Question on TAD

LAMMPS LAMMPS Mailing List Mirror
1

Dear LAMMPS users and developers,

I have 3 questions about TAD method, regarding simulation of vacancy migration in nano-crystalline Ni in long-time scales.
I intend to simulate conventional vacancy hoping using TAD method, and I had the following questions:

1- tmax: (reciprocal of lowest expected pre-exponential factor (time units))

Is this parameter the same as V_min (temperature independent pre-exponential factor described in Voters’ paper_Sorensen and Voter, J Chem Phys, 112, 9599 (2000)) ?
and if the answer is yes; in literature of vacancy hopping, it is said that V_min should be between 1e-13 and 1e-12, so its reciprocal would be between 1e12 and 1e13. Can I put this very big value as the tmax in my code?

2- I have written my code in a way that it just looks for atomic displacement more than 2 Angstrom (which can be described in a reverse state as a vacancy diffusion);

How can I dump my system, whenever this event occurs? (I need id, and positions of each atom, when atomic displacement more than 2 A occurs)

3- it is said in LAMMPS manual that N is the number of time steps to run;

Is it the total time steps for the entire simulation?, or is it the time steps for each basin-constrained MD simulation and LAMMPS will perform the same exact number of time steps after each transition occurs?
(e.g., I want to simulate vacancy hopping for 1 ms, and with del(t)=0.001 ps, do I have to continue my simulation for 1e9 consecutive time steps? or I have misunderstood the manual…)

(+) 3- Also, in the TAD section, I have performed NEB calculations;

Can LAMMPS give me the activation barrier of the events that are desired (atomic displacement more than 2 A), simultaneously with TAD simulation?
and if the answer is yes, how can I have the migration barriers of all the desired events? (something like dumping the id, atomic positions and migration barrier of the occurred event at the same time)

I have attached my codes, including TAD counterparts. Thanks in advance for your time.

*** (The LAMMPS version which the restart files are created: 12 Mar 2015-ICMS on win 7 64-bit OS)

in02.cpu.gb (2.16 KB)

2

Dear LAMMPS users and developers,

I have 3 questions about TAD method, regarding simulation of vacancy
migration in nano-crystalline Ni in long-time scales.
I intend to simulate conventional vacancy hoping using TAD method, and I had
the following questions:

1- tmax: (reciprocal of lowest expected pre-exponential factor (time units))

Is this parameter the same as V_min (temperature independent pre-exponential
factor described in Voters' paper_Sorensen and Voter, J Chem Phys, 112, 9599
(2000)) ?
and if the answer is yes; in literature of vacancy hopping, it is said that
V_min should be between 1e-13 and 1e-12, so its reciprocal would be between
1e12 and 1e13. Can I put this very big value as the tmax in my code?

1. tmax is in ***time units***. If you are using LAMMPS metal units,
then time is in units of picoseconds, not seconds, as implied by your
question. 1 picosecond is 1e-12 seconds.

2- I have written my code in a way that it just looks for atomic
displacement more than 2 Angstrom (which can be described in a reverse state
as a vacancy diffusion);

2. Sounds like you did not read the doc page: "Any dump files defined
in the input script will be written to during a TAD run at timesteps
when an event is executed. "

How can I dump my system, whenever this event occurs? (I need id, and
positions of each atom, when atomic displacement more than 2 A occurs)

3- it is said in LAMMPS manual that N is the number of time steps to run;

Is it the total time steps for the entire simulation?, or is it the time
steps for each basin-constrained MD simulation and LAMMPS will perform the
same exact number of time steps after each transition occurs?

Total time steps for the entire simulation.

(e.g., I want to simulate vacancy hopping for 1 ms, and with del(t)=0.001
ps, do I have to continue my simulation for 1e9 consecutive time steps? or I
have misunderstood the manual...)

That depends on whether you mean 1 ms at Thi or Tlo.

(+) 3- Also, in the TAD section, I have performed NEB calculations;

Can LAMMPS give me the activation barrier of the events that are desired
(atomic displacement more than 2 A), simultaneously with TAD simulation?
and if the answer is yes, how can I have the migration barriers of all the
desired events? (something like dumping the id, atomic positions and
migration barrier of the occurred event at the same time)

RT*M ("See the neb command for a full description of the NEB statistics. ")

3

Dear Aidan,

Thanks for your reply,

3- it is said in LAMMPS manual that N is the number of time steps to run;

Is it the total time steps for the entire simulation?, or is it the time
steps for each basin-constrained MD simulation and LAMMPS will perform the
same exact number of time steps after each transition occurs?

Total time steps for the entire simulation.

(e.g., I want to simulate vacancy hopping for 1 ms, and with del(t)=0.001
ps, do I have to continue my simulation for 1e9 consecutive time steps? or I
have misunderstood the manual…)

That depends on whether you mean 1 ms at Thi or Tlo.

I want to continue my simulation for 1 ms in Tlo, then should I have to use 1e9 time steps with timestep=0.001?

Regards
Afshin