Electric-LC Resonators: Principles and Applications

• Electric-LC resonators are electromagnetic structures that separate capacitive and inductive energy storage to achieve lumped, subwavelength resonance behavior.
• Design techniques such as interdigital capacitor loading and re-entrant geometries enable precise tuning of resonance frequencies and improved metamaterial homogeneity.
• Their ability to support strong light–matter interactions underpins applications in hybrid photon–magnon systems and axion haloscope experiments.

An electric-LC resonator (ELCR) is a resonant electromagnetic structure in which electric energy is concentrated in a capacitive region and magnetic energy is stored in an inductive current path, so that the mode is well described by a lumped $LC$ picture despite subwavelength overall dimensions. In planar metamaterial form, the canonical electric-LC (ELC) resonator consists of two symmetric inductive loops linked by a central capacitive gap and exhibits a predominantly electric resonance under appropriate illumination; in three-dimensional realizations, including THz double-metal circuits and re-entrant cavities, the same principle appears as a spatial separation of capacitor and inductor regions. Across these implementations, ELCRs are used to reduce electrical size, tailor homogenization, confine quasi-static fields, and engineer coupling to intersubband plasmons, magnons, or axion-conversion channels [(Withayachumnankul et al., 2010); (Jeannin et al., 2019); (McAllister et al., 2016); (Maurya et al., 6 May 2026)].

1. Canonical definition and field topology

The generic ELCR resonance follows the lumped relation

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$

In the canonical planar ELC geometry, two symmetric inductive loops are connected by a single central capacitive gap. When a normally incident plane wave has its electric field polarized perpendicular to that gap, charge accumulates across the gap and currents flow around the two loops in opposite directions. By mirror symmetry, the net magnetic dipole moment cancels, so the response is predominantly electric with negligible magnetic or magnetoelectric contribution. In the quasistatic limit, this ELC behaves as a lumped resonator whose inductance is set by the two loops and whose capacitance is set by the central gap. For that specific symmetric two-loop geometry, neglecting loss, the resonance is

$f_0=\frac{1}{\pi\sqrt{2LC}},$

whereas the general single-resonator form is

$f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$

The distinction reflects the current symmetry of the canonical ELC rather than a different physical principle (Withayachumnankul et al., 2010).

The same field topology reappears in volumetric ELCRs. In a 3D THz LC circuit, quasi-static electric energy is concentrated between parallel metallic plates that form capacitors, while magnetic energy circulates around a bent inductive wire. In a re-entrant cavity, the capacitor is the small post-to-lid or ring-to-lid gap and the inductor is the azimuthal current path around the post or ring and the cavity wall. A persistent feature across these variants is spatial disentanglement of electric and magnetic energies: $E$ is strongly localized in a small capacitive volume, whereas $B$ occupies an inductive loop or a larger surrounding volume (Jeannin et al., 2019, McAllister et al., 2016).

2. Planar ELC resonators as electric metamaterial inclusions

The metamaterial implementation studied on FR4 uses a single-layer copper ELC array with square unit cell $a=14\ \mathrm{mm}$, line width $w=0.4\ \mathrm{mm}$, gap width $g=0.4\ \mathrm{mm}$, and shared dimensions $b=0.8\ \mathrm{mm}$, $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$0; the baseline central-gap length is $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$1. The substrate is FR4 epoxy of thickness $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$2, measured dielectric constant $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$3, and loss tangent $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$4; the metal is $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$5 copper. Arrays of $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$6 unit cells were fabricated and characterized in free space using two facing horn antennas under normal incidence with the incident $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$7-field aligned to excite the central gap (Withayachumnankul et al., 2010).

Experimentally, $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$8 shows a transmission dip at the electric resonance. For the baseline ELC$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$9, the measured resonance is $f_0=\frac{1}{\pi\sqrt{2LC}},$0, close to the simulated $f_0=\frac{1}{\pi\sqrt{2LC}},$1. The corresponding effective-medium ratio is $f_0=\frac{1}{\pi\sqrt{2LC}},$2, with $f_0=\frac{1}{\pi\sqrt{2LC}},$3. Effective-parameter retrieval using the Chen et al. method, based on simulated $f_0=\frac{1}{\pi\sqrt{2LC}},$4 and $f_0=\frac{1}{\pi\sqrt{2LC}},$5 with an effective slab thickness equal to the unit cell $f_0=\frac{1}{\pi\sqrt{2LC}},$6, shows the electric resonance in $f_0=\frac{1}{\pi\sqrt{2LC}},$7 and a negative-$f_0=\frac{1}{\pi\sqrt{2LC}},$8 band around resonance. In the ideal picture, $f_0=\frac{1}{\pi\sqrt{2LC}},$9, consistent with the intended absence of magnetic response (Withayachumnankul et al., 2010).

