strange relativistic (sra, srals) DOS

Posted on : December 15, 2021 (Wed) 21:03:49

by Rafal Topolnicki

Dear AkaiKKR Administrator,

I try to compute the electronic structure of (TaNb)0,67(MoHfW)0,33 alloy.
For nrl I have a reasonable DOS and band-structure (see file below). However once the relativistic effects are included (sra, srals) there is a strong spike in total DOS around Fermi level and the bandplot gets totally distorted. Only total DOS is affected (black line in figures) - when I sum up (gray line) the contributions from different species (weighted according to sample composition) the result seems to be correct - however such sum cannot be done in case of band structure so this is not really a solution.
I attach my input file.

Is there a way to fix that? What I’m I doing wrong? Any help would be very much appreciated.

This is what I tried so far and didn’t help:
older version of the code
Code compiled with ifort and gfortran
Different lattice constant
Different XC functionals (with and without ASA mode)
Both larger and smaller value of ewidth and edelt
Different values of pmix and bzqlty
Tried magnetic calculations (system is nonmagnetic)
mxl=4
in specx.f i’ve changed the tolerance to tol/1d-7/ and increased msex and mse0 to 501 and 75 respectively.

Attachments (figures, input, output, spc file, potential)
http://www.ifd.uni.wroc.pl/~topolnicki/akaikkr/


Inputfile:
======================
c----------Ta33 Nb33 Mo11 Hf11 W11----------------------------
c---- Ta 73 Nb 41 Mo 42 Hf 72 W 74
c----
go alloy_potential
c------------------------------------------------------------
c brvtyp a c/a b/a alpha beta gamma
bcc 6.1767 , , , , , ,
c------------------------------------------------------------
c edelt ewidth reltyp sdftyp magtyp record
0.001 0.8 sra pbeasa nmag init
c------------------------------------------------------------
c outtyp bzqlty maxitr pmix
update 30 1000 0.005
c------------------------------------------------------------
c ntyp
1
c------------------------------------------------------------
c type ncmp rmt field mxl anclr conc
X 5 0 0.0 3
73 0.335
41 0.335
42 0.11
72 0.11
74 0.11
c------------------------------------------------------------
c natm
1
c------------------------------------------------------------
c atmicx atmtyp
0a 0b 0c X
c------------------------------------------------------------
c----------Ta33 Nb33 Mo11 Hf11 W11----------------------------
c---- Ta 73 Nb 41 Mo 42 Hf 72 W 74
c----
dos alloy_potential
c------------------------------------------------------------
c brvtyp a c/a b/a alpha beta gamma
bcc 6.1767 , , , , , ,
c------------------------------------------------------------
c edelt ewidth reltyp sdftyp magtyp record
0.001 0.8 sra pbeasa nmag 2nd
c------------------------------------------------------------
c outtyp bzqlty maxitr pmix
quit 30 1000 0.005
c------------------------------------------------------------
c ntyp
1
c------------------------------------------------------------
c type ncmp rmt field mxl anclr conc
X 5 0 0.0 3
73 0.335
41 0.335
42 0.11
72 0.11
74 0.11
c------------------------------------------------------------
c natm
1
c------------------------------------------------------------
c atmicx atmtyp
0a 0b 0c X
c------------------------------------------------------------
c----------Ta33 Nb33 Mo11 Hf11 W11----------------------------
c---- Ta 73 Nb 41 Mo 42 Hf 72 W 74
c----
spc alloy_potential
c------------------------------------------------------------
c brvtyp a c/a b/a alpha beta gamma
bcc 6.1767 , , , , , ,
c------------------------------------------------------------
c edelt ewidth reltyp sdftyp magtyp record
0.001 0.8 sra pbeasa nmag 2nd
c------------------------------------------------------------
c outtyp bzqlty maxitr pmix
quit 30 1000 0.005
c------------------------------------------------------------
c ntyp
1
c------------------------------------------------------------
c type ncmp rmt field mxl anclr conc
X 5 0 0.0 3
73 0.335
41 0.335
42 0.11
72 0.11
74 0.11
c------------------------------------------------------------
c natm
1
c------------------------------------------------------------
c atmicx atmtyp
0a 0b 0c X
c------------------------------------------------------------
c--P->Gamma->H->P->N->H->Gamma->N
c--inne oznaczenie wspolrzednych punktow - zobaczyc przyklady w AkaiKKR
500
1/2 1/2 1/2
0 0 0
0 1 0
1/2 1/2 1/2
1/2 1/2 0
0 1 0
0 0 0
1/2 1/2 0
======================



Best regards,
Rafal Topolnicki

[Re:03] strange relativistic (sra, srals) DOS

Posted on : January 11, 2022 (Tue) 21:43:08

by Hitoshi GOMI

Hello Rafal Topolnicki,

> all calculation parameters should be consistent across all calculations (go, dos, spc)
> while you suggest having different values of ewidth for go (0.6) and dos (0.8).

Basically, the same parameters should be used for go and dos calculations.
However, ewidth can be changed.
The ewidth for the go calculation needs to be chosen carefully to make the calculation converge.
On the other hand, the dos calculation just reads the result of the go calculation, so you are free to specify the range you want to plot.
In the default settings, you can specify the ewidth of 4/3 of the go calculation for the dos calculation to see the bottom of the band in the go calculation.
Please confirm "complex energy mesh" in the output from each calculation.



> why do we see no peak when summing up contributions from individual atoms? Both atom-wise and total-dos should describe the same electronic structure.

It is not surprising that PDOS and TDOS do not match.
Because PDOS does not include the interstitial region, while TDOS does.



> I would also like to study my system under high pressure.

