A Biphase Tandem Redox Electrolyte for Efficient Quantum Dot‐Sensitized Solar Cell

A new biphase tandem redox electrolyte system is designed by immobilizing imidazolium iodide on Co, N‐doped carbon electrode in conjunction with polysulfide redox mediator in aqueous solution. Due to its biphasic feature, this redox system is enabled to separate the redox reactions in different phases, and hence, prevent the intertwined problem of redox reactions that occurred in one‐phase tandem redox system. Investigations show that the biphase redox system can efficiently accelerate electron reduction, and thereby inhibit charge recombination. Consequently, quantum dot (QD)‐sensitized solar cell fabricates using this biphase tandem redox together with CdSe QD‐sensitized photoanode achieved a high cell efficiency of 6.85% at 100 mW cm2 simulated AM 1.5, showing about 33% increment as compared to the traditional pure polysulfide‐based device (5.14%). The encouraging results indicate that the utilization of biphase tandem electrolytes can be an efficient way to improve the performance of sensitized solar cells.


Introduction
The development of cost-effective and stable photovoltaics to realize the eco-friendly utilization of solar energy has been recognized to be one of the most efficient ways to alleviate energy and ecological problems.Among those photovoltaics, the use of semiconductor quantum dots (QDs) with a narrow bandgap as a photosensitizer to develop QD-sensitized solar cells (QDSSCs) has attracted considerable interest as the lowcost third-generation photovoltaics, due to the superior intrinsic properties of QDs of high molar extinction coefficients, easily tunable bandgaps, large intrinsic dipole moments as well as possible multiple carrier generation. [1]To date, various strategies have been applied to enhance the cell performance of QDSSCs, which have led to a significant improvement in the overall energy conversion efficiency from less than 1% to over 16%. [2]Nevertheless, since the cell efficiency of QDSSC is still far behind its theoretical thermodynamic efficiency of QDSSC (%44%), challenges to enhance the cell efficiencies of QDSSC remain.
Generally, a typical QDSSC consists of a quantum dotsensitized photoanode, a counter electrode (CE), and redox mediator electrolyte, being arranged into sandwich structure.When light is irradiated, QD sensitizer is excited to produce electron-hole pair, which is followed by electron injection from QD to the large bandgap semiconductor (e.g., TiO 2 ).After that, the oxidized QD is regenerated by electron donation from redox mediator in electrolyte, while the produced oxidized form of mediator is then transported to CE and further reduced at the CE to complete the electric circuit.This indicates that the redox electrolyte is closely related to the charge transportation between photoanode and CE, QD regeneration at photoanode as well as the catalytic reduction reaction at CE, and thus plays a very important role in determining the performance of the resultant device.Therefore, much effort has been devoted to obtain an appreciable electrolyte system for improving the performance of QDSSC. [3]Although various redox mediators including [(CH 3 ) 4 N] 2 S/[(CH 3 ) 4 N] 2 S n , ferricyanide/ferrocyanide, Co 2þ /Co 3þ , Fe 2þ /Fe 3þ , etc., have been investigated, S 2À /S n 2À redox couple is acknowledged to be the most suitable redox mediator for QDSSC especially in liquid electrolyte, due to its good compatibility with QD sensitizers, superior hole-extraction ability from QDs and remarkable cell performance. [4]To further improve the performance of QDSSC, modification of the S 2À /S n 2À