Priority 6.1.ii: FP6-2002-Energy-1ATHLET


Project no.
019670

Project acronym
ATHLET

Project title

Advanced Thin Film Technologies for Cost Effective Photovoltaics

Instrument: Integrated Project

Thematic Priority 6.1.ii

Publishable Final Activity Report

Period covered: from 01.01.2006 to 31.12.2009Date of preparation: 15.02.2010

Start date of project: 01.01.2006Duration: 48 Month

Project coordinator name: Prof. Dr. M.-Ch. Lux-Steiner

Project coordinator organisation name: Hahn-Meitner-Institut Berlin GmbH

Revision 1

1Project Execution......

1.1Objectives......

1.2Challenges......

1.3Project structure and partners......

1.4Progress within the project duration......

1.4.1SPI - High Efficiency Solar Cells......

1.4.2SP2 - Thin Film Module Technology......

1.4.3SP3 - Chalcopyrite specific heterojunctions and TCOs......

1.4.4SP4 - Thin film silicon large-area modules on glass......

1.4.5SP5 - Device analysis and modelling......

1.4.6SP6 - Sustainability, Training and Mobility......

1.5General Project Information......

2Dissemination of Knowledge......

1Project Execution

1.1Objectives

The main objective of the project is to accelerate the decrease inthe cost/efficiency ratiofor thin film PV modules towards 0.5 €/WP. It focuses on technologies based on amorphous, micro- and polycrystalline silicon as well as on I-III-VI2-chalcopyrite compound semiconductors. The work oriented along the value chain focuses on large area chalcopyrite modules with improved efficiencies and on the up-scaling of silicon based tandem solar cells. This is complemented by a range of activities from the demonstration of lab scale cells with higher efficiencies to the work on module aspects relevant to all thin film solar cells. An important aspect is the analysis and modelling of materials, processes and devices. Accompanying sustainability assessment gives advice to the consortium on successful implementation strategies.

1.2Challenges

Thin film photovoltaics have a higher potential for cost effective production in the economy of scale than the technologies on the market today. In order to benefit from this potential, production capacities must grow faster than the established technologies. Main obstacle for a fast growth is the degree of maturity. This concerns all aspects from the fundamentals to the industrial implementation. Accordingly, this project addresses a range of issues. The most important scientific and technical objectives are given below:

to improve front and back contacts in view of long-term stability, conductivity, transparency (TCO), as well as the related deposition methods (in-line compatible technologies),

to optimise semiconductors as well as interfaces and specific buffers aiming at stable and highly efficient solar cells (materials engineering, source materials, deposition techniques/ parameters),

to optimise encapsulation materials as well as processes based on glass and flexible non-glass materials (damp/heat stability, costs),

to develop high band gap alloys (potential of voltage increase, top cells for tandems) and explore cost-effective tandem devices (technical feasibility),

to scale up novel, cost-effective processes (quality, reliability, throughput, cost),

to set up a new virtual EU laboratory for device analysis and modelling of solar cells, to supply outstanding highly sophisticated and well-matched analytical methods for materials and devices and to develop modelling tools for performance optimisation (cross-linking of analyses),

to identify machinery requirements for production and to enable European manufacturers to improve and supply machinery for large-area manufacturing. The focus of the process development is on throughput, yield, quality and cost,

to identify and solve performance-related problems arising from the rigid glass substrates as well as from flexible substrates (physical/chemical properties, type of glass, metallic and polymeric foils, cost-effectiveness),

to identify suitable in-line compatible patterning methods for super- and substrate modules, to develop alternative monolithic series interconnection methods (quality, throughput),

to identify potentials for the reduction of energy consumption, material usage and waste, to develop improvement strategies,

to assess societal benefits and risks from large-scale technology implementations and to elaborate strategies for a more sustainable energy supply in Europe,

to provide training and to promote mobility for students and young scientists.

1.3Project structure and partners

The topics of the project are organised in six sub-projects. Two of them are mainly driven by industry partners. Sub-project “Chalcopyrite Specific Heterojunctions” aims on the optimisation of large area CIS modules in terms of materials and cell efficiencies, whereas suppliers for production equipment are developing suitable machinery for large are modules based on “micromorph” technology in the sub-project “Thin Film Silicon Large Area Modules on Glass”.

