A study of oligothiophene–acceptor dyes in p-type dye-sensitized solar cells

Abstract

Two new dyes, E1 and E2, equipped with triphenylamine as the electron donor, oligothiophene as the linker and different electron acceptor groups, have been designed and synthesized as photosensitizers for p-type dye-sensitized solar cells (p-DSCs). A systematic study of the effect of molecular structures on the observed photophysical properties, the electron/hole recombination process, the overall performance and the interfacial charge separation was carried out. Transient absorption spectroscopy (TAS) shows that the E1 dye with a napthoilene-1,2-benzimidazole (NBI) unit as the acceptor has a longer lifetime in the reduced state than the E2 dye with a malononitrile subunit on the NiO surface.

---

Introduction

Representing an inverted device structure with respect to the conventional dye-sensitized solar cells (n-DSCs),¹ p-type dye-sensitized solar cells (p-DSCs) were first reported in 1999 by Lindquist and co-workers.² The p-type device demonstrated cathodic photocurrent through an erythrosine-sensitized porous nickel oxide electrode. Recently, this type of solar cell has attracted intense research interest from the research community, since it can be used in tandem dye-sensitized solar cells^3,4 (t-DSCs) as well as in photoelectrocatalytic solar fuel production devices.^3,5–11 The configuration of p-DSCs is quite similar to n-DSCs, where p-DSCs consist of a p-type semiconductor substrate,^12–15 a photosensitizer,^2,3,16–26 a redox-active electrolyte^27–29 and a counter electrode. As one of the crucial components, the photosensitizer undertakes the task of light harvesting and electrons extraction from the p-type semiconductor substrate (or, alternatively the process can be regarded as hole injection). The first breakthrough result of p-DSCs by Sun and co-workers in 2008 ³⁰ was achieved by employing a triphenylamine-based photosensitizer with an anchor–donor–linker–acceptor (anch–D–π–A) configuration. After this work, much effort has been devoted to the optimization of the dye structure using different subunits.^{3,23,26,31–33} An ideal photosensitizer is also expected to boost the efficiency of corresponding solid state devices.³⁴ Furthermore, in analogous solar fuel devices, the photosensitizer also represents the heart of the device to generate the energy-rich electrons to be converted into an energy-rich substance, such as hydrogen gas. Therefore, the development and detailed investigation of new photosensitizers for p-DSSCs are very central for the overall improvement of the ability to convert solar light to electricity or fuels. Design–synthesis–study–optimization of new photosensitizers is a classic and efficient strategy to seek new and efficient dyes for molecular solar cells and solar fuels. In this work, we have designed and synthesized two new organic dyes (E1 and E2) with two carboxylic acid (–COOH) anchoring groups, triphenylamine as a donor unit, oligothiophene as a linker entity and different electron-withdrawing groups as electron acceptors. Taking the organic P1 dye as the reference photosensitizer, we have systematically studied the effects of different dye structures on the solar cell performance based on photophysical properties, electrochemical and photoelectrochemical performance and transient absorption characteristics to guide future dye design. The dye structures of E1, E2 and the reference dye P1 are shown in Fig. 1.

Fig. 1 Molecular structures of the E1, E2 and P1 dyes.

Experiments

General

All commercial chemicals were used as received without further purification. ¹H and ¹³C NMR spectra were recorded on Bruker AVANCE 400 NMR and Bruker AVANCE 500 MHz spectrometers in CDCl₃, at ambient temperature, using TMS as the internal standard. Chemical shifts are reported in ppm relative to an internal standard of a residual chloroform peak (δ ppm = 7.27 ppm for ¹H NMR and 77.00 ppm for ¹³C NMR spectra, respectively). UV-vis spectra were recorded using an Ocean Optics HR2000 spectrophotometer. Electrochemical experiments were performed using a CH Instruments electrochemical workstation (model 660A). Dry solvents were obtained from an MB SPS-800 dry solvent dispenser system. Chromatographic separations were carried out on a silica gel 60 Å (35–63 μm). Cobalt(II/III) tris(4,4′-di-tert-butyl-2,2′-dipyridyl) hexafluoridephorsphate (herein referred to as Co^2+/3+(dtbpy)₃) was provided by Dyenamo AB (Sweden).

Electrochemistry

The measurements were carried out at room temperature with a conventional three-electrode configuration consisting of a platinum working electrode, a platinum wire counter electrode and a non-aqueous Ag/AgNO₃ reference electrode. The potentials were referenced using an internal ferrocene standard.

Solar cell fabrication

Mesoporous, nanostructured NiO films with a thickness of 3 μm (0.25 cm²) were prepared on FTO by screen-printing, forming the p-type semiconductor electrodes. The printed NiO films were sintered at 450 °C for 30 min. All films used in the study originated from the same NiO nanoparticle and paste batch. The NiO electrodes were sensitized in a 0.2 mM dichloromethane (DCM) solution of the dyes for 2 h. All of the sensitized NiO electrodes were washed with ethanol (EtOH) and dried. The electrodes were then sealed with a platinized FTO electrode using a 25 μm thick hot-melt film (Surlyn, Solaronix) through heating the system to 120 °C. The electrolyte used contained 0.05 M Co²⁺(dtbpy)₃, 0.05 M Co³⁺(dtbpy)₃, 0.1 M LiClO₄ in a mixed solvent consisting of propylene carbonate (PC) and acetonitrile (MeCN) (1 : 1, v/v). The electrolyte was filled into the devices via pre-drilled holes in the counter electrodes using a vacuum procedure. Finally, the holes were sealed with a Surlyn sheet and a thin glass slide by heating.

Solar cell characterization

J–V characteristics were recorded using a Keithley source/meter under AM 1.5G simulated sunlight of 100 mW cm⁻² light intensity from a Newport 300 W solar simulator. Incident photon-to-current conversion efficiency (IPCE) was obtained using monochromatic light from a system consisting of a xenon lamp, a monochromator and transmittance filters. Light filters were used for calibration of the light spectrum. Both systems were calibrated using a certified reference solar cell (IR-filtered silicon solar cell, Fraunhofer ISE, Freiburg, Germany). All solar cells were illuminated from the photocathode side and the samples were masked during the measurements with an aperture area of 0.25 cm² (0.5 × 0.5 cm²). Electron lifetime and extracted charge were determined using a custom-made “toolbox setup” using a green-light-emitting diode (Luxeon K2 star 5 W, λ_max = 530 nm) as the light source.

Transient absorption spectroscopy

The femtosecond transient spectroscopy setup has been described in detail elsewhere.³⁵ In brief, the femtosecond transient absorption spectrometer includes a Coherent Legend Ti : sapphire amplifier (1 kHz, λ = 800 nm, fwhm 100 fs). The output is split to form pump and probe beams. The desired pump wavelengths were obtained using a TOPAS white light source, and, by using neutral density filters, the energy of each pulse could be kept between 200 and 400 nJ. The white light continuum probe was obtained by focusing part of the 800 nm light on a moving CaF2 plate. Polarization of the pump was set at the magic angle, 54.7° relative to the probe. The instrument response time depends on the pump and the probe wavelengths, but is typically about 150 fs. The NiO samples were translated during the experiment to avoid photodegradation. For each scan of signal vs. time, the transient absorption for 1000 laser shots was integrated for each delay time. Between 5 and 10 scans were averaged depending on the signal amplitude. The group velocity dispersion of the probe white light was fitted to a third-order polynomial by use of a homemade MATLAB program.

Nanosecond transient absorption spectroscopy was applied and the sample was excited by a Q-switched YAG-laser/MOPO combination (Spectra Physics) that delivered ca. 10 ns pulses at 10 Hz repetition frequency. The pulses were attenuated to 2 mJ per pulse over ca. 0.3 cm². The spectrometer (Edinburgh Instruments) was coupled to a probe light from a pulsed Xe-lamp (long pass filtered at 380 nm) at a 90 degree angle from the laser excitation, while the sample was placed at 45 degrees with respect to both beams for sensitized NiO films or in 1 cm cuvettes for the solution samples. Kinetic traces were recorded using a PMT and a digital oscilloscope. Transient spectra were detected with a CCD (ANDOR).

