- Email: [email protected]
Rational design of triphenylamine dyes for highly efficient p-type dye sensitized solar cells
Dyes and Pigments 105 (2014) 97e104
Contents lists available at ScienceDirect
Dyes and Pigments journal homepage: www.elsevier.com/locate/dyepig
Rational design of triphenylamine dyes for highly efficient p-type dye sensitized solar cells Linna Zhu a,1, Hong Bin Yang b,1, Cheng Zhong c, Chang Ming Li a, * a
Institute for Clean Energy & Advanced Materials, Southwest University, Chongqing 400715, PR China School of Chemical and Biomedical, Nanyang Technological University, 70 Nan-yang Drive, Singapore 637457, Singapore c Department of Chemistry, Hubei Key Lab on Organic and Polymeric Optoelectronic Materials, Wuhan University, Wuhan 430072, PR China b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 5 December 2013 Received in revised form 26 January 2014 Accepted 27 January 2014 Available online 12 February 2014
Two new sensitizers based on triphenylamine-dicyanovinylene have been synthesized and used for ptype dye-sensitized solar cells. The best performance amongst solar cells was achieved by the dye with a ter-thiophene bridge ligand between carboxylic acid group and the triphenylamine part (with power conversion efficiency of 0.19%, short circuit current of 4.01 mA cm2, open circuit voltage at 144 mV, and fill factor of 0.33). Results indicate that the ter-thiophene groups in the dyes strongly affects both charge recombination and hole injection in the photoelectrode. In addition, the hexyl chains on the bridged thiophene rings also help to avoid dye aggregation on the nickel oxide film and block I in electrolyte from approaching the surface of nickel oxide, which leads to a reduction in the charge recombination between nickel oxide semiconductor and electrolyte. This study suggested that modification of the bridge moiety between triphenylamine and the carboxylic group by increasing thiophene units is a promising way for preventing charge recombination and increasing the power conversion efficiency. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Triphenylamine Dye-sensitized solar cell NiO Thiophene Charge recombination Power conversion efficiency
1. Introduction Dye sensitized solar cells (DSSCs) as one of the most promising photovoltaic devices have attracted much attention due to their low material cost, device flexibility and simple manufacturing process compared to the silicon-based photovoltaic devices [1,2]. DSSCs using sensitized n-type semiconductors such as TiO2 as a photoelectrode (n-DSSCs) are mostly studied and have achieved high photon conversion efficiency (PCE) [3e6]. As their inverse model, ptype DSSCs (p-DSSCs) injecting holes to the valence band of a p-type semiconductor such as NiO from the excited state of a sensitizer are rarely studied, and their overall efficiency is much lower than nDSSCs [7e11]. Fundamentally and practically, it is very necessary to investigate p-DSSCs for inexpensive tandem solar cells, in which both electrodes are photoactive to offer great potential for significant improvement of the existing solar cell efficiency [12e14]. In p-DSSCs, the dye molecules absorbs photons and generates holes and electrons, followed by hole-transport from the acceptor to the donor, then injecting through the carboxylic group into the
* Corresponding author. E-mail addresses: [email protected], [email protected] (C.M. Li). 1 The two authors contribute equally to this article. http://dx.doi.org/10.1016/j.dyepig.2014.01.024 0143-7208/Ó 2014 Elsevier Ltd. All rights reserved.
valence band (VB) of NiO semiconductor. Fast charge recombination between the injected holes and the reduced sensitizer is a major challenge to greatly limit the power conversion efficiencies [15e17]. Rational design of the dye molecules could help to overcome this limitation. The organic sensitizer has advantages such as high molar absorption coefficients, facile molecular design and cost-effectiveness [18]. Since Lindquist and co-workers’ first demonstration for generation of cathodic photocurrent with an erythrosine-sensitized NiO cathode [19], numerous organic sensitizers such as coumarin, perylene monoimide and porphyrin derivatives have been explored [20e22]. “Pushepull” dyes with an anchoring group on the triphenylamine donor were first developed by Sun [7]. The p-DSSCs based on these dyes exhibit high incident photon-to-current conversion efficiencies (IPCEs), among which 4(bis-{4-[5-(2,2-dicyanovinyl)-thiophene-2-yl]-phenyl}amino)benzoic acid (P1) shows the highest efficiency (h) (w0.15%) [7,11]. Inspired from this work, various “pushepull” dyes have been developed. Diketopyrrolopyrrole (DPP) based novel dyad gives promising efficiency of 0.07% with I/I-3 electrolyte, and a PCE of 0.18% is obtained with cobalt complex as a redox shuttle [9]. Yen et al. increased the amount of anchoring groups on triphenylamine donor to facilitate hole injection, resulting in h of 0.08%e0.09% [23]. Organic dyads comprising a perylene monoimide (PMI) dye connected to a naphthalene diimide (NDI) or a fullerene (C60) with a
98
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
cobalt electrolyte in p-DSSCs has accomplished a PCE of 0.14% [24]. The performance has been further increased to 1.3% when Tris(1,2diaminoethane)Cobalt(II)/(III) based electrolytes were introduced [25]. More efficient sensitizers by varying the length of the oligothiophene bridge ligands between triphenylamine donor and PMI acceptor unveil long-lived charge separated excited states to result in the most efficient p-DSSCs up to date (h ¼ 0.43%) [9]. However, most of the reported arylamine dyes show PCE still lower than 0.1% [26,27]. Quite recently we have synthesized a novel dye T1 based on a triphenylamine donor and dicyanovinyl acceptor (Scheme 1) [28]. Compared to P1 reported by Sun et al. [7], a thiophene ring is inserted between triphenylamine and carboxylic acid group. Results show that T1 outperforms P1 under the same test condition. In light of this work, herein we design and synthesize two new dyes T3 and T4, in which the conjugation length between triphenylamine and carboxylic group further increased by inserting ter-thiophene and fluorene units respectively (Scheme 1). Hexyl chains on the bridged thiophene rings in T3 and T4 are introduced to maintain solubility, which are also expected to play an important role in preventing dye aggregation and blocking of I in electrolyte approaching the NiO surface for charge recombination. The influences of different linkages toward the photovoltaic properties of these organic dyes were investigated in detail to provide valuable knowledge for rational design of high performance dyes for p-DSSCs. 2. Experimental details The 1H NMR and 13C NMR spectra were recorded on a BRUKER AVANCE 300 MHz NMR Instrument in CDCl3, CD3COCD3 and DMSO-d6, using tetramethylsilane as an internal reference. GC-Ms was recorded on a GCMS-QP2010 Plus Spectrometer. MALDI-TOF was performed on a Bruker Autoflex instrument, using 1,8,9trihydroxyanthracene as a matrix. Elemental analyses of carbon, hydrogen, and nitrogen were performed on a Carlorerba-1106 microanalyzer. UVeVis absorption spectra were recorded on Shimadzu 160A spectrophotometer. Electrochemical experiments were performed using a CH Instruments electrochemical
workstation (model 660A). The potentials are quoted against the ferrocene internal standard. 2.1. Synthesis The synthetic procedures for intermediates of T3 and T4 are shown in the electronic Supplementary information. T3 and T4 were synthesized by Suzuki Coupling reaction between brominated triphenylamine and thiophene boronic acid before finally undergoing a Knoevenagel reaction with malononitrile as reported for dye T1 (Synthetic routes shown in Scheme 2) [28]. The final products were fully characterized by 1H and 13C NMR spectroscopy, MALDI-TOF spectrometry and elemental analysis as well (detailed synthetic routes, see ESI). 2.1.1. Synthesis of dye T3 (500 -(4-(bis(4-(5-(2,2-dicyanovinyl) thiophen-2-yl)phenyl)amino)phenyl)-30 ,300 -dihexyl-[2,20 :50 ,200 terthiophene]-5-carboxylic acid) To a solution of compound 7 (200 mg, 0.22 mmol) in degassed toluene (120 mL) were added malononitrile (44 mg, 0.66 mmol) and 0.1 mL Et3N. The mixture was stirred at 90 C overnight under an Ar atmosphere. After cooled to room temperature, water (100 mL) was added, the solution was extracted with CHCl3, and dried over anhydrous sodium sulfate. The solvent was evaporated, the residue was purified by column chromatography over silica gel using EtOH/CHCl3 (1:7, v/v) as the eluent to give the final product T3 (170 mg, 77%) as a dark red solid. 1H NMR (300 MHz, CD3OCD3): d (ppm): 8.35 (brs, 1H), 7.93 (brs, 1H), 7.76e7.65 (m, 8H), 7.48e7.17 (m, 12H), 2.82 (brm, 4H), 1.71 (brm, 4H), 1.32 (brm, 12 H), 0.89 (brm, 6H). 13C NMR (75 MHz, CD3OCD3): d (ppm): 180.17, 168.43, 158.27, 155.29, 151.55, 148.33, 141.81, 141.38, 133.90, 132.51, 130.01, 127.17, 126.68, 126.18, 125.30, 124.47, 124.10, 114.41, 113.75, 75.09, 35.28, 31.50, 30.25, 30.17, 29.71, 29.46, 22.44, 13.50. MALDI-TOF-MS: m/z 1019.55 (Mþ). Anal. calcd for C59H49N5O2S5 (%): C, 69.45; H, 4.84; N, 6.86. Found: C, 69.80; H, 4.76; N, 6.39. IR, n (cm1): 2225 (C^N), 1699 (C]O), 1600, 1560 (aromatic rings).
