Steric and Solvent Effect in Dye-Sensitized Solar Cells Utilizing Phenothiazine-Based Dyes

1 Beijing National Laboratory for Molecular Sciences, Key Laboratory of Polymer Chemistry and Physics of Ministry of Education, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China 2Dyeing, Printing and Textile Auxiliaries Department, Textile Research Division, National Research Center, Dokki, Cairo 12622, Egypt 3 Chemistry Department, Faculty of Science, King Abdulaziz University, P.O. Box 80203, Jeddah 21589, Saudi Arabia


Introduction
The increasing consumption of fossil and the more serious crisis of environment pollution have led us to the search for new and renewable energy sources.Solar energy is widely recognized as the most promising candidate in helping solve this problem.Dye-sensitized solar cells (DSSCs), also known as Grätzel cells, offer a viable alternative to conventional all-inorganic solar cells because of their lower production cost.During the past decades, DSSCs have attracted significant attention as an alternative to silicon solar cell because they use environmentally benign materials through low cost process and exhibit commercially realistic energyconversion efficiency [1][2][3].The photon-to-current conversion of DSSCs is achieved by ultrafast electron injection from a photoexcited dye into the conduction band of the TiO 2 , followed by dye regeneration and hole transportation to the counter electrode.To enhance the performance of the DSSC, extensive research has been performed on semiconductor nanocrystalline TiO 2 electrodes [4][5][6], dye molecule [7][8][9][10], electrolytes [11][12][13][14], and counter electrodes [15].
Compared with the rare and expensive metal complexes, organic dyes have the advantages of being eco-friendly, having flexible and diverse form of molecular structures, lower cost, generally high molar extinction coefficients, and easier preparation and purification.Metal-free organic dyes have been widely investigated recently, many of which exhibited an energy-to-electricity conversion efficiency close to that of N719 [16][17][18][19][20][21].
The design of organic dyes for DSSCs is important to improve the value of the short-circuit current ( sc ) and the open-current voltage ( oc ).The -conjugation of the  sc must be improved so the organic chromophore can harvest light energy to a large extent.An effort of reducing the rate of charge recombination is necessary for the  oc to minimize current leakage.
Whether the low voltages from organic dyes may be due to fast recombination kinetics is still being debated [22,23].A strong interaction between organic dye and iodine in the electrolyte has been proposed to be responsible for accelerating recombination [24,25].Also, the lack of electron donating moiety in the oxidized organic dye has also been proposed to have an influence on fast recombination reaction, compared with N719 with electron donating NCS ligand [26].In addition, dye aggregation due to planar or sterically less-hindered structure has also been proposed as one of major factors for the lowering voltage [27,28].Thus, a molecular structure of organic dye must be designed to reduce recombination and/or prevent dye aggregation.Recently, heterocyclic phenothiazine (PT), originally used in drug applications [29,30], has been adopted as a novel electron donor in organic dye because PT contains electronrich sulfur-nitrogen heteroatoms and its ring has sterically hindered nonplanar butterfly conformation.In addition, the incorporation of PT derivatives in organic dye backbone is expected to inhibit molecular aggregation.
Most studies have focused on enhancing the properties of the photoelectrode, specifically the dye, which is responsible for most light absorption.Some of the important properties that need to be considered when designing sensitizers for DSSCs are the geometric structures, molecular orbital energy, absorption profiles, and aggregation states of the dye [31,32].
In this study, the effect of THF and EtOH on the absorption and solar cell performance was studied and evaluated.The synthesis and application of three dyes composed of Nsubstituted phenothiazine units (PT1-3) shown in Scheme 1 as sensitizers in dye-sensitized solar cells have been investigated.A cyanoacrylate group was attached to one side of the compound, acting as an electron acceptor.We envisioned that the nonplanar conformation of phenothiazine can reduce the rate of charge recombination and molecular aggregation.The synthetic procedures of PT1-3 are described in Scheme 1.All structures have been confirmed by their NMR, MS, and spectroscopic data.The photovoltaic performance and electrochemical properties of devices based on the three dyes using THF and EtOH solvents were evaluated by different spectroscopic techniques.

