Compound Cavity Design Method

• Compound cavity design method is a strategy that couples distinct resonant subsystems to create functionalities like spectral selectivity, slow light, and tunable coupling not achievable with a single cavity.
• It spans diverse applications from plasmonic nanophotonics and microwave engineering to laser systems, demonstrating controlled trade-offs between efficiency, transmission, and modal behavior.
• Analytical and computational frameworks such as FDTD, equivalent-circuit modeling, and topology optimization are key to precisely tuning the interactions and modal dynamics in compound architectures.

Searching arXiv for the cited topic and related papers. “Compound cavity design method” denotes a family of cavity-engineering strategies in which performance is obtained not from a single uniform resonator, but from the deliberate coupling of multiple resonant subsystems, functional subregions, or boundary conditions. In the cited literature, the term appears across nanophotonics, microwave engineering, accelerator RF, fiber and semiconductor lasers, optical reference cavities, illumination optics, and acoustic scattering analysis. The common thread is the use of compound structure—embedded subcavities, coupled sub-rings, dielectric-isolated cavity partitions, multiport resonant bodies, re-entrant compensation rings, cavity-plus-coupler composites, or cavity-plus-virtual-boundary decompositions—to realize spectral selectivity, slow light, variable coupling, thermal compensation, DC biasing, harmonic mode locking, or controllable scattering that a single cavity would not provide as readily (Dai et al., 2010, Cohen et al., 2017, Liu et al., 2016, Feng et al., 2021, Sekutowicz et al., 2 Jul 2025, Zhang et al., 2012, Jin et al., 2018, Toshimitsu et al., 2023).

1. Conceptual scope and defining characteristics

In the cited work, the phrase does not denote one universal algorithm. Some papers explicitly propose a “compound” cavity or grating architecture, while others do not use that exact label but present a design logic that is directly applicable to compound or multi-section resonators. This suggests that the most accurate encyclopedia-level definition is functional rather than taxonomic: a compound cavity design method is a method that treats the cavity as a coupled system of distinct electromagnetic, optical, or acoustic subsystems whose interaction is itself the principal design variable.

Several recurring structural interpretations appear. In a plasmonic–dielectric compound grating, the cavity is an embedded plasmonic MDM resonance cavity inside a dielectric Si grating, and the useful response comes from interference between a dielectric surface mode and an SPP-induced cavity mode (Dai et al., 2010). In a split 3D microwave cavity, the cavity is partitioned into two dielectric-isolated conducting halves that simultaneously preserve the $\mathrm{TE}_{101}$ mode and act as DC electrodes (Cohen et al., 2017). In a variable-channel cavity combiner, the compound nature lies in the resonator coupled to many identical inputs and one output, with matching preserved through the relation $\beta_0=n\beta_i+1$ (Liu et al., 2016). In a figure-8 compound-ring-cavity filter, the cavity is a multi-loop fiber network whose transfer function depends on three interacting feedback loops rather than one round trip (Feng et al., 2021).

The same logic appears in ostensibly different settings. The DESY superconducting photoinjector treats the cathode recess and backplate opening as a local RF sub-geometry coupled into a 1.6-cell cavity, so that local reshaping can reduce field enhancement without materially disturbing the global accelerating mode (Sekutowicz et al., 2 Jul 2025). The low-thermal-noise optical reference cavity uses a ULE spacer, re-entrant fused silica mirrors, and fused silica rings so that thermomechanical coupling becomes a controllable degree of freedom rather than a parasitic effect (Zhang et al., 2012). The inverse-designed multimode cavity coupler treats the relevant object as the waveguide–coupler–cavity system, not the bare cavity alone (Jin et al., 2018). The acoustic cavity-scattering formulation introduces a virtual boundary enclosing the cavity and couples an interior integral equation to a half-space hybrid representation, effectively turning open-cavity analysis into a coupled-domain construction (Toshimitsu et al., 2023).

