Presentations

Bernhard Dimmler, Würth Solar

The new industry targets set by EPIA have foreseen an annual market of about 160GW in 2020, with 8-90GW being installed in that year in Europe. Although different assumptions exist on the market share of the various technologies (c-Si, Thin Film, CPV and emerging), everyone agrees that there will be a balanced share among all of them. EPIA estimates that Thin Film (TF) will cover more than 30% of the total market in 2020. This represents an annual production of 50GW of TF (70 times the production in 2008).

This could be possible from a technological point of view as the industry has demonstrated over the last 30 years, with c-Si module prices following a price experience factor (PEF) of 22% (module price decrease of 22% each time there is a doubling of the cumulative installed capacity). Based on this fact and experience from other sectors like the Flat panel display industry, we can expect similar PEF for TF modules and we could expect TF module prices below 1€/Wp in the near future.

Besides cost reduction, one of the major challenges for TF technology is to increase its efficiency in order to reduce the balance of system (BOS) cost. The European Photovoltaic Technology Platform foresees that by 2030 all TF technologies will be above 15% efficiency and this could go up to 25% in the case of CIGS.

Peter Rigby, Umicore

To summarise, we could see that materials (Gallium, Indium, Selenium, Tellurium) will be available for PV sector demand in the medium term, however:

• improved cooperation with other sectors like the copper industry is needed in order to avoid price volatility and
• optimised extraction, refining-recycling yields will need to be developed and applied to ensure being able to meet demand in the long run.

When looking at material availability in the future, it is crucial to evaluate the development of other technologies which compete for the same material. Some of the conclusions for certain materials are summarised below:

• Gallium: Is largely available but may well suffer from sporadic price volatility due to refined capacity lagging demand.
• Indium: Shows no problem of availability. However, to meet such a high demand, serious efforts will need to be made in increasing yields in both extraction & refining as well as production scrap collection logistics.
• Selenium: Reserves are large but recycling needs to be optimised. No real shortage is foreseen.
• Tellurium: Availability could be an issue in the future depending on the technology choices in the copper industry and the willingness to optimise recovery.

Stephen Carney, LINDE

A cost target for glass can be set at 1€/m² compared with the current cost (13-21 €/m²). To make this target possible the glass industry needs to look at different aspects in order to make a significant cost reduction:

• Integration of in-line processes like AR coating and TCO layers will help considerably. Other processes like toughing will soon be possible also.
• The replacement of solar grade sand as raw material by other sand with higher iron would help to keep the cost down. For this, glass chemistry needs to be manipulated in order to keep low iron in the final glass.
• The choice of the furnace type will have a major impact on the investment cost and energy needs for the operation.
• The choice of the process (float glass Vs. roll to roll) will determine the thickness of the glass which has a direct impact on the transmittance, as well as the use of oxidised glass chemistry which has an environmental impact. Roll to roll has a limitation as glass thinner than 2.2mm could not be treated with today’s technology. 

All these aspects are crucial in order to keep the cost of glass low as it already represents a significant part of the total module cost and this percentage may increase in the future due to the fact that other processes (wafering, cell production, raw materials, lamination, etc.) continue to optimise and reduce the final cost of a Wp.

Gert Doucet, Dupont

The materials (back sheet, aluminium paste, encapsulants, silver paste and glass) represent about 10% of the cost of the system. Up to 30% of poly-silicon is included.  Therefore innovation in this field (optimising packaging and BOS design, etc.) will be important to bring costs down.

The material choice significantly influences the reliability and performance of PV components and systems, and therefore their PV systems lifetime. An increase in the lifetime from 20 to 40 years would contribute to an LCOE decrease of 20% which is critical to the success of PV technology.

In terms of availability, the major challenge is to have punctual production capabilities to supply material under the very high demand conditions. Dupont has foreseen that, excluding silicon, more than 25b$ (85b$ including silicon) needs to be invested in material production in order to match the demand foreseen by EPIA in 2020.

Gases, Olivier Blachier, Air Liquide Solar

Looking at the supply of gases and chemicals, it does not seem to be a major technological issue in terms of expanding capabilities; however, access to the credit may represent a big bottleneck.

For a cost-efficient supply of gases and chemicals, the way forward lies in the creation of production clusters in which the production of feedstock, wafer, cells, modules and other materials like glass are produced in the same site.  Already existing clusters for c-Si (e.g. Thalheim, Wuxi and Osaka) have shown to be a very good solution for logistics which can significantly reduce the cost and time of production.   This is still not the case in Thin Film where major sites are not larger than 500MW. In order to the reach the 12% target (a production of 50GW of Thin Films annually), more than 10 campus of higher dimensions (1-2GW) would need to be in place.

Regarding the cost of gases, this represents 3% in c-Si and about 11% for Si- Thin Films. This means that innovation in this area will have major contributions to the final cost of Thin Film modules. Some of these activities should focus on reliable onsite generation and recovery technologies, as well as on precursors in CVD & PVD.

