There are two issues worth noting here:
- since you are not using a pair style that considers the particle diameter when computing forces and does not compute a torque, changing the diameter has no impact on the simulation. You probably don’t want to use the harmonic/cut pair style, which computes interactions between point particles, but rather use one of the granular pair styles.
- with respect to pair style harmonic/cut please also note the discussion here Recently added pair_style harmonic/cut seems to be wrong
so you probably want to wait until the next LAMMPS patch release is available before using it for proper simulations.