[Pw_forum] Yet another question about graphene bandstructure

Hadi Arefi hadi.arefi at tyndall.ie
Tue Mar 11 02:12:07 CET 2014


Hi Julia,

I mean the manual K-points for calculating the bands along high symmetry points: For example from Gamma to K you should move with a finer step. You have just 3 points for each directions if I am right which is not enough. Go for 10-15 points at least. Also for the scf automatic k-points, you must do a convergence test not just pick a random number!

To know why you cannot stay with default ecutrho for any type of pseudo-potential, you need to study a bit about the basics of QE and different type of pseudo-potentials and their differences (I assume you did but maybe you need more). It might be ok in some cases to use default value but only in the case of norm conserving pseudos it's a right guess by default. There are sometimes suggestive values for rcut/ecutro in the pseudo file and in your case the values of 37/146 are suggested.  

 Before start studying any particular system, you need to do convergence test for all system-dependent parameters. Choose logical guesses(not any values) might be ok for start or test but definitely not good for do comprehensive studies.


Best,
H.

________________________________________
From: pw_forum-bounces at pwscf.org [pw_forum-bounces at pwscf.org] on behalf of jbobak [jbobak at uvic.ca]
Sent: Monday, March 10, 2014 10:35 PM
To: PWSCF Forum
Subject: Re: [Pw_forum] Yet another question about graphene bandstructure

Hi Hadi,

Thank you so much for all the help!

Two quick questions, when you say the K-mesh is not dense enough do you
mean in the original SCF calculation (18x18x1) or do you mean I should
choose more points along the Gamma->K->M->Gamma path?

I have tried a variety of values for ecut (37, 47, 57, 67), but I have not
been specifying ecutrho -- why is the default value not appropriate in
this case? I thought 4 x ecut was ok for pseudopotentials that are not
ultrasoft.

Thank you again,
Julia

> Dear Julia,
>
> They are of course in the reciprocal space no matter you want them to be
> or not! If your K-mesh is dense enough (which is not in your current
> input) and along the right paths (Which still aren't:
> GAMMA-to-M-to-K-to-GAMMA), then there are some parameters you need to
> check such as smearing/K-points/ecut/ecutrho convergence.
>
> Your smearing looks high to me and your ecut is too much for a PAW pseudo
> potential that you are using. Also you need to specify ecutrho as well
> because the default value is used in the case of norm conserving one.
>
> HTH.
>
> Hadi.
>
> -----Original Message-----
> From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On
> Behalf Of jbobak
> Sent: 10 March 2014 18:14
> To: PWSCF Forum
> Subject: Re: [Pw_forum] Yet another question about graphene bandstructure
>
> Hi Hadi,
>
> Thank you for your quick response! I have tried that K-path also (I remain
> confused as to whether K-paths are in real or reciprocal space, so I tried
> both) with a similar result.
>
> Julia
>
>> Dear Julia,
>>
>> I guess your K-path is not entirely coincides with the high symmetry
>> points for graphene. Make sure you span the following routes with
>> enough
>> K:
>>
>> 0    0       0
>> 0.666        0.333   0
>> 0.5  0.5     0
>> 0    0       0
>>
>>
>> HTH,
>>
>> Hadi.
>> -----Original Message-----
>> From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org]
>> On Behalf Of jbobak
>> Sent: 10 March 2014 17:04
>> To: pw_forum at pwscf.org
>> Subject: [Pw_forum] Yet another question about graphene bandstructure
>>
>> Hi all,
>> I've searched the forum for this question, but nobody seems to be
>> having quite the same issue, so I expect I'm doing something uniquely
>> stupid. I am trying to calculate the band structure of graphene (using
>> Espresso
>> 5.0.3) for practice before I use a more complicated system. I do an
>> SCF calculation with just the unit cell (2 atoms) with automatic
>> k-points 18
>> 18 1 0 0 0. Then I do a band structure calculation with k-points as
>> follows:
>>
>> K_POINTS {tbipa}
>> 7
>> 0.00000 0.00000 0.00000 1
>> 0.28868 0.16666 0.00000 1
>> 0.57735 0.33333 0.00000 1
>> 0.43301 0.41666 0.00000 1
>> 0.28868 0.50000 0.00000 1
>> 0.14435 0.25000 0.00000 1
>> 0.00000 0.00000 0.00000 1
>>
>> I then run bands.x, followed by plotband.x which I run interactively.
>> Everything works fine in that I get no errors, but the bandstructure
>> plot is wrong (see attached pdf). The bands should cross at the K
>> point. I think the problem has something to do with the k-points,
>> which is why I have reproduced them above, but I have also attached
>> the relevant input and output files. Could it be because the
>> high-symmetry k-points I selected are not present in the automatically
>> selected grid?
>>
>> Thank you so much for any help,
>> Julia Bobak
>>
>> Research Assistant
>> University of Victoria, Victoria, BC, CANADA
>>
>> _______________________________________________
>> Pw_forum mailing list
>> Pw_forum at pwscf.org
>> http://pwscf.org/mailman/listinfo/pw_forum
>>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>


_______________________________________________
Pw_forum mailing list
Pw_forum at pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum




More information about the users mailing list