Can a Self-Referenced Resonator Make Dk Measurement More Reliable?

Resonant methods can resolve very small frequency shifts. Deciding what caused the shift is harder. A dielectric sample changes the electric field around a resonator, but so can the way the sample sits on the fixture. Thickness, a narrow air gap, a warmer lab bench or a disturbed connection may alter the same trace the engineer uses to calculate Dk.

That is the attraction of a self-referenced dual-mode resonator: two resonances are measured in one physical structure, and their different responses can help separate material loading from some shared disturbances. A 2026 study by Bijari and Nooki demonstrated a compact version of this idea. The more demanding question for RF material testing is how far that evidence goes—and what remains to be measured before a Dk result can be trusted in a stackup calculation.

A Frequency Shift Is a Measurement, Not Yet a Dk Value

In a resonant dielectric measurement, the material under test perturbs the resonator's electromagnetic field. The resulting change in resonance is interpreted through an analytical model, a full-wave model or a calibration against known materials. When dielectric loss is required, the quality factor may also be used, provided conductor, coupling and fixture contributions are treated separately. [2, 3]

Consider what happens between two measurements. A slightly different clamp force changes the sample-to-sensor contact. A thin coupon may bow, leaving a local air gap. The measured thickness may be an average rather than the thickness within the field-sensitive region. Temperature affects both the sample and the resonant structure. Peak fitting can be precise while any of these effects shifts the answer.

For a material engineer, the job is therefore twofold: first identify the electromagnetic response, then establish that the model attributes it to the material rather than the measurement arrangement. The second task is where reference structures, fixture design and uncertainty analysis earn their place.

One Resonator, Two Modes: How the Reference Works

Abolfazl Bijari and Sara Nooki's 2026 device uses a stub-loaded double split-ring resonator in a defected ground structure (DGS). Both resonant modes belong to the same fabricated device, rather than to two nominally identical sensors placed side by side. [1]

The lower-frequency mode interacts comparatively weakly with the dielectric sample and serves as a reference. The higher-frequency mode puts more of its electric field into the sensing region and responds more strongly to the sample. Comparing their frequencies gives the designer another observable, potentially less dependent on a perturbation that shifts both modes in nearly the same way.

The qualification is in that last phrase. The modes do not have identical fields, and a weakly sensitive reference mode is not a sample-independent clock. Temperature, deformation and loading can affect each mode differently. Whether the frequency difference is more stable than a single peak has to be established for the particular disturbance under study.

What the 2026 Experiment Establishes

Published in Scientific Reports on 15 October 2026, the study reports a fabricated prototype on Rogers RO4003 and measurements on several solid dielectric samples. The authors report 1.41% normalized sensitivity, an R² of 0.9814 for their fitted response, and agreement with full-wave simulation. [1] These are useful results for a proof of concept, but they describe different properties of the method.

The distinction is important in practice. A calibration model may fit the tested samples closely and still produce a biased Dk for an unfamiliar specimen. A strong frequency response to dielectric loading is valuable, but the same strong field interaction can make seating and thickness harder to ignore. Sensitivity and calibration fit should be reported alongside the conditions under which they were obtained.

Temperature Drift: What Would Common-Mode Rejection Actually Buy?

Imagine following both resonance frequencies while the sensor warms. If each shifts by almost the same amount, their difference will move less than either peak alone. If their temperature coefficients differ, the differential signal will drift too. This is the physical basis for self-referencing; it also tells us exactly how to test the claim: record both modes over the same controlled temperature sequence and compare the raw and differential drift.

The accessible published abstract of Bijari and Nooki [1] does not give such a temperature-sweep result or a measured drift-reduction factor. The sensitivity and fit statistics cannot fill that gap. Temperature compensation has been studied in a separate dual-mode triangular-resonator design, but that experiment belongs to a different sensor architecture. [4]

Reseating the Sample Is a Different Repeatability Test

There is an easy way to make a resonator measurement look repeatable: leave the sample in place and run the sweep again. That checks short-term instrument and fitting variation. It says little about what happens when the coupon is removed, turned over, cleaned, repositioned or clamped a second time.

For sheet materials, a meaningful repeat-mount study should include independent unload-and-reload cycles. Keep track of orientation, seating force or fixture setting, specimen thickness and any visible bow. Report the spread of the extracted Dk, not only the scatter of the resonance frequency with an untouched sample. The two distributions answer different questions.

The published abstract for the 2026 DGS study does not provide an independent remounting distribution. That limits what can be concluded about day-to-day material acceptance testing. Prior dielectric-characterization work provides the broader context: fixture geometry and loading procedure are part of the measurement, especially when the electric field is concentrated near a specimen boundary. [3]

Thickness and Air Gaps: Errors a Reference Mode May Leave Behind

A common drift can sometimes be subtracted. A change that perturbs one field more strongly than the other is another matter. In a planar DGS sensor, a thin air gap above the sensitive region may change the high-frequency mode appreciably while making only a small change to the reference mode. The differential result then contains a fixture error, despite the presence of an internal reference.

