[CP2K:2224] Re: Mulliken charge problem

[Teodoro Laino]

If you would explain properly what is your problem.. than you may 
possibly have more chances of getting a reply!

What is the problem?
- Is the problem that a and b both for PBE or BLY differ by 0.002 Angstrom
or
- is the problem that PBE and BLYP give different results
or
- is the problem that both PBE and BLYP are totally off compared to the 
experimental value

??

About the problem of not getting b=3^0.5*a for orthorombic keep in mind 
that cell optimization suffer like all optimization of
finding local minima.. if your starting  geometry/cell  is particularly 
wrong you may easily end-up with an high energy local minima.

Teo

obo wrote:
> Nobody here who can help me?
>
> On 11 Aug., 11:10, obo <bo... at matcalc.de> wrote:
>   
>> thank you!
>> these Basis Sets works better. the error with the poisson and the
>> cholesky error is vanished.
>> but I think the relaxation of the cell is still wrong. after
>> relaxation of the hexagonal cell, the two vector have not the same
>> length. with the pbe functional the lengths are only a=4.816 b=4.818
>> aangstrom and with the BLYP 4.845 and 4.847!!!
>> The biggest variation is in the O-Si-O angles, which range from 105°
>> to 115°!
>> The problem with the orthorhombic cell is still present. The relation
>> b=3^0.5*a is not fulfilled and so the Mulliken charges are also wrong.
>> For example I've got the following charges:
>>
>>   Atom     |    Charge
>>
>>      1  Si    0.957166
>>      2  Si    0.955747
>>      3  Si    0.919798
>>      4  O    -0.467908
>>      5  O    -0.488780
>>      6  O    -0.459461
>>      7  O    -0.473715
>>      8  O    -0.454923
>>      9  O    -0.487922
>>     10  Si    0.957148
>>     11  Si    0.955763
>>     12  Si    0.919796
>>     13  O    -0.467886
>>     14  O    -0.488787
>>     15  O    -0.459514
>>     16  O    -0.473733
>>     17  O    -0.454871
>>     18  O    -0.487918
>>   Total       0.000000
>>
>> On 10 Aug., 09:11, Matthias <matthia... at psi.ch> wrote:
>>
>>     
>>> cp2k/tests/QS/BASIS_MOLOPT
>>>       
>>> Matthias
>>>       
>>> On Aug 10, 9:05 am, obo <bo... at matcalc.de> wrote:
>>>       
>>>> thanks for your answer.
>>>> do you know where i can download a BASIS SET file with these MOLOPT
>>>> basis sets?
>>>>         
>>>> Oliver
>>>>         
>>>> On 10 Aug., 09:01, Matthias <matthia... at psi.ch> wrote:
>>>>         
>>>>> The overlap matrix for such basis sets becomes ill-conditioned. Please
>>>>> note, augmented basis sets as employed for molecular systems are
>>>>> usually not suited for condensed systems and shouldn't be used for
>>>>> such systems. You may try the MOLOPT basis sets e.g. DZVP-MOLOPT-SR-
>>>>> GTH-q4 for Si and DZVP-MOLOPT-SR-GTH-q6 for O for your quartz systems.
>>>>>           
>>>>> Matthias
>>>>>           
>>>>> On Aug 10, 8:39 am, obo <bo... at matcalc.de> wrote:
>>>>>           
