Inherent structures of phase-separating binary mixtures: nucleation, spinodal decomposition, and pattern formation

Anomalous properties in the potential energy landscape of a monatomic liquid across the liquid-gas and liquid-liquid phase transitions

As a liquid approaches the gas state, the properties of the potential energy landscape (PEL) sampled by the system become anomalous. Specifically, (i) the mechanically stable local minima of the PEL [inherent structures (IS)] can exhibit cavitation above the so-called Sastry volume, vS, before the liquid enters the gas phase. In addition, (ii) the pressure of the liquid at the sampled IS [i.e., the PEL equation of state, PIS(v)] develops a spinodal-like minimum at vS. We perform molecular dynamics simulations of a monatomic water-like liquid and verify that points (i) and (ii) hold at high temperatures. However, at low temperatures, cavitation in the liquid and the corresponding IS occurs simultaneously and a Sastry volume cannot be defined. Remarkably, at intermediate/high temperatures, the IS of the liquid can exhibit crystallization, i.e., the liquid regularly visits the regions of the PEL that belong to the crystal state. The model liquid considered also exhibits a liquid-liquid phase transition (LLPT) between a low-density and a high-density liquid (LDL and HDL). By studying the behavior of PIS(v) during the LLPT, we identify a Sastry volume for both LDL and HDL. The HDL Sastry volume marks the onset above which IS are heterogeneous (composed of LDL and HDL particles), analogous to points (i) and (ii) above. However, the relationship between the LDL Sastry volume and the onset of heterogeneous IS is less evident. We conclude by presenting a thermodynamic argument that can explain the behavior of the PEL equation of state PIS(v) across both the liquid-gas phase transition and LLPT.

Microscopic origin of breakdown of Stokes-Einstein relation in binary mixtures: Inherent structure analysis

Aqueous binary mixtures often exhibit dramatic departure from the predicted hydrodynamic behavior when transport properties are plotted against composition. We show by inherent structure (IS) analysis that this sharp composition dependent breakdown of the Stokes-Einstein relation can be attributed to the non-monotonic variation in the average inherent structure energy of these mixtures. Further IS analysis reveals the existence of a unique ground state, stabilized by both the formation of an optimum number of H-bonds and a favorable hydrophobic interaction at this composition. The surprisingly sharp turnaround behavior observed in the effective hydrodynamic radius also owes its origin to the same combination of these two factors. Interestingly, the temperature dependence of isothermal compressibility shows a minimum at the particular composition. Extensive studies on water-dimethyl sulfoxide and water-ethanol mixtures using two different force-fields of water reveal many features that are nearly universal. A justification of this quasi-universal behavior is provided in terms of a mode-coupling theory (MCT) of viscosity, which can serve as the starting point of a remarkable correlation observed with the nearest neighbor structure, as captured by the first peaks of the radial distribution function, and the slowdown in the intermediate scattering function at intermediate wavenumbers. Therefore, the formation of the local structure captured through IS analysis can be correlated with the MCT.

Anomalous viscoelastic response of water-dimethyl sulfoxide solution and a molecular explanation of non-monotonic composition dependence of viscosity

Amphiphilic molecules such as dimethyl sulfoxide (DMSO) and its aqueous binary mixtures exhibit pronounced nonideality in composition dependence of several static and dynamic properties. We carry out detailed molecular dynamics simulations to calculate various properties including viscosity of the mixture and combine the results with a mode coupling theory analysis to show that this nonideality can be attributed to local structures that are stable on a short time scale but transient on a long time scale to maintain the large scale homogeneity of the solution. Although the existence of such quasistable structures has been deciphered from spectroscopy, a detailed characterization does not exist. We calculate stress-stress autocorrelation functions (SACFs) of water-DMSO binary mixtures. We employ two different models of water, SPC/E and TIP4P/2005, to check the consistency of our results. Viscosity shows a pronounced nonmonotonic composition dependence. The calculated values are in good agreement with the experimental results. Fourier transform of SACF provides frequency-dependent viscosity. The frequency-dependent viscosity (that is, viscoelasticity) is also found to be strongly dependent on composition. Viscoelasticity exhibits sharp peaks due to intramolecular vibrational modes of DMSO, which are also seen in the density of states. We evaluate the wavenumber dependent dynamic structure factor and wavenumber dependent relaxation time. The latter also exhibits a sharp nonmonotonic composition dependence. The calculated dynamic structure factor is used in mode coupling theory expression of viscosity to obtain a semiquantitative understanding of anomalous composition dependence of viscosity. Both the self-diffusion coefficients and rotational correlation times of water and DMSO molecules exhibit nonmonotonic composition dependence.

A simple transferable adaptive potential to study phase separation in large-scale xMgO-(1-x)SiO2 binary glasses

A simple transferable adaptive model is developed and it allows for the first time to simulate by molecular dynamics the separation of large phases in the MgO-SiO2 binary system, as experimentally observed and as predicted by the phase diagram, meaning that separated phases have various compositions. This is a real improvement over fixed-charge models, which are often limited to an interpretation involving the formation of pure clusters, or involving the modified random network model. Our adaptive model, efficient to reproduce known crystalline and glassy structures, allows us to track the formation of large amorphous Mg-rich Si-poor nanoparticles in an Mg-poor Si-rich matrix from a 0.1MgO-0.9SiO2 melt.