[QE-users] Negatively charged isolated molecule

Tue Mar 19 14:16:32 CET 2019

Dear Ernane

> Is there a way of estimating this Jellium contribution and somehow discount
> it from the total energy?

> Is this necessary?

Periodic boundary conditions are fantastic for a lot of things (FFT,  
plane waves, scaling, etc,), but particular care, which is not  
necessary in the case of typical GTO calculations, must be given to  
the calculation of electrostatic energy. In particular, the efficient  
Ewald partition of Coulomb interactions in PBC diverges in the case of  
charged species. a diffuse countercharge (aka jellium) is therefore  
necessary, but it is also a source of error, as shown by Makov and  
Payne a lot of time ago.

> The assume isolated schemes deal with this and already provide a correct
> energy?

There are a lot of correction schemes, but AFAIK it is not possible to  
calculate "additive" total energy values in PBC for charged species,  
that is, additive in the sense that the energy difference of A and A+  
calculated in a given supercell is a meaningful estimate of the  
ionization energy of A.

If you want to read something, this (and the references therein) may  
be a good starting point

PHYSICAL REVIEW B 77, 115139, 2008

HTH
Giuseppe

Ernane de Freitas Martins <ernanefmg at gmail.com> ha scritto:

> Dear all,
>
> regarding energy calculations of charged systems in general, I also have
> doubts about the Jellium contribution to the total energy. The unique
> mention to Jellium I found is the phrase in the input description saying
> that a background Jellium is added for charged calculations. Thus I have
> some doubts.
>
> Is there a way of estimating this Jellium contribution and somehow discount
> it from the total energy?
>
> Is this necessary?
>
> The assume isolated schemes deal with this and already provide a correct
> energy?
>
> As I have mentioned, I didn't find further information about Jellium in QE
> documentation.
>
> Could some one clarify that to us or suggest a paper on this?
>
> Cheers,
>
> Dr. Ernane de Freitas Martins
> Postdoctoral researcher
> IF - USP
> São Paulo, SP - Brazil
>
> Em seg, 18 de mar de 2019 06:28, Laurens Siemons <laurenssiemons at hotmail.be>
> escreveu:
>
>> Dear QE-users,
>>
>> I always thought that it is not correct to calculate energies of isolated
>> ionic species under PBC due to the introduction of a Jellium background
>> inside the vacuum  which has physically no meaning. But after reading this
>> post I assume that I am wrong and that you can perform calculations on
>> ionic species in a vacuum with QE?
>>
>> With kind regards,
>> Laurens Siemons
>> PhD, UAntwerp (Belgium)
>> ------------------------------
>> *Van:* users <users-bounces at lists.quantum-espresso.org> namens Nattino
>> Francesco <francesco.nattino at epfl.ch>
>> *Verzonden:* zaterdag 16 maart 2019 9:02
>> *Aan:* Quantum Espresso users Forum
>> *Onderwerp:* Re: [QE-users] Negatively charged isolated molecule
>>
>> Dear Ernane,
>>
>>  As Giuseppe already pointed out, many anionic species are actually
>> unbound with standard density functionals. The continuum solvation model
>> helps to achieve convergence because the dielectric embedding stabilizes
>> the localized electronic configuration.
>>
>> A way to circumvent the issue and to obtain the energy of carbonate in
>> vacuum could be the following: you calculate the energy of the system for
>> decreasing values of the dielectric constant and you extrapolate the energy
>> to the vacuum dielectric constant (epsilon=1).
>>
>> Best regards,
>>
>> Francesco Nattino,
>> EPFL
>>
>> On Mar 15, 2019 7:30 PM, Michal Krompiec <michal.krompiec at gmail.com>
>> wrote:
>>
>> Dear Ernane,
>> Have you thought of using a more sophisticated method (like GW) on [CO3]-
>> to calculate its EA? This would give you the energy of [CO3]2- in vacuum.
>> Best,
>> Michal Krompiec
>> University of Southampton & Merck KGaA
>>
