CNC Machining Technology in Microfluidics

• 1
  What CNC machining technology in microfluidics actually removes
• 2
  Which machine and spindle suit micro channels
• 3
  Tolerances, aspect ratio and what the wall can take
• 4
  Material choice changes the cut and the assay
• 5
  Deburring, bonding and leak testing
• 6
  Machining versus molding versus casting for microfluidic parts

CNC machining technology is subtractive. A rotating cutter or a single-point tool enters a solid block and removes material along a programmed path. In microfluidics the block is usually a polymer or metal plate, and the removed material forms the channel network. Nothing is built up, so the channel is defined by the tool edge, not by a mask or a mold.

That difference matters at small scale. Photolithography and soft lithography build channels by curing a polymer against a master, so a new design needs a new master. Milling cuts a fresh geometry straight from the CAD file. You can change one channel width between two parts and cut both on the same day.

The trade-off is surface quality. A milled floor carries tool marks that a molded or lithographic channel does not. Those marks change flow resistance and can trap bubbles. So the question is rarely whether CNC can cut a microfluidic feature. It is whether the finish you get is good enough for the fluid behavior you need.

For most chips, it is. Channels at 200–500 μm wide with Ra 0.8–1.6 μm behave predictably for reagent routing, mixing and droplet generation. Below 100 μm the finish and the burr become the limiting factor, not the machine positioning.

Channel width sets the tool. A 200 μm channel needs a cutter smaller than 200 μm, often 100–150 μm, so the tool can pass without rubbing both walls. Tools that small deflect. Cutting force has to drop, which means shallow axial depth and light radial engagement.

High-speed spindles do the work here. At 30,000–60,000 rpm a 0.5 mm tool keeps a sane chip load while the surface speed stays high. A slow spindle forces you to feed slowly, and slow feed at small depth per tooth rubs the material instead of cutting it.

Rigid, thermally stable machines matter more than raw axis count for this work. Our compact travels of 500 × 500 × 450 mm and 500 × 310 × 200 mm are the frames we use for micro features, because a smaller envelope holds position better against spindle heat. For larger manifolds, the 750 × 1,150 × 550 mm travel takes the same tooling strategy.

Coolant choice is not neutral. Flood coolant clears chips but can lift thin walls. Air blast with a light mist suits 0.2–0.5 mm walls in PMMA and POM. On metals, flood coolant with a high-pressure through-tool option keeps deep, narrow slots from packing chips.

We hold ±0.005 mm (±0.0002 in) on critical features. That number covers channel width, wall thickness and the position of a port relative to a channel. It does not mean every surface on the part is held that tight. Tighter tolerance on a non-functional face only adds cost and time.

Aspect ratio is the real constraint. A channel 300 μm wide and 1,000 μm deep is a 3:1 slot. The tool needs to reach the bottom without snapping and still clear chips. Past roughly 5:1 in plastics, tool breakage and taper start to dominate, and the floor gets wider than the entry.

Wall thickness between channels behaves the same way. A 250 μm wall between two 300 μm channels will flex under cutting load. If the wall is load-bearing or must stay optically clear, design it thicker or accept a slower, lighter pass sequence with an intermediate stress-relief step.

Depth control on the floor is where finish and flatness meet. A 20 μm depth error across a 50 mm chip changes the hydraulic resistance of a long channel by a few percent. If your assay depends on precise residence time, specify floor depth and floor flatness separately, not just channel width.

PMMA is the usual starting point. It cuts cleanly, is optically clear and bonds by solvent or thermal methods. It also scratches and stress-cracks if you take heavy passes. Keep depth of cut light and use a sharp two-flute tool.

COC and COP hold up better to some solvents and have lower water absorption. They are more prone to crazing at the cut edge, so tool sharpness matters more than speed. PEEK machines well and takes autoclave cycles, but it is expensive and abrasive on tooling.

Metals enter the picture when the chip is a mold insert or a metal manifold. Aluminum 6061 and stainless 316L both machine to the tolerances above. Stainless work-hardens, so avoid dwelling in the cut. Titanium TC4 (Ti-6Al-4V) and Inconel are only worth it when corrosion or temperature demands them.

For mold inserts, hardness drives the decision. Tool steel and 17-4PH give longer insert life than aluminum, at higher cost and longer lead time. If you only need a few hundred parts, aluminum inserts are usually enough.

A machined channel has a burr at the top edge. In a 300 μm channel that burr can bridge a third of the width. Deburring has to be controlled, not aggressive. Bead blasting at low pressure, tumbling with fine media and controlled chemical polishing all work; each removes some material, so leave stock for it.

Bonding decides whether the chip works. Thermal bonding needs flat, clean surfaces and matched thermal expansion. Solvent bonding softens the surface and can round the channel edge if the solvent reaches the channel. Plasma activation followed by thermal bonding gives the strongest bond with the least channel distortion.

Surface roughness affects bond strength as much as flow. Ra 0.2–0.8 μm gives more contact area than an as-machined Ra 1.6–3.2 μm surface, but it also traps more contamination if cleaning is incomplete. Rinse and dry before bonding, and verify with a contact-angle check.

Leak testing closes the loop. Pressure-decay testing at 1–2 bar catches most bonding failures. Dye penetration finds channel-to-channel leaks that pressure alone may miss. Every part we ship gets 100% inspection before shipment, with reports on request.