AQS simulations

[jx291605427]

Hi, congratulations on your work! Looking forward to your published articles.
My current research also focuses on the glass dynamic using the deep learning method. But my unfamiliarity with LAMMPS really hindered my research progress. Do you have any example Lammps input file about the athermal quasi-static (AQS) simulations (V. Bapst, https://doi.org/10.1038/s41567-020-0842-8)?

[nihil39]

Hi, I didn't perform the AQS part of that article, sorry. I can give you basic LAMMPS files on how to set the thermal simulations.

Regarding that article, did you have a theoretical background on how the glassy phase temperature (T =0.50) was estimated?

Thank you

[jx291605427]

I think the glassy phase transition temperature was estimated from the E-T (Energy versus Temperature) curve or V-T (Volume versus Temperature) during the cooling process. The temperature of the transition point (transition of two slopes) on the curve can be identified as the phase transition temperature. Hope this helps.

[nihil39]

Thanks, I presumed that but I can't find information in the article and I didn't get a reply from the authors.

For the AQS simulation maybe in the future I will investigate further, for now I'm just applying a different NN on the article's dataset (well, a subset of that dataset). I will let you know.

Other questions:

1. What about the highest state point T = 0.56, they start the simulation at T = 0.5495 which is lower than the target temperature? I can't understand if there is a glassy transition or not for that configuration.
2. Is there a way to estimate the relaxation time without using the self intermediate scattering function but only the mean square displacement? I'm only using the msd in my simulations

[jx291605427]

OK, look forward to your follow-up work.

Can you give me some information about the research team you belong to？ I recently plan to find a research team to apply for being a visiting student. I hope to discuss with you more deeply if possible.

Reply to your questions:

1. I think the simulation starts at T = 0.5495 only when the target temperature is lower than T = 0.5495. For the target temperature above T = 0.5495, the equilibration process just runs at that temperature. The glassy phase only occurs below the glassy phase temperature (T =0.50).
2. I am afraid not. As far as I know, the published papers all used the self intermediate scattering function.

Thanks.

[nihil39]

I am currently writing my M.Sc. thesis, the main focus is on the Neural Network applied on the dataset generated by the article we are discussing. We aim to predict the mean square displacement for type A particles.

I will let you know about future research teams if I know something.

I have some extra question:

1. To your knowledge, is the cooling in the isothermal-isobaric (NPT) ensemble strictly necessary to the glass formation?

2. I managed to get stable starting configuration at the desired Temperature and Pressure following the isothermal-isobaric process described in the article, but I don't understand this part "We then switch to a canonical ensemble to further stabilize the temperature and pressure and obtain an equilibrated system.". In the canonical ensemble (NVT) the pressure obviously varies and cannot be fixed. I reach my T,P point with the NPT ensemble and then I evolve the system with the NVT ensemble which makes the pressure change? This is not a key point I guess, because msd values are the expected one but still a question.

Thank you

PS

If you let me know your name (in private) I can add you to the acknowledgements of my thesis

[jx291605427]

This question also puzzled me!

I guess the NPT ensemble is not necessary when the density (N/L^3) is given （e.g. Schoenholz. et. al, A structural approach to relaxation in glassy liquids. ）.

The main purpose of NPT is to adjust the density of the system. After the density is stable, the system is switched to the NVT ensemble, and continue to adjust the temperature to obtain the equilibrium phase. Although the pressure changes in the NVT ensemble, the density of the system is constant.

Hence， when the density is given, only the NVT ensemble is necessary.

Hope this help.

We can discuss this more in detail by email. My email address: xjiang@hnu.edu.cn