Repository logo

Infoscience

  • English
  • French
Log In
Logo EPFL, École polytechnique fédérale de Lausanne

Infoscience

  • English
  • French
Log In
  1. Home
  2. Academic and Research Output
  3. EPFL thesis
  4. Accelerating path integral evaluation of equilibrium and kinetic isotope effects
 
doctoral thesis

Accelerating path integral evaluation of equilibrium and kinetic isotope effects

Karandashev, Konstantin  
2018

Investigating the effect of isotope substitution on equilibrium and kinetic properties of molecules has become an important tool for estimating the importance of nuclear quantum effects. In this work, we discuss calculating both equilibrium and kinetic isotope effects, i.e., the isotope effects on a system's partition function and a reaction's rate constant. With the help of Feynman's path integral formalism, both quantities can be estimated using standard Monte Carlo methods that scale favorably with system's dimensionality; improving efficiency of such approaches is the main focus of this work.

First of all, we developed a novel procedure for changing mass stochastically during an equilibrium isotope effect calculation, and evaluated the numerical benefits of combining it with two popular approaches for calculating isotope effects, using either direct estimators or thermodynamic integration. We demonstrate that the modification improves statistical convergence of both methods, and that it additionally allows to eliminate integration error of thermodynamic integration. The improved methods are tested on equilibrium isotope effects in a model harmonic system and in methane.

Then we turn our attention to kinetic isotope effect calculations with the quantum instanton approximation, a method whose path integral implementation belongs among the most accurate approaches for evaluating reaction rate constants in polyatomic systems. To accelerate quantum instanton calculations of kinetic isotope effects, we combine higher-order Boltzmann operator factorization with virial estimators, allowing us to speed up both the convergence to the quantum limit and statistical convergence of the calculation. We estimate the overall resulting acceleration using H+H2/D+D2 as a benchmark system, and then apply the accelerated method to several kinetic isotope effects associated with the H+CH4=H2+CH3 exchange.

Last but not least, we explored ways to improve on the quantum instanton approximation for reaction rate constants. To that end, we review quantum instanton and Hansen-Andersen approximations, and propose a combined method, which, as the Hansen-Andersen approximation, has the correct high-temperature behavior, and at the same time, as the quantum instanton approximation, has more flexibility by allowing the dividing surface for the reaction to split into two surfaces at low temperatures. The properties of the combined method are tested on symmetric and asymmetric Eckart barrier.

  • Files
  • Details
  • Metrics
Type
doctoral thesis
DOI
10.5075/epfl-thesis-8211
Author(s)
Karandashev, Konstantin  
Advisors
Vanicek, Jiri  
Jury

Prof. Lothar Helm (président) ; Prof. Jiri Vanicek (directeur de thèse) ; Prof. Michele Ceriotti, Prof. Jeremy Richardson, Prof. O. Anatole von Lilienfeld (rapporteurs)

Date Issued

2018

Publisher

EPFL

Publisher place

Lausanne

Public defense year

2018-04-24

Thesis number

8211

Total of pages

142

Subjects

path integrals

•

isotope effects

•

chemical kinetics

•

quantum instanton

•

isotope fractionation

•

reaction rate constant

•

transition state theory

•

Monte Carlo

•

thermodynamic integration

•

free energy perturbation

EPFL units
LCPT  
Faculty
SB  
School
ISIC  
Doctoral School
EDCH  
Available on Infoscience
April 25, 2018
Use this identifier to reference this record
https://infoscience.epfl.ch/handle/20.500.14299/146167
Logo EPFL, École polytechnique fédérale de Lausanne
  • Contact
  • infoscience@epfl.ch

  • Follow us on Facebook
  • Follow us on Instagram
  • Follow us on LinkedIn
  • Follow us on X
  • Follow us on Youtube
AccessibilityLegal noticePrivacy policyCookie settingsEnd User AgreementGet helpFeedback

Infoscience is a service managed and provided by the Library and IT Services of EPFL. © EPFL, tous droits réservés