One important feature of the PV project market is that the suppliers of the actual power (and thus cash) generating technology – the module manufacturers – are not really in the business of providing Engineering, Procurement and Construction (EPC) construction services on a turn-key basis. This is in contrast to the windpower markets, where developers and investors regularly turn to turbine manufacturers like Vestas, Enercon, Suzlon and others to build the project, using their turbine technology on a turn-key basis.
In doing so, the construction services can be combined and aligned with warranties, extended warranties, and often maintenance services, which in many cases are also supplied by the turbine manufacturers.
Investors and banks like this approach because it frequently brings a company with a large balance sheet – and a thorough understanding of the technology – into the project.
For solar PV projects, the picture looks somewhat different. Sharp, Q-Cells, First Solar and other module manufacturers do not offer this kind of service, and confine themselves to supplying the modules. This paves the way, or rather makes it necessary, for other EPC providers to be introduced into a project. From a bank or investors perspective, these EPC providers need to have the experience to undertake such projects and, more importantly, have the financial muscle to back up any warranties.
The EPC agreements for solar PV projects themselves provide for many of the customary provisions in construction agreements but, in addition, contain a number of clauses which relate to the supply and function of the PV modules, and thus look to address the cash flow coming into the project. In most cases, the EPC contractor would buy the modules from the manufacturer, build them into the project and then sell the project turn-key.
This means that the EPC provider assumes the risk for – and of – the modules, usually for a limited period of one to two years. After that period (and sometimes right from the beginning) the principal has any warranty under the module supply agreement assigned to him, so that any claim is then taken up with the module manufacturer directly, rather than the EPC provider.
Sometimes, the EPC contractor does not buy the modules, rather this is done by the principal, who then hands the modules to the EPC provider – who in turn builds them into the project. Such an approach greatly favours the EPC contractor as it takes a large portion of risk away. It also makes the structure of the project difficult from a bankability point of view, and would normally necessitate some kind of interface agreement; this would regulate the responsibilities and liabilities of the various parties involved in a project; and would prevent the principal having to go from one service or equipment provider to another in search of the right counterparty (and possibly prevent a situation arising where the prinicpal ends up being sent from one to the other, unable to identify the right party).
A recent development, mostly seen in Southern Europe, has put additional pressure on EPC contractors and is probably – at least in part – due to a number of banks feeling uncomfortable about the risks associated with solar PV projects. Under this new development, banks have asked to see a long term output or performance guarantee included in the EPC agreement – before they accept an agreement as bankable. Some of these guarantees are supposed to run for up to 20 years.
This could impose a very heavy burden on the EPC provider, and one which they would almost be incapable of offering. They may also have put this kind of liability on their balance sheet – with all the negative consequences that this can result in.
O&M – and output – guarantee
Some of the Operation and Maintenance (O&M) issues relating to solar PV projects are very similar to those arising on other projects, but some significant differences must be noted.
While in wind projects the technical availability of power generation benchmark usually mentioned is 97%, no such agreed figure yet exists for PV projects. Why is this? It's easy to see how the technical availability of a wind turbine can be established; it has to be ready for generation on 355 days per year. For solar PV projects, the issue is more complex. First, it has to be decided at which level availability is measured; should this relate to the entire project, to a panel of PV modules, or each individual module?
It would seem impossible to guarantee the availability of the entire project where the failure of a single module would mean that the project is not available in its entirety; in fact the likelihood that one of the thousands of modules in a large project stops working at some point is very high. Also, the availability has to be defined very clearly. If, for example, there is dirt or snow on part of a PV module, and this stops the module from performing fully, does it mean that the module or panel is available for generation, or not?
In reality most O&M agreements, rather than having availability guarantees, specify that the contractor should only guarantee reaction time. The issue here is that the mere presence of an engineer on site does not necessarily guarantee that the project will become fully-operating immediately, and close attention should be paid to the obligations of the O&M contractor on site.
Finances and costs
Energy derived from PV still has cost issues, and is therefore not a technology for every site and country. Even if prices continue to come down over the next few years, which should happen if the silicon bottleneck is truly overcome and as market-driven technology results in better and more efficient products, the costs of producing one kWh of energy from PV may still be higher than for traditional energy and other renewable sources (especially if external factors such as the cost to the environment of fossil fuels continue to not be internalised).
The investment costs for PV projects currently range from between around €5500-€6500 per kWp installed capacity, depending of course on the technology used and the country. This is about four times higher than the costs for investing in onshore wind energy.
Such investment currently only makes sense in countries that either have fantastic solar radiation, or a very robust incentive system for driving the technology take up in the marketplace. Germany, for example, is not a particularly sunny country compared to many places, but the feed-in-tariff (FiT) of more than €0.45/kWh has led to Germany playing host to about half of the world's total PV projects.
