Please consider a diffraction grating, I think something in the 50 lines/mm should work, you can also do star spectroscopy with this thing. You need to find the right line density so the first spectral band covers exactly half the screen height, about 900 pixels.
Of course, the star tracking algorithm needs to be updated so it works even when the stars have the two rainbow spikes, if it cannot do that already. The stacking algorithm also needs to work with this and not refuse to stack due to "star trails".
This is mostly a sofware feature, apart from finding the proper diffraction film, which I am not able to locate to test this. You need a 50mm x 50mm piece, with line densities in the 10 to 100 lines/mm to see what works best. We might be able to do real science with the S50.