An investigation of the roles furan versus thiophene π-bridges play in donor – π-acceptor porphyrin based DSSCs †

Dye-sensitized solar cells (DSSCs) continue to attract interest due to their lower cost production compared to silicon based solar cells and their improving power conversion efficiencies. Porphyrin-based sensitizers have become an important sub-class due to their strong absorption characteristics in the visible region, convenient modulation of properties through synthetic manipulation and class-leading power conversion efficiencies. In this article, we report the synthesis and characterization of two porphyrinbased dyes and their application as sensitizers in DSSCs. A thiophene and a furan moiety have been incorporated into the push–pull architecture as a π-bridge, allowing the systematic investigation of how these moieties influence the physical properties of the dyes and the performance of their resulting DSSCs. A significant difference in PCEs has been observed, with the furan containing dye (PorF, PCE = 4.5%) being more efficient than the thiophene-based analogue (PorT, PCE = 3.6%) in conjunction with the iodide/triiodide redox electrolyte.


Introduction
Photovoltaics (PV) offer an attractive renewable energy technology and there is a continued interest in this field due to the potentially inexhaustible energy source provided by sun. 1 During the last twenty years, interest in dye-sensitised solar cells (DSSCs) as promising PV technology has increased, due to their ease of fabrication, relatively low cost and interesting design possibilities in terms of tunable transparency and colour.Since the development of the first working device in 1991 by Grätzel and O'Reagan, 2 there has been a huge effort directed towards increasing the power conversion efficiency (PCE) of devices by optimising their components, with a particular emphasis being placed on the sensitiser, as it determines the light harvesting and absorption properties of the device.Ruthenium-based dyes, such as N719, have monopo-lised the early progress in this field, however, metal-free organic dyes have become increasingly attractive due to their lower cost, together with higher molar extinction coefficients, easier synthesis, purification and structural modification. 35][6] In particular, the position of these bands, together with the redox properties, 7,8 can be tuned through the insertion of a metal in the central cavity 9 and through the substitution of the porphyrin periphery, which can be done by functionalising the β-positions and the meso-positions.While zinc is the most preferred metal due to its low-cost and ability to broaden the optical absorption of porphyrins, 10 the types of substituent that could be added to the peripheral positions to tune the properties of porphyrin derivatives is enormous.
Since the first report of a porphyrin-based sensitiser providing a PCE of 2.6%, 11 significant effort has been directed towards understanding the role substituents play on modulating the device characteristics of porphyrin based DSSCs.Functionality attached to both βand meso-positions have been investigated, with the latter typically providing the best performing dyes. 12However, the performance of this class of dyes is generally lower than that of the most efficient Ru-based dyes, due to their poor injection yields. 13,14To circumvent this issue, the development of meso-substituted porphyrin dyes featuring a D-π-A architecture represents a step forward regarding the performance of these molecules in DSSCs. 15,16In particular, substituting the porphyrin core with electron donating moieties and electron withdrawing units attached at opposite sides of the porphyrin ring, results in the generation of an intrinsic dipole moment which favours the electron injection of the dye into the conduction band of TiO 2 .Finally, the introduction of alkylated aromatic units [17][18][19] at the two side mesopositions has led to a further improvement of performance, due to the minimisation of charge recombination arising from the dye aggregation. 20,21Thanks to these developments, the PCE of porphyrin-sensitised devices has grown rapidly and has provided the new record PCE of 13%. 22Nevertheless, despite the high efficiencies reported for some porphyrin dyes, the synthesis of unsymmetrical derivatives is problematic and generally results in low yielding multi-step procedures leading to the formation of several side-products. 23,24n this article, we report the synthesis of two porphyrinbased dyes (Fig. 1) and their employment as sensitisers in DSSCs.Both the molecules incorporate three triphenylamine (TPA) moieties as donor units and differ for the presence of either a thiophene or a furan ring in the π-bridge (A 3 B-type). 12he three bulky TPA units have been incorporated into the dyes in view of their combined electron-donating character and ability to negate dye aggregation on the TiO 2 surface. 12hiophene and furan residues have been incorporated into the π-bridge, allowing the role these heterocycles have in controlling device performance of this class of dyes.Thiophene and its derivatives have been extensively studied as π-spacers in sensitisers for DSSCs, due to their smaller resonance energy (29 kcal mol −1 for thiophene, 36 kcal mol −1 for benzene) which enhances the conjugation by lowering the energy required for charge transfer (CT). 25,26The employment of furan as π-spacer should further improve the dye performance, due to its lower aromatic energy (16 kcal mol −1 ).It has also been verified that furan shows comparable optical properties and charge carrier mobility with thiophene 27 and better solubility when incorporated in polyconjugated systems. 28oreover, the higher oxidation potential of furan should improve the hole location, hence the stability of the dye, 29 and its PCE when compared to its thiophene analogue. 27,30,31[34][35]

