Cyclic voltammetric deposition of discrete nickel phosphide clusters with mesoporous nanoparticles on fluorine-doped tin oxide glass as a counter electrode for dye-sensitized solar cells

Abstract

A monolayer of widely spaced nickel phosphide clusters with mesoporous nanoparticles was directly formed on fluorine-doped tin oxide glass by cyclic voltammetric deposition as a counter electrode for a dye-sensitized solar cell (DSC). Cyclic voltammetry included the anodic dissolution of Ni-rich regions following the cathodic deposition, leading to the formation of discrete clusters with mesoporous nanoparticles. After annealing at 500 °C, the nickel phosphide could be characterized as Ni₅P₄ and its electrocatalytic behavior was evaluated by cyclic voltammetry and electrochemical impedance in an iodide/triiodide system. The mesoporous Ni₅P₄ catalyst prepared by the cyclic voltammetric method exhibits good electrocatalytic ability towards the iodide/triiodide redox couple as a result of its low charge-transfer resistance and diffusion impedance. The photoelectron conversion efficiency of a DSC employing the Ni₅P₄ counter electrode could reach 7.6%, which is higher than that of a DSC employing a Pt nanocluster counter electrode (7.2%). The successful utilization of ultra-low loading of the Ni₅P₄ catalyst for DSCs makes the application of such material more economically viable.

---

Introduction

Dye-sensitized solar cells (DSCs) have received much research attention because of their low cost and high efficiency at converting light into electricity.¹ The counter electrode is one of the most important components affecting the photovoltaic properties of DSCs.² Pt-based materials have been extensively used as an effective catalyst for counter electrode due to their high electrocatalytic activity and stability towards the iodide/triiodide redox couple.^3–7 However, the Pt-based materials are expensive and limited in supply, making them more difficult to be used in mass production of DSCs. Thus, it is highly important to replace Pt with alternative materials without the sacrifice of decreasing activity and stability in comparison to Pt-based catalysts. Pt-free catalysts such as carbon materials, oxides, nitrides, carbides, sulfides, and phosphides have been introduced into DSCs as the catalysts.^8–21 Among these Pt-free catalysts, nickel phosphides have become a focus of interest because of their magnetic and mechanical resistance properties, and good corrosion and wear resistances.²²

As a catalyst material in counter electrode, large surface area, high electrical conductivity, and an open porous structure are required for fast transport of electrolyte and electron.^23–25 Thus, the nickel phosphides with tailored configuration have become an important issue for the success of the counter electrode. It has been reported that Ni₅P₄ exhibits a decent catalytic activity for the iodide/triiodide redox couple and the catalytic activity can be improved significantly when Ni₅P₄ and mesoporous carbon are combined into one composite (Ni₅P₄/C) due to the reduced charge-transfer resistance and diffusion impedance.²⁶ The Ni₁₂P₅ particles have been successfully embedded into the graphene to form the composite which is beneficial to the electrocatalytic activity, electrolyte diffusion, and electrical conductivity for the counter electrode.²⁷

Spray-coating and doctor-blade coating methods have been utilized to fabricate the nickel phosphide/carbon composites on the FTO surface as a counter electrode for DSCs.^26,27 The resultant composite films of nickel phosphide/carbon with thickness up to several micrometers are required to achieve a reliable cell performance. In DSCs, some counter materials have low electrocatalytic activity for iodide/triiodide redox reaction. Thus, thick film is required to compensate for the lack of electrocatalytic activity, but it is unfavorable to back-illuminated DSCs. Electrochemical deposition method allows to deposit a thin layer of nickel phosphide spheres on FTO surface.²⁸ The formation of nickel phosphide nanostructures on FTO during electroplating process is mainly affected by voltage, current density, and bath parameters such as pH, temperature, composition, and deposition time. The electrocatalytic activity of nickel phosphides depends not only on their structure, but also on the composition. The nickel phosphide alloy deposited by cathodic method may have Ni-rich regions in the bulk alloy.²⁹ Periodically applying an anodic voltage (reverse voltage) following the cathodic deposition may remove the Ni-rich regions, leading to the formation of porous nickel phosphide alloy.²⁸ The periodic voltage reversal method has been demonstrated as an effective way to deposit the nickel sulfides on FTO glass to replace the Pt/FTO counter electrode.³⁰ The potentiodynamic deposition has been used to prepare the mesoporous CoS as a counter electrode for DSC.³¹ The reverse sweep could facilitate the formation of CoS with excellent electrocatalytic activity for iodide/triiodide redox reaction.³¹

