About the SA Blog Network



Opinion, arguments & analyses from the editors of Scientific American
Observations HomeAboutContact

Sunshine is free, so can photovoltaics be cheap?

The views expressed are those of the author and are not necessarily those of Scientific American.

Email   PrintPrint

Here’s how to make a solar cell from silicon: take one solid block of doped silicon, saw it into thin wafers, layer said semiconductors beneath a panel of transparent glass, connect them to a metal electrode that can channel away the electrons knocked loose by incoming photons and turn it into a photovoltaic device. That process has at least two flaws: such silicon is expensive, contributing more than half to the final price of a solar photovoltaic, and sawing it turns as much as half of that silicon into wasted grit.*

As a result, solar costs as much as $4 per watt by the time it’s installed on your roof or in a large-scale power plant, says Arun Majumdar, the first director of the new Advanced Research Projects Agency for Energy, or ARPA-E. "If you can reduce that to $1.50 per watt it can enable scaling," or widespread adoption of the clean, renewable electricity source, he told at last week’s ARPA-E summit.

And a company called 1366 Technologies may have found a way to do just that by growing a nearly pure wafer directly from melted silicon rather than forming an ingot that is then sawed.


The idea could cut wafer costs by as much as 80 percent, says Craig Lund, director of business development for the company that takes its name from the solar constant—the watts of solar radiation that hit each square meter of the atmosphere. From 1366′s furnace in Lexington, Mass., the company makes an "as-created fully functioning solar cell." In fact, silicon wafers currently contribute more than half of the high price of sil

That may make silicon photovoltaics, which are the most efficient currently at turning sunlight into electricity, as cheap as thin-film solar cells, whose advantage is cost but which are not as good at creating electric current. In fact, rapidly decreasing cost for solar power means some experts expect such distributed electricity generation to cost the same or less than electricity from today’s grid by as soon as 2015.

After all, "more solar energy reaches Earth in every year than the combined total of all the energy in all the fossil fuels still in the ground," noted Chris Rivest, co-founder of SunPrint, whose technology prints thin film solar cells made from cadmium-telluride, at the ARPA-E summit. "By subsidizing corn, we made it a national priority to make food cheap. Why not do that with silicon? Slightly cheaper glass may not be sexy today but when solar uses more glass than construction you can bet it will be."

Editor’s Note: David Biello is the host of a forthcoming series on PBS, tentatively titled "The Future of Electricity." The series will explore how transformation is coming to how we use and produce electricity, impacting the environment, national security and the economy. Detroit Public Television provided the video outtake embedded in this post.

 * Clarification (3/11/10): This sentences has been changed to clarify the expense of silicon wafers in regard to photovoltaics.

Rights & Permissions

Comments 16 Comments

Add Comment
  1. 1. JamesDavis 1:39 pm 03/10/2010

    So, what is he waiting for? Combine David Biello’s liquid metal battery and his solar cell together and get them to market and into homes and businesses.

    Link to this
  2. 2. sethdayal 2:10 pm 03/10/2010

    A mass produced no tech single pane skylight at Home Depot costs $200 a sq meter.

    A 120 watt solar panel is really nothing more than a sq meter of solar cells worth maybe $120 today glued to that skylite.

    If the skylite leaks the solar panel dies just like the mess it leaves on your ceiling.

    Until super efficient solar cells like nuclear fusion get down to a reasonable price, the current ones even free glued to that skylight will keep prices above $2 a watt or over 30 cents a kwh average installed.

    The current fire sale prices on solar panels are a result of the collapse of the Spanish solar market, it won’t last.

    Link to this
  3. 3. fyngyrz 2:14 pm 03/10/2010

    I am *so* tired of these announcements that NEVER turn into commercially available products. We’ve been promised cheap, efficient solar cells forever. Printed ones, thin ones, nano-tubed ones, coated ones, doped ones, reflector-trough embedded ones… none of it ever reaches my doorstep. And I live in a very high sun area (Montana plains) where payback (sunny hours) is high. You want to impress? Commercialize. Otherwise, you’re just like the other guys – full of hot air.

    Link to this
  4. 4. Soccerdad 3:27 pm 03/10/2010

    The day of solar will arrive when people no longer need a government subsidy – either on the installed cost or on the sell back of power – in order to make it feasible. Until then the subsidy simply perpetuates the current uneconomical technology. And the subsidies make investment in solar plants more risky as the manufacturer is reliant upon government as his partner. A more unreliable partner cannot be found, as demonstrated by the failed Spanish solar experiment.

    Link to this
  5. 5. doug l 3:46 pm 03/10/2010

    Interesting news and potentially very profitable news. I appreciate ARPA-E most when the research it supports is simultaneously hooked-in to the commercial and industrial research end of things, and all the better if it helps the US to maintain its manufacturing capacity of strategic materials such as these kinds of wafers and the process from which they are derived.

    Link to this
  6. 6. InquiringConstructivist 4:08 pm 03/10/2010

    I’m not sure David Biello should be the one to report on such technologies from the ARPA-E conference, as in the accompanying video. He seems completely unaware that two companies (Schott and Evergreen) in Massachusetts have already been making wafers from molten silicon for almost a decade. If he did, he would have been smart enough to ask what the difference in process is between 1366 and the string ribbon people.

    Link to this
  7. 7. jtdwyer 12:15 am 03/11/2010

    If this method does work and is applicable to computer chips, the price of memory and other chips would decline similarly. Of course it would also decimate the world’s silicone chip manufacturers who have $millions in manufacturing equipment on the books as well as R&D investments, so be prepared for possible stock market impact, if it works for computer chips…

    Link to this
  8. 8. jtdwyer 12:18 am 03/11/2010

    InquiringConstructivist – You might want to also inform the ARPA-E folks, since that also appear to be completely unaware.

