
Exercise 7

In this exercise you will play with different MD options.
The system under study is a supercell of 8 atoms of Si in
the diamond structure. In order to make the runs fast,
we use a single-Z basis set with short orbitals (energy shift
300 meV). 

Verlet dynamics

Go to the directory Ex7/verlet. Edit the si8.fdf file
and study it. It is an input file for a simple 
(microcanonical) MD simulation with the Verlet algorithm.
The initial positions correpond to the ideal crystal structure.
The inicial velocity is 300 K.  The time step is 3 fs, and the
simulation is run for 200 steps.

Run the simulation and analyze the result. Look at the si8.MDE file,
and plot:

- total energy and KS energy versus time. Check the energy conservation
- temperature versus time.  Why is the average temperature smaller
  than the initial one?

Nose dynamics

In directories nose-X-Y you have input files for different Nose runs,
for the same si8 system.
X indicates the target temperature, and Y the mass of the Nose thermostat.

Run each of the symulations, and analyze the results, again plotting the
different physical quantities versus time. Look at how the results
change with different target temperatures and Nose masses


Cell vectors relaxation

In directories Relax-CG and Relax-PR, you have input files to relax the
lattice vectors of the same Si8 system. In this case, the initial
configuration is such that the lattice vectors are distorted with
respect to those of the diamond structure.  Relax-CG does the simulation
using conjugate gradients minimizations, whereas Relax-PR uses
Parrinello-Rahman dynamics, with the MD.Quench option.

Run both cases. Plot the evolution of the lattice constant and
angles versus time. Which method is more efficient?


Parrinello-Rahman dynamics.

In directory PR you have an input file for a Parrinello-Rahman run.
The system starts from a distorted geometry. The target pressure is
cero.  Run the simulation and analyze the results. In particular,
look at the evolution of the lattice constants and angles
versus simulation time. If you have time, you can run another 
simulation with a target pressure of 10 GPa, and see the differences.


