You are on page 1of 26

MSE 4107

Molecular Dynamics

Jahirul Islam
Lecturer
Dept. of Materials Science and Engineering
Khulna University of Engineering & Technology
Email: jahirul@mse.kuet.ac.bd
Introduction
❑ What is molecular dynamics (MD)?
The molecular dynamics (MD) is numerical method for
studying many-particle systems such as atoms, molecules,
clusters. One of the most used of all modeling and
simulation techniques in materials research, yielding
information about structure and dynamics on the atomic
scale.

❑ Used extensively in materials science, chemical physics, and


biophysics/biochemistry

❑ First reported MD simulation:


Alder + Wainwright (1957): Phase diagram of a hard-sphere
gas

Department of Materials Science and Engineering, KUET


Classical vs. Quantum Description
❑ Compare inter-particle distance to de-Broglie wavelength

→ Motion of atoms and ions is classical


➢ Quantum effects become significant only when the particle
wavelength, λ, is comparable with the interatomic distance (1–3Å).
Otherwise, the use of the easier Newton’s equations of motion is well
justified.
➢ For most elements and at higher temperatures, therefore, atomic
dynamics can be predicted using Newton’s equations of motion.
Department of Materials Science and Engineering, KUET
Atomic Model in MD
➢ In MD, an atom is basically approximated as a sphere with
point mass at the center as schematically drawn in Fig.
➢ This implies that the electron’s role is totally neglected.
➢ Thus, the computation becomes drastically simple because
there is no need to do anything about electrons.
➢ Compared with the ab-initio, MD is extremely fast and thus can
handle much bigger systems.

Figure 1: Schematic representation


of atomic model in MD methods.

Department of Materials Science and Engineering, KUET


Atomic Model in MD
There is, however, a negative consequence.
➢ Because of the presence of electrons, which is the origin of
interatomic potential, is totally neglected, potential between
atoms should be generated empirically to carry out MD.
➢ Availability of potential is limited especially for multicomponent
systems, and the accuracy of a potential is always under
question.
➢ Length scale is still not macroscopic, and time scale is also
limited to nanoseconds.
➢ No electromagnetic properties can be obtained.

Department of Materials Science and Engineering, KUET


Basics of Molecular Dynamics

➢ Molecular dynamics (MD) is based on the ideas of classical


mechanics as applied to atomic and molecular systems: the
total forces on all the atoms are calculated and Newton’s
equations are solved to determine how the atoms move in
response to those forces.
➢ The equilibrium and time-dependent properties of the system
are then calculated from the motions of the atoms.
➢ Newton’s second law states that the force on a particle
equals its mass times its acceleration,

Department of Materials Science and Engineering, KUET


Basics of Molecular Dynamics
➢ From classical mechanics, we know that the force on a particle
is equal to the negative of the gradient of the potential energy
with respect to the position of that particle. The force on atom i
is

where U is the potential energy and the gradient is taken


with respect to the coordinates on atom i, i.e.,

➢ Therefore, if we know the potential of a system as a function of


interatomic distance, then we can have forces on atoms and thus
can solve the above equation for time evolution of the system.

Department of Materials Science and Engineering, KUET


Potentials
➢ When atoms are near enough to feel each other, the balance
between attraction and repulsion takes place, and it is
determined by so-called interatomic potential.
➢ In an N-atom system where N is the number of atoms, an atom i
interacts with all other atoms at the same time (i.e., there are
interactions of two atoms, three atoms, and so forth at any
moment): Potential energy,

where the ՛ indicates that the i = j terms and the i = j = k terms are not
included in the second and third sets of sums, respectively.

Department of Materials Science and Engineering, KUET


Potentials

➢ This expression is shown schematically for a system of N = 4 atoms


in Figure 5.1:

Figure 5.1: Hierarchy of interactions


between atoms: pair interactions.

Department of Materials Science and Engineering, KUET


Potentials

Figure 5.1: Hierarchy of interactions between atoms: pair interactions.


(a) The first term in Eq. (5.2), v1, represents the effect of an external field
(for example, an electric, magnetic, or gravitational field) on the atoms.
(b) Fig 5.1b represents a sum over the interaction of all pairs of atoms, as
indicated in Fig 5.1b. φij(ri, rj) is a function of the atomic positions that
represents the interaction between the pair of atoms (i, j), located at ri
and rj, respectively.
(c) Fig 5.1c arises from the interaction of triplets of atoms, v3(ri, rj, rk).
These types of interactions are called three-body interactions. Any
interactions that involve more than two atoms at a time are called
many-body interactions.
(d) Fig 5.1d is representative of a four-body interaction, i.e., v4(r1, r2, r3, r4).

