• Wednesday, September 9, 2009 Latest Update: 10:00AM

Greentech Solar

Nanosolar Boosts Cells Efficiency, Starts Mass Production

The CIGS thin-film solar maker has posted two white papers about its technology and products, and it has named some of its customers who have inked $4.1 billion of contracts.

After staying mum for most of the year, Nanosolar said it has made a leap into mass production and improved its cells' efficiency.

The San Jose, Calif.-based company touted these accomplishments to show that it has made a significant progress since it first announced the start of commercial production in December 2007.

That December announcement has drawn skepticism and even ridicule from competitors and analysts because the company kept refusing to divulge details of its technology, factory capacities or production rate, and it revealed little about customers or projects that would make use of its solar panels.

A Photon magazine article published in January this year criticized Nanosolar for making bold claims but having little to show for since its December 2007 announcement. Photon titled the article, "Nanosolar: No news from the world champion in blowing smoke."

Nanosolar's solar cells make use of a compound of copper, indium, gallium and selenium (CIGS) to convert sunlight into electricity. The materials and its manufacturing process are relatively new to the market, which is dominated by crystalline silicon cells.

The startup, founded in 2002, now said it already began producing its solar cells at its San Jose factory earlier this year. Nanosolar's CEO, Martin Roscheisen, declined to disclose the capacity of the solar cell factory, though back in August 2008 he talked about expanding its manufacturing capability by building a 430 megawatts plant (see Nanosolar Confirms $300M Funding).

The company also said Wednesday that it has completed a 640-megawatt factory to assemble the cells into panels in Luckenwalde, Germany, though this achievement isn't new and is timed to coincide with an inauguration ceremony attended by Germany's minister of environment and other public officials.

Back in February this year, Roscheisen told Greentech Media that the solar panel factory had already been completed and was in operation, though it was running in one-shift mode (see Nanosolar Broke round on 1MW Power Plant, Launched German Panel Factory).

"All tools in our Luckenwalde factory are up and running since the end of last year.  As a result, our latest panels are all being assembled there now as opposed to in our much smaller San Jose panel-assembly line," Roscheisen wrote in an email in February.

Nevertheless, completing the solar panel factory is a big deal for Nanosolar, because its San Jose panel assembly line wasn't fully automated like the one in Luckenwalde. Nanosolar wants to sell its panels to developers of large-scale power plants, so the company has to be able to produce in high volumes and do it cheaply.

Roscheisen declined to divulge the company's production costs. He pointed out in an email that the company has signed $4.1 billion worth of contracts, and named Beck Energy, EDF Energies Nouvelles, AES Solar, Juwi and NextLight Renewable Power among its customers.

That amount of contracts surpassed the roughly $2 billion of contracts announced by fellow CIGS company Solyndra, which broke ground on a 500-megawatt factory less than a mile from its headquarters in Fremont, Calif., last Friday (see Solyndra: Fab 2 Construction Begins).

Nanosolar wouldn't disclose when it is due to deliver on its contracts. The company said it's currently producing its cells and panels at a rate of roughly 1 megawatt per month. The company's panels have received certification from the International Electrotechnical Commission, a validation of the products' performances that is crucial for winning customer acceptance.

Securing bank financing remains a hurdle for the company and its customers, Roscheisen said. The credit crunch has plagued the solar market worldwide over the past year. Convincing banks to finance projects built with newer technologies is a challenge – bankers often prefer equipment by established and larger companies that are more likely to be around to honor its 20-year warranties.

One of its customers, Beck Energy, started building a 1-megawatt project on a former landfill in Luckenwalde last October.  Roscheisen declined to say when the project is scheduled for completion.

Nanosolar has posted two white papers to describe its solar cell technology and compare its panels to those made by First Solar. Tempe, Ariz.-based First Solar makes cadmium-telluride solar panels and is the largest thin-film maker in the world. Both Nanosolar and First Solar make what are commonly called thin films, which use little or no silicon.

The company said Wednesday that its solar cells could convert 16.4 percent of the sunlight that hits them into electricity. The National Renewable Energy Laboratory has confirmed that efficiency, which is higher than figures provided by rival CIGS thin-film makers such as Solyndra, Miasole and HelioVolt (click on company names to see their reported efficiencies).

