[QE-users] Unit for the output of average.x

Dingfu Shao dingfu.shao at gmail.com
Fri Oct 26 21:53:52 CEST 2018


Hi, Giovanni,

Yes, I confirm the fermi energy should be right. I actually checked this
twice. In the scf and nscf calculations, I used occupation = 'fixed', and
got the " highest occupied level (ev):    -1.0818" from the output of scf.
Then for the calculation of pp.x, I did two tests:

1. set emin = -65 and emax = -0.5 for plot_num =10

2. set emin = -65 and not set emax, since according to the input
description, for plot_num =10, if emax is not specified, emax=E_fermi.

Both tests leads to same result.

Moreover, since I used occupation =  'fixed', no states above EF is
calculated. Therefore, even if I set a wrong EF, it should only leads to a
smaller integrated value compared to the total number of electrons. But now
what I got a  integrated value of 29, which is much larger than the 26
electrons in MoS2 monolayer.

For your convenience, I put my input files here. I used QE-6.3 for those
calculations.

MoS2.nscf.in:


 &control
    calculation='nscf'
    restart_mode='from_scratch',
    prefix='MoS2',
    pseudo_dir = '../pp',
    outdir='./tmp/'
    disk_io='low'
       /
 &system
    ibrav=4 ,
    a = 3.13079 ,
    c = 20,
    nat = 3 , ntyp= 2,
    ecutwfc= 60.0,
    ecutrho= 600.0,
    occupations='fixed',
 /
 &electrons
    electron_maxstep=300
    conv_thr =  1.0d-7
    mixing_beta = 0.7
 /

&ions
   /
 &cell
 /

ATOMIC_SPECIES
 Mo 95.94  Mo.pz-spn-rrkjus_psl.0.2.UPF
 S  32.07 S.pz-n-rrkjus_psl.0.1.UPF


ATOMIC_POSITIONS (crystal)
Mo       0.000000000   0.000000000   0.000000000
S        0.333333333   0.666666667   0.077787748
S        0.333333333   0.666666667  -0.077787748

K_POINTS {automatic}
100 100 1 0 0 0



MoS2.pp.in:

&inputpp
    prefix='MoS2',
    outdir='../tmp',
    filplot = 'MoS2.charge'
    plot_num= 10
    emin = -65
/


MoS2.avg.in:

1
MoS2.charge
1.D0
2000
3
37.7945


Note the awin=37.7945 in MoS2.avg.in might not be appropriate, but it will
not influence the planar average.


Thank you very much.

Ding-Fu



> *Giovanni Cantele* giovanni.cantele at spin.cnr.it
>
> <users%40lists.quantum-espresso.org?Subject=Re:%20Re%3A%20%5BQE-users%5D%20Unit%20for%20the%20output%20of%20average.x&In-Reply-To=%3CAE0AD77F-2745-4B15-9DF1-E8491D64AAF9%40spin.cnr.it%3E>*Fri
> Oct 26 21:08:23 CEST 2018*
>


> Are you sure about you Fermi level? How did you get it?
>


> GIOVANNI
> Inviato da iPhone

>* Il giorno 26 ott 2018, alle ore 19:33, Dingfu Shao <dingfu.shao at gmail.com <https://lists.quantum-espresso.org/mailman/listinfo/users>> ha scritto:
*

>

>* Dear Giovanni and Paolo,
*

>

>* Thanks very much for your suggestions.
*

>

>* I reconsidered the definition of the local density of states (say, LD(x,y,z,E)). Since it is "local", the unit of it should be states/eV/bohr^3 or electrons/eV/bohr^3. Therefore, the integration of it within an energy window should lead to the charge density in this energy window: electrons/bohr^3 . Therefore, if we choose the energy window from the lowest energy to the Fermi energy, we should get exactly the total charge density.
*

>

>* So I did some tests following Giovanni's suggestion, using a simple case of momolayer MoS2, which has "number of electrons  =   26.00". The area of its xy plane is S. For those tests, the previous scf and nscf calculations are the same, with LDA USPP,  the occupation= 'fixed', and k-points of 40*40*1 for scf and 100*100*1 for nscf.
*

>

>* 1. I  calculated the total charge density (rho(x,y,z)) using pp.x with plot_num = 0, and then calculate the planar average of it (rho_avg(z)). Then I integrated it by \int (S* rho_avg(z)) dz and I got 25.89. It is close to 26.00, maybe a more accurate value can be obtained by a calculation with denser k points.
*

