Dye-sensitized solar cells with hole-stabilizing surfaces: “ inorganic ” versus “ organic ” strategies †

Two 2,2 0 :6 0 ,2 00 -terpyridine ligands ( 9 and 10 ) incorporating second-generation diphenylamino-dendrons have been synthesized and characterized; one ligand contains chromophoric benzothiadiazole domains. Using the ‘ surface-as-ligand, surface-as-complex ’ strategy, zinc( II )-containing sensitizers [Zn(L anchor )(L ancillary )] 2+ with carboxylic or phosphonic acid anchors ( 1 and 2 , respectively) have been assembled and tested in n-type DSCs. The solid-state absorption spectra of dye-functionalized electrodes show a broad spectral response for all the dyes with enhanced intensity for those containing the benzothiadiazole units. However, the [Zn(L anchor )(L ancillary )] 2+ dyes perform poorly, exhibiting very low values of the short-circuit current density ( J SC ) and open-circuit voltage ( V OC ). The external quantum e ﬃ ciency (EQE) spectra con ﬁ rm that electron injection occurs, but EQE max is # 3%. Non-optimal positioning of the thiadiazole domain in the dye probably contributes to the poor performances. Screening of DSCs containing FTO/TiO 2 photoanodes without adsorbed dye shows that they generate small short-circuit current densities and open-circuit voltages which contribute signi ﬁ cantly to parameters reported for badly performing dyes. An organic dye 11 , structurally similar to 10 and containing a 2-cyanoacrylic acid anchor, is also reported. This exhibits a broad and intense spectral response between 300 and 600 nm, and shows e ﬃ cient electron injection over a broad wavelength range. DSCs containing 11 are stable over a 17 day period and show global e ﬃ ciencies of 3.93 – 4.57% ( ca. 70% with respect to N719 set at 100%). Ground state DFT calculations reveal that the HOMO in each of [Zn( 1 )( 9 )] 2+ , [Zn( 2 )( 9 )] 2+ , [Zn( 1 )( 10 )] 2+ , [Zn( 2 )( 10 )] 2+ and 11 is localized on the peripheral diphenylamino units, allowing for hole-transfer to the reduced electrolyte. In 11 , a major contribution from the 2-cyanoacrylic acid anchoring group appears in the LUMO manifold; however, while the LUMO in each zinc( II ) dye is localized on anchoring ligand 1 or 2 , it is concentrated close to the metal centre which may contribute to poor electron injection.


