A nanophotonic resonator or nanocavity is an optical cavity which is on the order of tens to hundreds of nanometers in size. Optical cavities are a major component of all lasers, they are responsible for providing amplification of a light source via positive feedback, a process known as amplified spontaneous emission or ASE. Nanophotonic resonators offer inherently higher light energy confinement than ordinary cavities, which means stronger light-material interactions, and therefore lower lasing threshold provided the quality factor of the resonator is high. Nanophotonic resonators can be made with photonic crystals, silicon, diamond, or metals such as gold. For a laser in a nanocavity, spontaneous emission (SE) from the gain medium is enhanced by the Purcell effect, equal to the quality factor or Q {\displaystyle Q} -factor of the cavity divided by the effective mode field volume, F = Q / V mode {\displaystyle F=Q/V_{\text{mode}}} . Therefore, reducing the volume of an optical cavity can dramatically increase this factor, which can have the effect of decreasing the input power threshold for lasing. This also means that the response time of spontaneous emission from a gain medium in a nanocavity also decreases, the result being that the laser may reach lasing steady state picoseconds after it starts being pumped. A laser formed in a nanocavity therefore may be modulated via its pump source at very high speeds. Spontaneous emission rate increases of over 70 times modern semiconductor laser devices have been demonstrated, with theoretical laser modulation speeds exceeding 100 GHz, an order of magnitude higher than modern semiconductor lasers, and higher than most digital oscilloscopes. Nanophotonic resonators have also been applied to create nanoscale filters and photonic chips
Differences from classical cavities For cavities much larger than the wavelength of the light they contain, cavities with very high Q factors have already been realized (~125,000,000). However, high Q {\displaystyle Q} cavities on the order of the same size as the optical wavelength have been difficult to produce due to the inverse relationship between radiation losses and cavity size. When dealing with a cavity much larger than the optical wavelength, it is simple to design interfaces such that light ray paths fulfill total internal reflection conditions or Bragg reflection conditions. For light confined within much smaller cavities near the size of the optical wavelength, deviations from ray optics approximations become severe and it becomes infeasible, if not impossible to design a cavity which fulfills optimum reflection conditions for all three spatial components of the propagating light wave vectors. In a laser, the gain medium emits light randomly in all directions. With a classical cavity, the number of photons which are coupled into a single cavity mode relative to the total number of spontaneously emitted photons is relatively low because of the geometric inefficiency of the cavity, described by the Purcell factor Q / V mode {\displaystyle Q/V_{\text{mode}}} . The rate at which lasing in such a cavity can be modulated depends on the relaxation frequency of the resonator described by equation 1.
R 2 = ( a v g P 0 ) / τ p + β / ( τ p τ r 0 / F ) + ( β N 0 ) / ( ( τ r 0 / F ) P 0 ) ( 1 / τ total − 1 / ( τ r 0 / F ) ) ( 1 ) {\displaystyle R_{2}=(av_{g}P_{0})/\tau _{p}+\beta /(\tau _{p}\tau _{r0}/F)+(\beta N_{0})/((\tau _{r0}/F)P_{0})(1/\tau _{\text{total}}-1/(\tau _{r0}/F))\,\,\,\,\,\,\,\,(1)}
Where τ r 0 {\displaystyle \tau _{r0}} is the intrinsic carrier radiative lifetime of the bulk material, a {\displaystyle a} is the differential gain, v g {\displaystyle v_{g}} is the group velocity, τ p = Q / ω L {\displaystyle \tau _{p}=Q/\omega _{L}} is the photon lifetime, ω L {\displaystyle \omega _{L}} is the lasing frequency, β {\displaystyle \beta } is the spontaneous emission coupling factor which is enhanced by the Purcell effect, and 1 / τ total = F / τ r 0 + 1 / τ n r {\displaystyle 1/\tau _{\text{total}}=F/\tau _{r_{0}}+1/\tau _{nr}} where τ n r {\displaystyle \tau _{nr}} is the non-radiative lifetime. In the case of minimal Purcell effect in a classical cavity with small F = Q / V mode {\displaystyle F=Q/V_{\text{mode}}} , only the first term of equation 1 is considered, and the only way to increase modulation frequency is to increase photon density P 0 {\displaystyle P_{0}} by increasing the pumping power. However, thermal effects practically limit the modulation frequency to around 20 GHz, making this approach is inefficient. In nanoscale photonic resonators with high Q {\displaystyle Q} , the effective mode volume V mode {\displaystyle V_{\text{mode}}} is inherently very small resulting in high F {\displaystyle F} and β {\displaystyle \beta } , and terms 2 and 3 in equation 1 are no longer negligible. Consequently, nanocavities are fundamentally better suited to efficiently produce spontaneous emission and amplified spontaneous emission light modulated at frequencies much higher than 20 GHz without negative thermal effects.
Materials and designs
Nanocavities made from photonic crystals are typically implemented in a photonic crystal slab structure. Such a slab will generally have a periodic lattice structure of physical holes in the material. For light propagating within the slab, a reflective interface is formed at these holes due to the periodic differences in refractive index in the structure. A common photonic crystal nanocavity design shown is essentially a photonic crystal with an intentional defect (holes missing). This structure having periodic changes in refractive index on the order of the length of the optical wavelength satisfies Bragg reflection conditions in the y {\displaystyle y} and z {\displaystyle z} directions for a particular wavelength range, and the slab boundaries in the x {\displaystyle x} direction create another reflective boundary due to oblique reflection at dielectric boundaries. This results in theoretically perfect wave confinement in the y {\displaystyle y} and z {\displaystyle z} directions along the axis of a lattice row, and good confinement along the x {\displaystyle x} direction. Since this confinement effect along the y {\displaystyle y} and z {\displaystyle z} directions (directions of the crystal lattice) is only for a range of frequencies, it has been referred to as a photonic bandgap, since there is a discrete set of photon energies which cannot propagate in the lattice directions in the material. However, because of the diffraction of waves propagating inside this structure, radiation energy does escape the cavity within the photonic crystal slab plane. The lattice spacing can be tuned to produce optimal boundary conditions of the standing wave inside the cavity to produce minimal loss and highest Q {\displaystyle Q} . Beside those conventional resonators, they are some examples of rewritable and/or movable cavities, which are accomplished by a micro infiltration system and by a manipulation of single nanoparticles inside photonic crystals. Metals can also be an effective way to confine light in structures equal to or smaller than the optical wavelength. This effect is emergent from the confined surface plasmon resonance induced by the resonating light, which, when confined to the surface of a nanostructure such as a gold channel or nanorod, induces electromagnetic resonance. Surface plasmon effects are strong in the visible range because the permittivity of a metal is very large and negative at visible frequencies. At frequencies higher than the visible range, the permittivity of a metal is closer to zero, and the metal stops being useful for focussing electric and magnetic fields. This effect was originally observed in radio and microwave engineering, where metal antennas and waveguides may be hundreds of times smaller than the free-space wavelength. In the same way, visible light can be constricted to the nano level with metal structures which form channels, tips, gaps, etc. Gold is also a convenient choice for nanofabrication because of its unreactivity and ease of use with chemical vapour deposition.
A planar nanocavity consists of an absorptive semiconductive film no more than a few nanometers thick over a metal film also a few nanometers thick. Incident light is absorbed and reflected off of both layers, the absorbed light then resonates between the two interfaces, transmitting some light back at after each cycle. Germanium is commonly used for the absorptive layer, while gold, aluminum, and aluminum oxide are used as alternatives as well. Planar nanocavities are commonly used for thin film interference, which occurs when incident light waves reflected by the upper and lower boundaries of a thin film interfere with one another forming a new wave. An example of this is the colorful patterns produced by thin layers of oil on a surface. The difference in colors is due to minute differences in the distance reflected light travels whether it reflects from the top or bottom boundary of the oil layer. This difference is called the optical path difference, the difference in distance between the top and bottom reflection paths, which can be calculated with equation 2:
OPD = 2 n d cos ( θ ) ( 2 ) {\displaystyle {\text{OPD}}=2nd\,{\text{cos}}(\theta )\,\,\,\,\,\,\,\,(2)}
OPD = m λ ( 3 ) {\displaystyle {\text{OPD}}=m\lambda \,\,\,\,\,\,\,\,(3)}
Where n {\displaystyle n} is the refractive index of the absorptive material, d {\displaystyle d} is the thickness of the absorptive film, and θ {\displaystyle \theta } is the angle of reflection. As expressed in the equation 3, the optical path length difference (OPD) can be related to wavelengths which constructively interfere in the thin film. As a result, light which enters the film at different angles interferes with itself varying amounts, produces an intensity gradient for narrowband light, and a spectrum gradient for white light.