Thomas Pellkofer, Applied Materials

The price-experience factor of 20% for the crystalline silicon PV industry originates from the benefits of economies of scale, automation, increased efficiency and decreased wafer thickness and kerf losses. Bearing in mind that the value chain of thin film is more integrated than the one for c-Si, there might be even more opportunities for economies of scale in the thin film industry, which could lead to an even higher price-experience factor.

To reach the targets of the paradigm shift scenario, a price-experience factor of 25% for thin film is needed. Glass coating only has a price-experience factor of 17%. However, in industries that are closely related to the thin film industry, the price-experience factor is a lot higher. For example, in the flat panel display sector using TFT-LCD panel technology (which includes PVD, the price-experience factor is about 35%). Especially enlarging the area for PVD offers great opportunities to drive costs down further.

However, BOS costs for thin film represents a large share of the total system cost and this tends to increase as we decrease the module cost. Thus, efforts to increase efficiency will remain of the utmost importance.

Gerold Buechel, Oerlikon

Oerlikon’s roadmap to reach grid parity translates itself into a goal to reach 0.54€/Wp by 2010. For Si TF, plasma-enhanced chemical vapour deposition can contribute to lowering the costs of thin film modules. PECVD offers high quality (low ion bombardment, uniformity and no contamination), high throughput (high deposition rates) and high productivity (high capacity equipment). Moreover, PECVD has predictable downtimes and the highest yield and uptime possible.

PECVD with very high frequency (VHF) is necessary because it leads to lower ion bombardment which, in turn can limit defect density so that better quality material can be provided. The challenge here is that a higher excitation frequency might cause non-uniformity on large scale devices which makes it impossible to apply on large modules.

Important developments with respect to equipment material have been, the improvements of the throughput and the quality of the absorber material, more isothermal heating, reducing contamination as well as the footprint of the equipment device.

Claus Kuhn, Manz

Based on the calculations of Manz Automation, the 12% target for 2020 is feasible for the laser equipment industry. They estimate that each year 50 fabs with a size of 100-120 MW should be installed. Each fab should have about 3 to 6 scribing systems and 1 or 2 edge ablation systems.

To maximise the throughput of the system and to reduce costs is of the utmost importance. Therefore, decisions need to be made on the level of the motion platform and the laser optics; as for instance, increasing the accuracy and speed of the motion platform, improving the dynamics of the processes, implementing a modular structure, introducing a fully integrated laser optic module with a long lifetime, etc.

Considering the latest developments at Manz Automation, these goals have been achieved. Throughput can be increased 4 times and lines with 3 systems (1 for each process) can be developed. Moreover, costs have been reduced. Finally, local manufacturing for local markets is an important strategy.

Ronald Lange, 3S

At the back end of PV module production, the following steps take place: the encapsulation or lamination and the post-lamination stages, such as edge trimming and framing and module testing. Encapsulation plays a major role on determining the quality and lifetime of the module and it is, in general, irreversible. Post-lamination steps require automation and the main challenge is the reliability of the module. Module testing is important to determine the Wp and likewise the cost of the module. The difficulty here is to obtain spectral matching and stability.

Thin film currently makes up about 15% of the PV market (from which CdTe is the most common, 73%). More than 80% of thin film modules make use of PVB as an encapsulant and 92% are glass-on-glass modules (compared to only 8% of flexible thin films).  The future conditions for thin film technology are promising, considering that thin film has good properties for applications such as BIPV (glass facades) and the combination of thin film and crystalline silicon technologies in heterogeneous cells (HET cells) is possible. The main application will be in large solar farms using large, frameless, glassless modules that are adapted to cope with mechanical forces.

Taking into account the balanced relationship between costs, stability and efficiency, cost reductions should be dealt with carefully. In the short term, the cycle time of lamination can be reduced. This cannot happen without increasing the quality in the first place, for example by introducing in- or on-line testing after lamination, RFID tags and light-soaking experiments during the lamination process. Other aspects, like the replacement of glass, an alternative application of encapsulants and other techniques to significantly reduce the cycle time can be implemented only in the long term.

Arnulf Jäger-Waldau, EC DG JRC

According to the International Energy Agency (IEA), the demand for water and energy will increase considerably (30% and 50% respectively by 2030). For the manufacturing of PV modules, both water and energy are needed.

Increased efficiency of photovoltaic devices and optimised production (reducing energy requirements) has led to an enormous decrease in the energy pay-back time (EPBT) over the past decades to reach approximately 1,5 years to even less than 1 year for good locations (high irradiance). However, the expected further decrease (until 2020) in EPBT is relatively small. Hence, in order to reach the 12% target (160GW annual production), a large amount of energy (about 1% of the EU electricity production in 2020) will be required for the manufacturing chain.

Moreover, when conventional energy sources are used for the manufacturing of PV modules, water requirements will remain high, as for conventional power generation, a large amount of water is needed. The water needs to attain the 12% target would be equivalent to the water needs for about 140 to 240 TWh of thermoelectricity. 

This indicates that further reductions in the EPBT and the water needs related to PV manufacturing are top priorities, not only because of environmental but also economic reasons.