Synthesis

tert-Butyl 4-(phenyl(4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)methyl)³⁶ (5), 5,5′′′′-dibromo-3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophene^37,38 (10) and 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-7H-benzo[de]benzo[4,5]imidazo[2,1-a]isoquinolin-7-one (12) were prepared according to a literature procedure.³⁹

4-Bromo-benzoic acid tert-butyl ester (1). Under nitrogen, potassium tert-butoxide (8.4 g, 75 mmol) was added to a round flask containing 60 mL dry tetrahydrofuran (THF). The temperature was lowered to 0 °C by adding ice; this mixture was added drop-wise to a dry THF solution (60 mL) of 4-bromobenzyl chloride (15.0 g, 68 mmol) for about one hour in duration while the temperature was not allowed to exceed 5 °C. After the addition was complete, the solution was stirred continuously at room temperature for 2 h. The mixture was extracted with diethyl ether, and the collected organic phase was dried over MgSO₄. The liquid product is almost pure with a quantitative yield. ¹H NMR (500 MHz, CDCl₃) δ ppm 7.87 (d, J = 8.8 Hz, 2H), 7.57 (d, J = 8.8 Hz, 2H), 1.62 (s, 9H).

4-(N-Phenylamino)benzoic acid tert-butyl ester (2). Aniline (4.58 g, 49 mmol) and the ester 1 (8.5 g, 33 mmol) were dissolved in 100 mL dry toluene. After the solution was degassed, potassium tert-butylate (8.2 g, 73 mmol), a Pd₂(dba)₃ catalyst (381 mg, 0.33 mmol) and [HP(t-Bu₃)]BF₄ (153 mg, 0.53 mmol) were added. Subsequently, the resulting suspension was degassed and stirred for 2 hours at 40 °C. After the reaction was completed, the solution was poured into water and the aqueous phase was acidified with hydrochloric acid. The aqueous phase was extracted with dichloromethane (DCM) and the organic phase was dried with MgSO₄. After the solvent was removed by rotary evaporation the residue was passed through a silica column for purification to remove the catalytic residues. The pure product compound 2 was obtained by crystallization in hexane with 85% yield (7.57 g). ¹H NMR (500 MHz, CDCl₃) δ ppm 7.9 (d, J = 8.8 Hz, 2H), 7.35 (t, J = 8.1 Hz, 2H), 7.18 (d, J = 8.1 Hz, 2H), 7.07 (t, J = 7.5 Hz, 1H), 7.01 (d, J = 8.8 Hz, 2H), 5.99 (s, b, NH), 1.61 (s, 9H).

N,N-Di(4-benzoic acid tert-butyl ester)phenylamine (3). The diphenylamine compound 2 (5.5 g, 20.42 mmol) and ester compound 1 (5.75 g, 22.36 mmol) were dissolved in 60 mL dry toluene. After the solution was degassed, potassium tert-butylate (4.8 mg, 42.8 mmol), a Pd₂dba₃ catalyst (231 mg, 230 μmol) and [HP(t-Bu₃)]BF₄ (100 mg, 370 μmol) were added. Then, the resulting suspension was degassed and stirred for 2 hours at room temperature and 1 hour at 80 °C. Subsequently, the mixture was poured into water and the aqueous phase was neutralized using hydrochloric acid. The aqueous phase was extracted with DCM and the organic phase was dried using MgSO₄. After the solvent was removed by rotary evaporation the crude product was purified by a silica gel column using petroleum ether (PE)/DCM (3 : 1) to afford 8.2 g (90%) of the product as a white powder. ¹H NMR (500 MHz, CDCl₃) δ ppm 7.88 (d, J = 8.8 Hz, 4H), 7.35 (t, J = 8.1 Hz, 2H), 7.19 (t, J = 8.1 Hz, 1H), 7.15 (d, J = 8.1 Hz, 2H), 7.08 (d, J = 8.8 Hz, 4H), 1.61 (s, 18H).

N,N-Di(4-benzoic acid tert-butyl ester)-4-bromo-phenylamine (4). N,N-Di(4-benzoic acid tert-butyl ester)phenylamine (3) (5.75 g, 12.90 mmol) was dissolved in dried dimethyl formamide (DMF) (50 mL) and cooled on an ice bath. A solution of N-bromosuccinimide (2.75 g, 15.4 mmol) in dried DMF (10 mL) was slowly added to the reaction mixture, then the ice bath was removed and the mixture was stirred at room temperature overnight. DMF was evaporated under vacuum and the resulting solid residue was further crystallized from EtOH to afford 5.4 g (80%) of compound 4 as a white solid. ¹H NMR (500 MHz, CDCl₃) δ ppm 7.9 (d, J = 8.8 Hz, 4H), 7.44 (t, J = 8.8 Hz, 2H), 7.08 (d, J = 8.8 Hz, 4H), 7.02 (d, J = 8.8 Hz, 2H), 1.61 (s, 18H).

tert-Butyl 4-(phenyl(4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)methyl) (5). A mixture of compound 4 (3.5 g, 6.7 mmol), PdCl₂(dppf) (0.59 g, 0.8 mmol), diborane pinacol ester (2.37 g, 9.3 mmol) and KOAc (2.5 g, 26.8 mmol) in 50 mL DMF was heated to 90 °C for 12 h. After cooling, the solvent was removed by vacuum distillation. The reaction mixture was dissolved in DCM and washed with water and then dried over anhydrous MgSO₄. After removing the solvent, the residue was purified by column chromatography (PE : EtOAc, 8 : 1, v/v) to afford 2.48 g (65% yield) of product. ¹H NMR (500 MHz, CDCl₃) δ ppm 7.88 (d, J = 8.7 Hz, 4H), 7.76 (t, J = 8.3 Hz, 2H), 7.11 (m, 6H), 1.61 (s, 18H), 1.37 (s, 12H). ¹³C NMR (500 MHz, CDCl₃) δ ppm 165.36, 150.51, 149.03, 136.24, 130.84, 126.54, 124.44, 123.09, 83.84, 80.78, 28.24, 24.87.

4-Hexylthiophene-2-boronic acid (6). 3-Hexylthiophene (10 g, 59.4 mmol) in 100 mL dry THF was cooled to −78 °C in dry ice. Butyl lithium (26.2 mL, 26.2 mmol) was added dropwise by a syringe and the mixture was stirred at −78 °C for 2 h. 2-Isopropoxy-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (14.6 mL, 71.5 mmol) was added quickly and the mixture was left to heat up to room temperature and stirred overnight. The solvent was then evaporated under vacuum to afford a colorless oil which was dissolved in 100 mL DCM, washed with water (3 × 80 mL), dried over MgSO₄, filtered off and finally the solvent was evaporated under vacuum. Chromatography (silica, petroleum ether/ethyl acetate 20/1) afforded the desired colorless oily product (13.1 g, 75%). ¹H NMR (500 MHz, CDCl₃) δ ppm 7.5 (s, 1H), 7.23 (s, 1H), 2.65 (d, J = 7.8 Hz, 2H), 1.64 (m, 2H), 1.29–1.38 (m, 18H), 0.9 (t, J = 6.8 Hz, 3H).

3′,4,4′-trihexyl-4′′-pentyl-2,2′:5′,2′′-terthiophene (7). 2,5-Dibromo-3,4-dihexylthiophene (3 g, 7.3 mmol) and 2-(4-hexylthiophen-2-yl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (6) (5.1 g, 17.3 mmol) were dissolved in 60 mL THF. The resulting solution was carefully degassed, and a Pd(PPh₃)₄ catalyst (422 mg, 5 mol%) and Na₂CO₃ (2 M, 24 mL) were added. Next, the reaction mixture was carefully degassed and stirred under a refluxing condition overnight. After the above reaction time water was poured into the mixture and the organic layer was separated using diethyl ether. The combined organic phases were dried over MgSO₄ and the solvent was removed by rotary evaporation. The crude product was purified by column chromatography (n-hexane) to give a product (3.47 g, 83%) as a bright yellow oil. ¹H NMR (500 MHz, CDCl₃) δ ppm 6.98 (s, 2H), 6.90 (s, 2H), 2.7 (t, J = 8.4 Hz, 4H), 2.63 (t, J = 7.7 Hz, 4H), 1.53–1.7 (m, 10H), 1.29–1.48 (m, 22H), 0.92 (m, 12H).

