[gmx-users] kinetic energy distribution with and without constraints

[unitALX]

Hello!
I was thinking about the use (or not) of LINCS constraints and v-sites in
simulations, and I had a "qualitative" question which I could not understand
from reading the papers. It seemed to me that bond stretching and angle
bending are atomic motions which would consume energy (though perhaps very
little in a well equlibrated system). Is the energy absorbed into bond
lengthening and angle deflection negligible? But if we use LINCS and vsites,
it seems like for the same energy introduced during the first equilibration
step, this same energy would be redistributed over a system with fewer
degrees of freedom to absorb it. 

For example, let's say atom A is bound to atom B, which is bound to atom C;
 and then atom C has a non-bonded collision with atom D. With LINCS off and
no v-sites, I imagine that some of the collision energy would be absorbed
into B-C bond length deviation and/or A-B-C bond angle deviation. But with
LINCS and v-sites on, some fraction of the collision energy can't be
absorbed by these deformation modes, so does atom C then experience a higher
effective force from the same collision? And since C's bond to B cannot
compress, does this same initial collision also then transmit more force
onto atom B, and so on?

--
View this message in context: http://gromacs.5086.x6.nabble.com/kinetic-energy-distribution-with-and-without-constraints-tp5014591.html
Sent from the GROMACS Users Forum mailing list archive at Nabble.com.