Talk:Solar cell/Archive 4

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5

Archive 6

New Solar Cell Efficiency Record (July 2007)

DARPA's VHESC consortium led by Delaware has achieved 42.8%: | UD-led team sets solar cell record, joins DuPont on $100 million project. I don't think it's been independently verified yet, though.

Perhaps a table of efficiency records by date would be an interesting addition? — RJH (talk) 19:34, 1 August 2007 (UTC)

Efficiency	Date	Group
42.8%	July 2007	High Performance Solar Power Program
40.7%	December 2006	Boeing's Spectrolab, Inc.
32.3%	November 1999	National Renewable Energy Laboratory[1]
16.6%	April 1997	AstroPower, Inc.[2]
14.6%	1995	AstroPower, Inc.

There's already a chart published by NREL showing the record laboratory cell efficiencies on the main page. Obviously it doesn't include the very latest, but in time NREL will likely incorporate it. As for your table, the records attributed to Astropower are only records for the specific type of multicrystalline silicon they were using -- even in the mid- to late-'90s, there were commercially produced solar cells getting better efficiencies than those. And that is the crux of the matter when it comes to tabulating record efficiencies: What constitutes a "record"? If you comb the academic literature, you'll find people attaching long chains of qualifiers to their project so they can call it a record (e.g., "record high-efficiency for planar screen-printed solar cells using material X and processed using technique Y"). If such a table is to be included, I would suggest using the categories and figures published a few times a year in the academic journal Progress in Photovoltaics. Of course, that would require somebody with access to the journal to compile and maintain the table (I have access, but no time at the moment), and even though it keeps the number of categories to a minimum you'll still be talking about a lot of page area covered with tables. Considering the questionable value of laboratory efficiency records to anybody but scientists and academics, I think the existing NREL chart does the job just fine.--Squirmymcphee 14:37, 3 August 2007 (UTC)

Link Removal

The link that I removed was WP:LINKSPAM as the sole point of the linked to page was to sell a collection of journal articles. Admittedly, a tag of vandalism was a little harsh, but it looked like the link was added by a single purpose account, so ... Burzmali 13:32, 12 September 2007 (UTC)

Beware of Greek Scientists?

Hi,

I recently removed a section added by User:Prester John. I will attempt to explain my reasoning. The article that was referenced stated "[Solar panels] may be harming the environment more than convenional sources of energy, according to a study by scientists." The article only states that the scientists are based in Greece, not whom they are, where they published the work, nor any details therein. I removed this to stay in line with WP:VER; As in a scientific context, WP:VER recommends the use of scientific journals when making claims such as this one. (See : WP:VER#Science article in the popular press)

Regards User A1 01:16, 6 November 2007 (UTC)

Silicon energy to weight or cost ratio

Q: What is the total amount of Si used making solar panels in kg/kW? I would expect the numbers to be different for monocrystalline silicon, polycrystalline silicon, ribbon silicon, and other methods. I know some of the Si is used in the cutting process and is recycled, how much? The reason for asking this is to read between the lines when a news article states the number of metric tons of Si they are going to be making with a new plant. —Preceding unsigned comment added by (talk) 14:59, 26 November 2007 (UTC)

A: A wafer based Si cell will be about 1mm-2mm thick. A square meter of 1mm thick Si based solar cells will therefore contain about 23 kg of Si. However when using the most efficient monocrytalline wafers there will also be substantial losses when cutting square cells from the circular wafers. hAl (talk) 16:03, 26 November 2007 (UTC)

Whoa! Nobody uses 1-2-mm-thick silicon wafers for anything, not even in the IC industry! A wafer used for a microchip will tend to be in the range of 500-600 µm, and one used for a solar cell is about 250 µm (though that's changing fast -- just a few years ago it was 300-350 µm). On the whole, the PV industry uses about 10 g/Wp after accounting for waste. The highest-efficiency silicon cells and ribbon cells use around 7 g/Wp, while the worst cells use around 12 g/Wp.--Squirmymcphee (talk) 23:02, 26 November 2007 (UTC)