5,5′′-dibromo-3′,4,4′-trihexyl-4′′-pentyl-2,2′:5′,2′′-terthiophene (8). Bromination of the thiophene compound 7 was executed using the same procedure as previously described for compound 4. ¹H NMR (500 MHz, CDCl₃) δ ppm 6.83 (s, 2H), 2.65 (m, 4H), 2.58 (t, J = 7.7 Hz, 4H), 1.50–166 (m, 10H), 1.30–146 (m, 22H), 0.93 (m, 12H).

3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophene (9). Suzuki coupling was used to couple 2-(4-hexylthiophen-2-yl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (6) and compound 8 by a procedure similar to the one used to synthesize compound 7 to give compound 9 with 76% yield as a bright yellow oil. ¹H NMR (500 MHz, CDCl₃) δ ppm 6.99 (m, 4H), δ ppm 6.92 (s, 2H), 2.75 (m, 6H), 2.64 (m, 6H), 1.57–1.72 (m, 12H), 1.28–1.45 (m, 36H), 0.93 (m, 18H).

5,5′′′′-dibromo-3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophene (10). Bromination of thiophene 9 was performed using the same procedure as previously described for compounds 4 and 8. ¹H NMR (500 MHz, CDCl₃) δ ppm 6.98 (m, 2H), 6.84 (s, 2H), 2.73 (t, 7.8, 6H), 2.59 (t, 7.7, 6H), 1.53–1.71 (m, 12H), 1.30–1.50 (m, 36H), 0.85–0.92 (m, 18H). ¹³C NMR (500 MHz, CDCl₃) δ ppm 142.48, 140.30, 140.09, 135.42, 134.35, 129.88, 129.68, 128.42, 126.59, 108.72, 31.71, 31.67, 31.53, 30.57, 30.44, 29.69, 29.58, 29.31, 29.24, 29.06, 28.95, 28.28, 22.71, 22.67, 14.14.

Synthesis of 4-bromo-1,8-napthoilene-1,2-benzimidazole (11). To a 100 mL round-bottom flask containing 4-bromo-1,8-naphthalic anhydride (5 g, 18 mmol) in 60 mL of acetic acid, (2.4 g, 22.2 mmol) ortho phenylenediamine was added. The mixture was refluxed overnight. After the reaction time, the mixture was poured into ice water. The resulting precipitate was filtered and purified by recrystallization from acetone. A bright yellow powder of the product was obtained (6.3 g, quantitative). NMR spectra show an exact 50 : 50 mixture of the two isomers. The polarity of these two isomers is very similar and is hard to purify; therefore, the mixture was directly used in the next reaction. ¹H NMR (500 MHz, CDCl₃) δ ppm 8.87 (d, J = 7.3 Hz, 1H), 8.8 (d, J = 7.3 Hz, 1H), 8.63 (m, 2H), 8.56 (d, J = 8.1 Hz, 1H), 8.46–8.53 (m, 3H), 8.08–8.12 (m, 2H), 7.88–7.93 (m, 4H), 7.5 (m, 4H).

4-(4,4,5,5-Tetramethyl-1,3,2-dioxaborolan-2-yl)-7H-benzo[de]benzo[4,5]imidazo[2,1-a]isoquinolin-7-one (12). The isomeric mixture of 11 (2 g, 5.72 mmol), PdCl₂(dppf) (1 g, 1.4 mmol), diborane pinacol ester (2 g, 8 mmol) and KOAc (2.25 g, 22.8 mmol) in 50 mL DMF was heated to 90 °C and was kept at that temperature for 12 h. After cooling, the solvent was removed by vacuum distillation. The reaction mixture was dissolved in DCM, washed with water and then dried over anhydrous MgSO₄. After removing the solvent, the residue was purified by long column chromatography (PE : EtOAc; 30 : 1 to 10 : 1; v/v) to afford 0.59 g (52% yield) of pure product. ¹H NMR (500 MHz, CDCl₃) δ ppm 9.08 (d, J = 8.5 Hz, 1H), 8.86 (d, J = 7.3 Hz, 1H), 8.78 (d, J = 7.3 Hz, 1H), 8.53–8.59 (m, 1H), 8.39 (t, J = 7.3 Hz, 1H), 7.87–7.93 (m, 1H), 7.84 (t, J = 8 Hz, 1H), 7.46–7.53 (m, 2H), 1.5 (s, 12H). ¹³C NMR (500 MHz, CDCl₃) δ ppm 160.79, 149.60, 143.89, 135.95, 135.67, 132.84, 131.81, 130.14, 127.46, 126.86, 126.78, 125.77, 125.28, 125.13, 120.53, 119.91, 115.86, 84.67, 25.02.

Di-tert-butyl-4,4′-((4-(3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-5′′′′-(7-oxo-7H-benzo[de]benzo[4,5]imidazo[2,1-a]isoquinolin-4-yl)-[2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophen]-5-yl)phenyl)azanediyl)dibenzoate (13). A double Suzuki coupling was performed in one step, in which compound 5 (58 mg, 0.1 μmol), compound 12 (40 mg, 0.1 μmol) and compound 10 (110 mg, 0.1 μmol) were added to 10 mL THF. The resulting solution was carefully degassed, and a Pd(PPh₃)₄ catalyst (32 mg, 10 mol%) and 2 M aqueous sodium carbonate solution (1 mL) were added. Then, the reaction mixture was carefully degassed and stirred under a refluxing condition for 24 h. After the reaction, water was poured into the mixture and the organic layer was separated using diethyl ether. The combined organic phases were dried with anhydrous MgSO₄, and the solvent was removed by rotary evaporation. The crude product was purified using chromatography (PE : EtOAc 3 : 1) to afford 103 mg (68% yield) of the red product. ¹H NMR (500 MHz, CDCl₃) δ ppm 8.95 (d, J = 7.3 Hz, 1H), 8.77 (d, J = 7.3 Hz, 1H), 8.60–8.64 (m, 1H), 8.27 (d, J = 8.6 Hz, 1H), 7.90–7.96 (m, 6H), 7.84 (dd, J = 7.9 Hz, 1H), 7.51–7.55 (m, 2H), 7.42 (m, 2H), 7.12–7.21 (m, 7H), 7.05 (s, 2H), 7.02 (s, 1H), 2.67–2.90 (m, 10H), 2.46 (m, 2H), 1.08–1.79 (m, 57H), 0.87–0.99 (m, 24H), 0.8 (m, 3H). ¹³C NMR (500 MHz, CDCl₃) δ ppm 165.36, 165.28, 160.58, 150.55, 149.51, 145.46, 143.92, 142.54, 140.37, 140.22, 140.13, 139.62, 139.16, 137.01, 136.79, 134.43, 133.96, 133.76, 132.66, 131.99, 131.90, 130.73, 130.26, 130.08, 129.97, 129.93, 129.62, 128.61, 127.62, 127.58, 127.39, 127.22, 126.43, 125.88, 125.67, 125.54, 125.48, 122.99, 122.83, 122.60, 120.97, 120.02, 115.96, 80.82, 31.73, 31.67, 31.54, 31.44, 30.92, 30.66, 30.54, 30.48, 29.63, 29.55, 29.46, 29.28, 29.19, 29.08, 28.87, 28.30, 23.86, 22.69, 22.64, 222.60, 22.47, 20.82, 14.13, 13.99.

