Interface Engineered Binary Platinum Free Alloy-based Counter Electrodes with Improved Performance in Dye-Sensitized Solar Cells

The high cost and platinum dissolution issues of counter electrodes (CEs) are two core problems for the development of dye-sensitized solar cells (DSSCs). In this work, different CEs based on binary alloy Ru81.09Co18.91, Ru80.55Se19.45 and Co20.85Se79.15 nanostructures for DSSCs were successfully synthesized and investigated by a facile and environmentally friendly approach. Here, we found that the Co20.85Se79.15 alloy CE-based device yields the higher photoelectric conversion efficiency of 7.08% compared with that (5.80%) of the DSSC using a pure Pt CE because of the larger number of active sites with improved charge transferability and reduced interface resistance by matching work function with the I3‒/I‒ redox electrolyte. The inexpensive synthesis method, cost-effectiveness and superior catalytic activity of the Co20.85Se79.15 alloy may open up a new avenue for the development of CEs for DSSCs in the near future.

thereby generating very large numbers of active sites for the reduction reaction of I 3 -ions. Nevertheless, the work functions of some bimetallic PtM x alloys, such as PtCo, PtRu, or PtPd do not match the potential of the I 3 -/I -based electrolyte, leading to unsatisfactory electrocatalytic activity and electron transport [17][18][19] . Meanwhile, the high cost of Pt metal is still a crucial obstacle for the further development of DSSCs.
In recent years, perovskite solar cells have attracted considerable attention worldwide with a corresponding gradual decrease in the number of studies focused on bimetallic Pt-free alloy electrocatalysts for DSSCs. From this perspective, the development of Pt-free polymetallic alloys has become a very important subject in the electrocatalysis field. Yin et al. synthesized a novel N-doped-carbon coated CoSe 2 on a 3D carbon cloth as a photocathode for DSSCs, which exhibited a good photoelectric efficiency (8.40%) and cycle stability 20 . A carbon shell coated CoSe 2 nanoparticles catalyst-based DSSC was also reported, which gives the good conversion efficiency of 7.54% 21 . Wang et al. reported a Co 0.85 Se and Ni 0.85 Se CE -based DSSCs with high efficiency of 9.40 and 8.32%, respectively, which were synthesized by in situ growth using a hydrothermal method 22 . These results indicate that transition metal selenides have superior catalytic activities than the pristine Pt CE for DSSCs. In this case, for the purpose of reducing the fabrication cost and improving their electrocatalytic activity and stability in DSSCs, the Pt-free bimetallic alloys including RuCo, RuSe and CoSe -based CEs were synthesized by a simple electrodeposition approach. The exact compositions of Ru 81.09 Co 18.91 , Ru 80.55 Se 19.45 and Co 20.85 Se 79. 15 can be confirmed by Energy Dispersive Spectrometer (EDS) and X-ray photoelectron spectroscopy (XPS) characterization. As a result, the power conversion efficiency of Co 20.85 Se 79. 15 CE-based DSSCs reached 7.08%, compared with the 5.80% value of pure Pt CE-based DSSCs. The improved catalytic performance can be attributed to the matching work function, a large number of active sites and reduced interface resistance.

