Cavity-Enhanced Photoluminescence

• Cavity-enhanced PL is defined as the modification of an emitter’s radiative properties by resonant cavities, enabling enhanced spontaneous emission and spectral filtering.
• It leverages mechanisms such as the Purcell effect, local field concentration, and selective out-coupling in platforms ranging from dielectric to plasmonic nanostructures.
• This technique supports practical applications in quantum photonics, tunable light sources, and optical sensing by engineering radiative states and emission dynamics.

Searching arXiv for recent and foundational papers on cavity-enhanced photoluminescence. Cavity-enhanced photoluminescence (PL) denotes photoluminescence generated by emitters whose optical environment is modified by a resonant cavity so that excitation, spontaneous-emission pathways, spectral distribution, temporal dynamics, and far-field out-coupling no longer follow those of the same emitters in an unpatterned film, bulk region, or free space. In the literature, this umbrella includes dielectric and plasmonic nanocavities, whispering-gallery and Fabry–Pérot resonators, photonic crystal cavities, microrings, metasurfaces, and microcavities operating from weak-coupling Purcell engineering to strong- and quantum-strong-coupling regimes. Depending on platform, the cavity can increase the local density of optical states, concentrate the pump field, select specific excitonic or defect-related channels, spectrally filter emission, or redirect photons into collection-friendly or waveguided modes (Abdelraouf et al., 2024, Calusine et al., 2015, Jiang et al., 2023, Herrera et al., 2016).

1. Physical basis and operative definitions

At its most conventional, cavity-enhanced PL is described as a cavity-induced modification of the local photonic density of states that increases spontaneous emission into a resonant optical channel. Several papers explicitly frame this as a Purcell effect. In the hybrid quasi-BIC metasurface with nanocrystalline silicon (nc-Si) quantum dots, destructive interference suppresses radiative leakage, electromagnetic energy becomes strongly confined in the nanoresonator, and the local density of optical states is increased, which strengthens light–matter interaction with the nc-Si emitters and boosts spontaneous emission into the resonant output channel (Abdelraouf et al., 2024). In the SOI G-center microring, the ideal on-resonance Purcell factor is written as

$$F_P = \frac{3}{4\pi^2}\left(\frac{\lambda}{n}\right)^3\frac{Q}{V},$$

and the measured $\sim 5$-fold zero-phonon-line enhancement is interpreted as Purcell-enhanced spontaneous emission into the resonant mode (Lefaucher et al., 2022). For the Eu$^{3+}$ fiber Fabry–Pérot microcavity, the Purcell-factor expression is given as

$$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$

with an inferred $F_P = 380 \pm 40$ for the ${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$ channel (Vasilenko et al., 10 Jun 2026).

A second, equally important usage is excitation-side enhancement. In 3C-SiC photonic crystal cavities, cavity-enhanced photoluminescence excitation is achieved by tuning a narrow-linewidth laser to the cavity resonance, so that the defect ensemble experiences a much larger local excitation intensity for the same incident laser power. The local field enhancement is expressed as $|E|^2 \propto \eta Q/V_m$, and the more explicit coupled-mode relation

$$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$

is used to account for the measured increase in PL and ODMR (Calusine et al., 2015). In plasmonic nanocavities, the enhancement can be the product of local excitation-field enhancement and radiative-decay enhancement; for room-temperature localized excitons in WSe$_2$, the TEPL paper summarizes this through

$$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$

highlighting that “cavity-enhanced PL” need not be reducible to spontaneous-emission-rate enhancement alone (Lee et al., 2020).

A third usage concerns spectrally and channel-selective emission. In TMDC-integrated ultrahigh-$\sim 5$0 silica microcavities, PL collected through a tapered fiber waveguide differs markedly from free-space emission, indicating selective coupling of the microcavity to specific excitonic channels (Kanzawa et al., 5 May 2026). In organic cavity QED, the HTC framework shows that dark vibronic polaritons can be invisible in absorption yet still radiate efficiently through photon leakage, so enhanced PL near the lower polariton does not imply a correspondingly strong absorption line (Herrera et al., 2016). This suggests that cavity-enhanced PL is best treated as a spectroscopic category defined by cavity-mediated modification of emission observables, not by any single microscopic mechanism.

The quality factor is often reported through the standard relation

$\sim 5$1

used explicitly in the carbon-nanotube photonic-crystal-cavity work (Watahiki et al., 2012). Yet the literature also shows that high $\sim 5$2 is neither necessary nor sufficient for large PL enhancement: reflective dielectric cavities in hBN arrays yield $\sim 5$3 PL enhancement with only a slight lifetime shortening from $\sim 5$4 ns to $\sim 5$5 ns, which the authors interpret as weak Purcell enhancement and primarily improved collection efficiency (Zeng et al., 2022).