>

>* 2. I calculated the integrated local density of states  ( ILD(x,y,z)) from -65 eV (this energy is below the lower band) to Fermi energy using pp.x with plot_num =10, and and then calculate the planar average of it ( ILD_avg(z) ). When I integrated it by \int (S* ILD_avg(z)) dz, I got 29.25, which is much larger than the total number of electrons of 26.
*

>

>*  So it seems the the test 2 doesn't correctly reflect the  reality. I am not sure it is due to something happened with plot_num = 10 in pp.x, or just I understand this incorrectly.
*

>

>* Any suggestions? Thank you very much!
*

>

>* Best,
*

>

>* Ding-Fu
*

>

>

>

>

>

>

>>* From: Giovanni Cantele <giovanni.cantele at spin.cnr.it <https://lists.quantum-espresso.org/mailman/listinfo/users>>
*

>>* To: Quantum Espresso users Forum <users at lists.quantum-espresso.org <https://lists.quantum-espresso.org/mailman/listinfo/users>>
*

>>* Cc:
*

>>* Bcc:
*

>>* Date: Fri, 26 Oct 2018 09:51:40 +0200
*

>>* Subject: Re: [QE-users] Unit for the output of average.x
*

>>* Dear Ding-Fu,
*

>>* as far as I remember there is a surface factor that you need to adjust units.
*

>>* For sure on the ascissa axis the coordinate is in bohr.
*

>>* The planar average give you back a quantity with the same units as the averaged quantity
*

>>* (e.g. if you star from charge density in electrons/bohr^3 you get an averaged electron density in electrons/bohr^3),
*

>>* being defined as (let us suppose that you average in the plane defined by a1 and a2 vectors):
*

>>* rho_avg(z) = ( 1 / S ) * integral( dx dy rho(x,y,z) )
*

>>* That means that if you perform
*

>>*  integral( dz rho_avg(z) )
*

>>* you get
*

>>* number of electrons / S
*

>>* If you need number of electrons than just multiply by S with
*

>>* S = cross_product( a1, a2 )
*

>>* (in bohr^2)
*

>>* Just try, better if you do it with the total charge density, to check if the integral returns you
*

>>* the number of electrons.
*

>>* I’m sorry but I cannot check directly if I remember correctly at the moment, but
*

>>* this should work.
*

>>* Giovanni
*

>>* --
*

>>* Giovanni Cantele, PhD
*

>>* CNR-SPIN
*

>>* c/o Dipartimento di Fisica
*

>>* Universita' di Napoli "Federico II"
*

>>* Complesso Universitario M. S. Angelo - Ed. 6
*

>>* Via Cintia, I-80126, Napoli, Italy
*

>>* e-mail: giovanni.cantele at spin.cnr.it <https://lists.quantum-espresso.org/mailman/listinfo/users>
*

>>*             gcantele at gmail.com <https://lists.quantum-espresso.org/mailman/listinfo/users>
*

>>* Phone: +39 081 676910
*

>>* Skype contact: giocan74
*

>>* Web page: https://sites.google.com/view/giovanni-cantele <https://sites.google.com/view/giovanni-cantele>
*

>>* On 26 Oct 2018, at 03:31, Dingfu Shao <dingfu.shao at gmail.com <https://lists.quantum-espresso.org/mailman/listinfo/users>> wrote:
*

>>* Dear QE developers and users:
*

>>* I am wondering what should be the unit of the planar average data got from the average.x
*

>>*  I am calculating the planar average of charge density within a energy window. What I did is firstly using pp to get the integrated local density of states (ILDOS) of that energy window with plot_num=10, then using average.x to get the planar average.
*

>>* In this case, what is the unit of the second column (say, rho(z)) of  the output file? I thought since the DOS has a unit of states/eV, the integration of DOS within a energy window should get some states or electrons. Then the unit of rho(z) should be electron/bohr. But seems it is not. In my case the energy window I concerned contains one electron, However, if I directly integrate  rho(z), I can only get a very small value. If I assume the unit is electron/(bohr^3), the integretion of  rho(z)*A is also smaller than one (here A is the area of xy plane).
*

>>* Can you help me about it? Thank you very much!
*

>>* Best,
*

>>* Ding-Fu
*

>>

>>

>>* Ding-Fu Shao, Ph. D.
*

>>* Department of Physics and Astronomy, University of Nebraska-Lincoln
*

>>* Lincoln, NE 68588-0299
*

>>* Email: dingfu.shao at gmail.com <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>
*

>>* https://lists.quantum-espresso.org/mailman/listinfo/users <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>
*

>* https://lists.quantum-espresso.org/mailman/listinfo/users <https://lists.quantum-espresso.org/mailman/listinfo/users>*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20181026/7e807013/attachment.html>


More information about the users mailing list