Results and discussion
High-magnification FESEM images of various alloy CEs electrodeposited on the surface of the FTO glass are shown in Fig. 1(a-f). The Ru 81.09 Co 18.91 , Ru 80.55 Se 19.45 , and Co 20.85 Se 79.15 alloy CEs exhibited uniformly distributed surface morphologies compared to those of pure metal (Ru, Se, and Co) CEs. The alloy CEs showed a smaller average particle diameter and an almost homogeneous distribution ( Fig. 1(a-c)), whereas a larger grain size and non-uniform surface appearance were observed for the pure metal CEs ( Fig. 1(d-f)). XRD results of Co 20.85 Se 79.15 alloy was characterized as shown in Figure S1, the peaks at 37.62°, 51.57° match well with (211) and (311) crystalline faces of CoSe 2 (JCPDS PDF#09-0234), respectively. The peak at 51.52° can be well indexed to (200) face of Co (JCPDS PDF#15-0806). The peak at 33.33°, 61.57° can be well indexed to (101) and (103) face of Se (JCPDS PDF#27-0601). In order to confirm the composition of the metal alloys, the CEs were analyzed by the EDS method as shown in Figure S2 and the corresponding quantitatively compositional results were shown in Table 1, for which the atomic molar ratios of the alloys were found to be 1:0 The X-ray photoelectron spectroscopy (XPS) spectra were also utilized to characterize the formation of different alloys as shown in Fig. 2. From Fig. 2 Fig. 2(b), the two peaks located at 280.5 and 281.2 eV were attributed to the bonding results of Ru-O and Ru-Ru. The Co 2p spectrum in Fig. 2(c) can be distributed into shakeup satellites (Sat.) and spin-orbit doublets. The first double peak at 780.7 and 796.6 eV, and the second speaks at 782.0 and 797.8 eV belong to Co 2+ and Co 3+ , respectively [21][22][23] . The shakeup double peaks indicated the formation of the Co-Co bond [21][22][23] . Fig. 2(e) exhibits the Ru 3d spectrum of Ru 80.55 Se 19.45 alloy while two peaks located at 280.3 and 281.1 eV were assigned to the bonding results of Ru-Se and Ru-Ru. The doublet peaks at 54.5 and 55.5 eV in Fig. 2(f) can be ascribed to the bonding results of Ru-Se and Se-Se 21,24 . Fig. 2(h) shows the Co 2p spectrum of Co 20.85 Se 79.15 alloy. Similarly, the first doublet peaks at 780.9 and 796.5 eV, and the second speaks at 783.2 and 798.6 eV suggest the bonding results of Se-Co-Se and Co-O, respectively [21][22][23]25 . From Fig. 2(i), the two peaks at 53.8 and 54.6 eV were related to the bonding results of Co-Se and Se-Se, respectively 21,24 . The results indicated that the Co 20.85 Se 79.15 alloy can be composed of CoSe 2 , Co 2 O 3 , Co and Se. It is believed that the alloy CEs can provide a higher number of active sites for the adsorption of I 3 ions and their reduction reaction (I 3 -+ 2e -= 3I -), thereby accelerating the electrocatalytic reaction on the electrolyte/CEs interface as well as the transport of electrons. The smaller crystalline size of the alloy CEs may be explained in terms of the crystal growth occurring during the electrodeposition process. The multiple crystal nuclei on the alloy surface would contribute to the formation of smaller nanocrystals 26 . Figure 3(a,b) display the photocurrent density-voltage plots of different CE-based DSSCs with and without sunlight illumination, respectively. Figure 3(c) shows the corresponding device configuration of the binary alloy CEs in DSSCs. The photoelectric parameters of the DSSCs are also shown in Table 2 15 , which offers a much higher number of active sites for adsorption and reduction of I 3 ions. Therefore, the reduced electron-transfer resistance between the electrolyte and the Co 20.85 Se 79.15 CE can efficiently accelerate the reduction reaction and the electron transport, resulting in the enhanced efficiency of the DSSCs. Furthermore, a dark current, the characteristic feature associated with recombination reactions between photogenerated electrons at the conduction band of TiO 2 and the electrolyte (I 3 -), was investigated as shown in Fig. 3(b). Clearly, the Co 20.85 Se 79.15 alloy exhibited the smallest dark current density compared to that of pristine Pt, other alloys (Ru 81.09 Co 18.91 and Ru 80.55 Se 19.45 ) and pure metals (Ru, Se and Co), demonstrating suppression of the electron-loss reaction. The result can be explained by the fact that the Co 20.85 Se 79.15 alloy can accelerate the kinetics of the reduction reaction, namely, I 3 -+ 2e -= 3Iand the regeneration rate of dye molecules. Therefore, the Co 20.85 Se 79.15 alloy possesses advantages, such as a higher number of active sites, matched energy levels and  www.nature.com/scientificreports www.nature.com/scientificreports/ reduced interface resistance, resulting in a fast electrolyte (I 3 -) reduction at the electrolyte/CE interface instead of the recombination with electrons originated from the TiO 2 conduction band.
The electrocatalytic activities of the alloy CEs towards I 3 reduction were also investigated by electrochemical methods as shown in Fig. 4. Figure 4(a) shows cyclic voltammogram (CV) of the different CEs where two couples of redox peaks can be seen in the CV plots. The redox peaks on the left side of the figures correspond to the I 3 -→ Iprocess (red 1 : ) while the ones on the right correspond to the I 2 → I 3 process (red 2 : 26 . Since the I 3 -→ Ireduction process is the dominating reaction in I 3 -/I --based electrolyte systems, we mainly focus on the left redox peaks. The plot shows that the redox peaks of alloy CEs are much stronger than those of the pristine metals (Ru, Se and Co), suggesting that the alloy CEs have a superior catalytic activity toward I 3 reduction. Table 3 shows electrochemical parameters obtained from CV plots and EIS at different CEs where the J red1 and E red1 represent the peak current density and potential of reduction reaction (red 1 ), respectively. Because J red1 is a key parameter for estimating the electrocatalytic activity of CEs 27 , clearly, the J red1 decreases in the order Co 20 15 alloy. Moreover, the Randles-Sevcik theory was used to investigate the ion diffusion at the electrolyte/CE interface. The electron diffusion coefficient, D n , was determined from the equation J red1 = kn 1.5 AD n 0.5 C 0 v 0.5 where C 0 is the I 3 -/Iion concentration, v is the scan rate, A is the active area of the CEs, n is the number of electrons involved in the reduction process and K is a constant 28,29 . As a result, the calculated D n values of Ru 81.09 Co 18.91 , Ru 80.55 Se 19.45 , Co 20.85 Se 79.15 , Ru, Se, Co and Pt were 2.47×10 -5 , 2.22×10 -5 , 3.67×10 -5 , 0.37×10 -5 , 0.06×10 -5 , 0.03×10 -5 and 3.05×10 -5 , respectively. The improved D n of the Co 20.85 Se 79.15 alloy denotes the faster diffusion kinetics of I 3 ions between electrolyte and CEs. The higher number of active sites and the ligand effect between Co and Se are also expected to contribute the enhanced catalytic properties of the corresponding alloy CEs. Figure 4(b) shows the relationship between the square root of the scan rates and the peak current density where CV plots obtained for the CEs at different scan rates are shown in Figure S4. Obviously, the peak current density of reduction and oxidation increased almost linearly with the scan rate, indicating that the electrochemical reaction is controlled by the diffusion behavior of I 3 ions at the electrolyte/ CE interface 30 . In order to examine the internal electron transfer kinetics at the electrolyte/CE interface, Nyquist EIS curves of symmetrical devices, consisting of two identical CEs and the electrolyte, were measured and displayed in Fig. 4(c,d), respectively. The charge-transfer resistance (R ct ) between electrolyte and CEs determined by   Fig. 3(c). A smaller R ct denotes a rapid electron transport kinetics at the electrolyte/CE interface (Table 3). Therefore, the catalytic reactions are effectively accelerated by the Co 20.85 Se 79.15 alloy CEs and can be further confirmed by examining the interfacial electron lifetimes (τ e = 0.5πf peak where f peak represents the peak frequency in the Bode plots displayed in Fig. 4e) as shown in Fig. 4(e) and Table 4 shows electrochemical parameters obtained from the EIS and Tafel polarization plots based on CE/electrolyte/ CE structures 31 80 μs), respectively. Lower τ e values indicate the faster reduction kinetics of I 3 ions, yielding the improved electrocatalytic ability of the CEs. These results are in good agreement with the CV plots. Based on the Tafel polarization plots in Fig. 4(f), the exchange and diffusion-limited current densities (J 0 and J lim , respectively) can be calculated from the equations 9 J 0 = RT/nFR ct and J lim = 2nFCD n /l, respectively (    Table 2. Photovoltaic parameters of different CEs based DSSCs under AM 1.5 G sunlight (100 mW cm -2 ) illumination. J sc : Short-circuit current density, V oc : Open-circuit voltage, FF: fill factor, η: Photo-electric conversion efficiency.
