LOD and LOQ Calculator
Calculate limit of detection LOD=3sigma/slope and limit of quantification LOQ=10sigma/slope.
What LOD and LOQ Actually Mean
Every instrument produces a signal even when no analyte is present — electronic noise, stray light, reagent impurities. The limit of detection is the concentration whose signal is confidently distinguishable from that background, and the limit of quantification is the concentration you can report as a number rather than merely as present.
The multipliers are conventions with a statistical basis. Three standard deviations above the blank gives roughly 99% confidence that a signal is real rather than noise — but with about 30% relative uncertainty, far too imprecise to quote a value. Ten standard deviations brings relative uncertainty down to around 10%, which is generally accepted as the threshold for reporting a number.
| Term | Meaning | Units |
|---|---|---|
| σ | Standard deviation of the blank signal, from repeated blank measurements | Signal units |
| m | Slope of the calibration curve — the method's sensitivity | Signal per concentration unit |
| LOD | Lowest concentration reliably distinguished from blank | Concentration |
| LOQ | Lowest concentration reportable as a quantitative value | Concentration |
Notice that both terms sit in the equation for a reason. A method improves either by reducing noise (smaller σ) or increasing sensitivity (larger m). Doubling the slope halves the detection limit just as effectively as halving the noise.
Reporting Results Near the Limits
| Measured concentration | How to report | Why |
|---|---|---|
| Below LOD | “Not detected” or “< LOD” | Signal indistinguishable from blank |
| Between LOD and LOQ | “Detected but below LOQ” | Analyte is present, but the number is unreliable |
| Above LOQ | Report the numerical value | Uncertainty is acceptable for quantification |
Reporting a number between LOD and LOQ is a common and serious error. The analyte is genuinely there, but the value carries such large uncertainty that quoting it implies a precision the measurement does not have.
Worked Examples
Common Mistakes
In that band you can say the analyte is present but not how much. Relative uncertainty is too large for the number to mean anything — report it as detected below LOQ.
Standard deviation from three or four replicates is itself highly uncertain. Validation protocols typically require at least seven to ten independent blank measurements.
Instrument detection limit measures the detector alone. Method detection limit includes sample preparation, dilution, and extraction losses, and is usually considerably higher.
It depends on matrix, day-to-day conditions, calibration range and operator technique. It must be re-established for each method and matrix, not quoted from a brochure.
Frequently Asked Questions
Formula Explorer connections
Interpretation: This analytical relationship converts an instrument signal, separation measure or optical response into concentration, identity or performance. Assumption: Calibration, blank correction, linear range, path length, matrix effects and instrument settings must match the sample and method.