Re: [QE-users] Convergence with interstitials

[Mostafa Youssef via users]

Dear Tejas,

There are many tests to check and test here.  But one thing that comes to
my mind is the pseudopotentials you used. you could try to use another set
of pseudopotentials (especially for N). Try the recommended set here:
https://www.materialscloud.org/discover/sssp/table/efficiency.

Also one needs to make sure that W lattice constant is close to experiment.

Lastly, I would reduce the starting magnetization on W to 0.1 and on N to
0.3 or so.  No rationale here, just intuition.

Regards,
Mostafa
AUC
___
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users

Re: [QE-users] Convergence with interstitials

[Mostafa Youssef via users]

Dear Tejas,

Did you test giving an initial magnetic moment to N instead of W?

For example:

starting_magnetization(2)= 0.5

Regards,

Mostafa

AUC
___
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users

Re: [QE-users] a problem related to the electric field calculations

[Mostafa Youssef via users]

Dear Samare,


You are performing 'scf' calculation for Ti-doped Fe2O3 under electric
field.  This gives only the electron's response to the field.  If you are
interested in ions response (and in fact the ion-electron response) , you
may need to use 'relax'. Of course make sure to relax the zero field
Ti-doped hematite first before you start electric field calculations.

For plain PBE for Ti, you may end up with a metallic solution. Ti wants to
be in oxidation state +4 and the leftover electron may (or may not) spread
over the whole cell causing metallic solution. In this case of metallic
solution, electric field calculations will not converge.  You may check the
state of the last electron by plotting DOS  or visualizing the defect state
in 3D.

Also, the field is large. Add it gradually as I described in my previous
post.

I never used the pseudos you are using. But the ones here:
https://www.materialscloud.org/discover/sssp/table/efficiency
are usually very good. Although Fe pseudo is very computationally intensive.

If you change the occupations to 'smearing' and add a tiny smearing (0.004
Ry), this can accelerate the convergence. Again this is after you make sure
that the last electron in Ti-doped hematite is not leading to a metallic
solution.

Hope this helps


Regards,
Mostafa
AUC
___
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users

Re: [QE-users] a problem related to the electric field calculations

[Mostafa Youssef via users]

Dear Rostami,

As I explained in a recent thread, the raw dipole printed out by pw.x is
not meaningful. One needs to post-process the results. Please see here:
https://www.mail-archive.com/users@lists.quantum-espresso.org/msg39910.html

Few other observations:
1- Your k-points mesh does not look consistent with neither the hexagonal
cell of hematite nor the trigonal one.  [I assume you study alpha Fe2O3 and
not the gamma phase)
2-  I would apply the field gradually. Meaning start with a smaller value,
say 0.1 a.u., relax, and take the resulting structure and apply a
higher field , 0.2 a.u. and so on. Usually this works well for
defective systems.
3- From my experience nberrycyc = 1 or 2 is enough.
4- If you provide an input file, we can help more.

This paper and its supplemental materials can be helpful.

https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.119.126002



Regards,
Mostafa
AUC
___
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users

Re: [QE-users] Finite electric field polarisation

[Mostafa Youssef via users]

Dear Terry,

The dipole results need post-processing. One needs to subtract n x
polarization quantum to obtain a meaningful polarization.  n is an
integer.

Consider a simple dielectric like MgO at zero field. pw.x may give you
different dipoles in all directions, although physically in all directions
the net dipoles are zero. If you post-process as I indicated above you get
zero dipole as it should.

A good starter on this topic is this paper:
https://www.sciencedirect.com/science/article/abs/pii/S0022459612003234
A free arXiv version of it is also available.



Mostafa
AUC
___
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users

[QE-users] Convergence with interstitials

[Mostafa Youssef via users]

Dear Tejas,

I see two issues:

1) Your supercell size is too small. This corresponds to a high
concentration of N. I would go for 4x4x4 conventional cells of W. That is
128 atoms.

2) A magnetic moment on N is needed since the odd number of electrons is
because of N.


Regards,
Mostafa
AUC
___
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users

Re: [QE-users] magnetism in hematite (alpha-Fe2O3)

[Mostafa Youssef]

Dear Yasser,

Thank you for your response and suggestion. I agree that DFT+U would
improve the electronic structure of Fe-oxides, but the problem I'm facing
is more related to symmetry. But to confirm this, I did actually try a
quick DFT+U  calculation and I still have the same problem that some oxygen
ions in the conventional cell develop antiferromagnetic ordering.
 I would not be worried if all oxygen ions did this (both in the
conventional and primitive cells). But the fact that only some of the
oxygen ions (in the conventional cell only) exhibit this ordering is what
makes me concerned about the way I'm setting up the system. So still any
suggestions are welcome.


Regards,
Mostafa
___
users mailing list
users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users

[QE-users] magnetism in hematite (alpha-Fe2O3)

[Mostafa Youssef]

 Dear all

Hematite (alpha-Fe2O3) crystallizes in the corundum structure.  The
primitive cell is rhombohedral containing 2 formula units, whereas the
conventional cell is hexagonal containing 6 formula units. I'm trying to
model the antiferromagnetic phase of hematite with spin ordering (+--+) in
the primitive cell. Relaxation (or variable cell relaxation) works nicely
for the primitive cell and I obtain a reasonable final structure and
magnetic moments on iron ions and zero magnetic moment in oxygen ions. This
is consistent with literature. Note that all iron ions are equivalent
crystallographically and all oxygen ions are crystallographyically.
However, when I try to relax (or vc-relax) the corresponding conventional
cell, some symmetry breaking occurs and few iron ions develop magnetic
moment different than the rest. Additionally some oxygen ions develop
magnetic moment as well.  The final result is antiferromagnetic still, but
of course the structure is not consistent with prior reports and is not
consistent with primitive cell calculations.  Also the energy/Fe2O3 for the
conventional cell is higher than the primitive cell. I tried many things to
avoid this strange-looking local minimum such as:

1- Using cg instead of david algorithm.
2- Different starting magnetization
3- Different initial experimentally determined atom positions
4- Different pseudopotentials
5- Applying smearing
6- Gamma-centered vs. shifted k-points
7-Using ibrav = 4  and listing a and c, vs. using ibrav = 0 and providing
cell_parameters list

but nothing really seems to work.
Any suggestions are thoughts arevery appreciated. I provide below the input
files for both primitive and conventional cells and I'm using version
6.2.1. I also provide the final magnetization of the ions in the
conventional cell.

Best Regards,
Mostafa Youssef
AUC -  Egypt

# Input file for the primitive cell

   calculation =   'vc-relax'   ,
   verbosity = 'high' ,
   restart_mode = 'from_scratch' ,
   prefix='fe2o3' ,
   outdir='/data/mostafa/Fe2O3/fix_magnet_strucutre2/scratch/' ,
   pseudo_dir = '/home/mostafa/data/Fe2O3/PP' ,
   tstress = .true.  ,
   tprnfor= .true.  ,
   nstep = 2000 ,
   etot_conv_thr = 7.7D-6  ,
   forc_conv_thr = 4.0D-5 ,
 /
 
   ibrav = 0 ,
nat  =  10 ,
   ntyp  =  3 ,
   nbnd  = 96 ,
  ecutwfc  =  90 ,
  ecutrho  =  1080 ,
  nosym =.true.
  occupations = 'fixed',
  nspin = 2 ,
  starting_magnetization(1)=1.0,
  starting_magnetization(2)=-1.0,
  tot_magnetization = 0
 /
 
diagonalization='david',
mixing_mode = 'plain' ,
mixing_beta = 0.5,
startingwfc = 'random',
conv_thr  =  1.0d-9  ,
  /

ion_dynamics='bfgs'
/

   cell_dynamics = 'bfgs' ,
   cell_dofree = 'all',
   press_conv_thr = 0.5 , ! This is the default
 /
CELL_PARAMETERS  angstrom
 5.42000.0.
 3.087050004.454943000.
 3.087050001.616616004.15127400
ATOMIC_SPECIES
Fe1  55.845 Fe.pbe-spn-rrkjus_psl.0.2.1.UPF
Fe2  55.845 Fe.pbe-spn-rrkjus_psl.0.2.1.UPF
O15.999 O.pbe-n-rrkjus_psl.0.1.UPF
ATOMIC_POSITIONS crystal
Fe1   0.14500  0.14500  0.14500
Fe1   0.85500  0.85500  0.85500
Fe2   0.35500  0.35500  0.35500
Fe2   0.64500  0.64500  0.64500
O 0.75000  0.45800  0.04200
O 0.95800  0.25000  0.54200
O 0.45800  0.04200  0.75000
O 0.25000  0.54200  0.95800
O 0.04200  0.75000  0.45800
O 0.54200  0.95800  0.25000
K_POINTS automatic
  4 4 4  1 1 1

