Abstract
Nuclear magnetic resonance (NMR) spectroscopy is one of the most commonly
used tools in the analysis of chemical structures. In NMR the nuclear
spin-states of a molecule are probed, yielding information about the chemical
environment of each nucleus. As the spins of nearby nuclei interact, concrete
information about the relative positions of nuclei are also obtained. Due to
the widespread use of NMR spectroscopy, there is a keen interest in theoretical
predictions of the relevant parameters, the nuclear shielding constants
and indirect nuclear spin-spin coupling constants (SSCC). Though highly
advanced quantum chemical methods have been developed, the calculation
of NMR parameters with quantitative accuracy is far from trivial. In this
thesis I address some of the issues that makes accurate calculation of NMR
parameters so challenging, with the main focus on SSCCs.
High accuracy quantum chemical calculations are today typically based
on coupled cluster (CC) theory. The CCSD(T) method is very popular and
is known to give accurate results for a wide range of molecular properties,
including nuclear shielding constants, but not SSCCs. An alternative to
CCSD(T) for calculating SSCCs could be the CC3 method, but no programs
were available to perform such calculations. As part of this thesis the CFOUR
program has therefore been extended to allow the calculation of SSCCs using
the CC3 method. CC3 calculations of SSCCs have then been performed for
several molecules, including some difficult cases. These results show that
CC3 captures important correlation effects ignored by the previous standard
method, CCSD. Indeed, CC3 is shown to accurately reproduce gas-phase
experimental results for the SSCCs of acetylene.
In any molecular property calculation the one-electron basis set must be
chosen with care as the basis set must be able to describe not only Coulomb
interactions but also, e.g. the magnetic interaction between an electron and
a nucleus. SSCC calculations require the use of specially modified basis
sets if converged results are to be obtained. The convergence of two series
of SSCC optimized basis sets, the (aug-)pcJ-n, n=1,2,3 and the (aug-)ccJpVXZ,
X=D,T,Q,5, have been compared for systems where both contact
and non-contact contributions to the SSCC are important. It was found that
the triple zeta basis sets with diffuse functions, i.e. aug-ccJ-pVTZ or augpcJ-
2, can provide well converged results. However, these basis sets have
considerable contraction errors for one-bond SSCCs, so it might be desirable
to perform the calculations with the totally uncontracted versions of the basis
sets when calculating this kind of SSCCs.
When comparing experimental and theoretical results, the effect of molecular
vibrations must be included. The calculation of vibrational corrections
to NMR parameters has been reviewed as part of this thesis. A study of the
basis set convergence of vibrational corrections to nuclear shielding constants
has also been performed. The basis set error in vibrational correction calculations
is split amongst errors in the molecular force-field parameters and
errors in the shielding surface. Finally, accurate vibrational corrections to the
nuclear shielding calculations of noble-gas dimers are presented. While these
corrections are small compared to the total shielding, they have a large effect
on the chemical shift. By combining these results with relativistic effects
accurate predictions of the chemical shifts of these dimers are obtained.
used tools in the analysis of chemical structures. In NMR the nuclear
spin-states of a molecule are probed, yielding information about the chemical
environment of each nucleus. As the spins of nearby nuclei interact, concrete
information about the relative positions of nuclei are also obtained. Due to
the widespread use of NMR spectroscopy, there is a keen interest in theoretical
predictions of the relevant parameters, the nuclear shielding constants
and indirect nuclear spin-spin coupling constants (SSCC). Though highly
advanced quantum chemical methods have been developed, the calculation
of NMR parameters with quantitative accuracy is far from trivial. In this
thesis I address some of the issues that makes accurate calculation of NMR
parameters so challenging, with the main focus on SSCCs.
High accuracy quantum chemical calculations are today typically based
on coupled cluster (CC) theory. The CCSD(T) method is very popular and
is known to give accurate results for a wide range of molecular properties,
including nuclear shielding constants, but not SSCCs. An alternative to
CCSD(T) for calculating SSCCs could be the CC3 method, but no programs
were available to perform such calculations. As part of this thesis the CFOUR
program has therefore been extended to allow the calculation of SSCCs using
the CC3 method. CC3 calculations of SSCCs have then been performed for
several molecules, including some difficult cases. These results show that
CC3 captures important correlation effects ignored by the previous standard
method, CCSD. Indeed, CC3 is shown to accurately reproduce gas-phase
experimental results for the SSCCs of acetylene.
In any molecular property calculation the one-electron basis set must be
chosen with care as the basis set must be able to describe not only Coulomb
interactions but also, e.g. the magnetic interaction between an electron and
a nucleus. SSCC calculations require the use of specially modified basis
sets if converged results are to be obtained. The convergence of two series
of SSCC optimized basis sets, the (aug-)pcJ-n, n=1,2,3 and the (aug-)ccJpVXZ,
X=D,T,Q,5, have been compared for systems where both contact
and non-contact contributions to the SSCC are important. It was found that
the triple zeta basis sets with diffuse functions, i.e. aug-ccJ-pVTZ or augpcJ-
2, can provide well converged results. However, these basis sets have
considerable contraction errors for one-bond SSCCs, so it might be desirable
to perform the calculations with the totally uncontracted versions of the basis
sets when calculating this kind of SSCCs.
When comparing experimental and theoretical results, the effect of molecular
vibrations must be included. The calculation of vibrational corrections
to NMR parameters has been reviewed as part of this thesis. A study of the
basis set convergence of vibrational corrections to nuclear shielding constants
has also been performed. The basis set error in vibrational correction calculations
is split amongst errors in the molecular force-field parameters and
errors in the shielding surface. Finally, accurate vibrational corrections to the
nuclear shielding calculations of noble-gas dimers are presented. While these
corrections are small compared to the total shielding, they have a large effect
on the chemical shift. By combining these results with relativistic effects
accurate predictions of the chemical shifts of these dimers are obtained.
| Originalsprog | Engelsk |
|---|
| Forlag | Department of Chemistry, Faculty of Science, University of Copenhagen |
|---|---|
| Status | Udgivet - 2016 |