Cheaper, More Efficient Solar Cells

[rhoenix]

Wednesday, March 21, 2007
Cheaper, More Efficient Solar Cells

A new type of material could allow solar cells to harvest far more light.

By Kevin Bullis

Much more efficient solar cells may soon be possible as a result of technology that more efficiently captures and uses light. StarSolar, a startup based in Cambridge, MA, aims to capture and use photons that ordinarily pass through solar cells without generating electricity. The company, which is licensing technology developed at MIT, claims that its designs could make it possible to cut the cost of solar cells in half while maintaining high efficiency. This would make solar power about as cheap as electricity from the electric grid.

The effort uses a type of material called a photonic crystal that makes it possible to "do things with light that have never been done before," says John Joannopoulos, a professor of physics at MIT who heads the lab where the new designs for solar applications were developed. Photonic crystals, which can be engineered to reflect and diffract all the photons in specific wavelengths of light, have long been attractive for optical communications, in which the materials can be used to direct and sort light-borne data. Now new manufacturing processes could make the photonic crystals practical for much-larger-scale applications such as photovoltaics.

StarSolar's approach addresses a long-standing challenge in photovoltaics. Silicon, the active material that is used in most solar cells today, has to do double duty. It both absorbs incoming light and converts it into electricity. Solar cells could be cheaper if they used less silicon. If the silicon is made thinner than it is now, it may still retain its ability to convert the photons it absorbs into electricity. But fewer photons will be absorbed, decreasing the efficiency of the cell.

MIT researchers developed sophisticated computer simulations to understand how thin layers of photonic crystal could be engineered to capture and recycle the photons that slip through thin layers of silicon. Silicon easily absorbs blue light, but not red and infrared light. The researchers found that by creating a specific pattern of microscopic spheres of glass within a precisely designed photonic crystal, and then applying this pattern in a thin layer at the back of a solar cell, they could redirect unabsorbed photons back into the silicon.

Today's solar cells already reflect some of the light that passes through the silicon. But the photonic crystal has distinct advantages. Conventional solar cells are backed with a sheet of aluminum. The photonic crystal reflects more light than the aluminum does, especially once the aluminum oxidizes. And the photonic crystal diffracts the light so that it reenters the silicon at a low angle. The low angle prevents the light from escaping the silicon. Instead, it bounces around inside; this increases the chances of the light being absorbed and converted into electricity.

As a result, the photonic crystal can increase the efficiency of solar cells by up to 37 percent, says Peter Bermel, CTO and a cofounder of StarSolar. This makes it possible to use many times less silicon, he says, cutting costs enough to compete with electricity from the grid in many markets. The savings would be especially large now, since a current shortage in refined silicon is keeping solar-cell prices high and slowing the growth of solar-cell production.

The company plans to work with existing solar-cell makers, applying its photonic crystals with a machine added to the solar-cell makers' assembly lines, Bermel says. But StarSolar needs to choose a large-scale manufacturing technique that will allow it to produce the photon crystals inexpensively. What's needed is a way to cheaply arrange two materials in an orderly three-dimensional pattern. For example, microscopic spheres of glass would be arranged in rows and columns inside silicon. Currently, techniques such as e-beam lithography can be used, but that's too slow for large-scale manufacturing.

Shawn-Yu Lin, professor of physics at Rensselaer Polytechnic Institute, has developed a method for manufacturing eight-inch disks of photonic crystal--a measurement considerably larger than what can be done with conventional techniques. The method, which employs optical lithography similar to that used in the semiconductor industry, works best for a type of solar cell that concentrates light onto a small chunk of expensive semiconductor material. Such a device would require a relatively small amount of photonic crystal compared with conventional solar cells. Lin says the technique could be applied for more-conventional solar panels, although it would be expensive.

Another potentially less-expensive method, called interference lithography, creates orderly patterns in the photonic-crystal materials. The method is fast and uses machines that are far less expensive than those used for conventional optical lithography. It also requires fewer steps than Lin's existing process, so he says it could be far cheaper. Such methods have been developed by Henry Smith, professor of electrical engineering at MIT, who was not involved with the StarSolar-related work. Smith says his interference-lithography method could be used to build templates for imprinting photonic-crystal patterns on large areas.