2. Resonator classes and emitter platforms

The experimental landscape spans dielectric, plasmonic, and hybrid cavities, and the emitter classes include semiconductor QDs, impurity-bound excitons, color centers, rare-earth complexes, colloidal QDs, and excitons in two-dimensional semiconductors. The following table organizes representative implementations and the dominant cavity function emphasized in each case.

Platform	Emitter system	Dominant cavity function emphasized
Dual quasi-BIC hybrid metasurface	nc-Si QDs in a-Si/Sb$\sim 5$6S$\sim 5$7	high-$\sim 5$8 confinement, post-fabrication tunability, amplified NIR PL
Bullseye cavity	single Cl donor-bound exciton in ZnSe QW	small mode volume and near-Gaussian far field
Ultrahigh-$\sim 5$9 silica microsphere	monolayer WSe$^{3+}$0, WS$^{3+}$1	photothermal control and excitonic-channel selectivity
Fiber Fabry–Pérot microcavity	dinuclear Eu$^{3+}$2 complex	selective enhancement of $^{3+}$3
Polymer Fabry–Pérot thin-film cavity	FNDs with NV centers; hBN nanoparticles	scalable spectral reshaping, decay-rate enhancement, sensing improvement
Plasmonic nanocavity	localized or dark excitons in WSe$^{3+}$4; monolayer WS$^{3+}$5	ultrastrong near-field concentration, directionality, anti-Stokes activation
Photonic crystal cavity or microring	SWCNTs, Ky5 centers in 3C-SiC, G centers in SOI	mode-matched enhancement, cavity-assisted excitation, ZPL control

Dielectric resonators dominate where spectral purity, zero-phonon-line enhancement, or compatibility with integrated photonics is central. Examples include L3 photonic crystal slabs for SWCNTs, 3C-SiC L3 and H1 cavities for Ky5 centers, SOI microrings for G centers, and silica microdisks for CIS/ZnS QDs (Watahiki et al., 2012, Calusine et al., 2015, Lefaucher et al., 2022, Sun et al., 2017). Open-access microcavities provide a different operating point: the Eu$^{3+}$6 work uses a cryogenic, tunable fiber-based Fabry–Pérot resonator aligned to the $^{3+}$7 coherent transition, while the polymer thin-film cavities provide centimeter-scale, low-cost Fabry–Pérot structures compatible with fluorescent nanodiamonds and hBN nanoparticles (Vasilenko et al., 10 Jun 2026, Tibben et al., 5 Sep 2025).

Plasmonic architectures are used when extreme field confinement, sub-diffraction localization, or directional beaming is prioritized over narrow linewidths. The triple-sharp-tips TEPL nanocavity induces and controls localized exciton emission in WSe$^{3+}$8 at room temperature within a $^{3+}$9 nm area, and the nanoparticle-on-mirror geometry activates anti-Stokes PL through resonant excitation of a dark exciton (Lee et al., 2020, Mueller et al., 2023). The bent nanowire-on-mirror cavity instead emphasizes wavevector-space shaping of WS$$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$0 PL (Chaubey et al., 2022). Hybrid metasurfaces occupy an intermediate category: the a-Si/Sb$$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$1S$$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$2 quasi-BIC platform integrates emitters and a low-loss phase-change active modulator in the same resonant unit cell (Abdelraouf et al., 2024).

3. Enhancement channels and measured observables

Brightness enhancement is the most visible outcome, but the magnitude and meaning of “enhancement” vary across platforms. In the quasi-BIC nc-Si metasurface, the largest measured amplified PL reaches 15 times relative to a plain $$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$3 nm-thick a-Si film when $$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$4 nm, with other geometries giving enhancement factors between 4 and 12 (Abdelraouf et al., 2024). In the ZnSe bullseye cavity, the average saturation intensity rises from $$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$5 counts/s in bulk to $$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$6 counts/s in the cavity, corresponding to an average 15.8-fold brightness enhancement (Jiang et al., 2023). For SWCNTs on silicon photonic crystal nanocavities, the conservative lower bound on PL enhancement is $$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$7 (Watahiki et al., 2012). In 3C-SiC, resonantly excited L3 cavities yield $$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$8 relative to thin film, while silica microdisks produce more than 20 times PL enhancement at room temperature and 35 times at 20 K for CIS/ZnS QDs (Calusine et al., 2015, Sun et al., 2017).

