Scribd
Upload
18 views

Assignment 2

Uploaded by

Komal Kumar
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
18 views

Assignment 2

Uploaded by

Komal Kumar
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

Assignment 2

Komal Kumar
21110106

Solution 1- Differences between Monte Carlo and Molecular Dynamics Simulations

Methodology
MD: Molecular Dynamics (MD) simulations follow a deterministic approach by solving
Newton's equations of motion for particles, calculating their trajectories based on initial
positions, velocities, and interatomic forces. These forces are determined using
empirical or theoretical force fields that describe particle interactions. MD provides a
time-resolved picture of the system, capturing how particles move and interact
dynamically over time, making it useful for studying time-dependent processes and the
evolution of material properties.

MC: Monte Carlo (MC) simulations use a stochastic approach, relying on random
sampling to explore the configuration space of a system. Instead of tracking particle
motion over time, MC generates configurations based on probabilistic rules. It employs
statistical sampling, accepting or rejecting proposed configurations according to criteria
like energy or entropy. This method allows MC to efficiently explore a broad range of
configurations, making it particularly useful for complex systems where calculating
precise trajectories is difficult.

Outcomes

MD : Molecular Dynamics (MD) simulations provide valuable insights into


time-dependent properties such as diffusion coefficients, viscosity, and time correlation
functions. They generate realistic particle trajectories, allowing for the simulation of
processes like phase transitions, chemical reactions, and material deformation over
time. This makes MD ideal for studying the dynamic behavior of materials and systems.

MC: Monte Carlo (MC) simulations are well-suited for calculating thermodynamic
properties at equilibrium, such as partition functions and free energies. They excel in
efficiently sampling configurations of complex systems, even in high-dimensional
spaces where MD simulations may be computationally expensive, making MC ideal for
studying equilibrium behavior.

Metropolis Algorithm in MC Simulations


Metropolis Algorithm is a special monte carlo method used to generate sample
configurations of a system. Here’s are basic steps in which this algorithm functions:

a) Initiation: Start with an initial configuration of the system (e.g., positions of


particles).
b) New configurations: Randomly perturb the current configuration to generate a
new proposed configuration.
c) Calculation of Energy difference: Compute the change in energy (ΔE) between
the proposed configuration and the current one.

Corresponding to the calculated energy difference, there is a certain acceptance


criteria for a sample configuration to contribute to the probability which goes like:

● If the proposed configuration has a lower energy (ΔE < 0), it is accepted.
● If the proposed configuration has a higher energy (ΔE > 0), it is accepted
with a probability given by: P= exp⁡(−ΔE/kT), where k is Boltzman constant
and T is absolute temperature

Role in Generation of Sample Configurations: The Metropolis algorithm ensures that


MC sampling respects detailed balance, which is essential for accurately representing a
system's thermodynamic properties.
By accepting or rejecting new configurations based on their energy, the algorithm
efficiently explores the configuration space, allowing the system to sample both
low-energy and higher-energy states. As more configurations are generated and
sampled, the system gradually converges to equilibrium, following a Boltzmann
distribution of states, which enables precise calculations of equilibrium thermodynamic
properties.

Solution-2

Fix mol/swap command in Lammps


The `fix mol/swap` command in LAMMPS is used to swap molecules in a
simulation using Monte Carlo method. It works by picking two random molecules
and trying to exchange their identities. Whether the swap is accepted depends
on the energy change, usually following the Metropolis rule. This command helps
make the simulation more efficient, especially when dealing with mixtures or
systems where molecules can change type or react. It helps the system explore
more configurations and can be useful for studying things like phase changes or
chemical reactions.
For ε= 0.1, results are as
followers:

Since the interactions are not


that strong between both kinds
of atoms, the miscibility is not
good. Both polymers do not like
to mix with each other to much
extent which is clear from the
image.
The no. of molecules is also
constant due to higher value of
energy difference and thus lower
probability of mol/swap.
For ε= 1.0

Here in this, intramolecular (interactions between same kind of molecules) and


intermolecular(interactions between different kinds of molecules) are the same
which makes the miscibility better than the last case which can also be observed
from the simulation result. Here, the energy difference for miscibility has been
lowered and thus enhancing the miscibility. Here, we can say that mol/swap
probability is 0.5
For ε= 1.1

As, it is visible from the simulation result that both polymers are completely
miscible in this case due to stronger intermolecular interactions.

Solution-3
(a)
Initial Frame Final Frame

Density Variation

b)
Crack Formation in Silica

Initial Frame Final Frame

Water Absorption
Initial Frame Final Frame
Plot for Energy vs time

Here, it can be seen that energy achieves a global minima at around 20-22 steps
. It happens due to metropolis theorem as it selects all the energy lowering steps
and assigns a non zero probability of energy enhancing configurations. The
number of water molecules absorbed are about 15-25 as seen visually.

Solution-4

After running the script on lammps, I got many states of the system. Here are three
which demarcate the transition. Visually, the black pixel represents an atom with +1 spin
and white represents -1 spin.

Thus, for T>>Tc


For T=Tc

For T<Tc
Tc is defined as Curie temperature where magnetic properties of a material change and
it transforms from paramagnetic to diamagnetic. At Tc, the material has zero
magnetization.

b)

Looking at the plots, it can be concluded that energy starts stabilizing near Tc. From the
Ising Model, the energy at low external magnetic fields can be given as:
As at higher temperatures, all the neighboring charges produce H=0 due to opposite
charges.
From the magnetization vs Temp plot, the maximum magnetization at 2kT/J and starts to
stabilize at about 3kT/J.

Solution 5: Here is the Energy plot obtained in running the Cassandra MC :


a)

Here the system chooses the metropolis algorithm to sample the configurations with
which, energy starts decreasing very fast in the start and then starts to stabilize after
certain steps as clear from the plot shown.

You might also like