Sensitizer Activated Nanostructured Solar Cells

Final Report Summary - SANS (Sensitizer Activated Nanostructured Solar Cells)

Executive Summary:
This project aimed to deliver a new generation of nanostructured molecular photovoltaics using biomimetic principles, leading to sensitizer activated solar devices that produce reliable and affordable solar electricity (under all light conditions) together with long term stable efficiency and low embodied energy – a credible solution for a low carbon society. These sensitizer activated solar cells (SANS) are rooted in the Dye Solar Cell (DSC) science invented by the EPFL partner, and the objectives of this work was to take that core technology to a new dimension. The solar cell development addressed the three critical sub-systems incorporated in the device in a coherent manner: the light absorbing antenna, the electron-transporting mesostructured oxide and the electrolyte/hole-transporter sub-system, all tied together with the advancement of knowledge. The SANS team has been active in thin film sensitizer based solar cell science for over 20 years and are currently leading the field, illustrated by over 450 publications between them in this area over the last 4 years. Due to the breadth of expertise of the consortium our approach has addressed the entire photovoltaic system, in an ambitious yet realistic and achievable project.
The primary objective was to gain a predictable jump in long-term power output of sensitizer activated solar cells through novel multifunctional nanomaterials and molecular architectures utilizing materials design, development and processing. The overriding goal of the project was to realise a verified 14% efficient lab based cell, and a module with a lifetime equivalent to 20 years European outdoor usage.

Specific highlights include the realisation of highly stable solvent free electrolytes for dye sensitized solar cells, with the solar cells approaching the stability target of 20,000hrs continuous exposure to simulated sun light. The realisation of a new family of multi-chromophoric dyes which deliver good operation (~ 10% efficiency) with a blue coloured dye-sensitized solar cell, ideally suited for building integrated solar cell applications. For the mesoporus electrodes , we have developed a new family of mesoporous materials, namely mesoporous single crystals, which promise to offer much enhanced properties for dye-sensitzied solar cells, and potential broader applications in energy storage devices. Finally, throughout the last 18 months of the project the partners discovered remarkably efficient operation from perovskite absorbers employed in mesoscopic solar cells. Perovskites were not foreseen nor described within the description of work. However, we responded dynamically and substituted one of the previous high risk tasks, “Photoluminescence up conversion” with a task to incorporate the perovskite absorbers into the SANS workplan. This then subsequently delivered solid-state sensitized solar cells with over 14% efficiency, which was the project end efficiency target efficiency. All these achievements were possible due detailed probing and understanding of the fundamental operating principles and the information gained was fed back into the materials and device development work packages, enabling the directed advances. Throughout the project over 80 scientific publications were generated, a number of patents filed and a scientific workshop and conference successfully organised and executed. Commercially, Dyesol is progressing with commercialisation of sensitized solar cells and a new technology company, Oxford PV ltd spun-out from Oxford University, is scaling up the processing of solid-state perovskite solar cells. In addition, Dynemo, a spin-out from Uppsala, is selling sensitized solar cell materials to the research community.

Project Context and Objectives:
The SANS project has had a hugely successful with major advances in all aspects of the work. The team have pulled together to create the most significant advances in sensitized solar cell technologies observed for many years. The SANS project is coherently advancing all the critical technical components of sensitized solar cells, which include new stronger absorbing dyes and quantum dot absorbers, novel advanced mesoporous electrodes, more efficient, more easily processed and more stable electrolytes and hole-transporters. All the development is performed with the intention to rapidly transfer the advancements to industry, in a manufacturable platform. Specific highlights include the realisation of highly stable solvent free electrolytes for dye sensitized solar cells, with the solar cells approaching the stability target of 20,000hrs continuous exposure to simulated sun light. The realisation of a new family of multi-chromophoric dyes which deliver good operation (~ 10% efficiency) with a blue coloured dye-sensitized solar cell, ideally suited for building integrated solar cell applications. For the mesoporus electrodes, we have developed a new family of mesoporous materials, namely mesoporous single crystals, which promise to offer much enhanced properties for dye-sensitzied solar cells, and potential broader applications in energy storage devices. Finally, throughout the last 18 months of the project the partners discovered remarkably efficient operation from perovskite absorbers employed in mesoscopic solar cells. Perovskites were not foreseen nor described within the description of work. However, we responded dynamically and substituted one of the previous high risk tasks, “Photoluminescence up conversion” with a task to incorporate the perovskite absorbers into the SANS workplan, and this delivered solid-state sensitized solar cells with over 14% efficiency, which was the project end target efficiency.

The three critical sub-systems were addressed through six key science and technical objectives:
1. Creation of a new generation of highly absorptive, efficient and stable molecular sensitizers.
Utilising theoretical modelling feedback design and then demonstrate innovative molecular and nano-scaled antennae materials with unprecedented performance and stability. Confer highly efficient and spectrally broad light harvesting and supramolecular hole transport properties to the ruthenium based sensitizers by appropriate molecular engineering of their ligands.
2. Exploration of revolutionary new concepts of photon capture by nanocrystals, including near IR light capture Drawing on results of existing studies, the objective was to evaluate and determine the credible benefits of quantum dot technology. The known purpose of the quantum dot-based antenna is to provide efficient capture of near IR light to expand the spectral range of the device. Throughout the project the partners discovered extremely efficient operation of solar cells employing a new family of organometal halide perovskite absorbers, and this section was expanded to encompass those new materials in addition to quantum dots.
3. Implementation of novel and judiciously structured mesoscopic semiconducting metal oxide electrodes. Provide means for enhanced and vectorial electron transport in the mesoscopic semiconductor film combined with an effective ‘electrostatic shield’ for electron transport, through chemical control of the nanostructure, complemented by electronic bulk and surface modifications to the nanoparticulate metal-oxide. Improve the meso-morphology for optimized hole-transport and quasi-solid-state electrolyte infiltration, including incorporation of nanowire and gyroid morphologies.
4. Realization highly efficient and stable devices incorporating innovative non-volatile and quasi-solid-state electrolytes and organic hole transporters. Improve the conductivity of quasi-solid-state gel-electrolyte systems through use of supramolecular modified nanoparticulate materials in the electrolytes. Utilise the modified oxide morphology to study new solid-state hole conductors with energy levels better matched to the sensitizer redox potential than the standard iodide/triiodide couple, thereby improving the photovoltage. Improve the infiltration of hole-transporter and quasi-solid-state electrolyte into the mesoporous electrodes. Compare the transport and interfacial behaviour of the two electrolyte systems.
5. Develop a deep understanding of the physical and electrochemical processes occurring in cell operation. Pursue through combinatorial chemical and materials characterisation, photoelectrochemical, and electronic analytical techniques as well as computer simulation tools, interfaces modelling, electron transport and recombination studies to more completely understand the processes affecting performance and long term stability in SANS operation.
6. Demonstration of module scale stable photovoltaic operation of the new materials by realizing 20,000 hrs equivalent operation under AM1.5 simulated sun light. Achieve commercially viable longevity with the high performance systems via intensive and directed lifetime testing analysis and optimization of lab scale cells, while remaining compatible with large area module fabrication procedures.

Project Results:
The SANS project has been successful over its 36 month period, with major advances in all aspects of the work. The team have pulled together to create the most significant advances in sensitized solar cell technologies observed for many years. The SANS project has coherently advanced all the critical technical components of sensitized solar cells, which include new stronger absorbing dyes and quantum dot and perovskite absorbers, novel advanced mesoporous electrodes, more efficient, more easily processed and more stable electrolytes and hole-transporters. All the development has been performed with the intention to rapidly transfer the advancements to industry, in a manufacturable platform. Here, the key results are summarised following their work packages and tasks:

WP1 - Antenna System

Task 1.1 Ruthenium sensitizers (CNR-ISTM/EPFL):
Ruthenium sensitizers are traditionally the most efficient solar cells sensitizers, due to their good light harvesting (absorption onset at ca. 700 nm), coupled to optimal interfacial and stability properties. In search for high efficiency/high stability dyes, we have investigated two families of novel ruthenium dyes and complemented our study by computational analyses of the new dyes against existing sensitizers to provide insight into the reasons underlying the dye performance.
To maximize the DSSC efficiency we have investigated: a) new ruthenium dyes with an extended response into the red/near IR region, thus improving the short circuit phorocurrent (Jsc) output; and b) new ruthenium dyes with improved interfacial properties to achieve high open circuit voltage (Voc).
a) Synthesis, characterization, modeling and photovoltaic characteristics of trans-dithiocyanate Ru(II) dyes.
To shift the absorption onset towards the red/near IR region our strategy was to design and synthesize a highly pure quaterpyridine ligand (1) and the corresponding trans-Ru(II) complex (3) in high yield, thanks to a new and reliable microwave-assisted synthetic pathway. Along with the complex 3 complete characterization and DSSCs photovoltaic characterization, we carried out DFT/TDDFT computational analyses of 3 in solution and adsorbed onto TiO2 to clarify the effect of protonation of the terminal carboxylic groups on the electronic structure and optical properties and the properties of the dye-sensitized hetero interface, respectively.
The new complex shows an absorption onset at ca. 900 nm, thus substantially enhancing the solar cell light-harvesting capability. This translates into a Jsc values as high as 19.1 mA/cm2, which was further pushed to 19.8 mA/cm2 by co-sensitization with the organic D35 dye. The rather low Voc (0.55 V) penalizes the overall cell efficiency. This was interpreted as being possibly due to increased recombination between TiO2-injected electrons and the dye cation, in relation to the trans-NCS arrangement, as revealed by computational analyses.

