
Speeding up precise diagnosis and reducing the cost of drug discovery and delivery are crucial for self-driving laboratories, where tasks often involve characterizing biomolecular interactions and optimizing stochastic gradient descent for deep neural networks. During the pandemic, the demand for high-throughput characterization has witnessed a remarkable upswing in micro-scale thermophoresis. However, the lack of a theoretical description of this non-equilibrium transport phenomenon in aqueous media remains a major obstacle that hampers its biological application scenarios.
In this work, we propose a holistic model that reveals multi-scale coupling mechanisms in aqueous systems, where the coupling is governed by Onsager's reciprocal relations. By performing the model-coupling analysis, we found that:
The temperature dependence of the Soret coefficient stems from the competition between long-range bulk effects and short-range interfacial effects.
The size dependence of the Soret coefficient relies on the volume-surface area scaling of the short-ranged interactions, which is rather system-specific and depends on the details of the particle surface chemistry.
An intricate competition between electrostatic and hydration entropy interactions determines the Soret coefficient of nano-sized polystyrene beads, while further shrinking down the size, the Seebeck and viscosity contributions become pronounced for T4 lysozyme suspensions.
These findings enhance our understanding of different interactions at the molecular level, which is relevant for point-of-care diagnostics, binding affinity determination, purification, protein-folding studies, and drug delivery. Last but not least, this type of non-equilibrium transport phenomenon provides nice playgrounds for deep learning, where neural networks can be visualized in the context of the energy landscape.
Publications: Pu D, Panahi A, Natale G, Benneker A. A Mode-Coupling Model of Colloid Thermophoresis in Aqueous Systems: Temperature and Size Dependencies of the Soret Coefficient. Nano Letters. 2024. https://doi.org/10.1021/acs.nanolett.3c04861
Comments