QuantumATK Forum

QuantumATK => Scripts, Tutorials and Applications => Topic started by: pshinyeong on July 6, 2023, 09:13

Title: Question about Boron diffusion in bulk silicon tutorial
Post by: pshinyeong on July 6, 2023, 09:13

Hello,

I was looking at tutorial titled "Boron diffusion in bulk silicon" (link = "https://docs.quantumatk.com/tutorials/boron_diffusion_in_si/boron_diffusion_in_si.html" ) to perform Adaptive Kinetic Monte Carlo (AKMC) simulation. I realized that silicon was initially geometrically optimized but with the two atoms (which is all the atoms there is in the primitive cell) fixed. Does performing a geometry optimization calculation with all the atoms fixed have any meaningful purpose? I didn't understand why the Silicon atoms were fixed here.

Thanks.

+++ This was a bad question, lattice vectors were unfixed. Problem solved.

Title: Re: Question about Boron diffusion in bulk silicon tutorial
Post by: Anders Blom on August 7, 2023, 19:54

This old example is quite simplistic and was written before we had more fancy constraint. The purpose of the setup as written was indeed to optimize the cell size (lattice vectors) without disrupting the space group symmetry. Today you can do that with a space group constrain, but actually a better way to do that for a cubic material is to run a quick equation of state curve and find the lattice constant giving the minimum energy.