Another promising technique is self-assembly, in which the chemical and physical properties of material building blocks are engineered so that they arrange themselves in orderly patterns on a surface. For example, Chekesha Liddell, professor of materials science and engineering at Cornell University, has engineered building blocks in the shape of peanuts and the caps of mushrooms that line up in rows because of the way they fit together and the tug of short-range forces between them. She says this could be useful for assembling photonic crystals for solar cells.

With such approaches available, Bermel says that StarSolar hopes to have a prototype solar cell within a year and a pilot manufacturing line operating in 2008.

Well, this is certainly interesting. I've just begun looking into making houses more energy-efficient, and solar power made more efficient and cheaper to manufacture is very promising.

[His Divine Shadow]

We're still never getting above ~1kw per square meter though, as thats about as much energy the sun puts in a single spot.

[Wyrm]

We're still never getting above ~1kw per square meter though, as thats about as much energy the sun puts in a single spot.

True, but assuming a home electrical system rated at 200 A, and typical American lines voltage of 120 V, that's 24 kW for a home. Divide by the efficiency of the panel (37%) and you get the area of solar panel you would need to power this home during the day, about 65 m² of solar panel area is needed. You can find that on a typical rooftop.

It's only effective during clear, sunny days, and it has to be clear of snow and bird crap. But that's a given.

Though I'm wondering how much energy and nasty chemicals are used to create these panels in the first place, as well as typical lifetimes.

[dragon]

Nice even though I like the idea of the spray on solar cells that get energy even during cloudy days and from infrared.
link
granted going to be awhile before they hit the market.

[dragon]

damn ment to hit preview. Even if spary on don't work thin printable ones are already being sold. http://www.nanosolar.com/

[Lost Soal]

We're still never getting above ~1kw per square meter though, as thats about as much energy the sun puts in a single spot.

True, but assuming a home electrical system rated at 200 A, and typical American lines voltage of 120 V, that's 24 kW for a home. Divide by the efficiency of the panel (37%) and you get the area of solar panel you would need to power this home during the day, about 65 m² of solar panel area is needed. You can find that on a typical rooftop.

It's only effective during clear, sunny days, and it has to be clear of snow and bird crap. But that's a given.

Though I'm wondering how much energy and nasty chemicals are used to create these panels in the first place, as well as typical lifetimes.

The article says it can increase the efficiency BY UP TO 37%. The typical efficiency ratings are 12-15% while the world record is 40.1%
So using the typical figures, you get 17.81-20.55%

[Wyrm]

The article says it can increase the efficiency BY UP TO 37%. The typical efficiency ratings are 12-15% while the world record is 40.1%
So using the typical figures, you get 17.81-20.55%

Fine. That still means you can power your home on a good day completely with solar on a 134 m² panel with typical efficiency. Even if you can't afford this much roofage to solar cells, it'll still cut down on the amount of power you draw from the grid.

[Tricit]

Fine. That still means you can power your home on a good day completely with solar on a 134 m² panel with typical efficiency. Even if you can't afford this much roofage to solar cells, it'll still cut down on the amount of power you draw from the grid.

I recall one day talking to someone (on a game chatroom) that claimed he actually was paid (one way or another) money from the local energy companies for leaking energy into their grid. I am not so sure if just anyone can do this, but it sounds like a way to make easy money, provided you can afford the startup costs. An interesting note, he lived in California while I was talking to him. Maybe it varies by states?

[Sam Or I]

I think the price is bigger news than being slightly more efficient.

[Master of Ossus]

I recall one day talking to someone (on a game chatroom) that claimed he actually was paid (one way or another) money from the local energy companies for leaking energy into their grid. I am not so sure if just anyone can do this, but it sounds like a way to make easy money, provided you can afford the startup costs. An interesting note, he lived in California while I was talking to him. Maybe it varies by states?

No. National law requires power companies to pay people for power that they produce through renewable sources the same amount that it would have cost the power company to produce that electricity. Of course, given the costs of photovoltaic cells, now, it's essentially impossible to be better than the power company in the production of electricity, so you're looking at taking a loss if you try to make money by loading up your roof with solar panels--it's just a way for people who already put panels up to help defray their costs by giving them something to do with the power they produce but don't use.

