Assign sensible (non-zero) frequencies for photoionization fit type 0

[chvogl]

Is your feature request related to a problem? Please describe.
In PR #251 carsus assigns not only zero values for the photoionization cross sections but also for the frequencies for CMFGEN fit type 0. This causes problems in TARDIS because it leads to a zero division error in the calculation of the collisional ionization rates.

Describe the solution you'd like
The frequency grid should start at the threshold for photoionization. Also, the size of the cross-section table should be kept as small as possible to avoid storing unnecessary zeros. Two frequency points should be sufficient.

Describe alternatives you've considered
In principle, using any non-zero value for the frequencies also works but it might be a source of confusion.