Forgot your password?
typodupeerror
Power Science

Solar Cell Achieves 40% Efficiency 632

Posted by kdawson
from the too-cheap-to-meter dept.
Fysiks Wurks found on the U.S. Department of Energy website news of a breakthrough in solar energy efficiency. From the article: "...with DOE funding, a concentrator solar cell produced by Boeing-Spectrolab has recently achieved a world-record conversion efficiency of 40.7 percent, establishing a new milestone in sunlight-to-electricity performance." A page linked from Wikipedia's article on solar energy calculates the land area that would need to be covered by solar collectors at 8% efficiency to meet the world's energy needs (using 2003 figures). At 40% efficiency, it looks like a square 265 miles on a side in the American southwest would do it.
This discussion has been archived. No new comments can be posted.

Solar Cell Achieves 40% Efficiency

Comments Filter:
  • Re:Cost is the issue (Score:5, Informative)

    by GreyPoopon (411036) <gpoopon@@@gmail...com> on Wednesday December 06, 2006 @03:47AM (#17125674)
    Very nice, but I'd rather see a reduction in cost per watt than an increase in efficiency.
    By reducing the number of solar collectors needed or the area that needs to be covered, the installation costs are significantly reduced. The article indicates that this new technology could yield systems with installed costs of as little as $3 per watt.
  • Re:Cost is the issue (Score:2, Informative)

    by rgravina (520410) on Wednesday December 06, 2006 @03:50AM (#17125696)
    I'd rather see a reduction in cost per watt than an increase in efficiency

    Aren't the two related?

    Also, FTFA:
    This breakthrough may lead to systems with an installation cost of only $3 per watt, producing electricity at a cost of 8-10 cents per kilowatt/hour, making solar electricity a more cost-competitive and integral part of our nations energy mix.

  • where the facts? (Score:5, Informative)

    by Anonymous Coward on Wednesday December 06, 2006 @04:07AM (#17125808)
    So it's a bit unclear what the article means by 40% efficient as the article seems to confuse the concentrator part of the solar cell with the multi-junction part. The concentrator doesn't make the device more efficient at converting solar radiation into electrical power, it just concentrates the light so you don't have to use as large of a device. The idea being that the solar cell material is expensive but the optics are relatively cheap, so you might as well focus as much light on the device as it will absorb and still function.

    The multi-junction part comes from the idea that you can, using a solar cell, only extract as much energy from a photon as the size of something called the band gap of the material that the cell is made from. At the same time, a solar cell can only absorb photons with energies higher than the band gap. If the bandgap is small, as it is in silicon, then you can absorb most of the suns rays, but you can only get about 1 electronVolt of energy out of each one no matter how much energy the photon has. Since the bulk of photons emitted by the sun have more than 1 electronVolt of energy Si solar cells waste alot of the energy in sunlight as heat. If you make the solar cell out of a semiconductor with a larger bandgap then you absorb fewer photons (more of the solar spectrum lies below the critical energy for absorption) but you extract more energy from each photon. So, for a solar cell made from one material there is a sweet spot in terms of the bandgap that maximizes the energy extracted. Multi-junction cells try to overcome this by combining multiple devices with different bandgaps so that you can maximize both the total number of photons converted to electricity and the energy extracted from each photon.
  • Re:Cost is the issue (Score:5, Informative)

    by dch24 (904899) on Wednesday December 06, 2006 @04:13AM (#17125838) Journal
    In addition, 40.7% is just a bump up from 39% [boeing.com], which (apparently) Spectrolab has been achieving for the better part of the year. They may be very close to high-volume production. Direct photovoltaic solar generation is an immediate revenue source, but solar energy can be directly applied for other processes, the most notable being desalination [wikipedia.org].
  • by Anonymous Coward on Wednesday December 06, 2006 @04:15AM (#17125860)
    Hey MichaelSmith,

    Here's the link [wikipedia.org] you forgot.
  • Re:transport losses? (Score:5, Informative)

    by Anonymous Coward on Wednesday December 06, 2006 @04:18AM (#17125884)
    And hydrogen transports just as easily as oil via the same infrastruture.

