Bill of materials key to solving next-generation solar module failures

Bram Hoex, professor and deputy head of school (Research) of the School of Photovoltaic and Renewable Energy, UNSW Australia, tells PV Tech Premium that most failure modes in n-type modules can be avoided by having the right bill of materials. However, there are no 25-year-old TOPCon or HJT modules that are representative of the current technology to prove the longevity of either technology in the field.

Furthermore, since many of the newly discovered failure modes cannot be detected by standard industry testing, there is an urgent need for cell-level testing to become an industry-wide norm, especially given the rapidity with which PV manufacturers are transforming brand-new developments into end products.

Ultimately, Hoex warns against making technology choices motivated by upfront high efficiency alone, without sufficiently factoring in the negation of benefit if there is a swift decline in performance in the early years of operation. Moreover, he fears that some new US manufacturers, without the extensive in-house testing capabilities of Chinese giants, may get caught out, as they are not aware of the specific sensitivities of the new cell technologies nor the effort required to select the adequate module materials.

Cell-level testing

UNSW performs testing at the level of a PV cell, an approach that manufacturers tend to avoid as there are few standards encouraging it. The university’s work is picking up much attention and it wants to bring this capability to the rest of the market. Progress is already being made with one of its cell-level tests already accepted as a SEMI test standard. This rapid test for potential-induced degradation (PID), which is 100 times quicker than previous tests, was adopted in March this year.

UNSW is now working with the major Chinese manufacturer Canadian Solar to develop faster and more detailed cell-level tests that can be reproduced as standard.

Typically, removing cells from a module without damaging the sample is challenging, but by doing the test directly to investigate specific areas such as the metallisation, for example, the cause of the defects may be uncovered.

“It is this methodology that allowed UNSW to discover that aluminium in the front paste is the main issue for TOPCon,” says Hoex.

TOPCon metallisation concern

Unlike previous generation solar cells, the key concern for TOPCon degradation lies in its front metallisation. The paste used contains a large amount of aluminium particles to improve contact with the boron-diffused emitter, which makes the paste sensitive to corrosion.

There have been rapid developments in this form of paste, and although the corresponding sensitivities to corrosion oscillated between better and worse for each development, these changes were not always picked up in manufacturing companies’ standard quality controls.

“The Tier 1 companies knew what was happening; however, they did not disclose this information because it was not in their best interest to do so”

“That’s where the big risk lies,” says Hoex. “The Tier 1 companies knew what was happening; however, they did not disclose this information because it was not in their best interest to do so. They accept the sensitivity at the cell level and use a glass-glass module with high-quality encapsulants, ensuring a stable PV module.”

The problem becomes most pressing if, without any knowledge of the differences in cell sensitivity, new production lines or newcomer manufacturers start producing TOPCon modules under the assumption that it is similar to the older passivated emitter and rear cells (PERC) technology.

Academics have broadcast for years that n-type technology such as TOPCon is more stable than p-type. However, this greater stability referred to degradation regarding boron-oxygen defects and light and elevated temperature-induced degradation (LeTID), but other potential issues such as corosion-sensitive metallisation were overlooked, says Hoex.

A further drawback is that metal-related degradation is very likely to speed up when there are multimode failures. For example, if PID – corrosion caused by sodium penetrating a solar cell – is combined with humid conditions, this can cause further problems. These issues would likely not be picked up in a standard PID test of just 200 hours. However, a PID test followed by a normal damp heat test might pick this issue up.

Speed is of the essence given that innovations in PV tend to make it into end products within six months of development, meaning there is limited time to test for and pick up new defects.

In terms of solutions, last month UNSW produced a paper showing that using the paste in combination with laser-assisted firing decreases sensitivity towards sodium-related degradation as it enables the paste to be used without aluminium and still get a strong contact with boron. This technique, known as laser-enhanced contact optimisation (LECO), was developed by German firm Cell Engineering initially to improve the contacts in underfired solar cells.

However, the technique is still new and UNSW has identified some concerns that Hoex is unable to disclose at present, which suggests laser-assisted firing does not spell the end of these reliability concerns based on metallisation.

Industry awareness

Other than metallisation, certain additives in polyolefin elastomer (POE) can also be very detrimental to the cell. This is a cause for concern because if a manufacturer changes POE vendor or starts using a different paste, it may find the paste suddenly sensitive to additives in the POE. Even those opting for a conservative approach by using the most expensive encapsulate can still have these issues.

Not only is there a widespread misconception that n-type solar cells are intrinsically more stable, but some of the key areas of stability, such as boron-oxygen and LeTID are no longer a problem for p-type technology, adds Hoex.

“I’m very confident that other companies have seen this as well, but again, they have no incentive to publish this,” he says. “They will just disqualify the supplier, or that type of POE and move on to a bill of materials that gives them stable PV modules.”

Even leading Chinese companies have publicly mentioned this. Chinese manufacturer Tongwei,for example, at last year’s PV CellTech event in San Francisco organised by Solar Media,said that it took a whole year to achieve a stable TOPCon product, and warned newcomers not to expect it to be as simple as PERC since the cell is more sensitive. Likewise, Tongwei’s CTO referred to the sensitivity of TOPCon’s front metallisation.

