Stereolithography (SLA)

How stereolithography works

Stereolithography turns a puddle of liquid into a solid object by writing with light. The resin in the vat is a photopolymer — a soup of small reactive molecules (monomers and oligomers) plus a photoinitiator that, when it absorbs a photon of the right wavelength, splits into a radical and kicks off a chain reaction. That reaction links the small molecules into a rigid three-dimensional network. Where the light has been, the liquid becomes solid; everywhere else it stays liquid. SLA exploits this with surgical precision: shine light on exactly the cross-section of one layer, and you have manufactured that layer.

Stack the layers and you have a part. There are two machine geometries:

• Top-down (free-surface) SLA. The classic industrial layout. The build platform sits just below the resin surface; a laser scans the top of the liquid; the platform then dips down by one layer thickness and a recoater blade sweeps a fresh film of resin over the top. The original Chuck Hull machines and most large-frame industrial SLA work this way.
• Bottom-up (inverted) SLA. The desktop layout used by Formlabs, and by every cheap MSLA/DLP printer. The build platform hangs upside-down and dips into a shallow vat with a transparent, non-stick film (FEP or nFEP) at the bottom. Light comes up through the film, cures one layer against it, and the platform lifts a layer thickness — peeling the new layer off the film. Each peel is a small mechanical event that limits speed and is the chief failure mode.

Whether the light is a steered laser dot (laser SLA), a whole-layer projector image (DLP), or a whole-layer LCD mask (MSLA), the physics of curing is identical. The only question is how light gets delivered to the right pixels.

The governing physics: the Jacobs cure equation

How deep does each flash of light cure? That is the central design equation of SLA, derived by Paul Jacobs at 3D Systems and used to tune every resin and every layer:

Cure depth (Jacobs equation):

    Cd = Dp · ln(E / Ec)

  Cd  = cured depth                       [µm or mm]
  Dp  = penetration depth of the resin    [µm]  (depth where intensity falls to 1/e)
  E   = exposure delivered (energy/area)  [mJ/cm²]
  Ec  = critical exposure to start curing [mJ/cm²]

Light attenuates exponentially inside the resin (Beer–Lambert):

    I(z) = I0 · exp(−z / Dp)

Curing only happens where the accumulated energy exceeds Ec, so
the gel front sits at the depth where E·exp(−z/Dp) = Ec → z = Dp·ln(E/Ec).

Two numbers fully characterize a resin's response to light: Dp (how deep light reaches) and Ec (the energy threshold to start a cure). You measure them by curing single-layer "windowpanes" at several exposures, plotting cured thickness against the natural log of exposure, and reading Dp off the slope and Ec off the x-intercept. Because cure depth grows with the logarithm of exposure, you get diminishing returns: doubling the cured depth requires squaring the exposure ratio E/Ec, not doubling it.

In practice you set per-layer exposure so the cure depth is slightly greater than the layer thickness — a deliberate overcure of typically 20–50 μm — so each new layer welds firmly to the one beneath it. Too little overcure and layers delaminate; too much and you cure resin into voids and lose downward-facing detail. The same equation explains a notorious artifact, print-through (light bleeding past the intended layer into the resin below), and why dark pigments and UV absorbers are added to resin to shrink Dp and tighten Z-resolution.

Worked example: exposure for a 50 μm layer

Take a typical 405 nm desktop resin with measured constants and find the laser/LCD exposure needed to cure a 50 μm layer with 25 μm of overcure (so a target cure depth of 75 μm):

Given:
  Dp = 120 µm          (penetration depth)
  Ec = 7 mJ/cm²        (critical exposure)
  Target Cd = 75 µm    (50 µm layer + 25 µm overcure)

Rearrange Jacobs:   E = Ec · exp(Cd / Dp)
                    E = 7 · exp(75 / 120)
                    E = 7 · exp(0.625)
                    E = 7 · 1.868
                    E ≈ 13.1 mJ/cm²

For an MSLA printer with UV-LED irradiance at the film of I ≈ 4 mW/cm²:
  exposure time t = E / I = 13.1 mJ/cm² ÷ 4 mW/cm² ≈ 3.3 s per layer.

A 60 mm-tall part at 50 µm layers = 1200 layers.
  Cure-only time = 1200 × 3.3 s ≈ 66 min,
  plus ~3–6 s of lift/peel/settle per layer → roughly 2–3.5 h total.

Notice that on an MSLA or DLP machine this time is independent of how much area you cured — one tiny ring or a plate packed with forty rings both flash for 3.3 s per layer. That is the killer economic feature of masked SLA: throughput is set by height, not volume. On a laser SLA machine the opposite holds — the galvo must trace every square millimeter, so a full plate takes far longer than a single small part.