>>>>>> I have tested different Basis Sets. With confined Basis Sets the
>>>>>> problem becomes better but there is a difference anymore.
>>>>>> With an augment Basis Set I get the following error:
>>>>>>             
>>>>>>  *** 08:35:30 ERRORL2 in cp_fm_cholesky:cp_fm_cholesky_decompose
>>>>>> err=-300   ***
>>>>>>  *** condition FAILED at line 92
>>>>>>             
>>>>>> I have reada about linear dependences in the BASIS SET. Can somebody
>>>>>> explain me what this means?
>>>>>> thanks
>>>>>> Oliver
>>>>>>             
>>>>>> On 3 Aug., 10:23, obo <bo... at matcalc.de> wrote:
>>>>>>             
>>>>>>> Dear cp2k Users,
>>>>>>> I've made some calculations with alpha quarz. At first I have
>>>>>>> optimized the hexagonal cell with 9 atoms. I've got the initial
>>>>>>> coordinates from a cell optimization in another ab initio programm.
>>>>>>> The relaxation in cp2k was sucessfull and the mulliken charges from
>>>>>>> the Si atoms are approximately the same. Now I have constructed an
>>>>>>> orthorhombic cell with 19 atoms. I've constructed the basis from the
>>>>>>> coordinates of the relaxed hexagonal cell. The length of the cell
>>>>>>> vectors are also calculated from the hexagonal cell vectors(a, 3^0.5,
>>>>>>> c).
>>>>>>> When I start the cell optimization, I get the following mulliken
>>>>>>> charges:
>>>>>>>   Atom     |    Charge
>>>>>>>      1  Si    1.370421
>>>>>>>      2  Si    1.349422
>>>>>>>      3  Si    0.726644
>>>>>>>      4  O    -0.611734
>>>>>>>      5  O    -0.644209
>>>>>>>      6  O    -0.595641
>>>>>>>      7  O    -0.609976
>>>>>>>      8  O    -0.586622
>>>>>>>      9  O    -0.647799
>>>>>>>     10  Si    1.473611
>>>>>>>     11  Si    1.489633
>>>>>>>     12  Si    0.829598
>>>>>>>     13  O    -0.609031
>>>>>>>     14  O    -0.568553
>>>>>>>     15  O    -0.591643
>>>>>>>     16  O    -0.610546
>>>>>>>     17  O    -0.600818
>>>>>>>     18  O    -0.562758
>>>>>>>   Total       0.000000
>>>>>>>               
>>>>>>> At the end of the cell optimization, the relation between the cell
>>>>>>> vectors (a and 3^0.5*a) are also wrong.
>>>>>>> What is my mistake??? Can somebody help me?
>>>>>>>               
>>>>>>> The input file:
>>>>>>>               
>>>>>>> &FORCE_EVAL
>>>>>>>   METHOD Quickstep
>>>>>>>   STRESS_TENSOR NUMERICAL
>>>>>>>   &DFT
>>>>>>>      &POISSON
>>>>>>>        PERIODIC XYZ
>>>>>>>        POISSON_SOLVER PERIODIC
>>>>>>>        &EWALD
>>>>>>>         EWALD_TYPE SPME
>>>>>>>        &END
>>>>>>>      &END POISSON
>>>>>>>     &MGRID
>>>>>>>       CUTOFF 650
>>>>>>>       RELATIVE_CUTOFF 85.0
>>>>>>>     &END MGRID
>>>>>>>     &QS
>>>>>>>      WF_INTERPOLATION PS
>>>>>>>      EXTRAPOLATION_ORDER 4
>>>>>>>     &END QS
>>>>>>>     &SCF
>>>>>>>       EPS_SCF 1.E-6
>>>>>>>       CHOLESKY OFF
>>>>>>>       MAX_SCF 50
>>>>>>>       SCF_GUESS ATOMIC
>>>>>>>       &OUTER_SCF
>>>>>>>         DIIS_BUFFER_LENGTH 5
>>>>>>>         OPTIMIZER DIIS
