Within the framework of CrystalClear, a large European project on the development of low-cost, highly efficient silicon solar modules, a new manufacturing process has been developed that allows modules to be made using very thin solar cells. Wim Sinke explains.
An important goal of the current development of solar modules is to enable solar electricity production at costs that are competitive with retail prices, thus achieving so-called grid parity. Although Table 1 (shown on page 66) which is taken from the Strategic Research Agenda of the EU PV Technology Platform, does not refer to a specific photovoltaic module technology, it is obvious that wafer-based crystalline silicon – because of its very large market share – will play an important role in the overall picture for the short and medium term. For grid parity to be reached, the turn-key system price needs to be brought down roughly into the range of c2–2.5 per Watt-peak (Wp). Considering the fact that a turn-key system comprises modules as well as the so-called balance-of-system, which includes cabling, electronics, mounting materials, and labour – and acknowledging the difference between price and cost – it has been inferred that module manufacturing costs need to be lowered to typically c1/Wp to reach grid parity. It is no coincidence that this is the main project target of CrystalClear, a 5-year joint effort of a consortium of European companies, research institutes and university groups involved in crystalline silicon PV technology.
Based on an analysis of actual manufacturing practices made in the framework of the CrystalClear project, the cost of wafer-based crystalline silicon modules was estimated to be typically c2/Wp of module power at the end of 2005. This number is for the relatively small production volumes that were then common, at roughly 30 MWp per year. A detailed analysis of the possibilities for further cost reduction has shown that it is possible to reach a cost level of c1/Wp, using a combination of technological innovations on the levels of silicon feedstock, wafers, cells and modules, and of economies of scale in production.
The second factor implies the introduction of cell designs and processes that give high efficiencies. In addition, low module-related losses, for example those related to cell interconnection, and close cell packing in the module, can help to reduce the cost per Wp.
Finally, high throughput and high yield processes need to be compatible with the use of very thin wafers and with high efficiencies.
The role of conductive adhesives
Within the CrystalClear programme, several approaches towards low-cost crystalline silicon modules have been designed and are now under final development. These so-called ‘overall technologies’ represent distinct cell and module concepts and are based on different silicon feedstock and wafer options. The cell concepts can be divided into front and rear-contact designs and all-rear contact designs based on – among others – metallization wrap-through (MWT) and emitter wrap-through (EWT) technology. In the MWT design the cell still a metal front electrode, although of a distinctly different design, but the collected current is transported to the rear through an array of small metallized wires in the wafer. In the EWT design, on the other hand, the front metal grid has been deleted altogether. Current collection relies on the conductivity of the very thin silicon top layer, the so-called emitter. Since this conductivity is modest, the current is tranported to the rear through a much denser array of wires.
The MWT and EWT concepts have in common that both the negative and the positive pole can be copntacted at the rear of the cell: one directly and the other because it is ‘wrapped through’ to the rear. Rear-contact cell designs require a dedicated interconnection scheme, or rather, allow the use of an elegant new interconnection technology. They may be interconnected using ‘smart tabs’ or using a conductive pattern integrated in the module back sheet. Figure 1, below, shows a schematic view of rear-contact cell structures.
Since all technologies in the CrystalClear roadmap are based on large area, very thin cells (typically 0.12 mm), low stress interconnection methods are required. The soldering process normally used to connect the flexible metal strips, called tabs, to the cells results in a rigid interconnection and residual stresses due to thermal expansion coefficient mismatch between silicon and the solder. During service, the temperature cycles seen by the interconnection may result in damage to the silicon. The interconnection will result in micro-cracking of the silicon and eventually pull-out of silicon from the cell. This will be observed in a decreasing module efficiency and ultimately, in failure. Alternatively, cracks can develop in the solder itself resulting in an increase in electrical resistance through the interconnection and therefore reduced module efficiency.
This is where the use of conductive adhesives, which can be used as an alternative to solder joints, come in. The required curing temperatures are well below 150oC and thus substantially lower than the temperatures involved in processing of lead-based solder alloys which are around 220oC. The lower temperatures involved during the manufacturing process for the electrical contacts between the solar cells and the interconnecting tabs or the patterned back sheet results in lower mechanical stresses after cooling to room temperature and thus, in improved handling of very thin cells. Moreover, because the temperatures used for curing the conductive adhesives are comparable to those used for module lamination, it is possible develop a ‘single shot’ module manufacturing process in which cell interconnection and module lamination are combined in single process step. In the case of rear contact cell designs and interconnection using a patterned back sheet, this leads to a very elegant and potentially rapid module manufacturing sequence. Small dots of conductive adhesive are applied onto the patterned back sheet, which may be combined with a punctured ethylene vinyl acetate (EVA) foil – a material commonly used in module lamination – cells are positioned using a pick-and-place technique and the module lay-up is completed by an additional EVA sheet and a cover glass. After this, the actual module is made by heating the lay-up under pressure and vacuum in a laminator.
In addition to benefits related to module manufacturing, conductive adhesives also have the advantage of being lead-free. This fits well in the trend to further improve the environmental quality of solar modules.
To be a successful replacement for solder, the conductive adhesive has to meet a number of criteria. Any increase in resistance of the interconnection must be minimized so that the power output of a module made with conductive adhesive is similar to that for a soldered module. The mechanical strength of the interconnection must also remain sufficient to allow manipulation for further processing and to be able to survive possible stresses imposed during the 20-year or more service life of the module. The interconnection must be able to survive testing protocols, for example IEC 61215, without significant loss of module power output. Finally, the adhesive must also be available in a form that allows efficient processing when making the module.
