Thursday, June 12, 2008

First Solar, PrimeStar Solar and Cadmium Risks (FSLR; GE)

There has been quite a bit of misinformation flying around the blogosphere on the risk to human health from FSLR's use of cadmium in their photovoltaic cells. Cd is comparatively stable, FSLR's compound, CdTe, more so. There is some risk when it is exposed to fire but the fire has to be hot.* See note below.

The biggest risk is political. If one of the German ground arrays that use FSLR cells were to suffer a grass fire, it is probable that no cadmium would be released. It is also probable that FSLR would have a tough sell convincing the populace and politicians to allow any further deployment of CdTe. Cd is one of the six hazardous materials subject to the EU's "Restriction of Hazardous Substances Directive"

The National Renewable Energy Lab did a lot of the development work on CdTe with FSLR.
Here are a few of their publications on the subject, reading the first one will make you smarter on this subject than 99% of all bloggers (I know, not a high bar):

Cadmium Facts and Handy Comparisons
Ken Zweibel, National Renewable Energy Laboratory
Vasilis Fthenakis, Brookhaven National Laboratory

1. How much cadmium (Cd) is encapsulated in cadmium telluride (CdTe) modules?
2. Does the compound CdTe pose the same health risk as elemental Cd?
3. How much Cd is contained in nickel cadmium (NiCd) batteries, and how does this compare with the amount of Cd contained in PV modules?
4. How much Cd is used in CdTe modules per kW of power produced?...
7. Where does Cd come from?...
10. Can the CdTe in PV modules harm our health or the environment?
Toxic compounds cannot cause adverse health effects unless they enter the body in harmful doses. The only means by which people might be exposed to compounds from a finished PV module are by accidentally ingesting flakes or dust particles or inhaling dust and fumes. The thin CdTe/CdS layers are stable and solid and are encapsulated between thick layers of glass. Unless the module is ground to a fine dust, dust particles cannot be generated. The vapor pressure of CdTe at ambient conditions is zero. Therefore, it is impossible for vapors or dust to be generated when using PV modules.

See Question 11 for a discussion of fires as a potential source of exposure.

The only issue of some concern is the disposal of the well-encapsulated, relatively immobile CdTe at the end of the module’s useful life. Today's CdTe PV end-of-life or broken modules pass Federal (TCLP-RCRA) leaching criteria for non-hazardous waste. Therefore, according to current laws, such modules could be disposed of in landfills. However, recycling PV modules offers an important advantage, and the industry is considering this as it moves toward large and cost-effective production. The recycling issue is not unique to CdTe. The disposal of current x-Si modules, most of which incorporate Pb-based solder, presents similar concerns. Recycling the modules at the end of their useful life completely resolves any environmental concerns.

11. Do CdTe modules present health risks during a fire?
The flame temperature in typical U.S. residential fires is not high enough to vaporize CdTe. Flame temperatures in roof fires range from 800o-900oC, and flame temperatures in basement rooms range from 900o-1000oC.1 The melting point of CdTe is 1041oC, and evaporation starts at 1050oC*. Sublimation* occurs at lower temperatures, but the vapor pressure of CdTe at 800oC is only 2.5 torr (0.003 atm). The melting point of CdS is 1750oC, and its vapor pressure due to sublimation at 800oC is only 0.1 torr.

Preliminary studies at Brookhaven National Laboratory2 and at the GSF Institute of Chemical Ecology in Germany showed that CdTe releases are unlikely to occur during residential fires or during accidental breakage. The thin layers of CdTe and CdS are sandwiched between the glass plates, and at typical flame temperatures (800o-1000oC), these compounds would be encapsulated inside the molten glass, so any Cd vapor emissions are unlikely. In any case, the fire and other sources of emissions within the burning structure are expected to pose incomparably greater hazard than any potential Cd emissions from PV systems.

"Cadmium Use in Photovoltaics: The Perceived Risks and the Scientific Evidence"
References and Resources

General Overview

For a good general overview of the scientific research regarding cadmium, photovoltaics, and ES&H issues, please refer to the following materials:

Fthenakis, V.M; Fuhrmann, M.; Heiser, J; Wang, W. (2004). "Experimental Investigation of Emissions and Redistribution of Elements in CdTe PV Modules During Fires." Prepared for the 19th European PV Solar Energy Conference, Paris, France, June 7-11, 2004. (PDF 244 KB)

Abstract: This study is based on glass-glass CdTe PV modules, which are the only ones in the market. Pieces of commercial CdTe photovoltaic (PV) modules, sizes 3.8 cm by 30.5 cm, were heated to temperatures up to 1100°C to simulate exposure to residential and commercial building fires. It was found that a small amount of cadmium compounds (0.4 to 0.6% of the Cd content) was carried to the edges by the flow of EVA decomposition products. The pathway for this loss was through the perimeter of the sample before the two sheets of glass fused together. In actual size PV modules, the ratio of perimeter to area is 13.5 times smaller; thus the actual Cd loss during fires would be insignificant (‹0.04% of the Cd content).

Finally, since GE just announced they had achieved a majority shareholder position in PrimeStar Solar, here's PrimeStar's release from Sep. '07:

Our Responsible Use of Cadmium in PV

We use microscopic amounts of a heavy metal, cadmium, as an essential part of our modules. Is our product still “green” (environmentally attractive)?
The short answer is, yes, because our modules greatly improve the environment.
1. We remove cadmium from the atmosphere, water, and earth
a. Because coal and liquid fuel plants emit cadmium, but PV avoids the burning of coal and liquid
b. Because we recycle our cadmium in a safe, long-lived product, so that it never enters the
c. Because a coal plant emits more cadmium than we even have cadmium in our modules on a per kWh basis (and our Cd is not emitted).
d. Because cadmium is present in zinc refinery tailings as a waste, and if it is not used in ours or a
similar long-lived product, it will uncontrollably enter the environment (Ref 3).
e. Because CdTe PV modules use the least energy to make among all types of commercial PV
modules (Ref. 5), thus causing less coal to be burned and less cadmium to be emitted during
module manufacturing. Crystalline silicon PV modules use more energy to manufacture than CdTe PV, and consequently, they emit more cadmium in the atmosphere during their life cycles, than CdTe PV (Ref 5). However, these emissions are orders of magnitude lower than the emissions from fossil fuels that PV displaces.
f. CdTe PV modules have the lowest Energy Payback times among all commercial types of PV
modules (Ref. 6)
2. Our product is safe for the consumer
a. Cadmium telluride is a solid and melts at 1150 C.
b. CdTe is not as soluble as elemental cadmium and thus not as dangerous in terms of exposure to skin.
c. There is very little cadmium in our modules – 2500 times less than in NiCd batteries on a per
kilowatt hour basis (Ref. 2).
d. Our solid CdTe is isolated from the environment between two sheets of glass.
e. Fire tests show that essentially no cadmium escapes from our glass/CdTe/glass modules in a fire – the module edges close up (Ref. 4)....

*If I remember correctly (this is Boy Scout chemistry merit badge stuff, so don't take it as Gospel), a gasoline fire burns at 900C (about 1650F) to 1250C (about 2280F).

A truck carrying PV modules, colliding with a gasoline tanker could reach temperatures high enough to sublimate and even vaporize CdTe.

Which for some reason reminds me of Gary Larson's Far Side cartoon of the crash between "Bob's Assorted Rodents" and "Al's Small Flightless Birds":

The cat being a short-seller who hadn't pulled the trigger.

Again, the risk seems to be political. I can't help you quantify that.