<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
Several possible reasons:<br>
<br>
1. The force field parameters aren't perfect, so there is some inherent<br>
disagreement between simulation and reality. What is the expected value for<br>
this force field? </blockquote><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<br>
2. You're using the isothermal compressibility for water. If your system is<br>
pure liquid benzene, I'd think you would want to use the compressibility for<br>
benzene. I don't know how big the difference would be off-hand, but at least<br>
you'd be convinced that your simulation was set up properly.<br>
<br>
3. Although it won't matter a huge amount, to what temperature does the<br>
experimental density correspond? Usually these parameters are given at 25C (298<br>
K). The temperature you've used is 300 K. Again, a minor point, but one worth<br>
doing correctly in a simulation even though the such a change would not account<br>
for the magnitude of difference you're currently seeing.<br>
<br>
-Justin<br>
<br></blockquote></div>Hi Justin,<br> Thanks for your suggestions! <br> The OPLSAA model for liquid benzene could get a density of 0.873+/-0.001 g/cm^3 (JACS,1990,112,4768) at 298K and 1 atm. I have done some test and maybe I have found what is the problem. In the previous simulation, I applied long range dispersion corrections for energy
and pressure with "DispCorr = EnerPres". When I apply the long range dispersion corrections only for energy with "DispCorr = Ener" (the temperature is set to be 298K and the pressure is set to be 1 bar), the density of the system is 0.883, which is closely to the experimental value and the expect value of this model. <br>
Now, I have anther question. After the simulation, I want to calculate the hear capacity of liquid benzene. So I using this command:<br>g_energy -f *.edr -s *.tpr -o energy.xvg -b 10000 -nmol 600 -nconstr 12<br> And these are the results:<br>
Statistics over 5000001 steps [ 10000.0000 through 20000.0000 ps ], 11 data sets<br>All statistics are over 500001 points<br><br>Energy Average Err.Est. RMSD Tot-Drift<br>-------------------------------------------------------------------------------<br>
Potential 19.7926 0.036 0.435187 -0.0870648 (kJ/mol)<br>Kinetic En. 29.7264 1.7e-05 0.360236 -7.68964e-05 (kJ/mol)<br>Total Energy 49.519 0.036 0.575006 -0.0871418 (kJ/mol)<br>
Temperature 297.999 0.00017 3.61127 -0.000770665 (K)<br>Pressure 1.15978 0.0036 171.048 -0.00963121 (bar)<br>Box-X 4.45033 0.0015 0.0105071 -0.0038387 (nm)<br>
Box-Y 4.45033 0.0015 0.0105071 -0.0038387 (nm)<br>Box-Z 4.45033 0.0015 0.0105071 -0.0038387 (nm)<br>Volume 88.1423 0.089 0.624413 -0.228234 (nm^3)<br>
Density 883.013 0.89 6.25236 2.28197 (kg/m^3)<br>Enthalpy 29712.2 21 345.005 -52.2857 (kJ/mol)<br><br>Temperature dependent fluctuation properties at T = 297.999. #constr/mol = 12<br>
Isothermal Compressibility: 0.000107512 /bar<br>Adiabatic bulk modulus: 9301.25 bar<br>Heat capacity at constant pressure Cp: 218.791 J/mol K<br>Thermal expansion coefficient alphaP: 0.000135136 1/K<br><br>I got a hear capacity at constant pressure with 218.791 J/(mol K). However, the experimental value is 135.98 J/(mol K), and the expect value of the model is 130.54 J/(mol K) (JACS,1990,112,4768). In the new simulation, I used LINCS for all-bonds. So I think the #nconstr should be 12 in the g_energy command. Is this value for #nconstr right? If it is wrong, what value should I use. Or I should not use long range dispersion correction for energy? Are there other mistakes with my parameters?<br>
<br>Best regards!<br>Cenfeng Fu<br>