Geometry optimization by using LIBXC caused no SCF iteration with error NaN

[Marco]

Hello,

I looked at your input file and everything seemed okay. The reason you need 
to include a &HF section in your input file for M06-2X is that the M06-2X 
functional is a global hybrid functional which incorporates 54% HF 
exchange. It is my understanding that the minimum required input 
specification in the &HF section for a calculation employing a hybrid 
functional is the &SCREENING keyword. Take a look at the input files 
provided in the regtest-libxc directory (for example the 
"H2O-hybrid-b3lyp_libxc_uks.inp" file). 

~/input/cp2k/cp2k-2.4.0/tests/QS/regtest-libxc/

I looked at some of the source code files related to the calling of the XC 
functionalities. In particular, the xc_libxc.f file states the following as 
a comment at the beginning of the file.

!>      WARNING: In the subroutine libxc_lsd_calc, it could be that the
!>      ordering for the 1st index of v2lapltau, v2rholapl, v2rhotau,
!>      v2sigmalapl and v2sigmatau is not correct. For the moment it does 
not
!>      matter since the calculation of the 2nd derivatives for meta-GGA
!>      functionals is not implemented in CP2K.

If this is the case, then it is possible that a meta-GGA functional (which 
generally is a function of the laplacian of the density) such as M06-2X may 
not work properly in CP2K. Sorry I could not be of more help.

Best regards,
Marco

On Monday, April 14, 2014 10:28:37 AM UTC-4, Jingyun Ye wrote:
>
> Hi, Marco
>
> I tried to combine HF with M06-2X to run the job. But the same error was 
> obtained. Why we need to include HF to M06-2X? 
> I attached input and output files, you may find something important 
> information in it. 
>
> Thanks very much. 
>
> On Monday, April 14, 2014 12:33:06 AM UTC-4, Marco wrote:
>>
>> Hi,
>>
>> Something that comes to mind is the inclusion of HF exchange in going 
>> from the PBE functional to the M06-2X functional. Since your input file 
>> parameters were not shown, I want to mention that you should probably 
>> include a &HF section in your input file (
>> http://manual.cp2k.org/cp2k-2_5-branch/CP2K_INPUT/FORCE_EVAL/DFT/XC/HF.html) 
>> when using the M06-2X functional. If you need examples of the &HF section, 
>> I would look into the regtest files for sample inputs. I usually grep for 
>> the concepts of interest. For example, to get examples of the use of the 
>> &HF section, I would go into the QS directory:
>>
>> cd /cp2k-2.5.1/tests/QS/
>>
>> and then grep for &HF
>>
>> grep -i '&HF' */*.inp
>>
>> The grep will provide a list of files which make use of the &HF section.
>>
>> Something like the following may be useful for specifying the HF exchange 
>> parameters in calculations that make use of HF exchange such as the M06-2X 
>> functional:
>>  &XC
>>     FUNCTIONAL_ROUTINE NEW
>>     DENSITY_CUTOFF 1.0E-10
>>     GRADIENT_CUTOFF 1.0E-10
>>     TAU_CUTOFF 1.0E-10
>>     &HF
>>       PW_HFX_BLOCKSIZE 50
>>       &INTERACTION_POTENTIAL
>>         CUTOFF_RADIUS 3.0
>>         POTENTIAL_TYPE TRUNCATED
>>         T_C_G_DATA ./t_c_g.dat
>>       &END INTERACTION_POTENTIAL
>>       &MEMORY
>>         EPS_STORAGE_SCALING 1.0E-01
>>         MAX_DISK_SPACE YYYY
>>         MAX_MEMORY YYYY
>>         STORAGE_LOCATION ./STORAGE/
>>       &END MEMORY
>>       &PERIODIC
>>         NUMBER_OF_SHELLS -1
>>       &END PERIODIC
>>       &SCREENING
>>         EPS_SCHWARZ 1.0E-08
>>         EPS_SCHWARZ_FORCES 1.0E-06
>>         SCREEN_P_FORCES TRUE
>>       &END SCREENING
>>     &END HF
>>     &XC_FUNCTIONAL B3LYP
>>     &END XC_FUNCTIONAL
>>     &XC_GRID
>>       USE_FINER_GRID
>>       XC_DERIV SPLINE2_SMOOTH
>>       XC_SMOOTH_RHO NN10
>>     &END XC_GRID
>>   &END XC
>>
>> Best regards,
>> Marco
>>
>> On Sunday, April 13, 2014 2:33:23 PM UTC-4, Jingyun Ye wrote:
>>>
>>> Hi, all
>>>
>>> I am trying to use M06-2X functional in CP2K (PBE potential)  to compare 
>>> the binding energy of CO2  with Gaussian results( also use M06-2X 
>>> functional, same basis sets). 
>>>
>>> But the calculation stops  at the SCF every time with the following 
>>> information. But if I change the functional M06-2X to PBE, the job runs 
>>> well. 
>>>
>>> Anyone know what's the problem? Any suggestion will be help. 
>>> Thanks very much.
>>>
>>>  
>>>  SCF WAVEFUNCTION OPTIMIZATION
>>>
>>>   ----------------------------------- OT 
>>> ---------------------------------------
>>>
>>>   Allowing for rotations:  F
>>>   Optimizing orbital energies:  F
>>>   Minimizer      : CG                  : conjugate gradient
>>>   Preconditioner : FULL_ALL            : diagonalization, state selective
>>>   Precond_solver : DEFAULT
>>>   Line search    : 2PNT                : 2 energies, one gradient
>>>   stepsize       :    0.15000000
>>>   energy_gap     :    0.00100000
>>>
>>>   eps_taylor     :   0.10000E-15
>>>   max_taylor     :             4
>>>
>>>   mixed_precision    : F
>>>
>>>   ----------------------------------- OT 
>>> ---------------------------------------
>>>
>>>   Step     Update method      Time    Convergence         Total energy   
>>>  Change
>>>   
>>> ------------------------------------------------------------------------------
>>>
>>>   Trace(PS):                                  209.9999999999
>>>   Electronic density on regular grids:                   NaN             
>>>     NaN
>>>   Core density on regular grids:              210.0000000000       
>>>  0.0000000000
>>>   Total charge density on r-space grids:                 NaN
>>>   Total charge density g-space grids:                    NaN
>>>
>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140414/5ec33b73/attachment.htm>