Two-photon polymerization is a laser writing process that cures photoresin only where a femtosecond beam is tightly focused, giving true 3D sub-micron freeform structures; it works today and its historical throughput ceiling is now being broken by massively parallel multi-focus writing.
Summary
Two-photon polymerization (2PP, also called direct laser writing, DLW) exploits the fact that two-photon absorption scales non-linearly with intensity. A femtosecond laser is focused into a photosensitive resin, and polymerisation is confined to a voxel around the focus where the intensity is high enough for two photons to be absorbed effectively simultaneously. Everything the beam passes through on the way in stays liquid. Scanning the focus in three dimensions therefore builds arbitrary 3D geometry with sub-micron features, without layer supports or masks, and unpolymerised resin is washed away afterwards. The same non-linear confinement is what allows printing inside closed volumes, including inside living cells, where injected biocompatible resist has been cured into submicron structures such as a 10 micrometre elephant, cell-tracking barcodes, diffraction gratings and microlasers.
The parameters that decide the technology are voxel size, throughput and fidelity. Voxel size is set by wavelength, numerical aperture, resin chemistry and dose; short-wavelength sources help, and a 520 nm femtosecond source has produced microfluidic membrane channels down to 1 micrometre in commercial resins OrmoComp and FemtoBond. Throughput is the historic weakness because writing is serial, one voxel at a time. Fidelity is limited by physico-chemical effects (shrinkage, proximity dose accumulation, diffusion) that make the printed part deviate from the CAD model, and by the photopolymer itself: viscoelasticity and temperature-dependent stiffness cause hysteresis and drift in printed mechanisms.
All three parameters are moving. On throughput, holographic multi-focus writing with an LCoS spatial light modulator plus galvanometric scanning has reached more than 400 simultaneous foci at 1.49 x 10^8 voxels/s, roughly eight times the ~49-focus count previously treated as the SLM ceiling, with a companion DMD plus microlens-array system reaching a 1,600-focus array. A cheaper static route uses a single diffractive optical element to write 29 spots across two planes 1.8 micrometres apart simultaneously, giving 1 mm2 of four-layer woodpile in 90 s. On scale, 2PP is being hybridised with single-photon tomographic volumetric additive manufacturing so that centimetre-scale bodies with tens-of-micrometre features carry sub-micron 2PP detail in the same resin without intermediate processing. On fidelity, neural networks are being trained on experimental and theoretical data to predict deviations and pre-correct the write path.
A less obvious capability is that dose is a design variable, not just an on/off switch. Modulating the degree of conversion locally changes refractive index and stiffness, enabling grayscale and “(3+1)D” printing, now supported by an open-source slicer with a refractive-index calibration methodology. The same trick controls how much functional filler is incorporated: local dose modulation sets iron oxide nanoparticle loading in soft magnetic nanocomposites, yielding a microscale soft gripper and a bistable bit register with millimetre-scale elastic deformation.
Evidence base
- Holographic multi-foci 2PP with an LCoS-SLM plus galvanometric scanning reached more than 400 simultaneous foci at 1.49 x 10^8 voxels/s, roughly 8x the ~49-foci count previously framed as the SLM ceiling, with a companion DMD plus microlens-array system reaching 1,600 foci; the binding constraint moves to proximity effects and dose uniformity.
- A single static diffractive optical element generated 29 write spots across two planes 1.8 micrometres apart, fabricating four-layer woodpile structures at an effective rate of 1 mm2 in 90 s (arXiv, 17 April 2026).
- Stacked prism X-ray lenses made by 2PP were produced significantly faster and with higher geometric fidelity than by previous methods, with improved measured efficiency, but printing time and assembly remain unsolved (arXiv, 10 April 2026).
- Custom polymer microstructures, including a 10 micrometre elephant, cell-tracking barcodes, diffraction gratings and microlasers, were printed at submicron resolution inside living cells using injected biocompatible photoresist (16 June 2025).
- Plasmon-triggered 2PP localised quantum dots at a gold bipyramid tip, producing room-temperature Rabi splitting of 349.3 meV and coupling strength of 175.68 meV in a 3-QD system (5 November 2025).
- QuantumChem-200K released over 200,000 organic molecules annotated with eleven quantum-chemical properties including two-photon absorption cross sections, explicitly to unblock data-driven photoinitiator discovery for TPP (23 November 2025).
- Printed micro-optical mechanisms suffer hysteresis and drift from photopolymer viscoelasticity and temperature-dependent stiffness, requiring closed-loop control with an integrated Hall sensor and micromagnet (24 February 2026).
Open questions
- Do proximity effects and dose non-uniformity degrade feature fidelity as focus counts scale from tens to hundreds and thousands, and can learned pre-correction be applied per-focus in real time?
- For the stacked prism X-ray lens case, is printing time or post-print assembly the dominant cost, and does either scale acceptably to a full telescope aperture?
- Can data-driven photoinitiator design from datasets such as QuantumChem-200K actually deliver initiators with materially higher two-photon absorption cross sections, and how much writing speed does that buy?
- Are printed polymer mechanisms stable enough for deployed instruments, or does the viscoelastic drift documented in the lens scanner force integrated sensing into every precision application?
Assessment drafted 2026-08-31 from up to 18 KB sources using the technology-scorecard framework; scores are a draft read pending review.