Engineering efficient solar cells and producing them at low costs is the goal of every solar cell maker. While companies are usually eager to divulge efficiency numbers, they tend to remain mum about their production costs.

The 16.4 percent is the best Nanosolar can produce, though it is generally higher than the average cells that roll out of production lines. The company said the median efficiency of its cells hover between 11 percent and 12 percent.

Nanosolar also claims that when the cells are assembled into a panel, the collective efficiency remains essentially the same. That's a different than the manufacturing processes for the silicon and other types of panels could deliver today. Typically, a solar panel's efficiency is lower than that of its individual cells because the cells don't all perform uniformly when they are put together.

Nanosolar said it sorts cells before assembling them to avoid mixing poor performing cells with the good ones, and that helps to achieve the same cell/panel efficiency.

First Solar said its panels on average have 10.9 percent efficiency. First Solar also is known for being able to make panels cheaply, and has been happy to disclose that its manufacturing cost is about 87 cents per watt.

The company makes its cells by printing CIGS onto an aluminum foil in what's commonly called a roll-to-roll process. The technology can produce cells more quickly than some of the competing methods, the company said.

The aluminum, which isn't expensive, is conductive and serves as the bottom layer of electrode to help ferrying electricity produced by the cells, the company said (see Nansolar's white paper on its cell technology).

In its white paper about its panels, Nanosolar said it encases its cells with two pieces of tempered glass. This approach makes for a more durable panel and larger-sized panels compared with First Solar's panels, which uses tempered glass for the back.

Nanosolar said its manufacturing process allows it to easily make panels of different sizes and power ratings, which range from 160 watts to 220 watts.

The company is introducing panels that are 2 meters in length and 1 meter wide, Roscheisen wrote in an email. First Solar's panels are 1.2 meters by 0.6 meters, with power ratings ranging from 70 watts to 77.5 watts.

Image of Nanosolar CIGS cells via the company.

Comments [20]

  • StevePluvia 09/9/09 12:44 PM

    Comments on Nanosolar PR:

    1.  Nanosolar is in pilot” production at 1mw/month.  I question whether they know their production capacity if they’re producing at such a tiny rate given the 2yrs its been since they reportedly went into (ahem) “commercial” production. 

    2.  The $4 billion backlog is pure BS.  Nobody waits for unproven product when they can get cheap product that’s bankable elsewhere.  At 1mw/month Nanosolar has no demonstrated ability to produce enough product to fill $4b of orders (that’s like 20yrs of backlog at current run rate).  For a company in search of credibility, spewing ~optisolar~ type backlogs is the wrong move.

    3.  It sounds like nanosolar moved to a glass-glass samich to resolve the flex encapsulation problem. I suspect that’s what cost them the last 2yrs of delay.

    Reply
  • russ 09/10/09 5:19 AM

    @StevePluvia - Very correct on the backlog being vaporware - what banker is going to loan even one thin dime on something which is yet to be produced - let alone proven!

    These guys have 4 billion USD of foolish bankers lined up? No chance!

    Reply
      • StevePluvia 09/10/09 11:28 AM

        @russ—see the Nanosolar panel connector system; will it work with microinverters?  Or any of the new inverter technologies?  Makes me wonder if the Nanaosolar connector design will be obsolete before product hits the market? 

        Nanosolar connector design is found on linked whitepaper:  http://www.nanosolar.com/sites/default/files/NanosolarUtilityPanelWhitePaper.pdf

      • StevePluvia 09/10/09 11:37 AM

        @russ testing those flimsy Nanosolar cells for sorting/matching before module assembling must be a ~nightmare~; then there’s the task of handling that flimsy lil thing for assembly…  No wonder they’re at 1mw/month after 2yrs….

  • anon 09/14/09 4:23 PM

    @ecd: learn to read.  16% efficiency is representative of their best cells.  you need many cells to make a panel.  it appears the company can make many 11% production cells to put into commercial panels and some 16% cells to demonstrate the potential of their technology.