    Link to this
  9. 9. JamesDavis 7:58 am 03/11/2010

    Do any of you people who faithfully read SciAm’s articles believe that the scientists and governments who contribute to the articles do not read any of SciAm’s articles? If they did, they would be better connected to what the other one is doing. If any of them do read the articles, then they are laughing at us for being foolish enough in believing that they are actually trying to clean up our environment by providing to the people technology that could do that in half a decade. All of this technology that is spoke of, we faithful readers know that some of it is over a hundred years old. Why are they not producing that proven technology and getting it out to the people and tweeking it to profection like they did the fossil fuel car? It seems really odd to me that we have all this great and wonderous technology that none of us can use. Do any of you faithful readers feel the same way?

    Link to this
  10. 10. candide 3:27 pm 03/11/2010

    The "fossil fuel car" has been tweaked for almost 100 years, the internal combustion engine goes back ebven longer. It has not been quite as fast as you seem to imply.

    " It seems really odd to me that we have all this great and wonderous technology that none of us can use"

    None of us? Quite an exaggeration.

    Germany has the highest percentage of photovoltaic installations currently in use. The USA has many in use, though a low percentage. Denmark also has a lot of photovoltaics (in addition to other clean energy) and is well on its way to being energy independent.

    Solar is still on the beginning part or Moore’s curve – prices have been steadily dropping while output and efficiency have been increasing. We have not yet reached the magic crossover.

    Perhaps we would have proceeded faster if politics had not been involved – like removing the solar panels from the White House (symbolic – both ways) and if more research dollars had been invested.

    Many other countries are moving ahead with solar development, particularly China. They, obviously, think it is worth it and will lead to a positive outcome.

    Link to this
  11. 11. jtdwyer 6:58 pm 03/11/2010

    candide – Moore’s law applied to the miniaturization of electronics components, specifically the number of functional transistor circuits that can be implemented on a circuit board. It is based on the fact that as circuit size decreases, processing speed increases. I may be wrong, but I don’t think it can be extended to solar panels just because they both are commonly implemented on a silicon semiconductor substrate. If so, solar panels have been on the ‘beginning part of the curve’ for an awful long time compared to computer processors.

    Link to this
  12. 12. InquiringConstructivist 3:49 pm 03/13/2010

    Here’s an article from 1997 about making poly wafers directly from melt, specifically Evergreen Solar’s process:
    Solar Energy Materials and Solar Cells
    Volume 48, Issues 1-4, November 1997, Pages 179-186
    R. L. Wallacea, , J. I. Hanokaa, A. Rohatgib and G. Crottyb
    "Thin silicon string ribbon"
    Again, deceptive of 1366 to allow David Biello and Sci Am to deceive themselves into thinking that skipping the ingot-cutting is a new thing.
    I’ve toured a Schott plant, and cell efficiencies are 22-26% (not to be confused with module efficiencies).

    Link to this
  13. 13. Michael Hanlon 12:20 am 03/15/2010

    Here’s a dumb idea. Sputter coat silicon onto ferrous atoms under a magnetic field’s influence. That should line up the crystal structure so that two plane pressure causes the crystals to crack along the plane exposing the ferrous atoms. Sputter copper conducter onto the exposed ferrous side of the constructs and a voila, three step cheap disposable, er, recycleable solar cells roll off the production line. If the magnetic fields are applied right, even curved cells could be made. Send royalties to …

    Link to this
  14. 14. landdev 12:14 pm 03/25/2010

    By the middle of this century electrical transmission lines will have gone the way of the dinosaur.

    Ben Koshkin

    Link to this
  15. 15. jtdwyer 4:28 pm 03/26/2010

    landdev – I thought that was a quote from Nikola Tesla…

    Link to this
  16. 16. DavidHuieGreen 2:44 pm 04/23/2011

    “If any of them do read the articles, then they are laughing at us for being foolish enough in believing that they are actually trying to clean up our environment by providing to the people technology that could do that in half a decade. All of this technology that is spoke of, we faithful readers know that some of it is over a hundred years old. Why are they not producing that proven technology and getting it out to the people and tweaking it to perfection like they did the fossil fuel car? It seems really odd to me that we have all this great and wondrous technology that none of us can use. Do any of you faithful readers feel the same way? ”

    Sometimes I do but then I remember most of these articles only mention the potential and the promise. They are not big on discussing the problems with perfection and production.The fossil fuel car is not perfect but it is much better than some earlier versions. Even there we can look at Henry Ford who decided against building a perfect can and on building an affordable one. With millions of people using them and millions of people pointing out their flaws, improvements came hot and heavy.Some of these technologies have so many problems they can’t compete on the initial start-up. It is possible they would out-strip the others if enough were working on them, but they are just too far behind right now.

    One thing which bothers me to contemplate is that the problems which held back a given technology 40 years ago may have been solved elsewhere but the reasons for abandoning a line of inquiry may have been forgotten–all that is remembered is that there WAS a problem which stopped them from becoming feasible, not what it was and that it has been solved.

    There are more paths to take than there are people to walk them. As a result, many of these promising technologies are abandoned until the right person takes up the challenge.

    One of the many things I like about Scientific American is that it reminds us of articles it carried years ago. Maybe some of the alternative solutions will be taken up by people who see solutions.

    It would make an interesting industry to research abandoned concepts, recording why they were abandoned and then solutions which might revive them. We have hundreds of millions of people who might have the answer if they just had a share in the rewards–either praise or money.

    Link to this

Add a Comment
You must sign in or register as a member to submit a comment.

More from Scientific American

Email this Article