The accuracy of liquid-liquid phase transition temperatures determined from semiautomated light scattering measurements

The synthetic-method determination of liquid-liquid coexistence curves using semiautomated light scattering instrumentation and stirred samples is based on identifying the coexistence curve transition temperatures (T-cx) from sudden changes in turbidity associated with droplet formation. Here we use a thorough set of such measurements to evaluate the accuracy of several different analysis methods reported in the literature for assigning T-cx. More than 20 samples each of weakly opalescent isobutyric acid+water and strongly opalescent aniline+hexane were tested with our instrumentation. Transmitted light and scattering intensities at 2 degrees, 24 degrees, and 90 degrees were collected simultaneously as a function of temperature for each stirred sample, and the data were compared with visual observations and light scattering theory. We find that assigning T-cx to the onset of decreased transmitted light or increased 2 degrees scattering has a potential accuracy of 0.01 K or better for many samples. However, the turbidity due to critical opalescence obscures the identification of T-cx from the light scattering data of near-critical stirred samples, and no simple rule of interpretation can be applied regardless of collection geometry. At best, when 90 degrees scattering is collected along with transmitted or 2 degrees data, the accuracy of T-cx is limited to 0.05 K for near-critical samples. Visual determination of Tcx remains the more accurate approach in this case. (C) 2010 American Institute of Physics. [doi:10.1063/1.3469778]