>> On Fri, 15 Mar 2019 at 18:22, Ernane de Freitas Martins <
>> ernanefmg at gmail.com> wrote:
>>
>> Dear Giuseppe,
>>
>> I really appreciate your answer. Thank you very much for using your time
>> to answer my question.
>>
>> I'll think on your suggestion about trying hybrid functionals. The point
>> is that I need to estimate the solvation energy for carbonate ion using the
>> environ module, then I'll need to run a vacuum calculation using the same
>> functional I'm already using rVV-10).
>>
>> Thank you again for replying.
>>
>> Atenciosamente,
>>
>>
>> Dr. Ernane de Freitas Martins
>> Postdoctoral researcher
>> IF - USP
>> São Paulo, SP - Brazil
>>
>> Em sex, 15 de mar de 2019 15:04, Giuseppe Mattioli <
>> giuseppe.mattioli at ism.cnr.it> escreveu:
>>
>>
>> Dear Ernane
>> Your question contains part of the answer! Carbonate ion (CO3 2-) is
>> not stable outside water, and calculations of its properties in gas
>> phase are likely not so meaningful, but in the case of model
>> thermodynamics cycles (e.g. Born-Haber). The excess negative charge is
>> unbound when not stabilized by a strongly polar solvent, and this is
>> likely responsible for instabilities in the construction of the
>> Kohn-Sham potential along scf iterations. Moreover, this happens on
>> top of the strong delocalization error you experience when you use a
>> standard GGA exchange-correlation functional, when the
>> self-interaction of strongly localized electrons in the J[n] Coulomb
>> potential is not cancelled by a same term in the semi local exchange
>> potential. You may minimize this latter source of error by using a
>> hybrid GGA-EXX functional such as B3LYP, where the non local
>> Hartree-Fock part of the exchange functional can recover part of the
>> delocalization error, but you are not free yet from the instability of
>> carbonate in gas phase.
>> HTH
>> Giuseppe
>>
>> Ernane de Freitas Martins <ernanefmg at gmail.com> ha scritto:
>>
>> > Hello,
>> >
>> > I'm experiencing a problem to run a negatively charge molecule in quantum
>> > espresso. The system is CO32-.
>> >
>> > I try both vacuum and solvated (environ) calculations. The solvated one
>> > works fine.
>> >
>> > The problem is the calculation in vacuum. It never give the first ionic
>> > step because the SCF accuracy never reaches the convence criterion.
>> >
>> > I tried many different solutions (increase cutoffs and box size, use
>> assume
>> > isolated, decreasing and changing the mixing scheme and etc) and nothing
>> > works.
>> >
>> > The unique calculation that works fine for vacuum is the one with a box
>> > size of 7.9 x 7.9 x 7.9 A. I really don't understand why it only works
>> for
>> > this specific box size.
>> >
>> > I ran several other charged systems (+1, +2 and -1 total charge) and all
>> of
>> > them worked fine. The problem appears for -2 total charge in vacuum.
>> >
>> > Would some of you kindly help me in this?
>> >
>> > Cheers,
>> >
>> > Dr. Ernane de Freitas Martins
>> > Postdoctoral researcher
>> > IF - USP
>> > São Paulo, SP - Brazil
>>
>>
>>
>> GIUSEPPE MATTIOLI
>> CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
>> Via Salaria Km 29,300 - C.P. 10
>> I-00015 - Monterotondo Scalo (RM)
>> Mob (*preferred*) +39 373 7305625
>> Tel + 39 06 90672342 - Fax +39 06 90672316
>> E-mail: <giuseppe.mattioli at ism.cnr.it>
>>
>> _______________________________________________
>> users mailing list
>> users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
>>
>> _______________________________________________
>> users mailing list
>> users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
>>
>>
>>
>> On Mar 15, 2019 7:30 PM, Michal Krompiec <michal.krompiec at gmail.com>
>> wrote:
>> Dear Ernane,
>> Have you thought of using a more sophisticated method (like GW) on [CO3]-
>> to calculate its EA? This would give you the energy of [CO3]2- in vacuum.
>> Best,
>> Michal Krompiec
>> University of Southampton & Merck KGaA
>>
>> On Fri, 15 Mar 2019 at 18:22, Ernane de Freitas Martins <
