Synthesis and Characterization of Two New Organic Dyes for Dye-Sensitized Solar Cells

Abstract In the present study, two new organic dyes based on indigo were prepared and used as sensitizers in dye-sensitized solar cells. To this end, indoxyl was utilized as the electron-donor and acrylic acid and cyanoacrylic acid were used as the electron-acceptor anchoring groups. These dyes were purified and characterized by analytical techniques. Spectrophotometric evaluations of the prepared dyes in solution and on a nano-anatase-TiO2 substrate were investigated. Additionally, oxidation potential measurements were also carried out. Finally, dye-sensitized solar cells were fabricated to determine the photovoltaic behavior and conversion efficiency of each dye. [Supplementary materials are available for this article. Go to the publisher's online edition of Synthetic Communications® for the following free supplemental resource(s): Full experimental and spectral details.] GRAPHICAL ABSTRACT


INTRODUCTION
The improvement of solar energy-to-electricity conversion efficiency has continued to be an important research area of dye-sensitized solar cells (DSSC). [1,2] The photosensitizer (dye) is one of the critical components in DSSC to improve the solar energy-to-electricity conversion efficiency. [3] Metal-free organic dyes, because of their many advantages, such as high molar extinction coefficients, convenience of customized molecular design for desired photophysical and photochemical properties, inexpensiveness with no requirement for the expensive ruthenium metal, and environment friendliness, are suitable as photosensitizers for DSSC. [4] Many organic dyes including coumarine, [5] polyene, [6] hemicyanine, [7] thiophene, [8] and indoline [9] dyes that exhibit relatively high performances in dye-sensitized solar cells have so far been designed and developed. Because fo ever-increasing energy crisis, much research has been focused in recent years on the synthesis of new sensitizers and=or on the study of the mechanism of energy conversion. [10] The results suggest that smartly designed metal-free organic dyes are potentially capable of being good competitive candidates as photosensitizers in DSSCs because of the advantages mentioned previously. [2,[10][11][12] Such dyes are usually referred to as donor-acceptor p-conjugated dyes of the general formula D-p-A (where D is a donor part, p is a system of p conjugation, and A is an electron-withdrawing anchoring group). These dyes are attached to the surface by coordinate bonding with the Lewis acid sites of the nanocrystalline TiO 2 to act as light-harvesting sensitizers in dye-sensitized solar cells. [5][6][7][8][9][10] In the present study, two novel metal-free organic dyes based on indigo (in fact indoxyl or half indigo) as the electron donor, naphthyl residue as the p-conjugation system, and acrylic or cyanoacrylic acid as the donor groups are proposed and investigated for the first time. Indigo is both a pigment and a dye of antiquity and is still a current commercial interest.
Could this commercial interest be extended to the use of indigo in high-tech DSSCs? Seeking the answer to this question formed the basis for undertaking the present investigation. Actually, there are several good reasons to pursue the utilization of indigo in DSSCs. Sir Issac Newton illustrated the importance of indigo by selecting the word indigo to represent a special kind of blue color when he dispersed white light into a colored spectrum by the aid of a prism. Generally, blue pigments and dyes are very rare in nature and are very difficult and expensive to produce artificially.
For a colorant to be of blue color, it must absorb all the medium to long wavelength ranges of the visible spectrum (i.e., 2=3 of the visible light spectrum), leaving the short wave range, which would collectively produce a blue color. Absorbing two thirds of the visible spectrum is, undoubtedly, very beneficial in dye-sensitized solar cells. Additionally, indigo is an interesting cross-conjugated planar molecule of extreme thermal, chemical, and photochemical durability, which are also considered to be very advantageous in DSSCs. However, it must be noted that indigo itself gives very low light-to-electricity conversion efficiencies (i.e., 7-carboxylic acid indigo tested in our laboratories gave a very low conversion efficiency of 0.01%). It is the goal of this investigation to show for the first time how by reacting indigo with a simple naphthyl residue as the p-conjugation system and a simple acrylic acid or cyanoacrylic acid electron-acceptor anchoring groups, the conversion efficiency values change so dramatically that further research is advisable.
The synthesized dyes were then purified and characterized. The spectrophotometric properties of the prepared organic dyes in tetrahydrofuran (THF) solvent and on a nano-anatase TiO 2 substrate were examined. The absorption maxima and the intensities of the resultant dyes were also obtained. Dye-sensitized solar cells were then fabricated utilizing these metal-free organic dyes and their photovoltaic behaviors were determined. Schematic representation of the route for synthesis of the prepared metal-free organic dyes is given in Fig. 1.

