Middle 

East Technical University – Department of Metallurgical & Materials Engineering

Simulation of Finite Systems
Prepared and Presented by :

Asst.Prof.Dr. Caner Şimşir

2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 1/35

 Simulation of Finite Systems
 In almost all methods used to model
materials, the system will be described by a
set of discrete objects of some sort. Those
objects might be atoms and the goal may be
to calculate the cohesive energy by
summing interatomic interaction potentials.
 Learning how to calculate these sums is
thus fundamental to essentially all materials
modeling and simulation.
 We are trying to model a macroscopic
system that contains large numbers of
objects. Modeling the behavior of all those
atoms would be computationally impossible.
 To approximate the (effectively) infinite
systems, we use various boundary
conditions, mostly based on introducing a
repeating lattice.
 How one sums the interactions between the
objects within the framework of these
boundary conditions is the focus of this
chapter.

2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 2/35

 Sums of Interacting Pairs of Objects
 Consider that we want to calculate the
  
potential energy (U) by summing up rij  rj  ri j
individual interactions.
 φij= φji since interaction are energies. i
 For a system of 4 particles :  
ri rj
U    r12     r13     r13     r14     r23     r24     r34 

  1/2
rij   rij  rij 
 Now, for a system of N objects,
1 N N
U   `  rij  O

2 i 1 j 1
 Since there are N(N-1)/2 interactions :  2
r12 
N 1 N r23
U      rij  1 
i 1 j i 1 r13
 3
 or even simpler form :
 r24
r14 
U     rij  4
r34
i j i

2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 3/35

 Perfect Crystals

 Consider that we want to calculate the interaction energies (or forces) between a set of
objects in a periodic array.
  
 
n n
1
U cell   ` R  rj  ri
2 R i 1 j 1

 ` denotes i=j term for for the central cell (R=(0,0,0)) is not included in the sum.

1
u  U a  U cell
n
2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 4/35
 Perfect Crystals
 For a perfect 2D square lattice with cell
parameter a. Then, the lattice vectors will
have the general form of:

R  a(n1 xˆ  n2 yˆ )
 where n1 and n2 are integers in between -∞
to +∞. Then the sum takes the form of:
 

  
R n1  n2 

 It is easier to calculate the sum by

considering shells of atoms around the
central atom.

 For the first 4 nearest neighbors (a, √2a, 2a,

√5a), the expansion of Ucell is:

U cell  2  a   2  
2a  2  2a   4  
5a  ...

2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 5/35

 Cutoffs
 Interaction terms decrease with distance, so φ in farther shells
adds a smaller value to the sum. Thus, it would be wise to define a
cutoff distance to reduce computation.
 The impact of using a cutoff distance can be estimated as follows :
by ignoring the lattice structure, the number of atoms on a
spherical shell :
N  V  4 r 2  r
 Then the contribution of this shell to total interaction energy (δU)

 U  4 r   r   r
is: 2

 Total contribution of excluding atoms beyond rc can be estimated

by integrating the equation from rc to infinity :

U  4  r 2  r   dr
rc
 Most of the interatomic potential are in the form of 1/rn so :
 
4 3 n 4 3 n
U  4  r 2n
 dr  r  rc
rc
3 n rc 3 n
 The sum converges for any interatomic potential with n>=4.
 Other methods are necessary for n<=3.

2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 6/35

 Periodic Boundary Conditions
 The goal in most applications is to describe
the properties of a macroscopic system,
which can be considered as infinitely large
considering the size of the system.
 Thus, we need a method to mimic the
essentially infinite system using a system of
manageable size.
 Usage of Periodic Boundary Conditions is a
common approach to do this.
 The objects of interest are put in a finite-sized
volume, referred as simulation cell, which is
then replicated throughout the space.
 Whatever the properties of an object in a cell,
all replicas have of the same object have the
same properties.
 Any crystal lattice can be used as a
simulation cell.
 The correlations inhibit long-wavelength
fluctuations, which can make calculations on
larger scales inaccurate.
 Moreover, simulation cell must commensurate
the system of study, i.e., symmetries of the
structure at the periodic boundaries.

2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 7/35

 Implementation
Checking the cutoff distance (rc) by calculating the
distance is not computationally effective:
 Placing a square box with a size of 2*rc and calculating
distances only for the objects within the box.
 Creating a neighbor list of objects at the beginning,
updating the list when an objects if an object moves a
larger distance than ru-rc.
 Using the minimum image convention.

2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 8/35

 Periodic Boundary Conditions
 The goal in most applications is to describe
the properties of a macroscopic system,
which can be considered as infinitely large
considering the size of the system.
 Thus, we need a method to mimic the
essentially infinite system using a system of
manageable size.
 Usage of Periodic Boundary Conditions is a
common approach to do this.
 The objects of interest are put in a finite-sized
volume, referred as simulation cell, which is
then replicated throughout the space.
 Whatever the properties of an object in a cell,
all replicas have of the same object have the
same properties.
 Any crystal lattice can be used as a
simulation cell.
 The correlations inhibit long-wavelength
fluctuations, which can make calculations on
larger scales inaccurate.
 Moreover, simulation cell must commensurate
the system of study, i.e., symmetries of the
structure at the periodic boundaries.

2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 9/35

 Summary and Suggested Reading
Suggested Reading :

 “Understanding Molecular Simulations” by D.Frenkel and B. Smit

2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 10/35

 Middle East Technical University – Department of Metallurgical & Materials Engineering

Thanks for listening.
Prepared and Presented by :

Asst.Prof.Dr. Caner Şimşir

2019 Fall METE 443 – Computer Modelling & Simulation in Materials Science & Engineering 11/35