In order to further investigate the internal mechanism of the CEs, the work functions of the different CEs were determined by SKPM using a gold probe (5.1 eV) as a standard reference, and are shown in Fig. 6   www.nature.com/scientificreports www.nature.com/scientificreports/ -5.25, -4.94, -5.33, -5.40 and -5.55 eV whereas the corresponding value for the pure Pt electrode is -5.01 eV, respectively. Therefore, the work functions of the Co 20.85 Se 79.15 alloy CE show a better match with the potential (-4.90 eV) of the I 3 -/Iredox electrolyte 18 compared with those of the other CEs, thus resulting in improved electrocatalytic performance. The good match of the Co 20.85 Se 79.15 CE work function can be attributed to the ligand effect of the Co and Se transition metals, which would reduce the bond energy between atoms and free electrons 32 . As a result, the electronic configurations of Co and Se atoms near the surface are readjusted in such a way that electrons are prone to participate in the electrolyte (I 3 -) reduction process. Furthermore, the charge transport resistance is defined as the difference between the work function of the CE and the potential of the I 3 -/Iredox electrolyte of DSSCs 11,33 . For this reason, a lower energy drop would efficiently accelerate the electron transport from electrocatalyst CEs to I 3 -11,33 , which is in good agreement with the above EIS analysis. In this case, the superior electrocatalytic activity of the Co 20.85 Se 79.15 CE toward the I 3 electrolyte can definitely hinder the recombination reaction between I 3 and excited electrons at the conduction band of nanocrystalline TiO 2 , thereby creating an increased photogenerated current density.
Scanning Kelvin Probe Microstructures are also employed to characterize the surface nanostructures of various CEs as shown in Fig. 7(a). Relatively uniform distribution of Co 20.85 Se 79.15 alloy CEs was observed, which is in agreement with the SEM. To verify the effect of alloy CEs on the internal transfer kinetics of DSSCs, the Nyquist and Bode phase curves (under light illumination) of DSSCs based on different CEs are displayed in Fig. 7(b,c) and the inset in Fig. 7(b) shows the equivalent circuit where R tr and R CE represent the charge-transfer  Table 4. Electrochemical parameters obtained from the EIS and Tafel polarization plots based on CE/ electrolyte/CE structures. www.nature.com/scientificreports www.nature.com/scientificreports/ resistances at the dye-sensitized TiO 2 /electrolyte interface and electrolyte/CE interfaces for DSSCs, respectively. The electrochemical parameters obtained by fitting the impedance spectra are summarized in Table 5. The DSSC based on the Co 20.85 Se 79.15 alloy CE exhibited a lower R tr (3.06 Ω cm -2 ), suggesting a fast I 3 to Ireduction kinetics at the electrolyte/CE interface, where the rapid accumulation of Iions can then facilitate their diffusion to the dye-sensitized TiO 2 photoanode/electrolyte interface. Furthermore, the photogenerated electrons in the nanocrystalline TiO 2 photoanode of the Co 20.85 Se 79.15 alloy DSSC showed a longer τ e value compared with those of other devices (Table 5). This result demonstrates that dye molecules are rapidly regenerated by Iions, thus enabling fast transport of photogenerated electrons in the mesoporous TiO 2 nanocrystal photoanode, resulting in the superior catalytic activity of the Co 20.85 Se 79.15 alloy CE. Furthermore, photovoltaic parameters and synthetic technology comparisons of the various Pt-free transition metal selenides CEs based DSSCs 34-39 were provided as shown in Table 6.