A useful caution follows from the broad RF review: “RF Cavity Design” does not explicitly define a compound cavity design method by that name, but it does provide the general framework—specification, mode choice, geometry, figures of merit, equivalent circuit, beam interaction, and multi-gap optimization—from which compound or coupled-cavity design naturally follows (Jensen, 2016).

2. Canonical architectural patterns

The literature supports several stable architectural patterns.

Domain	Compound structure	Primary objective
Plasmonic nanophotonics	Si grating + embedded Ag/SiO$_2$/Ag MDM cavity	EIT-like transparency, slow light, high transmission
3D microwave cavity	Two dielectric-isolated cavity halves	DC bias without ruining the microwave mode
High-power RF combiner	Resonant cavity + many identical inputs + one output	Matching with variable channel count
Fiber laser filter	Figure-8 CRC with three couplers and four fiber sections	Ultra-narrow SLM comb filtering
SRF photoinjector	1.6-cell cavity + cathode recess sub-geometry	Reduce local peak field near cathode opening
Optical reference cavity	ULE spacer + FS mirrors + FS rings	Low thermal noise and tunable CTE zero crossing
On-chip photonics	Bare multimode cavity + compact coupler region + waveguide	Multimode critical or near-critical coupling
Acoustic scattering	Open cavity + virtual boundary + half-space hybrid representation	Stable cavity-scattering analysis

One pattern is broad–narrow pathway interference. The plasmonic–dielectric compound grating uses a low-loss dielectric direct pathway and a narrow plasmonic indirect pathway; the authors characterize the resulting transparency as an EIT-like effect arising from Fano-like interference/coupling (Dai et al., 2010). The multimode cavity coupler similarly uses a compact scattering region to tune several external coupling rates of a pre-existing cavity without redesigning the cavity body from scratch (Jin et al., 2018).

A second pattern is functional partitioning of the cavity conductor. In the split microwave cavity, the seam is placed parallel to the dominant surface-current flow of the $\mathrm{TE}_{101}$ mode so that the seam behaves mainly as a capacitive discontinuity rather than an interruption of an inductive current path (Cohen et al., 2017). In the DESY gun, the cathode opening is treated as a local field-shaping boundary condition inside the cavity rather than as an external accessory (Sekutowicz et al., 2 Jul 2025).

A third pattern is multi-loop or multiport resonant networking. The figure-8 CRC contains two local sub-rings and one global compound loop, and its narrow comb spectrum is determined by the combined loop phases rather than a single round-trip condition (Feng et al., 2021). The variable-way cavity combiner is likewise a compound resonator in network form, with $n$ identical input couplers and one output coupler linked by an equivalent-circuit matching relation (Liu et al., 2016).

A fourth pattern is local optimization under global modal constraints. The DESY cavity changes only the cathode-channel edge from a circular to an elliptical contour while keeping the resonant frequency change below $2\ \text{kHz}$ and field flatness at $99\%$ (Sekutowicz et al., 2 Jul 2025). The photonic-crystal slab work similarly optimizes selected hole positions and radii to control both target frequency and radiative loss across the band gap rather than optimizing $Q$ alone (Monim et al., 22 Dec 2025).

3. Analytical and computational frameworks

Compound cavity design is closely tied to the choice of mathematical representation. Different subfields emphasize different state variables, but the governing strategy is similar: write the coupled subsystems in a form where the interaction terms are explicit and then use those terms as design knobs.

For slow-light plasmonic–dielectric gratings, the principal observable is group delay derived from transmission phase,

$$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$

with spectra and phase obtained numerically by FDTD (Dai et al., 2010). The figure-8 CRC uses signal-flow graphs and Mason’s rule,

$$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$

to reduce a three-loop fiber cavity to an exact transfer function in terms of coupler gains and propagation factors (Feng et al., 2021). The variable-channel cavity combiner uses normalized equivalent-circuit analysis to derive the matching condition

$\beta_0=n\beta_i+1$0

and then links coupling directly to power dissipation through

$\beta_0=n\beta_i+1$1

(Liu et al., 2016).

