The Relentless Rise of Disruptive Photovoltaics

By: Scott Stephens, Material Matters Volume 4 Article 4

United States Department of Energy Solar Energy Technologies Program 1000 Independence Ave., SW Washington, D.C., 20585 
Email: scott.stephens@ee.doe.gov

Introduction

Continuous technological innovation has sustained over 30 years of cost reductions and exponential market expansion for photovoltaics. This article aims to provide a technological overview of the industry, to highlight some potential technology trends, and to convince the reader that this aggressive growth is likely to continue for the foreseeable future.

back to top

Market Growth

In 2008, the world market for Photovoltaics (PV) grew by 70% and resulted in approximately 6.9 gigawatts (GWp) of manufacturing capacity which translates to nearly $40 Billion dollars worth of system-level newly added capacity. This growth is a continuation of the roughly 30% compound annual growth rate the industry has sustained for the past 30 years. Although the growth for 2009 is projected to be sharply lower and significantly negative in terms of revenue growth, the long-term growth prospects for the industry remain very promising primarily due to the potential of future cost reductions.

back to top

Technology Developments

Today, approximately 90% of cumulatively installed PV modules are based on crystalline silicon (c-Si) technology. The upstream manufacturing steps of this process are closely related to the standard wafering processes used in the integrated circuit industry, but with greater emphasis on throughput and feedstock efficiency. The process begins with “solar grade” polysilicon that typically is between 99.999% and 99.9999% pure, and in most cases containing fewer than 1 part per million (ppm) of boron and phosphorus, and less than 5 ppm total for all other metals. This polysilicon feedstock is cast into an ingot either by pulling a cylindrical single crystal from a melt or by freezing a multicrystalline block in a ceramic crucible. The next step involves using a wire saw to slice the ingot into square wafers that are typically six inches in diameter and approximately 160-180 μm thick. Cell processing steps involve etching away the surface damage caused by the wire saw, diffusing either phosphorus or boron to form a pn junction, surface treatments to prevent charge carrier recombination and increase light capture, and applying metal contacts to collect the separated charge carriers. Today, there are numerous device architectures being manufactured, each with slight variations in this process and the addition of a few unique steps. Lastly, the finished cells are tiled, strung together (primarily in series to increase the voltage output), encapsulated between various polymer layers (the most common being one based on Ethylene-Vinyl Acetate or EVA on the front and a white back sheet often made of polyvinylfluoride), and encased with a front sheet of glass and supported along the edges with an aluminum frame. The final product (see Figure 1) is known as a panel or module, which is typically rectangular in shape, approximately 1-2 square meters in size with an output of around 150-250 watts at approximately 30-40 volts. Currently, most c-Si modules are rated to convert approximately 14-15% of incident light into electrical energy under standard testing conditions (25 °C, 1,000 W/m2 energy incident on surface) although some manufacturers use advanced cell designs that allow modules efficiency ratings to approach 20%.

Figure 1. An exploded view of a typical rigid photovoltaic cell. Published with permission from Hisco.


Thin film modules offer an alternate approach that allows manufacturers to reduce material and processing costs. The three leading thin film technologies are based on amorphous or nanocrystalline silicon (a-Si, nc-Si), CdTe, and Cu(In,Ga)Se2 (CIGS) as the light absorbing layer. PV cells are created by depositing submicron layers of these light absorbing direct bandgap materials onto a glass or metal foil substrate (Figure 2). Deposition techniques for the absorbing layer vary by technology type and include plasma-enhanced chemical vapor deposition for a-Si and nc-Si, close space sublimation or vapor phase transport for CdTe, and evaporation, sputtering, ink printing, or electroplating for CIGS. Other layers in thin film devices are deposited via sputtering or chemical bath. If all deposition steps achieve large area uniformity, the entire module can be deposited on a single large area substrate with cells formed via mechanical or laser etch steps, which is a process known as monolithic integration.

Figure 2. An exploded view of a typical flexible photovoltaic cell. Published with permission from Hisco.


