The device at the center of the study is a Circular InterDigital Capacitor Sensor, or CIDCS, a one-port microwave component that doubles as both a resonator and a radiator. Its sensing element is a circular array of interlocking metal fingers, the interdigital capacitor, printed on one side of an FR4 substrate, with a ground conductor on the reverse. Interdigital capacitors are prized in microwave sensing because their tightly spaced fingers confine strong electric fields into a compact footprint, making them acutely responsive to whatever material is brought near them. In this design, a microstrip transmission line feeds the structure electromagnetically, and the whole sensor is tuned to resonate at a well-defined frequency.

The research team, Swati Todi and Poonam Agarwal of the university’s MicroSystems Lab, fabricated two prototypes that differed in exactly one respect. Prototype-I carried a complete metallic ground plane behind the interdigital structure, while Prototype-II had a partial ground plane, leaving the substrate exposed directly beneath the sensing fingers. Every other dimension was held constant, ensuring that any difference in performance could be attributed to the ground configuration alone. The researchers also built a rigorous electromagnetic simulation campaign in CST to quantify field distributions, radiation patterns, directivity, gain, and radiation efficiency for both versions before any liquid ever touched the devices.

The simulations told a story of trade-offs. Removing the ground beneath the IDC pushed the resonant frequency down dramatically, from 6.345 gigahertz to 2.149 gigahertz, while simultaneously raising the electric field strength around the fingers, boosting surface current, and increasing both electric and magnetic energy densities. The physical explanation is intuitive: with a full ground plane present, most of the electromagnetic field is trapped between the patch and the ground, locked inside the dielectric substrate, and only weak fringing fields extend upward toward the sample. Strip that ground away and the fringing fields surge, spilling out of the surface where they can interact with whatever liquid sits above the fingers. The catch was radiative: radiation efficiency collapsed from 26 percent in the fully grounded prototype to just 5 percent in the partially grounded one, and the front-to-back ratio fell from 15.91 to 2.08.

To contain liquid samples during testing, the team bonded small circular wells carved from polydimethylsiloxane, the flexible biocompatible polymer known as PDMS, onto the sensor surface. Cavity size mattered more than one might expect. Because most of the sensing action occurs in the fringing fields between adjacent fingers, the researchers systematically tested cavities with diameters of 4, 6, and 8 millimeters, filling each with carefully measured volumes of glucose dissolved in deionized water at concentrations spanning 0.0 to 0.5 grams per milliliter. The 6-millimeter cavity proved optimal, and it was there that the contrast between the two prototypes reached its peak.

The experimental results were striking. With the 6-millimeter cavity, the partially grounded Prototype-II outperformed its fully grounded sibling by 20.3 percent in sensitivity measured through resonant frequency shifts and by a remarkable 49.41 percent in sensitivity measured through return loss changes. The linearity of the sensor’s response also improved, with the linear regression coefficient rising by 2.26 percent for frequency-based readings and 0.324 percent for return-loss-based readings. As glucose concentration climbed across the tested range, the partially grounded sensor tracked the changes with a consistency that the fully grounded version could not match.

Reliability was put to the test as well. The team repeated their measurements three times for each configuration and computed standard deviations, mean absolute deviations, coefficients of variation, and limits of detection. The coefficients of variation came out at 1.68 percent and 0.16 percent for the fully grounded device and 5.64 percent and 0.26 percent for the partially grounded one, indicating high precision with negligible relative variability. The estimated limits of detection confirmed that both prototypes could resolve low glucose concentrations, with the partially grounded design again holding the edge in raw responsiveness.

The physics behind the improvement can be understood through perturbation theory, the framework that governs how resonant frequencies shift when a dielectric material perturbs an electromagnetic field. The stronger the electric field in the sensing region, the larger the perturbation caused by the sample, and the bigger the resulting frequency shift. Removing the ground conductor beneath the interdigital capacitor simultaneously strengthens the fringing fields, reduces parasitic capacitance between the fingers and the ground, and alters the effective capacitance and inductance of the structure, all of which conspire to amplify the sensor’s reaction to subtle permittivity changes. The fully grounded prototype, by contrast, buries its field energy in the substrate where the sample cannot reach it.

What distinguishes this work from earlier studies is the rigor of the comparison. Partial and defected ground planes have been used liberally across the microwave sensor literature to shrink devices, widen bandwidths, and improve impedance matching, but most reports optimize a single design without quantifying what the ground modification actually bought. Here, both prototypes share identical sensing geometry on identical FR4 substrates, so the measured sensitivity gains are a direct, controlled measurement of the ground plane’s role. The authors note that most prior glucose sensors also rely on FR4, reinforcing the practical relevance of their approach for low-cost implementations.