Thickness enters the model in a similar way. The resonator samples a finite volume of dielectric, not the part number printed on a laminate datasheet. Field penetration, the coupon's lateral coverage and the actual local thickness determine how the material loads the modes. Depending on the geometry, a fixture dielectric, surface bow or a gap may matter as much as the uncertainty in the nominal Dk. [3, 5]

These effects must be assessed for the method being used. A planar DGS sensor and a split-post dielectric resonator (SPDR) distribute their fields differently and follow different fixture assumptions. An air-gap result from one arrangement cannot simply be transferred as a correction factor to the other. The relevant geometry belongs in the field model and the uncertainty budget. [5]

Sensitivity, Stability and Accuracy Need Separate Evidence

The 1.41% figure describes the authors' normalized sensitivity, not a 1.41% error band on Dk. Likewise, R² = 0.9814 characterizes the calibration fit; it does not say how accurately an unknown sample will be measured. If this method is used for supplier qualification or lot acceptance, the reader needs the reference-material comparison, repeatability, environmental stability and uncertainty—not just the most attractive statistic in the abstract.

Where the Dual-Mode Sensor Fits Among Established Methods

A compact planar sensor is appealing when access, cost or integration makes a larger fixture inconvenient. That does not automatically make it the right reference for every dielectric specification. IPC-TM-650 2.5.5.15, for example, describes split-post dielectric resonator (SPDR) measurements of base-material Dk and loss tangent from 1 to 20 GHz under its defined test conditions. Broadband methods and fabricated transmission-line methods answer different material and circuit questions. [3, 5]

The most informative comparison would put the same conditioned specimens through the dual-mode sensor and an established reference method at overlapping frequencies. It would include thickness measurements, independent sample placements, bias and scatter across the sample set, and a stated uncertainty budget. Until those data are available, the 2026 DGS result is best described as an experimentally demonstrated sensing architecture—not a replacement accuracy standard. [3, 5]

What a Laboratory Would Still Need to Validate

• Define the material state and quantity. Specify the frequency, specimen conditioning, actual thickness and orientation. State whether the deliverable is Dk only or Dk and Df.
• Challenge the fixture. Repeat independent removal and reseating cycles, vary orientation or clamping condition within a defined range, and track the extracted Dk distribution.
• Measure drift instead of assuming cancellation. Record both resonances across a controlled temperature/time sequence, then compare compensated and uncompensated results.
• Test the geometry model. Deliberately vary sample thickness, lateral position and local standoff. Quantify how those changes propagate into the inferred permittivity.
• Establish an external reference and fabrication spread. Compare characterized materials with an independent method, report bias and uncertainty, and repeat the study on separately built sensors rather than one prototype.

These are not extra tests for the sake of a longer qualification report. Each isolates a different error source. Together they show whether a promising resonator response can be converted into a reliable material decision in another laboratory or another production week.

What This Means for RF PCB Material Specifications

The practical lesson extends beyond this sensor. A Dk number used in an RF stackup should arrive with its measurement method, frequency, specimen condition, thickness and uncertainty. A supplier's material Dk, a design Dk inferred from a transmission line, and the effective permittivity of a finished microstrip are related quantities, but they are not interchangeable labels. [3]

That distinction changes the conversation at design release. Is the value being used to approve a laminate lot, tune a transmission-line model or verify a finished board? Each purpose calls for different evidence. UltroNiu's Rogers PCB manufacturing and RF PCB manufacturing pages describe the manufacturing context; the dual-mode sensor discussed here is an independent research development, not an UltroNiu-qualified measurement capability.

For an RF PCB project with tight dielectric or impedance requirements, the useful starting package is the laminate specification, operating frequency, proposed Dk/Df test method, stackup and critical RF geometry. These can be reviewed together through an Engineering Review before manufacturing assumptions are frozen.

Questions RF Engineers Ask About Self-Referenced Dk Sensing

Will two modes cancel temperature drift?

Only the shared component of the drift can be suppressed. Different temperature responses of the two modes leave a residual shift. A controlled sweep of both frequencies is needed to establish whether compensation improves the extracted Dk.

Does R² = 0.9814 establish Dk accuracy?

It describes the fit of the reported calibration data. Accuracy requires comparison with justified reference values, analysis of bias and an uncertainty estimate for samples beyond that fitted set.

Can the same resonator measure dielectric loss, Df?

Potentially, but a resonance-frequency shift mainly informs permittivity. Extracting Df requires reliable quality-factor measurements and separation of dielectric loss from conductor, radiation, coupling and fixture loss. The 2026 sensitivity result is not a Df validation.

Why does a full unload-and-reload test matter?

Because a stationary coupon does not expose changes in local contact, air gap, orientation and fixture seating. Those effects can determine whether two engineers obtain the same Dk from the same material.

Engineering References

1. Bijari, A., and Nooki, S. (2026). A compact dual-mode single-resonator DGS microwave sensor for self-referenced permittivity characterization. Scientific Reports. DOI: 10.1038/s41598-026-76422-y.
2. Coonrod, J. (2019). Characterizing Circuit Materials at mmWave Frequencies. Microwave Journal / Rogers Corporation.
3. Al-Hinaai, A.-M., Paquette, M. M., and Hamdalla, M. Z. M. (2026). Microwave Techniques for Characterizing Low-Loss Solid Dielectrics: Measurement Physics, Resolution Limits, and Validation. Electronics 15(19), 4437.
4. Laribi, H., Moez, K., Rouhani, H., and Mirzavand, R. (2023). Temperature Compensated Dielectric Constant Sensor Using Dual-Mode Triangular Structure. IEEE Sensors Journal 23(16). Separate architecture and validation study.
5. IPC-TM-650 Method 2.5.5.15 (2022). Relative Permittivity and Loss Tangent Using a Split-Post Dielectric Resonator.