QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: qweasel on January 27, 2017, 01:07

Title: How to manipulate MoS2 single layer into a rectangular device structure?
Post by: qweasel on January 27, 2017, 01:07

Building an MoS2 device is not the same as graphene device because:

• The unit cell is not conventional, and converting it to conventional is not possible (the Super lattice -> Conventional button is gray).
• When MoS2 is imported from the database, the C axis is perpendicular to the sheet plane. So I cannot make device from bulk.
• If I rotate it (Coord -> Rotate) 90deg around y, then 60deg around x, the sheet gets messed up (see attach).
• If I enlarge the cell prior to rotation the sheet is not messed up. However, Neuman condition for the Poisson solver is not available: I can only choose between periodic and dirichlet (and A, B direction are not really front-back, bottom-top).

So what is the correct way of making a rectangular MoS2 device?

Title: Re: How to manipulate MoS2 single layer into a rectangular device structure?
Post by: Daniele Stradi on January 27, 2017, 09:27

1. To obtain a rectangular cell, open Bulk Tools -> Supercell the Builder, change the crystalline system to "Base-Centered Orthorombic" and click transform.
2. You can use the "Swap axes" tool in Bulk Tools -> Swap axes
3. I guess this is because the rotated cell is not commensurate with the unit cell

To make an MoS2 device, apply 1. and 2., then repeat the cell along the C direction (Bulk Tools -> Repeat), then use "Device from Bulk".

Title: Re: How to manipulate MoS2 single layer into a rectangular device structure?
Post by: regmi403 on August 12, 2024, 07:44

The same problem is for PdSe2 monolayer. How can we convert PdSe2 device from bulk?

Title: Re: How to manipulate MoS2 single layer into a rectangular device structure?
Post by: Anders Blom on August 12, 2024, 21:03

That depends a lot on which version of PdSe2 you are using.

For instance, the one in https://iopscience.iop.org/article/10.1088/2053-1583/ad0f2a is already in an orthogonal lattice so that would be very easy to manipulate.

A hexagonal case (like the 1T structure) requires a bit more manipulation, but you can always turn it into an orthorhombic supercell using a transformation in the Builder (assuming the C axis is parallel to the vdW bonds) as shown in the attachment.