Molecular dynamics insights into the Debye process of 1-propanol

Abstract

The dielectric response of mono-alcohols exhibits a strong Debye peak generally attributed to the dynamics of hydrogen-bonds (HB) supramolecular structures through a mechanism that remains unclear in many aspects. In this letter, we use standard all-atom molecular dynamics simulations to investigate this phenomenon in 1-propanol, a prototypic monoalcohol, over a wide temperature range covering a significant change in dielectric permittivity. We obtained dielectric spectra showing a Debye peak in good agreement with experimental data, which we decomposed into the self and cross parts of the dipolar correlations. The latter extends over a few molecular distances and contributes increasingly to the Debye peak upon cooling. To investigate its physical origin, we analyzed the HB structures by identifying clusters from simulation snapshots. Below 300~K, the dielectric permittivity was shown to arise almost entirely from intra-cluster cross-correlations. Furthermore, by tracking the dipole decorrelation of groups of molecules initially belonging to the same cluster, we found that supramolecular structures play a key role in stabilizing cross-correlation over time scales longer than the relaxation of individual molecules.

Figures (5)

• Figure 1: (a) Dielectric loss spectra obtained from simulations (total and self, shown as colored lines) compared with experimental results sillren2014liquid (dashed grey lines). (b) Focus on the comparison between simulations and experiments near 200 K. (c) Relaxation time and (d) static permittivity plotted as functions of the inverse temperature, with experimental data sillren2014liquid shown in grey.
• Figure 2: (a) Relative orientation and distance dependence of static cross-correlations at 240 K. (b) $r$-dependent Kirkwood correlation factor. (c) Relaxation spectra at 220 K when considering virtual cavities of increasing radius $r_\mathrm{c}$. The inset shows the pair distribution function of the center of charge. (d) Dipole correlation function at $220$ K for the self (black) and the cross part from shells of increasing radius (colors). (e) Total relaxation time as a function of the cavity size for each temperature.
• Figure 3: (a) Relaxation spectra corresponding to the self dipole orientational correlation function (dashed lines) and to the HB correlation function (solid lines), at 200, 260, and 340 K. (b) Fraction of molecule involved in a HB as a donor (D) or acceptor (A) (blue circles) or as A and D (orange triangles), as a function of the inverse temperature. (c) Distribution of H-bond cluster sizes and (d) mean cluster size for each temperature.
• Figure 4: (a) Same as fig. \ref{['fig2']}a, restricted to molecules within the same cluster as the reference molecule (left) or outside of it (right). (b,c) $r$-dependent Kirkwood correlation factor showing the intra-cluster contribution at 240 K (b) and at different temperatures (c). (c) Static intra-cluster cross-correlation originating at different temperatures. (d) Finite-size Kirkwood correlation factor including all cross-correlations ($G_\mathrm{K}(r_\mathrm{c})$, in blue) and restricted to intra-cluster (in orange). (e) Same as fig. \ref{['fig1']}a with the spectra (in dashed lines) at 340, 240, and 200 K, obtained by restricting the cross-correlations to the molecules within the same cluster as the reference molecule at $t=0$ (shown in red in the drawing).
• Figure A1: (a) Probability distribution $P_n(0\cap t)$ of finding $n$ molecules in common between a cluster at time zero and at time $t$. (b) For each temperature, the mean number of molecules in common normalized by the time zero value, as a function of the elapsed time normalized by the self relaxation time. The back curve indicates the individual HB correlation function.