“…In these computational experiments we started from optimum HLIS using Al2O3/TiO2 as optical materials and 5nm steps on 4J IMM solar cells for ambient glass and both spectra used in previous sections. We considered two scenarios, namely, a thickness variation suffering from a random Gaussian error with standard deviation of 20% -as reported in (Dong et al, 2018)and a more favorable situation with a 10% standard deviation. The results of such calculations are summarized in Table 1.…”