DSSC Detail

ECN-C--05-078
NANOCRYSTALLINE DYESENSITIZED SOLAR CELLS HAVING MAXIMUM PERFORMANCE

NANOMAX Final Technical Report
J.M. KROON
AUGUST 2005
Acknowledgement/Preface
The work described in this report has been carried out in order of the European Commission, contract no. ENK6-CT2001-00575, ECN project no. 7.4623
ii
ECN-C--05-078
ENK6-CT2001-00575 NANOMAX
Publishable Report
PUBLISHABLE REPORT
Contract No: Acronym: Title :
ENK6-CT2001-00575 NANOMAX Nanocrystalline Dye-Sensitized Solar Cells Having Maximum Performance
Contact (Co-ordinator): Dr. Jan Kroon ECN Solar Energy P.O. Box 1 1755 ZG Petten The Netherlands Tel. +31 224 564734 Fax +31 224 568214 Email: j.kroon@ecn.nl
Partners: Energy research Centre of the Netherlands (ECN) Cracow University of Technology (CUTECH) The Imperial College of Science, Technology and Medicine (ICL) IVF Industrial Research and Development Corporation Alberts-Ludwig Universitt Freiburg (FMF) Ecole Polytechnique Fdrale de Lausanne (EPFL) Greatcell Solar SA (GSA) REPORTING PERIOD: from 1 January 2002 to 31 December 2004 PROJECT START DATE: 1st January 2002 DURATION: 36 months Date of issue of this report: 28-02-2005 NL PL UK S G CH CH
Project funded in part by the European Community under the FP5ENERGIE Programme
Publishable Report
Table of Contents
1. 2.
2.1 2.2
Executive Summary Objectives and strategic aspects

Socio economic objectives/strategic aspects Scientific and technical objectives 2.2.1 General Goal 2.2.2 Scientific objectives 2.2.3 Technical objectives
4 6 6 6 6 6 7
3.
3.1 3.2
Scientific and technical performance

Summary of specific project objectives Overview of the technical progress 3.2.1 Workpackage 1 (Concept development) 3.2.2 Workpackage 2 (Materials Development) 3.2.2.1 Nanocrystalline materials 3.2.2.2 Sensitizer 3.2.2.3 Electrolytes 3.2.2.4 Counterelectrodes 3.2.3 Workpackage 3 (Cell manufacturing, proof of principles) 3.2.3.1 Concepts ST and Aa 3.2.3.2 Concept Ab (TCO less concept) 3.2.3.3 Other concepts (variation of Concept B) 3.2.4 Workpackage 4 (Advanced Characterisation and modelling) 3.2.4.1 Limits to Efficiency 3.2.4.2 Highlights of WP4 research activity 3.2.5 Workpackage 5 (Long term Stability) 3.2.5.1 Stability issues in DSC 3.2.5.2 EPFL stability results 3.2.5.3 Masterplate results 3.2.5.4 Summary stability data 3.2.6 WORKPACKAGE 6 (Technical evaluation, cost analysis and module demonstration) 3.2.6.1 Introduction 3.2.6.2 Cost analysis for Concept Aa 3.2.6.3 Estimation of the costs for Nanomax low-cost concepts in relation to Concept Aa 3.2.6.4 Modules based on Concept Aa 8 9 10 12 12 14 16 16 17 17 19 20 21 21 22 24 24 27 28 30 31 31 32 34 34 36 36 37 38
3.3
Assessment of results and conclusions 3.3.1 General Conclusions 3.3.2 Dissemination and use of results
3.4
References
Publishable Report
Publishable Report
1.
Executive Summary
Objectives and problems to be solved: The research on Dye-sensitised Solar Cells (DSC) has been triggered by the discovery in 1991 that sensitising a nanocrystalline titanium dioxide photoelectrode with a specific organometallic dye produced a photoelectrochemical cell with significant current. This increased the performance of electrochemical solar cells by a factor of 10. Since then, under full sun (AM1.5) illumination, overall conversion efficiencies up to 11 % have been demonstrated for small (0.25 cm2) single cells by EPFL (Lausanne, Switzerland). Power conversion efficiencies of around 8 % have been reported for cells more than > 1 cm2. Most of the research work on DSC for the last several years has been restricted to the optimization of cells with a standard photoelectrode design, i.e. a single sensitising dye adsorbed on nanocrystalline titanium dioxide in terms of long-term stability and efficiency. Important progress has been made on technology development and stability, in large part due to work on electrolyte composition and sealing, but no significant progress has been made in efficiency. Main factors that were limiting device performance and requiring new cell and materials concepts were: a) inefficient light absorption by existing sensitizer dyes in the near infrared (up to 800-900 nm). With optimised dyes, an increase of photocurrent from currently 15-20 mA/cm2 to at least 23 mA/cm2 seems feasible; b) Sub-optimum photovoltage output. Current state of the art yields output voltages of ~750 mV, less than half the energy of incident photons, indicating improvements in photovoltage could up to double cell efficiencies. The photovoltage is limited primarily by the photon electron density and the surface recombination losses at the metal oxide/electrolyte interface. An increase in photovoltage from, currently, 750 mV to 900 mV seems realistic if surface recombination can be reduced significantly. Several announcements in literature called for imagination and exploration of new innovative ideas. New concepts, both for cell design and materials, are necessary to boost the efficiency from present 7 - 8 % to ~15% in the near future. Description of work In NANOMAX the following activities were planned: Evaluation of photoelectrode designs and materials and production of a large variety of test cells. In order to increase the photovoltage, combination of thinner layers with efficient light scattering materials. Study of multiple layer structures of oxides with different bandgap energies and surface properties Increase in photocurrent by better optical efficiency and broader absorption spectra of the sensitising dyes. Preparation of nano-crystalline oxides with reduced electron surface recombination by modifying the oxide surface (capped oxides) and mixing of oxides and optimizing surface selectivity by using near-infrared (NIR) enhanced sensitising dyes. Expected results and exploitation plans: As a project goal for NANOMAX, the following outcomes were sought: Preparation of surface selective and IR-enhanced dyes as well as nano-crystalline oxides with reduced electron recombination. Better understanding of loss mechanisms in state-of-the art cells and for new materials. Achievement of certified 12% cell efficiency on an active area of 2.5 cm2 using innovative new cell concepts and realistic extrapolation of the experimental results towards 15 % efficient DSC cells. Investigation of the upscalability with as an outcome a demonstrator module (100 cm2) having 9% efficiency
Publishable Report
Demonstration of long-term cell stability being superior to state-of-the-art cells (>10 years outdoor target) 40% reduction of production costs per Watt peak in comparison to existing DSC cell concepts
Achievements of the project: Part of the consortium followed a standardized approach for the processing of DSC on so-called masterplates with a common design and size to create a basis for compatibility of testing conditions and a comparable format for measurements and processing data. Three "baselines" at ECN, ISE and IVF to construct these masterplates are available in the consortium. Exploratory research in the direction of new concepts was carried out first on laboratory single cells of smaller areas. Several scientific exchanges were made for sharing information and exchanging ideas and knowhow. A number of young people got trained in the DSC field and improved networking contacts between partners as a result of Nanomax, has been made. This has also resulted in further collaboration in FP6 projects. These can be regarded as very important non-technical achievements of the project. The most important research highlights are summarized below: A maximum power conversion efficiency under full sunlight of 11 % for areas < 0.2 cm2 and 10.1 % for a cell with an active area of 1.3 cm2. Although this value is not yet certified by a calibration institute, the result can be considered as the highest efficiency ever measured for a DSC with an active area > 1 cm2. Maximum efficiencies up to 7.5 % were obtained for cells on masterplates with sizes of 2.5 cm2 using a photoelectrode design consisting of a double layer of transparent and scattering TiO2. Besides the hunt for higher efficiencies several new areas are explored in the direction of development of very novel concepts, fabrication protocols for new scatterlayers, metal oxide blocking layers, low temperature processes for Platinum deposition, electrical modelling, advanced characterization and long term stability with the following results: New Ru- containing sensitizing dyes with enhanced optical absorption in the visible part of the spectrum has been become available and were successfully applied in DSC Novel scatterlayers consisting of TiO2 with spherical voids are developed and applied in cells. Substantial improvement in performance was observed after comparison with cells lacking these voids. Optical modelling of this scatter phenomenon was accomplished. Proof of the novel concept "TCO-less design": 3.5 % efficiency has been achieved on masterplate level A novel co-sensitization strategy based upon the deposition of a secondary metal oxide layer between the absorption of the first and the second sensitizer dyes was explored Protocols of making metal oxide blocking layers on TiO2 resulting in retardation of recombination dynamics and improvement of the photovoltage of the device (5 % relative) Detailed understanding of interfacial behaviour of TiO2 layers of different origin in DSC made use of advanced characterization methods like transient optical studies, impedance spectroscopy and electrochemical analysis A 2D-electrical model was used for the simulation and evaluation of new concepts in terms of performance A low temperature process for deposition of Platinum on TCO using the POLYOL method resulted in counterelectrodes with low Charge-Transfer resistances ( <1 Ohmcm2) An >8 % efficient cell retaining over 98 % of its initial performance after 1000 h accelerated tests under thermal stress at 80oC in the dark has been demonstrated. Negligible device degradation was observed for 1000 h visible light soaking at 60 oC. A technical evaluation and cost analysis of the winning concept Aa and alternative low cost concepts have been performed and modules of different designs were demonstrated with a maximum active area efficiency of 5.5 % at full sun illumination. A large number of (joint) scientific publications
Publishable Report
2.
2.1
Objectives and strategic aspects

Socio economic objectives/strategic aspects
The dye-sensitized solar cell (DSC) technology is an upcoming thin-film PV technology that seems suitable for cheap mass production. All present knowledge indicates that DSC modules could reach the presented cost targets (photovoltaic modules aiming at 1.0 Euro/Wpeak by 2010). The project aim was to evaluate the full potential of this technology in terms of solar efficiency, long-term stability, materials development and next generation cell fabrication concepts. The generated know-how will stimulate the introduction of DSC on a large scale. The project was to demonstrate that through innovative solar technologies cost efficient photovoltaics are possible with relatively low investment costs. The main message from the project to the public was therefore to be that the general introduction of solar electricity is a good choice. In the nearer future, the costs of next generation photovoltaic modules could be brought down to a level that is comparable to the current price from the electricity grid. The project goal was to facilitate the political and technical introduction of cost efficient photovoltaic technologies within the presented range of cost target (system cost of 7 Euro/Wpeak and 3 Euro/Wpeak for the short and medium term, respectively and <1 Euro/Wpeak for the long term). 2.2 Scientific and technical objectives 2.2.1 General goal
The main objective of NANOMAX was to prove that at least 12% stabilised solar efficiency is feasible for DSC, if next generation cell concepts are applied. New innovative concepts, both for cell design and materials, were seen to be necessary to boost the small cell efficiency from present below 10 % to 15% in the future. Major improvements in the photocurrent and particularly the photovoltage were needed. 2.2.2 Scientific objectives
DSC are photovoltaic devices which are based on molecular dyes attached to nanoporous nanoparticulate TiO2, They efficiently make use of the dye's capability to inject photoexcited electrons into the conduction band of wide band gap semiconductor materials. It has been demonstrated that DSC can generate photocurrents with quantum efficiencies approaching unity over the whole visible spectrum. The factors limiting device performance and requiring new cell and materials concepts are: 1. Inefficient light absorption by existing sensitizer dyes in the near infrared (optimally up to 800900 nm). With optimised dyes, an increase of photocurrent from currently 16 mA/cm2 to 23 mA/cm2 seems feasible. 2. Sub-optimum photovoltage output. Current state of the art yields output voltages of ~750 mV, less than half the energy of incident photons, indicating improvements in photvoltage could up to double device efficiencies. Photovoltage output is limited primarily by the photon electron density, the surface recombination losses at the metal oxide/electrolyte interface, and series resistance losses. An increase of photovoltage from currently 750 mV to 900 mV seems realistic, if the surface recombination could be significantly reduced. Scientific methods identified to address these issues were: Introduction of band-edge engineering by making use of nano-crystalline metal oxides different from titanium dioxide (ZnO, SnO2, WO3, etc.), capped TiO2 and mixed oxides. Development of new sensitizer dyes with strong near IR absorption to be used in combination with existing dyes. Optimisation of light scattering in the photoelectrode, thereby increasing the optical density, i.e., the injected electron density.
Publishable Report
Improved electron transport through the metal oxide film, thereby reducing recombination losses. Alternative redox couples, minimising voltage losses at the dye / redox electrolyte interface. Retardation of interfacial recombination dynamics through optimised interface design. 2.2.3 Technical objectives
Technical objectives of the project were: Preparation of surface selective and IR-enhanced dyes as well as nano-crystalline oxides with reduced electron recombination. Better understanding of loss mechanisms in state-of-the art cells and for new materials. Achievement of certified 12% cell efficiency on an active area of 2.5 cm2 using innovative new cell concepts and realistic extrapolation of the experimental results towards 15 % efficient DSC cells. Demonstration of 10% efficient module (100 cm2 area) Demonstration of long-term cell stability being superior to state-of-the-art cells (>10 years outdoor target) 40% reduction of production costs per Wpeak in comparison to existing DSC cell concepts
Publishable Report
3.
3.1
Scientific and technical performance

