CdTe Thin Film Panel Degradation - a Time Bomb for FSLR?

During FSLR's first quarter conference call, two quality issues related to its CdTe thin film panel surfaced.

The first one may be called "bad batch". During certain period in 2008-2009 , some modules have a manufacturing defect which causes them to stop working some time after installation in the field. FSLR has paid for the warranty claim and even loss from lost electricity production. As a result, FSLR incurred hefty charges in its quarterly report. FSLR stated the most warranty claims had been processed and the problem is largely over.

The second issue is unrelated to the first one, it is the higher degradation (panel efficiency goes done due to aging and weathering) rate at hot climate. But the higher degradation rate may not be only limited at hot places.

When the CEO Ahearn was pressed during the CC, he said he is not clear about the situation as obviously there are not many data points. However, Ahearn said it may be prudent to add a point (1%) to existing degradation rate (FSLR says 0.5% a year) because of the mix change. By mix change, he implied that more modules would be installed in hot areas in future. Later, FSLR said in another press release that it has not found any performance abnormality in regard to the degradation.

Officially FSLR stated it has NREL done test for its module's degradation, and the results are almost identical to c-Si modules, i.e., about 0.5% yearly. However, there are other sources saying the actual degradation may double or triple the official 0.5% number. More degradation can occur at hot climate. There is one source said field professionals found unusually high degradation rate for CdTe's panel though it did not provide actual data. It is not clear whether it is related to the 'bad batch' mentioned above. That's also where several analysts cautioned: it may be FSLR's Achilles' heel and future liability can be huge.

So this is a murky situation as there are no rigorous studies (large scale, long term) available, especially in hot climate. Researching panel efficiency degradation is time and money intensive no matter which route you take: field or lab. Length of NREL's accelerated degradation tests is equal to 16.5 years field. But someone may object to its methodology as accelerated may not be the same as the field. For field studies, other perils abound. You need to measure lots of parameters and every measurements can have relative errors of a few percent, like the irradiance data is quite local sensitive (and expensive to measure). And one has to isolate the panel from the pv system, which is quite difficult given the year-to year small changes. Besides, the yearly degradation is not a constant, and known to fluctuate year-by-year. Of course, the weather differs from year to year and so is the degree of weathering. There are more issues like dusts, shading, angle, degradation in various electric components etc. Therefore, the signal can easily get lost among the white noises.

It is beneficial to mention FSLR's history and some past studies on CdTe's stability/degradation to have a better perspective. First Solar was founded in 1999 after True North Partners bought Solar Cell Inc (SCI). SCI was founded by Harold McMaster in 1990 as McMaster found that solar cells can be made cheap simply by coating the glass. Being an expert on glass, he tried different thin-film coatings on glass and finally settled on CdTe. By 1997, SCI built a prototype machine to make CdTe TF solar cells. As early as 1997-1998, Colorado School of Mines (CSM) started to test the module stability based on the cell structures supplied by SCI. So far many studies showed that the degradation of the back contact (BC) structure appears to be the main culprit of cell degradation. Typically, Cu is used as extrinsic dopant in the BC and the diffusion of Cu is identified as a primary degradation mechanism. Humidity is cited as an important stress factor. In addition,  impurity diffusion and changes in doping profiles also affect cell stability. In theory, the degradation can be modeled as increased shunt resistance (Rsh) and serial resistance (Rs) in the electricity circuit of a solar cell. The take-out message from the history of FSLR's CdTe TF cell is that it is far shorter than that of the traditional c-Si cell which can trace back to the 1950s. However, CdTe cell is not as flaky as some naysayers contend and the fact that degradation occurs mostly at BC instead of the light-absorbing CdTe layer suggests the degradation is likely slow and moderate.

There are some limited-scale studies available on CdTe cell degradation. Due to the reasons cited above, we should keep an open mind to these results. A two-year field study in Mexico claims the degradation rate is about 2% in 2 years (http://www.physics.colostate.edu/groups/photovoltaic/PDFs/dec02rpt.pdf). Another 8-year field study stated degradation rate for an unknown CdTe cell is 1.3% a year (http://www.photonenergysys.com/osterwald%20wcpec.pdf). In the same report (osterwald), the degradation rate for SCI (later First Solar) CdTe cell is found to be 1.84%/yr in a 10-year test. Joe del Cueto (footnote) reported a degradation of 3.3% over a 5.5 year period. Of course, there are some reports showing higher degradation rate for some CdTe cells. However, the limited test data seem to suggest that degradation for CdTe cells is higher than c-Si cells, but not much higher. FSLR's CEO's "add 1%" comment appears to be a reasonable estimate.

If we use 11% as a starting CdTe efficiency, the 10-yr efficiency will be 9.45% using a 1.5% degradataion, and 10.46% using a 0.5% degradataion. At year 25, efficiency will be 7.54% and 9.7% respectively. Therefore, the extra 1% degradation does cause significant difference in power output over the cell's lifespan, about 10% difference at year 10 and 23% at year 25. That's why there are reports that FSLR now installs extra panels for its solar farms to account the extra degradation and to ensure future electricity output not to drop below the design number. Since there are PPA contracts signed between FSLR and utility companies, FSLR has to be conservative - it does not want the degradation issue to be the spark for future lawsuits.

Overall, concerns on FSLR's panel degradation are well-founded even if  the extra degradation is only a paltry 1%. However, it appears not to be 'catastrophic' or 'astronomical' as some alarmists claimed. FSLR has already added 1 cent per watt to the cost to deal with the warranty issue related to the degradation. The 1 cent extra, which in my view, seems to be not enough to cover the degradation issue. Unless there will be further, more comprehensive studies released, investors should keep this degradation problem in mind.

--

Joe del Cueto, "Striving for a standard protocol for preconditioning or stabilization of polycrystalline thin film photovoltaic modules," in SPIE Solar Energy + Technology Conference, San Diego, 2009.