January 02, 2008
Photovoltaic Prices To Drop In Half By 2010?
Skyrocketing demand has kept up the prices for solar photovoltaics for several years running. However, the Earth Policy Institute expects rising production capacity to finally cause a big decline in photovoltaics cost in the next few years.
The average price for a PV module, excluding installation and other system costs, has dropped from almost $100 per watt in 1975 to less than $4 per watt at the end of 2006. (See data.) With expanding polysilicon supplies, average PV prices are projected to drop to $2 per watt in 2010. For thin-film PV alone, production costs are expected to reach $1 per watt in 2010, at which point solar PV will become competitive with coal-fired electricity. With concerns about rising oil prices and climate change spawning political momentum for renewable energy, solar electricity is poised to take a prominent position in the global energy economy.
Regarding competitiveness with coal: There are the not so minor details of where and when. Certainly photovoltaics become cost competitive in Arizona before Colorado and in Colorado before Alberta or England. So in the more northern climes and in cloudier areas the prices of photovoltaics will have to drop much further before becoming competitive. Also, photovoltaics will compete on June 21 in the northern hemisphere years before they compete on March 21, let alone December 21. Plus, we need really cheap electric power storage before day time photovoltaic energy will help us much during the night time. So keep in mind all the caveats and short-comings of solar power when you read rosy scenarios about solar energy.
The company which many observers think has the best chance to cause this big cost decrease is Nanosolar. CEO Martin Roscheisen says Nanosolar can get their production costs below $1 per watt.
- the world’s first printed thin-film solar cell in a commercial panel product;
- the world’s first thin-film solar cell with a low-cost back-contact capability;
- the world’s lowest-cost solar panel – which we believe will make us the first solar manufacturer capable of profitably selling solar panels at as little as $.99/Watt;
- the world’s highest-current thin-film solar panel – delivering five times the current of any other thin-film panel on the market today and thus simplifying system deployment;
- an intensely systems-optimized product with the lowest balance-of-system cost of any thin-film panel – due to innovations in design we have included.
The printed thin film process with which Nanosolar has just started commercial production looks like the ticket. They avoid the costs of the thick polysilicon crystals and supposedly can produce at fast speed using a printing technology.
Nanosolar has nanoparticles and the ability to process long rolls of metal and put photovoltaic materials on continuously.
The San Jose-based Nanosolar developed a proprietary ink that is based on “nanoparticles” of a material called copper indium gallium selenide (CIGR), which can be printed on metal foil, which is cheaper and 20 times more conductive than stainless steel.
Other companies that also specialise in 'thin-film solar' technology also use CIGRs, but require a vacuum chamber to disperse the particles. Nanosolar says its method of printing is cheaper and more effective. It can literally produce huge rolls of the product that are metres wide and up to kilometers long.
But Nanosolar is already sold out into 2009. If their process turns them a big profit during this time they obviously can and will ramp up. So how quickly will they ramp up? Will they run into troubles running their manufacturing process continuously?
If the Democrats put their money (I mean the tax payers money) where their mouth is and start subsidizing electric battery R & D, then by 2010 we will almost certainly have excellent batteries that give plug-in hybrid cars a pure electric range of 60 miles. And by 2020 the range of these batteries will almost certainly be raised to 300 miles.
If the Democrats declare a Bronx Project for energy and for their four years spend $100 billion per year for energy R & D including batteries, nuclear power, solar, then we will see some results.
But meanwhile nuclear power is also very important, because solar power is not enough.
That's right. Once electric storage catches up to electricity generation, the dam will burst. All thoughts of a "hydrogen economy" will disappear with the new reality of the "electric economy."
The idea of using EVs as grid load levelers becomes absurd given reasonable electric storage technology. Nope. Everything depends on electric storage. Give us that, and the whole problem instantly becomes tractable.
I agree with Martin about storage holding us up. And the news is looking brighter on that front.
Altairnano just completed 2 battery packs for AES that stores 1 MW per pack. AES has bought 2 for a pilot application. This sounds like solid news; Altairnano is small but productive and has very advanced battery technology; and AES is an enormous utility rather than an imaginary 'someday' firm.
Try as I might, the V2G concept always seemed ???? to me. IMO the batteries in hybrids and EVs will always be more expensive per storage unit than for centralized storage at the utility because the vehicle batteries have to be lighter but also be tougher to take road shocks and weather. V2G thus puts extra wear on the most expensive batteries. Not a good strategy.
And V2G requires extra electronics to send power to the grid - electronics more expensive than when simply drawing power from it. But sometimes things unexpectedly work. I wish them luck.
I have jousted with Randall for years about massive government funding for battery research. In general I thought it won't speed up anything. He differed. Sobeit. Nuclear is really a political problem, it isn't as if the technology is new and unproved.
"photovoltaics will compete on June 21 in the northern hemisphere years before they compete on March 21, let alone December 21."
