- Email: [email protected]
Cost−effective alloy counter electrodes as a new avenue for high−efficiency dye−sensitized solar cells
Electrochimica Acta 158 (2015) 397–402
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Costeffective alloy counter electrodes as a new avenue for highefficiency dyesensitized solar cells Benlin He a,b , Qunwei Tang a,b, * , Liangmin Yu a, * , Peizhi Yang c a
Key Laboratory of Marine Chemistry Theory and Technology, Ministry of Education, Ocean University of China, Qingdao 266100, PR China Institute of Materials Science and Engineering, Ocean University of China, Qingdao 266100, PR China Key Laboratory of Advanced Technique & Preparation for Renewable Energy Materials, Ministry of Education, Yunnan Normal University, Kunming 650092, PR China b c
A R T I C L E I N F O
A B S T R A C T
Article history: Received 22 October 2014 Received in revised form 19 January 2015 Accepted 30 January 2015 Available online 31 January 2015
Pursuit of costeffective and efficient counter electrodes (CEs) has been a persistent objective for dyesensitized solar cells (DSSCs). Aiming at reducing fabrication cost without sacrificing power conversion efficiency of DSSCs, here we report the successful design of binary Pt–Ni alloy CEs by a simple cyclic voltammetry technique. Due to the rapid charge transfer ability and electrocatalytic activity, the power conversion efficiency of the DSSC employing binary PtNi0.75 alloy CE has been elevated to 8.59% in comparison with 6.98% from Ptbased solar cell. The impressive results along with simple synthesis highlight the potential application of lowPt alloys in robust DSSCs. ã 2015 Elsevier Ltd. All rights reserved.
Keywords: Dyesensitized solar cell Counter electrode Binary alloy Electrochemical deposition
1. Introduction Fossil fuels such as coal, oil, and natural gas have occupied energy market for hundreds of years. However, the combustion of these nonrenewable fuels can release sulfides, carbides and dusts which will damage ecology and pollute environment. By addressing these issues, it is a prerequisite to develop renewable, green, and environmentfriendly energy resources to resolve energy and environment crisis [1]. Among various energy candidates, dyesensitized solar cells (DSSCs) [2–5], electrochemical devices converting solar energy into electrical power honored by a high efficiency and no environmental impact, are promising solutions to global energy and environmental problems because of clean, high efficiency, good durability, and relatively simple fabrication. Since the first prototype reported by Grätzel in 1991 [6], DSSCs have attracted growing interests and great achievements have been made. However, it is still premature for their commercialization. Until now, the most limiting factor in the development of commercial DSSCs has been their cost [7,8]. The historically high prices for Pt feedstock, a traditional counter electrode (CE) material, have meant that a cell could not be fabricated at a cost low enough to compete with conventional silicon solar cells. The
* Corresponding authors. Tel.: +86 532 66781690; fax: +86 532 66782533. E-mail addresses: [email protected] (Q. Tang), [email protected] (L. Yu). http://dx.doi.org/10.1016/j.electacta.2015.01.194 0013-4686/ ã 2015 Elsevier Ltd. All rights reserved.
task of a CE is to collect electrons from external circuit and to reduce triiodide (I3) into iodide (I), therefore, an efficient CE electrocatalyst should display good chargetransfer ability and excellent electrocatalytic activity. Other candidates such as carbonaceous materials [9,10], conducting polymers [11,12], or their composites [13–15] present either modest electrocatalysis, unsatisfactory electronconduction, or large interfacial resistance. More importantly, such CEs have a fast attenuation in electrocatalytic activity and longterm stability [16]. Therefore, it is a prerequisite to develop costeffective but stable CEs before DSSCs are becoming a commercial reality. Alloy materials have established themselves as the alternative electrocatalysts for fuel cell applications in the past two decades [17,18]. Aiming to increase the catalytic activity of the electrode kinetics and to lower the cost of electrocatalysts, it has been shown that alloying of Pt metal with transition metals could be an efficient route to meet the cell requirements. However, the employment of Ptfree or lowPt alloys in DSSCs is still at preliminary stage. Previous studies in our group have revealed that binary M–Pt (where M = Co, Ru) alloys have exceptional electrocatalysis toward I/I3 redox couples [19,20]. However, Co species in Co–Pt alloys are easily oxided and Ru sources are relative price high. In searching for other robust M–Pt alloys, here we report the synthesis and characterization of a new class binary Ni–Pt alloy CEs. The resultant Pt–Ni alloys show supercatalytic behavior toward I/I3 redox couples, allowing the rapid interconversion
398
B. He et al. / Electrochimica Acta 158 (2015) 397–402
between I3 ! I. The Pt–Ni alloy CEs are expected to significantly enhance the electron conduction, charge-transfer ability, and therefore power conversion efficiency of the DSSCs. 2. Experimental 2.1. Preparation of binary Pt–Ni alloy CEs The feasibility of this strategy was confirmed by a cyclic voltammetry (CV) method: A cleaned FTO glass substrate was used as a working electrode, and the CV curves were recorded from 0.4 to +0.3 V and back to 0.4 V for 10 cycles. Before the measurement, the supporting electrolyte consisting of 2 mM H2PtC16, 1 mM of Ni(NO3)2, and 5 mM of HCl aqueous solution was degassed using nitrogen for 10 min.
2.3. Electrochemical characterizations The electrochemical performances were recorded on a conventional CHI660E setup comprising an Ag/AgCl reference electrode, a CE of platinum sheet, and a working electrode of FTO glass supported Pt–Ni alloy. The CV curves were recorded from 1.0 to +1.4 V and back to 1.0 V. Before the measurement, the supporting electrolyte consisting of 50 mM M LiI, 10 mM I2, and 500 mM LiClO4 in acetonitrile was degassed using nitrogen for 10 min. Electrochemical impedance spectroscopy (EIS) measurements were also carried out on the CHI660E Electrochemical Workstation in a frequency range of 0.01 Hz 106 kHz and an ac amplitude of 5 mV at room temperature. The resultant impedance spectra were analyzed using the Zview software. Tafel polarization curves were recorded on the same Workstation by assembling symmetric cell consisting of Pt–Ni alloy CE|redox electrolyte|Ni–Pt alloy CE.
2.2. Assembly of DSSCs A layer of TiO2 nanocrystal anode film with a thickness of 10 mm was prepared by a sol-hydrothermal method [21] and a layer of TiO2 nanocrystal anode film with a thickness of 10 mm and an active area of 0.25 cm2 was prepared by coating TiO2 colloid onto conducting glass using a doctor blade technique, followed by sintering in air at 450 C for 30 min. Resultant anodes were further sensitized by immersing into a 0.50 mM ethanol solution of N719 dye ([cisdi(thiocyanato) N,N’ bis(2,2’ bipyridyl 4carboxylic acid)4tetrabutylammonium carboxylate]). The DSSC was fabricated by sandwiching redox electrolyte between dyesensitized TiO2 anode and FTO supported Pt–Ni alloy CEs. A redox electrolyte consisted of 100 mM of tetraethylammonium iodide, 100 mM of tetramethylammonium iodide, 100 mM of tetrabutylammonium iodide, 100 mM of NaI, 100 mM of KI, 100 mM of LiI, 50 mM of I2, and 500 mM of 4tertbutylpyridine in 50 ml acetonitrile.
