Abstract
This thesis presents the investigation of atmospheric molecular clusters using
computational methods. Previous investigations have focused on solving
problems related to atmospheric nucleation, and have not been targeted at
the performance of the applied methods. This thesis focuses on assessing
the performance of computational strategies in order to identify a sturdy
methodology, which should be applicable for handling various issues related
to atmospheric cluster formation. Density functional theory (DFT) is applied
to study individual cluster formation steps. Utilizing large test sets of
numerous atmospheric clusters I evaluate the performance of different DFT
functionals, with a specific focus on how to control potential errors associated
with the calculation of single point energies and evaluation of the thermal
contribution to the Gibbs free energy.
Using DFT I study two candidate systems (glycine and pinic acid) for atmospheric
cluster formation. Glycine is found to have a similar potential as
ammonia in enhancing atmospheric nucleation. Pinic acid molecules form
favourable clusters with sulfuric acid, but with formation free energies which
are too low to explain observed nucleation rates. Pinic acid could thereby
enhance the further growth of an existing cluster by condensing on the surface.
Conclusively, I find that the performance of a single DFT functional can
lead to an inadequate description of investigated atmospheric systems and
thereby recommend a joint DFT (J-DFT) approach. The approach considers
the approximate exchange-correlation functional as a source of random
errors, and thereby utilize the average of several functionals to describe individual
atmospheric cluster formation steps. The J-DFT approach thereby
compensates for the error that a given functional might introduce and represents
a swift alternative methodology to investigate the formation of atmospheric
molecular clusters
computational methods. Previous investigations have focused on solving
problems related to atmospheric nucleation, and have not been targeted at
the performance of the applied methods. This thesis focuses on assessing
the performance of computational strategies in order to identify a sturdy
methodology, which should be applicable for handling various issues related
to atmospheric cluster formation. Density functional theory (DFT) is applied
to study individual cluster formation steps. Utilizing large test sets of
numerous atmospheric clusters I evaluate the performance of different DFT
functionals, with a specific focus on how to control potential errors associated
with the calculation of single point energies and evaluation of the thermal
contribution to the Gibbs free energy.
Using DFT I study two candidate systems (glycine and pinic acid) for atmospheric
cluster formation. Glycine is found to have a similar potential as
ammonia in enhancing atmospheric nucleation. Pinic acid molecules form
favourable clusters with sulfuric acid, but with formation free energies which
are too low to explain observed nucleation rates. Pinic acid could thereby
enhance the further growth of an existing cluster by condensing on the surface.
Conclusively, I find that the performance of a single DFT functional can
lead to an inadequate description of investigated atmospheric systems and
thereby recommend a joint DFT (J-DFT) approach. The approach considers
the approximate exchange-correlation functional as a source of random
errors, and thereby utilize the average of several functionals to describe individual
atmospheric cluster formation steps. The J-DFT approach thereby
compensates for the error that a given functional might introduce and represents
a swift alternative methodology to investigate the formation of atmospheric
molecular clusters
| Original language | English |
|---|
| Publisher | Department of Chemistry, Faculty of Science, University of Copenhagen |
|---|---|
| Number of pages | 111 |
| Publication status | Published - 2014 |