Synthesis and electrochemical properties of

Journal of Electroanalytical Chemistry 787 (2017) 36–45

Contents lists available at ScienceDirect

Journal of Electroanalytical Chemistry journal homepage: www.elsevier.com/locate/jelechem

Synthesis and electrochemical properties of potassium 5-trifluoromethyl-1,3,4-thiadiazole-2-thiolate/disulfide redox couple Grégory Hersant a,1, Amer Hammami a,2, Michel Armand b,3, Benoît Marsan a,⁎ a

Département de Chimie, Université du Québec à Montréal, C.P. 8888, Succ. Centre-ville, Montréal, QC, H3C 3P8, Canada Laboratoire International sur les Matériaux Électroactifs, CNRS-UdM, UMR 2289, Département de Chimie, Université de Montréal, C.P. 6128, Succursale Centre-ville, Montréal, QC, H3C 3J7, Canada

b

a r t i c l e

i n f o

Article history: Received 3 September 2016 Received in revised form 16 December 2016 Accepted 2 January 2017 Available online 3 January 2017 Keywords: 5-Trifluoromethyl-1,3,4-thiadiazole-2-thiolate Disulfide Redox couple EMITFSI-based electrolyte Cyclic voltammetry ITO-CoS

a b s t r a c t A new redox couple, composed of the reduced species potassium 5-trifluoromethyl-1,3,4-thiadiazole-2-thiolate 2 and the oxidized species 5,5′-bis (2-trifluoromethyl-1,3,4-thidiazole) disulfide 3, was synthesized and characterized chemically and electrochemically when dissolved in the ionic liquid 1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl) imide (EMITFSI). Platinum (Pt), glassy carbon (GC) and cobalt sulfide-coated indium tin oxide glass (ITO-CoS) electrodes were used. The diffusion-controlled redox processes were shown to be electrochemically irreversible. At higher electrolyte temperature, the current densities are enhanced, and the potential difference between the anodic and cathodic peaks is decreased, in conjunction with a decrease in the viscosity of the electrolytic medium. The best electrocatalytic performance was obtained with the ITO-CoS electrode, followed by GC and Pt electrodes. A greater thiolate/disulfide ratio favors larger anodic current densities, whereas a smaller ratio leads to higher cathodic current densities. Preliminary stability studies of the ITO-CoS electrode confirmed the very good electrochemical and mechanical stability of this electrode considered as a serious candidate to replace platinum as the counter electrode in dye-sensitized solar cells (DSSCs). The electrolytic media − studied in this work exhibits a very low absorption of visible light, contrarily to the conventional I− 3 /I mediator, offering the possibility of illumination of the device via the counter electrode. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Solar energy is a clean, free and renewable energy that must, however, be harvested and converted. One possibility is to convert directly solar energy into electricity by using a dye-sensitized solar cell (DSSC), which has been developed by Grätzel and his team in 1991 [1]. This type of cell uses a wide band gap n-type semiconductor, such as TiO2 (3.2 eV for the anatase phase). To increase the visible light absorptivity of the mesoporous material, one monolayer of dye molecules, generally consisting of a ruthenium complex, are adsorbed onto the surface of the TiO2 particles. Upon illumination, electrons of the dye molecules are ⁎ Corresponding author. E-mail addresses: [email protected] (G. Hersant), [email protected] (A. Hammami), [email protected] (M. Armand), [email protected] (B. Marsan). 1 Present address: Vice-décanat à la Recherche, Faculté des Sciences et de Génie, Pavillon Alexandre-Vachon, Université Laval, Sainte-Foy, QC, G1V 0A6, Canada. 2 Present address: Groupe Kemplus, 6020 rue Vanden Abeele, St-Laurent, QC, H4S 1R9, Canada. 3 Present address: CICenergigune, Parque Tecnologico de Alava, Albert Einstein 48, Ed, CIC, 01510 Minano, Spain.

http://dx.doi.org/10.1016/j.jelechem.2017.01.008 1572-6657/© 2017 Elsevier B.V. All rights reserved.

promoted from their ground state S to an excited state S*. Subsequently, a very fast charge transfer (10−15 s) occurs between the excited state and the conduction band of TiO2, leaving the dye in its oxidized state (S* → S+ + e−). Dye molecules are regenerated by electron transfer from the reduced species of a redox couple (mediator), generally the − + 2S+ → I− I−/I− 3 couple: 3I 3 + 2S. The electrons injected into the TiO2 conduction band reach the external circuit, via an ohmic contact, to the counter electrode where they reduce the oxidized species of the − redox couple (I− 3 ) to regenerate the reduced species (I ). DSSCs suffer from several drawbacks. First, oxidized species might be reduced by the electrons injected into the TiO2 conduction band (back reaction process) [2]. Second, dye molecules may degrade under − redox couple sustained illumination conditions [3]. Third, the I− 3 /I may react with the platinum counter electrode (to form PtI4) [4] and corrode silver electrical contacts [5]. This redox couple has a reddish dark brown color, which is responsible for the strong absorption of visible light by the electrolyte, and forces illumination of the device through the photoanode with reduction of the global conversion efficiency. A significant driving force (~0.6 V) exists for the regeneration of dye molecules, which constitutes an energy loss: the redox potential of the sensitizer cation and of the I−/I− 3 mediator are ~1.0 V and ~0.4 V