b) Ruthenium dyes for high open circuit voltage DSSCs.
While the dye has a direct influence on Jsc, through its absorption spectrum, the dye effect on Voc is in principle only indirect. One among the various origins for the Voc modulation can be the electrostatic potential, related to the charge distribution and the dipole moment, of the dye itself. For positive dipoles, i.e. with the negative pole near the TiO2 surface, an upward shift of the conduction band can be achieved, while leaving the recombination behavior largely unchanged. Very recently, a series of new neutral thiocyanate-free sensitizers incorporating two 2-(4-(5-hexylthiophene))-pyridyl 4-(trifluoromethyl)-pyrazolate ancillaries and one 4,4’-dicarboxylic acid 2,2-bipyridyl ligand, coded TFRS-2, have shown Voc values of 810-830 mV in optimized DSSCs, delivering almost 10% efficiency. This observation is remarkable and represents the first case in which high Voc comparable to that of the prototypical N719, was obtained with a dye bearing a single 4,4’-dicarboxylic acid 2,2-bipyridyl ligand.
We have thus investigated the origin of the high Voc in the TFRS-2 by combining impedance spectroscopy and computational analyses. We investigated the novel homoleptic TFRS-2 dye against the standard heteroleptic C101 and Z907 dyes. Our results show that despite the three ruthenium dyes show a similar dye loeading, the TFRS-2 has Voc 20 to 40 mV higher than the other two dyes. This was interpreted as a dipole-induced TiO2 conduction band shift due to the different dipole direction in TFRS-2 (negative shift) and C101/Z907 (positive shift).

c) Computational analyses.
Computational analyses were intensively exploited in dye design (see above) and in providing the required feedback in establishing dye structure/functioning relations for optimal solar cell efficiency. As mentioned in a) and b) we have been able to substantially improve both Jsc and Voc, although in different dyes. Here we investigated the three major processes affecting charge generation in DSSCs, i.e. electron injection from the dye excited state to TiO2, recombination between TiO2-injected electrons and the oxidized dye and, indirectly, recombination between TiO2-injected electrons and the electrolyte, the latter two representing the major source of losses in DSSCs.
• Interfacial properties of the black dye.
Understanding how the interfacial properties of ruthenium dyes affect their performances in DSSC devices is important for further progress in the exploitation of this prototypical class of sensitizers. Here, we report on the modelling of the structural, electronic and optical properties of the N749 ruthenium sensitizer, also known as black dye, which has demonstrated as one of the best performing ruthenium dyes, with top efficiency exceeding 11%. We have shown that despite the low-lying excited state dye, electron injection in this system is quite efficient due to a “hot injection” mechanism which allows electron injection from higher lying dye excited states prior to relaxation in the lowest excited state.
• Recombination in C101-based DSSCs
One of the most successful heteroleptic Ru(II) sensitizers is the C101 dye, [Na Ru (4,4′- bis(5-hexylthiophene-2-yl) -2,2′-bipyridine) (4-carboxylic acid-4′-carboxylate-2,2′-bipyridine) (NCS)2], which was introduced in 2008 by EPFL. Thanks to its high molar extinction coefficient and to a corresponding improvement of charge-collection efficiency, it allowed demonstration of 11 % efficient solar cell. Very recently, the HIPC SANS partner presented an experimental study of the interaction of the top performing C101 dye with polycrystalline TiO2 anatase in (101) and (001) orientations. According to this work, the C101/anatase (001) exhibited the following differences compared to the C101/anatase (101): (i) enhancement of the DSC voltage, (ii) smaller surface concentration of the dye and (iii) about six-times slower recombination kinetics, traced by transient absorption spectroscopy of the back electron transfer in pure solvent.
Hence, the motivation for a computational modeling of the system was to acquire further knowledge about the C101/titania interaction, which would address these observations. Based on the calculated adsorption modes and electronic structure for the fully interacting C101 dye-sensitized (101)- and (001)-TiO2 models, we have modeled the recombination process between TiO2-injected electrons and the oxidized dye. We found that the same dye shows two different adsorption modes on the (101) and (001) TiO2 surfaces, leading to a different recombination behavior.
• Nature of the Electronic Trap States in TiO2 Nanocrystals
In DSCs, electron collection at the electrode competes with recombination at the TiO2/electrolyte interface, slow electron transport can limit the charge-collection efficiency and eventually lead to an overall diminished DSC’s conversion efficiency. The absence of a significant electric potential gradient in the semiconductor film, electron transport in mesoporous TiO2 is believed to proceed by diffusion. For nanostructured TiO2 films commonly employed in DSCs, diffusion coefficient values are orders of magnitude smaller than those observed for TiO2 single crystals, suggesting a high concentration of electron-trapping sites in the semiconductor film.
We have investigated the nature of trap states in realistic anatase TiO2 NCs of ca. 3 nm diameter, considering the effect of the adsorption of donor ligands and the interaction between two sintered TiO2 NCs across different interfaces. We found that trap states exist at the bottom of the conduction band even for perfectly crystalline TiO2 nanoparticles, which are localized within the central part of the NC, mainly at the intersection of the (100) and (101) surfaces. These states are pushed away by saturating (water) ligands. No additional trap states are found at the grain boundaries between two sintered nanocrystals. Our study basically tells us we cannot get entirely rid of trap states in TiO2 mesoporous films.

Task 1.2 Porphyrin Sensitizers (CNR-ISTM)
A large body of work has been devoted into the SANS project to the search of new photosensitizers to replace the Ru(II) metal complexes with more environmental friendly constituents. In this respect, porphyrins comprise a family of organic chromophores which are attractive for application in DSSCs owing to their high molar extinction coefficients, ease of modification, photochemical stability, low toxicity and potential low-cost. Porphyrin chromophores exhibit two absorption bands: the Soret- and the Q-bands, which occur in the blue (420 nm) and red (650 nm), respectively. A magnesium porphyrin acts as the main antenna system into the photosynthetic system II operating in plants, imparting the characteristic green color to plants. With suitable molecular control, the porphyrin bands can be shifted and broadened, potentially facilitating an increase in the total photon absorption and, therefore, the achievable photocurrent in DSSCs.
Several push-pull porphyrins have been designed, computationally screened and synthesized starting from the parent [5-(4’-carboxy-phenylethynyl)-15-(4”-N,N-dimethylamino-phenylethynyl)-10,20-bis(3,5-di-tert-butylphenyl) porphyrinate] Zn(II) compound. Syntheses of all the Zn(II)-porphyrinates investigated were carried out by the Sonogashira coupling, starting from the 5,15-diiodo or 5,15-dibromo Zn(II)-porphyrinates. In particular, symmetrical Zn(II) porphyrinates 1, 2, 3, 3’ and 5 were synthesized starting from the 5,15-diiodo Zn(II)-porphyrinate by a one-step reaction with the donor or the acceptor phenyl ethynyl moiety, whereas the push-pull Zn(II)-porphyrinate 9 was synthesized by a two-step method starting from the 5,15-dibromo Zn(II)-porphyrinate, by introducing first the donor and then the acceptor phenyl ethynyl moiety. Finally compound 4 was obtained by hydrolysis of 3 with LiOH in a THF/H2O mixture. By selective ligand substitution on both the donor and acceptor side we produced a new family of compounds by a novel synthetic route, which exhibited good light harvesting properties but did not deliver improved photovoltaic performances compared to the parent compound for which we obtained 4.5 % efficiency. In particular, by the wide class of investigated compounds, we showed how the difficulties of extending the dye light harvesting retaining the efficient push-pull concept.
An appealing strategy to overcome the synthetic challenges of typical meso-substituted porphyrins was explored, in which -pyrrolic-substituted porphyrins were synthesized, characterized and tested in DSSCs. An optimized synthetic route based on an microwave-enhanced Sonogashira coupling approach was implemented, which allowed us to obtain a variety of products in good yield. Computational analyses revealed a similar potential as dye sensitizers as meso-substituted porphyrins in terms of interfacial properties, although a reduced light harvesting was found for this class of less conjugated porphyrins. As a matter of fact, under our non optimized conditions, a slightly higher efficiency (4.7%) was achieved for the best -pyrrolic-substituted porphyrin against the best meso-substituted system, highlighting the potential of this class of synthetically simple dye sensitizers.
Further increase of the optical response was indeed achieved by extending the size of the conjugated core, i.e. with anthracene-fused porphyrins. For these systems the onset of absorption was substantially pushed to longer wavelengths, up to ca. 1000 nm, substantially enhancing the dyes light-harvesting capability. We were however not able to exploit the extended light absorption into increased photocurrent generation, mainly because to extend the light absorption a reverse excited state charge flow was introduced in the dye, which negatively affected the electron injection step.

