Imagine running the same enzyme assay three weeks in a row. Same protocol, same reagents, same pipettes. The first two runs give clean, reproducible kinetics. The third shows activity down by a third, with no obvious cause. You check the enzyme stock, recalibrate the plate reader, and remake the substrate. Nothing changes.

Eventually someone notices that the water purification cartridge is overdue for replacement. Water is the most heavily used reagent in almost every laboratory, and the one we think about least. Pristine and contaminated water look exactly the same, so problems hide in plain sight. Understanding water quality turns a mysterious failure into a controllable variable.

Laboratory water is not a single substance. It is a family of specifications. Standards bodies such as ASTM, ISO and CLSI define grades by measurable limits. A common scheme starts with Type III, used for rinsing glassware and feeding autoclaves. Type II suits general buffers and microbiological media. Type I, or ultrapure water, reaches a resistivity of 18.2 MΩ·cm at 25 °C.

The key is matching the grade to the sensitivity of your measurement. A titration measuring millimolar concentrations tolerates trace ions that would ruin an ICP-MS run measuring parts per trillion. HPLC and LC-MS need low organic content, because dissolved organics concentrate on the column and later emerge as ghost peaks. Cell culture and molecular biology add biological requirements. They need low endotoxin and freedom from the nucleases that degrade RNA.

Using the highest grade everywhere sounds safe, but it has costs. Ultrapure systems consume expensive cartridges. Type I water is also such an aggressive solvent that it can pull ions out of its containers. A better approach takes three steps. Identify your most sensitive measurement. Work out which contaminant it responds to. Then choose the grade that controls that contaminant.

Takeaway

Water purity is not a ladder to climb as high as possible but a specification to match. Ask what your measurement is sensitive to, then choose the grade that controls it.

Purified water starts degrading the moment it leaves the system. Ultrapure water readily dissolves almost anything it touches. Exposed to air, it absorbs carbon dioxide, which forms carbonic acid. Resistivity can fall from 18.2 to around 1 MΩ·cm within minutes. For a general buffer this rarely matters. For trace ion analysis, it already does.

Containers are the next source. Glass leaches sodium, boron and silica, especially soda-lime glass. Many plastics release plasticisers, antioxidants and mould-release agents, which show up as organic carbon or spectroscopic artifacts. Wash bottles are notorious, because their soft plastic is squeezed constantly and old water gets topped up with new. Reservoirs and tubing where water sits still grow biofilms. These thin bacterial layers shed cells and endotoxin into whatever passes through.

The practical defence is a set of simple habits. Draw water fresh for sensitive work instead of storing it. Discard the first portion dispensed after the system has been idle. Use containers suited to your analysis, such as PFA or polypropylene for trace metals, and rinse them with the same water you will use. Label wash bottles with fill dates, and never top them up.

Takeaway

Purity describes water at a particular moment and place, not a permanent label. Every container, every minute of air exposure, and every hand that touches it becomes part of your experimental system.

Contaminated water looks identical to pure water, so verification depends on instruments. Conductivity, or its inverse, resistivity, is the first line of defence. Ions carry current, so fewer ions mean higher resistivity. Most purification systems measure it inline and continuously. The test is fast and cheap, but it misses anything without a charge. Neutral organic molecules, weakly ionised silica and bacteria barely register.

Total organic carbon (TOC) analysis fills part of that gap. A typical method oxidises organics with ultraviolet light and measures the carbon dioxide produced, often through the change in conductivity it causes. Ultrapure systems commonly target TOC below 5 to 10 parts per billion. Rising TOC is an early warning that a cartridge is exhausted or that something is leaching into the water.

Biological tests cover the remaining risk. To count bacteria, filter a sample through a membrane and incubate it on a low-nutrient medium such as R2A agar. Results are reported as colony-forming units per millilitre. Endotoxins are bacterial cell-wall fragments that trigger immune responses in cell culture. They are measured with the Limulus amebocyte lysate (LAL) assay. Keep a simple log of all three tests so that drift becomes visible before it ruins an experiment.

Takeaway

Each test sees only one kind of contaminant, so confidence comes from combining them. A single good number tells you what is absent, not that the water is clean.

Robert Boyle insisted that experiments be described and controlled carefully enough for others to repeat them. Water belongs in that description. It is a reagent with a specification, a shelf life and a history, even when it comes from a tap on the wall.

Treat it that way, and a whole class of mysterious failures disappears. Choose the right grade, protect it from recontamination, and verify it with measurements. Reliable discoveries rest on the variables you remembered to control.