Introduction
Dye-sensitized solar cells (DSCs) were developed around twenty years ago and, even at an early stage, exhibited relatively high efficiencies. 1Improvements in DSC design have increased the efficiency of the conversion of photons to electrical current. 2,3onversion efficiency depends on optimizing the photoresponse of the dye, the electron injection, and the rate of oxidized dye regeneration.5][6] In a Grätzel n-type DSC, the LUMO of the dye must be higher than the level of the TiO 2 conduction band 7 (À4.0eV), while the HOMO of the dye should be slightly lower than the HOMO of the redox couple (À4.9 eV for the archetypal I À /I 3 À redox shuttle). 8Design of transition metal-containing sensitizers and molecular organic dyes must take account of all the above criteria. 3,9e and others are developing sensitizers for DSCs which incorporate low cost Earth-abundant metals to replace rare elements such as ruthenium. 10Copper(I) sensitizers [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26] feature strongly, but we have also reported the successful use of {Zn(tpy) 2 } 2+ -containing dyes (tpy ¼ 2,2 0 :6 0 ,2 00 -terpyridine), albeit with low photon-to-current conversion efficiencies. 27,28The lability of both copper(I) and zinc(II) complexes permits sequential assembly of photoactive dyes on a semiconductor surface using the 'surface-as-ligand, surface-as-complex' methodology; 17 ligand exchange is rapid for copper(I) but slower for zinc(II).Fig. 1 summarizes the strategy for {Zn(tpy) 2 } 2+ sensitizers, involving initial treatment of an FTO/TiO 2 surface with an anchoring tpy ligand (e.g. 1 and 2, Scheme 1), followed by treatment of the surface-as-ligand with a zinc(II) salt, and nally reaction of the surface-as-complex with a chromophorefunctionalized tpy ancillary ligand such as 3-7 (Scheme 1). 27lthough ligand exchange in [Zn(tpy) 2 ] 2+ complexes is slow on the NMR timescale, 27 once assembled, [Zn(tpy) 2 ] 2+ domains anchored to the n-type semiconductor are stable.This stepwise procedure contrasts with the preparation of heteroleptic dyes by the HETPHEN 29 approach, 19,21 or by the use of ligand exchange reactions involving homoleptic metal complexes. 10Advantages of the in situ stepwise strategy are a reduction in the number of synthetic steps in dye preparation and a consequential atom and nancial economy 16,17 for the photoactive materials.
Unlike copper(I)-containing dyes which exhibit metal-toligand charge transfer (MLCT) bands in the visible region, zinc(II) complexes characteristically possess absorption spectra dominated by intra-ligand transitions.In sensitizers based on {Zn(tpy) 2 } complexes, the zinc(II) ion acts as 'glue' between the anchoring and ancillary ligand domains.The advantage of the assembly principle shown in Fig. 1 is that it facilitates screening of a wide range of ligand combinations.In order that the photoresponse of the dye incorporates the visible region, it is essential that the ancillary ligand contains a chromophore.Consequently, the formation of the surface-bound {Zn(tpy) 2 } unit is readily monitored by an optical change. 27,28n contrast to the highly promising performance of porphyrinato zinc(II) dyes which can achieve power conversion efficiencies of up to 13%, 30,31 DSCs containing {Zn(tpy) 2 } complexes with anchoring ligands 1 and 2, and ancillary ligands 3-7 27,28 exhibit very low efficiencies, in part due to inadequate light-harvesting in the visible.A common strategy for enhancing light absorption is by extending the p-conjugation 32 or by combining electron-donating and electron-accepting moieties in the same conjugated ligand framework (so called 'push-pull' dyes). 33Here, we report the development of {Zn(tpy) 2 } sensitizers by moving from rst generation ligands 3-7 to second generation analogues 8 and 9, and to the related ligand 10 which contains a benzothiadiazole spacer (Scheme 2).[34][35][36] Experimental General A Bruker Avance III-500 NMR spectrometer was used to record 1 H and 13 C NMR spectra, and chemical shis were referenced to  Ground state density functional theory (DFT) calculations were performed using Spartan 14 (v.1.1.8) 38at the B3LYP level with a 6-31G* basis set in vacuum.Initial structure minimization was carried out at a molecular mechanics or PM3 level.

DSC fabrication
DSCs were made using a similar method to that reported. 28Full details are given in the ESI.†

Ancillary ligand synthesis and characterization
We have recently reported the benecial effects of incorporating rstor second-generation hole transporting dendrons into the periphery of 2,2 0 -bipyridine-based copper(I) sensitizers in DSCs, 13 and the use of a co-adsorbant to alleviate steric crowding of these dyes on the semiconductor surface. 16In this and related work, 12 we have found Hartwig-Buchwald aminations to be a reliable method of coupling diarylamino building blocks to 4bromophenyl-functionalized 2,2 0 -bipyridine or 1,10-phenanthroline domains.We have applied an analogous strategy to prepare rst generation ligands 3-7 (Scheme 2). 27,28However, aer several attempts to prepare the second generation ligand 8 (Scheme 2) using a Hartwig-Buchwald amination of 4,4 0 -bis(N,N-diphenylamino)diphenylamine with 4 0 -(4-bromophenyl)-2,2 0 :6 0 ,2 00 -terpyridine, we were unable to isolate a pure product.We therefore turned our attention to the synthesis of the methoxy-decorated ligand 9.
Hartwig-Buchwald amination of 4,4 0 -bis(N,N-bis(4-methoxyphenyl)amino)diphenylamine 11 with 4 0 -(4-bromophenyl)-2,2 0 :6 0 ,2 00terpyridine (Scheme 3) ESI † yielded 9 in 43% yield aer workup.The highest mass peak in the electrospray mass spectrum of 9 (m/z ¼ 930.9) was assigned to the [M + H] + ion, and high resolution ESI MS also conrmed the molecular mass.The 1 H and 13 C NMR spectra of a THF-d 8 solution of 9 were consistent with the C 2symmetric structure shown in Scheme 2, and were assigned by COSY, NOESY, HMQC and HMBC methods.The aromatic region of the 1 H NMR spectrum is shown in Fig. 2; the methoxy protons give rise to a singlet at d 3.75 ppm.NOESY cross-peaks between the pairs of protons H B3 /H C2 , H C3 /H D2 and H D3 /H E2 (see Scheme 3 for labels) allowed protons in the phenyl rings to be unambiguously assigned.The assignments were conrmed using the HMBC spectrum starting with the high-frequency 13 C NMR signal for C E4 (d 157.1 ppm) which showed a strong correlation to H E2 , and similarly for cross-peaks C D4 /H D2 and C D1 /H D3 .
The synthetic route to compound 10 is summarized in Scheme 4. The secondary amine 10a required for a Hartwig-Buchwald amination with 4 0 -(4-bromophenyl)-2,2 0 :6 0 ,2 00 -terpyridine was prepared starting from the Boc-protected bis(4bromophenyl)amine.Substitution of the bromo-groups for boronic acid pinacolate esters, followed by a Suzuki coupling [39][40][41] with two equivalents of 7-(4-bromophenyl)-N,N-diphenylbenzo[c] [1,2,5]thiadiazol-4-amine resulted in the formation of the Bocprotected precursor to 10a.Attempts to carry out the deprotection using excess triuoroacetic acid failed, but heating the Boc-protected precursor 42 at 210 C in the absence of solvent gave 10a which was used in the Hartwig-Buchwald step without further purication.Compound 10 was isolated as a red solid in 29% yield.The base peak (m/z 943.1) in the MALDI-TOF mass spectrum was assigned to the [M + H] + ion.The 1 H and 13 C NMR spectra were assigned using the COSY, NOESY, HMQC and HMBC spectra, with NOESY cross-peaks between pairs of protons H B3 /H C2 and H E5 /H D3 allowing unequivocal assignments of the protons in arene rings C, D and E (see Scheme 4).

