High boiling point solvent-based dye solar cells pass a harsh thermal ageing test

Solar Energy Materials & Solar Cells 144 (2016) 457–466

Contents lists available at ScienceDirect

Solar Energy Materials & Solar Cells journal homepage: www.elsevier.com/locate/solmat

High boiling point solvent-based dye solar cells pass a harsh thermal ageing test Thomas Stergiopoulos a,b,n, Athanassios G. Kontos b, Nancy Jiang c, Damion Milliken c, Hans Desilvestro c, Vlassis Likodimos b,d, Polycarpos Falaras b,nn a

Institute Department of Physics, University of Oxford, Clarendon Laboratory, Parks Road, Oxford OX13PU, UK Institute of Nanoscience and Nanotechnology, NCSR Demokritos, 15310 Agia Paraskevi Attikis, Athens, Greece c Dyesol Limited Company, Queanbeyan, NSW 2620, Australia d Solid State Physics Department, Faculty of Physics, University of Athens, Panepistimioupoli, 15784 Athens, Greece b

art ic l e i nf o

a b s t r a c t

Article history: Received 21 July 2015 Received in revised form 17 September 2015 Accepted 25 September 2015

Dye solar cells (DSCs) have emerged as one of the most efficient third-generation photovoltaic (PV) technologies, whose commercialization is mainly hampered by the lack of sufficient long-term stability compared to conventional PV devices. In this work, it is demonstrated that solvent based DSCs using tetraglyme as a non-nitrile, high boiling point, organic solvent for the iodide/triiodide redox shuttle, can pass a harsh accelerated thermal ageing test of 3000 h light soaking followed by additional 2000 h thermal ageing at 85 °C. Electrochemical and spectroscopic analysis on thermal degradation effects revealed that a conduction band edge shift towards more negative potentials for tetraglyme-DSCs underlies the enhanced photopotential of aged cells, compensating for the thermally induced photocurrent reduction due to slight triiodide loss. The tetraglyme-based solar cells (in contrast to cells based on methoxypropionitrile-MPN) showed exceptional stability, compatible with the established IEC61646 protocol for thin film PVs, keeping ca. 90% of their initial performance under 1 sun illumination. Quite notably, the cells even increased their initial efficiency by 4% when illuminated under 0.1 sun. This is the first time in literature that such a stability record is accomplished for solvent based DSCs utilizing commercially available and cost-effective materials. & 2015 Elsevier B.V. All rights reserved.

Keywords: Dye solar cells Ageing Thermal stability High boiling solvent Redox electrolyte

1. Introduction Although the sensitization of semiconductors by organic dyes dates back to the pioneer work of Prof. Heinz Gerischer in the late 1960s [1], it was only 20 years later when a breakthrough in the development of nanocrystalline solar cells was achieved by Brian O’Regan and Michael Grätzel using mesoporous, high surface area TiO2 film electrodes sensitized by a panchromatic dye in conjunction with the I
/I
3 redox couple [2]. Since then, the power conversion efficiency of dye solar cells (DSCs) consolidated at the level of 10–12% (with certified efficiencies of 11.9% under standard 1 sun AM1.5 G illumination conditions) [3] and quite surprisingly major advances came up (with efficiencies higher than 13%) again after about 20 years of stillness, when cobalt-based electrolytes [4] or organometal halide perovskite sensitizers [5,6] were employed. n Corresponding author at: Institute of Nanoscience and Nanotechnology, NCSR Demokritos, 15310, 15310 Agia Paraskevi Attikis, Athens, Greece. nn Corresponding author. E-mail addresses: [email protected] (T. Stergiopoulos), [email protected] (P. Falaras).

http://dx.doi.org/10.1016/j.solmat.2015.09.052 0927-0248/& 2015 Elsevier B.V. All rights reserved.

In addition to existing data confirming high power conversion efficiencies, the characterization of DSCs in terms of life time and stability lays the basis for further industrial developments and practical applications. In fact, despite the great perspectives for further efficiency enhancement in the very near future of either liquid DSCs based on the Co2 þ /Co3 þ redox mediator [7] or solidstate perovskite solar cells [8–10], long-term stability remains the key challenge for DSC’s commercialization. Although no formal figure of merit for qualifying the DSCs lifetime is currently available, 10% loss in efficiency is considered acceptable [11], perhaps up to 20% [12]. That can be evaluated by established accelerated ageing tests for photovoltaic (PV) devices, simulating a lifetime of 20 years outdoor operation [13]. It should be noted that stability has been scarcely addressed for the newly developed DSC systems [14–16], whereas, conventional DSCs employing the “old-fashioned” iodide/triiodide redox shuttle seem most efficient in terms of achieving long operating lifetime [17]. The stability of DSCs depends critically on the interplay of diverse degradation processes that may occur in each of the different cell components, i.e. dye (desorption, ligands substitution, isomerisation) [18,19], electrolyte (decomposition, solvent leakage,

458

T. Stergiopoulos et al. / Solar Energy Materials & Solar Cells 144 (2016) 457–466

additives evaporation, triiodide loss) [20,21], cathode (desorption of electrocatalyst, inactivation due to poisoning effects) [22,23] and current collecting grids (corrosion) [24]. The occurrence or not of these detrimental reactions on stressed DSCs depends on the ageing conditions and certainly the underlying degradation effects can be accelerated in the presence of UV light as well as the ingress of O2 and/or H2O into the cell due to imperfect sealing [11]; more details on the existing degradation mechanisms identified up to now can be found in our recent paper [25]. All the degradation routes lead to significant performance losses due to severe reduction of distinct electrical parameters (photopotential, photocurrent and fill factor), or a combination of them, defining the overall conversion efficiency. Even though certified ageing tests exist for the PV industry according to international standards such as IEC 61646 [26], currently, there are no corresponding established tests for DSCs. However, three stress tests can be applied for the evaluation of DSC stability, i.e. a) light soaking for 1000 h at  60 °C under 0.8– 1 sun illumination, b) thermal ageing at temperatures between 55 and 95 °C for 1000 h in the dark and c) thermal cycling from
40 °C to þ85 °C [27]. It should be noted that the latter thermal cycle test is not common for DSCs; the authors are only aware of two literature reports, which however showed excellent stability for both solvent and ionic liquid based DSCs, indicating the robustness of the cell chemistry and seals to temperature extremes [28,29]. Thus, prolonged light soaking and/or thermal ageing in the dark have been the most prominent stress tests thus far used for the assessment of DSC long term lifetime/stability. In particular, DSCs employing various robust dyes combined with non-volatile electrolytes readily sustained their performance after 1000 h continuous solar light (1 sun) illumination [30,31]. Extensive tests, carried out at Dyesol, demonstrated an impressive stability of DSCs under even more prolonged light soaking; i.e. over 6450 h [32] or even 25,600 h of continuous light soaking at 55–60 °C under resistive load [29]. On the other hand, long-term thermal stability of DSCs remains a highly challenging task at elevated temperatures (480 °C), where thermal stress can accelerate degradation reactions or even create new degradation pathways [33]. Promising thermal stability has been reported at 85 °C for DSCs employing ionic liquid based [34] or gelled electrolytes [35]. Ionic liquid electrolytes, employed in industrial DSCs at Dyesol and stored in the dark under open-circuit conditions for over 1000 h at 80 °C, resulted in a 22% decrease of cell performance [29]. However, ionic liquids are quite expensive (when compared with usual molecular solvents) and rather difficult to synthesize and purify [35], compromising to some extent the inherent advantage of DSC as a low cost PV technology. Significant stability under thermal stress has been also observed in DSCs utilizing electrolytes with typical low-cost organic solvents (such as methoxypropionitrile, MPN) [36], especially when the cells were frit sealed [37]. However, considerable deterioration of the cell performance was evidenced in many cases [18,20], implying that the integrity of the MPN-based solar cells under ageing seems more susceptible to the quality of the sealing. For instance, a 31% performance loss was observed in MPN-cells upon thermal stress at 80 °C for 1000 h [29], while much higher degradation (more than 70% power conversion efficiency loss) was reported for similar DSCs subjected to ageing at 80 °C for 2000 h in the dark (after a successful light soaking test under 0.8 sun for 2000 h) mainly due to substantial triiodide loss [38]. To mitigate the ensuing DSC degradation, a non-nitrile high boiling point organic solvent, tetraglyme (TGL) (Fig. S1 and Table S1), was originally applied to dissolve the iodide/triiodide redox shuttle, resulting in significantly improved durability of the cells at high temperatures; the efficiency of the TGL-cells was reduced by only 20% after a harsh ageing test of 2000 (light) þ2000 (dark at