[dragon]

I recall one day talking to someone (on a game chatroom) that claimed he actually was paid (one way or another) money from the local energy companies for leaking energy into their grid. I am not so sure if just anyone can do this, but it sounds like a way to make easy money, provided you can afford the startup costs. An interesting note, he lived in California while I was talking to him. Maybe it varies by states?

No. National law requires power companies to pay people for power that they produce through renewable sources the same amount that it would have cost the power company to produce that electricity. Of course, given the costs of photovoltaic cells, now, it's essentially impossible to be better than the power company in the production of electricity, so you're looking at taking a loss if you try to make money by loading up your roof with solar panels--it's just a way for people who already put panels up to help defray their costs by giving them something to do with the power they produce but don't use.

Yup with current solar panels, the great big thick, inefficient ones that several years to pay themselves off by reducing your power bill. Thats assuming they don't get damaged and other maintenance costs. Now newer ones such as the ones made by nanosolar have a payback of only a few months due to price and effinicy. Plus these ones are far more durable and even puntures won't mess up a whole segment like the old ones.

I now a few people in California and Arizona that has had the old ones for a while now and they have finely paid themselves off and they sell the extra back to the company but its small change. I think the last time I talked to the one in AZ he said makes a whopping 5 dollars from the power company. Granted he has no electrical bill himself.

Hell theres even a golf course in CA that us its own power and has enough energy left over to provide for several nearby houses. So in the long run in places where theres lots of sun then solar cells are a good idea. But in places of little sun and bad weather. Such as Hurricane alley having to replace solar cells every year would not be a good investment.

[Master of Ossus]

Hell theres even a golf course in CA that us its own power and has enough energy left over to provide for several nearby houses. So in the long run in places where theres lots of sun then solar cells are a good idea. But in places of little sun and bad weather. Such as Hurricane alley having to replace solar cells every year would not be a good investment.

If this is true, then why wouldn't the power companies simply set up their own solar cells on their properties? Why would they, instead, rely on their current generating technologies if they are not efficient enough to compete with solar?

[dragon]

Profit margins is the key word. A couple hundred sq meters is enough to povide power for several homes which can be fit on a small area. A power plant on the other hand generates power for tens of thousands of users and businesses. So inoder to provide enough power using solar power companys would need a very large area in which to set them up.

[dragon]

An example of how many is the bill sign by CA governor that will generate 3GW through solar power but will take 1 million solar panels on house tops. So as a power company what would you rather have a single power plants or millions of solar panels.

link

One million more solar panels
Californians are leading the way by supporting renewable energy. This week they passed a bill, signed by Gov. Arnold Schwarzenegger on Monday, that calls for the installation of one million rooftop solar panels on homes, businesses, farms, schools and public buildings by 2018.

The solar systems would generate 3,000 megawatts of power and reduce emissions of greenhouse gases by 3 million tons, equivalent to taking 1 million cars off the state's highways and making California the third biggest solar producer after Japan and Germany.
The California Public Utilities Commission in January approved a $2.9 billion program to help pay for the solar program. The money will come from funds earmarked for solar energy and from gas and electric utility rates. Reuters

[Darth Wong]

Profit margins is the key word. A couple hundred sq meters is enough to povide power for several homes which can be fit on a small area. A power plant on the other hand generates power for tens of thousands of users and businesses. So inoder to provide enough power using solar power companys would need a very large area in which to set them up.

They would also need a contingency plan for overcast days. Making the entire power grid dependent upon solar would be idiotic unless one can somehow devise gigantic and safe batteries.

[dragon]

Profit margins is the key word. A couple hundred sq meters is enough to povide power for several homes which can be fit on a small area. A power plant on the other hand generates power for tens of thousands of users and businesses. So inoder to provide enough power using solar power companys would need a very large area in which to set them up.

They would also need a contingency plan for overcast days. Making the entire power grid dependent upon solar would be idiotic unless one can somehow devise gigantic and safe batteries.

Also a big power plant won't get damaged by minor things like say a hail storm.

What they are needed are orbital power farms with energy beamed down in microwave form. Granted that idea has been floating around since the 60's and they still are no where close to doing it.

[Admiral Valdemar]

They would also need a contingency plan for overcast days. Making the entire power grid dependent upon solar would be idiotic unless one can somehow devise gigantic and safe batteries.