Scheme 1. Molecular structures of T1, T2, T3 and T4.
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
99
Scheme 2. Synthetic routes for dye molecules T3 and T4, respectively.
2.1.2. Synthesis of dye T4 (7-(4-(bis(4-(5-(2,2-dicyanovinyl) thiophen-2-yl)phenyl)amino)phenyl)-9,9-dihexyl-9H-fluorene-2carboxylic acid) The synthetic procedure for dye T4 is just the same as that for dye T3 described above. 1H NMR (300 MHz, CDCl3): d (ppm): 8.18e 8.12 (m, 2H), 7.86e7.77 (m, 4H), 7.72e7.62 (m, 10H), 7.43e7.41 (m,
2H), 7.29e7.21 (m, 6H), 2.06 (m, 4H), 1.14e1.06 (m, 12 H), 0.78e0.73 (m, 6H), 0.66 (m, 4H). 13C NMR (75 MHz, CD3OCD3): d (ppm): 167.9, 156.3, 153.5, 152.5, 151.9, 149.5, 146.5, 146.4, 142.7, 141.0, 140.3, 138.4, 133.5, 129.3, 129.2, 128.8, 128.7, 127.9, 127.2, 125.4, 123.6, 122.1, 120.5, 115.3, 114.7, 76.0, 56.3, 40.8, 24.6, 23.2, 14.2. MALDITOF-MS: m/z 937.68 (Mþ). Anal. calcd for C60H51N5O2S2 (%): C,
100
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
76.81; H, 5.48; N, 7.46. Found: C, 76.93; H, 5.21; N, 7.16. IR, n (cm1): 2222 (C^N), 1704 (C]O), 1597, 1573 (aromatic rings). 2.2. Photophysical measurements UVevis absorption of samples was measured using a spectrophotometer (UV-2450, Shimadzu). For transient absorption spectroscopy and time-resolved emission experiments, an LP920-KS instrument from Edinburgh Instruments, equipped with an iCCD camera from Andor and a photomultiplier tube. 2.3. Fabrication and characterization of p-DSSCs Before fabricating NiO nanoparticle films onto FTO glass, a compact p-type photocathode blocking layer (compact NiO layer) was deposited by spin-coating nickel acetate (þ98%, aldrich) in ethanol (99.7%, Merck) solution (0.2 M) at 2000 rpm on FTO glass and sintered at 150 C for 15 min. The nickel acetate layer was thermally decomposed by the following reaction during the high temperature sintering process: Ni(CH3COO)2 þ 4O2 / NiO þ 4CO2 þ 3H2O, and formed a compact NiO layer to improve the contact between FTO and the NiO nanoparticle film. The p-type NiO nanoparticle film was prepared by two cycle “doctor-blading” the precursor solution onto the substrate, and after each “doctor-blading” process the film was sintered at 450 C for 30 min. NiO precursor solution was prepared by dissolving anhydrous NiCl2 (1 g) and F108 (1 g) into a mixture of Mill-Q water (3 mL) and ethanol (6 mL). The film thickness measured by FE-SEM was about 1.6 mm. The NiO electrode with a 0.25 cm2 geometric area was immersed in a 0.3 mM dye solution at room temperature for 16 h, followed by rinsing with dichloromethane and drying in air. The solar cells were assembled with a platinized FTO counter electrode using a thermoplastic frame (Surlyn, SX1170-25 mm thick). The redox electrolyte (containing 0.8 M LiI and 0.15 M I2 in acetonitrile) was introduced through a pre-drilled hole at the counter electrode, which was sealed afterwards. The current density (J)-voltage (V) characteristics were measured using a Keithley 2420 m in dark and under illumination of a sun 2000 solar simulator (Abet) with 100 mW cm2 AM 1.5G spectrum. The intensity of the solar simulator was calibrated by standard Si photovoltaic cell. Incident photon-to-electron conversion efficiency (IPCE) measurements were performed without bias illumination with respect to a calibrated silicon diode. The monochromic light was supplied by xenon light illuminating through a Cornerstone monochromator. A chopper was placed after the monochromator and the signal was collected by Merlin lock-in radiometry after amplification by the current preamplifier. Electrochemical impedance spectroscopy (EIS) was carried out with Solartron 1260 þ 1294 impedance analyzer under illumination at open-circuit potential in dark with AC amplitude of 10 mV over a range from 20000 to 0.02 Hz. 3. Results and discussion 3.1. Photophysical properties of T3 and T4 The UVevis absorption of T1, T3 and T4 in dichloromethane (CH2Cl2) are shown in Fig 1, and exhibiting two bands, the one at around 380 nm corresponding to the pep* transition of the whole conjugation system, and the one located at longer wavelength (w500 nm) ascribed to the charge transfer process. The absorbance of T3 and T4 in the visible region is at longer wavelength than that of T1, due to their increased conjugation length. However, the absorbance of T4 in the region around 410 nm is particularly lower than that of T3, which probably comes from the blue shift of pep*
Fig. 1. UVevis absorption of dyes T1, T3 and T4 measured in CH2Cl2 solution (0.03 mM).
transition in T4 as a result of the shorter conjugation length of fluorene group in comparison to ter-thiophene groups in T3. Generally, the molar extinction coefficient of T3 and T4 is higher than T1, indicating more photons harvested (Table 1). Cyclic voltammograms (CVs) were performed in CH2Cl2 solution, using 0.1 M tetrabutylammonium hexafluorophosphate (TBAPF6) as the supporting electrolyte. The HOMO orbitals of all the dyes studied are below the top of NiO valence band (5.04 eV), and the LUMO orbitals are above that of the redox mediator (4.15 eV), suggesting sufficient driving force for hole injection and dye regeneration of these dyes. The optical band gap of T1, T3 and T4 were determined from the intersection of absorption and emission spectra in CH2Cl2 (E0-0 ¼ 1240/l). Detailed optical parameters and electrochemical properties obtained from the measured CVs are summarized in Table 1. 3.2. Theoretical calculations Density functional theory (DFT) calculations were performed in Gaussian 09 at B3LYP/6-31þG(d) level to analyze the electron density distribution of the frontier orbitals of the dye molecules [29]. The molecular-orbital energy diagram of T1, T3 and T4 were shown in Fig. 2. The HOMO orbitals are mainly located on the triphenylamine part, and the LUMO orbitals are mainly distributed on the dicyanovinylene groups and the neighboring bridge ligands. Notably, the calculated results reveal that the HOMO orbital distribution in dye T3 is more evenly and efficiently to ensure fast intramolecular hole transfer from the acceptor to the donor as observed in dye T1. Contrarily, in T4, the HOMO orbital is located on most of the conjugation moieties but only slightly distributed on the fluorene part, leading spatially a little far distance from NiO semiconductor for more difficult hole injection from HOMO orbital of T4 to NiO. Therefore, upon photoexcitation, T3 could offer better charge injection at the metal oxide/dye interface. 3.3. Photovoltaic performances Typical currentevoltage characteristics of p-DSSCs sensitized with T3 and T4 on double-layered NiO film with I/I-3 electrolyte measured under standard AM 1.5 G conditions (100 mW cm2) are shown in Fig 3. Dye T1 was also evaluated under the same condition for comparison. It turned out that dyes dissolved in the DMF-CH2Cl2 mixture (V:V ¼ 1:5) showed the best performance with I/I 3 electrolyte. Dye T3 sensitized cells exhibits a power conversion
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
101
Table 1 Optical and electrochemical properties of dyes T3 and T4 (dye T1 was also evaluated for comparison). Dyes T1 T3 T4 a b c d e f
labs (3 (104M1 cm1))a
lem (nm)
E0-0 (eV)b
Excited state lifetime (ns)c
Eon set.ox (V vs FCþ/FC)d
EHOMO (ev)e
ELUMO (eV)f
374 (3.50) 387 (4.78) 363 (5.24)
644 587 664
2.25 2.24 2.15
3.14 3.17 2.91
0.42 0.42 0.38
5.53 5.53 5.54
3.28 3.29 3.39
506 (4.01) 504 (4.65) 514 (5.23)
Absorption spectra were measured in CH2Cl2 solution. E0-0 was determined from the intersection of absorption and emission spectra in dichloromethane. Excited state lifetime was determined from the transient emission spectra. Oxidation potentials of dyes were measured in CH2Cl2 with 0.1 M TBAPF6 with a scan rate of 50 mV s1. EHOMO was calculated by (5.1 þ E(onset.ox vs FCþ/FC)). ELUMO was calculated by EHOMO þ E0-0.
efficiency (h) of 0.19% with a high short circuit current density (Jsc) of 4.01 mA cm2 and an open circuit potential (Voc) of 0.144 V; T1 under the same condition shows h of 0.11% with Jsc of 2.82 mA cm2 and Voc of 0.125 V; while T4 delivers the lowest h of 0.06% with Jsc of only 1.69 mA cm2 and Voc of 0.123 V. Detailed data of these pDSSCs in DMF-CH2Cl2 mixture are summarized in Table 2, demonstrating that T3 outperforms the other two dyes in regard to open-circuit photovoltage and short-circuit current, which are in good agreement with their IPCE results (Fig 3(b)). Meanwhile, dyes dissolved in single solvent DMF or CH2Cl2 were also investigated, but only poor performances achieved (Fig. S2). As shown in JeV curves and IPCE spectra, by inserting the terthiophene bridge ligand between triphenylamine and the carboxylic group, both Jsc and Voc of the photoelectrode are significantly enhanced, while the inserted fluorene group leads to a prominent decrease of Jsc. This is in accordance with the calculated orbital distributions and the excited state lifetime discussed above. In general, the IPCE (also photocurrent) of a photoelectrode is determined by three parameters in terms of IPCE(l) ¼ LHE(l) Binj hc, which are the light harvesting efficiency (LHE), the charge injection efficiency (Binj) from the excited dye to the photoelectrode and the charge collection efficiency (hc) related to charge loss by recombination during charge transport. The light harvesting efficiency of different dyes sensitized NiO photoelectrodes were studied by UVe vis absorption spectra. The absorption of the dyes on NiO films together with the light harvesting efficiency (LHE) spectra (LHE is defined as 1e10A, where A is the absorbance of the photoelectrode) of the sensitized electrodes were shown in Fig. 4 [30].
