Overview

Teaching: min
Exercises: min

Questions

Objectives

Tutorial

1. Setup
2. The qd_build workflow
3. The fitting workflow
4. Video recording

0. Setup

Back to TOC

In your working directory, copy the folder containing all the files required for this tutorial:

cp -r /projects/academic/cyberwksp21/Instructors_material/rpascazio/exercises/ .

Please refer to the following documentations:

1. The Building a Quantum Dot Model assignment is based on the CAT documentation and on the Building a Quantum Dot Model Tutorial;
2. The Forcefield Optimization assignment is based on the Auto-FOX documentation and on the Forcefield Optimization Tutorial.

1. Building a Quantum Dot Model

Back to TOC

We aim to create nanocrystals from a charge-balanced Cd68Se68 core. Use the corresponding CdSe.xyz file in your 1_qd_build directory to:

1. Replace a 30% fraction of the Se ions in the model with Cl dummies in a file called CdSe_30Cl.xyz with random distribution and use the file to replace the Cl on the surface with stearate anions.
2. Replace a 20% fraction of the Cd ions in the model with Na dummies in a file called CdSe_20Na.xyz with clustered distribution and use the file to replace the Na on the surface with oleylammonium cations. (Suggestion: their SMILES strings are available on PUBCHEM).

2. The fitting workflow

Back to TOC

We aim to fit the classical forcefield parameters of a CsPbBr3 core from a previous QM-MD trajectory using the Adaptive Rate Monte Carlo (ARMC) algorithm.

1. Modify the yaml script to fit the CsPbBr3_MD.xyz trajectory (you can find them in your 2_fitting directory). Use the oxidation states of the atoms (respectively 1.0, 2.0, -1.0 for Cs, Pb, Br) as starting parameters for their charges, and modify the guess_rdf.py script in the scripts subdirectory to obtain the starting values for the sigmas. (Beware of the units and of the constraints!). Run the parametrization for around 5/10 steps (Suggestion: use the logfile to check the number of last iteration). Use the python scripts in the scripts subdirectory to plot the errors of the ARMC and the Radial Distribution Functions (RDFs) of the best step from the armc.hdf5 file. What do the plots show? Why? (Suggestion: take a look at the errors in the logfile!)
2. Adjust the parameters to obtain a “better” fitting and repeat the same procedure for 5/10 more steps (don’t forget to rename your previous MM_MD_workdir beforehand). What do the plots look like now? How do you expect them to change over time?

3. Video recording

Back to TOC

Key Points