Analysis of Nonuniformity of Methanol-Water Mixture in Sub- and Supercritical State Via MD Simulation

Analysis of Nonuniformity of Methanol-Water Mixture in Sub- and Supercritical State via MD Simulation

 

Takumi Ono, Masaki Ota, Yoshiyuki Sato and Hiroshi Inomata

Research Center of Supercritical Fluid Technology, Tohoku University, Sendai, Japan

Abstract

  Alcohol-water mixtures exhibit interesting anomalies in their macroscopic properties.These features arise from inhomogeneous mixture caused by hydrogen bonding and this factor is influenced by temperature and pressure effects.  We carried out molecular dynamics simulation to evaluate the nonuniformity from the macroscopic point of view.  The remarkable nonuniformities of mixture were characterized by the distribution of number density of molecules in local domain at low methanol composition at 300 °C or less.

 

Introduction

  Alcohol-water mixtures are scientifically intriguing and important for many fields ranging from basic molecular research to widespread industrial.  Simple alcohol such as methanol or ethanol and water mix incompletely, which is a phenomenon described in terms of negative excess entropy [1, 2].  This well-known effect has been attributed to hydrophobic headgroups creating ice-like or clathrate-like structures in the surrounding water.  Related to the inhomogeneous mixing, furthermore, distinct solvation regimes as a function of alcohol concentration have been experimentally reported [3].  This indicates that alcohol-water mixtures have nonuniformity at the molecular level.  The fact that even sub- and supercritical water the hydrogen bonding still exists is reported [4, 5].  It is expected that inhomogeneous structure attributed by the hydrogen bonding is also influenced by temperature and pressure effects.  For the purpose of solution structure analysis of methanol-water mixture and effect of temperature and pressure, some physical properties of alcohol-water mixture were measured at high temperature and pressure including sub- and supercritical condition [6-8].  Since the hydrogen bonding structures affect the density, our group has measured the density of methanol-water mixture and performed molecular dynamics simulation for understanding of the relationship between the macro and micro properties at 300 ~ 400 °C, high pressure condition [9].  However, still no consensus exits on the precise microscopic picture at the molecular level.  In particular, the study of degree of nonuniformity in mixture is scarce.  In this work, we calculated the distribution of molecular arrangement of methanol-water mixtures in ambient, sub- and supercritical states using molecular dynamics simulation to discuss the nonuniformity at the molecular level.

Computational details

  The molecular models used in this study were the 3-sites flexible models proposed by Liew et al. [10] for water and by Honma et al. [11] for methanol.  Both models can represent the PVT and saturated properties of the fluid in the liquid and supercritical state.  The three sites for methanol are oxygen (O), hydrogen (H) and methyl group (-CH3).  Both flexible models were formulated as a sum of intra- and intermolecular potential by using the angular form of the Toukan-Rahman (TR) potential [12].  The intermolecular potential used a Lennard-Jones (LJ) 12-6 potential and a Coulombic potential.  Unlike interactions were computed using the Lorentz-Berthelot combination rule.

                                                                                        (1)

                                                                                                  (2)

The LJ and Coulombic potential parameters for water and methanol were those given in the original studies [10, 11].  All simulations were performed with NVT-ensembles for several methanol mole fractions.  The conditions in the simulation were set in the 25, 300, 400 °C and the number density N / V was set to the experimentally determined densities (0.1, 25MPa) [9, 13] to allow isobaric analysis at any given temperature.  The total number of molecules was 2048.  The temperature was initially (~ 60 ps) controlled by momentum scaling and afterwards (60 ps ~) by the Nosè-Hoover thermostat for computational stability.  The equation of motion was integrated using the velocity Verlet technique.  Time steps were 1 fs for the intermolecular motion and 0.2 fs for the intramolecular motion.  Statistical sampling in the simulation was done for 600 ps after 100 ps equilibration.