Apparently, the absorbance of the dyes sensitized photoelectrode descends in the order of T3>T1>T4, especially at the region around 400 nm, and the LHE spectrum follows the same trends (Fig. 4(b)). It is noted that the absorption of these dyes on NiO films is quite different from that measured in solution (the absorption of the dyes follows T4>T3>T1 in solution), which probably due to the much lower dye loading amount of T4 on NiO film, as compared to dye T1 and T3 (Table 2). To determine the dye loading amount, the dye molecules were desorbed in tetrabutylammonium hydroxide (Bu4NOH)/DMF (0.05 M) solution for 12 h. The calculated dye loading amount of T1, T3 and T4 are 30.6, 27.1 and 24.4 nmol cm2, respectively. Obviously, for the T4 sensitized cell, the photoelectrode performance is limited by dramatically reduced dye loading amount on NiO film. According to the definition of IPCE, the differences in IPCE cannot be completely ascribed to the differences in LHE. The calculated Binj hc value is about 0.66 and 0.74 for T1 and T3 dyes sensitized photoelectrodes in the visible region [31]. Considering the same charge transfer resistance of NiO electrodes, we argue that dye T3 sensitized photoelectrode has higher hole injection efficiency at the NiO/dyeeelectrolyte interface and less charge recombination loss for transporting to external circuit. On the other hand, for dye T4, lower dye loading amount as well as the lowest excited state lifetime restricts the overall performance by refining the hole injection efficiency. To confirm the higher hole injection efficiency and the less charge recombination loss in T3 sensitized photoelectrode, impedance spectra (IS) of devices sensitized by T1 and T3 at different bias voltage in dark and at open-circuit voltage bias under
Fig. 2. Molecular orbital energy diagram of the dyes T1, T3 and T4.
102
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
Fig. 3. JeV curves (a) and IPCE spectra (b) of T1, T3 and T4 sensitized photoelectrodes. (Both spectra were obtained by dissolving the dyes in DMF:DCM 1:5 (V:V) mixed solution).
illumination were measured (Fig. 5). In dark with external bias, holes are injected from the external, followed by recombination at the metal oxide/electrolyte interface via oxidation of I ions in the electrolyte. Thus recombination resistance (Rrec) at the metal oxide/ electrolyte interface and the free charge lifetime could be obtained by analyzing the IS with equivalent circuit (as shown in Figure S3) [32,33]. Nyquist plots of NiO electrodes sensitized with T1 and T3 with 0.1 V bias are shown in Fig. 5(a), in which the semicircle diameter for the measured devices is T3>T1 in dark, implying a larger resistance of T3 against interfacial hole transfer from the NiO valence band to I than T1 based devices. The Rrec values derived from the impedance data with transmission line model are plotted in Fig. 5(b) as a function of the bias voltage. The Rrec of T1 and T3 decreases with increased bias voltage due to the increased hole concentration in the valence band. The fact that the Rrec of T3 is larger than that of T1 at the same external bias in dark suggests a
Table 2 Performances of p-type DSSCs based on T1, T3 and T4 under 1 Sun, AM 1.5G illumination, and dye loading amount of three kinds of dyes on photocathodes. Dyes Jsc(mA/cm2) Voc(mV) FF (%) PCE (%) Dye Loading Amount (nmol/cm2)a T1 T3 T4
2.82 4.01 1.69
125 144 123
31 33 29
0.11 0.19 0.06
30.6 27.1 24.4
a To determine the dye loading amount, the dye-loaded NiO photocathodes with an area of 1.0 cm2 were immersed into 3 mL 0.05 M Bu4NOH/DMF) for 12 h under dark condition to desorb the dye molecules. By comparing the UVeVis spectra of the desorbed dye solution with corresponding dye (0.003 mM) dissolved in 0.05 M Bu4NOH/DMF, loading amounts of the three dyes on NiO were calculated.
Fig. 4. (a) UVevis absorption of dye molecules on NiO films, and (b) LHE spectra of NiO photoelectrode sensitized with T1, T3 and T4 (dyes dissolved in DMF-CH2Cl2mixed solvents). The absorbance of NiO film was subtracted from the absorbance of dyes on NiO films.
higher energy barrier for recombination of hole in valence band with I in T3 sensitized cell at the same hole concentration. Considering the dye loading amount of these dyes (T1>T3), it could be concluded that dye T3 could more efficiently block I approaching the NiO surface for charge recombination, due to the alkyl chains on thiophene rings in T3. The hole lifetime is plotted in Fig. 5(c) as a function of bias voltage, which is acquired by the inversion of frequency of the maximum imaginary impedance component at the medium frequency (1/2pfmax) of the semicircle. It has been reported that the charge carrier lifetime is inversely proportional to fmax, and thus a lower fmax indicates a longer effective charge lifetime to indicate a lower charge recombination [34]. Obviously, T3 sensitized electrode has the longest hole lifetime, which is consistent with the highest recombination resistance discussed above, and the synergistic effects result in the high Voc of T3 sensitized p-DSSCs. Under illumination at open-circuit voltage bias, there is no net current flow through the cell. All the injected holes from the excited dyes are recaptured by I before being extracted to the external circuit. Meanwhile, the reduced dye is regenerated by I 3 . Thus, the semicircle of the IS in the intermediate frequency range (Fig. 5(a)) also represents the transfer resistance at the NiO/dyeeelectrolyte interface, which is the apparent combination effect of hole injection, charge recombination and the reduced dye regeneration processes [28]. The charge transfer resistance at the NiO/dyee electrolyte interface of the devices obtained from the semicircle of
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
103
excited dye molecules. Consequently, better hole injection of T3 compared to T1 results in higher I concentration and hole concentration in VB of NiO, and further leads to lower charge transfer resistance of T3 photoelectrode at the NiO/electrolyte interface under illumination. Moreover, we also tested the photovoltaic performances of T3 in the presence of different concentration of deoxycholic acid (DCA) (Figure S4). Result shows that the introduction of coadsorbent gradually decreased Jsc and Voc with increased DCA concentration. This is very likely that the amount of dye molecules adsorbed on NiO surface was reduced by the competitive coadsorption of DCA, thus leading to significant loss of light harvesting [36]. The result here could also indicate that the design motif of the dye structures are rationale and can efficiently avoid dye aggregation on NiO film. 4. Conclusions In brief, two new dyes based on triphenylamine and dicyanovinylene as electron donors and acceptors were synthesized to demonstrate rational design of dyes for favorable synergistic effects in nDSSCs. Compared to T1 reported earlier, the bridge ligand between triphenylamine and the anchoring carboxylic group are increased by incorporating ter-thiophene and fluorene units respectively, to produce dye T3 and T4. The best performance among solar cells with dyes based on an arylamine structure was achieved by dye T3 (h ¼ 0.19%, Jsc ¼ 4.01 mA cm2, Voc ¼ 144 mV, and FF ¼ 0.33). Results indicate that the ter-thiophene groups in T3 strongly affects both charge recombination and hole injection in the photoelectrode. In addition, the hexyl chains on the bridged thiophene rings could also help to avoid dye aggregation on the NiO film and block I in electrolyte from approaching the surface of NiO, which leads to reducing the charge recombination process between NiO semiconductor and electrolyte. In contrast, dye T4 introducing fluorene unit between triphenylamine and carboxylic acid shows even poorer efficiency compared to T1, perhaps mainly because of the relatively smaller dye loading amount and poorer hole injection efficiency of T4. This study suggested that modification of the bridging moiety between triphenylamine and the carboxylic group by increasing thiophene units is a promising way for preventing charge recombination and as a result increasing the power conversion efficiency. Acknowledgments
Fig. 5. (a) Electrochemical impedance spectra on NiO electrodes sensitized with the P1, T1 and T3 dyes with 0.1 V under dark condition and under 1 sun at open-circuit voltage, (b) dependence of Rrec and (c) hole lifetime on the applied bias voltage.
the IS measured at open-circuit voltage under illumination are 198 U (T1) and 105 U (T3), which are smaller than that measured at open-circuit voltage bias in the dark (300 U (T1) and 342 U (T3)) (Table S2). Similar to the n-DSSCs, the disparity of charge transfer resistance at NiO/dyeeelectrolyte interface in the dark and under illumination originates from the differentiation of the local I concentration. In dark, I is generated at the counter electrode and penetrates into the NiO film by a diffusion process [35]. Under illumination, I is formed “in situ” by dye regeneration at the NiO/ electrolyte interface. As discussed above, higher LUMO level of the dyes (3.28 eV (T1) and 3.09 eV (T3)) than the redox mediator (4.15 eV) suggests sufficient driving force for dye regeneration. Thus the I concentration could be determined by the reduced dye concentration, which is relevant to the injection ability of the
This work is partially financially Institute for Clean Energy & Advanced Materials, Southwest University, Chongqing, P.R. China. L. Zhu thanks the National Natural Science Foundation of China (No. 51203046). Hong Bin Yang is grateful for financial support from the Agency of Science, Technology and Research (A*Star), Singapore under SERC Grant No. 122 020 3053. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.dyepig.2014.01.024 References [1] O’Regan B, Grätzel M. A low-cost, high-effieciency solar cell based on dyesensitized colloidal TiO2 films. Nature 1991;353:737e40. [2] Yella A, Lee H-W, Tsao HK, Yi C, Chandiran AK, Nazeeruddin M, et al. Science 2011;334:629e34. [3] Ince M, Cardinali F, Yum J, Martínez-Díaz MV, Nazeeruddin MK, Grätzel M, et al. Convergent synthesis of near-infrared absorbing, “push-pull”, bisthiophene- substituted, Zn(II)Phthalocyanines and their application in dye sensitized solar cells. Chem Eur J 2012;18:6343e8.
