What's the 7 f-orbital name?

229 views
Skip to first unread message

chao yang

unread,
Aug 28, 2021, 9:26:45 AM8/28/21
to DMFTwDFT
Dear everyone,
         Now I want use DMFTwDFT to study the correlation effect of f electrons. I'm trouble with the name of the 7 f-orbital names in DMFTwDFT, below is the values I will do, are these names are correct?

"cor_orb": [
        ["d_z2", "d_x2y2", "d_xz", "d_yz", "d_xy"],["fy3x2",   "fxyz",   "fyz2" ,   "fz3" ,  "fxz2" ,  "fzx2" ,   "fx3"]]
    ],  # DMFT orbitals, other orbitals are treated by HF


Thanks for any suggestions!
Best,
Chao 

Uthpala Herath

unread,
Aug 28, 2021, 5:44:41 PM8/28/21
to chao yang, DMFTwDFT
Hello Chao,

You can use f1,f2, … 

We might implement a more verbose method in a future update. 

Best,
Uthpala 

--
You received this message because you are subscribed to the Google Groups "DMFTwDFT" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dmftwdft+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dmftwdft/79d24d6f-b51d-4c11-a2d2-3e1b71d1290en%40googlegroups.com.
--
Uthpala Herath
PhD Candidate
Department of Physics and Astronomy
West Virginia University
Morgantown, WV 26505

chao yang

unread,
Aug 29, 2021, 12:09:28 AM8/29/21
to DMFTwDFT
Dear prof.  Uthpala,
      I want study the correlation effect of Yb, and set the d and f orbitals as the correlated orbitals. However the INPUT.py file seems not right. Could you give me some suggestions?
      Please find the details of INPUT.py at the end of the text. 


Best
Chao

$ more INPUT.py 
################   Input parameters for DFT+DMFT calculations   ##################

###### Main loop parameters ###########

p = {
    "Niter": 1,  # Number of DFT+DMFT iterations
    "Ndft": 1,  # Number of DFT iterations
    "Nit": 15,  # Number of DMFT iterations
    "n_tot": 24.0,  # Number of total electrons
    "nf": 7.0,  # Number of target Nd
    "nspin": 1,  # Number of total spins
    "atomnames": ["Yb","Yb"],  # The name of atoms
    "orbs": ["d","f"],  # The name  of orbitals
    "L_rot": [0, 0],  # Whether rotate local axis or not
    "cor_at": ["Yb1","Yb1"],  # Correlated atoms, put symmetry atoms in the same list
    "cor_orb": [
          ["d_z2","d_x2y2","d_xz","d_yz","d_xy"],["f-3","f-2","f-1","f0","f1","f2","f3"]
    ],  # DMFT orbitals, other orbitals are treated by HF
    "U": [5.0],  # Intra-U for each cor_at
    "J": [1.0],  # Hund's coupling
    "alpha": [0.2],  # Double counting parameter
    "mix_sig": 0.2,  # Mixing parameter for Sigma
    "q": [20, 20, 20],  # Number of k-points for Wannier k-sum
    "ewin": [-4, 8.1],  # Energy Window with respect to Fermi energy
    "noms": 400,  # Number of Matsubara frequencies for k-sum
    "dc_type": 1,  # Vdc type
    "mu_iter": 100,  # Steps for the chemical potential convergence
    "Nd_qmc": 0,  # 0: Use Nd_latt, 1: Use Nd_imp
    "path_bin": "DMFTwDFT-master-nscc/bin/",  # Path to bin files
}


$ more slurm.out 

Calculation type : Non-charge self-consistent DFT+DMFT calculation

Traceback (most recent call last):
  File "DMFTwDFT-master-nscc/bin/RUNDMFT.py", line 234, in <module>

Calculation type : Non-charge self-consistent DFT+DMFT calculation

Traceback (most recent call last):
  File "DMFTwDFT-master-nscc/bin/RUNDMFT.py", line 234, in <module>