Lithium-ion is the best you'll get for now, though there are nano-technologically enhanced capacitor units being looked at for future applications that use buckminster fullerenes for increased storage.

You could convert the excess to hydrogen, but making hydrogen is very inefficient anyway, not to mention dangerous.

[Sikon]

While 1.37 kW/m^2 of sunlight spread over an area of πr^2 intersects earth, the surface area of this spherical planet is 4πr^2, where r is earth's radius. Overall, average sunlight intensity on the ground is reduced by a factor of 4 aside from the atmosphere, varying with latitude. With all factors including periodic cloud cover, the average location in the U.S. or Europe receives about 20 MJ/m^2 in the average 24-hour period, a bit more than 0.2 kW per square meter.

For example, a 12% efficient solar cell on a roof typically produces about 0.03 kW per square meter, for the average over 24 hours a day and over 365 days a year. Movement to track the sun and/or concentrators would result in more per unit area of solar cells, but that is not the case for most systems. Most news publications misleadingly use advertised rated power, which is based on generation at peak illumination that is a number of times higher.

Around 93% of solar cells produced are between 12% and 15% efficiency, costing on average about $480 per square meter.

More efficient solar cells exist, like 30% efficient solar cells since a decade ago (1994) and 41% efficient solar cells recently, but, so far, the most efficient types have been restricted to less common applications due to much higher cost per unit area. There also exist solar cells costing far less than the typical $500/m^2, but the cheapest existing ones have so low inefficiencies as to be more uneconomical in most generation.

As a result, the photonic crystal can increase the efficiency of solar cells by up to 37 percent, says Peter Bermel, CTO and a cofounder of StarSolar.

A 37% efficiency increase would be like making a 12% efficient solar cell become up to 16% efficient or nominally making a 41% efficient solar cell become up to 56% efficient, except things like this tend to be more complicated. For example, exceptionally efficient multi-layer solar cells already do more to absorb a range of wavelengths than ordinary solar cells, and this resulting in 56% efficiency with broad-spectrum sunlight seems unlikely.

Shawn-Yu Lin, professor of physics at Rensselaer Polytechnic Institute, has developed a method for manufacturing eight-inch disks of photonic crystal--a measurement considerably larger than what can be done with conventional techniques. The method, which employs optical lithography similar to that used in the semiconductor industry, works best for a type of solar cell that concentrates light onto a small chunk of expensive semiconductor material. Such a device would require a relatively small amount of photonic crystal compared with conventional solar cells. Lin says the technique could be applied for more-conventional solar panels, although it would be expensive.

Reading between the lines, it sounds like this is probably substantially more expensive per unit area than regular solar cells, as one would expect from the added manufacturing step, which is the case for other high efficiency techniques so far. With that said, they apparently expect to more than make up for that by using cells with concentrators (reflectors and/or lenses), making the cost per unit area of the solar cell itself not quite as important as it would be otherwise, while mostly keeping the efficiency benefits. That's a possibility, and some past solar arrays have concentrated sunlight onto the solar cells, though they are less common.

This makes it possible to use many times less silicon, he says, cutting costs enough to compete with electricity from the grid in many markets. The savings would be especially large now, since a current shortage in refined silicon is keeping solar-cell prices high and slowing the growth of solar-cell production.

There is no fundamental shortage of silicon, a large portion of most minerals, especially SiO2 sand, but there may be fluctuations in the supply of refined silicon. Silicon refined to high purity is somewhat expensive, in contrast to metallurgical grade silicon metal that costs about $1.70 per kilogram. The semiconducting portion of solar cells can vary between 1 and 300 microns thickness depending upon type, utilizing around 0.002 to 0.7 kilograms of silicon per square meter. While that's not much, what can make the silicon still a significant component of the total production expense for solar cells costing typically ~ $500/m^2 is that refining silicon to far more extreme purity for semiconductor applications costs orders of magnitude more than metallurgical grade.

Research like that described in the article is beneficial, adding to the technological options for varying applications, whether terrestrial or on satellites. But the article's estimate that this technique could change solar cell economics enough to compete well with electricity from the grid is very questionable. It is common for enthusiastic articles of the news media like this to make hidden assumptions.

Let's look at the economics of current terrestrial solar power, with a focus on solar roofs because it is a popular proposal with the public and some governments, as illustrated by the billions of dollars for subsidies in the California Million Solar Roofs Initiative.

