Relaxation

[Mohammed]

Dear all,

Attached is the image of an already relaxed molecule on top of a Si-SiO2 substrate. I ran the optimization with a k-point sampling of (1,1,1). According to my understanding, because my structure size is fairly small, the reciprocal lattice should be even smaller, hence my choice of (1,1,1) k-points. My question is was this a right choice or not ?

[Anders Blom]

No it's opposite of course - the smaller the cell, the more k-points you need.

[Mohammed]

I am sorry then I mixed it up. So for this structure which k-point sampling would you propose, because I previously tried it with (4,4,4) and it takes an extremely large amount of simulation time.

[Anders Blom]

Why 4 in the C direction?

It's impossible to guess, you should ideally check by increasing. But 4 is at least reasonable (in A and B).

[Mohammed]

So it can be (4,4,1) because in the c-direction its large ? Just to point out, this is just a test molecule which will be eventually replaced by a longer chain of benzene rings attached to the phosphorus. In that case would (4,4,1) be sufficient or should I increase the k-points in the c-direction.

I am asking this because eventually after relaxation, transport calculations will be done in the c-direction, and from what I read in other posts is that the no. of k-points in the transport direction should be high. And that is what is confusing me regarding the choice of k-point sampling. 

[Anders Blom]

No, I mean you should learn why you need k-points in the first place, then it should be obvious that you only need 1 in C. It is quite important to understand what the parameters you use for a calculation mean (and not just get magic numbers from someone else), otherwise you are not in control of your own science.

For the system you have transport cannot be in the C direction but you can of course always turn the system around.

[Mohammed]

Thank you very much for your help, greatly appreciated