Dear fellow lammper,

I am trying to figure out what are the output units for the pressure (and pressure tensor) for a solvent where the input units are “real”. I expected that the output unit of pressure would be the same as the input unit , namely atmospheres - but the results are about 560 times larger than I expect for self-consistency. My guess is that the output units are not the same as the input ones. If someone can tell me (a) what the output units for pressure and (b) the pressure tensor are - I would be very grateful. This may be in front of my nose - but I can’t see it.

Further details of my simulations so far are below.

Thanks in advance, and Best Regards,

Donal

Further Details

To double check the parameterization of a forcefield (correct density, specific heat capacity, viscosity at standard temperature and pressure) for a molecular (highly polar) solvent, I did a series of simulations in the (i) NPT ensemble (the resulting density is within one percent of the experimental value) , (ii) NVT ensemble (computed specific heat capacity also within a percent of the correct value)) , but (iii) in the NVE ensemble (where the same density was used as in (i)) I get a mean value of the pressure of 560 (mysterious units) after almost .25 ns run. I expected a pressure of one atmosphere - which would be consistent with (i) where a thermostat and barostat gave the correct density. Using a smaller integrating time step does not change significantly the mean value. The forcefield is a slightly modified cff91 field.

I actually want to calculate the solvent viscosity via linear response theory (using time correlations functions of the pressure tensor)- but before I do that, I need to be sure that the pressure itself is correct.