80 °C) hours [38]. In this work, we extended the life-time ageing tests to a total of 5000 h (3000 h light soaking followed by 2000 h thermal ageing), at a temperature as high as 85 °C (such temperatures could be frequently met in tropical or desert locations) [27]. This is in agreement with current IEC protocols for thin film PVs [26] that adopt testing at 85 °C as the most reasonable temperature for accelerated lifetime experiments. By improving the sealing of the devices and protecting the cells from electrolyte leakage and/or triiodide loss, we achieved a nearly 90% stability of the initial performance at 1 sun, meeting the stability standards of conventional PVs [11]. Quite notably, the efficiency of TGL-cells increased by 4% (in comparison with their initial performance) after thermal ageing when the solar cells were illuminated at 0.1 sun. This could be a significant advantage for DSCs' performance in hot climates, as recently demonstrated in systematic studies of their long term stability under high temperatures and low irradiance levels pertinent at real outdoor conditions [39]. A combination of experimental techniques including linear sweep voltammetry, UV–vis absorption, electrochemical impedance, and micro-Raman spectroscopy were applied to identify the underlying degradation mechanisms, unveiling that a favorable conduction band edge shift for the TGL-cells effectively compensated for the diffusion limitations of the cell photocurrent induced by a moderate triiodide loss at high temperatures. This is the first time that such a harsh thermal test has been successfully passed by DSCs employing low-cost conventional organic solvents, which are readily accessible for evaluation to the whole research community of dye solar cells.

2. Materials and methods 2.1. Solar cell devices assembly DSCs of 0.88 cm2 active area were manufactured by Dyesol using standard Dyesol materials. The photoelectrode consisted of DSL 18NR-AO titania paste printed on TEC-15 glass, with both TiCl4 underlayer and overlayer, further sensitized with the N719 ruthenium dye. The cathode was composed of PT1 platinum paste, also printed on conductive glass [32]. I
/I
3 -based electrolytes were prepared with 1-propyl-3-methylimidazolium iodide (PMII, 499%, Merck), iodine (I2, 99.8%, Aldrich), guanidinium thiocyanate (99.9%, Fluka) and an imidazole derivative stabiliser. The solvent was tetraglyme (TGL, 99%, Sigma-Aldrich), while methoxypropionitrile (MPN, 99%, Fluka) was used as a reference. The devices were assembled with thermoplastic primary and hermetic secondary seals. 2.2. Ageing conditions Cells were initially illuminated using high pressure sodium lamps corresponding to a light intensity of 40.8 sun and a cell temperature of close to 55 °C for 3000 h with resistors applied to maintain near MPP conditions. These cells, denoted as TGL-ref and MPN-ref cells, respectively, did not present any deterioration in their electrical parameters, having virtually the same efficiency as fresh cells. Then, the above cells were further thermally aged at 85 °C for an additional 2000 h in the dark under open-circuit conditions, denoted as TGL-aged and MPN-aged cells, respectively. 2.3. Characterization Current–voltage (I–V) measurements were performed by illuminating the DSCs from the photoelectrode side; a Xe lamp in combination with AM 1.5 G and 400 nm (UV) cut-off optical filters (Oriel) was employed in order to provide solar simulated light

T. Stergiopoulos et al. / Solar Energy Materials & Solar Cells 144 (2016) 457–466

(1 sun, 1000 W m
2). Light piping was avoided by the use of a large black metal mask with an aperture area of 0.15 cm2. To work with lower light intensities, neutral density filters (Oriel) were employed. I–V characteristics in the range of 0.1–1 sun were obtained using linear sweep voltammetry at a scan speed of 50 mV s
1 on an Autolab PGSTAT-30 potentiostat. Chronoamperometric measurements were carried out under successive on–off cycles of illumination on the cells in order to access their stability and reproducibility as well as ionic diffusion limitations. In addition, open circuit voltage decay (OCVD) measurements were performed. The cells were illuminated to a steady voltage and the illumination was interrupted with a shutter. Then, the decay profiles were recorded with a sampling interval of 20 ms. Voltammetry experiments were also conducted under 1 sun illumination at an extended cell voltage range from
1.5 to þ 1.5 V, including both forward and reverse bias conditions in order to study the behavior of limiting current with respect to possible changes of redox shuttle concentrations [40]. Electrochemical Impedance Spectroscopy measurements were carried out using a PGSTAT-30 potentiostat and its built-in frequency response analyzer under dark conditions and at applied voltages near maximum power point and Voc (i.e. in the potential range from
0.25 to
0.85 V with a potential step of 0.025 V). UV–visible diffuse reflectance spectra were recorded on a 3200 Hitachi spectrophotometer equipped with a 60 mm integrating sphere using BaSO4 as reference. Transmittance of light at 450 nm from the inactive edge regions of the cells, which are filled with electrolyte was recorded with a home-made setup (see inset of Fig. S4). It consists of a LED source, a condenser, a lens focusing the light at a spot of less than 0.5 mm in diameter and a power meter. Micro-Raman spectra were recorded on open-circuit cells, in the backscattering configuration using a Renishaw inVia Reflex system equipped with an Ar þ ion laser at 514.5 nm excitation. The scattered light was filtered by a dielectric edge Rayleigh rejection filter with cut-off at 100 cm
1 and analyzed with a 1800 lines mm
1 diffraction grating. A 2.5 mW laser beam was focused on the sample surface using a long working distance (8 mm) objective with magnification 50 of a Leica DMLM microscope. The nominal spot size of the laser beam is about 1 μm and the power density on the sample was controlled to 0.25 mW μm
2 with a filter. Spectral analysis was carried out by nonlinear least-squares fitting of the Raman peaks to a mixture of Lorentzian and Gaussian line shapes.

3. Results 3.1. Solar cells performance before and after ageing J–V curves under 1 sun illumination. Six well sealed solar cells (Fig. S2) were initially light soaked at 0.8–1 sun for 3000 h at maximum power point. These cells, denoted hereafter as TGL-ref (3 cells) and MPN-ref (3 cells), respectively, did not present any decline in their electrical parameters, retaining essentially the same efficiency as pristine solar cells. Subsequently, three DSCs from each type of electrolyte were thermally stressed at 85 °C for 2000 h in the dark at open-circuit voltage, denoted hereafter as TGL-aged and MPN-aged cells, respectively. The current density–voltage curves for both reference and aged cells were recorded under 1 sun AM1.5 G illumination, excluding UV light using a 400 nm cut-off filter. Representative J–V curves are shown in Fig.1, while the average solar cells parameters (open-circuit voltage Voc, short-circuit current density Jsc, fill factor FF, and power conversion efficiency η) are summarized in Table 1. The corresponding parameters for all the tested cells are included in Table S2, following the protocol recently established for J– V data presentation [41].

459

Fig. 1. Representative J–V curves obtained under 1 sun AM 1.5 G illumination for the TGL and MPN cells before and after thermal stress at 85 °C. Table 1 Solar cell parameters (mean values and standard errors from 3 solar cells for each solvent) of the reference and aged TGL and MPN-based DSCs, under 1 sun (AM 1.5 G) illumination. Electrolyte

Jsc (mA cm
2)

Voc (V)

FF

η (%)

TGL-ref TGL-aged MPN-ref MPN-aged

9.8 70.4 8.0 70.3 13.4 70.1 10.1 70.2

0.50 7 0.01 0.58 7 0.01 0.56 7 0.01 0.45 7 0.01

0.59 7 0.01 0.53 7 0.01 0.677 0.01 0.58 7 0.01

2.8 70.1 2.5 70.1 5.1 70.1 2.6 70.1

As-fabricated TGL based solar cells with an optimized configuration for stability purposes, gave an initial power conversion efficiency (η) of 2.8%, which after 5000 h of ageing (3000 h of light soaking plus 2000 h of thermal ageing at 85 °C) dropped (by a relative  10%) to 2.5%. For comparison, following a similar stress test, solar cells employing the standard MPN solvent lost 49% of their initial performance. This is the first time in literature (according to our knowledge) that DSCs employing a I
/I
3 -based liquid electrolyte and incorporating a low-cost conventional solvent, pass such a harsh thermal stress (ageing) test. J–V curves under 0.1 sun illumination. Remarkably, when recording the J–V curves under 0.1 sun illumination, the efficiency of the TGL-ref cells increased up to 4.8% (from 2.8% at 1 sun), implying significant limitations under 1 sun illumination (Table S3). However, the most striking observation was that the TGL-aged cells not only sustained their initial performance at 0.1 sun illumination but they even slightly increased their efficiency (by ca. 4%) after thermal ageing. On the contrary, MPN-cells kept losing their initial efficiency even when assessed under low light illumination conditions, displaying a relative loss of 45%. Despite the excellent thermal stability for DSCs employing a recently developed Dyesol proprietary high stability solvent for both a Ru2 þ based (Z907) and a pure organic (Y123) dye [25], this is the first time that 0.1 sun efficiency of a DSC system increased after ageing for 5000 h. 3.2. Optical characterization In spite of the marked thermal stability of TGL-based cells at high temperatures, physical insight to the underlying mechanism of (slight) loss of initial efficiency is essential to further improve the lifetime of DSCs. The overall performance of TGL-cells diminished by about 10%, mainly due to the decrease of Jsc by 18% and a concomitant reduction of FF by 10%, after thermal stress. On the

460

T. Stergiopoulos et al. / Solar Energy Materials & Solar Cells 144 (2016) 457–466

Fig. 2. Absorbance spectra of representative (a) TGL and (b) MPN reference and thermally aged DSCs.

of the I
3 absorbance tail at the 550 nm spectral range, which effectively increased light absorption by the sensitized photoelectrode. Triiodide losses were directly verified by transmittance measurements conducted by pointing a light beam on the cell edges filled with electrolyte only (Fig. S3). The relative increase of the transmittance at 450 nm (by the electrolyte only area) [42] allowed an estimate of triiodide loss of the order of 30% for both TGL and MPN cells. Triiodide loss was further confirmed for the aged cells by in-depth micro-Raman measurements that monitored the variation of the 112 cm
1 Raman mode of free I
3 ions when focusing the laser beam on the electrolyte and collecting the Raman signal through the counter electrode side (see below).

3.3. Diffusion limitations Fig. 3. Comparative linear sweep voltammograms for representative TGL and MPN cells under 0.1 sun illumination. Scan rate: 50 mV s
1.

other hand, Voc surprisingly increased by 14%, partially compensating for the Jsc and FF losses, in marked contrast to the MPN-cells (Table 1) as well as most DSCs employing Ru2 þ -based dyes, where high temperature ageing invariably leads to the reduction of both Voc and Jsc [25]. Electrolyte leakage and triiodide loss have been identified as the main factors causing the loss of DSC performance upon high temperature stress, especially for cells employing nitrile solvents like MPN [25,38]. Although no sign of electrolyte bleaching could be detected by visual inspection on both MPN and TGL-aged cells under an optical microscope [38], indicative of the improved sealing process, UV–vis spectroscopy indicated moderate triiodide loss for the thermally stressed cells. Fig. 2 shows the corresponding absorbance spectra obtained by transforming the diffuse reflectance data on the active areas of the different DSCs to Kubelka–Munk units. TGL based aged cells presented a weak blueshift of the absorption maximum by 5 nm. This effect has been attributed to slight dye modification, such as detachment of -SCN ligands [42]. Furthermore, a weak reduction of the absorbance could be identified below 450 nm for both TGL and MPN-cells, indicative of a decrease in the absorption from the triiodides in the electrolyte [43]. Diminution of triiodide optical absorption was also observed from the MPN based cells whereas, quite notably, the intensity of the dye absorption at the peak maximum slightly increased for the aged cells. This may be attributed to the decrease

Triiodide loss, even if not extensive, may severely affect the correct operation of the aged cells, since depletion of I
3 species could provoke serious limitations in mass transport through the electrolyte [38]. To explore this effect, linear sweep voltammetry measurements were conducted by sweeping the voltage from forward to reverse bias (Fig. 3) [44,45]. Current saturation under reverse bias was clearly observed for all cells due to diffusion limitations. MPN-ref and MPN-aged DSCs showed diffusionlimited current (JDL) values significantly higher than the corresponding Jsc ones, indicating no serious diffusion limitation by I
3. For the TGL-ref cells, JDL was comparable to Jsc, indicative of moderate diffusion limitations. On the other hand, in the case of TGL-aged cells, JDL was lower than Jsc, showing that diffusion limitation was a significant reason for the 1 sun photocurrent decrease. Quite notably, JDL decreased by about 50% for both types of cells after ageing, while a 30% triiodide loss was calculated by the transmittance measurements from the electrolyte only area of the cells (Fig. S4). The above discrepancy could be explained by the fact that some I
3 species become immobilized within the active DSC area and thus are not available for cell operation. Similar conclusions were drawn when recording the photocurrent (Jsc) vs. the power density of light (Pin) incident on the cells (Fig. S4). MPNcells did not exhibit any linearity problems in contrast with TGLbased cells, where pronounced deviations from linearity occurred, in the case of TGL-aged cells even under 0.5 sun illumination, in qualitative agreement with the voltammetry results in Fig. 2.

