[QE-users] Oscillating volume in VC-MD

[Nicola Marzari]

Dear Andrii,

maybe the small oscillations are not so important, depending on what you 
want to achieve. If you are really keen on investigating further, you 
could do a fairly short run using MD with fixed cell, and average the 
stress tensor over that run - 1/3 of the trace would tell you how 
compressed or expanded your cell seems to be. Then, knowing the bulk 
modulus (ok to use the zero-T one you get from an equation of state) you 
would know how much to isotropically strain the cell so that on average 
it will not want to contract or expand.

nicola


On 22/11/2021 19:49, Andrii Shyichuk wrote:
> Dear Nikola,
> 
> Thank you for the suggestion.
> 
> The system is a lutetium oxide cell of 80 atoms, a moderately symmetric 
> crystal.
> 
> You are absolutely right, the initial geometry has the pressure 
> components of 0.01-0.03 kbar.
> Does it make sense to try and optimize it until the pressure is zero?
> 
>  From this paper I can tell that Nose-Hover approach is a kind of 
> thermostat, which (apparently) is not implemented in QE for VC-MD.
> https://aip.scitation.org/doi/pdf/10.1063/1.463940
> I will look into it, but I was hoping for a QE-based solution.
> 
> Also, I've found the Wentzcovitch paper to be a very interesting read,
> https://journals.aps.org/prb/pdf/10.1103/PhysRevB.44.2358
> 
> Here it is stated that Wentzcovitch cell dynamics uses the Lagrangian 
> from that paper:
> https://lists.quantum-espresso.org/pipermail/users/2007-August/007052.html
> 
> I've ran some tests with Wentzcovitch cell dynamics to see if it helps.
> Apparently, not quite: the cell also oscillates, albeit at a much lower 
> frequency.
> Actually, the Wentzcovitch dynamics (cell_dynamics='w') n_step-volume 
> curve looks like an oscillating envelope of the Parrinello-Rahman 
> dynamics (cell_dynamics='pr') n_step-volume curve.
> 
> Best regards.
> Andrii
> 
> 
> 
>> Date: Thu, 18 Nov 2021 17:21:10 +0100
>> From: Nicola Marzari <nicola.marzari at epfl.ch>
>> To: users at lists.quantum-espresso.org
>> Subject: Re: [QE-users] Oscillating volume in VC-MD
>> Message-ID: <93c39f9e-473d-1c31-c0bb-00eb19123c36 at epfl.ch>
>> Content-Type: text/plain; charset=UTF-8; format=flowed
>>
>>
>>
>> Dear Andrii,
>>
>>
>> this might be related to the fact that the more harmonic a system is,
>> the more difficult it is to thermalize it - since it's anharmonicity
>> that shuffles excitations.
>>
>> [If you have a system of harmonic oscillators, the populations are good
>> quantum numbers, i.e. they are constant of motions. A good starting
>> point is the literature about Nose-hoover chains, disucssing these
>> difficulties in thermalization.]
>>
>> In your case, you have low-ish temperatures, and you start from a cell
>> that has probably a small non-zero pressure, so it oscillates for a long
>> time like a pendulum around its equilibrium volume, with very weak
>> couplings.
>>
>>                 nicola
>>
>>
>>
>> On 18/11/2021 16:45, Andrii Shyichuk via users wrote:
>>> Dear Users,
>>>
>>> I am running a VC-MD calculation where the cell volume oscillates in a
>>> nice sinusoidal way, with the period of about 250 fs.
>>> The oscillations are in the range of 99.5-101.5% of the initial volume
>>> for 300K, and within 99-102.5% for 500K.
>>>
>>> That happens both with 5 fs (dt=103.353) and 0.5 fs (dt=10.33534) time
>>> steps.
>>> The tolp was 10 in the first case and 2 in the second case.
>>>
>>> I use ion_temperature = 'rescaling',? cell_dynamics = 'pr', ion_dynamics
>>> = 'beeman' and nraise = 1.
>>>
>>> I've tried searching the mailing archive and found nothing relevant.
>>>
>>> Is it normal? Shouldn't the volume converge to a certain value?
>>>
>>>
>>> Thank you in advance.
>>> Best regards.
>>> Andrii Shyichuk, University of Wroc?aw
>>> _______________________________________________
>>> 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
>>
>>
>> -- 
>> ----------------------------------------------------------------------
>> Prof Nicola Marzari, Chair of Theory and Simulation of Materials, EPFL
>> Director, National Centre for Competence in Research NCCR MARVEL, SNSF
>> Head, Laboratory for Materials Simulations, Paul Scherrer Institut
>> Contact info and websites at http://theossrv1.epfl.ch/Main/Contact
>>
>>


-- 
----------------------------------------------------------------------
Prof Nicola Marzari, Chair of Theory and Simulation of Materials, EPFL
Director, National Centre for Competence in Research NCCR MARVEL, SNSF
Head, Laboratory for Materials Simulations, Paul Scherrer Institut
Contact info and websites at http://theossrv1.epfl.ch/Main/Contact