# Conventional cell


   calculation =   'vc-relax'   ,
   verbosity = 'high' ,
   restart_mode = 'from_scratch' ,
   prefix='fe2o3' ,
   outdir='/data/mostafa/Fe2O3/fix_magnet_conventional_str3_gamma/scratch/'
,
   pseudo_dir = '/home/mostafa/data/Fe2O3/PP' ,
   tstress = .true.  ,
   tprnfor= .true.  ,
   nstep = 2000 ,
   etot_conv_thr = 7.7D-6  ,
   forc_conv_thr = 4.0D-5 ,
 /
 
   ibrav = 0 ,
nat  =  30 ,
   ntyp  =  3 ,
   nbnd  = 176 ,
  ecutwfc  =  90 ,
  ecutrho  =  1080 ,
  !!!nosym =.true.
  occupations = 'fixed',
  nspin = 2 ,
  starting_magnetization(1)= 1.0,
  starting_magnetization(2)= -1.0,
  !!starting_magnetization(3)= 0.0,
  tot_magnetization = 0
 /
 
diagonalization='david',
mixing_mode = 'plain' ,
mixing_beta = 0.5,
startingwfc = 'random',
conv_thr  =  1.0d-9  ,
  /

ion_dynamics='bfgs'
/

   cell_dynamics = 'bfgs' ,
   cell_dofree = 'all',
   press_conv_thr = 0.5 , ! This is the default
 /
CELL_PARAMETERS  angstrom
5.0288000107 0.00 0.00
   -2.514453 4.3550685598 0.00
0.00 0.0013.7302999496

Re: [Pw_forum] Problems computing cholensky

[Mostafa Youssef]

Dear Laurens,


When I get this error, it is typically due to a parallelization issue.
A quick fix that worked for me is using  -ndiag 1 , so for example:

pw.x_location  -ndiag 1  output_file.

Check the manual and forums for more details.


Regards,
Mostafa Youssef
AUC, Egypt.
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Zr supercell

[Mostafa Youssef]

Dear Krishnendu,

If you model the alloy using a unit cell, then you are modeling a very high
Nb concentration ordered alloy.

Take HCP unit cell having two Zr atoms. If you replace one with Nb, then
you get 50%-50% Zr-Nb.

There is no way to avoid forming a supercell of Zr (Say 54 atoms) and then
replace one with Nb.


Regards,
AUC, New Cairo, Egypt
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Problems Computing Cholesky for VC-Relax of MnO2

[Mostafa Youssef]

Dear Lance,


When I get this error, it is typically due to a parallelization issue. A
quick fix would be using  -ndiag 1 , so for example:

pw.x_location  -ndiag 1  output_file.

Check the manual and forums for more details.

As a side, I highly doubt that  these  Mn and O  pseudopotentials converge
at 37 Ry, some testing may be needed to confirm.

Regards,
Mostafa
AUC, New Cairo, Egypt
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] SCF calculations for molecules in electric fieldsare not converged

[Mostafa Youssef]

Dear Jibiao,

If you would like to use modern theory of polarization (MTP) version of
applying the electric field and reach 1 Ry a.u., then you can do this
gradually. So converge 0.1 a.u., keep the wave functions, and use them as
initial guess to 0.2 a.u.. Repeat with steps of 0.1 a.u. until you reach 1
a.u.

An alternative would be using a classical saw-tooth potential to apply the
field using tefield tag (instead of lelfield). This may be a bit easier to
converge. I would experiment with both approaches.

The classical saw-tooth implementation allows smearing, whereas MTP
implementation does not. But you do not really need smearing for isolated
molecules, you can use fixed occupations or even occupations read from
input file. See example05 of PW examples.

If you, however, insist on applying smearing with MTP, you can comment out
the places in the PW/src files where there is an If statement to check that
this is an insulator (i.e. fixed occupations) and recompile the code.  You
need to check this carefully on your own. The key idea is that smearing is
sometimes useful with insulators and there is no harm in applying it with
MTP as long as the final solution remains an insulator. But at the time
when MTP was implemented in PWscf it was thought that smearing immediately
implies a metal which is of course problematic for MTP.  Again, I do not
recommend this option unless you really need it and you make sure that
things work consistently.

Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] SCF calculations for molecules in electric fields are not converged

[Mostafa Youssef]

Dear Jibiao Li,

A field of 1 Ry a.u. is giant and atypical, so one should not be surprised
do not converge.  Recall,  1 in Ry a.u.=36.3609*10^10 V/m.

I also noticed that you impose fixed occupations and then apply Gaussian
smearing. Those are in contradiction of each other. Apparently the code
enforces fixed occupations and ignores the smearing.


Mostafa
AUC, New Cairo, Egypt
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Total & absolute magnetization in 2-atom unit cell and corresponding supercell

[Mostafa Youssef]

Dear AT Raji,

A sample input file will help clarifying the issue.

My guess is that you are dealing with a system with complex potential
energy landscape and multiple local minima.  If you know that the total
magnetization has to be zero and you are using smearing, then try giving
small magentization as intial guess (start_magnetization). You can also
constrain the total magnetization by tot_magnetization =0.

Mostafa
AUC, New Cairo, Egypt
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Need help in spin polarized calculations

[Mostafa Youssef]

Dear Sayan,

In the case of metallic systems, it is recommended to use smearing. With the 
smearing applied, you can use start_magnetization to give an initial guess for 
the magnetic moment.  pw.x will try self-consistently to find the final 
magnetization that leads to minimum energy (Achieving both goals 
simultaneously).  No need for the tot_magnetization in this case.


Regards,
Mostafa

___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Need help in spin polarized calculations

[Mostafa Youssef]

Dear Sayan,


1- nspin=2 works whether you use LDA or GGA.

2- start_magnetization works when you apply smearing (occupation tag). Using 
this tag you give an initial guess and let the code to find the final 
magnetization associated with some nearby local energy minimum.  
tot_magnetization works with fixed occupation.  Here you ask the code to 
maintain the overall magnetization you specified. There is no guarantee that 
this will lead to the ground state. But it is useful to do many simulations 
each with certain total magnetization and compare their final energies.

You can learn a lot about the practical aspects of these tags/key words by 
going over the examples of pw.x and previous forum posts.

Regards,
Mostafa
AUC / MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Fwd: PAW pseudo for Mn

[Mostafa Youssef]

Dear Mohammadreza,


Have a look at this link:

http://materialscloud.org/sssp/


Also here:
http://qe-forge.org/gf/project/pslibrary/


Regards,
Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Phonon under pressure

[Mostafa Youssef]

Dear Subhodip,

I'm no expert in high pressure physics and one needs to know the specifics of 
the system.

How much pressure are you applying ?  If too much the whole pseudopotential 
approach cannot be very helpful.  Are you able to get reasonable phonons at 
zero pressure ? Do you anticipate phase transitions as a function of pressure ?
By clarifying these questions, you may get help form someone well-versed in 
these topics.


Regards,
Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Phonon under pressure

[Mostafa Youssef]

Dear Subhodip,

If you maintain the lattice constant resulting from vc-relax then the pressure 
effect should be there in your phonon calculations. You apply the pressure by 
constraining the volume.


Regards,
Mostafa
MIT

___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] tot_charge in nano structures

[Mostafa Youssef]

Dear Zara,

Have a look at :

http://journals.aps.org/prx/abstract/10.1103/PhysRevX.4.031044


Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Default color for xcrysden isosurface

[Mostafa Youssef]

Resolved.  Clicking the box "Render +/- isovalue"  restores the default red and 
blue colors.



Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Default color for xcrysden isosurface

[Mostafa Youssef]

Dear all,

This is an Xcrysden usage question, but since most QE users use Xcrysden, I'm 
posting here as well.

I'm using xcrysden version 1.5.60 on windows using cygwin.  The default color 
for 3D isosurface is yellow.  However, in older versions it was red (+) and 
blue (-).  For consistency purpose, I want to use the same red and blue  which 
were the default in previous versions. I would appreciate it if someone still 
has one of the older versions  and can tell me the exact RGB ingredients of the 
default red and blue.

