Temperature increase progressively with qeq method

Hi, all

Thanks for the reply of Axel and Ray. Follow Axel’s advice, i run two simulations with qeq and fixed charge, respectively, using the same potenial and time step 0.0002fs, to see how the total energy is been affected. a have attached the plot, (etotal and “ecoul+elong” from qeq and fixed charge), it shows that during the 200ps NVE simulation, in the case of qeq, etotal increase from -360 to -320 ev, and “ecoul+elong” also increase, while the total charge of the system converges to e -14. But in the case of fixed charges, both etotal and “ecoul+elong” remain constant(fluctuating around the average). As for the system temperature, it increases from 300k to 450k in the case of qeq, but remains around 300 in the case of fixed charge. What does this tell me, How can i achieve a state like, NVE ensemble for the calculation of vibrational spectrum? Any idea wil be appreciated. Thanks for your help in advance.

Best Regards

Jiasen Guo

fix charge2.jpg




Several Qs:

  1. Did you really mean a time step of 0.0002fs? If yes, that is meaninglessly small. A time step size of 0.1fs would be small enough.

  2. Did you have the same system sizes? Why are the energies from fixed charge and qeq runs so different? What are the charge values for fixed charge and qeq runs?