V2O5 as an inexpensive counter electrode for dye sensitized solar cells

In pursuit of an abundant, inexpensive and stable counter electrode as an alternative to platinum for dye-sensitized solar cells (DSSCs), we report a new, low-cost substitute material. Here for the first time, we demonstrate that V2O5 can be used as a counter electrode material in DSSCs. We note that the efficiency of DSSCs with commercial V2O5 and hydrothermal treated V2O5 are upto 1.2% and 1.6%, respectively. The results indicate that, with optimization, V2O5 can be a promising choice to replace platinum from a cost perspective. The innovation of new economical counter electrodes offers a potential way to cut down the industrial costs which is crucial for large-scale production and commercial applications of DSSCs.


Introduction
Global warming is a major concern of today's world. Hence, there is an urgent need for environmentally sustainable energy technologies. Solar energy is the source of an enormous quantity of clean, renewable energy; it has attracted much attention as photovoltaic (PV) cells can directly convert solar energy into electrical energy with low carbon footprint. Despite being an abundant source of energy and promising industry, only approximately 0.015% of global electricity is generated from solar energy, however this could be massively extended by further reductions in cost and development of new markets for PV. In this context, dye-sensitized solar cells (DSSCs) have emerged as an interesting low-cost photovoltaic technology [1]. Furthermore, the other key attributes of DSSC technology are its simple manufacturing procedures, environmental friendliness, lightweight, flexibility, semi transparency and good performance in diverse light conditions [2].
DSSCs have stimulated great research interest since the first report of high-efficiency from O'Regan and Grätzel [1]. Although laboratory device efficiencies have reached 12% [3], to make this available for mass production, the cost of manufacture has to be reduced. DSSCs mainly consist of a photoanode, made of a few micron thick semiconductor such as TiO 2 on a conductive substrate such as fluorine doped tin oxide coated glass (FTO), an organic sensitizer (dye, e.g. N719) or inorganic sensitizer (quantum dots, e.g. CdS, PbS), an electrolyte (typically I 3 − /I − or Co 2+ /Co 3+ redox couples) and a counter electrode (CE) made of platinum (Pt) or a carbon material on a conductive substrate. Each component of the device decides the efficiency and cost of the resulting solar cell. In order to achieve economic viability, one of the challenges in DSSC technology is to find a suitable replacement for the precious metal Pt. Pt is a preferred CE because of its excellent electrocatalytic activity and high electrical conductivity [4]. However, there are some reports which claim that Pt reacts with I 3 -in the electrolyte to form PtI 4 [5] and in addition the catalytic activity decreases upon the exposure to dye solution and in the presence of I 2 based electrolyte [6]. Additionally, Pt is a rare and expensive metal. To reduce the fabrication cost, substantial research efforts have been placed on the exploration of alternative and cheap counter electrodes. To date carbon materials [7][8][9][10], conductive polymers [11][12][13], inorganic compounds such as sulfides [14][15][16], carbides [17], nitrides [18,19], phosphides [20], tellurides [21], and composites [22][23][24] have been explored as effective alternative counter electrodes. Though some of these materials have reached the efficiencies comparable to or slightly inferior to that of existing Pt, there still remain concerns of ready availability, eco-friendliness, stability, economical viability and large scale production cost. Therefore, the fact remains that very few alternative materials are capable of replacing the expensive Pt in fabricating economical and highly efficient DSSCs. Despite having a number of strengths such as durability, environmental friendliness, stability, high melting temperatures, and high electrical and thermal conductivities, oxides [25,26] are less studied compared to other inorganic materials. Recently tungsten oxide [26,27] has been reported as an efficient low cost CE. Vanadium based oxides belongs to transitional metal oxides which are of particular interest because of their low cost, abundance in nature, good electrical conductivities, good catalytic activity and excellent atmospheric stability. It is known that V 2 O 5 is already considered as a promising material in technologies as cathodes in rechargeable ion batteries [28], optical electrical switches [29], spintronic devices [30], gas detectors [31], capacitors [32], and electro-chromic devices [33]. In this present work, we choose V 2 O 5 as a material of interest for counter electrodes in DSSCs, because of its distinctive features: inexpensive, readily available, has excellent physical and chemical properties, and reasonable electrical conductivity. There are a few reports on the implementation of V 2 O 5 in DSSC technology such as an electron blocking layer [34] or a compact layer [35], and a composite of V 2 O 5 in solid state DSSCs [36,37]. To the best of our knowledge, this is the first time ever that V 2 O 5 to be used as a CE in DSSCs. In this report, we show that the photovoltaic performance of hydrothermally treated V 2 O 5 (H-V 2 O 5 ) shows better photoconversion efficiency compared to commercial V 2 O 5 (C-V 2 O 5 ) and has the potential to compete with expensive Pt counter electrodes. It is worthwhile to mention here that, the adhesion issues associated with C-V 2 O 5 during film process prompted us to treat this material hydrothermally to alter its particle sizes and improve the film process. So proper care has been taken in terms of hydrothermal synthesis conditions (temperature and duration) to avoid formation of any kind of hydrate or mixed valent vanadium oxide phases.