2

Current density (mA/cm )

T. Stergiopoulos et al. / Solar Energy Materials & Solar Cells 144 (2016) 457–466

16

MPN-ref -1 sun

MPN-aged- 1 sun

12 8 4 0 0

400

800

1200

1600

2000

2400

the light-on step at 400 s onwards) of both reference and aged TGL cells were very close to those of the fresh MPN cells, pointing to saturation of the photocurrents due to diffusion limitations at high sunlight and their superior performance under low light illumination conditions. A very similar behavior of the photocurrents under on–off illumination was observed for cells based on Co(II)–Co(III) polymer gel electrolytes in [47], which presented diffusion current limitations at high illumination conditions, too. The above achievements of very stable and reproducible DSCs with reasonably high efficiencies, provides confidence to accelerate development and scale-up of the preferred low-cost and easily produced DSC technology architecture. Based on this high boiling solvent (TGL) based electrolyte, work is now in progress to optimize the device performance.

2

Current density (mA/cm )

Time (s)

461

TGL-aged-1sun TGL-aged-0.1 sun

3.5. Electron transfer kinetics

8 4 0 0

400

800

1200

1600

2000

2400

2

Current density (mA/cm )

Time (s) 16 12 8 MPN-ref-1 sun TGL-ref-1 sun TGL-ref-0.1 sun

4

MPN-aged-1 sun TGL-aged-1 sun TGL-aged-0.1 sun

0 0

5

10

15

20

Time (s) Fig. 4. Photocurrent measurements under on–off illumination at 1 sun for (a) MPN-based cells, and (b) TGL-based cells at 1 and 0.1 sun. (c) Details of the transient photocurrent decay immediately after switching-on the illumination.

3.4. On–off stability performance In order to further investigate the reproducibility of fresh and aged devices, their photoresponse has been studied upon applying repeated on–off cycles of light illumination [46,47], as shown in Figs. 4a and b for MPN and TGL based cells, respectively. The results are in very good agreement with the corresponding J–V characteristics (Fig. 1 and Tables S2 and S3). The photocurrents reached the values recorded in the corresponding J–V curves and remained perfectly stable during successive on–off illumination cycles, pointing to an excellent stability and reproducibility of all cells. For the TGL based cells, the onset of 1 sun illumination resulted in significant photocurrent decay for the first  5 s followed by stabilization to 70% of their initial values. This drop can be quantitatively accounted for by the corresponding deviation of the TGL cells’ photocurrents at 1 sun illumination to that expected from a linear response of Jsc to the incident illumination intensity (see Fig. S4). This indicates considerable current diffusion limitations that impair the TGL cells photocurrent values at 1 sun (see Section 3.3), contrary to the behavior of the same cells at 0.1 sun and that of the MPN ones, where such transient decay is essentially suppressed. In fact, as shown in Fig. 4(c), the initial photocurrents (immediately after switching on 1 sun illumination-time counted after

In addition, open circuit voltage decay (OCVD) measurements were performed in order to study electron transfer kinetics. The decay profiles Voc vs. t were recorded and the reciprocal of the voltage derivative was taken and normalized to the thermal voltage. The electron lifetime was estimated in accordance to [48] by Eq. (1) and plotted in Fig. 5a and b, for illumination at 1 and 0.1 sun, correspondingly.  
1 k T dV oc τn ¼ B ð1Þ e dt The form and matching of the data in Fig. 5a and b, for different illumination conditions show that the lifetime presents an exponential dependence on Voc in a broad range. For the aged TGL based cells, the electron lifetime considerably increases in respect to the fresh ones, both under 1 and 0.1 sun illumination. Thus, the ageing of these cells suppresses the recombination pathway at the electrode electrolyte interface resulting in about 100 mV negative shift of the voltage attained for the same τn values. The behavior is totally different for the MPN based cells, where the electron lifetime is considerably smaller than the TGL- ones and further decreases by thermal ageing. These results corroborate well with the changes in the open circuit voltage of the cells (Table 1, S1 and S2). 3.6. Electron dynamics at the multi-interfaces Electrochemical Impedance Spectroscopy (EIS) was applied to investigate the corresponding stress effects on the cell’s electron 14 12 10

τn (s)

12

8 6 4 2

MPN-fresh-1 sun MPN-aged-1 sun TGL-fresh-1 sun TGL-aged- 1 sun

30 25 20

τn (s)

TGL-ref-1 sun TGL-ref-0.1 sun

16

15 10 5

-0.60

MPN-fresh- 0.1 sun MPN-aged-0.1 sun TGL-fresh 0.1 sun TGL-aged-0.1 sun -0.55

-0.50

-0.45

-0.40

-0.35

-0.30

-0.25

-0.20

-0.15

Voc (V) Fig. 5. Electron lifetime derived from OCVD curves for fresh and aged MPN and TGL based cells at (a) 1 and (b) 0.1 sun.

462

T. Stergiopoulos et al. / Solar Energy Materials & Solar Cells 144 (2016) 457–466

Fig. 6. (a) Diffusion resistance Rdif and (b) resistance at the counter electrode/electrolyte interface RPt/el as a function of the applied potential (Vappl) for representative TGL and MPN cells. Note that applied potentials are given with positive sign but they still refer to standard forward bias conditions.