Four of the sub-projects are mainly driven by research institutions. “High Efficiency Solar Cells” aims on efficiencies beyond the state-of-the-art for the technologies in the project. Activities comprise also new cell concepts, i.e. tandem solar cells based on chalcopyrite materials and the use of foil substrates for flexible solar cells. Vacuum free processes, i.e. electro-deposition for PV materials are developed and evaluated.

The sub-project “Thin Film Module Technology” focuses on module aspects. Topics are isolated substrates, contact technologies, Encapsulation, serial interconnection and demonstration.

A wide range of optical, electrical and structural analysis techniques are provided to the consortium by the sub-project “Device Analysis and Modelling”. Device simulations act as interpretation tools for measurement data. The objective is to get a better understanding of the structural and chemical properties of the cells and to provide a data base for the project.

The aim of sub-project “Sustainability, Training and Mobility” is to ensure that the work undertaken will have a positive impact on energy production, quality of live and the environment. Beneath the socio-economic impact, the training and mobility of the participating scientists are supported.

The consortium is composed of seven industrial partners, ten research institutes and seven partners from the higher education. The partners reflect the different technologies in the project and they have complementary expertise.

Helmholtz Zentrum Berlin für Materialien und Energie (HZB) acts as the co-ordinator of the project and provides its expertise in CIS and in thin film polycrystalline silicon technology.HZB is supported by scientists from the Freie Universität Berlin. Research on CIS technology is also domain of the Zentrum für Sonnenenergie- und Wasserstoff-Forschung (ZSW), thin film polycrystalline silicon is in the focus at the Interuniversity MicroElectronics Center (IMEC). A further technology present in the project is known as the micromorphous technology – tandem solar cells based on microcrystalline and amorphous silicon. This cell type is under research at the pioneering Ecole Polytechnique de Lausanne (EPFL), at the Forschungszentrum Jülich (FZJ) and in part at the Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas (CIEMAT). Flexible thin film solar cells offer advantages in application as well as in processing. Flexible cells on basis of CIS are investigated at the Swiss Federal Laboratories for Materials Testing and Research (EMPA). Work on vacuum free deposition processes, mainly on electro-deposition, is done at the Photovoltaic Energy Development and Research Institute in Paris (IRDEP).

A range of complementing research partners are contributing to common and different aspects of the various technologies. Universities of Gent (UGent), Lubljana (ULjub) and Prag (IPP) are performing device analysis and modelling. Process modelling of PECVD reactors is done at the University of Patras (UPat).

Development of advanced module technology is important for all thin film technologies for deriving a final product of proven quality. Energy research Centre of the Netherlands is dealing with this aspects.

University of Northumbria (UNN-NPAC) and Institute for Futures Studies and Technology Assessment (IZT) are accessing the technologies developed in this project in terms of their socio-economic impacts.

The industrial partners are aiming on the production of advanced solar cells and modules and setting up improved production equipment and facilities. The following partners represent the different technology paths in the project.

AVANCIS, former Shell Solar and Sulfurcell (SCG) are producers of large area modules based on compounds of the CIS family. Innovative upcoming products are flexible PV solar cells. They are developed by Solarion using polymer foils as substrate. Applied Materials (AMAT) and Oerlikon Balzers are both suppliers of production equipment for coatings. Both companies are developing PECVD systems for the deposition of thin silicon layers. Schott Solar is known for its solar modules based on silicon wafers and amorphous silicon. The company provides TCO-glass and selected functional layers for thin film cells and modules. Pre-industrial TCOs are also developed and delivered to the project partners by Saint Gobain Recherce (SGR).

1.4Progress within the project duration

1.4.1SPI - High Efficiency Solar Cells

Lighttrapping was found to be most important in silicon based devices but might become a critical issue for thinner CIGS devices as well. For this reason we studied the light trapping in all ATHLET thin film technologies by applying the rough interfaces as source for light scattering in identical µc-Si single junction solar cells. We revealed significant light trapping in plasma-textured polycrystalline silicon, but also severe optical losses due to seed layer and BSF. Interestingly, light trapping was concluded to be also present in several CIGS devices just by the intrinsically rough growth of the absorber. This conclusion is of high importance for further developments of several thin film technologies.