    Reply
  • ECD Fan 09/17/09 1:56 PM

    Hey anon:  My best cells are 100% efficient, unfortunately I don’t have many of those to offer to you (since they are the best, I keep them for myself).  I can offer you 1%-efficient cells at $2.60 per Watt, hopefully they will work for your projects. 

    Is Nanosolar selling 11%-efficient modules or 16%-efficient modules?  If the modules are 11% efficient, why are they bullshitting about 16%-cells.  Nanosolar customers won’t be offered cells, they will be offered modules.  And, unfortunately for Nanosolar, First Solar’s FS-280 is more efficient that their module, yet it costs about 90c to make.

    Reply
      • StevePluvia 09/17/09 3:07 PM

        ECD, the term “champion cell” is fairly standard verbiage in the PV industry.  It means essentially under the best conditions with current equipment they can produce 16% efficiency.  The best conditions for the best cell efficiency are typically not the best conditions for commercial production (ok never).  Thus pv mfgrs find a balance where their production line can produce volume at a consistent efficiency.

        Nanaosolar has stated their champion cells produce at 16% and their average commercial cells will be about 11%.  If you read their white papers you’ll find the efficiency across the roll differs.  Therefore, they cut cells from the roll, then sort them into common bins so the assembled modues have clls performing at very close levels.  This is frankly very clever.  It allows Nanosolar to utilize the entire roll width and produce some 9%, some 10, some 11 and some 12% modules. 

        If Nanosolar can crank up their current commercial line to somewhere around 350mw or more, I suspect they’ll be pretty cost competitive.  350mw/yr is roughly 1mw/day—right now they’re at 1mw/month….

  • ECD Fan 09/18/09 6:23 AM

    Steve:  All I am saying is that NREL made a big mistake years ago when they encouraged the industry to use that fake “champion-cell” metrics (the famous chart is cited all over the place).  For example, Unisolar claimed 12.2%-efficiency for their-triple junction laminates in 1985.  24 years later, Unisolar’s triple-junction modules (PVL-136), which constitute their main line, are 6.3%-efficient:

    http://ecdfan.blogspot.com/2009/06/progress-unisolar-way.html

    Oh, by the way, apparently Unisolar does a lot of “sorting” as well.  I have “theorized” that Unisolar employs “cherry-picking” to create their PVL-144 product (6.7%-efficient):

    http://ecdfan.blogspot.com/2009/06/unisolar-efficiencies-mystery-of-pvl.html

    The question I have about Nanosolar is taking them so long?  If they truly had a grasp of their processes, by now they should have been running a GW capacity.  First Solar, despite all my misgivings about them, at least knows how to execute and scale.

    Reply
  • StevePluvia 09/18/09 6:59 AM

    ECD, my personal theory is Nanosolar had an encapsulation problem and had to revert to glass/glass to solve the problem which required them to change their business plan and production line, resulting in the delay.  That said, Nanosolar has accomplished some very impressive and very, very difficult milestones to date.  My guess—they will manage to ramp to commercial scale within the time ECD goes belly up.  BTW you never responded re FSLR’s Tuscon system performance.  What’s their degradation problem?

    Reply
      • Thom 09/18/09 12:39 PM

        StevePluvia, what is this FSLR degradation problem?  Do you have a link?

      • StevePluvia 09/18/09 1:47 PM

        Thom, I don’t think there is a FSLR degradation problem, (ECD made that claim) all the info I’ve seen shows above avg LID performance from FSLR modules.  It sounded to me like ECD was just unfamiliar with the phenomena of LID.

      • ECD Fan 09/19/09 4:43 PM

        To Steve & Thom:  I suggest you copy and paste the text from the Tucson pdf files into an Excel spreadsheet and calculate generated kWh per solar intensity for each year and then calculate the annual degradation.  You are interested in the four observation sets SGS-125TF-n FS/XN 134.4 KW Cd-Tl (where “n” is 1 to 4).  You will see what I mean.

        The Tucson files are available here:  http://www.tep.com/Green/Business/Solar/progress.asp
        (the ones under the “Summary of GreenWatts Program” heading):

        http://www.tep.com/Green/GreenWatts/docs/Renewable2008Sept.pdf
        http://www.tep.com/Green/GreenWatts/docs/Renewable2007Dec.pdf
        http://www.tep.com/Green/GreenWatts/docs/Renewable2006Dec.pdf
        etc.

