Ultimate Efficiency Limits for PV Cells
October 27, 2011, International Solar Energy Technology Conference, Santa Clara, CA—Eli Yablonovitch from UC Berkeley described the physics of photovoltaics. The physics define the ultimate efficiency limits in solar cells.
For semiconductor solar cells, it’s important to optimize activities at the junctions, because holes and electrons combine at the surface. However, a PN junction is optional, because you only need a 4 percent ion dopant density gradient to separate holes and electrons. This gradient only costs you about 1 mV in output voltage.
Adding selected contacts enables single carriers to enter and exit the semiconductor material. The challenge now is managing bulk and surface recombination. First, you have to create an optimal fill factor that maximizes both voltage and current. One limiting factor is the thermal dynamics of devices. Second, is to maximize current by light trapping. Research shows that are textured surface on both top and bottom of the device can increase the internal light path length more than 50 times. Regular shapes do not improve light trapping as much as non-ergotic shapes which trap the most light. Patterned surface textures may be more efficient than random roughness.
Voltage losses come from physical limits and entropy. The loss of directional information causes a 0.28 eV drop in output. The Schockley-Queisser limit suggests that the ultimate limit for direct, single cell conversion is about 34 percent. This theoretical limit also says that any absorber must also be a emitter.
Therefore, an optimal solar cell would not only convert incoming light into an electron-hole pair, but also effectively work as 20 to 30 suns from internal reflections before luminescence. This level of performance requires an internal reflectivity of over 95 percent.
The current record for a single cell in direct light is 28.4 percent at an open collector voltage of 1.11 volts. Full panels achieve 26.4 percent conversion efficiency at 1.03 volts. A single-junction flat-plate collector will achieve 30 percent in the near future. Any collector attempting to get over 25 percent efficiency will need to manage the photons and a god solar cell is also a good LED. The record will go to a lattice-matched dual junction cell.
Existing technologies and expected efficiencies are: AlGaAs 41 percent but dissolves in acid, or 39 percent that doesn’t dissolve in acid. Peel films will raise the upper limits. Si is getting between 15 and 23 percent and is used in 90 percent of current installations. CuInGaSe is promising 13 percent, and poly CaTe has 11 percent in production. Flat plate GaAs could get as high as 43 percent and other III-IV materials are also good candidates.