The 1-2mm thickness I actually took from the article itself. hAl (talk) 15:52, 27 November 2007 (UTC)

The only place I can find that thickness in the article is in the section on Sliver cells, which are made in a completely different manner than regular solar cells (and are still not in commercial production, at least not in any significant volume). Elsewhere, the article states 180-240 µm.--Squirmymcphee (talk) 18:12, 27 November 2007 (UTC)

That is an incomplete answer hAl. As stated, the question was how many total kg of Si per peak kW? To see why this is useful ask yourself this: If I was going to use $10 Billion to build a nuclear power plant or to build a solar Si production plant, which is more profitable? Why?

Reddwarf2956 (talk) 21:24, 26 November 2007 (UTC)

The original question does not ask about the price or cost of the Si nor does it take into acount the efficiency of the cells so what you want to know is probably something else that was originally asked. hAl (talk) 22:21, 26 November 2007 (UTC)

I think knowing how much silicon is used per peak watt is helpful in understanding how many MWp of PV a polysilicon plant will support, but I don't understand how it would be useful in answering your sample question. How does the number of MWp the plant will support affect the plant's profitability? It doesn't.--Squirmymcphee (talk) 23:02, 26 November 2007 (UTC)

Squirmy,

Where did you get the info you stated on the first "23:02, 26 November 2007?" It sounds right, but it needs cite info which Wiki-folks can handle.

On the second reply, http://www.eere.energy.gov/news/news_detail.cfm/news_id=10211

This link states some numbers which may help you see Squirmy. I was thinking in the sence of the owner of either a $10 billion nuclear power plant genterating 1 GW and selling the electric power, or using the $10 billion to see how many metric tons(MT) Si operation can be built. If the MT's of Si is used to make solar panels, as to grade suggest, how many GW can be made and sold in one year? Now do you see? It is like simple interest compared to compound interest when you take in more years. 25 years is the normal life span of a "plant" no matter what kind it is.

For the editor, can you please state the above better for the page? I am sure no one has stated a cost / benefit comparison with coal, nuclear, and solar cell production like this. Thanks

Reddwarf2956 (talk) 05:46, 27 November 2007 (UTC)

A problem with determining cost is that a solar energy plant does not really produces the solar power it generates from the sun. It has also transformer/conversion losses and most likely losses in temporary storage of energy because solar energy is typically energy that is not generated 24/7 and the generated energy may not match the energy demands. Just knowing how much the cells cost and how much they generate does not give you a good enough total picture for a solar power plant. hAl (talk) 15:52, 27 November 2007 (UTC)

On the first point, you can use the Sun Screen II report as a reference, though I have several offline references as well. If I have time in the next few days I'll try to add it to the article.

On the second point, I think you're confusing the issue even more. The first time you posted your question you made it sound like a financial question -- which option would be more profitable? Now you seem to be asking which one will result in the greater number of GW. Those are each very different questions. How many GW of PV modules you could produce from the silicon is completely irrelevant to the profitability question. The cost to build a silicon feedstock plant is about $50,000-100,000/tonne, though, so with $10 billion you could build a 100,000-200,000 tonne plant (by way of comparison, current global feedstock capacity is approaching 40,000 tonnes). At 10 g/Wp, this would support 10-20 GWp in annual PV module production. At current PV system prices, though, building those modules and deploying them in the field would require an additional $90-180 billion (though in reality an investment of that size would most likely go a long way toward reducing downstream costs, so I would expect the actual size of the additional investment to be less than that -- how much less is impossible to estimate).--Squirmymcphee (talk) 18:12, 27 November 2007 (UTC)

Ormosil; economics

I deleted a reference in the silicon section to Ormosil. This seemed to be an advertisement, it wasn't referenced, and most notably, it is a form of silica, not silicon. Geoffrey.landis (talk) 01:53, 7 December 2007 (UTC)