Definitions of Nonuniformity

  We divided into the computational cell to cubes of X Å and counted the number density n of methanol and water molecules of each divided cells to evaluate the nonuniformity of mixture. Figure 1 shows the schematic diagram of the evaluation of nonuniformity for methanol-water mixture.  We performed above sampling at 50 fs intervals and calculated the standard deviation of the n of the molecules in the divided cells to evaluate the nonuniformity of the methanol-water mixtures.

Figure 1 Schematic diagram of the evaluation of nonuniformity for methanol-water mixture.

 

Results and discussions

  Figure 2 and 3 show n for each molecules in bulk and standard deviation of n for each molecules as a function of methanol composition in a 8 Å divided cubic cell (X = 8Å).  The standard deviations of each component at 25 °C - 0.1 MPa and water molecule at

Figure 2 X_m dependence of the number density n of for methanol (open symbols) and water (filled symbols) in bulk: 25 °C-0.1MPa (diamonds),   300°C-25MPa (triangles), 400 °C-25MPa (circles).

Figure 3 X_m dependence of the standard deviation s of number density n distribution for water and methanol at (a) 25 °C-0.1MPa,  (b) 300°C-25MPa, (c) 400 °C-25MPa: methanol (open diamond) and water (filled diamonds).

300 °C - 25 MPa have maximum around methanol composition X_m = 0.2 ~ 0.3 although n value in bulk monotonously varied with composition.  However this maximum was not observed in methanol at 300 °C - 25 MPa and both molecules at 400 °C - 25 MPa.  The fact that effects of hydrophobic hydration, so-called "ice-like" or "clathrate-like", are remarkable in low X_m has been reported [14], hence this maximum suggests that inhomogeneous spatial distributions of the molecules caused by the inhomogeneous solution structure exist from the microscopic point of view.  In addition, even under high temperature and pressure, such as 300 °C - 25 MPa, the inhomogeneous local structures proved to be remaining.  On the other hand, we assumed that these non-uniform local structures cannot be maintained at 400 °C - 25 MPa because of intense thermal motion of molecules.

References

[1] F. Franks, D.J.G. Ives, Quarterly Reviews, 20 (1966) 1-&.

[2] F. Franks, J.E. Desnoyers, Water Science Reviews, 1 (1985) 171-232.

[3] T. Sato, A. Chiba, R. Nozaki, J. Chem. Phys., 112 (2000) 2924-2932.

[4] Y. Ikushima, K. Hatakeda, N. Saito, M. Arai, J. Chem. Phys., 108 (1998) 5855-5860.

[5] M.M. Hoffmann, M.S. Conradi, J. Am. Chem. Soc., 119 (1997) 3811-3817.

[6] E. Bulemela, P. Tremaine, S. Kawa, Fluid Phase Equilib., 245 (2006) 125-133.

[7] C.B. Xiao, H. Bianchi, P.R. Tremaine, Journal of Chemical Thermodynamics, 29 (1997) 261-286.

[8] D.T. Safarov, A.N. Shakhverdiev, High Temp., 39 (2001) 395-400.

[9] T. Ono, S. Kobayashi, T. Hoshina, Y. Sato, H. Inomata, Fluid Phase Equilib., 302 (2011) 55-59.

[10] T. Honma, C.C. Liew, H. Inomata, K. Arai, J. Phys. Chem. A, 107 (2003) 3960-3965.

[11] T. Honma, S. Kuzuhara, C.C. Liew, H. Inomata, Fluid Phase Equilib., 194 (2002) 271-280.

[12] K. Toukan, A. Rahman, Physical Review B, 31 (1985) 2643-2648.

[13] M. Iglesias, B. Orge, J. Tojo, Fluid Phase Equilib., 126 (1996) 203-223.

[14] M. Kiselev, S. Noskov, Y. Puthovski, T. Kerdcharoen, S. Hannongbua, Journal of Molecular Graphics & Modelling, 19 (2001) 412-416.