Summary of specific project objectives
The following key deliverables were defined in the proposal in order to follow the project plan and reach the objectives.
Del.1 Deliverable title Delivery date2
D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 D11 D12 D13 D14
Nano-crystalline metal oxide materials suitable for testing of new cell concepts Summary from technical evaluation of cell concepts in terms of processing and costs Summary from electrical and optical modelling of new cell concepts Winning cell concept(s) showing realistic potential of 15 % efficiency Patent application for winning cell concept Nano-materials and dyes for winning cell concept Data of fundamental loss factors for cell efficiency determined by advanced characterisation methods Cell (at least 5 cells) with certified 10% efficiency using winning cell concept Long-term stable electrolyte composition for winning concept Cell (at least 5 cells) with certified 12% efficiency using winning cell concept Demonstrator module (10 cm x 10 cm) having 10% efficiency and using winning cell concept Process manual including realistic optimisation procedures for 15% efficient cells DSC specified accelerated ageing procedure Final cost calculation (materials and manufacturing costs) under the assumption of a 10 MWpeak per year pilot-line
8 11 11 11 11 20 24 24 24 35 35 35 35 35
1 Deliverable numbers in order of delivery dates: D1 Dn 2 Month in which the deliverable will be available. Month 0 marking the start of the project, and all delivery dates being relative to this start date.
Publishable Report
3.2
Overview of the technical progress
The project was organised as follows: WP 1: Concept development WP 2: Materials development and acquisition WP 3: Cell manufacturing / Proof of principles / Verification WP 4: Advanced characterisation and modelling WP 5: Long-term stability WP 6: Technical evaluation, cost calculation and module demonstration The project structure was determined by the work packages (WP 1-6), the project planning is ruled by the strategic topics, i.e. the new cell concepts and materials chosen in the project. Every strategic topic was addressed in the same structural way. This seems most convenient in order to split the responsibilities according to the specific strengths of the different partners. Also work like materials preparation could be performed simultaneously for the various topics, which was a cost-effective manner.
WP 1
Concept development Work package leader: EPFL
WP 2
Materials development Work package leader: ECN Strategic topics New cell concepts, new materials
No
WP 3 Cell manufacturing / Proof of principle / Verification Work package leader: FMF Winning concept(s) Cell with potentially 15% efficiency
Yes
WP 6 Technical evaluation, cost analysis and module demonstration Work package leader: IVF
WP 4 Advanced characterisation and modelling Work package leader: ICL
WP 5 Long-term stability Work package leader: ECN
Demonstrators:
1. Cell with stabilised 12% efficiency 2. Module with 10 % efficiency
Scheme 1. Project and management structure
A review of the work performed during the whole project is given on a workpackage basis.
Publishable Report
3.2.1
WORKPACKAGE 1 (Concept development)
The objective of WP1 was to define (a) winning cell concept(s) having the potential to achieve efficiency > 12 %. EPFL served as the activity leader of WP1, which ended in the first year as envisaged in the proposal. However the decisions taken in WP1, were regularly being monitored and followed throughout the project period. This work package started on month 0 and ended by month 11. During the project kick-off meeting held at ECN, it was decided to hold a workshop in May 2002 to particularly focus on this important WP, which had a very decisive role in the project outcome. In the proposal, 3 novel photoelectrode designs were presented in the proposal apart from the Standard photoelectrode design (ST): Concept (Aa) Thin oxide semiconductor layer with an efficient back scattering layer Concept (Ab) TCO-less design, having a porous metal layer used as a secondary electrode. Concept (B) Multilayer device using surface selective dyes with different bandgaps A schematic representation of the four concepts are shown in Scheme 1:
Nc-oxide + dye(s) TCO
Standard (ST)
white ZrO2
Glass
Nc-oxide + dye(s) TCO
Thin (Aa)
Glass
Porous Ti (metal) white ZrO2 Nc-oxide + dye(s)
TCO less (Ab)

