initial temperature

Hello,

I 'm trying to simulate clay layers and to constraint all degrees of freedom of the layers, I decide to consider them as immobile and don’t use any integration for them, but I should keep their temperature similar to the mobile parts which are integrated by NVT and for this I 'm using temp/berendsen, but the problem is for the start point it gives the error that immobile parts need to have initial temperature, so I was wondering if somebody tell me what the most compatible way is to assign an initial temperature to this system.

I use something like this:

timestep 0.1 # fs
fix 1 others nvt temp 150.0 150.0 100
fix 2 layers temp/berendsen 150.0 150.0 100.0
run 50000

Thanks in advance.

Well, if you keep the layers totally fixed, they're essentially stuck at 0K. If you want them to have a temperature, you'll have to allow them to move. Could you not use something like fix spring/self to stick the layer atoms to their starting positions?

Hello,

I 'm trying to simulate clay layers and to constraint all degrees of
freedom of the layers, I decide to consider them as immobile and don't use
any integration for them, but I should keep their temperature similar to the
mobile parts which are integrated by NVT and for this I 'm using
temp/berendsen, but the problem is for the start point it gives the error
that immobile parts need to have initial temperature, so I was wondering if
somebody tell me what the most compatible way is to assign an initial
temperature to this system.

using temp/berendsen on atoms that are not time integrated is pointless.

if particles don't move, they don't have a temperature regardless of
what velocity you have assigned.
if you want your nvt integrated area to be not affected by the
immobile layers, just add an additional
layer in between that is thermostatted.

axel.

Axel,
Yes, but the problem is I want to have interaction between inter-layer molecules and immobile layers’ molecules and if I don’t assign any temperature to the immobile layers, the system will loose its heat too.

Axel,
Yes, but the problem is I want to have interaction between inter-layer
molecules and immobile layers' molecules and if I don't assign any
temperature to the immobile layers, the system will loose its heat too.

you are not paying attention: if you don't time integrate atoms, they
*are* at 0K.
you either need to introduce an additional layer or use a less drastic
method to keep those atom in place.

axel.