G. Hersant et al. / Journal of Electroanalytical Chemistry 787 (2017) 36–45 − vs. NHE, respectively [5]. Finally, due to the use of the I− 3 /I mediator, the maximum photovoltage reached in conventional DSSCs is ~ 0.7– 0.8 V. Another type of solar cell, called electrochemical photovoltaic cell (EPC), consists of an n-type (photoanode) and/or a p-type (photocathode) semiconducting material in contact with an electrolytic medium, which can be solid (polymer), gel or liquid, containing one redox couple. The electrolytic medium is also in contact with an auxiliary electrode (anode or cathode) which is generally a conductive glass substrate (indium tin oxide, ITO, or fluorine-doped tin oxide, FTO) coated with a catalyst. The operating principle of an EPC primarily involves the illumination of the semiconductor via the transparent auxiliary electrode. If the energy of light radiation (hν) is greater or equal to the semiconductor band gap energy (Eg), electron (e−)-hole (h+) pairs are generated. The photogenerated holes (valence band) of an n-type semiconductor will migrate to the semiconductor∣electrolyte interface and oxidize the reduced species of the redox couple present in the vicinity of the electrode (photo-oxidation: A + h+ → A+). After being promoted to the semiconductor conduction band level, the electrons will reach the external circuit via an ohmic contact to the counter electrode where they will reduce the oxidized species that have migrated: A+ + e− → A. Gerischer and Gobrecht [6] have made the first demonstration of the conversion of light energy into electricity using an EPC; they employed ∣SnO2. Howevthe following configuration: n-CdSe∣aqueous Fe(CN)3−/4− 6 er, it was shown that the low-band gap semiconductor∣aqueous electrolytic medium junction generates a phenomenon called photocorrosion (dissolution of the semiconductor, leading to cell instability) [7,8]; the use of an organic solvent was then considered. However, the volatility of such solvents, a particular problem under illumination (increase of the cell temperature), produces an increase of pressure that may lead to mechanical failure of the device. On the other hand, organic solvents can dissolve the adhesive material used to seal the cells. Another option is to use polymer electrolytes composed of a polymer matrix complexing an alkali metal salt. The nature of these conducting materials has been highlighted by Wright in the seventies [9]. Vijh and Marsan [10] were the first to use a solid polymer electrolyte in contact with a polycrystalline semiconductor; the configuration of the solar cell was n-CdSe∣ modified PEOy-M2S/xS ∣ ITO, where M = Li, Na or K, y = 16, 30 or 60 and x = 1, 3, 5 or 7. Inorganic polysulfides were employed because they may undergo preferential photo-oxidation and thereby protect n-CdSe from photocorrosion. Modified PEO represents a co-polymer based on polyethylene oxide, (CH2CH2O)n, and the y index indicates that the polymer electrolyte has y monomer units (CH2CH2O) per M2S molecule. The weak electrolyte ionic conductivity (~10−6 S cm−1 at 50 °C) and the corresponding low cell energy conversion efficiency have encouraged Philias and Marsan to pursue the research with a new cell configuration, n-CdSe∣modified PEO12-CsT/0.1T2 ∣ITO, where T− stands for 5mercapto-1-methyltetrazole ion and T2 is the corresponding disulfide form [11]. The addition of T2 (oxidized species) improves the kinetics of the T−/T2 redox couple, especially for the reduction process. The choice of a cyclic and unsaturated redox couple, instead of an inorganic polysulfide, was motivated by the stabilization of the reduced species through delocalization of the negative charge. This situation promotes dissociation of the CsT salt in aprotic and polymeric media, and the reduction process leading to a more positive redox potential (potentially enhancing the device photovoltage). Nevertheless, the difficulty of establishing an effective semiconductor ∣ polymer electrolyte junction, particularly when the semiconductor is very porous, has prompted researchers to consider alternatives such as gel electrolytes. These systems, which result from the incorporation of a liquid electrolyte into a polymer matrix, have been described for the first time by Feuillade and Perche [12]. Renard et al. [13] have developed a series of gel electrolytes with the following optimal composition: PVdF (24%) and DMSO/DMF (40/60% v/v) (76%)/1.34 M CsT/0.13 M T2. 24% represents the mass ratio between the polymer and the polymer-solvent

37

mixture. PVdF is polyvinylidene fluoride, (\\CH2\\CF2\\)n, DMSO is dimethyl sulfoxide and DMF is dimethylformamide. This gel shows an interesting ionic conductivity of 7.1 × 10−3 S cm−1 at 25 °C. However, under extended illumination, drying of the gel was observed. The use of highly conductive and low volatility liquid electrolytes, such as ionic liquids, is of great interest. Indeed, ionic liquids possess a high concentration of ions, they are liquid over a wide temperature range and with a negligible vapor pressure [14,15]. Among them, salts formed by an imidazolium cation and an organic anion are chemically very stable [16,17]. A large number of ionic liquids based on the 1-ethyl-3methylimidazolium (EMI+) cation present a large electrochemical stability window (N 4 V) and a high ionic conductivity at 25 °C (N10− 3 S cm− 1). The imidazolium cation coupled with the bis(trifluoromethanesulfonyl)imide (TFSI−) counter ion has been used in solar cell devices [18] and Li-ion batteries [19]. The use of aromatic or ring unsaturated derivatives compounds different from CsT and T2 can be considered in EPCs. As an example, 2mercapto-5-methyl-1,3,4-thiadiazole (McMT) derivatives have been the subject of redox mechanistic studies in aprotic polar solvent (CH3CN). The work of Shouji and Buttry [20] was done in the field of secondary lithium batteries which have similarities with EPC problematics. Later, Hammami et al. [21] showed that thiolate/disulfide redox couples like 2-mercapto-1,3,4-thiadiazole derivatives, for instance, could be good candidates for application in EPCs or DSSCs. In 2010, Wang et al. [22] presented the first example of a triiodide/iodide-free redox couple based on a thiolate/disulfide redox couple (1-methyl1H-tetrazole-5-thiol tetramethylammonium salt and its dimer form) for application in a DSSC device. Using this iodide-free redox electrolyte in conjunction with a sensitized heterojunction has allowed reaching an unprecedented efficiency of 6.4% under standard illumination test conditions. Noteworthy, the work of Tian et al. [23], who have employed the 2-mercapto-5-methyl-l,3,4-thiadiazole ion (McMT−)/disulfide (BMT) redox couple, demonstrated again the great potential of thiolate/disulfide redox couples in DSSCs. The EPC developed in our laboratory uses a mixture of aromatic or unsaturated thiolate and disulfide as redox couple. Photo-oxidation of the thiolate species into disulfide takes place at the n-type semiconductor electrode while the reverse reaction occurs at the cathode. In order to improve the efficiency of the electroreduction process, which limits the cell performance, Marsan and Bourguignon [24] have developed a catalyst based on a new methodology for cobalt sulfide electrodeposition onto an ITO substrate. Burschka et al. [25] have reported the replacement of platinum by CoS as the counter electrode in DSSCs which showed the great potential of this catalyst for DSSCs working with a thiolate/disulfide redox couple; the electrode stability is still to be improved. The present work reports the study of the electrolyte based on a thiolate/disulfide redox couple derived from 5-trifluoromethyl1,3,4-thiadiazole-2(3H)-thione (TfmTT) 1 dissolved in an ionic liquid, 1-ethyl-3-methylimidazolium bis(trifluoromethanesulfonyl) imide (EMITFSI). In the first part of this study, the synthesis and the chemical characterization of 1, potassium 5-trifluoromethyl1,3,4-thiadiazole-2-thiolate (TfmTT− K+) 2 and its disulfide form, 5,5′-bis(2-trifluoromethyl-1,3,4-thiadiazole) disulfide (BTfmT) 3, are presented. A potassium salt has been preferred over a cesium salt because it is less expensive. Fig. 1 shows the structure of the three compounds. Contrary to Tian et al. [23], 2-mercapto-1,3,4-thiadiazole derivatives (salt and disulfide) presented in this study were substituted with a trifluoromethyl group instead of a methyl group. This functional group may strongly influence the redox processes due to inductive effect on the thiadiazole ring. For instance, Antonello et al. [26] have shown the influence of a series of electron-donating and -withdrawing substituents on the reduction of para-substituted diaryl disulfides derived from 1-mercaptobenzene. One major influence of the electron-withdrawing groups, such as NO2, CN and F, is on the cathodic peak potential

