A Quarterthiophene-Based Dye as an Efficient Interface Modifier for Hybrid Titanium Dioxide/Poly(3-hexylthiophene)(P3HT) Solar Cells

This work focused on studying the influence of dyes, including a thiophene derivative dye with a cyanoacrylic acid group ((E)-2-cyano-3-(3′,3′′,3′′′-trihexyl-[2,2′:5′,2′′:5′′,2′′′- quaterthiophene]-5-yl) acrylicacid)(4T), on the photovoltaic performance of titanium dioxide (TiO2)/poly(3-hexyl thiophene)(P3HT) solar cells. The insertion of dye at the interface improved the efficiency regardless of the dye used. However, 4T dye significantly improved the efficiency by a factor of three when compared to the corresponding control. This improvement is mainly due to an increase in short circuit current density (JSC), which is consistent with higher hole-mobility reported in TiO2/P3HT nanocomposite with 4T dye. Optical absorption data further revealed that 4T extended the spectral response of the TiO2/P3HT nanocomposite, which could also enhance the JSC. The reduced dark current upon dye insertion ensured the carrier recombination was controlled at the interface. This, in turn, increased the open circuit voltage. An optimized hybrid TiO2/P3HT device with 4T dye as an interface modifier showed an average efficiency of over 2% under-simulated irradiation of 100 mWcm−2 (1 sun) with an Air Mass 1.5 filter.


Introduction
Hybrid nanoporous metal-oxide polymer photovoltaic devices have been intensively studied for more than two decades, as these devices offer potential advantages relative to organic acceptors, such as low cost, facile synthesis via wet chemical processing, control of heterojunction morphology, and the potential for higher physical and chemical stabilities [1]. A metal-oxide nanoparticle (TiO 2 , ZnO) percolation network with thickness in the submicron scale provides a stable and transparent backbone network for free carrier transport in this type of solar cells [2]. However, the power conversion efficiency (PCE) of these hybrid devices is limited due to several reasons, including interfacial carrier recombination [3,4] at the interface, poor mobilities in the metal-oxide polymer nanocomposite, and poor spectral response of the polymer [5][6][7][8]. Typically, the nanoporous metal oxides are the electron acceptors and the π-conjugated polymers are the donors [9][10][11] in hybrid metal-oxide polymer solar cells. The electron transfer from a donor into an acceptor produces a large proportion of charge carrier pairs across the donor/acceptor interface. In that situation, the Coulombic attraction of these bound charge carrier pairs limit the device performance by feeding the recombination effects at the interface [8,[12][13][14]. It has been shown that engineering the metal-oxide polymer interface can improve the PCE of hybrid solar cells [12,[15][16][17]. Using nanolayers of absorber materials could improve the spectral response and reduce the interfacial recombination [18,19]. Organic dye molecules were also widely investigated as the interface modifier for metal-oxide polymer solar cells. In addition to a number of natural dyes [20][21][22][23][24], N719 and Z907 are two of the most common Ruthenium-based dyes successfully used as an absorber material in highly efficient dye-sensitized solar cells [25,26]. These dyes were also efficiently used as an interface modifier in solid-state hybrid solar cells and were found to improve the spectral response by participating in carrier generation, limiting the recombination [27] at the interface and hence improving both short circuit current density (J SC ) and the open circuit voltage (V OC ).
Planells M. et al. reported a series of thiophene derivative dyes with a cyanoacrylic acid group with conjugation length from one to five thiophene units (1T to 5T) as interface modifiers at TiO 2 /P3HT solar cells [16]. These dyes are metal ion-free dyes and have an electron-rich thiophene group. It was found that the dyes improve the V OC due to a dipole moment at the interface [16,28]. Oligothiophenes are discrete, monodisperse molecules, and are distinct from polythiophene, which inherently exists as a distribution of molecular weights. A pure carboxylated oligothiophene can be isolated from any unfunctionalized oligomers via column chromatography and recrystallization [28]. Such organic semiconducting oligothiophenes have been intensively investigated and widely used in organic photovoltaic (OPVs) due to the presence of excellent charge transport properties and tunable optical/electrochemical properties [29]. These tunable electrochemical properties were successfully investigated with variation of thiophene unit and show energy gap reduction when increasing the number of thiophene units from 1T to 5T [16,30].
The 4T dye at the interface was found to increase the hole-mobility in TiO 2 /P3HT polymer nanocomposite by a magnitude of 10 times compared to the corresponding untreated nanocomposite. This is due to passivation of surface traps by the dye, as well as improved packing of the polymer with the nanocrystals through effective inter-chain interactions of 4T with P3HT [8]. The molar extinction coefficient (MEC) is an important parameter in defining the amount of material to be loaded on an electrode for maximum energy conversion, particularly at thin layers of the acceptors. It was also reported that the dyes with higher MEC can improve the stability of dye-based solar cells [31]. Given that the 4T dye can improve the performance of TiO 2 -P3HT solar cells by involvement in photocurrent generation, the amount of dye molecules at the interface needs to be optimized. It was found that the MEC of 4T dye is higher than that of N719 and Z907 dyes. This work enhanced the performance of hybrid TiO 2 /P3HT polymer solar cells by optimizing the device fabrication conditions with dyes and investigates the role of 4T dye at the metal-oxide polymer interface in enhancing the performance of hybrid TiO 2 /P3HT polymer solar cells.

