[wordup] Response from Alan Zelicoff on "Payback Time for PV Panel"

[Adam Shand]

I wrote to Alan, the author of "Saving Energy without Derision" with 
two questions about his book, specifically about the amount energy 
required to manufacture a PV panel and if that energy would ever be 
recovered during the average life of a PV panel.  After getting his 
response and doing some more informed Google'ing I found a wealth of 
information and discussion on this very topic.  Perhaps the best 
resource I found was this URL:

http://www.autobahn.mb.ca/~het/energy/pv_fqa.html#Section12

It was also pointed about by several of you in response to the previous 
posting, that as industrial scale green/renewable energy sources become 
available the any energy debt from the creation of a PV panel becomes 
less of an issue.

For those that are interested here are his responses.

Adam.

>> On Dec 22, 2004, at 5:39 PM, Adam Shand wrote:
>> Hi,
>>
>> I recently found your free book and have spent the last hour or so 
>> browsing through it.  Thanks for a great resource, I hope to put some 
>> of your advice to the test when I get back home to New Zealand in a 
>> few weeks.
>>
>> I have a couple of questions for you, I hope I haven't missed the 
>> answers by only skimming your book.
>>
>> The first is just minor curiosity, how do you find the light emitted 
>> from the modern fluorescent lights?  My only experience with 
>> fluorescent lighting is horrible memories from school and work places 
>> where you got horrible flickers and an incomplete spectrum of light.  
>> I assume the modern compact fluorescent lights are much better?
>>
>> Second, one of the critiques of solar power that I've never been able 
>> to prove or disprove is that the amount of energy required to 
>> construct a solar panel is more then the energy that a solar panel 
>> will generate over an average lifetime.  Thus they can be a very 
>> efficient way of "transporting" energy but will always run at a net 
>> loss.
>>
>> Do you know, or have any idea where to look, to prove this one way or 
>> another?
>>
>> Thanks again for your book.
>>
>> Adam.
>
> On 26 December 2004, at 7:08PM, Alan Zelicoff wrote:
>
>
> Hi Adam:
>
> Let me tackle the second (and harder) question first: what is the 
> energy "payback time" of a PV panel?
>
> It is difficult to do completely independent research on this question 
> because of one practical "complexity" in computing energy consumption 
> in making solar cells: the basic raw material (rather highly purified 
> silicon) is actually the scrap (waste) from the silicon chip industry 
> (which makes VERY purified silicon).  The energy investment in VERY 
> purified silicon is probably proprietary, and in any case is more than 
> what would be required to make the rather (but not "very") purified 
> silicon necessary for PV cells.
>
> So, I turned to my colleagues at Sandia National Laboratories (where I 
> used to work) in their Photovoltaic (PV) department.  (Edited by Adam 
> <snip>) ... the PV department the engineers are well respected (and I 
> know many of them personally and have found them reliable on just 
> about every issue that I've raised with them).  Here's what they say 
> (I took this from a website):
>
> 'There has been a bit of discussion in this group on the energy 
> payback time
> for solar cells. The following paragraph was taken from:
>
> P.A. Basore and J.M. Gee, "Crystalline-Silicon Photovoltaics: 
> Necessary and Sufficient", Proc. 1st World Conference on Photovoltaic 
> Energy Conversion, Hawaii, 2254-2257 (1994)
> The authors were from Sandia National Labs, a US defence organisation.
>
> Energy Payback
>
> Although manufacturing cost is the ultimate test of feasibility for
> large-scale deployment of photovoltaics, it is also important that a PV
> technology not require more energy than it produces during its life. 
> The
> amount of energy required to convert readily available raw materials 
> into a
> turnkey flat-plate photovoltaic system based on crystalline silicon is
> approximately 400 kWh/m2. Of this total energy input, only 40% is in 
> the form
> of electricity consumed at the PV module manufacturing plant. The rest 
> is
> energy required to fabricate the balance-of-system components, 
> construct the
> factory, and to refine the aluminum, glass, and chemicals used in
> manufacturing the system. This total energy input is returned by the 
> system in
> just 24 months of outdoor operation. A PV system with a service life 
> of 20
> years returns ten times the energy used to manufacture it.
> The energy required to purify polysilicon feedstock from quartz sand 
> (200
