Novel Red Light-Absorbing Organic Dyes Based on Indolo[3,2-b]carbazole as the Donor Applied in Co-Sensitizer-Free Dye-Sensitized Solar Cells

Three novel organic dyes (D6, D7 and D8), based on indolo[3,2-b]carbazole as the donor and different types of electron-withdrawing groups as the acceptors, were synthesized and successfully applied in dye-sensitized solar cells (DSSCs). Their molecular structures were fully characterized by 1H NMR, 13C NMR and mass spectroscopy. The density functional theory (DFT) calculations, electrochemical impedance spectroscopy analysis, UV–Vis absorption characterization and tests of the solar cells were used to investigate the photophysical/electrochemical properties as well as DSSCs’ performances based on the dyes. Dye D8 showed the broadest light-response range (300–770 nm) in the incident monochromatic photo-to-electron conversion efficiency (IPCE) curve, due to its narrow bandgap (1.95 eV). However, dye D6 exhibited the best device performance among the three dyes, with power conversion efficiency of 5.41%, Jsc of 12.55 mA cm−2, Voc of 745 mV and fill factor (FF) of 0.59. We also found that dye aggregation was efficiently suppressed by the introduction of alkylated indolo[3,2-b]carbazole, and, hence, better power conversion efficiencies were observed for all the three dyes, compared to the devices of co-sensitization with chenodeoxycholic acid (CDCA). It was unnecessary to add adsorbents to suppress the dye aggregation.


Introduction
In the past three decades, scientists have paid much attention to the dye-sensitized solar cells (DSSCs). They have been regarded as an alternative energy source since the first report in 1991, due to their ability to convert solar energy to electricity at a low cost and excellent photovoltaic performance [1][2][3][4]. To date, Ru-complex and Zn-porphyrin sensitizers have achieved high power conversion efficiencies (PCEs) of 11.9% and 13.5%, respectively [5,6]. However, the high cost, limited resource of ruthenium and the complex synthesis procedure of zinc porphyrin impede their further application in DSSCs. Therefore, worldwide scientists have been involved in developing metal-free organic dyes with high PCE, due to their flexible structural modification, simple synthesis, low toxicity and low cost [7,8]. Recently, DSSCs based on metal-free organic dyes have achieved a very high PCE of 14% [9]. However, those dyes have complex structures and are difficult to synthesize. In addition, they need particular electrolytes and a co-sensitizer to improve open circuit voltage and suppress the dye aggregation to achieve high PCE. These reasons work against the commercialization of dyes. Therefore, the organic dyes featuring simple synthesis and without the necessity for using a co-sensitizer to improve PCE in the DSSCs could be a promising research field.
Among various metal-free organic dyes, the dyes with donor-π-bridge-acceptor (D-π-A) structures exhibit outstanding performance with simple molecular structures.
Among various metal-free organic dyes, the dyes with donor-π-bridge-acceptor (Dπ-A) structures exhibit outstanding performance with simple molecular structures. For the typical D-π-A dyes, the structural modifications of the donor and acceptor can effectively tune the energy levels and improve intramolecular charge transfer (ICT) from D to A, therefore providing a direct way for acquiring high PCE [10]. Many kinds of promising donors have been reported, such as carbazole [11][12][13][14][15], di (1-benzothieno) [3,2-b:2′,3′-d]pyrrole [16,17], triphenylamine [18], dithieno [3,2-b]pyrrolobenzotriazole [19], and so on. Recently, indolo [3,2-b]carbazole featuring a larger conjugated plane than triphenylamine and carbazole has been reported as an excellent donor to construct organic optoelectronic functional materials, due to its outstanding hole-donating ability [20,21]. With its high thermal and chemical stability, indolo [3,2-b]carbazole was proven to be a good candidate as the D unit of organic dyes [22]. We previously reported a series of efficient D-D-π-A organic dyes based on indolo [3,2-b]carbazole as the first donor [23,24], triphenylamine as the second donor, and thiophene cyanoacetic acid as the acceptor, showing efficient photovoltaic performance with a highest PCE up to 6.34%. That study indicated that indolo [3,2-b]carbazole was a potential donor group in D-D-π-A-type organic dyes. However, it was rarely reported that indolo [3,2-b]carbazole was used as a single donor group. The study focus on the behaviors of indolo [3,2-b]carbazole-based D-π-A and D-A-π-A dyes is also an interesting field. In addition, the light-response ranges based on previously reported dyes were not satisfactory, because they can only absorb the visible light, which leads to comparably low PCE, compared with those near-infrared-absorbing organic dyes. To expand the light absorption to longer wavelengths, an excellent acceptor should be introduced to the dyes to enhance the D-π-A effects. We found that 2,1,3-benzothiadiazole as promising electron acceptor plays a crucial role in power conversion efficiency [25,26]. Because of its strong electron withdrawing ability, benzothiadiazole could lower the energy level of the lowest unoccupied molecular orbital (LUMO) and further reduce the energy gap of sensitizers [27,28]. Therefore, benzothiadiazole may be a suitable acceptor and is easily connected by some well-performed donors to enhance dye performances with simple dye structures.
In this work, we synthetized three dyes (D6, D7 and D8) by adopting indolo [3,2-b]carbazole as single donor to simplify synthesis and thiophene cyanoacetic acid, benzothiadiazole benzoic acid and benzothiadiazole thiophene cyanoacetic acid as respective acceptors to modify the bandgap of dyes. The structures of D6-D8 were shown in Figure 1. As the results show, the introduction of benzothiadiazole effectively expanded the light absorption of dyes to 690 nm. In addition, the dye aggregation was also suppressed by the two alkyl chains of indolo [3,2-b]carbazole, demonstrating better PCE even in the co-sensitizer-free DSSCs. Finally, the systematic study of their photophysical and electrochemical properties, as well as the resulting DSSCs performances based on the three dyes, was also reported.