4,4′-((4-(3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-5′′′′-(7-oxo-7H-benzo[de]benzo[4,5]imidazo[2,1-a]isoquinolin-4-yl)-[2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophen]-5-yl)phenyl)-azanediyl)dibenzoic acid (E1). To a DCM solution of 13 (100 mg, 1.40 mmol), trifluoroacetic acid (243 μL, 3.1 μmol) was added (5 mL). The mixture was stirred at room temperature for 12 h. The reaction mixture was neutralized using triethylamine. After being extracted with DCM, the organic phase was dried over anhydrous MgSO₄. The crude product was purified by column chromatography (petroleum ether/ethyl acetate as the gradient eluent) to gain a light-red product (67 mg, 72%).¹H NMR (500 MHz, CDCl₃) δ ppm 8.96 (d, J = 7.3 Hz, 1H), 8.85 (d, J = 7.3 Hz, 1H), 8.61 (m, 1H), 8.27 (d, J = 8.5 Hz, 1H), 8.03–8.16 (m, 4H), 7.96 (m, 1H), 7.90 (d, J = 7.6 Hz, 1H), 7.83 (dd, J = 7.6 Hz, 1H), 7.46–7.54 (m, 4H), 7.25 (m, 7H), 7.01–7.09 (m, 3H), 2.68–2.93 (m, 10H), 2.46 (m, 2H), 1.26–1.83 (m, 42H), 1.19 (m, 6H), 0.92 (m, 15H), 0.8 (m, 3H). ¹³C NMR (500 MHz, CDCl₃) δ ppm 171.43, 160.53, 151.48, 149.46, 144.96, 143.64, 142.57, 140.37, 140.28, 140.23, 140.13, 139.67, 139.33, 136.81, 136.77, 134.43, 134.18, 133.83, 132.64, 131.94, 131.86, 131.80, 131.63, 130.95, 130.84, 130.66, 130.47, 130.10, 129.96, 129.94, 129.56, 128.61, 128.29, 127.61, 127.23, 126.36, 125.95, 125.55, 123.78, 122.89, 122.78, 120.76, 119.95, 115.95, 31.76, 31.73, 31.69, 31.56, 31.46, 30.94, 30.67, 30.55, 30.50, 29.66, 26.58, 29.49, 29.30, 29.22, 29.11, 28.91, 28.33, 22.71, 22.68, 22.66, 22.64, 22.49, 14.16. HR-MS ES (m/z): calcd for [C₉₄H₁₀₅N₃O₅S₅], 1515.6658; found, 1515.6735.

3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-[2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophene]-5-carbaldehyde (14). 3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophene (9) (1.3 g, 1.42 mmol) and DMF (126 μL, 1.63 mmol) were cooled in 1,2-dichloroethane (10 mL) on an ice bath. Phosphorous oxychloride (152 μL, 1.63 mmol) was added dropwise to the solution and was refluxed overnight. The resulting suspension was poured into a cold, saturated aqueous sodium acetate solution (20 mL) and allowed to stir for 2–3 hours. Extraction was completed with dichloromethane (3 × 20 mL), the extracted solution was dried with MgSO₄ and concentrated by rotary evaporation. Silica column chromatography purification of the residue (hexane : ethyl acetate 5 : 1) yielded 14 as a yellow solid (780 mg, 58%). ¹H NMR (500 MHz, CDCl₃) δ ppm 10.05 (s, CHO, 1H), 7.07 (s, 1H), 6.99–7.05 (m, 3H), 6.93 (s, 1H), 2.98 (t, 7.8, 2H), 2.85 (t, 7.8, 2H), 2.72–2.81 (m, 6H), 2.65 (t, 7.7, 2H), 1.57–1.78 (m, 12H), 1.31–1.54 (m, 36H), 0.84–1 (m, 18H). ¹³C NMR (500 MHz, CDCl₃) δ ppm 181.46, 153.31, 145.19, 143.68, 142.42, 140.82, 140.18, 139.52, 136.29, 136.21, 135.52, 133.52, 131.20, 130.62, 129.37, 129.01, 128.77, 128.57, 127.95, 127.09, 120.07, 31.75, 31.64, 31.56, 31.46, 30.58, 30.55, 30.45, 30.23, 29.89, 29.67, 29.36, 29.32, 29.30, 29.08, 28.51, 28.37, 28.29, 22.72, 22.69, 22.63, 14.14.

5′′′′-Bromo-3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-[2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophene]-5-carbaldehyde (15). Bromination of 14 was performed as previously described for compound 4 with a yield of 87%. ¹H NMR (500 MHz, CDCl₃) δ ppm 10.04 (s, CHO, 1H), 7.06 (s, 1H), 7.03 (s, 1H), 6.98 (s, 1H), 6.85 (s, 1H), 2.95–2.98 (m, 4H), 2. 85 (t, 7.8, 2H), 2.69–2.78 (m, 6H), 2.56–2.63 (m, 2H), 1.56–1.77 (m, 12H), 1.24–1.51 (m, 36H), 0.84–0.97 (m, 18H).

Di-tert-butyl 4,4′-((4-(5′′′′-formyl-3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-[2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophen]-5-yl)phenyl)azanediyl)dibenzoate (16). The Suzuki coupling reaction was repeated under the same conditions as described above for compound 13 with a yield of 78%. ¹H NMR (500 MHz, CDCl₃) δ ppm 10.04 (s, CHO, 1H), 7.94 (d, 7.8, 4H), 7.39–7.45 (m, 2H), 7.13–7.23 (m, 6H), 7.01–7.09 (m, 4H), 2.98 (t, 7.8, 2H), 2.57–2.88 (m, 10H), 1.57–1.78 (m, 30H), 1.26–1.52 (m, 36H), 0.84–0.99 (m, 18H). ¹³C NMR (500 MHz, CDCl₃) δ ppm 181.46, 165.34, 153.33, 150.57, 145.46, 145.14, 143.40, 142.42, 140.85, 140.22, 139.58, 139.15, 137.06, 136.21, 133.94, 133.55, 130.93, 130.25, 129.38, 129.05, 128.78, 128.66, 128.25, 127.97, 126.65, 126.44, 125.68, 123.00, 122.83, 80.77, 31.76, 31.73, 31.69, 31.62, 31.54, 31.44, 30.94, 30.56, 30.23, 29.87, 29.65, 29.31, 29.21, 29.04, 28.87, 28.50, 28.25, 22.71, 22.66, 22.60, 14.14.

4,4′-((4-(5′′′′-formyl-3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-[2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophen]-5-yl)phenyl)azanediyl)dibenzoic acid (17). The compound was synthesized using the same procedure as described above for E1. ¹H NMR (500 MHz, CDCl₃) δ ppm 11.38 (s, b, 2COOH), 10.02 (s, CHO, 1H), 8.01–8.09 (m, 4H), 7.42–7.49 (m, 2H), 7.16–7.25 (m, 6H), 6.98–7.08 (m, 4H), 2.96 (t, 7.8, 2H), 2.67–2.87 (m, 10H), 1.56–1.77 (m, 12H), 1.23–1.47 (m, 36H), 0.83–0.98 (m, 18H). ¹³C NMR (400 MHz, CDCl₃) δ ppm 180.73 (s), 170.28 (s), 169.89, 152.56, 150.20, 144.26, 143.99, 141.45, 139.88, 139.26, 138.64, 138.27, 135.83, 135.19, 135.07, 132.98, 132.53, 130.70, 129.37, 128.32, 127.84, 127.69, 127.25, 127.01, 125.19, 122.97, 121.89, 114.73, 76.30, 30.66, 30.62, 30.47, 30.40, 30.37, 29.88, 29.53, 29.39, 29.16, 28.57, 28.41, 28.22, 28.15, 28.02, 27.98, 27.46, 27.28, 21.63, 21.58, 21.53, 13.16, 13.09.

4,4′-((4-(5′′′′-(2,2-dicyanovinyl)-3′,3′′,4,4′′,4′′′,4′′′′-hexahexyl-[2,2′:5′,2′′:5′′,2′′′:5′′′,2′′′′-quinquethiophen]-5-yl)phenyl)azanediyl)dibenzoic acid (E2). A mixture of compound 17 (100 mg, 78.3 μmol) and an excess of malononitrile (10 mg, 152 μmol) was poured into a 10 mL mixture solvent of chloroform/ethanol (2 : 8, v/v). Then, a few drops of triethylamine were added and refluxed overnight. The crude sample was purified by silica gel chromatography with a ratio of EtOAc : MeOH of 100 : 1 to 10 : 1 to afford 57 mg (55% yield) of product. ¹³C NMR (400 MHz, CDCl₃): δ ppm 8.06 (d, J = 8.7 Hz, 4H), 7.84 (s, 1H), 7.45–7.52 (m, 2H), 7.20–7.26 (m, 6H), 7.12 (s, 1H), 7.04–7.09 (m, 2H), 2.57–2.94 (m, 12H), 1.56–1.78 (m, 12H), 1.26–1.53 (m, 36H), 0.84–0.98 (m, 18H). ¹³C NMR (400 MHz, CDCl₃) δ ppm 171.47, 156.39, 151.53, 147.33, 147.13, 144.27, 143.51, 141.44, 140.52, 139.74, 139.38, 138.20, 136.87, 134.03, 133.44, 131.89, 131.05, 130.49, 129.21, 129.15, 128.88, 128.72, 127.64, 127.09, 126.39, 123.48, 122.78, 115.38, 114.01, 73.34, 31.73, 31.69, 31.55, 31.50, 31.42, 31.35, 30.93, 30.54, 30.27, 30.18, 29.61, 29.46, 19.28, 29.22, 29.14, 29.08, 28.88, 28.37, 22.68, 22.66, 22.61, 22.57, 14.15, 14.08. HR-MS ES (m/z): calcd for [C₈₀H₉₇N₃O₄S₅], 1323.6083; found, 1323.6072.

