I am trying to simulate Ni-Ti alloy. There is no available potential I can use directly. I tryed two ways:
there is eam_database in tools folder. I tried to compile a Ni-Ti eam potential. At first, I compiled a CuAl eam potential and compare with my results that computed with CuAl.eam.alloy which already exist
in potential folder. But the results are quite different. So, I wonder if there is something more that I should pay attention before I compile create.f in eam_database?
I tryed MEAM potential, and I already got some parameters from some papers. But I still can not get alpha for Ni-Ni, Ti-Ti, and Ni-Ti. it said it could be computed form Bulk modulus, but I don’t know how.
I saw a function in a paper that alpha=(9B*omiga/Ec)-2 , and omiga is the equilibrium atomic volume. Does it mean that we can calculate based on radius of the atom (Ni or Ti), but how to calculate for
alpha Ni-Ti ? also, I am still confuse about the meaning rho0 (I) that listed. How can I get this for Ni-Ti ? Should I care about all the parameters that listed on the manual?
These are not really LAMMPS questions - they are Qs about
creating and using potentials, for alloys. That is an art
unto itself and there is no guarantee 2 different alloy potentials
should give the same answers. It depends on how they
were derived and what they were fit to. Ditto for MEAM.
You need to read papers or talk to people who derive potentials.
2. I tryed MEAM potential, and I already got some parameters from some
papers. But I still can not get alpha for Ni-Ni, Ti-Ti, and Ni-Ti. it said
it could be computed form Bulk modulus, but I don't know how.
I saw a function in a paper that alpha=(9B*omiga/Ec)-2 , and omiga is
the equilibrium atomic volume. Does it mean that we can calculate based on
radius of the atom (Ni or Ti),
You are correct, for single elements, alpha=sqrt(9B*omega/Ec).
but how to calculate for alpha Ni-Ti ?
Simplest way to get alpha for Ni-Ti is to obtain omega, B, and
Ec for Ni-Ti in the NaCl (B1) structure from ab-initio calculations,
and to specify lattce(1,2)='b1' as a reference structure for
Ni-Ti in the potential file. Then, the MEAM potential will exactly
follow Rose Equation of state for Ni-Ti in the NaCl with the
given alpha. Other reference structures are supported too.
also, I am still confuse about the meaning rho0 (I) that listed. How can I get this for Ni-Ti ? Should I care about all the parameters that listed on the manual?
Thanks very much. But what is default values of rho0(I) for Ni-Ti? I just saw single elements is 1.0. Is there default values for alloy? or you mean I can start with values for single elements?
you can start with values for single elements, and then modify them to
adjust alloy properties.
Rho0 parameter only comes into play for cross-element interactions.
E.g., if you have only "Ni" in your simulation, value of "rho0(Ni)"
does not affect results at all. For simulation with "Ni" only,
"rho0(Ni)" can be anything.
If you have both Ni and Ti in your simulation, then rho0(Ni) and
rho0(Ti) are significant, but only their ratio. I.e. *for simulations
with only Ni and Si* the values of "rho0(Ni)=4, rho0(Ti)=6" will give
the same results as "rho0(Ni)=2, rho0(Ti)=3".
Thanks very much. I am trying to modify them. It really helps a lot! Now I still have couple of questions:
1. what is omega for alloy? If I got re(1,2) for Ni-Ti, can I use re(1,2) as diameter of the alloy atom, and calculate omega according to sphere volume equation?
2. I am trying to simulate B19' structure for Ni-Ti alloy. But there is no reference structure in lammps. Can I just use lattce(1,2)='b2'?
2.) You can choose to use b2 as a reference structure for the Ni-Ti
interaction, i.e. you can set lattce(1,2)='b2'. You can do that since
b2 is one of the alloy reference structures available in the LAMMPS
MEAM implementation:
1.) If you choose *b2* as the reference structure for Ni-Ti potential,
then Omega, B, and Ec should be per-atom volume, bulk modulus, and
the cohesive energy of the Ni-Ti in the *b2* structure. Then you
obtain alpha=sqrt(9B*omega/Ec), and, consequently, MEAM method will,
by definition, exactly reproduce (postulate) Rose equation of state
for the Ni-Ti in the *b2* structure. Some of the parameters, e.g.
screening constants, can than be adjusted to better reproduce heat of
formation of other structures, e.g. B19 compound. We followed that
approach here:
except that we used "b1" as the reference structure uniformly for all
unlike element pairs. In our case, since we used "b1" as the reference
structure, Omega, B, and Ec were per-atom volume, bulk modulus, and
the cohesive energy of the unlike element pairs in the "b1" structure.