Materials and Methods
All the chemicals used in this study were purchased from Sigma-Aldrich (Saint Louis, MO, USA) and J&K Chemical Ltd (Beijing, China). The solvents used in synthesis were purified using standard processes. The reactions in this study were carried out in

Materials and Methods
All the chemicals used in this study were purchased from Sigma-Aldrich (Saint Louis, MO, USA) and J&K Chemical Ltd (Beijing, China). The solvents used in synthesis were purified using standard processes. The reactions in this study were carried out in the N 2 atmosphere. The separation of the compounds was carried out on column chromatography using silica gel (200-300 mesh).
Nuclear Magnetic Resonance (NMR) spectra ( 1 H and 13 C) characterization were tested on a Bruker 500 MHz spectrometer in CDCl 3 and DMSO-d 6 . Mass spectra were carried out on a Bruker microTOF-Q. The absorption spectra of dyes were recorded on a HP 8453 spectrophotometer. Electrochemical properties were studied by cyclic voltammetry measurements on a CHI604D electrochemical workstation, using Bu 4 NPF 6 as supporting electrolyte, dry dichloromethane as solvent and N 2 as protection gas. The scan rate was 50 mV s −1 . The ferrocene/ferrocenium (Fc/Fc + ) redox couple was used as the internal standard. The current voltage (J-V) tests of the DSSCs were carried out on Keithley 2400 source meter under simulated AM 1.5 G (100 mW cm −2 ) illumination with a solar light simulator. Incident monochromatic photo-to-electron conversion efficiency (IPCE) was measured from 300 to 800 nm by a Spectral Products DK240 monochromator. The electrochemical impedance spectra (EIS) were carried out on a CHI604D electrochemical work station under dark conditions.

Fabrication and Characterization of DSSCs
Nanocrystalline TiO 2 films were prepared by screen printing two kinds of TiO 2 nanoparticles (a 12 µm nanoporous layer and a 4 µm scattering layer) over the conductive side of the conducting glass. The active area of the TiO 2 was 0.196 cm 2 . The photoanode was prepared by immersing the TiO 2 film into a 0.3 mM dye solution in THF for 24 h under the dark. After the dye adsorption, the photoanode was washed with dichloromethane (DCM) and then dried. The electrolyte contained acetonitrile-valeronitrile (85:15, v/v), composed of 0.6 M BMII (1-butyl-3-methylimidazolium iodide), 0.05 M LiI, 0.03 M I 2 , 0.5 M 4-tert-butylpyridine and 0.1 M guanidinium thiocyanate. The electrolyte was inserted into the interspace between the photoanode and cathode from the two holes predrilled on the back of the counter electrode. At last, the holes were sealed with a Surlyn film and a thin glass.

