Water is easy. It flows at the plunger's speed, lets go of the tip cleanly, and forgives a class that is only roughly right. Everything else is where liquid class optimization earns its keep. The difficult liquids, the thick ones, the volatile ones, the foamers, the clingers, each fail in a characteristic way, and each responds to a characteristic set of parameter moves. Optimization is much faster when you start from the right recipe for the family instead of discovering it symptom by symptom. What follows is a set of starting points, not final answers: clone the closest validated class, apply the moves for the family, then measure and refine as the optimization loop demands.
Viscous liquids: glycerol, oils, concentrated sugars, bead slurries
Viscosity is resistance to flow, and every viscous-liquid problem traces back to the plunger outrunning the fluid. Aspirate too fast and you pull air in behind the liquid; dispense too fast and you leave a slug behind in the tip. So the whole recipe is patience.
- Flow rate: drop it hard. Ten percent of the corresponding water free-dispense value is a widely used starting point for something as thick as glycerol, and you adjust up from there only if the transfer stays clean.
- Settling time: lengthen it, on both aspirate and dispense, so the column has time to catch up and equalize before the tip moves. Delays of several seconds are normal here.
- Swap speed: slow the tip's exit from the liquid so a thick film has time to drain back rather than riding out on the tip.
- Dispense mode: prefer a surface dispense that touches the well over a jet, because a viscous slug does not jet cleanly.
Expect viscous classes to be strongly tip-specific and to need per-instrument tuning; the same intent needs different numbers on different hardware, and viscosity amplifies the difference.
Volatile liquids: ethanol, isopropanol, acetone, other low-boiling solvents
Volatile liquids fight you in two ways at once: they evaporate off the tip, so volume vanishes between aspiration and dispense, and they drip because they tend to have low surface tension. The recipe is about speed and sealing the column.
- Pre-wet the tip: aspirate and dispense back into the source once before the real run. The first dry-tip transfer is always the worst offender for volatiles, and pre-wetting brings it into line with the rest.
- Blowout volume: increase it. Volatiles need vapor room and extra force to leave the tip clean.
- Air transport volume: a trailing air gap holds the low-surface-tension column back from the opening so it does not drip in transit.
- Work fast and keep exposure short: every second the tip sits in open air is evaporation, so plan the sequence to minimize the time between aspirate and dispense. Chilling the source toward refrigerator temperature meaningfully slows evaporation when the assay allows it.
Volatiles are the family where the environment matters most; the same class can read differently on a warm afternoon, so record the temperature and expect to re-verify if the room drifts.
Involatile organic solvents: DMSO and its neighbors
DMSO is the deceptive one. It is an organic solvent, so people reach for solvent-aware caution, but it is not volatile and it is not especially viscous at room temperature, so it behaves much more like water on speed than like ethanol. Start from a water or aqueous class rather than a volatile one, keep the flow rate near the default, and pay attention instead to the solvent-specific issues: it is hygroscopic, so it pulls water from the air and its behavior drifts if a stock sits open, and it stresses some tip and seal materials. In compound management, where DMSO is ubiquitous, the volumes are often small, so the low-volume cautions below matter more than any solvent handling.
Foaming and protein-rich liquids: serum, plasma, detergents, surfactant buffers
Foam is the enemy here, because a layer of bubbles hides the true liquid level and ruins the next aspiration. Everything about the recipe is gentleness.
- Flow rate: keep it low so the stream does not whip air into the liquid on dispense.
- Dispense at the surface or below it: a wet dispense that touches or submerges into the liquid, rather than a jet from height, avoids the splashing that generates foam.
- Mixing: minimize it, and when you must mix, stay below roughly eighty percent of the well volume so you do not draw air at the top of the stroke.
- Blood products specifically: they clot and they are non-homogeneous, so gentle handling and, where the instrument offers it, clot detection protect both the sample and the tip.
Low-cohesion and low-surface-tension liquids
These liquids do not hold themselves together in the tip; they want to fall out. The core moves are a bigger air buffer and less time to form a drop.
- Air transport volume: increase it for a larger trailing buffer behind the liquid.
- Settling time: shorten it, so the liquid does not sit long enough to bead up and drip.
- Blowout: tune it so the tip clears without spraying, since an over-forceful blowout on a thin liquid splashes.
Low volumes near capillary scale
Below roughly twenty microliters, and especially below five, capillary forces dominate and the ordinary parameters stop behaving. A class that is perfect at 200 microliters can be useless at 2.
- Tip size: move to the smallest tip that covers the volume, because tip geometry interacts strongly with low volumes and a large tip's dead space swamps a tiny transfer.
- Following: turn liquid-level following off and keep the tip at a fixed, shallow, known position, because level detection is unreliable at this scale.
- Blowout: tune it carefully to prevent capillary over-aspiration, where the tip keeps drawing liquid past the target on its own.
- Consider the mechanism: at the smallest volumes, air-displacement pipetting hits its limits, and positive displacement or non-contact dispensing may be the honest answer rather than a heroic class.
Using these recipes
A recipe gets you to a clean-looking transfer quickly; it does not get you to a verified class. Apply the family moves, watch the transfer, then measure precision and set the correction curve as the optimization loop requires. The families overlap, too: a detergent-loaded buffer is both foaming and low-surface-tension, a bead slurry in DMSO is both viscous and solvent, so read the properties, combine the moves, and change one thing at a time when they conflict.
Difficult liquids are not unpredictable, they are just specific. Match the recipe to the family, then let measurement turn the recipe into a class.