You are on page 1of 11


D−D−π−A-Type Organic Dyes for Dye-Sensitized Solar Cells with a
Potential for Direct Electron Injection and a High Extinction
Coefficient: Synthesis, Characterization, and Theoretical
Supawadee Namuangruk,† Ryoichi Fukuda,‡,§ Masahiro Ehara,*,‡,§ Jittima Meeprasert,† Tanika Khanasa,∥
Somphob Morada,∥ Tinnagon Kaewin,∥ Siriporn Jungsuttiwong,∥ Taweesak Sudyoadsuk,∥
and Vinich Promarak*,∥,⊥

National Nanotechnology Center (NANOTEC), Thailand Science Park, Patumthani 12120, Thailand
Institute for Molecular Science and Research Center for Computational Science, 38 Nishigo-naka, Myodaiji, Okazaki 444-8585,
Japan Science and Technology Agency CREST, Sanboncho-5, Chiyoda-ku, Tokyo 102-0075, Japan

Center for Organic Electronic and Alternative Energy, Department of Chemistry and Center for Innovation in Chemistry, Faculty of
Science, Ubon Ratchathani University, Warinchumrap, Ubon Ratchathani 34190, Thailand

S Supporting Information

ABSTRACT: A series of organic sensitizers with the direct
electron injection mechanism and a high molar extinction
coefficient comprising double donors, a π-spacer, and
anchoring acceptor groups (D−D−π−A type) were synthesized and characterized by experimental and theoretical
methods for dye-sensitized solar cells. (E)-2-Cyano-3-(5″-(4((4-(3,6-di-tert-butylcarbazol-9-yl)phenyl)dodecylamino)phenyl)-[2,2′:5′,2″-terthiophene]-5-yl)acrylic acid showed performance with a maximal incident photon to electron
conversion efficiency of 83%, Jsc value of 10.89 mA cm−2,
Voc value of 0.70 V, and fill factor of 0.67, which correspond to an overall conversion efficiency of 5.12% under AM 1.5G
illumination. The molecular geometry, electronic structure, and excited states were investigated with density functional theory,
time-dependent density functional theory, and the symmetry-adapted cluster-configuration interaction method. The double
donor moieties not only contribute to enhancement of the electron-donating ability, but also inhibit aggregation between dye
molecules and prevent iodide/triiodide in the electrolyte from recombining with injected electrons in TiO2. Detailed assignments
of the UV−vis spectra below the ionization threshold are given. The low-lying light-harvesting state has intramolecular charge
transfer character with a high molar extinction coefficient because of the long π-spacer. Our experimental and theoretical findings
support the potential of direct electron injection from the dye to TiO2 in one step with electronic excitation for the present D−
D−π−A sensitizers. The direct electron injection, inhibited aggregation, and high molar extinction coefficient may be the origin
of the observed high efficiency. This type of D−D−π−A structure with direct electron injection would simplify the strategy for
designing organic sensitizers.

Since the report on dye-sensitized solar cells (DSSCs) with a
dramatic increase in the light-harvesting efficiency by O'Reagan
and Grätzel in 1991,1 this type of solar cell has attracted
considerable and sustained attention as it offers the possibility
of low-cost conversion of the photoenergy.2 To date, a DSSC
with a validated efficiency (η) record of >11% has been
obtained with ruthenium complex dyes such as the black dye.3
Though there is still room for improvement of the efficiency of
ruthenium-based DSSCs,4 ruthenium dyes are nevertheless
costly and hardly attained and normally have a moderate
absorption intensity.5 Enormous effort is also being dedicated
to developing efficient dyes suitable for their modest cost, ease
© 2012 American Chemical Society

of synthesis and modification, large molar extinction coefficient,
and long-term stability. Ruthenium-free dyes or organic dyes
meet all these criteria; therefore, there has been remarkable
development in organic dye-based DSSCs in recent years,6 and
efficiencies exceeding 12% have been achieved using dyes which
have broad, red-shifted, and intense spectral absorption in the
visible light region, 400−800 nm.7
Although remarkable progress has been made in the organic
dyes as sensitizers for DSSCs, optimization of their chemical
Received: May 9, 2012
Revised: November 15, 2012
Published: November 20, 2012
25653 | J. Phys. Chem. C 2012, 116, 25653−25663

thus. a systematic study of a series of dyes has not been done.29 confirmed that the SAC-CI method is potentially useful for theoretical design and investigation of novel dyes for DSSC. which is generated after electron injection. In terms of molecular structure. The electrochemical properties can be reasonably explained by the orbital energy levels. Further studies using various computational methods other than DFT will confirm such theoretical findings for modeling dyes with high efficiency. has to be lower than the I−/I3− redox potential for oxidation and regeneration of the dye.26 However. Previous works28.19 have reported that DSSCs with various triphenylamine (TPA) and carbazole (CAZ) based dyes have shown η up to 9. In this study.11−13 Most of the organic dyes used for highly efficient DSSCs have a long π-conjugated spacer between D and A. By the orbital model. but also to instability of the organic dyes due to the formation of excited triplet states and unstable radicals under light irradiation. Chem. however. Phys.24 The cyanoacrylic acid moiety acts as the electron-accepting and anchoring group. Organic dyes that are constituted by donor (D). because of their electron-rich nitrogen heteroatom. indicating the importance of their further investigation in DSSCs. the molar extinction coefficient can be enhanced with a lower tendency to aggregate and better thermal stability compared with those of the D−π−A structure.1%. and electronic excited states. The overall conversion efficiency by a 25654 dx. TPA and CAZ derivatives are also of interest in their hole-transport capabilities. and their ground-states geometries and MOs were calculated by DFT with the B3LYP method. πconjugation.1021/jp304489t | J. respectively.20 Incorporating a triarylamine group into the dye molecule increases the physical separation of the dye cation from the electrode surface. We found that one charge transfer state is obtained in the visible region.16 These studies suggested that organic dyes with the 2D−π−A structure may achieve better performance than the simple D−π−A structure. except for the Ru-based N3 dye. many groups18. to calculate electronic excited states. introduction of long π-conjugation segments gives prolonged rodlike molecules.17 In our design (Chart 1). for which many computational studies have been conducted. This feature is frequently interpreted by a simple orbital model as follows: the energy level of the lowest unoccupied molecular orbital (LUMO) of the dye must be higher than the CB of TiO2. were reported.9 Stability of the dye in its excited and ionized states is important for efficiency and durability. The density functional theory (DFT) calculations can predict the molecular geometry of the dyes in the electronic ground state. we also used the symmetryadapted cluster configuration interaction (SAC-CI) method. The energy level of the cationic dye. where two D moieties are connected to a π-spacer.8. experimental and empirical techniques are not enough. which may facilitate the recombination of the electrons to the triiodide and magnify aggregation between molecules. diphenylamine (DPA) and 3. while unoccupied MOs tend to localize in each moiety. Structure of the Dyes separation and collection.The Journal of Physical Chemistry C Article structures is still needed for further improvement in performance.6-ditert-butylcarbazole (t-Bu2CAZ) moieties are employed as the aromatic and additional donors. The two functional moieties are connected by oligothiophene segments rather than p-phenylene and vinylene because of their chemical and environmental stability as well as their electronic tenability. C 2012.doi.30 We found that the D−D moiety has a nonplanar structure.6. quantum chemical calculations are essentially needed for giving a guiding principle of molecular designs. the energy level of the highest occupied molecular orbital (HOMO) of the dye must be lower than the I−/I3− redox potential. the so-called type II mechanism. and broad ranges of conversion efficiencies were achieved by using the D−π−A structure. recently.14 On the basis of such findings. The known requirements for efficient dyes are complicatedly correlated in terms of molecular conformations. The close π−π aggregation can lead not only to self-quenching and reduction of electron injection into TiO2. There are rare reports on organic dyes with a structure of this type. organic dyes with a 2D−π−A structure. In this study. Synthesized dyes UB1−3 were characterized by experiment. avoiding unfavorable aggregation of dyes is also necessary to achieve high efficiency. great flexibility in molecular design is essentially important. Excited states can be calculated by time-dependent (TD) DFT. electronic therefore.21 Our D−D moiety has a nonplanar conformation in the ground state.23.15. and acceptor (A) moieties have such flexibility for designing. Some valence-occupied MOs have slightly delocalized character.16.6-ditert-butylcarbazol-9-yl)phenyl)-N-dodecylaniline donor moiety (D−D). in which an electron is directly injected from the ground state of the dye to the CB of TiO2 in one step. we report newly designed D−D−π−A-type dyes UB1−3 (Chart 1) for DSSCs with theoretical and experimental investigations. We expect that an improvement could be made by introducing an additional donor moiety into the aromatic amine donor to form the D−D−π−A structure. Additionally. Recently. the energy level of electronically excited dyes has to be higher than the energy of the conductive band (CB) of the TiO2 electrode for efficient electron injection.10 Furthermore. thereby forming a ((3.22 which should prevent the aggregation of the dye molecules and enhance the efficiency. Analyzing the Kohn−Sham molecular orbitals (MOs) is also helpful to understand the fundamentals of the electronic structure. Since the DSSC processes include the excitation and exchange of electrons. These requirements occasionally seem to conflict with each other. The observed ultraviolet−visible (UV−vis) spectra were assigned by the SAC-CI calculation. In terms of electronic structure. Electronic properties of dyes and the electron-injection mechanism are intensively studied using DFT/TDDFT methods. therefore.8. a high absorption coefficient in a wide range of visible light wavelength is required for light-harvesting efficiency. it may achieve a high rate of charge Chart 1. using theoretical methods rationalizes the design of new dyes. 25653−25663 .27 which is a wave-function-based theory. 116.25 To realize the design based on the new concept of dyes in DSSCs. has been receiving increasing attention because such a mechanism makes the sensitizing procedure simpler and more efficient.