Synthesis
Synthetic routes to the three dyes D6, D7 and D8 are depicted in Scheme 1. Compound 1 was prepared via substitution reaction of 5,11-Dihydroindolo[3,2-b]carbazole with 2-ethylhexyl bromide and NaOH in DMSO, according to the literature [29]. Compound 1 was prepared by controlling the amount of N-bromosuccinimde (NBS) to provide monobrominated compound 2. Then, key intermediate compounds, such as pinacol boronic ester 2, were afforded by coupling bis(pinacolato)diboron with compound 3. Compounds 5, 6, 3a, 3b and 3c were synthesized by Suzuki coupling reaction. The dyes D6 and D8 were synthesized by condensation of aldehydes 3a and 3c with piperidine cyanoacetic acid. Dye D7 was synthesized by hydrolysis of compound 3b in sodium hydroxide solution. The NMR and mass spectra of D6-D8 are shown in Figures S1-S6.

2-bromo-5,11-bis(2-ethylhexyl)-5,11-dihydroindolo[3,2-b]carbazole (2)
In a 100 mL flask, compound 1 (2.3 g, 4.8 mmol) was dissolved in 40 mL THF solution and stirred at 0 • C for 15 min. Under light-shielding conditions, (937 mg, 5.2 mmol) Nbromosuccinimide (NBS) was added portion-wise. After 2 h of reaction, the temperature was gradually raised to room temperature, and the reaction was continued for 24 h. The reaction mixture was poured into water and extracted with DCM, and the combined extracts were washed with water and dried with anhydrous NaSO 4 . The solvent was removed with a rotary evaporator, and the residue was isolated by silica gel column chromatography with petroleum ether (PE) as eluent to afford compound 2 as a yellowgreen solid (yield: 81%, 2.17 g). 1

Photophysical Properties
The UV-Visible absorption spectra of dyes D6, D7 and D8 in DCM solution with concentration of 1 × 10 −5 M are presented in Figure 2, and the corresponding data are listed in Table 1. The shorter wavelengths located at 300-370 nm are ascribed to the localized ππ* transitions of the conjugated backbone, while the band of 400-600 nm is attributed to the intramolecular charge transfer (ICT) transition from donor to acceptor [30,31]. The maximum absorption wavelength (λmax) for dyes D6, D7 and D8 are 500 nm, 463 nm and 529 nm, respectively. Dyes D6-D8 show molar extinction coefficient(ε) with values of around 4.59 × 10 4 M −1 cm −1 , 1.58 × 10 4 M −1 cm −1 and 1.73 × 10 4 M −1 cm −1 , respectively. The remarkably high ε of dye D6 is due to the introduction of a thiophene group, which expand the conjugate plane. Therefore, the light absorption properties of dye D6 have been improved [32,33]. In addition, a larger ε allows a thinner TiO2 film, which shortens the Scheme 1. Synthetic routes of the dyes.

Photophysical Properties
The UV-Visible absorption spectra of dyes D6, D7 and D8 in DCM solution with concentration of 1 × 10 −5 M are presented in Figure 2, and the corresponding data are listed in Table 1. The shorter wavelengths located at 300-370 nm are ascribed to the localized π-π* transitions of the conjugated backbone, while the band of 400-600 nm is attributed to the intramolecular charge transfer (ICT) transition from donor to acceptor [30,31]. The maximum absorption wavelength (λ max ) for dyes D6, D7 and D8 are 500 nm, 463 nm and 529 nm, respectively. Dyes D6-D8 show molar extinction coefficient(ε) with values of around 4.59 × 10 4 M −1 cm −1 , 1.58 × 10 4 M −1 cm −1 and 1.73 × 10 4 M −1 cm −1 , respectively. The remarkably high ε of dye D6 is due to the introduction of a thiophene group, which expand the conjugate plane. Therefore, the light absorption properties of dye D6 have been improved [32,33]. In addition, a larger ε allows a thinner TiO 2 film, which shortens the electrolyte diffusion distance in the film and, thus, reduces the charge recombination during transport [24]. Furthermore, dye D8 shows the broadest absorption band of the dyes, and the λ max red-shifted about 87 nm and 21 nm, in comparison with dyes D6 and D7, respectively, due to its stronger ICT transition.