Task 1.3 Quantum dots progress (IIT)
This task focused on the development of light antenna, based on colloidal quantum dots (QDs) with special unique architecture; a core semiconductor, covered by an epitaxial layer of another semiconductor, with optional alloying composition. The notation core/shell or core/alloyed-shell will be used in the forthcoming discussion. QDs are attracting large scientific and technological interest owed to their electronic and optical tunability with variation of size, shape and composition. Furthermore intriguing physical properties emanate from the size dependent, such as large oscillator strength, polarization, adaptation of carriers' lifetime or multiple exciton generation.

The surface of a colloidal QD plays the key role in almost each property of the QD, e.g. in light emission and solubility. There exist covalently unsaturated, ′′dangling′′, bonds on the QD surface, their energy levels invariably lying within the energetically forbidden gap of the bulk solid. These surface states act as traps for electrons or holes and impair the electrical and optical properties of the material. Therefore, it would be beneficial to provide capping of the QDs inside an inorganic shell, with organic molecules ligating only the outer surface. The shell material provides a sink for the surface layer defects, thereby pushing the surface electronic states out of the mid-gap region. Moreover, it opens the possibility of changing the exciton excited state dynamics and manipulating the wave functions, thus altering the optical properties of colloidal QDs.
The goals of the WP1.3 were developing effective synthesis procedures of PbSe and PbSe/PbS QDs, which should meet specific requirements for the solar cell applications. The QD conduction-band energy should exceed the conduction-band energy of titania, i.e. the QD band-gap energy should be 1.1-1.4 eV to create a large-enough driving force for efficient electron extraction. Moreover, the QDs should be stable in ambient conditions throughout the whole fabrication process of a PV device, including its final encapsulation, and sufficient amount of QDs should be available for the fabrication of solar cells.
As the first, preliminary stage, we have developed a high-yield synthesis of ultra-small core PbSe QDs 2-2.5 nm in diameter. The development of the synthesis included finding the appropriate temperature regime, precursors, and their optimal stoichiometric ratios for facilitating the reaction chemical yield at small QD sizes. Using the developed method we were able to prepare PbSe QDs with relatively narrow size distribution (~15%) and high chemical yields of 30-60%.
The appropriate coating procedure of a PbSe core with a PbS shell was developed with regard to specific properties of the core QDs; namely, we have found that the ultra-small PbSe QDs are sensitive to elevated temperatures. Therefore, it was necessary to find optimal temperature conditions of the coating reaction that would ensure, on the one hand, the core QD stability in the reaction solution and, on the other hand, perfect coating and high quality of the produced core/shell QDs. The optimal reaction temperature of 70 °C required the choice of a more reactive sulfur precursor than trioctylphosphine sulfur (TOPS) that was used previously. The new precursor, bis(trimethylsilyl) sulfur (TMS2S), stayed reactive enough at this temperature to ensure fast reaction and coating of large amounts (0.2-0.3 μmol) of PbSe QDs. Special precautions were taken to prevent the PbS co-nucleation during the coating reaction. With that end in view, a diluted solution of TMS2S was prepared and injected dropwise into the reaction mixture. It should be noted again that the benefits of the core/shell heterostructure include not only protecting PbSe against oxidation, but also passivating the dangling bonds in PbSe QDs and, in this way, eliminating the mid-gap states, which can act as electron trap-states. The prepared PbSe and PbSe/PbS QDs had a relatively high PL QY of ~60%.
Room-temperature and temperature-dependent steady-state photoluminescence investigation showed that PbSe/PbS QDs are much more stable towards oxidation than PbSe QDs. We found that the PL intensity of PbSe QDs quenched after ~10 minutes of air exposure, while that of PbSe/PbS QDs remained stable for at least 50 min of air exposure and then decreased gradually. This 50-minute interval is a “window of opportunity”, which gives enough time for PbSe/PbS QDs to be processed in a PV cell before the oxidation process begins.
The small PbSe/PbS QDs having a core/shell configuration were incorporated in solar cell having a heterojunction configuration. Preliminary results showed IPCE plots which are suggestive of the occurrence of carrier multiplication (i.e. multiple excitation generation) under excitation in the UV-Vis boundary region. This observation is also supported by the superlinear lifetime occurring under high intensity excitation. Also, the power conversion efficiency of these heterojunction solar cells reached 5% at 0.1 sun, getting close to the current breaking record in PbX QDs.

Comment on MS2.
In terms of MS2 – 18 months – Zn(II)-porphyrins “Realise effective NIR absorption in Zn(II)-porphyrin sensitizer with 900 nm absorption onset”, this was achieved by the anthracene-fused porphyrins. We have also demonstrated, however, that despite to enhanced light-absorption, it was difficult to extract the corresponding photocurrent.

Task 1.4 Dual mode harvesting (EPFL)
Under this task we have designed and developed a novel series of blue-coloured diketopyrrolopyrrole (DPP) based sensitizers for DSC applications. Traditional dye-sensitized solar cells have a significant opportunity in building integrated photovoltaics (BIPV) where differing levels of transparency and colour are demanded. The primary colours of choice are green or blue tinted glazing, or neutral density. For neutral density, there is a possibility of developing dyes which only absorb in the infrared and hence can be fully transparent to visible light, while still generating significant power in the NearIR region of the spectrum. Here, the successful development of very efficient (>10%) blue dye based DSSCs is a significant achievement (only 5% efficiency was previously possible for such a coloured solar cell. Additionally, the push towards the nearIR is another step closer to an all infrared absorbing solar cell.