Materials and Synthesis
2.1.1.General Considerations.Reagents, catalysts, ligands, and solvents were purchased reagent grade and used without further purification.The details of the synthetic procedure of the PT1-3 dyes were prepared according to the literature procedures, as shown in Scheme 1 [33][34][35].N-alkylation of phenothiazine followed by Vilsmeier-Haack formylation and then Knoevenagel condensation between the carbaldehyde and cyanoacetic acid in the presence of ammonium acetate affords the desired dyes.Column chromatography was made using silica gel 60, mesh 70−230.TLC silica gel plates, 60 F 254 , were also used.Lithium perchlorate (LiClO 4 , 99%) and 4-tert-butylpyridine (TBP) were obtained from Sigma-Aldrich.Acetonitrile (AN) solvent as purchased from Siyou Inc., Tianjin, China.The substance I 2 (AR) was acquired from the Beijing Chemical Reagent Company.FTO glasses and TiO 2 paste (DSL-18NR-T), with diameter of 20 nm, and the electrolyte components were obtained from Heptachroma Inc., Dalian, China.

Characterization and Measurement.
UV-Vis absorbance spectra were recorded on a lambda 35 UV/Vis Spectrometer (Perkin Elmer).All dyes were dissolved in THF and EtOH solution as bath solvent, and the concentrations of solutions used in the absorption experiment were set as 5 × 10 −5 M. 1 H NMR and 13 C NMR spectra were measured at room temperature on a 400 MHz (Bruker) spectrometers, respectively, using DMSO-d6 or CDCl 3 as the solvents.The electrospray ionization mass spectra (ESI-MS) were characterized on an APEX IV Fourier transform ion cyclotron resonance mass spectrometer (Bruker).

Fabrication of DSSCs.
The TiO 2 paste (DSL-18NR-T, diameter: 20 nm) was created on FTO glasses by screen printing.The obtained FTO glasses were sintered at 150 ∘ C for 15 min in a muffle furnace and then at 450 ∘ C for 30 min sequentially.The thickness of the TiO 2 film is approximately 3.5 m after one screen printing.Through reciprocating the above steps three times, the TiO 2 film can reach up to 10.5 m thick.After the last heated process, the TiO 2 electrodes were cooled to 100 ∘ C and then immersed into dye bath solution.After 24 h, the sensitized TiO 2 photoanodes were taken out from the dye solution, eluted by methanol to remove the excess dye from the TiO 2 surface, and were kept in vacuum drying oven at room temperature overnight.The counter electrode with a Pt film (thickness: approximate 50 nm) on the FTO glasses was prepared by magnetic sputtering.The obtained dye-adsorbed photoanodes (test area: ca.0.25 cm 2 ) and the Pt counter cathodes were assembled to form DSSCs and isolated by 25 m thick hot-melt ionomer film (Surlyn, Dupont).The electrolytes were then injected into solar cells by preset pores in the counter electrode.Finally, the devices were sealed by Surlyn film to enhance the stability of the DSSCs.

Dye Bath Solutions.
In this study, EtOH and THF bath solutions of PT1-3 dyes (5 × 10 −4 M) were prepared.The devices were soaked in EtOH and THF solution sensitizers for 24 h, respectively.The obtained FTO glasses were sintered at 150 ∘ C for 15 min in a muffle furnace and then at 450 ∘ C for 30 min sequentially.The thickness of the TiO 2 film was approximately 3.5 m after one screen printing.The TiO 2 films were cooled to 100 ∘ C, and then immersed into different dye bath solutions.After 24 h, the sensitized TiO 2 photoanodes were taken out from the dye solution and eluted by methanol to remove the excess dye from the TiO 2 surface.The sensitized TiO 2 photoanodes were then immersed in 0.1 M solution of (CH 3 ) 4 NOH base in ethanol overnight to extract the dye from the TiO 2 film.The film area was 0.64 cm 2 and thickness 10.5 m and it was extracted in 9 mL of the (CH 3 ) 4 NOH base in ethanol solution.The absorption spectra of the dyes in the extraction solution were then obtained.

