Transferable intermolecular potentials for carboxylic acids and their phase behavior

Abstract

in Wiley InterScience (www.interscience.wiley.com).Transferable step potentials are characterized for 39 carboxylic acids. The reference potential is treated with discontinuous molecular dynamics, including detailed molecular structure. Thermodynamic perturbation theory is used to interpret the simulation results and to provide an efficient basis for molecular modeling and characterization of the attractive forces. Four steps are used for representation of the attractive forces with only the first and last steps varied independently. The two middle steps are interpolated such that each site type is characterized by three parameters: the diameter, r, the depth of the inner well, e 1 , and the depth of the outer well, e 4 . The depths of the attractive wells are optimized to fit experimental vapor pressure and liquid density data. Generally, the vapor pressure is correlated to an overall 43% average absolute deviation (% AAD) and the liquid density to 5% AAD. The deviations tend to be largest for the higher molecular weight acids. These deviations are larger than the errors previously encountered in characterizing organic compounds, but carboxylic acids present exceptional challenges owing to their peculiar dimerization behavior. Simultaneous correlation of vapor pressure, vapor compressibility factor, and phase equilibria of water þ carboxylic acids place several constraints on the nature of the potential model, with the parameters of the present model representing a reasonable tradeoff. In other words, our model represents minimal deviations for vapor pressure, vapor compressibility factor, and phase equilibria of all acids simultaneously while varying the parameters r, e 1 , e 4 , e CC (dimerizing site bonding energy), e AD (acceptor-donor bonding energy), and K HB (hydrogen bonding volume) for the acid O¼ ¼ and OH site types. The present model is characterized by one acceptor and one dimerizing site on the carbonyl oxygen and one acceptor and one donor site on the hydroxyl oxygen. The acceptor and donor are capable of interacting with water while the dimerizing site is not. With this model, the saturated vapor compressibility factor of acids with seven or fewer carbons is near 0.5 while higher carbon ratios lead to a compressibility factor approaching 1.0. To compensate for the high vapor pressure deviations of the transferable potential model, a correction is introduced to customize the molecule-molecule self interaction energy. This adaptation results in deviations of 3.1% for vapor pressure of the pure acid database. To validate the behavior of the model for carboxylic acids in mixtures, 33 binary solutions were considered. Acids in this database ranged from formic to hexadecanoic. The average absolute deviation in bubble pressure for Additional Supporting Information may be found in the online version of this article.Correspondence concerning this article should be addressed to J. R. Elliott at dickelliott@uakron.edu aqueous acid systems is 4.4%, 10.5% for acid þ acid systems, and 4.7% for acid þ n-alkane s...

Dimerization of Carboxylic Acids: An Equation of State Approach

“…Here it is worth mentioning that such low molecular weight acids present a large deviation form ideal gas behavior (Z significantly deviates from unity) due to strong dimerization even at low pressures and even below 1 atm. 3 It is worth mentioning that the heat of vaporization and the compressibility factors are not satisfactorily described if the formation of dimers is not accounted for, in accordance with previous literature studies. 3,29 This is clearly shown in Figure 5, where the NRHB calculations are compared with the literature data for two cases: modeling acetic acid assuming a hydrogen bonding behavior similar to that of alcohols (one proton donor and one proton acceptor on every molecule that are able to cross associate and form oligomers), which is a frequently used approach 25 , and restricting the formation of hydrogen bonds to the formation of cyclic dimers.…”

“…For this reason, the use of experimental data for other properties, such as enthalpies of vaporization and vapor phase compressibility factors has been suggested. 3,24,29 However, for all hydrogen bonding fluids, such difficulty mainly arises from the fact that the parameters that characterize the strength of the various intermolecular interactions (physical or hydrogen bonding), are highly correlated to each other. 38,39 Consequently, the estimation of the association (dimerization) parameters (E dm and S dm ) for carboxylic acids is probably the most important step towards the parametrization of such fluids.…”

“…The cooperativity in the establishment of the second bond results in the formation of stable dimers and, thus, the formation of linear oligomers is not favored. 3,4,5 Accounting for hydrogen bonding through an equation of state approach is usually performed using thermodynamic models of two broad families, i.e. models that are based on the lattice fluid theory [6][7][8][9][10][11][12] , and all the variants of SAFT (Statistical Associating Fluid Theory) equation of state [13][14][15] , which are based on the perturbation theory of Wertheim.…”

“…[27][28][29][30] Most often, in order to test the validity of a model, its predictions are compared solely to phase equilibrium data. 3,[20][21][22][23][24][25][26] Such kind of predictions are very useful in designing applications, where models are applied to predict the phase behavior of multicomponent mixtures. However, in some cases, the accurate prediction of thermal properties, such as the mixing enthalpies, is also important for a successful prediction of the energy demands.…”

DynamO: a free ${\cal O}$(N) general event‐driven molecular dynamics simulator

“…Indeed, Alder et al [26] reported MD simulations for systems of hard spheres using an event-driven algorithm in 1957, 7 years before Rahman's soft potential simulations. Since then, many discrete potential force fields have been developed for a wide range of systems, including granular materials, [27] simple molecular systems (e.g., SPEADMD [28] ) such as mixtures of hydrocarbons, [28] ethers, [28] alcohols, [29] amines, [29] and carboxylic acids, [30] block copolymer micelles [31] and organized mesophases, [32] and detailed models (e.g., PRIME [33] ) for polypeptide [34,35] and protein [36] solutions.These models not only offer qualitative insight to these systems but also provide quantitative predictions for such properties as vapor-liquid phase equilibrium. [28] In addition, the models are detailed enough to realistically capture protein structure, but sufficiently efficient to examine folding and aggregation.…”