38

G. Hersant et al. / Journal of Electroanalytical Chemistry 787 (2017) 36–45

The synthesis of the potassium salt 2 was carried out using potassium carbonate in methanol:

Fig. 1. Structure of compounds 1, 2 and 3.

which is shifted to more positive values compared to non-substituted or electron-donating (OCH3, NH2) aromatic disulfides. In the second part of this paper, solution of 2 (TfmTT− K+) in EMTFSI, and solution of 2 and 3 (TfmTT−/BTfmT) in EMTFSI, have been studied by cyclic voltammetry at platinum (Pt), glassy carbon (GC) and ITO-CoS electrodes. Platinum is the conventional cathode used in DSSCs whereas carbon electrodes show interesting kinetics [27]. Glassy carbon is widely used as an electrode material; McCreery has published interesting studies on carbon electrodes and their properties [28]. The TfmTT−/BTfmT ratio and temperature of the electrolytic medium were varied in the present study. Finally, UV–vis spectroscopy measurements have been carried out to compare the absorbance, in the visible part of the solar spectrum, of the electrolytic medium with that of a I−/I− 3 solution in EMITFSI.

2. Experimental 2.1. Chemicals for synthesis of the redox molecules and ionic liquid 2-amino-5-trifluoromethyl-1,3,4-thiadiazole, 99.3% from Alfa Aesar, was the starting material for the synthesis of 1, 2 and 3. Lithium sulfide, Li2S, 99.98% from Aldrich, sodium nitrite, NaNO2, 98.4% from Anachemia, and HBr, 48% (p/p) in H2O from Aldrich, were used for the synthesis of 1. Potassium carbonate, K2CO3, ACS grade 98% from Acros Organics, was used for the synthesis of 2. Iodine, I2, 99.99% from Aldrich, was used to oxidize 2 into 3. Sodium thiosulfate (Na2S2O3), ACS grade 98% from Anachemia, sodium chloride (NaCl), N 99% from Alfa Aesar, and magnesium sulfate (MgSO4), 99% from Anachemia, were used for the work-up of the reactions. Dichloromethane (CH2Cl2), ACS grade 99.5% from Anachemia, methanol (MeOH), ACS grade 99.5% from Anachemia, and ether (CH3CH2O), grade ACS 99.0% from Aldrich, were used for the different syntheses. Bromoethane, ≥99% from Aldrich, N-1-methyl imidazole, 99% from Aldrich, and lithium bis(trifluoromethanesulfonyl) imide, LiTFSI from 3M, were used to synthesize EMITFSI (see Supporting information for the procedure). EMII, 97% from Aldrich, was used for the absorbance measurements. All the chemicals were used as received, without any further purification.

2.2. Synthesis of 1, 2 and 3 The synthesis of 5-trifluoromethyl-1,3,4-thiadiazole-2(3H)thione 1 involved, in a first step, the formation of the 2-bromo-5trifluoromethyl-1,3,4-thiadiazole intermediate in nanopure water via the diazonium salt of 2-amino-5-trifluoromethyl-1,3,4thiadiazole. Then, substitution of the bromide function by a thiol function was done using an excess of lithium sulfide (3 eq) in ether:

The disulfide form of 2 was prepared by oxidation of the potassium salt with iodine in water:

The detailed synthesis procedures and chemical characterizations are given in the Supporting Information. 2.3. Cyclic voltammetry measurements Cyclic voltammetry (CV) measurements were made with a threeelectrode electrochemical cell under positive argon atmosphere (grade 4.8 from Praxair) in a glove box, using a Solartron multipotentiostat (model 1470) interfaced with a PC; the softwares CorrWare (version 2.80) and CorrView (version 2.70) from Scribner Associates were used to run the analyses and to collect the data. The working electrode was a platinum electrode (Bioanalytical Systems MF-2013; surface area: 0.025 cm2), a glassy carbon electrode (Bioanalytical Systems MF-2012; surface area: 0.071 cm2) or an ITO-CoS electrode (surface area: 0.25 cm2). Pre-treatment of platinum and glassy carbon electrodes, and the procedure to prepare the ITO-CoS electrode are described in the Supporting Information. The working electrode was set at open-circuit potential for 10 min prior to the CV measurements. The auxiliary electrode (1 cm2) was a platinum foil (99.9%, 0.1 mm thick, Aldrich) previously passed through a flame, rinsed with nanopure water and dried for 15 min in a vacuum desiccator. A silver wire (99.9%, Acros Organics), mechanically cleaned with an emery paper and rinsed with nanopure water, was used as a pseudo-reference electrode. Cyclic voltammetry studies were performed on solutions of 2 (75 mM) in EMITFSI, solutions of 3 (25 mM) in EMITFSI, mixtures of 2 (75 mM) and 3 (25 mM) in EMITFSI, and mixture of 2 (83.3 mM) and 3 (16.7 mM) in EMITFSI at 25 °C unless otherwise stated. All solutions were degassed for 10 min with argon prior to the measurements and kept under inert atmosphere throughout the experiments. The maximum potential range applied to the working electrode was from 0.6 V to 1.3 V vs. Ag; the scan rate was 50 mV s− 1 unless otherwise stated. 2.4. Ionic conductivity measurements Electrolyte ionic conductivities were obtained using electrochemical impedance spectroscopy and a small volume (b 1 mL) thermostated conductivity cell (WTW, D82362 Weilheim). This cell uses two platinized platinum electrodes between which the solution resistance was measured. Cell constant (10.3–10.5 cm−1) was determined periodically with a standard KCl solution (Alfa Aesar, 0.117 M, batch no. F18S025, 0.015 S cm−1 at 25 °C) to ensure accurate values. A thermostated bath was employed to control the solution temperature (programmable temperature controller, Polysciences model 9112); a stabilization time of 15 min was used before each measurement. Measurements were made at 25 °C for solution of 2 (75 mM) and 3 (25 mM) in EMITFSI, and at 25, 50 and 70 °C for solution of 2 (83.3 mM) and 3 (16.7 mM)


39

in EMITFSI. For the measurements, frequency was swept between 1 MHz and 1 Hz using a frequency response analyzer (Solartron, model 1255B) connected to a multipotentiostat (Solartron, model 1470); a signal amplitude of 10 mV was applied at a dc bias of 0 V. Data acquisition and analysis were performed, respectively, with ZView (version 2.70) and Zplot (version 2.80) softwares from Scribner Associates under Windows environment. 2.5. Diffusion coefficient measurements Diffusion coefficient of thiolate (reduced) species 2 was determined at 25 °C in two mixtures of 2 and 3 in EMITFSI, using a platinum electrode and potential step chronoamperometry. The electrode was set at the open-circuit potential for 10 min and then at 1.14 V vs. Ag (potential step) for 50 s. The Cottrell equation was used to obtain the diffusion coefficients by relating the anodic current i to the time t following application of the potential: iðt Þ ¼ n F A D1=2 C=π1=2 t 1=2

ð1Þ

where n is the number of electrons exchanged, F is Faraday's constant (96,485 C mol−1), A is the electrode surface area (cm2), D is the diffusion coefficient (cm2 s−1) and C is the electroactive species bulk concentration (mol cm− 3). The slope of the i vs. t− 1/2 curve allows determination of D values. The i vs. t−½ curves are shown in Fig. S2 of the Supporting information. 2.6. Viscosity and absorbance measurements Viscosity of the solution of 2 (83.3 mM) and 3 (16.7 mM) in EMITFSI was obtained at 25, 50 and 70 °C using an electromagnetic viscosimeter (viscosity monitoring and control electronics, VISCOlab 4000, Cambridge Applied Systems) connected to a thermostated bath (Hake). Absorbance measurements of solutions of 2 (75 mM) and 3 (25 mM) in EMITFSI, and of 1-ethyl-3-methylimidazolium iodide (EMII, 163 mM) and I2 (10 mM) in EMITFSI, were performed at room temperature using a spectrophotometer (Ultrospec 100 pro). A 1 mL semi-micro quartz cell (VWR International) with an optical path of 1 cm was employed. 3. Results and discussion 3.1. Electrochemical properties of TfmTT− and of TfmTT−/BTfmT solutions in EMITFSI Fig. 2a, b and c depicts cyclic voltammograms (cycle number 5) of a solution of 2 (TfmTT− K+, 75 mM) in EMITFSI performed at 50 mV s−1 on platinum, glassy carbon and ITO-CoS electrode, respectively. The voltammograms show an oxidation and a reduction peaks that are however less well defined at the platinum electrode. The anodic peak corresponds to the oxidation of thiolate species 2 into their radical (RS− → RS• + e−). Combination of two radicals gives the disulfide. Scanning the potential from open-circuit to more cathodic values without reaching the anodic potentials allows us to demonstrate the relationship between the oxidation and reduction peaks; several cycles were carried out and the second one is presented in Fig. 2. For the three electrodes, after the first cycle, the reduction current has decreased drastically. This experiment indicates that the oxidation and reduction peaks are connected to each other; hence, the species formed during oxidation is responsible for the observed reduction peak. Fig. 2. Cyclic voltammograms (5th cycle for the full scan, 2nd cycle for the reduction scan) of a solution of 2 (TfmTT− K+, 75 mM) in EMITFSI: (a) platinum electrode, (b) glassy carbon electrode, and (c) ITO-CoS electrode. The initial scan direction was oxidizing (full scan) or reducing (reduction scan); scan rate = 50 mV s−1.

