Category Semiconductor Band Gaps

Solar energy is gaining considerable attention from Wall Street and countries looking to achieve energy independence. Solar energy represents one of the most significant energy solutions to help eradicate our addiction to oil. Despite the tremendous success offered with solar photovoltaic (PV), more research is required to sustain further deployment and achieve energy independence. Some semiconductor materials used to develop photovoltaic devices are scarce and may limit PV from achieving mass penetration. Let’s review the current solar PV market to better understand the dynamics of this market.

Figure 1 PV Production by Year

Figure 1 demonstrates the rapid market growth of solar PV and Solarbuzz is astute to point out some critical data points: cumulative PV deployment is still less than 1% of global electric usage, PV industry faces capacity constraints, and Germany and Spain account for 47% and 23% of total PV deployment in 2007. With the significant growth in both the production and deployment of solar PV devices, the stock price of some of the leading PV suppliers have appreciated dramatically even despite a recent pull back in the beginning of the year.

Figure 2 PV Production of Leading Suppliers

Despite the turbulence on Wall Street in 2008 with the NASDAQ down 14% year-to-date, and Dow Jones Industrial Average down 7.3% YTD, investor appetite for clean technology stocks remains robust. First Solar (FSLR), a leading supplier of thin film solar PV remains in positive territory and is up nearly ten-fold from its IPO in November 2006. Thin film PV offers a cost advantage over traditional crystalline PV cells. PV devices employ various elements with different band gap properties to achieve improving solar efficiencies. (See our post on semiconductor band gaps: What’s Pushing Solar Energy Efficiency?, October 1st, 2007)

Figure 3 Market Capitalization Solar PV Suppliers

There are several elements used in thin film PV production. Among the elements used include cadmium and tellurium (CdTe), copper, indium, and selenium, (CuInSe), and copper, indium, gallium, and selenium (CIGS). These various elements are used to improve operating efficiencies and lower production costs of PV devices. In general, crystalline PV devices have higher solar efficiencies, but cost more due to their material thickness of 200-to-300 microns. Whereas, thin film PV are usually about 3 microns deep offering significantly lower production costs. However, SunPower (SPWR) the leading polycrystalline silicon PV supplier offers the highest solar efficiency a rating of 22.7% that started shipping in 2007.

Figure 4 FSLR and SPWR Solar PV Production

FSLR and SPWR are the two leading PV players as measured by Wall Street in terms of market valuation. The cost-efficiency tradeoff between these two PV suppliers offers an interesting framework to evaluate the solar PV market.

Figure 5 PV Cost-Efficiency

The stock market appears to be betting on FSLR given its market capitalization of $22 billion and trading at 43 times 2007 revenues of $504 million. FSLR employs CdTe in its solar modules. In several postings on Seeking Alpha starting back in November 2007, Anthony and Garcia de Alba have provided valuable insight into material constraints in the production of PV devices.

Tellurium is a rare metalloid element that is used in producing semiconductor materials because it does not conduct electricity. Tellurium is recovered as a by-product in refining and processing of gold and copper as well as other ores. Tellurium was primarily used to create metal alloys that enable easier machining of end products.

Because of its unique properties, Tellurium and cadmium (CdTe) have been used in thin film PV production since the 1980’s. According to a comprehensive study by Fthenakis and earlier work by Moskowitz “The Life Cycle Impact Analysis of Cadmium in CdTe PV Production”, CdTe is deposited on a thin film substrate using electrodeposition, chemical surface deposition, and vapor transport deposition. FSLR reports in their 10K that they employ a proprietary vapor transport deposition process for CdTe PV production.

A thin film of CdTe is deposited on a substrate at a thickness of 3 microns. According to the Fthenakis and Moskowitz, back in the 1980’s, a 10 megawatt (MW) PV facility employing vapor transport deposition of CdTe uses 3,720 kilograms (kg) of CdTe to achieve a10% efficiency at 3 microns. A one-one bond of CdTe with an atomic weight of Cd at 112.41 and Te at 127.60 suggests Te comprises 53% of the weigh of CdTe. With 3,720 kg of CdTe used at 10MW, the amount of Tellurium used is estimated at 1,978 kg or 197.8 kg/MW.

The electrodeposition CdTe process using a mixture of cadmium sulfate and tellurium dioxide used 880 kg of tellurium dioxide, which amounts to approximately 696.8 kg of Te for 10 MW PV productions. The electrodeposition CdTe process would equate to about 69.7 kg of Te per MW. For a 100 MW PV production approximately 7 tons of Te are consumed.

One would assume the PV production process would improve significantly from the 1980’s and the amount of Te consume would decline with improving efficiencies. This would suggest that FLSR at 200 MW PV capacity in 2007 would consume somewhere between 14 and 38 metric tons of tellurium. This figure is significantly higher than the estimates derived from the FSLR tellurium posts on Seeking Alpha that are closer to10 tons per 100 MW (100 kg/MW).