A recurrent interpretive issue concerns the retrieved permeability. Near resonance, $f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$0 exhibits anti-resonant artifacts and negative imaginary parts. The reported interpretation is that these are known consequences of spatial dispersion and inhomogeneity in periodic media rather than evidence that the ELC has become a genuinely magnetic resonator. This distinction is important because ELCR homogenization is limited not only by resonance strength but also by lattice-scale diffraction and nonlocality (Withayachumnankul et al., 2010).

3. Interdigital loading, resonance lowering, and homogenization

A central design strategy for compact planar ELCRs is to replace the single parallel-strip gap capacitor with an interdigital capacitor (IDC). For identical $f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$1, $f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$2, and substrate, the IDC capacitance is

$f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$3

with

$f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$4

and the stated elliptic-integral approximations for $f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$5. The baseline single gap corresponds to $f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$6 and length $f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$7, so

$f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$8

and therefore

$f_0=\frac{1}{2\pi\sqrt{LC_{\mathrm{tot}}}}.$9

With the loop inductance fixed, the resonance scales as

$E$0

The design rationale is direct: increasing $E$1 lowers $E$2, which reduces the electrical size $E$3 at resonance and improves metamaterial homogeneity (Withayachumnankul et al., 2010).

Two practical IDC variants were examined. In variant “a,” the IDC gaps are perpendicular to the incident $E$4-field and the ports are at the outermost fingers; in variant “b,” the gaps are parallel to the incident $E$5-field and the ports are at the terminal bars. Variant “a” aligns the strong local IDC field with the illumination and generally produces stronger capacitive coupling. Although variant “a” breaks mirror symmetry geometrically, the two loop areas remain equal, so magnetic dipoles still cancel (Withayachumnankul et al., 2010).

Design	IDC parameters	Measured $E$6, $E$7
ELC$E$8	$E$9, $B$0	$B$1, $B$2
ELC1a	$B$3, $B$4, gaps $B$5	$B$6, $B$7
ELC1b	$B$8, $B$9, gaps $a=14\ \mathrm{mm}$0	$a=14\ \mathrm{mm}$1, $a=14\ \mathrm{mm}$2
ELC2a	$a=14\ \mathrm{mm}$3, $a=14\ \mathrm{mm}$4, gaps $a=14\ \mathrm{mm}$5	$a=14\ \mathrm{mm}$6, $a=14\ \mathrm{mm}$7
ELC2b	$a=14\ \mathrm{mm}$8, $a=14\ \mathrm{mm}$9, gaps $w=0.4\ \mathrm{mm}$0	$w=0.4\ \mathrm{mm}$1, $w=0.4\ \mathrm{mm}$2

The strongest measured downshift is ELC2a, from $w=0.4\ \mathrm{mm}$3 for the baseline to $w=0.4\ \mathrm{mm}$4, approximately a $w=0.4\ \mathrm{mm}$5 reduction in resonance frequency. Over the same comparison, $w=0.4\ \mathrm{mm}$6 improves from $w=0.4\ \mathrm{mm}$7 to $w=0.4\ \mathrm{mm}$8, approximately a $w=0.4\ \mathrm{mm}$9 increase, and $g=0.4\ \mathrm{mm}$0 decreases from about $g=0.4\ \mathrm{mm}$1 to about $g=0.4\ \mathrm{mm}$2. The reported homogenization criterion is $g=0.4\ \mathrm{mm}$3, equivalently $g=0.4\ \mathrm{mm}$4; IDC loading moves the array further into that regime, reducing diffraction from individual cells (Withayachumnankul et al., 2010).

The data also show a trade-off. As capacitance increases, more electric energy is confined within the IDC and substrate, which reduces apparent resonance strength and raises the effective loss tangent on FR4. A plausible implication is that capacitance engineering alone does not guarantee a superior effective medium unless substrate loss and coupling strength are co-optimized. The original design guidance therefore recommends $g=0.4\ \mathrm{mm}$5 to maximize capacitance density, orientation with IDC gaps perpendicular to the incident $g=0.4\ \mathrm{mm}$6-field to maximize the frequency reduction for a given footprint, and lower-loss substrates when possible (Withayachumnankul et al., 2010).