Photophysical and electrochemical properties of 9 and 10
The solution absorption spectra of ancillary ligands 9 and 10 are compared in Fig. 3. Compared to the absorption spectrum of 4 0phenyl-2,2 0 :6 0 ,2 00 -terpyridine, 43 that of 9 extends into the visible region as a result of the introduction of 4 0 -(4-(N,N-di(4methoxyphenyl)amino)phenyl) substituents and consistent with the spectrum of the rst generation analogue 4 (Scheme 2).Compounds 4 and 9 exhibit typical intra-ligand charge transfer (ILCT) bands with the tertiary amine functioning as an electron donor. 44Introduction of the benzothiadiazole unit in 10 extends the spectral response.The broad and intense absorption with l max ¼ 489 nm (3 max ¼ 26 100 dm 3 mol À1 cm À1 ) arises from charge transfer 45,46 and the lower energy bands from p* ) p transitions.The spectrum is consistent with those of the related compounds. 45,46ompounds 9 and 10 are electrochemically active and cyclic voltammetric (CV) data are presented in Table 1 and in Fig. S1; † processes are reversible unless otherwise stated.The reduction process at À1.96 V in 10 (absent in 9) arises from reversible reduction of the benzothiadiazole unit. 45,47The reversible oxidations are centred on the diphenylamine units.Of the oxidations observed for 9 or 10, the lowest potential process (+0.22 V) is for 9 which is consistent with the of the electron-releasing methoxy substituents.In order to support these conclusions, density functional theory (DFT) calculations were run at the B3LYP/6-31G* level.The orbital compositions of the HOMO and LUMO of each of 9 and 10 are shown in Fig. 4. The HOMO in 9 is delocalized over the 4,4 0 -bis(N,N-bis(4methoxyphenyl)amino)diphenylamino dendron (Fig. 4a), and the compositions of the HOMOÀ1 and HOMOÀ2 are similar (Fig. S2 †).The LUMO is based on the tpy domain (Fig. 4b) and the associated reduction process is presumably outside the solvent accessible window.In 10, the orbital contributions to the HOMO manifold (Fig. 4 and S2 †) are similar to those on 9, and both the LUMO (Fig. 4d) and LUMO+1 are localized on the benzothiadiazole units.

