Navigant Consulting's Paula Mints explains several lessons that the PV industry should recall from its past that will shape its future; she also reports on key themes coming out of the EPIA's recent thin-film conference in Munich.

It's been a long and winding road for all photovoltaic technologies, from kilowatt annual demand in the 1970s to multi-gigawatt annual demand in 2010. Demand in 2010 over 2009 is expected to surge >100% to ~16GWs, and will represent 40% of total industry cumulative demand. With over 80% of industry cumulative volume installed between 2005-2010, it may seem expedient to ignore previous PV industry history. However, in an incentive driven industry this would be unwise, because:


  • Gigawatt level demand still requires incentives and the industry has not learned how to do without them. Many lessons and some warnings can come from observing the incentive-less past.
  • Technology development history can show new entrants what to expect, what can be changed, and what cannot be overcome.
  • Pricing history is what it is, and it continues to repeat.
  • The industry is currently destabilized by rapid, massive growth and needs to learn how to maneuver in this brave new gigawatt level world. This is not dissimilar to changes in the past wherein the industry has had to learn to behave differently.
  • And finally, history repeats, and will continue to repeat itself in all aspects (and not just for PV).


And Now, On with the Show

From embracing the past to engaging a promising future, attendees and speakers at EPIA's 3rd Thin Film Conference (Nov. 9 in Munich) were realistic yet positive about the status of thin-film technologies in the current competitive environment.

During the opening session, EPIA vice-president Winfried Hoffmann offered his insights on future advances for photovoltaic technologies, and updated attendees on the status of Applied Materials' thin-film endeavors. (Yes, the company is still selling thin-film manufacturing tools; it recently achieved 11% efficiency on 1.43m2 for its tandem junction technology.) Other highlights from his presentation:

  • Crystalline silicon prices of €0.60-€0.80/W, and thin-film prices of €0.30-€0.70/W by 2020;
  • 100μm wafers, and efficiencies approaching 25%-30% by 2030;
  • Thin-film will have a 40% market share by 2020.

For thin films, which are lower in efficiency that crystalline technologies, costs increase with area, while power costs are constant and module costs are lower, Hoffmann noted.

Arnulf Jager-Waldau, representing the European Commission Joint Research Center (JRC), spoke about the impact of regulations on the deployment of thin films in Europe and Asia, noting that there are very few countries with distinguishing regulations for thin films. In China there are upfront subsidies for manufacturers, fixed subsidies for off- and on-grid solar, but a national feed in tariff is still under discussion. Concerning the current tender process, 280 companies bid, with the large state owned companies winning the first round at bids of 0.72-0.99 RMB/kWh. The results of the second round of bidding have not been announced. Thailand's 10-year FiT rate will decrease by -19% from 8THB/kWh (approximately US$0.27) to 6.5THB/kWh (~$0.22). Thailand is targeting 500MWp by 2020, which is small when considered alongside Germany, Italy and other European countries, and even in comparison with Ontario, Canada and expectations for the US market. A feed-in tariff has been proposed for Malaysia, and one is under discussion in Vietnam.

Concerning Europe's "20/20/20" mandatory targets, member states were required to submit plans (NREAPs). Out of 27 required plans, 22 have been submitted and are currently under review in Brussels. Under the EU Electricity Directive 2009/72, grid operators may be required to give priority to renewables.

Jager-Waldau also made several sobering, but important points during his presentation:

  • In the Czech Republic, from March 2011, only PV systems on roofs will receive the FiT.
  • In Germany, the priority access for renewables has been suspended since 2009. And most systems can be shut off by the grid operator if deemed necessary.
  • In 2010, PV accounted for 17GWp in Germany, or 18% of the summer peak and 45% of the summer load. There is no framework to allow this to be absorbed, and grid operators are reluctant to increase approvals.
  • By 2020 in Germany, it is expected that the amount of PV on the grid will exceed the summer peak.
  • In 2010, PV and wind were 50% of Germany's maximum winter peak. In 2020 they are expected to exceed the winter peak load.