Concentrating PV (CPV) modules (Figure 3) reduce the impact of the high cost of semiconductors by shrinking the cell size and using an inexpensive optic to concentrate light onto a high performance cell. Concentration ratios range from 2-1,000 times the light produced by the sun, but are generally clustered around 3 ranges. Low concentration PV modules typically use c-Si cells, concentrate light approximately 1.5-3x, and are placed on a fixed mount that is optimally tilted for a given latitude. Mid-concentration ratios range from 5-30x with linear optics that focus light onto a row of high-performance c-Si cells. This linear configuration only requires tracking the azimuth (hour angle) motion of the sun that can allow lower costs.

Figure 3. An exploded view of a typical CPV module with a Fresnel lens. Published with permission from Hisco.


High concentration PV generally ranges from 200-800x, requires 2-axis tracking, and uses very expensive yet highly efficient multijunction cells. High concentration is the most typical approach to CPV as the cell's cost contribution to the overall system price is minimized. Commercial multijunction cells currently approach 40% conversion efficiency and the 2-axis tracking allows modules to output more power in the early morning and late afternoon. However, there are also numerous drawbacks to CPV; additional costs and reliability concerns are incurred with the use of optics and a tracker, the cells only "see" a small region of the sky around the sun and so the diffuse component of the solar spectrum (~150 W/m2) is lost, and most designs require each tracker to be sufficiently spaced to avoid shading.

Currently the U.S. Department of Energy's (DOE) Solar Energy Technologies Program ("Solar Program") funds research and product development efforts across all three of these general classifications of photovoltaics. Relative to the global energy industry, none of these PV technologies have achieved an efficient or significant manufacturing scale, and additionally, each of these technologies holds promise for disruptive or transformational improvements beyond what is currently considered state-of-the-art. Thus, current cost structures are not the best indicator of which photovoltaic technology is most promising in the >5 year timeframe. Furthermore, these technologies vary significantly in their performance with respect to parameters such as solar resource, temperature, power density, weight, modularity, etc. These differences are highlighted in Figure 4 suggesting that there may be room in the market for multiple types of PV even as the industry matures. Existing product diversity in other technologies such as batteries or displays provides a strong precedent for such sustained market segmentation.

Figure 4. Significant differences in materials and processes exist across the three primary types of PV technologies. Differences in system attributes and performance across these technologies suggest that end markets may support multiple technology platforms into the future.


Since 2005, the Solar Program's stated goal for PV has been to achieve grid parity in residential, commercial, and utility markets and to install more than 5GW of peak capacity in the U.S. by 2015. These targets require a 50-70% reduction in levelized cost of electricity (LCOE) and a sustained 30% compound annual growth rate (CAGR) in the photovoltaic industry's production relative to a 2005 baseline. LCOE is defined as the ratio of the net present value of a PV system's costs to its generated energy ($/kWh) over the life of the system. This equation, along with a non-exhaustive list of technology drivers of the LCOE, are shown in Figure 5.

Figure 5. The Levelized Cost of Electricity (LCOE) is calculated from the Net Present Value (NPV) ratio of a system's price to its energy generation. A non-exhaustive list of significant LCOE drivers is shown along with their interdependencies.


Historically, as shown in Figure 6, modules have represented slightly more than 50% of an installed system price and thus were the focal point of the Solar Program's R&D program.

Figure 6. DOE's current PV research objectives target cost reductions at the system level.


Additionally, over 80% of the installed system costs scale directly with module efficiency. Until this decade, exploring ways to increase cell and module efficiency was the primary driver for advancing the technology. However, as cell efficiencies began to rise more slowly and PV systems became more competitive with grid-tied electricity, attention shifted to total installed system costs. Over the past three years, the DOE's Solar Program has shifted to a system-driven approach to address all significant LCOE drivers. In addition to module efficiency and cost, other key drivers include installation, labor costs, inverter lifetime, annual maintenance costs, and capital costs. A recent and continuing collapse in module prices has further justified this approach. Specific examples of these system cost drivers include; module level power conversion (either DC-DC or DC-AC) that can reduce wiring costs, enable system design flexibility, and increase array shade tolerance. Reducing “fixed” system costs is achieved by moving to larger modules, higher voltages, or partially pre-assembled mounting structures, and improving system energy yield by using anti-soiling coatings or mounting high efficiency modules on inexpensive 1-axis trackers.

back to top

Future Promise

Moving forward, the PV industry appears poised for dramatic growth in the U.S. and worldwide. Over the past 30 years module prices have maintained a remarkably stable progress ratio of 80% as shown in Figure 7.