116. UB2. All dyes can be crystallized in deep colored solids. TCO30-8.09 −0. T1. The present theoretical elucidations for the experimental findings contain several new physical insights. angles between the connecting moieties denoted in Figure 1. Although the t-Bu and dodecyl groups may affect the planarity of the molecules. The adsorption energy was calculated as UB3-dye-based cell was 5.35 As a model of the DSSC prototype system. we studied the cluster model (TiO2)38−dye Table 1.34 These computations were performed by GAUSSIAN 09. a Newport 818-UV silicon photodiode for power density calibration. Figure 1. Ground-State Geometry and Molecular Orbitals.1 M guanidinium thiocyanate. the nonplanarity exists at CAZ and the phenyl of DPA (D1−D2) due to steric repulsion between the H atoms with a dihedral angle of ∼63°. the thiophene (π-spacer). 0. The synthetic routes to prepare dyes UB1−3 are shown in Scheme S1 in the Supporting M tert-butylpyridine in a 15/85 (v/v) mixture of benzonitrile and acetonitrile was filled through the predrilled hole by a vacuum backfilling method.67 5. Computational Details. were also fabricated for comparison. In all the computations. as the sensitizer.10 −17.37 The full details of the computation are given in the Supporting Information. it reaches 89.7% of the efficiency of an N3-dye-based cell. six devices were fabricated and measured for consistency. The measured current density−voltage data were averaged from forward and backward scans with a bias step and a delay time of 10 mV and 40 ms. the TiO2 electrode with a cell geometry of 0. Eads = E[(TiO2 )38 −dye] − E[(TiO2 )38 ] − E[dye] (1) where E[(TiO2)38−dye] is the total energy of the (TiO2)38− dye system and E[(TiO2)38] and E[dye] are the total energies of the (TiO2)38 cluster and dye molecule. ■ RESULTS AND DISCUSSION Synthesis and Characterization.doi. this conjugation provides a significant factor for the strong photoabsorption of these molecules as seen later in the oscillator strength. Another phenyl group of DPA (D2).33 The UV−vis absorption spectra were calculated by SAC-CI with a doubleζ plus polarization (DZP) class basis.12%.The Journal of Physical Chemistry C Article systems with the TD CAM-B3LYP method and fully relaxed geometries that were optimized by the PBE/DNP method36 with DMol3. the t-Bu groups substituted at CAZ and dodecyl side chain substituted at DPA are replaced by hydrogen atoms.5 cm2 was treated with an aqueous solution of 4 × 10−2 M TiCl4 at 70 °C in a water saturation atmosphere.p) in CH2Cl2 dihedral anglea UB1 UB2 UB3 D1−D2 Ph−Phb D2−T1 T1−T2 T2−T3 T3−A −63. yet the conjugation holds across this N atom.87 −8. Solaronix) TiO2 layer were screen-printed on TiCl4treated fluorine-doped tin oxide (FTO).20 40. The molecules have a bent structure due to the N atom of the diphenylamine unit. The TiO2 electrodes were immersed in the dye solution (3 × 10−4 M N3 in ethanol and 5 × 10−4 M organic dyes in CH2Cl2) in the dark at room temperature for 24 h to stain the dye onto the TiO2 surfaces. C 2012. Dihedral Angles (deg) between Moieties for the Ground State of the Dyes Optimized by B3LYP/6-31G(d. and then cooled to 80 °C. For each dye. A = cyanoacrylic acid.96 0. heated to 450 °C for 30 min.1021/jp304489t | J. and the detailed synthetic procedures and characterization data are provided therein as well. D2 = diphenylamine unit. and T3 = thiophene units. Solaronix) and a scattering (Ti-Nanoxide R/SP. respectively. Solaronix) as a spacer between the electrodes. 25653−25663 . We showed the rational design in which theoretical and experimental techniques cooperate is important to improve the performance of organic DSSCs.34 40. All measurements were performed using a black plastic mask with an aperture area of 0. Solaronix) were used for transparent conducting electrodes. An electrolyte solution of 0.28 1.180 cm2 and no mismatch correction for the efficiency conversion data.38 a D1 = carbazole unit. Chem.31 The incident photon to electron conversion efficiency (IPCE) of the device under short-circuit conditions was determined by means of an Oriel 150 W Xe lamp fitted with a Cornerstone 130 1/8 m monochromator as a monochromatic light source. the DFT/TDDFT calculations of a (TiO2)38−dye cluster suggested that the electron injection proceeds directly from the dye to TiO2 by photoabsorption. 0. and the averaged cell data were reported. The optimized S0 state structures of UB1−3 are shown in Figure S1 in the Supporting Information. As we have expected. the thiophene units have a zigzag 25655 dx. Prior to dye sensitization. The colors of these solid products changed from orange to dark red as the number of thiophene units in the molecule increased from UB1. In UB2 and UB3. Schematic view of the ground-state (S0) structure for UB3.12 41.25 −63. respectively. The ground-state (S0) geometries of UB1−3 were optimized by the B3LYP/631G(d. Fluorine-doped SnO2 conducting glasses (8 Ω/sq. to UB3.62 −1. and 0. T2.6 M LiI.55 −20. bPh−Ph = dihedral angle between two phenylene rings of the diphenylamine moiety. The experimental details are described in the Supporting Information. and the A moieties are nearly planar.p)32 method in CH2Cl2 solvent of the conductor-like polarizable continuum model (C-PCM).82 −63. The double nanostructure thick film (∼1 μm thickness) consisting of a transparent (TiNanoxide 20T/SP. Table 1 shows the dihedral ■ EXPERIMENTAL AND COMPUTATIONAL METHODS Experimental Methods. Phys.5 × 0.03 M I2. and a Keithley 6485 picoammeter. The direct injection caused by intense absorption of visible light may be the origin of the high efficiency of the present dyes. The reference cells with the same device configuration based on the N3 dye. We also performed calculations of a model of dye adsorption on TiO2. The dye-adsorbed TiO2 photoanode and Pt counter electrode were assembled into a sealed cell by heating a gasket Meltonix 1170-25 film (25 μm thickness.27 0.