WP2 - Metal Oxide Architectures

Task 2.1 Oxide films template by amphiphilic diblock and triblock copolymers and electronic doping (Oxford/UCAM/HIPC/Dyesol)
a) The synthesis of titania films templated by Pluronic was explored systematically and optimized for dye-sensitized solar cell application. The deliverable D2.1 was achieved at M12. The milestone MS11 (‘Templated electrode with net 2% increase in efficiency vs. reference electrode” delivery at M12) was reached. The actual increase was 3.5%, 2.5% and 1.9 % at 0.1 0.5 and 1 Sun, respectively for the optimized 10-layers P-film referenced to the performance of Dyesol paste film of the same thickness. High-quality (crack-free) films of 2-3 µm in thickness were prepared, which significantly larger thickness than earlier reported in the literature (typically 1 µm or less).
b) Novel advanced structures were developed towards bicontinuous mesoporous titania and mesoporous single-crystals of anatase. The mesoporous single crystals (MSCs) represent a new family of mesoporous TiO2 which had not been previously realized despite much effort from the battery community. This material promises to be much less defective, and much easier processable than the previous nanoparticle state-of-the-art. In addition we have developed a route to deposition of uniform films of MSCs which do not require subsequent sintering, and only require one drying step at 150 ̊C. This is a step change improvement for solar cell manufacturing since it obviates most of the capital expenditure required for drying and sintering, and in principle increases the practical efficiency of manufacturable solar cells. (The state of the art lab based cells have two sintering steps with an intermediate TiCl4 treatment. The TiCl4 treatment and 2nd sintering step are unpractical at industrial scale, so the peak cell efficiencies are correspondingly reduced.)
c) Doping was investigated with deliverable D2.2 (at M12). Bulk doping via hydrothermal synthesis produced noticeable effects on the TiO2 properties. Commercial Dyesol 18NRT and pigment grade TiO2 are chosen in comparison with 1% Y3+ doped TiO2 in DSC performance. The particle size distributions were measured by Malvern laser particle size analyser. Bulk doping with 1% Y3+ boosted solar conversion efficiency to 6.4% compared with the undoped value of 5.1%. The 1% Y3+ doped TiO2 achieved net increase of 2.4% efficiency under full sun compared to the baseline 18NRT TiO2 material. The 1% Y3+ doped TiO2 has a higher open circuit voltage and short circuit current compared to undoped TiO2. The outperforming current/voltage characteristics of the Y3+ doped TiO2 as compared to 18NRT is caused not only by larger Voc, but also by other effects introduced by doping. More specifically, high porosity and pore size of Y3+ doped TiO2 film give better pore filling which has increased current. Pore size or forming open pore channels, not surface area and porosity, become vital parameters to influence solid state device performance. The 1% Y3+ doped TiO2 has also shown excellent long term stability under continuous light soaking without using UV filter. 18NRT TiO2 has dropped to 1% efficiency from 6.2% initially after 1200 hours light soaking without UV filter. While the 1% Y3+ doped TiO2 has still kept 4.2% from initially 5.6%, after 2000 hours light soaking without UV filter. 1% Y3+ doped TiO2 is much less photocatalytic reactive compared to 18NRT, it is even less reactive than a commercial pigment TiO2. The results from the stability study have shown that Dyesol UK has achieved MS12 again in solvent based DSC.
d) TiO2 films were usually prepared by supramolecular templating with amphiphilic copolymers using titanium ethoxide as the Ti-precursor in a sol–gel method. The synthetic strategy using this kind of supramolecular templating almost always leads to porous structures, typically mesoporous ones with highly ordered framework structures. However, within a systematic screening of templates we have discovered and important exception to this rule. Specifically, the poly(hexafluorobutyl methacrylate) as the structure-directing agent provided surprisingly dense (non-porous) films. The films were quasi-amorphous, but crystallized partly to anatase upon heat treatment. Compact non-porous TiO2 films exhibited the roughness factor smaller than 20 even for approximately 1 µm thick film. Films were grown on glass and FTO and are able to cover even rough surfaces of the substrates perfectly, which proves the thixotropic properties of the Ti-precursor gel. These films promise interesting application, particularly in the solid-state DSCs, see below, where the can be applied as alternatives to spray pyrolysis, electrodeposition or ALD.
e) Subsequent studies of the dense films (entry 2.1 d) confirmed that electrochemical techniques are suitable for the characterization of porosity or dense packing of such titania layers. Cyclic voltammetry using ferri/ferrocyanide, methylviologen or spiro-OMeTAD as the model redox probes indicated selectively the pinholes, if any, in the layer. The pinhole-free films on FTO represent excellent rectifying interface, at which no anodic Faradaic reactions occur in the depletion state. The flatband potentials of the electrodeposited films, determined from Mott-Schottky plots, are comparable to those of anatase single crystal. The values of sol-gel films are upshifted by ca. 0.2-0.4 V, but still follow the Nernstian pH dependence. The optimized buffer layer embodies interplay of quasi-amorphous morphology, responsible for the electrochemical blocking function, and the calcination-induced crystallinity, responsible for the fast electron injection and transport in the conduction band. The latter manifests itself by reversible charging of chemical capacitance and band-gap trap states of TiO2 in its accumulation state. The blocking layers in DSC prevent back reaction (recombination) of photoinjected electrons with the oxidized form of mediator or with the hole-transporting medium. The minimization of the recombination current is important particularly in the solid-state DSCs. For their proper function, a non-porous blocking underlayer of TiO2 must be deposited on top of FTO to prevent shunting of electrons from the FTO support to the hole-transporter. Two types of pinholes in the blocking layers are classified, and their effective area is quantified. Certain films of the thicknesses of several nm allow distinguishing the depletion layer formation both in the TiO2 film and in the FTO substrate underneath the titania film. The excellent blocking function of thermally oxidized Ti, electrodeposited film (60 nm), and atomic layer deposited films (>6 nm) is documented by the relative pinhole area less than 1 %. However, the blocking behavior of electrodeposited, and atomic layer deposited films is strongly reduced upon calcination at 500oC. The blocking function of spray-pyrolyzed films is less good, but also less sensitive to calcination. The thermally oxidized Ti is well blocking and insensitive to calcination.
f) Raman spectroscopy is one of the top important analytical techniques to characterize titania films. Consequently, some effort was focused to deeper understanding of these spectra. The Raman characterization of titania, including Raman characterization under electrochemical control (Raman spectroelectrochemistry) was upgraded by theoretical (DFT) modelling of the structural changes upon cathodic bias. To address these problems, we employed the isotope labeling (16, 17, 18-O, 6-Li, 7-Li) which allowed detailed Raman assignment to be carried out. In this way, we have succeeded in assigning 20 vibrations (out of 42 theoretically predicted) for the lithiated orthorhombic phase, which is considerably more than it was reported so far. The Raman features in anatase and rutile were addressed in detail at temperatures down to 5 K. The combination of experimental and theoretical Raman frequencies with the corresponding isotopic shifts allowed us to address various still-open questions about the second-order Raman scattering in rutile, and the analysis of overlapping features in the anatase spectrum. The isotope-labeled titania was further used for the characterization of titania/gas interface using IR spectroscopy and the EPR characterization, in which the 17-O labeled titania finds logically the unique use thanks to its nuclear spin of 17-O. These studies are addressing the fundamental differences between titania/gas (vacuum) and titania/electrolyte solution interfaces. At the moment, there are serious contradictions, concerning, e.g. the band edge positions (flatband potentials) which are crucial for voltage optimization of Graetzel solar cells.

Task 2.2 Oxide films from nanotubes, nanofibres, nanowires and nanosheets (Elmarco/Oxford/UCAM)
a) New synthetic protocols were developed towards titania nanofibers employing innovative polymeric templates with easier detemplating by calcination. The deliverable D2.3 was achieved at M12. The tasks for deliverable D2.3 were fulfilled. As specified in Tables 2.4 - 2.8 most of the prepared materials have their specific surface areas higher than 25 m2/g (threshold for D2.3); moreover, they strongly depend on calcination. The average pore sizes also match the projected threshold (100 nm). This is documented by SEM images. Besides titania, also doped TiO2 and SnO2 fibres were prepared.
b) Preparation of composite electrodes focused at fibre/nanoparticle assemblies in the composite layers, both TiO2 and SnO2. The solar performance of multilayer mesoporous TiO2 films sensitized with N-945 dye scales linearly for 1 – 3 layer films, but reaches plateau value for more than 8 layers. The solar conversion efficiency of 5.05 % was found for a 2.3 μm thick mesoporous TiO2 film consisting of 10 layers. A net improvement from 4.96 % to 5.51 % (that is 11 % relative difference) for a film made of templated film/nanofiber composite as referenced to pure templated film sensitized by C106 dye. This conclusion was confirmed by a net 9 % improvement for a similar system of nanocrystalline bottom layer from electrospun nanofibers, four mesoporous layers, and one nanocrystalline anatase scattering top layer sensitized with the N945 dye. Obviously, the composite of electrospun nanofibers in templated mesoporous films represents further boosting of efficiency, if we reference the performance of optimized film to that from non-organized particles. The relative improvement of 0 – 2% - 11 % in the series of non-organized particles – templated film – composite film has been achieved. The deliverable D2.4 was successfully reached at M18. Screening of multiple combinations of materials demonstrated that stable films containing up to 80 % of nanofibers can be obtained. The highest solar efficiency of the composites (Table 2.9) was obtained for 20% ET11 composite. The actual value of efficiency of the composite electrode (7.53%) is outperforming that of pure nanoparticle film (7.23%). This confirms improved electron transport (see D2.4) although the effect in not too big, and quite often it is hardly traceable at this stage of research.
To eliminate problem of electron capturing and recombination with dye/electrolyte within extremely open mesoporous structure, which probably hinders the increase of solar conversion efficiency for 8 and more mesoporous TiO2 layers, electrospun nanocrystalline fibrous TiO2 was incorporated into mesoporous TiO2 thin film. TiO2 with fibrous morphology was found to be beneficial for the performance of the corresponding photoanode in dye-sensitized solar cell (DSC). Obviously, its wirelike structure suitably interconnects mesoporous network and thus increases the electron collection efficiency from the TiO2 layer to the F-doped SnO2 (FTO) electrode. Performance of the DSC with 2.5 µm bimodal TiO2 photoanode reached 5.35%. Among others, the performances of DSCs are limited by the charge recombination taking place mainly at the FTO/TiO2 interface. Due to the porous structure of TiO2 films the electrolyte solution easy penetrates to the FTO. The physical contact between the electrolyte and the FTO surface causes the charge recombination resulting in a considerable loss of photoelectron conversion efficiency in DSCs. Therefore, the recapture of the photoinjected electrons with redox mediator should be avoided. FTO coverage with a thin compact TiO2 underlayer was found to be the effective way to reduce the contact surface area for the bare FTO substrate and the redox electrolyte (so-called blocking effect, see Task 2.1). Besides the blocking effect, the compact layer can improve the adhesion of the FTO/TiO2 interface as well and creates more electron pathways from the porous layer to FTO and subsequently increases the electron transfer efficiency.
c) Investigation of nanosheets provided significant results for the (001)-oriented anatase nanocrystals, which were prepared hydrothermally in HF medium. The enhancement of solar cell voltage is demonstrated by comparing with standard (101) oriented nanoparticles. Of particular interest is the markedly slower recombination kinetics for nanosheets (Fig. 2.25 right chart) compared to nanoparticles. The voltage enhancement is attributed to the negative shift of flatband potential for the (001)-face. The electronic structure of the TiO2 models shows a conduction band energy upshift for the (001)-surface ranging between ∼50 and ∼110 meV compared with the (101) surface. This conclusion rationalizes earlier works on similar systems, and it indicates that careful control of experimental conditions is needed to extract the effect of band energetic on the current/voltage characteristics DSCs. Nanosecond flash excitation was applied to the sensitized TiO2 film. Monitoring the transient absorbance signal allowed to follow the time course of the oxidized dye produced upon ultrafast electron injection from the photoexcited sensitizer into the conduction band of TiO2. The back electron transfer is by a factor of 6 slower for the (001)-nanosheets compared to the same process on (101)-nanoparticles. These experimental findings were supported by theoretical modelling using first-principles computational investigation on the adsorption mode of the dye C101, on anatase TiO2 models exposing the (001) and (101) surfaces. TDDFT excited-state calculations provided the same optical band gap, within 0.01 eV, for the (001)- and (101) models. Two dominant adsorption modes for C101 dye adsorption on the (001) and (101) surfaces were found, which differ by the binding of the dye carboxylic groups to the TiO2 surfaces (bridged bidentate vs monodentate), leading to sizeably different tilting of the anchoring bipyridine plane with respect to the TiO2 surface. The different adsorption mode leads to a smaller dye coverage on the (001) surface. For the energetically favored adsorption modes, we calculated a larger average spatial separation, by 1.3 Å, between the dye-based HOMO and the semiconductor surface in (001) and (101) TiO2 models. In terms of simple nonadiabatic electron-transfer considerations, our model predicts a retardation of the charge recombination kinetics, in agreement with the experimental observations.
d) Mesoporous TiO2 microbeads are successfully applied in solid-state dye-sensitized solar cells, reaching 3.5% efficiency.

Task 2.3 “Core-shell” Nanostructures and surface treatments (Oxford, UCAM, HIPC, Dyesol)
a) The reference titania core was investigated by Raman spectroscopy upgraded by isotope labelling and theoretical simulations. This work provided platform for investigation of surface treatments: the surface isotope exchange can be followed by high-resolution FTIR on model molecules.
b) The growth of core-shell structures focused at Al2O3 shell, but the protocols of growth of other oxides (Y2O3, MgO, Al2O3 and NiO) was developed, too. Electrochemically doped shell was prepared using Y3+ and Nb5+ as the intentionally inserted impurities in TiO2. The materials with Y-containing shell provided best values of the solar cell voltage. The core-shell structure with 1% Y3+ doped shell exhibited improved solar conversion efficiency to 6.3% compared to the value for pure titania core 5.1% (cf. also discussion of task 2.1 entry (c)).

WP3 - Electrolyte/hole transporter development (NCSRD)

Task 3.1 Cell manufacture and evaluation (Oxford/EPFL/NCSRD/UCAM/Dyesol)
Taking into account the revolution that came up with the use of perovskites during last years, pioneered by the two groups of the SANS consortium (OXF and EPFL), special attention was given in the manufacture of highly efficient solar cells employing the CH3NH3PbI3 perovskite. Thus:
a) Solar cells gave 11.7% under 1 sun AM1.5G illumination when fullerene monolayers modified the mesoporous alumina scaffold while a similar efficiency (11.9%) was obtained when a proton-doped hole transport material (spiro-OMeTAD) was employed au lieu of standard Li+ doped spiro (which gave 10.9%). A breakthrough in alumina cells (efficiency of 15.6%) came when a better electron acceptor (a graphene composite with titania) replaced compact titania layer which exhibited low conductivity. This value was the highest ever reported at the time, for a perovskite solar cell.
b) Solar cells were also fabricated without the use of alumina scaffold since alumina is an insulator; to this end, the perovskite was simply deposited ontop of the titania layer, fabricating planar heterojunction solar cells for the first time in literature; these cells produced record efficiencies of 12.3%.
c) Taking into advantage of the ambipolar nature of the perovskite semiconductor, we also fabricated planar solar cells of inverted structure where the perovskite was deposited ontop of a PEDOT:PSS p-type layer and the junction was formed between the perovskite and PCBM (which is one of the most effective electron transporters). The cells produced efficiencies of 10%.
d) To better control the crystallization process and thus enhance the surface coverage of the perovskite, a two step deposition-conversion was employed; first a PbI2 film was deposited on titania mesoporous layer and then the film was introduced in a saturated solution of methylammonium iodide in isopropanol. This process produced certified efficiencies of 14.1% for a perovskite sensitized mesoscopic solar cell employing the CH3NH3PbI3 as the absorber.
e) The most promising material from WP2 was also used here; mesoporous titania single crystals with unique electron transport properties were incorporated in standard perovskite solar cells, delivering efficiencies higher than 7.5%.

Task 3.2 Improved pseudo-solid-state electrolytes (EPFL/NCSRD/Dyesol)
Improved pseudo-solid-state electrolytes were synthesized and then incorporated in typical dye sensitized solar cells:
a) Mixtures of ionic liquids were prepared and used for the development of composite redox electrolytes by blending a standard ionic liquid EMiDCA (1-ethylimidazolium dicyanamide) with various iodine-based ionic liquids based on the methylimidazolium cation (DMII, EMII, PMII, BMII and HMII). The electrochemical properties along with the conductivity of the as-fabricated electrolyte were determined; promising values as high as 5-10  10-7 cm2/s and 2-4 mS/cm were estimated for the diffusion coefficients and conductivity, respectively. Solar cells based on the above electrolytes attained efficiencies over 4% under 1 sun with the highest being 5.5%, attained by DMII. Quite notably, these efficiencies further increased up to 6.5%, when the cells were illuminated by 0.1 sun.
b) A large variety of solidifiers were employed using organic polymers (such as polyethylene oxide), organic ionic plastic crystals and inorganic nanoparticles (like a mixture of anatase/rutile TiO2) . The solar cells attained efficiencies of 4-5.5%, which are very significant if we take into account the solid nature of the electrolytes. Particularly, novel solidified electrolytes based on the EMiTCM ionic liquid, endowed with silica, were fabricated, presenting promising efficiencies of the order of 4.5-5.0% and over 6.3% under 1 and 0.1 sun AM1.5G illumination conditions. Solidification of the ionic liquid electrolytes with amorphous silica nanoparticles revealed that electrolytes with the lowest viscosity loose only a marginal percentage of their initial efficiency, and most importantly that those with the greater viscosity increased their efficiency by more than 20%. The presented results pave the way towards the exploitation of ionic liquids based on tricyanomethanide ion in liquid and solidified redox electrolytes for dye-sensitized solar cells and other related applications such as CO2 capture. The solar cells incorporating the silica-based electrolytes presented significant efficiency stability (retaining the 90%of their initially efficiency) upon storage for more than 4000h.
c) Two different biopolymers, agarose and gelatin, which promote environmentally friendly applications, were used to solidify ionic liquid based electrolytes, and the resulting systems, which were completely solidified, were incorporated in dye sensitized solar cells, showing an efficiency of the order of ~ 3.5%. Additionally, the solar cells incorporating the agarose-based electrolytes presented efficiency stability upon storage, maintaining ~80% of their initial values for more than 600h.
d) Silica nanoparticles were also used to solidify a liquid electrolyte employing the novel Co2+/Co3+ redox couple for the first time in literature. The solidified electrolyte gave very promising efficiencies of more than 4% under low light illumination conditions.

Task 3.3 Solid-state hole-transporters (Oxford/EPFL)
Using cobalt dopant in Spiro-OMeTAD hole transport materials the efficiency of solid-state DSC reached 7.2% reaching the objective of the D3.4 which is >7% PCE. UOXF and UCAM have developed an optical probe to determine the pore filling fraction of the hole-conductor 2,2-7,7-tetrakis-N,N-di-pmethoxyphenylamine-9,9-spirobifluorene, and using this technique they concluded that pore filling is not a limiting factor in the fabrication of thick ss-DSCs. A new set of hole transporting materials and dopants were synthesized, which are under investigation for solid state dye sensitized solar cells. With the perovskite absorbers, the solid-state cells have excelled far beyond expectations and now represent the most efficient emerging PV technology in existence, having shot past amorphous silicon.

Task 3.4 Stable QD sub-systems with novel electrolytes (Oxford/EPFL/NCSRD/UCAM)
High efficient core/shell near Infra-red QDs in QDs/TiO2 heterojunction solar cell of the structure PbSe/PbS QDs/TiO2 were fabricated, which produces a light to electric power conversion efficiency (η) of 4% with short circuit photocurrent (Jsc) of 17.3 mA/cm2.

Task 3.5 Photovoltaic response from IR to visible up-conversion (Oxford/EPFL). The perovskite cells have exceeded expectations and surpassed the 14% efficiency target of the SANS project.

WP4 - Operational Mechanisms
Task 4.1 Testing the basic elements (Oxford/EPFL/NCSRD/CNR-ISTM/UU/GSA)
The TiO2 electrode: TiO2 underlayers by spray-pyrolysis provide better blocking behaviour for SANS devices than those obtained by sol/gel methods. For mesoporous films, increasing the average TiO2 particle size of the mesoporous films was found to lead to over an order of magnitude faster charge transport, at the cost of lower surface area. The experiments give evidence that transport limiting traps are located at the grain boundaries. Mesoporous microbeads have high surface area and faster electron transport, probably due to favourable orientation of the nanocrystals in the beads. In mesoporous anatase TiO2 single crystals such grain boundaries are absent and charge transport was found to be one order of magnitude faster than in the sintered nanoparticle films, even without thermal sintering.
Sensitizers: The binding strength of dyes on the TiO2 surface in electrolyte solution was found to range from 10,000 M-1 (D149) to 200,000 M-1 (Z907), suggesting significant concentration of dye in the electrolyte. Accurate quantum chemical methods have been developed to calculate energy levels, dipole moment and the adsorption configuration on the TiO2 surface of both ruthenium-based sensitizers and conjugated D−π−A organic sensitizers. Calculations show that the rhodanine-acetic binding group in weakly coupled and gives non- adiabatic electron injection, while cyano-acrylic acid give strong coupling with TiO2 and adiabatic injection. Calculations also demonstrate that positive charge localized on adjacent sensitizer molecules is a major cause for the Stark shift observed in transient absorption spectroscopy.For dye-sensitized solar cells with cobalt-based electrolytes complete monolayer coverage of dye on the mesoporous TiO2 is necessary for good photovoltaic performance, while partial dye coverage leads to poor performance due to fast electron recombination.
Electrolytes and hole conductors: Cobalt-based electrolytes were solidified using silica nanoparticles, giving good performance (75% compared to reference liquid electrolyte) and stability in solar cells. D−π−A organic sensitizers yielded excellent performance in ionic liquid iodide based electrolytes. The role of Li-TFSI in solid-state SANS devices with molecular hole conductors was identified a new family of p-dopants, protic ionic liquids, was developed, resulting in efficient solid-state solar cells.

Task 4.2 Understanding photovoltaic operation, combining theoretical modelling and state-of-the-art experimental techniques (Oxford/EPFL/NCSRD/CNR-ISTM/UU/GSA)
Electron injection is a deciding step for the performance of highly efficient SANS solar cells. The veritable injection efficiency can only be high if the subsequent recombination with the oxidized dye is suppressed. Measurements of both the injection and recombination dynamics are therefore very important to understand these limiting factors and to come up with recommendations for improvement of the solar cells. We have evaluated measurements. Furthermore, a supramolecular approach for suppressing rapid recombination by addition of a donor to the dye was evaluated and proved to be successful, resulting in a higher voltage at the maximum power point.
Ionic additives in the electrolyte affect the optical and electronic properties of a dye-sensitized TiO2 heterointerface. For instance, the relative position of the energy bands of TiO2 and the dye can be shifted. This can be accurately modelled using quantum chemical calculations. Also effects of energy transfer and co-sensitization were evaluated using calculations.
Filling of the pores in the mesoporous TiO2 with solid hole conductors was evaluated: with increasing pore filling fractions, the lifetime of electrons was extended, but no change in electron transport kinetics. A pore filling fraction of at least 60% is required to achieve optimized performance in solid-state dye-sensitized solar cells. It is possible to optimally fill mesoporous films of up to 5 µm in thickness.
Perovskite solar cells: the organolead triiodide perovskite appears to be a nearly ideal absorber for SANS, with high absorption below 800 nm. Time resolved PL studies show that the electron and hole diffusion lengths in perovskite absorbers are on the order of 1 micron in length, identifying why highly efficient charge collection at high voltages is feasible. A small addition of chloride ions to the organolead triiodide perovskite results in a striking increase in the electron-hole diffusion length, predominantly arising from a substantial inhibition of non-radiative electron-hole recombination. A mechanism for UV light induced degradation of TiO2 – based perovskite solar cells has been found. By avoiding direct excitation of TiO2, Al2O3 based cells are much more stable. Using functionalised mesoporous titania with a fullerene self-assembled monolayer (C60SAM) efficient perovskite solar cells were obtained using spiro-OMeTAD as the hole transporter. Finally, an analytical model for the solid-state perovskite solar cell has been developed which describes the operation of the solar cell as a p-i-n heterojunction device.

WP5 - Scale-up & stability

Task 5.1 Cost analysis (GSA)
This Task involved evaluation of the cost of producing SANS devices using various materials developed by the SANS partners.
Cost Analysis was done on various stages of SANS project. The initial step required evaluation of the manufacturing costs of the individual materials employed for the SANS devices. Subsequently, the costs to incorporate these materials into a SANS device were evaluated, along with the required materials volumes. Finally, the device performance characteristics were assessed, along with their lifetime characteristics, to ascertain the overall cost to produce energy using the specified device design and material set.
Material costs took into account both direct and indirect costs associated with production of the final materials which may subsequently enter into the production process for the SANS devices, and included the following items: Bill of Materials (BOM), Labour, Quality Control (QC), Utilities, Facilities, Plant/Equipment, Overheads, and Margin. Material costs were determined per g or mL of raw materials, based on a production scenario of 200,000 m2 per annum.
Based on the cost analysis of various materials and processes for SANS devices it was concluded that
1. The present high-performing organic Y123 sensitiser system, identified by the SANS partners, which is found to be compatible with both Co-based liquid electrolyte redox couples and solid state hole conductors, requires further development of the synthetic routes to eliminate the use of dangerous and difficult to scale intermediate chemicals such as butyl-lithium which is pyrophoric in nature and needs handling under inert conditions such as nitrogen/vacuum line. In addition, considerable process development is required to decrease synthetic steps for manufacture of materials and considerably improve production yields. Good progress has been made on refining and developing a more suitable industrialised synthesis route.
2. YD2-oC8 porphyrin based dye, was eliminated from consideration due to scalability issues and poor durability results for variants of porphyrin dyes studied separately in SANS project. Due to the above fact there was a little incentive for continuing to pursue YD2-oC8.

Device production related costs took into account the same items as detailed above and any associated process materials (such as dye solution solvents for example). In addition to the material costs, for which usage data was derived from SANS project data and information provided by SANS partners, device manufacturers also need to consider customer warranty, which must be provided for in the cost structure, as product liability factors throughout the lifetime of the final product need to be taken into account. Panel costs were calculated according to customarily used designs for SANS systems, with the basic assumptions of an 80% active area for SANS and sfSANS and a 90% active area for ssSANS. At the used 200,000 m2 production throughput figure, the manufacturing portion of the final panels was around 20%. The final estimated panel costs, per square meter were calculated for stage 1 and stage 2 material sets.
From the panel cost and performance data levilised cost of energy (LCOE) was calculated by summing all the costs incurred during the lifetime of the generating technology, including: Panel Cost, Balance of Systems (BOS), Operational Costs, and Decommissioning Costs. These costs were then divided by the units of energy produced during the lifetime of the array and expressed as Euros per kilowatt hour (€/kWhr) for both stage 1 and stage 2 material sets. A more conventionally used parameter cost per peak-Watt ($/Wp) was also analysed for very large production area.
A number of recommendations were made as an outcome of the Cost Analysis investigation:
• As device durability is such a key parameter in cost analysis, and appropriate data was not readily available for many of the materials, particular focus should be given during the next period of the SANS project to ensuring adequate long-term stability data can be acquired.
• The best combinations of materials in order to ensure adequate stability and performance result in a low overall panel efficiency. Thus, effort towards improving the stability of known well-performing materials, and/or towards improving the performance of known high-stability materials was a desirable route for future improvement.
• ssSANS had the least data available, and particular effort should be applied to this Material Set to better quantify its performance and especially durability capabilities, in light of its low material usage and thus potential for low-cost.
• At very large areas, conducting elements (either sheet conductors or metallic conductors) became a large component of the overall panel cost, and attention was warranted for approaches to reduce the cost of these elements by improved design or new materials, although this work lay outside the scope of the SANS project.

Task 5.2 Lifetime enhancement (Oxford/EPFL/NCSRD/Dyesol/GSA)
This Task involved testing the various SANS devices developed within the project, and assessing their durability characteristics. Analysis of the failure modes lead to improvements in system stability.
The Stage 1 and Stage 2 Material Set list was identified as per MS33 and MS34 from various SANS Partner results, focussing on maximum achievable efficiency in Stage 1 material set and highest durability in Stage 2 material set. SANS devices produced from these material set were categorised into three distinct classes, each of which has significant implications in relation to industrialisation:
SANS: traditional sensitiser activated nanostructures, which employ solvent containing electrolyte systems;
sfSANS: solvent free sensitiser activated nanostructures, which employ liquid electrolyte systems comprised of ionic liquids; and
ssSANS: solid state sensitiser activated nanostructures, which employ hole conductors of a solid nature.
Toward the end of the project material set 3 was identified by SANS partners as per MS35. A key paradigm shift for the final 12 months of the SANS project was to move towards industrially scalable and viable material sets, to pave the way for rapid product development activities.
Various SANS devices produced and their results outcome during this project are summarised in the following.
1. Solvent Containing Liquid SANS: Test cells fabricated by Dyesol using 2 different liquid electrolytes (by varying the electrolyte solvent, namely MPN and HBS were aged under 2 different conditions: light (0.8 sun) and thermal (80 0C) stress for 2000h. The efficiency of the initial MPN cells was about 5 % and dropped down to 1.3% after thermal ageing due to severe Jsc reduction. The drastic decrease of Jsc was attributed to triiodide loss due to leakage/evaporation and/or I3- depletion from the cathode (I3- could be trapped somewhere inside the pores of TiO2). No dye degradation or desorption was detected. On the contrary, the HBS cells have demonstrated <10% performance loss after 1000 hours at 85 °C.
2. Ionic Liquid SANS: Replacement of the low viscosity but toxic EMITCB ionic liquid with a similarly performing and durable alternative has proven challenging.
3. Solid Stage SANS: The durability of ssSANS devices under light soaking is very encouraging, but considerable further investigation is required to ascertain the role of oxygen and dopants in changing the performance in positive and negative ways during device operational lifetimes, as well as assessing durability under more stressful conditions such as elevated temperature.
4. A baseline material set, utilising a sensitising ruthenium dye antenna system, an iodine based electrolyte liquid hole conductor, and a mesoporous nanoparticulate titania metal oxide film, has been tested under accelerated conditions to provide confidence of a 20 year outdoor lifetime. This is a fantastic testemount to the potential stability of sensitized solar cells.

Task 5.3 Materials scale-up (Dyesol)
This Task involved scaling-up of promising SANS materials.
1. Dyesol continues to prepare capabilities and infrastructure for scale-up of SANS materials and devices.
2. Y123 scalability is now much more promising, but YD2-oC8 synthesis requires attention for improved scalability.
3. Spiro-OMeTAD and variants show suitable scalability.
4. Co-complex electrolytes are promising, but require considerably more rigorous evaluation for compatibility with HBS for durable SANS device manufacture.

WP6 - Demonstrator

Task 6.1 Demonstrator module (GSA)

WP6 had the objective to:
“Create a working demonstrator solar cell module incorporating the materials developed within the SANS project”.
There was a single task in this WP and has the aim of creating industrial scale cells and developing prototype modules and technology demonstrators. Baseline, Stage 1 and Stage 2 module were all based on Ti metal flexible substrate. The modules contains 12 active strips on a titanium metal working electrode, with each strip being 7 mm wide and 56 mm in length, interconnected in a parallel configuration with the use of silver current collection ‘fingers’. The device size is 105 cm2 with an active area of 47 cm2.
• Initially with standard state-of-the-art baseline materials a demonstrator module was made. The results from this module achieved Milestone 38 which states “base line modules with less than 30% drop in efficiency as compared to small cell performance”. The much larger Baseline Material Set Module provided over 90% of the conversion efficiency of the small test cell reference device using the same material set.
• The next step was to produce a demonstrator module by using Stage 1 Test Cell material set. The results show that the Stage 1 Module has a better dye solar cell performance compared to the Stage 1 Test Cell rather than dropping performance when transitioning from a small active area test cell to a large active area module. The results from the Stage 1 Material Set Module have achieved Milestone 39 which states “stage 1 modules with less than 30% drop in efficiency as compared to small cell performance”. The much larger Stage 1 Material Set Module exceeded the conversion efficiency of the small test cell reference device using the same materials.
• Further a demonstrator module was produced by using stage 2 Test Cell material set. Test results of Stage 2 Module and Stage 2 Test Cell show that the Stage 2 Module has a better dye solar cell performance compared to the Stage 2 Test Cell rather than dropping performance when transitioning from a small active area test cell to a large active area module.
• Finally, a module was made by using Stage 3 material set. Stage 3 Module was made on conducting glass substrate. Consideration to use a glass module was because glass substrate is a good barrier for moisture and oxygen. Also glass DSC is more appealing for BIPV application, such as used in windows. The module contains three active strips on an FTO glass working electrode, with active area of 42 cm2. The strips were interconnected in a parallel configuration with the use of silver current collection ‘fingers’. Stage 3 Module gave a higher current and efficiency than Stage 2 Module. Comparing to Stage 3 test cell, Stage 3 Module had slightly lower efficiency than Stage 3 Test Cell. That meets Milestone 41 which states ‘Stage 3 modules showing enhanced performance over stage 2 module with less than 30% drop in efficiency as compared to small cell performance’.Transitioning from a small active area test cell to a large active area module with marginal performance drop boosts confidence in the scalability of the SANS technology.

Potential Impact:
The SANS project set out to revolutionise the global scope of low-cost photovoltaics. To a large extent, the project has been unexpectedly successful in this ambitious aspiration. To put the societal impacts into perspective, it is worth reflecting on the concept of energy, and how it is sourced and utilised, as well as the impacts it has upon the human population, in both negative and positive ways. When viewed from this angle, the potential societal impacts of the outputs from the SANS project are vast. An available source of energy is at the root of most indicators for wealth and wellbeing. Energy is used for creating goods, producing food, heating homes, purifying water and transportation. Modern society is founded upon oil, gas and coal, without which we would still remain a preindustrial society. Moving forward, replacements to coal, gas and oil must be found since they are not an infinite resource and the burning of fossil fuels has a negative impact upon local air quality and global warming. Solar power has the prospects to easily deliver the power demand for the planet, and in fact is the only technology which is capable of doing so far into the future, provided a sufficiently cheap and efficient technology emerges. With the relative ubiquitousness of electrification, solar photovoltaics are the most readily integratable and accessible form of solar energy harvesting.

Within the solar photovoltaic space, the market is presently >90% dominated by so-called “1st generation” PV. 1st-gen PV, a.k.a. silicon PV (in either single-crystalline or poly-crystalline form), is an old technology dating from the 1950’s, and has had a long development cycle to reach its present status. Significant government funding schemes, such as subsidies for end-user capital investment, subsidies for module production costs, or subsidies for PV generated energy, have been essentially propping up this technology for the better part of the last two decades. While this has spurred great interest in this element of the “renewable energy” space, it has also masked the true costs, in both energy and economic terms, of the rapid growth in the PV industry. However, notwithstanding this rapid growth, PV energy generation still only accounts for a measly single digit percentage of global energy production. While more “newcomers” are entering the commercialised technology space in the form of 2nd-generation PV (CdTe, CIGS, amorphous silicon, etc.), these technologies have not yet reached superior commercial prospects except in narrowly defined roles such as low-cost CdTe cells for remote solar farm applications. What is missing from the equation to date is a true low cost, widely deployable, easily manufactured and rapidly scalable technology. Incumbent PV technologies of the 1st and 2nd generations simply cannot meet these requirements for a variety of reasons, such as energy intensive manufacture, or resource limitations of raw materials, for instance.

The work within SANS has made a significant advance of 3rd generation PV technologies and will in the future lead to an extraordinary advance in well-being and societal wealth, only previously witnessed with the discovery of oil. 3rd generation PV technologies are truly globally scalable, with no resource limitations, highly scalable and cheap manufacturing routes for both materials and panels, and extremely flexible deployment options valid for both traditional markets (solar farms, rooftop panels, etc.), as well as more fundamentally market-changing applications such as building integrated photovoltaics (BIPV). Although there is still a long road towards such globally significant deployment, the initial foundational framework for the inevitable success of this novel technology has been securely laid within the SANS project and its impressive technological gains. The overriding goal of the project was to realise a verified 14% efficient lab based cell, and a module with a lifetime equivalent to 20 years European outdoor usage. With these sorts of performance metrics (efficiency, lifetime), the low-cost material/production technologies developed within the SANS project are positioned for rapid industrialisation and the inevitable change they will bring to the energy marketplace and human society will be a welcome breath of fresh, low-CO2, low pollution, air.

Exploitation of the SANS results has already begun, and is only set to accelerate in the coming years. Dyesol maintains a core business element of raw material supply for 3rd generation photovoltaics. Since the wrap-up of the SANS project, Dyesol has doubled the product listing of its materials catalogue, releasing a new product range which includes materials such as precursors for solid-state perovskite SANS, and dopants for solid-state SANS hole transport materials, amongst other materials. Many of these new materials owe aspects of their heritage to the SANS project, for instance intermediaries for the highly stable dye Y123, or alkyl-ammonium halide precursors for perovskite based solar cells. As the premier supplier of large-scale, high quality SANS materials, Dyesol is well positioned to both contribute to, and capture value from, the imminent rapid expansion of this incredibly exciting technology, and in particular the more recent advances which are, in part, attributable to the SANS project.

Dyesol continues the commercialisation of dye-sensitised solar cells and related SANS technologies, and has successfully advanced the production feasibility of SANS DSC technology on various substrates including metal foil. The stability work on both doped TiO2 nanoparticles and solidified electrolytes is under consideration for feeding directly into these commercial activities. Production of perovskite based SANS technology is an expanding area of commercial interest for Dyesol, and Dyesol and EPFL have joint IP arising out of the SANS project which covers key doped titania materials and underpins this activity. Beyond this, Dyesol has revised its business plan to accommodate the commercial opportunities represented by the changing playing field in 3rd generation photovoltaics, a not insignificant contributing element to which was the outputs of the SANS project.

