Stability of ruthenium/organic dye co-sensitized solar cells: a joint experimental and computational investigation

Abstract

We carried out a joint experimental and theoretical study of the stability of dye-sensitized solar cells using a mixture of black dye (N749) and a Y1 organic co-adsorbent. The aim of this work was to investigate the stability of these highly efficient sensitizers. The co-sensitized device showed remarkable stability under the conditions investigated (1100 h of about 1 Sun light soaking at 55 °C). The partial desorption of the organic co-adsorbent was identified as a possible cause of the slight reduction in the photocurrent value. Theoretical investigations on the dye-sensitized semiconductor surface using DFT methods showed a considerably lower adsorption energy for Y1 compared with N749, in particular upon oxidation, possibly leading to desorption of the dye under working conditions. The devices using only the Y1 organic dye consistently showed a considerably lower stability.

---

1 Introduction

The depletion of the Earth's natural resources and environmental concerns have led to a need to develop efficient alternative energy technologies. Of these technologies, photovoltaic devices that directly convert visible light to electrical power are now seen as a viable solution. Dye-sensitized solar cells (DSSCs) are an emergent photovoltaic technology^1–4 and have the highest reported conversion efficiencies to date (13%).⁵ In addition to high efficiency, long-term stability is another requirement for the possible commercialization of this technology.

A standard DSSC consists of a wide band gap titanium dioxide (TiO₂) metal oxide layer sintered over a transparent conducting fluorine-doped tin oxide electrode. A monolayer of photoactive dye is adsorbed on the nanocrystalline TiO₂ surface. A redox shuttle [I⁻/I₃⁻ or Co(II)/Co(III) complex] solution fills the interstitial space between the sensitized semiconductor and a platinized counter electrode. In these cells, the photo-excited dye transfers an electron to the TiO₂ conduction band and the oxidized dye is then reduced by the redox mediator in the electrolyte. The injected electrons percolate through the metal oxide network, reach the fluorine-doped tin oxide electrode and are then transferred to the platinized counter electrode through the external circuit. At the counter electrode, the electrons regenerate the redox shuttle and close the circuit. Dyes have a critical role in DSSC devices in determining efficient charge generation and separation at the interface with the semiconductor. Ruthenium-based dyes,^6–8 porphyrin dyes^9–12 and organic dyes^13,14 are the most studied classes of sensitizers in this field of research.

Light absorption in the visible and near-infrared (IR) part of the solar spectrum is a key requirement of a DSSC sensitizer. In addition to structural engineering of the dye, co-sensitization with multiple dyes having complementary light absorption has been explored with the aim of enhancing the spectral response in the near IR region.^15–18 A peculiar application of co-sensitization was reported by Han et al.¹⁹ in 2012. To outweigh competitive light absorption by the I⁻/I₃⁻ electrolytes,²⁰ they co-adsorbed the Y1 organic dye (Scheme 1), which has a strong absorption at 400 nm, with the panchromatic dye [Ru(H₃tcterpy)(NCS)₃] (tcterpy = 4,4′,4′′-tricarboxy-2,2′:6′,2′′-terpyridine), known as N749 or black dye (Scheme 1), resulting in an improved certified efficiency of 11.4%.¹⁹

Scheme 1 Chemical structures of Y1 dye, Y2 dye and N749 dye, with the tetrabutylammonium counter ion. R = C₄H₉.

Another important requirement of a sensitizing dye is the grafting capability at the semiconductor surface. Three fundamental properties essentially depend on the anchoring moieties: (i) the electronic coupling between the dye and the metal oxide and thus the charge injection efficiency;^21,22 (ii) the orientation and packing of the adsorbed dyes, which influences the recombination processes;^23,24 and (iii) the long-term stability of the device.^25,26

For these reasons, the binding behaviour of the sensitizer on anatase TiO₂ surfaces has been investigated in detail. For an organic dye bearing one carboxylic acid, the anchoring group most often used, FT-IR analysis and theoretical calculations have shown that the most stable anchoring mode is the dissociative bridging bidentate mode.^27–31 For ruthenium-based sensitizers, which generally have two or more anchoring sites, different configurations are possible depending on the geometry of the complex. In this instance the preferred adsorption geometry also seems to be the bridging bidentate mode, sometimes coexisting with the deprotonated monodentate mode.^32–38

The stability of a correctly assembled and sealed DSSC device under working conditions (1 Sun light soaking, T > 50 °C, variable moisture rate) mainly depends on the degradation and/or desorption of the sensitizer in the electrolyte solution, which dramatically affects the photocurrent production.^39–41 The electrolyte composition plays a fundamental part in the degradation pathways at the dye/semiconductor interface. In particular, the substitution of thiocyanate with a 4-tert-butylpyridine additive or with nitrile-based solvent molecules has been observed in ruthenium-based dyes containing NCS ligands, which leads to significant reductions in performance;^42–44 this reaction only takes place at significant rates at increased temperatures (80–100 °C). Solvent molecules can introduce important modifications to the geometry of dye adsorption, affecting the stability of the interaction and potentially leading to desorption of the dye.^45,46

We report here a joint experimental and theoretical investigation of the stability of DSSC devices sensitized with a black dye and Y1 co-adsorbent (Scheme 1), as reported by Han et al.¹⁹ The DSSC devices were subjected to 1 Sun illumination at 55 °C. The photoconversion efficiencies were measured on fresh and aged devices after 300 and 1100 h of light soaking. As a first approximation of the reasons for performance degradation, the stability of dyes anchored on a TiO₂ surface was theoretically investigated. Our results show that the devices maintain significant stability under the investigated conditions. The slight decrease in the overall efficiency resulting from the reduction in the photocurrent may be due to desorption of the organic dye because of a weaker interaction at the TiO₂ interface on photoexcitation and during the charge injection processes.

2 Experimental section

2.1 Materials and preparation of solar cells

All the chemicals used in this work were stored in a dry room (relative humidity about 5%) and were used without further purification. All materials were purchased from either Dyesol or Sigma-Aldrich.

TCO sheet glass electrodes (Dyesol TEC15 2.3 mm, 15 Ω sq⁻¹) were used as substrates for the nanocrystalline TiO₂ films. A thin TiO₂ underlayer (blocking layer) was deposited by immersing the TCO/glass substrate in a TiCl₄ aqueous solution for 30 min at 70 °C and then sintering at 500 °C for 1 h. Mesoporous TiO₂ thin films (about 15 μm thickness, 0.88 cm²) were deposited onto the conducting glass by screen-printing using a Dyesol 18NR-AO TiO₂ paste and then sintering at 500 °C for 1 h. The thickness of the films was measured with a DEKTAK 150 surface profiler (Veeco Instruments Inc.). The dye solutions used were either a solution of black dye (Dyesol) or the Y1 co-adsorbent, synthesized according to the previously published procedure,¹⁹ at 0.2 mM concentration in dry ethanol, or a mixture of both at a 1 : 1 ratio. The nanoporous TiO₂ films were immersed in these solutions at 25 °C for 24 h.

The working and counter electrodes (platinum-coated conducting glass; Dyesol) were separated by an 80 μm thick Bynel spacer and sealed by heating the polymer frame. A Dyesol EL-HSE iodine-based high-stability electrolyte using 3-methoxypropionitrile (MPN) as a solvent was inserted into the cells via vacuum backfilling and patches of thermoplastic material and glass were used to close the fill hole. The devices were secondarily sealed by a Dyesol two-part thermal cure epoxy compound.

2.2 Methods

The UV-visible absorption spectra were recorded in acetonitrile (ACN) solution on a PerkinElmer Lambda 800 spectrophotometer.

The cell performance was evaluated by recording the photocurrent density–photovoltage (J–V) characteristics and by the incident photon to current efficiency (IPCE). The J–V characteristics of the DSSCs were recorded using a computer-controlled digital source meter (Keithley 2420) and a 150 W metal halide lamp calibrated by a silicon reference cell and corrected for any spectral mismatch on a clear day under full Sun outdoor illumination (close to AM 1.5). The incident light intensities were adjusted with wire mesh attenuators to collect data at various light levels. The IPCE was determined using a Peccell PEC-S20 Action Spectrum measurement system in DC mode. Calibration was carried out using a silicon photodiode (Model S1337-1010BQ).

Long-term light soaking was undertaken under the light of a metal halide lamp at about 1 Sun and a constant cell temperature was maintained at about 55 °C. All the cells were maintained close to their maximum power during long-term testing by a resistive load.

2.3 Computational details

All the calculations were performed by DFT implemented in the Gaussian09 program suite.⁴⁷ The ground state geometry optimizations were obtained using the B3LYP hybrid functional⁴⁸ together with the 6-31G* basis set⁴⁹ in ACN solution. Solvent effects (ACN) were included using the conductor-like polarization solvation model.⁵⁰

We used a neutral stoichiometric (TiO₂)₃₈ cluster to model the dye/TiO₂ interface,^51–53 This was obtained by appropriately “cutting” an anatase slab and exposing the majority (101) surface.⁵⁴ The most widely accepted bidentate bridging mode was used to model the Y1 and Y2 organic dye@TiO₂ complexes.²⁷ The N749@TiO₂ complex was modelled through dissociated monodentate binding involving two carboxylic groups.^55,56 The initial conformations for ACN and MPN on TiO₂ were modelled by keeping about 2.0 Å distance between the surface Ti atoms and the N atom of the solvent molecule. The equilibrium geometries for these complexes in ACN solution were obtained using the B3LYP functional and a 6-31G* basis set for all atoms except ruthenium, for which a DZVP basis set was used.

The adsorption energies (ΔE_ads) of dyes adsorbed on the TiO₂ surface were calculated from the optimized geometries in solution according to the equation ΔE_ads = E_Com – (E_Dye⁻ + E_TiO2–H⁺). E_Com is the total energy of the dye@TiO₂ complex, E_Dye⁻ is the energy of the deprotonated dye adsorbed on the TiO₂ surface and E_TiO2–H⁺ is the energy of the TiO₂ surface with the dissociated proton from the dye. The energies for the deprotonated dye and the protonated TiO₂ fragments at the complex geometry were calculated by single-point calculation in ACN solution using same level of theory as used for the complexes.

3 Results and discussion

3.1 Experimental results

We fabricated three different kinds of device to differentiate between the effects of each sensitizer on the overall stability of the solar cells: (1) a device using only the Y1 organic sensitizer; (2) a device using only the N749 dye; and (3) a device using a mixture of both sensitizers (Table 1). Co-sensitization with N749–Y1 resulted in much higher efficiencies than the individual Y1 and N749 dyes: 2.71, 4.94 and 5.75% efficiency, respectively (Table 1). The efficiencies obtained in this work were considerably lower than those reported previously,¹⁹ probably due to the thinner TiO₂ films (15 vs. 25 μm), the larger active surface area (0.88 vs. 0.25 cm²) and the high stability of the electrolyte. Our electrolyte used MPN as the solvent, which has a higher boiling point (164–165 °C) than ACN (81–82 °C), but has been shown to have a lower performance.^57–59

Table 1 Photovoltaic parameters (assessed at 1/3 Sun) for DSSC devices. The measurements were made on both fresh and aged (about 1 Sun light soaking at about 55 °C) samples. The percentage difference (Δ) between the measured parameters and the fresh devices are also given

Dye	Time (h)	Jsc (mA cm^−2)	Voc (mV)	FF	η (%)
Y1	0	2.41	557	0.67	2.71
	300	2.45	561	0.66	2.73
	Δ (%)	+1.3	+0.7	−0.9	+1.0
	1100	2.12	510	0.68	2.19
	Δ (%)	−12.3	−8.9	+1.1	−19.3
N749	0	5.19	565	0.56	4.94
	300	5.11	609	0.56	5.28
	Δ (%)	−1.6	+7.9	+0.5	+6.8
	1100	5.08	590	0.58	5.23
	Δ (%)	−2.0	+5.0	+2.7	+5.8
N749 + Y1	0	5.62	599	0.57	5.75
	300	5.33	636	0.59	5.98
	Δ (%)	−5.1	+6.1	+3.2	+3.9
	1100	5.26	610	0.60	5.74
	Δ (%)	−6.5	+1.5	+5.1	−0.3

---

By using this co-sensitization approach, we obtained an increase in the J_sc values compared with the device based on N749 alone, which is consistent with the increase in the light harvesting capability in the blue region (350–500 nm) of the solar spectrum (Fig. 1). This is readily converted into a gain in the IPCE in the same spectral region (see ESI†).

Fig. 1 Absorbance spectra of Y1 (red line), N749 (dotted black line) and an equimolar dye mixture of N749–Y1 (broken blue line) in ACN solution.

We also observed a significant increase (about 30 mV) in the V_oc values, mainly due to the shielding effect of the Y1 co-adsorbent on the TiO₂ surface, which effectively suppressed the recombination processes and enhanced the lifetimes of the electrons.¹⁹ A key part in this regard is played by the butoxyl chains on the donor part of the molecule, which are able to keep the oxidized I₃⁻ ions far away from the semiconductor surface and to efficiently act as an anti-aggregation agent.⁶⁰ Despite the lower efficiencies, the trends of the measured parameters are consistent with those in the original paper.¹⁹

The stability of the investigated devices was assessed with J–V measurements at 1/3 Sun (Table 1; Fig. 2). All the devices remained stable after 300 h of light soaking. We observed slight increases in the performance (about 1% for Y1, 4% for Y1–N749 and 7% for N749), mostly due to a slight increase in the V_oc value. The N749-sensitized device remained stable after 1100 h of light soaking in agreement with previous investigations of ruthenium dyes under similar test conditions.^57,61,62 The organic Y1 dye, however, showed a substantial decrease with respect to the initial efficiency (−19.3%) as a result of a parallel reduction in both the J_sc and V_oc values. We observed an intermediate behaviour for the device using the mixture of N749–Y1 dyes, with a negligible decrease in efficiency, suggesting that the performance of the device is slightly affected by the lower long-term stability of the Y1 co-adsorbent compared with N749.

Fig. 2 Stability of the photovoltaic parameters during ageing tests (300 h and 1100 h under about 1 Sun light soaking and about 55 °C). Measurements were made at 1/3 Sun illumination.

3.2 Theoretical calculations

One of the main reasons for the lack of long-term stability is the degradation or desorption of the dye. We carried out a computational investigation on the adsorption geometries and energies on an anatase TiO₂ cluster to determine the different stabilities of the two dyes. Our aim was to try to understand whether there was a correlation between the adsorption behaviour of the sensitizers and their long-term stability in the DSSC devices. We modelled the dyes in both their neutral form and after the removal of one electron to simulate the stabilities of the neutral and oxidized dyes; oxidation occurs after the charge injection process.

We simulated the N749 and Y1 dyes and then extended our analysis to the Y2 co-adsorbent used in the original work¹⁹ with slightly lower performances (Scheme 1). To assess the competition from the solvent molecules in binding the under-coordinated TiO₂ sites, we also simulated the adsorption of the ACN and MPN molecules as these are the solvents most commonly used in liquid electrolyte solar cells. The N749 dye is an anionic complex⁶³ and we therefore modelled two forms: the anionic form and the neutral form considering one tetrabutylammonium (TBA) counter ion. This was placed near to the thiocyanate ligands, where a partial negative charge is mainly localized. Fig. 3 shows the optimized geometries for all the considered systems.

Fig. 3 Optimized geometries of Y1, Y2, N749 dye, N749–TBA, ACN and MPN.

Organic co-adsorbent dyes. The optimized Y1 and Y2 dyes show fairly planar geometries, except for a dihedral angle of 24° between the phenyl and thiophene planes (Fig. 3). This distortion from planarity does not prevent the conjugation extending over the whole donor–π spacer–acceptor (D–π–A) molecular structure, as observed by Azar and Payami.⁶⁰ This is clearly visible from the HOMO–LUMO plot of the Y1 and Y2 dyes (Fig. 4).

Fig. 4 HOMO–LUMO isodensity plot of Y1 and Y2 dyes.

In a similar manner to related push–pull sensitizers, the HOMO is localized on the donor and π-bridge moieties, whereas the LUMO is mainly located over the cyanoacrylic acceptor group. The high overlap between the HOMO and the LUMO, which are mainly involved in the lowest energy transition, may account for the high molar extinction coefficients measured for the two dyes (3.0 × 10⁴ and 3.1 × 10⁴ M⁻¹ cm⁻¹ for Y1 and Y2, respectively).¹⁹

We noted a slight reorganization of both molecules during the optimization of the oxidized Y1 and Y2 dyes (Fig. 5). In particular, we calculated the planarization of the dye structures in conjunction with a reduction in the C–C bond length between the phenyl and the thiophene ring (from 1.46 to 1.42 Å for both molecules). Analysis of the spin density plots (Fig. 5) for Y1^ox shows that the hole is highly delocalized along the whole backbone of the dye. For the Y2^ox dye, however, the hole is mainly localized on two carbon atoms on the spacer and acceptor parts of the molecule, with a significant contribution on the cyanoacrylic moiety as well. This different behaviour between the two molecules may be due to the different substituents on the donor part of the molecules. The electron-donating effect of the alkoxy groups for Y1 is probably able to better stabilize the positive charge on the oxidized structure. The higher localization of the hole in the acceptor region of the Y2 dye molecule may account for the slightly lower photovoltaic performance compared with Y1 (ref. 19) as a result of an increase in the recombination probability of the injected electrons with the oxidized dye.

Fig. 5 Spin density plots (purple surface) for the oxidized Y1 and Y2 dyes.

To simulate adsorption of the dyes onto the TiO₂ anatase (101) surface, dye molecules were chemisorbed using a dissociative bridging bidentate anchoring mode, as proposed previously.^27,64 The two oxygen atoms of the COO⁻ anchor were bound to five coordinated titanium atoms, whereas the acidic proton was bound to a vicinal oxygen atom. Fig. 6a and b show the equilibrium geometries of these complexes; the main geometrical parameters for the dye–TiO₂ interfaces are given in the ESI.†

Fig. 6 Optimized geometries of (a) Y1@TiO₂, (b) Y2@TiO₂, (c) N749@TiO₂, (d) [N749-TBA]@TiO₂, (e) ACN@TiO₂ and (f) MPN@TiO₂.

Both dyes show a fairly strong interaction with the semiconductor surface with short bond distances (within 2.0 and 2.2 Å) between the carboxylic oxygen and the coordinating titanium atoms. Further inspection of the optimized structures reveals that the dye geometry does not show any substantial changes on adsorption onto the semiconductor. We next simulated the Y1 cation adsorbed on the TiO₂ surface. Interestingly, compared with the neutral dye@TiO₂ complex, we noted a partial elongation (about 0.05 Å) of both the Ti–O bonds of the anchor, suggesting a slightly less stable interaction between the dye and the semiconductor surface on oxidation. Inspection of the spin density plot of Y1^ox@TiO₂ (see ESI†) again reveals an efficient delocalization of the hole along the whole dye backbone, as found for the isolated dye in solution. The delocalization of the hole justifies the slight changes in the anchoring geometry. Further inspection of the complex structure highlights the same reorganization of the molecular structure as observed for the dye in solution, with a planarization of the phenyl and thiophene dihedral angle.

N749 in solution and adsorbed on TiO₂. The same computational procedure was carried out for the fully protonated N749 ruthenium dye. Fig. 3 and 7 (for oxidized compounds) show the optimized geometries.

Fig. 7 Spin density plots (purple surface) for the oxidized N749 and N749-TBA systems.

Fig. 7 also shows the spin density plots. The hole is highly localized on the metal centre for both systems. The main geometrical variations on the oxidation of the complexes relate to the Ru–N bonds, which are generally contracted by about 0.06 Å as a result of the need to stabilize the hole. No effect is observed on the terpyridine ligand, which bears the anchoring carboxylic groups. It is worth noting that for the N749 dye the hole is localized away from the anchor, in contrast with the Y1 and Y2 dyes. This is an important factor in retarding electron–hole recombination and is one of the reasons for the impressive photovoltaic performance of the N749 dye.

The adsorption geometry of the N749 dye on TiO₂ has been studied in detail previously. Bauer et al.⁶⁵ suggested a bidentate binuclear coordination on the basis of an experimental FT-IR investigation. Different results have been obtained from theoretical analyses depending on the computational approach used.^{55,56,66–69} Some of us have reported^55,56 that anchoring through two dissociated monodentate carboxylic groups results in the most stable form in a similar manner to results obtained for heteroleptic ruthenium dyes.⁷⁰ Tateyama et al.,⁶⁹ using molecular dynamics studies, confirmed this binding mode as the most favoured, although single anchoring through protonated carboxylic acids may be present to a lesser extent.

We focused here on only the most stable adsorption mode through two dissociated monodentate carboxylic groups (Fig. 6c and d). As expected, a strong interaction was found between the dye and the TiO₂ surface, with Ti–O bond distances <2 Å. A strong contribution to the stabilization of the anchoring is given by the formation of hydrogen bonds between the carboxylate oxygen and the protons transferred to the TiO₂ surface. This is also suggested by the reduction in the stability of the anchoring in the presence of protic solvents such as water.^45,71 The best results in terms of device stability were obtained with amphiphilic dyes, such as Z907, which are able to effectively insulate the dye/semiconductor interface from the approach of water.⁷²

Optimized geometries for the oxidized N749@TiO₂ complexes show only a marginal or no difference in the lengths of adsorption bonds between the carboxylate oxygen and the surface titanium atoms. In a similar manner to the isolated N749 dye in solution, we can see from the spin density plot that the hole in the oxidized complex is completely localized on the ruthenium centre and thus it has a negligible effect on the terpyridine ligand and the carboxylic groups. A detailed list of the geometrical parameters for the different investigated species is reported in the ESI.†

ACN and MPN solvent molecules. Fig. 2 gives the optimized geometries for the ACN and MPN solvents. To understand the behaviour of these solvents at the TiO₂ interface, one molecule each of the ACN and MPN electrolyte solvents was adsorbed at a five-fold coordinated Ti atom of the surface (Fig. 6e and f). Both solvents interact with the surface via a nitrogen atom and stabilize at a distance of 2.31 Å. In reasonable agreement with a previous investigation,⁷³ we found a “standing” structure for the adsorbed solvent molecules, with angles between the Ti, N and C atoms (of the nitrile groups) of about 170°.

Adsorption energies. Table 2 gives the calculated adsorption energies for the investigated complexes. The results show a comparable binding between the organic Y1 and the Y2 dyes (−71.1 and −69.3 kcal mol⁻¹, respectively). The N749 dye, in both the anionic form and neutral form, with the TBA counter ion, has a much stronger adsorption, with energies of about −120 kcal mol⁻¹. The large difference between the binding energies of the Y1/Y2 and N749 dyes are easily explained by their different adsorption geometries (Fig. 6). The Y1 and Y2 dyes are anchored with one COO– group on the TiO₂ surface, whereas the N749 dye is adsorbed via two COO– moieties. Further stabilization of the black dye adducts occurs via the formation of hydrogen bonds between the carboxylate oxygen atoms and the dissociated hydrogen atoms on the TiO₂ surface. The electrolyte solvents show considerably lower adsorption energies, of almost one order of magnitude, with respect to the sensitizing dyes and have negligible competition for the semiconductor binding sites.

Table 2 Adsorption energies of the investigated compounds on the anatase (101) TiO₂ surface

Dye@TiO2	ΔEads (kcal mol^−1)
Y1	−71.1
Y1^ox	−58.9
Y2	−69.3
N749	−118.0
[N749]^ox	−110.1
[N749-TBA]	−121.3
[N749-TBA]^ox	−120.4
Solvent@TiO2	ΔEads (kcal mol^−1)
ACN	−13.8
MPN	−8.5

---

Interestingly, the higher adsorption energy of the N749 dye suggests a stronger tendency to graft onto the surface during the sensitization process. At the same concentration of the Y1 and N749 dyes in the dye-bath solution, the ruthenium dye is probably preferably adsorbed, while the small organic dye fills the remaining binding sites. This may explain the improved photovoltaic performance of the device sensitized with the dye mixture because the presence of the Y1 dye does not limit the adsorption of the more efficient black dye. The light absorption and charge transfer properties of Y1 instead have an additive effect on the overall conversion efficiency of the co-sensitized devices. This assumption is also consistent with the degradation behaviour of the DSSC devices. The cells sensitized with the mixture of both dyes show only a slight decrease in efficiency with light soaking (−0.3%) in a similar manner to the stability behaviour of the individual N749 dye.

Charge transfer from the photo-excited dye to the TiO₂ conduction band results in the localization of holes on the oxidized dye. The electron-deficient donor formed temporarily may weaken the interaction between the adsorbed dye and the TiO₂ by withdrawing electron density from the anchor moiety. Therefore it is also important to look at the adsorption energies for the oxidized dye@TiO₂ complexes. Our results show an adsorption energy for the oxidized Y1^ox@TiO₂ complex of −58.9 kcal mol⁻¹, which is considerably lower (−17%) than that for the neutral system. However, the oxidized N749@TiO₂ complexes show a negligible reduction in binding energies (−6.7% or −1% depending on the model used). The hole is strongly localized on the metal centre in the ruthenium dye, with negligible effects on the terpyridine ligand and the carboxylic anchoring groups, reflecting the slight reduction in adsorption energy. In the Y1 dye, the hole delocalizes along the dye structure, showing a partial lengthening of the anchoring bond distances and a consistent reduction in the adsorption energy.

These data suggest that one of the main causes of the observed lower long-term stability for the measured devices under light soaking and thermal stressing conditions is the desorption of the Y1 dye, in particular during oxidation. The decrease in conversion efficiency for the Y1-sensitized devices is mainly a result of a reduction in the photocurrent, which is consist with previously published work on dye desorption.^39,40 Significant improvements in the long-term stability of organic dyes have been obtained when using di-anchoring sensitizers,^74,75 suggesting that the stability of the dye/semiconductor interaction is a key issue in developing efficient and robust dyes for DSSC applications.

4 Conclusions

We investigated the stability of DSSC devices using black dye and the Y1 organic co-adsorbent under light soaking for 1100 h at 1 Sun at 55 °C. The devices using the N749 sensitizer or the dye mixture showed remarkable stability, whereas a 20% reduction in the photovoltaic efficiency was measured for the isolated Y1 organic dye. A theoretical investigation of the adsorption energies and geometries of these sensitizers on an anatase TiO₂ cluster was carried out to evaluate the stability of the dye anchoring as a possible cause of the performaces degradation. In agreement with the experimental results, we found a considerably higher adsorption energy for the N749 dye with respect to the Y1 dye as a result of the double carboxylic anchoring. The latter also showed a significant reduction in the binding energy upon oxidation, suggesting dye desorption as a reliable explanation for the lower stability of the device under stress conditions.

Acknowledgements

The authors thank the joint DST-EU Project ‘ESCORT’ (FP7-ENERGY-2010, contract 261920) for financial support of this work. M.C. thanks DST, New Delhi for funding Project no. DST/TMC/SERI/FR/92.

Notes and references

1. M. Grätzel, Nature, 2001, 414, 338–344.
2. M. Grätzel, Acc. Chem. Res., 2009, 42, 1788–1798.
3. A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo and H. Pettersson, Chem. Rev., 2010, 110, 6595–6663.
4. B. O'Regan and M. Grätzel, Nature, 1991, 353, 737–740.
5. S. Mathew, A. Yella, P. Gao, R. Humphry-Baker, F. E. Curchod Basile, N. Ashari-Astani, I. Tavernelli, U. Rothlisberger, M. K. Nazeeruddin and M. Grätzel, Nat. Chem., 2014, 6, 242–247.
6. M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphry-Baker, E. Mueller, P. Liska, N. Vlachopoulos and M. Grätzel, J. Am. Chem. Soc., 1993, 115, 6382–6390.
7. M. K. Nazeeruddin, P. Pechy and M. Grätzel, Chem. Commun., 1997, 0, 1705–1706.
8. P. G. Bomben, K. C. D. Robson, B. D. Koivisto and C. P. Berlinguette, Coord. Chem. Rev., 2012, 256, 1438–1450.
9. W. M. Campbell, A. K. Burrell, D. L. Officer and K. W. Jolley, Coord. Chem. Rev., 2004, 248, 1363–1379.
10. A. Kay and M. Grätzel, J. Phys. Chem., 1993, 97, 6272–6277.
11. L.-L. Li and E. W.-G. Diau, Chem. Soc. Rev., 2013, 42, 291–304.
12. S. Namuangruk, K. Sirithip, R. Rattanatwan, T. Keawin, N. Kungwan, T. Sudyodsuk, V. Promarak, Y. Surakhot and S. Jungsuttiwong, Dalton Trans., 2014, 9166–9176.
13. A. Mishra, M. K. R. Fischer and P. Bäuerle, Angew. Chem., Int. Ed., 2009, 48, 2474–2499.
14. M. Zhang, Y. Wang, M. Xu, W. Ma, R. Li and P. Wang, Energy Environ. Sci., 2013, 6, 2944–2949.
15. S.-Q. Fan, C. Kim, B. Fang, K.-X. Liao, G.-J. Yang, C.-J. Li, J.-J. Kim and J. Ko, J. Phys. Chem. C, 2011, 115, 7747–7754.
16. L. H. Nguyen, H. K. Mulmudi, D. Sabba, S. A. Kulkarni, S. K. Batabyal, K. Nonomura, M. Grätzel and S. G. Mhaisalkar, Phys. Chem. Chem. Phys., 2012, 14, 16182–16186.
17. J.-H. Yum, E. Baranoff, S. Wenger, M. K. Nazeeruddin and M. Grätzel, Energy Environ. Sci., 2011, 4, 842–857.
18. A. Yella, H.-W. Lee, H. N. Tsao, C. Yi, A. K. Chandiran, M. K. Nazeeruddin, E. W.-G. Diau, C.-Y. Yeh, S. M. Zakeeruddin and M. Grätzel, Science, 2011, 334, 629–634.
19. L. Han, A. Islam, H. Chen, C. Malapaka, B. Chiranjeevi, S. Zhang, X. Yang and M. Yanagida, Energy Environ. Sci., 2012, 5, 6057–6060.
20. G. Boschloo and A. Hagfeldt, Acc. Chem. Res., 2009, 42, 1819–1826.
21. P. Persson, M. J. Lundqvist, R. Ernstorfer, W. A. Goddard and F. Willig, J. Chem. Theory Comput., 2006, 2, 441–451.
22. M. J. Lundqvist, M. Nilsing, S. Lunell, B. Åkermark and P. Persson, J. Phys. Chem. B, 2006, 110, 20513–20525.
23. J. Wiberg, T. Marinado, D. P. Hagberg, L. Sun, A. Hagfeldt and B. Albinsson, J. Phys. Chem. C, 2009, 113, 3881–3886.
24. M. Pastore and F. De Angelis, Phys. Chem. Chem. Phys., 2012, 14, 920–928.
25. F. Odobel, E. Blart, M. Lagree, M. Villieras, H. Boujtita, N. El Murr, S. Caramori and C. Alberto Bignozzi, J. Mater. Chem., 2003, 13, 502–510.
26. A. Abbotto, N. Manfredi, C. Marinzi, F. De Angelis, E. Mosconi, J.-H. Yum, Z. Xianxi, M. K. Nazeeruddin and M. Grätzel, Energy Environ. Sci., 2009, 2, 1094–1101.
27. C. Anselmi, E. Mosconi, M. Pastore, E. Ronca and F. De Angelis, Phys. Chem. Chem. Phys., 2012, 14, 15963–15974.
28. M. Pastore and F. De Angelis, ACS Nano, 2010, 4, 556–562.
29. H. Tian, X. Yang, R. Chen, R. Zhang, A. Hagfeldt and L. Sun, J. Phys. Chem. C, 2008, 112, 11023–11033.
30. P. Chen, J. H. Yum, F. De Angelis, E. Mosconi, S. Fantacci, S.-J. Moon, R. H. Baker, J. Ko, M. K. Nazeeruddin and M. Grätzel, Nano Lett., 2009, 9, 2487–2492.
31. W. H. Howie, F. Claeyssens, H. Miura and L. M. Peter, J. Am. Chem. Soc., 2008, 130, 1367–1375.
32. M. K. Nazeeruddin, R. Humphry-Baker, P. Liska and M. Grätzel, J. Phys. Chem. B, 2003, 107, 8981–8987.
33. P. Falaras, Sol. Energy Mater. Sol. Cells, 1998, 53, 163–175.
34. C. Pérez León, L. Kador, B. Peng and M. Thelakkat, J. Phys. Chem. B, 2006, 110, 8723–8730.
35. T. Moehl, H. N. Tsao, K.-L. Wu, H.-C. Hsu, Y. Chi, E. Ronca, F. De Angelis, M. K. Nazeeruddin and M. Grätzel, Chem. Mater., 2013, 25, 4497–4502.
36. F. De Angelis, S. Fantacci, A. Selloni, M. Grätzel and M. K. Nazeeruddin, Nano Lett., 2007, 7, 3189–3195.
37. K. S. Finnie, J. R. Bartlett and J. L. Woolfrey, Langmuir, 1998, 14, 2744–2749.
38. F. Schiffmann, J. VandeVondele, J. r. Hutter, R. Wirz, A. Urakawa and A. Baiker, J. Phys. Chem. C, 2010, 114, 8398–8404.
39. A. Hinsch, J. M. Kroon, R. Kern, I. Uhlendorf, J. Holzbock, A. Meyer and J. Ferber, Progress in Photovoltaics: Research and Applications, 2001, 9, 425–438.
40. P. M. Sommeling, M. Späth, H. J. P. Smit, N. J. Bakker and J. M. Kroon, J. Photochem. Photobiol., A, 2004, 164, 137–144.
41. M. I. Asghar, K. Miettunen, J. Halme, P. Vahermaa, M. Toivola, K. Aitola and P. Lund, Energy Environ. Sci., 2010, 3, 418–426.
42. A. R. Andersen, J. Halme, T. Lund, M. I. Asghar, P. T. Nguyen, K. Miettunen, E. Kemppainen and O. Albrektsen, J. Phys. Chem. C, 2011, 115, 15598–15606.
43. H. T. Nguyen, H. M. Ta and T. Lund, Sol. Energy Mater. Sol. Cells, 2007, 91, 1934–1942.
44. P. T. Nguyen, R. Degn, H. T. Nguyen and T. Lund, Sol. Energy Mater. Sol. Cells, 2009, 93, 1939–1945.
45. E. Mosconi, A. Selloni and F. De Angelis, J. Phys. Chem. C, 2012, 116, 5932–5940.
46. A. F. Bartelt, R. Schütz, C. Strothkämper, I. Kastl, S. Janzen, D. Friedrich, W. Calvet, G. Fuhrmann, D. Danner, L.-P. Scheller, G. Nelles and R. Eichberger, Appl. Phys. Lett., 2014, 104, 223902–223903.
47. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. J. Montgomery, T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez and J. A. Pople, Gaussian, Inc., Wallinford CT, 2009.
48. A. D. Becke, J. Chem. Phys., 1993, 98, 1372–1377.
49. G. A. Petersson and M. A. Al-Laham, J. Chem. Phys., 1991, 94, 6081–6090.
50. M. Cossi and V. Barone, J. Chem. Phys., 2001, 115, 4708–4717.
51. M. J. Lundqvist, M. Nilsing, P. Persson and S. Lunell, Int. J. Quantum Chem., 2006, 106, 3214–3234.
52. F. De Angelis, A. Tilocca and A. Selloni, J. Am. Chem. Soc., 2004, 126, 15024–15025.
53. P. Persson, R. Bergström and S. Lunell, J. Phys. Chem. B, 2000, 104, 10348–10351.
54. A. Vittadini, A. Selloni, F. P. Rotzinger and M. Grätzel, Phys. Rev. Lett., 1998, 81, 2954–2957.
55. S. Fantacci, M. G. Lobello and F. De Angelis, Chimia, 2013, 67, 121–128.
56. A. R. Srimath Kandada, S. Fantacci, S. Guarnera, D. Polli, G. Lanzani, F. De Angelis and A. Petrozza, ACS Appl. Mater. Interfaces, 2013, 5, 4334–4339.
57. D. Kuang, C. Klein, S. Ito, J. E. Moser, R. Humphry-Baker, N. Evans, F. Duriaux, C. Grätzel, S. M. Zakeeruddin and M. Grätzel, Adv. Mater., 2007, 19, 1133–1137.
58. A. Fukui, R. Komiya, R. Yamanaka, A. Islam and L. Han, Sol. Energy Mater. Sol. Cells, 2006, 90, 649–658.
59. R. Katoh, M. Kasuya, A. Furube, N. Fuke, N. Koide and L. Han, Sol. Energy Mater. Sol. Cells, 2009, 93, 698–703.
60. Y. T. Azar and M. Payami, Phys. Chem. Chem. Phys., 2014, 16, 9499–9508.
61. R. Harikisun and H. Desilvestro, Sol. Energy, 2011, 85, 1179–1188.
62. D. Shi, N. Pootrakulchote, R. Li, J. Guo, Y. Wang, S. M. Zakeeruddin, M. Grätzel and P. Wang, J. Phys. Chem. C, 2008, 112, 17046–17050.
63. M. K. Nazeeruddin, P. Pèchy, T. Renouard, S. M. Zakeeruddin, R. Humphry-Baker, P. Comte, P. Liska, L. Cevey, E. Costa, V. Shklover, L. Spiccia, G. B. Deacon, C. A. Bignozzi and M. Grätzel, J. Am. Chem. Soc., 2001, 123, 1613–1624.
64. M. Pastore and F. De Angelis, J. Phys. Chem. Lett., 2012, 3, 2146–2153.
65. C. Bauer, G. Boschloo, E. Mukhtar and A. Hagfeldt, J. Phys. Chem. B, 2002, 106, 12693–12704.
66. J. Chen, F.-Q. Bai, J. Wang, L. Hao, Z.-F. Xie, Q.-J. Pan and H.-X. Zhang, Dyes Pigm., 2012, 94, 459–468.
67. K. Sodeyama, M. Sumita, C. O'Rourke, U. Terranova, A. Islam, L. Han, D. R. Bowler and Y. Tateyama, J. Phys. Chem. Lett., 2012, 3, 472–477.
68. S.-H. Liu, H. Fu, Y.-M. Cheng, K.-L. Wu, S.-T. Ho, Y. Chi and P.-T. Chou, J. Phys. Chem. C, 2012, 116, 16338–16345.
69. Y. Tateyama, M. Sumita, Y. Ootani, K. Aikawa, R. Jono, L. Han and K. Sodeyama, J. Phys. Chem. C, 2014, 118, 16863–16871.
70. F. De Angelis, G. Vitillaro, L. Kavan, M. K. Nazeeruddin and M. Grätzel, J. Phys. Chem. C, 2012, 116, 18124–18131.
71. H. Pettersson and T. Gruszecki, Sol. Energy Mater. Sol. Cells, 2001, 70, 203–212.
72. P. Wang, S. M. Zakeeruddin, J. E. Moser, M. K. Nazeeruddin, T. Sekiguchi and M. Gratzel, Nat. Mater., 2003, 2, 402–407.
73. M. Sumita, K. Sodeyama, L. Han and Y. Tateyama, J. Phys. Chem. C, 2011, 115, 19849–19855.
74. A. Abbotto, V. Leandri, N. Manfredi, F. De Angelis, M. Pastore, J.-H. Yum, M. K. Nazeeruddin and M. Grätzel, Eur. J. Org. Chem., 2011, 2011, 6195–6205.
75. X. Ren, S. Jiang, M. Cha, G. Zhou and Z.-S. Wang, Chem. Mater., 2012, 24, 3493–3499.

---

Footnote

† Electronic supplementary information (ESI) available: IPCE graphics, detailed list of geometrical parameters for dyes and solvents adsorbed on TiO₂, spin density plots for oxidized Y1 and N749 on TiO₂. See DOI: 10.1039/c4ra09472g

---