Regarding creating a temperature profile.


Depending on the goals of the simulation you could in principle use the fix ttm or the one we developed to create a temperature gradient within the electronic subsystem and this would be translated trhough electron-phonon coupling into a temperature gradient in the atomic system.

Best wishes,
Artur Tamm

Axel Kohlmeyer writes:

Dear Axel and Artur,

Thanks a lot for your reply. I was trying to run the following code for assigning different temperature in different region.

read_data Data.R1
region e1 block INF INF INF INF 0 25
region e2 block INF INF INF INF 26 50
region e3 block INF INF INF INF 51 75
region e4 block INF INF INF INF 76 100
group ef1 dynamic all region e1
group ef2 dynamic all region e2
group ef3 dynamic all region e3
group ef4 dynamic all region e4

velocity ef1 create 670 25674 dist gaussian
velocity ef2 create 680 25674 dist gaussian
velocity ef3 create 690 25674 dist gaussian
velocity ef4 create 700 25674 dist gaussian

But it was showing “Cannot zero momentum of no atoms”. I checked the atom number in each group and there are many atoms in each group. Can you please suggest what’s the wrong with the code segment? Thanks in advance.
Best regards

what you are doing is complex and inefficient. just use a thermostat
bias via compute temp/profile. that is what it was made for. the
manual explains how to do this.


Thanks, I’ll check it out.


It looks like you defined dynamic groups; looks like at some point your model might’ve moved a lot and one of those regions was left with zero atoms, it then defined a zero atom group. If you were checking the group count from the log that would only give you the initial group counts I think.