Coating code wrong/broken; needs upgrade

Cannot handle the case when the outer layer on the HR coating has a higher index of refraction than the second layer.

Need to copy correct code from CryogenicLIGO/Sensitivity/coatings.

added bug label

changed weight to 5

there are 2 specific cases where this comes up:

1. GaAs:AlAs
   • here the substrate might be silica (n=1.45)
   • The AlAs is not so stable, so the GaAs needs to be the outer layer and it has a higher index that AlAs
   • the gwinc code always tries to start with a low index cap
   • this code (https://git.ligo.org/rana-adhikari/CryogenicLIGO/tree/master/Sensitivity/GWINC) has a modified set of coating codes which do this correctly
2. a-Si:SiO2 (or perhaps SiN or Al2O3 instead of SiO2)
   • for a substrate (like crystalline Si - n=3.5), the cap layer should be low index
   • the bottom layer should also be low index, since there's no reason to have a high index (e.g. a-Si, n = 3.5) layer next to a high index substrate
   • so the code needs to handle an odd number of layers in this case (this is also possible in the above Matlab gWINC link)

Can you give me some hints about how this bug has been fixed in Voyager gwinc?

So for example, I looked at the getCoatLayers and getCoatRefl functions, and it appears that both gwincs make the assumption that the top layer is low index. But perhaps you get around that by setting the top layer thickness to zero? Does it break if you do that?

oops - I think the AlGaAs one has only been fixed in the 40m/Coatings repo

Has this issue been fixed? I am about to do a calculation with an AlGaAs coating?

No, the issue has not yet been fixed in pygwinc. It might be fixed in the 40m/Coatings code.

As I understand it, the pygwinc coating noise calculation is correct for the coating designs that it accepts (even number of layers, low-index cap), but needs to be generalized to handle the designs we would actually use for AlGaAs (odd number of layers, high-index cap).

added noise label

Adding to this, the current code also does not seem to be able to handle the case where the substrate has a higher index of refraction than the high-index material (raises Exception in getCoatDopt). This comes up for silicon substrates with standard silica-tantala coatings. Would it be worthwhile to see whether this can be fixed (possibly as a special case) in the current code?

@rana-adhikari can you check if this is still an issue with the latest release, which includes @anchal.gupta's new coating code?