Top Page > Browsing
Problems in structure optimization of a slab system
Date: 2022/01/20 12:18
Name: Hiroaki Tanaka   <hiroaki-tanaka@issp.u-tokyo.ac.jp>

Dear users,

I want to perform structure optimization calculation of TaAs slab system but have got abnormal results.

The following is the input file (some comments and computer-dependent parameters such as System.CurrentDirectory and data.path are removed).
----
System.Name TaAs

Species.Number 2
<Definition.of.Atomic.Species
Ta Ta7.0-s3p2d2f1 Ta_PBE19
As As7.0-s3p2d2 As_PBE19
Definition.of.Atomic.Species>

Atoms.Number 24
Atoms.SpeciesAndCoordinates.Unit FRAC
<Atoms.SpeciesAndCoordinates
1 Ta 0.22666000 0.55000000 0.55000000 6.5 6.5 45 0 45 0 0 off
2 As 0.21068607 0.05876778 0.05888104 7.5 7.5 45 0 45 0 0 off
3 Ta 0.27701973 0.55622598 0.04040039 6.5 6.5 45 0 45 0 0 off
4 As 0.26067812 0.05893749 0.54014795 7.5 7.5 45 0 45 0 0 off
5 Ta 0.32706621 0.04066971 0.04161937 6.5 6.5 45 0 45 0 0 off
6 As 0.31069547 0.54013741 0.54018591 7.5 7.5 45 0 45 0 0 off
7 Ta 0.37703091 0.04177235 0.55815628 6.5 6.5 45 0 45 0 0 off
8 As 0.36069594 0.54017070 0.05978427 7.5 7.5 45 0 45 0 0 off
9 Ta 0.42666000 0.55000000 0.55000000 6.5 6.5 45 0 45 0 0 off
10 As 0.41068607 0.05876778 0.05888104 7.5 7.5 45 0 45 0 0 off
11 Ta 0.47701973 0.55622598 0.04040039 6.5 6.5 45 0 45 0 0 off
12 As 0.46067812 0.05893749 0.54014795 7.5 7.5 45 0 45 0 0 off
13 Ta 0.52706621 0.04066971 0.04161937 6.5 6.5 45 0 45 0 0 off
14 As 0.51069547 0.54013741 0.54018591 7.5 7.5 45 0 45 0 0 off
15 Ta 0.57703091 0.04177235 0.55815628 6.5 6.5 45 0 45 0 0 off
16 As 0.56069594 0.54017070 0.05978427 7.5 7.5 45 0 45 0 0 off
17 Ta 0.62666000 0.55000000 0.55000000 6.5 6.5 45 0 45 0 0 off
18 As 0.61068607 0.05876778 0.05888104 7.5 7.5 45 0 45 0 0 off
19 Ta 0.67701973 0.55622598 0.04040039 6.5 6.5 45 0 45 0 0 off
20 As 0.66067812 0.05893749 0.54014795 7.5 7.5 45 0 45 0 0 off
21 Ta 0.72706621 0.04066971 0.04161937 6.5 6.5 45 0 45 0 0 off
22 As 0.71069547 0.54013741 0.54018591 7.5 7.5 45 0 45 0 0 off
23 Ta 0.77703091 0.04177235 0.55815628 6.5 6.5 45 0 45 0 0 off
24 As 0.76069594 0.54017070 0.05978427 7.5 7.5 45 0 45 0 0 off
Atoms.SpeciesAndCoordinates>
Atoms.UnitVectors.Unit Ang
<Atoms.UnitVectors
58.280 0.000 0.000
0.000 3.437 0.000
0.000 0.000 3.437
Atoms.UnitVectors>

scf.XcType GGA-PBE
scf.SpinPolarization off
scf.maxIter 1000
scf.EigenvalueSolver band
scf.Kgrid 1 11 11
scf.restart off
scf.Mixing.type simple
scf.Init.Mixing.Weight 0.005
scf.max.mixing.weight 0.100

ESM.switch on1
ESM.buffer.range 4.5

MD.Type Opt # Nomd|Opt|DIIS|NVE|NVT_VS|NVT_NH
<MD.Fixed.XYZ
1 0 0 0
2 0 0 0
3 0 0 0
4 0 0 0
5 0 0 0
6 0 0 0
7 0 0 0
8 0 0 0
9 1 1 1
10 1 1 1
11 1 1 1
12 1 1 1
13 1 1 1
14 1 1 1
15 1 1 1
16 1 1 1
17 0 0 0
18 0 0 0
19 0 0 0
20 0 0 0
21 0 0 0
22 0 0 0
23 0 0 0
24 0 0 0
MD.Fixed.XYZ>
MD.maxIter 2

