Photonic Curing: Mechanisms & Applications

• Photonic curing is a process that uses precisely tuned optical pulses to locally induce phase changes and reactions without imposing a long-duration thermal load.
• It enables advanced applications such as defect activation in silicon and rapid, controlled polymerization in additive manufacturing with significant energy and time efficiency.
• Multiple modalities—including nanosecond laser annealing, hybrid single-photon-assisted two-photon polymerization, and DLP curing—demonstrate its diverse mechanisms and dynamic process control.

Photonic curing is the use of tailored light fields to drive desired reactions or phase changes in space and time. In one formulation, it is the use of short, intense optical pulses to deposit energy locally and transiently, thereby driving a reaction without imposing a long-duration thermal budget on the entire substrate; in another, it is the use of patterned or focused optical fields to drive polymer network formation and sculpt optical functionality such as refractive-index contrast, phase profiles, and voxelized solidification (Andrini et al., 2023, Heinrich et al., 13 Mar 2026). Recent work places the term across at least three technically distinct regimes: sub-melting nanosecond pulsed laser annealing for defect activation in silicon, single-photon-assisted two-photon polymerization for accelerated additive manufacturing, and single-photon DLP curing of photopolymers where refractive index and diffraction evolve dynamically during exposure (Unlu et al., 2024, Camposeo et al., 2022).

1. Scope and representative implementations

The recent literature treats photonic curing as a process class rather than a single device architecture. In silicon, nanosecond pulsed laser annealing functions as a photonic curing process that activates G-centers—telecom-wavelength emitters—in high-purity float-zone silicon. In additive manufacturing, a continuous-wave blue source at 405 nm can pre-excite a photocurable resin by single-photon absorption, after which a focused femtosecond beam at 780 nm provides the missing energy necessary to reach the polymerization threshold through two-photon absorption. In DLP-based micro-optics, UV projection drives a time-dependent refractive-index increase and a forming phase grating whose diffraction orders can be monitored in real time (Andrini et al., 2023, Unlu et al., 2024, Heinrich et al., 13 Mar 2026).

System	Light delivery	Reported outcome
High-purity FZ Si	532 nm, 4 ns pulses, 5 pulses at 5 Hz on 7×7 µm² squares	G-center activation at 1279 nm below melting
Acrylate resin for 2PP	405 nm sub-threshold pre-sensitization + 780 nm, 70 fs writing beam	150 nm lateral resolution at 10× shorter exposure
PR48 photopolymer under DLP	4.2–20 mW/cm² UV projection with 40 µm pixels	S-shaped refractive-index evolution and dynamic diffraction

These implementations share localized energy delivery, but they do not share a single microscopic mechanism. In silicon, the relevant variables are temperature excursions, thermal gradients, diffusion lengths, and structural metastability. In polymerization, the relevant variables are initiator excitation, radical generation, oxygen inhibition, autoacceleration, vitrification, and the spatial transfer function of the optical system. A plausible implication is that “photonic curing” is best understood as a controlled non-equilibrium processing regime rather than as a synonym for any one annealing or printing modality.

2. Governing physics: transient heating, threshold crossing, and spatial confinement

In sub-melting nanosecond laser annealing of silicon, optical absorption at 532 nm deposits heat within the first few hundred nanometers due to the large absorption coefficient in Si. The absorbed volumetric power density is modeled as

$Q(z,t) = (1 - R)\,\alpha\, I_0(t)\, e^{-\alpha z},$

with reflectivity $R = 0.371$ and absorption coefficient $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$ at 532 nm. The thermal evolution follows

$$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$

with $\rho = 2330\ \mathrm{kg/m^3}$. Short pulses create steep thermal gradients and rapid heating/cooling, with relaxation back to ambient in less than $1\ \mu\mathrm{s}$ and order-of-magnitude cooling rates of $10^8$–$10^9\ \mathrm{K/s}$ (Andrini et al., 2023).

In single-photon-assisted two-photon polymerization, the dose picture is explicitly hybrid. The one-photon absorption rate is

$R_{1PA} = \sigma_1 I_1,$

the excited photoinitiator population obeys

$\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$

and the two-photon absorption rate is

$R = 0.371$0

The exposure metrics are $R = 0.371$1 and $R = 0.371$2, and the paper uses the threshold condition

$R = 0.371$3

The blue-light step is kept below the one-photon polymerization threshold so that polymerization remains confined to the focal volume where the femtosecond beam supplies the remaining dose (Unlu et al., 2024).

