• Monday, June 22, 2009 Latest Update: 9:16AM

Greentech Solar

Whither Thin Film?

Something is happening in Europe today that could put a damper on thin film’s meteoric rise: banks and project lenders are refusing to finance projects using thin-film modules, says GTM Research's Senior Solar Analyst Daniel Englander.

Thin-film PV is the game changer, right? By nearly all relevant metrics – cost, performance, and operational behavior – it does just as well as, and sometimes better than, traditional crystalline silicon PV. It's cheaper and simpler to produce, attains higher output in both low light and high temperature conditions, and – for some technologies – boasts conversion efficiencies to rival some multicrystalline modules. Over the past few years, thin-film PV has shown it can play in almost all markets – from residential rooftops in Marin County to utility-scale projects in Bavaria to subway stations in Brooklyn. GTM Research even forecasted that thin film would comprise roughly 50 percent of the incremental demand market in 2012.

But something is happening in Europe today that could put a damper on thin film's meteoric rise: banks and project lenders are refusing to finance projects using thin-film modules. At a conference I recently attended and spoke at in London, I heard from a number of Spanish, Italian, French and German bankers who are routinely passing on thin-film projects. "Why go to thin film when we have plenty of polysilicon applications and we're making a killing?" one asked, rhetorically. "Our [technical due diligence] engineers have very exacting standards," said another. "The banking market is shut for new technologies," said a third – openly taunting the handful of thin-film execs in the room.

The credit crisis is driving bankers' hesitancy to finance thin-film projects. Capital today is both scarce and expensive, forcing bankers to pass on projects they might have financed only nine months ago. Today only low risk, "gold plated" projects are receiving financing. These are projects with quality sponsors, experienced developers and EPC firms that use "bankable" technology. In other words, the projects that receive capital are those with the lowest risk profile. Technology components comprise roughly 85 percent of a project's cost, and the module comprises roughly 50 percent of that cost center, which means that the module represents the largest cost risk component of a PV project.

One of the biggest issues in technology selection is durability, and the biggest aspect of durability is performance degradation. As Spire CEO Roger Little so ably instructed me in front of 200 people at a conference a few months back, thin-film modules have a tendency to degrade more and at a faster rate than c-Si modules. Whether this is true or not, it is certainly playing into the risk calculations of the European banks. At a higher level, operational experience with thin film is much lower than with c-Si, meaning there are more long-term questions that have yet to be answered by thin film in the field. By contrast, some c-Si projects built in the mid-1980s are still operating within their expected performance range.

Combining questions of risk with scarce, expensive capital means that – in the words of one Italian project developer – "banks are the technology choosers and they are continually exerting their desire to control the technology." In Europe this is resulting in banks passing on thin-film projects in favor of relatively less-risky c-Si projects.

That's not to say that all is lost for thin film. To the contrary, in the American PV market, thin-film projects are popping up like mushrooms after a spring rain. The U.S. downstream PV market is much more price sensitive than the European market where project revenues are set by feed-in tariffs in excess of $0.40/kWh. By contrast, commercial power purchase agreements in the U.S. might sell electricity for half that price, though they still need to reach internal rates of return found in Europe. Today those are in the range of 11 percent to 15 percent.

Developers are shaving capital costs to hit this mark and are routinely turning to thin-film modules from companies like First Solar to achieve upfront cost goals. That utilities and state and local governments are the U.S. market's two dominant counterparties today only improves the risk profile of commercial projects. The same is true for the dominant developers and EPCs in the U.S. market, many of whom have decades of operational experience in power generation, construction, and operations and maintenance.

There is another a possible explanation for European reticence. A rumor has made the rounds in the last few months that European banks are only financing European modules. This situation is most pronounced in Germany where many c-Si module manufacturers are struggling to support high cost structures in today's low price environment. Germany's powerhouse semi-state-owned banks, like LBBW and KfW-Ibex, may be giving preferential treatment to domestic modules in a bid of support. Or perhaps they're just suckers for quality German engineering. Either way, this means thin-film projects will continue to have a rough go of it in Europe for many months to come. 