Examples/applications Nanophotonic circuit designs are similar in appearance to microwave and radio circuits, minimized by a factor of 100,000 or more. Researchers have made nano-optical antennas which emulate the design and functionality of radio antennas. There are a number of important differences between nanophotonics and scaled down microwave circuits. At optical frequency, metals behave much less like ideal conductors, and also exhibit plasmon-related effects like kinetic inductance and surface plasmon resonance. A nantenna is a nanoscopic rectifying antenna, a technology being developed to convert light into electric power. The concept is based on the rectenna which is used in wireless power transmission. A rectenna functions like a specialized radio antenna which is used to convert radio waves into direct current electricity. Light is composed of electromagnetic waves like radio waves, but of a much smaller wavelength. A nantenna, an application of a nanophotonic resonator, is a nanoscale rectenna on the order of the optical wavelength size, which acts as an "antenna" for light, converting light into electricity. Arrays of nantennas could be an efficient means of converting sunlight into electric power, producing solar energy more efficiently than semiconductor bandgap solar cells. It has been suggested that nanophotonic resonators be used on multi core chips to both decrease size and boost efficiency. This is done by creating arrays of nanophotonic optical ring resonators that can transmit specific wavelengths of light between each other. Another use of nanophotonic resonators in computers is in optical RAM (O-RAM). O-Ram uses photonic crystal slab structure with properties such as strong confinement of photons and carriers to replace the functions of electrical circuits. The use of optical signals versus electrical signals is a 66.7% decrease in power consumption. Researchers have developed planar nanocavities that can reach 90% peak absorption using interference effects. This result is useful in that there are numerous applications that can benefit from these findings, specifically in energy conversion
Resonance and light confinement A nanophotonic resonator stores electromagnetic energy in one or more discrete optical modes. Its resonant frequencies and field distributions are determined by the device geometry, boundary conditions, constituent materials and surrounding dielectric environment. Resonators may confine light through total internal reflection, distributed Bragg reflection, a photonic band gap, surface plasmon resonance, destructive interference between radiation channels, or combinations of these mechanisms. Two central parameters are the optical quality factor Q {\displaystyle Q} and the effective mode volume V m o d e {\displaystyle V_{\mathrm {mode} }} . For a resonance at angular frequency ω 0 {\displaystyle \omega _{0}} with energy-decay rate κ {\displaystyle \kappa } , the quality factor is commonly expressed as Q = ω 0 / κ {\displaystyle Q=\omega _{0}/\kappa } . In wavelength units, it is approximately Q = λ 0 / Δ λ {\displaystyle Q=\lambda _{0}/\Delta \lambda } , where Δ λ {\displaystyle \Delta \lambda } is the full width at half maximum of the resonance. A high quality factor corresponds to a narrow linewidth and long photon lifetime, while a small mode volume corresponds to spatial concentration of the electromagnetic field. The combination of high Q {\displaystyle Q} and small V m o d e {\displaystyle V_{\mathrm {mode} }} strengthens light–matter interaction. In the weak-coupling regime, the spontaneous-emission enhancement of a resonant electric-dipole emitter is described by the Purcell effect and, for ideal spectral and spatial alignment, scales with Q / V m o d e {\displaystyle Q/V_{\mathrm {mode} }} . The enhancement is reduced when the emitter is detuned from the cavity, displaced from the field maximum, oriented incorrectly with respect to the cavity polarization, or broadened by dephasing.
