Micro additive process guide

Two Photons Polymerization Micro 3D Printing: A Step-by-Step Method for Long Overhangs

This process builds micro parts by curing resin only at a femtosecond laser focus. The guide is for engineers who need overhanging micro features without support removal. You will learn the parameter window, the failure modes, and the point where CNC becomes the cheaper route.

Get Instant Quote Talk to an Engineer

Overhang without supports150–200 μm cantilever exampleShell + solid modesCleanroom-friendly resins

Two photons polymerization micro 3d printing of a micro part for medical use

On this page

Key takeaways

Overhangs are a slicing problem, not a chemistry problemA long cantilever fails because each layer is written after the one below it, so the focal point drifts in partly cured resin.

Local print-size changes fix most of itSmaller voxel spacing near the overhang anchors the first few layers, then larger blocks build the bulk faster.

Shell plus solid beats pure shellA thin shell alone buckles under its own weight past roughly 100 μm of unsupported span.

Know when to stopBelow 50 μm features with tight tolerances often belong on a mill, not a laser.

Mechanism

How Two Photons Polymerization Micro 3D Printing Cures Resin

A femtosecond laser is focused into a photoresist. Two photons must arrive at the same spot within a few hundred attoseconds for the initiator to absorb them. Outside that tiny focal volume, single photons pass through without starting a chain reaction. That is why the cured voxel is smaller than the diffraction limit of the beam.

Typical voxel diameters run from 100 nm to a few micrometres, depending on laser power, exposure time, and the photoinitiator. Raise the power and the voxel grows. Shorten the dwell time and it shrinks. The process writes in three dimensions, so a layer can be placed above empty space as long as the surrounding material holds it.

Long overhangs break that assumption. Each new line is written into liquid resin that has already been partially exposed by scattered light. The resin there is not solid, but it is not virgin either. Its stiffness is a fraction of the fully cured value, so the focal point wanders and the line lands offset from where the slicer planned.

The drift compounds with distance. A 1,000 μm cantilever written as a single pass can end up curved, wavy, or detached at the free end. Support structures can hold it, but removing them from a 50 μm feature with a probe is slow and often damages the part.

1
Voxel size drives everything100 nm to a few micrometres, set by power and dwell time.
2
Partially cured neighbors matterScattered light stiffens nearby resin and shifts the focal spot.
3
Support removal is the hidden costAt 50 μm features, manual probing is slow and risky.

Design rules

Which Geometries Suit an Overhang Without Supports

Not every micro part needs this method. A cantilever, a bridge between two posts, or a curved arm that leaves the build plate are the clear candidates. So is anything where the support scar would sit on a functional surface, such as a microfluidic channel wall or an optical facet.

A flat plate with no overhang does not need it. Neither does a part under 200 μm in total height. In those cases a standard raster scan at constant voxel size is faster and easier to qualify.

The rule of thumb we use: if the unsupported span exceeds five times the local cross-section width, plan for local print-size adjustment. If it exceeds twenty times, expect to add a scaffold that you later dissolve rather than mechanically break off.

Also check the aspect ratio of the free end. A 50 μm wide cantilever that is 50 μm thick behaves differently from a 50 μm wide ribbon that is only 5 μm thick. The thin ribbon cools and shrinks faster between passes, which adds curl on top of the focal drift.

1
Good candidatesCantilevers, bridges, curved arms, any surface where support scars are unacceptable.
2
Poor candidatesFlat plates, parts under 200 μm tall, open lattices with no load path.
3
Span ruleOver 5× local width: adjust print size. Over 20×: use a dissolvable scaffold.

Parameters

Parameter Windows That Keep Long Overhangs Straight

Laser power between 10 mW and 30 mW covers most acrylate and epoxy resists. Start at 15 mW and move in 2 mW steps. Below 10 mW the line often fails to attach to the layer beneath. Above 30 mW the voxel balloons and you lose the sub-micrometre edges that justified the process.

Write speed runs from 0.1 mm/s to 10 mm/s. Slow speeds give stronger anchoring but let scattered light pre-stiffen a wider halo. For an overhang's first three to five layers, drop to 0.5–1 mm/s so the anchor is solid. After that, raise toward 5 mm/s to build the bulk.

Voxel spacing, also called hatching distance, should be about 60–80% of the measured voxel diameter. Tighter spacing overlaps too much and warps the line. Looser spacing leaves uncured gaps that collapse under the next layer.

Resist choice matters less than people expect. A high-initiator resist cures faster at the focus, which helps anchoring, but it also scatters more. If your overhang drifts, try lowering the initiator concentration before you touch the laser.

1
Power10–30 mW; start at 15 mW, step by 2 mW.
2
Speed0.5–1 mm/s for anchor layers, up to 5 mm/s for bulk.
3
Hatching distance60–80% of measured voxel diameter.