DSC fabrication and performance
Heteroleptic surface-bound [Zn(L anchor )(L ancillary )] 2+ complexes were assembled in a stepwise manner on an FTO/TiO 2 electrode using our 'surfaces-as-ligands' strategy. 17,28The electrode was initially immersed in a DMSO solution of 1 27 or 2 48 (Scheme 1), and then the ligand-functionalized electrode was dipped into an EtOH solution of ZnCl 2 to give a surface bound [Zn(L anchor )Cl 2 ] complex (Fig. 1).Finally, the electrode was dipped for ca.43 hours in either a CH 2 Cl 2 solution of 9 or a THF solution of 10, the solvent being chosen according to the solubility of the ancillary ligand.During this dipping cycle, the initially colourless TiO 2 layer changed to yellow-orange.The colour persisted when the electrode was dried, consistent with the formation of a surfacebound [Zn(L anchor )(L ancillary )] 2+ complex (Fig. 1).We note that the protonation state of surface-bound 1 and 2 is ambiguous and formulation of the dyes as [Zn(1)(9)] 2+ , [Zn(1)(10)] 2+ , [Zn(2)(9)] 2+ and [Zn(2)(10)] 2+ assumes full protonation.
A set of electrodes prepared using TiO 2 without a scattering layer was prepared in the same manner as those used in the DSCs, and their solid state absorption spectra were recorded.The spectra of dyes [Zn(1)(9)] 2+ and [Zn(2)(9)] 2+ are essentially identical, as are those of [Zn(1)(10)] 2+ and [Zn(2)(10)] 2+ (Fig. 5), consistent with the fact that, above 350 nm, the ancillary ligands dominate the absorption.The benecial effects of the thiadiazole units are seen in the enhanced absorbance above 400 nm, with maxima at 472 nm for [Zn(1)(10)] 2+ and 465 nm for [Zn(2)(10)] 2+ .These maxima correlate with the charge transfer band at 489 nm in the solution spectrum of ligand 10 (Fig. 3).Fig. 5 also shows the solid-state absorption spectrum of an electrode with adsorbed dye N719, conrming that while the   spectral response of the dyes is enhanced by incorporating the thiadiazole domains, it remains inferior to that of N719.The DSC measurements were made using sealed and fully masked [49][50][51] cells with an I À /I 3 À redox couple.Parameters of cells containing the zinc(II)-based dyes were compared with those of a DSC containing N719 (Table 3 and 4 Good ll factors (ff) of ca.70%.were obtained for all the DSCs (Table 2 and 3), but values of V OC are low (ca.400 mV, Table 2 and  3, and Fig. 7).Values of J SC were <1 mA cm À2 (compared with 15-17 mA cm À2 for N719).The DSCs show very low EQE values (#3%, Fig. 6), and in contrast to the enhanced absorption of the dyes containing ligand 10 versus 9, the trends in values of EQE max are   reversed (compare Fig. 5 and 6).This is consistent with the lower values of J SC for the dyes containing ancillary ligand 10 (Table 2  and 3).We rationalize this observation in terms of the positioning of the thiadiazole domain in the ancillary ligand.Although it leads to enhanced absorption in the visible region, its electronwithdrawing effect reduces electron transfer from the ancillary ligand over the metal centre and subsequently reduces electron injection into the TiO 2 .
The choice of anchoring ligand makes little difference to the performance of the dyes containing ancillary ligand 10.However, a combination of ancillary ligand 9 with phosphonic acid anchor 2 leads to slightly better dye performance than with carboxylic acid anchor 1 (Table 1).Both J SC and (to a lesser extent) V OC are enhanced on going from carboxylic to phosphonic acid anchor.However, for all the dyes, performance was disappointingly poor, although better than with rst generation ancillary ligands 4-7. 28n order to verify the validity of the DSC measurements, parameters of two blank cells (i.e.FTO/TiO 2 without adsorbed dye) were recorded (Table 4).A comparison of data in Tables 2-4 demonstrates that there are negligible differences between the parameters obtained for the blank cells and some of the DSCs containing the zinc(II) dyes.The results underline the difficulties of measuring parameters for poorly performing sensitizers.However, we note that the EQE spectra of dyes [Zn(1)(9)] 2+ , [Zn(2)(9)] 2+ , [Zn(1)(10)] 2+ and [Zn(2)(10)] 2+ conrm that electron injection (albeit small) does occur.Only for [Zn(2)(9)] 2+ was the efficiency higher than the blank cell and this is due to a higher value of J SC .