Other platforms target narrower or more specialized optical channels. The Eu$$F_{\text{P}} = \frac{6}{\pi^{3}}\bigg(\frac{\lambda}{n}\bigg)^{2}\frac{\mathcal{F}}{w_{0}^{2}}\eta_E \eta_t,$$9 microcavity increases the fraction of excitations ending in $F_P = 380 \pm 40$0 photons from $F_P = 380 \pm 40$1 to $F_P = 380 \pm 40$2, and the inferred underlying ideal Purcell factor is $F_P = 380 \pm 40$3 (Vasilenko et al., 10 Jun 2026). The SOI G-center microring enhances the $F_P = 380 \pm 40$4 nm zero-phonon line by a factor of $F_P = 380 \pm 40$5 after normalization to the phonon-sideband background (Lefaucher et al., 2022). In hBN $F_P = 380 \pm 40$6 arrays, the reflective dielectric cavity yields about 7-fold PL enhancement and 18% ODMR contrast (Zeng et al., 2022). In polymer cavities, hBN nanoparticles show brightness increases of up to $F_P = 380 \pm 40$7 and PL decay-rate enhancement up to $F_P = 380 \pm 40$8, while fluorescent nanodiamonds exhibit up to 2.9-fold PL decay-rate enhancement and a 4.8 times improved magnetic field sensitivity for 20 nm FNDs (Tibben et al., 5 Sep 2025).

Time-domain observables show that increased brightness does not enforce a single dynamical signature. In ZnSe, the fast decay time changes from $F_P = 380 \pm 40$9 ps in bulk to ${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$0 ps in the cavity, yielding a measured Purcell factor of ${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$1 (Jiang et al., 2023). In polymer cavities, the experimental Purcell factors for FNDs remain ${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$2 in the resonant regime (Tibben et al., 5 Sep 2025). By contrast, the G-center microring shows no sizeable decrease of the average lifetime despite enhanced ZPL emission, leading to the conclusion of low radiative yield ${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$3 in heavily implanted material (Lefaucher et al., 2022). The TMDC microsphere exhibits the opposite behavior for cavity-coupled emission collected through the taper: lifetimes increase from ${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$4 ns and ${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$5 ns in free space to ${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$6 ns and ${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$7 ns in the cavity-coupled channel, which the authors interpret as evidence for selective coupling to a different excitonic state rather than simple radiative-rate enhancement (Kanzawa et al., 5 May 2026). The InGaN LED “cavity plasmon” work likewise reports longer, not shorter, spontaneous-emission decay times in the plasmonic samples (Li et al., 2024).

These examples make a central methodological point: cavity-enhanced PL can refer to increased integrated counts, spectral narrowing, enhanced emission into a chosen transition, a faster decay rate, a slower decay associated with state selectivity, or improved collection into a specific optical mode. The metric must therefore be specified case by case.

4. Spectral shaping, tunability, and directional out-coupling

A major contemporary theme is post-fabrication control. In the hybrid quasi-BIC metasurface, the two resonances are localized mainly in the top Sb${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$8S${}^5\mathrm{D}_0 \rightarrow {}^7\mathrm{F}_0$9 ellipse and bottom a-Si ellipse, respectively, and crystallization of Sb$|E|^2 \propto \eta Q/V_m$0S$|E|^2 \propto \eta Q/V_m$1 redshifts the resonances because of the higher refractive index of c-Sb$|E|^2 \propto \eta Q/V_m$2S$|E|^2 \propto \eta Q/V_m$3. The large index contrast $|E|^2 \propto \eta Q/V_m$4 enables post-fabrication tuning, with experimentally observed redshifts of 46 nm and 60 nm for the two resonances. Geometric modulation of the gap $|E|^2 \propto \eta Q/V_m$5 tunes the quasi-BIC resonance wavelength from 857 nm at $|E|^2 \propto \eta Q/V_m$6 nm to 752 nm at $|E|^2 \propto \eta Q/V_m$7 nm, a 105 nm tuning range; the all-optical phase-change tuning shifts the PL peak from 791 nm in the amorphous state to 815 nm in the crystalline state, giving a 24 nm range (Abdelraouf et al., 2024). This suggests a useful distinction between geometry-induced tuning, which is fixed by device dimensions, and reversible stimulus-driven tuning after fabrication.

Spectral reshaping can also arise from thermal control. In the TMDC-integrated ultrahigh-$|E|^2 \propto \eta Q/V_m$8 silica microcavity, cavity heating shifts the resonance according to

$|E|^2 \propto \eta Q/V_m$9

with $$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$0. The resulting local temperature rise from $$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$1 K to about $$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$2 K is then used in a Varshni model,

$$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$3

to explain the measured PL redshift. For WSe$$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$4, tuning the pump by about $$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$5 Å across the resonance region shifts the PL peak from $$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$6 nm to $$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$7 nm, corresponding to about $$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$8 meV (Kanzawa et al., 5 May 2026).