SLA vs other 3D printing processes

	SLA (laser)	MSLA / DLP	FDM (FFF)	SLS (powder)
Mechanism	UV laser cures resin	LCD/projector cures resin	Extrude molten filament	Laser sinters powder bed
Typical layer thickness	25–100 μm	25–100 μm	100–300 μm	80–120 μm
Min. feature (XY)	~100–150 μm	~50–150 μm	~400–800 μm	~300–500 μm
Surface finish	Excellent (smooth)	Excellent (slight pixelation)	Visible layer lines	Grainy / matte
Supports needed	Yes (sacrificial)	Yes (sacrificial)	Yes (sacrificial)	No (powder self-supports)
Material toughness	Brittle (thermoset)	Brittle (thermoset)	Tough (thermoplastic)	Tough (nylon)
Throughput driver	Scanned area	Part height only	Path length + volume	Bed height + recoat
Post-processing	Wash + UV cure	Wash + UV cure	Remove supports	Depowder, blast
Best at	Fine detail, smooth surfaces	Cheap high-detail batches	Cheap functional parts	Durable functional parts

Real machines, resins, and numbers

System / spec	Type	Layer / resolution	Notes
3D Systems SLA 750	Industrial laser SLA	50–150 μm; dual 4 W lasers	750 × 750 × 550 mm build, high-throughput dual-beam
Formlabs Form 4	Desktop MSLA (LFS)	~50 μm; 4K mono LCD, ~5 s/layer flash	Industry-standard desktop; dental, jewelry, engineering resins
Generic 8K monochrome MSLA	Hobby/prosumer	~19–43 μm pixel pitch	Sub-$300 machines now hit ~20 μm XY pixels
Carbon DLS / CLIP	Continuous DLP	Continuous Z (no discrete layers)	Oxygen "dead zone" lets the part pull up continuously — 25–100× faster
Standard resin	Acrylate photopolymer	Tensile ~50–65 MPa; elongation ~6–12%	Cheap, high detail, brittle, low heat-deflection (~50–80 °C)
Tough / ABS-like resin	Toughened acrylate	Tensile ~45 MPa; elongation ~24–48%	Functional prototypes; trades detail and stiffness for impact
Castable resin	Wax-filled acrylate	Burns out cleanly <0.05% ash	Jewelry & dental investment casting patterns
Biocompatible resin	Certified (e.g. Class IIa)	Sterilizable	Surgical guides, dentures, hearing-aid shells

A few load-bearing figures: 405 nm is the dominant cure wavelength for modern resins (the photoinitiators and the cheap UV LEDs both peak there). Critical exposure Ec for common resins sits around 5–15 mJ/cm²; penetration depth Dp around 80–200 μm. Volumetric cure shrinkage of acrylate photopolymers is roughly 5–7%, which is why thin walls warp and large flat surfaces cup — anti-shrink chemistry and careful orientation are how you fight it. Continuous-DLP systems like Carbon's CLIP exploit an oxygen-inhibited "dead zone" a few tens of microns thick at the window, eliminating the peel step and printing tens to hundreds of millimeters per hour in Z.

Where SLA is used

• Dental. The single biggest production use of resin printing today. Surgical guides, orthodontic models for clear aligners (millions of arches printed for Invisalign-style workflows), night guards, temporary crowns, and full dentures. Detail and dimensional accuracy matter more than toughness, which is SLA's exact strength.
• Jewelry and investment casting. Castable resin patterns are printed at 25 μm layers, then burned out of an investment mold so molten metal can be cast in their place — fine filigree no wax carver could cut by hand.
• Hearing aids and audiology. Nearly every in-ear hearing-aid shell made in the last fifteen years is SLA-printed from a 3D ear scan; this was one of the first true mass-customization additive industries.
• Engineering prototypes and master patterns. Smooth, accurate concept models, fit-check parts, and masters for silicone molding / vacuum casting.
• Microfluidics and optics. Smooth surfaces and fine channels make SLA a favorite for lab-on-a-chip prototypes and even transparent resin lenses (with polishing).

Common misconceptions and pitfalls

• "SLA parts are production-strength plastic." They are thermosets, usually brittle, and most standard resins are notch-sensitive and UV-sensitive — they yellow and embrittle in sunlight over months. Treat untoughened SLA parts as detailed prototypes, not load-bearing production components, unless you have specifically chosen an engineering resin and validated it.
• Skipping or rushing the wash and post-cure. A green part left coated in uncured resin stays tacky, weak, and a skin sensitizer. The standard recipe is drain → wash in isopropyl alcohol (≥90%) → dry → UV post-cure. Under-curing leaves it weak; massively over-curing makes it brittle and yellow. Each resin has a calibrated post-cure time and temperature.
• Peel-force and orientation failures. On bottom-up printers the cured layer must peel off the FEP film every layer. Large flat layers parallel to the film generate huge suction and rip off the supports or cloud the film. The fix is to tilt the part 20–45° so each layer's cross-section is small and the peel front is a moving line, not a whole plane.
• Under-supporting overhangs and "islands." Because the part hangs in liquid, any feature that starts in mid-air with nothing cured below it (an "island") will simply float away and cure as debris stuck to the film — often wrecking the next layer. Slicer auto-support plus manual island checks are mandatory.
• Trapped resin / suction cups. Hollow models printed without drain holes trap liquid resin (heavy, wasteful) or create a vacuum cup against the film that exceeds peel force and fails the print. Add at least two drain holes (≥3 mm) to any hollowed part.
• Pouring resin or contaminated IPA down the drain. Uncured resin is toxic to aquatic life. Cure all waste solid under UV/sunlight before disposal; never let liquid resin enter wastewater.
• Confusing pixel/spot size with feature size. A 20 μm LCD pixel does not mean 20 μm features. Light bleed, the logarithmic cure law, and resin shrinkage push the smallest reliable positive feature to roughly 100–200 μm. Engraved details thinner than a few pixels simply fill in and vanish.