Electrochemical deposition of nickel phosphide is preferred for its simple procedure under room temperature. In the present work, the cyclic voltammetric technique involving potentiodynamic deposition and dissolution processes is employed to deposit largely spaced nickel phosphide clusters with mesoporous nanoparticles on FTO as a transparent counter electrode for DSCs. The high transparency of counter electrode is crucial in back-illuminated DSCs. Photovoltaic properties of DSCs using nickel phosphide/FTO counter electrode were characterized and compared with those of DSC using Pt/FTO counter electrode.

Experimental

The discrete nickel phosphide clusters were deposited on FTO glass (3.1 mm thick, 13 Ω □⁻¹, Nippon Sheet Glass) using cyclic voltammetric technique at room temperature in a homemade three-electrode cell. FTO glass with an exposure area of 1.5 cm × 1 cm, Pt foil (2 cm × 2 cm), and SCE (saturated calomel electrode) were used as the working, auxiliary, and reference electrodes, respectively. The deposition bath consisted of NiSO₄·6H₂O (0.27 M), NiCl₂·6H₂O (0.13 M), H₃PO₃ (0.3 M), and H₃PO₄ (0.35 M).²² The pH value of bath was adjusted to about 1.5 using 1 M NaOH prior to deposition. During cyclic voltammetric deposition, the voltage was linearly swept in the range of 0.1 to −0.8 V vs. SCE at a rate of 10 mV s⁻¹. The deposition bath was stirred at room temperature (about 26–29 °C) by a magnetic stir bar at 200 rpm. The number of deposition cycles was varied to obtain a reliable nickel phosphide catalyst for DSC application. After deposition, the nickel phosphide-coated FTO glass was rinsed with deionized water followed by heat-treated at 500 °C for 1 h in nitrogen atmosphere. For comparison, a nanocluster Pt counter electrode with high electrocatalytic activity was fabricated by a two-step dip coating process described in literature.⁶

The surface morphology of counter electrodes was examined with the field-emission electron microscope (FE-SEM, Auriga). The elemental analysis was carried out by use of an energy dispersive spectrometer (EDS, Auriga) coupled to the FE-SEM. The internal microstructure of nickel phosphide clusters scraped from FTO surface was characterized by the transmission electron microscopy (TEM, Jeol JEM-2010). The optical transmittance of catalyst-coated FTO electrodes was measured by means of an ultraviolet-visible spectrometer (UV-vis, PerkinElmer Lambda 35) in the wavelength range of 300 to 800 nm. The electrocatalytic performance of catalyst-coated FTO electrodes was evaluated by cyclic voltammetry (Keithley, 2400) in an acetonitrile solution containing 0.5 M LiClO₄, 50 mM LiI and 10 mM I₂ using a Pt foil and Pt wire as auxiliary and reference electrodes, respectively. The scan rate was 25 mV s⁻¹ and the voltage range was −0.4 to 0.4 V vs. Pt reference electrode.

Nanostructured TiO₂ photoanode was prepared by screen printing the TiO₂ slurry onto FTO glass and sintered at 450 °C for 1 h in air.³² The thickness of resultant TiO₂ film was approximately 15 μm measured by the profiler. The adsorption of dye on the TiO₂ surface was achieved by soaking the TiO₂ photoanode in an ethanol solution containing N719 dye for 12 h at room temperature. A dye-adsorbed photoanode was assembled with a counter electrode by using a sealing plastic film (Surlyn, about 50 μm thick) to form a sandwich-type DSC. An electrolyte containing 0.6 M 1-propyl-2,3-dimethylimidazolium iodide, 0.1 M lithium iodide, 0.05 M iodine, and 0.5 M 4-tert-butylpyridine in acetonitrile solvent was then infiltrated into the pores in electrode and space between the two electrodes. The active area of DSCs was 0.16 cm² (0.4 cm × 0.4 cm). The photocurrent–voltage characteristics of DSC under one-sun illumination (Yamashita Denso, YSS-E40; AM1.5, 100 mW cm⁻²) were measured by scanning DSC from the open-circuit voltage to the short-circuit condition (0 V) at a scan rate of 10 mV s⁻¹ using a source meter (Keithley, 2400) controlled by a personal computer through an IEEE 488 interface (GPIB). Electrochemical impedance spectroscopy (EIS) was carried out at open-circuit condition using a potentiostat/galvanostat (CH Instruments, CHI 608) with ac amplitude of 10 mV in a frequency range of 0.1–1 × 10⁵ Hz. For EIS measurements, the sandwich-type cell consisting of two identical Pt-coated or nickel phosphide-coated FTO electrodes, between which a Surlyn film was inserted as a spacer, was used. The cell was filled with the same electrolyte as the DSC.