Porous TiN
Glass
Nc-oxide + dye 1+ dye 2 Eg < Eg > transparent ZrO2 Nc-oxide + dye 1
Multilayer (B)
TCO Glass
Scheme 1 Subsequently, EPFL had invited and received more new ideas for including in a discussion paper, which was circulated among partners before the workshop. The meeting was later held at EPFL, Lausanne on May 23 24, 2002. Following were the main objectives of the meeting: To review the progress of WP 1 To discuss and analyse various ideas contained in the proposal and in the discussion paper (attached to the minutes of this meeting) To choose only the best and most promising ideas for the project (output of WP 1)
10
Publishable Report
At least 5 new sets of contributions were received from various partners to be listed in the discussion paper. All these new cell ideas and concepts have been thoroughly discussed and analysed along with the concepts already presented in the project proposal. It was important for the project to limit the number of ideas and the main aim of WP1 was to narrow down the various concepts to three or four, such that the partners would be able to carry out their activities in a well focused direction as decided collectively during this meeting. Table 1 presents a brief tabulation of the most important ideas and the allocation of responsibilities reached during this meeting:
Table 1: Collection of ideas after discussion WP1 on 23rd of May, 2002, Lausanne
Idea/concept Hinsch Aa
Add remarks - Masterplate level glass frit (high currents) vs. Surlyn/Bynel - Rutile vs. ZrO2 scattering layers: delivery of Rutile paste to ECN and FMF for masterplate fabrication - Nanofibres Hinsch Ab Proof of concept on single laboratory cells Michael Grtzel - Blocking hydrophobic dye (EPFL) - Atomic Layer Depostion at TUD: Conformal deposition of blocking layers on laboratory cells - New redox systems (Co+/Co2+) - Proton control Tests on single cells and masterplates CUT FMF (airspheres) IVF ICUK-blocking layers
RESPONSIBLE ECN/FMF-ISE EPFL EPFL FMF
General Discussion
EPFL/ECN/FM F - Porous Titanium substrates; test by EPFL on single CUT/ EPFL cells CUT/ECN - POLYOL-Pt Airspheres as scattering centres; Japanese materials to EPFL be tested by EPFL Exploration on nanofibres ICUK/IVF ICUK/FMF/EC - Single cells/Masterplates (Mg-acetate) N - Oriented metal oxides ICUK/IVF - Two discrete materials (low priority) - Carbon rods + TiO2 EPFL/ICUK/IV - Dye cocktails F EPFL/ICUK EPFL/ECN/ ECN - Thinner Surlyn (from 50 micron to 25 ??) ECN/EPFL - Glass sealed plates to EPFL for test of materials EPFL/ECN - Dye adsorption without O2 exposure High concentration Dye solution, Dye N712 (fully deprotonated) to be tested on single cells and masterplates
The meeting further identified all the important factors affecting the DSC efficiency and ways to ameliorate the cell performance. Achievement of higher open circuit voltage (Voc), low dark current, better fill factor and an optimum H+ balance inside the cell were some of the ongoing priorities at EPFL. Ways and means to achieve higher fill factor and current densities have been identified. Syntheses of new dyes, absorbing most of the visible wavelengths and even a portion of near IR, were achieved in EPFL during the first year.
11
Publishable Report
3.2.2
WORKPACKAGE 2 (Materials Development)
The main objective of this WP was to provide all the materials necessary for proof of principles of the various new concepts and for advanced characterisation (WP 3 and 4). The materials that were needed for testing the concepts as defined in WP1 are various nanocrystalline metal oxides, new dyes, electrolytes and novel additives etc. In this section, a summary is given of the activities that are carried out in this WP aimed at the delivery of materials to be tested in the WPs 3-6. 3.2.2.1 Nanocrystalline materials Nanocrystalline TiO2 Four partners, i.e. ECN, Greatcell, EPFL and FMF/ISE are able to make screenprintable TiO2 pastes for the manufacturing of small cells and masterplates. Greatcell has scaled up manufacture to 500g batches. The colloids are synthesised following hydrothermal sol-gel procedures via either acidic or basic routes. Particle sizes can be varied by changing the synthesis conditions. The selfmade TiO2 colloids as well as commercial colloids like Millennium and P25 are transformed into screenprintable pastes, based on formulations containing terpineol and ethylcellulose as organic media/binders. An analysis of the influence of acid/base conditions employed in the synthesis of TiO2 nanoparticles upon the performance of DSC fabricated from these particles was done and published. It was found that base peptization results in a reduction in film electron density under negative applied bias, correlated with slower interfacial recombination losses and a higher device open circuit voltage [1]. EPFL demonstrated that optimal cell results were obtained with TiO2 particles having a basic particle size of 20 nm and increasing layer thickness to more than 15 micron when using N719 as the sensitising dye. In addition 5-micron thick scattering layers were used. The preparation procedure of the pastes containing the 20 nm particles was published and transferred to the project partners during an instruction workshop held at EPFL, December 2003. Scattering layers For the scatterlayers, powders are used with particle sizes between 100 and 400 nm. Most of these materials are commercially available. The powders that are used in a number of experiments are: ZrO2 (Degussa), TiO2-anatase (Fluka), TiO2-Rutile (Bayer), TiO2-Anatase (ECN-made) and TiO2anatase (CCI, Japan). In addition, spherical voids as light scattering centres in a TiO2 film were realized with polystyrene particles of diameter of 400 nm. A cross section of a TiO2 films containing spherical voids are shown in Figure 1.
1.0m
5.0m
Figure 1. Spherical voids are left behind in the film of TiO2 when the polystyrene spheres are burnt out during the heating or the sintering of the films are 450oC. From ref.[2].
12
Publishable Report
A proof of principle of this concept was demonstrated by a direct comparison of the IV characteristics of devices derived from transparent TiO2 films and films containing hollow scattering spheres [2].
Table 2. Comparison of IV data of two cells consisting of screenprinted transparent TiO2 layers with and without hollow scattering centres. Dye: N719; electrolyte 0.6 M tert-ButylammoniumIodide (tBAI), 0.1 M LiI, 0.1M I2,0.5 M tert-Butylpyridin (tBP) in acetonitrile (ACN)
Isc/mAcm-2 nc- TiO2 Transparent TiO2 12 Hollow Scattering 13 spheres
Voc/mV 740 740
FF(%) 61 69
(%) 5.4 6.7
Electrical Impedance Measurements demonstrate that the films can successfully be used for devices with highly viscous electrolytes. Optical modelling was done on the various scatter concepts and materials. Conformal coatings of nanocrystalline electrodes A wet-chemical method was developed for coating TiO2 particles with other metal oxides, with higher band gap properties, acting as insulator barriers with the aim to retard recombination dynamics and increase Voc. Different metal oxides (Aluminium Oxide, ZrO2, SiO2 and MgO) have been used to compare the different properties [3]. Optimum deposition conditions were found to depend on properties of underlying film, attributed primarily to differences in the acid/base properties of the underlying film. The blocking character of conformal coatings worked best on acidic films and poorer quality films. Studies focussed on Aluminium oxide and MgO coatings. MgO coating could result in higher performances than Aluminium oxide but not reproducibly. The exact composition of the Aluminium Oxide films may vary depending on residual hydrated species and hydroxides. Another method to coat TiO2 particles with alumina or other oxides was investigated using atomic layer deposition (ALD) Suitable expertise for ALD was available in Delft University of Technology, The Netherlands. Dr. A. Goossens of TU Delft, who is familiar with this type of work, had kindly agreed to execute this work as a sub-contractor (6 person months in Delft). Several ALD Aluminium Oxide treated 5m TiO2 coated transparent electrodes were received from Delft, where the shell thickness varied between 1.2 6 nm (based on 0.2 monolayers/cycle with a monolayer thickness of 0.3nm). They have been used for optimising the experimental parameters. However, the desired beneficial effect was not observed. The study revealed the following facts. Initial tests on the plates revealed that a second sintering at 450C had no effect on the function of the Aluminium Oxide layer indicating that in this case the original coating was Aluminium oxide. The electrolyte 1376 (modified with 0.05M LiI in acetonitrile/valeronitrile) proved most effective with respect to the observed current density in the film. The dye Z-907 (see for structure figure 3) demonstrated the best chemisorption behaviour on the Aluminium oxide film, seen from solar cell IV data. Cyclic-voltametry has also been implemented to test the effectiveness of the Aluminium oxide layer. The results however could not be fitted to the IV and IPCE data. The efficiency of the Aluminium oxide treated cells was inversely proportional to the layer thickness. In all the cells, produced with the Aluminium Oxide blocking layer, slightly higher potentials were reported. However the overall current density in the film was reduced which in turn reduced the efficiency. The reduction in the current was almost certainly a direct result of the tunnelling distance through the Aluminium Oxide film. The IV curves also confirmed this observation with an exponential reduction in both the current and the efficiency, as the blocking layer thickness was increased. The ALD showed no positive effect in achieving higher DSC efficiency of the liquid junction cells, and was not continued further.
13
Publishable Report
Mixed oxides Within the framework of this project several mixed oxides have been synthesized. Mixed oxides of AlxOy -TiO2 as well as MgO-TiO2 were synthesized using sol gel synthesis similar to that of the synthesis of TiO2. Titanium isopropoxide and aluminium sec butoxide were used in the ratio of 1:1. Magnesium methoxide with titanium isopropoxide in the ratio 1:4 was synthesized as well. In both cases, peptization was done at 190oC. The idea behind the synthesis of these oxides was to increase the position of the conduction band in the semiconductors such that the open circuit voltage could potentially be increased. Potential application of these new mixed nanocrystalline oxides have not been fully explored. UV degradation of photosensitive dye such as methylene blue on MgO-TiO2 gave evidence of no photo bleaching over a period of 24 hours under UV. Preliminary tests carried out with some of these materials (MgO-TiO2) in DSC resulted in very low currents. Other oxides There has been wide interest in replacing TiO2 with other metal oxides such as Nb2O5 and ZnO. The most interesting replacement has been in the use of SnO2. SnO2 has a band gap and conduction band edge similar to that of TiO2 but shows faster transport of electrons. To date, no one has been able to improve on efficiencies obtained using TiO2, even when the SnO2 electrode is coated with a wide band gap metal oxide such as MgO [4]. A comparison of charge transport and recombination dynamics in dye-sensitized solar cells employing nanocrystalline TiO2 and SnO2 films has been made, and the impact of these dynamics upon photovoltaic device efficiency will be given in section 3.2.4.2. New morphologies: Synthesis and characterisation of TiO2 and ZnO nanorods. A range of strategies was investigated to fabricate TiO2 nanorods and evaluate the electron transport function of these rods. Some success was obtained in solution routes to TiO2 nanorods as illustrated below. However, attempts to fabricate mesoporous films from such rods were unsuccessful so far, and therefore devices were not fabricated. Transport studies of vacuum deposited orientated TiO2 nanorods indicated upto 3 fold acceleration of electron transport dynamics. In Figure 2 SEM pictures of TiO2 and ZnO nanorods are shown.
Figure 2. SEM image of anatase TiO2 nanorods obtained from 23-hour treatment of rutile TiO2 powder. rods fabricated from anatase (left) and ZnO nanorods (right)
3.2.2.2 Sensitizer In most of the experiments, Ru(dcbpy)2(NCS)2 with 2 protons, (N719), is used as the standard red dye and was obtained from different suppliers, i.e: 10 gram gifted from partner Greatcell, and commercially via Solaronix and INAP (earlier batch from the LOTS-DSC project) and EPFL. Other sensitising red dyes which were made available by EPFL and Greatcell are Ru(dcbpy)2(NCS)2 with 0 (N712), 1 and even 1.5 H+. By adjusting the number of protons in the dye, it is possible to modulate the current density and Voc to optimum values. It has been further shown that improved performances could be obtained after repetitive purification of N-719 dye using a Sephadex column (three times).
14
Publishable Report
Furthermore, the hydrophobic dye Z907 (see Figure 3) is of interest in view of its promising longterm stability at elevated temperatures in combination with special formulated electrolyte combinations [5]. The presence of the aliphatic chains apparently prevents/inhibits water adsorption to TiO2 and could retard recombination with I3-.
C9H19
HOOC N N Ru N HOOC N C S N C S N
C9H19
Figure 3. The molecular structure of Z907
While improvements in performance have been obtained using self-assembly assisting agents are remarkable (as will be shown later) it will be very difficult to achieve efficiencies much beyond 11% with the standard N-719 sensitizer unless the redox system is changed. Alternatively if the present iodide based redox system is maintained, introducing panchromatic sensitizers to enhance the spectral response of the dye towards near IR region may boost the efficiency of cells further. To give 15 % conversion efficiency, the dyes should be designed to yield at least 23 mA/cm2 short circuit current under full sunlight and fill factor as well as open circuit voltage values similar to those that are presently obtained. Several new Ru-photosensitizers were introduced containing tetradentate ligands to shift the absorption more to red/infrared part of the spectrum compared to N719. However, some of these Ru complexes containing tetradentate ligands did not show the expected improvement in solar cell performance so far, mainly due to solubility limitations. Most recently, new Ru-sensitizers with improved light harvesting for DSC have been synthesized by the EPFL group by extending the -conjugation of the hydrophobic ligand and endowing it with electron donating alkoxy groups These new dyes were demonstrated as a highly efficient, stable sensitizers for nanocrystalline solar cells. Two examples of these new class of sensitizers, i.e. Z910 and K-19, with enhanced optical absorption are shown in Figure 4 [6,7]. The K-19 dye shows excellent yield and stability, when tested with various electrolytes. This dye will be briefly revisited, in the WP5 section, when discussing issues related to cell stability.
COOH N HOOC N NCS
Ru
N N
NCS
OCH 3
OCH 3
Figure 4. The molecular structures of Z910 (left) and K19 (right) with an extended -sytem in the ligand
15
Publishable Report
Many more dyes were developed and tested, during this project period. Additionally, dyes in the lass of porphyrins have been obtained from other laboratories and performance was evaluated in DSCs. It has been demonstrated that the most efficient porphyrin was Zn-metalloporphyrin, in the presence of an electrolyte containing I-/I3-, which achieved very efficient sensitization yielding close to 75% incident photon-to-current efficiencies (IPCE) at the Soret peak and 50-60% at the Q band peaks in a 7 micron TiO2 film. Under standard AM 1.5 sunlight, the metalloporphyrin sensitized solar cell exhibited a short circuit photocurrent density of 9.7 0.5 mA/cm2, an open circuit voltage of 660 50 mV and a fill factor of 0.75 0.05, corresponding to an overall conversion efficiency of close to 5%, making it the most efficient metalloporphyrin sensitized solar cell reported to date [8]. 3.2.2.3 Electrolytes The electrolyte mixtures that have been used in high-efficiency cells mainly consist of alkylimidazolium iodides or Tetraalkyl ammoniumiodides, LiI, I2 and surface additives such as tertbutylpyridin (TBP), Benzimidazole derivatives (benzimidazole, N-Methylbenzimidazole) in solvents such as acetonitrile, propionitrile or mixtures of acetonitrile and valeronitrile. Novel additives A number of additives have been discovered, which are able to facilitate self-assembly of the dye molecules dye at the TiO2 electrode surface, rendering it more impermeable resulting in reduced dark current of the cell. Examples are carboxylic (like hexadecylmalonic acid) and phosphonic acid derivatives (like decylphosphonic acid). In addition, several cationic species have been investigated. Amongst those, guanidinium and imidazolium ions, had shown the most promising results in terms of efficiency improvements (see Figure 5a and b). Replacing for instance LiI by guanidinium thiocyanate in standard electrolyte compositions reduces the dark reduction current at the mesoporous TiO2 electrode significantly, resulting in an increasing cell voltage (> 800 mV) and overall efficiency of the device (5-10 % relative). Another additive that has shown to be successful when added to a highly purified N-719 dye is tetrabutyl ammonium chenodeoxycholate (Figure 5c), a bile acid derivative when added to a highly purified N-719 dye, produced an 11% efficient DSC.
R N R N R + N R R R
ABT+
-O
a) R = H, Me, Et X- = I-, SCN-
c)
HO H3C CH3
CH3
N + N R
H H H
X-
b) R = Me, C6,C10 X- = I-, SCN-
HO H
Figure 5. Molecular structures of a) Guanidinium and b) imidazolium slats and c) tetrabutyl ammonium chenodeoxycholate
3.2.2.4 Counterelectrodes As the standard counterelectrode, SnO2:F substrates coated with a catalytic amount of platinum are used. The standard counterelectrodes of the masterplates are obtained by screenprinting a platinum based paste, after which a heat treatment is applied to obtain homogeneous layers of Pt on SnO2:F. Another way is to apply the Pt-layer by solvent casting of organic platinum containing solution and subsequent heat treatment. Charge Transfer (CT) resistances are < 1 .cm2.
16
Publishable Report
At the CUTECH, an alternative method to manufacture Platinum counter electrodes was developed because of its potential cheapness and productiveness. In this method, called POLYOL, a salt of metal (i.e. PtCl4 or K2PtCl6) is diluted in a mixture of organic solvents and by a reduction process carried out at temperatures ranging from 90 to 160 oC, Pt is precipitated directly on the supporting material. The advantage of this process is formation of a fine, highly nanoparticulate sized layer of selected metal on the substrate at a relatively low temperature. It has been shown that a reaction temperature of 90 oC is in principle enough to obtain sufficient catalytic activity to prepare counterelectrodes with a suitable CT resistance (RCT) to be applied in high efficiency solar cells. The lowest Charge Transfer resistance was around 0.5 cm2. A subsequent heat treatment at 200 o C lowers the RCT to 0.36 cm2 while further increasing the temperature could lead to crystal growth of the Platinum and increase of RCT . The biggest advantage of this method compared to the standard methods of deposition is that much less Pt-precursor is used (up to 10 times). In addition, screenprintable pastes have been prepared using the nanoplatinum powders from the POLYOL process. Preliminary results indicate that relatively high amount of platinum (35.17 g/cm2) on the CE is necessary to reach a RCT of 1.4 cm2 while immersion in a POLYOL solution leads to RCT of 0.36 cm2 with a Pt coverage on the CE of only 3.7 g/cm2. Further research is in progress at CUTECH to optimise the screenprintable paste to make fully use of the benefits that the POLYOL method offers with respect to other standard preparation methods. 3.2.3 WORKPACKAGE 3 (Cell manufacturing, proof of principles)
The main objective of WP3 was to develop, characterise and test the various cell concepts that were proposed during the WP1 meeting in Lausanne May 2002. It was already decided at the kick-off meeting to go for a standardised approach in the manufacturing of the cells. As standard approach, 5 test-cells (active area 2.5 or 4 cm2) are produced simultaneously on a common plate (master plate) so that the proposed concepts from WP1 can be tested.
Figure 6. "Masterplate" containing five single DSC with an active area of 50 x 5 mm2.
The masterplate fabrication is presently carried out at ECN, FMF/ISE and IVF via semi-automated baselines. It was also decided that exploratory research in the direction of new concepts is carried out first on laboratory single cells of smaller areas since not all the partners (EPFL, ICL and CUT) have the equipment to make cells on masterplates. Several experiments were carried out to optimise the processing conditions for the masterplate fabrication, to optimise the performance of concepts ST and Aa by variation of the materials (TiO2, scatterlayers, Dye, Electrolyte, capped oxides) and to prove the principles of Concept Ab. The most important project achievements are summarized below: 3.2.3.1 Concepts ST and Aa I-V characteristics of single cells A large number of small test cells (< 1 cm2) were prepared to test the materials as described in section 3.2.1 and to apply improved fabrication protocols. Concept Aa (See Scheme 1) was chosen as the main working horse.