Comments [31]

  • ECD Fan 06/22/09 2:13 PM

    It is true that “thin-film modules have a tendency to degrade more and at a faster rate than c-Si modules,” just look at the warranties of thin-film vs c-Si modules. Unisolar, of course, is one of the worst offenders. NREL has documented 1.22% average annual linear degradation for its current triple junction product (see page 9 at http://www1.eere.energy.gov/solar/review_meeting/pdfs/exp_5_osterwald_nrel.pdf ), meaning that even Unisolar’s puny 20-year, 80%-rated-power warranty is not likely to hold.  This European performance study may also be of interest:  http://ecdfan.blogspot.com/2009/05/kwhkw-myth-part-iv.html
    At least First Solar has a 25-year warranty.

    There are several European thin-film producers, of course. Sunfilm AG comes to mind, for example.

    Reply
  • ECD Fan 06/22/09 2:14 PM

    And Flexcell, of course, a division of Qcells.

    Reply
  • JoeJoe 06/22/09 2:18 PM

    If German banks have a risk aversion to thin-film, why not go with high quality Chinese c-Si? Your last paragraph addresses this tangentially but your title throws a mixed message. Is this strictly a technology issue or are German banks looking out for German producers? I’d suspect the latter… Have for a while now…

    What’s the word on the pace of installations in Germany? Will they fall above or below their 1 - 1.5GW corridor?

    Reply
  • Steve Pluvia 06/22/09 2:54 PM

    EU buyers favoring domestic producers?  Udderly shocking, unless you’d done a shred of business in any EU country, in which case you’d know they always do business this way.  According to Danny boy I guess First Solar will show zero sales in the EU this quarter.  Umm yea.

    Reply
  • Solvida 06/23/09 1:03 PM

    A weekend long experiment I participated in during 1999 revealed that an equally nameplate rated brand new UniSolar module<> DC Battery system accrued more amp hours per rated watt than Siemens and AstroPower equivalent modules. I hypothesized that UniSolar beefed something up in their module output in order to compensate for accelerated degradation over time. Got an A-  and celebrated at the pub. I miss college

    Reply
      • Steve Pluvia 06/23/09 1:29 PM

        Unisolar is hardly a thin film producer.  They really just operate a pilot line producing insignificant volume.  Given that every inch they produce is mfg’d then sold at a loss, I doubt any developer or bank would green light their product for commercial use.

  • Solvida 06/23/09 1:57 PM

    85000 modules on 1 rooftop in Spain means somebody has been convinced of the commercial viability of USO product. That being said, it was the “gold rush” in Spain and perhaps pre contract diligence was less than adequate.  Regardless, someone opened up the vaults to fund this almost 12mWp rooftop project.

    Reply
  • ECD Fan 06/23/09 5:09 PM

    Solvida:  That weekend-long experiment was rigged.  As you know, Unisolar’s laminate outperform their rated power in the first few hours and weeks of exposure to sun, while they are rapidly degrading, after which they settle to their rated power under STC for a few years, starting to degrade noticeably again after the 6-7th year, in likely violation of their puny 20-year, 80% rated power (based on NREL data).

    That rooftop (lent to investors by GM/Opel) in Spain may have been simply a vehicle for revenue double-counting.  The details are at the bottom in this post:

    http://ecdfan.blogspot.com/2009/05/agts-mistake-and-possible-double.html

    Reply
      • rooferguy 06/23/09 10:38 PM

        UniSolar is great for school rooftops.  It’s almost indestructible.  Kids can skateboard on it!  Of course the output isn’t what its cracked up to be, but kids and rocks can crack it!  But a real pain in the ass to install—very, very labor intensive (“peel and stick….not”).

  • Steve Pluvia 06/23/09 10:04 PM

    ECD Fan, nice work on your site.  If you’re still looking for fully installed costs from FSLR, look at comments from Southern California Edison in the June 18th 2009 pr:

    “SCE said the cost for the PV panels it owns is estimated to be $3.50 per installed watt, which is about half the cost to install PV panels on homeowners roofs in the California solar incentives program. That’s because of economies of scale.”  All of the FSLR commercial projects in the U.S from this point forward will be installed for below $3.50/watt.