Results and discussion

Fig. 1 illustrates the formation of discrete nickel phosphide clusters on FTO glass through the cyclic voltammetric deposition. The nucleation of nickel phosphide strongly depends on the overpotential during electrochemical deposition. At higher overpotential, the nucleation rate increases, which results in the formation and growth of smaller nuclei. Thus, the overlapping effect may cause clusters to grow into a compact film composed of closely spaced grains. In this work, the cathodic deposition voltage was linearly increased to obtain the discrete clusters. During cathodic process, the limit voltage was set at −0.8 V vs. SCE to provide a certain overvoltage for deposition of nickel phosphide without a significant hydrogen evolution and overlapping effect. The Ni-rich regions may be formed in the nickel phosphide alloy during cathodic deposition, which tend to dissolve more easily than the bulk alloy in anodic process.²⁹ Thus, the limit voltage during anodic process was set at 0.1 V vs. SCE to dissolve the Ni-rich regions in the bulk alloy, leading to the formation of porous nickel phosphide deposit. After cyclic voltammetric deposition for several cycles, largely spaced nickel phosphide clusters with mesoporous nanoparticles could be formed on the FTO glass as a transparent counter electrode for DSC application. There are many factors such as voltage window, concentration, temperature, and scan rate affecting the composition and structure of the nickel phosphide alloy during cyclic voltammetric deposition. In this work, these factors were held constant, and attention was focused on the number of deposition cycles.

Fig. 1 Schematic illustrating the formation of discrete nickel phosphide clusters on FTO glass through the cyclic voltammetric deposition.

Fig. 2 shows the SEM images of bare FTO, nickel phosphide/FTO, and Pt/FTO glass electrodes. The surface morphology of bare FTO electrode shown in Fig. 2a has a tetragonal structure with markedly uneven appearance. Clearly, largely spaced nickel phosphide clusters with size ranging from 50 to 150 nm could be formed on the FTO surface by cyclic voltammetric deposition for 100 cycles (Fig. 2b). Most of the FTO surface is not covered with nickel phosphide clusters, indicating that the loading amount of nickel phosphide catalyst is very small. Previous reports indicated that the nickel phosphides prepared with potentiostatic (constant voltage) or galvanostatic (constant current) methods show a compact film structure with closely spaced spheres.^22,28,29 A very compact structure with very small pores is an important drawback because the small pore makes it difficult for the electrolyte to enter into the interior of the catalyst layer, only the surface can be more easily reached by the electrolyte. The difference in surface morphology could be attributed to the applied voltage protocol because the cathodic overvoltage influences both the nucleation and growth of grains during electrochemical deposition. Potentiostatic and galvanostatic depositions lead to the continuous growth of the grains resulting in the formation of closely spaced spheres,^29,33 while the cyclic voltammetric deposition restricts the growth of grains due to the anodic dissolution of Ni-rich regions in the anodic process, leading to the formation of clusters with smaller nanoparticles on the FTO surface. For comparison, a nanocluster Pt counter electrode was fabricated by a two-step dip coating process and its surface morphology is shown in Fig. 2c. The FTO surface is uniformly covered with a layer of closely spaced Pt nanoclusters. Counter electrode employing two-step dip coating method exhibits not only ultralow Pt usage but also good catalytic performance similar to a smooth Pt sheet electrode.³⁴

Fig. 2 SEM images of (a) bare FTO, (b) nickel phosphide/FTO, and (c) Pt/FTO glass electrodes. Nickel phosphide clusters were deposited by cyclic voltammetric method for 100 cycles and heat-treated at 500 °C in nitrogen atmosphere. Insets show high magnification images.