At least so far, the economics of such are actually very poor, if factors like capacity factors, discount rates, etc. are taken into account.

As one illustration, a solar roof system with 2000 watts rated power costs around $20000 total. (About $10000 per 1000 watts of rated power is common, such as this other illustration). If one read about the preceding 2 kW system in a news article, it would commonly be described by the rated power, the peak power. However, as described by the company, it produces on average 240 kilowatt-hours monthly, which is only 0.33 kW average power generation, the result of the factors discussed at the start of this post.

For ~ $20000 expense, the average power generation of 330 watts is not a lot. It is not huge compared to even the difference in power consumption between a bunch of typical household incandescent lightbulbs of 60 to 100 watts each versus the compact fluorescent equivalents that are 15 to 25 watts for two to several dollars each.

That solar roof system is marketed as satisfying the electrical needs of a home. Such may be true during times of good sunshine if the residents have an exceptionally low electricity consumption lifestyle. However, the average household uses far more power, not only for purposes like hot water and air conditioning but also indirectly, buying goods and services produced using far greater industrial and commercial electricity usage

The average U.S. household uses 890 kilowatt-hours per month (2001) in their home. Meanwhile, since most electricity use is industrial and commercial rather than residential, total U.S. electricity usage is the equivalent of 3100 kilowatt-hours (2005) per month per household. The ~ 109 million households do not directly utilize that much electricity, but that's the result for the ~ 340 billion kilowatt-hour monthly total U.S. electricity consumption divided by the number of households.

For perspective, at X fraction of $20000 per solar roof system providing 240 kilowatt-hours a month, generating Y fraction of current total U.S. electricity consumption with solar roofs would cost (X * Y) * $24 trillion in capital cost. To provide all energy usage in the manner of a solar-powered hydrogen economy sometimes suggested would require much more solar power than the preceding for electricity generation alone.

The cost difference between eliminating fossil fuels entirely versus eliminating them only in electricity generation is related to how total U.S. energy usage is far greater than electricity usage alone, currently 99 quadrillion Btu and equivalent to 13 quadrillion Btu respectively. Those figures and the ratio between them can change in future scenarios, but one sees the basic idea.

In contrast, for perspective aside from the sociopolitical factors preventing such, about $0.4 trillion in capital cost could make all U.S. electricity generation emissions-free with a net effective cost under $2 per month per household, through the manner described elsewhere; to replace non-electrical energy usage by synthesizing fuel and plastics with zero net CO2 emissions would be around $1.6 trillion extra in power plant capital costs, simultaneously added to by other expenses but also countered by eliminating existing importation and extraction costs.

Here's a description of an existing solar-powered house that includes production of hydrogen by electrolysis for the resident's fuel-cell car:

His energy bill is $0.00
By Jared Flesher, Correspondent of The Christian Science Monitor
Thu Mar 15, 4:00 AM ET

Mike Strizki lives in the nation's first solar-hydrogen house. The technology this civil engineer has been able to string together – solar panels, a hydrogen fuel cell, storage tanks, and a piece of equipment called an electrolyzer – provides electricity to his home year-round, even on the cloudiest of winter days.

Mr. Strizki's monthly utility bill is zero – he's off the power grid – and his system creates no carbon-dioxide emissions. Neither does the fuel-cell car parked in his garage, which runs off the hydrogen his system creates.

It sounds promising, even utopian: homemade, storable energy that doesn't contribute to global warming.

[...]
This is how it works

On sunny days, solar panels on the roof of Strizki's detached garage generate more than enough electricity to power his home. The excess electricity powers a device inside the garage called an electrolyzer, which transforms a tank of water into its base elements – oxygen and hydrogen.

The oxygen is released into the atmosphere, while the hydrogen is stored in 10 1,000-gallon propane tanks on Strizki's property. In the winter, when the solar panels collect less energy than the home needs, that hydrogen is piped to an air-conditioner-size fuel cell, located just outside the garage, which generates electricity.

The final piece of the equation is "The New Jersey Genesis," a hydrogen fuel-cell car Strizki helped design and now maintains for the New Jersey Department of Transportation. He can fill up the Genesis with hydrogen from his electrolyzer and drive it pollution free.
[...]

Source.

The solar power arrays provide free energy or cheap energy if one looks only at continuing cost after installation.