In general, as the volume decreases, the bandwidth increases.
Therefore, if you want to calculate high pressure, you need to increase the ewidth.
When calculating high pressures, I recommend changing the lattice parameters little by little and see how it goes.



P.S.
If you don't deal with lanthanides or actinides, I feel it is better to use xml=2. I'm not sure because I don't deal with atoms with large atomic numbers...

Hitoshi GOMI

[Re:02] strange relativistic (sra, srals) DOS

Posted on : January 05, 2022 (Wed) 18:39:35

by Rafal Topolnicki

Dear Hitoshi GOMI,

Thank you very much for your fast and accurate answer. This was super helpful.
I did as you recommended and the calculation converged to correct DOS (no peaks). I tested values of ewidth myself prior to posting the topic but didn't go as low as 0.6.
While reading the BBS and some manuals I had an impression that all calculation parameters should be consistent across all calculations (go, dos, spc) while you suggest having different values of ewidth for go (0.6) and dos (0.8). Only very recently I found a post regarding this issue
http://gomisai.blog75.fc2.com/blog-entry-723.html

There is still something I don't understand. If the problem is due to the core electrons, why do we see no peak when summing up contributions from individual atoms? Both atom-wise and total-dos should describe the same electronic structure.

I would also like to study my system under high pressure. The lattice constant is then 5.614. For this system I used even smaller ewidths (0.6, 0.5, 0.45) but peak is still present around Fermi level. For even lower values of ewidth (0.4, 0.3, etc) there is no convergence and resulting DOS makes no sense - probably contour energy does not include all valence states.
Following ideas form http://gomisai.blog75.fc2.com/blog-entry-709.html I tried to use large ewidth (up to 3) to include core states, but then the calculations does not converge (different pmix tested) - I got message
***msg in spmain...new ew and ez generated:
Is there a solution for this problem (not having a peak in the DOS)? Should I look more densely for ewidth for which contour goes exactly between the core and valence states? Or perhaps the high pressure conditions are beyond the capabilities of the method?

Best regards,
Rafal Topolnicki

[Re:01] strange relativistic (sra, srals) DOS

Posted on : December 15, 2021 (Wed) 23:47:02

by Hitoshi GOMI

Hello Rafal Topolnicki,

If we use ewidth=0.8, the nrl calculation succeeds, but the sra(ls) calculation fails due to the core electron state near the bottom of ewidth.
I have successfully completed all nrl, sra and srals calculations with ewidth=0.6 (go) and ewidth=0.8 (dos and spc).

Please try the following inputs.

Hitoshi GOMI

c----------Ta33 Nb33 Mo11 Hf11 W11----------------------------
c---- Ta 73 Nb 41 Mo 42 Hf 72 W 74
c----
go data/bccTaNbMoHfW_srals
c------------------------------------------------------------
c brvtyp a c/a b/a alpha beta gamma
bcc 6.1767 , , , , , ,
c------------------------------------------------------------
c edelt ewidth reltyp sdftyp magtyp record
0.001 0.6 srals pbeasa nmag 2nd
c------------------------------------------------------------
c outtyp bzqlty maxitr pmix
update 9 200 0.023
c------------------------------------------------------------
c ntyp
1
c------------------------------------------------------------
c type ncmp rmt field mxl anclr conc
X 5 0 0.0 3
73 0.335
41 0.335
42 0.11
72 0.11
74 0.11
c------------------------------------------------------------
c natm
1
c------------------------------------------------------------
c atmicx atmtyp
0a 0b 0c X
c------------------------------------------------------------




c----------Ta33 Nb33 Mo11 Hf11 W11----------------------------
c---- Ta 73 Nb 41 Mo 42 Hf 72 W 74
c----
dos data/bccTaNbMoHfW_srals
c------------------------------------------------------------
c brvtyp a c/a b/a alpha beta gamma
bcc 6.1767 , , , , , ,
c------------------------------------------------------------
c edelt ewidth reltyp sdftyp magtyp record
0.001 0.8 srals pbeasa nmag 2nd
c------------------------------------------------------------
c outtyp bzqlty maxitr pmix
update 9 200 0.023
c------------------------------------------------------------
c ntyp
1
c------------------------------------------------------------
c type ncmp rmt field mxl anclr conc
X 5 0 0.0 3
73 0.335
41 0.335
42 0.11
72 0.11
74 0.11
c------------------------------------------------------------
c natm
1
c------------------------------------------------------------
c atmicx atmtyp
0a 0b 0c X
c------------------------------------------------------------




c----------Ta33 Nb33 Mo11 Hf11 W11----------------------------
c---- Ta 73 Nb 41 Mo 42 Hf 72 W 74
c----
spc data/bccTaNbMoHfW_srals
c------------------------------------------------------------
c brvtyp a c/a b/a alpha beta gamma
bcc 6.1767 , , , , , ,
c------------------------------------------------------------
c edelt ewidth reltyp sdftyp magtyp record
0.001 0.8 srals pbeasa nmag 2nd
c------------------------------------------------------------
c outtyp bzqlty maxitr pmix
update 9 200 0.023
c------------------------------------------------------------
c ntyp
1
c------------------------------------------------------------
c type ncmp rmt field mxl anclr conc
X 5 0 0.0 3
73 0.335
41 0.335
42 0.11
72 0.11
74 0.11
c------------------------------------------------------------
c natm
1
c------------------------------------------------------------
c atmicx atmtyp
0a 0b 0c X
c------------------------------------------------------------
c--P->Gamma->H->P->N->H->Gamma->N
500
1/2 1/2 1/2
0 0 0
0 1 0
1/2 1/2 1/2
1/2 1/2 0
0 1 0
0 0 0
1/2 1/2 0
c------------------------------------------------------------