Fig. 3 shows the TEM images of nickel phosphide cluster prepared by cyclic voltammetric deposition for various cycles followed by annealing at 500 °C in nitrogen atmosphere. The nickel phosphide cluster deposited for 100 cycles shown in Fig. 3a is composed of several mesoporous nanoparticles with an average size of approximately 25 nm. The lattice image shown in the inset of Fig. 3b indicates that the lattice spacing of around 0.55 nm corresponds to the d-spacing of (002) crystal planes of Ni₅P₄ (JCPDS no. 89-2588). Interestingly, the Ni₅P₄ prepared by cyclic voltammetric deposition has small particle size and porous structure, resulting from the periodically potentiodynamic deposition and dissolution of Ni-rich regions in the bulk alloy. Porous structure may facilitate the electrolyte transport through nanoparticle, leading to an increase in the effective surface area for electrocatalytic reaction. Large active surface area is required to compensate the small amount of Ni₅P₄ catalyst supported on the FTO surface. Fig. 3c shows a large nickel phosphide cluster prepared by cyclic voltammetric deposition for 200 cycles. Clearly, the larger cluster composed of aggregated nanoparticles could be obtained with increasing the number of deposition cycles. As shown in Fig. 3d, the porous structure of nickel phosphide cluster remains even the number of deposition cycles is beyond 100. Larger clusters make the transport of both electron and electrolyte more difficult and require more time to reach the interior region of the clusters.

Fig. 3 (a) TEM image and (b) enlarged TEM image of nickel phosphide cluster prepared by cyclic voltammetric deposition for 100 cycles. The lattice image is shown in the inset (b). (c) TEM image and (d) enlarged TEM image of nickel phosphide cluster prepared by cyclic voltammetric deposition for 200 cycles.

Fig. 4 shows the UV-vis absorption spectra of bare FTO, Ni₅P₄/FTO, and Pt/FTO glass electrodes. The transmittance of FTO glass electrode is reduced after coated with Pt or nickel phosphide catalysts. The transmittance of Ni₅P₄/FTO electrode prepared by cyclic voltammetric method for 100 cycles could reach as high as 71% at the wavelength of 550 nm, which is slightly lower than that of bare FTO electrode (79%) and higher than that of Pt/FTO electrode (58%). The film thickness and coverage of catalyst layer influence the transparency of counter electrodes. It was reported that the nickel sulfide film with the thickness of 100 nm can reach the high transparency.³⁵ A monolayer of discrete Ni₅P₄ nanoclusters with very low coverage on FTO accounts for the high transmittance of Ni₅P₄/FTO glass. Transparent counter electrodes are key to the fabrication of bifacial DSCs.^36,37 In this work, cyclic voltammetric method features easy fabrication of transparent catalyst layer with well-dispersed porous Ni₅P₄ nanoclusters on the FTO surface, which is available for application in back-illuminated DSCs.

Fig. 4 UV-vis absorption spectra of bare FTO, Ni₅P₄/FTO, and Pt/FTO glass electrodes. The inset shows the photographs of, from left to right: bare FTO, Ni₅P₄/FTO, and Pt/FTO.