Spain appears to be following the German lead, having introduced a FiT of almost the same level, and, the fact that it benefits from better solar radiation conditions effectively makes the country the new El Dorado for solar PV projects; the Spanish Government appears to have noticed this trend, and is reputedly discussing measures to limit the applicability of the high FiT – as well as other measures designed to cool the market down.
Summary and outlook
The solar PV market still suffers from an astute shortage of silicon and, as a consequence, modules. Not only are prices high but there are not as many projects as there should be. All this could change in a few years though, and the solar PV market should gain a greater share of the renewable energy sector.
In many respects the solar PV market will always stand alone as manufacturers of solar PV modules don't suffer from the same critical position as wind turbine manufacturers do. The PV market will, for the foreseeable future, have a role to play for independent EPC contractors; and the documentation and contracts which regulate their – and everybody else's – role in a project are still relatively new and in the process of being standardised.
One thing is clear: PV projects alone cannot solve the world's energy problems. One advantage of PV is that the power curve is very predictable. No electricity is generated at night (or in the early hours of the morning and late in the evening) and peak PV generation usually occurs at noon. So on the one hand PV can only be part of the total mix of renewable energy projects. But it offers a great opportunity for investment with very little technology risk.
Cost implications for PV modules II – next generation solar
Many in the industry believe that cost reduction processes in silicon processing will eventually make crystalline silicon cells competitive with traditional generating technologies (without the need for subsidies – see cost implications for PV modules I, renewable energy focus September/October, page 25). At the same time, however, many also consider that the creation of crystalline solar cells is an inherently costly process, and that the best way to achieve cost competitive solar is with a new generation of solar cells.
The theory behind the thin-film technologies is that only a small amount of semiconductor material is required to exhibit the solar PV effect and create electricity. The traditional method of slicing solid ingots of silicon into wafers has its limitations, mostly due to breakages of thin wafers, and is essentially nearing the limits of ‘thinness’. These thin wafers – down to around 180 microns – use relatively large amounts of material which is not necessary for the PV effect, and therefore significant research has been focused on how to create thinner layers which still exhibit this photoelectric effect.
There are many different methodologies available to achieve the deposition of thin films of material, each with their advantages and disadvantages. And different developers have been working with different substrates, particle sizes and semiconductor materials.
Providing a look into all the different technologies under development is beyond the scope of this article but it is worth mentioning that one of the key challenges that developers have had to overcome is cell uniformity. Uniformity of the reactive layers in solar PV cells is essential, so that the cells provide a useful output when included as a module. This is because any slight differences in the solar PV cell structure make the cells exhibit slightly different electrical properties.
These slight differences lead to back currents between cells when wired together, and this effect is greatly exaggerated when many cells are wired up in series to create a module with a useful voltage output (i.e. a crystalline silicon module may comprise up to 36, 0.5 V cells to create a useful output ofaround 12 V; large installation inverters for grid connections can require up to 200-600 V).
Thin film deposition technology has been around for some time, with common examples including vacuum deposition for CDs and sputtering for hard disks. However, it has been a challenge to scale these processes for larger surface areas as required for solar cells. Many developers have thus developed proprietary deposition methods, a number of which are variants on the discussed techniques:
A first example has to be First Solar, which is prominent in Cadmium Telluride (CdTe) solar cells. CdTe cells exhibit relatively high efficiencies of around 10%, and First Solar has developed a technique called High-Rate Vapour Transport Deposition (see figure 1) which allows the company to deposit rapidly the semiconductor layers onto an inflexible glass substrate (pane by pane). First Solar recently announced contracts which priced its output at US$1.87/W (compared to crystalline silicon solar cells at around US$2-US$3/W – depending on the cell efficiency), which indicate costs of around US$1/W.
The second most well known supplier of thin film solar cells is United Solar Ovonic, commonly known as Uni-Solar. Uni-Solar, a division of Energy Conversion Devices Inc., creates triple-junction amorphous-silicon (a-Si) cells. These cells are based upon the deposition of three reactive sets of silicon layers, which absorb different bandwidths and enable the cells to have a relatively high (for a-Si) efficiency of over 7%. Uni-Solar's cells are created using a proprietary vapour deposition process, which deposits the layers onto a flexible metal backing (see figure 2).
Deposition onto a flexible metal backing is technically more difficult than onto a rigid glass substrate, but this methodology allows for continuous roll-to-roll processing, which can potentially reduce costs further. Uni-Solar currently sells its cells for around US$3/W and, based upon margin performance, Ambrian estimates costs to be around US$2.3/W, although this should decrease as the company ramps production from 60 MW pa currently to 300 MW pa by 2010.
