Copper(i)-based dye-sensitized solar cells with sterically demanding anchoring ligands: bigger is not always better

The synthesis and characterization of ((6,60-diphenyl-[2,20-bipyridine]-4,40-diyl)bis(4,1-phenylene)) bis(phosphonic acid), 2, are described. Compound 2 has been incorporated as an anchoring ligand in copper(I)-containing dyes in n-type dye-sensitized solar cells (DSCs), combined with 2,20-bipyridine (bpy), 6-methyl-2,20-bipyridine (6-Mebpy), 6,60-dimethyl-2,20-bipyridine (6,60-Me2bpy), 4,40-di(4bromophenyl)-6,60-dimethyl-2,20-bipyridine (3) or 4,40-di(4-bromophenyl)-6,60-diphenyl-2,20-bipyridine (4) as ancillary ligands (Lancillary). Dyes were assembled on mesoporous TiO2 using an on-surface assembly strategy which relies on ligand exchange between surface-anchored Lanchor and [Cu(Lancillary)2] ; H NMR spectroscopy was used to confirm that the bulky phenyl substituents did not hinder ligand exchange. Comparison of values of the open-circuit voltages (VOC), short-circuit current densities (JSC) and external quantum efficiency (EQE) spectra for DSCs with model dyes [Cu(2)(bpy)] , [Cu(2)(6-Mebpy)] and [Cu(2)(6,60-Me2bpy)] + confirm that methyl-substituents in Lancillary are beneficial. Performance data for DSCs with dyes [Cu(1)(3)], [Cu(1)(4)], [Cu(2)(3)] and [Cu(2)(4)] where 1 is the anchor ((6,60-dimethyl-[2,20-bipyridine]-4,40-diyl)bis(4,1-phenylene))bis(phosphonic acid) show that dyes with anchor 2 (phenyl substituents in the 6and 60-positions) give relative conversion efficiencies #10% with respect to standard dye N719 set at h 1⁄4 100%; this compares with relative efficiencies of up to 34.5% for the dyes [Cu(1)(3)] and [Cu(1)(4)]. The performance of [Cu(2)(3)] can be improved by the addition of the co-adsorbant chenodeoxycholic acid. Although the phenyl (versus methyl) substituents lead to enhanced light absorption to lower energies, dyes with anchor 2 quickly bleach when exposed to the I /I3 electrolyte; bleaching also occurs after soaking in solutions of LiI. The dye can be regenerated by treatment of a bleached electrode with Lancillary, or with [Cu(NCMe)4] + followed by Lancillary.


Introduction
In the last few years, there has been signicant progress in the development of n-type dye sensitized solar cells (DSCs) incorporating copper(I)-based dyes 1,2 in place of the ruthenium(II) sensitizers used in conventional Grätzel DSCs. 3 This move away from metals scarce on Earth is mirrored in the development of copper(I) complexes for organic light-emitting devices (OLEDs) 4 and light-emitting electrochemical cells (LECs). 5,6Compared to efficiencies of >12% for ruthenium-or porphyrin-based dyes or metal-free organic dyes, 7 photoconversion efficiencies for copper(I)-based DSCs only recently surpassed 3% (relative to 7.63% for reference dye N719) for masked and sealed cells. 8In 2014, Boujtita, Odobel and coworkers 9 reported a dramatic improvement in efficiency, achieving 4.66% (relative to 7.36% for N719) for the copper(I) dye shown in Scheme 1 in sealed DSCs.Factors that contribute to this impressive efficiency are the use of a 2,2 0bipyridine (bpy) ancillary ligand bearing peripheral triethylamino domains and conjugated spacers, a bpy anchoring ligand with bulky mesityl groups in the 6,6 0 -positions, and the addition of a co-adsorbant (chenodeoxycholic acid, cheno).
Bis(diimine)copper(I) complexes are labile making isolation of heteroleptic complexes in solution difficult; unless steric factors dictate otherwise, statistical mixtures of homo-and heteroleptic species are obtained. 10The HETPHEN approach which relies on bulky substituents 11 is successfully used to prepare heteroleptic copper(I) dyes prior to adsorption on the electrode surface. 9,12In contrast, we have introduced an onsurface strategy in which TiO 2 functionalized with an anchoring ligand, L anchor , acts as a 'surface-as-ligand' to bind a metal ion ('surface-as-complex') followed by an ancillary ligand, L ancillary . 13,14This versatile approach allows facile screening of heteroleptic dyes comprising carboxylic or phosphonic acid anchoring ligands with different ancillary ligands, 8,[13][14][15][16][17][18][19] and we have also used the method to assess the different performances of dyes with phosphonic acid or phosphonate ester anchoring domains. 20To stabilize tetrahedral [Cu(N^N) 2 ] + complexes in which N^N is a bpy or phen (phen ¼ 1,10-phenanthroline) against oxidation to square planar [Cu(N^N) 2 ] 2+ , it is necessary to introduce substituents in the 6,6 0 -or 2,9-positions, respectively, of bpy or phen. 21For L anchor , we have focused on 6,6 0dimethyl-2,2 0 -bipyridines and shown that 1 (Scheme 2) is the most effective anchor of the carboxylic and phosphonic acids screened so far. 13,16The electronic properties of bpy-based L ancillary are readily tuned by variation in the 4,4 0 -and/or 6,6 0substituents.Bromo or iodo-substituents in the 4,4 0 -positions appear particularly benecial, 8,16 and phenyl groups in the 6,6 0positions lead to enhanced absorption at higher wavelengths. 16he latter is an important point, since a major challenge for further development of copper(I) dyes is to broaden the spectral response in the red-end of the visible region. 9,16e now report the preparation and characterization of the phosphonic acid anchor 2 (Scheme 2) with phenyl substituents in the 6,6 0 -positions of the bpy domain, and compare its performance with the dimethyl analogue 1 (Scheme 2) when combined with ancillary ligands 3 or 4. We also assess the need for 6,6 0 -substituents in L ancillary in the [Cu(N^N) 2 ] + dye when L anchor contains sterically demanding phenyl substituents.

