That is to say, as an atom is being displaced 0.1 angstrom from group A to group B (disassociation of a hydrogen, lets say), rather than saving each step from the Scan run as a separate opt=ts file, can I do this during the actual scan itself? Thanks.
The procedure you are using is not very clear. If you run a scan it is not necessary to use the ts option to perform the TS search. The scan ( if you do a relaxed one) if the bond distance is the reaction coordinate will give you the TS as the top of the "hill" of the E(R) function. The only problem could be a refinement of the R value to locate exactly the top of the "hill" This can be done calculating the hessian, checking that it has only a negative eigenvalue and star the TS refinement using as force constant atrix this hessian by means the opt=(RCFC,TS) option and reading the checkpoint file conataining the hessaina matrix previouly computed. The problem with the bond breaking if the two moieties end with an unpaired electron is a physical one. In fact this process can not be described by the singlet single Slater determinant. This is an intrinsic shortcoming of this description because the single singlet determinant force the two electron to accupay the same spatial orbital whereas at long distance each electron should be in different orbital. In this case you need a multidterminat approach to tackle the problem. However if you have a redistrubution of the bonds, lets say a braking of some bond and the formation of another bond or a very strong polar bond so that the zwitterionic solution is the most stable you could be lucky and a single determinant approach could give you a reasonable answer. I suggest you to run a check of the wavefuntion toward the singlet triplet instability. If you are using the Gaussian set of programs there is an option "Stable" that performs this check.. In the manual you will find all the needed informations. Good Luck !