104
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
[4] Yang S, Wu K, Chi Y, Cheng Y, Chou P. Tris(thiocyanate) Ruthenium(II) sensitizers with functionalized dicarboxyterpyridine for dye-sensitized solar cells. Angew Chem Int Ed 2011;50:8270e4. [5] Wu W, Yang J, Hua J, Tang J, Zhang L, Long Y, et al. Efficient and stable dyesensitized solar cells based on phenothiazine sensitizers with thiophene units. J Mater Chem 2010;20:1772e9. [6] Marszalek M, Nagane S, Ichake A, Humphry-Baker R, Paul V, Zakeeruddin SM, et al. Tuning spectral properties of phenothiazine based donorepeacceptor dyes for efficient dye-sensitized solar cells. J Mater Chem 2012;22:889e94. [7] Qin P, Zhu H, Edvinsson T, Boschloo G, Hagfeldt A, Sun L. Design of an organic chromophore for p-type dye-sensitized solar cells. J Am Chem Soc 2008;130: 8570e1. [8] Qin P, Linder M, Brinck T, Boschloo G, Hagfeldt A, Sun L. High incident photonto-current conversion efficiency of p-type dye-sensitized solar cells based on NiO and organic chromophores. Adv Mater 2009;21:2993e6. [9] Favereau L, Warnan J, Pellegrin Y, Blart E, Boujtita M, Jacquemin D, et al. Diketopyrrolopyrrole derivatives for efficient NiO-based dye-sensitized solar cells. Chem Commun 2013;49:8018e20. [10] Powar S, Wu Q, Weidelener M, Nattestad A, Hu Z, Mishra A, et al. Improved photocurrents for p-type dye-sensitized solar cells using nano-structured nickel(II) oxide microballs. Energy Environ Sci 2012;5:8896e900. [11] Li L, Gibson EA, Qin P, Boschloo G, Gorlov M, Hagfeldt A, et al. Double-layered NiO photocathodes for p-type DSSCs with record IPCE. Adv Mater 2010;22: 1759e62. [12] Nattestad A, Mozer AJ, Fischer MKR, Cheng YB, Mishra A, Bauerle P, et al. Highly efficient photocathodes for dye sensitized tandem solar cells. Nat Mater 2010;9:31e5. [13] Odobel F, Pellegrin Y. Recent advances in the sensitization of wide-band-Gap nanostructured p-type semiconductors. Photovoltaic and photocatalytic applications. J Phys Chem Lett 2013;4:2551e64. [14] Odobel F, Le Pleux L, Pellegrin Y, Blart E. New photovoltaic devices based on the sensitization of p-type semiconductors: challenges and opportunities. Acc Chem Res 2010;43:1063e71. [15] Mori S, Fukuda S, Sumikura S, Takeda Y, Tamaki Y, Suzuki E, et al. Chargetransfer processes in dye-sensitized NiO solar cells. J Phys Chem C 2008;112: 16134e9. [16] Morandeira A, Boschloo G, Hagfeldt A, Ham-marström L. Photoinduced ultrafast dynamics of coumarin 343 sensitized p-type-nanostructured NiO films. J Phys Chem B 2005;109:19403e10. [17] Borgström M, Blart E, Boschloo G, Mukhtar E, Hagfeldt A, Hammarström L, et al. Sensitized hole injection of phosphorus porphyrin into NiO: toward new photovoltaic devices. J Phys Chem B 2005;109:22928e34. [18] Chang YJ, Chou PT, Lin SY, Watanabe M, Liu ZQ, Lin JL, et al. High performance organic materials for dye-sensitized solar cells: triarylene-linked dyads with a 4-tert-Butylphenylamine donor. Chem Asian J 2012;7:572. [19] He J, Lindström H, Hagfeldt A, Lindquist SE. Dye-sensitized nanostructured ptype nickel oxide film as a photocathode for a solar cell. J Phys Chem B 1999;103:8940e3. [20] Morandeira A, Fortage J, Edvinsson T, Le Pleux L, Blart E, Boschloo G, et al. Improved photon-to-current conversion efficiency with a nanoporous p-type
[21]
[22] [23] [24]
[25]
[26]
[27]
[28]
[29] [30]
[31]
[32] [33]
[34]
[35]
[36]
NiO electrode by the use of a sensitizer-acceptor dyad. J Phys Chem C 2008;112:1721e8. Morandeira A, Boschloo G, Hagfeldt A, Hammarström L. Coumarin 343eNiO films as nanostructured photocathodes in dye-sensitized solar cells: ultrafast electron transfer, effect of the I 3 /I redox couple and mechanism of photocurrent generation. J Phys Chem C 2008;112:9530e7. Zhu H, Hagfeldt A, Boschloo G. Photoelectrochemistry of mesoporous NiO electrodes in Iodide/triiodide electrolytes. J Phys Chem C 2007;111:17455e8. Yen Y, Chen W, Hsu C, Chou H, Lin JT, Yeh MP. Arylamine-based dyes for ptype dye-sensitized solar cells. Org Lett 2011;13:4930e3. Le Pleux L, Smeigh AL, Gibson E, Pellegrin Y, Blart E, Boschloo G, et al. Synthesis, photophysical and photovoltaic investigations of acceptorfunctionalized perylene monoimide dyes for nickel oxide p-type dyesensitized solar cells. Energy Environ Sci 2011;4:2075e84. Powar S, Daeneke T, Ma MT, Fu D, Duffy NW, Götz G, et al. Highly efficient ptype dye-sensitized solar cells based on tris(1,2-diaminoethane)cobalt(II)/(III) electrolytes. Angew Chem 2013;52:602e5. Leliège A, Blanchard P, Rousseau T, Roncali J. Triphenylamine/Tetracyanobutadiene-Based D-A-D p-conjugated systems as molecular donors for organic solar cells. Org Lett 2011;13:3098e101. Yuen MY, Kui SCF, Low KH, Kwok CC, Chui SSY, Ma CW, et al. Synthesis, photophysical and electrophosphorescent properties of fluorene-based platinum(II) complexes. Chem Euro J 2010;16:14131e41. Zhu L, Yang H, Zhong C, Li CM. Modified triphenylamine-dicyanovinyl-based donoreacceptor dyes with enhanced power conversion efficiency of p-type dye-sensitized solar cells. Chem Asian J 2012;7:2791e5. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, et al. Gaussian 09, revision A.02. Wallingford CT: Gaussian, Inc.; 2009. Mozer AJ, Griffith MJ, Tsekouras G, Wagner P, Wallace GG, Mori S, et al. ZnZn porphyrin dimer-sensitized solar cells: toward 3-D light harvesting. J Am Chem Soc 2009;131:15621e3. Cai L, Tsao HN, Zhang W, Wang L, Xue Z, Grätzel M, et al. Organic sensitizers with bridged triphenylamine donor units for efficient dye-sensitized solar cells. Adv Eng Mater 2013;3:200e5. Wang Q, Moser JE, Grätzel M. Electrochemical impedance spectroscopic analysis of dye-sensitized solar cells. J Phys Chem B 2005;109:14945e53. Fabregat-Santiago F, Bisquert J, Palomares E, Otero L, Kuang D, Zakeeruddin SM, et al. Correlation between photovoltaic performance and impedance spectroscopy of dye-sensitized solar cells based on ionic liquids. J Phys Chem C 2007;111:6550e60. Yang H, Guai GH, Guo C, Song Q, Jiang SP, Wang Y, et al. NiO/Graphene composite for enhanced charge separation and collection in p-type dye sensitized solar cell. J Phys Chem C 2011;115:12209e15. Wang M, Chen P, Humphry-Baker R, Zakeeruddin SM, Grätzel MM. The influence of charge transport and recombination on the performance of dyesensitized solar cells. ChemPhysChem 2009;10:290e9. Wan Z, Jia C, Duan Y, Zhou L, Lin Y, Shi Y. Phenothiazineetriphenylamine based organic dyes containing various conjugated linkers for efficient dyesensitized solar cells. J Mater Chem 2012;22:25140e7.