dynamics (recombination and conduction band edge shifts). Impedance spectra, recorded in the dark, presented three standard depressed semicircles (spectra not shown) for all cells [49], which were effectively analyzed by a simplified form of the transmission line model proposed for DSCs [50]. Fig. 6a shows the triiodide diffusion resistance (Rdif) as a function of the applied potential (Vappl), determined from the impedance analysis. An increase of Rdif was thereby confirmed for both TGL and MPN-cells, reflecting the thermally induced triiodide loss and the concomitant diffusion limitation, most prominent for the TGL-aged cells (from about 27 to more than 57 Ω), due the more viscous character of the tetraglyme. Likewise, the interfacial resistance at the cathode/electrolyte interface (RPt/el) was also affected by the loss of triiodide (Fig. 6b); for the TGL cells, RPt/el increased from 6 to about 10 Ω upon ageing, whereas in the case of the MPN electrolyte the corresponding variation was rather weak. The increase of both Rdif and RPt/el upon ageing is at the origin of the FF decrease at higher light levels for both types of aged cells. The effects of ageing were most sensitively reflected by the chemical capacitance (Cm) of the TiO2 film. Fig. 7 displays the apparent capacitance (Cm) of the TiO2/electrolyte interface against the actual voltage drop at the photoelectrode (denoted as VF) [51]. The voltage VF was calculated by subtracting the voltage drop at the series resistance Vseries ¼IRseries from the applied potential across the cell, where Rseries is the sum of Rs, Rpt and Rdif [51]. Marked shifts of the Cm(VF) curves was thereby derived for the MPN and TGL-aged cells in opposite directions, implying significant shifts of the TiO2 conduction band edge and/or variations in the electron trap density and distribution induced in the photoelectrodes after thermal stress [52]. In particular, the TGL cells presented a shift by about 80 mV towards more negative potentials (more positive cell voltages), while those based on the MPN-electrolyte exhibited a shift by 110 mV towards more positive potentials upon ageing. These shifts correlate perfectly with the corresponding Voc variations under 1 sun illumination (Table 1), and the electron kinetics behavior in Fig. 6, disclosing that a major effect of high temperature thermal ageing, besides triiodide loss, is the generation of significant conduction band shifts (even though in opposite directions for the two types of DSC). To justify the observed band shifts and discriminate between semiconductor’s electronic properties and recombination dynamics, we plotted the recombination resistance against Vecb that is the voltage at the equivalent band position [49]. This analysis allowed comparing directly the recombination resistances of the cells, independently of the Cm
VF shifts. The resulting Rrec(Vecb) curves,

Fig. 7. Variation of the capacitance of the TiO2 photoelectrode against the actual potential of the photoelectrode (VF) for representative TGL and MPN-based cells.

shown in Fig. 8, prove that recombination was hardly affected upon ageing, particularly for the MPN cells (Fig. 8b). In the case of TGL-cells (Fig. 8a), recombination was likewise marginally reduced upon ageing. It was thus confirmed that the observed Voc variations for the thermally aged cells were mainly the result of conduction band shifts rather than changes in the recombination dynamics. In a very recent paper by Flasque et al., thermally aged DSCs employing MPN-based electrolytes were also found to present a drop of photovoltage due to downshifts in the TiO2 conduction band edge [53]. The authors attributed the degradation mechanism to the growth of a solid electrolyte interface layer that wraps the TiO2 nanoparticles, consisting of S, N and I atoms coming from decomposed nitrile solvent and redox species. In fact, trapping of iodide in the interfacial layer, whose thickness increased significantly upon thermal ageing at 85°, was also considered as the main source of iodide depletion justifying photocurrent drops of the degraded solar cells, too. 3.7. Spectroscopic (micro-Raman) monitoring of cells Micro-Raman spectroscopy was subsequently employed to investigate the diverse aging behavior of TGL and MPN cells [32,38]. Fig. 9a displays comparative micro-Raman spectra of the

T. Stergiopoulos et al. / Solar Energy Materials & Solar Cells 144 (2016) 457–466

463

Fig. 8. Recombination resistance (Rrec) as a function of Vecb for representative (a) TGL and (b) MPN-based cells, before and after thermal stress.

Fig. 9. (a) Micro-Raman spectra for representative TGL and MPN cells, obtained via the Pt counter electrode side before and after thermal stress, at low frequencies. (b) Relative integrated intensity ratio profiles of the 112 cm
1 triiodide mode relative to the 143 cm
1 anatase band for the different cells at open-circuit.

MPN and TGL cells at low frequencies (100–600 cm
1) by focusing the laser beam on the region fully occupied by the electrolyte accessing the cell from the Pt counter electrode side. A list of the main observed Raman modes together with their assignment is given in Table S4. At this frequency region, the DSC Raman spectra are dominated by the most intense TiO2 anatase Eg mode at 143 cm
1 and the symmetric stretching mode of I
3 ions at 112 cm
1 [54], whose relative intensity decreased after thermal

stress, reflecting loss of triiodide. This was systematically explored by in-depth micro-Raman experiments, conducted by focusing the laser beam inside the DSC from the Pt back side with a x50 long distance objective lens. Fig. 9b shows the integrated intensity ratio profiles of the 112 cm
1 triodide Raman mode relative to that of the 143 cm
1 one vs. the in-depth focusing distance z. Positive values of depth (z 40) correspond to the laser focusing deeper in the cell, thus towards the glass/TiO2 interface, while zo0, towards

464

T. Stergiopoulos et al. / Solar Energy Materials & Solar Cells 144 (2016) 457–466

the electrolyte side. Notice that the in-depth resolution is limited (i.e. the scattering volume is much larger than 1 mm3) due to multi-refractions and reflections of the laser beam at the air– glass–electrolyte–film interfaces. It was thus confirmed that the thermally aged cells presented a strong decline of the maximum relative intensity of the 112 cm
1 triiodide mode for both cells, though most pronounced for the MPN ones, in accordance with the triiodide losses detected by the optical measurements. Furthermore, ageing resulted in a distinct increase of the background luminescence signal, mainly for the MPN but also for the TGL based cells. This effect has been previously observed for DSCs subjected to prolonged outdoor ageing [55] and reverse bias stress [56], pointing to the generation of luminescent species within the electrolyte of the stressed DSCs. Accumulation of these species on the surface of the TiO2 photoelectrode may shift its conduction band edge, leading to the Voc variation after thermal stress. Charging of the photoelectrodes could be in principle accounted by the intercalation of electrolyte degradation products [57], whose formation may be promoted-in the presence of water and/or oxygen inside the cells, by precipitation due to electrolyte leakage or desorption from the electrodes [25]. In the case of MPN cells, the formation of a solid electrolyte interface layer may justify the positive shifts of the photoelectrode potential [50], contrary to TGL based cells, where negative charging takes place, indicative of the different nature of polar precipitates from the electrolyte. These contrasting effects on DSC photoelectrodes employing different electrolytes can be attributed to the different properties of the solvents, namely their boiling point, donor number and dielectric constant [58]. Extended Raman measurements were performed by focusing the laser beam at the dye/TiO2 interface of the DSCs before and after thermal stress [59]. Representative Raman spectra are shown in Fig. 10, after subtracting the broad luminescence background by a polynomial fitting interpolation routine. Background elimination permits direct comparison between fresh and aged cells in terms of the different peak frequencies. The Raman modes of the dye before and after thermal ageing showed only minor differences, implying that no significant degradation took place in the dye molecular configuration. However, in the case of TGL-aged cells, a marked increase of the 2170 cm
1 –C ¼N stretching Raman mode was observed [42] (encircled in the corresponding spectrum of Fig. 10), with comparable intensity to the corresponding modes of the thiocyanate (NCS
) dye ligands at 2100 and 2130 cm
1. The 2170 cm
1 band cannot be correlated with that of the guanidinium thiocyanate additive SCN
modes, which is present in the

