Add ion mob slot number to the data for the user before loading Bruker timsTOF
data.

There is one big problem with drift times: the fact that we have the
timsframe reader produces mass spectra that do not have Bruker 1/K0 values
but we do expect them all over the place. Thus, any later integration that
involves the DT dimension will crash. Really, we should be able to:

either have a single mobility value associated with a unit.
either have a member data that says that "yes, the data are for Bruker
timsTOF, but we have no 1/K0."