    Bzzt! Wrong answer. Hydrogen requires a completely different infrastructure that has never been massively developed. Transporting hydrogen trapped in a hydrocarbon is feasible and could use the same infrastructure, but hydrogen itself is a much more complicated issue. You either need to cryogenic cooling or you need to build infrastructure that has low hydrogen diffusion and low hydrogen embrittlement (and probably very high pressure to move a significant energy density of hydrogen around if you go the gaseous path). People who want hydrogen for various industries tend to steam reform it from hydrocarbons instead of using this oil infrastructure you think can transport hydrogen.
  • by Nasarius (593729) on Wednesday December 06, 2006 @04:47AM (#17126032)
    Contact your local power company. Many (such as LIPA) will pay for a large percentage of your costs.
  • Re:where the facts? (Score:2, Informative)

    by swebster (530246) on Wednesday December 06, 2006 @04:54AM (#17126062) Homepage

    The concentrator doesn't make the device more efficient at converting solar radiation into electrical power, it just concentrates the light so you don't have to use as large of a device.
    Actually, that's not quite right. Higher light intensity does make the cells more efficient. It's one of the advantages of using concentrator cells.
  • Re:transport losses? (Score:2, Informative)

    by poopdeville (841677) on Wednesday December 06, 2006 @05:02AM (#17126106)
    And yet somehow we have no problems tranporting oil to non-oil producing regions.

    We don't?

    Wait, we do. And that's the prime economic reason developing alternative energy strategies is in the US's (and everyone else's) best interests, despite our reliance on our current profits in the energy market.

  • Figures a bit out (Score:4, Informative)

    by tttonyyy (726776) on Wednesday December 06, 2006 @05:18AM (#17126186) Homepage Journal
    According to this site [clean-energy.us], estimated world demand was 13.9 trillion kilowatt-hours in 2001.

    13.9 trillion kW/h / 8776 (hours/year) = 1.58TW

    This figure is comparable to the statement in the wikipedia that 2001 average world consumption was 1.7TW in 2001 [wikipedia.org]. So our sources agree within a reasonable margin.

    According to the wikipedia, the energy density from solar energy reaching the surface as a global average is 170 W/m2 [wikipedia.org]. At 40.7% efficient, that's 69.2W/m2.

    Using the lower figure of 1.58TW calculated above, you'd need 22.8 x 10^9 square meters, or approximately 8800 square miles of solar cells to meet 2001 world demand. (Or "just" 1900 square miles to meet the peak US demand of ~3 trillion kWh in the late 90s). Of course, these areas halve if sited in an area of the US where the solar energy density is 375 W/m2 (4000 square miles for world demand, 860 square miles for US demand).

    Neither correspond to the whopping (265x265) 70000 square miles the article summary claims. Sorry kdawson, looks like you're a magnitude out!
  • Re:transport losses? (Score:5, Informative)

    by dbIII (701233) on Wednesday December 06, 2006 @05:19AM (#17126190)
    The big point with photovoltaics is you can stick a little panel just about anywhere and not worry about line losses or being on the grid at all - plus the lead time to set something up is very low - buy a panel and get an electrician to wire it up. The big problem with photovoltatics is it doesn't scale up - so for a really big facility you are better off with something that does like a thermal solution or very large water turbines if you are lucky enough to have somewhere to put them. Having a lot of cheap mirrors putting heat on some expensive photovoltaics gets halfway there.

    If you are just going to put bare panels somewhere it makes more sense to stick them on the top of existing poles instead of in some big facility since they act as discrete units anyway. Once they get rolled out there really isn't much that has to be done with them - the photovoltaics that existed when Einstein was young probably still work.

    Personally I think we are already seeing the start of one of the major potential uses for photovoltaics - appliances that don't have to be plugged into the grid. If the prices come right down things like solar mobile phone chargers may well become mainstream.

  • Re:Cost is the issue (Score:5, Informative)

    by hankwang (413283) * on Wednesday December 06, 2006 @05:23AM (#17126212) Homepage
    It wasn't all that long ago that the electricity needed just to melt the silicon was more energy than the cell would generate throughout it's entire lifetime (they do degrade over time).