Offering a utility's perspective into Italy's booming market for PV systems was Ricardo Lama from Enel Distribuzione. In Northern Italy, PV installations are primarily BIPV or retrofit, while in southern Italy PV installations are primarily large-field or multi-megawatt. Enel receives 2000 requests a month for high-voltage connections, and he noted there is no efficient management of network saturation -- in many regions the connection requests are larger than the load. Peak load in Italy is 55GW, while connection requests are significantly higher at 150GW. Since 2008, of >33,000 requests to connect, only 7% (2.2GWp) have made it through the process from acceptance to authorization to connection. To help match generation with load, in overloaded areas a fee is paid to compensate for network overload, and this effort has shown some success, he said.

In the operation of the distribution grid, undesired islanding, voltage regulation, and high-voltage network saturation are critical, Lama said, and there is little room for embedded generation. Smart grid technology, he offered, is the solution to the problems associated with connecting renewable technologies to the grid. To be specific, he defined "smart grid" as "an electricity network that can intelligently integrate the behaviors and actions of all users connected to it in order to ensure sustainable, economic, and secure electricity supply" -- in sum, "build networks that can manage the load, not host the load." Transmission grids are already smart enough, he said, so the focus should be on smart distribution, and of course building new transmission. One fundamental problem is discovering how to build a transmission and distribution system around the capability of customers to change their behavior. Lama also pointed out that governments and others need to invest in R&D while also financing some degree of failure.

Three afternoon panels on the subjects of bankability, module prices generation costs, and bottlenecks for thin-film technologies made important, though in some cases overlapping, points:

  • Concerning technology development, 1-100MWp manufacturing lines are not bankable, asserted Matthias Fawer from Sarasin Bank; 1GWp is considered the size for manufacturing economies of scale. Concerning systems, a growing number of investors are focusing on lower risk/return profile, that is, proven quality with existing references.
  • Oversupply and the commoditization of PV (seeing prices and not technology) are current issues, claimed Marco Martorana from Unicredit Leasing. Banks are less influenced by good experiences, and more strongly influenced by bad experiences. The standard list for bankability is: corporate data, market data (project reference list), product data, and production data.
  • There is no clear technology winner, asserted Enel's Ricardo Amoroso, adding that Enel Greenpower has 6GWp of green installed power.
  • 10 good projects are needed to compensate for one bad project, related Commerzbank's Christian Junior. And the cost of dismantling a system has an influence on generation costs.
  • Said Lars Stolt from Solibro: categorically, 10% conversion efficiency is not high enough.
  • Manfred Bachler of Phoenix Solar made a variety of important points:
    - Phoenix has begun adding 500kWp to 1MWp of non-bankable (i.e., risky) technologies and modules to larger (~5MWp) systems to prove bankability.
    - The higher the fixed costs, the more efficiency matters, independent of project size.
    - In his experience, a-Si outperforms in the summer and underperforms in the winter. Shading can be used in proper system design. And thin-film technologies with 10% conversion efficiency do not need more space than 13% c-Si (though this does not hold true for sloped roofs).
    - Inverter O&M costs have significantly increased in 2010. The price from inverter manufacturers is too high, he said, and cannot be avoided for central inverters because the banks insist on long-term O&M contracts.
  • Cost/watt is still the key metric in the current subsidized environment, said David Wortmann from First Solar. He noted that the substrate (glass) is ~40% of panel costs, and that there are challenges to reducing this.

Asked more than once about the impending indium shortage, Bernhard Dimmler from CIGS maker Wurth Solar said there is none. If the price of indium increased 5×, he noted, it would add ~2% to the manufacturing cost of the technology, and even a 10-fold increase in the price of indium would have a marginal impact on manufacturing costs. Most CIGS manufacturers are working on reducing use of indium by replacing it with gallium, and current research is focusing on replacing both indium and gallium.

Paula Mints is principal analyst, PV Services Program, and associate director in the energy practice at Navigant Consulting. E-mail: pmints@navigantconsulting.com.