Daniela Schreiber, EuPD

Recently, increasing competitive pressure and a downturn in demand have triggered falling prices. The question arises as to where the large amount of planned capacity can still be installed. Because the German market cannot take up all the excess planned production capacity, new emerging markets should be developed.

Thin film is becoming more and more recognised as a product category on its own, rather than a low-cost substitute of c-Si devices. The commercialisation of thin film technologies has triggered product differentiation. The different characteristics of thin film compared to c-Si (with respect to hot temperature and low-light behaviour) indicate that thin film might serve certain new markets and market segments better than c-Si does. Hence, thin film cannot be considered as direct competition to c-Si applications, which normally seek higher efficiency solutions. However, it offers opportunities for the development of new markets and market segments that are necessary for the deployment of PV in general, as, for instance, building integration.

There are 141 companies involved in thin film. However, most of these companies still try to gain access to the market and only few companies are currently active on the market. One of them is FirstSolar, producer of CdTe, a material that accounts for more than 60% of the total thin film market.  Access to finance and developing a stable production process at low-cost are key elements for the future success of new thin film market entrants.

Claudia Lüling, Frankfurt Architecture Universiy

It was mentioned that there is in fact a market for BIPV although it is very small (less than 2% in Germany). To explain this, three factors should be taken into account, namely products, processes and people. Wishes and dreams of the clients, the challenge to keep costs within budget boundaries, availability of different products, building permission procedures, licensing to build the system and warranty of the system are all factors that might limit the development of BIPV.

In order to reach 12% PV by 2020, a considerable surface will be needed (for example 200 million m² in Germany). Non-residential roofs are an opportunity for the further development of BIPV. This is in contrast to residential housing; 80% of building projects in the residential building sector are related to adaptations of already existing houses.

To encourage BIPV, the three P’s (people, products and processes) should be addressed. Examples are the Sun-Area project (www.geoplex.de) meant to raise awareness amongst customers. Moreover, it is important to dispose of data on rooftop availability and to assure alignment between the ways of thinking (from Wp to m²). On the product level, it is important to convince architects of the possibilities of the specific BIPV products (for example with respect to light). Thin film technology can offer opportunities to the further deployment of BIPV, because it can be coloured, made flexible or transparent and is easily combined with other products. BIPV can, moreover, function as a shading device. Because shading devices are very much needed in the construction sector, there is also huge potential for BIPV.

Robert Kuban, Schott Solar

Due to the enormous decrease in the price of silicon, c-Si modules have become cheaper. The price is at 1.3 €/Wp with 50% of costs related to wafer production and 50% to the production of cells and modules.

To estimate the impact of low-price crystalline modules on the thin film industry, the trade-off between price and quality should be taken into account. A lower price can indeed indicate a lower quality. Hence, for its thin film modules, Schott Solar performs lifetime tests for power measurement with severe criteria (such as 2000 hours of damp-heat testing). Moreover, the energy yield (in kWh/kWp) as well as the performance ratio (in %) is measured throughout the lifetime of the module. It seems that the a-Si modules perform well compared to c-Si and that a lifetime of 20 years might be expected. Another important advantage over c-Si is thin film’s unlimited possibilities thanks to its specific characteristics (for example its flexibility).

For Si TF, material availability does not seem to pose a problem. Moreover the yield is very good. Improvements are possible (and necessary) in the area of efficiency and equipment costs. With respect to CdTe, the excellent yield is mainly offset by the lower sustainability of the material. CIGS has a good efficiency, but performs mediocre in all other fields (equipment costs, yield and material availability); as pointed out in the presentation.

Mariska de Wild-Scholten, ECN

To assess the sustainability of thin film products, 3 pillars need to be investigated, namely Life Cycle Cost (LCC), Social Life Cycle Assessment (SLCA) and Life Cycle Assessment (LCA). Here the focus is on the last one, namely the environmental impacts of the modules from cradle to grave.

The energy payback time (EPBT) has been analysed.  It was concluded that the EPBT is highly dependent on the location. Moreover, the carbon footprint (which equals the life-cycle CO2-equivalent emissions) has been investigated. As a result, the carbon footprint of PV (CdTe in this case) is much lower than that of electricity generation from coal with CCS (carbon capture and storage), but higher than nuclear electricity generation. The main contributor to the carbon footprint of PV modules is the electricity consumption for manufacturing and the glass used in modules. Low-iron solar glass requires more energy than architectural glass. Regarding the encapsulation, Ethylene-Vinyl-Acetate (EVA) has a much lower impact than polyvinyl butyral (PVB). For Si TF, the carbon footprint of the recycling is still unknown. With respect to Si TF, NF3 and SF6 are produced during PECVD chamber cleaning. These gases have a strong greenhouse effect and a considerably longer lifetime in the atmosphere than CO2.

In order to reduce the environmental impact, it is first of all important to know the magnitude and its breakdown. Then, specific measures can be implemented. These can be the reduction of energy consumption (especially for CIGS), material consumption, efforts to reduce and avoid life-cycle emissions of NF3 and SF6 (for Si TF) and CO2 (for all technologies) and efforts to increase module efficiency (especially for Si TF), module lifetime and recycling (PV-CYCLE).