Fig. 10. Extended micro-Raman spectra of representative TGL and MPN cells, obtained via the photoelectrode side, before and after thermal ageing. Spectra are corrected for the luminescence background. Arrows depict the 2130 cm
1 modes of the thiocyanate dye ligands, while circle highlights the 2170 cm
1 mode of -SCN groups detached from the dye in the TGL-aged cell.

electrolyte for both solvents. In fact, Raman spectra of GuSCN in powder form as well as in mixtures with electrolytes have shown signals solely in the 2070–2080 cm
1 range. The clear observation of this mode indicates the release of thiocyanate ligands from the N719 dye (probably exchanged by another ligand with higher binding constant) after thermal ageing, possibly forming I2 SCN species [42]. This assumption also justifies the corresponding shift in the absorption peak of the dye observed in Fig. 2(a) (further supporting that free SCN
is not coming from the GuNCS additive) and could be at the origin of the negative conduction band shift of TiO2 (since SCN
is an anionic ligand) for the aged TGL cells. Thiocyanates have been previously considered as the most sensitive part of Ru2 þ dyes like N719 and Z907, due to their relatively weak coordination with the central metal ion [60]. Despite the downward conduction band shift which promotes Voc, thiocyanate ligand loss could subsequently decrease dye regeneration efficiency for the TGL cells, since there is indication from ab-initio calculations that dye regeneration by I
ions takes place through SCN
ligands [61]. This could be an additional reason (besides diffusion limitations, identified in Fig. 3) for the reduction of Jsc upon ageing. It should be noted that electron injection could be also considered as a limiting step for the TGL-aged cells, since a 80 mV upward shift may substantially decrease the driving force for efficient injection of electrons from the N719 dye into TiO2 (the original driving force was estimated to be approximately 250 mV) [62].

4. Discussion Laboratory-scale dye solar cells were fabricated based on N719ruthenium dye and the iodide/triiodide redox shuttle with different electrolyte solvents. Replacing methoxypropionitrile (MPN) solvent with tetraglyme (TGL), which is a non-nitrile, high boiling point, organic solvent, an outstanding performance stability of about 90% under 1 sun was demonstrated for DSCs subjected to a light soaking stress of 3000 h at 55 °C followed by 2000 h harsh thermal ageing at 85 °C. TGL-aged cells even increased their initial efficiency by 4% under 0.1 sun illumination, in stark contrast to the MPN reference cells that typically lost half of their initial efficiency upon 5000 h ageing, independent of the illumination level. This is the first time in literature that such a stability record, compatible with international PV standards, is achieved for solvent based DSCs. Quite notably, this record is achieved with solar cells that utilize commercially available and relatively cost-effective materials, which can be readily tested and reproduced in other laboratories. The reason that thermal ageing does not impair significantly the efficiency of TGL-based cells is the increase of the open circuit voltage, which results from the enhanced electron lifetime and negative conduction band edge shift. On the other hand, the photocurrents are significantly reduced by diffusion limitations due to triiodide losses. However, such limitations are not an issue at 0.1 sun illumination, justifying the enhanced performance of the aged at such low illumination conditions commonly encountered in indoor applications. Even though an excellent result on the DSC endurance was accomplished, further stability improvements can be pursued by effectively mitigating triiodide losses (causing diffusion limitations) and SCN
ligand substitution (leading to dye regeneration limitations) as well as elucidating the electrolyte chemistry and degradation products that affect TiO2’s conduction band edge. Based on the results of this work, our research is now focused on better sealing materials and processes, as well as on novel tetraglyme like solvents with high dielectric constant [63], stability at temperatures as high as 85 °C and tolerance towards electrolyte leakage [25]. Thus, increased life time and stability are expected,

T. Stergiopoulos et al. / Solar Energy Materials & Solar Cells 144 (2016) 457–466

fact that will permit further developments and commercialization of the DSC technology.

Acknowledgments We gratefully acknowledge Dorothea Perganti for performing the on–off chronoamperometry and the open circuit voltage decay measurements. The research leading to these results has received funding from the European Union Seventh Framework Programme [FP7/2007–2013] under grant agreement 316494 (DESTINY) project. Additional financial support is also acknowledged by the European Social Fund and Greek national funds through ARISTEIA (AdMatDSC-1847) operational research programs.

Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.solmat.2015.09.052.

References [1] H. Gerischer, M.E. Michel-Beyerle, F. Rebentrost, H. Tributsch, Sensitization of charge injection into semiconductors with large band gap, Electrochim. Acta 13 (1968) 1509–1515. [2] B. O’Regan, M. Grätzel, A low-cost, high-efficiency solar cell based on dyesensitized colloidal TiO2 films, Nature 353 (1991) 737–740. [3] M.A. Green, K. Emery, Y. Hishikawa, W. Warta, E.D. Dunlop, Solar cell efficiency tables (version 41), Prog. Photovolt.: Res. Appl. 21 (2013) 1–11. [4] 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, M. Grätzel, Porphyrin-sensitized solar cells with cobalt (II/III)–based redox electrolyte exceed 12 percent efficiency, Science 334 (2011) 629–634. [5] J. Hyuck Heo, S. HyukIm, J. Hong Noh, T.N. Mandal, C.-S. Lim, J. Ah Chang, Y. Hui Lee, H.-J. Kim, A. Sarkar, Md. K. Nazeeruddin, M. Grätzel, S.II Seok, Efficient inorganic–organic hybrid heterojunction solar cells containing perovskite compound and polymeric hole conductors, Nat. Photonics 7 (2013) 486–491. [6] J. Burschka, N. Pellet, R. Humphry-Baker, P. Gao, M.K. Nazeeruddin, M. Grätzel, Sequential deposition as a route to high-performance perovskite-sensitized solar cells, Nature 499 (2013) 316–319. [7] J.-H. Yum, E. Baranoff, F. Kessler, T. Moehl, S. Ahmad, T. Bessho, A. Marchioro, E. Ghadiri, J.-E. Moser, C. Yi, Md. K. Nazeeruddin, M. Grätzel, A cobalt complex redox shuttle for dye-sensitized solar cells with high open-circuit potentials, Nat. Commun. 3 (2012) 631, http://dx.doi.org/10.1038/ncomms1655. [8] H.–S. Kim, C.–R. Lee, J.–H. Im, K.–B. Lee, T. Moehl, A. Marchioro, S.–J. Moon, R. Humphry-Baker, J.–H. Yum, J.E. Moser, M. Grätzel, N.–G. Park, Lead iodide perovskite sensitized all-solid-state submicron thin film mesoscopic solar cell with efficiency exceeding 9%, Sci. Rep. 2 (2012) 591, http://dx.doi.org/10.1038/ srep00591. [9] M.M. Lee, J. Teuscher, T. Miyasaka, T.N. Murakami, H.J. Snaith, Efficient hybrid solar cells based on meso-superstructured organometal halide perovskites, Science 338 (2012) 643–647. [10] M. Liu, M.B. Johnston, H.J. Snaith, Efficient planar heterojunction perovskite solar cells by vapour deposition, Nature 501 (2013) 395–398. [11] M.I. Asghar, K. Miettunen, J. Halme, P. Vhermaa, M. Toivola, K. Aitola, P. Lund, Review of stability for advanced dye solar cells, Energy Environ. Sci. 3 (2010) 418–426. [12] J. Baxter, Commercialization of dye sensitized solar cells: present status and future research needs to improve efficiency, stability, and manufacturing, J. Vac. Sci. Technol. A 30 (2012) 020801. [13] M. Grätzel, Photovoltaic performance and long-term stability of dye-sensitized mesoscopic solar cells, C. R. Chim. 9 (2006) 578–583. [14] W. Ziang, W. Huang, U. Bach, L. Spiccia, Stable high efficiency dye-sensitized solar cells based on a cobalt polymer gel electrolyte, Chem. Commun. 49 (2013) 8997–8999. [15] M. Grätzel, Nanostructured photosystems for the generation of electricity from sunlight, in: Presentation at E-MRS 2013 Spring Meeting, 2013. [16] T. Leijtens, G.E. Eperon, S. Pathak, A. Abate, M.M. Lee, H.J. Snaith, Overcoming ultraviolet light instability of sensitized TiO2 with meso-superstructured organometal tri-halide perovskite solar cells, Nat. Commun. 4 (2013) 2885, http://dx.doi.org/10.1038/ncomms3885. [17] A. Hagfeldt, G. Boschloo, L. Sun, L. Kloo, H. Pettersson, Dye-sensitized solar cells, Chem. Rev. 110 (2010) 6595–6663. [18] P.M. Sommeling, M. Späth, H.J.P. Smit, N.J. Bakker, J.M. Kroon, Long-term stability testing of dye-sensitized solar cells, J. Photochem. Photobiol. A: Chem. 164 (2004) 137–144.