conclusion
In summary, tunable and Pt-free CEs based on binary alloys (Ru 81.09 Co 18.91 , Ru 80.55 Se 19.45 , and Co 20.85 Se 79.15 ) have been synthesized by a simple electrodeposition approach. The results indicate that Co 20.85 Se 79.15 alloy CEs possess outstanding electrocatalytic properties towards I 3 reduction, which can be attributed to their higher number of active sites, reduced interfacial resistance, and matched work function with the I 3 -/Iredox electrolyte. The Co 20.85 Se 79.15 alloy-based CE device displays a higher PCE of 7.08% compared with that of a pure Pt CE (5.80%) as well as preferable stability. Although the obtained alloy composition and performance could be further optimized, the easy synthesis method and hopeful efficiency indicate that electrochemical technologies have significant potential for the development of low-cost, high efficiency and stable DSSCs.

experimental Section
Synthesis of RuCo/RuSe/CoSe alloy and pristine Pt CEs. The RuCo, RuSe and CoSe alloys were synthesized by electrochemical co-electrodeposition on a cleaned fluorine-doped tin oxide (FTO, sheet resistance 12 Ω sq -1 , purchased from Sunlaite) glass substrate, using a galvanostatic approach on the electrochemical workstation. First, a solution (A: 3 mM RuCl 3 , 2 mM CoSO 4 ; B: 3 mM RuCl 3 , 2 mM SeO 2 ; C: 2 mM CoSO 4 , 3 mM SeO 2 ) and 100 mM LiCl were dispersed by ultrasonic waves for 30 min. Then, the deposition was carried out in a three-electrode system equipped with an FTO substrate (working electrode), a Pt electrode (counter electrode) and Ag/AgCl (reference electrode). The procedure was performed at a current density of 0.25 mA cm -2 for 600 s. Finally, the synthesized alloy CEs were rinsed with deionized water and dried at 80 °C in a vacuum furnace. For comparison, pristine metal CEs were also prepared with a 5 mM RuCl 3 , 5 mM CoSO 4 , 5 mM SeO 2 under the same conditions, respectively. The pristine Pt CE was also prepared by cyclic voltammetry in the range of -0.8~0.6 V by using 5 mM H 2 PtCl 6 solution, the scan parameter was controlled at 10 mV s -1 for 5 cycles.
Electrochemical characterization. A CHI760E electrochemical workstation equipped with a three-electrode device was utilized to assess the electrocatalytic performance of the prepared CEs. Cyclic voltammetry (CV) measurements were carried out in an auxiliary electrolyte consisting of 500 mM LiClO 4 , 10 mM I 2 and 50 mM LiI in acetonitrile at scan rates of 25, 50, 75, 100 and 125 mV s -1 , respectively. Electrochemical impedance spectroscopy (EIS) measurements were performed on symmetrical dummy solar cells with identical CE structures at frequencies ranging from 0.01 to 10 5 Hz and amplitude of 10 mV in air, respectively. Tafel polarization plots were also recorded on symmetrical cells at a scan rate of 10 mV s -1 . Then, EIS measurements of the DSSCs were performed with an amplitude of 10 mV, under sunlight illumination.

Fabrication of DSSCs.
A TiO 2 nanoparticle film-based on the FTO glass substrate was prepared according to the procedure described in our previous work. Specific as follows: a mixed solution of 20 mL ethanol and www.nature.com/scientificreports www.nature.com/scientificreports/ 8 mL tetrabutyl titanate was magnetically stirred for 30 min. Then, the acquired solution was added to a solution consisting of 5 mL deionized water and 40 mL acetic acid with magnetic stirring for 2 h. Afterward, the solution underwent a hydrothermal process at 230 °C for 12 h. Finally, the photoanode was obtained after washing the products and spin-coating TiO 2 nanoparticle on FTO glass. The thickness and active area of the photoanode film were controlled to 10 μm and 0.25 cm 2 , respectively. Then, the prepared TiO 2 photoanode was sensitized in a 0.50 mM N719 ethanol solution for 18 h. Afterward, the DSSCs were obtained by assembling the dye-sensitized TiO 2 photoanode and CEs with Surlyn tape, followed by injection of the     Table 6. Photovoltaic parameters comparisons of the reported Pt-free transition metal selenides counter electrode based-DSSCs.