Why bubbles are a design problem rather than an accident

Designing channels so that gas cannot lodge in them.

Bubbles lodge in a manifold through two entirely predictable mechanisms, gas rising to whatever local high point a channel offers and a sharp corner pinning a bubble in place against the flow trying to carry it away, and a channel that has not been deliberately shaped to deny both a foothold will collect bubbles reliably.

Buoyancy finds the high point, a corner holds it there

A bubble is lighter than the liquid it displaces, so it feels the same upward push that lifts anything less dense than its surroundings, and inside a channel that push sends it toward the highest point in the local geometry, just as air collects at the top of any container. Once there, a second mechanism can stop it leaving. A sharp internal corner gives the bubble's surface a stable shape to pin itself against, with the same surface tension that makes a droplet round now holding the bubble's boundary fixed. One local high point paired with one sharp corner is enough to make a bubble effectively permanent, and a pinned pocket occupying only part of a channel's cross-section can still choke off a large share of the flow past that point.

Pushing the air to an open edge

Laying an adhesive screen protector flat on a phone and pressing from the centre traps air readily, since air caught beneath has nowhere obvious to go and ends up pinned in a stubborn bubble. The standard technique works the film down from one edge with a squeegee, pushing a continuous front of air toward the nearest open edge so it always has a clear path out. A bubble-free manifold follows the same principle: every point capable of collecting gas is connected by a continuously rising path to an actual vent, so gas arriving there has somewhere to go.

That is why manifold layouts are so often drawn with a deliberate overall gradient. A channel that dips and then rises again creates a local high point for buoyancy to send a bubble toward, and if that point is not connected onward to a vent, the bubble stays however vigorously liquid flows beneath it. A channel that slopes continuously in one direction, with no unvented high points, removes the buoyancy half of the trapping mechanism entirely. The gradient has nothing to do with appearance and everything to do with which way gas will drift once liquid flows through the finished part.

When the gas is supposed to be there

The segmented-flow design mentioned earlier in this set, where a plug of gas is introduced on purpose between liquid samples, reverses the goal. There the geometry is shaped to keep the plug moving with the flow, and the gradients and vents that banish an unwanted bubble elsewhere would work against a channel whose job depends on the gas plug staying in the stream. Whether a channel treats gas as a contaminant to evict or as a working fluid to keep moving has to be settled before any of these rules can be applied to it.

For every other channel, preventing bubbles is a drawing-stage decision: continuous slope, rounded internal corners, and a vent at every genuine high point. A longer prime cycle, or hoping ordinary flow will eventually carry every bubble along, cannot substitute for that geometry. A channel shaped this way denies gas both the resting place buoyancy carries it toward and the pinning corner it would need to stay. Reviewing a drawing with two questions in mind, where the local high point is and whether it leads to a vent, turns bubble prevention into a check a reviewer can apply feature by feature before the block is cut, and a design review that only asks whether a corner looks small enough to ignore is checking the wrong property.