Figure 6 Te Production

Let’s proceed with the conservative figure of 100 kg/MW (10 tons at 100 MW) to assess the tellurium constraints. Tellurium production is a by-product of gold, copper and other ores. We have found Te production estimates ranging from 132 metric tons (MT) to 300 MT per annum. In a National Renewable Energy Laboratory (NREL) report Assessment of Critical Thin Film Resources in 1999 estimated Te production between 200 and 300 metric tons per year in 1997 and indicated under utilization of capacity for the production of tellurium.

Let’s compare our conservative estimate of 100kg/MW Te usage for FSLR to the optimistic production forecast of 300 MT to evaluate capacity constraints for FSLR. With 300 MT (300,000 kg) global Te production and FSLR using 80% of the Te production, capacity of PV tops out at 2,400 MW (2.4 GW).

The U.S. electric energy usage in 2006 was 4,059.91 billion kilowatt hours (KWH) which translates into 463,460 MW (divide 4060 by 365 days x 24 hours). So without significant investment into research and development for PV FSLR could be constrained at 2,400 MW representing only 0.5% of the U.S. electric usage in 2004. Further more, if FSLR were to be constrained at 2.4 GW annual production, revenues ($2.60 per watt Q4/07) would peak at approximately $6.24 billion, a price-to-sales multiple of 3.4x with its market capitalization of $22 billion.

However, in comparison to leading companies in energy, pharmaceuticals, technology and finance, FSLR’s market capitalization is relatively small. Perhaps with improving production processes, FSLR could reduce the amount of Te per panel and improving mining and metal refinement process could increase Te production to expand the market for CdTe thin film PV devices.

Figure 7 Market Capitalization of Leading Companies

The bottom line is that more research and investment into alternative energies is required to ameliorate the world from being held hostage to oil and hydrocarbon fuels that are directly linked to rising CO2 levels and climate change.

With the Solar Decathlon is taking place in Washington DC today promoting commercially available solar energy technology and oil prices pushing to an all time high of $84/barrel, where are we in terms of cost parity with hydrocarbon fuels. Advances in solar energy technology, lower production costs, and rapidly rising oil costs could bring solar energy into parity with hydrocarbon fuels within five years.

The cost of solar energy remains prohibitively high at about $0.45 per kilowatt-hour (KWH) for an installed system. (a kilowatt-hour equals a 100-watt light in operations for 10 hours and equals 3,412.14 British Thermal Units – BTU) As a reminder, there is significant variance in cost per KWH ($0.23-to-$0.68) because a solar energy system is dependent upon daily available sunlight, which varies from about two hours a day in Alaska to almost six hours in Arizona. Solar energy costs are based on data from solar photovoltaic (PV) supplies SunPower (SPWR) and Sharp Solar.

In comparison to hydrocarbon fuels, one gallon of gasoline has an energy value equal to approximately 125,000 BTUs and cost around $3.00/gallon. The energy value of a gallon of gasoline equates to 36.6 KWH, therefore its price is about $0.08 per KWH. So there is significant disparity between gasoline at $0.08/KWH and solar energy at $0.45.

Fortunately, there are a number of initiatives seeking to improve the economics of solar energy. When we measure the cost of solar energy in dollars per watt we find a complete solar energy systems cost around $8.90/watt. (If we amortize the $8.90/watt over a twenty year expected life of a solar energy system, it amounts to $0.45/KWH) To improve the economic value of solar energy we can lower production costs and/or improve solar efficiency in terms of raising the KWH of electric produced per solar panel.

From the Lewis Group. at Caltech, we learned about several technological approaches to improving solar energy efficiency. Among some of the leading solar technologies are multi-junction semiconductor cells and solar concentrators. In a multi-junction cell, individual cells are made of layers, where each layer captures part of the sunlight passing through the cell. One of the key limits to solar efficiency are the semiconductor band gaps where excess solar energy is lost to heat instead of being converted to electric. Solar concentrators amplify available sunlight and increase the watts per square meter per solar panel.

Last year the Department of Energy reported that solar energy achieved a new record in efficiency performance. The news release, New World Record Achieved in Solar Cell Technology. suggesting that solar concentrators and multi-junction PV cells could surpass the Shockley-Queisser single-junction solar efficiency limit of approximately 31% for single layer solar cell. The multi-junction, concentrator solar cell produced by Boeing-Spectrolab (a Boeing company BA) achieved a world-record conversion efficiency of 40.7% in lab experiments while commercial products operate at 26.5%-to-28.3% efficiency range.

The DOE believes breakthroughs in solar efficiency such as these may lead to systems with an installation cost of only $3 per watt, which translates into a cost of $0.15/KWH. In addition, solar PV cell suppliers are diligently working to reduce production costs. SunPower is targeting a 23% solar efficiency as a goal to reduce its solar energy system cost by 50% by 2012. Semiconductor solar cells benefit from economies of scale, which could reduce cost by 20% for each doubling of production. These solar initiatives should act to reduce solar energy costs.

