Top Page > Browsing
negf-vcrelax
Date: 2017/01/24 15:16
Name: hamed   <dihi6213@gmail.com>

dear all

I want to calculate I-V for my device. The channel is different from Leads(C0 ≠ L0).

1-Should i add or merge C0 with L0 and make a bigger C0?(for example (new C0 = (L0L0C0R0R0))

2-How should i relax or vc-relax this device?

thanks
メンテ
Page: [1]

Re: negf-vcrelax ( No.1 )
Date: 2017/01/24 19:10
Name: Artem Pulkin

1. Yes, it is generally a good idea.
2. You should embed your "L0L0C0R0R0" into periodic boundary conditions and relax it as usual. I.e. relax something like "R0C1L0L0L0C0R0R0". Without NEGF, of course.
メンテ
Re: negf-vcrelax ( No.2 )
Date: 2017/01/25 01:06
Name: hamed  <dihi6213@gmail.com>

Dear sir


At first thanks for your answer

|-Do you think that i must follow the below steps for optimization or not?please explain

1-VC-relax of Leads(R0,L0)
2-VC-relax of center(L0L0C0R0R0)
3-VC-relax of device (Lead.center.Lead)(L0.L0L0C0R0R0.R0)

||- Sorry but i did not understand " relax something like "R0C1L0L0L0C0R0R0". Without NEGF, of course."

your assistance in this matter is greatly appreciated
メンテ
Re: negf-vcrelax ( No.3 )
Date: 2017/01/26 01:17
Name: Artem Pulkin

Regarding 1): yes.

Regarding everything else:

The point is that the NEGF calculation is too slow and too computationally demanding for relaxing your structures. To relax, you typically use supercell models with your defect. It is fairly simple if L0 = R0, i.e. you have the same lead material at both sides. In this case you, literally, change NEGF to BAND in the input file and do the relaxation.

Now, if your L0 is not the same as R0, you have to relax two defect at once. The first one is

L0L0C0R0R0

And the second one is

R0R0C1L0L0

You join them into a supercell like this

[L0L0C0R0R0][R0R0C1L0L0]

and relax it. Why do you have to do it? Because you have periodic boundary conditions in the supercell. I.e. your leftmost and rightmost units (L0 in the above case) have to be same. It is not the case if you relax your transport model [L0L0C0R0R0].

Artem
メンテ
Re: negf-vcrelax ( No.4 )
Date: 2017/01/27 00:32
Name: hamed  <dihi6213@gmail.com>



thanks for your help and guidance
メンテ
Re: negf-vcrelax ( No.5 )
Date: 2017/01/27 01:09
Name: kim



What is the difference between C0 and C1?
メンテ
Re: negf-vcrelax ( No.6 )
Date: 2017/01/27 20:55
Name: Artem Pulkin

It is a very relevant question. For example, if both L0 and R0 are inversion-symmetric, you may use a mirror-symmetric C0 as a C1. However, in a more general case you invent one more defect between the same materials (but different terminations). It should be as chemically neutral as it is possible.
メンテ

Page: [1]

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