Materials and Methods
Solar Cell Fabrication: The solar cells were made using indium tin oxide (ITO) coated glass substrates (12 mm × 12 mm, 10 Ω/square). All the chemicals and solvents used in this work were purchased from Sigma Aldrich. The cleaned ITO substrates were first spray-coated with a diluted solution of titanium (iv) isopropoxide and acetylacetone mixture [8] in ethanol at a substrate temperature of 500 • C and baked at the same temperature for 30 min in order to form a~50 nm-thick dense/blocking TiO 2 layer. Thereafter, a mesoporous TiO 2 layer was spin-coated on top of the dense TiO 2 with the solution (240 mg ml −1 ) of TiO 2 paste (18NRT) (Dyesol, Queanbeyan, Australia) [30,32,33] dissolved in tetrahydrofuran [8] and allowed to sinter at 450 • C for 30 min [34]. As in previous studies [8,11,35]   shows that the polymer uptake and visible light absorption of the electrode treated with 4T dye was 123 much higher than the electrodes treated with N719 and Z907 dyes. This is probably due to higher 124 MEC of 4T and better compatibility between thiophene based dye and poly(3-hexyl thiophene). The 125 combination of thiophene units of 4T with poly(3-hexyl thiophene) increased the overall thiophene 126 units in the π-conjugated system, which led to the red-shifted and broadened absorbance spectrum under the UV/visible region [37]. The influence of the concentration of the interface modifier on the device performance was examined with various concentrations of 4T dye. In each situation, the dye solutions were prepared using a 1:1 volume ratio solvent mixture of acetonitrile with tert-butanol [8,30,32]. After dye dipping, the electrodes were washed in 1:1 volume ratio mixture of acetonitrile with tert-butanol to remove excess dye in the nanoporous layer [8]. The dye-modified electrodes were first dip-coated with 2.5 mg ml −1 P3HT (Merck KGaA, Germany) and then spin-coated with (25 mg ml −1 ) P3HT solution dissolved in chlorobenzene. Next, 100 nm of Gold top contact was deposited as described in [3,18,19] by thermal evaporation under high vacuum through an Edwards E306 thermal evaporator (Moorfield, Cheshire, UK). Finally, the fabricated solar cell devices were allowed to anneal process with nitrogen medium at 120 • C for 10 min in order to improve the interfacial characteristics [36].
Optical Characterization: Absorbance spectra of the dye-coated TiO 2 films were recorded using a JENWAY 6800 UV/Vis. Spectrophotometer (OSA, UK), which was controlled using Flight Deck software. The thickness of TiO 2 and P3HT layers were recorded by field emission scanning electron microscopy (FESEM, ZEIS Sigma, UK) Electrical Characterization: The electrical characterization of both polymer and solid-state solar cells were tested, and the current-voltage curves were recorded with a computer-controlled Keithley 2400 source meter unit under the conditions of dark and 100 mW/cm 2 illuminations of the solar simulator (SCIENCE TECH, Ontario, Canada) with an Air Mass (AM) 1.5 spectral filter. The external quantum efficiency (EQE) measurements were carried out with a Monochromator (Newport, CA, USA) and a calibrated silicon photodiode (Newport, CA, USA).