40


The peak current densities increase with the geometric electrode surface area: ITO-CoS N GC N Pt, which indicates a better catalytic activity in that order. As shown in the scanning electron microscope image of Fig. S1 of the Supplementary information, the ITO-CoS electrode is very porous, which is not the case for GC and Pt. This explains the large capacitive current density observed in the CV of ITO-CoS. Fig. 3a, b and c shows cyclic voltammograms (cycle number 5) of EMITFSI, solution of 2 (TfmTT− K+, 75 mM) in EMITFSI, and of 2 (TfmTT− K+, 75 mM) and 3 (BTfmT, 25 mM) in EMITFSI, at platinum, glassy carbon and ITO-CoS electrode, respectively. Fig. 3 clearly demonstrates that EMITFSI is almost electrochemically inactive in the selected potential window. Fig. 3 also shows that the addition of 25 mM of the disulfide form 3 to 75 mM of the thiolate compound 2 in EMITFSI doesn't result in the formation of an additional reduction peak. However, a significant increase in the cathodic current density and a lower increase in the anodic current density (the relative anodic current density increase is more important at the Pt electrode) are observed. Reduction of disulfide BTfmT species increases the concentration of TfmTT− ions already present in the solution. More species are then oxidized during the reverse cycle, increasing the anodic current density. The CVs therefore demonstrate that the cathodic peak is associated with the reduction of BTfmT. Fig. S3 of the Supplementary Information compares the CV of disulfide form 3 in EMITFSI with those of thiolate form 2, and of mixture of forms 2 and 3 in EMITFSI at the three electrode materials. We can notice that the absence of thiolate species 3 in the as-prepared solution mixture gives much lower current densities. The fifth cycles are presented in order to allow the electrochemical system to stabilize and to have the same basis of comparison. Between the first and second cycles (and sometimes between the first and fifth cycles, depending of the experiment carried out), a difference in the amplitude of the current density peaks (oxidation and reduction) was observed. From the 5th cycle, reproducible and stable CVs were obtained. For instance, at a platinum electrode, the anodic and cathodic current density is increased from 0.84 to 0.95 mA cm−2 and from − 0.72 to −0.80 mA cm−2, respectively. For a better comparison of the redox kinetics at the three electrodes, Fig. 4 depicts the cyclic voltammograms (cycle number 5) as a function of potential, of the solution containing the species 2 (TfmTT−, 75 mM) and 3 (BTfmT, 25 mM) in EMTFSI. Table 1 gives the values of ΔEp (separation between the anodic (Epa) and cathodic (Epc) peak potentials), mid-peak potential (Emp = (Epa + Epc) / 2), anodic peak current density (ja) and absolute value of cathodic peak current density (|jc |). The greater current densities and smaller ΔEp values on GC and ITOCoS electrodes relative to Pt indicate faster reaction rates at these electrodes. Several groups have shown that carbon electrodes, and glassy carbon in particular, are good catalysts for the reduction of aromatic disulfide molecules [29,30]. For instance, Liu et al. [29] have shown that reduction of such class of molecules in a non-aqueous medium is favored on graphite electrode compared to platinum. On the other hand, Borsari et al. [30] have compared reduction of aromatic disulfides on glassy carbon and gold electrodes, and have shown that there is no adsorption phenomenon on glassy carbon contrary to gold, and that the kinetics is diffusion-controlled. As depicted in Fig. 4, the ITO-CoS electrode is characterized by the highest anodic and cathodic current densities; the geometric surface area of the electrodes was used to obtain the current densities. As discussed above, the ITO-CoS electrode is very porous and hence the electrochemically active surface area is greater than the geometric area. Therefore, the current densities presented in Fig. 4 are overestimated for this electrode, if we compare the electrochemically active surface areas. However, it can be seen (Fig. 4) that the ITO-CoS

Fig. 3. Cyclic voltammogram (5th cycle) of solutions of 2 (TfmTT− K+, 75 mM) in EMITFSI, 2 (TfmTT− K+, 75 mM) and 3 (BTfmT, 25 mM) in EMITFSI, and EMITFSI: (a) platinum electrode, (b) glassy carbon electrode, and (c) ITO-CoS electrode. The initial scan direction was oxidizing; scan rate = 50 mV s−1.


41

Fig. 4. Cyclic voltammogram (5th cycle) of a solution of 2 (TfmTT− K+, 75 mM) and 3 (BTfmT, 25 mM) in EMITFSI. A platinum electrode, a glassy carbon electrode and an ITOCoS electrode were used as working electrodes. The initial scan direction was oxidizing; scan rate = 50 mV s−1.

electrode possesses the less positive anodic peak potential and the less negative cathodic peak potential, leading to the smallest ΔEp value. This result indicates lower overpotentials for the redox reactions at this electrode. Table 1 shows that the mid-peak potential of the redox processes at the ITO-CoS electrode (0.42 V vs. Ag) is more positive compared to glassy carbon (0.35 V vs. Ag) and platinum (0.26 V vs. Ag). 3.2. Influence of the temperature on the electrochemical properties For this study, a mixture of 2 (TfmTT−, 83.3 mM) and 3 (BTfmT, 16.7 mM) in EMITFSI was used. To approach the temperature conditions found in a solar cell, cyclic voltammetry measurements were carried out at 50 °C and 70 °C, and compared with those obtained at 25 °C at platinum, glassy carbon and ITO-CoS electrodes. The first part of this study was conducted at platinum and glassy carbon working electrodes; the cyclic voltammograms (cycle number 5) are presented in Fig. 5. The voltammograms show that, for both electrodes, an increase in temperature results in an increase in the amplitude of the anodic and cathodic currents, and a reduction of ΔEp. This decrease in the ΔEp value reflects the less positive anodic peak potential and the less negative cathodic peak potential with increasing temperature. Table 2 gives the values of jpa, |jc | and ΔEp obtained for each electrode material. When the temperature is increased from 25 °C to 70 °C, ΔEp is decreased by a value of 0.32 V on Pt and 0.26 V on GC, which may be caused by the decreasing medium viscosity and hence resistivity. Indeed, the viscosity is decreasing from 34 cP at 25 °C, to 25 cP at 50 °C and to 13 cP at 70 °C, whereas the ionic conductivity is increasing from 7.05 ± 0.03 mS cm−1 at 25 °C, to 8.09 ± 0.17 mS cm−1 at 50 °C and to 21.90 ± 0.20 mS cm−1 at 70 °C. The decrease in viscosity will enhance mobility of the electroactive species with an impact on charge Table 1 Separation between the anodic and cathodic peak potentials (ΔEp), mid-peak potential (Emp), anodic peak current density (ja) and absolute value of cathodic peak current density (|jpc |) at 25 °C for a solution of 2 (TfmTT− K+, 75 mM) and 3 (BTfmT, 25 mM) in EMITFSI. The values were determined from CVs of Fig. 4. Electrode

ΔEp (V)