Materials synthesis
Sigma Aldrich analytical grade V 2 O 5 powder was used as-received without further purification. An aqueous solution of vanadium oxide was prepared by dissolving C-V 2 O 5 in double distilled water under magnetic stirring. After complete dissolution, the resulting solution was transferred to a teflon-lined stainless steel autoclave of 30 ml volume for thermal treatment at 180°C. Once the reaction time was complete, the autoclaves were cooled down to ambient temperature naturally; the resultant yellow powders were collected and rinsed with distilled water and then with ethanol several times; the final products (H-V 2 O 5 ) were finally dried in an oven at 50°C for 6 h and used for further characterizations.

Preparation of TiO 2 working electrodes (WEs)
FTO glass substrates (Pilkington TEC glass™, sheet resistance: 11.7 Ω sq −1 , 2.2 mm in thickness, TEC-12) were cut into small pieces (2×2 cm). The substrates were first cleaned thoroughly with soapy water, and then ultrasonicated in deionised water followed by acetone and iso-propanol for 15 min at each step, and then dried in flowing nitrogen to further remove any solvent traces. After drying, the substrates were treated with 40 mM TiCl 4 for 1 h at 80°C to prepare a compact layer of TiO 2 . Commercially available TiO 2 pastes (DSL 18NR-T, Dyesol as the transparent layer and DSL 18NR-AO, Dyesol as the scattering layer) were screen printed on the compact layered FTO substrates whose active areas were approximately 1 cm 2 . The TiO 2 coated films were then successively heated in ambient air on a hotplate at 325°C for 10 min, at 375°C for 5 min, at 450°C for 15 min and at 500°C for 15 min, before being cooled down slowly to room temperature. Afterwards these electrodes were further treated with 40 mM TiCl 4 for 30 min at 80°C followed by sintering at 450°C for 30 min. As soon as the WE reach 80°C, these films were subsequently immersed in N719 (Dyesol) solution and kept in an air-tight glass container at room temperature under dark and dry conditions for a day. Dye sensitized photoanodes were washed with iso-propanol to remove the excess dye which was firmly attached to the surfaces of TiO 2 , followed by drying with nitrogen flow at room temperature.

Preparation of counter electrodes
FTO coated glass was cut into the same dimensions as the WE and a small hole was drilled to facilitate the injection of electrolyte, using a sand blasting unit, with a smooth finish given by ending the drilling with a portable glass hole drilling tool. The drilled FTO substrates were subjected to the same cleaning and drying procedures as was the case for the photoanode substrates. C-V 2 O 5 and H-V 2 O 5 pastes were prepared by mixing 4:1 ratio of V 2 O 5 powders with ethyl cellulose and few drops of α-terpineol to make an appropriate viscous paste to facilitate the screen printing process. Finally the V 2 O 5 films were prepared by screen printing on the asprepared FTO substrates followed by sintering in air at 450°C for 30 min at a heating rate of 5°C min −1 to remove the binders. Then the V 2 O 5 films were allowed to cool down to room temperature. For comparison, a Pt CE was prepared using a sputtering system on FTO substrates. Three kinds of DSSCs were therefore fabricated, i.e. using Pt, C-V 2 O 5 and H-V 2 O 5 as counter electrodes, and their performances were compared.

Fabrication of photovoltaic cells
The dye sensitized TiO 2 photoanodes and counter electrodes (sputtered Pt, V 2 O 5 screen printed) were assembled into a sandwich configuration with a 25 μm hot-melt spacer (SX1170-60, Solaronix) through the hotpress method by heating at 120°C for few seconds. After assembling the cells, the electrolyte solution (composed of 0.6 M 1-butyl-3-methylimidazolium iodide, 0.05 M iodine, 0.1 M lithium iodine and 0.5 M 4-tertbutylpyridine in acetonitrile) was then injected into the space between the WE and CE via a hole in the CE by capillary effect under vacuum. The hole was then sealed with cover-glass using surlyn. The effective cell area was 1 cm 2 . A lead based alloy was ultrasonically soldered for contacts at the edge of the electrodes to make a good current collector. Five cells were prepared for each material.