For example the current default is yellow and has RGB parameters:
R: 1.00
G:1.00
B:0.20

Your help is appreciated
Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Charge of a supercell with a vacancy

[Mostafa Youssef]

Dear David,

In experiments , neutral O2 gas leaves the TiO2 crystal causing the creation of 
oxygen vacancies behind.  However, these vacancies can have any charge state.  
Neutral in which the two electrons (reminiscent of O2- ion) set around the 
vacancy forming polarons, 2+ in which the two electrons leaves the vacant site 
and go somewhere else, or 1+ which is intermediate between the past two cases.

Have a look at the following paper for theoretical analysis of the oxygen 
vacancy in TiO2:
J. Phys. Chem. C 2010, 114, 21694–21704


For an experimental point view, have a look at figure 2.4 (and the associated 
discussion) in the following thesis:

https://scholar.google.com.eg/citations?view_op=view_citation=en=1Ga4WncJ=pubdate_for_view=1Ga4WncJ:u-x6o8ySG0sC



Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] substitution of positive charge into a negative charge.

[Mostafa Youssef]

Dear Mulwa Winfred,

Substituting cations such as Nd3+ on anion site (O-site) in TiO2 is not 
expected to be favorable energetically. One good reason  comes from the simple 
ionic picture. Nd3+ will repel the surrounding Ti4+.

However, in spite of this I doubt the value of 634 eV.  Recall that calculating 
the formation energy of Nd defects in TiO2 requires a reference state for Nd; 
for example Nd2O3.

Regards,
Mostafa Youssef
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] External E-field Calculation for Bulk

[Mostafa Youssef]

Dear Mohammad,

Yes you can simulate electric field effect on bulk 3D material. In pw.x input 
description check lelfield = .true.
Also check these two papers:

http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.89.157602

http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.89.117602


Regards,
Mostafa
NU/MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Problem with charged defect calculation

[Mostafa Youssef]

Dear Evan,

The postprocessing utility pp.x can calculate and plot the average potential as 
a function of x or y or z. This can help in performing the potential alignment.



Regards
Mostafa Youssef
MIT

___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Relax charged system

[Mostafa Youssef]

Dear Nkosinathi,


Please send atomic positions and cell vectors and I can test it.



Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] vc-relax not converge

[Mostafa Youssef]

I'm not very familiar with ESM model.  But certainly modeling Na+ in water (or 
any insulator) requires tot_charge=+1.  For MD , or relaxation nothing else is 
needed.



Regards,
Mostafa




___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] vc-relax not converge

[Mostafa Youssef]

Also note that you have one sodium atom and you did not set tot_charge to +1. 
So you are actually modeling Na+ and an electron.  It is very difficult to 
predict the outcome of this calculation (metallic solution vs. insulating one) 
in a vc-relax calculation.  If your goal is to model Na+, then tot_charge=+1 
guarantees an insulating solution and as such no large smearing is needed.



Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] vc-relax not converge

[Mostafa Youssef]

Dear Nkosinathi,


What is exactly not converging , the electronic minimization, or the ionic?


Few things to note:

1- 90 H2O molecules +ion is 181 nat not 91.
2- It is a bit strange to use mp smearing typical for metals with liquid water. 
But it is unlikely that this is the cause of the problem. And also the smearing 
(in Ry units) is very large for an insulating system like water.  Try 0.004 Ry.
3- Why 2Dxy ? For this glassy system I would relax all degrees of freedom.
4- I think ecutrho = 4 ecutwfc is a bit small for PAW potentials. Also it seems 
70 Ry is not enough for the Na PP you choose based on the criteria and plot 
detailed here:

http://materialscloud.org/sssp/results/Na_conv_patt.png




Regards,
Mostaaf
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] force system to charge transferring

[Mostafa Youssef]

Dear Zara,


Please have a look at this tutorial

http://theossrv1.epfl.ch/Main/OxidationStates



Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] vc-relax with homogeneous electric field

[Mostafa Youssef]

Dear Lei,


If your material is does not exhibit large electrostriction and piezoelectric 
effect , then you can ignore changes in lattice parameter.  Other wise you need 
to generate multiple curves (energy vs. volume) each curve has a different c/a 
ration. Where c is the direction parallel to the field (typically elongated), 
and a is perpendicular to the field , typically shortened.




Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Error in charge difference

[Mostafa Youssef]

Dear Ozan,


Make sure that all three files have same supercell vectors and same FFT grid.


Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Application of electric field to quantum dot

[Mostafa Youssef]

K_POINTS automatic
  1 1 1   1 1 1

This works.


Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Application of electric field to quantum dot

[Mostafa Youssef]

Dear Parham,

I never studied quantum dots , so I cannot give a particular recommendation. 
But for the electric field in Berry phase formalism I can say that Gamma point 
only is not implemented in pw.x but it the default and only option in cp.x.  In 
pw.x you can use one  kpoint but it has to be shifted (not gamma).

>From my  experience nberrycyc beyond 2 does not change any of the calculated 
>properties I have tested. Even the default 1 is great.

It is up to you to decide how much field you need to apply. But there is a 
"computational" maximum  that can be estimated by   Egap/(L x k)

Egap: band gap
L: supercell length
K: kpoints along the length L


Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] charge density espresso 5.0.1

[Mostafa Youssef]

Dear  Sakhraoui,


The format of .XSF is explained here:

http://www.xcrysden.org/doc/XSF.html


You can plot rho_up -rho_down using plot_num=6 option in pp.x

http://www.quantum-espresso.org/wp-content/uploads/Doc/INPUT_PP.html#idm6483344



Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Upper estimation of lattice parameter by LDA pseudopotential -regarding

[Mostafa Youssef]

Dear  Suresh,


I think the reason is the Ti pseudopotential.  This one has 4 valence electrons 
and freezes the rest.  This leads to larger ionic sizes and overestimation of 
the lattice constant.  Allowing more Ti valence electrons, means further 
relaxation for the outer electronic shell and better adoption to the crystal 
environment.  Try  a Ti PP that has 12 electrons.  Pslib from which you 
obtained this one, has also ultrasoft pseudos with 12 electrons.



Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] rectangular surface of Al2O3

[Mostafa Youssef]

Dear Andrey,


I'm not particularly familiar with Al2O3 surfaces, but in general cutting 
rectangular surfaces can be accomplished in many cases by cutting the same 
surface from primitive cell instead of the conventional well.

For example (111) is not rectangular in tetragonal Hf2O based on the 
conventional cell. Starting from the primitive cell, this surface  becomes 
(101) and the cut becomes rectangular.

I believe the primitive cell of Al2O3 is Rhombohedral.



Hope this helps.


Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Macroscopic Polarization

[Mostafa Youssef]

Dear Subhodip,

Check example04 provided with pw.x. You can find it  in:

/Espresso_folder/PW/examples/example04



>From README:
example04:
This example shows how to calculate the polarization via Berry
Phase in PBTiO3 (contributed by the Vanderbilt Group in Rutgers
University).

I would also recommend this paper:
https://arxiv.org/abs/1202.1831



Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] epsilon calculation in the case of a soft mode

[Mostafa Youssef]

Dear all,


I'm studying an oxide with an optical mode frequency that is imaginary before 
applying a.s.r. and becomes real once a.s.r is applied. In the same input file 
for dynmat.x I requested computing epsilon. But it seems that epsilon is 
calculated based on frequencies (w2) before applying a.s.r.

Is there a way to enforce epsilon calculations based on frequencies after a.s.r 
application ?


I'm using Q.E.5.2.0 but I believe recent versions do not have updates in this 
part of the code.



Thank you,
Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Band Offsets too large

[Mostafa Youssef]

Dear  Dae Kwang Jun

Equation (1) that you wrote is the correct one.

And you are right  bulk V_AlAs and bulk V_GaAs are obtained in a similar way as 
in the interface but you need bulk cells.  Their values can be anything , 
positive or negative it does not matter and  in fact they are arbitrary.


Regards,
Mostafa

___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Band Offsets too large

[Mostafa Youssef]

Dear Dae Kwang Jun


In addition for each bulk calculation one needs to calculate the bulk average 
electrostatic potential.  Then for each compound you refer its EVBM to the 
electrostatic potential calculated in  its bulk (not yet the interface).  Only 
then you can use the interface electrostatic potential according too:

EVBM_offset =  EVBM_AlAs   -  EVBM_GaAs +  del_V_interface_AlAS/GaAS

where EVBM_AlAs is referred to bulk V_AlAs and EVBM_GaAs is referred to bulk 
V_GaAs.