Coupling and spectral response Nanophotonic resonators are commonly coupled to bus waveguides, free-space beams, tapered fibers, antennas or nearby emitters. The measured or loaded quality factor contains contributions from intrinsic loss and external coupling. In a simple single-mode description,
1 Q l o a d e d = 1 Q i n t r i n s i c + 1 Q c o u p l i n g . {\displaystyle {\frac {1}{Q_{\mathrm {loaded} }}}={\frac {1}{Q_{\mathrm {intrinsic} }}}+{\frac {1}{Q_{\mathrm {coupling} }}}.}
The relation between intrinsic and coupling losses determines whether the resonator is undercoupled, critically coupled or overcoupled. At critical coupling, the rate at which energy enters through the coupling channel equals the intrinsic loss rate, which can produce a deep transmission minimum in a side-coupled waveguide geometry. Interference between a narrow cavity resonance and a broad transmission pathway can produce an asymmetric Fano resonance. Temporal coupled-mode theory is widely used to describe the excitation, decay and interference of resonant modes in waveguide-cavity and free-space systems. Coupled resonators can exhibit mode splitting, avoided crossings, supermodes and collective effects that are absent in an isolated cavity.
Resonator architectures
Photonic-crystal cavities Photonic-crystal nanocavities are formed by introducing a defect or gradual modulation into a periodic dielectric structure. Common slab-cavity geometries include L3 cavities, H0 cavities, H1 cavities and width-modulated line-defect cavities. One-dimensional nanobeam cavities use a periodic sequence of holes or corrugations along a narrow suspended or supported waveguide. Tapering the lattice period or hole dimensions can reduce scattering into radiation modes and increase the quality factor. Theoretical designs can reach quality factors substantially higher than experimentally measured values. Surface roughness, sidewall angle, lithographic errors, material absorption and contamination introduce scattering and absorption that limit the realized quality factor. Optimization methods, including genetic algorithms, adjoint optimization and topology optimization, have been used to modify nearby holes and dielectric boundaries to reduce radiation loss.
Whispering-gallery resonators Whispering-gallery wave resonators confine light through repeated total internal reflection around a curved dielectric boundary. They include microrings, microdisks, microspheres, microtoroids and bottle resonators. Their resonances are separated by the free spectral range, which depends on optical path length and group index. Integrated microrings and microdisks can be coupled to adjacent waveguides and are used for filtering, modulation, sensing, nonlinear frequency conversion and frequency-comb generation. Whispering-gallery cavities can achieve very high quality factors because the optical mode can remain far from etched surfaces. Reducing the resonator radius increases the free spectral range and footprint density, but can increase bending radiation and surface-scattering loss.
Plasmonic nanocavities Plasmonic nanocavities confine optical fields through collective electron oscillations at metal–dielectric interfaces. Typical geometries include nanoparticle-on-mirror cavities, bowtie antennas, metallic nanogaps, patch antennas and plasmonic whispering-gallery cavities. Their mode volumes can be far below the diffraction-limited dielectric volume, but absorption in the metal usually produces lower quality factors than those of low-loss dielectric cavities. Nanometre-scale plasmonic gaps can produce large local field enhancement and strong coupling to molecules, excitons and quantum emitters. Strong coupling has been observed at room temperature in nanoparticle-on-mirror and related nanocavity configurations.
Dielectric nanoresonators and bound states in the continuum High-index dielectric nanoparticles and metasurfaces support electric and magnetic multipolar resonances with lower absorption than noble-metal plasmonic structures. Interference between multipoles or between radiation pathways can suppress far-field leakage. A bound state in the continuum is an ideal non-radiating state embedded within the spectrum of radiating modes; practical structures generally realize finite-quality-factor quasi-bound states through controlled symmetry breaking, finite size or absorption. Quasi-bound-state resonances have been used to enhance lasing, harmonic generation, sensing and directional emission. Their quality factor is controllable through geometry, but extreme values increase sensitivity to fabrication errors, disorder and environmental perturbations.
