internal ERROR: LINEAR_RESPONSE_DIIS matrix is zero

[jacekwu]

Dear all

I am trying to run a DFPT calculation on a 2D InSe structure to obtain its piezoelectric tensor when I encountered this error:

Code: Select all

 internal ERROR: LINEAR_RESPONSE_DIIS matrix is zero, try to call with LRESET
           1          11           2
 (2.167438697314308E-016,-1.045076509888809E-034)
 (-9.135338366750134E-011,-4.088538600438221E-010)
 (-9.135338366750134E-011,4.088538600438221E-010)
 (9.279167646373218E-002,8.092095728910679E-020)
 (-9.135421110726070E-011,4.088539368739345E-010)
 (9.279167602859092E-002,-1.393411982901073E-010)

A search online revealed several similar cases but all without a reason nor solution to the problem. It will be great if someone who knows a thing or two about this issue would share his or her experience.

Below are the contents of my input:

INCAR
------------------------------------------------

Code: Select all

ADDGRID = True
AGGAC = 0.0
ALGO = ALL
EDIFF = 1e-10
EDIFFG = -0.005
ENCUT = 520
GGA = MK
IBRION = 8
ICHARG = 0
ISIF = 2
ISMEAR = -5
ISPIN = 2
KPOINT_BSE = -1 0 0 0
LAECHG = False
LASPH = True
LCHARG = True
LEPSILON = True
LORBIT = 11
LPEAD = True
LREAL = FALSE
LUSE_VDW = .TRUE.
LVHAR = True
LWAVE = False
MAGMOM = 4*0.0
NELM = 300
PARAM1 = 0.1234
PARAM2 = 1.0
POTIM = 0.1
PREC = Accurate
SIGMA = 0.1

KPOINTS
------------------------------

Code: Select all

pymatgen 4.7.6+ generated KPOINTS with grid density = 1373 / atom
0
Gamma
12 12 2

POSCAR
---------------------------

Code: Select all

In2 Se2
1.0
2.015906 -3.491652 0.000000
2.015906 3.491652 0.000000
0.000000 0.000000 25.659413
In Se
2 2
direct
0.000000 0.000000 0.945487 In
0.000000 0.000000 0.054513 In
0.333333 0.666667 0.105144 Se
0.333333 0.666667 0.894856 Se

POTCAR is combined from these two pseudopotential files: PAW_PBE In_d 06Sep2000 and PAW_PBE Se 06Sep2000

[jihwan_lee]

Is there any solution with this Error?

I also faced same error in my DFPT calculation:

internal ERROR: LINEAR_RESPONSE_DIIS matrix is zero, try to call with LRESET
1 52 2
(8.931425424911767E-016,7.331514824884541E-033)
(-1.257815733058013E-011,8.687594872110668E-012)
(-1.257815733058013E-011,-8.687594872110668E-012)
(1.446654939283351E-003,2.432155323741877E-021)
(-1.258068600029599E-011,-8.687432092087997E-012)
(1.446654872575469E-003,1.193010228610683E-011)

[henrique_miranda]

Dear user,

The DFPT routines do not work with LUSE_VDW = .TRUE.
wiki/index.php/VdW-DF_functional_of_Lan ... vist_et_al.

I recommend you try using the finite differences approach by changing IBRION = 8 to IBRION = 6:
wiki/wiki/index.php/IBRION