Racing to Photovoltaic Grid Parity

[San Francisco, CA USA]

Targeting $2 per Watt Photovoltaic (PV) system installed cost.

From Intersolar North America 2010

At Intersolar North America 2010 last week, Dr. Thomas Surek of Surek PV Consulting presented “The Race to Grid Parity: Crystalline Silicon vs. Thin Films” in the Market & Technology Outlook program. I found his independent views on the PV Market and research directions a breath of fresh air after four conference days.

Lured by the NREL (National Renewable Energy Laboratory) Best Research-Cell Efficiencies eye chart and silicon shortages of “PV yesterday”, Dr. Surek observed many huge venture capital “investments in thin films, concentrator PV, and next- generation technologies (anything “nano”)” were driven with “unreasonably short R&D-to- product cycles projected”. Typical commercial modules only achieve 50-70% of champion cell efficiencies with a practical limit of ~80%.

PV Market Perspective
Crystalline silicon technology dominates the worldwide PV Market although in 2009 Thin Films captured 16-17% market share on a MegaWatt-peak (MWp) basis led by First Solar, Inc. (NASDAQ:FSLR) with about 13%.

IsNA2010RacetoGridParityslide8 The PV seller’s market of 2008 has transitioned to a buyer’s market beginning with the drastic Spanish Feed-in Tariff cuts and followed by the global financial crisis. Reductions in PV module and system prices have confirmed the elasticity of PV demand though well designed and implemented government subsidy programs remain crucial for rapid PV deployment and market growth.

New Polysilicon capacity investments by incumbent producers, chemical industry entrants, and start-ups have switched 2008 silicon shortages to oversupply for the next few years even with double digit PV Market growth rates. Please see my popular Solar Polysilicon Oversupply until 2013? post from Intersolar North America last year for the Hemlock Semiconductor projections.

Two bullets from the PV “today” slide shown above are telling:

  • New start-ups face significant pricing pressures – must compete against lowest-cost silicon producers and lowest-cost thin films (e.g., First Solar)
  • Expect many delays, failures, and consolidations in industry, and decreasing venture investments for next phase(s)

IsNA2010RacetoGridParityslide9

On Grid Parity, Dr. Surek said:

I will argue that this Grid Parity, basically cost effective PV electricity in whatever region of the world or the US you live in, will be reached in the next several years, 2 to 5 years perhaps, and in some cases it is actually cost effective today even without subsidies.

Dr. Surek said:

With supply exceeding demand for at least the next several years, only the lowest cost producers will be competitive. Many of the new start-ups with manufacturing costs that are essentially close to First Solar’s selling prices are going to have problems unless they have either deep pockets or some strategic markets where they can put their products. And Concentrator PV is facing the same problem.

Crystalline Silicon PV
Recovering from the polysilicon shortage, the crystalline silicon PV Module Production Learning Curve is returning to the 80% baseline trend. “Module price decreases by 20% for every doubling of cumulative production.

While many evolutionary developments are possible to continue along the 80% Learning Curve, it won’t go on forever and the limit is likely around or just below a $1 per Watt.

IsNA2010RacetoGridParityslide26 In reviewing the potential for cost reductions across the Silicon PV Value Chain, the following random points caught my attention:

  • An important point: every 1% in efficiency (plus or minus) is worth ~$0.07-0.10 per Watt at the system level.
  • Polyilicon prices will continue to come down and are forecast to be about $30 per kg (kilogram) in 2012 per Richard Winegarner of Sage Concepts.
  • UMG Silicon I think had a better chance a few years ago, but I think that window of opporuntity is much smaller now, it may be a niche add on to the market.
  • “Rule-of-thumb”: Every 1% (absolute) in module efficiency gain is worth ~$20 per kg in polysilicon price

In his prediction for “Soon”, Dr. Tom Surek said:

The Silicon cost will probably be cut in half.
Overall the ingot growth and wafering will be roughly cut in half again.
Solar cell costs will still keep decreasing as well as the module cost.
And you’ll be below $1 to $1.5 and that’s a quite reasonable expectation in the next few years. And that will result in System costs in the $2 to $3 range.

Thin Film PV
A few quick bullets from the lightning review of leading Thin Film technologies.

Amorphous Silicon (a-Si)

  • Fundamental understanding of Microstructure (amorphous to microcrystalline)
  • High-rate deposition (10-100 A/svs. 1-3 A/s)
  • Major growth in manufacturing worldwide 2008-2010, with more than 5,000 MW manufacturing under construction or announced (“standard” equipment available)
  • Are there markets for all these products?