Characterization of counter electrodes and DSSCs
The phase purity and crystallographic structure of commercial and as-synthesized V 2 O 5 was determined using automated PANalytical X' Pert Diffractometer, Netherlands (using Ni filtered Cu K α radiation, λ=1.5406 Å). The diffraction patterns were recorded in the range of 10°-80°. Lattice parameters were extracted from XRD patterns at room temperature using FullProf program suite [38]. The surface morphology of C-V 2 O 5 and H-V 2 O 5 powder samples and also the counter electrodes were investigated by field emission scanning electron microscopy (SEM) (Quanta 650 FEG ESEM). The optical absorption of commercial and hydrothermal treated powder samples were measured by diffuse reflectance at room temperature using a Perkin Elmer Lambda 950 UV-vis-NIR spectrophotometer. Cyclic voltammetry measurements were performed at room temperature using a μAUTOLAB Type III Potentiostat set-up which was driven by the electrochemical software GPES. The photocurrent density and voltage characteristics of DSSCs were measured by computer-controlled digital source meter (Keithley 2400) under AM 1.5 one-sun illumination using a solar simulator (92 250 A, Newport, USA). All the fabricated cells were tested at a light intensity of 100 W m −2 with an active cell area of 0.25 cm 2 , which was defined by a mask. An electrochemical work station system (Autolab potentiostat equipped with FRA2 module) was employed to measure the internal resistances of the cells. The impedance spectra were collected under dark conditions by varying the frequency from 0.1 Hz to 1 MHz at an AC amplitude of 10 mV.

Results and discussion
3.1. Structure, morphology and optical properties of V 2 O 5 3.1.1. X-ray diffraction (XRD) The crystallographic structures of C-V 2 O 5 and H-V 2 O 5 were investigated by XRD. Figure 1 shows the XRD patterns of C-V 2 O 5 and V 2 O 5 after hydrothermal reaction at 180°C. All the observed XRD reflections are well indexed to standard JCPDS pattern: #89-0612. Although the overall XRD pattern of the sample that underwent hydrothermal reaction exhibits no obvious difference when compared with that of the C-V 2 O 5 , the relative intensities of XRD reflections are diminished by approximately three times. This indicates a decrease of the crystallinity and particle size of the V 2 O 5 sample after hydrothermal reaction. Furthermore, no other peaks of impurities were observed, indicating that phase was retained under hydrothermal conditions. Using LeBail profile matching analysis, XRD data are indexed with orthorhombic crystal system with space group P m m n and lattice parameters are calculated as a=11.5114(5) Å, b=3.5634(2) Å, c=4.3717(2) Å and unit cell volume V=179.33(2) Å 3 . The slight changes in the lattice parameters compared to standard JCPDS data could be due to changes in the crystallite size and corroborated with the subsequent morphological studies. The comparatively lower intensity of the H-V 2 O 5 sample in XRD results is ascribed to a reduction of particle size. The crystal domain size of the C-V 2 O 5 and H-V 2 O 5 particles estimated from the Debye-Scherrer formula are 760 nm and 340 nm respectively and this further suggests changes in the crystal size after hydrothermal treatment, which is additionally confirmed in the following SEM micrographs. Figure 2 shows the surface morphologies of C-V 2 O 5 , H-V 2 O 5 powders and counter electrodes made of these materials studied using FESEM. For direct comparison of commercial, hydrothermal treated powder samples, and their corresponding counter electrodes, the SEM micrographs were taken with the same conditions and magnification. Figure 2(a) presents the surface morphology of C-V 2 O 5 consisted of agglomerated particles that were fused to each other with a non-uniform size which falls in the range of 1-50 μm. The hydrothermal treatment to the commercial bulk V 2 O 5 facilitated to reduce the agglomeration of particles and transformed into a rod like structures as shown in figure 2(b) whose size ranging from a few nm-1 μm. The C-V 2 O 5 counter electrodes shown in figure 2(c) presence of some visible large clusters of several micrometres. Figure 2(d) displays the surface morphology of H-V 2 O 5 counter electrodes which is more homogeneous and consists of micro-rod like structures of diameter ∼100 nm. This is possibly due to less adhesion taking place between the C-V 2 O 5 powder and the FTO substrate, which thus produces more granular features compared to that of H-V 2 O 5 films.  From the SEM analysis, we can conclude that the H-V 2 O 5 film appears to be smoother than that of the commercial one and its effect expected to reflect in the photovoltaic properties.