Department of Materials Science and Engineering, KUET


Review: Chemical Bonds
Atomic interactions – different types of chemical bonds
❑ Primary bonds (“strong”)
➢ Ionic (ceramics, quartz, feldspar - rocks)
➢ Covalent (silicon)
➢ Metallic (copper, nickel, gold, silver)
(high melting point, 1000-5,000K)
❑ Secondary bonds (“weak”)
➢ Van der Waals (wax, low melting point)
➢ Hydrogen bonds (proteins, spider silk)
(melting point 100-500K)
❑ Ionic: Non-directional (point charges interacting)
❑ Covalent: Directional (bond angles, torsions matter)
❑ Metallic: Non-directional (electron gas concept)

Difference of material properties originates from different


atomic
interaction.
Department of Materials Science and Engineering, KUET
Pair Potentials
➢Pair potentials are the simplest form of a potential that considers
only these two-atom interactions and neglects all others.

where the ՛ indicates that the i = j


terms are not included in the
sums and the factor of ½
compensates for having two Figure 2.4 Pair interactions
copies of each interaction in the (arrows) in a five-atom system.
sums.
➢ Sometimes written as ➢ Another common way of writing
the sum of interactions

Department of Materials Science and Engineering, KUET


Pair Potentials
➢The potential functions that we describe next are models –
approximations to the real interactions. They will generally be
empirical, with parameters that must be determined in some way.

➢In the end, the models may not even be particularly accurate.

So why study these functions?


➢We will see that even when these potential functions are not
particularly good representations of the real interactions between
atoms, calculations based on them can shed light on important
materials processes, providing understanding of materials
structure and behavior not possible otherwise.

Department of Materials Science and Engineering, KUET


The Lennard-Jones potential
➢A typical example of pair potential is the Lennard-Jones
potential (Lennard-Jones 1924), 𝜑 𝑟 , expressed in terms of
interatomic distance, r, with the following two parameters:
𝜎 12 𝜎 6
𝜑 𝑟 = 4𝜖 −
𝑟 𝑟

where:
parameter ε is the lowest energy of the potential curve (≡ well
depth, cohesive energy)
parameter σ is the interatomic distance at which the potential is
zero as shown in Figure 2.5.

Department of Materials Science and Engineering, KUET


The Lennard-Jones potential
❑ Several features are noted
here when two Lennard-
Jones atoms approach each
other from a long distance:
➢ At r = ∞, 𝜑 𝑟 is zero.
➢ When they get closer, dipole–
dipole attraction takes place,
and the r-6 term best
describes this van der Waals
interaction.
➢ When they get even closer,
the Pauli repulsion due to Figure 2.5 Schematic
overlapping electron clouds representation of the Lennard-
takes place, and the arbitrary Jones pair potential.
r-12 term describes this steep
increase in repulsion.
Department of Materials Science and Engineering, KUET
The Lennard-Jones potential

➢ At equilibrium interatomic distance,


rm (= 21/6σ = 1.122 σ), two forces are
balanced to a net force of zero, and
the corresponding energy becomes
minimum (d𝜑/dr = 0).
➢ The energy passes through 0 at r =
σ and increases steeply as r
decreases further due to the Pauli
repulsion.

Figure 2.5 Schematic


representation of the Lennard-
Jones pair potential.

Department of Materials Science and Engineering, KUET


The Lennard-Jones potential

➢ The Lennard-Jones (LJ) potential is commonly used because it is


simple and yet provides a good description of central-force
interatomic interactions.
➢ While it was developed to describe the general interaction between
closed-shell atoms and molecules, the Lennard-Jones potential has
been used to model almost everything.
➢ This simple pair potential can express the atomic interactions of
noble gases (Ne, Ar, Kr, etc.), spherical molecules, and secondary
bonds very well.
➢ However, since many-atom effects (~10% of the total potential in
most materials) of three atoms and so forth are completely
neglected, this potential cannot be applied to metals,
semiconductors, and other solids.

Department of Materials Science and Engineering, KUET


The Mie Potential
➢ Improvements can be made to the Lennard-Jones potential
by adding additional parameters.
➢ One approach would be to have the exponents in the 1/r12
and 1/r6 parts of the potential be adjustable parameters, i.e.,
instead of r-12, perhaps r-11 would be a better description of
the short-range interactions in some systems.
➢ A variation of the Lennard-Jones form, called the Mie
potential (or the mn Lennard-Jones potential), takes this
approach, replacing 1/r12 with 1/rm and 1/r6 with 1/rn.
Writing the Mie potential as
1
𝜖 𝑚𝑚 m−𝑛 𝜎 𝑚 𝜎 𝑛
𝜙𝑚𝑛 𝑟 = −
𝑚 − 𝑛 𝑛𝑛 𝑟 𝑟

yields an equivalent form to the Lennard-Jones potential.

Department of Materials Science and Engineering, KUET


The Mie Potential
➢ The Mie potential thus has four parameters: σ, ε, m, and n. The
parameters σ and ε have the same meaning as in the standard
Lennard-Jones potential: φmn(σ) = 0 and φmn(rm) = - ε, where
the minimum is located at

𝑟𝑚 = (𝑚/𝑛)1/(𝑚−𝑛) 𝜎

➢ If m = 12 and n = 6, the standard Lennard-Jones potential


is obtained.
➢ The mn potential yields somewhat better results for some
systems, especially those with repulsive potentials that are
less steep than the 1/r12 term in the standard Lennard-
Jones potential .

Department of Materials Science and Engineering, KUET


Other Pair Potentials
➢ The basic form of any pair interaction will be similar to that of
the Lennard-Jones potential, with a short-range repulsion and
an attractive well at long range.
➢ Like the Lennard-Jones form, these potentials best describe
systems made up of closed-shell atoms, ions, and molecules.
Here we describe two of the more commonly used forms.
➢ The Born-Mayer potential, which is sometimes referred to as
the exponential-6 (or exp-6) potential, is
𝐶
𝜙 𝑟 = 𝐴𝑒 −𝛼𝑟 − 6
𝑟
where A, C, and α are positive constants that depend on the
identities of the interacting atoms. α, with units of 1/length,
governs the steepness of the repulsive wall.

Department of Materials Science and Engineering, KUET


Other Pair Potentials
➢ An advantage over the Lennard-Jones potential is that the
exponential function that describes the repulsive part of the
potential matches what we expect from the electronic distributions
of the interacting atoms and is thus more physically reasonable than
the very repulsive 1/r12 term in the Lennard-Jones potential.
➢ The long-range part of the potential is the van der Waals -1/r6
attraction.
➢ The Born-Mayer potential is often used for ionic systems, where the
electrostatic interactions between the ions are included. At times, it
is convenient to write the exp-6 potential as
𝜖 𝑟 6
𝑚
𝜙 𝑟 = 6𝑒 −𝛼 𝑟−𝑟𝑚 − 𝛼𝑟𝑚
𝛼𝑟𝑚 − 6 𝑟
where the position of the minimum is rm and the well depth is ε and
A = 6ε exp(αrm)/(αrm - 6) and C = αr7m ε /(αrm - 6).
Note that this form is restricted to values of rm and α such that αrm - 6 > 0.

Department of Materials Science and Engineering, KUET


Other Pair Potentials

❑ Arising from a simple theory of diatomic bond potentials,


the Morse potential takes the form

𝜙 𝑟 = 𝜖 𝑒 −2𝛼 𝑟−𝑟m − 2𝑒 −𝛼 𝑟−𝑟m

where, as usual, rm is the position of the minimum, 𝜖 is the well


depth, and α governs the shape of the potential.
➢ The Morse potential has a very soft repulsive wall and is
sometimes used to model interatomic interactions in metals.

Department of Materials Science and Engineering, KUET


Potentials for Ionic Solids
➢ Ionic solids typically consist of closed-shell ions with little
charge in the interstitial regions.
➢ In ionic solids, however, ions can be polarized or charged, and
the coulombic component can spread into a very long-range.
➢ To describe potentials for these solids, therefore, one must
consider two pair-interaction terms; short-range and long-
range potentials as follows:
𝑁 𝑁
1 𝑞𝑖 𝑞𝑗
𝑈 = ෍. ෍ ′ 𝜙𝑖𝑗 𝑟𝑖𝑗 + 𝑘
2 𝑟𝑖𝑗
𝑖=1 𝑗=1

where rij is the distance between the two ions, and qi and
qj are the charges on ion i and j, respectively. k is a
parameter whose value depends on the units.

Department of Materials Science and Engineering, KUET


Potentials for Ionic Solids
❑ The first term involves short-ranged interactions consists of
typically two contributions—one by the repulsive interactions
between electrons in close range and the other by the
attractive interactions of van der Waals nature. A common
choice for 𝜙ij is the Born-Mayer potential,

−𝛼𝑟
𝐶
𝜙𝑖𝑗 𝑟𝑖𝑗 = 𝐴𝑒 − 6
𝑟
❑ The second term for coulomb interaction is the long-range
interaction that decays as 1/rij .
𝑞𝑖 𝑞𝑗
𝑘
𝑟𝑖𝑗

Department of Materials Science and Engineering, KUET


Potentials for Ionic Solids

➢ One should be very careful when using ionic potentials since


anions such as oxygen ions have a tendency to change their
size and charge states, especially with transition metal ions.
➢ It is also known that potentials for the mixed systems such as
the Al–Al2O3 system do not work properly.
➢ The fixed-charge potentials incorrectly predict the cohesive
energy of ionic materials, and they cannot be used to
simulate oxidation at metal surfaces or metal/oxide
interfaces where the local ion charge can be significantly
different from that in the bulk oxide.

Department of Materials Science and Engineering, KUET


Potentials for ionic solids
➢ Unlike solids made up of neutral atoms, ionic systems have large
electric fields that can cause significant distortions of the electronic
distributions around the ions (polarization) as shown schematically in
Figure 5.7a.
➢ For perfect, highly symmetric crystal structures, these distortions may
not matter, as the electric fields are essentially radial. However, as the
ions in the lattice vibrate, or if they are in asymmetric positions (e.g.,
near defects), field gradients can be formed and distortions of the
local electronic structure can be large.
➢ Since these distortions change with structure, a fixed potential cannot
model changes in atom position very well, including lattice vibrations
(phonons).
Figure 5.7 (a) Schematic view of distortion of the
valence electron charge in an atom in an electric
field. The small black circle represents the nucleus
plus core electrons. (b) The shell model
representation of (a). A spherical charge whose
center is held by a harmonic potential (spring) to the
nucleus/core electrons.
Department of Materials Science and Engineering, KUET

You might also like