Question:
From: "LEE KNOPER" < Subject: Re: DC power, and other random thoughts Date: Saturday, December 14, 2002 8:19 PM [Regarding the energy required to manufacture a "solar cell" versus the energy output of the cell over its useful lifetime] The lastest claims from the manufacturers indicate the energy used in producing them is recouped in 3-5 years. [...]
This discussion pops up with some regularity. However, all of the usual energy payback analyses for photovoltaic (PV) cells are based on manufacturing processes that rely on fossil fuels as the sole energy input. Although this treatment is straightforward accounting practice, the prevailing fixation with it crowds out other considerations, and tends to limit public perceptions of solar technology potential. Recognizing the uncomfortable irony of this situation, Dr Joseph Lindmayer, a research scientist at Solarex Corporation, proposed during the 1974 energy crisis the construction of "solar breeder plants" to manufacture PV cells. Except for startup and standby energy, a fully implemented solar breeder facility would be powered by some of the very PV cells it manufactured. Lindmayer’s study affirmed the feasibility of this approach – and this was back when the conversion efficiency of garden variety PV cells was lower than it is today. The Solarex proposal greatly interested several Arizona legislators at the time. Arizona has silica deposits, as well as a "solar triangle" in the southwestern part of the state. The latter has a higher solar insolation than almost anywhere else in the US (and the melanoma statistics correlate). However, shortly thereafter the energy crisis abated, the legislature moved on to more appealing crises and Lindmayer’s proposal became little more than an historical footnote. [1] A Google search for key phrase "solar breeder" just now turns up only 28 hits (none in <m.s), which illustrates the relative obscurity of this concept. (Solarex actually made some progress converting its own PV plant into a solar breeder facility, but the eventual implementation was not 100%.) [2] Solar breeders have little appeal in the current (NPI) economic and political climate. Mentioning them in <m.s is topical mainly in the context of a post-collapse reconstruction. The collapse would have to be harsh enough to inhibit some of the more controversial influences of Big Government and Big Business during the reboot. In this case the post-collapse economic and political environment would be amenable to developing this kind of energy source. Discussion of solar breeders outside of this narrow scope probably would be more topical in a newsgroup like <alt.energy.renewable. (YMMV, of course.) Lee_K [1] "Solar Breeders for America," testimony of Dr Joseph Lindmayer, Solarex Corp (Rockville MD) at the US Senate Energy Committee hearings on S 2806 (The Energy Trust Fund), Washington DC, 28 Jan 1974. Also covered briefly in: "Solar Energy’s Big Push Into the Marketplace," by John Fialka, Washington Star-News, 17 Jun 1974. Congressional Record, Vol 120 No 89 (19 Jun 1974), p S.10964-6; includes the Star-News article cited above, as inserted into the Record by Sen Mike Gravel (D-AK). (Regrettably, my cites are limited to selected clippings from various sources.) [2] Renewable Energy Workshop, at URL <http://www.inesglobal.com/publication/ines_proceedings/WORKSHOP_1HTM/… R.HTM. —
Response:
I got into a long debate in this very news group on this topic several months ago. I had opined that the energy payback was much greater than solar producers would like to admit. Someone posted a report that one of the solar manufacturers published on the subject of energy payback. I have to admit, that even with my sometimes jaded few of this energy field (based on my years of working in the industry), 3-5 years is a reasonable payback period. While the report that was posted had many gaps, the heavy energy hitters were all covered. It left me reasonably convinced. Naturally, energy payback depends on the process used to great the substrate. IMHO, ASE America probably has one of the most energy efficient processes. The ASE nonagon EFG process has a very high silicon growth rate, and it doesn’t loose lots of silicon while cutting into cells. I’d bet this energy payback period is closer to 3 years. At the other end of spectrum are the ribbon based technologies like Evergreen Solar’s string ribbon, or Ebara Solar’s dendritic ribbon. These technologies are very inefficient in their silicon production rate. By this I mean their silicon furnaces must run for long periods in order to produce relatively little silicon. Both Evergreen and Ebara have not fared well commercially. NASDAQ recently threatened to drop Evergreen from its exchange if its share value didn’t increase. Ebara Solar recently lost its financial backing and its employees have dropped from 100 to about 20.