This doesn't really make sense. PV has no marginal generation cost, so once you own it you'll use it year-round. If you can justify it based on the whole-year production pattern, you can justify it.
"Plus, we need really cheap electric power storage before day time photovoltaic energy will help us much during the night time. So keep in mind all the caveats and short-comings of solar power when you read rosy scenarios about solar energy."
No one is suggesting that we use solar for 100% of our electrical needs. We could supply 25% of of our consumption with no storage at all. We have quite a ways to go before inadequate storage holds back solar.
V2G is a red herring - something that won't be important for load-levelling for many years. The important thing about PHEV/EV's is the load leveling effect of dynamically scheduled charging (especially for wind), something which can take place during the 23 hours of the day the typical vehicle is parked, and which requires no battery charge/discharge cycles. This only requires a smart meter, something which is being rolled out right now, especially in California.
We have entirely adequate batteries right now: see www.gm-volt.com . It will take 2-3 years to complete the engineering and production planning to produce new vehicles around them. Are they a trifle too expensive? Sure, but economies of scale will take care of that: whatever company GM chooses to produce their batteries will instantly become the largest li-ion supplier in the world.
As a practical matter, all the technology we need to handle peak-oil and climate change is here right now (though more R&D can always help - entirely adequate is, well, adequate, but ideal tech certainly would be...ideal). The main problem is capex lag.
Indium? It seems possible that indium will run out.
I do not like that.
Nanosolar didn't say they could get their costs to less than $1 a watt - he said he can profitably sell them (market price) at $1/watt. And this is true - because Nanosolar's cost is $0.30/watt.
The solar cells sold to Germany are $2/watt; that's $0.30 costs and $1.70 profit. Those profits are probably being plowed into ramping up production, but they won't (can't) stay that high forever. Free market competition eventually drives all prices to marginal cost (including the risk premium), so "1/3rd cheaper than coal" is already in the cards, even if the technology does not improve at all from here.
I think we're in pretty good shape for our energy future.
At this point we are probably too late for the massive government funding for battery research. However, the price of oil has gotten high enough to provide plenty of incentive for the private sector.
Agreed about hydrogen. Electrification is so much easier because it allows lots of smaller steps using existing infrastructure. We already have electric power running to every house and office and factory. PHEV cars will allow an intermediate step. Pure electric cars are doable now with existing electric power plants.
Hydrogen requires too big a coordinated shift with huge capital costs.
Perhaps I should be more clear: To make photovoltaics replace a substantial fraction of summer coal electric costs an order of magnitude less than building enough photovoltaics to replace the same fraction of coal electric in the winter. The photovoltaic panels that supply most of the electric power on a sunny summer day will supply very little on Dec. 21 on a cloudy day.
Was it you arguing with Nick Roush (if memory serves) on The Old Drum who has photovoltaic panels on a house in England? He can get 100 kwh in a summer period (I forget if day or week) in England and the same panels put out only 4 kwh in a similar time period in winter. I was struck by the factor of 25 drop-off. Obviously the drop-off is less severe further south where the days don't get as short.
CapEx: If that was the only obstacle prices would already be cheaper. Maybe in the case of solar CapEx just became the only obstacle now that NanoSolar is producing. But I seriously doubt that is the case with batteries.
What's your source for the 30 cents per Watt from NanoSolar?
Long run, humans should become able to transmute elements, so no element will "run out."
The idea of using EVs as grid load levelers becomes absurd given reasonable electric storage technology.
It will be "reasonable" to use storage to displace motor fuel long before storage becomes competitive with peaking generation. Among other things, this means that transportation will keep the price of suitable batteries bid up beyond what utilities can justify paying for quite some time. However, those batteries will be out there; the Priuses on the road already have on the order of 1 GWH of storage between them. If their cost can be offset by supplying services to the grid, value is created from something people would be paying for anyway.
And V2G requires extra electronics to send power to the grid - electronics more expensive than when simply drawing power from it.
No it doesn't. AC Propulsion's reductive inverter/charger uses the same electronics it uses for charging, which are also the electronics it uses for driving the motor. It also uses the motor windings as the ballast inductance for the inverter/charger. IIUC, all the magic is done in software.
I'm surprised you didn't pick up on the 30 cent cost. Your buddies at Peak Oil News had a thread on it. :)
The original source I can find is Popular Science. http://www.popsci.com/popsci/flat/bown/2007/green/item_59.html
A year or two ago the owner was hoping for 50-cents/watt, so I guess they beat that. I suppose we should take that with a small grain of salt though, since it is unconfirmed by Nanosolar (not that I expect them to).
"To make photovoltaics replace a substantial fraction of summer coal electric costs an order of magnitude less than building enough photovoltaics to replace the same fraction of coal electric in the winter. The photovoltaic panels that supply most of the electric power on a sunny summer day will supply very little on Dec. 21 on a cloudy day."
Well, there's no question solar is less useful in winter, and that the UK probably has a lower ratio of winter:summer sun than most.
