Related docs:
Research Interests
MS research
From grad school:
Chapters of my MS thesis on Stochastic Searches (all in Spanish):
More information about my Surface Chemistry projects and previous projects
-----------------Minima Stochastic Searches over Potential Energy Surfaces:
I also did some computational chemistry, applying and testing programs for the energy optimization of small molecules. The molecular mechanics program was originally written a while ago by Prof. M. Saunders, and after some home-made modifications, the program was adjusted to geometrically minimize homogeneous small clusters formed by weak interactions.
Additionally, the program as a force field, was capable to cover wide searches on the potential energy surface, meaning that, starting from a minima of the cluster it would search for other local minima that would have the same or lower energy. In other words, starting from a stationary point, the energy was kicked to a higher level and after landing, the optimization began to scan for new stationary points nearby. And the process was repeated again and again, until no more new local minima seem to apear. The force field basically deals with a collection of mathematical expressions (or potentials), as accurate as possible, trying to represent the 3D-location of every particle, the nature of their interactions, and how these affect the energy of the total system that these particles comprise, still based strictly on a Newtonian framework, without the hassle of the quantum mechanics complexity.
The energetic aspects that two or more particles, as a whole system, have in terms of their position, constitute an equilibrium between repulsion and atraction forces. finding a very narrow range of 3D position where the energy is at its lowest value, and at every direction outside of this range, the energy starts growing exponentially.
Nowadays several potentials are available to be incorporated in a Monte Carlo computational method, but there is always a trade-off between the number of parameters, mathematical complexity, CPU time, and simulation accuracy.
The ideal approach is to indentify those parameters that can be ruled out, or at least simplified under proper assumptions, so the final force field performance does not lose substantial precision while using reasonable computational resources. For the use of this modified program for the first time under these sort of noble gases clusters, a simple system was chosen in order to test its performance: argon small and medium size clusters. In principle, the force field paramenters can be adjusted in order to find applications on several other higher molecular systems, having stronger interactions like covalent or ionic bonding.
In this work i was responsible to write, develop and test several subprograms that were integrated as algorithms to the main force field program. Two different potentials and their parameters were tested, using noble gases clusters comprising 6 to 13 atoms. The role of specific parameters on the potential accuracy were also studied. Different values were given to one of this parameters, that was the result of original assumptions, with the intention to simplify mathematical manipulation. Doing this we could elucidate how this parameter dictated qualitatively and quantitatively the minima found.
The runs included total optimization as well as stochastic searches for the global minimum of every cluster, while recording the number of different local minima, their energy and their geometrical relation. Theses algorithms turned out to be very efficient and relatively fast. They covered a great part of the whole surface, recording a large number of minima, therefore increasing the chances of hunting down the global minimum of noble gases clusters of small to medium size.
-----------------More information about my surface chemistry projects and previous research projects