I want to calculate the diffusion constant of urea in my system from Velocity Autocorrelation Function using GROMACS.
So I used the commands:
g_velacc -acflen 1001 -nonormalize -mol -n index.ndx -s md.tpr -o vac_urea.xvg
and then integrated using
g_analyze -f vac_urea.xvg -integrate
The result comes to be
Calculating the integral using the trapezium rule
Integral 1 0.05891 +/- 0.00000
relative deviation of std. dev. cumulants from those of
set average standard deviation sqrt(n-1) a Gaussian distribition
SS1 1.206294e-04 3.905069e-03 1.234891e-04 19.754 331.033
Is the integral value 0.05891 or 1.20629e-04? In the gromacs manual, it says to multiply by 10^-6 and divide by 3 while in a forum it was suggested that it should be multiplied by a 333.33.
How should the diffusion constant be correctly calculated from these values?
Hi, I am not very familliar with the question but keep in mind that gromacs manual is an official document, so it should be the safest choice.
Thank you for the answer . Which is the value that I should choose? The integral value or the one mentioned below the average value?
First of all, don't get confuse with the multiplicative constants. You are dividing by 3 because you are averaging over three components of velocity. The other 10^x factor depends on your choice because it is just a unit conversion factor (default is nm^2 / s).
Because i have never used vel. autocorrelation in gromacs , i can't tell you exactly which value you should use. you can confirm it by doing integration manually.
You should use the value of the integral for the calculation of diffusion coefficient using Green-Cubo relation. What is given below is simply an average and other moments over the first set (SS1) or column in the .xvg file you provided.
Thank you Karol and Gourav!
Your explanations definitely make it more clear.
I'll try and compare the values that I got from MSD and see if there's a similarity
- Badges
- Science topic