Experimental
General 1 H and 13 C NMR spectra were recorded at 295 K on a Bruker Avance III-500 NMR spectrometer with chemical shis referenced to residual solvent peaks with respect to d (TMS) ¼ 0 ppm.Solid-state absorption spectra were recorded on a Cary 5000 spectrophotometer, and FT-IR spectra on a Perkin-Elmer Spectrum two spectrometer equipped with a UATR.Electrospray ionization (ESI) mass spectra were measured using a Bruker Esquire 3000 Plus mass spectrometer.A Biotage Initiator 8 reactor was used for microwave reactions.

Ligands and complexes
Ligands 1, 16   were synthesized as previously reported.

Compound 5
Compound 4 (247 mg, 0.399 mmol), [Pd(PPh 3 ) 4 ] (138 mg, 0.12 mmol) and Cs 2 CO 3 (325 mg, 0.999 mmol) were combined in anhydrous THF (20 mL) in a 10-20 mL microwave vial equipped with a stirrer bar under argon.Diethylphosphite (0.12 g, 0.11 mL, 0.88 mmol) was added by syringe before the vial was sealed and the reaction mixture was heated under microwave irradiation to 110 C for 150 min.The reaction mixture was ltered to yield a yellow solution prior to evaporation of the solvent under reduced pressure.The resulting yellow residue was then dissolved in CH 2 Cl 2 (20 mL) and stirred with decolourising charcoal for 10 min then ltered over Celite prior to removal of solvent under reduced pressure to produce an oily yellow residue which crystallized quickly.Upon addition of acetone (20 mL), a white precipitate formed.The white solid was collected by ltration, washed with Et 2 O (20 mL) and dried under a stream of air, yielding 5 as a white solid (161 mg, 0.220 mmol, 55.1%).M.p. 255 C. 1   48 h.Solvent was removed under reduced pressure leaving an off-white residue.This was added to glacial acetic acid (35 mL)  and concentrated aqueous HCl (1 mL) and the mixture was heated at reux for 24 h.Aer cooling to room temperature and standing for 24 h, a pale yellow powder precipitated and this was collected by ltration.Aer washing with water (10 mL), acetone (10 mL) and Et 2 O (20 mL), 2 was isolated as off-white solid (65.9 mg, 0.108 mmol, 65.5%).M.p. > 350 C. 1  The complex [Cu(bpy) 2 ][PF 6 ] was not isolated.The relevant dyebath for the DSC assembly (see below) was prepared using a 0.1 mM CH 2 Cl 2 solution of [Cu(MeCN) 4 ][PF 6 ] and two equivalents of bpy.