Optical absorption spectra
UV-vis-NIR spectroscopy is a useful technique to measure the optical band gap. The UV-vis absorption measurements were carried at room temperature in diffuse reflectance mode in the wavelength range 350-800 nm. The reflectance of the films prepared on glass substrates were measured by subtracting the reflectance of glass substrate as a reference. Figure 3(a) shows the reflectance spectra of commercial and as prepared V 2 O 5 powders. It is observed that both the specimens show a strong absorption around 495-570 nm. The absorption coefficient of both the specimens is calculated from their reflectance spectra by means of following relation.
where the reflectance falls between R max and R min due to absorption by the material and R is the reflectance of any intermediate energy photons. The optical band gap energy of a material can be determined by using the Tauc relation [39], shown below where hn is the incident photon energy, A is a constant and E g is the optical band gap of the material and n depends on the type of transition. n=1/2, 2, 3 and 3/2 correspond to allowed direct, allowed indirect, forbidden direct and forbidden indirect transitions, respectively. The direct optical band gap of the material can be determined by plotting h 2 ( ) a n versus hn as shown in figure 3(b), and then extrapolating the linear portion of the curve to the energy axis for zero absorption coefficients. The direct band gap of H-V 2 O 5 is found to be around 2.19 eV, which is higher than that of C-V 2 O 5 value 2.13 eV. The slight increase in band gap of hydrothermal treated powder compared to the commercial one can be ascribed to the decrease of crystallite size, as confirmed from XRD and SEM measurements.

Cyclic voltammetry studies
One of the key parameters in a DSSC is the electrocatalytic activity of the CE and it is well known that cyclic voltammetry (CV) measurements are an essential electrochemical characterization tool to elucidate the interfacial charge transfer properties of the liquid electrolyte on the electrode surfaces. The CV scans were recorded at a scan rate of 20 mV s −1 in a three electrode system in a freshly prepared acetonitrile solution containing 0.1 M lithium perchlorate, 10 mM lithium iodine and 1 mM iodine, which was purged with nitrogen prior to each measurement. The three electrode system comprised of Ag/AgCl as the reference electrode (RE), Pt wire as the (CE) and Pt-coated or V 2 O 5 screen printed FTO glass as a WE. Peak to peak separation (EPP) is an important parameter, which qualifies the catalytic activity of CEs. Figure 4 shows the CV curves of the triiodide/iodide redox couple for Pt, C-V 2 O 5 , and H-V 2 O 5 electrodes, respectively. Two distinct pairs of redox peaks were obtained in cyclic voltammograms for the Pt CE, indicating its high catalytic activity. The lower potential redox pair can be attributed to reaction 1 (below) whereas the higher potential pair can be represented by reaction 2: From the EPPs of the Pt CE and the V 2 O 5 counter electrodes, it is clear that the sputtered Pt is catalytically more active towards the reduction of I 3to I − however, the commercial and hydrothermal treated V 2 O 5 counter electrodes also show reasonable catalytic activity. It is to be noted that our visual observations by naked eye did not see any peeling off V 2 O 5 from FTO substrate after cyclic voltammetry measurements which show the adhesion of V 2 O 5 was reasonable.

Photovoltaic characteristics
The photovoltaic performance of assembled DSSCs based on C-V 2 O 5, H-V 2 O 5 and Pt counter electrodes are described under illumination conditions of AM 1.5 G (100 mW cm −2 ) with an active cell area of 0.25 cm 2 . The photoconversion efficiency (η) and fill factor (FF) of the DSSCs are obtained from J-V curves, which were calculated according to the relations: here V max and J max are voltage and current density for the maximum power output, V oc is the open-circuit voltage, J sc is the short-circuit current density and P in is the incident light power. Figure 5 displays the comparison of DSSC performance of C-V 2 O 5, H-V 2 O 5 and Pt counter electrodes and their photovoltaic characteristics are also summarized in table 1. The efficiency of the cell with Pt CE is 6.9%. The champion cells made of C-V 2 O 5 and H-V 2 O 5 counter electrodes exhibits the maximum efficiencies of 1.2% and 1.6% respectively. J sc for the device based on H-V 2 O 5 CE is higher than that of C-V 2 O 5, suggesting that the H-V 2 O 5 based CE has better surface area (as seen in SEM micrographs in figures 2(c) and (d)), which has enhanced the electrochemical performance.
Both C-V 2 O 5, and H-V 2 O 5 cells have increased series resistance (R s ) compared with the reference cell, and also have lower shunt resistances (R sh ), all of which reduce the FFs to 29% and 27% respectively. The lower series resistance in case of H-V 2 O 5 compared with C-V 2 O 5 , indicates the better adhesion of the material to the FTO substrate. It is clear that the J-V properties of the devices are sensitive to surface morphology of the CE.