529(1.73 × 10 )
[a] The maximum absorption wavelength (λmax) and molar extinction coefficients (ε). [b] E0-0 was calculated from E0-0 = 1240/λonset, where λonset was determined from the onset of absorption spectrum. The normalized absorption spectra of dyes D6-D8 adsorbed on 12μm TiO2 films are shown in Figure 2b. After being anchored onto the TiO2 films, the absorption spectra (300-750 nm) of the three dyes were wider, compared with the spectra in solution (300-700 nm), which is beneficial to light-harvesting and Jsc enhancement. Compared with the λmax in solution, the λmax of dyes D6-D8 on TiO2 film are blue-shifted by about 49, 1 and 8 nm, respectively. The blue-shift of absorption on TiO2 film is mainly attributed to the deprotonation of carboxylic acid and H-type aggregation [34].

Electrochemical Properties
Electrochemical properties of dyes D6-D8 were studied by cyclic voltammetry (CV) curves in dichloromethane solution, with TBAPF6 as the electrolyte. The CV curves are shown in Figure 3a, and the corresponding data are displayed in Table 1. The three dyes show quasar-reversible oxidation waves, which indicate that the electrochemical properties of these dyes are stable. The oxidative potential (Eox) versus normal hydrogen electrode (vs. NHE) of dyes D6-D8 are 0.87, 1.01 and 0.81 V, respectively, which are higher than the I − /I3 − potential (+0.4 V), indicating that the oxidized dyes can be regenerated effectively [11,[35][36][37]. The energy gaps (E0-0) of dyes D6-D8, calculated according to UV-  The normalized absorption spectra of dyes D6-D8 adsorbed on 12 µm TiO 2 films are shown in Figure 2b. After being anchored onto the TiO 2 films, the absorption spectra (300-750 nm) of the three dyes were wider, compared with the spectra in solution (300-700 nm), which is beneficial to light-harvesting and J sc enhancement. Compared with the λ max in solution, the λ max of dyes D6-D8 on TiO 2 film are blue-shifted by about 49, 1 and 8 nm, respectively. The blue-shift of absorption on TiO 2 film is mainly attributed to the deprotonation of carboxylic acid and H-type aggregation [34].

Electrochemical Properties
Electrochemical properties of dyes D6-D8 were studied by cyclic voltammetry (CV) curves in dichloromethane solution, with TBAPF 6 as the electrolyte. The CV curves are shown in Figure 3a, and the corresponding data are displayed in Table 1. The three dyes show quasar-reversible oxidation waves, which indicate that the electrochemical properties of these dyes are stable. The oxidative potential (E ox ) versus normal hydrogen Materials 2021, 14, 1716 9 of 16 electrode (vs. NHE) of dyes D6-D8 are 0.87, 1.01 and 0.81 V, respectively, which are higher than the I − /I 3 − potential (+0.4 V), indicating that the oxidized dyes can be regenerated effectively [11,[35][36][37]. The energy gaps (E 0-0 ) of dyes D6-D8, calculated according to UV-Vis absorption spectra, were 2.17, 2.36 and 1.95 eV, respectively. Therefore, the reductive potentials of dyes D6-D8, calculated according to the equation of E red [V] = E ox − E 0-0 , were −1.30, −1.35 and −1.08 V (vs. NHE), respectively, which are much higher than the conduction-band edge of TiO 2 (−0.5 V). This indicates that the driving force for the electron injection from dyes to the TiO 2 film is sufficient. Figure 3b

Theoretical Calculations
The structures of dyes D6-D8 have been further analyzed by using (B3LYP/6-31G (d, p) level). The electron distributions in the highest occupied molecular orbital (HO-MOs) and lowest unoccupied molecular orbital (LUMOs) of dyes D6-D8 are shown in Table 2. The HOMOs of dyes D6-D8 reside mostly on the indolo[3,2-b] carbazole, and the LUMOs reside over the thiophene, benzothiadiazole and cyanoacrylic acid acceptor. We can observe the significant overlapping between the HOMOs (indolo[3,2-b] carbazole) and LUMOs (thiophene and benzothiadiazole), which facilitates charge transfer transition from the donor to the acceptor. It is worth noting that the electron density of the excited state of dye D7 is not completely localized at the anchoring units. This unoptimal electronic matching result may influence the electron injection from dye D7 to the TiO2 and, thus, cause the unexpectedly lower Jsc recorded. The optimized ground-state geometries (Table 3) indicate that the indolo[3,2-b] carbazole unit presents a rigid planar structure. The dihedral angles (between indolo[3,2-b] carbazole to acceptor) of dyes D6-D8 are 21.8°, 36.0° and 32.7°, respectively. The dihedral angle between thiophene and phenyl (21.8° for dye D6 and 0.7° for dye D8) is smaller than that between two phenyls (36° for dye D7 and 32.7° for dye D8). This indicates that thiophene group can improve planarity and conjugation, which is good for dye ICT transition [38][39][40].