Regards,
Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Band Offsets too large

[Mostafa Youssef]

Dear Dae Kwang Jun,

You also need 2 bulk calculations for AlAs and GaAs to get their band edges 
with respect to to their bulk electrostatic potentials. Then you can use the 
interface potential to align the band edges of the two semiconductors.  It 
looks like you missed this step.



Mostafa
MIT


___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Band gap value through charged supercell calculation

[Mostafa Youssef]

Dear Evan,


I would consult literature for that matter.  It is very difficult and 
counterproductive to discuss this in an online forum.


Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Band gap value through charged supercell calculation

[Mostafa Youssef]

Correction:


I-A= E(N+1)+E(N-1)-2E(N)



Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Question about electron (pseudo-)charge density calculations in QE

[Mostafa Youssef]

Dear Reza,


If hydrogen can transform to Silicon than Quantum Espresso is an all Nucleon 
code that simulates Fusion :-)

Check your Pseudo !


Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Band gap value through charged supercell calculation

[Mostafa Youssef]

Dear Evan and Layla,
E(N+x)  where x is a real number can be calculated using tot_charge:

http://www.quantum-espresso.org/wp-content/uploads/Doc/INPUT_PW.html#idm6396048


The ionization potential is not well-defined in bulk as Layla described but the 
difference I-A is well-defined for bulk.

I-A= E(N+1)+E(N)-2E(N)


Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Band gap value through charged supercell calculation

[Mostafa Youssef]

Dear Perevalov,


The K-S gap in left panel of Fig.2  in the paper is not what you get directly 
from the occupations of the neutral cell. What is shown in the figure is 
calculated using equation 13  which uses eigenvalues from the neutral cell and 
occupations from charged cell. This way there will a dependence on carrier 
concentration.
I believe what you plotted  and found to be independent of "cell size" is K-S 
gap using both eignevalues and occupations of the neutral cell.


You mentioned;  "I understand that dependence on the supercell size is due to 
compensating charge background". In fact even if you correct for the 
compensating background , you will still observe dependence on the charge 
density for I-A  and K-S calculated with equation 13.  In the dilute limit of 
charged carriers you should converge to K-S gap of the neutral cell in the case 
of functionals that do not have exact exchange (LDA, GGA, BYLP, ...).  For 
hybrid functionals that contains exact exchange (PBE0, HSE, ...) there will be 
a difference  between I-A and K-S (neutral) even in the dilute limit.  This is 
also discussed in the paper you cited right before Fig. 2.

It is common, at least in semiconductor defects  studies , to regard I-A as 
"the" band gap of the material.  Some may agree , others do not.


For  monoclinic ZrO2,  the first order M-P correction was reported here:

http://journals.aps.org/prb/abstract/10.1103/PhysRevB.75.104112

Of course based on the lattice parameters and supercells that the authors 
reported.


In computing the K-S gap of a neutral cell I would use the tetrahedron method  
or fixed occupations (i.e no smearing) and a dense K-point mesh

 Regards,
Mostafa  Youssef
MIT




___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Full k-points in Brillouin Zone (turning-off all symmetries)

[Mostafa Youssef]

Thank you for clarification, Ryky!


Best Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Regarding Absolute band-edges

[Mostafa Youssef Mahmoud Youssef]

Dear Vipul,


This is an old problem in semiconductor physics. In summary you need 2 set of 
calculations. One for a surface (interface) and another one for bulk.

Key papers to review are:

http://journals.aps.org/prb/abstract/10.1103/PhysRevB.35.8154

http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.61.734


There are many modern examples as well. In my opinion absolute alignment to 
vacuum is more difficult than aliment of material 1 to material 2.


Best Regards,
Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Regarding polarisation from Berry phase

[Mostafa Youssef Mahmoud Youssef]

Dear Varadharajan ,



This paper clarifies a lot of the ambiguities associated with determining the 
spontaneous polarization. Particularly the discussion around Fig. 9.


http://arxiv.org/abs/1202.1831


Best,
Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] negative occupations in DFT+U for nonmagnetic semiconductor

[Mostafa Youssef]

Thank you for the clarification, Prof. de Gironcoli.  I think I misinterpreted 
the meaning of the matrix elements.

Best Regards,
Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] negative occupations in DFT+U for nonmagnetic semiconductor

[Mostafa Youssef]

Dear all,

I'm relaxing the structure of rutile TiO2 using ultrasoft pseudopotentials for 
Ti and O and PBEsol + U with U=5 eV on Ti.  For that purpose I use occupations 
= 'fixed'.  But I found that one of the Ti ions has negative occupations. To me 
this looks a bit strange since I'm not using any smearing.  The same result can 
be obtained by another set of PP. Improving the cutoffs or the k-point sampling 
does not resolve the problem. Any thoughts about the reason and possible 
remedies are appreciated.  I copy below the occupation matrices for the two Ti 
ions in the cell and the input file.


Thank you,
Mostafa Youssef
MIT

The matrices are:

atom1   Tr[ns(na)] =   2.70769
eigenvalues:
  0.142  0.177  0.182  0.417  0.436
eigenvectors:
  0.675  0.000  0.000  0.325  0.000
  0.000  0.000  0.500  0.000  0.500
  0.000  0.000  0.500  0.000  0.500
  0.000  1.000  0.000  0.000  0.000
  0.325  0.000  0.000  0.675  0.000
occupations:
  0.231  0.000  0.000  0.000 -0.129
  0.000  0.309 -0.127  0.000  0.000
  0.000 -0.127  0.309  0.000  0.000
  0.000  0.000  0.000  0.177  0.000
 -0.129  0.000  0.000  0.000  0.328
atom2   Tr[ns(na)] =   2.70768
eigenvalues:
  0.142  0.177  0.182  0.417  0.436
eigenvectors:
  0.675  0.000  0.000  0.325  0.000
  0.000  0.000  0.500  0.000  0.500
  0.000  0.000  0.500  0.000  0.500
  0.000  1.000  0.000  0.000  0.000
  0.325  0.000  0.000  0.675  0.000
occupations:
  0.231  0.000  0.000  0.000  0.129
  0.000  0.309  0.127  0.000  0.000
  0.000  0.127  0.309  0.000  0.000
  0.000  0.000  0.000  0.177  0.000
  0.129  0.000  0.000  0.000  0.328
N of occupied +U levels =5.415371

The input file:

   calculation =   'vc-relax'   ,
   verbosity = 'high' ,
   restart_mode = 'from_scratch' ,
   outdir='./' ,
   pseudo_dir = './' ,
   tstress = .true.  ,
   tprnfor= .true.  ,
   nstep = 2000 ,
   etot_conv_thr = 1.0D-7  ,
   forc_conv_thr = 4.0D-5 ,
 /
 
   ibrav = 0 ,
nat  =  6 ,
   ntyp  =  2 ,
   nbnd  = 48 ,
  ecutwfc  =  45 ,
  ecutrho  =  360 ,
  nosym =.true.
  occupations = 'fixed',
  lda_plus_u = .true.,
  Hubbard_U(1) = 5.0 ,
  Hubbard_U(2) = 0.0 ,
 /
 
diagonalization='david',
mixing_mode = 'plain' ,
mixing_beta = 0.7,
startingwfc = 'random',
conv_thr  =  1.0d-8  ,
  /

ion_dynamics='bfgs'
/

   cell_dynamics = 'bfgs' ,
   cell_dofree = 'shape',
   press_conv_thr = 0.5 ,
 /
CELL_PARAMETERS  angstrom
4.63202526698411180.0.
0.4.63202526698411180.
0.0.2.960893414442
ATOMIC_SPECIES
  Ti   47.867   ti_pbesol_v1.4.uspp.F.UPF
  O15.9994  o_pbesol_v1.2.uspp.F.UPF
ATOMIC_POSITIONS crystal
Ti   0.  0.  0.
Ti   0.5000  0.5000  0.5000
O0.3048741584772117  0.3048741584772117  0.
O0.6951258415227883  0.6951258415227883  0.
O0.8048741584772117  0.1951258415227883  0.5000
O0.1951258415227883  0.8048741584772117  0.5000
K_POINTS automatic
  4 4 6  1 1 1

___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] vc-relax problem

[Mostafa Youssef]

Dear David,


In principle, starting from a 2D material and applying vc_relax should end up 
with the 3D version of the material. That is in the graphene case you should 
end up with graphite upon complete convergence.  I doubt that this simulation 
is your intention here.

If you would like , instead, to study biaxial strain effect on doped graphene, 
then simple manual changes of the a and b of the supercell and relax  
calculation should do the job.