But let's look at the capital cost of the system:

[...]
The solar-hydrogen house took longer to complete than Strizki expected – a strict local zoning officer and the state permitting process caused delays, he says – but in October 2006, the system finally went online. The total cost, $500,000, was paid for in part with a $250,000 grant from the New Jersey Board of Public Utilities.
[...]
Strizki understands that few people can afford to pay hundreds of thousands of dollars for clean energy. Now that he's demonstrated his idea works, his goal is to make the system better and less expensive. (For example, the 10 propane tanks could be replaced by one high-pressure hydrogen tank buried underground.) With mass production, he believes he could get the price of the system, not including the solar panels, down to about $50,000. (A new solar panel system can cost as much as $80,000, Strizki says, but some states, including New Jersey, have offered rebates that cover up to 70 percent of the cost.) Strizki is seeking government grants and private donors for funding, and he's started a company, Renewable Energy International, which he hopes will one day market his product. He says he's already heard from potential customers: "We've been called by some A-list Hollywood types interested in powering their islands."

Source.

So it cost $0.5 million for one household, with residential use and not commercial/industrial use, with him expecting mass-production to drop the $0.5 million to $0.13 million.

Even his estimate of eventually $50,000 for the system aside from the solar panels plus $80,000 for the solar panels is $130,000 per household. The preceding figure is apparently neglecting the cost of the fuel-cell vehicle itself. It is also neglecting indirect electricity usage through purchased products and services, which is substantial as suggested by total U.S. electricity consumption today being 3.5 times that of residential usage alone, with commercial and industrial consumption.

Even without considering the preceding, the $130,000 figure alone is like multiple years salary even for the average person in a relatively rich country like the U.S., let alone the average person worldwide. That's why in the second global warming solutions thread I had illustrated a vastly different technique.

While renewables have low ongoing operations and maintenance expenses, the same is fairly much the case with nuclear power too, and the capital expense can be observed to be much less, relative to average generation. Here's a random illustration, in this case with Finland data:



While renewable energy has obvious environmental benefits, news articles often are misleading about economics in their enthusiasm.

Here's a random example of a common trend of looking at the rated capacity (peak power) rather than properly considering average generation: Doing a search for wind power news articles on the BBC website immediately found one describing 2 GW of wind power as equivalent to two coal-fired power stations. It also gave U.S. wind power generation as 11600 MW. Average generation is a lot less than rated peak generation, approximately varying with the third power of wind speed. Generation of U.S. wind power is better expressed a number of times less, e.g. the 1600 MW average that results from ~ 14.2 billion kilowatt-hours / (365 days * 24 hours in a year) = ~ 1.6 million kilowatts.

Let's look at the situation today:

Share of Total U.S. Electricity Generation by Fuel (2002)
Coal 50.0%
Renewables 3.0%
Hydro 6.9%
Nuclear 20.2%
Gas 17.6%
Oil 2.5%
Source: Energy Information Administration

U.S. Sources of Emission Free Generation (2001)
Wind .7 %
Solar < .1%
Geothermal 1.4%
Hydro 21.6%
Nuclear 76.2%

[...]
U.S. industry believes new nuclear capacity can be built at a capital cost of $1,000-1,200 per kilowatt, which is ...

• Competitive with gas-fired combined cycle plants at $600 per kilowatt with gas delivered at $4-5 per million Btu
• Competitive with new baseload coal-fired capacity

Source.

Here's an illustration for total U.S. energy consumption including electricity, as opposed to electricity alone:



Economic factors have kept solar power a tiny portion of total generation. Corporations would install large solar arrays and become rich selling power from them if doing so was a straightforward gain with existing hardware, but that is far from the case. Organizations install a proportionally small amount of solar power for reasons like environmental concerns and public relations, but the big picture so far is that illustrated previously.

The above isn't meant to be negative on solar power in general, nor other renewable energy, just to provide realistic perspective against some misconceptions that may result from careless enthusiasm. For example, from the preceding perspective, a government measure like the California Million Solar Roofs project is good for the environment in itself, yet it is somewhat unfortunate that the billions of dollars aren't spent in a manner that could provide a number of times greater environmental benefits per dollar with existing technology ... supporting nuclear power.