Characterization of the Photoelectrochemical and Photovoltaic Properties.
The photovoltaic performance of the devices was measured under AM 1.5 simulated sunlight illumination with a light intensity of 100 mW/cm 2 .The simulated sunlight source was YSS-50A (Yamashita DESO).The electrochemical impedance spectra (EIS) were implemented on the Autolab (Frequency range: 100 kHz to 10 mHz).Two parameters, namely, chemical capacitance (  ) and charge recombination resistance ( rec ), were obtained by fitting the EIS at different reverse biases with Z-View software according to the transmitting line mode.The incident photon-tocurrent efficiency (IPCE) was measured from 300 nm to 800 nm under SolarCellScan-100 (Beijing ZOLIX Corp.).

Results and Discussion
3.1.Synthesis.PT1-3 dyes were prepared according to the literature procedures [33][34][35][36][37].The sequence of reaction steps proceeds smoothly and efficiently to give a good yield of the product.Thus, N-alkylation of phenothiazine with different alkyl groups afforded the corresponding compounds 1a-c, which upon Vilsmeier-Haack formylation gave the carbaldehyde 2a-c.Knoevenagel condensation between 2a-c and cyanoacetic acid in the presence of ammonium acetate produced the desired phenothiazine dyes PT1-3.All chemical structures were characterized by their spectroscopic data.

Absorption Spectra.
The UV-Vis absorption spectra of all dyes in the EtOH and THF solutions are shown in Figure 1.Table 1 shows all the parameters; all the dyes exhibited broad absorption in the range of 250-350 nm and 370-525 nm.The short wavelength region at 280-340 nm is attributed to the localized - * and n- * transitions, whereas the long wavelength region in the range of 390-510 nm is attributed to the charge transfer transitional energy of the delocalized - * transition as a result of donor--acceptor system.Compared with THF, the absorption bands of all dyes are relatively blue-shifted using EtOH.This behavior would be attributed to the solvent effect.As PT1-3 dyes are donor--acceptor system, charge transfer absorption band is destabilized in protic solvent such as ethanol due to hydrogen bonding interactions between ethanol and the donor and/or acceptor moieties [38,39].On the other hand, THF is aprotic solvent and less polar than EtOH; such effect does not exist and thus transfer absorption band is more favorable in this solvent than EtOH.

Photovoltaic Performance.
Figure 2 shows the photocurrent-voltage curves and IPCE spectra for the PT1-3 sensitized DSSCs in the EtOH and THF adsorption solvent.The corresponding photovoltaic parameters are listed in Table 2.
The device performance of PT1-3 dyes indicates that the devices using EtOH show a higher  sc and a significantly higher  oc than those obtained using THF, with a higher light conversion efficiency of PT3 dye compared with those of PT2 and PT1.The reasons for this inferior performance for the dyes using THF solution as mentioned before would be attributed to the smaller amount of adsorbed dye in the TiO 2 semiconductor due to the propensity of dye molecules to exist more in THF solvent rather than to get adsorbed and also due to the hydrogen bond interaction of the oxygen of the tetrahydrofuran and the hydrogen of the carboxylic acid of the dyes [40,41].Meanwhile, it was also noted that dyes with more branched alkyl groups show higher  oc and  sc in both EtOH and THF dye-bath solvent.We will discuss these afterwards.
The incident photon-to-current efficiency (IPCE) spectra as a function of the wavelength for DSSCs based on the PT1-3 dyes in THF and EtOH is shown in Figure 4(c  of 80.5% at 460 nm for PT3 sensitized), whereas using THF only exceeded 50% from 425 nm to 475 nm (with the highest value of 55% at 430 nm PT3 sensitized).The higher IPCE for PT1-, 2-, and 3-sensitized solar cells using EtOH can be ascribed to the larger adsorbed amount of PT1-3 dyes compared with those obtained using THF, which is consistent with the obtained photovoltaic data for the PT1-3 sensitizers under the same conditions, wherein the  sc for the devices that use EtOH are much higher than those obtained using THF as dye-adsorption solvent.
The absorption spectra of all dyes extracted from the surface of TiO 2 film are shown in Figure 3.The absorption band displayed a blue shift of ca.1-38 nm with respect to those in the solutions.The blue shift appeared to be a result of deprotonation of the carboxylic acid when anchored onto the titanium oxide surface.Similar to the case of ethanol effect on the absorption compared with THF, it is anticipated that deprotonation of the carboxylic group would lower its acceptability for electrons and thus would lead to the observed blue shift upon anchoring with TiO 2 .Additionally, a possible formation of H-aggregation for these dyes in the surface of TiO 2 would also lead to the observed blue shift [42].

