Section 01 · The Field E-01
The limit decides what 'clean' means
A result below the detection limit is not zero — it is a statement about the method, not the air.
What the limit actually is
Every measurement has a floor. Below it, the instrument cannot distinguish a real signal from the noise inherent in the system — background electrical fluctuation, trace contamination on the sorbent, solvent impurities in the analytical blank. That floor is the detection limit, and it is not a fixed property of the compound being sought. It is a property of the entire analytical chain: the sampler geometry, the sorbent mass, the pump flow rate and duration, the desorption efficiency, the detector type, the column, the integration decisions made at a workstation after the run. Change any one of those and the limit moves.
This matters because a result reported as 'below the detection limit' — written variously as < LOD, ND (not detected) or BDL (below detection limit) — is commonly read as a statement about the air. It is not. It is a statement about what this method, run in this way, for this long, was capable of seeing. The air may contain the compound at a concentration the method cannot resolve from noise. The compound may still be there. The method simply cannot confirm it.
The detection limit (LOD) is conventionally set at the concentration corresponding to a signal three times the standard deviation of the blank signal. The quantification limit (LOQ) sits higher — typically at ten times that standard deviation — and marks the threshold below which the instrument detects something but cannot assign it a reliable number. A result between LOD and LOQ is real but imprecise: the compound is present, but the reported concentration carries a large relative uncertainty. Many laboratories report the LOQ as the practical working floor and do not report individual results between the two thresholds at all. Which convention a laboratory uses should be stated in the report. If it is not, the reader cannot interpret the table.
The sampler as gatekeeper
The sampler determines how much of the compound actually reaches the detector, and that mass — not the air concentration alone — is what the detector sees. A diffusive badge accumulates compound at a rate set by its uptake rate constant, measured in millilitres per minute. Run it for eight hours and you collect a given number of nanograms. Run it for two hours and you collect a quarter of that. If the compound is present at a concentration just above the LOD for an eight-hour deployment, a two-hour deployment of the same badge in the same air will produce a result below the LOD. Same air, different sentence.
A pumped tube has a different geometry: you set the flow rate and the duration, and the product of the two is the sampled volume. Double the volume and you halve the effective detection limit expressed as a concentration. This is why occupational hygienists running short-duration task-based measurements often face higher detection limits than those running full-shift samples. The task is brief; the volume is small; the limit is correspondingly coarser. The result 'not detected' in a ten-minute task sample and the result 'not detected' in an eight-hour personal sample are not the same statement, even if they carry the same three letters.
Duration is therefore a design choice before the measurement begins, and it should be made with the limit explicitly in mind. If the purpose is to demonstrate that a concentration is below a given benchmark, the sampler and its deployment time must be chosen so that the LOQ is already below that benchmark before sampling starts. A result cannot demonstrate compliance with a threshold that lies below the method's quantification limit. The measurement was never capable of answering the question.
What changes when you change the method
A concentration near the detection limit of one method may be well within the quantifiable range of another. Thermal desorption from a sorbent tube coupled to gas chromatography with mass-spectrometric detection is among the most sensitive mainstream techniques for volatile organic compounds. Photoionisation detector instruments, often used for real-time screening, carry detection limits typically one or two orders of magnitude higher for many compounds. A screening survey that finds nothing does not establish that nothing is there; it establishes that concentrations are below the screening instrument's limit — which may be hundreds of micrograms per cubic metre when the technique capable of seeing tens of nanograms per cubic metre has not been deployed.
Laboratory factors compound this. The extraction solvent, its purity, and the volume used during liquid desorption all influence the blank level and therefore the LOD. In thermal desorption systems, the condition of the cold trap, the transfer line temperature, and any breakthrough from overloaded tubes shift the effective limit run to run. This is why good laboratory practice requires that detection limits be calculated — or at minimum verified — from the actual blanks and calibration standards run in the same analytical batch as the field samples. A detection limit carried forward from a previous batch, or borrowed from a published method without verification, is a number that describes a different day's analysis.
Reading the report
When a measurement report lists results as 'not detected', three questions are worth asking before accepting the table at face value. First: what is the stated LOD or LOQ, in the same units as the benchmark being applied? If that number is not present, the report is incomplete. Second: was the sampler and duration chosen to achieve a detection limit below the relevant benchmark before sampling? A measurement designed around the wrong question cannot answer the right one. Third: are the blanks reported alongside the samples? The blanks are the direct evidence from which the limit is derived; a limit without its supporting blanks is an assertion, not a calculation.
A result that reads 'not detected' is meaningful only inside its method. Strip away the sampler type, the sampling duration, the sorbent, the analytical technique, and the batch-specific blank level, and the three letters ND carry no information at all. 'Clean', as a conclusion drawn from a measurement report, means clean to the resolution this method offered on this occasion — and the method's resolution is printed, or should be, in the column header that nobody reads.
Filed inSection 01 — The Field