Best Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Ultrasoft or Norm conserinvg PP for Ce

[Mostafa Youssef]

Dear Mehmet,

Thank you for the link to the PAW dataset and your advice regarding the 
convergence of DFT+U.  As much as I'm willing to test this dataset ,  my 
current investigations require Ultrasoft or Norm conserving PP because of the 
limited implementation of PAW calculations in Q.E. package.


Best Regards,
Mostafa
MIT

From: Mehmet Topsakal [mtops...@umn.edu]
Sent: Thursday, March 10, 2016 3:58 PM
To: PWSCF Forum; Mostafa Youssef
Subject: Re: [Pw_forum] Ultrasoft or Norm conserinvg PP for Ce

Dear Mostafa,

Please find our PAW potentials for rare-earth elements here :
http://www.vlab.msi.umn.edu/resources/repaw/index.shtml
and our paper
http://www.sciencedirect.com/science/article/pii/S0927025614005059

I've tested these for various different systems. I always compare QE and VASP 
results.
I have not noticed any discrepancy between my potentials and VASP.

You should be careful with the occupations. Usually it is necessary to enter 
occupations
by hand using starting_ns_eigenvalue(m,ispin,I)  
http://www.quantum-espresso.org/wp-content/uploads/Doc/INPUT_PW.html#idm6307200.
Otherwise calculations do not converge !!!

Bests



On Tue, Mar 8, 2016 at 1:39 AM, Mostafa Youssef 
<myous...@mit.edu<mailto:myous...@mit.edu>> wrote:
Dear all,


I wonder if there is a reasonably performing Ce PP,  US or NC , with 12 valence 
electrons and  non-linear core correction.


Thank you,
Mostafa
MIT

___
Pw_forum mailing list
Pw_forum@pwscf.org<mailto:Pw_forum@pwscf.org>
http://pwscf.org/mailman/listinfo/pw_forum



--
Mehmet Topsakal
Chemical Engineering and Materials Science,
University of Minnesota, Research Associate,
http://www.researchgate.net/profile/Mehmet_Topsakal
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Ultrasoft or Norm conserinvg PP for Ce

[Mostafa Youssef]

Dear all,


I wonder if there is a reasonably performing Ce PP,  US or NC , with 12 valence 
electrons and  non-linear core correction.


Thank you,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Spin-Polarized NEB

[Mostafa Youssef]

Dear Giuseppe,


The ethylene rotation example you mentioned is very interesting.  I wonder if 
there is a reference you recommend that carefully analyzed this case with DFT 
or other quantum chemical approaches.


Best Regards,
Mostafa
MIT



___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Oxidation state for dopants in TiO2

[Mostafa Youssef]

Dear Giacomo,

If the supercell is charge neutral which is the default, then one aims at 
simulating a neutral substitutional defect with respect to Ti. In this case V 
in 4+ oxidation state which corresponds in the naive fully ionic picture to V 
losing 4 of its valence electrons to oxygen and retaining the last one as you 
said.  The question arises whether this  electron is localized on the V site 
and it is really in 4+ oxidation state. The other alternative is that the 
electron delocalizes and the supercell contains V in 5+ oxidation state and an 
extra "free" electron. In my opinion the best  way to analyze this problem is 
to  start from the most positive oxidation state and systematically reduce it. 
In the case of V you can do 6 separate relaxation simulations starting from 
tot_charge=+1 which corresponds to V on 5+ oxidation state, all the way to 
tot_charge=-4 which corresponds to V in 0 oxidation state. After these 6 
simulations are done you can track the changes in the charge density and spin 
density when you go from the oxidation state q to the oxidation state q-1. In 
my experience V in oxides such as TiO2 can take oxidation states from 5+ to 2+. 
 If you compare 1+ and 2+ you will notice that the extra electron you add to 2+ 
to achieve 1+ never localizes on V and as such 2+ is likely the lowest 
oxidation state for V in these oxides.


If you do not want to do this lengthy analysis and you just want to check 
whether you have 4+ or not , check the spin density. V4+ will likely have a net 
magnetic moment close to 1 Bohr Mag.


Also there are many cases in literature where achieving certain oxidation state 
never happens because of electron (hole) delocalization. For example, if you 
try to model neutral hydrogen interstitial in ZnO or ZrO2, you will get 
interstitial proton and a free delocalized electron. See for example:
http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.85.1012


A final word of caution, analyzing the oxidation states of transition metal 
dopants and their changes by adding or removing electrons requires very dense 
and accurate grids for representing the charge (and spin) density. The reason 
is that the change of the localized charge (if any) on the transition metal 
defect while going from formal oxidation state q to q-1 is usually very low (in 
my experience in the order of 0.1 e). This observation was discussed in this 
article: http://www.nature.com/nature/journal/v453/n7196/full/nature07009.html

(It is also fun and instructive to follow the debate that this paper raised in 
literature!)

Best Regards,
Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] non-cubic dielectric tensor in a cubic crystal.

[Mostafa Youssef]

Hi Eduardo,


I also get a non-cubic dielectric tensor for cubic semiconductors when I allow 
the computation of the nonanalytic terms by specifying a q-vector and q is not 
equal 0.  But I do not fully understand what is going on here.  I think (but 
not very sure) if you do not specify q then you actually calculate the 
nonanalytic terms for q->0 and then you recover the correct cubic tensor with 
zone center phonons contribution.

On the other hand calculating epsilon_0 by applying electric field and relaxing 
the ions yields a reasonable cubic dielectric tensor.


I hope somebody can help in clarifying this.

Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] DOS Without 'Relax' ?

[Mostafa Youssef]

Dear Victor,


I think there is nothing fundamentally wrong with calculating DOS from 
snapshots obtained from finite T MD. I  also have seen this once in the 
literature from the work of Prof. A. Selloni group on TiO2/water interfaces 
somewhere around 2010.

 One could argue that even the single particle Kohn Sham DOS obtained for a 
relaxed structure are meaningless  but a lot of the researchers in 
computational materials pragmatically accept it as an insightful tool. My 
opinion is one can compute, think, and interpret but keep in mind that DFT is a 
ground state theory.

Mostafa
MIT


___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] optimized Vanderbilt pseudos

[Mostafa Youssef]

Dear Stephan,

The ONCV are norm conserving  so 4*ecutwfc should be fine.



Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Band gap hybrid functionals

[Mostafa Youssef]

Dear Phil and all,


Note that for a functional with exact exchange, the difference between 
ionization potential and electron affinity (I-A) is not the same as the 
Kohn-Sham single particle states  gap.  The consensus, as far as I know, is 
that I-A is more physical and relevant to comparison with experimental band 
gaps.  Calculating I-A  requires 3 scf calculations (neutral cell, +electorn, 
-electron).  The cases of +/- e require high smearing similar to that of metals 
(or you could use the tetrahedron method but I doubt it is implemented for 
hybrids). Also +/-  calculations require large cells and correction for a 
charged cell (Makov-Payne for example).  See section III-A in this paper for 
elaboration on computational details:

PHYSICAL REVIEW B 78, 235104 ?2008?

For PBE, LDA, ...  I-A is the same as  Kohn-Sham gap.

For elaboration on the theoretical aspects and why I-A is not the same as K-S 
band gap for the exact exchange, consult the work of A. J. Cohen and 
collaborators .

Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] total energy for charged defects in semiconductors

[Mostafa Youssef]

Dear  Manhong,

As far I know the incosistency of treating charged defects within PAW formalism 
discussed in  PHYSICAL REVIEW B 89, 045116 (2014) has not been resolved in any 
pseudopotential code. I  wish I can devote more time  to study this paper more 
carefully.

Regarding the second paper,  PHYSICAL REVIEW B 91, 024107 (2015), it suggests 
that the arbitrariness  in the reference of the electrostatic potential has a 
sizable effect on the pressure (stress) calculated by DFT codes for charged 
systems.  Figure 1 tells the whole story.  The authors suggested a method to 
correct for that given the bulk modulus (eleastic constants) and the 
deformation potential. of the host semiconductor.  However, the paper is of 
concern if you plan to do variable cell relaxation.  Most of the charged 
defects papers focus on fixed volume calculations claiming that this is a good 
way to represent dilute limit of defects.

In summary, if you plan to do fixed volume calculations, the second paper 
should not be of concern  (roughly speaking).  Resolving the inconsistency of 
the first paper is to be implemented. To avoid it one might use all electron 
code.