PowerFilm, an Iowa-based company listed on AIM, produces a-Si based cells with a flexible plastic backing, which achieve a conversion efficiency of 5%-5.5%. Deposition onto a flexible plastic backing is extremely difficult, not only due to the physical difficulties in maintaining an even and level surface, but also because plastic materials tend to emit gases or vapourised liquids when held under a vacuum – called out-gassing.
At present PowerFilm is the only company (to the author's knowledge) that creates a commercially-available thin film material on a flexible plastic substrate; it does this using a proprietary vacuum deposition process. PowerFilm currently sells into high-value applications for around US$5/W and makes around a 40% margin; however, costs should decrease below US$1/W when new production machinery comes on line, and the company begins to sell into the building integrated solar PV market (BIPV).
Other thin-film companies of note include:
- Sharp, one of the world leaders in crystalline solar, has also invested heavily in thin-film and has developed a tandem type cell which combines a-Si with microcrystalline silicon to create cells with 8.5% efficiency, and 1/100th the width of its crystalline silicon cells (see figure 3). Sharp has recently announced that it intends to build the world's largest solar PV cell manufacturing plant for its new thin-film technology — a 1 GW facility — which is anticipated to begin operation in 2010;
- Daystar is developing CIGS (Copper Indium Gallium Selenide), and is building a 25 MW production line expected to create cells with around 10% conversion efficiency on a flexible metal foil backing;
- Ascent Solar is also developing CIGS-based flexible cells. It plans to produce these on a flexible plastic using technology developed by its parent ITN Energy Systems (developer of CIGS deposition machines); it has not yet entered into commercial production;
- XsunX creates and sells a-Si thin film manufacturing equipment for both flexible and rigid substrates using Plasma Enhanced Vapour Deposition;
- In the unlisted world, there are a number of other thin film manufacturers worth noting (as well as many developers). A few of the most well known include Terra Solar (51%-owned by China Solar Energy), which manufactures a-Si cells on glass substrates; Fuji Electric Systems, which creates a-Si cells similar to the Uni-Solar product; and Global Solar (owned by Solon AG), which produces CIGS cells on a flexible metal substrate.
Photo-electrochemical, polymer, nano-crystal and hybrid cells
These types of cell, sometimes referred to as 3rd and 4th (hybrid) generation cells, are very different in nature to crystalline silicon and thin film cells (2nd generation) as they do not rely on a traditional p-n junction to encourage electron movement.
The most well-developed cell type in this new category is the dye-sensitised cell, a photo-electrochemical cell first developed by Michael Grätzel in 1991. These cells split the source of photoelectrons and the charge separation (p-n junction) in two separate stages and although these cells do not, at present, show great conversion efficiencies, they have the potential to be very low cost. This is because the key materials to create a cell are Titanium Dioxide (the main compound found in white paint), photo-sensitive dye and iodide (an Iodine salt), which are usually held in an ion-conducting paste between two layers of glass.
The most well-known developer of this new class of cell is Dyesol (listed in Australia), which is a major supplier of dye solar cell material. The company currently sells dyes, titania pastes and manufacturing materials to third parties, and develops its own cells and modules.
Another developer of this technology is Konarka Technologies, which has licensed the technology to G24 innovations. G24i is currently in the process of testing and refining the design of a 25 MW production line at its facility in Wales (see News Roundup on page 18).
AIM-listed ITM Power has also ventured into the development of the Grätzel cell. It believes its new class of ion-conducting materials will make a perfect host to the electrochemical compounds of the cell, thus negating the need for the ion-conducting paste and glass. Based upon ITM's materials' performance in other electrochemical processes, experts believe that this is likely to be the case. However, as discussed previously, the key hurdle is not creating one cell that works but creating large quantities of uniform cells that provide a meaningful power output in a module.
What next for PV?
Expect the next 10 years to see leaps and bounds in PV technology development and, ultimately, decreases in cost. These decreases will eventually bring the costs of PV generation below that of traditional generating technologies. When this point is reached, the market for PV cells will be almost limitless – who would want a new coal-fired power station when clean PV generation can be installed near the point of use? It is this almost limitless potential for the product that has driven huge investment into the industry, both angel and VC investment as well as public market, leading to the sky-high valuations of those companies with a commercial product.
There is no doubt that these are exciting times for the PV sector and (especially) the next generation of cells.
|About the authors |
Stefan Schmitz is a partner at Squire Sanders & Dempsey in London, UK;
John-Marc Bunce works for Ambrian.