RESULTS AND DISCUSSION
The organic dyes D1 and D2 were synthesized as schematically shown in Fig. 1. Components 1 and 2 were prepared in a way similar to that described in the literature. [10,11] Compounds 1 and 2 reacted with indoxyl to give the respective indigo components. Finally, D1 and D2 were synthesized through the reaction with LiOH respectively.
The wavelength of maximum absorption (k max ) and the molar extinction coefficients (e max ) for the two dyes in THF are listed in Table 1 and shown in Fig. 2, together with the k max of the corresponding dyes adsorbed on the TiO 2 film. The absorption peaks at around 548.5 nm for D1 can be assigned to an intramolecular charge transfer between the donor group and the acrylic acid group, [13] providing an efficient charge separation for the excited state. For the cyanoacetic acid-based dyes (D2), when an extra electron acceptor (-CN) was linked to the vinyl bridge, the k max had a bathochromic shift from 548.5 nm for D1 to 572 nm for D2. This shift of the maximum absorption peak arises from the fact that the stronger electronaccepting ability of cyanoacrylic groups intensifies the overall electron-withdrawing capability of the system and hence lowers the level of the lower unoccupied molecular orbital (LUMO), thus reducing the gap between the higher unoccupied molecular orbital (HOMO) and the corresponding LUMO state. [14] Upon dye adsorption on