Calculation type : Non-charge self-consistent DFT+DMFT calculation

Traceback (most recent call last):
  File "DMFTwDFT-master-nscc/bin/RUNDMFT.py", line 234, in <module>
    TB.Compute_cor_idx(cor_at, cor_orb)
    TB.Compute_cor_idx(cor_at, cor_orb)
  File "DMFTwDFT-master-nscc/bin/Struct.py", line 190, in Compute_cor_idx
  File "DMFTwDFT-master-nscc/bin/Struct.py", line 190, in Compute_cor_idx
    TB.Compute_cor_idx(cor_at, cor_orb)
    self.ncor_orb += len(self.TB_orbs[at])
  File "DMFTwDFT-master-nscc/bin/Struct.py", line 190, in Compute_cor_idx
KeyError: 'Y'
    self.ncor_orb += len(self.TB_orbs[at])
KeyError    self.ncor_orb += len(self.TB_orbs[at])
: 'Y'

Calculation type : Non-charge self-consistent DFT+DMFT calculation
.........

Franck

unread,
Aug 29, 2021, 6:49:09 AM8/29/21
to chao yang, DMFTwDFT
Dear Chao ,
to treat the effects of correlations, the electrons which interest us are the electrons of valence and in this case the electrons f of Yb. (you can correct me if I am wrong).
This said
1- I suggest you check the installation direction (you can just go to the bin and do pwd) normally you should have something like  "path_bin":/DMFTwDFT-master-nscc/bin/  or ~/DMFTwDFT-master-nscc/bin/ and not DMFTwDFT-master-nscc/bin/

2-) the parameters are not correct.        atomnames": ["Yb"],    "orbs": ["f"],    "L_rot": [0],   "cor_at": ["Yb1"]  "cor_orb": [[["f-3","f-2","f-1","f0","f1","f2","f3"]]], (I don't know exactly but he will be able to tell you how the f-electrons have been implemented in the code ans how to declare it.   regarding the U and J parameters you have to look at the literature or put the one obtained with VASP or SIESTA.  for this parameter,  "ewin": You have to represent the DOS and look at the energy window where there is a high weight of electrons.

3-) "beta": don't forget to give the right value to this parameter which is simply the inverse of the temperature in eV.  another point is that you should be aware of the type of calculation you need. single-shot and or self-consistence

Good luck,

Franck O.



Franck

unread,
Aug 29, 2021, 7:13:38 AM8/29/21
to chao yang, DMFTwDFT
oups.. something like this  "cor_orb": [ [["f-3"], ["f-2"], ["f-1"], ["f0"], ["f1"], ["f2"], ["f3"]]]

chao yang

unread,
Aug 29, 2021, 10:34:32 AM8/29/21
to DMFTwDFT
Dear Franck,
      I checked and solved  some of those problems. 
      1. Yb is very special, it's valence states are mainly d orbitals, and the f-orbitals are just below the Fermi level,  therefore I set the d and f orbitals as the correlated states.
      2.  "beta": it seems  this parameter  is  the inverse of the temperature in meV,  I want study the properties of Yb under room temperature, therefore I set beta=25.85, is it right ?
      3. but now I encounter new problems. This time I used the INPUT parameters as showing below:

$ more INPUT.py 
################   Input parameters for DFT+DMFT calculations   ##################

###### Main loop parameters ###########

p = {
    "Niter": 1,  # Number of DFT+DMFT iterations
    "Ndft": 1,  # Number of DFT iterations
    "Nit": 15,  # Number of DMFT iterations
    "n_tot": 24.0,  # Number of total electrons
    "nf": 7.0,  # Number of target Nd
    "nspin": 1,  # Number of total spins
    "atomnames": ["Yb","Yb"],  # The name of atoms
    "orbs": ["d", "f"],  # The name  of orbitals
    "L_rot": [0, 0],  # Whether rotate local axis or not
    "cor_at": [["Yb1"],["Yb1"]],  # Correlated atoms, put symmetry atoms in the same list
    "cor_orb": [["d_z2", "d_x2y2", "d_xz", "d_yz", "d_xy"],["f1","f2","f3","f4","f5","f6","f7"]],  # DMFT orbitals, other orbitals are treated by HF
    "U": [5.0],  # Intra-U for each cor_at
    "J": [1.0],  # Hund's coupling
    "alpha": [0.2],  # Double counting parameter
    "mix_sig": 0.2,  # Mixing parameter for Sigma
    "q": [24, 24, 24],  # Number of k-points for Wannier k-sum
    "ewin": [-4, 8.1],  # Energy Window with respect to Fermi energy
    "noms": 400,  # Number of Matsubara frequencies for k-sum
    "dc_type": 1,  # Vdc type
    "mu_iter": 100,  # Steps for the chemical potential convergence
    "Nd_qmc": 0,  # 0: Use Nd_latt, 1: Use Nd_imp