2.4. Photovoltaic measurements The photocurrent–voltage (J–V) curves of the assembled DSSCs were recorded on an Electrochemical Workstation (CHI600E) under irradiation of a simulated solar light from a 100 W xenonmercury arc lamp (CHFXM500 W, Beijing Trusttech Co., Ltd) in an ambient atmosphere. The incident light intensity was calibrated using a FZA type radiometer from Beijing Normal University Photoelectric Instrument Factory to control it at 100 mW cm2 (AM1.5 calibrated by a standard silicon solar cell). Each DSSC device was measured at least five times to eliminate experimental error and a compromise J–V curve was employed. 2.5. Other characterizations The morphologies of the Pt–Ni alloy CE were observed with a scanning electron microscope (SEM, S4800). The XRD data were
Fig. 1. Topview SEM photographs of (a) PtNi0.25, (b) PtNi0.50, (c) PtNi0.75, and (d) PtNi alloy CEs.
B. He et al. / Electrochimica Acta 158 (2015) 397–402
collected in a scan mode with a scanning speed of 10 min1 in the 2u range between 20 and 80 . XPS experiment was carried out on a RBD upgraded PHI5000C ESCA system (Perkin Elmer) with Mg Ka radiation (hn = 1253.6 eV).
characterization is carried out to determine the chemical structure and composition of PtNi0.75 alloy. As shown in Fig. 2b and Fig. 2c, the peaks for Pt4f and Ni2p are centered at 71.28 eV (Pt4f7/2), 71.45 eV (Pt4f5/2), and 852.5 eV (Ni2p3/2), confirming the metallic nature of assynthesized PtNi0.75 alloy [24]. It has been mentioned that the alloying of transition metals can favor the electronic perturbation of other metals [25] and therefore accelerate the electrocatalytic activity of the alloys for their electrocatalyst applications. The binary alloy CEs synthesized by a CV method are subjected to XRD measurements. As shown in Fig. 2d, XRD results indicate that the alloy CEs consist of alloy materials at 2u = 39.96 and FTO glass substrate marked with (*). Moreover, unalloyed Pt is also detected because of the appearance of diffraction peak at 2u = 46.24 . A golden rule in determining an efficient CE is to evaluate its electrocatalytic activity toward electrolyte [26]. As displayed in Fig. 3a, the CV peak shapes and positions of alloy CEs are similar to that of pristine Pt electrode, suggesting that the alloys have a similar electrocatalytic function to Pt CE. However, the peak current densities of CV curves have been elevated by alloying Pt with Ni in comparison with pristine Pt electrode. Considering the task of a CE is to collect electrons from external circuit and to reduce I3 to I, therefore, Red1 peak can be employed to assess the catalytic activity of alloy CEs. The ratio of Jox1/|Jred1| is a parameter to elevate the reversibility of the redox reaction toward I/I3 species [27]. The obtained values for pristine Pt, PtNi0.25, PtNi0.50, PtNi0.75, and PtNi are 1.155, 0.985, 0.990, 1.007, and 0.966,
3. Results and discussion SEM photographs in Fig. 1 suggest a high surface coverage and loading of alloys on FTO glass. Deep examination gives homogeneously spherical aggregations. However, the compact structures in PtNi0.25 and PtNi0.50 alloy CEs can hinder the diffusion of I/I3 redox couples at CE/electrolyte interface. By optimizing stoichiometry of Pt and Ni sources, a loose structure with an average aggregation size of 200 nm is obtained for PtNi0.75 alloy CE, providing channels for I/I3 species across the alloy layer [22]. At a high Ni dosage, such as PtNi alloy, the cracks and large interfacial resistance may limit electron migration from FTO back to alloy layer. Energydispersive Xray spectra (EDS), as shown in Fig. 2a, demonstrate that Pt and Ni elements as well as partial elements from FTO glass are detected, indicating a successful deposition of Pt and Ni on FTO substrate. The measured atomic ratios by EDS for PtNi0.25, PtNi0.50, PtNi0.75, and PtNi alloys are 1.000:0.266, 1.000:0.518; 1.000:0.752, and 1.000:1.083, respectively. The measured atomic ratios are relatively close to their stoichiometries, therefore the chemical formulae of the alloy CEs can be expressed according to the stoichiometric ratios [23]. XPS
C O Ni
Pt
b
PtNi0.25
Sn
71.45 eV
}
PtNi0.75 PtNi
cps 0
2
4
80
6
75
Energy (keV)
c
Pt4f
71.28 eV
PtNi0.50
Counts (a.u.)
a
399
70
65
Binding energy (eV)
852.5 eV
d
Ni2p
PtNi0.25
840
850
860
870
880
20
* *
40
* Pt (200)
Counts (a.u.)
Intensity (a.u.)
*
Pt-Ni alloy
PtNi0.50
2theta (degree)
PtNi0.75 PtNi
* *
60
Binding energy (eV) Fig. 2. (a) EDS of Pt–Ni alloy CEs, XPS spectra of (b) Pt4f and (c) Ni2p in PtNi0.75 alloy, and (c) XRD patterns of FTO glass supported alloy CEs.
80
400
B. He et al. / Electrochimica Acta 158 (2015) 397–402
Fig. 3. (a) CV curves of various CEs for I/I3 redox species recorded at a scan rate of 50 mV s1. (b) CV curves of PtNi0.75 alloy CE for I/I3 redox species at different scan rates. The insert gives the plots of peak current density versus scan rate. (c) A total of 100 consecutive CVs for the I/I3 system using the PtNi0.75 alloy CE at a scan rate of 50 mV s1. (d) Relationship between the cycle time and the redox peak current for the PtNi0.75 alloy CE.