Contents lists available at ScienceDirect
Dyes and Pigments journal homepage: www.elsevier.com/locate/dyepig
Rational design of triphenylamine dyes for highly efficient p-type dye sensitized solar cells Linna Zhu a,1, Hong Bin Yang b,1, Cheng Zhong c, Chang Ming Li a, * a
Institute for Clean Energy & Advanced Materials, Southwest University, Chongqing 400715, PR China School of Chemical and Biomedical, Nanyang Technological University, 70 Nan-yang Drive, Singapore 637457, Singapore c Department of Chemistry, Hubei Key Lab on Organic and Polymeric Optoelectronic Materials, Wuhan University, Wuhan 430072, PR China b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 5 December 2013 Received in revised form 26 January 2014 Accepted 27 January 2014 Available online 12 February 2014
Two new sensitizers based on triphenylamine-dicyanovinylene have been synthesized and used for ptype dye-sensitized solar cells. The best performance amongst solar cells was achieved by the dye with a ter-thiophene bridge ligand between carboxylic acid group and the triphenylamine part (with power conversion efficiency of 0.19%, short circuit current of 4.01 mA cm2, open circuit voltage at 144 mV, and fill factor of 0.33). Results indicate that the ter-thiophene groups in the dyes strongly affects both charge recombination and hole injection in the photoelectrode. In addition, the hexyl chains on the bridged thiophene rings also help to avoid dye aggregation on the nickel oxide film and block I in electrolyte from approaching the surface of nickel oxide, which leads to a reduction in the charge recombination between nickel oxide semiconductor and electrolyte. This study suggested that modification of the bridge moiety between triphenylamine and the carboxylic group by increasing thiophene units is a promising way for preventing charge recombination and increasing the power conversion efficiency. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Triphenylamine Dye-sensitized solar cell NiO Thiophene Charge recombination Power conversion efficiency
1. Introduction Dye sensitized solar cells (DSSCs) as one of the most promising photovoltaic devices have attracted much attention due to their low material cost, device flexibility and simple manufacturing process compared to the silicon-based photovoltaic devices [1,2]. DSSCs using sensitized n-type semiconductors such as TiO2 as a photoelectrode (n-DSSCs) are mostly studied and have achieved high photon conversion efficiency (PCE) [3e6]. As their inverse model, ptype DSSCs (p-DSSCs) injecting holes to the valence band of a p-type semiconductor such as NiO from the excited state of a sensitizer are rarely studied, and their overall efficiency is much lower than nDSSCs [7e11]. Fundamentally and practically, it is very necessary to investigate p-DSSCs for inexpensive tandem solar cells, in which both electrodes are photoactive to offer great potential for significant improvement of the existing solar cell efficiency [12e14]. In p-DSSCs, the dye molecules absorbs photons and generates holes and electrons, followed by hole-transport from the acceptor to the donor, then injecting through the carboxylic group into the
* Corresponding author. E-mail addresses: [email protected], [email protected] (C.M. Li). 1 The two authors contribute equally to this article. http://dx.doi.org/10.1016/j.dyepig.2014.01.024 0143-7208/Ó 2014 Elsevier Ltd. All rights reserved.
valence band (VB) of NiO semiconductor. Fast charge recombination between the injected holes and the reduced sensitizer is a major challenge to greatly limit the power conversion efficiencies [15e17]. Rational design of the dye molecules could help to overcome this limitation. The organic sensitizer has advantages such as high molar absorption coefficients, facile molecular design and cost-effectiveness [18]. Since Lindquist and co-workers’ first demonstration for generation of cathodic photocurrent with an erythrosine-sensitized NiO cathode [19], numerous organic sensitizers such as coumarin, perylene monoimide and porphyrin derivatives have been explored [20e22]. “Pushepull” dyes with an anchoring group on the triphenylamine donor were first developed by Sun [7]. The p-DSSCs based on these dyes exhibit high incident photon-to-current conversion efficiencies (IPCEs), among which 4(bis-{4-[5-(2,2-dicyanovinyl)-thiophene-2-yl]-phenyl}amino)benzoic acid (P1) shows the highest efficiency (h) (w0.15%) [7,11]. Inspired from this work, various “pushepull” dyes have been developed. Diketopyrrolopyrrole (DPP) based novel dyad gives promising efficiency of 0.07% with I/I-3 electrolyte, and a PCE of 0.18% is obtained with cobalt complex as a redox shuttle [9]. Yen et al. increased the amount of anchoring groups on triphenylamine donor to facilitate hole injection, resulting in h of 0.08%e0.09% [23]. Organic dyads comprising a perylene monoimide (PMI) dye connected to a naphthalene diimide (NDI) or a fullerene (C60) with a
98
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
cobalt electrolyte in p-DSSCs has accomplished a PCE of 0.14% [24]. The performance has been further increased to 1.3% when Tris(1,2diaminoethane)Cobalt(II)/(III) based electrolytes were introduced [25]. More efficient sensitizers by varying the length of the oligothiophene bridge ligands between triphenylamine donor and PMI acceptor unveil long-lived charge separated excited states to result in the most efficient p-DSSCs up to date (h ¼ 0.43%) [9]. However, most of the reported arylamine dyes show PCE still lower than 0.1% [26,27]. Quite recently we have synthesized a novel dye T1 based on a triphenylamine donor and dicyanovinyl acceptor (Scheme 1) [28]. Compared to P1 reported by Sun et al. [7], a thiophene ring is inserted between triphenylamine and carboxylic acid group. Results show that T1 outperforms P1 under the same test condition. In light of this work, herein we design and synthesize two new dyes T3 and T4, in which the conjugation length between triphenylamine and carboxylic group further increased by inserting ter-thiophene and fluorene units respectively (Scheme 1). Hexyl chains on the bridged thiophene rings in T3 and T4 are introduced to maintain solubility, which are also expected to play an important role in preventing dye aggregation and blocking of I in electrolyte approaching the NiO surface for charge recombination. The influences of different linkages toward the photovoltaic properties of these organic dyes were investigated in detail to provide valuable knowledge for rational design of high performance dyes for p-DSSCs. 2. Experimental details The 1H NMR and 13C NMR spectra were recorded on a BRUKER AVANCE 300 MHz NMR Instrument in CDCl3, CD3COCD3 and DMSO-d6, using tetramethylsilane as an internal reference. GC-Ms was recorded on a GCMS-QP2010 Plus Spectrometer. MALDI-TOF was performed on a Bruker Autoflex instrument, using 1,8,9trihydroxyanthracene as a matrix. Elemental analyses of carbon, hydrogen, and nitrogen were performed on a Carlorerba-1106 microanalyzer. UVeVis absorption spectra were recorded on Shimadzu 160A spectrophotometer. Electrochemical experiments were performed using a CH Instruments electrochemical
workstation (model 660A). The potentials are quoted against the ferrocene internal standard. 2.1. Synthesis The synthetic procedures for intermediates of T3 and T4 are shown in the electronic Supplementary information. T3 and T4 were synthesized by Suzuki Coupling reaction between brominated triphenylamine and thiophene boronic acid before finally undergoing a Knoevenagel reaction with malononitrile as reported for dye T1 (Synthetic routes shown in Scheme 2) [28]. The final products were fully characterized by 1H and 13C NMR spectroscopy, MALDI-TOF spectrometry and elemental analysis as well (detailed synthetic routes, see ESI). 2.1.1. Synthesis of dye T3 (500 -(4-(bis(4-(5-(2,2-dicyanovinyl) thiophen-2-yl)phenyl)amino)phenyl)-30 ,300 -dihexyl-[2,20 :50 ,200 terthiophene]-5-carboxylic acid) To a solution of compound 7 (200 mg, 0.22 mmol) in degassed toluene (120 mL) were added malononitrile (44 mg, 0.66 mmol) and 0.1 mL Et3N. The mixture was stirred at 90 C overnight under an Ar atmosphere. After cooled to room temperature, water (100 mL) was added, the solution was extracted with CHCl3, and dried over anhydrous sodium sulfate. The solvent was evaporated, the residue was purified by column chromatography over silica gel using EtOH/CHCl3 (1:7, v/v) as the eluent to give the final product T3 (170 mg, 77%) as a dark red solid. 1H NMR (300 MHz, CD3OCD3): d (ppm): 8.35 (brs, 1H), 7.93 (brs, 1H), 7.76e7.65 (m, 8H), 7.48e7.17 (m, 12H), 2.82 (brm, 4H), 1.71 (brm, 4H), 1.32 (brm, 12 H), 0.89 (brm, 6H). 13C NMR (75 MHz, CD3OCD3): d (ppm): 180.17, 168.43, 158.27, 155.29, 151.55, 148.33, 141.81, 141.38, 133.90, 132.51, 130.01, 127.17, 126.68, 126.18, 125.30, 124.47, 124.10, 114.41, 113.75, 75.09, 35.28, 31.50, 30.25, 30.17, 29.71, 29.46, 22.44, 13.50. MALDI-TOF-MS: m/z 1019.55 (Mþ). Anal. calcd for C59H49N5O2S5 (%): C, 69.45; H, 4.84; N, 6.86. Found: C, 69.80; H, 4.76; N, 6.39. IR, n (cm1): 2225 (C^N), 1699 (C]O), 1600, 1560 (aromatic rings).
