Y. Liu, Q. Liu, H. Zeng, J. Mater. Chem. A 2 (2014) 17730. L. Ellis-Gibbings, V. Johansson, R.B. Walsh, L. Kloo, J.S. Quinton, G.G. Andersson, Langmuir 28 (2012) 9431. A. Zrner, J. Kirstein, M. Dblinger, C. Bruchle, T. Bein, Nature 450 (2007) 705. S. Yuan, Q. Tang, B. He, L. Menand, H. Chen, Electrochim. Acta 125 (2014) 646. P. Xu, Q. Tang, B. He, Q. Li, H. Chen, Electrochim. Acta 134 (2014) 281. Z. Wang, Q. Tang, B. He, H. Chen, L. Yu, Electrochim. Acta 178 (2015) 18. Y. Duan, Q. Tang, Z. Chen, B. He, H. Chen, J. Mater. Chem. A 2 (2014) 12459. Z. Wang, Q. Tang, B. He, X. Chen, H. Chen, L. Yu, J. Power Sources 1 (2014) 175. F. Fabregat-Santiago, G. Garcia-Belmonte, I. Mora-Sero, J. Bisquert, Phys. Chem. Chem. Phys. 13 (2011) 9083. R. Kern, R. Sastrawan, J. Ferber, R. Stangl, J. Luther, Electrochim. Acta 47 (2002) 4213. Z. Huang, G. Natu, Z. Ji, M. He, M. Yu, Y. Wu, J. Phys. Chem. C 116 (2012) 26239.