and the contribution from each moiety is summarized in Table 2. ΔH−L.90 −4. which are decomposed as 38% (π-spacer) and 47% (A) for UB1. The experimentally predicted HOMO energy levels show very good agreement with the orbital energy levels obtained by B3LYP shown in Table 2.3 V above the conduction band of TiO2 to ensure efficient charge injection. T = thiophene unit.46 eV (UB1) > 2. Phys. which is a major characteristics of D−π−A-based dyes. respectively.85. Nb2O3.40 The energy level of the excited dye molecule should be >0. Thermal. The calculated HOMO− LUMO energy gap.02 −2.32 −1.85 0.07 V vs the normal hydrogen electrode (NHE)) is more positive than those of both UB2 (1.20.26 V vs NHE) are more negative than the CB-edge energy level of the TiO2 electrode (−0. dThe ground-state oxidation potential (Eox) or the first oxidation potential of the dyes adsorbed on a TiO2 film was converted to the oxidation potential of NHE by addition of 0.88 V vs NHE). Dyes UB1−3 in CH2Cl2 demonstrate multiple quasireversible oxidation processes and one irreversible reduction process. Therefore. 1.24 0. Electrochemical. 1.17 −2. 73%. 53% (π-spacer) and 41% (A) for UB2. These figures show that the HOMO and HOMO − 1 are delocalized over the D− D moiety. with no additional peak at lower potential on the cathodic scan (Epc) being observed. As a result. A = cyanoacrylic acid. and Electron Contribution of the HOMO and LUMO of UB1−3 Calculated by B3LYP/6-31G (d.04 eV (UB3). cCalculated from the onset oxidation potential (Eonset) of the dyes dissolved in CH2Cl2 with glassy carbon as the working electrode: HOMO = −(4. which are 91%. The Eox* potentials of these dyes (−1. 1.22 V.74. and SrTiO3 and their composites. the ground-state oxidation potentials (Eox) of the dyes were calculated. The electrochemical properties of the dyes were studied by cyclic voltammetry (CV) and differential pulse voltammetry (DPV). while the planarity in the D2−π−A moieties would contribute to the high absorption coefficient as the computational results show.60 −1.13/−2.76 −5.71 −5.76 −5. The Eox level of UB1 (1. Chem. 1. The CV and DPV curves are shown in Figure S4 in the Supporting Information. LUMO = HOMO + E0−0. and UB3. We expect the nonplanarity between the D1 and D2 moieties to inhibit aggregation of dyes on the TiO2 surface.The Journal of Physical Chemistry C Article Electrochemical and Thermal Properties.24 −1. Their multiple CV scans reveal identical CV curves. 1. all dyes have enough energetic driving force for efficient DSSCs using a nanocrystalline titania photocatalyst and the I−/I3− redox couple. The CV scans of dyes UB1−3 adsorbed on a TiO2 film show one quasireversible oxidation with potentials of 0. all dyes have a sufficient driving force for electron injection from the excited dyes to the conduction band of TiO2. which could hamper the dye regeneration.p)a electron contribution (%) dye UB1 UB2 UB3 LUMO HOMO LUMO HOMO LUMO HOMO energy (eV) ΔH−L (eV) D1 D2 T A −2. These values show that the distributions of the HOMO and LUMO of the dyes are wellseparated. C 2012.26 284 336 362 a Measured by DPV with the dyes dissolved in CH2Cl2 and glassy carbon as the working electrode.44 + Eonset). therefore. 116.87 −5.26 V vs NHE) these dyes become very attractive for other metal oxide semiconductors having conduction bands more negative than the conduction band of TiO2.07 1. the characterization of the HOMO and LUMO is the essential issue. and they are electrochemically stable molecules. This feature also enables the desirable efficient electron transfer with strong photoabsorption.39 0. and Electronic Properties of the Dyes dye E1/2(ox)a (V) E1/2(re)a (V) E1/2(ox)b (V) HOMO/LUMOc (eV) Eoxd (V) vs NHE Eox* e (V) vs NHE T5df (°C) UB1 UB2 UB3 0. The orbital densities of the HOMO are distributed over the D−D moiety.07.4 V vs NHE). tends to decrease as the effective conjugation length increases: 2. and all the numerical data are listed in Table 3.55 0. on the other hand. with high Eox* potentials (−1. but all are much more positive than the redox potential of the I−/I3− couple (0. calculated by B3LYP/6-31G(d.80. are delocalized across the π-spacer and A moieties. The Hartree−Fock MOs are used for excitedstate calculations and are qualitatively similar to those of Kohn−Sham MOs in the valence region. dye regeneration should be thermodynamically favorable and can be competed efficiently with the recapture of the injected electrons by the dye radical while the LUMO is localized on the A moiety. 25656 dx.22/−2.81 0. suggesting that the HOMO → LUMO transition can be regarded as an intramolecular charge transfer (ICT). Moreover.08/−2. D2 = diphenylamine unit. The excited-state oxidation potentials (Eox*) of the dyes were estimated from the groundstate oxidation potentials (Eox) of the dyes and the 0−0 excitation energy (E0−0) shown in Table 3. Orbital density analysis (the gross orbital population of the Mulliken charge) using the GaussSum program38 was carried out for the HOMO and LUMO of all dyes. 0.26 eV (UB2) > 2. Although this quite good agreement structure which maintains the near planarity of the molecules. such as ZnO.04 a D1 = carbazole unit.1021/jp304489t | J. signifying no oxidative coupling at the 3.6-positions of the peripheral CAZ due to the t-Bu substituents. The orbital densities of the LUMO. This type of electrochemical coupling reaction can be detected in some CAZ derivatives with unsubstituted 3. Table 2 lists the HOMO and LUMO energies of UB1−3 Table 2.6positions39 and might occur upon charge separation. Table 3. and therefore.91 −5.98 −1.5 V vs NHE) and the equivalent potential of the N3 dye (−0.94 2. resulting in a much improved cell efficiency.46 0 57 0 29 0 20 15 34 6 44 2 42 38 6 53 22 60 35 47 3 41 4 38 3 2.41 to achieve a high open-circuit voltage (Voc).23 −1. fMeasured by TGA. and 62% for UB1.71 −1.85.p). Energy Level. respectively. 25653−25663 .98 V vs NHE). HOMO−LUMO Energy Gap (ΔH−L).doi.26 2. and 60% (πspacer) and 38% (A) for UB3. eEstimated excited-state oxidation potential (Eox*) vs NHE calculated from the ground-state oxidation potential (Eox): Eox* = Eox − E0−0. From these oxidation potentials.03 0.03 V vs NHE) and UB3 (0. The relevant photoabsorption in these dyes is characterized as a HOMO → LUMO transition as shown later. UB2.23 to −1. and 0.74 V versus Ag/ Ag+.84 1. bMeasured by DPV with the dyes adsorbed on a TiO2 film as the working electrode.