The broader implication is a design principle rather than a single device. When a one-port microwave structure must serve primarily as a sensor, sacrificing antenna metrics such as directivity, gain, and radiation efficiency is a worthwhile price for a device that responds more sharply to the dielectric fingerprint of the liquid it examines. The researchers point toward future work refining the sensing geometry, exploring alternative sample placement, and extracting additional parameters such as Q-factor, phase, and signal-to-noise ratio for multi-dimensional analysis. For applications like glucose monitoring, where detecting minute variations in permittivity is the whole point, the message is clear: sometimes the ground beneath a sensor is better left partly removed.

Subject of Research: Effect of complete versus partial backside ground conductors on the sensitivity of a microstrip-fed circular interdigital capacitor microwave glucose sensor

Article Title: Sensitivity analysis of the microstrip-fed circular InterDigital capacitor sensor with complete and partial backside ground conductor

Article References: Sensitivity analysis of the microstrip-fed circular InterDigital capacitor sensor with complete and partial backside ground conductor. (n.d.). https://doi.org/10.3389/fanpr.2026.1813732

Image Credits: AI Generated

DOI: 10.3389/fanpr.2026.1813732

Keywords: microwave sensor, interdigital capacitor, partial ground plane, glucose sensing, permittivity, resonant frequency, PDMS microfluidics, FR4 substrate, sensitivity analysis, electromagnetic simulation, biosensor, antenna trade-off

The appeal of sweat as a diagnostic fluid has grown steadily over the past decade. Sweat contains glucose at concentrations that correlate with blood glucose, and it can be harvested passively through a wearable patch during exercise or induced perspiration. The difficulty lies in sensitivity. Glucose levels in sweat are far lower than in blood, so any sensor intended for this matrix must detect minute quantities of the sugar against a background of salts, lactate, urea and other interfering compounds. Conventional enzymatic sensors, which use glucose oxidase to catalyze the reaction, suffer from well-known weaknesses: enzymes degrade with heat, humidity and time, losing activity within days and making long-term wearable use impractical.

Non-enzymatic sensing sidesteps this fragility by letting an inorganic catalyst do the electrochemical work directly. Cobalt oxide, specifically the spinel phase Co3O4, has emerged as a strong candidate because its cobalt ions readily shuttle between oxidation states, mediating the oxidation of glucose at the electrode surface. The catch is that bare cobalt oxide is a relatively poor electrical conductor, which limits how efficiently the electrons generated by glucose oxidation can be collected. This is precisely where the carbon nanotubes enter the picture. Multi-walled carbon nanotubes form a highly conductive, three-dimensional scaffold with enormous surface area, and when cobalt oxide nanoparticles are dispersed uniformly across this network, the two materials complement each other: the oxide supplies catalytic sites while the nanotubes provide an electron highway.

The synthesis reported by the Chennai team is deliberately simple. A hydrothermal route, in which precursor materials react in water inside a sealed vessel at elevated temperature and pressure, was used to grow the nanocomposite. Hydrothermal methods are attractive because they produce crystalline nanoparticles at relatively low cost and without exotic reagents, and the resulting particles can be tuned in size and distribution through reaction conditions. The researchers then modified a graphite electrode with the nanocomposite, creating a working sensor surface. Graphite is inexpensive, widely available and mechanically robust, which matters for any device intended to be mass-produced as a disposable or wearable component.

Characterization formed a substantial part of the study, and the authors deployed an unusually complete analytical arsenal. X-ray diffraction confirmed the crystalline structure of the cobalt oxide spinel phase, while Fourier-transform infrared spectroscopy verified the chemical bonds and functional groups present in the composite. Field-emission scanning electron microscopy revealed the morphology of the material, showing cobalt oxide nanoparticles distributed across the nanotube network, and energy-dispersive spectroscopy confirmed the elemental composition. Brunauer-Emmett-Teller measurements quantified the surface area, a critical parameter for sensing because more accessible surface means more catalytic sites in contact with the analyte. Finally, X-ray photoelectron spectroscopy probed the surface chemistry and oxidation states of the cobalt, confirming that the active species needed for glucose electrocatalysis were present at the interface where sensing actually occurs.

