Magnetic forces steer tiny chemical plugs across a chip without channels

A tiny blob of dissolved molecules can now be steered, trapped, or sent circling around obstacles on a microfluidic chip using nothing more than electricity, a magnet, and carefully chosen electrode pairs — no pumps, valves, or physical channels required. That is the central result of a new computational study published in Discover Fluid Mechanics, in which researchers built a three-dimensional, time-resolved model of how a discrete sample plug behaves under redox-magnetohydrodynamics, or R-MHD, microfluidic flow. The work, led by Shirin Hesan and colleagues including Jörg König, Foysal Z. Khan, Christian Cierpka, and Ingrid Fritsch, addresses a question that previous computational studies had left untouched: what actually happens, moment by moment, to a small defined volume of solute as it is pushed through an open chamber by magnetic forces.

The physics behind the technique is elegant in its simplicity. When an ionic current density j flows through an electrolyte in the presence of a magnetic flux density B, the two interact to produce a body force, F = j × B, that drives the fluid along the direction of the resulting vector. In conventional magnetohydrodynamic pumping, this force is harnessed to move bulk fluid. The redox variant adds a chemical twist: rather than relying on dissolved redox species that can generate bubbles and corrode electrodes, the new study models devices whose pumping electrodes are coated with poly(3,4-ethylenedioxythiophene), or PEDOT. This conducting polymer hosts the faradaic charge transfer within its own film, converting electronic current to ionic current without gas formation or electrode degradation, and it allows the ionic current between electrodes to be treated through the bulk electrical properties of the supporting electrolyte rather than by tracking redox-species transport.

That simplification is not merely a computational convenience. Earlier simulations of R-MHD had to resolve the full electrochemistry of dissolved redox mediators, including heterogeneous electron-transfer kinetics at metal electrodes and the coupled diffusion, migration, and convection of redox-active solutes. By eliminating soluble mediators, the new model frees computational resources for the more complex electrode geometries and time-resolved plug dynamics that matter for lab-on-a-chip applications. It also sidesteps electrochemically generated concentration density gradients that would otherwise drive natural convection alongside the desired magnetic flow. The team implemented the coupled physics — magnetic fields, electric currents, laminar flow, and transport of diluted species — in COMSOL Multiphysics, using geometries and conditions that mirror real chip-based R-MHD devices.

The modeled chamber is small by any standard: an enclosed aqueous-electrolyte space measuring 3.0 centimeters by 1.7 centimeters by 429 micrometers in height, defined by a polydimethylsiloxane gasket on a microchip sitting above a 0.37 tesla permanent neodymium magnet. The coplanar parallel band pumping electrodes are roughly 900 micrometers wide, 1.5 centimeters long, and 28 micrometers thick, and the applied current was ±400 microamperes across electrode pairs at two separations, 2760 and 4441 micrometers. Into this chamber the researchers introduced an idealized cylindrical plug of 539 picoliters, with a 20-micrometer radius, an initial solute concentration of 0.10 molar, and a diffusion coefficient of 8.75 × 10⁻¹⁰ square meters per second, representative of small molecules in liquid electrolytes near room temperature.

A crucial strength of the study is that the simulated flow field was not left untested. The team validated it against full three-dimensional, three-component astigmatism particle tracking velocimetry measurements, taken both midway along the pumping path and at the ends of the electrodes. The agreement was close: along the uniform portion of the horizontal velocity profile between electrodes E1 and E2, the simulation gave 165 ± 1 micrometers per second while the experiment gave 167 ± 2 micrometers per second, within one standard deviation of each other. The vertical profiles showed the expected parabolic shape, with maximum speed at the chamber mid-height, reaching 166 micrometers per second for the narrower electrode pair and 141 micrometers per second for the wider pair, the difference arising because electrode spacing alters the current distribution between them.

The model also revealed the intricate three-dimensional flow structure near the electrode ends, where the spatial distribution of ionic current density varies widely in both magnitude and direction. Between the electrodes, the transverse current component dominates and produces a nearly flat, uniform velocity profile across the gap. At the electrode terminations, however, a significant x-directed current density emerges that interacts with the vertical magnetic field to generate a lateral body force, pushing fluid sideways around the electrode edge. The vertical current component, being aligned with the magnetic field, contributes no force at all — a detail that turns out to be central to how plugs behave when they drift over the electrodes themselves.