Results and discussions

An efficient synthetic strategy was applied to render the two new dyes and the corresponding synthetic routes are illustrated in Schemes 1–4. The synthesis of 13 required double Suzuki cross-couplings (5) between the boronic ester of arylamine, the boronic ester of naphthalic anhydride (NPA) (12) and polythiophene (10). The subsequent hydrolysis with trifluoroacetic acid (TFA) provided the target E1 dye. In order to make the corresponding dye E2, compound 15 was first synthesized by the formylation reaction of 9 and then by bromination of 14. Thereafter, a Suzuki coupling between 5 and polythiophene bromide 15 was performed, followed by hydrolysis with TFA. Finally, the aldehyde condensation with malononitrile using the Knoevenagel reaction gave the E2 product. The overall syntheses gave high yields of two new dyes E1 and E2 (Scheme 4 and 5). The syntheses of the intermediates and building blocks are explained in more detail in the Synthesis section.

Scheme 1 Synthesis of the triphenylamine compound 5. Reagents and conditions: (i) KOtBu, THF, 5 °C; (ii) Pd₂dba₃, [HP(tBu)₃]BF₄, KOtBu, toluene, rt; (iii) Pd₂dba₃, [HP(tBu)₃]BF₄, t-BuOK, toluene, 80 °C; (iv) NBS, DMF; (v) PdCl₂(dppf), bis(pinacolato)diboron, KOAc, DMF.

Scheme 2 Synthesis of the oligothiophene building block 10. Reagents and conditions: (i) n-BuLi, THF, pinacol isopropyl borate, −78 °C; (ii) Pd(PPh₃)₄, Na₂CO₃, toluene, reflux; (iii) N-bromosuccinimide (NBS), DMF; (iv) Pd(PPh₃)₄, Na₂CO₃, toluene, reflux; (v) NBS, DMF.

Scheme 3 Synthesis of the acceptor building block 12. Reagents and conditions: (i) acetic acid, reflux; (ii) n-BuLi, THF, pinacol isopropyl borate, −78 °C.

Scheme 4 Synthetic routes of the E1 dyes.

Scheme 5 Synthetic routes of the E2 dyes.

The absorption spectra and photophysical data are displayed in Fig. 2 and Table 1, respectively. All the dyes exhibit two absorption bands, at 300–400 and 400–550 nm. The absorption at 400–550 nm corresponds to a π–π* transition (see computational results below). E1 shows a hypochromic shift in contrast to P1 and E2 caused by the weaker electron withdrawing ability of the NPA component in E1 with respect to that of the malononitrile subunits in P1 and E2. Notably, a bathochromic absorption shift and a broader absorption band of E2 are observed as compared to the absorption spectrum of P1, which can be assigned to the larger conjugated system in E2, although both dyes contain the malononitrile entity as the acceptor unit. From the intersection of the normalized absorption spectra and emission spectra, E_0–0 energies of 2.25, 2.69 and 2.26 eV were obtained for the P1, E1 and E2 dyes in a DCM solution, respectively. A comparison of the solution UV-vis absorption and emission profiles of P1, E1, and E2 is provided in Fig. 2 and S3.† Fig. 2b shows the absorption spectra of the dyes adsorbed on a NiO film. Unfortunately, it is difficult to accurately quantify the amount of adsorbed dye because the desorption process of the dyes (used for quantification) from the NiO films totally destroys all these dyes. But taking a rough estimate from the absorption spectra of the dyes in solution and on a film, and the small difference in extinction coefficients (Table 1), we can conclude that E1 and E2 show lower dye loadings than P1 on a NiO film, probably due to their more bulky molecular structure The lower dye loading for E1 and E2 has to be taken into account in the following research and dye design.

Fig. 2 Absorption spectra of the E1, E2 and P1 dyes in DCM solution (a) and on NiO film (b).

Table 1 Optical and electrochemical parameters of the E1, E2 and P1 dyes

Dye	λabs^a (ε × 10^−4 M^−1 cm^−1) nm	λabs^b nm	E0–0 (1S*)^c eV	Eox (S^+/S)^d V	Ered (S/S^−)^d V	E (S*/S^−)^e V	E (S^+/S*)^e V
a The absorption spectra were recorded from a dichloromethane solution.b The absorption spectra were recorded on a NiO film (3.5 μm).c The transition energy (E0–0) was estimated from the intersection of the normalized absorption and emission curves.d All potentials are given vs. NHE and were obtained using a scan rate of 100 mV s^−1. The ground-state oxidation potential (Eox) and reduction potential (Ered) of the dyes were determined in a THF (E1 and E2) and MeCN (P1) solution containing 0.1 M (TBA)PF6 and using CV with a 100 mV s^−1 scan rate.e The excited state potentials were estimated from E0–0 and the recorded potentials using CV according to the below relations; E1/2(S^+/S*) = E1/2(S^+/S) − E0–0(^1S*), E1/2(S*/S^−) = E1/2(S/S^−) + E0–0(^1S*).
P1	353(3.6), 487(5.7)	475	2.25	1.35	−0.89	1.36	−0.9
E1	366(5.1), 404(4.5)	440	2.69	1.05	—	—	−1.64
E2	360(6.7), 500(4.0)	500	2.26	1.14	−1.05	1.21	−1.12

---

In order to get an insight into the electron distribution of the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) of the different dyes, density functional theory (DFT) quantum chemical and time-dependent DFT (TDDFT) calculations were performed using the B3LYP⁴⁰/6-31G* functional/basis sets for all atoms and without any symmetry constraints. In order to simplify the computational process, all hexyl groups were replaced by methyl units in the E1 and E2 structures. All calculations were carried out by means of the program system Gaussian09. The TDDFT excited state calculations were carried out in an ethanol solution using the MPW1K⁴¹ functional for E1 and E2. The HOMO–LUMO spatial distribution in the E1 and E2 dyes is shown in Fig. 3 and Table S1.† The simulated absorption spectra are shown in Fig. S5† and coincide well with the experimentally recorded spectra. According to the calculations, the HOMO is primarily localized on the oligothiophene groups, while the LUMO is exclusively located on the acceptor part of the sensitizers. From the calculation results, we can conclude that the lowest π–π* excited state in E1 and E2 has a substantial charge transfer character, setting up the dyes for efficient interfacial charge separation with p-type NiO.

Fig. 3 Frontier-orbital density distribution of the E1 and E2 dyes as calculated using the B3LYP/6-31G* functional/basis sets for all atoms, without any symmetry constraints (carbon in gray, nitrogen in blue, oxygen in red, sulfur in yellow and hydrogen in white). In order to accelerate the convergence of the electronic and geometric optimization, the long hexyl chains were replaced by methyl substituents.

