Finally, it is important to point out that while the Kleinhans and Mazur freezing point summation model defines the number of solute-specific coefficients to be used for each solute (three), the osmotic virial equation does not. In principle, it is possible to fit the osmotic virial equation to osmometric data with any number of osmotic virial coefficients, regardless of solute, and the fit should improve, even if only slightly, with
each added coefficient. Screening Library solubility dmso However, the model fit converges quickly (recall that the osmotic virial coefficients represent increasing orders of interactions between solute molecules), with each added coefficient contributing progressively less to the accuracy of the fit. Indeed, previous studies [14] and [55] have shown that for most solutes,
the second osmotic virial coefficient is sufficient to accurately capture non-ideal solution behavior, although some particularly non-ideal solutes such as proteins require a third osmotic virial coefficient [55]. Navitoclax cell line Furthermore, as noted by Prausnitz et al. [53], excessive coefficients (i.e. overfitting) may actually lead to a loss of accuracy when predicting the thermodynamic behavior of more complex, multi-solute solutions, due to the corresponding need for a greater number of mixing rules, each of which may have some uncertainty associated with it arising from assumptions made in its development. For these reasons, when curve-fitting the osmotic virial equation, the number of coefficients used (i.e. the order of the fit) should be limited to the minimum that gives an adequate fit. Pricket et al. [55] defined and applied a criterion based on the adjusted R2 statistic for determining the adequate order of fit for the osmotic virial equation. Liothyronine Sodium However, this criterion did not account for the fact that the osmotic virial equation must pass through the origin (i.e. the osmolality of pure water is zero). Furthermore, there exist other criteria that are
appropriate for establishing the order of fit. In this work, two criteria were applied to determine the number of osmotic virial coefficients required for both the molality- and mole fraction-based osmotic virial equations: the adjusted R2 statistic, taking into account regression through the origin, and confidence intervals on the osmotic virial coefficients. In summary, the specific objectives of this work are threefold. First, to provide revised osmotic virial coefficients for the molality- and mole fraction-based multi-solute osmotic virial equations for solutes of interest to cryobiology, using the relationship between osmolality and osmole fraction defined through water chemical potential and an improved and extended set of criteria for selecting the order of fit. Second, to provide coefficients for the freezing point summation model for all the solutes considered in the first objective using the same data sets.