Automation & robotics
The hard part is never the motor: microlitres, sloshing and settle time set the pace of all laboratory robotics.
In laboratory automation the bottleneck is almost never mechanics. It is the liquid: microlitre dispensing errors do not scale with set volume, plate transport pace is set by settle time after acceleration, and cell culture is governed by the asymmetry between the culture’s doubling time and a contaminant’s. Hence the cluster’s first recurring motif: automation is primarily the control of liquids and sterility, and only secondarily the motion of machines.
The second motif is interfaces. A single instrument is reliable; a chain of instruments is a distributed system where the schedule itself becomes an experimental variable and recovery from faults must be designed in advance. The third motif is scale-up: moving a process from flask to fermenter runs into heat removal, mixing time and oxygen transfer, which grow with geometry while the cell’s demands do not.
Start with liquid handling: the fixed microlitre error is the base unit from which everything else is derived.
Articles 9
| Liquid handling: the physics of moving microlitres A pipetting robot never measures volume directly — it moves an air column or a pressure wave, and every error source becomes relatively worse as the volume shrinks. |
| Gantry robots: motion above the wet deck A plate-moving gantry is not fast because its motors are fast — the real limit on speed is the free surface of the liquid inside the plates. |
| Wet-lab robotics: the orchestration layer Each instrument in an automated line works on its own; the engineering difficulty lives almost entirely in the connections between them. |
| Colony picking: a camera decides, a pin commits Picking converts a visible phenotype on a lawn into a physical culture, and every error of judgement becomes invisible the moment the pin touches the agar. |
| Automated cell culture: keeping a living product alive A culture is a warm nutrient bath that grows more slowly than its worst enemies — which is why sterility physics and surface-bound cells, not robotics, set the design. |
| Seed trains: arithmetic of an expansion chain The seed train exists because a vial holds six orders of magnitude too few cells — and every stage added to close that gap compounds both time and contamination risk. |
| Scale-up: why bigger vessels are different physics A bioreactor scaled up is not the same process made larger — transport processes scale with geometry while the cell's demands stay constant, and the gap is where bench success dies. |
| SCADA & MES: the control loop and its memory A bioreactor cannot be set and forgotten because the cells keep changing the plant they live in — and the batch record exists because a controlled process must be reconstructible. |
| Cloud laboratories: what the network actually carries A cloud lab moves a protocol description, not an experiment — and whether biology can be offered as a service is decided by specification, not by bandwidth. |