Emp (V vs. Ag)

jpa (mA cm−2)

|jpc | (mA cm−2)

Pt GC ITO-CoS

1.12 0.80 0.57

0.26 0.35 0.42

0.95 1.35 2.41

0.80 1.32 2.34

Fig. 5. Cyclic voltammogram (5th cycle) of a solution of 2 (TfmTT− K+, 83.3 mM) and 3 (BTfmT, 16.7 mM) in EMITFSI at 25, 50 and 70 °C: (a) platinum electrode, and (b) glassy carbon electrode. The initial scan direction was oxidizing; scan rate = 50 mV s−1.

transport within the electrolytic medium. This fact contributes to the increase in anodic and cathodic current densities (more important for the former) that has been observed at higher temperature. Anodic current is associated with oxidation of thiolate species which are charged species. Table 2 Separation between the anodic and cathodic peak potentials (ΔEp), anodic peak current density (ja) and absolute value of cathodic peak current density (| jpc |) at 25, 50 and 70 °C for a solution of 2 (TfmTT− K+, 83.3 mM) and 3 (BTfmT, 16.7 mM) in EMITFSI. The values were obtained from CVs of Fig. 5. T (°C)

Pt

GC

ΔEp (V)

jpa (mA cm−2)

|jpc | (mA cm−2)

ΔEp (V)

jpa (mA cm−2)

|jpc | (mA cm−2)

25 50 70

1.19 1.13 0.87

1.24 1.72 2.40

0.76 0.88 1.08

0.84 0.67 0.58

1.69 2.68 3.52

1.13 1.83 2.11

42


In a dissociated medium like EMITFSI, addition of charged species has an important impact on the viscosity. For instance, the addition of oxidized and reduced species in EMITFSI promotes a decrease of the medium conductivity, from 9.60 ± 0.02 mS cm−1 for EMITFSI to 7.05 ± 0.03 mS cm− 1 for the mixture of TfmTT− (83.3 mM) and BTfmT (16.7 mM) in EMITFSI at 25 °C. As the viscosity is decreased at higher temperature, more thiolate species reach the surface of the working electrode to give greater anodic current. A similar behavior was observed with the disulfide species (larger cathodic current at higher temperature). On the other hand, both reactions are kinetically activated at higher temperature. Table 2 shows that the cathodic current density is less improved when the temperature is raised, compared to the anodic current density. For instance, on Pt electrode, these improvements are −0.32 mA cm−2 (42%) and 1.16 mA cm−2 (94%), respectively, between 25 °C and 70 °C. On glassy carbon electrode, these values are −0.98 mA cm−2 (87%) and 1.83 mA cm−2 (108%), respectively. Fig. 6 presents the cyclic voltammograms (cycle number 5) of the above electrolytic mixture registered at 25 °C, 50 °C and 70 °C, at an ITO-CoS electrode. In this case, the anodic potential was limited to 0.55 V in order to observe more specifically the variation of the cathodic current with temperature as this material may replace Pt as the counter electrode in DSSCs. As in the case of platinum and glassy carbon electrodes, the temperature increase caused an increase in the amplitude of the ITO-CoS cathodic peak current density, from − 0.4 mA cm− 2 at 25 °C, to −1.28 mA cm−2 at 50 °C and to −2.20 mA cm−2 at 70 °C. Hence, the cathodic current density is improved by − 1.60 mA cm− 2 (400%), which is much more than the increase observed for Pt and GC electrodes. This is again an indication of the excellent electrocatalytic activity of ITO-CoS electrode toward the reduction of disulfide species.

3.3. Comparison of the electrochemical performance of mixtures of compounds 2 and 3 Tables 1 and 2 allow us to compare the electrochemical parameters of two mixtures of compounds 2 and 3 in EMITFSI at 25 °C, with different TfmTT−/BTfmT ratios: 75 mM TfmTT− + 25 mM BTfmT (ratio of 3/ 1, A), and 83.3 mM TfmTT− + 16.7 mM BTfmT (ratio of 5/1, B), at Pt and GC electrodes. The total redox concentration was kept constant, at

100 mM. The parameters ΔEp, Jpa and |jpc | were obtained from CVs of Figs. 4 and 5, and using the electrode geometric surface areas. Apart from the peak current densities, few differences are noticed for the same electrode dipped in mixtures A and B. Indeed, the ΔEp values are very similar. However, the anodic peak current densities are larger with mixture B (greater thiolate/disulfide molar ratio) whereas the cathodic peak current densities are more important with mixture A. For instance, the Jpa values are 1.24 mA cm−2 (Pt) and 1.69 mA cm− 2 (GC) for mixture B, and 0.95 mA cm−2 (Pt) and 1.35 mA cm−2 (GC) for mixture A. These results may be explained by the greater concentration of the reduced species (thiolate) with respect to the oxidized species (disulfide) in mixture B even if the diffusion coefficient of the thiolate is greater in mixture A (8.2 × 10− 7 cm2 s−1 compared to 7.7 × 10−7 cm2 s−1 for mixture B). The higher diffusion coefficient of the charged species 2 in mixture A is consistent with the trend observed in the ionic conductivity values: 7.46 ± 0.22 mS cm−1 (A) vs 7.05 ± 0.03 mS cm−1 (B) at 25 °C (greater mobility of the thiolate species in mixture A). On the other hand, the | jpc | values are 0.80 mA cm−2 (Pt) and 1.32 mA cm− 2 (GC) for mixture A, and 0.76 mA cm− 2 (Pt) and 1.13 mA cm−2 (GC) for mixture B. The more significant concentration of the oxidized species with respect to the reduced species in mixture A may explain the results. Regarding the ITO-CoS electrode, similar conclusions can be made if we compare Fig. 4 (Table 1) and Fig. 9 (introduced later): the Jpa value is greater for mixture B (2.99 mA cm−2 vs 2.41 mA cm−2 for mixture A), whereas the |jpc | value is more important for mixture A (2.34 mA cm−2 vs 2.22 mA cm−2 for mixture B). 3.4. Diffusion-controlled redox process Fig. 7 depicts the cyclic voltammograms (cycles number 5) of mixture A (TfmTT−/BTfmT, ratio of 3/1) registered at 25 °C and at several potential scan rates (20 to 200 mV s−1), at Pt, GC and ITO-CoS electrodes. All the CVs show well defined peaks (semi-infinite diffusion process), however less obvious for the oxidation reaction at the ITO-CoS electrode (Fig. 7c), with currents increasing with the scan rate. Furthermore, a shift of the oxidation and reduction peak potentials, respectively toward more anodic and more cathodic values, is observed with increasing the scan rate. This phenomenon reflects that the redox processes are not electrochemically reversible. When an irreversible redox couple is diffusion-controlled, a linear relationship exists between the peak current and the square root of the scan rate. This relation is known as a Randles-Sevcik-type equation [31]: ip ¼ 2:99 105 α1=2 AD1=2 C ν1=2

