Tuesday, May 08, 2007


How much energy is required to manufacture a photovoltaic cell?

I asked hakia and hakia replied "Wonderful question, hope I have some useful answers for you below."

The first answer was "about four years to payback." The explanation of the assumptions at A-to-Z of Materials notes that "To calculate payback, Dutch researcher Erik Alsema reviewed previous energy analyses and did not “charge” for the energy that originally went into crystalising microelectronics scrap." . Similarly, the calculation for thin-film PV systems includes the substrate, film-deposition process, and facility operation. Obtaining copper indium diselenide, cadmium telluride, or amorphous silicon is not charged.

Another article, from July 2005, claims
The primary variable cost incurred to manufacture photovoltaics is electricity, which is produced by, you guessed it, photovoltaics. Each photovoltaic cell will produce twenty times more energy in its lifetime than the amount of energy required in its manufacture, and this ratio continues to improve.

Inasmuch as the lifetime of service is estimated to be 30 years, "twenty times more energy in its lifetime" would mean a payback of 18 months, even lower than the previous estimates of three to four years which omitted easily identified (if not easily estimated) charges. So no, I don't accept "twenty times more." On the other hand, I was pleased to learn of a clever product: "sunslates", photovoltaic panels that replace roof tiles! "Most of the large-scale PV manufacturers are now making a "roof-shingle" style product, including BP Solar, GE Energy, Sharp, and Kyocera Solar," says Marianne Walpert, Vice President of Marketing and Sales at Pacific Power. The photos are of an Atlantis Energy Systems product.

After lots of non-answers, I finally got to the NREL report of which the first article was an extract. They cite more studies, and have a figure to show the payback time by component (module, frame, balance of system). Paybacks are on the order of 3 to 4 years with current technologies (3.3 years in an empirical study by Knapp and Jester, "including the energy to purify and crystallize the silicon"). The 18-month payback is for the module alone, excluding frame (energy-expensive aluminum) and other parts.

This is good news. With 12v DC I could power my adsl modem-router, charge my mobile phone, and a lamp or two (all of these are particularly useful in case of a blackout). While I fear the calculation does not include the batteries, wire, and regulator, it still seems likely to payback.

Technorati Tags: , , ,

Powered by ScribeFire.

StumbleUpon Toolbar Stumble It!
i actually adore your own writing way, very useful.
don't quit and also keep writing due to the fact that it simply just that is worth to look through it.
impatient to look at much of your current web content, enjoy your day!
This is assuming low gas/coal prices. Remember, PV cell manufacture takes energy not from PVs.
First, my apologies to Richard Fineman for taking years to approve the comment (above).
In response, this is not assuming anything about gas/coal prices: that is the point of using the energy pay-back framework. But PV cell manufacture does take energy from other sources, as he pointed out (to produce the first batch at the very least). That raises an interesting question about the possibility of bootstrapping production of solar converters. Reflective convergence (as used by BrightSource and others, e.g.) requires reflective surfaces: what energy is needed to produce that equipment?
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?