Poly-Si production technologies show a potential for improvement. The FBR (Fluidized Bed Reactor) is expected to increase its share against Siemens processing as shown in the figure below.
(Figure 1: Expected change in share of Poly-Silicon production technologies)
Other technologies such as Upgraded Metallurgical Grade – Silicon or UMG-Si are not predicted to yield significant cost advantages compared to conventional poly-Si technologies over the coming years.
Discussions with producers have revealed that major poly-Si quality reductions do not lead to significant cost reductions. Despite the possibility of using 6N material for p-type mc-Si cells, its application would not result in cost reductions as the price is not significantly lower than that of 9N . We foresee that a PV-grade poly-Si quality of between 8N and 9N will be sufficient for current and future p- and n-type cell concepts, another reason for this being the fact that feedstock purity is not the sole variable affecting wafer quality.
A significant improvement in cost reductions in the wafering process is expected due to the introduction of diamond wire sawing, especially for mono wafers. Diamond wire sawing is expected to become widespread for mono-Si wafering; however, the field is open with regard to mc-Si wafering. Other new wafer manufacturing techniques, especially kerf-less technologies, are not expected to gain notable market shares due to the maturity of the established sawing technologies. The Figure 2 below shows the expected share of wafering technologies in volume production. The roll out of diamond wire sawing technology requires synchronization with cell process development.
(Figure 2: Respective market share of wafering technologies for mono and mc-Si)
The ITRPV (International Technology Roadmap of Photo Voltaics) survey also showed that the industry uses various technologies with regards to recycling various consumables with the aim of continuing cost reductions. However, it will be necessary to standardize recycling processes to further decrease the production costs of wafer slicing both with regard to the necessary purification level and sufficient costs of ownership. While silicon carbide recycling is already used in numerous fabs, the re-usage of silicon, slurry and other consumables needs further development.
(Figure 3: Expected trend of wire diameters and grit particle size)
Material savings can be achieved by using thinner wafers or reducing kerf loss.
The main driver of kerf loss is the core wire diameter, as shown in Fig. 3. More sophisticated wire management and control is needed to implement thinner wires in mass production. At the same time, kerf loss can be reduced using smaller particle and diamond sizes. Diamond wire processes use bigger particles and there is therefore great potential to develop processes using smaller diamond sizes.
Consumables such as crucibles, graphite parts, slurry and sawing wires provide opportunities for cost reduction in crystallization and wafering. We foresee that prices will also be affected by the cost of these goods and price reductions of between 5% and 10% per year, as shown in Fig. 4, seem possible.The increasing share of diamond wiring means that the price level of diamond wires will need to reach 25% of today’s price by 2023.
(Figure 4 : Expected relative price reduction for key consumables in crystallization and wafering)