    Reply
      • rooferguy 06/23/09 10:50 PM

        SCE’s first installations were done at a price closer to $4.75/watt (per their presentation at SP 2008).  They will get down to $3.50 and below.  Costs are much less than for residential for three reasons:
        1. FirstSolar panels are cheaper than crystalline
        2. Economies of scale
        3. No utility bureaucracy since the utility itself is the “installer.”  No incentive paperwork, interconnection, inspections, overhead, sales and other indirect labor.  These costs amount to $1.50-2.00/w for smaller systems.

  • Sunfan 06/24/09 4:13 AM

    One reason for the EU-Banks decision could be, if TF module manufacturer are not cost competitive against CSi modules,  Today, how can they survive midterm – only First Solar has a real chance.
    The risk for an investor is to by a cheap TF module plant today, but what does this help
    if the manufacturer disappears due to losses he will make?  What will happen if
    the modules fail later (warranty issue)? So i think the most European anks are aware about this risk!

    Reply
  • ECD Fan 06/24/09 7:23 AM

    rooferguy:  Which SCE “presentation at SP 2008” are you referring to?  I am aware of the initial application to the CPUC (and subsequent docket filings), but I wasn’t aware of a separate presentation ...

    All-in system costs for large commercial rooftop projects for Unisolar appear to be over $6 per Watt (with possible exception of one SunEdison installation, at $5 per Watt)

    http://ecdfan.blogspot.com/2009/05/is-350-per-watt-all-in-reality.html

    Reply
  • Larry Kelley 06/24/09 1:08 PM

    Hello Everyone…..especially ECD Fan.
    ECD Fan, I like your blog. Very few people will take the time you do, to chases down all the information available about ECD.  Having been in Michigan for 64 years and having 1st hand knowledge of Stanford O, I have always been extremely dubious of their claims.  Since I am a high temperature furnace designer and crystal grower, I have always laughed at their method of high vacuum production. 300’ long vacuum chambers…..LOL LOL LOL.
    The guy running their company is as close to an idiot as I’ve ever seen. Only technically,,,,,,he took away about $2,000,000 from ECD last year.  He is extremely well politically connected also. Michigan’s Gov Granholm treats ECD like it’s a jewel in Michigan’s crown. I have a company near Muskegon Mi, and I make Ruby, Sapphire, Garnet, Emerald, Cubic Zirconia crystals and the equipment they are made in. http://www.shelbygemfactory.com  (look at the ‘factory pictures’ link).  I have developed and manufactured a new way to make multi-crystalline Si PV cells. With no slicing or grinding and no stupid Si ‘ribbon’ growth ( Evergreen.com), I can make PV cells 0.002”-0.005” thick. The cells have been tested by MIT, Arthur D Little and Miami Research Labs, and have been found to be ‘remarkable’ and ‘easy to manufacture’.  Using approximately 1/15 the Si and 1/15 the electricity, and about 1/10th the labor, along with an approximate cost of equipment (CAPEX) of about $250k/MWe, I think we will be making the world’s least expensive Si PV cells as soon as our factory gets finished. The projected manufacturing costs for a finished solar panel with anti-reflective glass and an aluminum rear support and blocking diodes is conservatively projected to be about US$ 0.70/watt. (Of course, if I had a salary of $2,000,000/year, the cost/watt would be a lot higher.) Sincerely,  Larry P Kelley   (JavaScript must be enabled to view this email address)

    Reply
      • Jim McGrath 06/24/09 3:17 PM

        Mr. Kelley.  I’m fascinated with your new PV product.  What efficiency level are you seeing with it?  Better than a-SI I expect.  I’m in Chicago area and my brother also owns a business in the Muskegon area.  I would love to come see your place sometime.  Best Regards,  Jim McGrath ((JavaScript must be enabled to view this email address))

  • Foontok 06/24/09 8:31 PM

    European banks favoring European manufacturers?  How unfair!
    Here in the US, our motto is, “Buy European.”  Or something like that.

    Reply
  • russ 06/25/09 3:54 AM

    European business has normally been a group process - many get a slice of the pie which keeps costs high and everyone alive if not fat.

