Airborne particles are important for public health, visibility, and climate. Predicting their concentrations, effects, and responses to control strategies requires accurate models of their formation and growth in air. This is challenging, as a large fraction is formed by complex reactions of volatile organic compounds, generating secondary organic aerosol (SOA), which grows to sizes important for visibility, climate, and deposition in the lung. Growth of SOA is particularly sensitive to the phase/viscosity of the particles and remains poorly understood. We report studies using a customdesigned impactor with a germanium crystal as the impaction surface to study SOA formed from the ozonolysis of α-pinene at relative humidities (RHs) up to 87% at 297 ± 2 K (which corresponds to a maximum RH of 70-86% inside the impactor). The impaction patterns provide insight into changes in phase/viscosity as a function of RH. Attenuated total reflectance-Fourier transform infrared spectroscopy and aerosol mass spectrometry provide simultaneous information on composition changes with RH. The results show that as the RH at which the SOA is formed increases, there is a decrease in viscosity, accompanied by an increasing contribution from carboxylic acids and a decreasing contribution from higher molecular mass products. In contrast, SOA that is formed dry and subsequently humidified remains solid to high RH. The results of these studies have significant implications for modeling the growth, aging, and ultimately, lifetime of SOA in the atmosphere.aerosol phase | SOA viscosity | oligomers | SOA composition | particle bounce A irborne particles negatively affect health (1) and visibility (2-4), and their effects on climate are substantial (2,3,(5)(6)(7). A large portion of atmospheric particulate matter comprises lowvolatility organics formed during gas phase oxidation of hydrocarbons, including biogenics. This organic contribution, which is typically more than 50% of the total particle mass (8-10), is often underpredicted by models. Although recent measurementmodel comparisons are in closer agreement (11-15), challenges remain because of the complexity of SOA. For example, despite improvements in modeled mass loading, there remain issues when trying to predict specific characteristics such as oxygen-tocarbon ratio and volatility (16) simultaneously.It is vital that models and measurements are in agreement for the right reasons to have robust predictive capability. A potential contributor to the remaining discrepancies could be the assumptions incorporated into the models. Specifically, SOA has been generally considered to be a low-viscosity liquid that equilibrates with the surrounding gases, with no condensed phase diffusion limitations on the scale of the time steps used in the models (17,18). However, there is recent evidence for solid or semisolid phases of SOA, such as much slower evaporation than expected (19,20), SOA composition that is consistent with irreversible condensational growth (21), viscosity measurements (22), time...