electrolyte by introducing additives was proposed. [5]he role of additives is mainly ascribed to their advantages for inhibiting charge recombination at photoanode/electrolyte interfaces due to the adsorption of additives on the QD-sensitized photoanode. [6]In addition, the additive has also been found to influence the charge transfer as well as the reduction reaction at CE.For instance, by adding different amines to polysulfide electrolyte, Beiraghdar et al. found that the charge transportation in the polysulfide electrolyte is improved, which contributes to the enhancement of the cell efficiency of CdS/CdSe QDSSC. [7]un et al. reported that the presence of polyethylene glycol additives in polysulfide electrolyte facilitates the catalytic reduction reaction of the redox couple and promotes the charge transfer from CE to electrolyte. [8]Dissanayake et al. reported that the addition of PbS QDs to the polysulfide electrolyte increased sulfide ion (S 2À ) conductivity due to the indirect ionic dissociation facilitated by PbS QDs. [9]n contrast, researches in dye sensitized solar cell (DSSC) have demonstrated that the use of different redox mediators to form tandem redox system can be an efficient strategy to enhance the performance of DSSC. [10]The benefit role of tandem redox system is ascribed to the synergistic effect of different redox mediators, being supposed to operate at photoanode and counter separately that caused the improved charge transfer and the suppressed charge recombination of the device.For instance, Cong et al. fabricated a 2,2,6,6-tetra-methyl-1-piperidinyloxy (TEMPO)-Co tandem redox system-based DSSC using TEMPO and Co(bpy) 3 2þ/3þ as the redox couples and achieved higher cell efficiency compared to the single Co-complex-based device. [11]ao et al. reported a tandem redox-based DSSC by adding tris(4-methoxyphenyl)amine (TPAA) into cobalt tris(bipyridine), which exhibits a 26% efficiency improvement compared with the cobalt tris(bipyridine) electrolyte. [12]On variation of the length of the substituted alkoxy chains of tris(4-alkoxyphenyl)-amine, they achieved a power conversion efficiency (PCE) up to 11.0%, showing an improvement of up to 50% as compared to the cells based on pure tris(bipyridine)cobalt-based electrolytes. [13]Very recently, Ayaz et al. reported that the addition of redox intermediate tris(p-anisyl)amine (TPAA) to an iodide/triiodide electrolyte improved the dye regeneration, and thus enhanced its cell efficiency. [14]Nevertheless, since the presence of multiple species in the same solution, the application of the so-called "tandem" electrolyte may lead to the intertwined problem of redox reactions, which is detrimental to the performance of the resultant device.
Here, we designed a new biphase tandem redox system by immobilizing imidazolium iodide on a Co, N-doped carbon electrode in conjunction with a well-known polysulfide dissolved in aqueous solution.As the redox species are confined in different phases, this biphase tandem redox system is much different from the former reported one-phase tandem electrolyte in DSSC, which can realize the separation of redox reactions in different phases and hence prevent the intertwined problem of redox reactions.Electrochemical investigations show that this biphase redox system can efficiently accelerate electron reduction, resulting in the inhibition of the charge recombination.While it was used to fabricate QDSSC together with CdSe QD-sensitized TiO 2 , a high cell efficiency of 6.85% at 100 mW cm 2 simulated AM 1.5 conditions was achieved, showing about 33% increment as compared to the traditional pure polysulfide based device (5.14%).