Scheme 1. Molecular structures of T1, T2, T3 and T4.
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
99
Scheme 2. Synthetic routes for dye molecules T3 and T4, respectively.
2.1.2. Synthesis of dye T4 (7-(4-(bis(4-(5-(2,2-dicyanovinyl) thiophen-2-yl)phenyl)amino)phenyl)-9,9-dihexyl-9H-fluorene-2carboxylic acid) The synthetic procedure for dye T4 is just the same as that for dye T3 described above. 1H NMR (300 MHz, CDCl3): d (ppm): 8.18e 8.12 (m, 2H), 7.86e7.77 (m, 4H), 7.72e7.62 (m, 10H), 7.43e7.41 (m,
2H), 7.29e7.21 (m, 6H), 2.06 (m, 4H), 1.14e1.06 (m, 12 H), 0.78e0.73 (m, 6H), 0.66 (m, 4H). 13C NMR (75 MHz, CD3OCD3): d (ppm): 167.9, 156.3, 153.5, 152.5, 151.9, 149.5, 146.5, 146.4, 142.7, 141.0, 140.3, 138.4, 133.5, 129.3, 129.2, 128.8, 128.7, 127.9, 127.2, 125.4, 123.6, 122.1, 120.5, 115.3, 114.7, 76.0, 56.3, 40.8, 24.6, 23.2, 14.2. MALDITOF-MS: m/z 937.68 (Mþ). Anal. calcd for C60H51N5O2S2 (%): C,
100
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
76.81; H, 5.48; N, 7.46. Found: C, 76.93; H, 5.21; N, 7.16. IR, n (cm1): 2222 (C^N), 1704 (C]O), 1597, 1573 (aromatic rings). 2.2. Photophysical measurements UVevis absorption of samples was measured using a spectrophotometer (UV-2450, Shimadzu). For transient absorption spectroscopy and time-resolved emission experiments, an LP920-KS instrument from Edinburgh Instruments, equipped with an iCCD camera from Andor and a photomultiplier tube. 2.3. Fabrication and characterization of p-DSSCs Before fabricating NiO nanoparticle films onto FTO glass, a compact p-type photocathode blocking layer (compact NiO layer) was deposited by spin-coating nickel acetate (þ98%, aldrich) in ethanol (99.7%, Merck) solution (0.2 M) at 2000 rpm on FTO glass and sintered at 150 C for 15 min. The nickel acetate layer was thermally decomposed by the following reaction during the high temperature sintering process: Ni(CH3COO)2 þ 4O2 / NiO þ 4CO2 þ 3H2O, and formed a compact NiO layer to improve the contact between FTO and the NiO nanoparticle film. The p-type NiO nanoparticle film was prepared by two cycle “doctor-blading” the precursor solution onto the substrate, and after each “doctor-blading” process the film was sintered at 450 C for 30 min. NiO precursor solution was prepared by dissolving anhydrous NiCl2 (1 g) and F108 (1 g) into a mixture of Mill-Q water (3 mL) and ethanol (6 mL). The film thickness measured by FE-SEM was about 1.6 mm. The NiO electrode with a 0.25 cm2 geometric area was immersed in a 0.3 mM dye solution at room temperature for 16 h, followed by rinsing with dichloromethane and drying in air. The solar cells were assembled with a platinized FTO counter electrode using a thermoplastic frame (Surlyn, SX1170-25 mm thick). The redox electrolyte (containing 0.8 M LiI and 0.15 M I2 in acetonitrile) was introduced through a pre-drilled hole at the counter electrode, which was sealed afterwards. The current density (J)-voltage (V) characteristics were measured using a Keithley 2420 m in dark and under illumination of a sun 2000 solar simulator (Abet) with 100 mW cm2 AM 1.5G spectrum. The intensity of the solar simulator was calibrated by standard Si photovoltaic cell. Incident photon-to-electron conversion efficiency (IPCE) measurements were performed without bias illumination with respect to a calibrated silicon diode. The monochromic light was supplied by xenon light illuminating through a Cornerstone monochromator. A chopper was placed after the monochromator and the signal was collected by Merlin lock-in radiometry after amplification by the current preamplifier. Electrochemical impedance spectroscopy (EIS) was carried out with Solartron 1260 þ 1294 impedance analyzer under illumination at open-circuit potential in dark with AC amplitude of 10 mV over a range from 20000 to 0.02 Hz. 3. Results and discussion 3.1. Photophysical properties of T3 and T4 The UVevis absorption of T1, T3 and T4 in dichloromethane (CH2Cl2) are shown in Fig 1, and exhibiting two bands, the one at around 380 nm corresponding to the pep* transition of the whole conjugation system, and the one located at longer wavelength (w500 nm) ascribed to the charge transfer process. The absorbance of T3 and T4 in the visible region is at longer wavelength than that of T1, due to their increased conjugation length. However, the absorbance of T4 in the region around 410 nm is particularly lower than that of T3, which probably comes from the blue shift of pep*
Fig. 1. UVevis absorption of dyes T1, T3 and T4 measured in CH2Cl2 solution (0.03 mM).
transition in T4 as a result of the shorter conjugation length of fluorene group in comparison to ter-thiophene groups in T3. Generally, the molar extinction coefficient of T3 and T4 is higher than T1, indicating more photons harvested (Table 1). Cyclic voltammograms (CVs) were performed in CH2Cl2 solution, using 0.1 M tetrabutylammonium hexafluorophosphate (TBAPF6) as the supporting electrolyte. The HOMO orbitals of all the dyes studied are below the top of NiO valence band (5.04 eV), and the LUMO orbitals are above that of the redox mediator (4.15 eV), suggesting sufficient driving force for hole injection and dye regeneration of these dyes. The optical band gap of T1, T3 and T4 were determined from the intersection of absorption and emission spectra in CH2Cl2 (E0-0 ¼ 1240/l). Detailed optical parameters and electrochemical properties obtained from the measured CVs are summarized in Table 1. 3.2. Theoretical calculations Density functional theory (DFT) calculations were performed in Gaussian 09 at B3LYP/6-31þG(d) level to analyze the electron density distribution of the frontier orbitals of the dye molecules [29]. The molecular-orbital energy diagram of T1, T3 and T4 were shown in Fig. 2. The HOMO orbitals are mainly located on the triphenylamine part, and the LUMO orbitals are mainly distributed on the dicyanovinylene groups and the neighboring bridge ligands. Notably, the calculated results reveal that the HOMO orbital distribution in dye T3 is more evenly and efficiently to ensure fast intramolecular hole transfer from the acceptor to the donor as observed in dye T1. Contrarily, in T4, the HOMO orbital is located on most of the conjugation moieties but only slightly distributed on the fluorene part, leading spatially a little far distance from NiO semiconductor for more difficult hole injection from HOMO orbital of T4 to NiO. Therefore, upon photoexcitation, T3 could offer better charge injection at the metal oxide/dye interface. 3.3. Photovoltaic performances Typical currentevoltage characteristics of p-DSSCs sensitized with T3 and T4 on double-layered NiO film with I/I-3 electrolyte measured under standard AM 1.5 G conditions (100 mW cm2) are shown in Fig 3. Dye T1 was also evaluated under the same condition for comparison. It turned out that dyes dissolved in the DMF-CH2Cl2 mixture (V:V ¼ 1:5) showed the best performance with I/I 3 electrolyte. Dye T3 sensitized cells exhibits a power conversion
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
101
Table 1 Optical and electrochemical properties of dyes T3 and T4 (dye T1 was also evaluated for comparison). Dyes T1 T3 T4 a b c d e f
labs (3 (104M1 cm1))a
lem (nm)
E0-0 (eV)b
Excited state lifetime (ns)c
Eon set.ox (V vs FCþ/FC)d
EHOMO (ev)e
ELUMO (eV)f
374 (3.50) 387 (4.78) 363 (5.24)
644 587 664
2.25 2.24 2.15
3.14 3.17 2.91
0.42 0.42 0.38
5.53 5.53 5.54
3.28 3.29 3.39
506 (4.01) 504 (4.65) 514 (5.23)
Absorption spectra were measured in CH2Cl2 solution. E0-0 was determined from the intersection of absorption and emission spectra in dichloromethane. Excited state lifetime was determined from the transient emission spectra. Oxidation potentials of dyes were measured in CH2Cl2 with 0.1 M TBAPF6 with a scan rate of 50 mV s1. EHOMO was calculated by (5.1 þ E(onset.ox vs FCþ/FC)). ELUMO was calculated by EHOMO þ E0-0.
efficiency (h) of 0.19% with a high short circuit current density (Jsc) of 4.01 mA cm2 and an open circuit potential (Voc) of 0.144 V; T1 under the same condition shows h of 0.11% with Jsc of 2.82 mA cm2 and Voc of 0.125 V; while T4 delivers the lowest h of 0.06% with Jsc of only 1.69 mA cm2 and Voc of 0.123 V. Detailed data of these pDSSCs in DMF-CH2Cl2 mixture are summarized in Table 2, demonstrating that T3 outperforms the other two dyes in regard to open-circuit photovoltage and short-circuit current, which are in good agreement with their IPCE results (Fig 3(b)). Meanwhile, dyes dissolved in single solvent DMF or CH2Cl2 were also investigated, but only poor performances achieved (Fig. S2). As shown in JeV curves and IPCE spectra, by inserting the terthiophene bridge ligand between triphenylamine and the carboxylic group, both Jsc and Voc of the photoelectrode are significantly enhanced, while the inserted fluorene group leads to a prominent decrease of Jsc. This is in accordance with the calculated orbital distributions and the excited state lifetime discussed above. In general, the IPCE (also photocurrent) of a photoelectrode is determined by three parameters in terms of IPCE(l) ¼ LHE(l) Binj hc, which are the light harvesting efficiency (LHE), the charge injection efficiency (Binj) from the excited dye to the photoelectrode and the charge collection efficiency (hc) related to charge loss by recombination during charge transport. The light harvesting efficiency of different dyes sensitized NiO photoelectrodes were studied by UVe vis absorption spectra. The absorption of the dyes on NiO films together with the light harvesting efficiency (LHE) spectra (LHE is defined as 1e10A, where A is the absorbance of the photoelectrode) of the sensitized electrodes were shown in Fig. 4 [30].