Electrochemical impedance analysis
Electrochemical impedance spectroscopy (EIS) studies were carried out in the dark at the bias 0.9 V to understand the effect of the electrocatalytic activities of different counter electrodes on the I 3reduction. The impedance can be described by real (Z | ) and imaginary (Z || ) components or its modulus (|Z|) and the phase shift (j). The EIS of the electrodes is shown in figure 6. The semicircle in the highest frequency region describes the electron/transport at CE/electrolyte interface [40], which is clearly seen in the case of Pt (inset (i)) and is not distinctly visible in case of C-V 2 O 5 and H-V 2 O 5 . The charge transfer resistance across CE/electrolyte (R ct ) interface is a crucial parameter on which the power conversion efficiencies of DSSCs rely. The R ct value of the reference electrode Pt is the lowest among all the counter electrodes and H-V 2 O 5 is comparatively very much less than that of C-V 2 O 5 indicating the higher catalytic activity of Pt than that of the other two counter electrodes. The CV measurements also showed the higher electocatalytic activity of H-V 2 O 5 than that of the commercial one. The life time of the electrons was calculated from Bode phase shown in figure 7 using the expression τ=1/ 2πf max, where f max is the peak frequency in the mid-frequency region [41]. The life time of the electron for Pt is 11.43 ms, for C-V 2 O 5 is 1.59 ms and for H-V 2 O 5 is 1.2 ms. The longer electron lifetime for Pt CE based DSSC indicated more effective suppression of the back reaction between the photo-injected electrons and I 3in the    electrolyte. The CE made of H-V 2 O 5 shows better electrocatalytic and photoconversion performances than that of C-V 2 O 5. There is a close connection between adhesion of the active material on the substrate and the efficiency of the devices. An improper adhesion of the active material on the FTO substrate affects the photoconversion efficiency [42]. The lower efficiencies of C-V 2 O 5 cells are possibly due to improper adhesion which results from the granular features of micron size and the lower surface area of the active material on the FTO substrate. Although the performance of these devices are comparatively low, these results shed light on the possibility of V 2 O 5 as a CE, which can significantly reduce the production cost by five orders relative to conventional Pt [43]. These results are encouraging as a driver towards more research efforts on V 2 O 5 as a CE in the near future. We believe that there is still a plenty of room to further optimize V 2 O 5 material in the form of high surface area, tuning the electrical conductivity by doping and its synthesis procedures to fabricate better DSSCs.

Conclusion
In summary, for the first time, we have demonstrated the utilization of V 2 O 5 as counter electrodes in DSSCs. The C-V 2 O 5 and H-V 2 O 5 materials were structurally characterized by powder XRD, which showed the purity of these materials. The surface morphologies of commercial, as-prepared powders, and the screen printed counter electrodes were examined by SEM, which revealed that hydrothermal treatment to C-V 2 O 5 has facilitated in reducing the particle size compared to the commercially available ones. Both specimens showed comparable catalytic activity for the tri-iodide reduction. EIS measurements further demonstrated that the electrocatalytic activity of H-V 2 O 5 counter electrodes based devices is higher than that of C-V 2 O 5 . Likewise H-V 2 O 5 showed better photovoltaic parameters such as V OC , J SC , and η of 0.68 V, 8.34 mA cm −2 and 1.6% when compared to C-V 2 O 5 . Since the DSSCs using these counter electrodes still yield low FFs and low efficiencies as compared to conventional Pt counter electrodes; further improvements need to be carried out. However this looks a promising avenue of investigation since it was found that photovoltaic performance is affected by the morphology of V 2 O 5 materials. By tuning the morphology of counter electrodes made of V 2 O 5 and improving the layer adhesion, one can likely enhance the device performance and certainly can push the cell efficiencies towards conventional Pt. This work opens a way for further development of more practical and cost effective, stable Pt-free counter electrodes materials for DSSCs.