4. Three-dimensional ELCR architectures

A 3D THz ELCR meta-atom embeds the capacitor and inductor in a genuinely volumetric circuit. Two parallel-plate capacitors are defined by overlaps between bottom square pads and top rectangular pads, each of area $g=0.4\ \mathrm{mm}$7 with $g=0.4\ \mathrm{mm}$8, while a bent top-metal wire connecting the two top plates forms the inductive loop. The active-region thickness that sets the capacitor spacing is $g=0.4\ \mathrm{mm}$9 in the first sample and $b=0.8\ \mathrm{mm}$0 in the second. The top pads are $b=0.8\ \mathrm{mm}$1 wider than the bottom pads to engineer fringing fields and break in-plane symmetry for free-space coupling. Arrays contain about $b=0.8\ \mathrm{mm}$2 resonators over $b=0.8\ \mathrm{mm}$3 with $b=0.8\ \mathrm{mm}$4 period, illuminated at $b=0.8\ \mathrm{mm}$5 incidence in TM polarization above a $b=0.8\ \mathrm{mm}$6 SiN spacer and an Au back mirror (Jeannin et al., 2019).

The equivalent circuit is an inductor $b=0.8\ \mathrm{mm}$7 in series with two identical capacitors $b=0.8\ \mathrm{mm}$8, with

$b=0.8\ \mathrm{mm}$9

Fitting the measured dispersion as the loop perimeter $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$00 is varied gives $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$01 from $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$02 down to $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$03, while $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$04, consistent with the parallel-plate estimate

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$05

the discrepancy being attributed to fringing fields and in-plane parasitics. Numerical field maps show that electric energy is confined almost entirely between the plates, with $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$06 normal to the metal surfaces, whereas magnetic energy is localized around the inductive wire. This geometry is therefore an ELCR in the strict lumped and spatially separated sense (Jeannin et al., 2019).

The 3D re-entrant cavity is the volumetric analogue emphasized in axion-haloscope research. It consists of a conducting cavity containing either a central post or a thin ring protruding toward the cavity lid and leaving a small gap. The gap acts as a parallel-plate capacitor with

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$07

while the surrounding azimuthal current path provides the inductance. Its lowest mode is described by

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$08

Because $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$09 to leading order while $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$10 changes weakly with $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$11, the resonance increases approximately as $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$12; as the gap becomes large, the mode evolves continuously into the empty-cavity TM$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$13 mode. For a $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$14 diameter, $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$15 long envelope, simulated post geometries tune from about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$16 to $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$17, and ring geometries from about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$18 to $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$19, by varying $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$20 from about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$21 to about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$22 (McAllister et al., 2016).

These 3D implementations establish that ELCR behavior is not restricted to lithographic planar meta-atoms. The defining feature is the controlled separation of capacitive and inductive energy storage, which permits resonances far below the full-wave scale associated with an empty cavity or an unstructured metallic element (Jeannin et al., 2019, McAllister et al., 2016).

5. Hybrid light–matter and photon–magnon coupling

In the THz ELCR platform, the absorbing medium is a semiconductor 2DEG placed exactly inside the capacitor overlaps, so that the strongly confined $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$23 field overlaps the intersubband dipoles and satisfies the polarization selection rule. The active region of the first sample comprises $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$24, modulation $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$25-doped with Si $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$26 from each QW and nominal $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$27 per QW. The intersubband plasmon is

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$28

with

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$29

The coupled ELCR–plasmon dispersion satisfies

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$30

and the minimum polariton splitting is

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$31

Here $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$32, $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$33, and $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$34 is the electric-energy overlap factor. Experimentally, the first ELCR sample yields $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$35 and $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$36; the minimum splitting is $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$37 with $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$38. In the optimized second sample, $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$39 and $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$40, with clear anticrossing and a polaritonic gap. The number of electrons participating is unusually small: about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$41 in the first sample and about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$42 in the second (Jeannin et al., 2019).

A distinct hybrid use appears in the planar cavity–magnonic system, where an ELCR is side-coupled to a $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$43 microstrip on a high-$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$44 substrate and integrated with a YIG thin film. In that work the ELCR is a copper patterned ring-like structure with outer radius $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$45, inner radius $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$46, track width $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$47, split gap $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$48, arm length $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$49, and lateral spacing to the microstrip $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$50. The resonator supports two orthogonal photon modes with

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$51

observed near $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$52 and $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$53. Rotating the ELCR by an in-plane angle $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$54 changes the projection of the microstrip near field onto the current loops and therefore the extrinsic damping,

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$55

with $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$56 and $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$57. At $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$58, only mode-1 is bright; at $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$59, mode-1 is dark and mode-2 is bright (Maurya et al., 6 May 2026).