However, the DMFTwDFT show me the errors below:

$ more slurm-3600167.out 

Calculation type : Non-charge self-consistent DFT+DMFT calculation

Wannier orbitals in correlated subspace:
Yb1 : ['f1', 'f2', 'f3', 'f4', 'f5', 'f6', 'f7']
Traceback (most recent call last):
  File "~/DMFTwDFT-master-nscc/bin/RUNDMFT.py", line 276, in <module>
    idx[d_orb.index(orb) + ispin * len(d_orb)] = loc_idx
ValueError: 'd' is not in list

Calculation type : Non-charge self-consistent DFT+DMFT calculation
.......
Calculation type : Non-charge self-consistent DFT+DMFT calculation

Wannier orbitals in correlated subspace:
Yb1 : ['f1', 'f2', 'f3', 'f4', 'f5', 'f6', 'f7']
Traceback (most recent call last):
  File "~/DMFTwDFT-master-nscc/bin/RUNDMFT.py", line 276, in <module>
ValueError: 'd' is not in list
ValueError: 'd' is not in list
    idx[d_orb.index(orb) + ispin * len(d_orb)] = loc_idx
ValueError: 'd' is not in list
yhrun: error: cn1276: tasks 0-23: Exited with exit code 1

could you give me some suggestions?

Best,
Chao

Uthpala Herath

unread,
Aug 29, 2021, 11:37:38 PM8/29/21
to chao yang, DMFTwDFT
Hello, 

Thanks Franck for helping with the responses. 
I would like to add some more comments to the responses. 

I haven't tried systems where we treat both d and f electrons at the same time so this is something I should test. Some suggestions to your INPUT.py involve:

cor_at": [["Yb1"],["Yb1"]],  # Correlated atoms, put symmetry atoms in the same list
    "cor_orb": [["d_z2", "d_x2y2", "d_xz", "d_yz", "d_xy"],["f1","f2","f3","f4","f5","f6","f7"]],  # DMFT orbitals, other orbitals are treated by HF
    "U": [5.0],  # Intra-U for each cor_at
    "J": [1.0],  # Hund's coupling
    "alpha": [0.2],  # Double counting parameter

TO

cor_at": [["Yb1"],["Yb1"]],  # Correlated atoms, put symmetry atoms in the same list
    "cor_orb": [["d_z2", "d_x2y2", "d_xz", "d_yz", "d_xy"],["f1","f2","f3","f4","f5","f6","f7"]],  # DMFT orbitals, other orbitals are treated by HF
    "U": [5.0, 5.0],  # Intra-U for each cor_at
    "J": [1.0, 1.0],  # Hund's coupling
    "alpha": [0.2, 0.2],  # Double counting parameter

In the meantime, if you send me your inputs I can test it on my end. 

Best,
Uthpala

Franck

unread,
Aug 30, 2021, 9:59:59 AM8/30/21
to chao yang, DMFTwDFT
Dear Chao,

2.  "beta": it seems  this parameter  is  the inverse of the temperature in meV,  I want study the properties of Yb under room temperature, therefore I set beta=25.85, is it right ?: Yes indeed beta is the inverse of temperature but in eV. you can do the basic conversion 1 eV = 11600 K.