the B1 bands show a bathochromic shift and increased ε. The observed E0−0 also agreed well with the calculated ΔH−L. and B3. a 25657 dx. the complete lists are given in the Supporting Information (Tables S4−S7). cThe 0−0 transition energy. Figure 2 shows that the HOMO has a large coefficient at Figure 2.18. SACCI reproduced the UV−vis spectrum satisfactorily to provide reliable assignment and characterization of the electronic spectrum.6-positions of CAZ are active in the oxidized (cationic) form of the dye. As the number of thiophene units in the molecules increases from UB1 to UB2 and UB3. the experimentally predicted LUMO energy levels agreed well with the calculated one. In solution. covering the region of 350−600 nm.45 The absorption spectra of UB1−3 adsorbed on the TiO2 films (Figure 3b) are broadened. Theoretical Absorption Spectra of the Dyes. may be caused by some error cancellations. and it was difficult to assign the spectra by TDDFT calculations only. The positions of B2 and B3 are nearly independent of the solvent polarity. Supporting Information). E0−0. dipole moment..6-positions of CAZ. The former is ascribed to the interaction of the anchoring groups of the dyes with the surface of TiO2 and is commonly observed in the spectral response of other organic dyes.doi. This also benefits the electrolyte diffusion in the film and reduces the recombination possibility of the light-induced charges during transportation.44 indicating excellent lightharvesting ability. oscillator strength. Moreover. therefore. bDyes adsorbed on a TiO2 film. and slightly blue-shifted (20−28 nm) compared to their solution spectra. where the states with f > 0. Those results suggest that dyes UB1−3 are thermally stable materials with a temperature at 5% weight loss (T5d) well over 284 °C (Figure S5 in the Supporting Information).42 UV−Vis Spectra. UB2 (red circles). This indicates that B1 is ascribed to an ICT transition from the D−D moiety to the A moiety.24 Measured in CH2Cl2 solution. orbital relaxation and reorganization effects are relatively small for organic molecules in comparison with those for transition-metal compounds. all the dyes display three strong absorption bands appearing at 450−600. The thermal properties were investigated by thermogravimetric analysis (TGA). where orbital relaxation is significant even for vertical detachment processes. and UB3 (blue inverted triangles). ε = 17 000 M−1 cm−1 in CHCl3)43 having a D−π−A structure bathochromically shifts in the ICT peak together with significant increases in the molar extinction coefficient (about 1. 330−340.1021/jp304489t | J.The Journal of Physical Chemistry C Article tively. ε of B1 of these dyes is moderately high. B2.31 2.05 are shown. while the ICT character is small for B2 and B3. the 3. i. The TDDFT calculations shown in Tables S1−S3 in the Supporting Information suffered severe functional dependence. and 297−300 nm.24. estimated from the intercept of the normalized absorption and emission spectra of the dyes in CH2Cl2: E0−0 = 1239. a hypsochromic shift of maximum wavelength (λmax). B1 shows a negative solvatochromic shift. in more polar solvents (Figure S6. ε of all dyes at the ICT band is considerably larger than that of the standard ruthenium dye N3 at 518 nm (ε = 13 000 M−1 cm−1). the 3.48-fold). It should be noted that the absorption spectrum of UB1 compared to that of (E)3-(5-(4-(diphenylamino)phenyl)thiophene-2-yl)-2-cyanoacrylic acid (λICT = 379 nm. UV−vis absorption spectra of UB1−3 in a dilute solution of CH2Cl2 and on TiO2 films are shown in Figure 3. respecTable 4. 25653−25663 . Chem.84/λintercept. and the characteristic data are summarized in Table 4. The theoretical absorption spectra calculated by SAC-CI are shown in Figures 4−6 for UB1−3 with the experimental spectra.e.46 The latter can be explained by the nonplanar structure of the tBu2CAZ unit of the dyes preventing aggregation via molecular stacking. The D−D−π−A structure is effective for a bathochromic shift in the absorption spectrum and enhances ε. and transition character for UB1 calculated by SAC-CI are summarized in Table 5. The protection of these positions by t-Bu groups is favorable for stability of the oxidized form of the dye according to the theoretical prediction. Phys. On the basis of frontier orbital theory. these bands can be assigned as localized π → π* transitions of CAZ and DPA. We concluded that the long-range correction is at least necessary for studying these compounds by TDDFT. The related excitation Optical Properties of the Dyes dye UB1 UB2 UB3 λabsa (nm) 457 463 472 εa (M−1 cm−1) 25 174 27 460 30 102 λabsb (nm) 429 435 452 λema (nm) 598 630 635 E0−0c (eV) 2. (a) Absorption spectra of the dyes in CH2Cl2 solution and (b) absorption spectra of the dyes on TiO2 films: UB1 (black squares). We denote these bands as B1. The greater maximum absorption coefficients of the organic dyes allow a correspondingly thinner nanocrystalline film so as to avoid the decrease of the film mechanical strength.26 2. 116. Figure 3. C 2012. ranging from 25 174 to 30 102 M−1 cm−1. The better thermal stability of the dye is important for the lifetime of the solar cells. Hartree−Fock MOs of UB1 relevant to the excited states.

this can be a serious disadvantage. In the experimental UV−vis spectrum. The details of the absorption spectra can be analyzed using the computational results.15 eV may be assigned to the 19th and 20th states calculated at 6. below the ionization threshold. are not favorable because molecular vibrations are Figure 5. we also evaluated the dipole moment (μ) for the ground and excited states. The SAC-CI results were convoluted with a Gaussian envelope: the fwhm was 0. and B4. The SAC-CI calculation can consistently assign these 25658 dx. we could not obtain high-intensity states around 6. Figure 6. (a) Experimental and (b) SAC-CI absorption spectra of UB2. in excellent agreement with experiment and characterized as a HOMO → LUMO transition that is ICT as explained in the previous section. Our SAC-CI and TD CAM-B3LYP calculations showed that UB1−3 have only one ICT excited state in the visible light region. (a) Experimental and (b) SAC-CI absorption spectra of UB3. S1 and S2. however.doi. while other absorption bands are composed of several electronic transitions.81 eV. such as conjugated macrocyclic compounds. with two shoulders.4 eV for the lowest state and 0. The spectra can be assigned in a similar manner. The higher excitations relate to the contributions of local π → π* transitions or a mixture of ICT and π → π* characters. The B2 band observed corresponds to the second state calculated at 4. C 2012. can be identified for UB1 at 600−200 nm.34 eV.1021/jp304489t | J.2 eV for the other states.71 eV (457 nm). For efficient light-harvesting properties. B2. From μ and Δμ.4 eV for the lowest state and 0. B3. This feature seems to be common to many other D−π− A-type dyes. The SAC-CI results were convoluted with a Gaussian envelope: the fwhm was 0. consequently.2 eV for the other states. The band structures of the absorption spectra of UB2 and UB3 shown in Figures 5 and 6 are very similar to that of UB1. B1. The fifth and sixth states correspond to the observed B3 band. very rigid π-frameworks.3 eV may be assigned to the 11th state calculated at 4. Within 20 states solved.11 and 6. This can be simply explained by the decrease of the HOMO−LUMO energy gap. state calculated corresponds to the S2 The lowest absorption in the visible light region contains only one electronic transition. The SAC-CI results were convoluted with a Gaussian envelope: the fwhm was 0. Our findings obtained by the SAC-CI calculations can be summarized briefly as follows: the lowest absorption that corresponds to the HOMO → LUMO transition with ICT character shifts to lower energy with increasing length of the thiophene chain. 25653−25663 .02 eV. it yields a narrow absorption band. 12th. Phys. (a) Experimental and (b) SAC-CI absorption spectra of UB1. The lowest excited state was calculated at 2. which may correspond to the B4 band of UB2 and UB3. structures of the absorption spectrum.0 eV. The S1 shoulder observed around 4. which are included in Table 5. four absorption bands. This excitation is responsible for the electron transfer from CAZ and DPA donors to the thiophene π-spacer and cyanoacrylic acid acceptor.2 eV for the other states.The Journal of Physical Chemistry C Article Figure 4. although the first absorption band is the most relevant in the present DSSC molecules. In this sense. For the characterization of the excited states.72 eV (455 nm) by SAC-CI. Chem.4 eV for the lowest state and 0. The broadness of the ICT band of UB1−3 is attributed to the Franck−Condon factor: the effects of molecular vibrations and solute−solvent interactions. respectively. The absorption center of B1 was observed at 2. the present computational conditions are insufficient for quantitative agreement of excited states in the higher energy region. the excited states with considerable oscillator strength are assigned to ICT in the low-energy region and the mixture of π → π* and ICT in the high-energy region. The B4 band observed at 5. Then the next. 116.

