Constant cooling rate and quenching via epsilon.

Dear lammps users,

My intention is to simulate a system at a constant cooling rate and study its collapse. I work in LJ units and in order to quench my system I turn on the attractive LJ by setting the LJ cut-off to 3. I am using fix nve together with fix langevin, and I vary T linearly from 1 to 0.1 in the course of my simulation. Is this reasonable, or does the LJ attraction interfere with cooling dynamics by somehow over-calculating the energy, or by having two separate ways of quenching - by reducing T and via epsilon?

Kind regards