Figure 7. Historically, the price of PV modules has fallen by 20% for every doubling of cumulative production.


Although there is less information available for installed system prices over this period, there are many reasons to expect a similar downward progression in total system price. The fundamental drivers for PV cost reduction are solar resources, technology, and market. Solar energy is the Earth's most abundant energy resource by a wide margin. Compared with other renewables, it is over ten times more abundant than wind, biomass, geothermal, tidal, and wave power combined. Compared with the world's current energy consumption, the amount of solar energy incident on the Earth is three orders of magnitude larger than the total world power demand. Moreover, it is more uniformly distributed across the Earth than any other energy source with most areas receiving several kWh worth of incident energy per square meter of land per day. In the past, this was critically important for the growth of PV as it enabled remote power production at a much cheaper price than building electrical transmission or transporting fuels. Today and into the future, this attribute remains important as this is roughly the power density required by most homes and buildings. Additionally, the distributed, grid-tied nature of PV generation fits within the constraints of the current transmission network, which is unable to move large amounts of power across long distances. The nature of PV technology promises continued cost reductions into the foreseeable future. The performance improvement and cost reduction of solid state photoelectric devices has, and continues to be, a result of technological and manufacturing improvements. Until recently, this was almost entirely due to the “learning by doing” improvements typical in all modular manufacturing of high tech products. This is likely to continue as the industry grows, but there is now greater attention being placed on lowering non-module costs by exploiting the modular nature of PV installations. Examples of this trend that go beyond traditional system design improvements include software-optimized system design and logistics management based upon standard rooftop or landscape drawings, developing automated installation equipment for ground mounted installations, or even integration of PV products with the manufacture of standard roofing products.

Finally, the abundant and distributed nature of the technology allows market penetration that is both incremental and non-uniform; these are requirements that any expensive, paradigm-shifting technology must conform to. In the mid 1990s, the number of newly added grid-tied solar installations surpassed off-grid systems in the world and marked the beginning of the technology's subtle entrance into the existing and enormous electricity market. Initially, the grid-tied market where PV is competitive with other available electricity was limited to regions with some combination of early adopters, strong incentives, progressive interconnection policy, low cost financing, high electricity rate, and to a lesser extent, a sunny climate. Over the past 15 years, prices have fallen as developments across all but the last of these market drivers have been strongly in favor of PV. There are a variety of reasons for these developments, but in general, they can be attributed to increased costs of traditional energy production and public support for an energy technology, which is inconspicuous, clean, and indigenous. Although cost remains the largest barrier to even greater adoption of PV, market barriers such as interconnection regulations and codes are increasingly a bottleneck. In the 5-15 year timeframe, the effect on the grid of PV's intermittent electricity production will need to be addressed as penetration levels pass 10-15% of the total local generation. It is important to note; however, that the production profile of PV is well correlated with most utility demand curves. In other words, photovoltaics produce power when utilities most need it. Thus, it is likely that promising developments such as smart grid technologies and distributive storage will significantly increase these penetration limits for distributive PV. In both the near and long term, the future for PV appears bright. Various industry experts have made analogies comparing the rise of the PV industry to the rapid growth in other technologies such as integrated circuits or flat panel displays. However, the low cost, long lifetime, and high volumes of materials and consumables associated with PV will ensure that the industry also bears some semblance to lower tech fields like roofing or windows. Ultimately, manufacturers and materials suppliers should anticipate some combination of these two development paths in order to harness the growth of this exciting industry.

back to top

Related Links