Synthesis
Porphyrin based dyes PorT and PorF were synthesized according to Scheme 1.The synthesis of the central meso-substituted porphyrins was performed according to Lindsey's method 36,37  (Scheme 1).4-(Diphenylamino) benzaldehyde 1 and 4-bromobenzaldehyde 2 in 3.5/1 molar ratio were reacted with an excess of pyrrole to afford compound 3 in modest yield.This was then quantitatively converted to the Zn-based porphyrin 4, by reaction with zinc acetate.A Suzuki cross-coupling reaction between 4 and the boronic acids 5a and 5b, was carried out affording aldehydes 6a and 6b, respectively.These were then reacted with cyanoacetic acid through a Knoevenagel condensation to obtain the final dyes PorT and PorF.

Optical and electrochemical properties
The UV-Vis absorption spectra of the two dyes recorded in solution (Fig. 2) show the typical behaviour of metalloporphyrins, consisting of an intense and sharp B-band and two weaker Q-bands. 38The former, at 438 nm for PorT and 426 nm for PorF, have an intensity of the order of 10 5 M −1 cm −1 , whilst the latter, at 565 nm and 609 nm for PorT and 553 nm and 597 nm for PorF, are one order of magnitude weaker.The absorption maxima of PorT and PorF are at 650 nm and 625 nm, respectively, corresponding to optical bandgaps (E G, OPT ) of 1.91 eV and 1.98 eV.
The solution-based electrochemical properties of both dyes were explored by cyclic voltammetry (CV, Fig. 3) and square wave voltammetry (SWV, ESI Fig. S1 †).Both dyes undergo pseudo-reversible one-electron reduction and oxidation.SWV allowed us to estimate the ionization potential (I P ), electron affinity (E A ) and fundamental band gap (E G,F ), respectively (Table 1).Both E red and E ox of PorF are more positive than those of PorT, of about 100 mV (−1.72 and 0.31 V for PorT and −1.62 and 0.40 V for PorF), resulting in more negative energy values for I P and E A .

Theoretical calculations
To gain insight in electronic properties of the two dyes, density functional theory (DFT) calculations were performed using Gaussian 09. 39The optimized geometry shows a similar conformation for both the molecules, with the dihedral angles between the porphyrin and the appended benzene moieties of about 60°.Although this limits the conjugation throughout the molecule, a twisted structure could effectively decrease the dye aggregation by obstructing the formation of π-stacks among the porphyrin units. 40Corresponding molecular orbitals for PorT and PorF are shown in Fig. 4. The HOMO is delocalised throughout the TPA donors and the porphyrin ring, while the LUMO is spread over the acceptor arm.This good separation is normally not found in porphyrin dyes having the π-bridge co-planar with the porphyrin core, 41,42    indicating that the absence of the intramolecular CT band is likely due to the twisted geometry of the molecule.Fig. 4 also shows that the spatial distribution of the LUMO+1 is confined on the porphyrin core, suggesting that an optical transition between the HOMO and the LUMO+1 would be more favourable.The estimated energy of the transition is 2.54 eV and 2.64 eV, for PorT and PorF, respectively, corresponding to a wavelength of 488 nm and 469 nm.The values are in good approximation with the experimental offset of the Soret bands (S 0 → S 2 ) of both the dyes.This suggests that the π-bridge does not have a significant role in the light absorption, which presumably occurs due to the presence of the porphyrin core.However, the π-bridge should play a significant role in the extraction of electrons upon dye excitation.
A summary of the properties of the two dyes is provided in Table 1.It is worth observing that the calculated properties are in good agreement with both the experimentally estimated ones and those calculated in previous work for similar molecules. 43

