Aggregation, Not Micellization: Perfluorooctanoic Acid, Perfluorobutanesulfonic Acid, and Potassium Perfluorooctanesulfonate Behavior in Aqueous Solution.

Surface tension, conductivity, and dynamic light scattering (DLS) measurements were used to examine the surface and bulk solution behaviors of three members of the PFAS family, perfluorooctanoic acid (PFOA), perfluorobutanesulfonic acid (PFBS), and the potassium salt of perfluorooctanesulfonic acid (PFOS). Measurements were carried out in solutions having variable (acidic) pH and in solutions buffered to pH = 8.0. Surface tension data show traditional soluble surfactant behavior, and results illustrate that PFOA, PFBS, and PFOS surface activity depends sensitively on solution phase pH. The tightly packed monolayers formed by PFOA in mildly acidic solutions imply that the surface pH of PFOA solutions is several units lower than bulk. Results from conductivity experiments generally show increasing conductivity with increasing bulk solution surfactant concentration. In pH = 8.0 solutions, changes in conductivity slope with surfactant concentration suggest the onset of micelle formation at concentrations <1 mM, markedly lower than reported in literature. In general, apparent critical micelle concentrations (CMCs) determined from conductivity data agree with similar predictions made from surface tension results. DLS measurements show that at concentrations close to the predicted PFAS CMCs, objects with diameters ≤10 nm start to form. However, unlike micelles, these objects continue to grow with increasing bulk solute concentration. These aggregates form structures having diameters of 50-150 nm. Aggregate size shrinks modestly as solution phase temperature increases, and this behavior is reversible. Cryo-EM images of PFOA solutions confirm a broad distribution of particles, supporting the DLS measurements. Findings reported in this work represent the first evidence that these three EPA-regulated PFAS surfactants form aggregates rather than micelles in solution. Findings also begin to reconcile differences in reported surface behaviors that have led to CMC predictions in the literature varying by more than an order of magnitude.