Fig. 5a shows the cyclic voltammograms of bare FTO and Ni₅P₄/FTO electrodes in iodide/triiodide electrolyte at a scan rate of 25 mV s⁻¹. Ni₅P₄/FTO electrodes were prepared by cyclic voltammetric method for various deposition cycles. The EDS analysis reveals that the atomic ratio of Ni to P in the nickel phosphide electrodes deposited at various cycles is found to be approximately 5/4, corresponding to a chemical formula Ni₅P₄. All Ni₅P₄ electrodes exhibit an apparent pair of redox peaks, resulting from the redox reaction between iodide and triiodide ions at the electrolyte–electrolyte interface. Compared to bare FTO, the presence of Ni₅P₄ catalyst significantly increases the peak current density, indicating that Ni₅P₄ has a much higher electrocatalytic activity towards redox of iodide/triiodide than bare FTO. Apparently, the electrocatalytic activity of Ni₅P₄/FTO electrode is considerably influenced by the number of deposition cycles. The peak current density increases with increasing the number of deposition cycles, it reaches a maximum value at 100 cycles, and then decreases with further increase in the cycles. The more the number of deposition cycles, the larger is the loading of Ni₅P₄ catalyst on the FTO surface. It is known that the peak current in a cyclic voltammogram is proportional to the active surface area of electrode and the diffusion coefficient of electrolyte. Larger catalyst loading may provide more surface area. Thus, an increase in peak current density corresponds to the increases in Ni₅P₄ catalyst loading on the FTO when the diffusion effect is not significantly changed by the nanoclusters. A decrease in peak current density beyond 100 cycles is attributed to the reduced electrocatalytic activity by the aggregation of particles. It is believed that the large aggregates are unfavorable for the transport of electrolyte because not all the surface area of aggregates can be easily accessed by the electrolyte. In addition, Ni₅P₄/FTO electrode prepared by cyclic voltammetric method for 100 cycles shows the smallest peak to peak separation (ΔE_p) compared with electrodes deposited other than 100 cycles. ΔE_p varies with the charge-transfer resistance, reflecting that the Ni₅P₄/FTO electrode prepared for 100 cycles has the lowest charge-transfer resistance. Although the FTO surface could be exposed to electrolyte through the porous structure of the Ni₅P₄ layer and the uncovered region, but a bare FTO electrode has a much poorer electrocatalytic activity than Ni₅P₄/FTO electrodes. Thus, the electrocatalytic performance of Ni₅P₄ catalyst should be mainly dependent on the Ni₅P₄ loading rather than the exposed surface area of FTO substrate.

Fig. 5 (a) Effect of deposition cycles on the cyclic voltammograms of Ni₅P₄/FTO electrodes in iodide/triiodide electrolyte at a scan rate of 25 mV s⁻¹. (b) Comparison between cyclic voltammograms obtained from the Pt/FTO and Ni₅P₄/FTO (deposited for 100 cycles) electrodes in iodide/triiodide electrolyte at a scan rate of 25 mV s⁻¹.

As revealed in Fig. 5b, the Ni₅P₄/FTO electrode deposited for 100 cycles exhibits an apparent pair of redox peaks like the Pt/FTO electrode. Better electrocatalytic property of the Ni₅P₄/FTO electrode in iodide/triiodide electrolyte is evident from high peak current density and small ΔE_p compared with Pt/FTO electrode. The improved electrocatalytic activity of Ni₅P₄/FTO electrode may be attributed to the largely spaced clusters with mesoporous nanoparticles that accommodate electrolyte solution and provide exceptional amounts of active sites in between the electrolyte–electrode interface for facilitating the electrocatalytic reactions. It has been reported that the kinetic and diffusion properties of nickel phosphide electrodes could be improved by introducing the carbon materials to the composite electrodes.^26,27 Graphene-supported nickel nanoparticles have shown better electrocatalytic performance compared to the unsupported nickel nanoparticles due to the improved charge-transfer resistance.³⁸ In this work, an ultrasmall amount of Ni₅P₄ without the need of carbon materials can reach a higher electrocatalytic activity than the Pt nanocluster catalyst. The stability of Ni₅P₄ catalyst for iodide/triiodide redox couple is evaluated by consecutive cyclic voltammetry tests and the result is shown in Fig. 6. The redox peak current densities decrease in the first 20 cycles and then decrease very slowly after 20 cycles. After an initial decay, the redox peak current densities of Ni₅P₄/FTO electrode in iodide/triiodide system remain almost unchanged in the subsequent cycles (over than 1000 cycles), indicating good electrochemical stability of Ni₅P₄/FTO electrode prepared by cyclic voltammetric method. It is known that Pt is durable under room temperature, but it starts to react with electrolyte under high humidity and high temperature. Such severe stability test for Ni₅P₄/FTO electrode will be evaluated in the future, making it more attractive for practical application in DSCs.

Fig. 6 Cyclic voltammograms of Ni₅P₄/FTO electrode obtained from consecutive cycles in iodide/triiodide electrolyte. Inset shows the variation of redox peak current densities associated with the number of cycles. Ni₅P₄/FTO was deposited by cyclic voltammetric method for 100 cycles.

