Langevin thermostat

Hi all LAMPPS users and reaserchers
Does it correct to use a langevin thermostat in coincidence with a PKA in order to study frenkel pair defect? In other words, when thermal spike occurs, the thermostat decreases temperature that increases duo to PKA and it fix the temperature in a constant value. this decreasing befalls in order of some pico-second. is it practically true? or the thermostat must be remove in order to the temperature of the crystal continues to decrease(although the crystal is connect to a source of heat in practice).
Best regard
S.M.Zamzamian

Hi all LAMPPS users and reaserchers
Does it correct to use a langevin thermostat in coincidence with a PKA in
order to study frenkel pair defect? In other words, when thermal spike
occurs, the thermostat decreases temperature that increases duo to PKA and
it fix the temperature in a constant value. this decreasing befalls in
order of some pico-second. is it practically true? or the thermostat must
be remove in order to the temperature of the crystal continues to
decrease(although the crystal is connect to a source of heat in practice).

​have you checked in the literature, what other people have done in this
regard for PKA simulations?​

axel.

Thank you Dr.Kohlmeyer for answering. Actually i did not find any paper that this point was be mentioned in it. Perhaps they did not use it, if so, what if a crystal connect to a source of heat all the time, even though a PKA begins to create damage. In this case, using a thermostat could be urgent, But is it true to cool down a crystal to a fix temperature after such a event?
Let me tell my question beyond the damage subject, and in field of heat transfer, thermodynamic or something like this. In a order of pico-second, is it possible that a crystal reaches to for example 300k (begin with 350k) using langevin thermostat? is it practically true?
Best regards
S.M.Zamzamian

Thank you Dr.Kohlmeyer for answering. Actually i did not find any paper
that this point was be mentioned in it. Perhaps they did not use it, if so,
what if a crystal connect to a source of heat all the time, even though a
PKA begins to create damage. In this case, using a thermostat could be
urgent, But is it true to cool down a crystal to a fix temperature after
such a event?

​papers will not answer direct questions, particularly they will not tell
you what doesn't work or what doesn't make sense. otherwise each paper
would be a thick and extremely boring book. you have to draw conclusions
from observing and comparing what is said and what is not said.

please note, that if you have a sample connected to a temperature
controlled object, then the correct way to model this is *not* to
thermostat the *entire* sample, but rather apply the thermostat to the
contact region. you also have to factor into your conclusions thermal
conductivity of the sample and total time of the simulation/incident. you
also have to consider how a langevin thermostat affects your dynamics, and
particularly how this impact is connected to the damping factor or time
constant.

Let me tell my question beyond the damage subject, and in field of heat
transfer, thermodynamic or something like this. In a order of pico-second,
is it possible that a crystal reaches to for example 300k (begin with 350k)
using langevin thermostat? is it practically true?

​this is *your* research and you have to ​make your own informed choices.
if you worry about whether a model is realistic, you will have to find
experimental evidence proving or disproving it. it most certainly is not
sufficient to get the opinion of somebody that you don't even know and
where you have no record of expertise in your specific case. the internet
is full of trolls who take please from messing with other people...
as for whether an algorithm is capable of doing something, there is the
approach of making empirical tests, or reading up on in and figuring it out
theoretically. in the case of the langevin thermostat, that is
embarrassingly simple to rationalize. in the limit of an infinitely short
time constant, the velocity damping will be so large, the thermostat will
turn into re-initializing the velocities to the desired target temperature
from a random distribution in every step.

axel