The electrochemical data from cyclic voltammetry (CV) results of the dyes are presented in Table 1 and Fig. S4.† All compounds exhibit a reversible one-electron redox process attributed to the oxidation of the arylamine moiety. The oxidation (E_ox) and reduction potentials (E_red) of the compounds were evaluated using the electrochemical data, and an estimate of the exited state energies (E_S*/S⁻ and E_S*/S⁺) is obtained through the combination of E_0–0 and the corresponding oxidation/reduction data. The first oxidation potential (E_ox) of P1, 1.35 V vs. the Normal Hydrogen Electrode (NHE), is obviously higher than that of E1 and E2, 1.05 V and 1.14 V vs. NHE, respectively. This means that the presence of the long oligothiophene linker significantly influences the oxidation potential of the dyes, making it less positive. Due to the weaker electron-withdrawing ability of the NBI unit in the E1 subunit as compared to that of the malononitrile moiety in E2, the E_red value of E1 is much more negative than that of E2. For E1, we cannot obtain a reliable reduction potential due to the weak and irreversible reduction behavior. However, from Table 1, we can note that the potentials of E_ox and E_S*/S⁻ for P1 and E2 are quite similar, as are the values of E_red or E_S*/S⁺. Consequently, we use the E_S*/S⁻ values for P1 and E2 and the E_ox value for E1 to get a rough estimate of the driving force of injection. Similarly, we use either the E_red or E_S*/S⁺ values to analyze the dye regeneration process. The E_ox or E_S*/S⁻ levels of the dyes range from 1.05 to 1.35 V vs. NHE, and are much more positive than the energy level of the NiO valence band (∼0.5 V vs. NHE). This makes the hole injection from the dye to NiO VB thermodynamically favorable. The E_red or E_S*/S⁺ levels are more negative than the redox potential of the cobalt redox system,^42,43 which implies an efficient dye regeneration by the electrolyte. Also, from electrochemical experiments, we can conclude that the electron-withdrawing groups induce a significant change in the reduction and oxidation potentials of the dyes, thus resulting in different absorption spectra.

In order to further evaluate the performance of these dyes in solar cell devices, p-DSC devices were fabricated. Fig. 4a shows the photocurrent–voltage (J–V) curves of the solar cells and the corresponding data have been collected in Table 2. p-DSCs based on E1 and E2 show lower photocurrents as compared to the devices based on P1, which can probably be attributed to the lower extinction coefficient and also lower dye loading of E1 and E2. The devices based on the dyes E1 and E2 show significantly higher open-circuit voltages (V_OC). The underlying reason can probably be attributed to reduced charge recombination of the NiO holes with the reduced dye and/or the electrolyte, benefiting from the more extended linker unit that may also block the surface from the electrolyte (see hole lifetime measurements below). As mentioned above, one of the performance bottlenecks is the low open-circuit potentials in p-DSCs; in this aspect our new dyes show an improved performance.

Fig. 4 (a) J–V curves and (b) IPCE spectra of devices based on the E1, E2 and P1 dyes using a Co^2+/3+(dtbpy)₃ electrolyte.

Table 2 Photovoltaic performance of DSCs based on E1, E2 and P1-sensitized NiO and a Co^2+/3+(dtbpy)₃ electrolyte^a

Dye	JSC mA cm^−2	VOC mV	ff	η %
a Light intensity: 100 mW cm^−2; solar cell area: 0.25 cm^2; thickness of NiO: 3 μm. Electrolyte: 0.05 M Co^2+(dtbpy)3, 0.05 M Co^3+(dtbpy)3, 0.1 M LiClO4 in a solvent mixture of PC and MeCN (1 : 1, v/v).
E1	0.93	320	0.44	0.130
E2	0.78	320	0.41	0.102
P1	1.18	280	0.30	0.099

---

As an electrolyte redox system, we used the Co^II/III–tris(4,4′-di-tert-butyl-2,2′-dipyridyl) (herein referred to as Co^2+/3+(dtbpy)₃) redox couple. As compared to a standard, iodine-based electrolyte (Fig. S9 and Table S3†), the lower solubility of the cobalt polypyridyl complexes in the electrolyte limits photocurrent; however, the Co^2+/3+(dtbpy)₃ system can offer a higher voltage than the I⁻/I₃⁻ redox couple due to the reduced hole recombination with the electrolyte which shifts the quasi-Fermi level positively.^5,24,44 The overall conversion efficiency of the devices based on E1 is 0.13% with a short-circuit photocurrent density (J_SC) of 0.93 mA cm⁻², an open-circuit voltage (V_OC) of 320 mV and a fill factor (ff) of 0.44. Under the same conditions, the reference P1-based devices rendered an efficiency of 0.099% with a J_SC of 1.18 mA cm⁻², a V_OC of 280 mV and an ff of 0.30.

Even though the E1-based devices render better performance as compared to the P1-based ones, the Incident Photon-to-Current Conversion Efficiency (IPCE) from E1-based devices are much lower than that from devices based on P1. The considerably higher IPCE for P1-based devices is in agreement with the higher short-circuit current in the J–V studies. However, the integrated J_SC of the P1-based device from IPCE measurements, 2.8 mA cm⁻², is much higher than the actual J_SC measured under 100 mW cm⁻² light illumination (1 sun), 1.18 mA cm⁻². In contrast, the integrated J_SC values for the E1-and E2-based devices from IPCE match the J_SC values obtained under 1 sun illumination. Thus, the devices based on the dye P1 experience much more serious recombination losses at higher light intensities. For E1 and E2, the long and bulky conjugated systems are designed to prolong the charge-separation lifetime, probably making the dye regeneration process more efficient and inhibiting charge recombination losses between the reduced dye/the redox couple and the injected hole in the NiO VB.

In order to further investigate and challenge the aforementioned hypothesis, photoelectrochemical measurements were performed to obtain hole lifetime and extracted charge of NiO in the different devices. Hole lifetime as a function of V_OC is shown in Fig. 5a. At a given V_OC value, we can note that the devices based on E1 and E2 show much longer hole lifetimes than the devices based on P1. This means that the recombination loss process in P1-based devices is much more pronounced than that in E1-and E2-based devices. This result is in good agreement with the hypothesis formulated above on the basis of photovoltaic and IPCE results. The dye design, the long and bulky conjugated linker system, in the dyes E1 and E2 really inhibits the recombination between injected holes and the redox couple. From the extracted charge plotted against open-circuit voltage (Fig. 5b), we can observe the influence of the different dyes on the valence-band position. As compared to devices based on P1, the corresponding devices based on E1 and E2 display a slightly positively shifted valence band energy, thus positively shifting the Fermi level also, which represents another factor that could give devices based on E1 and E2 a higher photovoltage than observed for devices based on P1. With the same electrolyte, devices based on E1 and E2 show a higher V_OC than the P1 dye, implying that the VB of E1-and E2-sensitized NiO also shifted positively as compared to that of the P1-based NiO electrode.

Fig. 5 (a) Hole lifetime and (b) extracted charge plotted against open-circuit voltage (V_OC) for the devices based on the dyes E1, E2 and P1.

In order to get insights into the hole injection process and the recombination process between the injected holes in the NiO VB and the reduced dye, nano-/femto-second transient absorption spectroscopy (TAS) was performed for the dyes in solution and adsorbed on NiO films (Fig. 6 and 7). For the sake of a clear comparison, P1 and E2 were first selected based on the argument that the two dyes have the same donor unit, the same acceptor entity and also similar light absorption spectra. In solution, TAS of E2 in DCM shows a ground state bleach around 520 nm, a stimulated emission signal around 620 nm as well as an absorption >650 nm immediately after excitation (Fig. 6). Within the first few picoseconds the stimulated emission undergoes a dynamic Stokes shift to 700 nm, which results in a broad net absorption band centered at 630 nm. The inset shows the kinetic traces of initial excited state relaxation on a few ps time scale and the subsequent decay to the ground state with a lifetime of 940 ps. TAS of P1 in MeCN has been reported, and our results are in agreement with those.⁴⁵ The data shows qualitatively the same behavior as E2, although the excited state absorption and stimulated emission are somewhat red-shifted, and the singlet excited state lifetime is about 34 ps. In ref. 45, the short excited state lifetime was attributed to photo-induced charge separation in P1 aggregates, forming P1˙⁻ and P1˙⁺. However, we probe in the region around 400 nm where P1˙⁻ would have shown strong absorption and the absence of induced absorption in our data around 400 nm (Fig. 6b) excludes significant P1˙⁻/P1˙⁺ formation. We instead tentatively suggest that twisting of the dicyanovinyl group in the excited state leads to an increased rate of decay to the ground state, similar to what has been proposed for related dyes.⁴⁶

Fig. 6 Transient absorption spectra of (a) E2 in DCM and (b) P1 in MeCN after excitation at 530 nm. The strong Rayleigh scattering at the excitation wavelength was omitted.

