Refactored 'solvate' and 'insert-molecules'

Refactored 'solvate' and 'insert-molecules'

The use of "VDW radii" in these tools was misleading in both
documentation and code. VDW radii are used if they are found, but they
used to be scaled down via a hidden option, and the user had to
magically know what kind of number to give to -vdwd. Documented what
is going on here.

Made that hidden option public, and renamed -vdwd to -radius. Renamed
variables named r to be more reflective of their use as additive
inter-atomic exclusion half-distances. This clarified what "mk_vdw"
was really doing, and that a call to it should not be living at the
end of a function that reads in a conformation from a file. Renamed
lots of things and documented more behaviour. Freed some memory
that should have been freed all along.

This exposed some horrible use of t_atoms. Fixed the code to always
allocate on the heap, call the initializer and call the destructor.

Change-Id: I3f8bbbed07442ccc6cd8e563168f5a68736c6b40