In DLP photopolymerization for micro-optics, the dynamic refractive index and optical function are governed by radical photopolymerization kinetics and by projector optics. The paper uses initiation, propagation, and termination rates in quasi-steady state,

$R = 0.371$4

and relates refractive-index increase to densification through the Lorentz–Lorenz relation,

$R = 0.371$5

For diffraction, the thin-phase relation is

$R = 0.371$6

These expressions place photonic curing in a thresholded, time-dependent, and spatially filtered regime rather than in a static dose-only description (Heinrich et al., 13 Mar 2026).

3. Sub-melting nanosecond photonic curing in silicon

A specific semiconductor implementation was demonstrated in high-purity float-zone Si wafers, n-type, resistivity $R = 0.371$7–$R = 0.371$8, with native carbon $R = 0.371$9. The implanted species was $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$0 at $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$1 with fluence $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$2. Conventional rapid thermal annealing used a PID-controlled SSI SOLARIS 150 system in $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$3 for $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$4 at $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$5, with tested temperatures of $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$6, $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$7, $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$8, $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$9, $$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$0, $$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$1, and $$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$2. Under these conditions in FZ silicon, no detectable G-center signatures were produced; only W-centers, with a ZPL at $$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$3 and phonon replicas, formed and then annealed out at high RTA temperatures above $$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$4 (Andrini et al., 2023).

The nanosecond pulsed laser annealing route used a high-power Q-switched Nd:YAG laser at $$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$5, pulse duration $$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$6, repetition rate $$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$7, and five pulses per treated site. The processed geometry was a set of discrete $$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$8 square regions patterned across the implanted sample. Studied fluences were $$\rho\, c_p(T)\, \frac{\partial T}{\partial t} = \nabla \cdot \big(k(T)\nabla T\big) + Q(z,t),$$9–$\rho = 2330\ \mathrm{kg/m^3}$0, with specific reported values at $\rho = 2330\ \mathrm{kg/m^3}$1, $\rho = 2330\ \mathrm{kg/m^3}$2, $\rho = 2330\ \mathrm{kg/m^3}$3, $\rho = 2330\ \mathrm{kg/m^3}$4, $\rho = 2330\ \mathrm{kg/m^3}$5, $\rho = 2330\ \mathrm{kg/m^3}$6, and $\rho = 2330\ \mathrm{kg/m^3}$7. The maximum energy was $\rho = 2330\ \mathrm{kg/m^3}$8 per $\rho = 2330\ \mathrm{kg/m^3}$9 pulse, corresponding to up to $1\ \mu\mathrm{s}$0. The treatment was kept below melting; no recrystallization or ablation was reported (Andrini et al., 2023).

The target defect was the G-center, described as a neutrally charged substitutional dicarbon pair coupled to an interstitial silicon atom, often denoted as a $1\ \mu\mathrm{s}$1 paired with $1\ \mu\mathrm{s}$2. Its zero-phonon line is at $1\ \mu\mathrm{s}$3 in the telecom O-band. Ensemble photoluminescence showed that ZPL intensity increased at low fluence, peaked near $1\ \mu\mathrm{s}$4, then decreased and plateaued at higher fluence. The reported optimal window for strong G-center activation with suppressed W-center emission was approximately $1\ \mu\mathrm{s}$5–$1\ \mu\mathrm{s}$6, corresponding to simulated local peak temperatures of approximately $1\ \mu\mathrm{s}$7–$1\ \mu\mathrm{s}$8. At $1\ \mu\mathrm{s}$9, the modeled surface peak temperature was approximately $10^8$0, still below silicon melting. Heating was confined to the top few hundred nanometers, and at $10^8$1 depth the temperature never exceeded approximately $10^8$2 (Andrini et al., 2023).

The optical metrics reported for activated ensembles were a lower-bound formation yield of $10^8$3 fabricated optically active emitters per implanted ion, FWHM values of approximately $10^8$4 at $10^8$5 and approximately $10^8$6 at $10^8$7, and a lifetime of $10^8$8 under $10^8$9 pulsed excitation for ensembles formed at $10^9\ \mathrm{K/s}$0. Debye–Waller factor and single-emitter brightness were not reported (Andrini et al., 2023).