In thermal-noise-limited optical cavities, the key design equation is the effective coefficient of thermal expansion. The re-entrant FS-ring design is organized around the sign and magnitude of the deformation coupling coefficient $\beta_0=n\beta_i+1$2, with the effective zero crossing written as

$\beta_0=n\beta_i+1$3

The design goal is to obtain negative $\beta_0=n\beta_i+1$4 by replacing ULE rings with FS rings (Zhang et al., 2012).

Inverse-designed multimode couplers move to distributed optimization over material degrees of freedom. The general minimax program uses

$\beta_0=n\beta_i+1$5

subject to radiative-$\beta_0=n\beta_i+1$6 constraints, while the practical critical-coupling surrogate maximizes the worst normalized stored energy,

$\beta_0=n\beta_i+1$7

(Jin et al., 2018). The photonic-crystal slab optimization instead uses a low-dimensional parametric cost,

$\beta_0=n\beta_i+1$8

where $\beta_0=n\beta_i+1$9 is a smooth geometric penalty enforcing minimum radius and wall-thickness constraints (Monim et al., 22 Dec 2025).

The acoustic scattering formulation is different in implementation but similar in philosophy. It combines a truncated layer potential on a finite wall segment with a truncated Sommerfeld integral, then introduces a virtual boundary $_2$0 around the cavity so that the open cavity can be analyzed through coupled boundary equations in $_2$1 and $_2$2 (Toshimitsu et al., 2023). This suggests that compound cavity design often requires compound representations: local physical-space discretization for near interactions and spectral or asymptotic treatment for the infinite exterior.

4. Microwave, RF, and accelerator implementations

In 3D microwave engineering, the split-cavity design shows how compound partitioning can add functionality without destroying the desired mode. The cavity is machined in two halves, separated by a dielectric, and used directly as DC electrodes. With the split parallel to the dominant surface-current flow of the $_2$3 mode, the room-temperature linewidth remains nearly constant even when the cavity is separated by about $_2$4, while at $_2$5 the split architecture maintains $_2$6 at $_2$7 and $_2$8 at $_2$9 (Cohen et al., 2017). A common misconception is that any seam necessarily ruins the cavity mode; the paper shows instead that seams aligned with current streamlines can preserve the target mode while suppressing other modes such as $\mathrm{TE}_{101}$0.

The variable-way high-power cavity combiner treats matching and dissipation as coequal design constraints. Its central design choice is to keep the input coupling coefficient fixed and retune only the output coupling according to

$\mathrm{TE}_{101}$1

In the 500 MHz example, $\mathrm{TE}_{101}$2, the operating range is $\mathrm{TE}_{101}$3, and the corresponding output couplings are $\mathrm{TE}_{101}$4, $\mathrm{TE}_{101}$5, $\mathrm{TE}_{101}$6, and $\mathrm{TE}_{101}$7, with $\mathrm{TE}_{101}$8 at 500 MHz after retuning in all four cases (Liu et al., 2016). The method is not just a matching rule; it is explicitly an efficiency-aware rule because $\mathrm{TE}_{101}$9 sets the cavity power loss ratio.

Accelerator SRF design provides a different compound-cavity lesson: local sub-geometry can dominate high-field limitations. In the DESY 1.6-cell photoinjector, the original circular cathode-channel edge produced a local wall field of $n$0 at an on-axis peak field of $n$1, whereas replacing that edge with an elongated elliptical contour reduced the local wall field to $n$2, with resonant frequency change below $n$3, field flatness still $n$4, no multipacting in the cathode region from $n$5 to $n$6, and only about a $n$7 increase in RF power dissipated at the cathode (Sekutowicz et al., 2 Jul 2025). The design objective is therefore not merely field minimization, but optimization of useful cathode field relative to parasitic wall field.

The CEPC cavity and HOM-coupler design generalizes the same principle to a five-cell SRF system. The cavity is a 650 MHz, five-cell superconducting cavity with cell-to-cell coupling $n$8, $n$9, and $2\ \text{kHz}$0, combined with five waveguide HOM couplers and one coaxial main coupler (Zheng et al., 2015). The paper separates spectral functions: the input coupler damps same-order modes, while waveguide HOM couplers damp higher-order modes using waveguide cutoff to reject the accelerating mode. This is a distinctly compound design philosophy: different attached subsystems are optimized for different modal families.