17
Publishable Report
This has resulted in non-certified AM1.5 power conversion efficiencies of 11.0 % in full sunlight for cells prepared at EPFL with an active area < 0.2 cm2 .The improved performance is a result of using a very pure N719 dye in combination with a self-assembly facilitating adsorbent, such as guanidinium thiocyanate (GuSCN) and improved cell preparation protocols. As a result, the Voc increased to 0.86 V without sacrificing the short circuit photocurrent. Therefore, GuSCN appears to block the dark current, i.e. the reduction of triiodide by conduction band electrons, leading to the observed augmentation in the Voc value. One of the aims of the consortium was to achieve a certified world record efficiency on cell areas > 1 cm2. Attempts to translate the high efficiency results from the small sized cells to single cells with active areas > 1 cm2 has resulted in a 1.3 cm2 device with a non-certified AM1.5 power conversion efficiency of over 10 % measured in the EPFL laboratories using well established standard measuring techniques[9]. This value is significantly higher than the 8.2 %, currently regarded as the world record efficiency for a DSC with an area > 1 cm2 [10]. The I-V curve for this top device is shown in Figure 7.
Figure 7. I-V characteristics of a 1.31 cm2 DSC top-cell prepared at EPFL, with antireflector coating; Dye: N719 GSA based (1H+) 4 times purified; Electrolyte: 0.6 M Butyl, Methyl Imidazolium iodide (BMII), 0.1 M Guanidinium Thiocyanate, 0.03 M I2, 0.5 M Tert-butylpyridin in acetonitrile/valeronitrile 85:15 volume ratio. MPP = Maximum Power Point. High transmission TCO-glass (Nippon Sheet Glass, 10 Ohm/square) was used provided with an antireflector coating. TiO2: 16 micron TiO2 (20 nm particle size) + 4 micron scattering TiO2 scattering layer (400 nm particle size).
I-V characteristics of masterplate cells Part of the consortium follows a standardized approach for the processing of DSC on so-called "masterplates" (see figure 6). Masterplates were constructed on semi-automated baselines now available at ECN, FMF/ISE and IVF, with the aim to do materials research on cells which reaches a high yield and reproducibility to investigate the small deviations that can be expected upon changing material parameters/production pathway. It should be noted that these masterplates are not suitable for practical outdoor use. The upscaling of DSC to larger area modules of at least 100 cm2 is subject of discussion in section 3.2.6.4. The cell design shown in Figure 6 was optimised to minimize resistance losses through the TCO substrates thereby maximizing the Fill Factor. A large number of masterplates has been made in this project. N719 dyes of different origin and purity, nanocrystalline TiO2 and Platinum screenprintable pastes have been prepared following different procedures. The best I-V data for masterplate cells using Concept Aa fabricated at ECN are shown in Table 3 using SnO2:F substrates as the front electrode from different suppliers varying in sheet resistance and transmission. LOF-TEC 8 glass (3mm, 8 / ) and LOF-TEC 15 glass (2.3 &3 mm 15 / ) are SnO2:F substrates commercially available (from Hartford Glass Company and provided as a gift by STI) while Sx TCO glass (1mm, 9 / ) is a research sample obtained as a kind gift from Solaronix SA (Aubonne, Switzerland).
18
Publishable Report
Cell assembly Front electrode TiO2 / scatter Dye Electrolyte CE: Sealant
: : : : : :
Different TCO's 9 m ECN B42/ 2 m ECNB38/ 6 m ZrO2 N719 from Solaronix 0.6 M HMII, 0.1M LiI,0.05 M I2, 0.5 M tBP in Acetonitril SnO2Sb/Pt (MPs1-4) and Pt-Polyol (MP5) sintered at 420oC Surlyn 1702
The transmission of the SnO2:F substrates in the wavelength region where the dye absorbs increases in the order TEC 8 < TEC15 < Sx TCO which is reflected in the measured current densities increasing from 14.3 mA/cm2 TEC8 to 15.0 mA/cm2 for Sx TCO (probably due to the use of thinner glass with lower absorption.
Table 3. I-V data at 1 sun:
MP. No 1 2 3 4
TCO TEC 8/3 mm TEC15/3 mm SxFTO/1 mm
Jsc(mA/cm2) Voc 14.3 14.8 15 (14.3*) 14.9 0.708 0.716 0.710 (0.725*) 0.723
FF 0.7 0.69 0.69 (0.69*) 0.69
(%) 7.09 7.30 7.35 (7.52*) 7.43
TEC15/2.2 mm 5 TEC15/2.2 14.8 0.715 0.69 7.30 mm * obtained for another device based on SxFTO measured at 0.95 sun equiv.
A maximum power conversion efficiency of almost 7.5 % was achieved for the highest transmission glass. Masterplate 5 containing a Pt counterelectrode prepared with the POLYOL method showed comparable high fill factors as obtained for devices prepared with conventional methods (see WP2 section). It should be noted that there appears a gap in performance between the smaller sized single cells as fabricated by EPFL and the cells fabricated on masterplates by the other partners. The effort to translate the higher efficiencies reported for small cells to masterplate level dimensions received much attention during the project. The reasons behind the wide difference between EPFL cells and the remaining partners have been investigated and identified. The reasons mostly pertain to differences in TiO2 pastes, particle size of the scattering layer, Ohmic losses in the glass plate, iodine/iodide concentration, dye purity and the addition of surface modifiers to the dye solution/electrolytes. Another issue that should be addressed is the variability in printed film quality. Pastes designed for small cells do not necessarily behave the same rheologically and when drying as printed pastes for larger widths (up to 1 cm). A loss of at least 1% efficiency is not uncommon due to claypan crazing during drying of thick pastes in larger areas. This indicates that results of the project for small cells are a more valid reflection of the status of the technology than those of larger cells and masterplates, for which further materials engineering effort will be required by industry. 3.2.3.2 Concept Ab (TCO less concept) In the novel concept Ab, nc-TiO2 layers are deposited on a glass substrate without SnO2:F resulting in a better optical efficiency. A secondary electrode, consisting of a porous metal, is deposited on top of the TiO2 layers and acts as the current collecting layer. Cells of this novel TCO-less design were made on masterplate level. Proof-of-principle was demonstrated for this concept with efficiencies of 3.6 % (see Figure 8). The concept has been further evaluated by IVF in WP6.
19
Publishable Report
TCO/Pt
10
-2 current density/mAcm
TiO2-particle with adsorbed dye molecules Electrolyte Sputtered Ti-layer; ca. 150nm
8 6 4 2 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 voltage/V
= 3.6%
glass
Figure 8. Setup of cells with a secondary electrode made of sputtered porous Ti (left) and the IV data @ 1 sun for Concept Ab cells prepared on masterplates
3.2.3.3 Other concepts (variation of Concept B) Co-sensitization Attempts to sensitize TiO2 electrodes with multiple dye cocktails with different absorbing properties have achieved only limited success to date. In this project, an alternative approach to achieving a molecular electron transfer cascade at a dye-TiO2 interface, employing two spatially separated layers of distinct sensitizer dyes to achieve a vectorial movement of the oxidized dye species away from the TiO2 surface. This novel co-sensitization strategy is based upon the deposition of a secondary metal oxide layer between the absorption of the first and the second sensitizer dyes, as illustrated in Figure 9.
E[V] vs SCE
TiO2
CB
eS /S
+
Al2O3
*
-0.8
S+/S*
0.0
+0.8
S+/S S+/S
I- / I3-
Figure 9. Charge transfer processes in multilayer co-sensitized nanocrystalline TiO2 films
An initial study was completed with two model dyes (Ru(dcbpy)2(NC)2 and RuPc) and reported in literature [11]. Transient absorption and photovoltaic data confirm this is a potentially attractive route to co-sensitization of nanocrystalline TiO2 films. Further development of this strategy employing blue and red absorbing sensitizer dyes therefore presents an attractive route to the panchromatic sensitization of DSC if the second coating can be deposited as a stable nanocrystalline oxide. Tandem concepts Besides the above mentioned co-sensitization strategy to extend the absorption range of the solar spectrum, tandem-cell like structures could help in achieving higher efficiency values than can
20
Publishable Report
maximally be obtained for a DDSC based on one sensitized electrode. Recently, tandem cells are described in literature consisting of two different sensitizer dyes in two different compartments of the cell [12]. It was shown that the total efficiency of the tandem device was higher than the efficiencies of the single compartments. In order to make fully use of the potential that a tandem structure can offer, sensitising dyes should be used with complementary absorbing properties and high IPCE in the relevant wavelength regions. First studies to assemble tandem like structures were carried out successfully by EPFL. A DSC was assembled with an inorganic solar cell absorbing far more in the (infra) red part of the solar spectrum thereby increasing the efficiency from 8 % for the single DSC to over 12 % for the tandem device structure. Further attempts to the realize tandem cell consisting of two DSC are ongoing at EPFL. 3.2.4 WORKPACKAGE 4 (Advanced Characterisation and modeling)
The objective of WP4 was to develop and test quantitative models of device performance for targetdirected development of next generation DSC. This work package focused on the following topics Evaluation of the factors limiting the efficiencies of different device concepts Evaluation of different materials developed for these device concepts Development of models of materials and device function and comparison against experimental data. Assessment of the efficiency and power potentials of different the different device concepts.
The research activities undertaken in this WP comprised the partners and tasks detailed in the project proposal. Partner P3 (ICL) led the WP with additional inputs from partners P2 (Cracow), P5 (FMF) and P6 (EPFL). ICLs research has included on transient optical studies of interfacial electron transfer dynamics (Task 4.2), numerical modeling of such dynamics (Task 4.4), device I/V modelling and transient photocurrent / photovoltage studies (Task 4.2). Cracow studies have focused on impedance analyses of novel Pt counterelectrodes (Task 4.1). FMF have focused on frequency domain studies of device function (Task 4.1) and light scattering studies (Task 4.3). EPFL have focused on I-V analyses and consideration of limitations to device efficiency (Task 4.4). All WP milestones and deliverables have been achieved according to the proposal schedule, as detailed in previous year reports. Consistent with the basic science focus of this WP, the research undertaken in this WP is leading to 14 publications in refereed journals, as detailed in attached the publication list. The work undertaken in this WP has focused particularly upon in situ characterisation of dye sensitized photovoltaic devices under operation, using not only established frequency domain techniques, but also transient optical studies including all key steps of interfacial charge separation and recombination dynamics and transient photovoltage and photocurrent studies of transport dynamics. These experiments have been correlated with I-V data in the dark and under illumination and with a numerical model of these I-V data based upon the non-ideal diode equation. These methodologies have been employed to characterise fundamental loss factors both in standard (ST) devices and a range of innovative device concepts as developed in this proposal. For standard devices the primary loss mechanism is confirmed as charge recombination from electrons in the metal oxide to the oxidised redox couple. 3.2.4.1 Limits to Efficiency Scientifically, there appears to be no reason why efficiencies of 15% or higher cannot be achieved in DSC. It has already been demonstrated that DSC can generate photocurrents with quantum efficiencies approaching unity over the whole visible spectrum. Factors currently limiting device performance are:
21
Publishable Report
Inefficient light absorption by existing sensitizer dyes in the near infrared (700 900 nm range). The latest dyes tested such as K19 are already showing good progress in this direction. Furthermore co-sensitization strategies employing dye multilayer structures (see WP3 above) with one dye optimised for far-red/near IR absorption show good promise this direction. With such optimised dyes, an increase of photocurrent from currently 16 mA/cm2 to up to 23 mA/cm2 seems feasible. Sub-optimum photovoltage output. Current state of the art devices yield open circuit voltages of ~750 mV, less than half the energy of incident photons. Experience with inorganic p-n junction solar cells shows that open circuit voltages of the best cells are typically 0.4 V smaller than the potential corresponding to the band gap, which suggests that improvements in the photo voltage of dye sensitized solar cells could substantially increase device efficiencies by up to 50%. Open circuit voltages are limited primarily by surface recombination losses at the metal oxide/electrolyte interface. Strategies to reduce interfacial recombination losses, including the use of inorganic and sensitizer dye barrier layers and optimisation of electrolyte composition, as detailed in WP3 above, have already demonstrated the potential for significant increases in cell potential to > 900 mV. Fill factor. Fill factors (FF) for DSC are limited in general by series resistance losses, lightintensity dependent recombination, non-ideal dark diode currents and, in some cases, shunt resistance losses. For optimised liquid-electrolyte devices, modelling of I-V plots using one- and two-diode models indicates that the primary limitations on the FF are the nonideal dark current behaviour and, for large cell areas, series resistance losses. Dark current losses are already being minimised by strategies designed to optimise VOC as discussed above, resulting in FF values of up to 0.75 for small area devices. For larger devices, such as masterplates, series resistance losses in the TCO result in slightly lower FF (0.67-0.7). A key challenge to optimisation of overall device efficiency is the inverse correlation between Isc and VOC often observed in empirical studies of device optimisation. Our studies of interfacial electron transfer dynamics now provide a sound fundamental understanding of this inverse correlation, and are providing new routes to overcome this limitation, allowing systematic device optimisation. Employing such strategies, it can reasonably be expected that performance parameters (at least for small area devices) of Isc = 22 mA cm-2, VOC = 0.9 V and FF = 0.75, leading to = 15 % should be achievable with existing device designs. Our studies are already indicating that the ultimate limits to efficiency may be associated with the issue of inhomogeneous interfacial energetics, and the resulting dispersive kinetics [13]. However innovative strategies, such as the use of supramolecular dyes are providing promising approaches to overcome such dispersive kinetics and may ultimately lead to device efficiencies significantly in excess of 15 %. 3.2.4.2 Highlights of WP4 research activity
Interfacial electron transfer kinetics We have completed a detailed study of interfacial electron transfer kinetics in complete DSC, and demonstrated a clear inverse correlation between electron injection kinetics from the sensitizer dye excited state and interfacial charge recombination kinetics [13]. For variations in electrolyte composition, this inverse correlation is shown to result from variations in the TiO2 density of states possibly due to anion impurities, hydrated species, variation in surface oxide stoichiometry, and related variations in surface crystallinity. For a device with optimised electrolyte composition, and therefore optimised device efficiency, the electron injection kinetics were found to be two orders of magnitude slower than previously expected, being just fast enough to compete with excited state decay to ground (see figure 10). The dispersive nature of these kinetics is shown to be consistent with a model of inhomogeneous energetic disorder, whilst the direct correlation between interfacial kinetics and device efficiency, with the appearance of optimum charge separation kinetics is shown to be in good agreement with theoretical device models.
22
Publishable Report
1.0
(i) (i) (ii)
Injection Yield
0.5
(ii)
0.0 10
0
Figure 10. Charge separation dynamics in a dye sensitized TiO2 film (ii) and a complete fully functioning solar cell (i). The smooth red lines are simulations of the experimental data based upon a model of inhomogeneous interfacial energetics. From ref. 13
10 10 10 10 time / picoseconds
10
Metal oxide function One of the previously considered promising routes to improving device performance is to replace the TiO2 electrode with a structurally analogous material with higher electron mobility such as SnO2. To date such studies have been unsuccessful. By a combination of device I/V modelling (see figure) and interfacial electron transfer studies, we have been able to develop a clear understanding of the correlation between device performance and electron mobility [4]. In particular we have found that the faster charge recombination losses resulting directly from the increased electron mobility can result in the appearance of new efficiency losses due to electron recombination to dye cations, and result in lower device performance. We have thus been able to identify clearly the limitations in the use of such materials, and in particular identified that such materials could only be used with alternative electrolytes with faster dye re-reduction kinetics. The TiO2 data in the light and dark show an excellent fit to the one-non-ideal diode model, for SnO2 the fit is very poor, consistent with the enhanced recombination losses to dye cations observed in our transient studies.
12 10 (a)
-2
Current / mAcm
8 6 4 2 0 -2 0.0 (b)
0.2
0.4
0.6
0.8
Voltage / V
Figure 11. Current-voltage characteristics for DSCs employing 4 m thick, non-scattering (a) TiO2 and (b) SnO2 nanocrystalline metal oxide films and matched dye loadings. Data collected under simulated AM1.5 irradiation (black lines) and in the dark (grey lines). Dashed lines are fits to the data based on the non-ideal diode equation. From ref. 4
23
Publishable Report
3.2.5
WORKPACKAGE 5 (Long-term Stability)
The main task in this WP was to test the long-term stability of the developed concepts by identifying critical stress factors and searching for dye-electrolyte combinations that permits longterm stable cells. Internationally accepted (accelerated) ageing tests have not yet been established for organic-based solar cells, such as the liquid-junction DSC, which is understandable considering the stage of development of these cells. In fact, the only thin film technology for which accelerated ageing tests have been developed into an IEC standard is amorphous silicon and even this standard has not be translated to product specifications (e.g. BIPV). These test protocols are normally developed for professional use including extreme test conditions like Annealing (85 0C, dark), Damp heat (85 0C, 85 % RH), Humidity Freeze (-40 to 85 0C), Thermal cycling, Light soaking. For the research carried out in this project, we therefore adapted individual accelerated lifetime tests to identify critical stress factors for DSC. The stability tests that are described in this section are performed with the intention to study DSC performance as a function of materials, compositions and processing schemes in the various baselines, and not to define the potential lifetime that can be obtained with DSC devices under outdoor conditions. No scientific outdoor tests were carried out in this project except product exposure studies by Greatcell to evaluate effects of cell imbalance. 3.2.5.1 Stability issues in DSC The next part of this technical report will be dedicated to issues affecting the stability of DSC, on a molecular level. Macro-level engineering or technological challenges such as poor cell sealing resulting in evaporation of solvents from the electrolytes, etc. are not factored here. At present, DSC has attained an efficiency exceeding 11 %, albeit with a volatile solvent based electrolyte. However, obtaining long-term stability at 80~85 C for the DSC had remained a major challenge for over 10 years, and has only recently been achieved by judicious molecular engineering of the sensitizer used in conjunction with a robust and non-volatile electrolyte as will be shown later in this section. A schematic presentation of the operating principles of the DSC is given in Figure 12. At the heart of the system (Figure 12A) is a mesoscopic oxide semiconductor film, which is placed in contact with a redox electrolyte or an organic hole conductor. The material of choice has been TiO2 (anatase) although alternative wide band gap oxides such as ZnO, and Nb2O5 have also been investigated. Attached to the surface of the nanocrystalline film is a monolayer of the sensitizer.
Conducting glass TiO2 Dye Electrolyte Cathode
Injection S* -0.5 0 E vs NHE 0.5 (V ) 1.0 e