The mechanistic interpretation is explicitly non-stationary. During nanosecond pulses, transient high temperature increases diffusion and reaction rates, but the brief duration $10^9\ \mathrm{K/s}$1 limits diffusion lengths through

$10^9\ \mathrm{K/s}$2

while the Arrhenius forms $10^9\ \mathrm{K/s}$3 and $10^9\ \mathrm{K/s}$4 describe the temperature dependence. The reported implication is that short, high-temperature excursions enhance the mobility of light carbon interstitials relative to silicon interstitials and enable capture at substitutional carbon sites before competing, more stable configurations dominate. This is presented as the basis for overcoming the structural metastability that limits G-center activation under conventional RTA (Andrini et al., 2023).

4. Hybrid single-photon-assisted two-photon polymerization

A polymer-based form of photonic curing is realized by combining sub-threshold single-photon absorption and focal two-photon polymerization. The resin was Sartomer PRO21905, an acrylate-based photocurable resin, containing Lucirin TPO-L at $10^9\ \mathrm{K/s}$5. TPO-L has high radical quantum yield, $10^9\ \mathrm{K/s}$6, for $10^9\ \mathrm{K/s}$7–$10^9\ \mathrm{K/s}$8 and also exhibits two-photon absorption. A continuous-wave blue source at $10^9\ \mathrm{K/s}$9 pre-excites the photoinitiator, while a mode-locked Ti:Sapphire source at $R_{1PA} = \sigma_1 I_1,$0, $R_{1PA} = \sigma_1 I_1,$1 repetition rate, and $R_{1PA} = \sigma_1 I_1,$2 pulse duration provides the missing energy to exceed the polymerization threshold only in the focal volume (Unlu et al., 2024).

The one-photon polymerization threshold dose was measured as $R_{1PA} = \sigma_1 I_1,$3, and all sensitization doses were kept below this threshold to avoid bulk curing. In point-exposure characterization, pure 2PP achieved a minimum lateral resolution of $R_{1PA} = \sigma_1 I_1,$4 at $R_{1PA} = \sigma_1 I_1,$5 using $R_{1PA} = \sigma_1 I_1,$6 average power and $R_{1PA} = \sigma_1 I_1,$7 exposure. With blue pre-sensitization, the same $R_{1PA} = \sigma_1 I_1,$8 voxel size was achieved with $R_{1PA} = \sigma_1 I_1,$9 femtosecond power and $\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$0 exposure. This is reported as a $\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$1 shorter exposure and $\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$2 lower femtosecond power. At the same $\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$3, pre-sensitization produced voxel sizes up to approximately $\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$4 larger than pure 2PP (Unlu et al., 2024).

The fitted voxel-growth laws distinguish the hybrid process from pure 2PP. Pure 2PP follows

$\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$5

whereas the dual-exposure process follows

$\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$6

The paper interprets the sum of two logarithms as evidence that the preparatory 1PA step and the nonlinearly confined 2PA step jointly contribute to onset and growth of polymerized features (Unlu et al., 2024).

The same logic was extended to a custom blue light-sheet-assisted 2PP printer. The $\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$7 writing beam was focused by a $\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$8 objective, and the blue light-sheet was generated from a $\frac{dn^*}{dt} = \sigma_1 I_1 N_0 - k n^*,$9 diode using a cylindrical achromatic doublet and a $R = 0.371$00 objective. The measured sheet waist was $R = 0.371$01 in air, approximately $R = 0.371$02 in resin with $R = 0.371$03, and the air Rayleigh range was approximately $R = 0.371$04, approximately $R = 0.371$05 in resin. The femtosecond focal FWHM was approximately $R = 0.371$06 at the rod surface (Unlu et al., 2024).

For 3D printing, the blue sheet used $R = 0.371$07 average power and $R = 0.371$08 exposure, still below $R = 0.371$09. A tall rectangular prism printed without the blue sheet required $R = 0.371$10 femtosecond power and $R = 0.371$11 exposure, and lower femtosecond powers failed. With blue-sheet sensitization at the same $R = 0.371$12 and $R = 0.371$13, larger voxels filled inter-voxel gaps and improved surface quality. Equivalent quality without blue required $R = 0.371$14 and $R = 0.371$15. The paper therefore reports a $R = 0.371$16 reduction in femtosecond power, a $R = 0.371$17 reduction in exposure, and a factor of $R = 0.371$18 reduction in the total light dose for comparable parts. A second 3D part printed successfully at $R = 0.371$19 and $R = 0.371$20 only with blue-sheet sensitization. Two layers were printed within a single $R = 0.371$21-thick sensitized sheet by shifting the $R = 0.371$22 focus axially by $R = 0.371$23 using SLM Fresnel masks, demonstrating axial resolution better than the sheet thickness by at least a factor of $R = 0.371$24 (Unlu et al., 2024).

A common misconception is that adding a one-photon exposure necessarily sacrifices two-photon depth sectioning. The reported design specifically keeps the $R = 0.371$25 dose sub-threshold, and polymerization occurs only in the three-dimensional intersection of the sensitized region and the femtosecond focal volume. This suggests that the speed advantage of 1PA and the axial confinement of 2PP can be combined without collapsing the process into bulk curing under the tested conditions (Unlu et al., 2024).

5. Dynamic refractive-index writing in DLP-cured micro-optics

In DLP-based photonic curing for micro-optics, the principal output is not only solidification but also a time-dependent optical function. The investigated photopolymer was PR48 Clear, composed by mass of Sartomer SR494 LM $R = 0.371$26, Allnex Ebecryl 8210 $R = 0.371$27, Rahn Genomer 1122 $R = 0.371$28, Esstech TPO+ $R = 0.371$29, and Mayzo OB+ $R = 0.371$30. The theoretical initial refractive index was $R = 0.371$31, matching measurements, and the maximum index rise was $R = 0.371$32. The Jacobs working-curve parameters were penetration depth $R = 0.371$33 and critical dose $R = 0.371$34 at an irradiance of $R = 0.371$35 (Heinrich et al., 13 Mar 2026).

The UV source was a Wintech4500 DLP system with pixels of $R = 0.371$36 including the inter-pixel dead zone. The “on row” mask comprised two adjacent pixel rows, producing $R = 0.371$37-wide illuminated stripes alternating with $R = 0.371$38 nominally dark stripes. Each pixel contained a circular non-reflective center, and the inter-pixel gap was also a dead zone. White-light interferometry of cured structures showed that the circular central dead zone reduced the plateau height by approximately $R = 0.371$39, while the inter-pixel dead zone showed near-zero height (Heinrich et al., 13 Mar 2026).

Time-resolved refractive-index metrology used a high-index prism with $R = 0.371$40 and a focused probe beam of approximately $R = 0.371$41 diameter. The critical-angle relation was

$R = 0.371$42

At $R = 0.371$43, the prism-interface refractive index followed the characteristic S-shaped curve of radical photopolymerization: an induction period dominated by oxygen inhibition, a rapid rise due to autoacceleration, and a slow approach to plateau as vitrification limits diffusion. In the experiment where UV was turned on at $R = 0.371$44, the onset of the rapid rise occurred after approximately $R = 0.371$45 of exposure, and the total refractive-index increase approached $R = 0.371$46 (Heinrich et al., 13 Mar 2026).

The intensity dependence of inhibition was quantified by the exponential fit

$R = 0.371$47

with tested irradiances of $R = 0.371$48–$R = 0.371$49. After the induction phase, the propagation rate was reported to scale linearly with irradiance,

$R = 0.371$50

These relations place oxygen inhibition and irradiance control at the center of process-window design (Heinrich et al., 13 Mar 2026).

Real-time diffraction monitoring treated the forming cure pattern as a volume phase grating. Diffraction built up after approximately $R = 0.371$51: the $R = 0.371$52th order began to decrease as energy redistributed into higher orders; the $R = 0.371$53st and $R = 0.371$54nd orders increased first; the $R = 0.371$55rd followed with slight delay. Deviations from an ideal sinusoidal thin phase grating appeared sequentially after approximately $R = 0.371$56 for the $R = 0.371$57nd order, approximately $R = 0.371$58 for the $R = 0.371$59rd, approximately $R = 0.371$60 for the $R = 0.371$61st, and approximately $R = 0.371$62 for the $R = 0.371$63th. A distinctive triple substructure—three maxima and two minima within each diffraction order—was reproduced experimentally and in simulation and traced to the discrete pixel structure and inter-pixel dead zones (Heinrich et al., 13 Mar 2026).

The Fourier optics model used

$R = 0.371$64

with phase

$R = 0.371$65

A baseline model using the measured S-shaped $R = 0.371$66 reproduced the $R = 0.371$67th and $R = 0.371$68st orders well up to approximately $R = 0.371$69 and the $R = 0.371$70nd order up to approximately $R = 0.371$71. An improved model added delayed polymerization in dark regions driven by scattered UV light and diffusion of radicalized oligomers, described phenomenologically by a delayed $R = 0.371$72 and related to the transport form

$R = 0.371$73

This improved agreement but still left a secondary peak in simulation not present in experiment, which the paper attributes to missing physics such as volumetric shrinkage, evolving scattering, and thermal gradients (Heinrich et al., 13 Mar 2026).

