Silicon Crystalline Cells

The idea of surface texturing and its utility is illustrated in Figure 6.4. It is seen that the “facet” makes two reflections necessary for backscattering of light, so that if the

photons. Energy runs vertically, with zero at energy of free electron outside the device. This could be called aPIN device, with an (I) insulator layer. Work function j shows highest energy of an electron in the device. Symbols x are electron affinities in the different regions ofthe device, while Ec and Ev, respectively, are the lowest electron energy and the highest filled valence band energy. The effect oflight is to separate the quasi-Fermi levels for electrons and holes, indicating nonthermal light-driven concentrations of electrons and holes. The resulting shift of quasi-Fermi energies gives the output voltage of the illuminated device [64].

single reflection occurs with 0.1 probability, the backscattering probability is reduced to 1%, a large improvement. The photon that enters the silicon now has a larger angle to the normal, making its path in the silicon longer before encountering the back surface, and also increasing its chance for internal reflection at that surface.

The reflection coefficient R is calculated at normal incidence by the formula R — (n-1 — П2)2/(П1 + П2)2. Takingvalues 3.5 and 1 forthe indexofrefractionfor silicon and air, respectively, R — 0.31, which is substantial, and an antireflection coating is usually applied as mentioned above. The large index of refraction also implies that

been applied to (100)-oriented single crystal Si cells, and a similar method to polycrystalline Si cells, by M. Green and coworkers [66].

light inside the silicon is easily internally reflected. In fact, since the critical angle 0C is given by sindc = 1/nsi is small, 0c = 16.6°, all light except that within 16.6° of the surface normal is internally reflected. The corresponding angle at the Si/SiO2 interface is 24.7°. Growing thermal oxide (quartz) on the back surface of the cell as indicated in Figure 6.4 reflects light to increase the light path in the silicon, and reduces the recombination rate of electrons and holes at the surface. Heating in a hydrogen atmosphere (“passivating”) further reduces recombination at Si/SiO2 interfaces, by filling “dangling bonds.” Any step to increase the lifetime of minority carriers is beneficial to the operation of the solar cell, where the open-circuit voltage (the separation of electron and hole quasi-Fermi levels as in Figure 6.2) arises as a competition between photogeneration and recombination ofcarriers. (Figure 6.4 is simplified, neither the PN junction nor the front antireflection layer is shown, but these are clearly shown in Figure 6.5.) In Figure 6.4 an array of small contacts to the rear is seen.

The most efficient single-crystalline silicon solar cell, evolving from Figures 3.17, 6.1-6.3, is shown in Figure 6.5 [67]. The faceting indicated in Figure 6.3 has been implemented, showing the N+P junction diffused into the faceted surface. The metal conductive fingers easily form ohmic contacts with the N+ layer, and then the upper surface is coatedwith a MgF2/ZnS antireflection double layer applied above a thin and

passivated native oxide (quartz) layer. This article also details a polycrystalline Si solar cell of 19.8% efficiency with a “honeycomb” texturing of the upper surface.

These authors found that completely enclosing the Si surface with thin thermal oxide to reduce recombination improved cell efficiency. However, the thermal oxide must be thin, on the order of 20 nm, so that the antireflection double layer applied above the oxide will still operate correctly. The polycrystalline version of this cell was grown on 1.5 V m, large-grained directionally solidified P-type silicon 260 pm in thickness. The use of diffused highly doped material just below metallic contacts, on front and back, suppresses recombination by repelling the minority carriers.

It was also reported that the Si/SiO2 interfaces could be improved, passivated to reduce recombination of electrons and holes, by exposure to atomic hydrogen. Thus, the record-efficiency cell is denoted “PERL” (passivated emitter, rear locally diffused cell).

A large installation of single-crystal solar cells is shown in Figure 6.6, located at Nellis Air Force Base in the United States. This array provides 15 MWof power, and is shown to track the sun’s motion in one direction.

The second example of a large solar panel installation, this time based on polycrystalline silicon (Dell Jones (2011) Regenesis Power, personal communica­tion.), is suggested on the cover of this book, the lower right image. This view is similar to panels in a 2 MW field of 185 W Mitsubishi Electric modules, covering 16 acres, installed at the Florida Gulf Coast University in 2008. These modules are rated at 13.4% efficiency, and each consists of 50 cells of dimension 15.6 cm x 15.6 cm. These cells are thus similar to the silicon cells described above.

6.1.2