M. Y.
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] DFT+U with static homogeneous finite electric field or Uniaxial tensile strain

[Mostafa Youssef]

Dear Yusuf,


I'm not a developer but from my limited experience,  I can tell that 
cell_dofree tag (the key to apply uniaxial strain) works well with DFT+U  (With 
NC and USPP, I'm not sure about PAW).  Finite electric field with DFT+U (NC and 
USPP) seems to work well too based on few tests I did on computing the 
dielectric constant for oxides with DFT+U.


Examples 08, 10 and VCSexample give you the basic elements for what you asked. 
But you need to combine and compose something out of these examples to meet 
your needs.


Mostafa
MIT


___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Total potential under Berry phase electric field

[Mostafa Youssef]

This question is resolved.  I understand a little better now that in small 
field regime one solves for polarized Bloch states and as such the charge 
density is indeed periodic.


Mostafa
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Total potential under Berry phase electric field

[Mostafa Youssef]

Dear all,

I plotted the total potential (bare + Hartee + xc) for MgO under finite 
electric field applied via Berry Phase approach. This is basically example 09 
in  CPV/examples and I run the same calculation also using PWscf.  Attached is 
the result. In particular the macroscopic average of the total potential is 
zero.  Is the total scalar potential really flat under electric field applied 
via Berry phase or is there something wrong with what the code prints as the 
total potential?

I'm still trying to understand how the electric field is applied in this case , 
in particular how the non-periodicity of the scalar potential is avoided.

Any thoughts or references are highly appreciated.

Mostafa Youssef
MIT


MgO_poential.pdf
Description: MgO_poential.pdf
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Finding the most stable structure

[Mostafa Youssef]

Dear Reza,

Exploring the potential energy surface of glassy systems is not a trivial 
problem not to mention impurities in glassy systems. Depending on how deep you 
want to explore this,  you may want to explore PLUMED package which is 
compatible with PWscf.

http://www.plumed-code.org/

For something quick, I would run an MD at very high temperature , take 100 
snapshots from the run , and relax them at zero temperature.

Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Temperature effect on band gap

[Mostafa Youssef]

Dear Reza,


In addition to what Lorenzo suggested in relation to the accuracy of DFT in 
describing the band gap,  there is a possibility that point defects can play a 
role in reducing the gap as the temperature increases.  You might need to 
inspect the literature more to see how the observed reduction in the band gap 
was explained for the material of interest.



Regards,
Mostafa
MIT

___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Force calculation time in relaxation under electric field

[Mostafa Youssef]

The last time report I quoted is with Hubbard U and this in fact what I 
typically find in PWscf relaxation calculation.

I will send you the two output files offline. Thank you very much for your help!


Regards,
Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Force calculation time in relaxation under electric field

[Mostafa Youssef]

Thank you for replying, Prof. Giannozzi.  I'm using version 5.2.0, the system 
has 135 atoms  of which 108 has U correction.   I'm running it on 32 processors 
with 1 pool,  a serial algorithm for Linear algebra to avoid "error in Cholesky 
decomposition",  and 1 task group.

I'm a bit surprised because the time report of the same system without electric 
field and with a denser Kpoint grid (2x2x2) looks like:

 init_run : 38.32s CPU 39.48s WALL (   1 calls)
 electrons:   9042.53s CPU   9168.56s WALL (   8 calls)
 update_pot   :  4.99s CPU  6.13s WALL (   7 calls)
 forces   :873.59s CPU949.08s WALL (   8 calls)
 stress   :   3211.03s CPU   3260.70s WALL (   8 calls)

Here the forces is just a tiny fraction of the electrons time.


Regards,
Mostafa
MIT


___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Force calculation time in relaxation under electric field

[Mostafa Youssef]

Dear all,


I performed ionic relaxation for 3x3x3 unit cells of a semiconducting oxide 
under electric field (lelfield =.true.) using ultrasoft PP and 1x1x3 MP K-point 
grid where 3 is in the direction of the electric field. These are GGA+U 
calculations. I noticed in the time report that force calculations time is 
becoming even more than the electrons time as shown below:

 init_run :  8.32s CPU  8.91s WALL (   1 calls)
 electrons:  42723.25s CPU  44995.21s WALL (   5 calls)
 update_pot   :  2.62s CPU  3.19s WALL (   4 calls)
 forces   :  66735.95s CPU  66943.64s WALL (   5 calls)

This is first time in which I find PWscf use more time in force calculations 
than in electrons. For the same system and without electric field , forces time 
would be about 0.2 of electrons time.  Is this a normal behavior and is there a 
recommended possibility to reduce force calculation time?

I attempted using Norm Conserving PP's but unfortunately I could not get the 
code pass the 1 ionic step because of the S matrix not being positive definite.

I appreciate your thoughts on this.

Regards,
Mostafa Youssef
MIT

___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Failed in tests and examples

[Mostafa Youssef]

Dear Fei

As the message indicates there is a problem in reading the norm conserving 
pseudopotential file. I think this may also be the reason why example01 did not 
work because it uses NC PP for Si.

Try the attached version of the H and O PP used in the H2O test. The attached 
worked for me. Or may be download a Si NCPP from Q.E website and try again 
example 1 (of course after modifying the name of the PP in the input file). Or 
even try example with Ultrasoft PP to see if it works at all with your compiled 
version of Q.E.


Also diff the attached with the ones you have to see what is the cause of the 
problem.


Mostafa
MIT


H.blyp-vbc.UPF
Description: H.blyp-vbc.UPF


O.blyp-mt.UPF
Description: O.blyp-mt.UPF
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Choice of Pseudopotential

[Mostafa Youssef]

Hi Elham,

Testing Pseudopotentials and choosing optimal ones is a daunting task that 
needs some serious work. In addition to the valuable advice you quoted from 
Prof. de Gironcoli, I'd add that converging total energy will lead you to a 
very safe e_cut and rho_cut but probably one can afford a little less accuracy 
than the very safe one. The way to test is to focus on the quantity you care 
about and try to check its convergence.


See for example this descent work on testing PP's and finding optimal cutoffs 
based on phonon frequency convergence and equation of state error with respect 
to all electron calculations (among other things):

http://materialscloud.org/sssp/

A similar example is GBRV library for which the criteria for convergence (as 
stated in the authors paper) is energy difference rather than absolute total 
energy:

http://www.physics.rutgers.edu/gbrv/


I hope this helps.
M. Y.
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Failed in tests and examples

[Mostafa Youssef]

Hi Fei Mao,

Please post the error message in the file  b3lyp-h2o.out. Apparently the test 
stopped when it encountered error message here.


M. Y.

MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Question on pseudopotential

[Mostafa Youssef]

Hi Sridhar,

Of course you can do this mix. Please have a look at:

http://www.quantum-espresso.org/faq/pseudopotentials/#2.3


M. Y.
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] nscf+lelfield -----> Wrong k-string? Error

[Mostafa Youssef]

Dear Mojtaba,

I have not encountered this error before, so I cannot comment on it.

But I think it is a little bit strange to apply the electric field in the nscf 
calculation only.  If you would like to calculate the polarization in the 
presence of the electric field , you need 2 scf calculations. The first at zero 
field but still with lelfield=.true., the second starts from the wave functions 
generated from the first but has the magnitude of the field you would like to 
apply.  Please see example 10 in PW/examples.



Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Serial vs. parallel runs of berry.in test case

[Mostafa Youssef]

Thank you very much for fixing this, Prof. Giannozzi !


Best Regards,
Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Serial vs. parallel runs of berry.in test case

[Mostafa Youssef]

Thank you, Prof. Giannozzi.  I tested one more time and found out that by 
removing tqr=.true. it works on 8 processors even with an FFT grid of 36x36x36. 
 With tqr=.true. , the FFT grid has to be multiples of 8.



Best Regards,
Mostafa Youssef
MIT

___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Serial vs. parallel runs of berry.in test case

[Mostafa Youssef]

Dear all,

I tried to run the standard tests of PWscf  both in series and in parallel. In 
series all is fine.  In parallel, I found the case of berry.in to be 
problematic as I got an error:

 Error in routine addusdens_r (1):
 WRONG CHARGE expected  44.00, found  44.404701: ions may be overlapping or 
increase ecutrho

And the code stops at the first scf step. When I double the energy and charge 
density cutoffs , it works  well in parallel. This behavior is reproduced in 
versions 5.1.1, 5.1.2, and 5.2.0.

Any thoughts on the possible reasons are appreciated.