>> ernanefmg at gmail.com> wrote:
>>
>> Dear Giuseppe,
>>
>> I really appreciate your answer. Thank you very much for using your time
>> to answer my question.
>>
>> I'll think on your suggestion about trying hybrid functionals. The point
>> is that I need to estimate the solvation energy for carbonate ion using the
>> environ module, then I'll need to run a vacuum calculation using the same
>> functional I'm already using rVV-10).
>>
>> Thank you again for replying.
>>
>> Atenciosamente,
>>
>>
>> Dr. Ernane de Freitas Martins
>> Postdoctoral researcher
>> IF - USP
>> São Paulo, SP - Brazil
>>
>> Em sex, 15 de mar de 2019 15:04, Giuseppe Mattioli <
>> giuseppe.mattioli at ism.cnr.it> escreveu:
>>
>>
>> Dear Ernane
>> Your question contains part of the answer! Carbonate ion (CO3 2-) is
>> not stable outside water, and calculations of its properties in gas
>> phase are likely not so meaningful, but in the case of model
>> thermodynamics cycles (e.g. Born-Haber). The excess negative charge is
>> unbound when not stabilized by a strongly polar solvent, and this is
>> likely responsible for instabilities in the construction of the
>> Kohn-Sham potential along scf iterations. Moreover, this happens on
>> top of the strong delocalization error you experience when you use a
>> standard GGA exchange-correlation functional, when the
>> self-interaction of strongly localized electrons in the J[n] Coulomb
>> potential is not cancelled by a same term in the semi local exchange
>> potential. You may minimize this latter source of error by using a
>> hybrid GGA-EXX functional such as B3LYP, where the non local
>> Hartree-Fock part of the exchange functional can recover part of the
>> delocalization error, but you are not free yet from the instability of
>> carbonate in gas phase.
>> HTH
>> Giuseppe
>>
>> Ernane de Freitas Martins <ernanefmg at gmail.com> ha scritto:
>>
>> > Hello,
>> >
>> > I'm experiencing a problem to run a negatively charge molecule in quantum
>> > espresso. The system is CO32-.
>> >
>> > I try both vacuum and solvated (environ) calculations. The solvated one
>> > works fine.
>> >
>> > The problem is the calculation in vacuum. It never give the first ionic
>> > step because the SCF accuracy never reaches the convence criterion.
>> >
>> > I tried many different solutions (increase cutoffs and box size, use
>> assume
>> > isolated, decreasing and changing the mixing scheme and etc) and nothing
>> > works.
>> >
>> > The unique calculation that works fine for vacuum is the one with a box
>> > size of 7.9 x 7.9 x 7.9 A. I really don't understand why it only works
>> for
>> > this specific box size.
>> >
>> > I ran several other charged systems (+1, +2 and -1 total charge) and all
>> of
>> > them worked fine. The problem appears for -2 total charge in vacuum.
>> >
>> > Would some of you kindly help me in this?
>> >
>> > Cheers,
>> >
>> > Dr. Ernane de Freitas Martins
>> > Postdoctoral researcher
>> > IF - USP
>> > São Paulo, SP - Brazil
>>
>>
>>
>> GIUSEPPE MATTIOLI
>> CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
>> Via Salaria Km 29,300 - C.P. 10
>> I-00015 - Monterotondo Scalo (RM)
>> Mob (*preferred*) +39 373 7305625
>> Tel + 39 06 90672342 - Fax +39 06 90672316
>> E-mail: <giuseppe.mattioli at ism.cnr.it>
>>
>> _______________________________________________
>> users mailing list
>> users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
>>
>> _______________________________________________
>> users mailing list
>> users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
>>
>> _______________________________________________
>> users mailing list
>> users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users

GIUSEPPE MATTIOLI
CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
Via Salaria Km 29,300 - C.P. 10
I-00015 - Monterotondo Scalo (RM)
Mob (*preferred*) +39 373 7305625
Tel + 39 06 90672342 - Fax +39 06 90672316
E-mail: <giuseppe.mattioli at ism.cnr.it>