Warning about linearisation energies in s.c. loop and minimum eigenvalue

[Gesgeron]

Hi, everybody,

During our calculation about MXene,some 2D early transition metal carbides like Ti2C.
The structure of Ti2C is from the relaxation of VASP.
During the elk calculation,I can't get rid of the warning:

Elk code version 4.3.06 started

Info(elk): current task :      0
Info(checkmt): reduced muffin-tin radius of species   1 (Ti) from   2.4000 to   1.1719
Info(checkmt): reduced muffin-tin radius of species   2 (C) from   1.8000 to   0.8790
Info(findsymcrys): atomic basis shift (lattice) :
   0.000000000       0.000000000     -0.4995543400    
See GEOMETRY.OUT for new atomic positions

Warning(linengy): could not find   5 linearisation energies in s.c. loop     1

Warning(occupy): minimum eigenvalue less than minimum linearisation energy :   -6.535589309      -2.287200000    
 for s.c. loop     1

Warning(linengy): could not find   5 linearisation energies in s.c. loop     2

Warning(occupy): minimum eigenvalue less than minimum linearisation energy :   -6.984179210      -2.287200000    
 for s.c. loop     2

Warning(linengy): could not find   5 linearisation energies in s.c. loop     3

Warning(occupy): minimum eigenvalue less than minimum linearisation energy :   -6.868855199      -2.287200000    
 for s.c. loop     3

The elk.in is following:

tasks
0

xctype
20

autolinengy
true

mixtype
3

maxscl
200

nempty
10

sppath
'.../elk-4.3.6/species/'

avec
   1.519767823       2.632315084       0.000000000    
   3.039535645       0.000000000       0.000000000    
   0.000000000       0.000000000       22.60601465    

atoms
   2                                    : nspecies
'Ti.in'                                 : spfname
   2                                    : natoms; atposl, bfcmt below
    0.66666669    0.66666666    0.55064953    0.00000000  0.00000000  0.00000000
    0.33333334    0.33333331    0.05064953    0.00000000  0.00000000  0.00000000

'C.in'                                  : spfname
   1                                    : natoms; atposl, bfcmt below
    0.00000000    0.00000000    0.49955434    0.00000000  0.00000000  0.00000000

ngridk
5 5 1

I try to add some tags, like trimvg, rgkmax, gmaxvr...but these doesn't work.

[mfechner]

Dear Gesgeron,

did you take the plain vasp strucutre ? Considering your lattice constants I guess yes, so you problem is maybe that the lattice constant in vasp are provided in Angstrom and in Elk in Bohrradius (a0). So you have to scale your ''avec'' lattice vectors to a0 by providing

scale
1.8897261246

which scales all lattice vectors equally to a0. Once you can rerun and everything should work now.

best regards
Michael

PS: Please remember to check convergence against RGKMAX, GMAXV, lmaxo, lmaxapw ..... its a little bit more demanding compared to vasp

[Gesgeron]

Thank you for your answer,Michael.
I added the tag "scale",the warning disappear. The elk shows as following :

Elk code version 4.3.06 started

Info(elk): current task :      0
Info(checkmt): reduced muffin-tin radius of species   1 (Ti) from   2.4000 to   2.2401
Info(checkmt): reduced muffin-tin radius of species   2 (C) from   1.8000 to   1.6800
Info(findsymcrys): atomic basis shift (lattice) :
 -1.0000000000     -1.0000000000     -0.4995543400    
See GEOMETRY.OUT for new atomic positions

Elk code stopped

I am the rookie of the elk, just like this process,Does it over?

[Brice KENGNI ZANGUIM]

Hello mfechner.

I had the same problem and after looking in this furum i saw that my scale tag was in Amgstrom intead of Bohr.

I have a question : what is the correlation between the interatomic distance and the fact that the code can not find the linerize energy ?

Thanks

[mfechner]

Actually this means your calculations finished you may find a detailed list of each iteration in

INFO.OUT, where at the end the converged total energy etc.. is listed. You may also check all other created files for further information. From here on you can now compute dos --> tasks 10 etc ....
you may want to check the examples for more details on the input requiered.

best regards
Michael

[Gesgeron]

Thank you, dear Michael~
I can run the ground state calculations now.

[mfechner]

Dear Brice,

the point is that the ''default'' potential are delivered with default linearization energies matching the radii with which they have been generated. Hence if you alter the radii strongly you may also have to change the linearization energies as it is now a different !

best regards
Michael