Hybrid resonators Hybrid resonators combine different confinement mechanisms or material platforms. Examples include dielectric cavities coupled to plasmonic antennas, semiconductor emitters transferred onto silicon nitride or lithium-niobate circuits, and diamond nanocavities containing color centers. Hybridization can combine the low loss of a dielectric cavity with the small mode volume of a plasmonic hotspot or the active properties of a separate emitter material.
Materials platforms Silicon provides high refractive-index contrast and mature fabrication, but at telecommunications wavelengths it exhibits two-photon absorption and free-carrier effects at elevated optical intensity. Silicon nitride has a broad transparency range, low linear loss and negligible two-photon absorption near 1.55 μm, making it widely used for high-Q resonators, nonlinear optics and frequency combs. III–V semiconductors such as gallium arsenide, indium phosphide and gallium nitride provide direct-bandgap optical gain and second-order nonlinearities. Diamond combines a wide transparency window, high thermal conductivity and optically active defect centers. High-Q diamond nanocavities have been fabricated from bulk single-crystal material for quantum optics, nonlinear optics and optomechanics. Lithium niobate, lithium tantalate and aluminum nitride provide electro-optic, piezoelectric and second-order nonlinear responses. Heterogeneous and transfer-printing processes allow materials with different optical, electrical and mechanical properties to be combined on one chip. Two-dimensional materials, carbon nanotubes, organic layers and phase-change materials can be positioned near cavity field maxima to add absorption, gain, electro-optic tuning or nonvolatile reconfiguration.
Nonlinear optics The circulating field inside a resonator can be much larger than the input field, lowering the external power needed to observe nonlinear optical effects. Resonator-enhanced processes include the Kerr effect, two-photon absorption, free-carrier dispersion, saturable absorption, Raman scattering, optical bistability, second- and third-harmonic generation, sum- and difference-frequency generation and four-wave mixing. In Kerr resonators, intensity-dependent refractive index changes shift the resonance and can produce bistability, self-phase modulation, parametric oscillation and frequency-comb generation. Dispersion engineering and phase matching determine which nonlinear interactions accumulate coherently. Dissipative Kerr solitons in high-Q microresonators provide coherent broadband optical frequency combs with applications in metrology, communications, spectroscopy and signal generation. Second-order nonlinear processes require non-centrosymmetric materials or effective symmetry breaking. Resonant enhancement has been demonstrated in III–V semiconductors, lithium niobate, aluminum nitride and dielectric quasi-bound-state structures. Bound-state-based nanoresonators can simultaneously increase field intensity and control far-field radiation of the generated harmonic. In silicon cavities, two-photon absorption can generate free carriers that alter both refractive index and absorption. The resulting response may enable switching, but carrier lifetime and thermal accumulation can limit repetition rate and introduce memory effects. Silicon microrings can also display regenerative oscillation when free-carrier and thermo-optic dynamics interact.
Optical switching, modulation and memory A resonance can be switched by changing the refractive index or absorption of the cavity material. Control mechanisms include optical pumping, carrier injection or depletion, the thermo-optic effect, the electro-optic effect, mechanical displacement and material phase transitions. Because the cavity converts a small resonance shift into a comparatively large change in transmitted or reflected power, resonators can reduce switching energy at the cost of narrower optical bandwidth and increased sensitivity to detuning. Photonic-crystal nanocavities containing InGaAsP have demonstrated all-optical switching with sub-femtojoule pulse energies and switching times of tens of picoseconds. The reported switching resulted from carrier-induced nonlinearity enhanced by the small cavity volume. Silicon resonators have also been used for all-optical control through carrier and thermal nonlinearities. Bistable cavities can function as optical latches or memory elements. Nonvolatile tuning can be introduced by integrating phase-change materials whose optical constants remain in a programmed state after the control pulse is removed. Such materials can compensate fabrication variation or configure resonator networks without continuous holding power.
