Performance gain and electro-mechanical design optimization of microneedles for wearable sensor systems

Minimally invasive microneedle (MN) is an emerging technology platform for wearable and implantable diagnostics and therapeutics systems. These short MNs offer pain-free insertion and simple operation. Among the MN technologies proposed to enhance interstitial fluid (ISF) extraction, porous and swellable (P-S) hydrogels absorb analyte molecules across the entire lateral surface. Currently, the design, development, and optimization of the MNs rely on empirical, iterative approaches. Based on theory of fluid flow and analyte diffusion through geometrically complex biomimetic systems, here we derive a generalized physics-guided model for P-S MN sensors. The framework (a) quantifies MN extracting efficiency \({\eta _\textrm{PS}}\) in terms of its geometric and physical properties, and (b) suggests strategies to optimize sensor response while satisfying the mechanical constraints related to various skin-types (e.g., mouse, pig, humans, etc.). Our results show that, despite the differences in geometry and composition, P-S MNs obey a universal scaling response, \({\eta }_\textrm{PS} \sim \zeta \left( \frac{\textrm{h}_\textrm{T} \textrm{l}_\textrm{n}^\textrm{2}}{\textrm{D}_\textrm{n}\textrm{s}} \right) ^\textrm{n}\) with \(\textrm{l}_\textrm{n}, \textrm{D}_\textrm{n}, \textrm{s}\) being MN length, diffusivity, and radius, respectively, and \({\zeta }\), \(\textrm{h}_\textrm{T}\) and \(\textrm{n}\) are the ratio between approximate vs. exact analytical solutions, the effective biofluid transfer coefficient between dermis and skin, and the exponent for the power-law approximation, respectively. These parameters quantify the biomolecule transfer through the dermis-to-MN interface at different scaling limits. P-S MNs outperform hollow MNs by a 2-6x enhancement factor; however, the buckling-limit of insertion defines the maximized functionality of the sensor. Our model, validated against experimental results and numerical simulations, offers a predictive design framework to significantly reduce the optimization time for P-S MN-based sensor platforms.