With that said, the economics of solar power might someday change. Nothing in the laws of physics strictly requires solar arrays to cost the current ~ $500 per square meter at ~ 12 to 15% efficiency.

When I have talked about the potential of future space solar power like giant reflectors, such is for considering a situation with inexpensive space access and a combination of differences. Most importantly, the reflector foil is assumed to be like that for some large solar sail proposals, inexpensive like garbage bags or weather balloons that cost literally orders of magnitude less than $500 per square meter semiconductors. Secondly, the solar intensity in space is a constant 1.37 kW/m^2 at 1 AU from the sun, instead of the approximately 0.2 kW/m^2 less reliable average on earth. Thirdly, upwards of 90% of intercepted sunlight can be reflected and concentrated by aluminized plastic or aluminum foil. And the structure is able to be subject to nearly no stresses and lightweight compared to terrestrial structures. Such potentially influences mass and cost a little like the difference between a 20ft $100,000 building and a 20ft $100 inflatable weather balloon. The net effect of all several factors combined is the potential of orders of magnitude difference in cost per unit of energy between such concentrated sunlight and electricity from solar roofs today.

As for terrestrial solar power, if large reasonably efficient arrays costing vastly less than $500/m^2 were obtained someday, one remaining issue would be the variability of solar power on earth. Currently, solar power is such a small portion of the electricity grid supply that its variability isn't important, with utilities regularly adjusting output from load following power plants like natural gas turbines to deal with far more major fluctuations in electricity supply/demand. But the variability of solar power would lead to new issues in the event that it ever became economical enough to dominate power generation.

Aside from different types of power generation, one method sometimes proposed to deal with electricity demand in the night and other periods of low generation (like cloudy days) is energy storage. Various energy storage technologies have been proposed for dealing with variability in renewable power, like pumped hydro, compressed air, and rechargeable batteries, among other techniques. However, if such had to be implemented, the added cost would further harm the economics of solar power. In the long-term, an extreme leveling technique in event of major usage of variable renewable power generation could be power transmission for thousands of miles instead of the current range of a few hundred miles and less. A concept sometimes proposed is a global network of superconducting cables including transmission from the night side to the sunlit side of the planet, eliminating the effect of local fluctuations.

The land area requirements of solar power is not a problem by itself, with the primary limiting factor rather being the cost of the solar arrays.

The world's total land area of 150 million square kilometers is 15 billion hectares, more than there are people. (A hectare is 10000 m^2, 2.47 acres). Land prices that are up to many millions of dollars per hectare in prime real estate in cities drop to several thousand dollars per hectare or less for the bulk of total land.

For example, at about 12 kWh/day*m^2 solar insolation for sun-tracking arrays in suitable desert locations, 12% or better efficiency leads to ~ 1.4+ kWh / day*m^2, proportionally like ~ 530 trillion kilowatt-hours annually per million square kilometers. For a rough illustration, the world's current electricity usage and total energy usage are equivalent to around 15 trillion kilowatt-hours and around 120 trillion kilowatt-hours respectively. The Sahara desert alone is 9 million square kilometers of barren land used for very little to nothing currently, and most continents have some large desert areas.

However, although land prices are not the problem at a few thousand dollars or less per hectare in suitable areas, today's 12+% efficient solar arrays cost ~ $50,000,000 per hectare, and that is what makes current solar power vastly more expensive than nuclear power.

[Sikon]

I forgot to comment on the following:

This week they passed a bill, signed by Gov. Arnold Schwarzenegger on Monday, that calls for the installation of one million rooftop solar panels on homes, businesses, farms, schools and public buildings by 2018.

The solar systems would generate 3,000 megawatts of power and reduce emissions of greenhouse gases by 3 million tons, equivalent to taking 1 million cars off the state's highways and making California the third biggest solar producer after Japan and Germany.
The California Public Utilities Commission in January approved a $2.9 billion program to help pay for the solar program. The money will come from funds earmarked for solar energy and from gas and electric utility rates.

If one has read all of my preceding post, one might suspect that the 3000 megawatt figure is based on peak generation a number of times greater than average generation. One might also suspect that the $2.9 billion figure is the amount of government expenditures for the incentives to homeowners, while expecting environmentally-concerned home buyers to pay the vast majority of vastly greater total cost at current solar roof prices.

One would be right in both regards.