"3D printing" covers several unrelated ways of building a part, and they are not interchangeable. Most people picture the desktop filament machines, which is the family least able to do this work. This page sets out how each family builds, what it resolves, and which of them can produce a sealed channel at micron scale.
Everything below either deposits material and fuses it, or solidifies a liquid or powder in place. That mechanical difference is what decides whether a part comes out as one solid or as bonded pieces, and it is why only part of this list is relevant to microfluidics.
A heated nozzle extrudes molten filament in beads, line by line, layer by layer. This is the desktop machine most people mean by 3D printing. Resolution is bounded by nozzle diameter and layer height, and the part is a stack of fused strands rather than a continuous solid.
Not usable for microfluidics. Seams between beads leave channel walls neither smooth nor reliably sealed, and the finish is opaque.
A laser traces each layer in a vat of liquid photopolymer, curing it in place. Resolution is set by the optics rather than by a nozzle, and the finished part is a single continuous solid. This is the family our process is built for.
Reaches microfluidic features. Clear resins, smooth walls, and no seam between strands.
The same chemistry as SLA, but a whole layer is projected at once by a digital micromirror device or an LCD panel instead of being traced by a laser. That makes them faster per layer, and resolution is set by the pixel pitch of the projector or screen. Purpose built microfluidic printers usually sit here.
Reaches microfluidic features. Same post processing problem as SLA, and the same solution.
A laser or a fusing agent sinters powdered polymer bed layer by layer, with unfused powder supporting the part. Excellent for functional mechanical parts with complex geometry and no support structures.
Not usable for microfluidics. The surface is porous and grainy by nature, and loose powder cannot be cleared from an enclosed micron scale channel.
Print heads spray droplets of photopolymer and cure them immediately, often with a dissolvable support material printed alongside. Very fine layers, multiple materials in one part.
Partially applicable. Resolution is there, but support material has to be cleared from inside the channels, which is the same clearing problem in a different form, and the machines are expensive.
Two photons converge to cure a point inside the resin volume, writing structures well below a micron. The finest additive resolution available.
Applicable but not scalable. Build volumes and write speeds put it in the research instrument category rather than the production one.
| Family | How it builds | Surface | Optical clarity | Sealed micron channels |
|---|---|---|---|---|
| FDM | Extruded beads | Ridged, seamed | Opaque | No |
| SLA | Laser cured resin | Smooth, continuous | Clear | Yes |
| DLP and LCD | Projected layer cure | Smooth, continuous | Clear | Yes |
| SLS and MJF | Sintered powder | Porous, grainy | Opaque | No, powder trapped |
| Material jetting | Jetted droplets | Smooth | Varies by material | Support must be cleared |
| Two photon | Point cure in volume | Sub micron | Clear | Yes, but not at scale |
Green means suitable for this work, oxblood rules it out, and grey is conditional. Characteristics described here are general properties of each family, not measurements of specific machines.
Only the resin families produce a clear, smooth, continuous part with a sealed channel inside it, which is why the whole conversation about printed microfluidics is a conversation about SLA, DLP and LCD. Every one of them ends with the same unsolved step: clearing the uncured resin and curing what is left.
Capital cost tracks the mechanism, not the marketing. The families that cannot make a microfluidic part are the cheap ones, and the families that make the finest features are sold the way production equipment is sold, on a quote.
| Family | Entry cost | Professional and industrial | The limitation you are paying around |
|---|---|---|---|
| FDM | $200 to $500 | $5,000 into six figures | No amount of money makes a bead into a sealed micron channel |
| SLA | Low thousands | $2,000 to over $200,000 | Resolution rises with the optics, and so does the price |
| DLP and LCD | Low thousands | Microfluidic specific machines quote only | The finest closed channels carry production equipment pricing |
| SLS and MJF | None; bureau access only | $5,000 into six figures | Porous surface and trapped powder, at any price |
| Material jetting | None at desktop scale | Five to six figures | Support material still has to be cleared from inside the channel |
| Two photon | None | Quote only, research instrument tier | Build volume and write speed, not resolution |
General ranges as published by 3dprinting.com's 2026 cost guide. Microfluidic specific DLP hardware such as the CADworks3D ProFluidics 285D carries no public list price; that page publishes 28.5 µm pixel resolution and 80 µm closed channels but quotes on request. These are industry guide figures rather than vendor quotations, and they are wider and softer than the equipment figures on our PDMS and cleanroom pages.
The pattern worth noticing: cost buys resolution, and nothing else on this list. It does not buy a way out of the finishing step. Every resin machine in every tier hands back a part still coated in uncured resin, and the price of the printer has no bearing on that.
The printing industry solved the geometry. What it did not solve was what happens to the part after the print, and that is the step we changed.
See what we built