The scaling behavior of ab initio molecular dynamics simulations for the different size bulk systems of liquid methanol is presented and thereby the characteristics of every system performing on either a local compute cluster or a supercomputer are analyzed. Additionally, the influence of different parameters on the quality of the infrared and Raman spectra is investigated using different simulation frameworks, including time step, convergence criteria, density functional approximation, and basis set. Both the maximally localized Wannier functions and the radical Voronoi tessellation approaches are employed to evaluate vibrational spectra from the trajectories. The results of infrared and Raman spectra are classified in two frequency regions, 500 to 1600 cm−1 and 2500 to 4000 cm−1, in order to compare and discuss the experimental spectra and the results derived from ab initio molecular dynamics simulations comprehensively. The outcome of this study guides future experimental and theoretical researchers in order to acquire a profound perception into vibrational spectra, which evolves the way of elucidating molecular structure.

Weighting methods applied to systems with many conformers have been broadly employed to calculate thermodynamic properties, structural characteristics, and populations. To better understand and test the sensitivity of conventional weighting methods, the conformational distributions of nicotine and its phosphorus-substituted derivatives are investigated. The weighting schemes used for this are all based on Boltzmann statistics. Classical Boltzmann factors based on the electronic energy and the Gibbs free energy are calculated at different quantum chemical levels of theory and compared to cluster weights obtained by the quantum cluster equilibrium method. Furthermore, the influence of the modified rigid-rotor–harmonic-oscillator (mRRHO) approximation on the cluster weights is investigated. The substitution of the nitrogen atom in the methylpyrrolidine ring by a phosphorus atom results in more monomer conformers and clusters being populated. The conformational distribution of the monomers remained stable at different levels of theory and weighting methods. However, going to dimers and trimers, we observe a significant influence of the level of theory, weighting method, and mRRHO cutoff on the populations of these clusters. We show that mRRHO cutoff values of 50 and 100 cm–1 yield similar results, which is why 50 cm–1 is recommended as a robust choice. Furthermore, we observe that the global minimum for ΔE 0 and ΔG varies in a few cases and that the global minimum is not always the dominantly occupied structure.