Thank you,
Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] DFT+U calculations

[Mostafa Youssef]

Dear Adnan,

Please check the two papers I cited  in my earlier message.  In  the 2014 J. 
Phys. Condens. Matt. paper , it was shown that U on Ti in SrTiO3  removes the 
the Gamma point instability and enhances the M-point instability.  In the other 
paper there is more discussion of the effect of U on TiO2 phonons.



Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] DFT+U calculations

[Mostafa Youssef]

Dear Muhammad ,

This is really tricky. Using DFT+U for 0K energies followed by DFT for phonons 
is a bit inconsistent.  I would rather be consistent and understand the 
limitations of the theory rather than trying to mix inconsistent parts together.

You could ignore the vibrational contribution as suggested by Jia Chen , 
although I think at high temperature this contribution might be significant. 
Alternatively you could resort to the extremely expensive hybrid functionals  
to do both 0K energies and phonons (within frozen phonon approximation). 
Another sound alternative is just to use DFT+U for 0K energies and phonons with 
the clear understanding that therer might be some error associated with the 
phonons part.  In fact, although literature clearly shows that DFT+U predicts 
qualitatively problematic phonon dispersion for soft-mode materials such as 
TiO2 , there has not been a systematic investigation of the performance of 
DFT+U in predicting vibrational free energy *differences* between a perfect 
crystal (TiO2) and a defective one (TiO2-x) for this class of materials.  I 
would be very interested if somebody can point out to a paper that addressed 
this.


Mostafa Youssef
MIT
<http://www.mail-archive.com/search?l=pw_forum@pwscf.org=from:%22Muhammad+Adnan%22>
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] DFT+U calculatoopns

[Mostafa Youssef]

Hi Adnan,


The effect of DFT+U on the phonon dispersion of materials with soft-modes such 
as TiO2 is large. Not only quantitatively but also qualitatively.  See for 
example:

http://www.sciencedirect.com/science/article/pii/S1567173910004682


and figure 8 in:

http://iopscience.iop.org/0953-8984/26/12/122203/


If your main goal is to study phonons of TiO2 , I would recommend LDA or 
PBEsol.  But if you insist on using DFT+U , then you can use PHONOPY + PWscf 
code of Q.E equipped with a *large* supercell of TiO2.  This should allow you 
to calculate phonon dispersion in frozen phonon approximation with Parlinski et 
al interpolation.  Please follow the instructions on PHONOPY website.


Mostafa YOussef
MIT
P.S. Somehow I read the title of your thread as DFT+U cartoons !
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] tot_charge flag with DFT+U calculations

[Mostafa Youssef]

Dear J. Park,

In principle there is no problem at all and there have been publications 
addressing charged defects based on Q.E. calculations employing DFT+U.  See for 
example:
http://pubs.acs.org/doi/abs/10.1021/jp1041316

The only practical detail worth mentioning in my opinion is that if you use 
fixed occupations then you are forced to specify the total magnetization of the 
super cell instead relying on self-constant determination of the magnetic 
moment.  One way to go around this is to use small smearing and then 
self-consistent determination of the total magnetic moment is possible (of 
course starting from some initial guess).

Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Can I obtain IR spectrum from CPMD trajectory in Quantum Espresso?

[Mostafa Youssef]

Dear Yin,

You already got great suggestions. I would like to add one more option to 
calculate IR from ab initio MD. You could do scf calculation on each 
configuration obtained via MD to calculate Maximally Localized Wannier 
Functions (MLWF). From MLWF you could calculate the dipole moment of the 
simulation cell or even for single molecules inside the cell at each time step. 
It is straightforward to obtain IR  once you have the dipole moment at each 
time step.

An example for this option can be found here:

http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.98.247401

Note that the authors of this paper implemented Wannier on the fly which I 
think is not available in Q.E. But the computational cost to re-track the 
trajectory and calculate MLWF should not be too expensive if one already 
affords the MD run.


Mostafa Youssef
MIT


___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] How to calculate charged system binding energy?

[Mostafa Youssef]

Dear Bahadir

First, charged slabs are problematic because their total energy does not 
converge with respect to vacuum thickness. You can test on a simple model.  
However, there is a trick to go around this by inserting a dopant far a way 
from the critical reaction zone.  For example suppose you want to study +1 
charged defect on ZrO2 surface.  Then you can insert one Y ion(typically 3+ and 
hence -1 with respect to Zr4+) .   This should generate a positive charge 
somewhere else in the slab and one hopes that this positive charge will 
localize correctly where you expect it to localize.  But one has to be cautious 
because this also generates a large dipole across the slab. One way to go 
around this by symmetrizing the slab such that dipoles cancel.


Second, even if charged slabs work, I think you are not conserving the charge 
when you calculate the binding energy. Because you mentioned you used  
tot_charge=-1 in the slab+molecule , slab only, molecule only. To me this will 
not conserve the charge when you calculate the binding energy (B.E.)

B.E. =(slab+molec.)  - (slab) - (molec)

Although I have seen papers defining binding energies that do not conserve the 
charge, I do not think this is meaningful.

Mostafa Y.
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] control strain calculations

[Mostafa Youssef]

>if you can please give me more details about how to control the number of cell 
>relaxed , because it is the same with
>different press-conv-thr . for example I want to relax only the c axis keeping 
>a and b fixes, by using cell-dofree='z' ,
>but I would give the number of c which i want to relax ? is that possible?!


Hi Imène,
I'm not very sure that I understand, but I guess if you use cell-dofree='z'
and vary press_conv_thr (to smaller values) and still get the same value of
the c-axis after vc_relax than that indicates that the results that you get
for the c-axis is not that sensitive to the stopping criterion imposed by
press_conv_thr.  If you the post the input file with more details then may be
somebody with better expertise in this can help.


Mostafa Y.
MIT

___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Is geometry optimization same to MD?

[Mostafa Youssef]

Hi Shaofeng,


Relaxation will give you "a" structure but most likely it is not a 
representative  structure. MD will give you a set of possibly meaningful and 
representative solvation structures. If you average over these  meaningful 
structures, you can get useful information about the solvation of arsenic ion.

Arsenic is anticipated to be ionized in water, so you will need a charged 
system but Q.E. will automatically apply a neutralizing background.

Have a look at some of the early DFT MD work on solvation to get a better idea 
about what can and cannot be done.
For example:

http://www.sciencedirect.com/science/article/pii/S0009261497006180

Mostafa Y.
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] control strain calculations

[Mostafa Youssef]

Hi Imène,


I guess the sopping criterion is  press_conv_thr.


http://www.quantum-espresso.org/wp-content/uploads/Doc/INPUT_PW.html#idp365664


Mostafa Youssef
MIT
<http://www.mail-archive.com/search?l=pw_forum@pwscf.org=from:%22Im%C3%A8ne+Ch%C3%A9ra%C3%AFr%22>
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] spin polarization

[Mostafa Youssef]

Hi Tayebeh

I think all what you need to calculate a helix of magntic moments for Co ions 
in  SrCO3 would be:

1- Noncollinear calculations.
2- A larger supercell to accommodate the helix.
3- Fully relativistic pseudopotentials (such as those in Pslib)
4- Define different Co species each with different direction for the magnetic 
moment
5- You might also wish to use vc_relax rather than relax with cell_dofree to 
fix the volume but vary the shape of the cell. Or just vc_relax if you want to 
get a quick answer.


I think all these ingredients are available and Pwscf is ready to tackle SrCO3 !


Mostafa Youssef
MIT

___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Ion

[Mostafa Youssef]

Hi Ashkan,

Click on the title your thread or alternatively search the forum with the 
keyword ion and you will find detailed answers to your question.

See for example:

http://www.mail-archive.com/pw_forum%40pwscf.org/msg09585.html


and also the manual entry:

http://www.quantum-espresso.org/wp-content/uploads/Doc/INPUT_PW.html#idp110032


Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] HSE Si Lattice Parameter

[Mostafa Youssef]

Hi Galvin,

Examining the energies and stresses you obtain from your calculations by the 
eye is not a very accurate way to obtain the lattice parameter that minimizes 
the energy and corresponds to a zero pressure. More accurate is to fit your 
results to an equation of state such as Birch-Murnaghan or Vinet.  The reason 
is that most likely the optimal lattice parameter lies in between the values 
you obtained from your scf results. As a bonus you get also other useful 
parameters form the fit like the bulk modulus.
However, make sure that all the data points you use in the fit correspond to 
hydrostatic stress by examining the stress tensor printed in output file. This 
is because all these equation of states assume hydrostatic stress in their 
derivation.