-
Maximum Voltage
excitation
Red Mediator Ox
Diffusion
s+
s*
S/S+
e
-
Figure 12 (A) Principle of operation of the dye-sensitized nanocrystalline solar cell. Photoexcitation of the sensitizer (S) is followed by electron injection into the conduction band of an oxide semiconductor film. The dye molecule is regenerated by the redox system, which itself is regenerated at the counter electrode by electrons passed through the load. Potentials are referred to the normal hydrogen electrode (NHE) The energy levels drawn match the redox potentials of the standard N3 sensitizer ground state and the iodide/triiodide couple. (B) Scheme to illustrate the catalytic cycle that the sensitizer performs during cell operation.
24
Publishable Report
Photo-excitation of the latter results in the injection of an electron into the conduction band of the oxide. The dye is regenerated by electron donation from the electrolyte, usually an organic solvent containing redox system, such as the iodide/triiodide couple. The regeneration of the sensitizer by iodide intercepts the recapture of the conduction band electron by the oxidized dye. The iodide is regenerated in turn by the reduction of triiodide at the counterelectrode, where the circuit is being completed via electron migration through the external load. The voltage generated under illumination corresponds to the difference between the Fermi level of the electron in the solid and the redox potential of the electrolyte. Overall the device generates electric power from light without suffering permanent chemical transformation. Kinetic criteria for long-term stability of the sensitizer. Figure 12 (B) illustrates the catalytic cycle that the sensitizer performs during cell operation. Critical for stability are any side reactions that may occur from the excited state S* or the oxidized state of the dye (S+), which would compete with electron injection from the excited dye into the conduction band of the mesoscopic oxide and with the regeneration of the sensitizer. These destructive channels are assumed to follow first or pseudo-first order kinetics and are assigned the rate constants k1 and k2. Introducing the two branching ratios, P1 = kinj /(k1 + kinj) and P2 = kreg /(k2 + kreg) where kinj and kreg are the first order or pseudo-first order rate constants for the injection and regeneration process, respectively, the fraction of the sensitizer molecules that survives one cycle is given by the product P1 x P2. After performing n cycles the remaining fraction of the sensitizer is
= (P1 x P2)n = [S]/[S]t = 0