New advanced TiO2 mesoporous single crystals, developed within SANS, have been licenced to and are being employed by Oxford Photovoltaics Ltd. (Oxford PV, a spin-out company from the UOXF) in their solid-state solar cell activity. This material has enabled removal of the sintering step in ssDSC manufacture, greatly broadening the substrate choice and reducing the capex (and footprint) required for manufacture. Oxford PV are also developing the processing of perovskite absorber based solid-state photovoltaic modules, which are exhibiting much higher efficiencies than the solid state DSSC. Within the SANS project, we have removed the sintering step of this perovskite technology, another important aspect for manufacture.

Dyenamo is the commercial platform of the Center for Molecular Devices (CMD), headed by Professor Anders Hagfeldt at Uppsala University, Sweden. CMD also includes two research groups: the Royal Institute of Technology (KTH) in Stockholm, and the DSSC activities at the research institute Swerea IVF AB. Dyenamo offers competence and in-house made high-quality materials for nanoscale solar energy converters, e.g. DSSC and solar fuels, by making research and competence from CMD directly accessible for research groups, companies, and investors. Examples of materials are: efficient organic sensitizers based on triphenylamine donor group such as D35; cobalt-based redox mediators; and methylammonium salt for perovskite solar cells. Dyenamo also markets patent and intellectual property services in the space surrounding SANS technology. http://dyenamo.se

Detailed plans for exploitation will not be reported here as they are exceptionally commercially sensitive. However, broad statements in relation to scope and positioning will be summarised to provide perspective. As discussed above, the realms of opportunity for truly low-cost PV (without subsidies), which is also widely and easily deployable, represent a considerable market opportunity. Both Dyesol and Oxford PV, as recipients either directly or via license of SANS intellectual property outputs, view the opportunities of BIPV as being the most appealing. BIPV is attractive from many angles. Energy use-wise, building consume the lion’s share of energy production. Being able to electrify the building envelope (roofs, facades, skylights, etc.), provides energy generation at the point of use. This eliminates the cost of building centralised energy generation facilities (irrespective of the nature of the energy generating technology), and, importantly, removes the associated transmission costs and accompanying energy losses. A double-digit percentage of energy is lost during transmission from source to end-users, and the economic costs of installation of transmission lines is comparable to generation itself. Further, building integrated PV technology eliminates the “double handling” of current building based PV systems, which are “building applied PV” (BAPV). Building integration replaces historically passive building components with photo-active equivalents. Such components provide the original purpose (insulation, weather proofing), but now also provide energy generation. This approach is highly cost effective as it utilises a common substrate (no additional glass, metal, concrete, etc.) is required for the BIPV panel as in BAPV. Further yet, the installation labour costs are dramatically reduced, as installation occurs during building fabrication, rather than afterwards as a bolt-on process. The installation costs can represent up to 65% of the total cost of a BAPV system.

While BIPV is the preferred exploitation model, it is of course not the only approach. BIPV installation requires either a new build, or a major refurbishment, in order to best leverage its economic advantage. As such, it will slowly, gradually and inexorably become the major sustained energy generating technology, but with building construction cycle turnarounds, this will of course take many tens of years. A number of exciting retrofit business opportunities exist, for which low-cost 3rd generation photovoltaics such as those developed within the SANS project are eminently suited. Of particular relevance are flexible PV opportunities, which by virtue of their lightweight nature, lend themselves to retrofit opportunities more than heavier, bulkier systems that may require structural modification. SANS technology lends itself exceptionally well to flexible technologies. Additionally, SANS technologies with efficiencies around 15% are competitive with incumbent 1st and 2nd generation PV systems, for such applications as solar farms or more traditional rooftop mounted arrays. Such market entry points are plausible while the new commercial offerings in the 3rd generation PV realm gain market acceptance and build consumer confidence in the technology, as these applications are more familiar to the existing “customer” base.

To effectively take the technologies developed within the SANS project into the market opportunities described above will require dedicated hard work in scale-up of both materials and panel fabrication, as well as considerable market development. Many of the best opportunities either have heavily guarded routes to market, or may be via unconventional channels. A great deal of business development is necessary for these opportunities to be realised. Likewise, the technology itself requires further proving, particularly at very large scale (e.g. 1 m2 panels), and also for industrially meaningful production throughputs (>100,000 m2 per annum). Both of these exercises are expensive in relation to capital, and a significant fund raising endeavour is already underway in order to secure the necessary investment required for these next steps. Certification to incumbent industry standards presents both a challenge and opportunity for SANS technology. While existing customers are familiar with the present standards, and will expect newcomers to the field to adhere to them, there will be inevitable changes required to standards in order to accommodate the new technologies which have different characteristics to prior offerings. This will be similar to the changes which were made to IEC standards for 2nd-generation “thin film” PV, which were historically tested under silicon-PV (1st generation) standards that were inappropriate for the newer technologies. Considerable future engagement work will be required between the various players involved in establishing standards. Also, it must be made clear that, although the efficiency and stability results from the SANS project were impressive, advances in these must be continued in order for SANS technology to remain competitive in the marketplace. To stop the key research and development efforts and assume industrial deployment will occur naturally is to abandon the technology at a critical juncture in its commercial development cycle.

In addition to the field of Solar Cells, the newly developed mesoporous metal oxides may have applications in energy storage and membrane technologies. SANS partners are currently investigating these fields and interested parties to see if exploitation and commercialisation is possible. Likewise, the perovskite structures will likely have applications in light emitting systems, lasers, and a host of related opto-electronic applications, many of which will remain fertile grounds for exploration for years to come.

Within the project we have organised one workshop, and one symposium at the EMRS conference which were both well attended and showcased the activity within the SANS project.

The SANS Excitonic Solar Cells Workshop was held in Switzerland from 12-15 March 2012.

A wide range of interesting topics were discussed at the event spanning the myriad of current research topics related to sensitized solar cells. This included the preparation of carefully structured oxides for light management with photonic structures, improving charge transport and voltage, and minimising recombination; spectroscopic analyses employed to monitor processes from cell degradation to electronic processes; the physics of the operation of devices as probed with various experimental techniques; theoretical modelling down to the molecular level; and more.

Within the various analytical approaches to studying excitonic solar cells it emerged that a broad range of materials systems were being investigated from organic to hybrid to inorganic, from quantum dot sensitized to dye sensitized systems, and from electrolytic to solid-state, each providing a different perspective on the development of solar cells and new ideas for those less familiar with their colleagues work. The talks and poster session were accompanied by free and lively discussion which is ever more important in such a broad and multidisciplinary area of research.

Some highlights of the meeting came from Michael Grätzel who presented recent work at EPFL on dye-sensitized solar cells (DSSCs) with world-record efficiency; Brian O’Regan presented a thermodynamic description of DSSCs; Gerald Meyer discussed time-resolved spectroscopic studies of the electronic processes at the TiO2 interface in DSSCs; Emilio Palomares analysed the efficiency losses in excitonic solar cells; Masaru Ken Kuno presented recent work on quantum dot and nanowire sensitized solar cells; Qing Wang analysed the processes in sensitized solar cells with impedance spectroscopy; and Annamaria Petrozza presented her studies on interface engineering in hybrid solar cells.

Approximately 50 delegates from both industry and academia, from all over Europe and beyond, participated in the event which was a fantastic platform for networking and everyone benefited tremendously from the wider interaction.

The second event, a Symposium, organised during the course of the project was staged at the EMRS Spring meeting in Strasbourg, France in May 2013. The symposium entitled “Organic and hybrid interfaces in excitonic solar cells: from fundamental science to applications” had over 160 abstracts submitted for the event and included contributions from presenters from all over the world including Japan, Korea, USA, Europe.

The Symposium covered a wide range of topics and focused on three main types of interface: organic-inorganic-bio. They provided the main framework to discuss in-depth concepts from fundamentals to applications in terms of chemical interactions, materials processability, morphologies, optoelectronics processes in multi-components hybrid systems and photovoltaic devices physics.

The aim of the symposium was to gather together researchers with different backgrounds in order to approach in a multidisciplinary way the keys issues involved in interfacial mechanisms which play a key role in the definition of the final performances of Excitonic solar cells (OPV, DSSC, Quantum Dots based SC, etc).

The event was one of 25 Symposia at the EMRS meeting which attracted over 1200 delegates from both industry and academia, from all over Europe and beyond.

We have published over 80 scientific papers in highly reputable journals (see list attached) and filed a number of patents which are being commercialised through our industrial contacts.

List of Websites:

http://sans-solar.eu/
Professor Henry Snaith
h.snaith1@physics.ox.ac.uk

Final Report Summary - SANS (Sensitizer Activated Nanostructured Solar Cells)

Share this page

Download