1st, wind, on average, is slightly stronger at night and during winter. More importantly, wind sites vary consistently in their seasonal and diurnal characteristics - it's possible to choose sites that balance out solar's patterns, and that in sum have much lower variance (intermittency) than you'd expect.
2nd, in the US (which has a more serious energy problem than the UK/Europe), my impression from reviewing NREL national insolation maps for both summer and winter is that the ratio of winter:summer in most of the US perhaps 40%, while the ratio of winter:summer electrical demand is perhaps 60%. That's not much of a mismatch.
3rd, part of the problem for the UK example (which actually was not a discussion of mine, IIRC) is that conventional silicon PV absorbs a rather smaller % of diffuse sunlight than of direct. Cloud filtering doesn't reduce insolation as much as you might think. Now, CIGS PV, like Nanosolar's, doesn't have that problem.
Oh, and 4th: wind also has a strong negative correlation with solar's weather pattern, i.e, when it's cloudy it tends to be windy. OTOH, when A/C demand is very high, there's a very, very high likelihood that it's clear and sunny. Works out nicely.
Those are excellent points on the "meeting load" argument, but they're bad for cost because that means you need an installed base of 1kW Solar AND 1kW Wind to replace 1kW of installed base of Coal (which burns well regardless of weather conditions). That would mean they would have to be HALF the price of coal to be competitive.
But for the most part I think it's a moot point. The future of wind energy is high-altitude wind, which blows steady and sure year round. Wind energy can meet the base load and solar can meet the day-time and summer costs, since most energy use is during the day and summer (your 2nd point). Time-of-day pricing will also encourage people to run energy-intensive applications at night when base load wind can run them.
"The future of wind energy is high-altitude wind, which blows steady and sure year round."
That's what Google seems to be working on. I'm not sure it's proven yet, though. Do you know of any working installations?
"they would have to be HALF the price of coal to be competitive. "
The cost premium to deal with intermittency probably wouldn't be high. It would be more expensive to push wind & solar over a combined 50% of kwh market share at the current state of technology, but we won't reach that point for a while, and the state of the art will have changed enormously.
1) Solar power peaks at noon. Whereas summer temperatures are quite high into the late afternoon and evening and electric demand peaks late in the afternoon. Heck, as a kid growing up in New Jersey I used to sleep at night with a fan above my head pulling air over me because it was so hot.
2) Solar power peaks on the first day of summer. Whereas it continues to get hotter in July and August.
3) But solar photovoltaics will help with declining natural gas production by reducing the need to burn natural gas to generate electricity at peak. The natural gas can just get burned when there is less sun.
Your name is too brief to serve as a primary key here or on The Oil Drum.
What do you think the cost premium is for intermittency? I think it will be considerable since intermittent energy sources require back-up capital equipment.
A major part of the afternoon peak is A/C. Ice storage can move the A/C demand peak to whatever time of day has the cheapest power, and a few tons of water may as well be free. The only reason this isn't already in widespread use is because consumers have been billed at a flat rate, and utilities have historically received a set return on whatever investment was required to meet demand. If you have ice storage, following the production peak is a piece of cake. Ditto with a significant fleet of PHEVs.
Solar thermal may be cheaper than PV for a while. If something like a solar Stirling dish system is built with some molten-salt heat storage, handling the afternoon and evening peak may be fairly cheap too.
No, I am not aware of any (non-prototype) high-altitude wind installations. I just meant that the vast majority of wind's potential power (and the only source of reliable wind) is high up. You can't bring the wind down to the surface, so if wind is ever going to be a real player in electrical generation, you have to go where the wind is. It's inevitable.
My statement of "The future of wind is high-altitude" could also be read as "The future of solar is daytime solar." Naturally no one tries to collect solar energy from the moon or stars at night, but that's essentially what we've been doing with wind to date.
"if wind is ever going to be a real player in electrical generation, you have to go where the wind is."
Wind is cheaper than natural gas now, and arguably the cheapest and (fastest) source available when you include externalities (the capital cost of new US coal generation is much more expensive than you might think). High-altitude wind installations may eventually be cheaper, but they're not necesssary to make wind competitive.
There's no question that solar insolation doesn't match demand perfectly. OTOH, much of the late afternoon peak is caused by people turning on A/C when they get home from work. That's the kind of behavior that even simple demand management (and programmable thermostats) can change - cheaper rates can move demand earlier.
The marginal cost of intermittency is very low right now. I don't see it being significant until solar passes at least 10% market share - phev charging begin to help absorb wind/solar intermittency in just 2-3 years. I see the marginal cost limiting solar to about 25-35% of market share in the medium long-term. In the longterm solar will become so cheap that it will overwhelm other sources - PV will simply become ubiquitous.
Demand management is greatly under-appreciated.
FYI: The solar-thermal scheme with cheap Fresnel reflectors and below-ground hot-water energy storage essentially eliminates the intermittency issue also.