Questions: Limit of Detection and Limit of Quantification

Question 1

An analyst measures 10 blank replicates and calculates σ_blank = 2.0 signal units. The calibration curve slope is 4.0 signal units per ppb. What is the LOD expressed in concentration units?

• A. 2.0 ppb — the LOD equals the standard deviation of the blank
• B. 6.0 ppb — the LOD is 3σ in signal units and does not need correction
• C. 1.5 ppb — calculated as 3σ / slope = (3 × 2.0) / 4.0
• D. 20 ppb — the LOD requires a 10σ margin above the blank

Answer: (C)

LOD = 3σ_blank / slope converts the detection threshold from signal units to concentration units. The 3σ signal threshold (3 × 2.0 = 6.0 signal units) must be divided by the calibration sensitivity (4.0 signal units/ppb) to get 1.5 ppb. Option B is the common error: stopping at 3σ in signal units without converting to concentration. Option D describes the LOQ, not the LOD. This calculation shows why high sensitivity (steep calibration slope) directly improves LOD — the same noise translates to a smaller concentration uncertainty.

Question 2

An environmental lab characterizes the LOD for mercury in drinking water as 0.1 μg/L using its standard spectrometer. A regulatory chemist then assumes this same LOD applies when analyzing mercury in coastal seawater samples. What error has been made?

• A. No error — the LOD is an instrument specification that does not change with sample composition
• B. The chemist has confused LOD with LOQ, and should use 0.33 μg/L instead
• C. The LOD is method- and matrix-specific; the high salt content of seawater can suppress analyte signal and increase blank noise, making the actual LOD much higher than 0.1 μg/L in that matrix
• D. LODs apply only to aqueous standards, not to environmental samples

Answer: (C)

The critical insight is that LOD is a property of the entire analytical method in a specific matrix — not just the instrument. Seawater's high ionic strength can suppress the mercury signal (matrix interference) and contribute additional background noise, both of which worsen the LOD in concentration units. The 0.1 μg/L value was determined in clean water; it cannot be assumed to transfer to a chemically different matrix without experimental validation. This is why LOD and LOQ must be determined in the actual sample matrix of interest.

Question 3

A result between the LOD and LOQ is typically reported as 'detected but below the limit of quantification' — meaning the analyte is present but cannot be reliably assigned a precise numerical concentration.

• T. True
• F. False

Answer: True

The LOD–LOQ gap is a zone where the signal is distinguishable from blank noise with ~99% confidence (so detection is valid) but measurement precision is too poor for a reliable number (often ±50% or worse). Reporting 'detected but below LOQ' honestly conveys both pieces of information: the analyte is present, but its concentration cannot be stated with acceptable accuracy. Reporting a number below the LOQ implies false precision. Reporting 'not detected' would be incorrect because the signal IS distinguishable from blank.

Question 4

The limit of detection is the lowest concentration that produces any measurable signal above zero in the analytical instrument.

• T. True
• F. False

Answer: False

This is the most common misconception about LOD. Below the LOD, the instrument almost certainly does produce a signal — the problem is that the signal cannot be reliably distinguished from the inherent variation of the blank. LOD is a statistical confidence threshold: the concentration at which signal is 3σ above the blank mean, corresponding to ~99% confidence (one-tailed) that the signal is not a random blank fluctuation. At concentrations below the LOD, real analyte signals exist but are buried in noise and cannot be reliably attributed to the analyte rather than baseline variability.

Question 5

Why is the limit of quantification (LOQ) set higher than the limit of detection (LOD), and what does this mean for how analytical results are reported?

Model answer: The LOD (3σ) establishes that the analyte is probably present, but at this signal level measurement uncertainty is enormous — often ±50% or more. The LOQ (10σ) sets a higher threshold where precision is acceptable for reporting a meaningful numerical concentration, typically ±10–20% relative standard deviation. This creates three reporting zones: above LOQ (report the number), between LOD and LOQ (report 'detected but below quantitation limit'), and below LOD (report 'not detected'). Using 'not detected' for a result below LOD does not mean the analyte is absent — it means it cannot be distinguished from blank noise at the method's sensitivity.

The distinction matters practically because regulatory decisions often hinge on whether a contaminant is 'detected' versus 'quantifiable' versus 'absent.' A result below LOQ but above LOD can still inform risk assessment (the analyte is present at low levels) even though it cannot be precisely quantified. Conflating LOD and LOQ — reporting 'not detected' when a signal sits between them — systematically underestimates environmental contamination.