Cadmium Telluride (CdTe)

  • Thin CdS and alternate buffer/window layers
  • CdTe growth; thinner layers (< 1μm)
  • Lowest manufacturing cost and module prices
  • Credible followers (Abound Solar, PrimeStar Solar, …)

CIGS (Copper Indium Gallium diSelenide)

  • Understanding of film growth, microstructures, defects, and device physics – complex but “tolerant” to processes
  • Process controls in manufacturing for uniform composition and thinner films (<1 μm); higher throughput and yield
  • Low-cost, moisture-resistant flexible packaging for roll-to- roll, light-weight products – this is key

In conclusion, Dr. Surek said:

Both crystalline silicon and thin films have credible pathways to installed systems at less than $3 per watt (“soon”) and heading to $2 per watt in a few years after that – both technologies can be winners.

There have been rumors of getting to a $1 per Watt and maybe an initiative coming out of Washington.

I am all in favor of a major initiative in PV along the lines of a ‘space initiative’ such as landing a man on the Moon or Mars.  But while the Moon and Mars are real targets, $1 per Watt PV systems are not realistic.

IsNA2010RacetoGridParityslide33

I know Intersolar North America has been having Login problems, but I still cannot access the Conference Proceedings after complaining for almost a week. Bummer.

(Full disclosure: I own some shares of FSLR stock.)

7 comments

  1. JoeJoe says:

    Good report Ed. Dr. Surek has a good handle on things. I’m surprised he says crystalline silicon will stop around $1? Is this price or cost? As Dr. Surek said, the roadmaps are challenging but straightforward. I would think a 10 year horizon on costs would indicate a limit much closer $.50/Watt. He must see the learning curve slowing down in parts of the production chain.

  2. baseman says:

    Talk to cell phone chip manufacturers and LCD television screen manufacturers about their cost curves with increased volume. Solar costs will go much much lower than those in the industry will lead you to believe. With over 100 manufacturers and capacity coming online quarterly, it will be the lowest cost per watt who wins….and that lowest cost is already targeted at $0.70/watt. With DOW developing solar shingles, First Solar already under $1.00/watt cost, and Nanosolar blowing away BOS costs with edge integrated connections, we don’t have long til under $2/watt is acheivable. Who would have thought in 2002 you could buy a 50″ flat screen plasma at HH Gregg for $640 today? Those who try to go against the tide will get swept away….

  3. admin says:

    Here is a link to the Dr. Surek’s PV Module Learning Curve Slide.

    Don’t forget the metric is total installed cost including module cost, BOS (Balance of Systems), and everything else. The installed price for a California 300 kW to 500 kW PV system in 1Q10 just dipped below $4 per Watt. SunCentric has an excellent CSI analysis.

  4. JoeJoe says:

    Hmmm… The 10 year roadmaps for wafers have thickness coming down to 100 um. Who knows what will happen to poly prices but poly costs should be able to get down to around $20/kg one way or another. Average crystal efficiency should climb up to at least 18% by the end of the decade. Selective emitters, copper instead of silver, all-back contacts, automatic sorting of wafers… CIGS? Lots of good stuff coming along.

    Thanks for the tip on the SunCentric analysis. I read a few of the reports. Those total days in process statistics are deplorable. It’s hard to believe. The incentive structures are way too complicated. I am baffled by the calls for more incentives. It will mean more days in process. The system costs will tend to float on top of all the incentives such that the end-user prices hover around $4/Watt no matter what how many incentive you pile in. If you strip out the dog damn incentives we’ll still be able to easily achieve installed prices of $4/Watt. With added volume we’ll be able to press prices below $4/Watt and really start getting somewhere. The days in process will fall ten fold. The utilities can pre-approve areas by substation, feeder or recloser for X amount of PV. Check… That’s a no brainer. The banks will quickly become comfortable with fast tracking 15 to 20 thousand dollar loans for PV. I don’t see any good reasons why it shouldn’t take much more time than a car loan takes. Check… The material costs are where they need to be. There’s plenty of labor available. What we need is a hero… A hero to lead us. She’s gotta be strong and she’s gotta be fast and she’s gotta go quest for the light.

    We are thinking about these incentives all wrong. We are treating symptoms rather than addressing the underlying inefficiencies which are easy enough to fix.

    Here’s the Photovoltaic IRR/NPV/SPP/LEC calculator I’m working on. This is my US version. The German version has accelerated depreciation, grace periods etc. but the NPV, IRR and SPP functions don’t work. If anyone has comments shoot them off to.

    bingbangvroom@yahoo.com

    http://thepiratebay.org/search/us%20residential%20photovoltaic/0/99/0

  5. JoeJoe says:

    Does anyone have a table of values for the self-consumption kicker. I’ve got the current value and the October value. How far does this kicker value go back?

  6. JoeJoe says:

    The German Wikipedia site had the kicker info I was looking for with a link to a pdf file. Your link provided some missing FiT rate data. Very helpful.. Thanks Slurry!

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>