Quantum optics and cavity quantum electrodynamics Nanophotonic resonators are used to control emission from quantum dots, color centers, atoms, molecules, two-dimensional materials and carbon nanotubes. In the weak-coupling regime, a resonator can increase emission into a selected mode, shorten radiative lifetime and improve collection into a waveguide or free-space channel. These effects are used to improve brightness and indistinguishability of single-photon sources. In the strong-coupling regime, the coherent emitter–cavity interaction rate exceeds relevant dissipative rates, producing hybrid light–matter states and vacuum Rabi splitting. Plasmonic nanocavities can approach this regime with extremely small mode volumes, while dielectric cavities generally provide lower loss and longer coherence times. Diamond color centers are investigated as spin–photon interfaces for quantum networking and sensing. Nanocavities increase the fraction of emission into the zero-phonon line and can interface the emitter with an integrated photonic circuit. Deterministic positioning, spectral matching and preservation of optical and spin coherence remain major integration challenges.
Cavity optomechanics and transduction In cavity optomechanics, the resonance frequency of an optical cavity depends on a mechanical coordinate. Radiation pressure, electrostriction or gradient forces allow photons to drive mechanical motion, while mechanical displacement modulates the optical field. Nanobeam and slab optomechanical crystals can co-localize optical and mechanical modes within the same periodic structure. Optomechanical resonators are used for displacement, force, mass and acceleration sensing; mechanical cooling; microwave-to-optical conversion; and studies of quantum mechanical motion. Hybrid piezo-optomechanical devices couple microwave fields to mechanical vibrations and then to optical cavity modes, providing a route to coherent conversion between superconducting circuits and optical networks.
Sensing and spectroscopy An analyte near a resonator changes the effective refractive index, absorption or scattering loss of the optical mode. The resulting resonance shift, linewidth change, mode splitting or phase response can be measured without fluorescent labels. Resonator sensors have been developed for gases, chemicals, nanoparticles, proteins, nucleic acids, temperature, pressure and acceleration. Sensitivity is influenced by field overlap with the analyte, resonance linewidth, optical power, noise and thermal stability. A high quality factor improves spectral discrimination, but a very long photon lifetime can reduce measurement bandwidth and increase susceptibility to thermo-optic drift. Hybrid plasmonic–photonic resonators can place an intense nanoscale field at the sensing surface while retaining a higher-Q dielectric mode. Optomechanical transduction can convert a small optical resonance shift into a mechanical-frequency shift. A high-Q optomechanical oscillator has been used to detect individual bovine serum albumin molecules through the optical spring effect.
Photonic and neuromorphic computing Resonators are used in photonic computing as wavelength-selective filters, modulators, weight elements, delay elements, nonlinear processors and optical memories. Arrays of microrings can perform wavelength-division multiplexing and weighted summation, while photonic-crystal cavities can provide compact nonlinear transfer functions. In optical neural networks, resonator-enhanced material nonlinearities have been investigated for activation functions and spiking dynamics. Candidate mechanisms include Kerr refraction, free-carrier dispersion, two-photon absorption, saturable absorption, optical bistability and phase transitions. The usefulness of a cavity-based activation element depends not only on the shape of its transfer function but also on insertion loss, operating energy, response time, thermal stability, fan-out and cascadability. A graphene–silicon photonic-crystal cavity has been investigated as a reconfigurable nonlinear activator by combining silicon Kerr nonlinearity with graphene saturable absorption. Such devices illustrate the use of cavity enhancement to reduce the optical energy required for neuromorphic nonlinear processing.
Tuning and reconfiguration Fabricated resonators frequently require post-fabrication tuning because nanometre-scale dimensional errors can shift the resonance by more than its linewidth. Reversible tuning methods include local heating, carrier injection, electro-optic modulation, optical pumping, mechanical strain, fluid infiltration and adsorption. Permanent trimming methods include oxidation, laser modification, material deposition and controlled etching. Thermo-optic tuning is comparatively simple but consumes static power and can introduce thermal crosstalk. Electro-optic tuning can be faster and is particularly effective in materials such as lithium niobate. Hybrid silicon–lithium-niobate platforms have demonstrated ring, disk and photonic-crystal resonators together with electro-optic cavity tuning. Cavity-enhanced photoelectrochemical etching has been used to permanently align multiple gallium-arsenide resonators with picometre-scale precision. The method selectively etches only resonators driven near resonance and can tune several devices using a common optical field.
