[lammps-users] meam.library-how to determine rozero and ibar

Dear LAMMPS developers,

I wanna ask you two questions about MEAM potential library file.

The 1st one: could I add a new line in the potentials/library.meam file for other elements?

The 2nd one: how to determine “rozero” and “ibar”.

I’ve noticed that your previous mail in the website described as follows:

“> --> Rozero is called the (relative) density scaling parameter. It is only important for multi-component systems and is irrelevant for single element systems.

> --> Ibar decides which formula is used on the gamma function, which combines the partially electron densities, before calculating the embedding energy.” 

References: http://lammps.sandia.gov/threads/msg11851.html

The last two parameters have confused me a lot. What’s the default value of rozreo for single element systems, 1.0 or other value? What’s the relationship between rho0(i) in the “lib\meam\meam_dens_final.F” file, Equ. 4.3 rho0 and rho chapter of “Numerical Tools for Atomistic Simulations, ?

“ibar”is a selection flag for computing G and dG in file “lib\meam\meam_dens_final.F”. ibar==0 or 1 or 2 or 3 or 4. Different values mean different G formula. I wanna know how to determine the value of ibar or which G formula I should choose. Are there some differences between multi-component systems and single elements systems while choosing the ibar value from 0, 1, 2, 3 and 4?

I have checked Baskes’s previous articles in the references, but these articles have no details about “rozero” and “ibar”. My intereted materials are Be and C.

I’m looking forward to your reply soon. Thanks very much!

Best regards,

CHIAO Kuaixiuan

GUCAS

, PRChina

PS:

[1] H. C. Huang, N. M. Ghoniem, J. K. Wong, and M. I. Baskes. Molecular-dynamics determination of defect energetics in beta-SiC using 3 representative empirical potentials.

Modelling Simul. Mater. Sci. , 3(5):615-627, 1995.

***Rho0 is the composition-dependent electron density scaling Equ.6

[2]M. I. Baskes. Determination of modified embedded atom method parameters for nickel.Mater. Chem. Phys., 50(2):152-158, 1997.

***where A is an adjustable parameter and p. is a density scaling parameter. For this work (f.c.c. reference structure) po= z= 12. Equ.7

These are questions Greg will need to answer.

Steve

2010/12/1 肇同学 <[email protected]...>:

Dear LAMMPS developers��

I wanna ask you two questions about MEAM potential library file.

The 1st one: could I add a new line in the potentials/library.meam file for other elements��

Yeah, just follow the library.meam format to input the required parameters of the elements you wanna add

The 2nd one: how to determine ��rozero�� and ��ibar��.

rozero needs to be specified for single element systems sometimes, e.g., your potential parameters for this single type of element come from the meam potential that is used for alloy system

ibar refers to specific functional form of G(T), check code “meam_dens_final.F” for details. The functional form of G is specified in the literature of the potential you wanna use
e.g.
ibar
c 0 => G = sqrt(1+Gamma)
c 1 => G = exp(Gamma/2)
c 2 => not implemented
c 3 => G = 2/(1+exp(-Gamma))
c 4 => G = sqrt(1+Gamma)

Yes, you can add new lines to the library.meam file; just follow the pattern in the file. If there are multiple entries in the file that use the same element name, LAMMPS will use the first one it finds, so ordering matters.

The rozero parameter is the value \rho_{i0} on the RHS in equation (4.5) in the reference you’re looking at; it’s the parameter defined as “an element-dependent density scaling.” For a single element system it should usually be 1.0.

Sorry about the lack of documentation of the ibar parameter. This parameter was taken from Mike Baskes original MEAM code, and the definition has remained the same. The options are listed in the comment in the G_gam() subroutine in meam_dens_final.F, but I’ll reproduce those here:

c Compute G(Gamma) and dG(gamma) based on selection flag ibar:

c 0 => G = sqrt(1+Gamma)

c 1 => G = exp(Gamma/2)

c 2 => not implemented

c 3 => G = 2/(1+exp(-Gamma))

c 4 => G = sqrt(1+Gamma)

I’ll try to include these options on the documentation page along with some other changes I hope to release soon.

Regards,

Greg

Greg Wagner

Sandia National Laboratories

P.O. Box 969 MS 9409 Livermore, CA 94551

Tel: (925) 294-2180 Fax: (925) 294-3410

Email: gjwagne@…3…