The broad RF review reinforces that this multi-section logic is not incidental. It treats the cavity as a structure whose useful behavior is determined by mode choice, $2\ \text{kHz}$1, $2\ \text{kHz}$2, shunt impedance, equivalent-circuit representation, beam loading, and, in multi-gap systems, Brillouin-diagram-guided phase synchronism (Jensen, 2016). This suggests that compound cavity design in RF is best understood as controlled mode coupling across cells, gaps, and attached subsystems.

5. Photonic, laser, and illumination implementations

The plasmonic–dielectric compound grating is one of the clearest examples of a compound cavity designed through modal interference. The structure is a 1D Si grating with an embedded Ag–SiO$2\ \text{kHz}$3–Ag MDM cavity. The all-plasmonic reference system yields a transparency window near $2\ \text{kHz}$4 with group index around $2\ \text{kHz}$5 but poor transmission because the transparent peak coincides with absorption $2\ \text{kHz}$6. Replacing the direct plasmonic pathway by a dielectric grating pathway produces an asymmetric transmission resonance with small absorption; after coupling optimization through Ag thickness, the reported performance reaches group index $2\ \text{kHz}$7 with transmission $2\ \text{kHz}$8, and about $2\ \text{kHz}$9 periods are sufficient for the finite structure to approximate the infinite-array response (Dai et al., 2010). The relevant interpretive point is not “EIT-like” versus “Fano-like.” The paper explicitly supports the more precise statement that the transparency is EIT-like in function and Fano-like in mechanism.

The figure-8 compound-ring-cavity filter shows the same compound logic in fiber form. Using three 95:5 couplers and fiber lengths $99\%$0, $99\%$1, the authors obtain a measured F8-CRC free spectral range of $99\%$2 and a measured passband of $99\%$3, narrow enough to force SLM oscillation in a main ring cavity with mode spacing about $99\%$4 (Feng et al., 2021). In the full laser, all four single-wavelength states are stable SLM with linewidths $99\%$5, $99\%$6, $99\%$7, and $99\%$8, and tunable dual-wavelength operation produces a $99\%$9 microwave signal. Here the compound cavity is a filter subsystem embedded inside a longer laser cavity.

The compound feedback InP mode-locked laser uses two equal-length 1.6 mm Fabry–Perot subcavities connected by two 50:50 MIRs: an active colliding-pulse mode-locking subcavity and a passive feedback subcavity (Lo et al., 2018). By varying only DC bias conditions, the optical mode spacing switches from $Q$0 to $Q$1, and the best $Q$2 state gives a pulse duration of $Q$3. The paper is careful, however, that the $Q$4 and $Q$5 states occur during the transition between normal passive mode locking and dual-wavelength optical mixing. A common overreading would be to treat all reported spacings as equally clean harmonic mode-locking states; the paper does not support that.

In integrated photonics, inverse-designed multimode cavity couplers treat the coupler region as the main design variable. The demonstrated devices achieve critical or nearly critical coupling between multi-ring cavities and a single waveguide at up to six widely separated wavelengths spanning the $Q$6–$Q$7 range (Jin et al., 2018). The compound object is the full waveguide–coupler–cavity system. This suggests that, in photonic compound-cavity design, cavity performance can be limited less by the cavity mode itself than by the interface through which several modes must be addressed simultaneously.

The photonic-crystal slab study moves from interface design to frequency-addressable cavity-family optimization. In GaAs slabs immersed in water, the method optimizes the quality factor of a standard L3 cavity from $Q$8 to $Q$9–$$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$0, with an addressable resonance-frequency range covering $$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$1 relative bandwidth and spanning more than half of the band gap (Monim et al., 22 Dec 2025). It also reports optimized H1 cavities and a scalarized frequency-loss objective. The paper does not explicitly formulate a compound-cavity framework, but it states that the method is of general use for arbitrary cavity geometries and materials. This suggests a plausible extension to multi-defect or coupled-defect cavity families, though that extension is not itself demonstrated.