The practical significance is direct: for diffractive optical elements, dose should reach the target phase depth quickly without prolonged dark curing that erodes refractive-index contrast; for GRIN optics, grayscale patterning should be pre-compensated using the measured or simulated transfer function that includes dead zones, projector PSF, and radical diffusion. The work therefore frames photonic curing as a dynamic optical-writing problem rather than as a simple exposure-to-solidification step (Heinrich et al., 13 Mar 2026).

6. Confinement, size effects, and in-situ monitoring

Photonic curing in polymers is strongly size dependent. A dedicated study on BisEMA droplets, E-Shell 600 droplets, and SU-8 films showed that the time needed for complete curing increases as the polymerization volume is decreased below a characteristic threshold that depends on the specific reaction pathway. The in-situ signal was the intensity of the same $R = 0.371$74 laser backscattered from the curing region, denoted IBS, acquired with temporal resolution of a few tens of milliseconds. The curing-time metric was defined as $R = 0.371$75 such that $R = 0.371$76; samples exposed for at least $R = 0.371$77 were structurally robust after rinse or develop (Camposeo et al., 2022).

For BisEMA oligomer with $R = 0.371$78 w/w DMPA photoinitiator, monitored in dewetting-derived microdroplets with heights from $R = 0.371$79 to $R = 0.371$80, $R = 0.371$81 for droplet heights of $R = 0.371$82–$R = 0.371$83 under the $R = 0.371$84 confocal beam. As height decreased below approximately $R = 0.371$85, $R = 0.371$86 increased markedly, reaching approximately $R = 0.371$87 at $R = 0.371$88. Below approximately $R = 0.371$89 droplet height, no polymerization was observed even after $R = 0.371$90 minutes of continuous UV exposure at $R = 0.371$91; the study attributes this to oxygen inhibition (Camposeo et al., 2022).

For E-Shell 600, a free-radical acrylate-based resin containing a phosphine oxide photoinitiator, ex-situ Raman spectroscopy showed much faster curing at $R = 0.371$92. The conversion factor was defined as

$R = 0.371$93

and reached a plateau after approximately $R = 0.371$94 at $R = 0.371$95. Spatially resolved conversion for a droplet with total height $R = 0.371$96 showed a dead layer near the air interface: conversion was present only for $R = 0.371$97, implying an inhibited region of approximately $R = 0.371$98 (Camposeo et al., 2022).

For SU-8, the governing size effect was photothermal rather than oxygen-limited. SU-8 polymerizes mainly via photothermally enabled cationic mechanisms and has $R = 0.371$99. The paper uses Beer–Lambert attenuation,

$\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$00

with $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$01 at $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$02, and estimates the local steady-state temperature rise as

$\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$03

with $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$04 and $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$05. For a $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$06 film, $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$07 and $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$08, which places the local temperature above $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$09 and yields polymerization in less than one minute. For a $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$10 film, $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$11, and exposure times had to be increased by approximately a factor of $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$12 to achieve curing in about one minute. The reported thickness threshold for faster curing was approximately $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$13 (Camposeo et al., 2022).

The phenomenological model for size-dependent conversion was

$\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$14

with

$\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$15

In the thin-film, early-stage limit, the study gives

$\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$16

leading to a size-dependent growth law for $\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$17. To reproduce the early-time non-monotonicity of the backscatter signal, the paper multiplies a refractive-index-growth term and a transient-absorption term:

$\alpha = 7.69\times 10^5\ \mathrm{m}^{-1}$18

These results directly contradict the common simplification that smaller volumes necessarily cure faster. In the reported systems, confined volumes can be reaction-limited or diffusion-limited because oxygen diffusion, dead-layer formation, or reduced photothermal heating dominate the initial stages (Camposeo et al., 2022).

Taken together, the reported literature defines photonic curing by three recurring properties. First, the relevant control variable is not only total dose but also the spatiotemporal structure of energy delivery: nanosecond pulses, sub-threshold sensitization plus nonlinear finishing, or pixelated UV projection. Second, the operative kinetics are non-stationary: rapid heating and cooling in silicon, excited-state reservoir dynamics in hybrid 1PA+2PP, and oxygen inhibition–autoacceleration–vitrification sequences in photopolymerization. Third, the final functional output may be defect activation, voxel solidification, refractive-index contrast, or diffraction response. This suggests that rigorous implementation of photonic curing requires coupled control of light delivery, transport, and in-situ observables rather than reliance on nominal exposure alone.