Fig. 7 Transient absorption spectra of (a) NiO–E1 (inset is the kinetic trace probed at 560 nm and fitting data), (b) NiO–E2 (inset is the kinetic trace probed at 580 nm and fitting data) and (c) NiO–P1 with LiClO₄ in PC (λ_ex = 536 nm with a laser pulse power of about 300 nJ per pulse).

When bound to NiO, P1 shows similar results to those in a previous publication.⁴⁵ The initial spectra display strong positive absorption bands at ∼420 nm and at ∼600 nm, significant for the P1˙⁻ state,⁴⁵ a ground state bleach around 500 nm and weak stimulated emission ∼680 nm (Fig. 7c). Although a large part of the hole injection appears to be ultra-fast (τ < 200 fs) the remaining excited P1 reveals slower injection components. On the time scale of 10–100 ps the signals decrease, red shift and broaden, which we attribute to the simultaneous, overlapping processes of slow hole injection from the remaining excited states ^1*P1 and charge recombination between P1˙⁻ and NiO⁽⁺⁾. In Fig. 7b instead, the transient absorption spectrum of NiO–E2 in the presence of LiClO₄ in PC is shown. The spectrum is non-distinct, but different from that in solution, with ground state bleaching and a transient absorption above 600 nm. A significant feature is the negative signal around 590 nm, attributed to stimulated emission that turns into a positive absorption during the first few ps. The spectrum after 10 ps is red-shifted compared to that in solution at the same time delay, and lacks the stimulated emission around 690 nm seen in the latter. We attribute this to formation of E2˙⁻ by hole injection into NiO. E2˙⁻ was too unstable under the conditions of spectroelectrochemistry to allow for determination of a reference spectrum (Fig. S6†). In order to confirm that the absorption feature matches that of E2˙⁻, E2 was excited with a 10 ns laser flash at 532 nm in solution with a sacrificial electron donor (triethylamine), which is expected to produce E2˙⁻. Resulting TAS at 3 μs (Fig. S8†) was indeed in good agreement with that for E2 on NiO, and the positive absorption feature at around 700 nm can therefore be assigned to E2˙⁻.

The normalized kinetic traces for P1 and E2 respectively probed at 620 nm and 700 nm are illustrated in Fig. 8. The traces were fitted with a sum of four exponential decays, giving the lifetimes presented in Table 3. Both molecules show rapid hole injection, with ca. 0.2–12 ps lifetimes (cf. also Fig. 7b inset that probes the rise of the 580 nm signal). The slower lifetime t₄ ∼ 100 ps leads mainly to decay of the signal and is instead assigned to recombination. Notably, the NiO–E2 system does not show a significantly longer charge separation lifetime as compared to that of the NiO–P1 system, indicating that the long π-conjugated linker in E2 actually does not prolong the charge separation when the dye is adsorbed on the NiO surface. This suggests that the dyes may lie down on the NiO surface, with a similarly close distance between the acceptor group and NiO.

Fig. 8 The kinetic traces for NiO–E1, NiO–E2 and NiO-P1 with LiClO₄ present (λ_ex = 536 nm).

Table 3 Time constants obtained from the multi-exponential fit of E1, E2 and P1 adsorbed on NiO in the presence of LiClO₄ in PC

Samples (probe wavelength)	τ1/ps (rise)	τ2/ps	τ3/ps	τ4/ps	τinf
NiO–E1 (700 nm)	0.326/35%	1.93/24%	20.2/12%	392/16%	13%
NiO–E2 (700 nm)	0.152/36%	1.01/17%	8.0/15%	62/31%	1%
NiO–P1 (600 nm)	0.215/39%	1.33/22%	11.5/27%	138/9%	3%

---

Subsequently, we also investigated TAS of E1 on NiO to compare with E2 and study the effect of different acceptor units on interfacial charge separation. The E1 dye shows a positive absorption from 550 nm to 750 nm after excitation at 536 nm (Fig. 7a), with non-distinct stimulated emission in the region of excitation. On the time scale of around 1 ps the stimulated emission is replaced by a positive absorption, and the isosbestic point blue-shifts, which is assigned to the formation of the reduced state E1˙⁻. The normalized kinetic traces (Fig. 8) both probed at 700 nm and time constants (Table 3) for E1 and E2 on NiO demonstrated that the E1 dye with a NBI acceptor unit showed much longer charge recombination than E2 with the common malononitrile subunit, revealing that the NBI unit could be more effective to stabilize the reduced dye radical than the malononitrile unit. The prolonged lifetime of the reduced dye, in principle, is beneficial to give effective dye regeneration. From photovoltaic data, the E1 dye gave a higher photocurrent than E2, which is in agreement with the conclusion.

Conclusion

In conclusion, we have designed and synthesized two oligomer organic dyes E1 and E2 using triphenylamine unit as the electron donors, oligothiophene unit with bulky alkyl chain as the linker entities and different electron withdrawing groups as acceptors. It is the first time to introduce a napthoilene-1,2-benzimidazole (NBI) unit as the electron acceptor group in organic dyes for p-DSCs. The narrower light absorption spectrum of E1 and the lower extinction coefficient of E2, in comparison with P1, result in lower photocurrents than for the P1 dye in solar cell devices. From charge-extraction and photovoltaic experiments, we can conclude that E1 and E2 shifts the Fermi level of NiO to more positive position in comparison with P1. Transient absorption spectroscopy (TAS) measurements were carried out in order to investigate the charge-separation kinetics. All three dyes have a similar ultrafast hole injection lifetime on the τ = 0.2–10 ps time scale. However, in spite of the longer linker system, the E2 dye does not show a prolonged dye–NiO⁽⁺⁾ charge separation lifetime when adsorbed on the NiO surface in comparison with the P1 dye. Only the bulky structure of E2 is beneficial to suppress the subsequent recombination loss process between injected holes and the electrolyte, according to hole lifetime measurements in an operative cell. Interestingly, the NBI unit in the E1 dye led to a prolonged dye–NiO⁽⁺⁾ charge separation lifetime in comparison with E2, that has the common malononitrile acceptor group. This implies that NBI is a good unit to stabilize the reduced dye, which will be beneficial to improve dye regeneration in solar cells. Therefore, we expect that efficient p-type dyes with NBI as the acceptor can be obtained, after the problem of a narrow light harvesting region of E1 is overcome.

Acknowledgements

This work was financially supported by the Swedish Research Council, the Swedish Energy Agency, the Knut and Alice Wallenberg Foundation, the Åforsk Foundation (no. 14-452), and the Stiftelsen Olle Engkvist Byggmästare. We would like to thank Prof. Licheng Sun for his valuable discussion and also his help with recording the HR-MS spectra at the Dalian University of Technology. L. Z. and B. X. acknowledge the China Scholarship Council (CSC) for the doctoral fellowship support. We also greatly thank Daniel Quentin (KTH) and Ming Cheng (KTH) for their help and laboratory experiments.