Introducing P1 x P2 = 1- and rewriting equation (1) in logarithmic form gives:
(3)
ln = ln(P1 x P2)n = n ln(1-)
(4)
For any sensitizer showing reasonable stability the product P1 x P2 must be very close to 1 and hence << 1. This allows using the first term of a Taylor expansion for the logarithm yielding ln(1) = - . Hence = 1 P1 x P2 = - ln /n (5)
Expressing P1 and P2 by the rate constants and taking into account that k1 << kinj and k2 << kreg, one obtains = 1 P1 x P2 = k1/ kinj + k2/ kreg = - ln ([S]/[S]t = 0)/n (6). Finally, as we are interested in the lifetime of the dye, we note that at t = ln ([S]/[S]t = 0) = -1. Therefore, k1/ kinj + k2/ kreg = 1/n (7).
Thus the simple result of this consideration is that at the time when the outside exposure of the cell reaches the sensitizer lifetime, the sum of the branching ratios for the two decomposition channels becomes equal to the reciprocal value of the turnover number of the dye. Using equation 5, the branching ratios that can be tolerated to reach a sensitizer lifetime of 20 years under outdoor operation can be calculated. The surface concentration of the currently employed ruthenium complexes is about 2 x 10-7 mol/cm2 and the photocurrent delivered in full sun is between 15 and 20 mA/cm2. Assuming for the photocurrent density, averaged over day and night and over the different seasons a value of 3 mA/cm2; the turnover frequency of the dye is derived to
25
Publishable Report
be 0.155 s-1. Hence each dye molecule adsorbed at the surface of the mesoscopic oxide film performs the cycle presented in scheme1 on the average about once in seven seconds. Twenty years corresponds to 631113472 seconds yielding a turnover number of 9.78 x107, i.e. about 100 million. The upper limit for the sum of the two branching ratios is therefore 1x 10-8. Recent kinetic measurements For most of the currently employed sensitizers the rate constant for electron injection from the excited state in the conduction band of the mesocopic TiO2 particles is in the femtosecond range. Taking for kinj a value of 1 x1013 s-1, a destructive side reaction with k1 < 105 s-1 could be tolerated. Ruthenium sensitizers of the N3 type readily satisfy this condition. They can undergo photoinduced loss or exchange of the thiocyanate ligand, which however occurs at a much lower rate than the 105 s-1 limit. It is also debatable whether this pathway is destructive as the product formed, often an iodo or solvato complex, still acts as a charge transfer sensitizer. In ethanolic solution prolonged photolysis of N3 dye leads to sulphur loss and formation of the cyanato-ruthenium complex probably via photoxidation by oxygen. However this reaction is not observed when the dye is adsorbed on oxide surfaces. Precise kinetic information has also been gathered for the second destructive channel involving the oxidized state of the sensitizer, the key parameter being the ratio k2/ kreg of the rate constants for the degradation of the oxidized form of the sensitizer and its regeneration. The S+ state of the sensitizer can be readily produced by chemical or electrochemical oxidation and its lifetime determined independently by absorption spectroscopy. Data from a recent study of the monosodium salt of the amphiphilic dye Ru(4-carboxylic acid-4-carboxylate)( 4,4-dinonyl-2,2-bipyridine)(NCS)2, coded as Z-907 are shown in Figure 13. Upon addition of the NOBF4 one electron oxidant, the Z-907 peaks shift to the red. The insert in figure 13 shows the spectrum of the oxidized sensitizer exhibiting two broad LMCT peaks with maxima at 594 and 768 nm. By monitoring the absorption change at 770 nm, the formation of the oxidized form of Z-907 is seen to occur over the first 810 min after the addition of the oxidant. The subsequent decay occurs with a lifetime of 75 minutes corresponding to k2 = 2.2 x 10-4 s-1. The regeneration rate constant for this sensitizer and a typical iodide/ triiodide redox electrolyte is at least 2x105 s-1. Hence the branching ratio is about 10-9 that is well below the limit of 10-8 admitted to achieve the 100 million turnovers and a 20-year lifetime for the sensitizer.
0.25
0.20
0.15
Absorbance / OD
60
40
20
0
0.10
0.05
bsorbance / mOD
500
700 600 Wavelength / nm
800
900
0.00
0
20
40 Time / min
60
80
Figure 13 The time dependence for the evolution and decay of the oxidized form of Z-907 in solution. The insert shows the spectrum of the oxidized form derived from the subtraction of the starting material at t0.
26
Publishable Report
3.2.5.2 EPFL stability results Many long-term tests have been performed with the N3 -type ruthenium complexes confirming the extraordinary stability of these charge transfer sensitizers. For example, a European consortium financed under the Joule program (LOTS-DSC JOR3-CT98-0261) has confirmed cell photocurrent stability during 10000 hours of light soaking at 2.5 suns corresponding to ca 56 million turnovers of the dye without any significant degradation. These results corroborate the projections from the kinetic considerations made above. A more difficult task has been to reach stability under prolonged stress at higher temperatures, i.e. 80 85 C. Recent stabilization of the interface by using selfassembly of sensitizers in conjunction with amphiphilic co-adsorbents has been particularly rewarding by allowing the DSC to meet for the first time, the specifications laid out for outdoor applications of silicon photovoltaic cells. A 8% efficiency DSC has been realized showing strikingly stable performance under both prolonged thermal stress and light soaking (see Figure 14)3. This high performance and stable device was realized by using a new electrolyte formulation (0.8 M 1-propyl-3-methylimidazolium iodide (PMII), 0.15 M I2, 0.1 M guanidinium thiocyanate (GuSCN), and 0.5 M N-methylbenzimidazole (NMBI) in 3-methoxypropionitrile) and the amphiphilic Ruthenium sensitizer, coded as K19 (see figure 10 of this report), which was grafted together with 1-DPA on the mesoporous titania film acting as a photoanode.
device A during continued one sun visible-light soaking at 60 oC.
Figure 14. Left: temporal evolution of photovoltaic parameters (AM 1.5 full sunlight) of device A during continued thermal aging at 80 C in the dark. Right: Temporal evolution of photovoltaic parameters (AM 1.5 full sunlight) of
This work has been accepted for publication ion in Applied Physics Letters [see ref.14]. The work was co-funded by several funding agencies as indicated in the acknowledgement of this paper. The part of the work that was funded under the NANOMAX project concerns the addition of guanidinium thiocyanate to the electrolyte as well as in the dye adsorption process. The basic electrolyte formulation was developed by EPFL together with licensee Maxell, and is protected by a patent. The specific electrolyte formulation described above was developed under Swiss Project CTI 7019.1.
27
Publishable Report
3.2.5.3 Masterplate results Since Surlyn doesnt ensure proper sealing for long-term tests under outdoor conditions, glassfrit material (by ISE) or Bynel (by ECN) was used to seal the two TCO glass plates that are commonly used in the masterplates with the sandwich configuration. Accelerated long-term tests for a number of cells containing different dye and electrolyte combinations have been performed at different temperatures up to 85 0C in the dark for up to 1000 hours and under simulator sunlight. A selection of test data for the two types of masterplates is summarized below: Glass sealed masterplates In order to have a hermetic sealing to prevent solvent leakage and inclusion of water and oxygen, not only the plates have to be sealed with glass frit but the injection holes as well. In Figure 15 an example of a glass frit sealed masterplate is shown and the final step of sealing the filling tubes using a micro flame.
Lochversiegelung mittels M ikroflamme
Figure 15. Photo of a glass sealed masterplate as has been used for the accelerated ageing tests. The final step, sealing of filling tubes by a micro flame is shown.
EPFL has delivered dye and electrolyte samples and the glass sealed masterplates were processed in the baseline of ISE. The used electrolytes had the following compositions: Robust 0.8 M PMII + 0.15 M I2, 0.1 M GSCN and 0.5 M NMBI in MPN Z71: PMII + 0.5 M I2 + 0.45 M NMBI H: 0.6 M HMII, 0.1 M I2, 0.5 M NMBI, in butoxypropionitrile Maxell type: 0.2 M I2, 0.5 M NMBI, 0.1 M GSCN in PMII:Ethylmethyl imidazoliumthiocyanate (65:35 solvent) The results after ageing are summarized in Table 4.
Table 4. Electrolyte dye combinations used in accelerated ageing test (glass sealed masterplates). Electrolyte Robust Z71 H Maxell type Dye K19 Z907 N719 N719 Initial efficiency (1 sun) Degradation after 85 oC Degradation after 1 sun ageing 7.1% 39 % (300 hours) < 5% (300 hours) (s. remark!) 2.7% < 5% (800 hours) 3.5% < 5% (500 hours) < 5% (500 hours) 3.4% < 5% (3000 hours)
Remark: The strong degradation of the cells consisting of the dye K19 and Robust electrolyte is in contradiction to the high stability results reported by EPFL for this electrolyte (see figure 14). In this case the manufacturing protocol, in particular coloration time, has still to be adapted from the EPFL method. The other tested electrolytes show very promising stability results.
28
Publishable Report
Bynel sealed masterplates Thermal annealing/illumination cycling Masterplates of the conventional "old" design (active area 4 cm2) were used to prevent fast corrosion of the silver lines by the iodine electrolyte. Hotmelts like Surlyn and Bynel are used for the ageing tests. A homemade TiO2 coated FTO substrate was used as the photoelectrode, while N719 (from Sx SA) was used as the sensitizing dye. The following electrolyte compositions have been used in the master plates: EL 1: 0.6 M BMII, 0.1 M LiI, 0.05 M I2, 0.5 M TBP in a mixture of acetonitrile/valeronitrile (AN/VN) (85/15 V/V) EL 2: 0.6 M HMII, 0.1 M LiI, 0.05 M I2, 0.5 M TBP in acetonitrile (AN) EL 3: 0.6 M DMPI, 0.1 M I2, 0.5 M NMBI in acetonitrile (AN) Electrolytes 1 and 2 are standard compositions, which usually result in the highest efficiencies. Electrolyte 3 was chosen to mimic the type of electrolyte that was currently used in long term stable devices as reported by EPFL using Z907 hydrophobic dye (Ref). The only difference is that in this case acetonitrile is used instead of Methoxypropionitrile (MPN) since MPN was not available in sufficient purity.
6 5 Eff (%) 4 3 2 1 0 0 500 time (h) 1000 1500 EL.1 EL.2 EL.3
Figure 16. Stability data for 3 cells depending on conditions: from 0-110 h, 260-328 h and 473-1162 h: 85 C, dark. From 110-260 h and 328-473 h: 45 C, 1 sun. Note: cell filled with EL 3 started leaking around 800 h
From Figure 16 it can be derived that the initial decrease in performance at 85C is reversible. Switching the testing conditions from (85C/dark) to (45C/1 sun illumination) results in a recovery of the cell performance (Figure 16, from 110 260 h). After 110 hours at 85 C/dark the cell efficiency is decreased to 59 % of the initial value for cells filled with electrolyte EL 1. A spectacular recovery to 89 % of the initial value after 26 h under 45C/1 sun occurs, increasing to 94 % after an extra 124 h under 45C/1 sun. After a second cycle of degradation in the dark at 85 C the performance recovers again during illumination at 45 C. It has been demonstrated that this recovery occurs only after visible light soaking, since storage of the devices in the dark at room temperature does not lead to recovery. The degradation/recovery behaviour is dependent on the applied electrolyte. In contrast to cells with electrolytes based on acetonitrile or other organic solvents, cells based on ionic liquids as a solvent hardly degrade at 85 C in the dark, though the cell efficiency of this system in sealed devices is substantially lower due to charge transport limitation . In sealed cells with 50 micron spacer distance between front and counterelectrodes efficiencies of approximately 1-2 % are reached using pure ionic liquids like HMII and I2. In open cells the distance between the electrodes is minimized, resulting in approximately 5 % efficiency. This result is a good proof of principle for the temperature stability of a DSC. The degradation pattern of the solvent-based electrolytes is dependent on the exact electrolyte composition. As can be seen in figure 16 the two cells filled with electrolytes EL 1 and 2 have a much stronger response to the cycling between annealing and illumination than the cell with
29
Publishable Report
electrolyte EL 3. The electrolyte that results in a more stable behaviour contains an increased concentration of iodine (0.1 M instead of 0.05 M) and does not contain LiI and TBP. It has been suggested by several people that the presence of LiI in the electrolyte plays a detrimental role in the stability of DSC. This will be further illustrated in the next section: Thermal annealing at different temperatures A number of masterplates were aged in the dark at three different temperatures for 1000 hours, i.e 55, 70 and 85 oC. Dye N719 (INAP). After thermal annealing, cells have been stored in the refrigerator for some time and subsequently illuminated in the STF4 test facility at around 1sun to study the recovery behaviour. Two electrolyte compositions have been compared: Electrolyte standard prepared by ECN: El.1: 0.6 M PMII (STI), 0.1 M LiI, 0.05 M I2, 0.5 M TBP (pur. EPFL), Propionitrile (EPFL) Electrolyte prepared by FMF 0.6 M HMII, 0.1 M I2, 0.5 M NMBI, Butoxypropionitrile The temporal evolution of the performance (expressed in MPP) for the aged masterplates containing the two different electrolyte compositions is shown in Figure 17.
25 20
10 8 6
MPP
15 10 5 0 0
A
A
500 1000
B4
B
1500 2000
2 0
2500
0
A
500 1000
B
1500 2000
C
2500
hours
hours
Figure 17. Ageing test tests carried out for masterplates:Region A: dark ageing at RT, 55 (red), 70(blue) and 85 oC(green), Region B: Storage of MP in fridge, Region C: Light soaking in STF4 The left graph refers to cells containing ECN standard electrolyte and the right graph to cells containing FMF electrolyte (efficiency=MPP/4)
It can be seen from the figure that the initial efficiency of cells with the standard electrolyte composition is up to 3 times higher than for the viscous BPN containing electrolyte, mainly due to charge transport limitation in the electrolyte. While a clear temperature dependence of cell degradation is observed for the ECN standard electrolyte containing LiI and tBP, this is not observed for the FMF electrolyte without these additives. Cells aged at 85 oC even showed a higher performance after 1000 hours as compared to cells aged at 55 and 70 oC ageing. After some initial decrease of the performance in the first 150 hours, the cells remain reasonably stable. Cells remain stable while stored in the refrigerator at 4 oC. Subsequent light soaking of these cells results in recovery as was also described in the former section. The recovery was most pronounced for cells containing the BPN electrolyte where final performance was even higher than measured directly after preparation. 3.2.5.4 Summary stability data The ageing experiments that were carried out for single cells (at EPFl and on masterplates) described in the former sections were dedicated to the stability of DSC, on an intrinsic, molecular level. Solvent leakage was ruled out since appropriate sealing methods with suitable materials (Bynel and glassfrit) were used in these experiments; even cells containing acetonitrile as lowboiling electrolyte solvent didn't show evaporation at 80 oC when properly sealed. Although the
30
Publishable Report
data in this summary doesn't give a complete picture, the research that has been done by the partners in this consortium, within but also outside the project has led to the following empirical observations and intermediate conclusions with respect to the choice of materials for long term stable DSC: - Choice of sensitizer: concerning long term stability at elevated temperatures, it has been postulated that the use of hydrophobic dyes like Z907 and K19 are preferred over the frequently used classical "red" dye N719. The main explanation for this is that water is more effectively excluded from the interface between dye and the semiconductor when using a more hydrophobic dye such as Z907 and K19. Although the role of surface water needs more investigation, this is often mentioned as the main reason for the observed voltage decline at high temperatures. Cografting of the sensitizers with amphiphilic additives like 1-decylphosphonic acid (DPA) [15] or Hexadecylmalonic acid (HDMA) [16] further improves the compactness of the hydrophobic monolayer and restricts access of water or other interfering impurities to the surface of the titania nanocrystals. The latter was clearly observed by decreased FTIR signals from surface adsorbed water in the mixed monolayers than for the amphiphilic sensitizer alone. - Choice of electrolyte: A large number of different electrolyte compositions in combination with different dyes have been tested over the past years. The highest efficiencies were usually obtained for cells with the red N719 dye and electrolytes containing Ionic liquids, LiI, Lewis Bases like tBP and low concentrations of iodine in a volatile solvent like acetonitrile. However, it has been shown (see Figure 16 and 17) that when these electrolytes are used in lifetime test, performed in dark and temperatures > 60 oC, stability is rather limited, although a spectacular recovery of the performance is observed when these cells are subsequently illuminated at around 1 sun at temperatures < 50 oC. Cells based on electrolytes with a higher content of I2, NMBI instead of TBP and avoiding LiI in conjunction with a amphiphilic sensitizer showed very good thermal stability, although the initial efficiencies are somewhat lower as compared to the highest efficiency cells that often contain LiI and/or TBP. The recently discovered voltage enhancing additive GuSCN seems also to act as a voltage stabilizer in cells containing the new dye K19 and so-called "robust" electrolyte, which have shown to be the most stable DSC up to now. - It has been mentioned above that apart from the cell composition itself, unwanted impurities like water and other unknown substances can have a serious (negative) impact on the chemical stability of the system. This means that the purity of the starting materials and well-defined processing conditions seems to be very important factors in getting reliable output. At this stage of development, many R&D laboratories have developed their own manufacturing protocols, using materials of different origin and purity and this is often the reason for lab-to-lab variations in efficiency and stability experiments. - Besides IV monitoring of aged and reference non-aged devices, analytical, electrical and optical methods are needed to understand more in detail the aging phenomena occurring in DSC. Apart from the intrinsic stability issues related to the materials used in DSC, long-term high temperature stability trials of larger modules carried out by Greatcell demonstrated that thermal expansion of the module resulted in loss of fill factor and, under, high drive conditions, imbalance between cells. This is regarded as an important risk when using multicell modules and requires a careful and smart module design to minimize the extrinsic effects that are associated with upscaling from cell to large area modules. 3.2.6 WORKPACKAGE 6 (Technical evaluation, cost analysis and module demonstration) 3.2.6.1 Introduction The objective of Nanomax has been to work on new concepts and materials for dye-sensitized solar cells in order to improve the efficiency and to reduce the costs. The most promising ones have been analysed from a manufacturing point of view in WP6 Technical evaluation, cost analysis, and module demonstration. This section summarises the main outcome of WP 6:
31
Publishable Report
Cost analysis of the winning concept Aa, i.e. a reflecting layer on top of the TiO2 (3.2.6.2) Estimation of the costs for alternative low-cost concepts in relation to Concept Aa (3.2.6.3) Module demonstration (3.2.6.4).
More information about the activities and outcome of WP 6 can be found in the IVF progress reports from 2003 and 2004. Various existing patents cover many of the cell/module concepts and process methods that have been dealt with in WP 6. No patent analysis has been made within Nanomax. Moreover, aspects related to availability and costs for required licences for a potential producer have not been taken into account in the cost analyses. 3.2.6.2 Cost analysis for Concept Aa Manufacturing costs/m2 based on the present material costs projected on a large-volume demand ECN[17] and FMF have performed thorough cost calculations for manufacturing of serialconnected dye PV modules for different production capacities (1-4 MWpeak/year) for Concept Aa. These have been further analysed and summarised by IVF. The two analyses are slightly different. The ECN work is based on a high level of labour and semi-automatic pieces of equipment whereas the FMF work is based on a higher level of automation. Moreover, the definitions of the cost categories are slightly different. Their common result, however, is that the production costs for devices based on concept Aa with the present material costs, projected on a large-volume demand, would be 120-150 /m2. The cost distributions (Figure 18) imply that the production costs of the technology mainly come from the materials (50-70 %), especially the TCO glass and the dye. Consequently, the major cost reduction in the dye PV technology can be obtained through reducing the material costs.
Screen-printing pastes 14 % Dye 15 % Framing 13 %
Screen-printing pastes 9 % Running costs 21 %
Running costs 40 %
Various materials 14 %
Investment cost 6 % TCO glass 24 % Dye 22 %
TCO glass 16 % Equipment 9 %
Various materials 5 %
(A)
Figure 18.
(B)
Overview of the production cost distributions based on analysis from (A) ECN (serialconnected thermoplastic-sealed devices, appr. 1 MWpeak/year) and (B) FMF (serialconnected glass-frit sealed devices, appr. 4 MWpeak/year).
32
Publishable Report
Potential for cost reductions According to Gavin Tulloch of Greatcell, the price of TCO glass will come down to half the present prices once a large-volume dye PV production is running. The high cost of the materials also highlights the importance on minimising waste (especially of the dye), using thin TiO2 films etc. the new EPFL group of hydrophobic organometallic Ru-dyes with higher extinction coefficients in the visible part of the spectrum than the standard N719 Ru-dyes (see EPFL progress report from 2004), opens for thinner TiO2 layers. Consequently, such dyes will lead to lower dye cost per device. Taking all of this into account, the project partners believe that it is realistic that the abovedefined manufacturing costs may be reduced with up to 40 %, i.e. production costs of 70-90 /m2. Manufacturing costs/Wp In Table 5, the manufacturing costs/Wpeak are given for various cell efficiencies that have been obtained within Nanomax in combination with the above-defined manufacturing costs per square meter device (120-150 /m2 with the present material costs and 70-90 /m2 with the forecasted future material costs). With the typical Nanomax masterplate efficiencies (7.5% active area efficiency) and the present materials costs (manufacturing costs of 120-150 /m2), the analysis shows values of 2.1-2.7 /Wp. With the efficiencies obtained on the Nanomax test cells at EPFL (10 % on cells > 1 cm2 and 11 % on cells <1 cm2), the manufacturing costs could end up below 2 /Wpeak with todays material costs. Based on the forecasted future costs of material, the manufacturing costs may well be below 1 /Wp.
Table 5. The manufacturing costs/Wpeak for the above-defined manufacturing costs/m2 in combination with key efficiencies obtained within Nanomax. The module efficiencies are calculated from a 75 % TiO2 coverage of the module area. Active area efficiency Manufacturing costs/m2 costs/m2 Manufacturing based on the forecasted based on the present material future material costs costs 70 /m2 90 /m2 120 /m2 150 /m2
7.5 % (ECN masterplate)
1.2 / Wp
1.6 / Wp
2.1 / Wp
2.7 / Wp
8% 1.2 / Wp (EPFL test cell using the stable electrolyte)
1.5 / Wp
2.0 / Wp
2.5 / Wp
10 % 0.9 / Wp (EPFL test cell >1 cm2)
1.2 / Wp
1.6 / Wp
2.0 / Wp
11 % 0.8 / Wp (EPFL test cell < 1 cm2)
1.1 / Wp
1.4 / Wp
1.8 / Wp
33
Publishable Report
3.2.6.3 Estimation of the costs for Nanomax low-cost concepts in relation to Concept Aa The screen-printed version of the TCO-less Concept Ab The manufacturing costs for the monolithic screen-printed TCO-less device (see progress report from IVF from 2004) holds the potential to become cheap compared to concept Aa. According to the above-reported cost analysis, the TCO glass represents approximately 20 % of the overall manufacturing costs of Concept Aa. The TCO-less design naturally avoids this cost. Further cost advantages may be obtained from the comparatively straightforward process methods. In a first
estimation, the investigated screen-printed TCO-less design may lead to 30 % lower costs
than for concept Aa. However, it is important to highlight that the manufacturing costs cannot be fully analysed until a current collector material and a solution for the serial connections have been identified. Moreover, it is highly uncertain whether the costs /Wpeak would be lower in relation to concept Aa since reduced performance is foreseen due to performance losses at high light intensities caused by diffusion limitations of the electrolyte and/or ohmic energy losses. The monolithic concept The monolithic device concept (see progress report from IVF from 2004) offers cost advantages in the manufacturing costs/m2 in relation to Concept Aa. Main reasons for this are the comparatively low number of process steps, the need of only one TCO substrate, and that the serial-connections are made with the carbon counter electrode. An initial estimation of the manufacturing costs indicates 15-20 % lower costs/m2 than the values for Concept Aa. On the other hand, the masterplate and module efficiencies presently obtained with the TCO-based monolithic concept are lower than the ones obtained on Concept Aa. In terms of costs/Wpeak, the two concepts therefore at present lead to approximately the same values. 3.2.6.4 Module demonstration The objective of the module demonstration of Nanomax has been to implement the concept, materials, and methods of the winning concept into procedures that have been developed outside Nanomax in order to demonstrate a module of approximately 100 cm2. The background knowledge about module fabrication has been used but not shared with the project partners. ECN and Fraunhofer have done this for the winning concept Aa. In addition, the IVF background knowledge of module fabrication of monolithic modules on TCO-glass has been used to prepare such modules with the applicable Nanomax components (TiO2 and dye). Modules based on Concept Aa ECN modules The ECN modules are made on TCO glass of the size 10x10 cm2. Each module consists of four serial-connected cells and the active TiO2 area of the individual cells is 16.7 cm2 (Figure 19). The modules have been characterised at full sun using the ECN Steuernagel simulator.
Figure 19. An ECN module (100 cm2) based on Concept Aa with four serial-connected cells.
34
Publishable Report
Fraunhofer modules The Fraunhofer modules are made on TCO glass of the size 30x30 cm2, i.e. much larger devices than the other Nanomax modules. Each module consists of 29 serial-connected cells and the active TiO2 area of the individual cells is 17.4 cm2 (Figure 20).
Figure 20. A Fraunhofer ISE module (900 cm2) with 29 serial-connected cells.
Monolithic modules from IVF The IVF monolithic modules are made on TCO-substrates of the size of 8 x 10 cm2. Each module consists of 8 cells in series (Figure 21) and the active TiO2 area of each cell is 8.9 cm2. The modules have been characterised at full sun and at 1/10 of a sun using the ECN Steuernagel simulator.
Figure 21. A monolithic IVF module (8 x 10 cm2) with 8 serial-connected cells.
Module results The yields of the different Nanomax modules are summarised in Table 2. Typically, the same opencircuit voltages and short-circuit density are obtained for the modules as for the masterplate. The challenge is to obtain as high fill-factors on the modules as on the masterplates. Whereas the masterplate design is optimised to minimise ohmic losses, and therefore obtain fill-factors above 70 %, the module design must be to compromised between minimising ohmic losses in the TCO and maximising the fraction active TiO2 area versus the total area. Hence, it is logical that the fill-factor
35
Publishable Report
for a module is less than the one obtained for a masterplate, especially at high light intensities. A glass with sheet resistance below 5 ohm/square is highly desirable since this would simplify to overcome this problem. As can be seen in the results, the fill-factors of the Nanomax modules are at present too low. This can likely, in addition to the above-mentioned fill-factor problem, be explained by the fact that the module preparation only took place during the last months of the project and that more effort is required to get the performance up. The exact reasons for the low factors are likely depending on the materials, process methods, and module designs of each partner.
Table 6. Overview of the performance of the Nanomax modules. Light intensity 1 sun Uoc/cell 0.68 V 0.69 V 0.78 V 0.71 V jsc 13.8 mA/cm2 9.7 mA/cm2 11 mA/cm2 1.3 mA/cm
2
ff 0.58 0.53 0.5 0.7
ECN concept Aa Fraunhofer 1 sun concept Aa IVF - 1 sun monolithic 1/10 sun
Active area efficiency 5.5 % 3.5 % 4.4 % 6.5 %
3.3
Assessment of results and conclusions

