Monolithic vs. two-piece microfluidic chips

OpenChip can build your chip three ways.

Monolithic — one sealed block, channels as internal voids.

Two-piece — split at a cut plane into a channel plate and a lid, bonded after printing. This is the default.

Glass slide — one open-faced part with a pocket milled for a glass microscope slide, which becomes the lid.

The short version

	Monolithic	Two-piece	Glass slide
Drainage	Hard, sometimes impossible	Trivial — open grooves	Trivial — open grooves
Leaks	Nothing to leak	A bond that can	A bond that can
Assembly	None	Align and glue	Drop in and bond
Optical access	Through the block	Through a printed lid	Through glass
Minimum channel	~0.5 mm realistically	Down to the printer’s limit	Down to the printer’s limit
Print time	One part	Two, but both thinner	One, and the thinnest
Ports	Anywhere	Anywhere	Only outside the pocket

If you are not sure, use two-piece. It is the default for good reason.

Why two-piece is usually right

The channels are open until the lid goes on. Uncured resin rinses straight out, which removes the single biggest failure mode in resin microfluidics — see Draining uncured resin.

It also lets you go smaller. Below about 0.5 mm, draining a sealed channel stops being reliably possible; as an open groove, the limit is just the printer’s resolution.

And it gives you options a sealed block does not: a glass lid for optical work, a membrane between two layers, or a lid you remove to clean and re-bond.

The cost is a bond, and a bond is a seam that can leak. Getting a good one is its own skill — see Bonding and sealing a two-piece chip.

When monolithic is right

Large channels, short paths. A chip with 1 mm channels and a direct route between two ports drains fine. If you can flush it clear in a minute, the argument against monolithic mostly evaporates.

High pressure. No seam means no seam to fail. If you are running near what the material will take, monolithic is stronger.

Complex 3D routing. Channels crossing over each other at different heights cannot be expressed as a single split plane. (OpenChip’s channel graph is planar at a single Z, so this is more a note about where you would need another tool.)

Nothing to align. No jig, no clamp, no cure time. For a quick one-off, it can be genuinely faster.

The split height

Where the cut plane sits, from the bottom of the block. 2 mm by default, the mid-plane of the default 4 mm chip.

Rectangular channels are placed so their ceiling sits at the cut plane, which gives you a flat lid with no features on it — the easiest thing to bond and the easiest thing to replace with glass.

Raise the split height for a thicker channel plate and a thinner lid; lower it for the reverse.

Capping with a glass slide

A printed lid is as flat as your printer is, which is not very. If you want to see into the channel — for a dye front, a droplet count, anything on a microscope — the lid is the thing in the way.

Glass slide replaces it with a plain microscope slide. The chip prints open-faced, with a pocket cut into the top the size of the slide plus a fit clearance, and the pocket floor is the plane the channels open onto. You rinse it out, drop the glass in, bond it, and the channels are closed by a flat optical surface that cost a few pence.

Set the slide’s size on the CAD panel: width, length and thickness are up front, and Advanced carries the fit clearance, a bond gap for adhesive, and an offset if the pocket should not sit in the middle of the chip. The defaults describe a standard 75 × 25 × 1 mm slide.

What it changes about your design

Ports have to live outside the pocket. A socket is bored down from the top face, so a port under the glass opens into the pocket floor and the slide bonds over it. Put them on the rim and run the channel out to them; port-under-slide names any that are covered.

A channel outside the pocket keeps its roof. Only the pocket opens the channels. The run out to a rim port prints as a closed tunnel, which is fine and necessary — just check it can drain. channel-outside-slide tells you which segments those are.

Round channels become half-round. A round section is centred on the open face, so the pocket takes the top half off and the glass closes a semicircle holding about half the area. The flow estimates assume the full circle. Use rect unless you want the groove shape — round-channel-under-slide is the reminder.

The chip has to be big enough. The standard slide does not fit the default chip. Fit the chip to the slide on the panel resizes it, leaving your minimum edge spacing as the rim.

Bonding the glass

Same choices as any two-piece bond, with one difference: glass does not cure. UV-curing adhesive works through it rather than in spite of it, and a resin bond line has to be cured from the printed side. Add a bond gap under Advanced if you want the pocket cut deeper than the glass is thick to leave room for the adhesive. See Bonding and sealing.

Checking it in the preview

Explode in the CAD toolbar separates the two halves so you can see the bonding faces. On a glass-slide chip it lifts the glass off the plate instead; on a monolithic chip the button is disabled.

The glass itself is drawn translucent in the preview so you can see what it covers. It is never exported: the STL holds the printed part only, because glass is bought, not printed.

Worth doing before every print. It is where you notice that a channel is open at the split plane where you did not mean it to be, or that the lid has a feature on it you did not expect.

What the checks do differently

The drainage checks — enclosed-cavity and undrained-region — only run on an enclosed chip. A two-piece split is open by construction, and so is the pocket on a slide chip, so they stay quiet on both.

That is not the checks being lenient. It is that the problem does not exist.

Switching later

Change Split type in the CAD panel at any time. The channel network is unaffected — it is the same graph whichever you pick — and the solid rebuilds.

Nothing is lost by trying both and looking at them.