>>>>>>>       &END OUTER_SCF
>>>>>>>     &END SCF
>>>>>>>     &XC
>>>>>>>        &XC_GRID
>>>>>>>         XC_DERIV SPLINE3
>>>>>>>        &END XC_GRID
>>>>>>>        &XC_FUNCTIONAL PBE
>>>>>>>        &END XC_FUNCTIONAL
>>>>>>>     &END XC
>>>>>>>   &END DFT
>>>>>>>   &SUBSYS
>>>>>>>     &CELL
>>>>>>>       PERIODIC XYZ
>>>>>>>       MULTIPLE_UNIT_CELL 1 1 1
>>>>>>>       ABC 4.805 8.3225 5.393
>>>>>>>       ALPHA_BETA_GAMMA  90.0 90.0 90.0
>>>>>>>     &END CELL
>>>>>>>     &COORD
>>>>>>> Si 1.1328078594000004 6.4061604696       3.5954303785000006
>>>>>>> Si 1.1329589190000002 1.9039591092000003 1.7978168006000006
>>>>>>> Si 2.6289910789000004 8.3163249591       5.3929132817
>>>>>>> O 1.6426232969999999  7.8451213509999995 4.154876766399999
>>>>>>> O 4.4371215705        1.6255884538000005 2.3576057410000004
>>>>>>> O 3.6202186447        7.155568958499999  0.5597043610000001
>>>>>>> O 1.6427913173999997  0.4650789015000002 1.237976828999999
>>>>>>> O 3.620111714099999   1.1547964305000002 4.8334310334
>>>>>>> O 4.4368490245        6.684213627799998  3.035570217399999
>>>>>>> Si 3.4358078594       2.2346604696       3.5954303785000006
>>>>>>> Si 3.435958918999999  6.054959109200001  1.7978168006000006
>>>>>>> Si 0.1269910789       4.144824959099999  5.3929132817
>>>>>>> O 3.9456232970000005  3.6736213510000013 4.154876766399999
>>>>>>> O 1.9351215704999998  5.776588453800002  2.3576057410000004
>>>>>>> O 1.1182186447000002  2.984068958500001  0.5597043610000001
>>>>>>> O 3.9457913174000008  4.616078901500002  1.237976828999999
>>>>>>> O 1.1181117141        5.305796430500002  4.8334310334
>>>>>>> O 1.9348490245        2.5127136278000006 3.035570217399999
>>>>>>>      &END COORD
>>>>>>>     &KIND Si
>>>>>>>       BASIS_SET DZVP-GTH-PBE
>>>>>>>       POTENTIAL GTH-PBE-q4
>>>>>>>     &END KIND
>>>>>>>     &KIND O
>>>>>>>       BASIS_SET DZVP-GTH-PBE
>>>>>>>       POTENTIAL GTH-PBE-q6
>>>>>>>     &END KIND
>>>>>>>     &KIND H
>>>>>>>       BASIS_SET DZVP-GTH-PBE
>>>>>>>       POTENTIAL GTH-PBE-q1
>>>>>>>     &END KIND
>>>>>>>     &TOPOLOGY
>>>>>>>      CENTER_COORDINATES T
>>>>>>>      &DUMP_PDB
>>>>>>>      &END
>>>>>>>        MULTIPLE_UNIT_CELL 1 1 1
>>>>>>>     &END TOPOLOGY
>>>>>>>   &END SUBSYS
>>>>>>> &END FORCE_EVAL
>>>>>>> &GLOBAL
>>>>>>>   PROJECT SiO-Cell
>>>>>>>   RUN_TYPE CELL_OPT
>>>>>>>   PRINT_LEVEL MEDIUM
>>>>>>> &END GLOBAL
>>>>>>> &MOTION
>>>>>>>   &GEO_OPT
>>>>>>>   MAX_ITER 200
>>>>>>>   OPTIMIZER BFGS
>>>>>>>   MAX_FORCE 5.0E-5
>>>>>>>   &END GEO_OPT
>>>>>>>   &CELL_OPT
>>>>>>>    EXTERNAL_PRESSURE [bar] 0.0
>>>>>>>    PRESSURE_TOLERANCE 15.0
>>>>>>>    MAX_ITER 900
>>>>>>>    KEEP_ANGLES TRUE
>>>>>>>    OPTIMIZER CG
>>>>>>>    &CG
>>>>>>>     &LINE_SEARCH
>>>>>>>       TYPE 2PNT
>>>>>>>     &END LINE_SEARCH
>>>>>>>    &END CG
>>>>>>>   &END CELL_OPT
>>>>>>> &END MOTION
>>>>>>>               
>>>>>>> thanks
>>>>>>> Oliver
>>>>>>>               
> >
>