[gmx-users] box vectors - regd

[Mark Abraham]

On 16/08/2012 5:46 AM, ramesh cheerla wrote:
> Dear Gromacs users,
>
>           I am using gromacs for simulations of a polymer,  for that  I
> am planing to see how  lattice parameters a , b & c are varying
> during simulation. Here lattice parameter "a" is the length of unit
> cell  along X- direction, "b" is the length of the unit cell along
> "Y" axis and "c"  is along Z -axis.
> For my polymer polymer chains are not exactly oriented  along Z-
> direction they are  a little bit tilted from the Z- axis.  "a" and "b"
> are  along "x" and "Y"  directions respectively so that I can get
> lattice parameters "a" and "b" just by dividing box lengths along
> those directions with the number of  unit cells in those directions.
> As the c- direction and Z- direction are not exactly same ( "c" is a
> little bit tilted from Z ) in this case I shouldn't divide the box
> length along Z - direction with the number of unit cells in that
> direction to get lattice parameter "c". Here my questions are:
> 1) How can I calculate exact  "C" lattice parameter from simulation
> data ? is there any way to get appropriate "c"?
> 2) How one can get box vectors XX YY ZZ XY XZ YX YZ ZX and  ZY  for
> total trajectory, as g_energy is giving only Box - X , Box-Y  and Box-
> Z but i need exact box vectors for valid lattice parameters
> calculations.

Sounds like g_energy is reminding you that you had a rectilinear 
simulation cell when you started, and still do. There are various ways 
to measure angles that will help you address your problem, if you check 
out manual sections 7.4 and 8.

Mark