Mostafa
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Finite electric field (smearing , DFT+U, and total energy)

[Mostafa Youssef]

Thank you very much for addressing my questions, Prof. Giannozzi ! I will keep 
the mailing list updated if I succeed in reasonably getting smearing work with 
finite electric fields.


Mostafa Youssef
MIT


___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Finite electric field (smearing , DFT+U, and total energy)

[Mostafa Youssef]

Dear PWscf users and developers,

I have 2 questions and a comment on finite electric field calculations in PWscf 
and your help and thoughts are very appreciated.


Questions:

(1)  PWscf refuses smearing in Berry phase calculations (lelfield=.true.). 
However, smearing could help in accelerating the convergence of some in-gap 
states for semiconductors that arise because of a defect or impurity. Is there 
a fundamental reason for why smearing should not be used with Berry phase 
calculations, or is it possible to modify PWscf to accept smearing?


(2) I recall Prof. Giannozzi in this mailing list mentioned that DFT+U is not 
tested with Berry phase calculations. The code does not complain about DFT+U 
with Berry phase and I have performed preliminary tests on metal oxides and 
found out that dielectric constant predicted using DFT+U and (lelfield=.true.) 
are reasonable compared to results obtained with other codes using DFTP(+U).  I 
wonder if somebody did more tests on DFT+U with Berry phase and would like to 
share their conclusion.


Comment:

(3)  The total energies that PWscf outputs in finite electric field 
calculations are in fact electric enthalpies.  One could check that the ( 
Harris-Foulkes estimate   +  (total dipole x filed)) is the what PWscf prints 
as the total energy. I think it would be nice to designate this output as the 
electric enthalpy espeically that this will be consistent with what PWscf 
prints out for calculations under pressure. (Of course with the implicit 
warning that the enthalpy inherits the uncertainty in the polarization quantum).


Thank you,
Mostafa Youssef
MIT


___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Negative Polarization !!!

[Mostafa Youssef]

Hi Mojtaba,

Polarization is a vector, so negative or positive just indicate its direction. 
So no worries about negative polarization. But in periodic boundary conditions 
you may still need to add (subtract) multiples of the polarization quantum in 
order to get the "reasonable" value of the spontaneous polarization of the 
material.

A favorite of mine which go over some of these difficulties is this  paper:

http://journals.aps.org/prb/abstract/10.1103/PhysRevB.71.014113


Mostafa Youssef
MIT


___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] PPs for V and Cs

[Mostafa Youssef]

Hi Carlo

Both pslib0.31 (http://theossrv1.epfl.ch/Main/Pseudopotentials)  and pslib1.0.0 
(http://qe-forge.org/gf/project/pslibrary/) have relativistic PP for those two 
elements.


Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Questions related to Co and Fe Pseudopotentials in pslib 1and 0.31

[Mostafa Youssef]

Thank you very much for your reply, Prof. Dal Corso. And of course many thanks 
for the development of the library!


Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Questions related to Co and Fe Pseudopotentials in pslib 1and 0.31

[Mostafa Youssef]

Dear all,

I have few questions related to the above mentioned PP's.

(1)  After generating the PAW and USPP versions of these PP (pslib1.0.0) using 
ld1 code provided with  q.e.5.1.2, I applied the test script.  The output seems 
to be fine in the case of USPP but for PAW extra few lines are printed. For 
example, in the case of LDA scalar-rel Fe, I get:

Making   Fe.pz-n-kjpaw_psl.0.2.4.test.in  ...  Done !
 3 2 3D   1( 7.00)   -0.41920   -0.41398   -0.00522  !
Making   Fe.pz-n-kjpaw_psl.1.0.0.test.in  ...  Done
Making   Fe.pz-n-rrkjus_psl.0.2.4.test.in  ...  Done
Making   Fe.pz-n-rrkjus_psl.1.0.0.test.in  ...  Done
Making   Fe.pz-spn-kjpaw_psl.1.0.0.test.in  ...  Done
Making   Fe.pz-spn-rrkjus_psl.1.0.0.test.in  ...  Done

And in the case of  Fe-rel I get:

Making   Fe.rel-pz-n-kjpaw_psl.0.2.4.test.in  ...  Done !
 3 2 1.5 3D   1( 4.00)   -0.42153   -0.41623   -0.00530  !
 3 2 2.5 3D   1( 3.00)   -0.41137   -0.40618   -0.00519  !
Making   Fe.rel-pz-n-kjpaw_psl.1.0.0.test.in  ...  Done
Making   Fe.rel-pz-n-rrkjus_psl.0.2.4.test.in  ...  Done
Making   Fe.rel-pz-n-rrkjus_psl.1.0.0.test.in  ...  Done
Making   Fe.rel-pz-spn-kjpaw_psl.1.0.0.test.in  ...  Done
Making   Fe.rel-pz-spn-rrkjus_psl.1.0.0.test.in  ...  Done

Similar output obtained for Co PAW.  Are these extra printed lines indicative 
of some error in the generation?


(2) The charge density cutoff recommended (in the UPF files) for these PP is 
about 12 *ecutwfc. This large cutoff for the charge density was also employed 
in a recent paper using Fe (pslib0.31) (http://arxiv.org/pdf/1502.01534.pdf) . 
However, when I did convergence test for the total energy on a unit cell of 
SrFeO3 , and SrCO3, and it looks like 4* ecutwfc is sufficient to achieve 
convergence. The convergence in the case of SrFeO3 (pslib0.3, USPP) can be 
found here (http://web.mit.edu/myoussef/www/Fe_16e_pslib.pdf).  The pink bound 
indicates that convergence was achieved within 25 mev/unit cell (5 meV/atom)  
at ecutwfc=85 Ry and ecutrho =4 *ecutwfc. It seems to me that the recommend 
ecutrho it too strict.  Are there other tests that one needs to perform to 
check the convergence of charge density cutoff?

(3) The recommended ecutwfc and ecutrho for this set of PP in Pslib do not 
change from PAW to USPP and do not depend on the functional. I did preliminary 
tests  and it seems so far that the values work for GGA also work for LDA 
consistent with the recommendation. I wonder if one could test convergence of 
ecutwfc and ecutrho for LDA only and apply the outcome to all other semilocal 
functionals, or are there cases where one encounters different convergence 
depending on the functional.


Thank you in advance for reading and help!
Mostafa Youssef
MIT




Fe_16e_pslib0.3.pdf
Description: Fe_16e_pslib0.3.pdf
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Problem With WC Pseudopotentials in "Relax" Process

[Mostafa Youssef]

Hi Mojtaba,



Providing an input file could help in resolving the problem you have.  Also 
performing few convergence tests on the cutoff for energy and charge density 
could help in getting fruitful results with the pseudopotential.





Mostafa Youssef

MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Positions of Dopant in Supercell

[Mostafa Youssef]

For 64 atoms cell you can test nearest neighbor(NN) , 2nd NN, and so on.  This 
can give you some local information about local arrangements of the defects.

If you reduce the cell to 32 and insert one dopant ion, then you are simulating 
an ordered array of defects because of the periodic boundary conditions.

In fact the interaction across the cell boundary applies also to the case of 64 
atoms above, but one hopes (and can possibly show) that the results represent 
the local interactions between the two defects within the cell rather than the 
image-image interaction across the cell boundary.


In my opinion, a clever way to study concentrated solutions of defects is to 
resort to model Hamiltonians along the lines of the cluster expansion approach. 
 This can work nicely if you care about neutral defects in semiconductors (eg: 
the oxygen vacancy in CeO2 as in http://arxiv.org/abs/1206.5429 ). For charged 
defects where long range interactions become important, there is no clean way 
(as far as I know) to get a model Hamiltonian for the interacting defects 
system. In the literature (especially that of CeO2) you may find few attempts 
for concentrated charged defects. See for example: 
http://journals.aps.org/prb/abstract/10.1103/PhysRevB.90.115120


Mostafa Youssef
MIT
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Normconserving Full_Relativistic Pseudopotentials

[Mostafa Youssef]

Hi Valentina,



There seem to be a set of "untested" norm conserving relativistic PP provided 
in pslib0.31 and 1.0. A ready to use compilation of these (version 0.31)can be 
found here:



http://theossrv1.epfl.ch/Main/Pseudopotentials





Other versions of the library can be found here

http://www.qe-forge.org/gf/project/pslibrary/frs/?action=FrsReleaseBrowse_package_id=41









Mostafa Youssef

MIT


___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum