Phase transition in Zirconia

Don’t forget adding the list to your reply. It increases your chances for more/better answers.
The problem you describe is a complex one and I hope you realize that. I am no expert in the specific details you are to face therefore my next lines only include comments of a much general nature.
First, before going all in when it comes to complexity you need to better understand how to split up such a problem into more fundamental bits. You talk about radiation cascades acting on some interfaces. Do you already know what do the interfaces look like and how to create them? From all the possible interfaces one can imagine, do you have an initial criteria to guide you on how to build those “relevant” interfaces for your particular needs? Second, when you talk about the cascade triggering a transition, is that an experimental observation or someone else have achieved this via computer simulations? Just because it has been observed experimentally does not mean you can reproduce it in the computer. Sometimes its possible but could be quite tough. I mean, do you know what are the important variables driving the transition; is it the temperature only?, does the cascade triggers local melting that leads to recrystallization and growth? Are the magnitudes of the local stresses driving the transition as well? Please, notice that I’m not trying to confuse you further but only attempting to unravel “the devil in the details”. When I talked about alternative methods to study phase transitions I imagined you were after the phase transformation of an homogeneous system, not quite the mix and mingle case that you are trying to study. For your case I would recommend to look in the lit (if you haven’t done it) for simulations of a similar nature, doesn’t have to be zirconia, but something with the general flavor of the problem you are trying to tackle. Not sure how much of an expert in the field you are but for learning, imitation is the first step to success. Don’t need to be a scientist to know this, this is exactly what children do since very early stages and of course, a bit of neuroplasticity goes a long way. Maybe another Lammpster will have less general and more particular comments for your specific needs.
Carlos

Dear Carlos,

Thanks for your reply, yes indeed I have done the same as you described, breaking the complex problem into smaller and simpler ones.

Don’t forget adding the list to your reply. It increases your chances for more/better answers.
The problem you describe is a complex one and I hope you realize that. I am no expert in the specific details you are to face therefore my next lines only include comments of a much general nature.
First, before going all in when it comes to complexity you need to better understand how to split up such a problem into more fundamental bits. You talk about radiation cascades acting on some interfaces. Do you already know what do the interfaces look like and how to create them? From all the possible interfaces one can imagine, do you have an initial criteria to guide you on how to build those “relevant” interfaces for your particular needs?

I am using LAMMPS as my main MD engine for a year and half now. I am investigating the modification of the primary cascades near atomic GBs (twin, tilt and twist grain boundaries), and yes I know exactly how to create them in almost all metallic elements or alloys with different lattice structures.

Second, when you talk about the cascade triggering a transition, is that an experimental observation or someone else have achieved this via computer simulations? Just because it has been observed experimentally does not mean you can reproduce it in the computer. Sometimes its possible but could be quite tough.

It is both experimental and theoretical. I should also add that I know how to initiate primary cascades using LAMMPS. In fact I have simulated 3 materials, each with 4 different interfaces, for at least 3 different knock-on atom energies with 8 trials, which resulted in 3 reports sent for review and probably publication.

I mean, do you know what are the important variables driving the transition; is it the temperature only?, does the cascade triggers local melting that leads to recrystallization and growth? Are the magnitudes of the local stresses driving the transition as well?

Typically, I use NVE ensemble and null the pressure in the desired Cartesian direction, to conserve the additional energy added due to introducing the knock-on atom (with some kinetic energy) and eliminating the local stress fields. Introducing knock-on atom also leads to local melting and local re-crystallization as well.

The only variable that causes phase transition in this case, based on my review, is temperature, and “phase duality” is what I am investigating on.

Please, notice that I’m not trying to confuse you further but only attempting to unravel “the devil in the details”. When I talked about alternative methods to study phase transitions I imagined you were after the phase transformation of an homogeneous system, not quite the mix and mingle case that you are trying to study. For your case I would recommend to look in the lit (if you haven’t done it) for simulations of a similar nature, doesn’t have to be zirconia, but something with the general flavor of the problem you are trying to tackle. Not sure how much of an expert in the field you are but for learning, imitation is the first step to success. Don’t need to be a scientist to know this, this is exactly what children do since very early stages and of course, a bit of neuroplasticity goes a long way. Maybe another Lammpster will have less general and more particular comments for your specific needs.

And of course, it is the fundamental of research to find similar (not identical) studies and try to imitate or improve the methods for your specific case (and if I could find one, then for sure, I would ask for that specific method I learned).

But I believe, what I asked for in this community, is simple in its nature. I have a hcp structure, that should transform into another lattice (that is bcc), when the temperature rises and passes a threshold value. The eam potential file used, clearly states that it is fitted for this phenomena.

So, I will appreciate for any comments that help me to find a solution.

Also, I asked for the possibility of creating structures with dual phases using LAMMPS (e.g. partially hcp and partially bcc), any comments?

Regards
Afshin

Ok. But why do you keep the volume constant when searching for the transition? Why not allowing the cell to change shape?
As to the dual phase creation that’s something to be done outside lammps. Remember that for the most part lammps is just an md engine. The input is on you.
Carlos

Ok. But why do you keep the volume constant when searching for the
transition? Why not allowing the cell to change shape?

​another term that is missing in this discussion is: "finite size effect".
​observing spontaneous and unnucleated phase transitions at typical MD
simulation system sizes is not at all easy.

True indeed and I should have noticed that in my very first email.
To Afshin, sometimes by making the box inertia small enough, one can get to
enhance the probability of observing a phase transition when running plain
vanilla md simulations. For equilibrium dynamics this is typically not
recommended as low box inertia can trigger wild fluctuations in the box
geometry and pressures. Yet, at the same time it allows the box coordinates
to quickly respond to structural instabilities such as the ones to be had
during phase transitions. How small is small enough? You have to play with
the params...
Carlos

Dear Carlos and Axel,

Thanks for your replies. I will play with parameters, lets see what happens… Also, I saw an article last night describing metadynamics to observe phase transition, which is apparently implemented in LAMMPS. I will investigate more, and if I get any result will comeback to you.

Regards
Afshin

Metadynamics on the cell vectors is one of the choices I had in mind when I mentioned other techniques to “induce” phase transitions. Last time I worked on it with Lammps (long ago) that was possible via the external plugin PLUMED which was distrubuted by other people and had to be installed independently. But I see there is now a fix colvars which probably can handle it.
Carlos