DSC fabrication and measurements
DSCs were made based on the method of Grätzel and coworkers. 24,25Solaronix Test Cell Titania Electrodes were used for the photoanodes.For the counter electrodes, Solaronix Test Cell Platinum Electrodes were used, and volatile organic impurities were removed by heating on a heating plate at 450 C for 30 min.
The dye-covered TiO 2 electrode and Pt counter-electrode were combined using thermoplast hot-melt sealing foil (Solaronix Test Cell Gaskets) by heating while pressing them together.The electrolyte (LiI (0.1 M), I 2 (0.05 M), 1-methylbenzimidazole (0.5 M), 1-butyl-3-methylimidazolinium iodide (0.6 M) in 3-methoxypropionitrile) was introduced into the DSC by vacuum backlling.The hole in the counter electrode was sealed with hot-melt sealing foil (Solaronix Test Cell Sealings) and a cover glass (Solaronix Test Cell Caps).
For each dye, duplicate DSCs were made.In the discussion, the optimal results are presented, but within experimental error, the performances of duplicate cells were similar.Measurements were made by irradiating from behind using a light source LOT Quantum Design LS0811 (100 mW cm À2 ¼ 1 sun).The power of the simulated light was calibrated using a reference Si cell.All DSCs were completely masked 26,27 before measurements were made.

External quantum efficiencies
External quantum efficiency (EQE) measurements were made using a Spe-Quest quantum efficiency instrument from Rera Systems (Netherlands) with a 100 W halogen lamp (QTH) and a lambda 300 grating monochromator (Lot Oriel).The monochromatic light was modulated to 3 Hz using a chopper wheel (ThorLabs).The cell response was amplied with a large dynamic range IV converter (CVI Melles Griot) and measured with a SR830 DSP Lock-In amplier (Stanford Research).

Synthesis and characterization of anchoring ligand 2
The synthetic route to 2 (Scheme 2) followed that reported for the 6,6 0 -dimethyl analogue 1 (ref.16) and is summarized in Scheme 3. The precursor 4 was prepared as previously described 22 and was converted to the phosphonate ester 5 before deprotection using aqueous HCl and glacial acetic acid.
Compound 2 was isolated in 65.5%.The compound is poorly soluble in many organic solvents, and the electrospray mass spectrum was measured using a solution of 2 in MeOH with added CF 3 CO 2 H.The base peak at m/z 621.4 arose from the [M + H] + ion.A higher mass peak at m/z 649.4 was consistent either with the presence of monoethylphosphonate ester or with diesterication of 2 through reaction with MeOH solvent (i.e.[M À 2H + 2Me + H] + , calc.649.2).The fact that no ethyl signals appeared in the 1 H NMR spectrum of 2 was evidence that deprotection was complete and that esterication occurred during mass spectrometric analysis.The 1 H and 13 C NMR spectra (see Experimental section) of a DMSO-d 6 solution of 2 were in accord with the structure shown in Scheme 3; spectra were assigned using COSY, HMQC and HMBC methods.
The Scheme 4 'Surface-as-ligand', 'surface-as-complex' approach to stepwise assembly of heteroleptic copper(I) dyes by on-surface exchange between an anchoring ligand (red) and ancillary ligand (blue) via reaction of the functionalized surface with a homoleptic copper(I) complex.
extended absorption towards the red end of the spectrum.This is highly desirable since enhancement of photon harvesting at lower energies is still greatly lacking in copper(I) dyes 9,18,28 reported to date.
Values of V OC and of J SC are similar (Table 1 and Fig. 3 1) are comparable with our previous measurements, 8,16 conrming the reproducibility of dye performances between different batches of DSCs.J-V curves recorded on the day of fabrication are shown in Fig. 5 and  ) and the spectra of these adsorbed dyes exhibit similar absorption maxima (Fig. 7).These are red-shied with respect to that of [Cu(1)(3)] + , and additional red-    32 These spectroscopic changes auger well for enhanced photon harvesting, but once dye-functionalized photoanodes are exposed to the I À /I 3 À electrolyte in DSCs, bleaching of dyes with anchor 2 occurs (Fig. 8); this mimics the behaviour of the model dyes described above.Dye bleaching was quantied by recording the solid-state absorption spectra of an electrode functionalized with [Cu(2)(3)] + before and aer exposure to I À /I 3 À electrolyte (Fig. 9, solid-blue to red curves).A similar effect was observed when an analogous electrode was treated with a solution of LiI in 3methoxypropionitrile (Fig. 9, hashed-blue to red curves).The results suggest that bleaching is caused by attack at the copper(I) centre by iodide ion.However, an important nding is that the on-surface dye is readily regenerated by soaking the bleached electrode in either an CH 2 Cl 2 solution of the ancillary ligand 3 or in an MeCN solution of [Cu(NCMe) 4 ][PF 6 ] followed by a solution of CH 2 Cl 2 solution of 3 (Fig. 9, red to green curves).
The copper(I) centre in [Cu(2)(4)] + (Scheme 5) is expected to possess a distorted tetrahedral coordination sphere, attened as a consequence of inter-ligand p-stacking interactions as conrmed in the solid-state structures of [Cu(4) 2 ] + (Fig. 10a) 16 and [Cu(dpp) 2 ] + . 32It is noteworthy that the solution absorption spectrum of [Cu(dpp) 2 ] + is independent of solvent (MeCN, MeCN/H 2 O, CH 2 Cl 2 ), 32 implying that coordinating solvents do not enter the copper(I) coordination sphere.Going from dpp to ligands 1-4 involves a change from a rigid phen backbone to a bpy unit that, even when chelated, can undergo signicant torsional deformation. 33We suggest that this is a contributing factor towards the sensitivity of [Cu(2)(4)] + towards iodide ion.Finally, while both [Cu(1)(4)] + and [Cu(2)(3)] + contain identical local copper(I) coordination environments (Fig. 10b), only DSCs containing [Cu(2)(3)] + bleach in the presence of I À .This suggests that 6-and 6 0 -phenyl substituents in the ancillary ligand help to protect the copper(I) centre; in contrast, placing them in the anchoring domain renders the dye susceptible to attack by I À .We are currently extending these studies to other dyes containing anchor 2.   À electrolyte and sealing.The coloured borders correspond to the colours used in Fig. 7.

