A multijunction PV device consists of several individual semiconductor junctions stacked together (also called subcells) and connected in series to obtain higher performance. With two subcells, a multijunction solar cell is commonly referred to as a tandem cell. In such devices, however, both the Isc and the fill factor (FF) are functions of the incident spectral irradiance, greatly complicating the determination of device performance at SRC . Because all subcells are in series, the same current must flow through each, and the output current is usually limited by the junction generating the smallest current (called the limiting subcell). A special case occurs when all subcells generate the same amount of current, which is termed current matched. The reason the FF can be a function of the spectral irradiance is that the current generated by a limiting subcell can cause a different subcell to operate in reverse bias and result in a stepped I-V curve in forward bias [1,44-46].
The only way to avoid measuring the wrong Isc and FF is to ensure the currents generated in each subcell under the test light source are equal to those that would be generated under the desired reference spectral irradiance. In general, this can only be accomplished if the spectral response for each subcell is known and if the spectral irradiance of the test light source can be adjusted [1, 44, 45]. Such a simulator suitable for standard multijunction performance is called a spectrally adjustable solar simulator. There are a number of ways an adjustable simulator can be realised, including multiple independently adjustable light sources and selective filtering . A number of procedures can be used to perform spectral matching in adjustable simulators. One procedure is an extension of the reference cell method with spectral mismatch corrections. An iterative process is used in which the spectral mismatch M for each subcell in the test is obtained, following a measurement of the simulator spectral irradiance. Based on these results, the simulator spectral irradiance is adjusted and the process is repeated. When all values of M are within an acceptable tolerance of one (usually 2-3%), the process is stopped and the I-V measurements are made [44, 45, 47]. This procedure will be standardised in the near future . Another procedure relies on a multiple source simulator in which each source can be adjusted independently without changing its spectral distribution. A system of linear equations is then used to obtain the current outputs of one or more reference cells that indicate when the simulator is correctly adjusted .
It should be noted that the spectral adjustment of the simulator does not reproduce the reference spectral irradiance; instead, it sets the simulator spectral irradiance so that the test device is operating as if it were illuminated by the reference spectrum.