NEW ORGANIC DYES 781
to a nano-TiO 2 surface, the wavelength of maximum absorption is bathochromically shifted by 19.5 nm and 23 nm for D1 and D2, respectively, compared to the corresponding spectra in THF solution, implying that dyes adsorbed on to the TiO 2 surface contain partial J-type aggregates. [15,16] The molar extinction coefficients of D1 and D2 in THF at their respective k max are also shown in Table 1, indicating that these novel dyes have good light harvesting abilities. [17] The fluorescent characteristics of the prepared dyes in THF are also depicted in Table 1. In THF solution, dyes show intense green fluorescence due to the charge transfer from the electron-donating entity to the electron-accepting entity. The fluorescence emission maxima of the synthesized dyestuffs D1 and D2 in the THF are 703 nm and 721 nm, respectively.
The oxidation potential (E ox ) of D1 and D2 was measured in acetonitrile by cyclic voltammetry. There are two distinct oxidative waves observed in the voltammogram. The first oxidative wave (I) is due to the oxidation of the external standard of ferrocene, whereas the second wave (II) is due to the electrochemical oxidation of dyes. The oxidation peak potential (Epa) for D1 and D2 can therefore be calculated to be 0.57 V and 0.52 V vs Fc=Fc þ (as an external standard ferrocene= ferrocenium redox couple), respectively. Although the standard E ox value is usually not easily obtained experimentally, it can be approximately estimated from the cyclic voltammetric peak potential, which equals it if the electrochemical oxidation was a reversible step. [18] Therefore, the E ox -E 0-0 level, where E 0-0 represents the intersection of normalized absorption and the fluorescence spectra in THF, was calculated. This is considered to correspond to the reduction potential. [19,20] . The E 0-0 of D1 and D2 were observed at 660 nm and 672 nm, corresponding to 1.88 V and 1.85 V, respectively. Therefore, the E ox -E 0-0 level of D1 and D2 were calculated to be À1.31 V and À1.33 V vs Fc=Fc þ in acetonitrile. All the excited-state oxidation potentials (LUMO) of D1 (À1.31 V) and D2 (À1.33 V) are more negative than the conduction band gap edge of TiO 2 [À0.5 V (vs NHE, normal hydrogen electrode)]. Provided that an energy gap [between dye LUMO and TiO 2 conduction bond (CB)] of 0.2 eV is necessary for efficient electron injection, [21] the driving force is sufficient for efficient charge injection. Thus, the electron injection process from the excited dye molecule to the TiO 2 conduction band and the subsequent dye regeneration are energetically permissible. The energy levels of the ground state (HOMO) of D1 (0.57 V) and D2 (0.52 V) are sufficiently more positive than the I À 3 =I À redox potential [0.42 V (ns. NHE)], [21] indicating that the oxidized dye formed after electron injection into the conduction band of TiO 2 could accept electrons from I À ions in the electrolyte, being thermodynamically favorable. Such electronic structures thus ensure a favorable exothermic flow of charge throughout the photoelectronic conversion.
Dye-sensitized solar cells (DSSCs) were constructed and compared to clarify the relationships between the sensitizing behavior of D1 and D2 dye molecules and their structures. The DSSCs utilized these dyes as sensitizers for nanocrystalline anatase TiO 2 . The detailed photovoltaic parameters are also summarized in Table 2.
According to the results shown in Table 2, under the standard global AM 1.5 solar condition, the conversion efficiencies of cells containing D1 and D2 are 3.11% and 3.45%, respectively. The larger conversion efficiency of D2 sensitizer is probably due to the stronger electron-withdrawing ability of the cynoacrylic acid group. The conversion efficiency of solar energy to electricity of the present organic dyes could be improved by extending the conjugated length of the organic dyes or by adding substituants on the naphthyl residue or by incorporation of a thiophene p-bridge. [22,23] The solar energy to electricity conversion efficiency of the presently synthesized organic dyes can further be improved by addition of substituents on the indigo or the naphthyl residue and=or by extending the p-conjugation length of such organic dyes. [12] Moreover, other related dyes without the indigo group show much lower g values of around 2%. [7] In our study, DSSCs based on indigo derivatives showed conversion efficiencies of 3.11 and 3.45%, respectively, even though they had much broadend spectral response and greater extinction coefficiences. The results show promise for future use in nanostructure dye-sensitized solar cells with respect to low material costs, high performance, and recyclability.

EXPERIMENTAL
All compounds used in this study were of analytical grade unless otherwise stated, and components 1 and 2 were prepared in a way similar to that described in the literature. [10,11] The FTIR measurements were carried out on a Bomene Canada instrument. NMR measurements were carried out on a 500-MHz Jeol instrument.

Synthesis of Dyes
Under an inert atmosphere of argon, a solution of 1 or 2 (9 mmol) in methanol (15 ml) was vigorously stirred with Na 2 CO 3 . After 1 h under stirring at 45 C, a dark violet residue was filtered and successively and intensively washed with methanol and cold water. The resulting precipitate was filtered and dried. To a solution of the resulting compounds (5 mmol) in dimethylsulfoxide (DMSO) (3.5 ml), 0.8 ml of 1 M LiOH was added at 0 C. The mixture was stirred overnight at room temperature, and the reaction was quenched by the addition of 30 ml of H 2 O. The resulting precipitate was filtrated, dried, redissolved in acetone, and concentrated in vacuo to give the corresponding dye D1 or D2, respectively. The dyes were then purified by silica-gel column chromatography (ethyl acetate-hexane ¼ 1:2).