Fabrication and integration Nanophotonic resonators are fabricated using combinations of thin-film growth, electron-beam or optical lithography, dry etching, wet etching, deposition, oxidation and wafer bonding. Suspended photonic-crystal cavities generally require a selective undercut to create vertical index contrast. Metallic nanogaps may be produced through aligned lithography, self-assembly, electromigration or nanoparticle placement. The smallest fabrication errors can dominate the loss of a high-Q cavity. Sidewall roughness scatters light into radiation modes, while surface states and contamination add absorption. In photonic-crystal slabs, deviations in hole radius and position perturb both the resonance frequency and the radiation-loss cancellation used to obtain high quality factor. CMOS-compatible processes have produced silicon photonic-crystal nanocavities with quality factors above 10 5 {\displaystyle 10^{5}} using optical lithography. This is lower than the best electron-beam-defined research devices but supports larger-scale integration with conventional silicon photonics. Heterogeneous integration methods include wafer bonding, direct growth, transfer printing and pick-and-place assembly. These methods allow an optimized resonator material to be combined with separate lasers, detectors, quantum emitters, electronics or mechanical structures. Inverse design can generate compact couplers and cavity boundaries subject to fabrication constraints, although the resulting structures may require careful validation against disorder and minimum-feature-size limits.
Characterization Resonances are characterized through transmission, reflection, scattering, photoluminescence or cathodoluminescence spectroscopy. The quality factor can be estimated from the linewidth when the spectrometer and laser have sufficient resolution. Time-domain cavity ring-down measurements instead determine the photon lifetime directly. Near-field optical microscopy and electron-energy-loss spectroscopy can map localized modes with subwavelength spatial resolution. Pump–probe measurements are used to determine switching dynamics, carrier lifetime and thermal relaxation. For emitter–cavity systems, lifetime measurements, polarization dependence, photon-correlation measurements and spectral detuning are used to distinguish Purcell enhancement, lasing and strong coupling. In nonlinear and thermal experiments, the measured resonance depends on scan direction, optical power and scan rate. Slowly sweeping a laser across a thermally unstable high-Q resonance can distort the lineshape, so calibrated power, scan speed and environmental stability are important for extracting intrinsic parameters.
Performance trade-offs and limitations Increasing the quality factor strengthens field buildup and narrows the resonance, but also increases photon lifetime and sensitivity to temperature, fabrication variation and laser-frequency noise. A narrow linewidth can limit data bandwidth unless the resonance is actively stabilized or dynamically tuned. Reducing mode volume increases field intensity but can place more energy near etched surfaces, metals or defects that add scattering and absorption. Dielectric resonators generally provide lower loss and higher quality factors, while plasmonic cavities provide smaller mode volumes and stronger local fields. Hybrid systems attempt to balance these properties but add fabrication complexity and interface loss. For active cavities, optical gain must overcome intrinsic and coupling losses; for quantum systems, dephasing and spectral diffusion can prevent the emitter from benefiting fully from a high-Q resonance. Nonlinear cavities also face competing timescales. Electronic Kerr responses can be ultrafast, while free-carrier and thermo-optic responses are slower and may cause pattern-dependent behavior. Resonance-enhanced devices can reduce operating energy, but practical systems must account for coupling loss, pump rejection, thermal control, fabrication yield and the optical power required to drive subsequent stages.
Emerging directions Research directions include inverse-designed cavities, topology-based confinement, bound states in the continuum, moiré photonic structures, dynamically reconfigurable resonators and large networks of mutually coupled cavities. Topological photonic-crystal nanocavities have been investigated as a route to localized modes with robustness derived from the surrounding band structure. Networks of tuned resonators are being studied for coupled-mode computing, synthetic dimensions, topological transport, frequency conversion and simulation of many-body systems. Scaling these systems requires reproducible r