Ps: as mentioned in my last comment it will be necessary to check if there is no difference between   ['f1', 'f2', 'f3', 'f4', 'f5', 'f6', 'f7']    et    ['f1'], ['f2'], ['f3'], ['f4'], 'f5'], ['f6'], ['f7']





Uthpala Herath

unread,
Aug 30, 2021, 10:10:20 AM8/30/21
to Franck, chao yang, DMFTwDFT
The difference between the bracketing would be to do the DMFT calculations without considering the f subshells (m) and only take the angular momentum quantum number (l). 
For instance, in the first method you will only find 3 columns in the self-energy/ Green's functions outputs for the matsubara frequency, real part of f-orbital and imaginary part of f-orbital. If you have them separated, each sub-orbital would have a real and imaginary column. 

- Uthpala

Franck

unread,
Aug 30, 2021, 10:16:27 AM8/30/21
to Uthpala Herath, chao yang, DMFTwDFT
Yes, exactly. That's why I ask him to set this parameter according to what he wants to analyze.

Uthpala Herath

unread,
Aug 30, 2021, 10:41:59 AM8/30/21
to Franck, chao yang, DMFTwDFT
Yes Frank that is correct. 

Chao, I wanted to ask how you got the values for n_tot (number of electrons in wannier subspace) and nf (occupancy)?
They are important for initializing the calculation. 

Thanks, 

Best,
Uthpala

chao yang

unread,
Aug 30, 2021, 7:00:21 PM8/30/21
to Uthpala Herath, Franck, DMFTwDFT
Dear Uthpala and Frank,
I set the two parameters with random number, just for test at the  first step. In my opinion,the value of n_tot should be 24, since the out-shell valence electron structures of Yb is d10f14. For the value of nf,It can be obtained by integrated the pdos of each orbital s. is it right?

Best,
Chao

Franck

unread,
Aug 30, 2021, 11:23:46 PM8/30/21
to chao yang, Uthpala Herath, DMFTwDFT
you need an approximate value of the real value. to determine this value you can quickly do a DFT+U calculation (vasp) and you will have in the OUTCAR file the value of the filling of each orbital. in the evening you use the script present in the electron_count.py. but personally I prefer to do it with vasp.

Uthpala Herath

unread,
Aug 31, 2021, 12:05:03 AM8/31/21
to chao yang, DMFTwDFT, Franck
Dear Chao, 

Yes, using Frank's suggestion you can obtain the necessary values for the input. 

I did some test scenarios to get a better understanding of the calculation at hand. Here's a summary. 

1. First, I wanted to see what the electronic configuration of Yb was and found it to be: 1s2 2s2 2p6 3s2 3p6 3d10 4s2 4p6 4d10 5s2 5p6 4f14 5d0 6s2 which differs from your d10 f14.

2. VASP offers two different POTCAR's for Yb, one with 24 valence electrons and Yb_2 which considers the last two electrons as valence and moves everything else to the core. 

    (i). With the first pseudo, I did the integration of DOS and found that the number of f electrons add up to be 14. For the same case, the d electrons were 0.33 which is minimal and establishes the 5d0 configuration. 
    (ii). With the second pseudo, the DOS integration for d states sum up to be 0.17 which is also close to zero. This further establishes the 5d0 state in the electronic configuration. 

    Based on their projected band structures, I used a relevant Wannier energy window to encompass each state for the calculations. 

3. In the initial case since f has all 14 electrons filled so I'm not sure if they electrons do anything interesting, please correct me if I'm wrong. 

Let me know what you think about these comments. This is very interesting material and I believe we all can learn a lot through these discussions. Please correct me if I'm wrong about the above as I have not worked with Yb before. 

Thanks, 

Best,
Uthpala

Franck

unread,
Sep 2, 2021, 5:32:20 AM9/2/21
to chao yang, DMFTwDFT
Dear Chao,

Can you send us your POSCAR?

Franck O.

--

Franck

unread,
Sep 2, 2021, 7:25:16 AM9/2/21
to chao yang, Uthpala Herath, DMFTwDFT

I wanted to check the occupation.

on the vasp site(https://www.vasp.at/wiki/index.php/Available_PAW_potentials), the "OPTIMAL" POTCAR is Yb_2 so I did the vasp calculation(with LORBIT=11) for each of the POTCAR [Yb, Yb_2, Yb_3]. here are the results for the orbital occupancy.


