Restrict rho in photopairproduction between 0 and 0.5

[Jean1995]

As discussed in #345, it would make sense to restrict rho between 0 and 0.5.
Since the differential cross section should be symmetric in rho, it is sufficient to only look at half of the differential cross section.
The total cross section for example could be multiplied by two.
This can improve accuracy and performance.

First, we need to check why the Tsai distribution in fact is not symmetric in rho. Otherwise this approach won't work.

[asandrock]

For nonrelativistic energies it is fine for the cross section to be slightly asymmetric in rho. If the energy is small, the different sign of the charge between electron and positron should make a difference; after all, the electron is attracted by the nucleus, while the positron is repelled. If its energy in the rest frame of the nucleus is small, that would make a difference.