Constraining two water droplets at constant distance between their COMs using pull code

In the previous results, the temperatures extracted from the velocities of the atom groups needed to be corrected based on the no. of constraints imposed.

The command used for extracting the temperature values from the trajectory velocities: gmx_mpi traj -f nvt.trr -s nvt.tpr -ot temp_SOL1.xvg -w -b 0 -e 5000

Correction needed: T_real = T_calculated * [3natoms / (3natoms - nconstraints)] (Reference: How can I obtain the temperature of a specific group?)

Here

natoms = no. of atoms in the water molecule = 4
nconstraints = no. of constraints = 6 (3 for SETTLES in the topology file of water molecule + 3 for the constraining all the bonds by LINCS in MDP file)
T_real = T_calculated x 2

The corrected temperature profile is shown in the image below which matches with the set temperatures.

For the second simulation, as the temperature of both droplets are maintained at 300K. there is no evaporation and no fluctuations. However, the droplet splits into 2 over a longer time.

Kindly suggest as to what might be the cause of the droplet splitting.

Thanks and regards,
Tonmoy Sharma