    Can’t blame banks on avoiding startups as the failure rate has to be well above 95%. New processes are not even a dime a dozen - and normally worth just about as much. The rare ‘new’ process actually makes it to the commercial stage.

    Reply
  • Rational Observer 06/26/09 12:58 PM

    Thin film PV is on shaky ground for a number of reasons - reliability, rarity of the materials used, toxicity, marginal (if any) cost advantage, low throughput, etc.  It only made recent gains because of the temporary Si feedstock shortage.  It is good to see that Europe is being rational and focusing on PV technology that can carry through for the long term.  One can only hope that the US will be equally diligent in it’s future investments in PV technology and not be swayed by the excessive thin-film hype out there.

    Reply
  • Kerstin Ladislav 06/28/09 8:27 AM

    FYI!
    Liebe Grüsse
    HG

    Reply
  • TF expert 06/30/09 7:15 PM

    Probably all less-competitive a-Si and newer CIS. FSLR probably still selling fine. Not all thin films are alike, tho their competitors and wave making stock touts might wish to think so.

    Reply
      • TF expert 06/30/09 7:17 PM

        and how ironic - REC has to recall all its modules for faults; and BP has a warehouse roof fire. O sure, bankers know best.

  • Mekhong Kurt 07/4/09 10:18 PM

    I’m not qualified to comment on the technical aspects of solar film, though I like the fundamental *idea* of it—in principle. I do understand that thin film (like any product) has to meet material costs goals, as well as having enough value to generate profits, which in this case is partly affected by material costs and, for that matter, manufacturing costs. I hope someone can come up with a film that is cheap, material wise; easy to make, which would make it less expensive to make than a product requiring a complex manufacturing process; and efficient enough to meet or exceed parity.

    Reply
  • VIKRAM 07/5/09 1:35 AM

    I need sugestion - In case we are thinking of a solar power plant,which type of solar plant is better. Is it PV based OR solar thermal.This is with respect to-
    1. Intial investment
    2. Number of useful years of power production
    3. Efficency of the plant

    Reply
  • JoeJoe 07/5/09 6:01 PM

    The numbers could go either way depending on which plants you compare.

    1. The recently completed Andasol 1 plant cost about $6/watt but the costs for Andasol 2/3 should be lower. The cheapest PV plant that I could find a cost for (Brandis) was $4.5/Watt… PV costs have been falling rapidly this year so cheaper plants are probably in the works.
    2. The life of the PV plant will depend on the technology you use. Same goes for the Solar Thermal plant. With the solar thermal plant you’d have to figure for the possibility of refurbishment extending the life of the plant if the economics justify the investment.
    3. Solar thermal plants will generally be much more efficient.
    Note 1. You need direct sunlight for solar thermal plants. This restricts the siting of CSP.
    Note 2. You need power lines for centralized plants. PV can potentially be used on-site so you can bypass some wires - this equates to a savings of about 4 cents per kWh.
    Note 3. The most important metric is the delivered costs of electricity.

    Reply
  • Steve Pluvia 07/5/09 8:29 PM

    JoeJoe:  You’re way off target on PV plant capex; many PV plants are constructed under $2/watt

    Reply
  • ECD Fan 07/6/09 12:02 PM

    Steve:  First Solar put $1.40 per Watt BOS as of Q1 of 2009 at the Analyst Day (slide 65).  Are you saying First Solar is selling the modules below 70c per Watt these days?

    Reply
      • Steve Pluvia 07/9/09 10:10 AM

        ECD Fan:  I meant $2 for capex of PV production plants, not PV installations; FSLR installed costs for systems they design are currently in the the $3.20/watt range +/- 20 cents.

  • JoeJoe 07/6/09 8:42 PM

    Under $2/Watt eh? Are you from the future Steve?

    Reply
  • JoeJoe 07/9/09 1:19 AM

    So Steve… Your $2/Watt estimate is very wrong as far as I can tell… You talk the big talk but sprutter when it comes to details. Please surprise me with an link to a finished project that comes in below the Brandis plant.

    Reply
      • Steve Pluvia 07/9/09 10:17 AM

        JoeSchmoe, re-read the comments; you’re confused.

.