Fig. 7 shows the Nyquist plots of symmetric cells using two identical Pt-coated or Ni₅P₄-coated FTO electrodes. The obtained spectra were fitted with the equivalent circuit shown in the inset of Fig. 7. The resistance element denoted as R_s in the high-frequency region can be ascribed to the sheet resistance of the bare FTO substrate. R_s can be obtained from the high-frequency intercept of the Nyquist plots on the real axis. All cells exhibited a similar value of R_s (close to 14.5 Ω), probably due to the same FTO substrate used. A depressed semicircle composed of two overlapped semicircles appears in the Nyquist plots. The semicircle in the high-frequency region represents the charge-transfer (R_ct) process at the electrode–electrolyte interface, and another one in the low-frequency region indicates the Nernst diffusion impedance (Z_d) process of I⁻/I₃⁻ redox species within pores. It is generally believed that the value of R_ct plays an important role in determining the photovoltaic performances of the DSCs. The charge-transfer resistances of cells with Pt, Ni₅P₄-100 (deposited for 100 CV cycles), and Ni₅P₄-200 (deposited for 200 CV cycles) electrodes are 14.9, 14.2, and 18.4 Ω, respectively. The cell consisting of two Ni₅P₄-200 electrodes has a large R_ct of 18.4 Ω, probably due to the large clusters composed of aggregated nanoparticles that decrease the active sites available for I⁻/I₃⁻ redox species. After decreasing the size of clusters by reducing the number of deposition cycles to 100, the R_ct of cell consisting of two Ni₅P₄-100 electrodes can be reduced to 14.2 Ω, which is even below that of cell consisting of a pair of Pt electrodes (14.9 Ω). This result highlights the superior electrocatalytic activity of Ni₅P₄ electrode prepared by cyclic voltammetric deposition for 100 cycles. The Nernst diffusion impedance values of the cells with Pt, Ni₅P₄-100, and Ni₅P₄-200 electrodes are 13.5, 9.7, and 27.3 Ω, respectively. The diffusion of I⁻/I₃⁻ is faster within the small porous Ni₅P₄-100 clusters compared with large Ni₅P₄-200 clusters. As a result, ultrasmall Ni₅P₄-100 clusters composed of mesoporous nanoparticles show superior electrocatalytic features for iodide/triiodide redox couple, including the lower charge-transfer resistance and diffusion impedance. The unique electrode structure provides more channels for the electrolyte to percolate into the interior of the clusters.

Fig. 7 Electrochemical impedance spectra of systematical cells with Pt/FTO or Ni₅P₄/FTO electrodes under open-circuit condition.

Fig. 8 shows the photocurrent–voltage (J–V) characteristics of DSCs employing the Pt/FTO and Ni₅P₄/FTO counter electrodes. The resulting photovoltaic parameters obtained from J–V curves are tabulated in Table 1. Photovoltaic properties of DSC with Ni₅P₄/FTO (prepared by cyclic voltammetric deposition for 100 cycles) are better than those of DSC with Pt/FTO under one-sun illumination. When a Pt/FTO is used as the counter electrode, the open-circuit voltage (V_oc) of DSC is 0.70 V, the short-circuit current density (J_sc) is about 14.1 mA cm⁻², the fill factor (FF) is 73%, and the photoelectron conversion efficiency (η) of cell reaches to 7.2%. When a monolayer of discrete Ni₅P₄ clusters is coated on FTO glass as a counter electrode, the V_oc, J_sc, and FF of cell are 0.72 V, 14.7 mA cm⁻², and 72%, respectively. The η value of DSC achieved by using the Ni₅P₄/FTO counter electrode is increased to 7.6%, indicating the high electrocatalytic activity of Ni₅P₄ catalyst. Porous nanoparticles formed by cyclic voltammetric method may provide large electroactive surface area and efficient ion transport pathways for facilitating the reduction and diffusion of triiodide ions. Cyclic voltammetric method was used to obtain Ni₅P₄ catalysts with very low catalyst loadings. The highest cell efficiency achieved is 7.6% using the Ni₅P₄/FTO counter electrode prepared by cyclic voltammetric method for 100 cycles, which is higher than that using the Pt nanocluster electrode. Thus, the Ni₅P₄ catalyst on FTO glass is a promising candidate for a low-cost and transparent electrode and is suitable for application in back-illuminated DSCs.

