Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops...

28
Kinetic Monte Carlo Triangular lattice

Transcript of Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops...

Page 1: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Kinetic Monte Carlo

Triangular lattice

Page 2: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Diffusion

D =Θ⋅DJ

DJ =12d( )t

1N

Δr R i t( )

i=1

N

2

Θ =

∂µkBT

∂ ln x=

N

N 2 − N 2Thermodynamic

factor

Self DiffusionCoefficient

Page 3: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Δr R i t( )

Δr R j t( )

DJ =12d( )t

1N

Δr R i t( )

i=1

N

2

Diffusion

D* =12d( )t

1N

Δr R i t( )2

i=1

N

Page 4: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Standard Monte Carlo to studydiffusion

• Pick an atom at random

Page 5: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Standard Monte Carlo to studydiffusion

• Pick an atom at random• Pick a hop direction

Page 6: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Standard Monte Carlo to studydiffusion

• Pick an atom at random• Pick a hop direction• Calculate

exp −ΔEb kBT( )

Page 7: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Standard Monte Carlo to studydiffusion

• Pick an atom at random• Pick a hop direction• Calculate• If ( >

random number) do the hop€

exp −ΔEb kBT( )

exp −ΔEb kBT( )

Page 8: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Kinetic Monte Carlo

Consider all hops simultaneously

A. B. Bortz, M. H. Kalos, J. L. Lebowitz, J. Comput Phys, 17, 10 (1975).F. M. Bulnes, V. D. Pereyra, J. L. Riccardo, Phys. Rev. E, 58, 86 (1998).

Page 9: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such
Page 10: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such
Page 11: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such
Page 12: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such
Page 13: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such
Page 14: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such
Page 15: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such
Page 16: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such
Page 17: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Wi = ν * exp −ΔEikBT

For each potential hop i, calculate the hop rate

Page 18: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Wi = ν * exp −ΔEikBT

For each potential hop i, calculate the hop rate

Then randomly choose a hop k, with probability

Wk

Page 19: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Wi = ν * exp −ΔEikBT

For each potential hop i, calculate the hop rate

Then randomly choose a hop k, with probability

Wk= random number

ξ1

Page 20: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Wi = ν * exp −ΔEikBT

For each potential hop i, calculate the hop rate

Then randomly choose a hop k, with probability

Wk= random number

ξ1

Wii=1

k−1

∑ <ξ1 ⋅W ≤ Wii= 0

k

W = Wii= 0

Nhops

Page 21: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Then randomly choose a hop k, with probability

Wk= random number

ξ1

Wii=1

k−1

∑ <ξ1 ⋅W ≤ Wii= 0

k

W = Wii= 0

Nhops

Page 22: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Time

After hop k we need to update the time

= random number

ξ 2

Δt = −1Wlogξ 2

Page 23: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Two independent stochasticvariables:

the hop k and the waiting time

Wii=1

k−1

∑ <ξ1 ⋅W ≤ Wii= 0

k

Δt = −1Wlogξ 2 €

W = Wii= 0

Nhops

∑€

Wi = ν * exp −ΔEikBT

Δt

Page 24: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Kinetic Monte Carlo• Hop every time• Consider all possible hops simultaneously• Pick hop according its relative probability• Update the time such that on average

equals the time that we would have waitedin standard Monte Carlo

Δt

A. B. Bortz, M. H. Kalos, J. L. Lebowitz, J. Comput Phys, 17, 10 (1975).F. M. Bulnes, V. D. Pereyra, J. L. Riccardo, Phys. Rev. E, 58, 86 (1998).

Page 25: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Triangular 2-d lattice, 2NNpair interactions

E r σ ( ) = Vo +Vpo intσ l +

12VNNpairσ l σ i

i=NNpairs∑ +

12VNNNpairσ l σ j

j=NNNpairs∑

l=1

Nlattice−sites

Page 26: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Activation barrier

ΔEkra = Eactivated−state −12E1 +E2( )

ΔEbarrier = ΔEkra +12E final −Einitial( )

Page 27: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Thermodynamics

Θ =

∂µkBT

∂ ln x=

N

N 2 − N 2

D =Θ⋅DJ

Page 28: Kinetic Monte Carlo - IllinoisKinetic Monte Carlo •Hop every time •Consider all possible hops simultaneously •Pick hop according its relative probability •Update the time such

Kinetics

D =Θ⋅DJ

DJ =12d( )t

1N

Δr R i t( )

i=1

N

2