Whether or not a PV module uses more energy to manufacture than it will
normally produce is an interesting and important question that I wanted
to know the answer to. So I posed it to the PV Users discussion list
along with part of Paul Hetrick's post on the subject.
I got a number of replies, but most concentrated on whether the dollar
value of the energy generated by the PV module would equal or surpass
its retail cost. To no one's surprise, the answer is usually no, unless
one is in a situation that would require an extension of the grid. In
that case, PV might very well be cheaper, often much cheaper.
So far, only one response has focused on the energy payback, as opposed
to the financial. For those interested, that post is attached.
Doug Fields
Subject: RE: embodied energy in PV modules
Date: Sun, 19 Dec 1999 22:37:45 -0500
From: "Alan Ristow" <[EMAIL PROTECTED]>
To: "ForestHaven" <[EMAIL PROTECTED]>, <[EMAIL PROTECTED]>
> We are currently having a discussion on another list about whether or
> not a PV module uses more energy in its manufacture than it produces in
> the course of its useful life. Does anyone have the answer to this?
The short answer is that PV modules have an energy payback period
(defined
as the amount of time required to generate an amount of electricity
equivalent to the energy used in their manufacture, and NOT to be
confused
with the financial payback period) of anywhere from about 8 months to 10
years. Even 10 years is less than the typical service lifetime of a PV
module.
The wide swing from minimum to maximum period is due to variations in
manufacturing technologies, cell efficiencies, and geographical
considerations. In general, amorphous silicon cells have the shortest
energy
payback period, and single crystal silicon cells the longest. As I
recall,
silicon production is dominant energy consumer in PV module manufacture
(it
is certainly the dominant cost factor in crystalline silicon
technologies).
Several excellent papers on this topic have been published in journals
and
conference proceedings. If anyone is interested, I can post a
bibliography
of the ones I have read, though I probably won't have time to do so
until
after the holidays.
> "But how fairly are they in tabulating the real amount of energy it
> takes to build them?
Most of the studies I have read are life cycle analyses, i.e., they take
into account all energy inputs from all phases of manufacturing. They
reach
the conclusions stated above.
> Solarex's buildings in Frederic MD, have huge banks of cells, but
> even if that plant is off the grid, it is not where most of the
> energy for manufacturing is used. Smelting Si and growing the
> crystals takes a lot.
Many PV companies, including Solarex, make their own silicon wafers.
They
don't purify the silicon themselves, but they cast the ingots and cut
the
wafers themselves. Thus, a significant portion of the energy used in
producing the wafers actually occurs in the PV manufacturer's facility.
On the flip side, since PV manufacturers often use reject silicon from
the
IC industry, their cells probably have more embodied energy than they
would
if they just bought "fresh" silicon for their wafer production
operations.
This will have to change soon, though -- rejects from the IC industry
are
getting to be in short supply, and if the PV industry doesn't develop
its
own independent sources of silicon soon there won't be a large enough
supply
to sustain further growth.
> Adding the final layers to a waffer and wiring it
> up does not take much energy, and I'm willing to bet that it's the
> only step they measure in the 2-3 year pay back figure.
Not true, as I mentioned above.
Happy holidays!
___________________________________
Alan Ristow
University Center of Excellence for Photovoltaics
Georgia Institute of Technology
Atlanta, Georgia