Directional control is equally prominent. The bent plasmonic nanowire-on-mirror cavity forces WS$$\frac{|E_c|^2}{|E_0|^2}=\frac{Q \eta \lambda n_o w_o^2}{4 n_c^2 V_m}$$9 emission into a narrow wavevector cone with radial angular spreading of $_2$0 and azimuthal spreading of $_2$1, compared with nearly isotropic emission for WS$_2$2 on bare gold (Chaubey et al., 2022). In the ZnSe bullseye cavity, simulations predict a nearly Gaussian far-field profile with about 37% of the radiation falling within the collection cone of the objective lens (Jiang et al., 2023). The G-center microrings use an external ring to scatter in-plane light toward the objective, while the Eu microcavity funnels emission into a well-defined cavity mode with $_2$3 (Lefaucher et al., 2022, Vasilenko et al., 10 Jun 2026). The dual-BIC metasurface extends the concept further by combining amplified PL with a tunable high-$_2$4 metalens that preserves more than 65% transmission efficiency across Sb$_2$5S$_2$6 phases and yields diffraction-limited focal spots with FWHM down to $_2$7m at 791 nm and $_2$8m at 803 nm and 815 nm (Abdelraouf et al., 2024).

Spectral filtering by cavity resonance can also be used as an emitter-selection tool. In polymer Fabry–Pérot cavities, the PL peak position of NV centers follows the cavity resonance, with average $_2$9 red-shifts by more than 50 nm for $$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$0–650 nm and a maximum shift of 80 nm for $$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$1 nm (Tibben et al., 5 Sep 2025). In photonic crystal cavities for SWCNTs, changing the lattice constant from $$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$2 nm to $$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$3 nm redshifts the cavity modes across the broad nanotube PL spectrum, which the authors explicitly connect to selective enhancement of a chosen chirality (Watahiki et al., 2012).

5. Quantum-emitter, excitonic, and polaritonic regimes

Cavity-enhanced PL is central to quantum-light-source engineering. The ZnSe bullseye cavity demonstrates more than an order of magnitude brighter emission than bulk impurity-bound exciton emitters while preserving antibunching, with raw $$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$4 and background-corrected $$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$5 (Jiang et al., 2023). The platform is explicitly framed as promising for spin–photon interfaces because the impurity-bound exciton has a spin ground state. The Eu$$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$6 dinuclear complex addresses a different quantum-optical requirement: selective enhancement of a weak but coherent optical transition. By boosting the $$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$7 line, the cavity improves the optical interface of a molecular rare-earth system that is already attractive for conditional ion–ion interactions and long optical coherence times (Vasilenko et al., 10 Jun 2026). In SOI microrings, enhancement of the telecom-band G-center zero-phonon line is positioned as a step toward deterministic single-photon sources for integrated quantum photonics (Lefaucher et al., 2022).

Excitonic cavity enhancement in low-dimensional semiconductors is more diverse. The ultrahigh-$$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$8 silica microsphere acts simultaneously as a photothermal actuator, a resonance thermometer, and a channel selector for WSe$$\langle \mathrm{EF}\rangle = \left|\frac{E_z}{E_0}\right|^2 \times \frac{\gamma^{\mathrm{rad}}}{\gamma_0},$$9 and WS$\sim 5$00 PL (Kanzawa et al., 5 May 2026). The WSe$\sim 5$01 nanoparticle-on-mirror work shows that plasmonic cavities can activate a normally weak excitation pathway: the out-of-plane near field couples to the dark exciton, and under $\sim 5$02 nm excitation the anti-Stokes PL enhancement factor reaches $\sim 5$03 in a representative cavity. Statistical data over 252 cavities show average anti-Stokes enhancement $\sim 5$04, rising to $\sim 5$05 when $\sim 5$06, establishing dark-exciton activation as the key ingredient for strong anti-Stokes PL enhancement (Mueller et al., 2023). The triple-sharp-tips TEPL cavity adds a different form of excitonic control: room-temperature localized-exciton emission in WSe$\sim 5$07 with TEPL enhancement as high as $\sim 5$08, PL energy fluctuations up to 40 meV, and imaging resolution better than 15 nm (Lee et al., 2020).

At the fully quantum-electrodynamic end, cavity-enhanced PL no longer means only brighter emission. In the planar microcavity quantum-dot system, the Jaynes–Cummings Hamiltonian

$\sim 5$09

is used to analyze the quantum strong-coupling regime, where manifolds with photon number $\sim 5$10 are populated and PL evolves from a peak around the cavity resonance to a doublet at the lower and upper polariton resonances in the low-density limit (Ishida et al., 2013). In organic cavity QED, the HTC model predicts dark vibronic polaritons with

$\sim 5$11

which are dark in absorption but can still emit through cavity leakage. The leakage PL spectrum is given by

$\sim 5$12

and the resulting theory explains strong PL near the lower polariton and near the bare molecular transition despite weak or absent corresponding absorption signatures (Herrera et al., 2016). A persistent misconception is therefore that cavity-enhanced PL should mirror absorption; these results show that, in dissipative vibronic and polaritonic systems, absorption and PL probe different matrix elements and different initial conditions.

6. Interpretation, limitations, and recurrent misconceptions

One recurrent misconception is that any PL increase directly measures a Purcell factor. Several studies explicitly argue otherwise. The hBN reflective dielectric cavity states that the enhancement is “mainly reflected in the improvement of the collection efficiency,” with only weak Purcell effect inferred from the small lifetime change (Zeng et al., 2022). The SWCNT photonic crystal cavity notes that the observed factor of at least 50 likely reflects a combination of increased emission rate into the cavity mode, improved directionality and collection efficiency, and spectral filtering by the cavity resonance (Watahiki et al., 2012). In 3C-SiC, the enhancement is strongly tied to excitation efficiency enhancement as well as collection efficiency enhancement (Calusine et al., 2015). A plausible implication is that CEPL should be parsed into excitation enhancement, radiative-rate modification, spectral redistribution, and out-coupling efficiency rather than collapsed into a single scalar figure of merit.

A second misconception is that stronger cavity coupling always shortens the PL lifetime. The TMDC microsphere data show much longer biexponential decay components in fiber-collected cavity-coupled emission than in free-space PL, which the authors interpret as selective coupling to long-lived or out-of-plane-dipole-related excitonic channels (Kanzawa et al., 5 May 2026). The InGaN/GaN polygonal metal microcavity likewise reports longer spontaneous-emission decay times in the plasmonic samples and interprets this as evidence for a distinct “cavity plasmon” coupling mechanism (Li et al., 2024). Conversely, the G-center microring shows enhanced ZPL intensity without a measurable shortening of the short lifetime component, leading to the conclusion that non-radiative decay dominates because of low radiative yield and parasitic defects (Lefaucher et al., 2022). Lifetime reduction therefore remains diagnostic only when non-radiative rates, ensemble averaging, and channel selectivity are controlled.

Loss, disorder, and material degradation are also pervasive limitations. In the quasi-BIC metasurface, the measured $\sim 5$13-factors in the crystalline phase are limited by scattering from rough sidewalls and interfaces (Abdelraouf et al., 2024). In heavily implanted SOI microrings, parasitic defects reduce $\sim 5$14, add background luminescence, and obscure intrinsic G-center dynamics (Lefaucher et al., 2022). The CIS/ZnS microdisk study shows that cavity-enhanced pumping can accelerate photooxidation and photobleaching under high excitation, even though the same cavity yields bright PL with almost absent degradation at sufficiently low power in vacuum (Sun et al., 2017). For hBN nanoparticles in polymer cavities, the authors stress that the observed behavior is more complex than simple cavity Purcell physics alone and likely involves plasmonic coupling, geometry changes from aggregation, and possibly charge-state dynamics (Tibben et al., 5 Sep 2025). These caveats matter because they define whether a device functions as a quantitative cavity-QED system, a practical brightness-enhancement platform, or a hybrid structure in which multiple mechanisms coexist.

Across the surveyed literature, the field converges on several application domains: tunable emitters, quantum photonic devices, optical communications, sensing, on-chip directional sources, spin–photon interfaces, anti-Stokes lasing, optical cooling, and scalable cavity-integrated quantum sensors (Abdelraouf et al., 2024, Tibben et al., 5 Sep 2025, Chaubey et al., 2022, Mueller et al., 2023, Jiang et al., 2023). What distinguishes recent work is not merely higher enhancement factors, but the co-integration of multiple functions in one architecture: emission amplification plus post-fabrication tuning in quasi-BIC metasurfaces, cavity-assisted thermal-state readout in TMDC microcavities, state-selective emission routing in fiber-coupled resonators, and broad-area manufacturability in polymer thin-film Fabry–Pérot devices. This suggests that the modern meaning of cavity-enhanced PL is progressively shifting from “brighter fluorescence in a resonator” toward a more general program of radiative-state engineering, where the cavity is designed as an active spectro-temporal and modal interface between the emitter and the photonic system.