Results
Figure 2a-c compares optical absorption spectra of dyes (4T, N719, and Z907) dissolved in tert-butanol and acetonitrile solution with 0.3 mM concentration, dye dip-coated nanoporous TiO 2 electrodes, and dye polymer dip-coated nanoporous TiO 2 electrodes. It is clear that the peak MEC of 4T dye is a factor of two higher than the other two standard dyes. P3HT has broader absorption spectrum in the visible region when compared to the metal-complex dyes, and the absorption spectrum of the 4T dye compliments the polymer absorption in the visible region. Figure 2 further shows that the polymer uptake and visible light absorption of the electrode treated with 4T dye was much higher than the electrodes treated with N719 and Z907 dyes. This is probably due to higher MEC of 4T and better compatibility between thiophene based dye and poly(3-hexyl thiophene). The combination of thiophene units of 4T with poly(3-hexyl thiophene) increased the overall thiophene units in the π-conjugated system, which led to the red-shifted and broadened absorbance spectrum under the UV/visible region [37].   To confirm the dimensions of the layer, we performed FESEM on the samples. A completed device was cut into two and the cross-section of the device was examined. The cross-sectional FESEM image of the completed device is shown in Figure 3. It clearly shows that thickness of the TiO 2 /4T/P3HT nanocomposite was about 780 nm, in which about 150 nm excess polymer layer could serve as an electron blocking layer (to block direct contact between TiO 2 nanoparticles and top contact).

143
2%, which was five-fold higher than that of the corresponding control TiO2/P3HT devices without 144 any dye treatment. This is mainly due to a four-fold increment in the JSC. This can be attributed to 145 increased hole-mobility of P3HT due to the insertion of 4T dye [8] at the interface between the 146 TiO2/polymer interface. Figure 4(b) shows that the dark current was significantly suppressed in dye-147 treated devices when compared with the corresponding control. This is an indication of reduced back 148 electron transfer [38]. The lowest dark current was observed in 4T dye-treated devices, and was three 149 orders of magnitude lower than that of the corresponding control device. This may suggest that the 150 metal complex dyes have a more beneficial effect in shifting up the TiO2 conduction band energy.

151
Then, the external quantum efficiency (EQE) spectra of the dye-treated and untreated TiO2/P3HT 152 devices were measured. Figure 5 illustrates the EQE spectra of all devices tested. The conversion 153 efficiency in polymer increased regardless of the dyes used.  , and subsequently the PCE. However, the devices with 4T dye-treated electrodes showed a maximum efficiency of about 2%, which was five-fold higher than that of the corresponding control TiO 2 /P3HT devices without any dye treatment. This is mainly due to a four-fold increment in the J SC . This can be attributed to increased hole-mobility of P3HT due to the insertion of 4T dye [8] at the interface between the TiO 2 /polymer interface. Figure 4b shows that the dark current was significantly suppressed in dye-treated devices when compared with the corresponding control. This is an indication of reduced back electron transfer [38]. The lowest dark current was observed in 4T dye-treated devices, and was three orders of magnitude lower than that of the corresponding control device. This may suggest that the metal complex dyes have a more beneficial effect in shifting up the TiO 2 conduction band energy. Then, the external quantum efficiency (EQE) spectra of the dye-treated and untreated TiO 2 /P3HT devices were measured. Figure 5 illustrates the EQE spectra of all devices tested. The conversion efficiency in polymer increased regardless of the dyes used. Figure 5 clearly shows that the influence of dyes N719 and Z907 carrier generations were minimal in the fabricated devices, while influence of 4T dye on carrier generation was dominant, with the peak external quantum efficiency over 60% higher than the peak absorption of 4T dye. This is probably attributed to the improved hole-mobility caused by the 4T dye and better compatibility of the oligothiophene dye with the poly(3-hexyl thiophene) polymer.