465

[19] N. Heo, Y. Jun, J.H. Park, Dye molecules in electrolytes: new approach for suppression of dye-desorption in dye-sensitized solar cells, Sci. Rep. 3 (2013) 1712, http://dx.doi.org/10.1038/srep01712. [20] D. Bari, N. Wrachien, R. Tagliaferro, S. Penna, T.M. Brown, A. Reale, A. Di Carlo, G. Meneghesso, A. Cester, Thermal stress effects on dye-sensitized solar cells (DSSCs), Microelectron. Reliab. 51 (2011) 1762–1766. [21] G.G. Xue, Y. Guo, T. Yu, J. Guan, X.R. Yu, J.Y. Zhang, J.G. Liu, Z.G. Zou, Degradation mechanisms investigation for long-term thermal stability of dyesensitized solar cells, Intern. J. Electrochem. Sci. 7 (2012) 1496–1511. [22] G. Syrrokostas, A. Siokou, G. Leftheriotis, P. Yianoulis, Degradation mechanisms of Pt counter electrodes for dye sensitized solar cells, Sol. Energy Mater. Sol. Cells 103 (2012) 119–127. [23] C.H. Lee, K.M. Lee, Y.L. Tung, J.M. Wu, Degradation analysis of thermal aged back-illuminated dye-sensitized solar cells, J. Electrochem. Soc. 159 (2012) B430–B433. [24] H. Matsui, K. Okada, T. Kitamura, N. Tanabe, Thermal stability of dye-sensitized solar cells with current collecting grid, Sol. Energy Mater. Sol. Cells 93 (2009) 1110–1115. [25] N. Jiang, T. Sumitomo, T. Lee, A. Pellaroque, O. Bellon, D. Milliken, H. Desilvestro, High temperature stability of dye solar cells, Sol. Energy Mater. Sol. Cells 119 (2013) 36–50. [26] IEC 61646, Thin-Film Terrestrial Photovoltaic (PV) Modules-Design Qualification and Type Approval, 2008. [27] H. Desilvestro, M. Bertoz, S. Tulloch, G. Tulloch, Packaging, scale-up and commercialisation of dye solar cells, in: K. Kalyanasundaram (Ed.), DyeSensitized Solar Cells, EPFL Press, Lausanne, Switzerland, 2010. [28] H. Arakawa, T. Yamaguchi, K. Okada, H. Matsui, T. Kitamura, N. Tanabe, Highly durable dye-sensitized solar cells, Fujikura Tech. Rev. 38 (2009) 55–60. [29] R. Harikisun, H. Desilvestro, Long-term stability of dye solar cells, Sol. Energy 85 (2011) 1179–1188. [30] A. Mishra, N. Pootrakulchote, M. Wang, S.J. Moon, S.M. Zakeeruddin, M. Grätzel, P. Bäuerle, A thiophene-based anchoring ligand and its heteroleptic Ru(II)-complex for efficient thin-film dye-sensitized solar cells, Adv. Funct. Mater. 21 (2011) 963–970. [31] D. Costenaro, C. Bisio, F. Carniato, G. Gatti, F. Oswald, T.B. Meyer, L. Marchese, Size effect of synthetic saponite-clay in quasi-solid electrolyte for dyesensititized solar cells, Sol. Energy Mater. Sol. Cells 117 (2013) 9–14. [32] V. Likodimos, T. Stergiopoulos, P. Falaras, R. Harikisun, J. Desilvestro, G. Tulloch, Prolonged light and thermal stress effects on industrial dyesensitized solar cells: a micro-Raman investigation on the long-term stability of aged cells, J. Phys. Chem. C 113 (2009) 9412–9422. [33] J.-H. Yum, R. Humphry-Baker, S.M. Zakeeruddin, M.K. Nazeeruddin, M. Grätzel, Heteroleptic ruthenium complex containing substituted triphenylamine holetransport unit as sensitizer for stable dye-sensitized solar cell, Nano Today 5 (2010) 91–98. [34] W. Kubo, T. Kitamura, K. Hanabusa, Y. Wada, S. Yanagida, Quasi-solid-state dye-sensitized solar cells using room temperature molten salts and a low molecular weight gelator, Chem. Commun. (4) (2002) 374–375. [35] B. Clare, A.S. Wardana, D.R. MacFarlane, Synthesis, purification and characterization of ionic liquids, Top. Curr. Chem. 290 (2009) 1–40. [36] J. Goldstein, I. Yakupov, B. Breen, Development of large area photovoltaic dye cells at 3G solar, Sol. Energy Mater. Sol. Cells 94 (2010) 638–641. [37] A. Hinsch, W. Veurman, H. Brandt, R. Loayza Aguirre, K. Bialecka, K. Flarup Jensen, Worldwide first fully up-scaled fabrication of 60  100 cm2 dye solar module prototypes, Prog. Photovolt. Res. Appl. 20 (2012) 698–710. [38] A.G. Kontos, T. Stergiopoulos, V. Likodimos, D. Milliken, J. Desilvestro, G. Tulloch, P. Falaras, Long-term thermal stability of liquid dye solar cells, J. Phys. Chem. C 117 (2013) 8636–8646. [39] A. Asghar, M. Emziane, H.K. Pak, S.Y. Oh, Sol. Energy Mater. Sol. Cells 128 (2014) 335–342. [40] H.J. Snaith, The perils of solar cell efficiency measurements, Nat. Photonics 6 (2012) 337–340. [41] H.A. Greijer, J. Lindgren, A. Hagfeldt, Resonance Raman scattering of a dyesensitized solar cell: mechanism of thiocyanato ligand exchange, J. Phys. Chem. B 105 (2001) 6314–6320. [42] G. Rowley, B.H. Farnum, S. Ardo, G.J. Meyer, Iodide chemistry in dye-sensitized solar cells: making and breaking I
I bonds for solar energy conversion, J. Phys. Chem. Lett. 1 (2010) 3132–3140. [43] S. Mastroianni, A. Lanuti, T.M. Brown, R. Argazzi, S. Caramori, A. Reale, A. Di Carlo, Reverse bias degradation in dye solar cells, Appl. Phys. Lett. 101 (2012) 123302. [44] A. Listorti, C. Creager, P. Sommeling, J. Kroon, E. Palomares, A. Fornelli, B. Breen, P.R.F. Barnes, J.R. Durrant, C. Law, B. O’Regan, The mechanism behind the beneficial effect of light soaking on injection efficiency and photocurrent in dye sensitized solar cells, Energy Environ. Sci. 4 (2011) 2501–3494. [45] M.I. Asghar, K. Miettunen, S. Mastroianni, J. Halme, H. Vahlman, P. Lund, In situ image processing method to investigate performance and stability of dye solar cells, Sol. Energy 86 (2012) 331–338. [46] (a) K. Prabakaran, A.K. Palai, S. Mohanty, S.K. Nayak, Aligned carbon nanotube/polymer hybrid electrolytes for high performance dye sensitized solar cell applications, RSC Adv. 5 (2015) 66563–66574; (b) K.-S. Ahn, Y. Yan, S.-H. Lee, T. Deutsch, J. Turner, C.E. Tracy, C.L. Perkins, M. Al-Jassim, Photoelectrochemical properties of N-incorporated ZnO films deposited by reactive RF magnetron sputtering, J. Electrochem. Soc. 154 (2007) B956–B959. [47] W. Xiang, W. Huang, U. Bach, L. Spiccia, Stable high efficiency dye-sensitized solar cells based on a cobalt polymer gel electrolyte, Chem. Commun. 49 (2013) 8997–8999.

