Use this URL to cite or link to this record in EThOS:
Title: Improving the accuracy of lattice energy calculations in crystal structure prediction using experimental data
Author: Gatsiou, Christina-Anna
ISNI:       0000 0004 5917 7601
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Crystal structure prediction (CSP) has been a problem of great industrial interest but also a fundamental challenge in condensed matter science. The problem involves the identification of the stable and metastable crystals of a given compound at certain temperature and pressure conditions. Computational CSP methods based on the lattice energy minimization have been successful in identifying experimentally observed crystals of an organic compound as local minima of the lattice energy landscape but not always with the correct relative stability. This is primarily controlled by the lattice energy model. The lattice energy model adopted in this work is based on the assumption that molecules are rigid, electrostatic interactions are modelled via distributed multipoles derived from the ab initio charge density of the gas phase conformation and an empirical pairwise exp-6 potential for the repulsion dispersion interactions. Based on the fact that the reliability of all computational models is based on their agreement with experimental evidence, the use of available experimental data for improving the lattice energy model is the main focus of this work. First the impact of different modelling choices -- choice of level of theory for electrostatics and parameters for the repulsion-dispersion term -- in the modelling of experimental structures, energies and relative stabilities is investigated. Results suggest that a reestimation of the repulsion-dispersion parameters is expected to produce parameters consistent with changes in the other lattice energy terms and bring the model closer to experiment, consequently improving predictions. An algorithm, CrystalEstimator, for fitting the exp-6 potential parameters by minimizing the sum of squared deviations between experimental structures and energies and the corresponding relaxed structures and energies is developed. The lattice energy of the experimental structures is minimized by the program DMACRYS. The solution algorithm is based on the search of the parameter space using deterministic low discrepancy sequences; and the use of an efficient local minimization algorithm. The proposed method is applied to derive transferable exp-6 potential parameters for hydrocarbons, organosulphur compounds, azahydrocarbons, oxohydrocarbons and organosulphur compounds containing nitrogen. Three different sets of parameters are developed, suitable for use in conjunction with three different models of electrostatics derived at the HF/6-31G(d,p), M06/6-31G(d,p) and MP2/6-31g(d,p) levels. A good fit is achieved for all the new sets of parameters with a mean absolute error in sublimation enthalpies less than 3.5 kJ/mol and an average rmsd15 less than 0.35 Å. Prediction studies are performed for acetylene, tetracyanoethylene and blind test molecule XXII and the generated lattice energy landscapes are refined with the new models. The observed experimental structures are predicted with better structural agreement but the same or higher ranking than those obtained by the previously used FIT parameter set.
Supervisor: Adjiman, Claire ; Pantelides, Costas Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available