Effect of a co-adsorbant on DSCs with [Cu(2)(3)]
The addition of a co-adsorbant such as cheno can signicantly enhance the DSC performance of bis(diimine)copper(I) dyes, particularly those with sterically demanding ligands. 9,35We therefore investigated whether the performance of [Cu(2)(3)] + (Scheme 5) could be improved by addition of cheno.The coadsorbant was added in the second step of the on-surface dye assembly (Scheme 4) because we have previously shown that competitive binding of cheno and anchoring ligand may occur if cheno is introduced in the rst soaking cycle. 35The second dye-bath comprised a 1 : 1 mixture of cheno and [Cu(3) 2 ][PF 6 ].Performance parameters for the DSCs with and without coadsorbant are compared in Table 3.The addition of cheno leads to a signicant gain in both J SC and V OC (Fig. 11) and an overall increase in DSC conversion efficiency (0.59 to 0.97%).However, the efficiency remains less than half that of the dye [Cu(1)(3)] + (Scheme 5).The increase in J SC is conrmed in the EQE spectra (Fig. 12) which show that the presence of cheno results in an increase in EQE max from 12.2% (l ¼ 500 nm) and 8.4% (l ¼ 570 nm), to 19.3% (l ¼ 500 nm) and 15.6% (l ¼ 570 nm).However, one day-old DSCs had bleached with concomitant deterioration in performance (Fig. 11 and 12), in contrast to the stability over a week or more of other bis(diimine)copper(I)-containing DSCs. 8,16,18,35g. 9 Solid-state absorption spectra (transmission mode) of electrodes functionalized with dye [Cu(2)(3)] + (duplicate electrodes, blue), after treatment with I À /I 3 À electrolyte or LiI solutions (red), and after dye regeneration (green).