--also, the end of the article had the statement: "those companies with large scale manufacturing technology for coating inexpensive substrates will be the lowest cost net electricity producers, even with cell efficiencies that are only in the 5–7% range." This is an opinion, not a fact, and it is not at all the consensus of opinions that efficiencies as low as 5-7% could result in cost-competitive production of electrical power. I modified this to "the most effective cells for low cost electrical production are not necessarily those with the highest efficiency, but those with a balance between low-cost production and efficiency high enough to minimize area-related balance of systems cost. Those companies with large scale manufacturing technology for coating inexpensive substrates may, in fact, ultimately be the lowest cost net electricity producers, even with cell efficiencies that are lower than those of single-crystal technologies." Geoffrey.landis (talk) 18:46, 7 December 2007 (UTC)

"Makers" section

I changed the list of solar cell-manufacturing nations so that only the most significant producers are named. I'm not sure it's desireable to list them all in the article -- here are the currently producing nations I can come up with off the top of my head: Japan, Germany, USA, China, Taiwan, Spain, Australia, Philippines, India, Canada, France, Norway/Sweden (same manufacturer, different parts of the process in different countries), Switzerland, Italy, Czech Republic, Russia, Saudi Arabia, South Africa, the Netherlands, and Belgium. Furthermore, if South Korea and Malaysia don't already have production, they will soon (the plant First Solar is building in Malaysia will be one of the largest in the world). That's only off the top of my head -- I've likely missed a few. Also, am I the only one who thinks it's a bit odd to have the cost/efficiency discussion under the heading "Makers" (which I would suggest be changed to "Manufacturers")?--Squirmymcphee (talk) 17:05, 12 December 2007 (UTC)

Difficult for the laymen

Not knowing much about electricity it is a bit hard for me to understand how efficient solar cells are. How many square centimeters would I need to generate i.e. 9V DC on a bright sunny day? 88.68.197.232 (talk) 17:57, 12 January 2008 (UTC)

conveniently, the solar intensity is about a kilowatt per square meter (at noon, on a clear day, on a surface flat on to the sunlight). So multiply the area in square meters by the efficiency, and you'll get the power output in kW (at noon, on a clear day, on a surface flat on to the sunlight).

If instead you want, say, kw-hr per square meter per day, try the maps on the NREL website for your location, or make a map from NREL data based on the criteria you select. Geoffrey.landis (talk) 22:08, 12 January 2008 (UTC)

Each cell might produce about 1V. So to produce 9V DC you would need about ten cells, connected in series. But they could be very small! The larger they are, they more current and power they could produce in sunlight. -69.87.199.222 (talk) 13:53, 11 April 2008 (UTC)

Solar cell, wafer and module

I suggest include information about the relations between solar cells, wafers and modules. --Mac (talk) 14:45, 22 January 2008 (UTC)

That's to be found on the photovoltaics article--Oldboltonian (talk) 09:43, 17 February 2008 (UTC)

Germanium

What about Germanium as a light absorbing material: it is cited as used in space, as a layer for TPV and as a substrate for multijunction stacks.--Oldboltonian (talk) 09:44, 17 February 2008 (UTC)

I don't recall off the top of my head how well germanium absorbs light over most of the solar spectrum, but in theory it should absorb more of the solar spectrum than silicon does. However, because of its low bandgap, most of the solar energy it absorbs will be converted to heat instead of electricity. Relative to silicon, it will produce more current, but a much lower voltage and a much lower power output. Throw in that it's a whole lot more expensive that silicon is, even at today's inflated silicon prices, and it's a bit of a non-starter. As you note, though, it is used for TPV, but that's because it's small bandgap makes it efficient at converting infrared photons to electricity. Those photons make up the bulk of the photons emitted from thermal sources (hence the "T" in TPV), but only a small fraction of those emitted by the sun.