TOPCon is “not intrinsically unstable technology”, but it runs the risk of getting a bad name, says Hoex – adding: “If you are not aware of these failure modes and do not take the right precautions, then you might end up with these failures like we demonstrated by a 65% reduction in performance after only 1,000 hours of damp heat testing. That’s a shocking result and it was comparable to what we see at cell-level testing.”

US newcomer fears

Massive multibillion-dollar Tier 1 manufacturers in China that are shipping 50GW+ a year have the means to carry out extensive in-house quality controls of their supply chain and don’t have to rely on sending modules off for IEC testing alone. Paradoxically, this means Hoex is most concerned about the proliferation of PV manufacturing in the US stemming from the Inflation Reduction Act (IRA), because companies may emerge that cannot afford extensive in-house quality control and could be caught out by changes in their supply chain.

“If those companies think that TOPCon is as robust as PERC, they will get some nasty surprises,” he says.

Front pastes, which are not yet mature, are being rapidly developed in pursuit of higher efficiencies, and these are being moved into production very quickly too.

“That’s also where our new manufacturers could be caught out,” says Hoex. “But most of the TOPCon products out there are done by Tier 1 companies, who are aware of this issue and have this high-quality control. So, I’m relatively optimistic that we don’t have a ticking time bomb.”

Nonetheless, the cells are more sensitive, and it’s not clear if IEC testing can pick up all these failure modes, so the key question remains about how these modules will fare in the field over 20-25 years.

The UNSW researchers are also investigating whether issues can arise even with a conservative bill of materials. Glass-glass modules can still have water ingress at the edges, causing degradation that spreads into the module.

“It’s comparable to PID, which was assumed to be solved in the mid-2010s,” says Hoex. “But if we look at the field of data in India, for example, which has some hot and humid climate zones, they had massive failures for PID in the late 2010s. Even though the modules were sold as PID-free, they were not PID-free, so there’s still a question mark there.”

PERC, on the other hand, is quite robust. In UNSW’s recent test on PERC solar cells’ front and rear exposed to sodium chloride, PERC was shown to be relatively stable with just 10% reduction, while TOPCon showed up to 80% reduction in efficiency.

A second-order but still significant potential failure mode for TOPCon, on which a few papers have been published, is degradation due to excess hydrogen resulting in an increase in surface recombination, typically referred to as “surface-related degradation”. The industry should keep an eye on this, says Hoex, as this failure mode takes a long time to develop and thus, is difficult to identify in accelerated testing. However, it is “nothing compared to these failures that we could see with metallisation”.

‘Mass application’ of HJT carries risk

For HJT technology, some of the challenges also lie with metallisation, but the metal pastes used for HJT are quite different compared to TOPCon. When it comes to screen-printed metallisation, curing at higher temperatures allows for far more organic matter to be burned off, and the pastes used for HJT can only be cured at 200 degrees Celsius or less, as opposed to 700 degrees or more for TOPCon.

As a result, In HJT, some of the binding materials used for the adhesion of the contact include non-silver compounds, which may be causing chemical reactions that result in the contact becoming more porous or even detached.

The soldering flux used to apply the metallisation can be “quite aggressive”, says Hoex, with components from the flux actually reacting with the heterojunction paste, which can cause heightened sensitivities and similar severe failures to those of TOPCon.

The soldering flux can even interact with the transparent conductive oxide (TCO) which is used to increase the lateral conductance of the front and rear of the HJT solar cell. Moreover, the industry is moving away from indium tin oxide (ITO) due to the scarcity of indium, despite it being known that alternative layers are typically more sensitive.

Heterojunctions are also very sensitive to contamination during the manufacturing process. Vacuum grippers are often used to handle the cells during manufacturing, yet “handler’s fingerprints” are occasionally visible after damp heat testing, indicating that the handlers left contamination on the solar cell, a problem which cannot be saved by any expensive encapsulant.

HJT cells are also sensitive to temperature and illumination and can be described as “metastable”, meaning that they have the potential to get better or worse over time; Hoex believes the full impact on the lifetime yield is not fully understood. Likewise, HJT’s sensitivity to UV degradation also raises questions about long-term stability that need to be further investigated.

There have also been multiple generations in heterojunction with HJT 1.0, 2.0, 3.0 and now 4.0 emerging, each improving on the previous generation in relatively quick succession without sufficient time to identify and mitigate potential new failure modes.  

However, unlike the rapid expansion of TOPCon, only a few manufacturers are putting large volumes of HJT products onto the market. Most are quite conservative in their bill of materials, which leads to decent levels of stability, says Hoex.

For example, Chinese firms Huasun and Risen, the largest producers of HJT, are both conservative with their bill of materials but are also able to demand a premium for their products, Hoex says.

Hoex is more concerned about the potential “mass application” of heterojunctions in the future since, in his view, the paste is even more sensitive than TOPCon paste.

Referring to both HJT and TOPCon modules, Hoex concludes: “The main thing I’m saying is stability is as important as efficiency. So, if your degradation rate increases significantly, then your high efficiency is worth nothing. You lose all the benefits. If your degradation rate goes up, then your LCOE may even be higher than a stable PERC module.”