ð2Þ

where ip is the peak current (A), α the transfer coefficient of the electrochemical process, A the electrode surface area (cm2), D the diffusion coefficient of the electroactive species (cm2 s− 1), C⁎ its equilibrium concentration (mol cm−3) and ν the scan rate (V s−1). In order to confirm the diffusional control with redox mixture A, Fig. 8 presents the plot of absolute value of the anodic and cathodic currents as a function of the square root of the scan rate for the three electrode materials. A linear relationship is clearly demonstrated in all cases, which confirm the diffusion control. Hence, oxidation currents are limited by the diffusion of thiolate species (TfmTT−) toward the surface of the working electrode, and reduction currents are limited by the diffusion of disulfide species (BTfmT). 3.5. Electrochemical stability of ITO-CoS electrodes Fig. 6. Cyclic voltammogram (5th cycle) of a solution of 2 (TfmTT− K+, 83.3 mM) and 3 (BTfmT, 16.7 mM) in EMITFSI at 25, 50 and 70 °C. An ITO-CoS electrode was used as the working electrode. The initial scan direction was reducing; scan rate = 50 mV s−1.

A preliminary stability study of ITO-CoS electrodes was carried out in redox mixture B (TfmTT−/BTfmT, ratio of 5/1) using cyclic voltammetry


Fig. 7. Cyclic voltammogram (5th cycle) of a solution of 2 (TfmTT− K+, 75 mM) and 3 (BTfmT, 25 mM) in EMITFSI at 25 °C and various potential scan rates: (a) platinum electrode, (b) glassy carbon electrode, and (c) ITO-CoS electrode. The initial scan direction was oxidizing.

43

Fig. 8. |ip | vs. v1/2 curves relative to cyclic voltammograms of Fig. 7: (a) platinum electrode, (b) glassy carbon electrode, and (c) ITO-CoS electrode.

44


Fig. 10. UV–vis absorption spectra of two electrolytic media: a solution of 2 (TfmTT− K+, 75 mM) and 3 (BTfmT, 25 mM) in EMITFSI, and a solution of EMII (163 mM) and I2 (10 mM) in EMITFSI.

Fig. 9. Cyclic voltammograms of a solution of 2 (TfmTT− K+, 83.3 mM) and 3 (BTfmT, 16.7 mM) in EMITFSI at 70 °C (5th and 300th cycles) at an ITO-CoS electrode. The initial scan direction was reducing; scan rate = 50 mV s−1.

at 70 °C to simulate the temperature reached by solar cells under illumination. It should be recalled that CoS is a serious candidate to replace platinum as the counter electrode in DSSCs (reduction of the mediator oxidized species). For this study, 300 cycles were recorded at a scan rate of 50 mV s−1, between 0.55 V and 0.15 V vs. Ag; the voltammograms are presented in Fig. 9. By comparing the 5th cycle with the 300th cycle, it can be concluded that the cathodic peak current density is decreased, but not to a great extent, with the number of cycles. Hence, the ITO-CoS electrode maintains a very good electrochemical activity. This also means that the electrode retains a good dimensional and mechanical stability at 70 °C. To confirm the stability of the ITO-CoS electrode under these conditions, a greater number of cycles (e.g., 5000 cycles) will be performed in our laboratory. However, the present study is a first indication of the good electrochemical and mechanical stability of the electrode under the operating conditions of a solar cell. 3.6. Visible spectrum of the electrolyte During illumination of an EPC (and ideally of a DSSC), photons must pass through both the counter electrode and the electrolytic medium before reaching the surface of the semiconductor electrode. Therefore, the electrolytic medium must possess a very good transparency in the visible region of the solar spectrum. Fig. 10 compares the qualitative UV–vis absorption spectrum of redox mixture A in EMITFSI (TfmTT−/BTfmT, ratio of 3/1) with that of a solution of EMII (163 mM) and I2 (10 mM) in EMITFSI. The I−/I2 redox couple is used in conventional DSSCs [32]. The data clearly show a much weaker absorption of redox mixture A in the visible region, which is very interesting for application in solar cells. 4. Conclusion The objective of the present work was to synthesize and characterize chemically a new redox couple, composed of the reduced species potassium 5-trifluoromethyl-1,3,4-thiadiazole-2-thiolate 2 and the oxidized species 5,5′-bis (2-trifluoromethyl-1,3,4-thidiazole) disulfide 3, and to study its electrochemical properties when dissolved in the ionic liquid EMITFSI. The electrochemical study was carried out by cyclic