Fig. 8 Photocurrent–voltage curves of DSCs employing the Pt/FTO and Ni₅P₄/FTO counter electrodes under one-sun illumination.

Table 1 Photovoltaic parameters obtained from the photocurrent–voltage curves of DSCs employing the Pt/FTO and Ni₅P₄/FTO counter electrodes

Counter electrodes	Jsc (mA cm^−2)	Voc (V)	FF	η (%)
Pt/FTO	14.1 ± 0.3	0.70 ± 0.01	0.73 ± 0.01	7.2 ± 0.2
Ni5P4/FTO	14.7 ± 0.3	0.72 ± 0.01	0.72 ± 0.01	7.6 ± 0.2

---

Conclusions

A catalyst layer of largely spaced Ni₅P₄ clusters on FTO glass was prepared by cyclic voltammetric deposition for 100 cycles followed by heat treatment as a counter electrode for DSCs. The Ni₅P₄ clusters were found to comprise mesoporous nanoparticles. In cyclic voltammetry, the anodic dissolution of Ni-rich regions following the cathodic deposition led to the formation of porous clusters with limited particle size on the FTO surface. The results clearly demonstrated that cyclic voltammetric technique is an effective method to the development of catalysts containing very low loading on FTO, whilst maintaining high electrocatalytic activity. The Ni₅P₄/FTO electrode exhibited higher transparency and electrocatalytic activity than the Pt/FTO electrode, making it available to the back-illuminated DSCs. Discrete Ni₅P₄ clusters composed of mesoporous nanoparticles provided more channels for facilitating the electrolyte transport and increasing the surface area accessible to electrolyte ions. Therefore, the charge-transfer resistance and diffusion impedance of Ni₅P₄/FTO could be very small, primarily due to the high peak current density and small peak to peak separation in the cyclic voltammograms. The photoelectron conversion efficiency of DSC with Ni₅P₄/FTO counter electrode (7.6%) was higher than that of DSC with Pt/FTO (7.2%). The Ni₅P₄/FTO counter with discrete clusters exhibited an excellent electrocatalytic activity and good electrochemical stability in iodide/triiodide system, which is a highly prospective material for counter electrode in DSCs.

Acknowledgements

The authors acknowledge financial support from the Ministry of Science and Technology, Taiwan (Project no. MOST 103-2221-E-151-057-MY3 and 103-3113-E-007-005).