Electrochemical testing was carried out in 0.1 molar phosphate buffer solution at pH 7.4, a standard electrolyte that mimics physiological neutrality. The team used a battery of electroanalytical techniques, each probing a different aspect of sensor behavior. Cyclic voltammetry mapped the redox activity of the electrode as the potential was swept back and forth. Differential pulse voltammetry and linear sweep voltammetry provided higher-resolution measurements of the glucose oxidation signal. Square-wave voltammetry added another sensitive detection mode, while electrochemical impedance spectroscopy measured how readily electrons moved across the electrode-electrolyte interface, a direct indicator of charge-transfer efficiency. Chronoamperometry, in which the current is monitored at a fixed potential over time, served as the workhorse for quantitative detection, since the current step produced when glucose is added scales with its concentration.

The performance figures are striking. The sensor achieved a sensitivity of 18.47 milliamperes per millimolar per square centimeter, a high value for a non-enzymatic system, with a detection limit of 0.1 millimolar. That detection limit is relevant to the glucose concentrations expected in sweat, which typically sit well below blood levels. Equally important for real-world deployment, the electrode demonstrated strong stability, repeatability and reproducibility. After thirty days of storage, the sensor retained more than 96 percent of its original response, a figure that speaks directly to the durability problem that plagues enzymatic alternatives. Reproducibility across multiple electrodes suggests the hydrothermal synthesis is controllable enough to yield consistent devices, a prerequisite for any commercial translation.

Perhaps the most consequential test came when the researchers moved from buffer solution to artificial sweat, a synthetic formulation designed to reproduce the salt content and pH of human perspiration. Real sweat is a hostile medium for electrochemistry: chloride ions, lactic acid, urea and ascorbic acid can all generate signals that masquerade as glucose or poison the catalyst. The MWCNT/Co3O4 electrode recovered 94 percent of the expected glucose signal in this matrix, demonstrating that the composite can function under conditions that approximate the physiological environment a wearable patch would actually encounter. While artificial sweat is not identical to the genuine article, and human trials remain a necessary next step, the recovery result provides meaningful evidence that the sensing chemistry survives outside the pristine laboratory beaker.

The broader context makes the result timely. Diabetes affects hundreds of millions of people worldwide, and the burden of daily finger-stick testing drives much of the noncompliance that leads to poor outcomes. Continuous glucose monitors based on subcutaneous sensors have transformed care for many patients, but they remain invasive, expensive and require regular replacement. A truly noninvasive alternative that reads glucose from sweat could lower the barrier to frequent monitoring dramatically, particularly in resource-limited settings. Non-enzymatic electrodes are especially well suited to this vision because they tolerate temperature swings, long storage and repeated use, all of which are unavoidable in a consumer wearable. The literature on cobalt oxide and carbon nanotube composites for glucose sensing is already substantial, and the present work builds on that foundation while pushing sensitivity and stability figures forward with an unusually accessible fabrication route.

Challenges remain before a sweat-based patch built on this chemistry reaches the clinic or the drugstore. Sweat glucose correlates with blood glucose but the relationship is dynamic and influenced by sweating rate, skin conditions and individual physiology, so calibration strategies and on-body validation will be essential. The researchers tested their electrode on graphite in artificial sweat rather than in a fully integrated flexible device on human subjects, and scaling from a laboratory electrode to a manufacturable, skin-conformal sensor involves additional engineering in fluid handling, electronics and data interpretation. Nevertheless, the combination of a low-cost hydrothermal synthesis, an inexpensive graphite substrate, enzyme-free operation, high sensitivity and thirty-day stability makes the MWCNT/Co3O4 nanocomposite a credible candidate platform for the next generation of noninvasive glucose monitors. As the field of wearable diagnostics accelerates, studies like this one show how careful materials engineering at the nanoscale can translate into practical devices that patients may one day simply wear, sweat into, and read.

Subject of Research: Non-enzymatic electrochemical glucose sensing using an MWCNT/Co3O4 nanocomposite for sweat-based monitoring

Article Title: Facile hydrothermal synthesis of MWCNT/Co3O4 nanocomposite for sweat-based non-enzymatic glucose sensing in artificial sweat

Article References: Perumal, K., & B. S., S. (2026). Facile hydrothermal synthesis of MWCNT/Co3O4 nanocomposite for sweat-based non-enzymatic glucose sensing in artificial sweat. Journal of Materials Science. https://doi.org/10.1007/s10853-026-13873-8

Image Credits: AI Generated

DOI: 10.1007/s10853-026-13873-8

Keywords: glucose sensing, non-enzymatic sensor, MWCNT, cobalt oxide, Co3O4, nanocomposite, hydrothermal synthesis, sweat, wearable biosensor, electrochemistry, noninvasive monitoring, diabetes