    I don't know about how long ago you are talking, but the Energy return on investment [wikipedia.org] varies between a factor 4 and a factor 17 for current solar cells, rather than a number below 1 as you are suggesting.

  • Re:Cost is the issue (Score:2, Informative)

    by catprog (849688) on Wednesday December 06, 2006 @05:31AM (#17126252) Homepage
    Current energy cost to manufacture solar cells are about 1.8 years now
  • Gallium Nitride (Score:5, Informative)

    by GanjaManja (946130) on Wednesday December 06, 2006 @05:54AM (#17126370)
    A student at The Univ. of California, Santa Barbara just presented research showing the use of multi-junction devices using Gallium Nitride. This is awesome because Nitride materials are very well suited for a HUGE amount of the sun's radiation, and since he managed to perfect a way of sticking several layers of differently absorbing Nitride Materials together in ONE device, we could theoretically see solar cells that absorb the Entire spectrum of the sun's rays in the near future!

    Here's some links:

    Indium-Gallium-Nitride can be made to absorb the entire spectrum of solar rays:
    http://www.lbl.gov/Science-Articles/Archive/MSD-fu ll-spectrum-solar-cell.html [lbl.gov]

    Tunnel Junctions - this is how you stick together many different layers of material, each layer with their own optimal absorption range (in terms of wavelength, aka. color):
    http://www.hitachi-cable.co.jp/ICSFiles/afieldfile /2005/11/28/review07.pdf [hitachi-cable.co.jp]
    (sorry, this is the best I could do, there was no simple paper explaining a tunnel junction. "tunnel" is for electron tunneling...)

    In essence, you have different layers that absorb only one range of wavelengths (colors of light), and whatever isn't absorbed goes straight through, and the next layer absorbs another range, etc. etc.

    As an aside, did you ever wonder how blue LEDs & lasers finally managed to get working? Nitrides paved the way for emission (and absorption) in a range of visible wavelengths, including blue. This is also why they're great for this application.
  • Re:transport losses? (Score:3, Informative)

    by Sique (173459) on Wednesday December 06, 2006 @06:19AM (#17126480) Homepage
    The best ones are the directly injected Diesel engines (like the TDI or HDI) with about 46%, then the directly injected gasoline engines (~43%). For a car you have to subtract the losses due to the transmission, the clutch and the tires, and you have to take into consideration that the engine runs most efficiently only in a narrow band of rotations per minute (around 40% of the max revs for four-stroke machines, about 70% of max revs for two-stroke machines).
  • Re:life span (Score:3, Informative)

    by BlackPignouf (1017012) on Wednesday December 06, 2006 @06:27AM (#17126522)
    You're not the first one to ask this question, and I won't be the first one neither to give this link:
    http://en.wikipedia.org/wiki/Photovoltaic?section= Energy_return_on_investment#Energy_return_on_inves tment [wikipedia.org]
    I tried to calculate energy payback-time for different cells, and got results ranging from 8 months to 2 and a half years.
    Even extreme PV-Cells bashers don't succeed in proving that payback-time exceeds 5 years, which still lefts you 3 times as much "free" energy.
  • Re:where the facts? (Score:5, Informative)

    by frostband (970712) on Wednesday December 06, 2006 @06:53AM (#17126660) Homepage

    It's not necessary to do it that way. The way these multi-junction cells work is you have several layers of different semiconductor materials (with varying band gaps as the parent said). The material with the largest band gap is on top and the band gap of the material decreases as you go down the layers of the device. If a photon is not absorbed in the first layer (meaning the photon doesn't have very high energy, since, as the parent also said, the photon must have greater than the band gap energy to be absorbed), then it continues on to the next layer to be absorbed, then the next layer. This way, you are extracting the maximum amount of energy out of each photon.

    That isn't a perfect explanation and any experts out there, please correct anything that's wrong.

