The Problem of Company
Dissolving a sample gets the analyte into solution, but it brings along everything else too. A drop of blood holds the drug you want to measure plus proteins, salts, fats, and sugars. A scoop of soil holds the pesticide plus humus, clay, and a zoo of other compounds. Any of those tag-alongs that fools the measurement is an interferent. The next job is separation: pull the analyte cleanly away from the crowd before the instrument ever sees it.
All separation rests on one idea: find a property where the analyte differs from the matrix, then exploit it. Does the analyte love oil while the matrix loves water? Does it stick to a special surface while the matrix washes past? Each method in this guide is a different answer to the same question — what is different about the thing I want?
Liquid–Liquid Extraction: Oil and Water Choose Sides
The oldest trick is liquid–liquid extraction. You shake your watery sample together with an oily solvent that does not mix with water — like shaking oil and vinegar in a jar. When they settle back into two layers, each compound has had to choose a side: water-lovers stay in the water, oil-lovers move into the oil. If your analyte loves oil and most of the matrix loves water, the analyte concentrates in the small oil layer, which you simply pour off — leaving the bulk of the interference behind.
How strongly a compound prefers one layer over the other is captured by its partition coefficient — basically the ratio of how much sits in oil versus water at equilibrium. A big difference in partition coefficients between analyte and matrix is what makes a clean split possible. When the analyte is shy about leaving the water, chemists often do several smaller extractions rather than one big one, because repeating the shake pulls out more in total than a single try.
Solid-Phase Extraction: A Sieve You Can Switch On and Off
The modern favourite is solid-phase extraction, usually shortened to SPE. Instead of two liquid layers, you pour your sample through a tiny cartridge packed with specially coated particles. The analyte sticks to the particles as the liquid flows past, while much of the matrix runs straight through and down the drain. Then you wash away whatever loosely clung on, and finally pour a small splash of a strong solvent through to release the analyte — now clean and crowded into a few drops.
- Condition: wet the cartridge with solvent so the coated particles are ready to grab the analyte.
- Load: pour the sample through; the analyte sticks while much of the matrix passes through.
- Wash: rinse with a gentle solvent to flush out weakly held interferents while the analyte stays put.
- Elute: pour a small amount of a strong solvent through to release the now-purified analyte into a tiny volume.
SPE has mostly replaced shaking jars of solvent because it uses far less solvent, throws away the matrix instead of you, and can be lined up in racks of dozens at once. And notice the bonus hiding in that last step: by trapping analyte from a large sample and releasing it into a few drops, SPE does not just clean the sample — it makes the analyte more concentrated than it started, which the next guide explores.
Cleanup: Tidying Before the Instrument
All of these separation steps share one umbrella name: cleanup — any step whose purpose is to remove matrix junk before measurement. The motive is partly honesty and partly self-preservation. Dirty extracts clog instruments, coat sensitive parts, and slowly poison the very detectors that cost a fortune. A few minutes of cleanup can save days of downtime, so good analysts treat it as protecting the instrument as much as protecting the result.
There is a tension to respect, though. Every cleanup step throws something away — and there is always a risk you throw away a little of the analyte with the junk. More steps mean a cleaner extract but also more chances to lose what you came for. The craft is doing just enough cleanup to tame the matrix, and no more.
Derivatization: Giving the Analyte a Costume
Sometimes the analyte is perfectly present but the instrument simply cannot see it well — it does not absorb light, or it will not turn into a gas, or it looks identical to its neighbours. Here you reach for derivatization: you react the analyte with a chosen reagent that latches on a new chemical group, deliberately changing the molecule into a form the instrument loves. It is like dressing a shy guest in a bright costume so the camera finally catches them.
For example, a colourless sugar can be reacted to produce a coloured product that absorbs light strongly, so a simple colour-reading instrument can now measure it. Or a heavy, sticky molecule can be tagged to make it light and volatile enough to travel through a gas instrument. The catch is that the disguise must be reliable: if the reaction does not go all the way every time, you are now measuring how well the costume fit rather than how much analyte was there — so derivatization trades a seeing problem for a consistency problem.