[QE-users] A few questions about cell dynamics

[laurent pizzagalli]

Dear Andrii

1) you cannot make a direct comparison between your NEB computed barrier 
and the MD-estimated transition temperature. It is not possible (I 
believe) to extract an energy barrier from a single MD run. But you can 
make several runs at different temperatures, and extract the barrier 
(Arrhenius plot). This way you can compare.
2) 10fs is a large timestep. The stability depends (not only) on the 
duration and temperature of your run. Since your run is quite short, 
this is maybe it is stable. If you have CPU time, I would suggest to be 
conservative and try a smaller timestep.
3) I do not understand your question. I think that the temperature set 
by tempw is equal to the temperature in Kelvin that you want.

hth

Best,

Laurent P.

On 18/10/2020 15:42, Andrii Shyichuk wrote:
> Dear Users,
>
> I study a certain defect in a crystal, where there are two states 
> (geometries), say, A and B.
>
> Structure B is a meta-stable state:
> - with PW vc-relax I get an optimized B structure;
> - structure A total energy is lower than B total energy, and B can in 
> principle morph into A;
> - NEB shows presence of energy barrier of the B>A process, but it also 
> shows that B is a local minimum.
>
> In order to check if structure B is stable under heating, I've ran PW 
> vc-md at different temperatures, and for lower temperatures B is 
> stable, while for higher temperatures B morphs into A, as expected.
>
> The questions:
>
> 1. How can I make a correspondence between the NEB barrier and the 
> MD-estimated transition temperature? Should I expect the transition to 
> happen when Ekin in MD is larger than E_barrier? (Both energies are 
> per cell.)
>
> 2. I use ion_dynamics = 'beeman', ion_velocities = 'default', 
> ion_temperature = 'rescaling' and nraise = 1. I've tested different 
> values of dt, and it seems that with dt = 206.623 (10 fs) my system 
> does not misbehave (the things go south with the atoms flying around 
> at dt = 500 and higher).
> Should I assume that I play it safe, or is it better to use a smaller dt?
> The transition takes about 500 fs, hence it's either long steps, or a 
> lot of steps.
>
> 3. How much does the temperature set by tempw ( tolp = 50 ) correspond 
> to the actual temperature? E.g. if I find that the morph happens at, 
> say, tempw = 1000+-50 and higher, how safe it is to say that the same 
> will happen at 1000+-X K and more in real life?
>
> Thank you in advance.
> Best regards.
> Andrii Shyichuk
> Department of Chemistry, University of Wrocław, Poland
> _______________________________________________
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
> users mailing list users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users

-- 
                                                       ,,,     __,
                                                      /'^'\   |__|
                                                     ( o o )  |
--------------------------------------------------oOOO--(_)--OO|o------
<Laurent.Pizzagalli at univ-poitiers.fr>
http://laurent.pizzagalli.free.fr/            Tel +33 549 49 74 99
------------------------------------------    Fax +33 549 49 66 92
Institut P'
Departement de Physique et de Mécanique des Matériaux
CNRS UPR 3346
Université de Poitiers
SP2MI
TSA 41123                                          .oooO
86073 Poitiers Cedex 9, FRANCE                     (   )   Oooo.
----------------------------------------------------\ (----(   )-------
                                                     \_)    ) /
                                                           (_/