We gratefully acknowledge

This calculated thickness is then used in Eq. [19] to ob-tain the approximate order for ~,j. Since the order should be half-integral, the APL function assumes the closest half-integral, and the order for each subsequent mini-mum to shorter k is increased by 1. Equat ion [9] is then used (with the appropriate corrections if necessary) to calculate the thickness corresponding to each minimum d = + Ads + Adr [9] 2(n 2- sin20)i/2 The thicknesses are then averaged and the standard e-viation determined. If the dispersion values are in error, then d from Eq. [18] will be incorrect, and if it is far off, the order Nj will be incorrect and the standard eviation will be larger. The possibi l ity of obtaining incorrect or-ders increases with film thicknesses, but if the disper-sion is correct, then this possibility is minimized