Conclusions
The effects of replacing the 6,6 0 -dimethyl substituents in phosphonic acid 1 by phenyl substituents have been investigated; although we have shown

Compound 2 Compound 5 ( 9 Scheme 2
Scheme 1 State-of-the-art copper(I) dye in DSCs reported by Boujtita, Odobel and coworkers.9 They were washed with EtOH and sintered at 450 C for 30 min, then cooled to z80 C and soaked in a 1.0 mM DMSO solution of 1 or 2 for 24 h at a constant temperature of 28 C.Each electrode was removed from the solution, washed with DMSO and CH 2 Cl 2 and dried at z60 C (heat gun).Each functionalized electrode was then soaked for 3 days in a 0.1 mM CH 2 Cl 2 solution of [Cu(bpy) 2 ][PF 6 ] (see above), [Cu(6-Mebpy) 2 ][PF 6 ], [Cu(6,6 0 -Me 2 bpy) 2 ][PF 6 ], [Cu(3) 2 ][PF 6 ] or [Cu(4) 2 ][PF 6 ] at a constant dye-bath temperature of 28 C. The electrodes were then removed from the dye-bath and washed with CH 2 Cl 2 .For experiments with cheno, the electrodes with adsorbed 1 or 2 were soaked for 3 days in a dye-bath (at a constant 28 C) consisting of a 1 : 1 mixture of 0.1 mM CH 2 Cl 2 solutions of cheno and [Cu(3) 2 ][PF 6 ] or [Cu(4) 2 ][PF 6 ].N719 reference electrodes were prepared by soaking Solaronix Test Cell Titania Electrodes in a 0.3 mM EtOH solution of dye N719 (Solaronix) for 3 days.The electrodes were taken out of the dye-bath, washed with EtOH and dried at z60 C (heat gun).

Solaronix
Test Cell Titania Electrodes Transparent were washed with EtOH and sintered at 450 C for 30 min, then cooled to z80 C and soaked in a 1 mM DMSO solution of 2 for 24 h.Each electrode was removed from the solution, washed with DMSO and CH 2 Cl 2 and dried at z60 C (heat gun).Each functionalized electrode was then soaked for 3 days in a 0.1 mM CH 2 Cl 2 solution of [Cu(3) 2 ][PF 6 ].The electrodes were removed from the dye-bath, washed with CH 2 Cl 2 and dried at z60 C (heat gun).Solid-state absorption spectra of the functionalized electrodes were recorded using a Cary-5000 spectrophotometer, before and aer being dipped into I À /I 3 À electrolyte (composition as above) or a 0.1 M solution of LiI in 3-methoxypropionitrile for 15 min; aer soaking, the electrode was washed with 3-methoxypropionitrile and EtOH, dried at z80 C (heat gun).Adsorbed dyes were regenerated as follows.One bleached electrode was soaked in a CH 2 Cl 2 solution of 3 (0.1 mM) for 15 min, was then removed, washed with CH 2 Cl 2 and dried with a heat gun at 80 C. The other bleached electrode was dipped into a MeCN solution of [Cu(NCMe) 4 ][PF 6 ] (2.0 mM) for 15 min, was then removed, washed with MeCN and dried with a heat gun at 80 C; the electrode was then dipped in a CH 2 Cl 2 solution of 3 (0.1 mM) for 15 min, before being washed with CH 2 Cl 2 and dried as above.

Fig. 11 J
Fig. 11 J-V curves for DSCs with the dye [Cu(2)(3)] + (Scheme 5) with and without cheno, and measured on the day of sealing the DSC and one day later.

Fig. 12
Fig. 12 EQE spectra for DSCs with the dye [Cu(2)(3)] + with and without cheno, and measured on the day of sealing the DSC and one day later.

3 À
+ can be improved by the addition of the co-adsorbant cheno.All dyes with anchor 2 bleach in the presence of I À /I electrolyte and bleaching also occurs when the dyes are exposed to iodide ion.The cause of bleaching appears to be displacement of L ancillary or stripping of copper from the surface-bound complex; anchored dyes are readily regenerated by soaking a bleached electrode in solutions of L ancillary , or in solutions of [Cu(NCMe) 4 ][PF 6 ] followed by L ancillary .We are now exploring the generalities of these observations with anchoring ligand 2, and regeneration of surface-adsorbed dyes aer dyes have bleached.The bleaching effects justify more vigorous investigations of alternative electrolytes such as the Co 2+ /Co 3+ based systems to replace I À /I 3 À in copper(I)-based DSCs.

17
Further development of copper(I) dyes must continue to address improved spectral response towards the red-end of the visible spectrum, and nd ways of increasing the short-circuit current density which is currently signicantly lower than for state-ofthe-art ruthenium(II) dyes such as N719.

Table 3
Performance data for masked DSCs with dye [Cu(2)(3)] + (Scheme 5) with and without co-adsorbant cheno; data are compared to a DSC with dye N719.Data for DSCs without cheno are taken from