Hello Discussion Dwellers,
I was wondering if someone might help me understand some unexpected behavior in the temperature/kinetic energy output from a dynamics simulation using Forcite. Specifically, I observed that the reported temperature and kinetic energies for a series of NVT dynamics simulations exhibited significant periodic fluctuations over time. Graphs of this behavior for a particular trajectory are attached to this message, but are characteristic of multiple trajectories.
So here are my questions:
Is the observed periodicity indicative that I've set-up my simulation improperly? If so, then does anyone have any suggestions for what I might have done wrong? I've included more details about these simulations below.
Simulation details:
This simulation was composed of a silica surface of ~57 x ~57 x ~20 Angstroms and vacuum slab of thickness ~20 Angstroms interacting with an adsorbate molecule. The silica structure used in this simulation was obtained from the "structures" folder in Materials Studio and minimized using Discover. To prepare the simulation box containing the silica structure, an adsorbate molecule was randomly selected from a box of adsorbate molecules, which was equilibrated at 298K for 5 ps with the COMPASS26 force field using Forcite. The randomly selected adsorbate was then placed 10 Angstroms above the surface along the Z axis and randomly placed along the X and Y directions and given a velocity of -1 Angs/ps (toward the surface). The silica atoms are each given random velocities. Since I couldn't find a way to assign random velocities to only one particular set of atoms while specifying the velocity for other sets of atoms, I worked around this by reading in the velocity assigned randomly to the silica atoms in a different simulation and then re-assigning those values to the silica atoms in the current simulation box. I had thought that the cause of this behavior might be attributed to how I was assigning velocities to the silica atoms, but then I found that I also observed similar periodicity in temperature/KE when I allowed Forcite to assign velocities randomly automatically.
---- Dynamics parameters ----
Ensemble : NVT
Temperature : 298.00 K
Control method : Nose
Q ratio : 1.0000000
Timestep : 1.00 fs
Number of steps : 10000
Duration : 10 ps
Initial velocities : From structure
---- Energy parameters ----
Forcefield : COMPASS26
Electrostatic terms:
Summation method : Ewald
Accuracy : 0.001 kcal/mol
Buffer width : 0.5 A
van der Waals terms:
Summation method : Ewald
Accuracy : 0.001 kcal/mol
Repulsive cutoff : 6 A
Buffer width : 0.5 A
48 inversion(s) have been deleted from the energy expression as they have no forcefield parameters.
---- Thermostat ----
Nose mass parameter, Q mass: 1715.35 kcal/mol*ps^2
---- Dynamics summary ----
Initial Final Average Std. Dev.
------- ----- ------- ---------
Tot. energy (kcal/mol) -40335.038 -39659.467 -39971.578 304.303
Pot. energy (kcal/mol) -41039.224 -40669.310 -40844.631 153.266
Kin. energy (kcal/mol) 704.186 1009.844 873.053 152.770
Tot. enthalpy (kcal/mol) -52804.347 -51769.807 -52262.460 529.087
Temperature (K) 244.670 350.872 303.344 53.080
Pressure (GPa) -0.666 -0.647 -0.657 0.023
Volume (A^3) 130051.216 130051.216 130051.216 0.000
Density (g/cm^3) 0.883 0.883 0.883 0.000