For space cells, where it is used as a substrate for multijunction stacks, its job is mechanical -- it doesn't contribute to the electrical operation of the device. Instead, it's just a piece of material to deposit the electrically active semiconductor materials upon. The reason germanium is used for that has to do with lattice matching, which is just a fancy way to say that the spacing between germanium atoms is similar to that of GaAs and the other III-V semiconductor materials often deposited on germanium substrates. (If you deposit a material on top of a substrate that is poorly lattice-matched you produce mechanical stresses that can introduce defects into the material and interfere with solar cell operation.)--Squirmymcphee (talk) 22:50, 17 February 2008 (UTC)

interesting links

I found interesting links some months ago:

• 45% efficiency in 2010
• 48% efficiency (industrializing)
• 60% efficiency
• 65% efficiency quantic cell
• photovoltaic cell - 87% theoretical efficiency

Van Rijn —Preceding unsigned comment added by 82.67.12.72 (talk) 23:13, 12 March 2008 (UTC)

Multi-junction cells

Throughout this article, there are references to "GaAs" cells. GaAs is only one of two or more junctions in the cell. A more proper name is multi-junction cells, or III-V cells. Any objection to this revision? Wefoij (talk) 20:18, 26 March 2008 (UTC)

That depends on which GaAs reference you're talking about -- sometimes GaAs is the only junction in the cell. Some of those references could be accurately and more generally replaced by references to III-V, but since single-junction III-V cells are almost always GaAs I would tend to leave those references alone. The section of the article title "Gallium arsenide (GaAs) multijunction" is mistitled, in my opinion. For one thing, as you say, GaAs is only one of the junctions in the cell. In addition, the name implies that the cells are made of multiple GaAs junctions, then goes on to describe cells that clearly are not. It would probably be better titled simply "Multijunction". Where GaAs references actually mean multijunction, they should be changed to multijunction.--Squirmymcphee (talk) 14:48, 28 March 2008 (UTC)

Circuit equations

The circuit equations added to the article a little while ago are unsourced, unexplained, and (in my opinion) overly complex. The first equation can be derived from the equivalent circuit shown next to it, though it is lacking the shunt resistance term. The remaining equations appear to be empirical equations geared toward estimating the I_o and series resistance (R_s) terms from easily measured data -- as far as I can see they can't be derived from first principles. Furthermore, G is omitted from the list of variables, but if it means what it conventionally does then equation 3 is incorrect. Finally, since these equations are empirical they aren't very useful without some context.

Can anybody -- perhaps the editor who posted the equations -- comment on whether they think these equations add anything to the article and why? In my opinion they add nothing, particularly since there is no context and no reference provided with which to find context. Since equation 1 is derived from basic physical principles and appears frequently in PV literature I think it could stay, but only if it is updated to reflect the shunt resistance term and some sample I-V curves are added to help explain it. I see no reason to keep the others, though.--Squirmymcphee (talk) 03:46, 8 April 2008 (UTC)

Agree: not helping the article.--Old Moonraker (talk) 06:49, 8 April 2008 (UTC)

Regular vandalism

It seems this article is vandalised on a very regular basis over a long period. Probalby by the same person. It might be a good idea to get an admin to protect this article from anonymous edits for a few months. hAl (talk) 14:59, 8 April 2008 (UTC)

"Simple Explanation"

I don't think this sounds encyclopedic at all. —Preceding unsigned comment added by 35.11.38.122 (talk) 22:50, 8 April 2008 (UTC)

Energy required to produce solar cells

Solar cells harvest energy from light. It takes energy to produce solar cells. Please improve the article by adding info about the energy required to produce the various kinds of solar cells. -69.87.199.222 (talk) 13:56, 11 April 2008 (UTC)

Waving by degrees

Hello, Having recently re-read this article, it is my opinion that far to many "Dr so and so from XYZ university with HJK industries has....". I am going to remove this information soon (~12hrs) as this does not contribute information to the article. It is my opinion, and rather strongly, that researchers should not be using this article to further their work. If readers are to be kept abreast of the latest developments the appropriate way to do this, which is also true in scholarly work, its to reference the work of these authors by citing them, not by advocating their names. User A1 (talk) 00:21, 21 April 2008 (UTC)