----

When I execute OpenMX (ver. 3.9.7 in ISSP supercomputer Ohtaka) using the above input, I got the following result.
At MD=1, the SCF was achieved at SCF= 679.
The following is the output on the console.
----
******************* MD= 1 SCF=679 *******************
<Poisson_ESM> Poisson's equation using FFT & ESM...
<Poisson_ESM> Boundary condition = vacuum | cell | vacuum
<Poisson_ESM> Total number of electrons = 0.000205031
<Set_Hamiltonian> Hamiltonian matrix for VNA+dVH+Vxc...
<Band> Solving the eigenvalue problem...
KGrids1: 0.00000
KGrids2: -0.45455 -0.36364 -0.27273 -0.18182 -0.09091 -0.00000 0.09091 0.18182 0.27273 0.36364 0.45454
KGrids3: -0.45454 -0.36363 -0.27273 -0.18182 -0.09091 0.00000 0.09091 0.18182 0.27273 0.36364 0.45455
<Band_DFT> Eigen, time=0.625056
<Band_DFT> DM, time=0.000001
1 Ta MulP 6.2249 6.2249 sum 12.4497
2 As MulP 7.7046 7.7046 sum 15.4091
3 Ta MulP 6.2347 6.2347 sum 12.4694
4 As MulP 7.7527 7.7527 sum 15.5054
5 Ta MulP 6.2423 6.2423 sum 12.4846
6 As MulP 7.7571 7.7571 sum 15.5142
7 Ta MulP 6.2415 6.2415 sum 12.4830
8 As MulP 7.7575 7.7575 sum 15.5150
9 Ta MulP 6.2430 6.2430 sum 12.4861
10 As MulP 7.7585 7.7585 sum 15.5170
11 Ta MulP 6.2444 6.2444 sum 12.4888
12 As MulP 7.7548 7.7548 sum 15.5096
13 Ta MulP 6.2429 6.2429 sum 12.4858
14 As MulP 7.7568 7.7568 sum 15.5136
15 Ta MulP 6.2417 6.2417 sum 12.4835
16 As MulP 7.7575 7.7575 sum 15.5150
17 Ta MulP 6.2434 6.2434 sum 12.4868
18 As MulP 7.7584 7.7584 sum 15.5168
19 Ta MulP 6.2443 6.2443 sum 12.4885
20 As MulP 7.7548 7.7548 sum 15.5096
..........
......

Sum of MulP: up = 168.00000 down = 168.00000
total= 336.00000 ideal(neutral)= 336.00000
<DFT> Total Spin Moment (muB) = 0.000000000000
<DFT> Mixing_weight= 0.001000000000 SCF_RENZOKU=-1
<DFT> Uele = -382.627808407585 dUele = 0.000000876885
<DFT> NormRD = 0.000095062581 Criterion = 0.000001000000
----

However, at MD=2 the result is obviously abnormal.
MulP becomes zero, and the sum of MulP does not correspond to the actual number of electrons (336).
----
******************* MD= 2 SCF= 3 *******************
<Poisson_ESM> Poisson's equation using FFT & ESM...
<Poisson_ESM> Boundary condition = vacuum | cell | vacuum
<Poisson_ESM> Total number of electrons = -322.891092535
<Set_Hamiltonian> Hamiltonian matrix for VNA+dVH+Vxc...
<Band> Solving the eigenvalue problem...
KGrids1: 0.00000
KGrids2: -0.45455 -0.36364 -0.27273 -0.18182 -0.09091 -0.00000 0.09091 0.18182 0.27273 0.36364 0.45454
KGrids3: -0.45454 -0.36363 -0.27273 -0.18182 -0.09091 0.00000 0.09091 0.18182 0.27273 0.36364 0.45455
<Band_DFT> Eigen, time=1.193357
<Band_DFT> DM, time=0.000000
1 Ta MulP -0.0000 -0.0000 sum -0.0000
2 As MulP -0.0000 -0.0000 sum -0.0000
3 Ta MulP -0.0000 -0.0000 sum -0.0000
4 As MulP -0.0000 -0.0000 sum -0.0000
5 Ta MulP -0.0000 -0.0000 sum -0.0000
6 As MulP -0.0000 -0.0000 sum -0.0000
7 Ta MulP -0.0000 -0.0000 sum -0.0000
8 As MulP -0.0000 -0.0000 sum -0.0000
9 Ta MulP 0.0000 0.0000 sum 0.0000
10 As MulP 0.0000 0.0000 sum 0.0000
11 Ta MulP 0.0000 0.0000 sum 0.0000
12 As MulP 0.0000 0.0000 sum 0.0000
13 Ta MulP 0.0000 0.0000 sum 0.0000
14 As MulP 0.0000 0.0000 sum 0.0000
15 Ta MulP 0.0000 0.0000 sum 0.0000
16 As MulP 0.0000 0.0000 sum 0.0000
17 Ta MulP -0.0000 -0.0000 sum -0.0000
18 As MulP -0.0000 -0.0000 sum -0.0000
19 Ta MulP -0.0000 -0.0000 sum -0.0000
20 As MulP -0.0000 -0.0000 sum -0.0000
..........
......