The illumination system based on an aspheric lens and a compound ellipsoidal cavity shows the compound-cavity idea in nonresonant beam shaping. The aspheric element creates a transmitted Flat-top core and a reflected outer double-Gaussian component; the latter is redistributed by a cavity formed from 50 confocal complex ellipsoid cavities obtained by a mapping equalization algorithm (Lu et al., 2023). With maximum reflection angle $$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$2, transmission angle $$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$3, and adjustment factor $$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$4, the design reports a minimum RMS focusing radius of $$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$5, minimum effective transmission reflection power density ratio $$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$6, aperture $$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$7, aperture angle $$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$8, and target-plane uniformity $$n_g=\frac{c\,\tau_g}{L}=-\frac{c}{L}\frac{d\phi(\omega)}{d\omega}, \qquad \tau_g=-\frac{d\phi}{d\omega},$$9 over a $$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$0 diameter at $$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$1. Although this is not a resonant cavity in the narrow RF sense, it is still a compound cavity in the paper’s own geometric and flux-partitioning sense.

6. Tradeoffs, limitations, and broader implications

The most consistent theme across the literature is tradeoff management. In the plasmonic–dielectric grating, increasing Ag thickness weakens coupling, lowers transmission, and increases group index until the transparency window disappears once $$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$2 exceeds roughly the Ag skin depth (Dai et al., 2010). In the cavity combiner, larger $$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$3 improves combining efficiency but makes the output coupler more difficult to realize; the practical recommendation is $$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$4 for high-power operation (Liu et al., 2016). In the reference cavity, fused silica mirrors lower thermal noise but require compound thermomechanical compensation to keep the effective zero crossing near room temperature (Zhang et al., 2012). In photonic-crystal slabs, pushing the resonance upward in normalized frequency enlarges the radiative cone and makes simultaneous frequency targeting and low loss more difficult (Monim et al., 22 Dec 2025).

Another recurring lesson is that compound architecture is rarely free. The split microwave cavity preserves the target mode, but the split reduces cryogenic $$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$5 from $$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$6 in the closed cavity to $$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$7 in the split cavity (Cohen et al., 2017). The multilayer superlattice extension of the plasmonic–dielectric grating preserves high group index of about $$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$8 in a five-layer structure but reduces transmission from about $$H=\frac{1}{\Delta}\sum_i P_i\Delta_i, \qquad \Delta=1-\sum_i Q_i+\sum_{l\neq m}Q_lQ_m-\cdots,$$9 for a single-layer grating to about $\beta_0=n\beta_i+1$00 (Dai et al., 2010). The H1 photonic-crystal optimization raises $\beta_0=n\beta_i+1$01 strongly but also increases mode volume from $\beta_0=n\beta_i+1$02 to $\beta_0=n\beta_i+1$03 (Monim et al., 22 Dec 2025).

Methodological limits are equally clear. The multimode cavity-coupler work mainly designs the coupler/interface rather than the cavity from scratch (Jin et al., 2018). The cavity-scattering boundary-element method is 2D, scalar, and rigid-boundary only; it is best viewed as a forward solver for open-cavity analysis, not yet a full compound-cavity optimization framework (Toshimitsu et al., 2023). The RF review likewise provides a conceptual foundation for compound design but does not name a single formal procedure (Jensen, 2016).

Taken together, the cited work suggests a broad technical synthesis. Compound cavity design is most successful when the cavity is decomposed into subsystems with clearly assigned functions—storage, transport, filtering, biasing, damping, compensation, or remapping—and when the coupling between those subsystems is treated as a first-class design variable rather than a perturbation. The specific mathematical machinery varies from FDTD phase extraction to Mason-rule network reduction, equivalent-circuit matching, FEM eigenmode tuning, adjoint topology optimization, finite-difference parametric descent, and hybrid integral equations. What remains constant