Apparently, the absorbance of the dyes sensitized photoelectrode descends in the order of T3>T1>T4, especially at the region around 400 nm, and the LHE spectrum follows the same trends (Fig. 4(b)). It is noted that the absorption of these dyes on NiO films is quite different from that measured in solution (the absorption of the dyes follows T4>T3>T1 in solution), which probably due to the much lower dye loading amount of T4 on NiO film, as compared to dye T1 and T3 (Table 2). To determine the dye loading amount, the dye molecules were desorbed in tetrabutylammonium hydroxide (Bu4NOH)/DMF (0.05 M) solution for 12 h. The calculated dye loading amount of T1, T3 and T4 are 30.6, 27.1 and 24.4 nmol cm2, respectively. Obviously, for the T4 sensitized cell, the photoelectrode performance is limited by dramatically reduced dye loading amount on NiO film. According to the definition of IPCE, the differences in IPCE cannot be completely ascribed to the differences in LHE. The calculated Binj hc value is about 0.66 and 0.74 for T1 and T3 dyes sensitized photoelectrodes in the visible region [31]. Considering the same charge transfer resistance of NiO electrodes, we argue that dye T3 sensitized photoelectrode has higher hole injection efficiency at the NiO/dyeeelectrolyte interface and less charge recombination loss for transporting to external circuit. On the other hand, for dye T4, lower dye loading amount as well as the lowest excited state lifetime restricts the overall performance by refining the hole injection efficiency. To confirm the higher hole injection efficiency and the less charge recombination loss in T3 sensitized photoelectrode, impedance spectra (IS) of devices sensitized by T1 and T3 at different bias voltage in dark and at open-circuit voltage bias under
Fig. 2. Molecular orbital energy diagram of the dyes T1, T3 and T4.
102
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
Fig. 3. JeV curves (a) and IPCE spectra (b) of T1, T3 and T4 sensitized photoelectrodes. (Both spectra were obtained by dissolving the dyes in DMF:DCM 1:5 (V:V) mixed solution).
illumination were measured (Fig. 5). In dark with external bias, holes are injected from the external, followed by recombination at the metal oxide/electrolyte interface via oxidation of I ions in the electrolyte. Thus recombination resistance (Rrec) at the metal oxide/ electrolyte interface and the free charge lifetime could be obtained by analyzing the IS with equivalent circuit (as shown in Figure S3) [32,33]. Nyquist plots of NiO electrodes sensitized with T1 and T3 with 0.1 V bias are shown in Fig. 5(a), in which the semicircle diameter for the measured devices is T3>T1 in dark, implying a larger resistance of T3 against interfacial hole transfer from the NiO valence band to I than T1 based devices. The Rrec values derived from the impedance data with transmission line model are plotted in Fig. 5(b) as a function of the bias voltage. The Rrec of T1 and T3 decreases with increased bias voltage due to the increased hole concentration in the valence band. The fact that the Rrec of T3 is larger than that of T1 at the same external bias in dark suggests a
Table 2 Performances of p-type DSSCs based on T1, T3 and T4 under 1 Sun, AM 1.5G illumination, and dye loading amount of three kinds of dyes on photocathodes. Dyes Jsc(mA/cm2) Voc(mV) FF (%) PCE (%) Dye Loading Amount (nmol/cm2)a T1 T3 T4
2.82 4.01 1.69
125 144 123
31 33 29
0.11 0.19 0.06
30.6 27.1 24.4
a To determine the dye loading amount, the dye-loaded NiO photocathodes with an area of 1.0 cm2 were immersed into 3 mL 0.05 M Bu4NOH/DMF) for 12 h under dark condition to desorb the dye molecules. By comparing the UVeVis spectra of the desorbed dye solution with corresponding dye (0.003 mM) dissolved in 0.05 M Bu4NOH/DMF, loading amounts of the three dyes on NiO were calculated.
Fig. 4. (a) UVevis absorption of dye molecules on NiO films, and (b) LHE spectra of NiO photoelectrode sensitized with T1, T3 and T4 (dyes dissolved in DMF-CH2Cl2mixed solvents). The absorbance of NiO film was subtracted from the absorbance of dyes on NiO films.
higher energy barrier for recombination of hole in valence band with I in T3 sensitized cell at the same hole concentration. Considering the dye loading amount of these dyes (T1>T3), it could be concluded that dye T3 could more efficiently block I approaching the NiO surface for charge recombination, due to the alkyl chains on thiophene rings in T3. The hole lifetime is plotted in Fig. 5(c) as a function of bias voltage, which is acquired by the inversion of frequency of the maximum imaginary impedance component at the medium frequency (1/2pfmax) of the semicircle. It has been reported that the charge carrier lifetime is inversely proportional to fmax, and thus a lower fmax indicates a longer effective charge lifetime to indicate a lower charge recombination [34]. Obviously, T3 sensitized electrode has the longest hole lifetime, which is consistent with the highest recombination resistance discussed above, and the synergistic effects result in the high Voc of T3 sensitized p-DSSCs. Under illumination at open-circuit voltage bias, there is no net current flow through the cell. All the injected holes from the excited dyes are recaptured by I before being extracted to the external circuit. Meanwhile, the reduced dye is regenerated by I 3 . Thus, the semicircle of the IS in the intermediate frequency range (Fig. 5(a)) also represents the transfer resistance at the NiO/dyeeelectrolyte interface, which is the apparent combination effect of hole injection, charge recombination and the reduced dye regeneration processes [28]. The charge transfer resistance at the NiO/dyee electrolyte interface of the devices obtained from the semicircle of
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
103
excited dye molecules. Consequently, better hole injection of T3 compared to T1 results in higher I concentration and hole concentration in VB of NiO, and further leads to lower charge transfer resistance of T3 photoelectrode at the NiO/electrolyte interface under illumination. Moreover, we also tested the photovoltaic performances of T3 in the presence of different concentration of deoxycholic acid (DCA) (Figure S4). Result shows that the introduction of coadsorbent gradually decreased Jsc and Voc with increased DCA concentration. This is very likely that the amount of dye molecules adsorbed on NiO surface was reduced by the competitive coadsorption of DCA, thus leading to significant loss of light harvesting [36]. The result here could also indicate that the design motif of the dye structures are rationale and can efficiently avoid dye aggregation on NiO film. 4. Conclusions In brief, two new dyes based on triphenylamine and dicyanovinylene as electron donors and acceptors were synthesized to demonstrate rational design of dyes for favorable synergistic effects in nDSSCs. Compared to T1 reported earlier, the bridge ligand between triphenylamine and the anchoring carboxylic group are increased by incorporating ter-thiophene and fluorene units respectively, to produce dye T3 and T4. The best performance among solar cells with dyes based on an arylamine structure was achieved by dye T3 (h ¼ 0.19%, Jsc ¼ 4.01 mA cm2, Voc ¼ 144 mV, and FF ¼ 0.33). Results indicate that the ter-thiophene groups in T3 strongly affects both charge recombination and hole injection in the photoelectrode. In addition, the hexyl chains on the bridged thiophene rings could also help to avoid dye aggregation on the NiO film and block I in electrolyte from approaching the surface of NiO, which leads to reducing the charge recombination process between NiO semiconductor and electrolyte. In contrast, dye T4 introducing fluorene unit between triphenylamine and carboxylic acid shows even poorer efficiency compared to T1, perhaps mainly because of the relatively smaller dye loading amount and poorer hole injection efficiency of T4. This study suggested that modification of the bridging moiety between triphenylamine and the carboxylic group by increasing thiophene units is a promising way for preventing charge recombination and as a result increasing the power conversion efficiency. Acknowledgments
Fig. 5. (a) Electrochemical impedance spectra on NiO electrodes sensitized with the P1, T1 and T3 dyes with 0.1 V under dark condition and under 1 sun at open-circuit voltage, (b) dependence of Rrec and (c) hole lifetime on the applied bias voltage.
the IS measured at open-circuit voltage under illumination are 198 U (T1) and 105 U (T3), which are smaller than that measured at open-circuit voltage bias in the dark (300 U (T1) and 342 U (T3)) (Table S2). Similar to the n-DSSCs, the disparity of charge transfer resistance at NiO/dyeeelectrolyte interface in the dark and under illumination originates from the differentiation of the local I concentration. In dark, I is generated at the counter electrode and penetrates into the NiO film by a diffusion process [35]. Under illumination, I is formed “in situ” by dye regeneration at the NiO/ electrolyte interface. As discussed above, higher LUMO level of the dyes (3.28 eV (T1) and 3.09 eV (T3)) than the redox mediator (4.15 eV) suggests sufficient driving force for dye regeneration. Thus the I concentration could be determined by the reduced dye concentration, which is relevant to the injection ability of the
This work is partially financially Institute for Clean Energy & Advanced Materials, Southwest University, Chongqing, P.R. China. L. Zhu thanks the National Natural Science Foundation of China (No. 51203046). Hong Bin Yang is grateful for financial support from the Agency of Science, Technology and Research (A*Star), Singapore under SERC Grant No. 122 020 3053. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.dyepig.2014.01.024 References [1] O’Regan B, Grätzel M. A low-cost, high-effieciency solar cell based on dyesensitized colloidal TiO2 films. Nature 1991;353:737e40. [2] Yella A, Lee H-W, Tsao HK, Yi C, Chandiran AK, Nazeeruddin M, et al. Science 2011;334:629e34. [3] Ince M, Cardinali F, Yum J, Martínez-Díaz MV, Nazeeruddin MK, Grätzel M, et al. Convergent synthesis of near-infrared absorbing, “push-pull”, bisthiophene- substituted, Zn(II)Phthalocyanines and their application in dye sensitized solar cells. Chem Eur J 2012;18:6343e8.