3.3.1 General conclusions
The proposed project aimed to develop and test new and improved ways to realise DSC that are believed to further enhance the efficiency and stability. During the project several new areas are explored in the field of new concepts and materials, fabrication protocols for TiO2 and scatterlayers, metal oxide blocking layers, strategies for co-sensitization, low temperature processes of platinum deposition and fundamental understanding of the working principles was gained by means of electrical and optical modelling and advanced characterization techniques. All the scientific objectives mentioned under section 2.2.2 have been met successfully in this project as can be seen from the large number of (joint) scientific publications that were made in the frame of this project. Furthermore, several scientific exchanges were made for sharing information and exchanging ideas and know-how. A number of young people got trained in the DSC field and improved networking contacts between partners as a result of Nanomax, has been made, which has also resulted in further collaboration in FP6 projects. These can be regarded as very important non-technical achievements of the project. The combined research efforts have led to the following technical achievements and should be related to the technical objectives as mentioned under section 2.2.3 : New Ru- containing sensitizing dyes with enhanced optical absorption in the visible part of the spectrum have become available and are successfully applied in DSC Protocols of making metal oxide blocking layers on TiO2 resulting in retardation of recombination dynamics and improvement of the photovoltage of the device. New concepts like TCO-less design (Concept Ab) have been developed and introduced. New scatterphenomena in TiO2 films have been discovered. A maximum power conversion efficiency under full sunlight of 11 % for areas < 0.2 cm2 and 10.1 % for a cell with an active area of 1.3 cm2. Although this value is not yet certified by a calibration institute, the result can be considered as the highest efficiency ever measured for a cell with an area > 1 cm2. A >8 % efficient cell retaining over 98 % of its initial performance after 1000 h accelerated tests under thermal stress at 80oC in the dark has been demonstrated. Negligible device
36
Publishable Report
degradation was observed for 1000 h visible light soaking at 60 oC. These very encouraging data will foster practical applications for DSC Advanced characterisation techniques have been developed for in situ characterisation of dye sensitized photovoltaic devices under operation, including transient optical studies including all key steps of interfacial charge separation and recombination dynamics and transient photovoltage and photocurrent studies of transport dynamics. These experiments have been correlated with J/V data in the dark and under illumination and with a numerical model of these J/V data based upon the non-ideal diode equation. These methodologies have been employed to characterise fundamental loss factors both in standard (devices and a range of innovative device concepts as developed in this proposal. For standard devices the primary loss mechanism is confirmed as charge recombination from electrons in the metal oxide to the oxidised redox couple. A technical evaluation and cost analysis of the winning concept Aa and alternative low cost concepts have been performed. Modules of different designs (Z-type, current collecting and monolitic) and sizes (from 100 up to 900 cm2) were demonstrated in the existing processing baselines with maximum active area efficiencies of 5.5 % @1 sun and 6.5 % at 0.1 sun
Although some technical objectives have not been met in a quantitative way, the consortium addressed all the scientific and technological issues sufficiently with success and good results 3.3.2 Dissemination and use of results
The market for DSC technology will be based on traditional PV applications. It is anticipated that DSC will become a competitive PV technology starting with consumer indoor products like calculators, watches, clockworks (although there will be strong competition with low-cost amorphous silicon), followed by off-grid solar home systems for lighting, radio and TV and gridconnected (power) PV. Separately, DSC will enable new types of PV applications to be developed due to its very specific character. This will for instance occur in the field of building integrated PV for which it will be combined with other functions, such as electrochromic windows in so-called tandem applications. The success of penetration into existing and new PV markets will depend on many aspects; the important ones are most likely (1) costs in euro/Wp as well as euro/m2 of product, (2) technical and environmental profile (3) added value for the consumer and architects and (4) ease of production and scale at which production plant becomes economically feasible. It is expected that DSC can be a competitive PV technology with regards to aspects (1), (3) and (4), whereas it's technical performances are still lacking behind existing PV technologies. To what extend the technical performance should improve before DSC becomes an attractive option is still a matter of discussion, since obviously technical performance is not the sole criterium. It's probably safe to state that the minimum set of criteria would be at least similar or better as compared to single junction amorphous silicon PV modules, i.e. module efficiencies > 6% with outdoor lifetime > 8 years (in case of solar home system market). NANOMAX was mainly concerned with aspect no. (2), i.e., the improvement of the technical performance of the dye sensitized solar cell to increase light conversion efficiency and durability. At the end of this project, typically 7-7.5% efficiencies can now be obtained on routine base for > 2.5 cm2 DSC laboratory devices with top efficiencies ranking up to 10.1 % (1.3 cm2) and 11 % (0.16 cm2). In this respect, DSC is reaching the same level of performance of some inorganic thin film PV technologies that are already commercially available. As regards durability, intrinsic chemical stability has been proven at elevated temperatures (up to 80 oC) and under light soaking (1 aun) for DSC containing specific combinations of dye and electrolyte. Another interesting feature of DSC long-term stability has been identified recently for DSC containing less stable combinations of Dye and electrolyte: The almost quantitative recovery of DSC cell performance after ageing in dark at 55 C or 85 C simply due to light exposure. This means that under realistic conditions, with
37
Publishable Report
natural daily dark/light cycling, the actual long-term stability is probably underestimated if based on individual stability testing procedures. This once again demonstrates the need for adequate stability tests specifically for organic solar cells. However, it is still too early to give reliable data on longterm outdoor stability of DSC laboratory devices. It is certainly not yet competitive with existing (inorganic) PV technologies. It is most probably sufficient for indoor applications at this stage of development. Indirect DSC applications for commercial exploitation Apart from the PV products based on DSC principles, the knowledge and infrastructure obtained through NANOMAX and related R&D projects will become available for commercial exploitation. This will be true for new dyes, screenprintable pastes, electrodes, electrolytes, production knowledge, new DSC concepts and DSC production machinery.
3.4
References
1. Acid versus base peptization of mesoporous nc-TiO2 films: functional studies in dyesensitized solar cells, S.Hore, E.Palomares, H.J.P. Smit, N.J. Bakker, P. Comte, P.Liska, K.R. Thampi, J.M. Kroon, A. Hinsch, J.R. Durrant, J. Mater.Chem, 15 (2005), 412-418 2. Scattering spherical voids in nanocrystalline TiO2 enhancement of efficiency in dye solar cells, S. Hore, P. Nitz, C. Vetter, C. Prahl, M. Niggemann and R. Kern, Submitted to Chem.Comm. 2004 3 a) Slow Charge Recombination in Dye-sensitized Solar Cells (DSC) Using Al2O3 Coated Nanoporous TiO2 Films. E. Palomares, J.N. Clifford, S.A. Haque, T. Lutz, J.R. Durrant Chem. Comm. 14, (2002) 1464-1465. Control of Charge Recombination Dynamics in Dye Sensitized Solar Cells by the use of 3 b) Conformally Deposited Metal Oxide Blocking Layer. E. Palomares, J.N. Clifford, S.A. Haque, T. Lutz, J.R. Durrant, J. Am. Chem. Soc. 125 (2003) 475-482. 4. Charge transport versus recombination in dye sensitized solar cells employing nanocrystalline TiO2 and SnO2 films, A. Green, E. Palomares, S.A. Haque, J.M. Kroon, J.R. Durrant, , submitted to J.Phys.Chem.B 2005 5. A remarkably stable quasi-solid state dye-senitized solar cell based on an amphiphilic ruthenium sensitizer and a polymer gel electrolyte, P. Wang, J-E. Moser, M.K. Nazeeruddin, T. Sekiguchi, M. Grtzel, Nature Materials 2 (2003), 402-408 6. Stable new sensitizer with improved light harvesting for nanocrystalline Dye-sensitized solar cells, P. Wang, S.M. Zakeeruddin, J-E. Moser, R. Humphrey-Baker, P. Comnte, V. Aranyos, A. Hagefeldt, M.K. Nazeeruddin, M. Grtzel, Advanced Materials, 20 (2004), 1806-1811 7. A High molar exctinction coefficient sensitizer for stable Dye-sensitized Solar Cells, P. Wang, C.Klein, R. Humphrey-Baker, S.M. Zakeeruddin, M. Grtzel, J. Am. Chem. Soc 127 (2005), 808-809 8. Application of metalloporphyrins in nanocrystalline dye-sensitized solar cells for conversion of sunlight into electricity, Md.K. Nazeeruddin, R. Humphrey-Baker, D.L. Officer, W.M. Campbell, A.K. Burrell, M. Grtzel, Langmuir 20(2004), 6514-6517 9. New concepts and materials for world-class dye-sensitized solar cells J.M. Kroon, N.J. Bakker, H.J.P. Smit, P. Liska, K.R. Thampi, M. Grtzel, A. Hinsch, S. Hore, J.R. Durrant, E. Palomares, H. Pettersson, T. Gruszecki, J. Walter, K. Skupien, G. Tulloch, Proceedings of the 17th EU Photovolatic Solar Energy conference, Paris, 7-11 June 2004, pg. 54-57 10. M.A. Green, K. Emery, D.L. King, S. Igari, W. Warta, Solar cell efficiency tables (Version 20). Progress in Photovoltaics: Research and Applications 2002; 10:355-360.
38
Publishable Report
11. Multistep electron transfer processes on dye co-sensitized nanocrystalline TiO2 films J.N. Clifford, E. Palomares, M.K. Nazeeruddin, K.R. Thampi, M.Grtzel, J.R. Durrant, J. Am. Chem. Soc. 126 (2004) 5670-5671. 12. Tandem dye-sensitized solar cell for improved power conversion efficiencies, M. Durr, A. Bamedi, A. Yasuda, G. Nelles, Applied Physics Letters, 84 (2004), 3497-3499 13. Charge separation versus recombination in dye-sensitized nanocrystalline solar cells: the minimization of kinetic redundancy, S.A. Haque, E. Palomares, B.M. Cho, A.N.M. Green, H. Hirata, D.R. Klug, J.R., Durrant, J. Am. Chem. Soc. 126 (2005) in press. 14. A stable, >8% efficient nanocrystalline dye-sensitized solar cell based on an electrolyte of low volatility, P. Wang, C.Klein, R. Humphrey-Baker, S.M. Zakeeruddin, M. Grtzel, accepted for publication in Applied Physics Letters 15a) Molecular-scale interface engineering of TiO2 nanocrystals: Improving the efficiency and stability of dye-sensitized solar cells, P. Wang, S.M. Zakeeruddin, R. Humphrey-Baker, , M. Grtzel, Advanced Materials, 15 (2003), 2101 15b) Enhance the performance of dye-sensitized solar cells by Co-grafting amphiphilic sensitizer and hexadecylmalonic acid on TiO2 nanocrystals, P. Wang, S.M. Zakeeruddin, P. Comte, R. Charvet, R. Humphrey-Baker, M. Grtzel, J.Pys. Chem. B 107 (2003), 14336-14341 16. Long-term stability testing of dye-sensitized solar cells, P.M. Sommeling, M. Spath, H.J.P. Smit, N.J. Bakker, J.M. Kroon, J. of Photochem. and Photobiol. A: Chemistry 164 (2004), 137-144; 169 (2005), 317-318 17. Dye-sensitized solar cell from laboratory scale to pre-pilot stage, M. Spth , J. van Roosmalen , P. Sommeling , N. van der Burg , H. Smit , D. Mahieu , N.J. Bakker , J.M. Kroon, Proceedings 3rd World Conference on Photovoltaic Energy Conversion, Osaka, 2003, pg. 196-199
39

