![]() Time) while the MSD of the solid grows exponentially. Solid matrix, the gas MSD levels off (rather than growing proportionally to The result, however, is that whenįix rigid/nvt is used for the gas (co2 or n2) and fix nvt is used for the My scripts run basically the same as you have described (both for CO2 andįor the commented out atom.cpp for N2). Is there any reason that fix rigid/nvt and fix rigid might not be working well together? Is there some better way to represent the electrostatics of N2 in LAMMPS that does not require a point charge at the COM (and will thus not require the use of fix rigid/nvt)? Any other test cases I can run to figure out the root of this problem? The use of fix rigid/nvt for the gas and fix nvt for the matrix, the matrix looks to be melted and flowing (reflected in the MSD). When I use fix nvt all (gas + matrix) the matrix simply vibrates around its equilibrium position and the gas diffuses as expected. The result, however, is that when fix rigid/nvt is used for the gas (co2 or n2) and fix nvt is used for the solid matrix, the gas MSD levels off (rather than growing proportionally to time) while the MSD of the solid grows exponentially. My scripts run basically the same as you have described (both for CO2 and for the commented out atom.cpp for N2). While CO2 can be successfully simulated without fix rigid/nvt, it provides a test case that can be extended to N2 (which contains the slightly more complicated zero mass point charge at the N2 center of mass). The speed up from fix rigid on CO2 and N2 will likely be nominal, but will be more consistent with my previous simulations. Any other ways to check that the temp is being calculated correctly? This suggests that fix rigid/nvt is monitoring the correct temperature. I was planning to get back to this soon, so if you find anything helpful and can share, please let me know.īy adding “compute gTemp gas temp” and adding “c_gTemp” to thermo_sytle custom output, I find that the gas temperature is equal to the total system temperature, but with slightly larger fluctuations. I was testing with only 33 N2 molecules before. I was planning to do some larger systems for comparison with DLPOLY so that my temperature fluctuations weren’t so large. I compared with DLPOLY with rigid molecules for a box of N2, and was getting reasonable agreement I thought, but I did not compare with fix nvt for flexible N2. It seemed that the fix/rigid was defining the temperature correctly, if I used the entire fluid (nitrogen + massless site) to define it (see “If you use a temperature compute with a group that includes particles in rigid bodies, the degrees-of-freedom removed by each rigid body are accounted for in the temperature (and pressure) computation, but only if the temperature group includes all the particles in a particular rigid body.”). I then “fixed” the entire fluid (both the nitrogen and the massless site)įix 1 fluid rigid/nvt molecule temp 300.0 300.0 0.5 Velocity nitrogen create 300.0 234324 dist gaussian mom yes rot yes So if my nitrogen was type 2 and my massless site was type 3 I defined 2 groups, one for the fluid (which encompassed the N’s as well as the massless site) and the other which encompassed just the N’s To accomplish running N2 with 2 atomic sites and 3 point charges, I commented out the three lines in atom.cpp that check for 0 mass. I will elaborate on what I was trying to do. The fixes used in each simulation are as follows:įix 2 gas rigid/nvt molecule temp 293.15 293.15 100Īm I using fix rigid/nvt incorrectly? I need to use fix rigid/nvt for N2 and would like to use it for CO2 as well, to be consistent with other simulations and to reduce degrees of freedom and increase simulation speed. The fix nvt results are similar to what is observed for single LJ center gases and the results make sense. However, using the test case of CO2, the use of “fix rigid/nvt” gives results that are inconsistent (and look incorrect) as compared to “fix nvt”. Eventually, I would like to simulate nitrogen (n2) as 2 atomic sites and 3 point charges with one charge on each atomic site and the third on the center of mass (represented as a massless particle with LJ eps and sigma = 0), similar to a previous thread ( ) which requires the use of fix rigid. I am attempting to simulate self diffusion of various gases in a flexible solid in the NVT ensemble by measuring the MSD.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |