Photonic-Crystal Resonators (PhCRs)

• Photonic-crystal resonators (PhCRs) are optical cavities that use periodic dielectric modulations to localize electromagnetic modes and create photonic bandgaps.
• They span diverse geometries—from slab and nanobeam to ring and fiber-based designs—that achieve high quality factors and minimized mode volumes for precise spectral control.
• Their advanced design facilitates nonlinear optical processes and integration with electronic and quantum systems, driving innovations in communications, sensing, and photonic circuits.

Photonic-crystal resonators (PhCRs) are optical cavities in which periodic dielectric modulation is used to localize electromagnetic modes, create photonic bandgaps, or impose mode-selective frequency splitting. In the literature, the term spans multilayer slab cavities, one-dimensional nanobeams, microrings with corrugated or sinusoidally modulated boundaries, circular resonators with concentric hole gratings, topological ring resonators defined by interfaces between distinct photonic crystals, and defect-free nanofiber Fabry–Perot structures. Across these implementations, the core figures of merit are the quality factor $Q$, mode volume, integrated dispersion, spectral selectivity, and compatibility with electrical, thermal, and quantum-emitter integration (Bushell et al., 2017, Pernice et al., 2014, Zhang et al., 2024).

1. Resonator classes and canonical geometries

A major PhCR family is the slab photonic-crystal cavity formed in a periodically patterned membrane. One representative realization is a multilayer slab photonic crystal cavity with a high-refractive-index semiconductor core, GaAs with $n=3.4$, sandwiched between lower-index cladding layers with variable refractive index $n_{\mathrm{clad}}$ from 1.0 to 3.4. In that structure, the periodic lattice of air holes extends through both the core and the cladding, and the analyzed high-$Q$ designs are an L3 cavity with three missing holes and a dispersion-adapted cavity based on a W1 waveguide with tailored hole displacements (Bushell et al., 2017).

A second major class is the one-dimensional photonic-crystal nanobeam cavity. The dumbbell resonator is a one-dimensional photonic crystal slot cavity in which two closely spaced holes are connected by a thin air slot, concentrating the optical field in a low-index region. In silicon, tapering is achieved by continuously reducing the radius of the holes from the mirror region towards the cavity center, whereas in aluminum nitride the lattice constant is varied in a nearly adiabatic parabolic transition to optimize confinement in a lower-index-contrast environment (Pernice et al., 2014). A related encapsulated Si/SiO$_2$ nanobeam cavity introduces a lattice offset $\Delta x_c$ between two central holes and varies only the first few holes surrounding the cavity in size and position, rather than applying a long “gentle confinement” taper across many holes (Vasco et al., 2019).

Ring-derived PhCRs constitute another broad category. In lithium niobate, photonic crystal ring resonators are implemented by sinusoidal modulation of the ring width at the inner boundary; in tantala, the photonic crystal is introduced as a nanopatterned modulation on the inner wall of a microring; and in hybrid SiN-on-LNOI devices, periodic corrugation on the inner periphery couples clockwise and counterclockwise modes into split supermodes (Zhang et al., 2024, Black et al., 2022, Peng et al., 1 May 2025). A distinct “edge-less” realization uses an azimuthally uniform, sub-wavelength sinusoidal nanostructure around a Ta$_2$O$_5$ ring to shift a single azimuthal mode without introducing discrete cavity edges (Yu et al., 2020).

Two additional geometries enlarge the scope of PhCRs. Circular photonic crystal resonators for telecom quantum light sources consist of a central disk surrounded by concentric rings of air holes with 12-fold symmetry, while a semiconductor topological photonic ring resonator is formed by embedding a hexagonal core of expanded unit cells inside a host of shrunken unit cells so that the interface itself becomes the resonator (Barbiero et al., 16 May 2025, Mehrabad et al., 2019). At a different scale, defect-free optical-nanofiber photonic-crystal Fabry–Perot resonators are fabricated as periodic nanograting “craters” on a 500 nm diameter nanofiber using a single high-power femtosecond pulse (Tanaka et al., 27 Jan 2026).

2. Confinement mechanisms, $Q$, and mode volume

A recurring assumption in slab photonic crystals is that high $Q$ requires high vertical index contrast and strong total internal reflection. The heterostructure study on GaAs and AlGaAs explicitly shows a counter-intuitive regime: due to the confinement provided by the photonic band structure in the cladding layers, it becomes possible to achieve $n=3.4$0-factors $n=3.4$1 with only a small refractive index contrast, even when $n=3.4$2 and $n=3.4$3–3.3. The mechanism is that when the photonic bandgap of the cladding overlaps the resonance frequency of the cavity, in-plane propagation of the cavity mode in the cladding is forbidden; finite-thickness claddings retain “pseudo-gap” behavior that inhibits coupling to extended Bloch states; and directional $n=3.4$4 analysis shows that in-plane propagation is strongly suppressed near $n=3.4$5 (Bushell et al., 2017).

For PhCRs more generally, $n=3.4$6 is used in both energy-decay and linewidth formulations. In the slab heterostructure work, the quality factor is measured via temporal decay as

$n=3.4$7

where $n=3.4$8 is stored energy and $n=3.4$9 is power dissipation. In the nanofiber Fabry–Perot work, the linewidth form

$n_{\mathrm{clad}}$0

is used, with measured linewidths as low as 16.6–32 MHz near 850–852 nm, a highest observed total $n_{\mathrm{clad}}$1, and intrinsic $n_{\mathrm{clad}}$2 (Bushell et al., 2017, Tanaka et al., 27 Jan 2026).

Nanobeam and slot-based PhCRs emphasize simultaneous high $n_{\mathrm{clad}}$3 and ultrasmall mode volume. In the dumbbell resonator, confinement arises from the slot effect and tapered Bragg reflectors, with simulated $n_{\mathrm{clad}}$4 above $n_{\mathrm{clad}}$5 in silicon and above $n_{\mathrm{clad}}$6 in AlN, while the best measured loaded $n_{\mathrm{clad}}$7 values are 20,000 in silicon and 33,000 in AlN. The effective mode volume is written as

$n_{\mathrm{clad}}$8

with reported examples $n_{\mathrm{clad}}$9 in silicon and $Q$0 in AlN (Pernice et al., 2014).

The encapsulated Si/SiO$Q$1 nanobeam cavity makes the same trade-off explicit in a compact integrated format. Its total quality factor is written as

$Q$2

and its linear and nonlinear mode volumes are

$Q$3

The optimized structure reaches $Q$4, $Q$5, $Q$6, in-plane transmission above 65\%, and a footprint of around $Q$7 (Vasco et al., 2019).

Hybrid silicon-on-lithium-niobate nanobeams and defect-free nanofiber PhCRs illustrate two distinct limits of confinement engineering. In the hybrid Si/LN platform, simulations yield radiation-limited $Q$8 with a mode volume about $Q$9 and 13–17\% of the optical energy in LN, whereas the nanofiber platform attains very high measured $_2$0 in a geometry with an estimated $_2$1 (Witmer et al., 2016, Tanaka et al., 27 Jan 2026). This suggests that the PhCR designation covers both ultrasmall-volume nanocavities and longer Fabry–Perot structures when the periodic photonic-crystal element is the dominant spectral and confinement mechanism.

3. Frequency engineering, mode splitting, and broadband loss

Mode-selective spectral engineering is a defining PhCR capability in microrings. In isotropic lithium niobate, the ring width is modulated as

$_2$2

and selective splitting occurs when $_2$3, with $_2$4 the azimuthal mode number. That platform demonstrates single-mode splitting up to 0.5 nm (62.5 GHz) with $_2$5 nm and maintains $_2$6 of order $_2$7. In anisotropic x-cut LN, a uniform modulation creates multiple unintended splittings because both effective index and perturbation strength vary around the circumference, so a gradient design is introduced that varies both mean ring width and modulation amplitude to keep the effective index and effective perturbation constant along the ring; the measured result is a single targeted splitting of about 60 pm while maintaining $_2$8-factors above $_2$9 in active racetrack devices with electrodes and SiO$\Delta x_c$0 cladding (Zhang et al., 2024).

A related control knob is the supermode splitting bandwidth in corrugated microrings. In the hybrid SiN-on-LNOI PhCR, periodic corrugation on the inner periphery couples clockwise and counterclockwise modes into two supermodes separated by a bandwidth $\Delta x_c$1 obeying

$\Delta x_c$2

The fabricated devices show intrinsic quality factor $\Delta x_c$3, supermode splitting bandwidth $\Delta x_c$4 GHz, electro-optic tunability of 0.85 pm/V, and a linear relation of 93.4 MHz/nm between splitting bandwidth and corrugation amplitude (Peng et al., 1 May 2025).

Fast inverse design of multimode PhCRs extends frequency engineering from individual split modes to full modal ladders. For tapered nanobeam cavities, a reduced one-dimensional model reproduces the optical response accurately across the spectral range of interest and enables optimization of the width profile so that all modes are equally spaced in frequency. The integrated dispersion is written as

$\Delta x_c$5

and, in the reported example, the optimization suppresses integrated dispersion from $\Delta x_c$6 THz to $\Delta x_c$7 GHz over 7 modes after only a small number of full three-dimensional solves (Talenti et al., 2022).

Spectral engineering in PhCRs is not loss-neutral across arbitrary wavelength ratios. A systematic study of grating-induced loss in photonic crystal microrings maps the full spectral response as a function of $\Delta x_c$8, where $\Delta x_c$9 is the grating period and $_2$0 the modal wavelength. It identifies a broad excess-loss region at $_2$1 attributed to vertical out-coupling into OAM-carrying states, strong narrow peaks at $_2$2 and $_2$3, and the conventional Bragg condition $_2$4 or $_2$5, which produces selective mode splitting with minimal extra loss. The practical guideline is explicit: operational wavelengths should avoid the major radiation-loss channels unless excess loss is intentionally assigned to unwanted modes (Pimbi et al., 20 May 2025).

4. Nonlinear optics: four-wave mixing, optical-parametric oscillation, and microcombs

High-$_2$6 multimode PhCRs have been developed as selective nonlinear oscillators rather than merely passive cavities. In bichromatic photonic crystal slab cavities in InGaP, a mode-dependent thermo-refractive tuning mechanism uses localized heating from a laser red-detuned from mode “0” to shift the frequencies of modes “S”, “0”, and “AS” differently and bring the triplet into exact spectral alignment, $_2$7. The measured loaded $_2$8 reaches $_2$9, stimulated four-wave-mixing conversion efficiency reaches $_5$0 dB ($_5$1) at 80–800 $_5$2W on-chip continuous pump power, spontaneous emission reaches GHz rates for less than 1 mW pump power, and the observed scaling follows the expected $_5$3 law for stimulated FWM and $_5$4 for the spontaneous regime (Marty et al., 2019).

The same selectivity is sharpened in the canonical FWM program, where the cavity is designed to support only the required number of modes. In a statistical study over more than 100 resonators and 10 parametric oscillators, the bichromatic geometry is compared with coupled-cavity and nanobeam designs. The lowest observed OPO threshold is approximately 40 $_5$5W, specifically 43 $_5$6W in the reported measurements, and OPO is not observed for $_5$7 or average photon lifetime below 150 ps because three-photon absorption clamps the intracavity energy before gain overcomes loss. The threshold scaling is written in terms of the cavity energy

$_5$8

together with a pump-threshold expression that depends on mode volume, nonlinear index, resonance frequency, and inverse square of the decay rate (Chopin et al., 2022).

Microring PhCRs use photonic-crystal bandgaps to make parametric oscillation wavelength-selective and designable. In tantala PhCRs, controlling the bandgap enables OPO generation across 1234–2093 nm with a 1550 nm pump and 1016–1110 nm with a 1064 nm pump, with pump-to-sideband conversion efficiency of $_5$9 and negligible additive optical-frequency noise across the output range. The frequency-selection mechanism is described using the integrated dispersion

$Q$0

with the bandgap compensating the phase-matching error for the selected signal–idler pair (Black et al., 2022).

PhCRs also modify the route to mode-locked states. In edge-less Ta$Q$1O$Q$2 PhCRs with $Q$3, a photonic-crystal-induced shift of the pump mode re-balances Kerr-nonlinear frequency shifts so that the flat state destabilizes directly toward a soliton pulse rather than a Turing pattern. The measured comb states show a $Q$4 envelope, terahertz-scale spacing confirmed by electro-optic modulation, and relative intensity noise below $Q$5 dBc/Hz up to GHz Fourier frequencies (Yu et al., 2020). In normal-dispersion photonic crystal ring resonators, the SIFF strategy adds auxiliary mode splittings at $Q$6 so that forward-propagating platicon states become deterministic; the reported optimal synchronization condition is $Q$7 in normalized units (Lucas et al., 19 Jun 2026).

The high-power limit of nonlinear selectivity is demonstrated in microresonators that combine weakly normal geometric dispersion with a photonic crystal bandgap that activates only a single OPO interaction. Those devices produce idler output power exceeding 40 mW and signal output power exceeding 30 mW, while maintaining side-mode suppression ratios greater than 40 dB. Four independent oscillators vary only the photonic crystal parameters to select different output waves, showing that wavelength programming can be accomplished without re-engineering the full resonator geometry (Brodnik et al., 10 Apr 2025).

5. Materials platforms, fabrication routes, and integration strategies

Material choice in PhCRs is inseparable from integration strategy. The all-semiconductor slab heterostructure addresses a practical limitation of air-clad slabs: poor thermal conductivity and the inability to integrate electrical contacts for current injection. By using semiconductor cladding layers such as AlGaAs and extending the air-hole lattice through the cladding, the design is compatible with standard epitaxial growth and processing and is presented as a route toward electrically pumped nano-cavity lasers (Bushell et al., 2017).

Hybrid material stacks are used when optical confinement, fabrication convenience, and active tuning must coexist. In the silicon-on-lithium-niobate nanobeam platform, only the 220 nm crystalline silicon film is patterned, then bonded to LN, and the resulting cavity combines high optical confinement in Si with useful overlap in LN. Patterned electrodes placed 600 nm from the nanobeam yield an electro-optic coupling rate $Q$8 GHz/V, equivalent to 4.3 pm/V, while keeping $Q$9, above the structural $Q$0 (Witmer et al., 2016).

The hybrid SiN-on-LNOI microring PhCR takes a different approach: the SiN is patterned, the LN is left unetched, and Au/Ti electrodes are patterned on the LN surface. This avoids direct LN etching, retains a high intrinsic quality factor, and adds voltage-driven frequency tuning and programmable mode splitting in a CMOS-compatible flow (Peng et al., 1 May 2025). Encapsulation can serve a similar systems role even without active electrodes: the Si/SiO$Q$1 nanobeam cavity is explicitly motivated by the mechanical, thermal, and fabrication-related weaknesses of free-standing, air-bridged nanobeams, and the fully encapsulated geometry is optimized for in-plane transmission and compactness (Vasco et al., 2019).

Integration can also proceed by redesigning the coupling architecture rather than the cavity material. In the metalens–PhCR metasystem, a broadband dielectric metalens is monolithically combined with an L3 photonic crystal cavity or photonic-crystal waveguide so that no single-mode silicon nanowire waveguide is required and no oxide cladding is needed. By mode-matching the metalens focal spot to the PhC input mode, the reported insertion loss is as low as $Q$2 dB, and the platform is described as mechanically robust and suitable for chemical and biosensors operating in air or solution (Xiao et al., 2021).

At the fiber scale, single-shot femtosecond laser ablation provides a distinct fabrication paradigm. The nanofiber resonator uses a phase mask to create a Bragg grating with a stop band at about 852 nm on a 500 nm diameter fiber, and the single-shot process is described as robust against mechanical vibrations. The resulting cavity is “defect-free” in the sense that the Gaussian intensity profile naturally forms a symmetric variation in ablation rather than an abrupt cavity defect (Tanaka et al., 27 Jan 2026).

6. Topological, quantum-emitter, sensing, and systems applications

One application direction uses PhCRs as structured environments for emitters and chiral transport. In the semiconductor topological photonic ring resonator, the interface between expanded and shrunken photonic crystals supports edge states inside the photonic bandgap, and the ring is probed with embedded InGaAs quantum dots. The degree of perturbation $Q$3 controls confinement: at $Q$4, the measured decay length transverse to the interface is $Q$5m, while at $Q$6 it increases to $Q$7m. The same work reports that missing-unit-cell defects have negligible effect on mode spectrum or $Q$8, and that circularly polarized dipoles excite one propagation direction depending on handedness (Mehrabad et al., 2019).

Telecom single-photon sources motivate a second line of development. Circular photonic crystal resonators incorporating InAs quantum dots demonstrate bright, Purcell-enhanced single-photon emission in the telecom C-band and electrically contacted devices with wide range tuneability in the telecom O-band. The geometry is specified by concentric rings of holes around a central disk, with hole coordinates

$Q$9

Simulations target Purcell factor $n=3.4$00 and collection efficiency up to about 90\% within NA = 0.65, while the measured devices show lifetimes of approximately 1.91 ns off cavity and 0.75–0.8 ns in the PhCR, $n=3.4$01 under above-band excitation, $n=3.4$02 under phonon-assisted excitation, detected count rates up to 20 million counts/s, and Stark tuning greater than 10 nm (Barbiero et al., 16 May 2025).

Sensing, optomechanics, and delay engineering form a third application cluster. The dumbbell nanobeam resonator is explicitly proposed for ultra-small opto-mechanical resonators, high-frequency operation, and sensing applications because the field maximum is concentrated in the air slot and the field gradients are large (Pernice et al., 2014). In two-dimensional photonic-crystal ring-resonator coupled-resonator optical waveguides, multiple resonant modes such as hexapole and octupole modes create multiple transmission bands, enabling multi-channel transfer on a single device. For example, with cavity spacing 4.320 $n=3.4$03m, the reported normalized group velocities are 0.0031 for the hexapole mode and 0.0034 for the octupole mode, values lower than the compared point-defect coupled-cavity waveguides and comparable to microring-resonator CROWs (Chauhan et al., 2018). This suggests that the systems role of a PhCR is not limited to high-$n=3.4$04 confinement; it can also be to impose compact, mode-resolved group-delay engineering.

A broader outlook treats PhCRs as a general platform for programmable soliton and frequency-comb generation. A 2025 perspective argues that PhCRs provide access to solitons “in a convenient and stable way,” and proposes a nanocomposite waveguide structure for optimization of group velocity dispersion, a pump-harmonic microcomb for $n=3.4$05–$n=3.4$06 self-referencing, and a two-microresonator network that retains the convenience of using a continuous-wave laser for microcomb generation. Because these elements are presented as proposals and design studies rather than solely as completed hardware, a cautious reading is warranted; a plausible implication is that the next research phase will combine PhCR-style local spectral control with broader cross-sectional dispersion engineering and multi-resonator architectures (Liu et al., 18 Aug 2025).