One air bubble ruins a measurement
Compressibility in a system that assumed everything was liquid.
A single trapped air bubble ruins a small-volume measurement because a plunger's exact travel is only converted into an exact volume delivered when everything ahead of it is genuinely incompressible, and the moment a pocket of air sits in that path some of the plunger's motion is spent squeezing the bubble smaller instead of pushing liquid out through the tip.
The promise a plunger relies on
A liquid-handling system works on a simple promise: move the plunger by a known distance and the liquid ahead of it is displaced by the matching, precisely calculable volume, because liquid barely shrinks however hard it is pushed. Air makes no such promise. It compresses readily under modest pressure, so a bubble anywhere in the liquid path acts as a small hidden spring that absorbs part of the plunger's travel. The instrument still moves its plunger by exactly the distance it was told to, the liquid at the far end moves by less, and nothing about the plunger's own motion reveals the shortfall.
A car's hydraulic brakes rely on the same promise. The fluid in the lines is close enough to incompressible that pressing the pedal a fixed distance squeezes the pads against the discs by a predictable amount every time. Let a little air into those lines, which is what a mechanic bleeding the brakes is trying to prevent, and the pedal develops a soft, spongy give before it engages anything, because the first part of every push is spent compressing the trapped air. The driver presses roughly as far as always while the stopping power at the wheels quietly drops, which is why bleeding the lines until no air escapes is treated as a safety procedure. A pipette or syringe pump with a bubble in its line fails in exactly this way, with a wrong volume in place of a soft pedal, and its control software has no foot on the pedal to feel anything going soft.
Twenty thousand times softer than water
Water is so stiff that raising its pressure by one atmosphere shrinks it by only about 0.005 percent of its volume. Air starting at atmospheric pressure, given the same one-atmosphere rise, halves in volume. Measured as stiffness, air at room pressure is roughly twenty thousand times easier to squeeze than water. A bubble filling even a small share of a line's internal volume can therefore absorb a share of plunger travel far out of proportion to its size, since almost all of the compression happening anywhere in that line happens inside the one pocket of air. A bubble of one microlitre in a line pushed to half an atmosphere above ambient would shrink by about a third of a microlitre, which is already a sizeable error against a dispense of a few microlitres.
An error that will not repeat
A worn seal or a mildly stretched tube produces an error that is consistent from one dispense to the next, which is inconvenient but correctable by calibration. A bubble is far less cooperative. It can grow or shrink between strokes as pressure in the line rises and falls, split into smaller bubbles or merge with another, and lodge in a different part of the flow path on different cycles depending on how it was carried along. Each change shifts how much of the plunger's travel is absorbed rather than delivered, so the volume error varies from stroke to stroke and cannot be calibrated away the way retention or a fixed mechanical offset can.
Catching a bubble by how the line pushes back
Keeping air out perfectly across thousands of cycles in a real machine is rarely achievable, so some systems watch the relationship between how hard the plunger has to push and how far it has travelled. A liquid path with no air in it resists a stroke with a characteristic stiffness that a path harbouring a compressible pocket does not match. A stroke that takes longer than expected to reach working pressure, or that yields under less force than a fully liquid-filled line should, is indirect evidence that something in the path is behaving like a spring, and a pressure sensor watching for that mismatch can flag a bubble that no inspection of the tubing would have caught in time.
Designing the air out
Removing a bubble after the fact is far harder than preventing it from forming or lodging in the first place, so every liquid-handling system that cares about small-volume accuracy is designed around keeping air out of the wetted path. Priming routines that flush the system before the first real measurement, tube geometry chosen to avoid pockets and dead corners where a bubble could sit undisturbed, and periodic checks built to catch a trapped bubble before it corrupts a run all follow from the same fact: a system built on an incompressible liquid path has no graceful way of coping with even one small pocket of a fluid that does not keep that promise. The next article in this set turns to exactly that problem, getting every trace of air out of a channel network before it is asked to measure anything at all.