Theoretical Calculations
The structures of dyes D6-D8 have been further analyzed by using (B3LYP/6-31G (d, p) level). The electron distributions in the highest occupied molecular orbital (HOMOs) and lowest unoccupied molecular orbital (LUMOs) of dyes D6-D8 are shown in Table 2. The HOMOs of dyes D6-D8 reside mostly on the indolo[3,2-b] carbazole, and the LUMOs reside over the thiophene, benzothiadiazole and cyanoacrylic acid acceptor. We can observe the significant overlapping between the HOMOs (indolo[3,2-b] carbazole) and LUMOs (thiophene and benzothiadiazole), which facilitates charge transfer transition from the donor to the acceptor. It is worth noting that the electron density of the excited state of dye D7 is not completely localized at the anchoring units. This unoptimal electronic matching result may influence the electron injection from dye D7 to the TiO 2 and, thus, cause the unexpectedly lower J sc recorded. The optimized ground-state geometries (  D8). This indicates that thiophene group can improve planarity and conjugation, which is good for dye ICT transition [38][39][40].

Photovoltaic Properties
The photovoltaic properties of dyes D6-D8 and N719 are performed under standard conditions (AM 1.5, 100 mW cm −2 ). The photocurrent-voltage (J-V) curves are shown in Figure 4a, and the corresponding data of short circuit photocurrent density (J sc ), opencircuit voltage (V oc ), fill factor (FF) and power conversion efficiency (PCE) are shown in Table 4. In the UV-Visible absorption spectra, dye D8 shows a broader absorption band, which may lead to a higher J sc . However, the J sc of the DSSCs based on dyes D6-D8 are 12.55, 10.76 and 10.17 mA cm −2 , respectively. It is clear that the dye D6based cell shows higher J sc than that of dyes D7 and D8, which is due to its higher molar extinction coefficients. Furthermore, the dye loading amount of dyes D6-D8 are 4.16 × 10 −7 mol cm −2 , 2.81 × 10 −7 mol cm −2 and 2.51 × 10 −7 mol cm −2 , respectively, which is consistent with the J sc of the three dyes. The V oc of the DSSC based on dyes D6-D8 is 745, 744 and 668 mV, respectively. Compared with the DSSCs of dyes D7 and D8, dye D6-based DSSC shows the highest PCE of 5.41%, due to its high J sc . Finally, these three dyes exhibit good power conversion efficiency, which indicates that indolo[3,2-b]carbazole group might be a good donor.  Table 4. In the UV-Visible absorption spectra, dye D8 shows a broader absorption band, which may lead to a higher Jsc. However, the Jsc of the DSSCs based on dyes D6-D8 are 12.55, 10.76 and 10.17 mA cm −2 , respectively. It is clear that the dye D6-based cell shows higher Jsc than that of dyes D7 and D8, which is due to its higher molar extinction coefficients. Furthermore, the dye loading amount of dyes D6-D8 are 4.16 × 10 −7 mol cm −2 , 2.81 × 10 −7 mol cm −2 and 2.51 × 10 −7 mol cm −2 , respectively, which is consistent with the Jsc of the three dyes. The Voc of the DSSC based on dyes D6-D8 is 745, 744 and 668 mV, respectively. Compared with the DSSCs of dyes D7 and D8, dye D6-based DSSC shows the highest PCE of 5.41%, due to its high Jsc. Finally, these three dyes exhibit good power conversion efficiency, which indicates that indolo [3,2-b]carbazole group might be a good donor.