43 0. respectively (Table 6).The Journal of Physical Chemistry C Article Table 5. moderate rigidness is also required for a high absorption coefficient and also stability of the dyes.92 12.287 0.44 0. therefore.58 0. D).55 4. a Table 6. relative η of D−π−A-type sensitizers in the range of 49−90% (see Table S8 in the Supporting Information for details). bObtained from dye adsorption measurements.50 0. the IPCE values of dyes UB1−3 are higher than that of the N3 dye.48 The chemisorption of all dyes onto the surface of TiO2 films was confirmed by FT-IR spectroscopy. eV (nm) ΔE.5 104. and (104.8 1.4 ± 3. cRelative η compared with the reference N3 value.99 11.89 9.71 (457) 3.16 (298) B3 4.36 (H → L + 4) − 1 → L + 12) → L + 12) + 0. the difference observed in performance can be directly related to the effect of the molecular volume and how much dye is absorbed. These profiles are typical for organic adsorbates into nanoporous inorganic matrixes.07 26.doi.91 −0.89 11.81 5. Phys.98 8.26 1. Dipole Moment (μ. The largest dye.12 5. Figure 7. The IPCE spectra of UB1−3 sensitizers plotted as a function of the excitation wavelength are broadened and red-shifted as the number of thiophene units in the molecule increases.67 0. the N3 dye shows a broader IPCE spectrum.72 4.4 97.66 0.8) × 1015.75 (331) B1 B2 4.137 0.70 3. dObtained from integration of the corresponding IPCE spectra.88 10. UB3 (blue triangles).31 −1.069 0.56 4. UB2 (red circles). a leading to a small incident monochromatic photon to current conversion efficiency for the UB1-based cell. such steric hindrance can be reduced in UB3 with longer π-conjugated spacers.453 0.02 4.15 (241) S1 S2 B4 2.10 5. The performance of a DSSC also depends on the total amount of dye present.70 0. parameters (average values) are summarized in Table 6.36 6. Dyes UB1−3 were used as the sensitizers for dye-sensitized nanocrystalline anatase TiO2 solar cells.25 cm2 working area on the FTO (8 Ω/sq) substrates.0) × 1015.32 (H − 8 → L) − 1 → L + 5) → L + 4) + 0.45 4. If the excited-state geometry of the molecule is significantly different from the geometry in the ground state. the dye uptake was determined.61 10.12 5.074 0.3 (288) 4. UB2.29 Experiments were conducted in identical conditions using a TiO2 photoanode with an approximately 11 μm thickness and a 0. of Figure 7.76 18.31 D.47 In all case. D).86 −1.3 55.61 3.1021/jp304489t | J. Flexibility of the present D−D−π−A framework could contribute to the broadness of the ICT band. Photovoltaic Properties. (97.67 0. At equilibrium.0 Jsc (mA cm−2) Voc (V) FF η (%) η/ηN3c (%) calcd Jscd (mA cm−2) 7.52 4. eV (nm)).8 2. Performance Parameters of DSSCs Constructed Using Dyesa dye UB1 UB2 UB3 N3 dye uptakeb (1015 molecules cm−2) 84.49.28 (H → L) − 2 → L + 1) + 0.34 (455) (309) (279) (272) (272) (258) (231) (203) (196) μ (D) Δμb (D) f 19. which is slightly higher than the IPCE values of both UB1 (81%) and UB2 (82%). and N3 (green tilted squares) dyes.45 6. The dye uptakes depend on the steric hindrance of the donor moiety around the carboxylic acid anchoring group. maximum uptakes of each dye are (84.70 0.58 11. which enhances the electron-injection yield in comparison with those of the other dyes.57 9. which is coincident with the result of the absorption spectra.31 (H − 1 → L + 12) − 8 → L) − 1 → L + 5) + 0. 116.5 ± 2. and UB3.62 7.43 (H − 4 → L + 5) The states with f > 0.1 ± ± ± ± 3. The characteristic vibration modes of the carboxylate group of all dyes are identical to those reported for other dyes. the Franck−Condon overlap becomes small and the absorption coefficient decreases.3 ± 1. Oscillator Strength ( f). C 2012.23 7. respectively. Under this dye-loading result. which is consistent with its wide absorption 25659 dx. and Transition Character of the Singlet Excited States of UB1 Calculated by SAC-CIa SAC-CI state exptl ΔE. and the ground-state dipole moment is 8.50 Because of their larger molar extinction coefficients. therefore.66 0.60 (H (H (H (H (H (H (H (H (H → L) → L + 5) + 0.76 0.71 0. showed the broadest absorption. which showed absorption peaks of both the dyes and TiO2 (Figure S8 in the Supporting Information).8) × 1015 molecules cm−2 for UB1.945 amplitude (orbital transition) 0. eV (nm) 1 2 4 5 6 11 12 19 20 2.50 This indicates that all dyes bind in the same way to TiO2. Absorption Energy (ΔE. the light-harvesting efficiency of UB1 is expected to be less than those of the other dyes.19 8.191 0. the rate of dye adsorption is initially rapid and eventually reaches a plateau (see Figure S7 in the Supporting Information).54 0. Dye Adsorption on a TiO2 Film.32 1. Chem. The higher IPCE value of UB3 is probably due to its ε.19 13. The corresponding IPCE plots and current density− voltage (J−V) characteristics are shown in parts a and b.05 are shown.67 0. Δμ = |μES| − |μGS|. and the resulting photovoltaic restricted and such compounds usually show sharp absorption bands.7 (264) 5. bValues show the changes in the dipole moments from the ground state (GS) to excited states (ESs).73 0.70 0. The IPCE spectrum of UB3 shows a high maximum value of 83%. Therefore.70 0.192 0.71 62 72 90 100 6. 25653−25663 .0 1.11 6.086 0. Dipole Moment Change (Δμ. (a) Photocurrent action spectra and (b) J−V characteristics of the DSSCs based on UB1 (black squares).

partly verifying the reported efficiency and the results found to be in agreement to within 3% (Table 6).12%. While the lower efficiency of the UB1based cell compared to the UB2.999. C 2012. this ε × dye uptake value shows excellent linear correlation with the observed Jsc as shown in Figure 8.88 mA cm−2.doi. The linear regression line is Jsc = 2. These transition characters and MOs of the (TiO2)38dye model system show that the sensitization mechanism in the prototypes is an interfacial direct charge transfer process corresponding to electron injection from the excited dyes to the conduction band of the TiO2 surface.52%). we model the prototype system with the dye adsorbed on the anatase (TiO2)38 cluster. involves photoexcitation to a dye excited state. 3.1021/jp304489t | J. 116. The length of the thiophene chain does not affect the efficiency of electron injection.12%) > UB2 (8. To obtain a further understanding of the electronic structure involved in the dye/semiconductor charge transfer process in the prototype system. and then an electron is transferred to the semiconductor. For the (TiO2)38−UB2 and (TiO2) 38−UB3 complexes. This relative efficiency (η/ηN3) enables the comparison of the solar cell performance in different experimental conditions. If we assume the ε value of free dyes is directly proportional to that of adsorbed dyes. This result indicates that the direct injection mechanism is identified in the present dyes.19 mA cm−2. where the intercept is approximately zero. Our dyes chemisorb on the TiO2 surface as suggested by FT-IR spectroscopy. Under standard AM 1. with the Ti−O bond distances ranging from 2. therefore.1 kcal/mol. Two Ti−O bonds form. indicating the strong coupling between the dye excited state and the conduction band states of TiO2.89 mA cm−2.20 Å. the potential for direct injection of these dyes was examined by theoretical calculations for the dye−TiO2 cluster model system in the next section. For the present dyes. the excitation from the HOMO − 1 to the LUMO + 12 and LUMO + 21 contributed to the strong absorption in the visible light region. The efficiency of the UB1based cell also reaches >62% of the efficiency of the N3-based cell even though UB1 has a lower IPCE value and narrower IPCE spectrum than both UB2 and UB3.056(ε × dye uptake) − 0. The measured Jsc values of these solar cells were also crosschecked with the Jsc values calculated from integration of their corresponding IPCE spectra (calcd Jsc). indirect injection. Chem. which shows that the use of D−D−π− A-type sensitizers improves the energy harvesting of the DSSC by decreasing aggregation and increasing the molar extinction coefficients.3. This length independence of the injection mechanism strongly suggests the direct injection mechanism for UB1−3 because such a mechanism may not explicitly involve the electronic states of the thiophene units. The absorption peaks were characterized in the gas phase using TD CAM-B3LYP52 with the 3-21G(d) basis. The HOMO − 1 is delocalized over the dye.5G 100 mW cm −2 illumination. Model of the Dye/Semiconductor Prototype. the UB3-sensitized cell shows the highest overall efficiency among the three dyes and gives a short-circuit photocurrent density (Jsc) of 10. we calculated only the chemisorption configuration which is formed by the two carboxylate oxygen atoms bonded with the two Ti atoms of the TiO2 surface. This linear correlation states that Jsc is determined only by the absorbance of a photon.The Journal of Physical Chemistry C Article spectrum.70 V. The η/ηN3 values of UB1−3 (Table 6) are competitive with or higher than those of the D−π−A-type sensitizers containing carbazole or diphenylamine as the donor (49− respectively. 25660 dx. while the LUMO + 12 and LUMO + 21 are distributed at the (TiO2)38−dye interface. and this is one origin of their excellent performance as sensitizers. CT bands are not clearly observed in Figure 3 for the adsorbed system. and UB3 on the (TiO2)38 cluster are −10.and UB3-based cells can be attributed to both the poorer spectral property and the lower dye content on the TiO2 film. thus. ε (mM−1 cm−1) multiplied by dye uptake (mol cm−2) leads to a dimensionless value. Phys. a one-step electron injection from the ground state of the dye to the conduction band of the semiconductor by photoexcitation. corresponding to an overall conversion efficiency (η) of 5. The MOs relevant to this transition and conduction band are shown in Figure 9. the amount of excited dyes is proportional to this ε × dye uptake value and also linearly correlates with Jsc. 5. The electronic structure and excitation energies in the present (TiO2)38−dye model systems were examined by the TDDFT calculations at the optimized geometries. open-circuit voltage (Voc) of 0. although the LUMO energy level and Eox* depend on the chain length. Linear relation between Jsc and ε × dye uptake.07 to 2. The better solar cell performance (highest η and Jsc) of the UB3-based cell than that of the other dyes in this series can be explained by the red shift of the absorption spectrum of UB3 compared to UB1 and UB2: this is better for the lightharvesting efficiency.71%).53 The results are shown in Table S9 in the Supporting Information. The efficiency of the UB3-based device reaches 90% of the efficiency of the standard ruthenium dye N3-based cell (η = 5. the zero intercept implies that no current density is observed for zero absorbance. see Table S8 in the Supporting Information for details). This value corresponds to the absorbance of the optical spectra. and fill factor (FF) of 0. UB2.04 with R2 = 0. The second is a direct mechanism. −13.9. This theoretical and experimental evidence altogether supports the direct injection mechanism for the Figure 8. This is the experimental evidence for direct injection. and −13. 4.67. which can be classified into two types. It is well-known that the mechanism of electron injection from the dye to the semiconductor can be theoretically elucidated by the study of the electronic structure of the dye adsorbed on the semiconductor. 25653−25663 . In (TiO2)38−UB1. indicating bidentate chemisorption.11 The first mechanism. The short-circuit photocurrent densities and efficiencies of the DSSCs are in the order UB3 (10.10%) > UB1 (7. indicating strong interactions between the dyes and TiO2 surface. a similar picture was obtained (see Figures S10 and S11 in the Supporting Information).51 The calculated adsorption energies of UB1. while a proton of the dye forms a hydroxyl group with a surface oxygen.89 mA cm−2.

