[salmon-user:00270] dft calculation on a nonorthogonal grid

[河合 涼太郎]

Dear Salmon developers and users,

 

My name is Ryotaro Kawai ,and I’m a graduate student in the School of Engineering in the Onishi Lab. at Tohoku university.

I am using SALMON for my research. I am currently doing dft calculations on a non-orthogonal grid, but I don't seem to be getting the energies, etc. correct. I can't seem to figure out why, so can you help me out? I have the code below.

 

 

 

&calculation

  theory = 'dft'

/

 

&control

 

  sysname = '1layer'

/

 

&parallel

  nproc_k = 2

  nproc_ob = 1

  nproc_rgrid = 1,1,1

/

 

&units

 

  unit_system = 'A_eV_fs'

/

 

&system

 

  yn_periodic = 'y'

 

   al_vec1 = 2.1313d0, 1.2305d0, 0.0d0

   al_vec2 = 2.1313d0, -1.2305d0, 0.0d0

   al_vec3 = 0.0d0, 0.0d0, 3.35d0

 

  nelem  = 1

  natom  = 2

  nelec  = 8

  nstate = 8

/

 

&pseudo

 

  file_pseudo(1) = './C_rps.dat'

 

  izatom(1) = 6

 

  lloc_ps(1) = 1

/

 

&functional

 

  xc = 'PZ'

/

 

&rgrid

  num_rgrid =12,12,12

/

 

&kgrid

 

  num_kgrid = 4, 4, 4

/

 

&scf

 

  nscf      = 300

  threshold = 1.0d-9

/

 

&atomic_red_coor

 

  'C'        .0           .0           .0           1

  'C'        .333333 .333333 .0           1

 

/

 

Sincerely,

 

Kawai Ryotaro

Graduate student, Tohoku university

[山田俊介]

 Dear Kawai-san,

Thank you for sharing your input file. Is this AA-stacking graphite?

I tried the following primitive lattice vectors instead of your input and it works well:

  al_vec1 = 2.4610,  0.0000000000,  0.0000000

  al_vec2 = 1.2305,  2.1312885187,  0.0000000

Here, the primitive lattice vectors are defined as follows:

a1 = (a, 0, 0),

a2 = (a/2, a*sqrt(3)/2, 0),

where a=2.461Å is the lattice constant.

In this definition, the x-axis is zigzag direction.

In your input, on the other hand, the lattice vectors were defined as

a1 = (a*sqrt(3)/2, a/2, 0),

a2 = (a*sqrt(3)/2, -a/2, 0).

Here, the y-axis is zigzag direction.

This definition is also valid but SALMON fails by that. 

We didn't recognize this bug and will fix it. Thank you very much for reporting this issue!

Sincerely,

Shunsuke

2023/10/30 19:45、河合 涼太郎 <rydber...@outlook.jp>のメール:

Dear Salmon developers and users,

 

My name is Ryotaro Kawai ,and I’m a graduate student in the School of Engineering in the Onishi Lab. at Tohoku university.

I am using SALMON for my research. I am currently doing dft calculations on a non-orthogonal grid, but I don't seem to be getting the energies, etc. correct. I can't seem to figure out why, so can you help me out? I have the code below.

 

<0DA93E9A040F415C8CB226C2FDACB57F.png>

[山田俊介]

Sorry, I just realized the cause of the error.

Pls use the following primitive lattice vectors:

   al_vec1 = 2.1313d0, -1.2305d0, 0.0d0

   al_vec2 = 2.1313d0, 1.2305d0, 0.0d0

In your previous definition, the order of the lattice vectors was swapped (a1,a2 —> a2, a1).

If the order is wrong, the volume of the lattice cell becomes negative. I think this is the cause of the error.

We will fix SALMON as it can warn when the cell velume is negative. Thank you!

Sincerely,

Shunsuke

2023/10/31 17:41、山田俊介 <sya...@ccs.tsukuba.ac.jp>のメール:

[Mizuki Tani]

Dear Yamada-san,

Defining that

a1 = (a*sqrt(3)/2, a/2, 0),

a2 = (a*sqrt(3)/2, -a/2, 0).

works well for mono-layer graphene, namely,

a3 = (0,0, 15Ang.)

Regards,

Mizuki Tani

2023年10月31日(火) 17:42 山田俊介 <sya...@ccs.tsukuba.ac.jp>:

--

===================================== 

谷　水城（Tani Mizuki）

MAIL：forultra...@gmail.com

=====================================

[山田俊介]

Dear Tani-san,

Hmm, thank you for the info. I will check that.

Sincerely,

Shunsuke

2023/10/31 18:11、Mizuki Tani <forultra...@gmail.com>のメール:

[河合 涼太郎]

Dear  Yamada-san and Tani-san,

 

I tried switching the coordinate system between al_vec1 and al_vec2, and the calculation worked. Thank you very much.

 

Sincerely,

Kawai Ryotaro

差出人: 山田俊介
送信日時: 2023年10月31日 18:22
宛先: salmo...@salmon-tddft.jp
件名: [salmon-user:00275] Re: dft calculation on a nonorthogonal grid