When the researchers tracked the sample plug through six simulation studies, distinct manipulation strategies emerged. With no pumping applied, the plug simply spread by diffusion, with the simulated broadening matching a simple diffusion-length estimate. With pumping active, the plug’s fate depended strongly on electrode configuration and geometry. Using the outer electrode pair, whose member E1 sits only 1496 micrometers from a chamber wall, the plug was directed primarily toward one electrode’s end rather than splitting. Using the more centrally placed inner pair, the plug traveled more slowly — consistent with the lower current density between those electrodes — and by 300 seconds it began moving equally toward both sidewalls, appearing to split before wrapping around the electrode ends and eventually returning toward the region between the pumping electrodes in a circulating pattern.

Initial plug position proved equally decisive. When the starting point was shifted 444 micrometers toward electrode E2, most of the plug moved toward one wall rather than splitting evenly. When the shift was increased to 1776 micrometers, placing the plug at roughly ten percent of the gap from E2, the flow directed the entire plug along the inner edge of that electrode in a tight turn, hugging the electrode tip and preventing any spread toward the opposite wall. Notably, some solute close enough to the electrode diffused over it into a quiescent zone where the body force is weak, became trapped in a circulating region, and was repeatedly redirected by the stronger forces along the electrode edges — demonstrating that placement alone can either promote mixing or hold a plug in a more stable position.

The physical explanation for this confinement lies in the geometry of the body force itself. Directly over the electrodes, the current is predominantly vertical and nearly parallel to the magnetic field, so the dominant cross-product contribution vanishes and only a smaller transverse component produces a weak force. Solute that drifts into this quiescent region is confined by the surrounding flow. The authors note that to transport rather than confine a plug, the inter-electrode gap should be substantially wider than the plug, with transit time accounted for so the plug clears the region over the electrodes. Diffusion proceeds regardless of the flow field, so confinement is never complete and some spreading is unavoidable. A plug in a low-velocity, high-shear region spreads more rapidly, whereas one in a high-velocity, low-shear region moves faster and disperses less.

The researchers are careful about the limits of what they have shown. The simulated flow fields are experimentally validated, but the predicted plug trajectories, spreading, and confinement have not yet been confirmed by direct observation, which the authors identify as a natural next step. The study is also scoped to a single-species plug without a stationary phase, and future work will address experimental tracking, multi-species transport, and pumping in a loop with a stationary phase immobilized on the chamber walls to enable chemical separations. Even so, because the model uses physical constants and geometries drawn directly from fabricated devices with PEDOT-modified electrodes rather than nondimensionalized parameters, the predicted transport maps onto realizable designs. The authors argue the findings are transferable to other geometries, since outcomes are governed by the local current-density distribution rather than the specific electrode shape — a claim with implications for chemical separations, on-chip mixing and reaction processes, and multiplexed analytical platforms where small, discrete sample volumes must be controlled without enclosed channels.

Subject of Research: Time-resolved modeling of sample plug transport under redox-magnetohydrodynamic microfluidic flow with parallel band pumping electrodes

Article Title: Influence of mass transport near parallel band pumping electrodes and walls on the fate of chemical species in a sample plug introduced onto a redox-magnetohydrodynamics microfluidics chip

Article References: Hesan, S., König, J., Khan, F. Z., Cierpka, C., & Fritsch, I. (2026). Influence of mass transport near parallel band pumping electrodes and walls on the fate of chemical species in a sample plug introduced onto a redox-magnetohydrodynamics microfluidics chip. Discover Fluid Mechanics, 2(1), Article 12. https://doi.org/10.1007/s44369-026-00012-8

Image Credits: AI Generated

DOI: 10.1007/s44369-026-00012-8

Keywords: redox-magnetohydrodynamics, microfluidics, sample plug transport, PEDOT electrodes, Lorentz force, finite element modeling, lab-on-a-chip, particle tracking velocimetry, ionic current density, chemical separations, COMSOL simulation, magnetohydrodynamic pumping