■ REFERENCES (1) O′Regan. 19. and electronic structure of the dye molecules were investigated by DFT and the SAC-CI method. Ho.D. Experiments were done at the Center for Organic Electronic and Alternative Energy (COEA). 414. M. Jsc value of 10. F.). DSSCs using these dyes exhibit efficiencies ranging from 3.-b. The computations were done at the Research Center for Computational Science in Okazaki. Zakeeruddin. which correspond to an overall conversion efficiency of 5.-J.. which shows a maximal IPCE value of 83%. Emery.6-di-tert-butylcarbazol-9-yl)phenyl)-N-dodecylaniline as the electron donor moiety (D−D)... The D−D moiety has a nonplanar structure that may inhibit unfavorable dye aggregation yet maintains the conjugation in the D−π−A moiety. 49−68. the conjugation in the D−π−A moiety is important for a high intensity of photoabsorption. M.12% under AM 1. and at the Nanoscale Simulation Laboratory in National Nanotechnology Center (NANOTEC). 25661 dx. (b) LUMO + 12. Hishikawa.. Wang. Noda. Yanagida.52% to 5. H. Zhang. This direct injection mechanism was also analyzed in detail for other (V. Angew. M.. S. Department of Chemistry and Center for Innovation in Chemistry. 7342−7345.. Grätzel. We acknowledge scholarship support from the Royal Golden Jubilee Ph. Faculty of Science. The strategy for designing dyes will be significantly simplified.. Chem. J. and S. 353. The best performance among these dyes was found in UB3. Japan. (2) (a) Grätzel. Chem.K. electronic excitations.E. and the Electricity Generating Authority of Thailand (EGAT).. Photovoltaics 2011. Shi.89 mA cm−2. M. (4) (a) Chen. Masaki. ■ ASSOCIATED CONTENT S Supporting Information * Figure 9. Tian.67.. Although CT bands were not clearly observed for these dye−surface systems. This material is available free of charge via the Internet at http://pubs. Jing. M. 2010. Warta. (b) Hagfeldt. Li.12 The direct mechanism for D−π−A-type dyes with a cyanoacrylic acid anchor is not well-known. S. Institute of Science. 95. A. Xia. Voc value of 0. 2006. Z. Grätzel. 2460−2462.. Suranaree University of Technology. These dyes exhibit high thermal and electrochemical stability. they are mainly catechol-based dyes. (a) HOMO − 1. Soc. The ground-state geometry. Commun. The electronic mechanism related to light-harvesting and charge transfer processes in the dye/semiconductor system was also simulated with the model prototype system of the dye adsorbed on the (TiO2)38 cluster.13 Corresponding Author ■ AUTHOR INFORMATION *E-mail: ehara@ims.-G.. 1995.1021/jp304489t | J.-Y. 737−740.. M. S. J. acknowledges support from JST-CREST. J. ■ ACKNOWLEDGMENTS This work was supported by the Thailand Research Fund (Grant RMU5080052) (M. S. 130. NMR spectra. (c) Ning. Chem. 338−344. and cyanoacrylic acid as the electron acceptor (A) for DSSCs. M. The transition characters for the absorption peak are assigned to S0 → S1. 3. Rev.-C. A. Energy Environ. Y.5G illumination. Thailand.P. D.23 (H → L + 21).M. Thailand.N. Humphry-Baker. the linear relation between Jsc and ε × dye uptake and the theoretical calculation supported a potential for direct electron injection for the present D−D−π−A sensitizers.-Y... 47. Chem. respectively. Japan. Program (RGJ) and Center for Innovation in Chemistry (PERCH-CIC) to T. 2008. W.. K. 84−92. Wu. MEXT. 25653−25663 . Thailand. Wang. the Strategic Scholarships for Frontier Research Network for Research Groups (Grant CHE-RESRG50) funded by the Office of the Higher Education Commission. The present dyes are exothermically adsorbed on the TiO2 surface with chemisorption.. R. and Cartesian coordinates for the computations. Nature 1991.-J. The photoabsorption is shown as an interfacial charge transfer process regarding electron injection from the excited dyes to the conduction band of TiO2. The CAZ and The authors declare no competing financial interest. B. X. M. J. and (c) LUMO + 21 of the (TiO2)38−UB1 system. Thailand.11 however. acknowledges the post-JENESYS program named EXODASS for expenses during work in Japan. Nature 2001. K.. K.70 V. electron injection of the present dyes. and FF value of 0. Am... providing the reliable and detailed assignment of the spectra in the energy region of 200−600 nm below the ionization threshold... oligothiophenes as πconjugated spacers (π). The direct mechanism for D−π−A-type dyes with a cyanoacrylic acid anchor is not well-known. (d) Maestri.. Chem.-G. SPIRE. P..doi. Wu. S. 10720−10728. Phys. (3) Green. (b) Gao. which is related to the linear combination of −0..).. Fu. Ubon Ratchathani University. 116. N. M. Present Address ⊥ School of Chemistry and Center of Excellence for Innovation in the Chemistry. and CMSI. The SAC-CI method could reproduce the UV−vis absorption spectra satisfactorily. Chen. Wang. This work suggests that the organic dyes based on a double donor moiety are promising candidates for improvement of the performance of the theoretical calculations. 1170−1181. characterization. while the higher absorption was assigned to the mixture of the ICT and π → π* excited states. Y. additional figures and tables. Experimental procedures. Grätzel.E. The low-lying light-harvesting state was assigned to the ICT state. C. C 2012. pvinich@sut. Nakhon Ratchasima 30000.24 (H → L + 12) + 0.. C. J. Y. Sci. ■ Notes CONCLUSIONS We have developed new D−D−π−A-type organic sensitizers using ((3. (c) Jiang.12% (90% of the efficiency of the N3-based cell).acs. 2008.. and the conversion efficiency is not so high. Prog. such a type of dye is quite promising because the enhancement of photoabsorption in the visible region directly increases Jsc.The Journal of Physical Chemistry C Article DPA moieties contributed to the sensitization for both the bare dyes and the prototype systems. Ed.

Gaussian Inc. 2008. 2011. Lett. Phys.. P.. A. Renz.. V. Engel.. P. C.. Chem... Rev. E. 2011. Petersson.. 125. (5) Tian... D. H. R. V. Tachiya..The Journal of Physical Chemistry C Article suttiwong. Yang.. T. 2009. (b) Hehre.. J. 9550−9556. (9) Tatay. et al. M. Tian. J.-M. Chang. H. Chem. 2009. S. Chem. K. M. 1613−1624. Chem. 9. H. Org. S.. Ge. Phys.. Chem. S. Fang. A.. 2011. Chem. Cai.. H. G. Mater. 2011. Acta 1973. Nakatsuji.. A. (21) Xu. Y. 1343−1349. Y... F. J. C 2010. F. T. Ciofini. 2799−2805.. Y. Tsai. Armaroli. Yang. K. Li. 114304. S. et al. (b) Gather.. M. J. E. R. W.. D. A. Pople. 8005−8013. Pople. Yamashita. 20. (26) (a) Labat. Comte. Clifford.. Lin.. Park. J.. Chem... J. Liska. Kawabuchi. P. Phys. Lett.. Scalmani.. Mater.. Ehara. 2881− 2888... W. Chem. Gavina. (b) Le Bahers. Kim. Tetrahedron 2007. Chem.. Commun. A. J.. C.. J. Zhai. 56. K.. H.. Soc. (b) Perdew. A. S.. J.. N.. Yoon. V. W.. Frisch. 25653−25663 . 3791−3797. 2012. J. T. 1978... 3. R.Asian. Y. (b) Fukuda. Perpete. J. Sun. 34.. Phys. J. 2008. K. F. 1772−1779. Phys... Y. SAC-CI method applied to molecular spectroscopy. Nazeeruddin.. Li. H. J.. Perepichka. Shieh. G. H.. H.-W. Hehre. Lee. Phys.. Seibt. J. Chem.. 114. pp 1−27.. Chem. Soc. Sun.. P. Lainé.. Phys. In Modern Theoretical Chemistry. Lin. Burn. N. C. T. Chou. Lee. J. G. 4887− 4889.. 114. J. Y. K. X. 2007. S.. 516. Sudyoadsuk. In Theory and Applications of Computational Chemistry: The First 40 Years. 73.. 123. Chem. R. J.. Chem. Schlegel. Rev. S. Q. (b) Agrawal. S. J. Long. 75. R. A. Soc. A. A. 2007. Ahn.. P. Phys. X. Bauerle. I. P. C 2008. Int. Zhu. K. Wang. Dev. 2006. W. B 2001. 2005. 12858−12859. (c) Sayama. Chem.. R. Phys. Nonomura. 2007.. E. 508−517. (37) (a) Delley. H.. Lee.. M.. Burke. J. (7) Yella. T. Phys. (10) Mishra. (19) (a) Koumura.. 1995. M. Hannongbua.. W. G. A 2006. (b) Tsao. 362−364. E. J. (33) (a) Improta.-H. B 2005. 2011. I. 1997. D. 273−279. B. Comput.. Chem.. English. D. 1971. 2006. K...-Y. 2009. Y. J. L.. Kaupp. R. Nakatsuji. Sun. Liang. Koumura. M. M. J. Meng. Ovied. (30) (a) Lee. Huang.. 11591−11599. Liu.. R.-P. M. 5648−565. Shen. Chem. D. Tian. Sanz. W... Am. Constable.. 329−333.. Dev.... 2005. Lee. Frisch. 3993−4003. M.. W. T. B 2002. 2009. B. (28) Ehara. Hara. Haque.01. Adamo. T. Batsanov. Commun. Mater. H. 111. Wu. Soc. M... M.-C. 323− 327. 3654−3666..-H. E. Mori.. G. X. Phys. P... J. K... M. Chem. Yum. Brinck.. O′Regan.. Ho... T. (c) Qu. G. M. J.. (40) Furube.-H. M. Y.-Y.. M. F. C 2009. Watanabe. Chem. Phys. 2006. S. Trucks.. C. Hsu.. Pei. J. Yeh. A. J. K. P. Inorg. S.. 1982. (d) Saha. 14290− 14298.. Nazeeruddin.. 2257−2261. Chen.... Lee. S. S. V. Mori. Theor. 37. M.. Satsuki. H. Hehre. Cheeseman. H. M.. S. K. English. C. K. Hara. Materials. M... (c) Poolmee. Chem...-H... W. Jungsuttiwong.. Barone. P. M.-T. Phys.. K.. Chem. 074504. Chem. J. J. Sugihara.. S. Liu. S.-H. 131. Chem. Zakeeruddin. M. Am. Langner.. Katoh. A 2012. Grätzel.. P. Lett. F. 63.... Lin. Chem. Jiu. 2. Stuhl. Robb. J. 416−424. W. Am. 2010. et al. Hratchian.. Chin. Rev. (15) (a) Chen. 785− 789.. P. Péchy. T. (c) Jacquemin. (22) Wong. S. (c) Velusamy.-M. Chem. Durrant. W. Ehara. (24) Justin Thomas. Holmes... V. 28. C. V. 270−283. R. Hannongbua. M. Mennucci.-W. Wu. J. 20. S. Cevey. B. Lett. J. 3−6.. 161−173. (b) Dunning. G. Thampi. Chem. Hara. Yi. Forneli. Scalmani. Chem.. J. Miyashita. H. Fungo. Chem.. W. 112. S.. Hasegawa. P. Lochbrunner.-H. K. Nakatsuji. (16) Ning. Yang. B.. R. 144316. Gabrielsson... 2010. 629−634. Chem. E. C.. X.-F. D. Tsukagoshi. Han. N.. (38) O′Boyle. B. 29. Chem. (b) Becke. 2474−2499.. K. Frisch.. B. 54. 98. C. 2007. Renouard. Phys. III. Peng.. Y.. M. ChemSusChem 2011.. Cui. Vidal-Ferran. (20) (a) Promarak. Phys... Phys. (36) (a) Perdew. Diau.-K. A 2007. J.. S. N. K. Petty.. Thompson.. C 2010. Bradley.-T.-H. 128. R. H.. Mater.. Angew. Phys. T. Long. J.. T. H. Tenderholt. Park.. W. Chem. F. E... C.. Sekiguchi. Suzuki... Sudyoadsuk.. Z.. K.. J.. (18) Hwang. Ehara. Chem.. Hu. K. E. S. 112. K.. 116. GAUSSIAN 09. 59. Barone. J. A. Zhang. J. S. S. J. 14256−14257.. Li. Mori. Fujisawa. Phys... Liu. J. Org. Mater. J. Y. Lett. Liang. Fischer. Phys. J. (b) Thangthong.-H. C 2010. U. E. 105.. et al. F. A. Wu. W. 48... 054103. 591−594.. Promarak.-M.-W. J. D. J. Nat. MacElroy.. T. (12) (a) Tae. V. (d) Wu. (c) Sánchez-de-Arms. (b) Tang. J.. 2008. Kim.. 2008... A... 1985... (b) Tanaka. 3823−3835. 3. D. (c) De Angelis. M. Saengsuwan. S. K. 130.. K. Chem. K. L. A.. 21. Chem. A. J. 115. J. Phys. (6) (a) Li. Kubo. 211−215... Lee. Chem. A.. A. (17) (a) Jiang.C. L. Nakatsuji. Chem.... K..-T. J. P. B. E. Chem. CRC: London. W. M. Scuseria. K. J. Soc. C. L. Scalmani... Rev. J. T. Phys. 334. 136. 874−880. 35. A. M. revision B.. Phalinyot.. Lee. Humphry-Baker. C 2008. R. Mater. Labat. C. Dykstra.. C.. S. M.. DeFrees.. Wu. 2000. (c) Hariharan. Huang.. H..-C. Wolschann. 2011. Jr. Y. M. 2007. Gordon. Hara. Ciofini. 77. K. F. Keawin. Instrum. W. G. P.. S. P. Photochem. L. 2008. R. G. S. Zakeeruddin. F... N.... Kim. H. J. Photobiol. (c) Kim. Lee. Chem. 2010. Thampi. J. Elsevier: Amsterdam.. S.. 126. C. Burke.. Jiang.. Chem. Duffy. P.. Furube. 3037−3044. Frenking. A. H. M.. T.-C. (14) (a) Ehret. J. I. B. 1173−1174. P. J.-L. R. 114. 4778−4785. Y. G. Chou.. Phys. J.. Chem. M. Kajihara. (34) (a) Hay.. 085403.-W. J. C 2010. Tian. Z. J. J. Sellon. Jiang.. 87−96. (e) Poolmee. 9960−9965. Phys. J. 7122−7124. N. Eds. Bach. L.1021/jp304489t | J. M.-K. Prachumrak. 1913−1922. 4531−4534. Chem.. Adamo. P. J. M. 65. 1830−1840.. S. J. J. J. (b) Lin. B.. Ed. J.. Acc.. San Miguel.. Grätzel. Ehara.. Verhees. K.. 114. (b) Delley.doi. J. H. Ditchfield. J. B. N. T. Takahashi. Chem. M. Hua. S. M. Rev. I. (d) Tan.-G. D. Phys... Lett.. Chem. Chem. (d) De Angelis. Barone.. R. J. 47. T. 130. K. Chem. Mater.. (39) Kamtekar. N. Abe. L. Vol. H. L. Kwon. V. Engels.-H. Plenum: New York. C. M. J. Zhang. Parr. Á . 1993.. 1396−1396. T. C 2011. Phys. K. M. 1979. 5483−5495. Hsu.. MacElroy. 479−485.... H. K. Cheng. 2006. Hsu. Org.. L. Chem. A.. Ko. S. Arakawa. N.. L..-C.. Phys. Palomares..-J. Bettington. J. C. 6272−6276.. 109. 127. 1990.-H. Funct. A. 213−222. D. 77. 46. Scuseria. (f) Suramitr. Würthner. Y.-H. W. Adv. H.. Otsubo. Pietro. (27) (a) Nakatsuji. 22513−22522. R. G. D... P. J. L. E. C. Chem. J. K.. Soc. Z.. J. A.. 17964−17974.. B 1988.. Marinado.. Chen. M.. K.: Wallingford. (11) (a) Jono. R.. Pauporté. Scalmani. Sci. L. 3.. P. J. S. 2007. Palomares. Costa. T. N. M. Pople. Polymer 2005. 924−937. Park. Phys. 2004. Tilocca. 2012. 4. Phys. Hua. Am.. Suga. C. P. (35) Frisch. C. M.. 92. H.. Lett. Theor. E.. Am. Shklover. 448. Chem. Tsao. H. 5.. T. Lee. T. Qin.. R. J. 2759−2767. pp 1099−1141.. Z. 78.. N. Ehara. (c) Daeneke. Spitler.. H. Tang. Pfister. (25) (a) Hagberg.. J. H. J. (e) Nazeeruddin. Kang... Z. C.. Zheng... S.. J. 82. Ichikawa. Commun. H. C. Keawin.. 133.. S. J. 114. P. Spiccia. Zhao. 5473−5481.. J.. Wong.-L. T. H. Wang. Yoo. H. K. 29. Marinado. Ed... 133. 2007. H. Balzani.. 3865−3868. K. Chem.. et al. M. H. Phys.. Liao. P. S. 2001. Tian.. J. Meunmart. Phys. C. Zheng. M. B. L. Phys.-M. (b) Wang... (b) Improta. J. (31) Koide. J. Rabinal. 1977. J. R. K. Ernzerhof. Wang. Commun. (23) (a) Ning. M.-H. M. 21786−21794. 592− 593.. 11101−11108. S. Chandiran. Xue.. M... 2005. CT.. Huang. Segawa.. Lett. Takimiya. N.. 11. 20. Chem. F.. 72. (d) Francl. H. M. G. Frisch.. Chem... Hagfeldt. 15024−1502. (13) Agrawal. 89−96. 1996. K... H.. Chem. Chem. Hagberg. (b) An.. 2004.. A.. 2010. O. 17. Chem. Chuang. Science 2011. Binkley... L. 6474−6481. Raghavachari. Am. N.. M. M.-T. P. Moehl. 2828−2831. J. (32) (a) Ditchfield. (29) (a) Fukuda. Lett. Schaefer. C. H. J. F. 2010. M. M. Y.. Karlsson. M.-Y. Chem. Lin. 17.. Kang. J. Kong. 182.Eur. W. J. 493. Nakatsuji. 11293−11301.. A. (b) Nakatsuji. 75.. S.. Pople. 839−845.. R. J.. Hay. M. T. S. 2011. (b) Yang. and Devices. C.. Yi. 116.. Wadt.. Chem. Ehara. Sun. E. 3280−3286. Ernzerhof. H. Murai. Bryce. Jung25662 dx. S.. Ciofini. Wang. 1972. L. Hsieh. (8) Fink. J. J. A. Chem. Y. C. B. Lin. A. P. D. A. C 2012. Thin Solid Films 2008. J. M. Harima. F. S. Hagfeldt. T. S. Phys. Sci. Org. Boschloo.. Organic Photovoltaics: Mechanisms. 724−729. K. Kroon. Chem. Fukuda. Lai. 1167−1170. 67.. M. 16. Zakeeruddin. W.. J. R. Meredith. K.

Bhattacharaya.. 4693−4700.. Phys. (c) Pal.. Ganguly. 104. Sarker. 2006. 5039−5048. Chem. Grätzel. B 1999. Dyes Pigm... N.-J. (48) Yaneva. Linb. (44) Grätzel... K. K. Sun.. S. Koumanova. 393.. Colloid Interface Sci. Gomez. S. Phys..1021/jp304489t | J. D. N. 20117−20126. M.-P. A. Hehre. Z. Yesudas. J. D. Phys. J.. S.. Handy. J.-Y. L. 293. 10. Chem. J. 181. J. Lin.. 1356−1363. A. Tekoglu. 47.. M. Ko.. Liu. (45) Khazraji. Shi. 116. C. Das. 274−281... Duanb. L. P. M. (51) (a) Srinivas. A. P. Rahman.doi. Langmuir 2010.. Advances in Composite Materials for Medicine and Nanotechnology. J.-D. J. Kang. Photochem.. K.-Q. (49) Erten-Ela. S.-K. Tan. Lett. Zhanga. P. 2008.-Q. E.. 2010. Chem. Hsu. Zhao. C 2011. M. S. Electron. S. (53) (a) Binkley. Y. K. J. J. M. Ed... 861−879.. Baik. X. H.. M. Can. Am. Han. C 2009. H. S. Icli. A.. Chem. M... Fei.. 103. 327− 330. Nath.. K. Soc. 199.-C. J. G... S. K. Chen.. Mulherin. M. 115. 10. 189. Demopoulos.. Am. Binkley. 2797−2803. Binkley. D. Francl. 8. J. Xue. Z. W. Batabyal. S. Jiaa. 25653−25663 . Tew... Langley. X. B. J. M. S. Soc. Chem. Meredith. (47) An. 1980.. 26. Y. A 2007.. H. (50) Wana. (b) Lee. J.. B. M. Y.. Elouatik. 102. M. S. Hehre. W. K.-S.. (46) Lu. 51−57. 83. Hehre. Phys.. Angew.. S. J. 939−947. K. J. C. Chem. 1987.. M.. C 2007. M. Prog. Pople.. P. D. S. Org. M. Hsu. Chem. K. S. J. Am. S. O. (52) Yanai. J. (b) Gordon. Huang. B. J. M. J. Z. Pietro. 2009. J. 749−756. 113... 2004. Chem. W.. J. Chem. Photobiol... W. J. J. Y.. Hotchandani. C 2012.. 426−431. C. 104. Chem. J.. InTech: New York. Nazeeruddin. Marszalek. T. Kamat. S.. S. Liu.. A. Photovoltaics 2006. Phys. 2011. Burn. (d) Dobbs.. W. (43) Shen.. 303−311. J. 25663 dx. Liang. V... 1982. J. Rao... 86−93. R. W. Soc. 429−442. T. Pradhan. T.. C. 187−197. 18785−18793. Curr. Vol. J. T.-S. C. Chem. Bhanuprakash. 14.. Hehre. B. Power Sources 2012. Appl. 111. J. Phys. 9575−9583.-C. P.. 1982. H. Xiang. Phys. 2009. Comput. Pople. Int. Defrees. P. (42) (a) Choi. (b) Tsai... V.The Journal of Physical Chemistry C Article (41) Lee.. J...-Y. (c) Pietro. Giribabu. Huang. Wang.. Kang. Y. J.