Many thanks for the helpful suggestions. Cheers, Jun <div class="gmail_quote">On Mon, Feb 2, 2009 at 3:37 PM, Teodoro Laino <<a href="mailto:teodor...@gmail.com">teodor...@gmail.com</a>> wrote: <blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> Hi Jun then you don't need at all the different thermostatting region at different temperatures. You can specify velocities in the section &VELOCITIES in &SUBSYS. These velocities act like restart quantities. You can (with a small external program) determine the velocities of the hotter atoms through a Boltzman distribution at temperature T1, and the velocities of the other atoms at temperature T2 (T1>T2). You can then input these quantities in the &VELOCITIES section and run your MD. Of course you may consider about running an NVE ensemble instead of the NVT (which you were mentioning in your previous message). NVE makes more sense to me in order to inspect heat transfer (but of course more expert people can comment on this point). Cheers Teo <div class="Ih2E3d"> Jun Cheng wrote: > Thanks a lot for rapid reply, Teo. > > What I am trying to simulate is that a few atoms or molecules in an > equilibrated system are excited to higher vibrational levels (e.g. due > to photo adsorption), and see how the system evolves afterwards. I am > thinking of assigning the excited species an high temperature > initially to mimic the situation of excitation by photo adsorption, > and then seeing how the system reacts. So, it is not necessary to keep > the excited species at the high temperature during MD simulations. > Please let me know what do you think. Thanks. > > Jun > > > > On Mon, Feb 2, 2009 at 3:09 PM, Teodoro Laino <<a href="mailto:teodor...@gmail.com">teodor...@gmail.com</a> </div><div><div></div><div class="Wj3C7c">> <mailto:<a href="mailto:teodor...@gmail.com">teodor...@gmail.com</a>>> wrote: > > > Dear Jun, > it is possible to define different thermostatting regions but all of > them have the same temperature. > It would not work (most of the time) having different regions at > different temperatures unless the regions > are not fully adiabatic. In fact, different regions at different > temperatures ( in case of energy flow between > the different regions) would never thermalize to the temperature you > asked for.. > > Cheers > Teo > > Jun wrote: > > Dear all, > > > > Is it possible to define different regions with different > temperatures > > in a MD simulation using cp2k? If yes, how to do that? > > Many thanks in advance. > > > > Jun > > > > > > > > > > > > </div></div></blockquote></div>