Dielectrocapillarity for exquisite control of fluids

Abstract

Spatially varying electric fields are prevalent throughout nature, such as in nanoporous materials and biological membranes, and technology, e.g, patterned electrodes and van der Waals heterostructures. While uniform fields cause free ions to migrate, for polar fluids they simply reorient the constituent molecules. In contrast, electric field gradients (EFGs) induce a dielectrophoretic force, offering fine control of polar fluids even in the absence of free charges. Despite their vast potential for optimizing fluid behavior under confinement, such as in nanoporous electrodes, nanofluidic devices, and chemical separation materials. EFGs remain largely unexplored at the microscopic level due to the absence of a rigorous first-principles theory of electrostriction. By integrating state-of-the-art advances in liquid state theory and deep learning, we reveal how EFGs modulate fluid structure and capillarity. We demonstrate that dielectrophoretic coupling enables tunable control over the liquid-gas phase transition, capillary condensation, and fluid uptake into porous media. Our findings establish "dielectrocapillarity" -- the use of EFGs to manipulate confined fluids -- as a powerful mechanism for controlling volumetric capacity in nanopores, holding immense potential for energy storage, selective gas separation, and tunable hysteresis in neuromorphic nanofluidics. Furthermore, by linking nanoscale dielectrocapillarity to macroscopic dielectrowetting, we establish a foundation for field-controlled wetting and adsorption phenomena of polar fluids across length scales.

Figures (5)

• Figure 1: Inhomogeneous electric fields arising from interdigitated electrodes strongly influence water's wetting behavior. Snapshot of an SPC/E water simulation in a hydrophobic slit with alternating positive and negative electrode patches. Either holding the electrodes at a 10 V potential difference or attributing a fixed charge of $\pm 0.05\,e/\mathrm{atom}$ causes the fluid to exhibit enhanced wetting at the walls, accompanied by strong lateral density oscillations. Cross-sections of the electrostatic potential in the constant charge setup are shown parallel to the surface (top right) and normal to the surface (bottom left).
• Figure 2: Reorganization of fluids under non-uniform electric fields. An applied electric field, $E^*(z) = E^*_{\rm max}\sin(2\pi z/\lambda)$, shown in (a), induces pronounced density variations in bulk supercritical water, as can clearly be seen from a snapshot of a molecular dynamics simulation. (b) Results from cDFT for the number ($\rho^{*}(z)$, top) and charge ($n^{*}(z)$, bottom) densities capture this behavior. It can clearly be seen that number density is locally depleted where the $|\nabla E|$ is large, and locally enhanced where $|\nabla E|$ is small. The same qualitative behavior is seen in (c) for a supercritical dipolar fluid, except that its response is symmetric, in contrast to water where local depletion depends upon the sign of $\nabla E$. (d) In contrast to both water and the dipolar fluid, an electrolyte is locally depleted in regions of low field strength due to electrophoretic forces (purple and green lines show cation and anion density, respectively, while the blue line shows the total density). Reduced units are described in the Methods section.
• Figure 3: Controlling liquid--vapor equilibrium with EFGs. (a) At $T^* > T^*_{\rm c,0}$ where $T^*_{\rm c,0} \approx 1.97$, increasing EFGs by independently varying $\lambda$ and $E_{\rm max}$ amplifies dielectrophoretic rise. (b) The local electrostrictive response of the dipolar fluid, as measured by the rise in the maximum density peak relative to zero field $\Delta\rho^*_{\rm max}$, is highly non-linear. Solid and dashed lines show the response of systems with dipolar interactions that are long-ranged (LR), e.g., polar molecules, and screened short-ranged (SR), e.g., colloids, respectively. The effect of LR interactions becomes pronounced for $\lambda \gg \sigma$. The SR fluid is a nearly identical dipolar fluid, but whose Coulomb potential is replaced by $\mathrm{erfc}(\kappa r)/r$ where $\kappa^{-1}=1.5\,\sigma$. (c) At an isotherm where $T^* < T^*_{\rm c,0}$, stable solutions for the density $\rho^\ast(z)$ under a sinusoidal electric field with $\lambda/\sigma = 1.7$, are shown for different values of the chemical potential. These results are used to investigate liquid--vapor coexistence. (d) Results in light pink show the binodal of the dipolar fluid in the absence of an electric field. At $E^*_{\rm max} = 4$, $T^*_{\rm c}$ shifts to a lower temperature, as seen in the binodal in dark purple. Solid symbols show results obtained from the multiscale cDFT approach, while crosses indicate estimates of $T^*_{\rm c}$ using the law of rectilinear diameters and critical exponents rowlinson2002book. Solid lines serve as a guide to the eye.
• Figure 4: Control of fluid uptake by dielectrocapillarity. (a) Schematic of a fluid in a slit pore, with $H/\sigma = 6.6$, in equilibrium with a reservoir at chemical potential $\mu=\mu_{\rm co,0}+\Delta\mu$. (b) Adsorption/desorption isotherms at $T^*/T^*_{\rm c,0} = 0.68$ obtained by varying $\mu$ for different $E_{\rm max}$ at fixed $\lambda/\sigma = 1.7$. Larger $E_{\rm max}$ promotes adsorption, while simultaneously decreasing hysteresis. For large enough $E^*_{\rm max}$, the transition becomes continuous. (c) Adsorption/desorption isotherms at $T^*/T^*_{\rm c,0} = 0.68$ obtained by varying $E_{\rm max}$ at fixed $\lambda/\sigma = 1.7$ for different $\Delta\mu$. Changing $E_{\rm max}$ can switch the pore between filled and empty states. The vertical dashed lines indicate the equilibrium transition, i.e., where both adsorbed "liquid" and "gas" states are stable.
• Figure 5: Connecting to dielectrowetting experiments. The electrostatic potential from interdigitated electrodes, shown schematically in (a), decays exponentially. Applying this potential symmetrically from both confining walls in a slit geometry with $H\approx 16\,\sigma$ enhances wetting of the solid-liquid interface, as can be seen in the density profiles (left) and changes in local compressibility (right) in (b) (both quantities are normalized by their bulk values).