voltammetry using two different thiolate/disulfide ratios (3/1 and 5/ 1), at different temperatures (25, 50 and 70 °C) and at Pt, GC and ITOCoS electrodes. Conductivity, viscosity and UV–visible absorbance measurements completed the study. The results showed that the redox reactions are irreversible and subject to a diffusion-controlled process. At higher electrolyte temperature, the anodic and cathodic current densities are enhanced, and the potential difference between the anodic and cathodic peaks is decreased, which is of great interest for application in solar cells. The increase in the conductivity of the electrolytic medium (the viscosity is decreased) contributes largely to this behavior. The best electrocatalytic performance was obtained with the ITOCoS electrode, followed by GC and Pt electrodes. It has been shown that when the proportion of thiolate species in the redox mixture is increased, the anodic current density is increased; the cathodic activity is improved when this proportion is decreased. Preliminary stability studies (300 voltammetric cycles) of the ITO-CoS electrode confirmed the very good electrochemical and mechanical stability of this electrode, which is considered as a serious candidate to replace platinum as the counter electrode in DSSCs. The withdrawing effect of the CF3 group on the thiadiazole ring further favors the reduction of disulfide molecules. The electrolytic media studied in this work exhibits a very low ab− sorption of visible light, contrarily to the conventional I− 3 /I mediator (EMII), which allows illumination of the device via the counter electrode in DSSCs. In a forthcoming publication, this electrolyte will be used in practical conditions of a DSSC with ITO-CoS as the counter electrode to compare the solar cell performance with the literature. Acknowledgements The authors are grateful to the Natural Sciences and Engineering Research Council of Canada (NSERC) (grant number DG-41924) and to DKS Co., Ltd. (Japan) for financial support. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jelechem.2017.01.008. References [1] M. Grätzel, B. O'Regan, Nature 353 (1991) 737–740. [2] J. Krüger, R. Plass, M. Grätzel, P.J. Cameron, L.M. Peter, J. Phys. Chem. B 107 (31) (2003) 7536–7539. [3] M.I. Asghar, K. Miettunen, J. Halme, P. Vahermaa, M. Toivola, K. Aitola, P. Lund, Energy Environ. Sci. 3 (4) (2010) 418–426. [4] E. Olsen, G. Hagen, S.E. Lindquist, Sol. Energy Mater. Sol. Cells 63 (3) (2000) 267–273. [5] J.F. Lu, J. Bai, X.B. Xu, Z. Li, K. Cao, J. Cui, M. Wang, Chin. Sci. Bull. 57 (32) (2012) 4131–4142.

G. Hersant et al. / Journal of Electroanalytical Chemistry 787 (2017) 36–45 [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21]

H. Gerischer, J. Gobrecht, Ber. Bunsenges. Phys. Chem. 80 (1976) 327–330. H. Gerischer, Electrochim. Acta 35 (11–12) (1990) 1677–1699. M. Grätzel, Coord. Chem. Rev. 111 (1991) 167–174. P.V. Wright, Br. Polym. J. 7 (1975) 319–327. A.K. Vijh, B. Marsan, Bull. Electrochem. 5 (1989) 456–461. J.M. Philias, B. Marsan, Electrochim. Acta 44 (17) (1999) 2915–2926. G. Feuillade, P. Perche, J. Appl. Electrochem. 5 (1) (1975) 63–69. I. Renard, H. Li, B. Marsan, Electrochim. Acta 48 (7) (2003) 831–844. J.S. Wilkes, M.J. Zaworotko, J. Chem. Soc. Chem. Commun. (1992) 965–967. J. Fuller, R.T. Carlin, H.C. De Long, D. Haworth, J. Chem. Soc. Chem. Commun. (1994) 299–300. K. Matsumoto, R. Hagiwara, R. Yoshida, Y. Ito, Z. Mazej, P. Benkič, B. Žemva, O. Tamada, H. Yoshino, S. Matsubara, Dalton Trans. 1 (2004) 144–149. P. Bonhôte, A.P. Dias, N. Papageorgiou, K. Kalyanasundaram, M.B. Armand, M. Grätzel, Inorg. Chem. 35 (5) (1996) 1168–1178. R. Kawano, H. Matsui, C. Matsuyama, A. Sato, M.A.B.H. Susan, N. Tanabe, M. Watanabe, J. Photochem. Photobiol. A Chem. 164 (1–3) (2004) 87–92. H. Sakaebe, H. Matsumoto, K. Tatsumi, J. Power Sources 146 (1) (2005) 693–697. E. Shouji, D.A. Buttry, J. Phys. Chem. B 103 (12) (1999) 2239–2247. A. Hammami, B. Marsan, M. Armand, G. Hersant, European Patent, Publication Number: EP2044647 B1, January 15, 2014.

45

[22] M. Wang, N. Chamberland, L. Breau, J.E. Moser, R. Humphry-Baker, B. Marsan, S.M. Zakeeruddin, M. Grätzel, Nat. Chem. 2 (5) (2010) 385–389. [23] a) H. Tian, X. Jiang, Z. Yu, L. Kloo, A. Hagfeldt, L. Sun, Angew. Chem. Int. Ed. 49 (2010) 7328–7331; b) H. Tian, Z. Yu, A. Hagfeldt, L. Kloo, L. Sun, J. Am. Chem. Soc. 133 (24) (2011) 9413–9422. [24] B. Marsan, B. Bourguignon, Canadian Patent 2,544,073 (2007); US Patent 7,626,117 B2, 2009. [25] J. Burschka, V. Brault, S. Ahmad, L. Breau, M.K. Nazeeruddin, B. Marsan, S.M. Zakeeruddin, M. Grätzel, Energy Environ. Sci. 5 (3) (2012) 6089–6097. [26] S. Antonello, K. Daasbjerg, H. Jensen, F. Taddei, F. Maran, J. Am. Chem. Soc. 125 (48) (2003) 14905–14916. [27] J. Wu, Z. Lan, J. Lin, M. Huang, Y. Huang, L. Fan, G. Luo, Chem. Rev. 115 (5) (2015) 2136–2173. [28] R.L. McCreery, Chem. Rev. 108 (7) (2008) 2646–2687. [29] M. Liu, S.J. Visco, L.C. De Jonghe, J. Electrochem. Soc. 136 (9) (1989) 2570–2575. [30] M. Borsari, M. Cannio, G. Gavioli, Electroanalysis 15 (14) (2003) 1192–1196. [31] A.J. Bard, L.R. Faulkner (Eds.), Electrochemical Methods: Fundamentals and Applications, second ed.John Wiley & Sons, New York, 2001. [32] F.T. Kong, S.Y. Dai, K.J. Wang, Adv. OptoElectronics (2007), Article ID 75384. (13 pages).