HOMO and LUMO characteristics of the zinc(II) dyes
Ground state DFT calculations on the zinc(II) dyes [Zn(2)(9)] 2+ and [Zn(2)(10)] 2+ were carried out to gain some insight into their poor DSC performances.In a detailed theoretical study of two representative bis(diimine) copper(I) sensitizers, we demonstrated that the calculated absorption spectra of the dyes depend upon the atomic orbital basis set (6-311++G** basis set on all atoms, 6-311++G** on copper and 6-31G* basis set on C, H and N, or 6-31G* basis set on all atoms) but that the inuence on the orbital compositions of the HOMOs and LUMOs are little altered. 52For a qualitative evaluation of the MOs of [Zn(1)(9)] 2+ , [Zn(1)(10)] 2+ , [Zn(2)(9)] 2+ and [Zn(2)(10)] 2+ , we have used a 6-31G* basis set on all atoms. 15n their ground state, each of the complexes exhibits a LUMO centred on the anchoring ligand (Fig. 8a and Fig. S3a-S5a †) although it is localized on the tpy domain rather than the carboxylic or phosphonic acid units which would be optimal for electron injection.The HOMO is localized on the peripheral diphenylamino groups of the ancillary ligand (Fig. 8b and Fig. S3b-S5b †) which should facilitate hole-transfer over these domains to the reduced electrolyte.In each complex, the characters of the lled MOs lying immediately below the HOMO are also dominated by contributions from the peripheral groups of the ancillary ligands.
From the zinc(II) dye containing 10 to organic dye 11 Although the poor performances of the zinc(II) dyes containing second generation ancillary ligands 9 and 10 were discouraging,   we decided to capitalize on the absorption properties imparted by dendron 10a by combining this domain with a 2-cyanoacrylic acid anchoring unit to give the organic dye 11 (Scheme 2).Compound 11 was prepared by the route shown in Scheme 5. 4-Bromobenzaldehyde and 10a were reacted under Hartwig-Buchwald amination conditions, and the resulting aldehyde treated with 2-cyanoacetic acid in the presence of piperidine in a Knoevenagel condensation. 53Compound 11 was isolated as a red solid in 56% yield.
The highest mass peak envelope in the negative mode high resolution ESI mass spectrum of 11 was observed at m/z 941.2470, consistent with the [M À H] À ion.The 1 H and 13 C NMR spectra were assigned using NOESY, COSY, HMQC and HMBC spectra.The 2-cyanoacrylic acid anchoring group was characterized by 13 C NMR resonances with the C]O and C^N groups at d 164.0 and 118.1 ppm, respectively.The 1 H NMR spectrum is shown in Fig. 9, and is consistent with the presence of the benzothiadiazole units and two generations of amino domains.The cyclic voltammogram of 11 was recorded in CH 2 Cl 2 and exhibited a reversible reduction at À1.91 V; this potential is close to that for the rst reduction of 10 (Table 1) and is assigned to a benzothiadiazole-based process.Within the solvent accessible window, compound 11 exhibits three oxidation processes at potentials similar to those for 10 (Table 1) and are, presumably, based on the diphenylamino-containing domains.These assignments are corroborated by the results of the DFT calculations discussed below.
The solution absorption spectrum of 11 (Fig. 3) is dominated by intense high energy bands assigned to p* ) p transitions and a broad CT band with maxima at 431 and 482 nm.The solid-state absorption spectrum of an FTO/TiO 2 electrode (which appears bright red by eye) consists of a very broad and intense CT band centred at 460 nm (Fig. S6 †) and shows extended spectral response in the red region with respect to the solution absorption (Fig. 3).
Two DSCs incorporating 11 as the sensitizer, and a reference electrode with N719 were fabricated.The performance data in Table 5 conrm the reproducibility of the DSCs and reveal that sensitizer 11 achieves photon-to-current conversion efficiencies that are z70% relative to 100% for N719.The J-V curves in Fig. 10 demonstrate good ll factors, and high values of both J SC and V OC .The DSCs were monitored over a 17 day period and remained stable; no bleaching of the cells was observed.The antenna structure in 11 and the presence of the electronwithdrawing thiadiazole domains contribute to effective V OC /mV ff/% h/% electronic communication and electron injection.The EQE spectra for the two DSCs containing 11 (Fig. 11) conrm with enhanced electron injection across a broad spectral region.The broad EQE spectrum correlates well with the solid-state absorption spectrum (Fig. S6 †).Moving the thiadiazole units closer to the anchoring domain should have benecial effects and we are currently investigating this and other structural modications as a means of enhancing dye performance.

HOMO and LUMO characteristics of organic dye 11
Ground state DFT calculations on 11 reveal that the LUMO and LUMO+1 are close in energy and their combined character is predominantly localized on the 2-cyanoacrylic acid anchoring domain and on the benzothiadiazole groups (Fig. 12a).The HOMO is delocalized as depicted in Fig. 12b, while the HOMOÀ1 possesses character mainly on the peripheral diphenylamino units.The orbital compositions of the HOMO and LUMO manifolds correspond to what is required, respectively, for efficient hole-transfer over the peripheral diphenylamino units to the reduced electrolyte, and localization of the electron on the anchoring domain aer excitation.A distinction between the characteristics of the lowest lying virtual MOs in dye 11 and the four zinc(II) complexes is the dominant contribution from the anchor domain in 11 (Fig. 12a) versus the localization of the LUMO close to the metal centre in the heteroleptic dyes (Fig. 8a).

Fig. 1
Fig.1Assembly of zinc(II)-complexes in a stepwise manner using the 'surface-as-ligand, surface-as-complex' methodology in which the FTO/TiO 2 surface (blue) is sequentially functionalized with an anchoring ligand (exemplified with a phosphonic acid anchoring domain, red ¼ tpy metal-binding domain), metallated with ZnCl 2 , and finally capped by an ancillary ligand (purple).
). Duplicate cells were measured for each dye and the efficiencies of the DSCs were remeasured 2, 3 and 10 or 17 days aer sealing.External quantum efficiency (EQE) spectra are shown in Fig. 6.For clarity, only the EQE spectrum of the DSC with the highest EQE max for each pair of cells is shown; the same trend in values is observed for both cells, differences falling within the limits of the errors of the measurements.Fig. 7 shows plots of the shortcircuit current density (J SC ) against open-circuit voltage (V OC ) for DSCs containing [Zn(1)(10)] 2+ and [Zn(2)(10)] 2+ .
)] 2+ , [Zn(2)(9)] 2+ , [Zn(1)(10)] 2+ and [Zn(2)(10)] 2+ have been assembled and tested in n-type DSCs.The solid-state absorption spectra of dye-functionalized electrodes conrm a broad spectral response for all the dyes with enhanced intensity for the dyes containing ancillary ligand 10.A change from the carboxylic acid anchor 1 to phosphonic acid anchor 2 has little effect on light absorption.The dyes perform poorly, with very low J SC and V OC .The EQE spectra conrm that [Zn(1)(9)] 2+ and [Zn(2)(9)] 2+ exhibit better electron injection than [Zn(1)(10)] 2+ and [Zn(2)(10)] 2+ .In part, this is probably a consequence of the non-optimal positioning of the thiadiazole domain in the dye.Interpretation of the DSC parameters for this set of dyes highlights a problem that is oen ignored when reporting global efficiencies of poor dyes; DSCs incorporating FTO/TiO 2 photoanodes without adsorbed dye generate small short-circuit current densities and open-circuit voltages which contribute to parameters reported for inefficient dyes.Organic dye 11 is structurally similar to ancillary ligand 10 and has an excellent spectral response in the range 300-600 nm.EQE data for 11 evidence efficient electron injection over a broad wavelength range, comparable with N719 except above 600 nm.DSCs containing 11 are stable over at least 17 days and show global efficiencies of 3.93-4.57%(ca.70% with respect to N719 set at 100%).Ground state DFT calculations demonstrate that each of [Zn(1)(9)] 2+ , [Zn(2)(9)] 2+ , [Zn(1)(10)] 2+ , [Zn(2)(10)] 2+ and 11 possesses a HOMO localized on the peripheral diphenylamino units.In 11, a dominant contribution from the 2-cyanoacrylic acid anchoring group appears in the lowest lying virtual MOs; in each zinc(II) dye, the LUMO resides on the anchoring ligand but is concentrated on the tpy domain close to the metal centre which may militate against good electron injection.

Fig. 11
Fig. 11 EQE spectra of DSCs containing 11 compared to N719, measured on the day of assembling the cells.

Table 3
DSC performance of [Zn(L anchor )(10)] 2+ .Relative efficiencies are with respect to 100% for N719 measured under the same conditions.Each experiment was measured with two cells (1 and 2 in the table)

Table 2
DSC performance of [Zn(L anchor )(9)] 2+ .Relative efficiencies are with respect to 100% for N719 measured under the same conditions.Each experiment was measured with two cells (labelled 1 and 2 in the table)

Table 4
DSC parameters for blank cell (FTO/TiO 2 photoanode).Relative efficiencies are with respect to 100% for N719 measured under the same conditions.Each experiment was measured with two cells

Table 5
DSC performance of 11.Relative efficiencies are with respect to 100% for N719 measured under the same conditions.Each experiment was measured with two cells J SC /mA cm À2