Device performance
The above dyes were tested as sensitisers in DSSCs by applying cobalt and iodide/triiodide redox electrolytes on 4 µm + 4 µm double layer TiO 2 films.The photovoltaic characteristics of these devices, A-D, measured at AM 1.5G irradiance (100 mW cm −2 ), are tabulated in Table 2.For dye PorF, Device A, the highest PCE obtained is 4.1% with a V OC of 0.700 V, a J SC of 7.7 mA cm −2 and a FF of 75%.In turn, device B exhibits V OC of 0.660 V, J SC of 8.9 mA cm −2 , and PCE of 4.5%.Furthermore, employing the same cobalt and iodide/triiodide based redox electrolytes with PorT dye, labelled as devices C and D, respectively, the photovoltaic characteristics are tabulated in Table 2.
The data show that having thiophene (PorT) closer to the surface of TiO 2 enhances the recombination compared to furan (PorF) dye, which is reflected in the lower V OC of PorT dye with both electrolytes.To the best of our knowledge, these results are in contrast with related studies on porphyrin-based dyes.However, these results are in good agreement with several studies on metal-free push-pull dyes 27,31,45 which showed an increase in V OC when the device is sensitised with a furan-based dye compared to its thiophene analogue.As suggested by the authors, this difference may be due to the enhanced suppression of the electron recombination between TiO 2 and electrolyte.
The incident photon-to-current conversion efficiency (IPCE) spectrum of the devices is shown in Fig. 5a (inset).The IPCE spectra of both the devices A and B have peak maxima at 450 and 620 nm and the tail goes up to 700 nm.The IPCE spectrum reflects closely the absorption spectra of dyes.The integrated current of IPCE spectra of devices is close to the measured current.The difference in the HOMO level of these dyes with the cobalt electrolyte redox energy level is large enough to provide a driving force for the dye regeneration.Transient photovoltage characterisation under open circuit voltage was performed to understand the differences in the photovoltaic parameters of devices with two different electrolytes.The plot of charge vs. potential for the devices A and B shows similar behaviour but at the same charge density the V OC is higher for cobalt-based electrolyte (Fig. 5b) which is consistent with the difference in the energy level of these two electrolytes.

Conclusions
In conclusion, two porphyrin dyes containing furan and thiophene as π-conjugated spacers have been synthesized.The systematic modification of the dyes has allowed us to investigate the role this has on the optical and redox properties of the dyes together with its influence on DSSC device performance.Modification of this type leads to a significant change in the dyes solution optical and redox properties, with PorT having a lower estimated optical band gap.PorF gave rise to the best performing DSSC devices in terms of power conversion efficiencies, which was independent of the electrolyte used.This work is in contrast with related studies that have indicated that thiophene-based bridges tend to give rise to superior power conversion efficiencies, and indicates that the role the structure of the π-conjugated spacer has in determining device characteristics of porphyrin based dyes is dependent upon the parent structure of the porphyrin.