For POTCAR Yb  we have this:
 total charge
# of ion       s       p       d       f       tot
--------------------------------------------------
    1        2.258   6.078  21.966  10.764  41.066

For POTCAR Yb_2  we have this:
 total charge
# of ion       s       p       d       tot
------------------------------------------
    1        0.482   6.205   0.286   6.973

For POTCAR Yb_3  we have this:
 total charge
# of ion       s       p       d       tot
------------------------------------------
    1        0.513   6.705   0.768   7.986

and rememeber that

n-tot=s+p+d+f
nf=f and  or nf=d

It's up to you to see which POTCAR is stable and would best describe your type of calculation.

for  Yb n-tot=41.066
for  Yb_2  n-tot=6.973
for  Yb_3 n-tot= 7.986




















Uthpala Herath

unread,
Sep 2, 2021, 8:35:09 AM9/2/21
to Franck, chao yang, DMFTwDFT
A slight correction to what Franck mentioned;

n_tot = number of electrons in Wannier subspace. Not total number of electrons. 

If you project the d or f orbitals in wannier, this would the number of those electrons. It doesn't consider the s or p electrons you don't mention in the wannierization input. 

The integration of electron charge is not perfect within DFT which is why the NELECT value is not the same as the charge summation in the OUTCAR. 

Best,
Uthpala

chao yang

unread,
Sep 3, 2021, 4:53:09 AM9/3/21
to DMFTwDFT
 Dear  Uthpala,
 I still cannot run the DMFTwDFT successfully, could you give me some suggestions?  You can find those files in the attachment.

Best,
Chao
dmftelectron.tar.gz

chao yang

unread,
Sep 4, 2021, 7:29:30 PM9/4/21
to DMFTwDFT
Dear  Uthpala,
I don't know what's going on, but it worked now. I will share with you the results after the DMFTwdft code running ended.

Best,
Chao

chao yang

unread,
Sep 6, 2021, 12:59:04 AM9/6/21
to DMFTwDFT
Dear  Uthpala,
Do you know the meaning of the matrix in the kgen.py (under the example LaNiO3_vasp/NCSC/bands) and how to obtain it ? 

Best,
Chao

$ more kgen.py 
.......
 KPoints=[[0,0,0],[0.5,0.5,0],[0.5,0,0],[0,0,0]]
   KPoints=[matrix([[0.66972822, -1.16000331, 0.48613733],[0.66972822, 1.16000331, 0.48613733],[-1.33945645, 0.00000000, 0.48613733]]).dot(array(KPoints[i])) for i in range(len(KPoints))]
   KPoints=array(KPoints)
   SKPoints=['$\Gamma$','X','M','$\Gamma$']
   nk_band=500
   klist, dist_K, dist_SK = Create_kpath(KPoints,nk_band)
   klist=[matrix([[0.66972822, -1.16000331, 0.48613733],[0.66972822, 1.16000331, 0.48613733],[-1.33945645, 0.00000000, 0.48613733]]).I.dot(array(klist[i][0,:])) for i in range(len(klist))]
   #print klist,dist_K
........

Uthpala Herath

unread,
Sep 6, 2021, 1:12:01 AM9/6/21
to chao yang, DMFTwDFT
Hello Chao, 

These are the k-path for plotting the spectral function. 
The easiest way to do this would be to use the postDMFT.py script with the -autokp flag provided you have a KPOINTS file with a k-path in the DMFT directory. 

E.g. 
$postDMFT.py bands -plotplain -autokp

I just noticed that I haven't added it in the documentation as this was implemented later. However, you can get more options with postDMFT.py bands -h.

Let me know if that helps. 

Best,
Uthpala

chao yang

unread,
Sep 7, 2021, 10:07:46 PM9/7/21
to DMFTwDFT
Dear  Uthpala,
Could you give me an example KPOINTS file? I launched the postDMFT.py with the KPOINTS below but it cannot read the KPOINTS  correctly. 