At low growth temperature (<550°C) solar cells on stainless steel foil with or without diffusion barrier reached efficiencies up to 12%. To improve cell performance further, higher temperatures are needed and thus a diffusion barrier against Fe diffusion to the absorber is required. With a SiOx diffusion barrier layer high efficiency ( 15%) CIGS solar cells on stainless steel substrate were achieved by an in-line co-evaporation process.The installation of closed loop control and improved evaporation sources at Solarion led to an increase in efficiency to currently 13.4% (confirmed by ISE Freiburg). Due to better homogeneity and process stability also the overall process yield has been improved significantly. Additionally improved deposition rates are feasible by new evaporation sources, but further work is needed. An electro-deposited ZnO:Cl layer performed similarly as the sputtered ZnO:Al layers of Solarion showing the potential of electro deposition for in-line processing of CIGS cells on foils. The CIGS tandem device development progressed in terms of top cell transparency of 55% at cell efficiency of 9.1% and simulation tools were developed to predict the respective tandem cell performance. Additionally, mechanically stacked tandem devices were prepared using CIGS wide gap material as bottom cell and a-Si as top cell and further experiments are planned.

Initial peak efficiency was 13.3 % for thick (> 3 µm bottom cell) tandem cells and 12.5% for around 2µm total absorber thickness at the beginning of the reporting period. During the last year of ATHLET we studied the interrelation between intermediate reflector and surface morphology of TCO, glass as well as the intermediate reflector itself on electrical and optical cell performance. Electrical performance could be improved by smoother surfaces, however, cell current decreased. This interrelation makes device optimization quite difficult. On the other hand high currents close to 15 mA for transparent a-Si top and 30 mA for µc-Si single junction devices could be achieved which was identified as one major milestone for high efficiency devices. Silicon deposition process was optimized in order to control plasma conditions and avoid or intentionally induce short and long term drifts of the process conditions. However, cell efficiency could not be improved further and best efficiency is still at 13.3% though at slightly reduced absorber thickness. Another approach on extremely thin tandem devices reduced the total process time to one half while keeping the stabilized efficiency nearly constant (9.8 -> 9.6 %).

At the end of the third year we showed a best cell efficiency of 8.9% in the high-temperature route by combining plasma texturing with heterojunction emitters to improve the current density and the Voc of our cells. Progress was achieved on cell level by thinning of the seed and back surface field layer to reduce optical losses and on module level by a new preparation procedure of solar modules which will be applied for best solar cells in future. Unfortuantely, both approaches have not yet lead to higher efficiency of cells and modules due to the problems with the plasma texturization reactor so light trapping was not applied. The best cell efficiency in the high-temperature route at the end of the project is therefore still the 8.9% on alumina substrates and 6.4% on glass-ceramic (1 cm2, active area).

The best results for solar cell in the intermediate temperature route on glass at the beginning of this reporting period were the following: = 3.2%, VOC = 407 mV, JSC = 11.9 mA/cm², FF = 67%.Investigation of light trapping by plasma texturization were performed, but due to the worse grain structure of silicon on glass as compared to the high temperature route the texturization led to shunting of the cells and inhomogeneous absorbers. A significant improvement in Voc was achieved by plasma hydrogenation and rapid thermal annealing leading to Voc values of 450 mV in several experiments. However, solar cells with improved efficiency were not achieved yet. The development of a suitable light trapping will be the major topic for poly silicon devices in general.

1.4.2SP2 - Thin Film Module Technology

In the first period of ATHLET, experiments on mechanical terminal contacting were performed using various contacting options. Then the objective was to select the options most viable for practical use in factory or at field site. Three good workable options have been identified for climate room testing: the press connector, the SMT nut connector and the strip connector, all combined with proper junction boxes. Up to now some significant differences appear in initial contact resistance as well as in contact resistance degradation rate. Two new methods developed, i.e. using the press connector or the SMT nut connector, show behavior quite better than observed for the more state-of-the-art strip connector, in particular for initial resistance. These two are also degrading less, but longer testing time is required for a final judgment and for discriminating between the two.