General information
Chemicals were purchased from Sigma-Aldrich, TCI and Alfa Aesar, and were used without further purification.All reactions were run under an argon atmosphere.Solvents were purified using a PureSolv solvent purifier system.NMR spectra were performed with either a Bruker AVIII 400 MHz or a Bruker AVIII 500 MHz spectrometer and all reported chemical shifts are relative to TMS.Mass spectra were performed by the National Mass Spectroscopy Facility (NMSF), Swansea University (UK).UV-vis spectra were recorded on a PerkinElmer Lambda 25 instrument.Optically determined band gaps (E OPT ) were estimated using the absorption edge of the longest wavelength absorption (λ) using E G,OPT (eV) = (1240/λ(nm)).Cyclic voltammetry measurements were undertaken using a CH Instruments 440A electrochemical analyser using a platinum working electrode, a platinum wire counter electrode and a silver wire pseudo-reference electrode.
Ferrocene was used as an external standard and all redox couples are reported versus the ferrocene/ferrocenium (Fc/Fc + ) redox couple, adjusted to 0.0 V.The solutions were prepared using dry DMF containing electrochemical grade tetrabutylammonium hexafluorophosphate (0.1 M) as the supporting electrolyte.The solutions were purged with nitrogen gas for 3 min prior to recording the electrochemical data.

Theoretical calculations
Density functional theory (DFT) calculations were performed using Gaussian 09, revision D.01. 39Global minimum states were confirmed by absence from imaginary frequencies under ground-state geometry optimization followed by vibrational frequency calculations.All calculations were conducted with Becke's three-parameter hybrid and For photovoltaic measurements of the DSSCs, a solar simulator equipped with a 450 W xenon light source (Osram XBO 450) with a filter (Schott 113) was employed, whose power was regulated to the AM 1.5 solar standard by using a reference Si photodiode equipped with a colour-matched filter (KG-3, Schott) to reduce the mismatch in the region of 350-750 nm between the simulated light and AM 1.5 to less than 4%.The measurement-settling time between applying a voltage and measuring a current density for the J-V characterization of DSSCs was fixed to 80 ms with a Keithley model 2400 digital source meter.The photocurrent action spectra were measured with an Incident Photon-to-Current Conversion Efficiency (IPCE) test system.The modulation frequency used was about 2 Hz and light from a 300 W xenon lamp (ILC Technology, USA) was focused through a computer controlled Gemini-180 double monochromator (John Yvon Ltd, UK).A white light bias was used to bring the total light intensity on the device under testing closer to operating conditions.A light mask was used on the DSSCs, so the illuminated active area of DSCs was fixed to 0.159 cm 2 .Synthesis 5-(4-Bromophenyl)-10,15,20-tris(N,N-diphenylaniline)porphyrin 3. Pyrrole (0.750 mL, 10.8 mmol) and 4-bromobenzaldehyde 2 (0.500 g, 2.70 mmol) were added to a solution of 4-(diphenylamino)benzaldehyde 1 (2.50 g, 9.15 mmol) in degassed DCM (150 mL).BF 3 •OEt 2 (0.300 mL, 2.43 mmol) was added, and the mixture was stirred at r.t. for 4 h.Afterwards, DDQ (1.75 g, 7.71 mmol) was added and the resulting mixture was stirred for further 1 h.Et 3 N (5 mL, 36.07 mmol) was then added and the mixture was stirred for further 30 min, filtered through a pad of silica and concentrated under reduced pressure.The crude compound was purified by column chrom- 5-(4-(5′-Formyl)thienyl)phenyl-10,15,20-tris(N,N-diphenylaniline) porphyrinato zinc(II) 6a.An aqueous solution of 2 M K 2 CO 3 (0.200 mL, 0.400 mmol) was added to a mixture of 4 (0.100 g, 79.4 µmol) and 5-formyl-2-thiopheneboronic acid 5a (18.6 mg, 0.119 mmol) in DME (10 mL) and the resulting mixture was degassed with N 2 for 30 min.Pd(dppf )Cl 2 (6.2 mg, 3.97 µmol) was added and the solution was stirred at 80 °C for 48 h.The mixture was then allowed to cool to r.t., poured into water (30 mL) and extracted with DCM (3 × 30 mL).The combined organic extracts were dried over MgSO 4 , filtered and then concentrated under reduced pressure.The crude compound was purified by column chromatography (SiO 2-Cyano-3-(5-(4-(10,15,20-tris(N,N-diphenylaniline))porphyrinato zinc(II)-yl)phenyl)thienyl acrylic acid (PorT).Compound 6a (0.100 g, 77.5 µmol) and cyanoacetic acid (29.7 mg, 0.349 mmol) were dissolved in a mixture of THF and acetic acid (12 mL, 1 : 1, v : v).Ammonium acetate (29.9 mg, 0.388 mmol) and zinc acetate dihydrate (68.0 mg, 0.310 mmol) were added and the resulting mixture was stirred for 4 h at 70 °C under N 2 .The mixture was allowed to cool to r.t. and water (50 mL) was added.After stirring for further 10 min at r.t., the precipitate was filtered off, washed with water (50 mL) and dried under vacuum.The crude compound was purified by column chromatography (SiO 2 , DCM : MeOH, 9 : 1) affording PorT (0.100 g, 95%) as a dark green solid.)porphyrinato zinc(II)-yl)phenyl)furyl acrylic acid (PorF).Compound 6b (0.100 g, 78.5 µmol) and cyanoacetic acid (30.0 mg, 0.353 mmol) were dissolved in a mixture of THF and acetic acid (12 mL, 1 : 1, v : v).Ammonium acetate (30.3 mg, 0.392 mmol) and zinc acetate dihydrate (68.9 mg, 0.314 mmol) were added and the resulting mixture was stirred for 4 h at 70 °C under N 2 .The mixture was allowed to cool to r.t. and water (50 mL) was added.After stirring for further 10 min at r.t., the precipitate was filtered off, washed with water (50 mL) and dried under vacuum.The crude compound was purified by column chromatography (SiO 2 , DCM : MeOH, 9 : 1) affording PorF (97.9 mg, 93%) as a dark green solid.m.p. 230-233 °C; δ H (400 MHz,