Results and Discussion
Co, N-doped carbon (Co@N-C) was prepared by carbonizing ZIF-67 at 800 °C in N 2 as described in the Experimental Section.After being deposited on Ti mesh to form a film electrode (Co@N-C/Ti), the film was further immersed in a 1-hexyl-3-methylimidazolium iodide (HMII) for the immobilization of HMII. Figure 1 shows the scanning electron microscopy (SEM) images of Co@N-C/Ti before and after HMII immobilization.As can be seen, the Co@N-C/Ti film is dense with smooth surface, but some small cracks are visible (Figure 1A).The enlarged SEM image shows that the film is of porous structure, composed of small particles having dodecahedral structure (Figure 1B).Upon the immobilization of HMII, the surface of the film becomes smoother and the small cracks in the film are considerably reduced, but its porous structure is still retained (Figure 1C,D).Elemental mapping analyses indicate that except the uniform distributed C, N, O, and Co, an additional I element was detected after the immobilization of HMII (Figure S1, Supporting Information), suggesting that HMII is immobilized on Co@N-C/Ti film.
The chemical composition of Co@N-C before and after HMII immobilization was analyzed by X-ray photoelectron spectroscopy (XPS).The survey spectra shown in Figure 2A demonstrate the coexistence of C, N, O, and Co elements.N 1s high-resolution XPS spectrum exhibits five peaks at 398. 5, 399.2, 400.5, 401.4,and 404.3 eV, which are assigned to pyridinic-N, Co-N, pyrrolic-N, graphitic-N, and oxidized-N, respectively (Figure S2, Supporting Information). [15]The Co 2p spectrum can be deconvoluted into three peaks corresponding to metallic Co (778.5 eV), CoO x (780.4 eV), and CoN x (781.5 eV) (Figure S2, Supporting Information). [16]These results together with the elemental mapping (Figure S1, Supporting Information), confirmed the formation of Co, N-doped carbon material.Compared to the pristine Co@N-C, we found that the immobilization of HMII leads to the appearance of additional new peaks at 619.0 and 630.6 eV, which is assigned to I 3d signal (Figure 2A). [17]This result together with the elemental mapping analyses (Figure S1, Supporting Information) confirmed the binding of HMII on Co@N-C.Figure 2B shows the high-resolution I 3d spectrum for the HMII immobilized sample with the peak positions indicating the different species identified by spectrum deconvolution.As demonstrated, the I 3d is composed of two pair of peaks centered at 632.3 eV, 621.1 eV, and 630.1 eV, 618.6 eV, being assigned to I 3d5/2, I 3d3/2 of I 2 , and I À , respectively. [18]The formation of I 2 is ascribed to the oxidation of I À by air during the immobilization process as it is processed in an ambient atmosphere.
The Co@N-C/Ti before and after HMII immobilization was employed as the CEs to construct QDSSC using CdSe QD-sensitized TiO 2 as the photoanode and sulfide/polysulfide (S 2À /S x 2À ) aqueous solution containing SiO 2 nanoparticles as the liquid electrolyte, and their photovoltaic performances were characterized by photocurrent density-photovoltage ( J-V ) measurements.Figure 3 shows the typical J-V curves of the QDSSC with Co@N-C/Ti before and after the immobilization of HMII, and their photovoltage (V OC ), photocurrent ( J SC ), fill factor (FF), and solar energy conversion efficiency are summarized in Table 1.For comparison, the cell performance of CdSe QDSSC using PbS as the CE was also included.The cell with bare Co@N-C/Ti achieved a PCE of 5.14%, which is comparable to the reference cell with PbS CE (η = 5.38%).It is worth noting that upon the immobilization of HMII, both J SC and V OC are significantly increased, which leads to a large increase in PCE from 5.14% to 6.85%, showing ca.33% increment.In controlled experiments, imidazole derivatives with different counter ions of 1-hexyl-3-methylimidazole chloride (HMIC) and 1-hexyl-3methylimidazole trifluoromethanesulfonate (HMIS) were also immobilized on Co@N-C/Ti instead of HMII to fabricate CdSe QDSSC.Cell performance investigations show that in contrast to HMII, the immobilization of HMIC or HMIS does not lead to obvious improvement in the cell efficiency (Figure S3, Supporting Information).Also, we found that the treatment of Co@N-C/Ti electrode with I À and/or I 2 instead of HMII does not lead to the improvement of the cell performance of the device.These results indicate that the presence of HMII is crucial for the performance enhancement of QDSSC.