respectively. The high peak current density, low peaktopeak separation (Epp, 0.528 V for PtNi0.75 and 0.632 V for Pt) [28], and good reversibility demonstrate the alloying of Pt with Ni can markedly enhance the reversibility of I3 $ I reaction, and PtNi0.75 alloy has superiority of catalyzing liquid electrolyte containing I/I3 redox couples. From the stacking CV curves of PtNi0.75 alloy CE at different scan rates, one can find an outward extension of redox peaks (Fig. 3b). By plotting peak current density corresponding to I3 $ I versus square root of scan rate, linear relationships are observed, indicating a diffusion controlled mechanism of redox reaction on alloy CE [29]. A total of 100 consecutive CV curves for the I/I3 system using the PtNi0.75 alloy CE at a scan rate of 50 mV s1 are shown in Fig. 3c, no obvious attenuation of peak current densities in Ox1/Red1 means the alloy CE is relatively stable for reducing I3 ions (Fig. 3d). Nyquist EIS plots of the symmetric dummy cells by two identical CEs are employed to demonstrate the intrinsic interfacial chargetransfer kinetics at CE/electrolyte interface. As presented in Fig. 4a, the Rct value of the PtNi0.75 alloy CE is 3.03 V cm2, smaller than 5.75 V cm2 for pristine Pt electrode (Table 1). The Rct of CEs follows an order of PtNi0.75 < PtNi0.50 < PtNi0.25 < Pt < PtNi. The sequence is consistent with the results of CV curves. Enhanced electrocatalytic kinetics and chargetransfer ability could accelerate dye regeneration and therefore elevate the Jsc of cells. Due to the participation of electrons in I3 reduction reaction, therefore
the electron lifetime (t ) at CE/electrolyte interface can be calculated from t = 1/(2pfp) and utilized to assess the reduction reaction kinetics [30], where fmax is the maximum frequency of the midfrequency peak in the Bode phase plots (Fig. 4b). The t at CE/electrolyte interface has an order of PtNi0.75 (42 ms) < PtNi0.50 (50 ms) < PtNi0.25 (61 ms) < Pt (74 ms) < PtNi (90 ms). A low t value refers to a rapid reduction reaction occurred at CE/electrolyte interface. Moreover, the exchange current density (J0, the slope for anodic or cathodic branches) and limiting diffusion current density (Jlim, the intersection of cathodic branch with Y-axis), in the Tafel polarization curves can also be uitilized to reconfirm the catalytic activity of CEs. As shown in Fig. 4c, both J0 and Jlim have an order of PtNi0.75 > PtNi0.50 > PtNi0.25 > Pt > PtNi. From equations J0 = RT/nFRct [31],Jlim = 2nFCDn/l [32], and Jred1 = Kn1.5ACDn0.5v0.5 [33] (R is the gas constant, F is Faraday’s constant, n is number of electrons participating in I3 reduction reaction, l is the distance between electrodes in a symmetric dummy cell, C is concentration of I3 ions), one can conclude that the results from Tafel polarization curves, EIS plots, and CV curves are in good agreement. Fig. 4d compares characteristic J–V curves of the DSSCs with pristine Pt and alloy CEs. The solar cell from PtNi0.75 electrode yields a maximum h of 8.59% (Voc = 0.716 V, Jsc = 17.50 mA cm2, FF = 68.6%), which are much higher than the photovoltaic parameters with pristine Pt CE and those in literatures [34,35], The solar cell from PtNi0.75 electrode. This may be attributed to a fact that the rapid
B. He et al. / Electrochimica Acta 158 (2015) 397–402
401
Fig. 4. (a) Nyquist and (b) Bode EIS plots and (c) Tafel polarization curves for symmetric dummy cells fabricated two identical CEs. (d) Characteristic J–V curves for the DSSCs with various CEs. The insert gives an equivalent circuit for the symmetric dummy cells.
interconvertion between I3 and I can accelerate electron generation from N719 dye and therefore elevate the electron flow from excited dye to conduction band of TiO2, therefore the electron density on TiO2 conduction band is markedly enhanced. The recorded h from PtNi0.75 alloy is impressive for the DSSCs with lowPt CEs. Critical requirements of the engines and vehicles drived by solar panels are high power conversion efficiency, costeffectiveness, fast startup, multiple start capability, and operational stability. The efficiency and expenses greatly depend on the design of CEs, in which Ptfree or lowPt materials are always required. However, a robust catalyst in CE is desirable for fast startup, multiple start, and stability [36]. As shown in Fig. 5a, the photocurrent of the cell device employing PtNi0.75 alloy electrode has an abrupt increase at full irradiation. No delay in time is detected in starting the cell, this
observation supports a fact that the alloy CEs are robust in catalyzing I/I3 redox couples. After five start–stop cycles, there is no failure in photocurrent density, however, the current density from Pt based solar cell has an apparent decrease in the first two cycles. These results suggest that PtNi0.75 alloy is a preferred CE material for efficient DSSCs. Additionally, the relationship between photocurrent density and time demonstrates cell stability on prolonged exposure to light irradiation (100 mW cm2). As displayed in Fig. 5b, the photocurrent densities for DSSCs with PtNi0.75 and Pt decrease by 6.75% and 14.69% over 5 h, respectively. In comparison with the solar cell with Pt electrode [37], the DSSC employing PtNi0.75 alloy electrode displays enhanced stability. In order to explore potential mechanism for enhanced catalytic activity and stability, the Gibbs energy of the possible dissolving reaction of PtNi0.75 alloy is calculated [38]. The values of Gibbs free
Table 1 Comparison of photovoltaic and electrochemical parameters as well as electron lifetime and transparency for the DSSCs. h: power conversion efficiency; Voc: opencircuit voltage; Jsc: shortcircuit current density; FF: fill factor; Rct: chargetransfer resistance; W: Nernst diffusion resistance corresponding to I/I3 couples; t : electron lifetime at CE/electrolyte interface. CEs
h (%)
Voc (V)
Jsc (mAcm2)
FF (%)
Rct (V cm2)
W (V cm2)
t (ms)
Pristine Pt PtNi0.25 PtNi0.50 PtNi0.75 PtNi
6.98 7.70 8.05 8.59 6.46
0.698 0.703 0.717 0.716 0.703
14.63 16.07 16.45 17.50 13.75
68.4 68.2 68.3 68.6 66.8
13.00 9.18 5.42 3.03 13.26
5.75 3.88 2.92 1.85 8.47
74 61 50 42 90
402
B. He et al. / Electrochimica Acta 158 (2015) 397–402
b 20
a
20
PtNi0.75 -2
Current density mA cm
-2
Current density (mA cm )
Pt
15
10
5
0
15
PtNi0.75
10
Pt
5
0 0
50
100 150 Time (s)
200
250
0
1
2
3
4
5
Time (h)
Fig. 5. (a) Startstop switches and (b) photocurrent stabilities of the DSSCs employing PtNi0.75 alloy and Pt electrodes. The on–off plots were achieved by alternatively irradiating (100 mW cm2) and darkening (0 mW cm2) the DSSC devices at an interval of 25 s and 0 V, whereas the photocurrent stabilities were carried out under sustained irradiation of 100 mW cm2.
