As TNG is based on the AREPO which is a moving mesh code, I wonder if i can use the ParticleIDs of the gas cell to trace the path of the gas(like we usually do in SPH simulation)?
The physically correct answer is no, and the reason is because mass is exchanged between gas cells. As a result, tracking a gas cell with a particular ID does not actually follow any particular (Lagrangian) parcel of mass.
The correct way to do this is then via the tracer particles - I suggest to check out that supplementary catalog, and the references given there.
For a short amount of time, i.e. between two adjacent snapshots, you could try to follow cells based on their IDs. There would be some error involved, but it would give some amount of information about how the gas is evolving (because AREPO is a moving mesh, the amount of mass exchange between cells is minimized, so that this technique makes some degree of sense, as opposed to a fixed grid code, where it wouldn't make any sense).
I see, thank you !