Liquid handling: the physics of moving microlitres

Every workflow in this cluster — plate movement, colony picking, cell culture, seed trains — reduces at some point to one primitive: a defined volume of liquid must move from A to B. The arms and schedulers get the attention, but the hard physics sits in the pipetting head, because the volume is never measured directly. It is produced indirectly, by a moving air column or a pressure wave, and the liquid’s own surface physics decides how much actually arrives.

The air cushion does the measuring

In air-displacement pipetting, the standard mechanism of deck liquid handlers, a stepper-driven piston moves inside a tip, leaving a cushion of air between piston and liquid. The design volume is the piston displacement; the delivered volume is that displacement minus everything the cushion does. It compresses under the hydrostatic pressure of the liquid column, expands or contracts with any temperature difference between tip and fluid, and yields part of its space to vapour when the liquid is volatile. These are absolute errors, roughly fixed in microlitres rather than proportional to the set volume. That is why accuracy specifications — tested gravimetrically at 100, 50 and 10 % of nominal volume under ISO 8655 — hold their percent figures at full volume and degrade steeply near the bottom of the range: a fixed few-microlitre bias is invisible at 1000 µL and dominant at 20. The remedy is not a better motor but a shorter air column: small transfers go to narrow-bore tips, and viscous or volatile fluids go to positive-displacement heads, where the piston touches the liquid directly and there is no cushion to betray you.

Where the liquid fights back

The calibration fluid is water; almost nothing in a wet lab is water. Viscous glycerol or protein solutions flow through the orifice at a rate inversely proportional to viscosity, so a piston timed for water outruns the liquid and pulls air. Foaming media defeat capacitive liquid-level detection, which locates the meniscus by the change in tip capacitance: foam is a false surface, and the tip aspirates gas. Wetting cuts both ways. A pendant droplet left at the orifice after dispense — held there by surface tension — can, at a narrow-bore tip, approach the transfer volume itself; hence the touch-off against the well wall that wicks the residual. The standard countermeasures are pure fluid mechanics: reverse pipetting, where an excess is aspirated and discarded with the residue; slower piston speeds; and pre-wetting the tip so the retained film stops changing between strokes.

Sound instead of contact

Acoustic droplet ejection removes the tip entirely. A focused ultrasound pulse at megahertz frequencies concentrates enough pressure at the meniscus to eject a single droplet, and its size is fixed by the capillary wavelength — set by surface tension, density and frequency. That is how a machine delivers a few nanolitres without any moving part near the liquid; nothing touches the fluid, so there is no carryover and no consumable. The same physics sets the limits: clean ejection demands a well-defined surface tension, so aqueous buffers and DMSO compound libraries transfer, while viscous, surfactant-containing or particulate fluids — cell suspensions, many protein solutions — scatter into spray or fail to eject at all.

The vocabulary for the rest of this cluster

Everything downstream inherits these constraints. Gantry speed is limited by liquid settle time; closed systems exist partly because the open transfer is where fixed errors and contamination enter together. The fixed-error term, the fluid-property requirements and the volume floor of each mechanism are the coordinates the rest of this cluster navigates by.