From: "LEE KNOPER" < Subject: Re: DC power, and other random thoughts Date: Saturday, December 14, 2002 8:19 PM [Regarding the energy required to manufacture a "solar cell" versus the energy output of the cell over its useful lifetime] The lastest claims from the manufacturers indicate the energy used in producing them is recouped in 3-5 years. [...] This discussion pops up with some regularity. However, all of the usual energy payback analyses for photovoltaic (PV) cells are based on manufacturing processes that rely on fossil fuels as the sole energy input. Although this treatment is straightforward accounting practice, the prevailing fixation with it crowds out other considerations, and tends to limit public perceptions of solar technology potential. Recognizing the uncomfortable irony of this situation, Dr Joseph Lindmayer, a research scientist at Solarex Corporation, proposed during the 1974 energy crisis the construction of "solar breeder plants" to manufacture PV cells. Except for startup and standby energy, a fully implemented solar breeder facility would be powered by some of the very PV cells it manufactured. Lindmayer’s study affirmed the feasibility of this approach – and this was back when the conversion efficiency of garden variety PV cells was lower than it is today. The Solarex proposal greatly interested several Arizona legislators at the time. Arizona has silica deposits, as well as a "solar triangle" in the southwestern part of the state. The latter has a higher solar insolation than almost anywhere else in the US (and the melanoma statistics correlate). However, shortly thereafter the energy crisis abated, the legislature moved on to more appealing crises and Lindmayer’s proposal became little more than an historical footnote. [1] A Google search for key phrase "solar breeder" just now turns up only 28 hits (none in <m.s), which illustrates the relative obscurity of this concept. (Solarex actually made some progress converting its own PV plant into a solar breeder facility, but the eventual implementation was not 100%.) [2] Solar breeders have little appeal in the current (NPI) economic and political climate. Mentioning them in <m.s is topical mainly in the context of a post-collapse reconstruction. The collapse would have to be harsh enough to inhibit some of the more controversial influences of Big Government and Big Business during the reboot. In this case the post-collapse economic and political environment would be amenable to developing this kind of energy source. Discussion of solar breeders outside of this narrow scope probably would be more topical in a newsgroup like <alt.energy.renewable. (YMMV, of course.) Lee_K [1] "Solar Breeders for America," testimony of Dr Joseph Lindmayer, Solarex Corp (Rockville MD) at the US Senate Energy Committee hearings on S 2806 (The Energy Trust Fund), Washington DC, 28 Jan 1974. Also covered briefly in: "Solar Energy’s Big Push Into the Marketplace," by John Fialka, Washington Star-News, 17 Jun 1974. Congressional Record, Vol 120 No 89 (19 Jun 1974), p S.10964-6; includes the Star-News article cited above, as inserted into the Record by Sen Mike Gravel (D-AK). (Regrettably, my cites are limited to selected clippings from various sources.) [2] Renewable Energy Workshop, at URL
<http://www.inesglobal.com/publication/ines_proceedings/WORKSHOP_1HTM/… R.HTM. —
Response:
Yes, according to the world’s most dedicated anti-PV internet crusader, Don Lancaster (Father of One Million Floating PCs). Here is Don’s documented answer to the question: Curious World: Don, what is the ratio of the lifetime energy output from a photovoltaic module to the energy consumed in it’s manufacture? Father Don: About five. Curious World: How can we factor in the energy consumed by the manufacture of the balance of PV systems (wire, inverter, etc)? Father Don: Energy equals money. Curious World: So if the cost of a photovoltaic module is about 40% of total system cost, that means a total PV system returns about twice the energy consumed in its manufacture? Father Don: <prolonged silence
Response:
Yes, according to the world’s most dedicated anti-PV internet crusader, Don Lancaster (Father of One Million Floating PCs). Here is Don’s documented answer to the question: Curious World: Don, what is the ratio of the lifetime energy output from a photovoltaic module to the energy consumed in it’s manufacture? Father Don: About five. Curious World: How can we factor in the energy consumed by the manufacture of the balance of PV systems (wire, inverter, etc)? Father Don: Energy equals money. Curious World: So if the cost of a photovoltaic module is about 40% of total system cost, that means a total PV system returns about twice the energy consumed in its manufacture? Father Don: <prolonged silence
He’s probably driving around hoping to catch his first glimpse of a home power system. If we’d all leave our doors open, he might finally get a fleeting look. Expect numerous and lengthy dire pronouncements immediately after. Wayne
Response:
From: "LEE KNOPER" < Subject: Re: DC power, and other random thoughts Date: Saturday, December 14, 2002 8:19 PM [Regarding the energy required to manufacture a "solar cell" versus the energy output of the cell over its useful lifetime] The lastest claims from the manufacturers indicate the energy used in producing them is recouped in 3-5 years. [...] This discussion pops up with some regularity. However, all of the usual energy payback analyses for photovoltaic (PV) cells are based on manufacturing processes that rely on fossil fuels as the sole energy input. Although this treatment is straightforward accounting practice, the prevailing fixation with it crowds out other considerations, and tends to limit public perceptions of solar technology potential.
<<<much irrelevent stuff clipped Does it really matter where the energy came (comes) from? Energy still has to be used to create the PV cells. At some point you get some net energy yield from them regardless of how they are made. The distinction between PV generated energy and energy generated by "fossil" fuels is sort of like the distinction made by the federal government regarding the social security "lockbox". Its a distinction without much difference.
Response:
no, it does not matter from an accounting standpoint where the energy comes from, only the amount of it. Obviously, we would like to see the energy come from renewable sources, but that will happen with time. — Steve Spence Subscribe to the Renewable Energy Newsletter & Discussion Boards. Read about Sustainable Technology: http://www.green-trust.org
From: "LEE KNOPER" < Subject: Re: DC power, and other random thoughts Date: Saturday, December 14, 2002 8:19 PM [Regarding the energy required to manufacture a "solar cell" versus the energy output of the cell over its useful lifetime] The lastest claims from the manufacturers indicate the energy used in producing them is recouped in 3-5 years. [...] This discussion pops up with some regularity. However, all of the usual energy payback analyses for photovoltaic (PV) cells are based on manufacturing processes that rely on fossil fuels as the sole energy input. Although this treatment is straightforward accounting practice, the prevailing fixation with it crowds out other considerations, and tends to limit public perceptions of solar technology potential. <<<much irrelevent stuff clipped Does it really matter where the energy came (comes) from? Energy still has to be used to create the PV cells. At some point you get some net energy yield from them regardless of how they are made. The distinction between PV generated energy and energy generated by "fossil" fuels is sort of like the distinction made by the federal government regarding the social security "lockbox". Its a distinction without much difference.
Response:
Does it really matter where the energy came (comes) from?
In terms of overally energy production or use, no. However, an entire PV production chain that is powered entirely by PV could be a compelling demonstration. It would be more symbolic than anything else, but that doesn’t mean it doesn’t have value on that basis.
Response:
Interesting thought, but keep in mind a silicon crystal growth furnace consumes about 10kW. A small facility would run about 10 of those. That’s a lot of panels!
Does it really matter where the energy came (comes) from? In terms of overally energy production or use, no. However, an entire PV production chain that is powered entirely by PV could be a compelling demonstration. It would be more symbolic than anything else, but that doesn’t mean it doesn’t have value on that basis.
