Collected from.

https://www.google.com/search?sxsrf=ALeKk032_LpTdkV_moWe0xsG9fy1_YIxPw:1588138457221&source=univ&tbm=isch&q=Muffin-tin+radius+of+solid+atoms&client=firefox-b-d&sa=X&ved=2ahUKEwjj5eLk9IzpAhWVzDgGHUeWCZEQsAR6BAgJEAE#imgrc=G-hmF7ssttRxBM

Hi,

I think you should use the full-potential linear muttn-tin orbital (FPLMTO) method, and the paper below is may be suitable to explain to you:

"Assessing Relativistic Effects and Electron Correlation in the Actinide Metals Th to Pu"

Can you give me the value of the muffin tin radius of Ga atom?

ELK code works based on linearized augmented-plane wave (LAPW) method. in ELK database, the muffin- tin radius for Ga is 2.40.

Thanks for the answer. But I've got the result.

The muffin-tin radius of Ga, r1(Ga)=2.45 a.u.

The muffin-tin radius of Ga, r1(Ga)=2.45 a.u. and the muffin-tin radius of As, r1(As) =2.17 a.u.

For the FLAPW method, the Rmt is in principal not so crucial, but you should do all calculations with the same radius.

You can choose the Rmt as large as possible but consider possible structural changes.

If you have different atoms, don't make their Rmt's too different (about 20%)

The muffin tin radius is an auxiliary concept within the framework of the APW-type of bandstructure models. It is first and for all a mathematical concept, where a sphere surrounding the atoms in the model, separates a region near the atomic nucleus where a spherical symmetry prevails, from the interstitial region where a plane wave description is used. From the physics point of view the muffin radius is arbitrary. You can give it any value, even 0. Overlapping muffin tine spheres are a mathematical nuisance. The price to pay however is numerical efficiency. Of course the practical choice has to be, the value where the numerics performs fastest. This has nothing to do with measurement in the sense of a usual experimental observation. From a lot of trial and error, a set of values can be obtained, where the numerics works fastest. If you call that measurement, let it be, but there is no real physics behind.

Thanks for your information C.M.J. Wijers