The State of the Photovoltaic Union

One might say we have come a long way since Alexandre-Edmond Becquerel’s discovery of the photovoltaic (PV) effect in 1839, Charles Fritts’ 1% efficient solar cell in 1883, and the 1954 Bell Labs PV cells which began supplanting batteries in satellites starting with the Vanguard 1 in 1958. Others would say that even Solar Junction’s 43.5% efficient concentrating photovoltaic (CPV) solar cell in 2011 is not enough. As the current tempest of activity in solar cell technology might attest, the search for the best PV solar panel and system grinds on toward the theoretical limits of efficiencies and grid-parity in cost per watt.

The photovoltaic effect

To understand the photovoltaic effect, it’s important to understand the photoelectric (PE) effect. The PE effect is the emission of an electron from a material as a result of absorbing a photon of sufficient energy. The energy of the emitted electron is proportional to the frequency of the photon. The PE effect ultimately led to the discovery of discrete allowable quantized energy states of photons by Albert Einstein in 1905.

The photovoltaic effect is a consequence of the photoelectric effect in optical semiconductor devices, which replace the vacuum of the PE effect with a conductor. Optical semiconductors tend to have a direct bandgap, which denotes that upon excitation from a photon, an electron may change energy state from the valence to conduction band across the forbidden bandgap without an additional change in momentum. Indirect bandgap materials cannot make the shift in electron energy state without an additional change in momentum, which is typically wasted heat.

For instance, a single junction PV solar cell may be implemented with p-type and n-type semiconductor material sandwiching an insulator. This is called a p-i-n junction for short. Upon excitation by a photon of an energy comparable to the junction bandgap, an electron moves from the valence band to the conduction band, while a hole is left behind in the valence band. Together, the electron and hole are called an electron-hole pair. For this to be the PV effect as opposed to the PE effect, the electron is not emitted from the surface of the material but conducted, in this case, by the n-type semiconductor and the cathode contact. The anode contact would be made at the p-type material for the corresponding positive hole charge.

When the PV solar cell is exposed to light, the potential difference between the two electrodes is proportional to the material’s bandgap, or the difference in energy between valence and conduction bands, which is usually expressed in electron volts (eV). Larger bandgap materials correspond to shorter wavelength, higher frequency photons such as ultraviolet and blue, whereas longer wavelength, lower frequency photons, such as red and infra-red, correspond to smaller bandgaps.

PV efficiency

Efficiency is defined as the ratio of useable steady-state electrical power from a PV solar cell to the total input power. Because solar power is used in outer space, solar-panel efficiencies are typically specified with an air mass (AM) coefficient that describes performance under standard conditions. AM0 is the spectrum outside of the Earth’s atmosphere. AM1.5 is the standard testing condition within industry for terrestrial applications.

The seminal calculation on the limit of PV efficiency was done by Shockley and Queisser in 1961. They based the approximation on the fundamental physical limits imposed by blackbody radiation, recombination, and spectral losses with an assumed single p-n junction, bandgap of 1.1 eV, AM1.5, single sun light intensity. Their approximation also held that the energy of a photon in excess of the bandgap is entirely lost, which may not be entirely true. Given these assumptions, a single silicon p-n junction would have a peak theoretical efficiency of 33.7%.

Many researchers have spent the last half-century studying additional factors that reduce PV performance as well as methods for mitigating the limiting assumptions outlined by Shockley and Queisser. Foremost among the methods for improving performance is the multi-junction devices that optimally tune for separate pieces of the available spectrum. Further increases in efficiency may come from concentrating the input light, infrared capture, hot electron capture, and down conversion to a common wavelength.

A study in 1980 applied an updated Shockley-Queisser limit to multi-junction and concentrated PV solar cells. The study showed a 68% limit for unconcentrated and 86% limit for concentrated solar cells. The units of sun light concentration are “suns” where 1 sun is equivalent to unconcentrated conditions and 100 suns is 100 times the intensity of the sun. Typical intensities for concentrated photovoltaic (CPV) cells are from 2 to 1000 suns. CPV performs well in direct sun light, but not as well as unconcentrated PV does with diffuse light environments. Currently, CPV is a small fraction of the solar market, though as the technology is developed and costs are reduced it has the potential to gain market share (see chart above).

Crystalline silicon solar cells use a single crystal substrate, similar to semiconductor integrated circuits (ICs). This mono-crystal wafer for the substrate tends to be an expensive and heavy component to the PV device. The familiar lithography, deposition and etching processes are available to manipulate the wafer into a PV solar cell, just as they might for a MOSFET, NPN transistor, or microprocessor. In the 1960s, the crystalline silicon solar cells were 11% efficient. As of June 2011, SunPower’s production single crystal silicon solar panels have been up to 24.2% efficient. As seen in the NREL figure, the highest efficiency single crystal silicon solar cell was an Amonix CPV at 27.6% in 2005.

Multicrystalline, or polycrystalline (p-Si) or ribbon, silicon PV solar cells are cheaper and less efficient than crystalline. Instead of the tightly controlled single crystal growth process, a casting of molten silicon into the desired form may be used to produce the multicrystalline substrate, which has been a popular method for optical semiconductor devices for some time now. As the silicon cools, it changes phase from liquid to solid. The faster the cooling, the smaller the crystals, often called “grains”, form within the mold. Due to grain boundary defects, the efficiencies leveled off at 20.4% in 2004.

An amorphous (uncrystallized) silicon (a-Si) PV cell, the most popular thin-film technology at about 15% market share, is a thin homogenous layer of silicon deposited rigid or flexible substrates, such as a glass and polymer. Because it is better at absorbing light, the amorphous silicon can be thinner than crystalline silicon PV cells. According to the NREL figure United Solar achieved 12.5% efficiency in 2009 with a triple junction amorphous/nanocrystalline silicon device (a-Si/nc-Si/nc-Si). Nanocrystalline silicon (nc-Si) is synonymous with microcrystalline silicon (µc-Si). Other varieties of amorphous silicon include silicon carbide (a-SiC), silicon germanium (a-SiGe), and silicon nitride (a-SiN). One drawback of the amorphous approach to PV solar cells is the increased degradation of the device performance over time.

Thin films have lower theoretical max efficiency, but superior manufacturability to the tradition semiconductor material processes. As the costs of production decreases, a lower efficiency is allowable to compete with the other solar panels on the often-cited dollar per watt ratio. Current thin film solar cell technologies leading in efficiency and notability are:

• 28.2% GaAs, single junction, by Alta Devices in 2011

• 20.3% CIGS, on glass, by ZSW Stuttgart in 2010

• 19.1% Si, 43µm thick, thin film transfer, by ISFH in 2011

• 12.8% CdTe, monolithic, by PrimeStar in 2011

•10.1% CZTSS, solution grown, by IBM in 2011

Closing remarks

With all the activity to support as many different facets of research, it comes as no surprise that solar PV is the fastest growing renewable energy resource and topped 40 GW worldwide with an annual production rate of 24 GW in 2010. As we climb the ladder to peak efficiency and price-to-performance ratios, it is important to remember that the modern PV solar cells of today are the culmination of over 172 years of research and development. As the NREL figure can attest, there is no easy solution and there were no “overnight” technologies to appear out of nowhere to make PV solar a reality. Each success has a lineage surrounded by a litany of paths that did not make the cut, either for future R&D money or perceived technical feasibility. As yet, there is no one-size-fits-all technology that meets everyone’s requirements, both economically and technically. However, there are a variety of commercial products with economic viability even within the current cost of energy. Each case requires individual analysis of the economic case for solar PV energy systems. Though as the price-to-watt ratio to pushes well below $1/watt for manufacturers, the question will no longer be “Is it economically feasible?” but “When can you install mine?” SPW