Question about building interface model

[fangyongxinxi]

Dear Sir,
First, the interface model building part helps me alot to my own work, I think ATK is the most powerful "model builder".  Thanks ATK.

I have two questions:
First,
When we put two surfaces into "first" and "second" surface, click "creat",
the software will give us a default interface model.
question is, why the software gives us this special model?

Second,
When we put two surfaces into "first" and "second" surface, click "select surface cells",
there will be alot of interface models for us to choose,
As my experience, the figure: "Number of Atoms---Mean absolute strain" ,
does not contain all the possible lattice match.
(you can try this with small  "grid size")
So, question is, why this figure does not contain all the possible matches?
or, what is the algorithm of this figure.

Thanks~

 

[Anders Blom]

1. The "default" model is the one with the lowest number of atoms and the smallest strain, but since these two parameters are in opposition to each other in many cases, it's supposed to be some compromise. One had better not just blindly accept the model, but check the strain (and indeed the surface shifts) very carefully.

2. There may be some possible matches that are missed, but they may also be "hidden" since several model may have very similar strain and atom numbers, so the plot points end up on top of each other. You can always make your own model by pulling the red and blue arrows in the surface plots at the top!

The algorithm is called the surface coincidence method and just tries all possible rotations and linear combinations of the two sets of lattice vectors. There is however a certain resolution, which can be changed if you know already that you are looking for a particular angle (or at least an angle in a particular interval). Cf. http://quantumwise.com/publications/tutorials/builder/179-commensurate-supercell-for-rotated-graphene-layers.

[njuxyh]

Dear Professor Blom:

because i want to construct an interface in which one of the layers rotated 30%. so would you please tell me how to build, because I cannot find the  Cf. http://quantumwise.com/publications/tutorials/builder/179-commensurate-supercell-for-rotated-graphene-layers.

1. The "default" model is the one with the lowest number of atoms and the smallest strain, but since these two parameters are in opposition to each other in many cases, it's supposed to be some compromise. One had better not just blindly accept the model, but check the strain (and indeed the surface shifts) very carefully.

2. There may be some possible matches that are missed, but they may also be "hidden" since several model may have very similar strain and atom numbers, so the plot points end up on top of each other. You can always make your own model by pulling the red and blue arrows in the surface plots at the top!

The algorithm is called the surface coincidence method and just tries all possible rotations and linear combinations of the two sets of lattice vectors. There is however a certain resolution, which can be changed if you know already that you are looking for a particular angle (or at least an angle in a particular interval). Cf. http://quantumwise.com/publications/tutorials/builder/179-commensurate-supercell-for-rotated-graphene-layers.

Cf. http://quantumwise.com/publications/tutorials/builder/179-commensurate-supercell-for-rotated-graphene-layers.

[Anders Blom]

Go to https://docs.quantumwise.com and you can browse the tutorials there, and find an updated version of this tutorial.