Electrochemical impedance spectroscopy (EIS) was performed to study the charge transfer processes in QDSSCs.Figure 4A shows the Nyquist plots of Co@N-C/Ti before and after HMII immobilization, while the EIS measurements were carried out using a symmetric dummy cell consisting of two identical electrodes.By fitting the EIS data with the equivalent circuit shown in the inset of Figure 4A, the interfacial charge transfer resistances (R ct ) were obtained, where R s is the equivalent series resistance, R ct1 , CPE1 and R ct2 , CPE2 represent the charge transfer resistance and chemical capacitance at the interfaces of  Table 1.Photovoltaic parameters of CdSe QDSSC with Co@N-C/Ti before Co@N-C and after HMII immobilization (HMII-Co@N-C), and PbS as the CE.Co@N-C/Ti and CE/electrolyte.Due to the high conductivities of carbon material and Ti metal, both electrodes give similar small R s and R ct1 (Table 2).Compared to the pristine Co@N-C/Ti electrode, however, we noted that the immobilization of HMII significantly decreased R ct2 .As R ct2 is inversely proportional to the charge transfer at CE/electrolyte, the lower R ct2 indicates that the electron reduction at CE is largely improved after the immobilization of HMII.
Figure 4B shows the Tafel polarization plots of Co@N-C/Ti electrode before and after HMII immobilization using a symmetric dummy cell similar to EIS investigations.From the intersection of the cathodic branch and the equilibrium potential, we obtained the exchange current density ( J 0 ), which gives a larger value for the HMII immobilized electrode compared to the bare Co@N-C/Ti electrode.Since J 0 is inversely proportional to the charge transfer resistance (R ct ) at the electrode by J 0 ¼ RT=nFR ct , where R and F represent the gas constant and Faraday constant, T is the absolute temperature, n is the charge transfer number of the reduction reaction at the electrode, the larger J 0 of HMII immobilized Co@N-C/Ti electrode indicates that the immobilization of HMII is beneficial for accelerating the electron reduction reaction at the electrode, in good agreement with the EIS investigations, as shown in Figure 4A.As a result, the improved electron reduction rate leads to the enhanced cell performance of QDSSC.
The above results indicate that the immobilization of HMII on Co@N-C/Ti electrode can accelerate the electron reduction at CE, which results in the enhancement of the cell performance of CdSe QDSSC.The beneficial role of HMII immobilization is ascribable to the formation of biphase tandem redox that improved electron transfer processes in QDSSC.Upon light illumination, CdSe is excited to produce an electron-hole pair.Followed by the injection of electrons from CdSe QDs into the conduction band of TiO 2 , the oxidized CdSe QDs are regenerated by S 2À in the aqueous electrolyte to produce S 2À n , whereas electrons are transferred to CE to induce a reduction reaction.When a bare Co@N-C/Ti electrode is used, S 2À n in aqueous electrolyte is reduced to S 2À at the electrode.While HMIIimmobilized Co@N-C/Ti electrode is used, however, the immobilized I 2 at CE is supposed to be reduced first to produce •I À 2 , which is then regenerated to I 2 via an electron exchange with oxidized species of S 2À n in polysulfide electrolyte, as shown in Figure 5. Due to its faster electron reduction of I 2 at CE and electron exchange between S 2À n and •I À 2 , the electron reduction rate is largely increased, and thus enhanced the photocurrent of CdSe QDSSC.In contrast, the increased electron reduction can further accelerate the hole transfer across QD/electrolyte interface and minimize the electron recombination at the photoanode, which leads to the improvement of the photovoltage.It needs to be pointed out that, the biphase tandem electrolyte system in our case is formed by confining the two pairs of redox couples in separated phases, which is much different from the previously reported one-phase "tandem" electrolyte in DSSC.The biphasic feature of our tandem electrolyte can separate the redox reactions in different phases, and hence prevent the intertwined problem   of redox reactions that is more favorable for enhancing the cell performance of the device.Furthermore, the stability of the HMII immobilized Co@N-C/Ti is evaluated by repeated cyclic voltammetry (CV) measurements.Figure 6 shows the CV curves of HMII immobilized Co@N-C/Ti in S 2À n =S 2À aqueous electrolyte under reduplicate CV measurement at a scan rate of 100 mV s À1 .As demonstrated, one can find that the CV curve remains almost unchanged during the continuous 200 cycles scanning.This indicates that the immobilization of HMII on Co@N-C/Ti is of high stability in polysulfide electrolytes.