Meanwhile, as energy costs rise with oil at $84/barrel, the cost of gasoline is expected to increase. If gasoline at the local filling station rises to $6.00/gallon, its cost in KWH jumps to $0.16/KWH. At that price gasoline would be above those achievable through solar energy advancements of $0.15/KWH.

Figure 1 Photovoltaic Device: Efficiency/Cost

With pricing on PV cells falling below $3.00/watt (or $0.15/ kilowatt-hour), solar energy will be close to parity with conventional hydrocarbons. If the cost of gasoline rises to $6.00 a gallon, it would cost $0.16/KWH while current cost of electric is about $0.11/KWH because of the low cost of coal. The bottom line is innovative research initiatives have the potential to deliver disruptive technologies that significantly changes the economics of delivering an energy solutions to all nations.

Further review:
Please see research at California Institute of Technology’s division of Chemical and Chemical Engineering at The Lewis Group.
Imperial College, Jenny Nelson’s Third generation Solar Cells.
DOE’s National Renewable Energy Lab (NREL) Third Generation Solar Photon Conversion (Exceeding the 32% Shockley(Exceeding the 32% Shockley—Queisser Limit)
Solar Energy Market Data Solarbuzz

From our post last week we visited some research at California Institute of Technology’s division of Chemical and Chemical Engineering The Lewis Group and its focus on energy conversion and trade-offs between costs and efficiencies for photovoltaic (PV) devices.

We are providing a brief and very oversimplified analysis of the physics behind solar energy, and to that end, an overview of semiconductor band gaps and new technological approaches to optimize solar energy conversion efficiency. Our prior discussions on energy conversion process, dealt with of the properties of various materials used in silicon and thin-film solar cells. (see Research at Caltech may provide clues to improving solar cell efficiency Energy bands and band gaps provide a way of understanding some of the limitation of solar energy efficiency.

Semiconductor Band Gaps
In semiconductors, the band gap refers to the separation between the valence band, where the electrons are shared among the atoms in the material, and the conduction band where electrons can travel. When the material absorbs light, the electrons, when excited to the band gap threshold level, are able to jump to the conduction band and in the process, create electrons and positively charged holes, that in turn generate electric current.

The energy bands refer to the large number of tightly spaced energy levels associated with crystalline materials. Because electrons behave in wave like movements, their movement interferes with neighboring atoms. The physical properties of semiconductor material are much then those of metals. The minimum energy required to free an electron in metal, ( photo-electric effect) is significantly less than that of semiconductors. For metals the energy bands are more or less continuous, the valence band and the conductive band overlap.

In semiconductor materials such as crystalline, energy bands are grouped into bands, hence energy bands. In crystalline material, the energy bands are split between an upper band called the conductive band and a lower band called the valance band. The energy band gaps in photovoltaic semiconductors are different depending upon materials used. The separations of energy bands in crystalline materials results in gaps between the energy bands and are referred to as the energy band gaps. There are discrete and direct band gap materials. Direct band gap refers to having the conductive band minimum energy value and the maximum valance band energy at the same value such as metals. Please see Bart Van Zeghbroeck’s Principles of Semiconductor Devices chapter 2 for a review of energy bands and semiconductor band gaps.

Figure 1 Semiconductor Band Gaps

With a host of physics equations, models, functions, and laws, one can calculate the energy band gaps for semiconductor materials. In addition, one can alter the band gap of the material through, heat, pressure, or introducing impurities into the material (doping the material) to change its physical properties. In general, the more direct the band gap, the more efficient the band gap material. However, research is working processes to improve efficiencies through doping the material and introducing new materials. The processes seek to narrow the band gap of the semiconductor and thereby improve their efficiencies.

One must remember the cost efficiency tradeoff in evaluating which semiconductor material is best in achieving efficiency and cost parameters. Again, the Shockley-Queisser limit of approximately 30%-to-32% for single junction solar cell and a thermodynamic limit of about 83% are also constraint to improving solar energy efficiency. Please see “Detailed Balance Limit of Efficiency of p-n Junction Solar Cells” (W Shockley and HJ Queisser, Journal of Applied Physics, 1961). Semiconductor materials have limited efficiency because excess photon energy generates heat and not electric current.

Research studies on improving solar energy efficiency are embracing numerous fronts including concentrators and multi-junction semiconductors to impact ionization and spectral conversion. For a quick review, please Imperial College, Jenny Nelson’s Third generation Solar Cells. Multi-junction semiconductors employ two or triple junction semiconductor material to capture excess photon energy. A good overview on solar energy is presented by the DOE’s National Renewable Energy Lab (NREL) Third Generation Solar Photon Conversion (Exceeding the 32% Shockley(Exceeding the 32% Shockley—Queisser Limit)

Figure 2 Approaches to PV Cell Efficiency

There should be considerable excitement given the potential that greater efficiency should be able to drive solar energy costs lower. With PV cells below $3.00/watt (or $0.15/ kilowatt-hour), solar energy with be close to parity with conventional electric rates which are about $0.11/KWH. The bottom line is innovative research has the potential to offer the disruptive technology that significantly changes the economics of delivering an energy solutions to all nations.