Notes and references

1. B. O'Regan and M. Gratzel, Nature, 1991, 353, 737–740.
2. T. N. Murakami and M. Grätzel, Inorg. Chim. Acta, 2008, 361, 572–580.
3. G. Calogero, P. Calandra, A. Irrera, A. Sinopoli, I. Citro and G. Di Marco, Energy Environ. Sci., 2011, 4, 1838–1844.
4. S. J. Cho and J. Ouyang, J. Phys. Chem. C, 2011, 115, 8519–8526.
5. Y.-L. Lee, C.-L. Chen, L.-W. Chong, C.-H. Chen, Y.-F. Liu and C.-F. Chi, Electrochem. Commun., 2010, 12, 1662–1665.
6. T. C. Wei, C. C. Wan and Y. Y. Wang, Appl. Phys. Lett., 2006, 88, 103122.
7. S. Thomas, T. G. Deepak, G. S. Anjusree, T. A. Arun, S. V. Nair and A. S. Nair, J. Mater. Chem. A, 2014, 2, 4474–4490.
8. A.-M. Alexander and J. S. J. Hargreaves, Chem. Soc. Rev., 2010, 39, 4388–4401.
9. M. Wang, A. M. Anghel, B. t. Marsan, N.-L. Cevey Ha, N. Pootrakulchote, S. M. Zakeeruddin and M. Grätzel, J. Am. Chem. Soc., 2009, 131, 15976–15977.
10. G. R. Li, J. Song, G. L. Pan and X. P. Gao, Energy Environ. Sci., 2011, 4, 1680–1683.
11. Y. Wang, M. Wu, X. Lin, Z. Shi, A. Hagfeldt and T. Ma, J. Mater. Chem., 2012, 22, 4009–4014.
12. M. Wu, X. Lin, Y. Wang, L. Wang, W. Guo, D. Qi, X. Peng, A. Hagfeldt, M. Grätzel and T. Ma, J. Am. Chem. Soc., 2012, 134, 3419–3428.
13. G. H. Guai, M. Y. Leiw, C. M. Ng and C. M. Li, Adv. Energy Mater., 2012, 2, 334–338.
14. L. Wang, E. W.-G. Diau, M. Wu, H.-P. Lu and T. Ma, Chem. Commun., 2012, 48, 2600–2602.
15. J. D. Roy-Mayhew and I. A. Aksay, Chem. Rev., 2014, 114, 6323–6348.
16. M. Wu and T. Ma, J. Phys. Chem. C, 2014, 118, 16727–16742.
17. E. Bi, H. Chen, X. Yang, W. Peng, M. Gratzel and L. Han, Energy Environ. Sci., 2014, 7, 2637–2641.
18. J. Yang, C. Bao, K. Zhu, T. Yu, F. Li, J. Liu, Z. Li and Z. Zou, Chem. Commun., 2014, 50, 4824–4826.
19. W. Ke, G. Fang, H. Tao, P. Qin, J. Wang, H. Lei, Q. Liu and X. Zhao, ACS Appl. Mater. Interfaces, 2014, 6, 5525–5530.
20. J.-Y. Lin, W.-Y. Wang, Y.-T. Lin and S.-W. Chou, ACS Appl. Mater. Interfaces, 2014, 6, 3357–3364.
21. C.-W. Kung, H.-W. Chen, C.-Y. Lin, K.-C. Huang, R. Vittal and K.-C. Ho, ACS Nano, 2012, 6, 7016–7025.
22. S. S. Djokić, J. Electrochem. Soc., 1999, 146, 1824–1828.
23. Q. Jiang, Y. Qiu, K. Yan, J. Xiao and S. Yang, MRS Commun., 2012, 2, 97–99.
24. Q. W. Jiang, G. R. Li and X. P. Gao, Chem. Commun., 2009, 6720–6722.
25. H. Su, M. Zhang, Y.-H. Chang, P. Zhai, N. Y. Hau, Y.-T. Huang, C. Liu, A. K. Soh and S.-P. Feng, ACS Appl. Mater. Interfaces, 2014, 6, 5577–5584.
26. M. Wu, J. Bai, Y. Wang, A. Wang, X. Lin, L. Wang, Y. Shen, Z. Wang, A. Hagfeldt and T. Ma, J. Mater. Chem., 2012, 22, 11121–11127.
27. Y. Y. Dou, G. R. Li, J. Song and X. P. Gao, Phys. Chem. Chem. Phys., 2012, 14, 1339–1342.
28. M.-S. Wu and J.-F. Wu, Chem. Commun., 2013, 49, 10971–10973.
29. J. Crousier, Z. Hanane and J. P. Crousier, Thin Solid Films, 1994, 248, 51–56.
30. H. Sun, D. Qin, S. Huang, X. Guo, D. Li, Y. Luo and Q. Meng, Energy Environ. Sci., 2011, 4, 2630–2637.
31. J.-Y. Lin and J.-H. Liao, J. Electrochem. Soc., 2011, 159, D65–D71.
32. M.-S. Wu, C.-H. Tsai and T.-C. Wei, Chem. Commun., 2011, 47, 2871–2873.
33. G. McMahon and U. Erb, J. Mater. Sci. Lett., 1989, 8, 865–868.
34. T.-C. Wei, C.-C. Wan, Y.-Y. Wang, C.-m. Chen and H.-s. Shiu, J. Phys. Chem. C, 2007, 111, 4847–4853.
35. J. Song, G. R. Li, C. Y. Wu and X. P. Gao, J. Power Sources, 2014, 266, 464–470.
36. Q. Tai and X.-Z. Zhao, J. Mater. Chem. A, 2014, 2, 13207–13218.
37. Y.-S. Wang, S.-M. Li, S.-T. Hsiao, H. Wei, S.-Y. Yang, H.-W. Tien, C.-C. M. Ma and C.-C. Hu, J. Power Sources, 2014, 260, 326–337.
38. R. Bajpai, S. Roy, N. Kulshrestha, J. Rafiee, N. Koratkar and D. S. Misra, Nanoscale, 2012, 4, 926–930.

---