[gmx-users] single component temperature coupling

[Chong Teng]

Dear gmx user,

I’m trying to study the thermal property of the liquid, and I have some trouble with equilibrium before I applied different temperature gradient.
First I used flat bottom posres function confine 2nm thickness of the solvents (acetonitrile) at both end of the 5, 5, 100nm box. Then I did the semiisotropic NPT 1ns equilibrium at 300K 1bar. I get temperature of all 4 components(acetonitrile, water, Na, Cl)  and pressure on all x, y, z stable. It looks perfect until I tried to run the NVT with temperature coupling only at both two ends of the  box. It turns out the temperature of the other 3 components in the middle (water, Na, Cl) keep dropping to around 290K.
Below is my t-coupling setting :

tcoupl                  = nose-hoover

tc-grps                 = ACN SOL NA CL   ;

tau_t                   = 10 -1 -1 -1

ref_t                   = 300 300 300 300

Could anyone tell me if there is anything I could make this system working? Or it is just wrong to use single tc-group in nvt ensemble.

Yours sincerely

Chong Teng