Monte Carlo Simulation of CsInCs Clusters

1
Monte Carlo Simulation of (CsI)nCs Clusters
Rich Wyrwas Jefferson Wu Dr. Matthew Wolf Dr.
Robert Whetten
2
Problem
Krückeberg,S.,et.al, Phys Rev.Lett., 2000, 85,
4494.
3
Electron Diffraction of Small Clusters (CsI)nCs

• N 30-39 Give similar diffraction patterns
• N 32 Gives a different pattern because of a
  different structure for N 65 atoms

Krückeberg, S.,et.al
4
Possible Structures for (CsI)nCs

• Rock Salt Structure
• Same Crystal Pattern as NaCl
• Same Structure as N30-39 Clusters

5
Possible Structures for (CsI)nCs

• Rhombic Dodecahedron
• Same Structure as CsCl
• N 32 cluster has 65 atoms
• 65 atoms can form a complete Rh.dodecahedron
• Magic Numbers
• n4k3-6k24k-1

6
Comparison of Cluster to Basic Structure
CsCl 8 Nearest Neighbors
Rock Salt 6 Nearest Neighbors
7
Electron Diffraction Patterns for Various
Structures

• Theoretical vs. Experimental Data
• There is good agreement between calculated and
  measured patterns
• Lower plot emphasis shift between CsI and Rock
  Salt structures
• Shaded area is the standard deviation

8
So What is the Problem?

• Big Contradiction
• Energies for each structure doesnt make sense

9
Where did they go wrong and how can we fix it?

• Potential Function?

Coulomb Born-Meyer
Charge-Dipole
Dipole-Dipole
MC parameter
Self-consistent Dipole Term
10
Monte Carlo Code

• Our code takes in to account for the Self
  Consistent Dipole
• MC Parameter is the position
• Randomly change and allow to relax
• Gather Statistics of Energy Structures
• From these Statistics approximate the Average
  Structure
• Electron Diffraction

11
Parallel Implementation (EP)
P0

• MC part
• Diffraction Part

P5
12
Progress

• Serial Code Runs
• Need to make correction to potential function
• Need to add in Diffraction Part
• Need to Parallelize