The long read: Criteria and implications for gallium-doping

There is a key rule for evaluating the market potential of any new PV technology, at least as posited by the team at Exawatt: It must deliver a significant increase in performance at acceptable cost, with minimal integration challenges. This rule remains a key tenet as Exawatt considers what directions the industry will move in over the next few years. Any technology that does not satisfy the rule may still have an important role to play in certain markets, especially in the space-constrained residential sector, but cannot be expected to become a driving force in the PV industry. This rule was set out in the article, “The Mainstreaming of Mono” (pv magazine 02/2019).

PERC, bifacial, check

The transition from a PV manufacturing landscape dominated by multi BSF cells to one dominated by mono PERC can be framed by the criteria contained within the Exawatt technology rule. The combination of rapid cost reductions for mono wafers (brought about by huge productivity increases), improvements in equipment for PERC cell manufacturing, and the significant efficiency improvements afforded by a move to mono PERC, all combined to ensure that production cost-per-watt for mono PERC modules could rival, and even beat, the lower-efficiency multi BSF incumbent. Meanwhile, integration posed minimal challenges for cell and module manufacturers, with PECVD or ALD deposition equipment and lasers for ablation simply being added to the end of existing cell lines, and so the seeds were sown for a rapid technology transition.

Bifacial modules are another example of Exawatt’s criteria in action. Significant performance benefits are available in the right deployment scenario, but until the widespread adoption of PERC, bifacial modules required a shift to n-type cell technologies that failed to satisfy the requirement for acceptable cost. With p-type PERC becoming mainstream, the manufacturing of bifacial cells and modules was possible with little cost increase, and with only very minor integration challenges for cell and module manufacturers.

In fact, the largest integration challenge for bifacial PERC has perhaps been at the system level, which explains why there was a period when manufacturers were pushing bifacial PERC modules while developers, although interested, were not yet entirely on board. As the integration issues at a system level – including modeling of energy yield, bankability and so forth – are being solved, bifacial’s market share is increasing rapidly, as part of a trend that is likely to continue.

For high-efficiency cell technologies typically based on n-type wafers, such as heterojunction (HJT) and tunnel oxide passivated contacts (TOPCon), the time when all the criteria are satisfied may be getting closer. Exawatt’s bottom-up cost modeling work suggests that GW-scale manufacturing for these technologies may approach cost-per-watt competitiveness with p-mono PERC in the 2023-25 time frame. Should cost-per-watt parity be achieved, the remaining challenge will be integration – not likely to be a significant issue for TOPCon, but liable to limit HJT to new capacity expansions.

Gallium doping

Since October 2019, major vertically integrated manufacturer JA Solar and wafer giant Longi have both acquired intellectual property licenses around gallium doping from chemicals producer Shin-Etsu (see pp. 44-47). Given this development, it is worth looking at this technology through the same lens discussed above – can gallium doping satisfy the necessary criteria for successful widespread adoption of a novel PV technology? Below, we consider each of the three criteria to get a feel for gallium doping’s chances of success.

Performance parameters

The most obvious driver for a move to gallium doping for p-type wafers is the removal of boron-oxygen (B-O) complexes, infamous for causing light-induced degradation (LID). It should be noted that solutions to the issue of LID in boron-doped wafers already exist.

“B-O LID mitigation processes are capable of completely eliminating the impact of B-O degradation in commercial cells,” Alison Ciesla of University of New South Wales points out. “There are a number of commercial tools available for this […] and the processes are actually very simple, so most or all manufacturers should be able to achieve this.”

That said, gallium-doped wafers eliminate the need for these additional manufacturing steps. This not only reduces cost and complexity in the manufacturing process, but may also be able to help mitigate light- and elevated temperature-induced degradation (LeTID), which occurs with both boron and gallium doping.

How might this work? The processes involved in LeTID are complex and are generally less well understood than B-O LID, but it seems clear that hydrogen plays an important role. Since elimination of B-O LID involves the use of hydrogen for passivation, this leads to a challenging balancing act around the use of hydrogen with boron-doped cells – eliminating LID while avoiding an increase in LeTID. Although striking a balance around the use of hydrogen is likely to be important for gallium-doped cells, it should be easier than for their boron-doped counterparts, allowing for a reduction in degradation.

There have also been reports of improved cell efficiency with gallium-doped wafers. In late April a press release from Longi claimed that its tests had shown cell efficiency increases in the region of 0.1% (absolute) for cells based on gallium-doped wafers, and that tests conducted by Chint and Aiko Solar had produced similar results. That said, others in the industry have expressed skepticism about efficiency benefits due to gallium doping, and the effect on mass-production cell efficiencies across the industry is not yet entirely clear.

Cost considerations

Increased cost is only likely to be an issue if the gallium-doping process leads to reduced productivity in ingot growth. Gallium has a low segregation coefficient, meaning that during ingot growth the dopant preferentially remains in the melt rather than the crystal, which can lead to significant variation in dopant density from one end of the ingot to the other.

In order to mitigate the potential variance in doping, ingot manufacturers may need to pull shorter ingots – a process change that could decrease productivity. However, there may be other parameters that can be varied in order to avoid a large decline, such as the ingot pull speed or the number of ingots pulled per crucible. At the time of writing, Exawatt has heard that productivity decreases for the current processes are minimal, and Longi is marketing gallium-doped wafers at the same price as its standard boron-doped wafers. Note that the cost of licensing agreements from Shin-Etsu has not been disclosed by either JA Solar or Longi.

Integration challenge

Assuming that solar wafer manufacturers have overcome the technical challenges around gallium doping without a cost increase, and that the resistivity of the gallium-doped wafers is comparable to that of their boron-doped counterparts, the integration of gallium-doped wafers into existing cell and module lines should be quite straightforward. As discussed, some process changes may be required in order to optimize balances around hydrogen passivation and LeTID mitigation strategies, but no significant retooling is likely to be needed.

N-type implications

With improvements to degradation, the possibility of increased cell efficiency, and easy integration into cell and module lines, it seems that – so long as ingot productivity can be maintained – gallium doping should be able to satisfy Exawatt’s three criteria for widespread adoption. But what might this mean for other technologies currently vying for manufacturers’ attention?

Although reduced degradation in p-type cells may remove one string from n-type’s bow, it does not look likely that it will significantly impact any forthcoming transition, which will come when TOPCon or HJT can themselves satisfy the criteria set out earlier in this article. Will n-type’s rise come before other solar wafer manufacturers feel the need to move to gallium doping? The answer is not yet clear, but moves by Longi and JA Solar toward gallium doping may well sharpen the minds – and the time frames – of their competitors eyeing any upcoming n-type transition.