Category Handbook of Photovoltaic Science and Engineering

The two-terminal series-connected configuration provides the fewest possibilities for interconnection of the devices. This configuration requires that the subcells be of the same polarity and that the photocurrents of the subcells be closely matched, since in this series connection the subcell with the least photocurrent limits the current generated by the entire device. This current-matching constraint, about which more will be said shortly, puts relatively tight constraints on the selection of bandgaps for the various junctions in this structure. Against these disadvantages, however, are critical advantages...

In contrast, in the three-terminal configuration the subcells are not electrically isolated; the bottom of each cell is electrically connected to the top of the cell beneath it. The fabrication of a monolithic three-terminal device is relatively straightforward, although more complex than the fabrication of a two-terminal device. The semiconductor structure must be designed to provide a layer for contact to the intermediate terminal, and to accommodate the processing steps necessary to put the intermediate terminal in place. With this intermediate terminal, the different subcells in the stack do not need to have the same photocurrents. Furthermore, in this three-terminal configuration, the different subcells in the stack may have different polarities, e. g...

8.1.3 Four-terminal

There are several ways to connect power leads to the junctions comprising a multijunction stack. These configurations, which provide for varying degrees of electrical isolation of the subcells, are illustrated in Figure 8.4c for a two-subcell stack. In the four-terminal configuration, each subcell has its own two terminals and is electrically isolated from the other subcells. This configuration has the advantage that it sets no constraints on the polarities (p/n vs n/p) of the subcells, or on their currents or voltages. However, the terminals and the electrical isolation between subcells in the four-terminal configuration would be inconvenient to accomplish monolithically, requiring a complicated cell structure and processing...

The multijunction approach requires that an incident photon with a given energy be directed onto the correct subcell. Perhaps the conceptually simplest approach would be to use an optically dispersive element such as a prism to spatially distribute photons with different energies to different locations, where the appropriate cells would be placed to collect these photons. This approach is illustrated in Figure 8.4a. Although conceptually simple, in practice the mechanical and optical complexities of this scheme make it undesirable for most circumstances. A preferable approach is to arrange the cells in a stacked configuration, as illustrated in Figure 8.4b, arranged so that the sunlight strikes the highest bandgap first, and then strikes the progressively lower-bandgap junctions...

Henry has calculated the limiting terrestrial one-sun efficiencies for conversion with 1, 2, 3, and 36 bandgaps; the respective efficiencies are 37, 50, 56, and 72% [16]. The improvement in efficiency from one to two bandgaps is considerable, but the returns diminish as more bandgaps are added. This is fortunate since the practicality of a device with five or more junctions is questionable. Note that the promise of the multijunction efficiency improvements will not be realized unless the bandgaps of the multiple junctions are correctly chosen; this choice will be discussed below in detail. Theoretical efficiency limits for multijunction devices based on thermodynamic fundamentals are presented in Chapter 4.

8.1.2 Wavelength Dependence of Photon Conversion Efficiency

As a prelude to the detailed examination of the design and performance of multijunction cells, it is useful to review briefly the fundamental factors that limit the efficiency of single-junction cells. Consider an ideal single-junction cell with characteristic bandgap Eg. A photon incident on this cell with photon energy hv > Eg will be absorbed and converted to electrical energy, but the excess energy hv — Eg will be lost as heat. The greater the excess, the lower the fraction of that photon’s energy will be converted to electrical energy. On the other hand, a photon of energy hv < Eg will not be absorbed and converted to electrical energy. Thus, the efficiency of photon conversion is a maximum efficiency at hv = Eg...

The PV industry currently services a wide range of terrestrial applications, from power for small consumer products to larger grid-connected systems. III-V solar cells are currently too expensive for most one-sun applications. While satellites represent an example of an application for which the extra cost is acceptable, for bulk electricity generation, a concentration of 400 suns or greater may be needed to achieve an acceptable cost. Multiple companies have now implemented multijunction III-V cells into terrestrial concentrator systems. Concentrator cells and systems are discussed in detail in Chapter 10.

Use of GaInP/Ga(In)As/Ge cells in high-concentration systems (500x and above) has the potential of generating electricity at 7 cents/kWh [15]...

8.1.1 Space Solar Cells

The higher efficiencies and radiation resistance of III-V cells have made them attractive as replacements for silicon cells on many satellites and space vehicles. Over the years, III-V mul­tijunction cells have largely replaced silicon cells on new satellite launches. The GaInP/GaAs/Ge cells are integrated into modules very much like single-junction solar cells, and have the added advantage of operating at high voltage and low current, as well as having excellent radiation resis­tance. They also have a smaller temperature coefficient than silicon cells, which implies better performance under the operating conditions encountered in space applications.

Space applications of GaInP/GaAs/Ge and other III-V solar cells are discussed in detail in Chapter 9.