A precise method for the spectral adjustment of LED and multi-light source solar simulators

Abstract

Solar simulators based on light-emitting diodes (LEDs) usually consist of many spectrally different LEDs, which in sum produce a sun-like spectrum. On the one hand, this results in the advantage of a high spectral tunability of these systems and on the other hand, however, also in the challenge of a high number of parameters which have to be set for the adjustment of a suitable simulator spectrum. Multijunction solar cells consisting of series-connected subcells are very sensitive to spectral irradiance conditions, which are affecting the current and the fill factor of the device. A precise adjustment of the simulator spectrum based on the spectral responsivity of the subcells is therefore essential for accurate multijunction measurements. Therefore, the number of spectrally different light sources used should be at least as high as the number of subcells in the device under test. However, for the measurement of multijunction devices, the much higher number of spectrally different light sources in common LED solar simulators results in a plethora of different simulator spectra, potentially suitable for the measurement. Furthermore, the nonlinear intensity characteristics of the utilized LEDs as well as the distance-dependent illumination uniformity of such solar simulators add complexity when aiming for a precise spectral adjustment. To tackle these challenges, a new spectral adjustment procedure which is based on a least square's solution algorithm and the definition of appropriate boundary conditions for the calculation of suitable simulator settings is introduced in this publication. Focusing on measurements carried out under constant illumination makes the presented method especially applicable for perovskite-on-silicon multijunction devices. Therefore, an adapted method for the determination of the solar simulator's spectral properties, considering thermal influences which are particularly relevant when carrying out continuous illumination measurements, is introduced in this work. The presented method is verified applying it on a Wavelabs SINUS 220 LED solar simulator by performing a measurement comparison on a multijunction solar cell with Fraunhofer ISE CalLab's well-established multilight source solar simulator.