The idea of water tolerant encapsulation, originating from x-Si technology was transferred to the use for f-Si. A particular dimension is given by the flexibility aspect; this necessarily introduces polymer encapsulants, also at the front side, that are never water tight. The essential issue is a good functionality - cost optimization by balancing encapsulation quality and PV technology robustness. The next step, the extrapolation for CIS technology could not be made: it turned out that no flexible CIS technology samples could be made available within the consortium, and thus deliverable DII.3.25 could not be realized. In order to have a comparison anyway, an experimental evaluation has been done on rigid f-Si technology provided from the outside-SPII partner FZJ. In this way a comparison could be made with a more robust technology in stead of with a more vulnerable one. These activities finalized the work package.

For the development of insulating coatings on metals a detailed analysis was performed about the influence of the thermal expansion coefficient of the substrate material, the substrate roughness, the influence of high-temperature CIGS deposition and a cleaning process between two SiOx deposition steps on the barrier properties. As one result it became clear, that the realisation of a perfect insulation barrier with a high barrier resistance and high breakdown voltage values becomes more and more difficult with increasing substrate area. Nevertheless, it was possible to realise the up scaling of insulating barriers to > 300 cm2 with a high barrier resistance and disruptive discharge voltage of > 100 kV/cm. However, occasional shunts could be found in each of the barriers.

1.4.3SP3 - Chalcopyrite specific heterojunctions and TCOs

From the different processes under development in SP III to replace the CdS buffer layer in a CIS-type module, potential candidates for near-future implementation in production lines have been evaluated. Although a significant progress has been made during the lifetime of ATHLET, different limitations and open questions impede the direct application of the developed solutionsat the end of the project:

  • A reliable CBD process for the deposition of Zn(S,O) layers has been developed. On Cu(In,Ga)(S,Se)2absorbers, deposition times are at least comparable to CBD-CdS and the same kind of deposition equipments can be applied. The highest performance of a Cd-free 30x30 cm² module within ATHLET has been demonstrated with this technique (13.5% aperture area efficiency). On CuInS2 absorbers, the process window for optimal device performance (7.4% peak efficiency)is not as wide as with the CdS standard process. Metastability of the device performance and the necessity to light-soak the modules to determine the module power are actually seen as the major hints for introduction of the CBD process in module production.
  • The difference between CdS-buffered devices and devices with sputtered buffer layer on CuInS2 absorbers can be small but is believed to still be statistically significant. Both (Zn,Mg)O and Zn(O,S) seem to perform well on cell level and monolithically interconnected module test structures. Results on 30x30 cm² were inferior (5.9% best efficiency)–but encouraging enough considering that experience was limited to a single batch of modules. Under the assumption that the ideal case is a flat alignment, a Mg-content of ~13% appears to be ideal. The optimal value for Cu(In,Ga)(S,Se)2absorbers is slightly lower, due to the lower band gap of the absorber.In the latter case, the up-scaling was already terminated in the third project year due to stagnation of progress. It remains unclear whether the limitation in efficiency is of principle nature or due to technical problems in up-scaling.
  • Anindium sulphide evaporation processfor Cu(In,Ga)(S,Se)2 absorbers was successfully transferred to a new labline evaporator with load lock and larger evaporation sources. Continuous processing over several hours could be demonstrated with average cell efficiencies close to 12%, which is an important requirement for industrial application. The up-scaling to 30x30 cm² was so far handicapped by equipment limitations. Nevertheless 12.1% best efficiency value and 12% average of the best five modules were achieved in a batch reactor, which is only slightly below the target values.
  • For the spray-based techniques, up-scaling of the USP method towards 10x10 cm² has been successful. The deposition mechanism has been studied and the investigations have been expanded towards ZnS layers. Some preliminary experiments with a first industrial ILGAR in-line machine (substrate size up to 30x30 cm²) have been performed. On the lab scale, mixed ZnS+In2S3 ILGARlayers led to inferior cell efficiencies than pure In2S3 or stacked ZnS/In2S3layers. Results have been compared with buffer layers deposited by ALCVD processes.

With the new 1% doped Zn:Al target, the quality of the reactively sputteredTCO films could be substantially improved. It was possible to achieve films with nearly identical performance like the reference films sputtered from a ceramic target. Applied to solar modules, this resulted in 13.4% or 12.8% best module efficiency for Cu(In,Ga)(S,Se)2 or Cu(In,Ga)Se2absorbers respectively. Furthermore, reactive TCO coating of a few modules with alternative Cd-free buffers was successful leading to efficiencies as high as 11.7 % with high photocurrents, suffering only from a lower fill factor.