Dye-Sensitized Solar Cell (DSSC) Assembly and Photovoltaic Characteristics of the Resultant Solar Cells
A nanocrystalline TiO 2 film was coated on a transparent glass support. Each dye was adsorbed by dipping a coated glass in a 5 Â 10 À5 M ethanolic solution of each dye containing 7% 4-tert-butylpyridine and 50 mM 3a,7a-dihydroxy-5b-cholic acid (cheno) for several hours. The visible bands in the absorption spectrum of the dyes after adsorption on the nano-TiO 2 film only appeared after the TiO 2 electrodes were dipped in the dye solution for at least 18 h. The presence of 4-tert-butylpyridine and cheno is necessary to avoid surface aggregation of the sensitizer (D1 or D2). Finally, the film was washed with a 1:1 acetonitrile-ethanol mixed solution. Acenonitrile-ethylenecarbonate (v=v ¼ 1:4) containing tetrabutyl ammonium iodide (0.5 mol dm À3 ) was used as the electrolyte. The dye-adsorbed TiO 2 electrode, the Pt counterelectrode, and the electrolyte solution were assembled into a sealed sandwich-type solar cell. [24,25] For each solar cell an action spectrum was measured under monochromatic light with a constant photon number (5 Â 10 15 photon cm À2 s À1 ). J-V characteristics were measured under illumination with AM 1.5 simulated sunlight (100 mW cm À2 ) through a shading mast (5.0 mm Â 4 mm) by using a Bunko-Keiki CEP-2000 system.

Electrochemical Measurements
Electrochemical measurements of the synthesized dyes were carried out in solution in acetonitrile. The oxidation potential (E ox ) was measured using three small-sized electrodes. An Ag quasireference electrode (QRE) was used as the reference. Platinum wires were used as the working and the counterelectrodes. All electrode potentials were calibrated with respect to ferrocene (Fc)=ferrocenium (Fc þ ) redox couplet. An acetonitrile solution (2 ml) of dyes containing tetrabutylammonium perchlorate (0.1 mol dm À3 ) and ferrocene (ca. 1 mmol dm À3 ) was prepared. The electrochemical measurements were performed at a scan rate of 100 mV s À1 . [21] CONCLUSIONS Two new metal-free D-p-A type organic dyes to be used as sensitizers in DSSCs were designed and synthesized based on indigo, which is known to adsorb two thirds of the visible spectrum and possess a high light fastness by employing a naphthyl residue as the p-conjugation system and acrylic acid or cyanoacrylic acid as the acceptor units. These dyes were synthesized and their structures were identified by the use of FTIR, 1 H NMR, elemental analysis, UV-visible spectroscopic, and fluorometric techniques. The spectrophotometric properties of the prepared organic dyes in solution in THF and on nano-anatase TiO 2 films were examined. According to the results, D2, containing a cyanoacrylic acid as the acceptor group, showed a more bathochromic shift than D1. In all cases, the absorption maxima of D1 and D2 applied on the surface of a TiO 2 film gave a bathochromic effect compared to the corresponding dye spectra in solution. Finally, the prepared dyes were utilized in constructed DSSCs and their photovoltaic behaviors were assessed. A solar energy to electricity conversion efficiency of 3.11 and 3.45 percent were achieved for D1 and D2, respectively. D2, containing a cyanoacrylic acid, gave greater conversion efficiency than the D1-containing acrylic acid as the acceptor unit. This is attributable to the stronger electron-withdrawing ability of the cyanoacrylic acid group. Detailed experiments and investigation of the interfacial charge transfer processes of these dyes are currently in progress, aiming to further increase the overall performances of DSSCs fabricated with this new group of dyes.
The goal of this investigation was to show that how a colorant of antiquity could be utilized in modern high-technology DSSCs. These novel dyes gave more than 300-folds (i.e., at least two orders of magnitude) increase in the quantum efficiency compared to 7-carboxlic acid indigo (Ca. g carboxylated indigo ¼ 0.01% to g D1 ¼ 3.11% and g D2 ¼ 3.45%). Furthermore, a conclusion can be drawn that these novel dyes have good light-harvesting properties, which can further be improved by the use of additional substituents and=or by utilizing more complex p-conjugated entities.