Sum of MulP: up = 0.00000 down = 0.00000
total= 0.00000 ideal(neutral)= 336.00000
<DFT> Total Spin Moment (muB) = 0.000000000000
<DFT> Mixing_weight= 0.100000000000 SCF_RENZOKU= 0
<DFT> Uele = 0.000000000000 dUele = 0.000000000000
<DFT> NormRD = 0.465499075158 Criterion = 0.000001000000
----

When I check TaAs.dat# file, the positions of the atoms are very close to those of the original input (the difference is about 0.05 Ang).

Is there any mistake in my input file, or do you have any suggestions to perform such calculations?

Best regards,
Hiroaki Tanaka
メンテ
Page: [1]

Re: Problems in structure optimization of a slab system ( No.1 )
Date: 2022/01/20 21:01
Name: Naoya Yamaguchi

Hi,

Although I've not reproduced your issue, I tried an example of `Al-Si111_ESM.dat` after I changed `scf.Mixing.type` from `rmm-diisk` to `simple`, and a similar issue occurs.
******************* MD= 1 SCF=43 *******************
<Poisson_ESM> Poisson's equation using FFT & ESM...
<Poisson_ESM> Boundary condition = vacuum | cell | metal
<Poisson_ESM> Total number of electrons = 0.074122538
<Set_Hamiltonian> Hamiltonian matrix for VNA+dVH+Vxc...
<Band> Solving the eigenvalue problem...
KGrids1: 0.00000
KGrids2: -0.33333 -0.00000 0.33333
KGrids3: -0.33333 0.00000 0.33334
<Band_DFT> Eigen, time=0.040691
<Band_DFT> DM, time=0.000000
1 Al MulP 9.9237 9.9237 sum 19.8475
2 Si MulP 5.7850 5.7850 sum 11.5699
3 Si MulP 3.3962 3.3962 sum 6.7924
4 Si MulP -0.0899 -0.0899 sum -0.1799
5 Si MulP -0.0172 -0.0172 sum -0.0344
6 Si MulP -0.0003 -0.0003 sum -0.0006
7 Si MulP 0.0000 0.0000 sum 0.0000
8 Si MulP -0.0000 -0.0000 sum -0.0000
9 Si MulP 0.0000 0.0000 sum 0.0000
10 Al MulP -0.0000 -0.0000 sum -0.0000
Sum of MulP: up = 18.99750 down = 18.99750
total= 37.99500 ideal(neutral)= 38.00000
<DFT> Total Spin Moment (muB) = 0.000000000000
<DFT> Mixing_weight= 0.027597253123 SCF_RENZOKU=-1
<DFT> Uele = 17.783295855037 dUele = 51.306705295423
<DFT> NormRD = 9.445352129352 Criterion = 0.000001000000

I guess that some kind of mixing type is not compatible with the ESM, and a workaround is to use `rmm-diisk` as with the example.
When I calculate your input with the following parameters, the issue didn't appear.
```
scf.Mixing.type rmm-diisk # or rmm-diisv
scf.Mixing.StartPulay 15
scf.Mixing.History 30
```

Regards,
Naoya Yamaguchi
メンテ
Re: Problems in structure optimization of a slab system ( No.2 )
Date: 2022/01/21 10:16
Name: Hiroaki Tanaka

Dear Naoya,

Thank you for your suggestion!
I will try "rmm-diisk" option.

Hiroaki Tanaka
メンテ
Re: Problems in structure optimization of a slab system ( No.3 )
Date: 2022/01/21 12:37
Name: Hiroaki Tanaka

I found that "rmm-diisk" option works well.
I could complete my structure optimization calculations.

Thank you for your help!

Best regards,
Hiroaki Tanaka
メンテ

Page: [1]

Thread Title (must) Move the thread to the top
Your Name (must)
E-Mail (must)
URL
Password (used in modification of the submitted text)
Comment (must)

   Save Cookie