DSSC Detail

Загружено:

Сведения о документе

Исходное описание:

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

DSSC Detail

Загружено:

Авторское право:

Доступные форматы

ECN-C--05-078

NANOCRYSTALLINE DYESENSITIZED SOLAR CELLS HAVING MAXIMUM PERFORMANCE

Contract No: Acronym: Title :

ENK6-CT2001-00575 NANOMAX Nanocrystalline Dye-Sensitized Solar Cells Having Maximum Performance

Executive Summary Objectives and strategic aspects

Scientific and technical performance

Objectives and strategic aspects

Scientific and technical performance

D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 D11 D12 D13 D14

Overview of the technical progress

WP 4 Advanced characterisation and modelling Work package leader: ICL

WP 5 Long-term stability Work package leader: ECN

Scheme 1. Project and management structure

WORKPACKAGE 1 (Concept development)

Nc-oxide + dye(s) TCO

Porous Ti (metal) white ZrO2 Nc-oxide + dye(s)

TCO less (Ab)

Nc-oxide + dye 1+ dye 2 Eg < Eg > transparent ZrO2 Nc-oxide + dye 1

RESPONSIBLE ECN/FMF-ISE EPFL EPFL FMF

WORKPACKAGE 2 (Materials Development)

Isc/mAcm-2 nc- TiO2 Transparent TiO2 12 Hollow Scattering 13 spheres

Voc/mV 740 740

(%) 5.4 6.7

Figure 3. The molecular structure of Z907

a) R = H, Me, Et X- = I-, SCN-

b) R = Me, C6,C10 X- = I-, SCN-

TCO TEC 8/3 mm TEC15/3 mm SxFTO/1 mm

FF 0.7 0.69 0.69 (0.69*) 0.69

(%) 7.09 7.30 7.35 (7.52*) 7.43

Figure 9. Charge transfer processes in multilayer co-sensitized nanocrystalline TiO2 films

(i) (i) (ii)

WORKPACKAGE 5 (Long-term Stability)

Injection S* -0.5 0 E vs NHE 0.5 (V ) 1.0 e

= (P1 x P2)n = [S]/[S]t = 0

ln = ln(P1 x P2)n = n ln(1-)

700 600 Wavelength / nm

device A during continued one sun visible-light soaking at 60 oC.

Screen-printing pastes 14 % Dye 15 % Framing 13 %

Screen-printing pastes 9 % Running costs 21 %

Investment cost 6 % TCO glass 24 % Dye 22 %

TCO glass 16 % Equipment 9 %

7.5 % (ECN masterplate)

8% 1.2 / Wp (EPFL test cell using the stable electrolyte)

10 % 0.9 / Wp (EPFL test cell >1 cm2)

11 % 0.8 / Wp (EPFL test cell < 1 cm2)

Figure 21. A monolithic IVF module (8 x 10 cm2) with 8 serial-connected cells.

ff 0.58 0.53 0.5 0.7

Active area efficiency 5.5 % 3.5 % 4.4 % 6.5 %

Assessment of results and conclusions

Вам также может понравиться