The incident photon-to-current conversion efficiency (IPCE) spectra for the cells is shown in Figure 4b. As can be seen, all the three dyes exhibit broad response which indicates that dyes D6-D8 can convert the visible light to photocurrent efficiently. The maximum spectra response band of dyes D6-D8 are at 710, 660 and 770 nm, respectively. The absorption range of dye D8-based DSSC is obviously broader than that of dyes D6 and D7, due to the incorporated benzothiadiazole-thiophene unit decreasing the E0-0. However, dye D6 shows strong response in the range from 350 to 550 nm, with a highest IPCE value of 72.55% at 480 nm. Dyes D7 and D8 display the maximum IPCE values of 68.4% at 500 nm and 47.74% at 525 nm, respectively. Therefore, the best IPCE performance is observed on the dye D6, among the three dyes. It mainly originates from its much higher molar extinction coefficient and more dye loading amount. These results are also in good accordance with the Jsc value obtained in above J-V measurements. The chenodeoxycholic acid (CDCA) co-sensitization was tested by blending dyes D6, D7 and D8 with different ratios of CDCA in the THF dye bath, respectively. The corresponding curve and data are shown in Figure 5 and Table 4, respectively. Generally, the PCE of most dyes can be improved by the addition of CDCA, because it can restrain the dye aggregation. However, for aggregation-free organic dyes, the addition of CDCA would reduce adsorption of the dye molecules on the TiO2 film, which would reduce the short-circuit current density and, hence, the power conversion efficiency [41]. The photovoltaic properties of the DSSCs based on dyes D6-D8 with no additives and with CDCA as co-sensitizer were investigated. As can be seen in Figure 5a,b, the decrease of Jsc is found  The incident photon-to-current conversion efficiency (IPCE) spectra for the cells is shown in Figure 4b. As can be seen, all the three dyes exhibit broad response which indicates that dyes D6-D8 can convert the visible light to photocurrent efficiently. The maximum spectra response band of dyes D6-D8 are at 710, 660 and 770 nm, respectively. The absorption range of dye D8-based DSSC is obviously broader than that of dyes D6 and D7, due to the incorporated benzothiadiazole-thiophene unit decreasing the E 0-0 . However, dye D6 shows strong response in the range from 350 to 550 nm, with a highest IPCE value of 72.55% at 480 nm. Dyes D7 and D8 display the maximum IPCE values of 68.4% at 500 nm and 47.74% at 525 nm, respectively. Therefore, the best IPCE performance is observed on the dye D6, among the three dyes. It mainly originates from its much higher molar extinction coefficient and more dye loading amount. These results are also in good accordance with the J sc value obtained in above J-V measurements.
The chenodeoxycholic acid (CDCA) co-sensitization was tested by blending dyes D6, D7 and D8 with different ratios of CDCA in the THF dye bath, respectively. The corresponding curve and data are shown in Figure 5 and Table 4, respectively. Generally, the PCE of most dyes can be improved by the addition of CDCA, because it can restrain the dye aggregation. However, for aggregation-free organic dyes, the addition of CDCA would reduce adsorption of the dye molecules on the TiO 2 film, which would reduce the shortcircuit current density and, hence, the power conversion efficiency [41]. The photovoltaic properties of the DSSCs based on dyes D6-D8 with no additives and with CDCA as cosensitizer were investigated. As can be seen in Figure 5a,b, the decrease of J sc is found in dyes D6-and D7-based DSSC after the co-sensitization. When the ratio of CDCA increases, both the J sc and PCE of these two dyes decrease. It can be explained that indolo [3,2b]carbazole and its long alkyl chain can restrain dye aggregation efficiently. Meanwhile, adopting CDCA as the co-adsorbent is unfavorable for these two dyes adsorptions on TiO 2 films. However, Figure 5c shows that J sc increases by adding 10 times of CDCA. This indicates that D8 presents aggregation and improved by the addition of CDCA. Meanwhile, adopting CDCA as the co-adsorbent is unfavorable for these two dyes adsorptions on TiO2 films. However, Figure 5c shows that Jsc increases by adding 10 times of CDCA. This indicates that D8 presents aggregation and improved by the addition of CDCA.