References

1. B. O’Regan and M. Gratzel, Nature, 1991, 353, 737–740.
2. J. He, H. Lindström, A. Hagfeldt and S.-E. Lindquist, J. Phys. Chem. C, 1999, 103, 8940–8943.
3. A. Nattestad, A. J. Mozer, M. K. R. Fischer, Y. B. Cheng, A. Mishra, P. Bauerle and U. Bach, Nat. Mater., 2010, 9, 31–35.
4. F. Odobel, L. c. Le Pleux, Y. Pellegrin and E. Blart, Acc. Chem. Res., 2010, 43, 1063–1071.
5. F. Odobel, Y. Pellegrin, E. A. Gibson, A. Hagfeldt, A. L. Smeigh and L. Hammarström, Coord. Chem. Rev., 2012, 256, 2414–2423.
6. F. Odobel and Y. Pellegrin, J. Phys. Chem. Lett., 2013, 4, 2551–2564.
7. G. Boschloo and A. Hagfeldt, Acc. Chem. Res., 2009, 42, 1819–1826.
8. E. A. Gibson, L. Le Pleux, J. Fortage, Y. Pellegrin, E. Blart, F. Odobel, A. Hagfeldt and G. Boschloo, Langmuir, 2012, 28, 6485–6493.
9. L. Li, L. Duan, F. Wen, C. Li, M. Wang, A. Hagfeldt and L. Sun, Chem. Commun., 2012, 48, 988–990.
10. Z. Ji, M. He, Z. Huang, U. Ozkan and Y. Wu, J. Am. Chem. Soc., 2013, 135, 11696–11699.
11. H. Tian, ChemSusChem, 2015, 8(22), 3736.
12. H. Kawazoe, M. Yasukawa, H. Hyodo, M. Kurita, H. Yanagi and H. Hosono, Nature, 1997, 389, 939–942.
13. A. Nattestad, X. Zhang, U. Bach and Y.-B. Cheng, J. Photonics Energy, 2011, 1, 011103–011109.
14. A. Renaud, B. Chavillon, L. Le Pleux, Y. Pellegrin, E. Blart, M. Boujtita, T. Pauporte, L. Cario, S. Jobic and F. Odobel, J. Mater. Chem., 2012, 22, 14353–14356.
15. D. Xiong, Z. Xu, X. Zeng, W. Zhang, W. Chen, X. Xu, M. Wang and Y.-B. Cheng, J. Mater. Chem., 2012, 22, 24760–24768.
16. S. Mori, S. Fukuda, S. Sumikura, Y. Takeda, Y. Tamaki, E. Suzuki and T. Abe, J. Phys. Chem. C, 2008, 112, 16134–16139.
17. A. Nakasa, H. Usami, S. Sumikura, S. Hasegawa, T. Koyama and E. Suzuki, Chem. Lett., 2005, 34, 500–501.
18. L. Le Pleux, A. L. Smeigh, E. Gibson, Y. Pellegrin, E. Blart, G. Boschloo, A. Hagfeldt, L. Hammarstrom and F. Odobel, Energy Environ. Sci., 2011, 4, 2075–2084.
19. Z. Ji, G. Natu, Z. Huang and Y. Wu, Energy Environ. Sci., 2011, 4, 2818–2821.
20. Z. Ji, G. Natu, Z. Huang, O. Kokhan, X. Zhang and Y. Wu, J. Phys. Chem. C, 2012, 116, 16854–16863.
21. C.-H. Chang, Y.-C. Chen, C.-Y. Hsu, H.-H. Chou and J. T. Lin, Org. Lett., 2012, 14, 4726–4729.
22. J. Warnan, J. Gardner, L. le Pleux, J. Petersson, Y. Pellegrin, E. Blart, L. Hammarström and F. Odobel, J. Phys. Chem. C, 2013, 118, 103–113.
23. Y.-S. Yen, W.-T. Chen, C.-Y. Hsu, H.-H. Chou, J. T. Lin and M.-C. P. Yeh, Org. Lett., 2011, 13, 4930–4933.
24. H. Tian, J. Oscarsson, E. Gabrielsson, S. K. Eriksson, R. Lindblad, B. Xu, Y. Hao, G. Boschloo, E. M. J. Johansson, J. M. Gardner, A. Hagfeldt, H. Rensmo and L. Sun, Sci. Rep., 2014, 4, 4282.
25. H. Tian, B. Xu, H. Chen, E. M. J. Johansson and G. Boschloo, ChemSusChem, 2014, 7(8), 2150–2153.
26. C. J. Wood, G. H. Summers and E. A. Gibson, Chem. Commun., 2015, 51, 3915–3918.
27. S. Powar, T. Daeneke, M. T. Ma, D. Fu, N. W. Duffy, G. Götz, M. Weidelener, A. Mishra, P. Bäuerle, L. Spiccia and U. Bach, Angew. Chem., Int. Ed., 2013, 52, 602–605.
28. I. R. Perera, T. Daeneke, S. Makuta, Z. Yu, Y. Tachibana, A. Mishra, P. Bäuerle, C. A. Ohlin, U. Bach and L. Spiccia, Angew. Chem., Int. Ed., 2015, 54, 3758–3762.
29. X. Xu, B. Zhang, J. Cui, D. Xiong, Y. Shen, W. Chen, L. Sun, Y. Cheng and M. Wang, Nanoscale, 2013, 5, 7963–7969.
30. P. Qin, H. Zhu, T. Edvinsson, G. Boschloo, A. Hagfeldt and L. Sun, J. Am. Chem. Soc., 2008, 130, 8570–8571.
31. A. Morandeira, J. Fortage, T. Edvinsson, L. Le Pleux, E. Blart, G. Boschloo, A. Hagfeldt, L. Hammarström and F. Odobel, J. Phys. Chem. C, 2008, 112, 1721–1728.
32. Z. Liu, D. Xiong, X. Xu, Q. Arooj, H. Wang, L. Yin, W. Li, H. Wu, Z. Zhao, W. Chen, M. Wang, F. Wang, Y.-B. Cheng and H. He, ACS Appl. Mater. Interfaces, 2014, 6, 3448–3454.
33. M. Gennari, F. Légalité, L. Zhang, Y. Pellegrin, E. Blart, J. Fortage, A. M. Brown, A. Deronzier, M.-N. Collomb, M. Boujtita, D. Jacquemin, L. Hammarström and F. Odobel, J. Phys. Chem. Lett., 2014, 5, 2254–2258.
34. L. Zhang, G. Boschloo, L. Hammarstrom and H. Tian, Phys. Chem. Chem. Phys., 2016 10.1039/c5cp05247e.
35. J. Petersson, M. Eklund, J. Davidsson and L. Hammarström, J. Phys. Chem. B, 2010, 114, 14329–14338.
36. M. Weidelener, A. Mishra, A. Nattestad, S. Powar, A. J. Mozer, E. Mena-Osteritz, Y.-B. Cheng, U. Bach and P. Bauerle, J. Mater. Chem., 2012, 22, 7366–7379.
37. A. A. El-Shehawy, N. I. Abdo, A. A. El-Barbary and J.-S. Lee, Tetrahedron Lett., 2010, 51, 4526–4529.
38. Y. Wei, Y. Yang and J.-M. Yeh, Chem. Mater., 1996, 8, 2659–2666.
39. J.-F. Lee and S. L.-C. Hsu, Polymer, 2009, 50, 5668–5674.
40. A. D. Becke, J. Chem. Phys., 1993, 98, 1372–1377.
41. B. J. Lynch, P. L. Fast, M. Harris and D. G. Truhlar, J. Phys. Chem. A, 2000, 104, 4811–4815.
42. E. A. Gibson, A. L. Smeigh, L. C. Le Pleux, L. Hammarström, F. Odobel, G. Boschloo and A. Hagfeldt, J. Phys. Chem. C, 2011, 115, 9772–9779.
43. E. A. Gibson, A. L. Smeigh, L. le Pleux, J. Fortage, G. Boschloo, E. Blart, Y. Pellegrin, F. Odobel, A. Hagfeldt and L. Hammarström, Angew. Chem., Int. Ed., 2009, 48, 4402–4405.
44. E. A. Gibson, A. L. Smeigh, L. Le Pleux, L. Hammarström, F. Odobel, G. Boschloo and A. Hagfeldt, J. Phys. Chem. C, 2011, 115, 9772–9779.
45. P. Qin, J. Wiberg, E. A. Gibson, M. Linder, L. Li, T. Brinck, A. Hagfeldt, B. Albinsson and L. Sun, J. Phys. Chem. C, 2010, 114, 4738–4748.
46. B. Zietz, E. Gabrielsson, V. Johansson, A. M. El-Zohry, L. Sun and L. Kloo, Phys. Chem. Chem. Phys., 2014, 16, 2251–2255.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra26310g

‡ These authors contribute equally.

§ Present address: Department of Chemistry, University of Isfahan, Isfahan 81746-73441, Iran.

¶ Present address: Institute of Chemical Sciences and Engineering, École Polytechnique de Fédérale de Lausanne, EPFL SB ISIC LSPM, CH G1 523, Chemin des Alambics, Station 6, CH-1015 Lausanne, Switzerland.

---