Procedure

Step-by-Step: Printing a Long Overhang

Order matters. Skipping the anchor step is the most common cause of a drooping cantilever.

1
1. Measure the voxel firstWrite isolated single lines at 10, 15, and 20 mW on a test substrate. Image them under SEM and record the width. Do not trust the datasheet value. Your resist, substrate, and objective all shift it.
2
2. Slice the overhang into three zonesMark the anchor zone (first 3–5 layers of the overhang), the transition zone (next 10–20 layers), and the bulk zone. Each zone gets its own spacing and speed.
3
3. Set the anchor zone small and slowUse 60% of your normal hatching distance and 0.5–1 mm/s. This is the only zone where you accept a longer print time. It prevents the free end from lifting.
4
4. Ramp the transition zoneIncrease hatching distance by 5% per layer and speed by 0.3 mm/s per layer. A linear ramp avoids a sudden stiffness change that can crack the interface.
5
5. Switch to shell plus solid for the bulkWrite a 2–3 μm shell around the outer surface, then fill the interior with solid blocks at 80% hatching distance. The shell carries the surface finish, the solid core carries the load.
6
6. Check for focal drift every 100 μmStop the job, image the free end, and compare it to the CAD. If it has moved more than 2 μm, reduce laser power by 2 mW before continuing. Do not compensate in software alone.
7
7. Develop, then rinse in two bathsUse the resist maker's developer. Two baths, 3 minutes each, with gentle agitation. A single bath leaves uncured resin in the shadow of the overhang, which later sags.
8
8. Cure under the specified wavelengthFlood cure at the wavelength the initiator absorbs. Undercured overhangs creep over days; overcurved ones crack. Follow the datasheet time, then measure again.

Decision table

When to Use TPP, DLW, or CNC for Micro Features

Pick the row that matches your smallest feature and your overhang span.

Method	Smallest feature	Overhang limit	Best for
Two photons polymerization	100–300 nm	~1000 μm with adjusted print size	Optical and microfluidic overhangs
Direct laser writing (DLW)	~1 μm	~300 μm without supports	Larger micro parts, faster build
Micro CNC milling	50–100 μm	No limit (rigid tool)	Metal parts, tight tolerances
CNC turning	20–50 μm	No limit	Axial micro features, round parts
Injection micro-molding	10–50 μm	No limit	Runs above 10,000 parts

The short verdict

Use two photons polymerization micro 3d printing when your smallest feature is under 50 μm and the overhang span is under 1,000 μm. Use CNC when you need metal, tighter than ±0.005 mm, or a finish below Ra 0.8 μm.

Get Instant Quote Talk to an Engineer

FAQs

Common questions

Why does my overhang droop even with supports?

Supports hold the structure, but they do not stop the focal drift caused by scattered light. The resin around the support is partly cured and stiffer than virgin resist, so the beam lands off target.

Reduce laser power by 2–3 mW in the overhang zone and lower the hatching distance to 60% of the voxel diameter. That fixes more droop cases than adding more supports.

Can I print a 1,000 μm cantilever in one pass?

Yes, with the zone method. Print the anchor zone at 0.5–1 mm/s, ramp the transition, and use shell plus solid for the bulk. The 1,000 μm cantilever we test with is 50 μm wide and 50 μm thick.

A single-pass job at constant speed usually curves. The free end is where the error shows first, so inspect it before you run the full batch.

What resist should I use for overhangs?

A standard acrylate or epoxy resist with a moderate initiator concentration works for most cases. High-initiator resists anchor faster but scatter more, which hurts long spans.

If your geometry has spans over 500 μm, test both and compare the free-end deviation in SEM. The difference is usually visible within one build.

How do I know when to switch to CNC instead?

Switch when the smallest feature is above 50 μm, when the part is metal, or when the tolerance is tighter than ±0.005 mm. TPP wins on feature size and overhang freedom; it loses on material choice and surface finish.

A 50 μm feature in aluminium on a micro mill is a routine job. The same feature printed in resin may need post-processing that costs more than the milling.

Does the overhang method work for metal parts?

Not directly. Two photons polymerization cures polymer. Metal micro parts are usually made by printing a polymer scaffold, then debinding and sintering, which shrinks the part by 15–20%.

Shrinkage changes your overhang geometry, so the anchor zone must be designed oversized. For tight metal micro features, micro milling is more predictable.

What inspection do you recommend for micro overhangs?

SEM imaging at the free end and at the anchor, plus a dimensional check on the overall span. Optical profilometry works for surface finish but struggles on vertical overhang faces.

Check the free end before and after final cure. If it moved between the two, the cure schedule, not the print, is the problem.