With YIG placed atop the ELCR, the magnon mode obeys the Kittel relation

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$60

where $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$61 and $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$62. The fitted photon–magnon couplings redistribute strongly with angle: $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$63 at $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$64, while $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$65 at the same angles. The measured channel-switching angle defined by the dissipation order parameter

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$66

is $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$67, with a symmetry-related model prediction at $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$68. Cooperativities extracted using the HWHM convention are all greater than unity in the measured hybridized configurations, demonstrating strong coupling in this planar platform (Maurya et al., 6 May 2026).

Taken together, these results show that ELCRs can be designed either to maximize $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$69-dominated interaction with intersubband matter excitations or to redistribute microwave magnetic interaction between competing photon channels. The common enabling mechanism is field localization imposed by capacitive and inductive substructures, rather than a particular operating frequency or material stack.

6. Axion haloscopes, design criteria, and common interpretive issues

In axion-haloscope applications, the re-entrant ELCR is used because its lumped resonance lies well below the empty-cavity modes of the same physical volume. The signal power is

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$70

with mode form factor

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$71

where

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$72

and

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$73

The geometric factor

$$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$74

sets $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$75, and the design figure of merit is $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$76. For a $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$77 diameter, $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$78 long cavity, simulations indicate that post or ring radii in the range of about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$79 to $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$80 of the cavity radius are typically optimal; beyond about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$81, gains in form factor are offset by losses in volume and geometric factor. With an $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$82 solenoid, $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$83 at $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$84, $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$85, and a quantum-limited SQUID amplifier with effective noise about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$86, the projected scan range is about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$87 to $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$88 in about $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$89 days at $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$90 (McAllister et al., 2016).

Across ELCR platforms, several design rules recur. In planar IDC-loaded ELCs, $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$91 maximizes capacitance density, and placing IDC gaps perpendicular to the incident $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$92-field produces the largest measured downward resonance shift. In THz ELCRs, matching $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$93 to the intersubband plasmon requires tuning the inductive loop perimeter while keeping $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$94 fixed by $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$95 and $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$96, and maximizing $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$97 requires placing the QWs exclusively inside the capacitor overlap. In re-entrant haloscope cavities, tuning is accomplished primarily by changing the gap $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$98, while ring geometries are especially effective when magnetic form factor $$f=\frac{1}{2\pi\sqrt{LC}}, \qquad \omega=\frac{1}{\sqrt{LC}}.$$99 must be enhanced by pushing the magnetic field outward [(Withayachumnankul et al., 2010); (Jeannin et al., 2019); (McAllister et al., 2016)].

Several common misconceptions are explicitly corrected in the literature. First, geometric asymmetry does not by itself imply magnetic response: the IDC-loaded planar variant “a” breaks mirror symmetry geometrically, yet equal loop areas preserve magnetic dipole cancellation. Second, anomalous retrieved $f_0=\frac{1}{\pi\sqrt{2LC}},$00 near resonance in periodic ELC arrays should not be read as direct evidence of intrinsic magnetic resonance, because anti-resonant artifacts and negative imaginary parts can arise from spatial dispersion and inhomogeneity. Third, stronger miniaturization is not unconditionally beneficial: FR4-based IDC loading lowers $f_0=\frac{1}{\pi\sqrt{2LC}},$01 and improves $f_0=\frac{1}{\pi\sqrt{2LC}},$02, but also raises effective loss tangent and weakens resonance strength; similarly, practical limits in the microstrip-coupled YIG platform arise from fabrication tolerances, connector and microstrip transitions, and finite conductor and dielectric losses. A plausible general implication is that ELCR optimization is inherently multiparametric, balancing resonance frequency, overlap, external coupling, and dissipation rather than minimizing size alone [(Withayachumnankul et al., 2010); (Maurya et al., 6 May 2026)].

From metamaterial inclusions to THz ultra-strong-coupling circuits and MHz axion haloscopes, the ELCR is therefore best understood as a family of resonators defined by lumped $f_0=\frac{1}{\pi\sqrt{2LC}},$03 behavior, spatially separated energy storage, and extreme control over field localization. Its implementations differ substantially in geometry and operating band, but they are unified by the same design logic: engineer capacitance and inductance independently enough that resonant frequency, field polarization, mode volume, and hybrid coupling can be specified with unusual precision.