Conclusion
In summary, we have designed a new tandem redox electrolyte system by confining two redox couples in different phases, being formed by immobilizing iodide-based redox couple in Co@N-C/Ti electrode together with a polysulfide in aqueous solution.While this tandem redox system was used to fabricate QDSSC, a high cell efficiency of 6.85% at 100 mW cm 2 simulated AM 1.5 was achieved, showing about a 33% increment as compared to the traditional pure polysulfide-based device (5.14%).The enhanced cell performance of QDSSC is ascribed to its beneficial role in the tandem electrolyte in improving electron reduction that enables the decrease of charge recombination.These results demonstrate that the formation of a biphase tandem electrolyte system is an advantage for enhancing the electron reduction at CE, and thus provides a way to construct a sensitized solar cell with enhanced solar-to-energy conversion efficiency.

Experimental Section
Preparation of Co, N-Doped Carbon Material: ZIF67 nanocrystals were synthesized as follows: typically, 5.82 g of Co(NO 3 ) 2 •6H 2 O and 13.14 g of 2-methyl imidazole were separately dissolved in 300 mL methanol.The two solutions were mixed together under vigorous stirring to form blue-purple precipitates.The precipitates were collected by centrifugation and thoroughly washed with methanol, followed by drying at 60 °C to obtain ZIF67 nanocrystals.The obtained ZIF67 was then put into a tubular furnace and carbonized under N 2 atmosphere at 800 °C for 3 h at a heating rate of 10 °C min À1 .After cooling to room temperature naturally, Co, N-doped carbon material was obtained, denoted as Co@N-C.
Preparation of CE: CEs were prepared by coating carbon material on Ti mesh.First, terpineol, titanium-isopropoxide, and ethyl cellulose were mixed to form a light yellow adhesive mixture.Then, Co@N-C was milled with the adhesive mixture to get a sticky carbon slurry paste.The slurry was deposited on a clean Ti mesh through a drop-coating method by dropping the carbon slurry paste onto Ti mesh.After drying at 120 °C, the electrode was calcined at 450 °C for 30 min under N 2 .
Immobilization of HMII: The immobilization of HMII was carried out by immersing Co@N-C/Ti into 1-hexyl-3-methylimidazolium iodide, which was followed by sequentially rinsed with distilled water, ethanol and dried in air, denoted as HMII-Co@N-C.
Fabrication of Quantum Dot Solar Cells (QDSSCs): CdSe QD-sensitized TiO 2 film electrode was prepared through a linker-assisted binding approach by immersing TiO 2 film in CdSe QD solution.QDSSCs were assembled into a sandwich structure with a CdSe-sensitized TiO 2 as the working electrode, a PbS as the CE, and polysulfide aqueous solution composed of 2.0 M S, 2.0 M Na 2 S, and 0.2 M KCl with 5% SiO 2 (Aladdin, %7 nm) as the electrolyte.
Characterizations: SEM was performed using a Hitachi S-4800F, while energy dispersive X-ray spectrometer was fitted to electron microscopes for elemental analysis.XPS analyses were performed on an X-ray photoelectron spectrometer (VG Scientific ESCALab 250-XI).The spectrometer was calibrated using the C 1s (284.8 eV) signal.Photocurrent-voltage ( J-V ) measurements were performed with an electrochemical interface system (Solartron SI1287) under the illumination of simulated sunlight (AM1.5, 100 mW cm À2 ).EIS was carried out using a symmetric dummy cell consisting of two identical electrodes.CV was performed in polysulfide aqueous solution on CHI660e electrochemical workstation with Pt wire as an auxiliary electrode and Ag/AgCl as a reference electrode.

Figure 1 .
Figure 1.SEM images of Co@N-C/Ti before A,B) and after HMII immobilization C,D).

Figure 2 .
Figure 2. XPS spectra of Co@N-C/Ti before (Co@N-C) and after HMII immobilization (HMII-Co@N-C).A) Survey spectra and B) high-resolution of I 3d with peak positions indicating different species identified by spectrum deconvolution.

Figure 3 .
Figure 3. J-V curves of CdSe QDSSC with Co@N-C/Ti before a) and after HMII immobilization c), and PbS b) as the CE.

Figure 4 .
Figure 4. Electrochemical impedance A) and Tafel polarization curves B) of Co@N-C/Ti before a) and after HMII immobilization b).

Figure 5 .
Figure 5. Schematic description of the electron transfer processes of QDSSC with biphase tandem redox electrolyte.

Figure 6 .
Figure 6.CV curves (200 cycles) of HMII immobilized Co@N-C/Ti in polysulfide electrolyte at a scan rate of 100 mV s À1 .

Table 2 .
Electrochemical parameters determined from electrochemical impedance spectra shown in Figure4A.