> kWh/kg) is not included in our estimate, because polysilicon material 
> is
> available as a reject byproduct of the IC industry, and thus qualifies 
> as a
> raw material for photovoltaic applications. If, in the future, the 
> supply of
> reject IC material becomes inadequate to meet PV demand, then the 
> energy
> required to refine the polysilicon should be considered. By the time 
> this
> might happen, the amount of silicon required per wafer will be less 
> than half
> what it is today, so that silicon purification would add at most 18 
> additional
> months to the energy payback time for the system.
>
> Cheers,
> Stuart Bowden
> Centre for Photovoltaic Devices and Systems
> University of New South Wales'
>
> and, consistent with this assessment is a separate source:
>
> "New Study Shows Siemens Solar Panels Energy Payback Time Is One to 
> Three Years
>
> CAMARILLO, Calif., Sept. 6 -/E-Wire/--  Siemens Solar,
> one of the world's leading manufacturers of solar panels,
> presented research findings on the energy payback time for
> photovoltaic modules.
>
> "The researchers calculated the energy consumed in the manufacture of
> photovoltaic modules currently in production at Siemens Solar. 
> Calculations
> included process energy, used in cell and module manufacturing as well 
> as
> the energy used in producing both direct and indirect raw materials.
> Historical and directly measured data were employed in deriving 
> process and
> embodied energy. Sources included utility bills, monthly production 
> data,
> measured energy consumption, and detailed bills of materials. The data 
> was
> used to measure the amount of energy required to make photovoltaic 
> (solar
> electric) panels, i.e. the "energy payback time."
> Energy payback time depends on both the energy content and the 
> installation
> details. The estimated break-even point is approximately two to three 
> years,
> which means that over its lifetime, a Siemens Solar panel generates 
> nine to
> seventeen times as much energy as is needed to create it. "
>
> So, it appears that the bottom line is that "recovery time" on 
> electricity (or even total energy) investment on PV panels is a couple 
> of years.  And, do note that one could have the same concern about 
> coal burning plants (huge amounts of steel go into making the boilers 
> in those), natural gas burning plants (ditto for the turbines) and 
> even nuclear plants (although in this case we're talking about CO2 
> release from the enormous amounts of "curing" concrete to make 
> containment vessels).
>
> Thus, electricity production plants of any kind (including PV, but not 
> limited to it) consume energy (some of which is electricity of course) 
> during the production process.  But, I have little doubt that that 
> investment is recovered rather quickly (and occupying perhaps 10 - 15% 
> of the lifetime of the PV panels).
>
> Now, on to the question of fluorescent light quality: the first 
> compact florescent bulbs were eery in their light spectrum (determined 
> mostly by the "temperature" of the gas in the bulb and whatever 
> absorbing materials are put on the glass tube.  But, over the past few 
> years, they have improved dramatically.  Most people notice a little 
> difference from the (very inefficient) standard light bulb, but they 
> get used to it quickly.  Thus, I think that what physicists call the 
> "color temperature" problem has been largely eliminated.  The problem 
> that remains is that it takes about 30 - 40 seconds for a compact 
> fluorescent to get up to full temperature (it takes only a fraction of 
> a second for a filament tungsten bulb to do so), and some folks are 
> very impatient.  Flickering is no longer a problem at all.
>
> Bottom line: if you can wait 30 - 40 seconds of less than optimal 
> lighting, compact fluorescent lights are a "no brainer" in those 
> lights that you use most frequently (like reading lights).  They don't 
> yet come in 3-way versions (well, they do, but they are expensive); 
> the standard compact fluorescent bulb can be had for around $3 USD in 
> most hardware and home supply stores when you buy them in packs of 3 
> or 4.  Recovery time for the investment of $3 (instead of 50 cents or 
> so for a standard bulb) depends on how many hours a day you need them 
> and your local cost of electricity, but it is on the order of a few 
> months.
>
> I enjoyed your questions.  May I use them (and my answers) in the next 
> version of "Saving Energy without Derision"?  I'll credit you for 
> asking them (they are quite thoughtful and I'm sure other readers 
> would learn from them) if you'd like.
>
> Best,
> ALZ .