        Here is what I do exactly:  Take the annual KWH generation for each line, average it across the 4 sets, divide by 134.4 to get to KWH/KW (which comes to 1722 for year 2004) and then divide it by the annual solar intensity (which is 2044 for year 2004), which results in a “performance ratio” (84% for 2004).  By year 2007, that performance ratio has dropped to 77%, which is severe degradation in my book.

        You will also see that the ASE’s performance ratio beats First Solar (by 3% in 2006, for example).  BP’s panels are the worst and severely underperform both ASE and FSLR (no news here).

        If anybody is interested in the Unisolar results from the Tucson study, I have them on my blog (they contradict ECD’s management claims, of course):

        http://ecdfan.blogspot.com/2009/05/kwhkw-myth-part-i.html

      • StevePluvia 09/20/09 1:24 PM

        ECD you have a simple math error in your calculation; you can’t compare year-to-year data (relative to annual solar hours) and conclude the difference is “degradation”  If one year has less sunny days than the avg solar hours, your calculation will show less kwh produced that year.  The lower KWH production would have been from weather (fewer sun days than avg) not LID (light induced degradation)....  Also, are you figuring the line loss from inverter a cabling?

      • ECD Fan 09/20/09 2:02 PM

        To Steve: No, Steve, I have NOT made that particular math error (after all, on my blog I have analyzed thin-film degradation data, mostly focusing on Unisolar, from numerous sources from many countries across three continents).  I have specifically adjusted for the ACTUAL solar irradiation (called “intensity” in the files).  I urge you to actually do the math, and then discuss.  Regarding the cabling - unless you can convince me that the desert rats are aggressively chewing on First Solar’s cabling, I am pretty sure the cabling has no effect on the DEGRADATION over time, but of course, has an effect on the nominal value of the performance ratio at any specific point of time.  I cannot say what is the source of the annual degradation in this particular study - it can be the wiring, the inverters, the modules, or even a faulty pyranometer (other studies have been much more careful in separating those, unfortunately, I haven’t seen one that specifically analyzes First Solar data).  But the inverters attached to the ASE modules are the same as the ones attached to the First Solar modules, and I suspect the wiring is the same.  And I don’t care for the LID term - the long-term degradation is not necessarily light-induced - it could also be from delamination, microfractures from the thermal stresses, etc, etc - I don’t care, because even the LID mechanism is poorly understood - all I care about is the annual KWH/KW, adjusted for the actual solar irradiation.

      • StevePluvia 09/20/09 2:13 PM

        ECD—Sorry, I see you used measured solar vs avg; I also see the decreased production you described.  Interesting stuff. Let me look at this a little more—thx for the links.

  • anon 09/19/09 7:44 AM

    @ECD: the crucial difference between your 100% cells and Nanosolar’s 16% cells is that your’s are not certified by NREL.  Whether or not the efficiency of a “champion cell” has any significance is a separate issue from insulting GTM’s journalists and/or Nanosolar’s integrity. 

    Anyway, I like how you shamelessly promote links to your blog next to outrageous comments about lies and bullshit.  Good strategy.

    Reply
  • ECD Fan 09/19/09 4:50 PM

    anon:  And I claim that NREL is doing disservice to the PV customers by allowing these “champion cells” data points to be used in promotional materials and in planted articles, as I have demonstrated by my “Unisolar progress” example, where even after 24 years, the actual module efficiency is still about 1/2 of the champion cell efficiency advertised in 1985.

    I don’t find anything shameless in promoting to the truth.  Do you?  The people who have suppressed the truth for over 24 years (and continue to do so) should be ashamed, not me.

    Reply
      • Eric Wesoff 09/19/09 8:23 PM

        Dear ECD Fan,
        When it comes to ECD - how are you defining “fan”
        Eric

  • ECD Fan 09/20/09 5:21 AM

    To Eric: That exercise is left to the readers of my blog.  Merriam-Webster, of course, defines it as “an enthusiastic devotee,” “fanatic,” and even as “an instrument for producing a current of air.”

    Reply
.