Best,
Chao

$more KPOINTS
coord.13.02
 30
l
r
0 0 0   
0 0.5 0.5   

0 0.5 0.5
0.5 0.5 1    

0.5 0.5 1
0 0 0  

0 0 0
0.5 0.5 0.5  


$./postDMFT.py  bands -plotplain -autokp

.............
Interpolating points on real axis...
Reading a file  ./ac/Sig.out
Interpolation complete.


Traceback (most recent call last):
  File "./postDMFT.py", line 1686, in <module>
    args.func(args)
  File "./postDMFT.py", line 738, in bands
    args.knames, ticks, discontinuities, args.kplist = self.readKPOINTS(args)
  File "./postDMFT.py", line 1486, in readKPOINTS
    tick_labels = np.array(re.findall("!\s*(.*)", KPmatrix[0]))
IndexError: list index out of range 

Uthpala Herath

unread,
Sep 7, 2021, 10:28:37 PM9/7/21
to chao yang, DMFTwDFT
Hello Chao,

You can you a KPOINTS file like the following:

KPOINTS generated by PyProcar
40 ! Grid points
Line_mode
reciprocal
0.000000 0.000000 0.000000 ! GAMMA
0.000000 0.500000 0.000000 ! X
0.000000 0.500000 0.000000 ! X
0.500000 0.500000 0.000000 ! M
0.500000 0.500000 0.000000 ! M
0.000000 0.000000 0.000000 ! GAMMA

If you have an EIGENVAL from a DFT calculation and place it within the DMFT directory, with the -compare flag you can additionally have the DMFT spectral function and DFT bands on the same figure. For this remember to set the DFT Fermi energy in the DFT_mu.out file. 

Best,
Uthpala 

Uthpala Herath

unread,
Sep 7, 2021, 10:29:44 PM9/7/21
to chao yang, DMFTwDFT
I missed the spaces in the previous email. This is better:

KPOINTS generated by PyProcar
40 ! Grid points
Line_mode
reciprocal
0.000000 0.000000 0.000000 ! GAMMA
0.000000 0.500000 0.000000 ! X

0.000000 0.500000 0.000000 ! X
0.500000 0.500000 0.000000 ! M

0.500000 0.500000 0.000000 ! M
0.000000 0.000000 0.000000 ! GAMMA

chao yang

unread,
Sep 8, 2021, 4:44:53 AM9/8/21
to DMFTwDFT
Dear Uthpala,
how to set the DFT Fermi energy in the DFT_mu.out file ?  Use the Fermi energy replace the Mu value or just add the Fermi energy at the next line?

Best,
Chao

$ more DFT_mu.out 
5.324012999999999884e+00

Uthpala Herath

unread,
Sep 8, 2021, 8:38:59 AM9/8/21
to chao yang, DMFTwDFT
Dear Chao, 

Yes, just replace the value that is there. For this case since we're done with the DMFT calculation DFT_mu.out will not be used anymore. That's why I made it in a way so that the Fermi energy for the post-processing (spectral function) calculation would be obtained by the DFT fermi energy value set in DFT_mu.out

Another note is that you should use the Fermi energy calculated from the DFT self-consistent calculation and not the non-self consistent band structure calculation. 

Best,
Uthpala

chao yang

unread,
Sep 8, 2021, 9:00:08 AM9/8/21
to DMFTwDFT
Dear Uthpala,
I'm sucessfully obtained the A_k spectral and DFT bands on the same fig. Due to the A_k.eps file is too large, I can't send it to you from email. And the parameters are crude, the results are unbeliveable. 

Next, since the f-orbital electrons are very close to the Fermi level, I need take it, however, I tried many times but still can't get the DMFT run sucessfully. may you help me ?
The attachment is the INPUT.py I used to get the A_k.eps only with d-orbital electrons.

Best,
Chao

INPUT.py

Uthpala Herath

unread,
Sep 8, 2021, 9:43:12 AM9/8/21
to chao yang, DMFTwDFT
Hello Chao, 

You can use the following input for f-orbitals. However, you'll need to select a pseudopotential that has f electrons in valence and that would define the n_tot and nf values. You would also need to plot the DOS or projected bands to find the wannier energy window that encompasses the orbitals. Both these values are important for the initialization. Sometimes they would eventually acquire the correct values but based on the initialization you may end up in some other local minimum. 