Fig. 1
Fig. 1 Chemical structure of dyes PorT and PorF.

Fig. 4
Fig. 4 Frontier molecular orbital representation of PorT and PorF calculated by DFT.

Fig. 5
Fig. 5 (a) Photocurrent density vs. voltage ( J-V) curves of devices A and B with PorT under standard AM 1.5G illumination.The inset is the IPCE of devices.(b) Charge vs. V OC curves.

a
Calculated using the equation: E G,OPT = 1240/λ onset .b Estimated using the equations: I P (eV) = −4.8− E ox and E A (eV) = −4.8− E red . 44c Values obtained by DFT calculation.d Calculated using the formula: E G,F = E A − I P .

Table 2
Photovoltaic parameters of devices A-D measured under standard AM 1.5G illuminations at 100 mW cm −2

Table 1
Summary of dyes PorT and PorF physical parameters State-of-the-art double layer mesoporous TiO 2 layer (4 μm thickness of 20 nm particle (DSL 18NR-T, DYESOL) plus 4 μm thickness of 400 nm light scattering particles (HPW-400NRD, CCIC)) were deposited on FTO conducting glass (Solar-4 mm, Nippon Sheet Glass Co, Ltd).The double layer TiO 2 film was sensitized by immersing it into a solution of the PorF and tert-butylpyridine in acetonitrile.Devices A and B were fabricated by using PorF with cobalt electrolyte and iodide/ triiodide redox electrolyte, respectively.Devices C and D were fabricated by using PorT with cobalt electrolyte and iodide/ triiodide redox electrolyte, respectively.An antireflection film (ARCTOP, Mihama Co.,) was attached on the photoanode side.