  • by kotku (249450) on Wednesday December 06, 2006 @07:33AM (#17126888) Journal
    A decentralized power generating system is good for national security. Imagine that every house had solar generating capacity. There could be a guaranteed minimum power capacity per house. Sure, the system would be degraded in the event of the base supply being knocked out but supply would be enough for critical services that people rely on, heating, cooking, water pumping etc. It may take a long while for the base supply to be re-integrated in the event of a coordinated strike/failure against public utilities.

    In Australia a few years ago there was a major disaster in the gas supply system that took a whole season to fix. The entire southern region was without gas for heating and cooking for weeks. Luckily the electricity system was still operational but a simultaneous failure would have resulted in a calamity.
  • Re:transport losses? (Score:5, Informative)

    by hcdejong (561314) <hobbes@ x m snet.nl> on Wednesday December 06, 2006 @07:44AM (#17126960)
    Nuclear power is an inefficient method to create a buffer. You'll need to run the reactor at a significant power level to keep the steam circuit hot enough that you can start generating immediately. Starting up a cold reactor takes hours, so you're better off not shutting it down at all.
    And even at low power levels, your fuel will keep fissioning merrily along, so in essence you're throwing away a finite resource. Also, your buffer will be significantly more expensive than the solar energy you're using as primary.

    If you have an abundant source of renewable energy, you're better off using some of that to drive a buffer. Hydro buffer plants such as Dinorwig (see elsewhere in this discussion) have been shown to work well.
  • Re:transport losses? (Score:4, Informative)

    by Dilaudid (574715) on Wednesday December 06, 2006 @07:50AM (#17126992)
    Over history oil prices have been inversely proportional to growth in GDP. Hence the parent logically concludes that reducing the price of energy will increase GDP i.e. the economy will grow. The people in the power industry may lose jobs - however the enormous reduction in costs in the manufacturing sector (boosting say the automobile manufacturing sector), the decreased cost of electricity (air conditioning unit sales would take off), and increased household disposable income (lower bills) would compensate this for most people.

    Grid power would fall in price - because a) there is reduced demand for it (everyone is using their own panels) and b) supply would increase (people can sell the excess electricity from their panels back).

  • Re:Cost is the issue (Score:5, Informative)

    by Alioth (221270) <no@spam> on Wednesday December 06, 2006 @09:19AM (#17127732) Journal
    At current prices, you'll need a little more than 6 months on your mortgage. Assuming you're in Britain, which by the usage of your language is probably reasonable...

    I bought an 80 watt peak solar panel in the summer, basically as a fun project and to investigate the practicality of generating some of my own electricity. Here is how it works out, using a monocrystalline panel (the most efficient panel commercially available at present):

    Peak power is produced only within about an hour or so each side of mid day on a bright, cloudless, hazeless sunny day.
    Three hours before or after mid day, the unit produces about 50% of peak.
    Five hours before or after mid day, the unit produces around 10-15% of peak
    At mid day, summer time haze with 10 miles visibility will cut output to around 80% of peak
    At mid day, with thin cirrus clouds (still bright sunshine), output is around 50%
    At mid day, on a bright cloudy day where shadows are still cast, output is around 15%
    At mid day, on an overcast day, output is generally 5% or less.
    In the winter, I've never seen the unit capable of producing more than about 25% of peak on the brightest winters day.

    All in all, the average output even in the summer will only be 5% of peak (because of night time, and cloudy days). Winter time is even worse. So if you want to make sure you have an average of 200 watts - which really isn't a lot, but if you can store it or put it back on the grid it'll make your house more or less neutral in terms of the electricity you use, if you have the normal domestic cycle of being out and not using much electricity during the day. To get that average of 200 watts, you'll need 4000 watts peak of solar panels.

    80 watt panels cost (in quantity) around £250 a piece. That'll cost you £12,500 *just* for the panels, without a grid tied inverter and storage system or installation (probably another 4 to 6 grand) - to get a measly average of 200 watts - i.e. just enough to power one Pentium 4 computer continuously. It's simply not worth doing at all unless you can put it back on the grid (not many electricity companies let you do that - yet), or store it in batteries - since if you have a normal domestic cycle, while your solar panels are producing near peak you will be away from the house and letting three or four thousand watts go wanting. You'll probably need three grand's worth of batteries if you can't sell back to the grid - and even deep cycle leisure batteries are going to need replacing at least once every 10 years. This is for a system which will only work reasonably well in the summer. In the winter, when the days are short and you need the most power, it'll hardly contribute anything - perhaps you'll get 50 watts average from £12,500 worth of solar panels.

