Bonding and sealing a two-piece chip

A two-piece chip is easy to print and easy to drain. The bond is where the work is.

Two failure modes, and they pull in opposite directions: too little adhesive and it leaks; too much and it wicks into the channels and blocks them.

Before you bond

Both pieces must be completely washed, dried and post-cured. Bonding a tacky surface gives a weak joint and traps uncured resin at the seam.

Check the bonding faces are flat. A slight warp from printing is common on a thin lid. Lapping on wet-and-dry paper on a flat surface — 800 then 1500 grit, light pressure, figure-of-eight — fixes it. Wash and dry again afterwards.

Test-fit dry. Hold them together against a light and look for gaps.

Methods

UV resin (the usual answer)

The same resin you printed with, cured in place.

Apply a very thin film to the lid, not the channel plate — this is the single most useful trick. Adhesive on the flat piece cannot run into a groove.
Spread it to an even, barely-wet film. A lint-free wipe or a squeegee edge.
Bring the pieces together with a slight sliding motion to exclude air.
Clamp lightly and evenly. Two flat plates and a couple of small clips.
Cure under UV, several minutes, both sides.

Advantages — same material, so no thermal or chemical mismatch; you already have it; strong.

Risk — wicking. Resin is low-viscosity and channels are capillaries. Too much and it is drawn straight in.

Solvent bonding

A solvent that softens the surface, so the two pieces fuse.

Results with photopolymers are inconsistent — it depends entirely on the resin’s chemistry, and manufacturers do not publish it. Some soften usefully in dichloromethane; many just craze.

If you try it, test on offcuts. Crazing is not always visible until the chip is under pressure.

Adhesive film

Double-sided pressure-sensitive adhesive, cut to the channel pattern.

Advantages — no wicking at all, no cure time, reversible if you use a removable adhesive.

Disadvantages — needs a cut pattern, so a laser cutter or a plotter; adhesive chemistry contacts your fluid; lower pressure rating.

Good for prototyping and for anything you want to open again.

Mechanical clamping with a gasket

A compressible gasket — silicone sheet, or an O-ring in a groove — and screws or clips.

Advantages — fully reversible, no adhesive in contact with the fluid, and you can take it apart to clean.

Disadvantages — bulky, needs fastener holes designed in (export OpenSCAD and add them), and the gasket must not intrude into the channels.

The right answer for anything you will disassemble.

Testing the bond

Before you put anything valuable through it:

Flush with water and look for external leaks.
Pressure test. Cap one port, push water in the other, hold at moderate pressure for a minute. A syringe by hand is enough to find the obvious failures.
Dye test. Flush with dyed water. Look for dye in an adjacent channel — that is an internal leak through a thin wall, not a bond failure, and it means the channels-too-close warning was right.
Check the channels are clear. Hold to a light. Adhesive that wicked in during bonding is visible as a change in the channel profile.

When it leaks

At the edge — insufficient adhesive, or a warped face. Re-lap and re-bond.

Between adjacent channels — the wall is too thin. That is a design problem, not a bonding one. Check channels-too-close and move a channel.

At a port — usually the fitting rather than the bond. See Ports and fittings.

Only under pressure — a partial bond. Any of the methods above will do this if the clamp was uneven. Re-do it with a flat plate on each side.