104
L. Zhu et al. / Dyes and Pigments 105 (2014) 97e104
[4] Yang S, Wu K, Chi Y, Cheng Y, Chou P. Tris(thiocyanate) Ruthenium(II) sensitizers with functionalized dicarboxyterpyridine for dye-sensitized solar cells. Angew Chem Int Ed 2011;50:8270e4. [5] Wu W, Yang J, Hua J, Tang J, Zhang L, Long Y, et al. Efficient and stable dyesensitized solar cells based on phenothiazine sensitizers with thiophene units. J Mater Chem 2010;20:1772e9. [6] Marszalek M, Nagane S, Ichake A, Humphry-Baker R, Paul V, Zakeeruddin SM, et al. Tuning spectral properties of phenothiazine based donorepeacceptor dyes for efficient dye-sensitized solar cells. J Mater Chem 2012;22:889e94. [7] Qin P, Zhu H, Edvinsson T, Boschloo G, Hagfeldt A, Sun L. Design of an organic chromophore for p-type dye-sensitized solar cells. J Am Chem Soc 2008;130: 8570e1. [8] Qin P, Linder M, Brinck T, Boschloo G, Hagfeldt A, Sun L. High incident photonto-current conversion efficiency of p-type dye-sensitized solar cells based on NiO and organic chromophores. Adv Mater 2009;21:2993e6. [9] Favereau L, Warnan J, Pellegrin Y, Blart E, Boujtita M, Jacquemin D, et al. Diketopyrrolopyrrole derivatives for efficient NiO-based dye-sensitized solar cells. Chem Commun 2013;49:8018e20. [10] Powar S, Wu Q, Weidelener M, Nattestad A, Hu Z, Mishra A, et al. Improved photocurrents for p-type dye-sensitized solar cells using nano-structured nickel(II) oxide microballs. Energy Environ Sci 2012;5:8896e900. [11] Li L, Gibson EA, Qin P, Boschloo G, Gorlov M, Hagfeldt A, et al. Double-layered NiO photocathodes for p-type DSSCs with record IPCE. Adv Mater 2010;22: 1759e62. [12] Nattestad A, Mozer AJ, Fischer MKR, Cheng YB, Mishra A, Bauerle P, et al. Highly efficient photocathodes for dye sensitized tandem solar cells. Nat Mater 2010;9:31e5. [13] Odobel F, Pellegrin Y. Recent advances in the sensitization of wide-band-Gap nanostructured p-type semiconductors. Photovoltaic and photocatalytic applications. J Phys Chem Lett 2013;4:2551e64. [14] Odobel F, Le Pleux L, Pellegrin Y, Blart E. New photovoltaic devices based on the sensitization of p-type semiconductors: challenges and opportunities. Acc Chem Res 2010;43:1063e71. [15] Mori S, Fukuda S, Sumikura S, Takeda Y, Tamaki Y, Suzuki E, et al. Chargetransfer processes in dye-sensitized NiO solar cells. J Phys Chem C 2008;112: 16134e9. [16] Morandeira A, Boschloo G, Hagfeldt A, Ham-marström L. Photoinduced ultrafast dynamics of coumarin 343 sensitized p-type-nanostructured NiO films. J Phys Chem B 2005;109:19403e10. [17] Borgström M, Blart E, Boschloo G, Mukhtar E, Hagfeldt A, Hammarström L, et al. Sensitized hole injection of phosphorus porphyrin into NiO: toward new photovoltaic devices. J Phys Chem B 2005;109:22928e34. [18] Chang YJ, Chou PT, Lin SY, Watanabe M, Liu ZQ, Lin JL, et al. High performance organic materials for dye-sensitized solar cells: triarylene-linked dyads with a 4-tert-Butylphenylamine donor. Chem Asian J 2012;7:572. [19] He J, Lindström H, Hagfeldt A, Lindquist SE. Dye-sensitized nanostructured ptype nickel oxide film as a photocathode for a solar cell. J Phys Chem B 1999;103:8940e3. [20] Morandeira A, Fortage J, Edvinsson T, Le Pleux L, Blart E, Boschloo G, et al. Improved photon-to-current conversion efficiency with a nanoporous p-type
[21]
[22] [23] [24]
[25]
[26]
[27]
[28]
[29] [30]
[31]
[32] [33]
[34]
[35]
[36]
NiO electrode by the use of a sensitizer-acceptor dyad. J Phys Chem C 2008;112:1721e8. Morandeira A, Boschloo G, Hagfeldt A, Hammarström L. Coumarin 343eNiO films as nanostructured photocathodes in dye-sensitized solar cells: ultrafast electron transfer, effect of the I 3 /I redox couple and mechanism of photocurrent generation. J Phys Chem C 2008;112:9530e7. Zhu H, Hagfeldt A, Boschloo G. Photoelectrochemistry of mesoporous NiO electrodes in Iodide/triiodide electrolytes. J Phys Chem C 2007;111:17455e8. Yen Y, Chen W, Hsu C, Chou H, Lin JT, Yeh MP. Arylamine-based dyes for ptype dye-sensitized solar cells. Org Lett 2011;13:4930e3. Le Pleux L, Smeigh AL, Gibson E, Pellegrin Y, Blart E, Boschloo G, et al. Synthesis, photophysical and photovoltaic investigations of acceptorfunctionalized perylene monoimide dyes for nickel oxide p-type dyesensitized solar cells. Energy Environ Sci 2011;4:2075e84. Powar S, Daeneke T, Ma MT, Fu D, Duffy NW, Götz G, et al. Highly efficient ptype dye-sensitized solar cells based on tris(1,2-diaminoethane)cobalt(II)/(III) electrolytes. Angew Chem 2013;52:602e5. Leliège A, Blanchard P, Rousseau T, Roncali J. Triphenylamine/Tetracyanobutadiene-Based D-A-D p-conjugated systems as molecular donors for organic solar cells. Org Lett 2011;13:3098e101. Yuen MY, Kui SCF, Low KH, Kwok CC, Chui SSY, Ma CW, et al. Synthesis, photophysical and electrophosphorescent properties of fluorene-based platinum(II) complexes. Chem Euro J 2010;16:14131e41. Zhu L, Yang H, Zhong C, Li CM. Modified triphenylamine-dicyanovinyl-based donoreacceptor dyes with enhanced power conversion efficiency of p-type dye-sensitized solar cells. Chem Asian J 2012;7:2791e5. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, et al. Gaussian 09, revision A.02. Wallingford CT: Gaussian, Inc.; 2009. Mozer AJ, Griffith MJ, Tsekouras G, Wagner P, Wallace GG, Mori S, et al. ZnZn porphyrin dimer-sensitized solar cells: toward 3-D light harvesting. J Am Chem Soc 2009;131:15621e3. Cai L, Tsao HN, Zhang W, Wang L, Xue Z, Grätzel M, et al. Organic sensitizers with bridged triphenylamine donor units for efficient dye-sensitized solar cells. Adv Eng Mater 2013;3:200e5. Wang Q, Moser JE, Grätzel M. Electrochemical impedance spectroscopic analysis of dye-sensitized solar cells. J Phys Chem B 2005;109:14945e53. Fabregat-Santiago F, Bisquert J, Palomares E, Otero L, Kuang D, Zakeeruddin SM, et al. Correlation between photovoltaic performance and impedance spectroscopy of dye-sensitized solar cells based on ionic liquids. J Phys Chem C 2007;111:6550e60. Yang H, Guai GH, Guo C, Song Q, Jiang SP, Wang Y, et al. NiO/Graphene composite for enhanced charge separation and collection in p-type dye sensitized solar cell. J Phys Chem C 2011;115:12209e15. Wang M, Chen P, Humphry-Baker R, Zakeeruddin SM, Grätzel MM. The influence of charge transport and recombination on the performance of dyesensitized solar cells. ChemPhysChem 2009;10:290e9. Wan Z, Jia C, Duan Y, Zhou L, Lin Y, Shi Y. Phenothiazineetriphenylamine based organic dyes containing various conjugated linkers for efficient dyesensitized solar cells. J Mater Chem 2012;22:25140e7.