    If solar panels were 1/10th of the price they are now - yes, it'd be worth it. I'm waiting for the breakthrough in price, not efficiency (if the efficiency brings the breakthrough in price all the better). Even a moderate sized south facing roof - I've calculated just my shed roof replaced with solar panels could produce 1kW peak - is large enough for a decent peak output using current monocrystalline panels. Price is everything. If I could get the panels at 10% of what they cost now, you bet my shed roof (my only south facing roof) would be covered by the spring. But at the current price point? It's simply not affordable for the meagre amount of electricity you get. It's a shame because the panels aren't visually intrusive and they are silent and almost maintenance free, unlike wind turbines. I really really want solar panels to be worthwhile - but at the moment - at current prices, they simply aren't.
  • Re:transport losses? (Score:5, Informative)

    by olyar (591892) on Wednesday December 06, 2006 @10:29AM (#17128702) Homepage Journal
    I'm building a new house right now and will be putting on solar panels. Its an easy decision these days - at least in the part of California where I live.

    I'll get about a quarter of the cost back in refunds from the power company right up front. The remaining cost (around $25k) will roll into my mortgage, which will increase it by around $100 per month (30 year mortage at 6%). My monthly electricity bills should be reduced by at least $150. It just makes sense.

    The fact that I care about the environment just makes it an even better deal.

  • Re:transport losses? (Score:3, Informative)

    by jandrese (485) <kensama@vt.edu> on Wednesday December 06, 2006 @11:10AM (#17129542) Homepage Journal
    Two things:
    1 Manufacturing solar cells is currently an extremely energy intensive process, it also creates a startling amount of toxic waste. Solar Cells need to be replaced on a regular basis as well.
    2. Solar Cells are extremely expensive. I don't know about these 40% jobs, but something tells me they won't be a lot cheaper than our current top of the line cells. We're talking about a project in the trillions of dollars to do what you describe.
  • Re:transport losses? (Score:3, Informative)

    by drinkypoo (153816) <martin.espinoza@gmail.com> on Wednesday December 06, 2006 @12:30PM (#17131082) Homepage Journal
    You don't need batteries if you just sell your excess power back to the grid. You don't get much money for it when you do that but you don't need to store it, either. Also, the beauty of that is that most of us use the most power at night, because most of us are gone during the day. At least where I live, in order to sell back power, you need a time-of-use meter anyway... If you have a time-of-use meter you get off-peak power at rock bottom prices because they want to encourage people to use it then, since they have excess power generation. Also in my area, getting a time of use meter is free, because again, they want people to be motivated to use off-peak power. All you have to do is pay for the sync converter (or whatever that unit is) to hook your inverter up to the mains and bingo! Of course you still have to upgrade your box and bring your wiring up to code, which as you point out, is a very serious issue if you can't do it yourself and is usually still pretty major even when you can.
  • Re:transport losses? (Score:3, Informative)

    by Maxo-Texas (864189) on Wednesday December 06, 2006 @02:10PM (#17133318)

    Right off the top you get $29,000 from other oregon and city taxpayers to buy your system. In any state without subsidies, that means over THIRTY YEARS longer payout period. Most solar cells degrade substantially by 20 years (10-20% less power output). Likewise, I'm betting like MOST states, that not just anybody can get those $29k in credits in Oregan. There is probably a fairly limited budget (a few million) and once that is gone each year, there are no more subsidies until the next calendar year.

    I'm unclear how I'm supposed to use Federal Depreciation if it is not for a business or rent house. Granting that I can somehow use it for personal income taxes tho... it sounds like it would be a deduction, not a credit. That would mean it would lower your INCOME by roughly $3 grand a year- which would mean it would lower your TAXES by $1 grand a year at most. That lowers your savings from $13,000 to $5,000. That means your system costs $8,000 more --- that's 10 years more to cover costs. So you are up to 16 years, using your figures, and assuming you are lucky/smart enough to be one of the people that gets the subsidies.