Response:
R+D and start up usually means that alot of panels are not used or sold… these can be used and I am sure that there could be taxation advantages too – depreciation and investment bollocks… Interesting thought, but keep in mind a silicon crystal growth furnace consumes about 10kW. A small facility would run about 10 of those. That’s a lot of panels! Does it really matter where the energy came (comes) from? In terms of overally energy production or use, no. However, an entire PV production chain that is powered entirely by PV could be a compelling demonstration. It would be more symbolic than anything else, but that doesn’t mean it doesn’t have value on that basis.
Response:
Even if its not all solar powered, what ever is will offset some cost.
R+D and start up usually means that alot of panels are not used or sold… these can be used and I am sure that there could be taxation advantages too – depreciation and investment bollocks… Interesting thought, but keep in mind a silicon crystal growth furnace consumes about 10kW. A small facility would run about 10 of those. That’s a lot of panels! Does it really matter where the energy came (comes) from? In terms of overally energy production or use, no. However, an entire PV production chain that is powered entirely by PV could be a compelling demonstration. It would be more symbolic than anything else, but that doesn’t mean it doesn’t have value on that basis.
Response:
Even if its not all solar powered, what ever is will offset some cost.
Actually, it won’t. If the cost of producing a solar panel is greater than the cost of producing an equivilent amount of power by some other means, then every solar panel you use in the production process drives the cost *UP*. Likewise, in a closed system, If it takes more than one panel to produce the amount of power used to make one, then every panel you divert to driving your production process sucks that much more energy OUT of the system…
Response:
Interesting thought, but keep in mind a silicon crystal growth furnace consumes about 10kW. A small facility would run about 10 of those. That’s a lot of panels!
Quite so. -100 square meter of sunlight is around 100 kW (peak). -If your cells are 15% efficient you need around 660 square meters for 100kW (peak). (500 square meters for 20% efficiency.) -If you want to run the furnaces 24/7 you need between 5 & 6 times as much area of panels. -(5*500 square meters = 2,500 square meters.) -(6*660 square meters = 3960 square meters.) As you say, that’s just for the silicon furnaces. Lights & the rest of the manufacturing process for the solar cells costs extra (as do the energy costs of inverters, regulators, storage cells, etc.) Karl Johanson
Response:
Even if its not all solar powered, what ever is will offset some cost. Actually, it won’t. If the cost of producing a solar panel is greater than the cost of producing an equivilent amount of power by some other means, then every solar panel you use in the production process drives the cost *UP*. Likewise, in a closed system, If it takes more than one panel to produce the amount of power used to make one, then every panel you divert to driving your production process sucks that much more energy OUT of the system…
Unfortunately your ideas fall down, unless you have some magical insight into the cost of living, of which fuel/energy costs are extremely high, for the next twenty to thirty years – something I am yet to see… no one yet has a crystal ball of that magnitude… eg: who would have thought that at the end of WWI we would be standing on the edge of the next World War, or around 30 years after Vietnam the USSR would fall…
Response:
Unisolar does not need the furnaces to make silicon wafers. http://www.uni-solar.com/PV%20Manufacturing.html
Interesting thought, but keep in mind a silicon crystal growth furnace consumes about 10kW. A small facility would run about 10 of those. That’s a lot of panels! Quite so. -100 square meter of sunlight is around 100 kW (peak). -If your cells are 15% efficient you need around 660 square meters for 100kW (peak). (500 square meters for 20% efficiency.) -If you want to run the furnaces 24/7 you need between 5 & 6 times as much area of panels. -(5*500 square meters = 2,500 square meters.) -(6*660 square meters = 3960 square meters.) As you say, that’s just for the silicon furnaces. Lights & the rest of the manufacturing process for the solar cells costs extra (as do the energy costs of inverters, regulators, storage cells, etc.) Karl Johanson
Response:
Since it doesn’t need silicon growing furnaces, do you know if Uni-solar is able to generate its own power?