Effect of the Adsorption Solvent on DSSC Performance.
The adsorption solvent has an indispensable role in the efficiency of solar cells that influences the dye loading on the TiO 2 surface and also has an important role in the formation of dye solvent-TiO 2 complex [43].The cell performances of the PT1, 2, and 3 dyes using EtOH and THF as dyeadsorption solvents are listed in Table 2.All the photovoltaic parameters, such as the  sc ,  oc , and FF, were affected by the adsorption solvent.THF provides lower photovoltaic values compared with the EtOH measured under AM 1.5 simulated sunlight illuminations with the light intensity 100 mW/cm 2 conditions.The use of THF as adsorption solvent decreases the photovoltaic efficiencies with respect to EtOH, which can be ascribed to the less dye loading from THF solution on the TiO 2 surface relative to EtOH, as shown in Figure 3.The use of THF as adsorption solvent decreases the photovoltaic efficiencies compared with EtOH case, which could be ascribed to the less dye loading from THF solution on the TiO 2 surface as shown in Figure 3. Compared with THF, EtOH seems favorable medium for dye loading onto TiO 2 owing to its high polarity that would facilitate better dye adsorption onto TiO 2 surface.The dye loading for PT1-3 in EtOH and THF is 439 nmol/cm 2 (PT1-EtOH), 432 nmol/cm 2 (PT2-EtOH), 337 nmol/cm 2 (PT3-EtOH), 288 nmol/cm 2 (PT1-THF), 266 nmol/cm 2 (PT2-THF), and 205 nmol/cm 2 (PT3-THF), respectively.Meanwhile, the conversion efficiency strongly depended on the dielectric constant of the dye-adsorption solvent.The amount of dye adsorption increased as the dielectric constant of the dye-adsorption solvent increased [44].The higher conversion efficiency obtained when EtOH was used as dye-adsorption solvent can also be attributed to the higher dielectric constant for ethanol ( = 24.5)compared with THF ( = 7.8).Table 2 shows that the  sc of the device using EtOH as dye-adsorption solvent are drastically higher than those using THF as dyeadsorption solvent.Thus, a higher IPCE value of DSSCs based on PT1-3 dyes was obtained using EtOH, which exceeded 80% from 425 nm to 500 nm, compared with using THF, International Journal of Photoenergy which only exceeded 50% from 425 nm to 475 nm, as shown in Figure 2(c).Meanwhile, the dye loading in case of PT3 was relatively less than PT1 and PT2 in both cases of THF and EtOH solution but the  sc values and the overall conversion efficiency were comparatively higher for PT3 more than PT1 and PT2 in both dye-bath solvents EtOH and THF, indicating that the steric hinder effect of PT3 dye with the branched aliphatic chain has the ability of decreasing the dye aggregation on the TiO 2 film as well as decreasing the electron recombination reaction rate at the interfaces between TiO 2 /dye and the electrolyte species, hence increasing the performance of the device.
To further elucidate the lower performance of the DSSC sensitized by PT1-3 that uses THF, the electron lifetime and EIS were implemented on the Autolab (Frequency range: 100 kHz to 10 mHz).Two parameters, namely, chemical capacitance (Cm) and charge recombination resistance ( rec ), were obtained by fitting the EIS at different reverse biases with Z-View software, according to the transmitting line mode.Figures 4(a) and 4(b) show the Nyquist plot and recombination resistance, where the applied voltage is 0.65 V under dark condition for the PT1-3 sensitizers using EtOH and THF as dye-adsorption solvents.In the Nyquist plots, a major semicircle was observed for all dyes, which is related to the electron recombination transport process at the interfaces between TiO 2 and the electrolyte/dye [45].The data listed in Table 3 is related to the electron lifetime (), the electrochemical capacity (), and the electron recombination ( rec ); for example, a smaller  rec indicates faster electron recombination and hence a larger dark current as shown in Figure 2(b).
The recombination resistance ( rec ) decreased in the order of PT3 (154.4Ω⋅cm 2 ) > PT2 (35.68 Ω⋅cm 2 ) > PT1 (726 mV), and PT1 (691 mV), much more than that using THF PT3 (726 mV), PT2 (676 mV), and PT1 (644 mV) as dyeadsorption solvent, and that could be ascribed to the bent and the sterically hindered conformation of phenothiazine with a branched and long aliphatic chain for PT3 > PT2 > PT1; the longer and branched aliphatic chain can prevent the direct contact between the electrolyte and the TiO 2 surface as well as inhibits or reduces dye aggregation.This behavior was clearly observed in case of using EtOH as the dye-adsorption solvent because of the higher dye loading and coverage on the TiO 2 surface and can thus reduce the charge recombination reaction rate.Figure 4(c) shows the electron lifetime for all devices.A larger value of  oc corresponds to the significantly longer electron lifetime for all sensitizers in EtOH  = 222.64ms, 56.09 ms, and 29.86 ms for PT3, PT2, and PT1, respectively, whereas, that of THF,  oc was comparatively lower than EtOH because of the comparatively lower electron lifetime  = 8452 ms, 20.84 ms, and 10.45 ms for PT3, PT2, and PT1, respectively, which can be attributed as mentioned above to the sterically hindered conformation of phenothiazine with a branched and long aliphatic chain of phenothiazine on one hand and on the other hand because of the higher dye loading when using EtOH more than that when using THF as dyeadsorption solvent.