143
2%, which was five-fold higher than that of the corresponding control TiO2/P3HT devices without 144 any dye treatment. This is mainly due to a four-fold increment in the JSC. This can be attributed to 145 increased hole-mobility of P3HT due to the insertion of 4T dye [8] at the interface between the 146 TiO2/polymer interface. Figure 4(b) shows that the dark current was significantly suppressed in dye-147 treated devices when compared with the corresponding control. This is an indication of reduced back 148 electron transfer [38]. The lowest dark current was observed in 4T dye-treated devices, and was three 149 orders of magnitude lower than that of the corresponding control device. This may suggest that the 150 metal complex dyes have a more beneficial effect in shifting up the TiO2 conduction band energy.

151
Then, the external quantum efficiency (EQE) spectra of the dye-treated and untreated TiO2/P3HT 152 devices were measured. Figure 5 illustrates the EQE spectra of all devices tested. The conversion 153 efficiency in polymer increased regardless of the dyes used.     present. The 4T layer was expected to obstruct hole transfer between P3HT and TiO2, and thus to 178 localize hole-polarons in the P3HT away from the TiO2 surface. Energy levels for TiO2 [18], 4T [16], 179 and P3HT [16,18,39] in Figure 6 were directly taken from literature. Further, Figure 4 shows that dye treatment, especially 4T dye, significantly suppressed the dark current and increased both the V OC and J SC under simulated solar irradiation, relative to the case of the device without dye. This is consistent with the schematic energy band diagram of the TiO 2 /4T/P3HT device shown in the Figure 6, where deep LUMO levels of 4T relative to P3HT were present. The 4T layer was expected to obstruct hole transfer between P3HT and TiO 2 , and thus to localize hole-polarons in the P3HT away from the TiO 2 surface. Energy levels for TiO 2 [18], 4T [16], and P3HT [16,18,39] in Figure 6 were directly taken from literature.

182
As in Figure 3(a), the extinction coefficient of the 4T dye was higher than that of N719 and Z907 183 dyes. To find the optimum dye concentration for maximum energy conversions, TiO2 electrodes with 184 different concentrations of 4T dye were studied. Figure 7 summarizes the variation of photon As in Figure 3a, the extinction coefficient of the 4T dye was higher than that of N719 and Z907 dyes. To find the optimum dye concentration for maximum energy conversions, TiO 2 electrodes with different concentrations of 4T dye were studied. Figure 7 summarizes the variation of photon conversion efficiencies (PCEs) of six cells fabricated with three different concentrations of 4T dye. The average power conversion efficiency was at maximum at 0.15 mM, with a champion efficiency over 2.0%. Table 2 compares the PCEs of solar cells in this work with recently reported TiO 2 /P3HT solar cells with various interface modifiers including dyes. The Table 1 clearly shows that the 4T dye-modified devices showed the best PCE of devices with pristine P3HT and nanoporous TiO 2 electrodes. It should be noted that our devices have TiO 2 nanoparticles which did not undergo TiCl 4 treatment.

Conclusions
Three different dyes, including a metal-free 4T dye as an interface modifier on TiO 2 /P3HT solar cells, were investigated. It was found that the commercial dyes N719 and Z907 improved the performance of the solar cells by improving the hole-mobility of the polymer and by reducing the back-electron transfer at the interface. Among all the dyes used, the insertion of 4T dye improved the efficiency by five-fold, which was higher when compared to other dyes used. Optimized nanoporous TiO 2 /P3HT solar cells with 4T dye yielded maximum efficiency over 2% under 1 sun illumination with an AM 1.5 filter. This is attributed to a combination of charge carrier generation due to 4T dye, as shown by EQE spectra data, and improved morphology and mobility of the P3HT caused by the 4T.

Conflicts of Interest:
The authors declare no conflict of interest.