energy (DrGm) were calculated for the possible reactions between Ni or Pt and triiodide species [Reaction 1: Pt (s) + 2I3 (aq) = PtI4 (s) + 2I (aq); Reaction 2: Ni (s) + I3 (aq) = NiI2 (s) + I (aq)]. The DrGm values for reaction 1 and reaction 2 are 45.9 and 76.4 kJ mol1, respectively. In this fashion, the reaction between Ni species and triiodides is easier. The alloying of Ni with Pt can form a competed reaction between Ni and Pt, therefore protecting the high catalytic activity of Pt species by sacrificing Ni species. The presented results demonstrate the superiority of PtNi0.75 alloy CE than pristine Pt electrode for DSSC application. 4. Conclusions In summary, binary Pt–Ni alloys have been fabricated by an electrochemical codeposition strategy free of any surfactant or template and employed as CE materials in DSSCs. It is demonstrated that PtNi0.75 alloy CE has an optimal chargetransfer ability and electrocatalytic activity toward I3 reduction. The DSSC employing PtNi0.75 alloy CE exhibits impressive power conversion efficiency of 8.59% in comparison with 6.98% from pristine Pt CE based device. The research presented here is far from being optimized but these profound advantages along with costeffectiveness, mild synthesis and scalable materials promise the binary Pt–Ni alloy CEs to be strong candidates in robust DSSCs. Acknowledgements The authors would like to acknowledge financial supports from Fundamental Research Funds for the Central Universities (201313001, 201312005), Shandong Province Outstanding Youth Scientist Foundation Plan (BS2013CL015), Shandong Provincial Natural Science Foundation (ZR2011BQ017), Research Project for the Application Foundation in Qingdao (13-4-198-jch), National Natural Science Foundation of China (51102219, 51342008, U1037604), National Key and National Key Technology Support Program (2012BAB15B02). References [1] J.M. Tour, C. Kittrell, V.L. Colvin, Nat. Mater. 9 (2010) 871–874. [2] N. Memarian, I. Concina, A. Braga, S.M. Rozati, A. Vomiero, G. Sberveglieri, Angew. Chem. Int. Ed. 50 (2011) 12321–12325. [3] A. Yella, H.W. Lee, H.N. Tsao, C. Yi, A.K. Chandiran, M.K. Nazeeruddin, E.W.G. Diau, C.Y. Yeh, S.M. Zakeeruddin, M. Grätzel, Science 334 (2011) 629–634.
[4] J.H. Wu, Y.M. Xiao, Q.W. Tang, G.T. Yue, J.M. Lin, M.L. Huang, Y.F. Huang, L.Q. Fan, Z. Lan, S. Yin, T. Sato, Adv. Mater. 24 (2012) 1884–1888. [5] X.X. Chen, Q.W. Tang, B.L. He, L. Lin, L.M. Yu, Angew. Chem. Int. Ed. 53 (2014) 10799–10803. [6] B. O’Regan, M. Grätzel, Nature 353 (1991) 737–740. [7] M.X. Wu, T.L. Ma, J. Phys. Chem. C 118 (2014) 16727–16742. [8] J. Liu, Q.W. Tang, B.L. He, J. Power Sources 268 (2014) 56–62. [9] J.D. Roy-Mayhew, D.J. Bozym, C. Punckt, I.A. Aksay, Nano 4 (2010) 6203–6211. [10] B. Tang, G.X. Hu, H.Y. Gao, Z.X. Shi, J. Power Sources 234 (2013) 60–68. [11] Q.W. Tang, H.Y. Cai, S.S. Yuan, X. Wang, J. Mater. Chem. A 1 (2013) 317–323. [12] J.H. Wu, Y. Li, Q.W. Tang, G.T. Yue, J.M. Lin, M.L. Huang, L.J. Meng, Sci. Rep. 4 (2014) 4028. [13] B.L. He, Q.W. Tang, T.L. Liang, Q.H. Li, J. Mater. Chem. A 2 (2014) 3119–3126. [14] B.L. He, Q.W. Tang, M. Wang, H.Y. Chen, S.S. Yuan, ACS Appl. Mater. Interfaces 6 (2014) 8230–8236. [15] J. Ma, L. Zhou, C. Li, J. Yang, T. Meng, H. Zhou, M. Yang, F. Yu, J. Chen, J. Power Sources 247 (2014) 999–1004. [16] F. Gong, H. Wang, X. Xu, G. Zhou, Z.S. Wang, J. Am. Chem. Soc. 134 (2012) 10953–10958. [17] X. Chen, Y. Jiang, J. Sun, C. Jin, Z. Zhang, J. Power Sources 267 (2014) 212–218. [18] Y. Bing, H. Liu, L. Zhang, D. Ghosh, J. Zhang, Chem. Soc. Rev. 39 (2010) 2184–2202. [19] B.L. He, X. Meng, Q.W. Tang, ACS Appl. Mater. Interfaces 6 (2014) 4812–4818. [20] H.Y. Cai, Q.W. Tang, B.L. He, P.J. Li, J. Power Sources 258 (2014) 117–121. [21] B.B. Hu, Q.W. Tang, B.L. He, L. Lin, H.Y. Chen, J. Power Sources 267 (2014) 445–451. [22] S. Cho, S.H. Hwang, C. Kim, J. Jang, J. Mater. Chem. 22 (2012) 12164–12171. [23] Y.Y. Duan, Q.W. Tang, B.L. He, R. Li, L.M. Yu, Nanoscale 6 (2014) 12601–12608. [24] K. Park, J. Choi, B. Kwon, S. Lee, Y. Sung, H. Ha, S. Hong, H. Kim, A. Wieckowski, J. Phys. Chem. B 106 (2002) 1869–1877. [25] C. Xu, J. Hou, X. Pang, X. Li, M. Zhu, B. Tang, Int. J. Hydrogen Energy 37 (2012) 10489–10498. [26] S. Thomas, T.G. Deepak, G.S. Anjusree, T.A. Arun, S.V. Nair, A.S. Nair, J. Mater. Chem. A 2 (2014) 4474–4490. [27] H. Tributsch, Coordin. Chem. Rev. 248 (2004) 1511–1530. [28] B.L. He, Q.W. Tang, J.H. Luo, Q.H. Li, X.X. Chen, H.Y. Cai, J. Power Sources 256 (2014) 170–177. [29] Z.B. Lv, J.F. Yu, H.W. Wu, J. Shang, D. Wang, S.C. Hou, Y.P. Fu, K. Wu, D.C. Zou, Nanoscale 4 (2012) 1248–1253. [30] J. Yoon, M. Jin, M. Lee, Adv. Mater. 23 (2011) 3974–3978. [31] M. Wu, X. Lin, Y. Wang, L. Wang, W. Guo, D. Qi, X. Peng, A. Hagfeldt, M. Grätzel, T.L. Ma, J. Am. Chem. Soc. 134 (2014) 3419–3428. [32] M.K. Wang, A.M. Anghel, B. Marsan, N.C. Ha, N. Pootrakulchote, S.M. Zakeeruddin, M. Grätzel, J. Am. Chem. Soc. 131 (2009) 15976–15977. [33] T. Daeneke, A.J. Mozer, T.H. Kwon, N.W. Duffy, A.B. Holmes, U. Bach, L. Spiccia, Energy Environ. Sci. 5 (2012) 7090–7099. [34] Y. Gong, C. Li, X. Huang, Y. Luo, D. Li, Q. Meng, B.B. Iversen, ACS Appl. Mater. Interfaces 5 (2013) 795–800. [35] V.-D. Dao, S.-H. Kim, H.-S. Choi, J.-H. Kim, H.-O. Park, J.-K. Lee, J. Phys. Chem. C 115 (2011) 25529–25534. [36] D. Papadias, S.H.D. Lee, Ind. Eng. Chem. Res. 45 (2006) 5841–5858. [37] G. Syrrokostas, A. Siokou, G. Leftheriotis, P. Yianoulis, Sol. Energy Mater. Sol. Cells 103 (2012) 119–127. [38] E. Olsen, G. Hagen, S.E. Lindquist, Sol. Energy Mater. Sol. Cells 63 (2000) 267–273.
