Thermochimica Acta 512 (2011) 268–272 Contents lists available at ScienceDirect Thermochimica Acta journal homepage: www.elsevier.com/locate/tca How reliable are extrapolations to infinite dilution of partial molar properties using Redlich–Kister fittings? Ângela F.S. Santos, Isabel M.S. Lampreia ∗ Departamento de Química e Bioquímica, Centro de Ciências Moleculares e Materiais, Faculdade de Ciências, Universidade de Lisboa, P-1749-016 Lisboa, Portugal article info Article history: Received 14 September 2010 Received in revised form 20 October 2010 Accepted 3 November 2010 Available online 12 November 2010 Keywords: Partial molar properties at infinite dilution Redlich–Kister equation Extrapolation methods Binary systems Uncertainties abstract An explicit method for uncertainty estimation associated with excess partial molar properties at infinite dilution calculated using standard uncertainties of Redlich–Kister parameters, fitted to data over the whole composition range, is presented for the first time. The application of this method to sets of accurate volumetric experimental data for aqueous water + ethanol binary mixtures, experimentally determined by different authors is made and the results are compared. A refinement of this method is also presented when, in very-diluted regions, Redlich–Kister lines deviate from experimental points more than their standard uncertainties. The procedure of uncertainty calculation is based on the Guide to the Expression of Uncertainty in Measurements (GUM). © 2010 Elsevier B.V. All rights reserved. 1. Introduction Reporting measurement uncertainty is fundamental in engi- neering and experimental sciences, principally when a measure- ment is intended to demonstrate an aspect of a scientific theory, or an innovative or improved method of measurement. The detection, description and analysis of errors are also goals to be achieved with the uncertainty evaluation. In addition, in some cases a preliminary uncertainty propagation analysis permits to identify the most sig- nificant contributions to the combined uncertainty of a property under measurement. From the anticipation of those contributions, we can gather important information on how the experiments might be developed or improved. The uncertainty associated with a result is also a measure of the degree to which that value is expected to agree with similar experimental determinations and is becom- ing a formal requirement for authors to report their measurements in some well-reputed journals. Hence, authors are coming under increasing pressure to report uncertainties as a means to demon- strate the quality of their results. In the last decades limiting partial molar values of different thermodynamic properties such as volume, isobaric expansion, isentropic compression and isobaric heat capacity have been obtained and tabulated. The efforts of many authors in estimat- ing those properties for homologous series of compounds and for compounds differing in polar head groups and branching [1–6] aim ∗ Corresponding author. Tel.: +351 217500995; fax: +351 217500088. E-mail address: milampreia@fc.ul.pt (I.M.S. Lampreia). at a better understanding of solute–solvent interactions on the one hand, and at the application of group contribution methods on the other hand [7–11]. The latter purpose permits to predict values for further compounds using the convenient additive feature of these properties. Two methods have been identified in the literature as prefer- entially used to reach the aforementioned goals. One of them, the classical method, applied to the very-diluted composition range is based on the extrapolation to infinite dilution of apparent molar values of the property of interest and the other makes use of model fitting coefficients, conveniently applied to the respective excess molar property, over the whole composition range, as is the case of the well-known Redlich–Kister (R–K) method in either of its two faces [12]. The calculation procedure to obtain uncertainties associated with limiting partial molar properties derived by the classical method is well established being mostly based on the standard uncertainty associated with the intercept which is statistically obtained from polynomial least-squares fitting of apparent molar properties as a function of molality [13–15]. Conversely some authors, including our group, have been using R–K expansions to obtain limiting partial molar values without presenting a clear explanation of the uncertainty claimed, either in implicit or in explicit form [16–19]. In this work we illustrate for the first time a reliable method of evaluating these standard uncertainties based on the standard uncertainties of the fitted R–K coefficients. This procedure brings out the possibility of data comparison and further adequate use on group contribution schemes. The formalism and methods applied were those outlined in the Guide to the Expression 0040-6031/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.tca.2010.11.010