Best,
Uthpala

INPUT.py

chao yang

unread,
Sep 9, 2021, 4:55:47 AM9/9/21
to DMFTwDFT
Dear Uthpala,
Since both the d-orbital and f-orbital electrons have distribution near the Fermi level, we should take them both  into consideration during the construction of wannier functions. This time I want study the correlation effects of the f-orbitals with the INPUT.py attached. However, the DMFT can't run succesfully. 
Could you give me some suggestions?
Please find the INPUT.py and the output file with the errors.

Best,
Chao
 
$ more INPUT.py 
################   Input parameters for DFT+DMFT calculations   ##################

###### Main loop parameters ###########

p = {
    "Niter": 1,  # Number of DFT+DMFT iterations
    "Ndft": 1,  # Number of DFT iterations
    "Nit": 15,  # Number of DMFT iterations
    "n_tot": 14,  # Number of total electrons
    "nf": 14,  # Number of target Nd
    "nspin": 1,  # Number of total spins
    "atomnames": ["Yb","Yb"],  # The name of atoms
    "orbs": ["d","f"],  # The name  of orbitals
    "L_rot": [0, 0],  # Whether rotate local axis or not
    "cor_at": [["Yb1"]],  # Correlated atoms, put symmetry atoms in the same list
    "cor_orb": [[["f1", "f2", "f3", "f4", "f5", "f6", "f7"]]],
    "U": [5.0],  # Intra-U for each cor_at
    "J": [1.0],  # Hund's coupling
    "alpha": [0.2],  # Double counting parameter
    "mix_sig": 0.2,  # Mixing parameter for Sigma
......

error tips:
...
  File "/THL8/home/biyan3/software/dmftwdft/DMFTwDFT-master-nscc/bin/Struct.py", line 236, in Compute_cor_idx
    self.cor_idx[idx] = 2 + idx1 + j
    self.cor_idx[idx] = 2 + idx1 + j
IndexError: index 7 is out of bounds for axis 0 with size 7
IndexError: index 7 is out of bounds for axis 0 with size 7
    self.cor_idx[idx] = 2 + idx1 + j
    self.cor_idx[idx] = 2 + idx1 + j
IndexError    self.cor_idx[idx] = 2 + idx1 + j
IndexError    self.cor_idx[idx] = 2 + idx1 + j
IndexError: index 7 is out of bounds for axis 0 with size 7
    self.cor_idx[idx] = 2 + idx1 + j
IndexError: index 7 is out of bounds for axis 0 with size 7
: index 7 is out of bounds for axis 0 with size 7
    self.cor_idx[idx] = 2 + idx1 + j
: index 7 is out of bounds for axis 0 with size 7
    self.cor_idx[idx] = 2 + idx1 + j
...
INPUT.py
slurm-3659134.out

Uthpala Herath

unread,
Sep 19, 2021, 12:57:09 AM9/19/21
to chao yang, DMFTwDFT
Dear Chao, 

As of now we cannot study treat both d and f orbitals at the same time. You would have to first treat d and then f. 

for d:

 "atomnames": ["Yb"],  # The name of atoms
    "orbs": ["d"],  # The name  of orbitals
    "L_rot": [0],  # Whether rotate local axis or not

    "cor_at": [["Yb1"]],  # Correlated atoms, put symmetry atoms in the same list
"cor_orb":   [[['d_z2'],['d_x2y2'],['d_xz'],['d_yz'],['d_xy']]]

or if you want to preserve d degeneracy:
"cor_orb":   [[['d_z2','d_x2y2'],['d_xz','d_yz','d_xy']]],

and then for f:

 "atomnames": ["Yb"],  # The name of atoms
    "orbs": ["f"],  # The name  of orbitals
    "L_rot": [0],  # Whether rotate local axis or not

    "cor_at": [["Yb1"]],  # Correlated atoms, put symmetry atoms in the same list
"cor_orb":   [[['f1'],['f2'],['f3'],['f4'],['f5'],['f6'],['f7']]]

for preserved f degeneracy: 
"cor_orb":   [[['f1','f2','f3','f4','f5','f6','f7']]]

Let me know if that works for you. 