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Costeffective alloy counter electrodes as a new avenue for highefficiency dyesensitized solar cells Benlin He a,b , Qunwei Tang a,b, * , Liangmin Yu a, * , Peizhi Yang c a
Key Laboratory of Marine Chemistry Theory and Technology, Ministry of Education, Ocean University of China, Qingdao 266100, PR China Institute of Materials Science and Engineering, Ocean University of China, Qingdao 266100, PR China Key Laboratory of Advanced Technique & Preparation for Renewable Energy Materials, Ministry of Education, Yunnan Normal University, Kunming 650092, PR China b c
A R T I C L E I N F O
A B S T R A C T
Article history: Received 22 October 2014 Received in revised form 19 January 2015 Accepted 30 January 2015 Available online 31 January 2015
Pursuit of costeffective and efficient counter electrodes (CEs) has been a persistent objective for dyesensitized solar cells (DSSCs). Aiming at reducing fabrication cost without sacrificing power conversion efficiency of DSSCs, here we report the successful design of binary Pt–Ni alloy CEs by a simple cyclic voltammetry technique. Due to the rapid charge transfer ability and electrocatalytic activity, the power conversion efficiency of the DSSC employing binary PtNi0.75 alloy CE has been elevated to 8.59% in comparison with 6.98% from Ptbased solar cell. The impressive results along with simple synthesis highlight the potential application of lowPt alloys in robust DSSCs. ã 2015 Elsevier Ltd. All rights reserved.
Keywords: Dyesensitized solar cell Counter electrode Binary alloy Electrochemical deposition
1. Introduction Fossil fuels such as coal, oil, and natural gas have occupied energy market for hundreds of years. However, the combustion of these nonrenewable fuels can release sulfides, carbides and dusts which will damage ecology and pollute environment. By addressing these issues, it is a prerequisite to develop renewable, green, and environmentfriendly energy resources to resolve energy and environment crisis [1]. Among various energy candidates, dyesensitized solar cells (DSSCs) [2–5], electrochemical devices converting solar energy into electrical power honored by a high efficiency and no environmental impact, are promising solutions to global energy and environmental problems because of clean, high efficiency, good durability, and relatively simple fabrication. Since the first prototype reported by Grätzel in 1991 [6], DSSCs have attracted growing interests and great achievements have been made. However, it is still premature for their commercialization. Until now, the most limiting factor in the development of commercial DSSCs has been their cost [7,8]. The historically high prices for Pt feedstock, a traditional counter electrode (CE) material, have meant that a cell could not be fabricated at a cost low enough to compete with conventional silicon solar cells. The
* Corresponding authors. Tel.: +86 532 66781690; fax: +86 532 66782533. E-mail addresses: [email protected] (Q. Tang), [email protected] (L. Yu). http://dx.doi.org/10.1016/j.electacta.2015.01.194 0013-4686/ ã 2015 Elsevier Ltd. All rights reserved.
task of a CE is to collect electrons from external circuit and to reduce triiodide (I3) into iodide (I), therefore, an efficient CE electrocatalyst should display good chargetransfer ability and excellent electrocatalytic activity. Other candidates such as carbonaceous materials [9,10], conducting polymers [11,12], or their composites [13–15] present either modest electrocatalysis, unsatisfactory electronconduction, or large interfacial resistance. More importantly, such CEs have a fast attenuation in electrocatalytic activity and longterm stability [16]. Therefore, it is a prerequisite to develop costeffective but stable CEs before DSSCs are becoming a commercial reality. Alloy materials have established themselves as the alternative electrocatalysts for fuel cell applications in the past two decades [17,18]. Aiming to increase the catalytic activity of the electrode kinetics and to lower the cost of electrocatalysts, it has been shown that alloying of Pt metal with transition metals could be an efficient route to meet the cell requirements. However, the employment of Ptfree or lowPt alloys in DSSCs is still at preliminary stage. Previous studies in our group have revealed that binary M–Pt (where M = Co, Ru) alloys have exceptional electrocatalysis toward I/I3 redox couples [19,20]. However, Co species in Co–Pt alloys are easily oxided and Ru sources are relative price high. In searching for other robust M–Pt alloys, here we report the synthesis and characterization of a new class binary Ni–Pt alloy CEs. The resultant Pt–Ni alloys show supercatalytic behavior toward I/I3 redox couples, allowing the rapid interconversion
398
B. He et al. / Electrochimica Acta 158 (2015) 397–402
between I3 ! I. The Pt–Ni alloy CEs are expected to significantly enhance the electron conduction, charge-transfer ability, and therefore power conversion efficiency of the DSSCs. 2. Experimental 2.1. Preparation of binary Pt–Ni alloy CEs The feasibility of this strategy was confirmed by a cyclic voltammetry (CV) method: A cleaned FTO glass substrate was used as a working electrode, and the CV curves were recorded from 0.4 to +0.3 V and back to 0.4 V for 10 cycles. Before the measurement, the supporting electrolyte consisting of 2 mM H2PtC16, 1 mM of Ni(NO3)2, and 5 mM of HCl aqueous solution was degassed using nitrogen for 10 min.
2.3. Electrochemical characterizations The electrochemical performances were recorded on a conventional CHI660E setup comprising an Ag/AgCl reference electrode, a CE of platinum sheet, and a working electrode of FTO glass supported Pt–Ni alloy. The CV curves were recorded from 1.0 to +1.4 V and back to 1.0 V. Before the measurement, the supporting electrolyte consisting of 50 mM M LiI, 10 mM I2, and 500 mM LiClO4 in acetonitrile was degassed using nitrogen for 10 min. Electrochemical impedance spectroscopy (EIS) measurements were also carried out on the CHI660E Electrochemical Workstation in a frequency range of 0.01 Hz 106 kHz and an ac amplitude of 5 mV at room temperature. The resultant impedance spectra were analyzed using the Zview software. Tafel polarization curves were recorded on the same Workstation by assembling symmetric cell consisting of Pt–Ni alloy CE|redox electrolyte|Ni–Pt alloy CE.
2.2. Assembly of DSSCs A layer of TiO2 nanocrystal anode film with a thickness of 10 mm was prepared by a sol-hydrothermal method [21] and a layer of TiO2 nanocrystal anode film with a thickness of 10 mm and an active area of 0.25 cm2 was prepared by coating TiO2 colloid onto conducting glass using a doctor blade technique, followed by sintering in air at 450 C for 30 min. Resultant anodes were further sensitized by immersing into a 0.50 mM ethanol solution of N719 dye ([cisdi(thiocyanato) N,N’ bis(2,2’ bipyridyl 4carboxylic acid)4tetrabutylammonium carboxylate]). The DSSC was fabricated by sandwiching redox electrolyte between dyesensitized TiO2 anode and FTO supported Pt–Ni alloy CEs. A redox electrolyte consisted of 100 mM of tetraethylammonium iodide, 100 mM of tetramethylammonium iodide, 100 mM of tetrabutylammonium iodide, 100 mM of NaI, 100 mM of KI, 100 mM of LiI, 50 mM of I2, and 500 mM of 4tertbutylpyridine in 50 ml acetonitrile.
2.4. Photovoltaic measurements The photocurrent–voltage (J–V) curves of the assembled DSSCs were recorded on an Electrochemical Workstation (CHI600E) under irradiation of a simulated solar light from a 100 W xenonmercury arc lamp (CHFXM500 W, Beijing Trusttech Co., Ltd) in an ambient atmosphere. The incident light intensity was calibrated using a FZA type radiometer from Beijing Normal University Photoelectric Instrument Factory to control it at 100 mW cm2 (AM1.5 calibrated by a standard silicon solar cell). Each DSSC device was measured at least five times to eliminate experimental error and a compromise J–V curve was employed. 2.5. Other characterizations The morphologies of the Pt–Ni alloy CE were observed with a scanning electron microscope (SEM, S4800). The XRD data were
Fig. 1. Topview SEM photographs of (a) PtNi0.25, (b) PtNi0.50, (c) PtNi0.75, and (d) PtNi alloy CEs.
B. He et al. / Electrochimica Acta 158 (2015) 397–402
collected in a scan mode with a scanning speed of 10 min1 in the 2u range between 20 and 80 . XPS experiment was carried out on a RBD upgraded PHI5000C ESCA system (Perkin Elmer) with Mg Ka radiation (hn = 1253.6 eV).
characterization is carried out to determine the chemical structure and composition of PtNi0.75 alloy. As shown in Fig. 2b and Fig. 2c, the peaks for Pt4f and Ni2p are centered at 71.28 eV (Pt4f7/2), 71.45 eV (Pt4f5/2), and 852.5 eV (Ni2p3/2), confirming the metallic nature of assynthesized PtNi0.75 alloy [24]. It has been mentioned that the alloying of transition metals can favor the electronic perturbation of other metals [25] and therefore accelerate the electrocatalytic activity of the alloys for their electrocatalyst applications. The binary alloy CEs synthesized by a CV method are subjected to XRD measurements. As shown in Fig. 2d, XRD results indicate that the alloy CEs consist of alloy materials at 2u = 39.96 and FTO glass substrate marked with (*). Moreover, unalloyed Pt is also detected because of the appearance of diffraction peak at 2u = 46.24 . A golden rule in determining an efficient CE is to evaluate its electrocatalytic activity toward electrolyte [26]. As displayed in Fig. 3a, the CV peak shapes and positions of alloy CEs are similar to that of pristine Pt electrode, suggesting that the alloys have a similar electrocatalytic function to Pt CE. However, the peak current densities of CV curves have been elevated by alloying Pt with Ni in comparison with pristine Pt electrode. Considering the task of a CE is to collect electrons from external circuit and to reduce I3 to I, therefore, Red1 peak can be employed to assess the catalytic activity of alloy CEs. The ratio of Jox1/|Jred1| is a parameter to elevate the reversibility of the redox reaction toward I/I3 species [27]. The obtained values for pristine Pt, PtNi0.25, PtNi0.50, PtNi0.75, and PtNi are 1.155, 0.985, 0.990, 1.007, and 0.966,
3. Results and discussion SEM photographs in Fig. 1 suggest a high surface coverage and loading of alloys on FTO glass. Deep examination gives homogeneously spherical aggregations. However, the compact structures in PtNi0.25 and PtNi0.50 alloy CEs can hinder the diffusion of I/I3 redox couples at CE/electrolyte interface. By optimizing stoichiometry of Pt and Ni sources, a loose structure with an average aggregation size of 200 nm is obtained for PtNi0.75 alloy CE, providing channels for I/I3 species across the alloy layer [22]. At a high Ni dosage, such as PtNi alloy, the cracks and large interfacial resistance may limit electron migration from FTO back to alloy layer. Energydispersive Xray spectra (EDS), as shown in Fig. 2a, demonstrate that Pt and Ni elements as well as partial elements from FTO glass are detected, indicating a successful deposition of Pt and Ni on FTO substrate. The measured atomic ratios by EDS for PtNi0.25, PtNi0.50, PtNi0.75, and PtNi alloys are 1.000:0.266, 1.000:0.518; 1.000:0.752, and 1.000:1.083, respectively. The measured atomic ratios are relatively close to their stoichiometries, therefore the chemical formulae of the alloy CEs can be expressed according to the stoichiometric ratios [23]. XPS
C O Ni
Pt
b
PtNi0.25
Sn
71.45 eV
}
PtNi0.75 PtNi
cps 0
2
4
80
6
75
Energy (keV)
c
Pt4f
71.28 eV
PtNi0.50
Counts (a.u.)
a
399
70
65
Binding energy (eV)
852.5 eV
d
Ni2p
PtNi0.25
840
850
860
870
880
20
* *
40
* Pt (200)
Counts (a.u.)
Intensity (a.u.)
*
Pt-Ni alloy
PtNi0.50
2theta (degree)
PtNi0.75 PtNi
* *
60
Binding energy (eV) Fig. 2. (a) EDS of Pt–Ni alloy CEs, XPS spectra of (b) Pt4f and (c) Ni2p in PtNi0.75 alloy, and (c) XRD patterns of FTO glass supported alloy CEs.
80
400
B. He et al. / Electrochimica Acta 158 (2015) 397–402
Fig. 3. (a) CV curves of various CEs for I/I3 redox species recorded at a scan rate of 50 mV s1. (b) CV curves of PtNi0.75 alloy CE for I/I3 redox species at different scan rates. The insert gives the plots of peak current density versus scan rate. (c) A total of 100 consecutive CVs for the I/I3 system using the PtNi0.75 alloy CE at a scan rate of 50 mV s1. (d) Relationship between the cycle time and the redox peak current for the PtNi0.75 alloy CE.
respectively. The high peak current density, low peaktopeak separation (Epp, 0.528 V for PtNi0.75 and 0.632 V for Pt) [28], and good reversibility demonstrate the alloying of Pt with Ni can markedly enhance the reversibility of I3 $ I reaction, and PtNi0.75 alloy has superiority of catalyzing liquid electrolyte containing I/I3 redox couples. From the stacking CV curves of PtNi0.75 alloy CE at different scan rates, one can find an outward extension of redox peaks (Fig. 3b). By plotting peak current density corresponding to I3 $ I versus square root of scan rate, linear relationships are observed, indicating a diffusion controlled mechanism of redox reaction on alloy CE [29]. A total of 100 consecutive CV curves for the I/I3 system using the PtNi0.75 alloy CE at a scan rate of 50 mV s1 are shown in Fig. 3c, no obvious attenuation of peak current densities in Ox1/Red1 means the alloy CE is relatively stable for reducing I3 ions (Fig. 3d). Nyquist EIS plots of the symmetric dummy cells by two identical CEs are employed to demonstrate the intrinsic interfacial chargetransfer kinetics at CE/electrolyte interface. As presented in Fig. 4a, the Rct value of the PtNi0.75 alloy CE is 3.03 V cm2, smaller than 5.75 V cm2 for pristine Pt electrode (Table 1). The Rct of CEs follows an order of PtNi0.75 < PtNi0.50 < PtNi0.25 < Pt < PtNi. The sequence is consistent with the results of CV curves. Enhanced electrocatalytic kinetics and chargetransfer ability could accelerate dye regeneration and therefore elevate the Jsc of cells. Due to the participation of electrons in I3 reduction reaction, therefore
the electron lifetime (t ) at CE/electrolyte interface can be calculated from t = 1/(2pfp) and utilized to assess the reduction reaction kinetics [30], where fmax is the maximum frequency of the midfrequency peak in the Bode phase plots (Fig. 4b). The t at CE/electrolyte interface has an order of PtNi0.75 (42 ms) < PtNi0.50 (50 ms) < PtNi0.25 (61 ms) < Pt (74 ms) < PtNi (90 ms). A low t value refers to a rapid reduction reaction occurred at CE/electrolyte interface. Moreover, the exchange current density (J0, the slope for anodic or cathodic branches) and limiting diffusion current density (Jlim, the intersection of cathodic branch with Y-axis), in the Tafel polarization curves can also be uitilized to reconfirm the catalytic activity of CEs. As shown in Fig. 4c, both J0 and Jlim have an order of PtNi0.75 > PtNi0.50 > PtNi0.25 > Pt > PtNi. From equations J0 = RT/nFRct [31],Jlim = 2nFCDn/l [32], and Jred1 = Kn1.5ACDn0.5v0.5 [33] (R is the gas constant, F is Faraday’s constant, n is number of electrons participating in I3 reduction reaction, l is the distance between electrodes in a symmetric dummy cell, C is concentration of I3 ions), one can conclude that the results from Tafel polarization curves, EIS plots, and CV curves are in good agreement. Fig. 4d compares characteristic J–V curves of the DSSCs with pristine Pt and alloy CEs. The solar cell from PtNi0.75 electrode yields a maximum h of 8.59% (Voc = 0.716 V, Jsc = 17.50 mA cm2, FF = 68.6%), which are much higher than the photovoltaic parameters with pristine Pt CE and those in literatures [34,35], The solar cell from PtNi0.75 electrode. This may be attributed to a fact that the rapid
B. He et al. / Electrochimica Acta 158 (2015) 397–402
401
Fig. 4. (a) Nyquist and (b) Bode EIS plots and (c) Tafel polarization curves for symmetric dummy cells fabricated two identical CEs. (d) Characteristic J–V curves for the DSSCs with various CEs. The insert gives an equivalent circuit for the symmetric dummy cells.
interconvertion between I3 and I can accelerate electron generation from N719 dye and therefore elevate the electron flow from excited dye to conduction band of TiO2, therefore the electron density on TiO2 conduction band is markedly enhanced. The recorded h from PtNi0.75 alloy is impressive for the DSSCs with lowPt CEs. Critical requirements of the engines and vehicles drived by solar panels are high power conversion efficiency, costeffectiveness, fast startup, multiple start capability, and operational stability. The efficiency and expenses greatly depend on the design of CEs, in which Ptfree or lowPt materials are always required. However, a robust catalyst in CE is desirable for fast startup, multiple start, and stability [36]. As shown in Fig. 5a, the photocurrent of the cell device employing PtNi0.75 alloy electrode has an abrupt increase at full irradiation. No delay in time is detected in starting the cell, this
observation supports a fact that the alloy CEs are robust in catalyzing I/I3 redox couples. After five start–stop cycles, there is no failure in photocurrent density, however, the current density from Pt based solar cell has an apparent decrease in the first two cycles. These results suggest that PtNi0.75 alloy is a preferred CE material for efficient DSSCs. Additionally, the relationship between photocurrent density and time demonstrates cell stability on prolonged exposure to light irradiation (100 mW cm2). As displayed in Fig. 5b, the photocurrent densities for DSSCs with PtNi0.75 and Pt decrease by 6.75% and 14.69% over 5 h, respectively. In comparison with the solar cell with Pt electrode [37], the DSSC employing PtNi0.75 alloy electrode displays enhanced stability. In order to explore potential mechanism for enhanced catalytic activity and stability, the Gibbs energy of the possible dissolving reaction of PtNi0.75 alloy is calculated [38]. The values of Gibbs free
Table 1 Comparison of photovoltaic and electrochemical parameters as well as electron lifetime and transparency for the DSSCs. h: power conversion efficiency; Voc: opencircuit voltage; Jsc: shortcircuit current density; FF: fill factor; Rct: chargetransfer resistance; W: Nernst diffusion resistance corresponding to I/I3 couples; t : electron lifetime at CE/electrolyte interface. CEs
h (%)
Voc (V)
Jsc (mAcm2)
FF (%)
Rct (V cm2)
W (V cm2)
t (ms)
Pristine Pt PtNi0.25 PtNi0.50 PtNi0.75 PtNi
6.98 7.70 8.05 8.59 6.46
0.698 0.703 0.717 0.716 0.703
14.63 16.07 16.45 17.50 13.75
68.4 68.2 68.3 68.6 66.8
13.00 9.18 5.42 3.03 13.26
5.75 3.88 2.92 1.85 8.47
74 61 50 42 90
402
B. He et al. / Electrochimica Acta 158 (2015) 397–402
b 20
a
20
PtNi0.75 -2
Current density mA cm
-2
Current density (mA cm )
Pt
15
10
5
0
15
PtNi0.75
10
Pt
5
0 0
50
100 150 Time (s)
200
250
0
1
2
3
4
5
Time (h)
Fig. 5. (a) Startstop switches and (b) photocurrent stabilities of the DSSCs employing PtNi0.75 alloy and Pt electrodes. The on–off plots were achieved by alternatively irradiating (100 mW cm2) and darkening (0 mW cm2) the DSSC devices at an interval of 25 s and 0 V, whereas the photocurrent stabilities were carried out under sustained irradiation of 100 mW cm2.