    Using your figures, for the 40 some-odd states where we only get the federal credits, the payout period is longer than the likely lifespan of the item.

    ---

    Now one thing we need to remember is this: Power is appreciating about as fast as inflation. That means your $813 bill today is going to be $1600 in 10 years and $3200 in 20 years. That works in your favor. It helps home buyers too. In 1995 my payment was $700 a month and rent was $655 a month. Today my payment is $700 a month and rent is $1100 a month. In 2015, my payment will be $700 a month and rent will be roughly $2200 a month.

    ---

    No.. It's not just an inverter. If you are connected to the grid you need a special circuit box that won't allow power to flow backwards into the grid unless the grid has power. I think it is called a reverse switchback circuit breaker.

    The inverter converts DC - > AC.
    The charge controller keeps the batteries charged and cycled at the proper levels so they have optimum lifespan.
    The switchback? circuit box (can't remember the name exactly) keeps your 1.21 gigawatts of power from killing electric company linemen.

    ---

    No need to be insulting about the AC ideas. If solar power drops by an order of magnitude it is suddenly a very clever and reasonable way. $900 of solar power and a simple $200 window unit and you don't need to run your central AC during the day at all. No need to upgrade your electrical system. No need to have charge controllers, batteries, etc.

    ---

    There are several companies on the edge of dropping the price by an order of magnitude. Two of them basically print solar material with nano-particles. If any of them succeed and solar power does drop to 10% of the current price and becomes unbelievably easy to install (picture unrolling a 50'x3' long strip on your roof that costs about $3000 and produces 1.5 mw of power during the day with no clouds and it has a plug on one end to hook to another strip- no panels, no sun tracking arrays- nothing to break- no cells to be corrupted by water leakage. cool stuff)

    As far as the other things- I'm already using efficient appliances, have a programmable thermostat, added radiant barrier, and use florescent lights.

    ---

    I'm pro solar power. It just isn't practical yet unless you can get other people to help you buy it through subsidies. That's only an option in some states.
  • by NatteringNabob (829042) on Wednesday December 06, 2006 @03:31PM (#17134926)
    once you count the infrastructure costs. I own an off-grid second home which is about 3000ft from the nearest power pole. The cost to extend the power to our house is estimated by PG&E at about $20/ft, so about $60,000 to get to our house, and that is *after* you have negotiated an easement over the neighboring properties. By contrasts, a complete off-grid systems run about $10000/KW, so you can have a nice 3KW system for about $30K, or 1/2 the price, and the 'generation' cost after that is the cost of replacing the lead/acid batteries, which, unfortunately, are still the best storage alternative. Yes, it only works in places where there is a lot of sunlight, and you still need a generator for night and winter months, and it helps a lot to have all florescent lights (which, fortunately has also improved dramatically). The fact of the matter is that once everything is factored in, solar already looks pretty good. If you factor in the cost of things like conquering oil producing states (as well as the cost of maintaining a military large enough to do so at any time), solar is an absolute bargain.
  • Re:transport losses? (Score:3, Informative)

    by Rei (128717) on Thursday December 07, 2006 @03:02PM (#17149834) Homepage
    Methane reformation releases CO

    Carbon monoxide is highly flammable. It has the same NFPA rating as diesel (2). Burn it, recapture the energy. With a good countercurrent heat exchanger, you approach 100% efficiency. You can never reach 100%, and eventually it's not worth the effort to try for more, but you can get close.

    It's less efficient than burning all of the natural gas to heat your water

    Of course it's less efficient than burning the natural gas when you only consider the heat; you're getting hydrogen, too, and that's where part of the CH4's energy ends up. Burn the carbon monoxide, and your net result is:

    H2
    CO2
    Heat

    "Wasted" energy will end up as heat. Yes, there will be no perfect heat exchanger (which is why I said "almost"), but countercurrent heat exchangers can do a very good job.

I cannot draw a cart, nor eat dried oats; If it be man's work I will do it.

Working...