Generation of Recommendable Values for the Surface Tension of Water Using a Nonparametric Regression

Recommended values for the surface tension of water are presented, generated based on all the experimental data available from literature using nonparametric regression. A procedure was used which consists of neither a single data selection nor a determination of a single correlation function. The procedure works without the need for uncertainties in the processed data and other factors that characterize the experimental data credibility, while the processed collection of data sets only needs to be sufficiently large. The equation γ = aτ^c(1 – bτ) with a = 233.58 mN·m^–1, b = 0.61594, c = 1.2527, and τ = 1 – T/647.096 closely approximates the tabulated recommended surface tension values. New experimental data for air–liquid water interfacial tension are presented, measured at atmospheric pressure at 33 temperatures in the temperature range from 272.6 to 343.5 K. The measurements were performed by the Wilhelmy plate method using a Krűss K100Mk2 tensiometer. The uncertainty at 0.95 confidence level of the presented data is estimated to be 0.03 mN·m^–1 and their root-mean-square deviation from the recommended data is 0.027 mN·m^–1. The obtained results substantially extend the database of high-quality data for the air–liquid surface tension of water available from the literature. The worst-case error level of the proposed recommended values is estimated to be 0.1 mN·m^–1.