Unisolar does not need the furnaces to make silicon wafers. http://www.uni-solar.com/PV%20Manufacturing.html Interesting thought, but keep in mind a silicon crystal growth furnace consumes about 10kW. A small facility would run about 10 of those. That’s a lot of panels! Quite so. -100 square meter of sunlight is around 100 kW (peak). -If your cells are 15% efficient you need around 660 square meters for 100kW (peak). (500 square meters for 20% efficiency.) -If you want to run the furnaces 24/7 you need between 5 & 6 times as much area of panels. -(5*500 square meters = 2,500 square meters.) -(6*660 square meters = 3960 square meters.) As you say, that’s just for the silicon furnaces. Lights & the rest of the manufacturing process for the solar cells costs extra (as do the energy costs of inverters, regulators, storage cells, etc.) Karl Johanson
Response:
Interesting thought, but keep in mind a silicon crystal growth furnace consumes about 10kW. A small facility would run about 10 of those. That’s a lot of panels! Quite so. -100 square meter of sunlight is around 100 kW (peak). -If your cells are 15% efficient you need around 660 square meters for 100kW (peak). (500 square meters for 20% efficiency.) -If you want to run the furnaces 24/7 you need between 5 & 6 times as much area of panels.
There is no requirement for 24 hour operation. By inserting that requirement, you implicitly insert the need for huge energy storage capability. The objective is to demonstrate that a closed energy PV system can exhibit growth, not maximize factory utilization. Photovoltaics are great for producing electricity for use in machinery and lighting, but I think they are exceedingly inappropriate if the primary use of the energy is heat energy. Solar concentrator furnaces would be more appropriate for such duties, although that too would violate the purpose of the PV demonstration system.
Response:
Danny Nicole Smith, who I am pretty sure is Bubba No-Name, asked: Since it doesn’t need silicon growing furnaces, do you know if Uni-solar is able to generate its own power?
Not only is a-Si less PV heat intensive in its manufacture, it uses much less silicon, less than a micron thickness (vs 450 microns for poly/mono?) Most of the references I’ve seen credit a-Si PV with much shorter energy payback periods than crystalline and semi-crystalline. Some of these estimates have error bars. For example, von Meier in ’94 shows a range of 6 months to 2.3 years for a-Si PV, vs 3.5 to 12 years for monocyrstalline silicon PV, and 1.7 to 6.4 years for polycrystalline silicon PV.
Response:
Unisolar does not need the furnaces to make silicon wafers.
from http://www.uni-solar.com/Our_Technology_a_Si.html "Using our proprietary thin-film, vapor-deposited amorphous silicon (a-Si) alloy materials, we have developed proprietary technology to reduce the materials cost in a solar cell" Anyone have an idea of energy required to take silicon all the way from solid to vapor instead of just melting it? -mod.
Response:
How’s business?
Unisolar does not need the furnaces to make silicon wafers. from http://www.uni-solar.com/Our_Technology_a_Si.html "Using our proprietary thin-film, vapor-deposited amorphous silicon (a-Si) alloy materials, we have developed proprietary technology to reduce the materials cost in a solar cell" Anyone have an idea of energy required to take silicon all the way from solid to vapor instead of just melting it? -mod.
Response:
… There is no requirement for 24 hour operation.
For a PV system building the more efficient monocrystaline solar cells a furnace is needed. The furnace has to be kept very hot for several days to weeks as the crystal is pulled. I think most of the diffusion furnaces that are usually used to dope the cells also need a relatively long time at a stable temperature. Very little of this equipment can be turned on and off in an hour. The vendor growing poly crystalline cells from silane gas on glass, plastic, or metal substrates, may not have to stabilize the temperature so long, but as a general issue, if you cannot operate for more than 6 hours at a time, the range of options for cell manufacture is rather strongly limited.
Response:
The silicon is not vaporized in most of the systems I am familiar with. It is grown from a Silane gas, SiH4. Sounds rather unstable to me. I think I will leave this process in the hands of experts.
Unisolar does not need the furnaces to make silicon wafers. from http://www.uni-solar.com/Our_Technology_a_Si.html "Using our proprietary thin-film, vapor-deposited amorphous silicon (a-Si) alloy materials, we have developed proprietary technology to reduce the materials cost in a solar cell" Anyone have an idea of energy required to take silicon all the way from solid to vapor instead of just melting it? -mod.
Response:
… There is no requirement for 24 hour operation. For a PV system building the more efficient monocrystaline solar cells a furnace is needed. The furnace has to be kept very hot for several days to weeks as the crystal is pulled.
True for standard sized monocrystaline ingots using CZ and even FZ manufacture. I think most of the diffusion furnaces that are usually used to dope the cells also need a relatively long time at a stable temperature. Very little of this equipment can be turned on and off in an hour. The vendor growing poly crystalline cells from silane gas on glass, plastic, or metal substrates, may not have to stabilize the temperature so long, but as a general issue, if you cannot operate for more than 6 hours at a time, the range of options for cell manufacture is rather strongly limited.
For the purposes of demonstrating a closed energy cycle, the "strongly limited" makes it challenging, not impossible. I believe amorphous silicon cells could be manufactured without the need to operate at night. And strictly speaking, dye-based PVs should be considered suitable for such a demonstration.
Response:
Interesting thought, but keep in mind a silicon crystal growth furnace consumes about 10kW. A small facility would run about 10 of those. That’s a lot of panels! Quite so. -100 square meter of sunlight is around 100 kW (peak). -If your cells are 15% efficient you need around 660 square meters for 100kW (peak). (500 square meters for 20% efficiency.) -If you want to run the furnaces 24/7 you need between 5 & 6 times as much area of panels. There is no requirement for 24 hour operation. By inserting that requirement, you implicitly insert the need for huge energy storage capability. The objective is to demonstrate that a closed energy PV system can exhibit growth, not maximize factory utilization.
I know very little about silicon growth furnaces. I figured running them by heating them up, reaching peak heat around noon, then them cooling down, wouldn’t be good for them. Maybe they can run on intermittent power though. Many industrial processes can. I also didn’t want anyone to confuse peak power with continuous equivalent. A 100 kilowatt grid connection could make somewhere around 5 to 6* times as much silicon as a 100 kilowatt photovoltaic system. (*perhaps a bit less than that, as running the solar cells intermittently wouldn’t imply charging losses in the storage system.) Photovoltaics are great for producing electricity for use in machinery and lighting, but I think they are exceedingly inappropriate if the primary use of the energy is heat energy. Solar concentrator furnaces would be more appropriate for such duties, although that too would violate the purpose of the PV demonstration system.
Well said. Karl Johanson
Response:
It’s also true for silicon casting furnaces, EFG, and ribbon furnaces.
… There is no requirement for 24 hour operation. For a PV system building the more efficient monocrystaline solar cells a furnace is needed. The furnace has to be kept very hot for several days to weeks as the crystal is pulled. True for standard sized monocrystaline ingots using CZ and even FZ manufacture. I think most of the diffusion furnaces that are usually used to dope the cells also need a relatively long time at a stable temperature. Very little of this equipment can be turned on and off in an hour. The vendor growing poly crystalline cells from silane gas on glass, plastic, or metal substrates, may not have to stabilize the temperature so long, but as a general issue, if you cannot operate for more than 6 hours at a time, the range of options for cell manufacture is rather strongly limited. For the purposes of demonstrating a closed energy cycle, the "strongly limited" makes it challenging, not impossible. I believe amorphous silicon cells could be manufactured without the need to operate at night. And strictly speaking, dye-based PVs should be considered suitable for such a demonstration.
Response:
… For the purposes of demonstrating a closed energy cycle, the "strongly limited" makes it challenging, not impossible. I believe amorphous silicon cells could be manufactured without the need to operate at night. And strictly speaking, dye-based PVs should be considered suitable for such a demonstration.
Actually the plastic semiconductors might make an interesting option for this type of evaluation. Most of our historic examples were not interested in energy input, just energy output per square.