Electrochemical Impedance Spectra (EIS) Analysis
In order to study the charge recombination and charge transfer process in the DSSCs, the electrochemical impedance spectroscopy (EIS) analysis is performed in the dark. The measurement is under a bias of −0.7 V, and the frequency range is 0.1 Hz-10 kHz. The corresponding Nyquist plots are shown in Figure 6a. The first semicircle at higher frequencies corresponds to the electron transport at the Pt/electrolyte interface, and the last one at lower frequencies corresponds to the charge recombination at the TiO 2 /dyes/electrolyte interface. The larger value of charge transfer resistance (R ct ) reflects the lower charge recombination, smaller dark current and larger open circuit voltage [42,43]. The values of R ct can be deduced by fitting curves using ZView complex nonlinear least-square regression software (AMETEK, Leicester, UK) The order of R ct values is D8 (63.7 Ω) < D7 (113.0 Ω) < D6 (155.5 Ω), which is consistent with the order of V oc values of D6 (668 mV) < D7 (744 mV) < D8 (745 mV). The electron lifetime (τ) of the dyes D6-D8 can be calculated from the Bode phase plots (Figure 6b) to support the trends of V oc . The electron lifetime (τ) is calculated from τ = 1/(2πf) [44,45], where f is the peak frequency in lower frequency. The order of τ is D6 (25.8 ms) > D7 (25.5 ms) > D8 (8.3 ms), which is in line with their V oc .

Electrochemical Impedance Spectra (EIS) Analysis
In order to study the charge recombination and charge transfer process in the DSSCs, the electrochemical impedance spectroscopy (EIS) analysis is performed in the dark. The measurement is under a bias of −0.7 V, and the frequency range is 0.1 Hz-10 kHz. The corresponding Nyquist plots are shown in Figure 6a. The first semicircle at higher frequencies corresponds to the electron transport at the Pt/electrolyte interface, and the last one at lower frequencies corresponds to the charge recombination at the TiO2/dyes/electrolyte interface. The larger value of charge transfer resistance (Rct) reflects the lower charge recombination, smaller dark current and larger open circuit voltage [42,43]. The values of Rct can be deduced by fitting curves using ZView complex nonlinear least-square regression software (AMETEK, Leicester, UK) The order of Rct values is D8 (63.7 Ω) < D7 (113.0 Ω) < D6 (155.5 Ω), which is consistent with the order of Voc values of D6 (668 mV) < D7 (744 mV) < D8 (745 mV). The electron lifetime (τ) of the dyes D6-D8 can be calculated from the Bode phase plots (Figure 6b) to support the trends of Voc. The electron lifetime (τ) is calculated from τ = 1/(2πf) [44,45], where f is the peak frequency in lower frequency. The order of τ is D6 (25.8 ms) > D7 (25.5 ms) > D8 (8.3 ms), which is in line with their Voc.

Stability Study
The study of stability of dyes is very important for the DSSCs' applications. The devices, for long term stability, were encapsulated and stored in dark conditions. The evolution of efficiency was measured by testing the PCE values of three DSSCs, based on dyes D6-D8, once a day. As we can see in Figure 7, nearly 90% of initial efficiency (η0) for all three DSSCs remained over 480 h, which suggests that our dyes are stable and, thus, suggesting application potential.

Stability Study
The study of stability of dyes is very important for the DSSCs' applications. The devices, for long term stability, were encapsulated and stored in dark conditions. The evolution of efficiency was measured by testing the PCE values of three DSSCs, based on dyes D6-D8, once a day. As we can see in Figure 7, nearly 90% of initial efficiency (η 0 ) for all three DSSCs remained over 480 h, which suggests that our dyes are stable and, thus, suggesting application potential.

Conclusions
Three novel metal-free organic dyes (D6, D7 and D8) with different acceptors were synthesized and applied in DSSCs. Dye D6, containing thiophene cyanoacetic acid acceptor, exhibits much higher ε and dye loading amount than the other two dyes, which indicates

Conclusions
Three novel metal-free organic dyes (D6, D7 and D8) with different acceptors were synthesized and applied in DSSCs. Dye D6, containing thiophene cyanoacetic acid acceptor, exhibits much higher ε and dye loading amount than the other two dyes, which indicates that dye D6 possesses better light-harvesting capability. By the adoption of alkylated indolo [3,2-b]carbazole, dye aggregation was efficiently suppressed. Hence, the power conversion efficiencies of the dyes without co-sensitization with CDCA are better, as compared with the co-sensitized devices. Because of the introduction of benzothiadiazole thiophene group, dye D8 shows the broadest light-response range (300-770 nm) in IPCE curve, due to the decrease of the E 0-0 . However, its lower ε and dye loading amount leads to poor photovoltaic performance. The DSSC-based dyes D6-D8 show over 90% of initial efficiency during 480 h. Finally, dye D6 exhibits the best performance among these dyes, with the highest power conversion efficiency of 5.41%, mainly due to its highest J sc (12.55 mA cm −2 ).