Since the quality of the interconnections and the encapsulation is decisive for the lifetime and reliability of solar modules innovations in this area require very thorough testing and are extremely rare. CrystalClear researchers have therefore built dedicated equipment to enable highly reproducible application and curing of adhesives.
With this equipment, strings of cells have been made that have subsequently been encapsulated by industrial partners. The cells used in the testing programme varied in silicon type, thickness and technology, but only front and rear contact cells and interconnection tabs were used. The strings produced using the conductive adhesives were encapsulated in modules adjacent to strings made with similar cells interconnected by soldering to allow a direct comparison of the two technologies during climate chamber, outdoor and mechanical tests.
The adhesive can be epoxy, acrylic or silicone-based with the conductive component provided by silver flakes added at 75%-85% by weight. During curing of the adhesive, the silver flakes are forced into contact with each other providing a conduction path through the adhesive. The choice of adhesive is determined in part by the processing necessary for curing. If the aim is to make only small modifications to current manufacturing practice, a system is needed which requires a minimum of adaptation of the cells and of the process used for interconnection. If the introduction of conductive adhesives is part of a more drastic innovation in module manufacturing, the window of opportunity becomes wider and other systems may be suitable or required. Furthermore, while curing of the adhesive can take place during lamination, it can also be implemented as part of a tabber-stringer process with curing by a heat source such as infrared lamps.
Of the several types of conductive adhesives that have been evaluated within CrystalClear, snap curable acrylic-based adhesives have been selected for the extensive testing reported here. ‘Snap’ refers to the very short heating time of only a few seconds in order to cure the adhesive at a temperature of 120o-150oC. Modules were manufactured using the conductive adhesive with industrially manufactured thin crystalline silicon solar cells. The best results were obtained for an adhesive with a high silver filler content of approximately 85%.
Although the contact resistance of the adhesive is much higher than for solder, it does not affect the module output power since it is not a limiting factor. The adhesive also resulted in peel strength similar to a soldered interconnection.
As examples of the range of test results obtained, Figures 2 and 3, below, give quantitative data for two important types of tests: thermal cycling and damp-heat. These tests are part of the IEC61215 test sequences that are required as part of module qualification.
As Figure 4 shows, degradation in fill factor of less than 2% was measured after 500 thermal cycles, while over 1000 hours of damp-heat testing resulted in a degradation of less than 4%. The fill factor, along with the open circuit voltage and the short-circuit current, determines the power conversion efficiency of the module. It can thus be interpreted as a measure of the electrical quality of the module and the components inside. Other modules showed no degradation at all after 1000 hours of damp-heat testing. The results comply with the IEC criterion of <5% degradation.
Testing on full-size modules to prove the suitability of the conductive adhesive in outdoor use is ongoing and, after some six months of exposure, no degradation is observed.
It’s a stick up
As the results from the programme show, it is possible to use an acrylic-based conductive adhesive as an alternative to soldering. The short curing time of the adhesive is comparable to the time necessary for making soldered interconnections, the interconnection’s electrical is comparable to that of a soldered interconnection at module level, and the durability of the interconnections has been found to be as good as for soldered modules using similar cells. Furthermore, the greater elasticity of interconnections made with conductive adhesive and a lower processing temperature as compared with soldering allows handling of very thin cells.
The use of very thin silicon wafers and hence, of very thin cells, is an important part of the strategy to reduce wafer silicon module manufacturing costs. By combining efficient silicon use with high cell and module efficiencies and high-throughput, high-yield processing it is possible to reduce module manufacturing costs to a1/Wp or even less. This cost level is compatible with the turn-key system price needed to achieve grid parity, the ultimate goal for all PV manufacturers.
Wim Sinke is Co-ordinator of the CrystalClear project, based at the Energy research Centre of the Netherlands (ECN), this article was produced with contributions from I.J. Bennett, P.C. de Jong, M.J.H. Kloos, C.N.J. Stam (ECN), A. Henckens, J. Schuermans (Emerson & Cumming), R.J. Gomez (BP Solar Spain), P. Sánchez-Friera, B. Lalaguna (Isofoton), H. Schmidt (SolarWorld Industries Deutschland GmbH), and E. van Kerschaver (IMEC).
You can contact Wim Sinke at email@example.com
This work was conducted as part of the Integrated Project CrystalClear and funded by the European Commission under contract no. SES6-CT-2003-502583.
CrystalClear is a large, five-year joint effort of a consortium of European companies, research institutes and university groups involved in crystalline silicon photovoltaic (PV) technology. It is an Integrated Project carried out in the 6th Framework Programme of the EU and the overall aims are:
- Research, development, and integration of innovative manufacturing technologies that allow solar modules to be produced at a cost of a1 per watt-peak (Wp) in next generation plants;
- Improvement of the environmental profile of solar modules by the reduction of materials consumption, replacement of materials and design for recycling;
- Enhancement of the applicability of modules and strengthening of the competitive position of photovoltaics by tailoring to customer needs and improving product lifetime and reliability.
Realisation of these aims is a necessary condition for the European PV industry to maintain and strengthen its position on the world market and for photovoltaics to fulfil the expectations and policy targets.
CrystalClear is running from January 2004 to December 2008 and has a total budget of a28 million. Of this amount a16 million is contributed by the EU and a12 million by the 16 partners:
German industry partners:
- BP Solar (ES)
- Deutsche Cell (DE)
- Deutsche Solar (DE)
- Isofotón (ES)
- Photowatt (FR)
- REC (NO)
- REC ScanWafer (NO)
- Schott Solar (DE)
- SolarWorld Industries (DE)
For more information visit CrystalClear.