energy (DrGm) were calculated for the possible reactions between Ni or Pt and triiodide species [Reaction 1: Pt (s) + 2I3 (aq) = PtI4 (s) + 2I (aq); Reaction 2: Ni (s) + I3 (aq) = NiI2 (s) + I (aq)]. The DrGm values for reaction 1 and reaction 2 are 45.9 and 76.4 kJ mol1, respectively. In this fashion, the reaction between Ni species and triiodides is easier. The alloying of Ni with Pt can form a competed reaction between Ni and Pt, therefore protecting the high catalytic activity of Pt species by sacrificing Ni species. The presented results demonstrate the superiority of PtNi0.75 alloy CE than pristine Pt electrode for DSSC application. 4. Conclusions In summary, binary Pt–Ni alloys have been fabricated by an electrochemical codeposition strategy free of any surfactant or template and employed as CE materials in DSSCs. It is demonstrated that PtNi0.75 alloy CE has an optimal chargetransfer ability and electrocatalytic activity toward I3 reduction. The DSSC employing PtNi0.75 alloy CE exhibits impressive power conversion efficiency of 8.59% in comparison with 6.98% from pristine Pt CE based device. The research presented here is far from being optimized but these profound advantages along with costeffectiveness, mild synthesis and scalable materials promise the binary Pt–Ni alloy CEs to be strong candidates in robust DSSCs. Acknowledgements The authors would like to acknowledge financial supports from Fundamental Research Funds for the Central Universities (201313001, 201312005), Shandong Province Outstanding Youth Scientist Foundation Plan (BS2013CL015), Shandong Provincial Natural Science Foundation (ZR2011BQ017), Research Project for the Application Foundation in Qingdao (13-4-198-jch), National Natural Science Foundation of China (51102219, 51342008, U1037604), National Key and National Key Technology Support Program (2012BAB15B02). References [1] J.M. Tour, C. Kittrell, V.L. Colvin, Nat. Mater. 9 (2010) 871–874. [2] N. Memarian, I. Concina, A. Braga, S.M. Rozati, A. Vomiero, G. Sberveglieri, Angew. Chem. Int. Ed. 50 (2011) 12321–12325. [3] A. Yella, H.W. Lee, H.N. Tsao, C. Yi, A.K. Chandiran, M.K. Nazeeruddin, E.W.G. Diau, C.Y. Yeh, S.M. Zakeeruddin, M. Grätzel, Science 334 (2011) 629–634.
[4] J.H. Wu, Y.M. Xiao, Q.W. Tang, G.T. Yue, J.M. Lin, M.L. Huang, Y.F. Huang, L.Q. Fan, Z. Lan, S. Yin, T. Sato, Adv. Mater. 24 (2012) 1884–1888. [5] X.X. Chen, Q.W. Tang, B.L. He, L. Lin, L.M. Yu, Angew. Chem. Int. Ed. 53 (2014) 10799–10803. [6] B. O’Regan, M. Grätzel, Nature 353 (1991) 737–740. [7] M.X. Wu, T.L. Ma, J. Phys. Chem. C 118 (2014) 16727–16742. [8] J. Liu, Q.W. Tang, B.L. He, J. Power Sources 268 (2014) 56–62. [9] J.D. Roy-Mayhew, D.J. Bozym, C. Punckt, I.A. Aksay, Nano 4 (2010) 6203–6211. [10] B. Tang, G.X. Hu, H.Y. Gao, Z.X. Shi, J. Power Sources 234 (2013) 60–68. [11] Q.W. Tang, H.Y. Cai, S.S. Yuan, X. Wang, J. Mater. Chem. A 1 (2013) 317–323. [12] J.H. Wu, Y. Li, Q.W. Tang, G.T. Yue, J.M. Lin, M.L. Huang, L.J. Meng, Sci. Rep. 4 (2014) 4028. [13] B.L. He, Q.W. Tang, T.L. Liang, Q.H. Li, J. Mater. Chem. A 2 (2014) 3119–3126. [14] B.L. He, Q.W. Tang, M. Wang, H.Y. Chen, S.S. Yuan, ACS Appl. Mater. Interfaces 6 (2014) 8230–8236. [15] J. Ma, L. Zhou, C. Li, J. Yang, T. Meng, H. Zhou, M. Yang, F. Yu, J. Chen, J. Power Sources 247 (2014) 999–1004. [16] F. Gong, H. Wang, X. Xu, G. Zhou, Z.S. Wang, J. Am. Chem. Soc. 134 (2012) 10953–10958. [17] X. Chen, Y. Jiang, J. Sun, C. Jin, Z. Zhang, J. Power Sources 267 (2014) 212–218. [18] Y. Bing, H. Liu, L. Zhang, D. Ghosh, J. Zhang, Chem. Soc. Rev. 39 (2010) 2184–2202. [19] B.L. He, X. Meng, Q.W. Tang, ACS Appl. Mater. Interfaces 6 (2014) 4812–4818. [20] H.Y. Cai, Q.W. Tang, B.L. He, P.J. Li, J. Power Sources 258 (2014) 117–121. [21] B.B. Hu, Q.W. Tang, B.L. He, L. Lin, H.Y. Chen, J. Power Sources 267 (2014) 445–451. [22] S. Cho, S.H. Hwang, C. Kim, J. Jang, J. Mater. Chem. 22 (2012) 12164–12171. [23] Y.Y. Duan, Q.W. Tang, B.L. He, R. Li, L.M. Yu, Nanoscale 6 (2014) 12601–12608. [24] K. Park, J. Choi, B. Kwon, S. Lee, Y. Sung, H. Ha, S. Hong, H. Kim, A. Wieckowski, J. Phys. Chem. B 106 (2002) 1869–1877. [25] C. Xu, J. Hou, X. Pang, X. Li, M. Zhu, B. Tang, Int. J. Hydrogen Energy 37 (2012) 10489–10498. [26] S. Thomas, T.G. Deepak, G.S. Anjusree, T.A. Arun, S.V. Nair, A.S. Nair, J. Mater. Chem. A 2 (2014) 4474–4490. [27] H. Tributsch, Coordin. Chem. Rev. 248 (2004) 1511–1530. [28] B.L. He, Q.W. Tang, J.H. Luo, Q.H. Li, X.X. Chen, H.Y. Cai, J. Power Sources 256 (2014) 170–177. [29] Z.B. Lv, J.F. Yu, H.W. Wu, J. Shang, D. Wang, S.C. Hou, Y.P. Fu, K. Wu, D.C. Zou, Nanoscale 4 (2012) 1248–1253. [30] J. Yoon, M. Jin, M. Lee, Adv. Mater. 23 (2011) 3974–3978. [31] M. Wu, X. Lin, Y. Wang, L. Wang, W. Guo, D. Qi, X. Peng, A. Hagfeldt, M. Grätzel, T.L. Ma, J. Am. Chem. Soc. 134 (2014) 3419–3428. [32] M.K. Wang, A.M. Anghel, B. Marsan, N.C. Ha, N. Pootrakulchote, S.M. Zakeeruddin, M. Grätzel, J. Am. Chem. Soc. 131 (2009) 15976–15977. [33] T. Daeneke, A.J. Mozer, T.H. Kwon, N.W. Duffy, A.B. Holmes, U. Bach, L. Spiccia, Energy Environ. Sci. 5 (2012) 7090–7099. [34] Y. Gong, C. Li, X. Huang, Y. Luo, D. Li, Q. Meng, B.B. Iversen, ACS Appl. Mater. Interfaces 5 (2013) 795–800. [35] V.-D. Dao, S.-H. Kim, H.-S. Choi, J.-H. Kim, H.-O. Park, J.-K. Lee, J. Phys. Chem. C 115 (2011) 25529–25534. [36] D. Papadias, S.H.D. Lee, Ind. Eng. Chem. Res. 45 (2006) 5841–5858. [37] G. Syrrokostas, A. Siokou, G. Leftheriotis, P. Yianoulis, Sol. Energy Mater. Sol. Cells 103 (2012) 119–127. [38] E. Olsen, G. Hagen, S.E. Lindquist, Sol. Energy Mater. Sol. Cells 63 (2000) 267–273.