Best,
Uthpala

Message has been deleted

Uthpala Herath

unread,
Sep 27, 2021, 12:54:01 AM9/27/21
to chao yang, DMFTwDFT
Hello Chao, 

Can you send me your QE inputs and INPUT.py please? I'll take a look. 

Thanks, 

Best,
Uthpala

On Mon, Sep 27, 2021 at 12:51 AM chao yang <bitan...@gmail.com> wrote:
Dear Uthpala,
Since DMFT cannot handle the SOC, I will use QE to do do the calculation with the scalar relativistic aproximation PP . And this time I will only take the d-orbitals to study the correlation effect.

And of course I get the problem now, it's a hard work for me.
.................
$ ls
dmft                       run-nscf.sh        slurm-3735371.out  slurm-3742534.out               SrVO3.mmn              SrVO3.pw2wannier90.out  SrVO3.wout           V.pbe-spnl-kjpaw_psl.1.0.0.UPF
INPUT.py                   run-pw.sh          slurm-3735378.out  Sr.pbe-spn-kjpaw_psl.1.0.0.UPF  SrVO3.nnkp             SrVO3.save              SrVO3.xml            wannier90x.sh
O.pbe-n-kjpaw_psl.0.1.UPF  run-qe.sh          slurm-3735466.out  SrVO3.amn                       SrVO3.nscf.in          SrVO3.scf.in            submit.sh            wannier-win
out.wannier90xSrVO3        slurm-3735334.out  slurm-3735568.out  SrVO3.chk                       SrVO3.nscf.out         SrVO3.scf.out           submit-wannier90.sh
para_com.dat               slurm-3735353.out  slurm-3737245.out  SrVO3.eig                       SrVO3.pw2wannier90.in  SrVO3.win               submit-x90.sh
$ cd dmft/
$ ls
$ Copy_input.py -structurename SrVO3 /THL8/home/biyan3/runfile/Yb/qe-dmftwdft/SrVO3_qe/
No Quantum Espresso results have been found in a /THL8/home/biyan3/runfile/Yb/qe-dmftwdft/SrVO3_qe/ directory!
No VASP results have been found in a /THL8/home/biyan3/runfile/Yb/qe-dmftwdft/SrVO3_qe/ directory!
No aiida results have been found in a /THL8/home/biyan3/runfile/Yb/qe-dmftwdft/SrVO3_qe/ directory!
No Siesta results have been found in a /THL8/home/biyan3/runfile/Yb/qe-dmftwdft/SrVO3_qe/ directory!
wannier90.chk must exist in a /THL8/home/biyan3/runfile/Yb/qe-dmftwdft/SrVO3_qe/ directory! Exiting!
..................
can you help me check it?

Best,
Chao

Uthpala Herath

unread,
Sep 27, 2021, 12:56:14 AM9/27/21
to chao yang, DMFTwDFT
Also, did you check the SrVO3 Quantum Espresso example found in https://github.com/DMFTwDFT-project/DMFTwDFT/tree/master/examples/SrVO3_qe ?

- Uthpala
Message has been deleted
Message has been deleted

Uthpala Herath

unread,
Sep 27, 2021, 10:54:02 PM9/27/21
to chao yang, DMFTwDFT
Hello Chao, 

I just checked and it works fine. Are you sure your QE calculations have completed successfully?

The command to copy is:

Copy_input.py -structurename SrVO3 ../

But if you are running the DMFT.py script this is all handled automatically. 

Best,
Uthpala

On Mon, Sep 27, 2021 at 10:49 PM chao yang <bitan...@gmail.com> wrote:
Dear Uthpala, 
I'm truly running the SrVO3  Quantum Espresso example. The problem is  I  cannot successfully run the script Copy_input.py  ( i.e. , Copy_input.py -structurename SrVO3 path-to-DFTwwannier90). 

Best,
Chao
Reply all
Reply to author
Forward
0 new messages