466

T. Stergiopoulos et al. / Solar Energy Materials & Solar Cells 144 (2016) 457–466

[48] (a) H. Wang, B. Wang, J. Yu, Y. Hu, C. Xia, J. Zhang, R. Liu, Significant enhancement of power conversion efficiency for dye sensitized solar cell using 1D/3D network nanostructures as photoanodes, Sci. Rep. 5 (2015) 9305; (b) A. Zaban, M. Greenshtein, J. Bisquert, Determination of the electron lifetime in nanocrystalline dye solar cells by open-circuit voltage decay measurements, Chem. Phys. Chem. 4 (2003) 859–864. [49] F. Fabregat-Santiago, G. Garcia-Belmonte, I. Mora-Seró, J. Bisquert, Characterization of nanostructured hybrid and organic solar cells by impedance spectroscopy, Phys. Chem. Chem. Phys. 13 (2011) 9083–9118. [50] F. Fabregat-Santiago, J. Bisquert, G. Garcia-Belmonte, G. Boschloo, A. Hagfeldt, Influence of electrolyte in transport and recombination in dye-sensitized solar cells studied by impedance spectroscopy, Sol. Energy Mater. Sol. Cells 87 (2005) 117–131. [51] F. Fabregat-Santiago, J. Bisquert, E. Palomares, L. Otero, D. Kuang, S. M. Zakeeruddin, M. Grätzel, Correlation between photovoltaic performance and impedance spectroscopy of dye-sensitized solar cells based on ionic liquids, J. Phys. Chem. C 111 (2007) 6550–6560. [52] M. Konstantakou, T. Stergiopoulos, V. Likodimos, G.C. Vougioukalakis, L. Sygellou, A.G. Kontos, A. Tserepi, P. Falaras, Influence of fluorine plasma treatment of TiO2 Films on the behavior of dye solar cells employing the Co (II)/(III) redox couple, J. Phys. Chem. C 118 (2014) 16760–16775. [53] M. Flasque, A.N. Van Nhien, J. Swiatowska, A. Seyeux, C. Davoisne, F. Sauvage, Interface stability of a TiO2/3-methoxypropionitrile-based electrolyte: first evidence for solid electrolyte interphase formation and implications, ChemPhysChem 15 (2014) 1126–1137. [54] A.G. Kontos, T. Stergiopoulos, G. Tsiminis, Y.S. Raptis, P. Falaras, In situ microand macro-Raman investigations of the redox couple behaviour in DSSCs, Inorg. Chim. Acta 361 (2008) 761–768. [55] N. Kato, Y. Takeda, K. Higuchi, A. Takeichi, E. Sudo, H. Tanaka, T. Motohiro, T. Sano, T. Toyoda, Degradation analysis of dye-sensitized solar cell module

[56]

[57]

[58]

[59]

[60]

[61] [62] [63]

after long-term stability test under outdoor working conditions, Sol. Energy Mater. Sol. Cells 93 (2009) 893–897. A. Quatela, A. Agresti, S. Mastroianni, S. Pescetelli, T.M. Brown, A. Reale, A. Di Carlo, Fabrication and reliability of dye solar cells: a resonance Raman scattering study, Microelectron. Reliab. 52 (2012) 2487–2489. B. Macht, M. Turrión, A. Barkschat, P. Salvador, K. Ellmer, H. Tributsch, Patterns of efficiency and degradation in dye sensitization solar cells measured with imaging techniques, Sol. Energy Mater. Sol. Cells 73 (2002) 163–173. E. Mosconi, A. Selloni, F. De Angelis, Solvent effects on the adsorption geometry and electronic structure of dye-sensitized TiO2: a first-principles investigation, J. Phys. Chem. C 116 (2012) 5932–5940. T. Stergiopoulos, M.C. Bernard, A. Hugot-Le Goff, P. Falaras, Resonance microRaman spectrophotoelectrochemistry on nanocrystalline TiO2 thin film electrodes sensitized by Ru(II) complexes, Coord. Chem. Rev. 248 (2004) 1407–1420. P.T. Nguyen, A.R. Andersen, E.M. Skoud, T. Lund, Dye stability and performances of dye-sensitized solar cells with different nitrogen additives at elevated temperatures – can sterically hindered pyridines prevent dye degradation? Sol. Energy Mater. Sol. Cells 94 (2010) 1582–1590. A. Hagfeldt, M. Grätzel, Molecular photovoltaics, Acc. Chem. Res. 33 (2000) 269–277. G. Boschloo, A. Hagfeldt, Characteristics of the iodide/triiodide redox mediator in dye-sensitized solar cells, Acc. Chem. Res. 42 (2009) 1819–1826. T. Stergiopoulos, A.G. Kontos, V. Likodimos, D. Perganti, P. Falaras, Solvent effects at the photoelectrode/electrolyte interface of a DSC: A combined spectroscopic and photoelectrochemical study, J. Phys. Chem. C 115 (2011) 10236–10244.