Conclusion
Three phenothiazine-based sensitizers, namely, PT1, PT2, and PT3, were synthesized, and the photovoltaic performances of the three dyes were tested in EtOH and THF solution as dye-adsorption solvents, and it was found that the three sensitizers were influenced by the dye-adsorption solvent condition.The highest conversion efficiency 3.78% was obtained in the case of using EtOH as dye-bath solvent.The conversion efficiency of the DSSC was found to strongly depend not only on the kind of solvent but also on the steric hindered effect of the dye structure.The DSSCs that used dye with higher steric hindered effect showed higher  sc value and higher overall conversion efficiency than the DSSCs that used dye with lower steric hindered effect.
The device that used THF as a dye-bath solvent showed lower  sc value and lower conversion efficiency than the device that used EtOH, due to increase in the dye aggregates at the TiO 2 surface during the adsorption process in case of using THF and consequence lowering the overall cell performance which is attributed mainly to less dye loading on the TiO 2 surface in case of THF, the lower dielectric constant for the THF, the hydrogen bond interaction of the oxygen of the tetrahydrofuran, and the hydrogen of the carboxylic acid moiety of the dye.This study revealed that designing of phenothiazine dyes with longer and branched aliphatic chain increases the steric hindered effect which was mainly responsible for increasing the electron lifetime and decreases the dye aggregation as well as increasing the electron recombination resistance rate at the TiO 2 -dye-electrolyte interface, hence, enhancing the overall cell performance.Moreover the dye-bath solvent has a significant role for the overall cell performance and choosing a compatible dye-adsorption solvent leads to enhancement of the performance of DSSCs.

2 .
The TiO 2 paste was made on FTO glasses by screen printing.

Figure 3 :
Figure 3: UV-vis absorption spectra of PT1-3 dyes extracted from the surface of TiO 2 film in EtOH and THF.

Figure 4 :
Figure 4: (a) Electrochemical impedance spectra (Nyquist plots) applied voltage is 0.65 V under dark condition, (b) recombination resistance ( rec ), and (c) lifetime for the DSSCs using EtOH and THF as a dye bath solvent, respectively.

Table 1 :
Absorption properties of PT1-3 dyes in the EtOH and THF solutions.
).The IPCE values of DSSCs based on PT1-3 dyes when using EtOH exceeded 80% from 425 nm to 500 nm (with the highest value

Table 3 :
Recombination resistance, electron life time, and the electrochemical capacity for the PT1-3-sensitizers using EtOH and THF as dye-adsorption solvents.Ω⋅cm 2 ) in EtOH, consistent with the values of the  oc that significantly increased using EtOH PT3 (779 mV), PT2