Detection limit

The limit of detection (LOD or LoD) is the lowest signal, or the lowest corresponding quantity to be determined (or extracted) from the signal, that can be observed with a sufficient degree of confidence or statistical significance. However, the exact threshold (level of decision) used to decide when a signal significantly emerges above the continuously fluctuating background noise remains arbitrary and is a matter of policy and often of debate among scientists, statisticians and regulators depending on the stakes in different fields.

In analytical chemistry, the detection limit, lower limit of detection, also termed LOD for limit of detection or analytical sensitivity (not to be confused with statistical sensitivity), is the lowest quantity of a substance that can be distinguished from the absence of that substance (a blank value) with a stated confidence level (generally 99%). The detection limit is estimated from the mean of the blank, the standard deviation of the blank, the slope (analytical sensitivity) of the calibration plot and a defined confidence factor (e.g. 3.2 being the most accepted value for this arbitrary value). Another consideration that affects the detection limit is the adequacy and the accuracy of the model used to predict concentration from the raw analytical signal.

As a typical example, from a calibration plot following a linear equation taken here as the simplest possible model:

where, ${\displaystyle f(x)}$ corresponds to the signal measured (e.g. voltage, luminescence, energy, etc.), "b" the value in which the straight line cuts the ordinates axis, "a" the sensitivity of the system (i.e., the slope of the line, or the function relating the measured signal to the quantity to be determined) and "x" the value of the quantity (e.g. temperature, concentration, pH, etc.) to be determined from the signal ${\displaystyle f(x)}$, the LOD for "x" is calculated as the "x" value in which ${\displaystyle f(x)}$ equals to the average value of blanks "y" plus "t" times its standard deviation "s" (or, if zero, the standard deviation corresponding to the lowest value measured) where "t" is the chosen confidence value (e.g. for a confidence of 95% it can be considered t = 3.2, determined from the limit of blank).

Thus, in this didactic example:

$${\displaystyle {\text{LOD for }}x={\frac {\left(f(x)-b\right)}{a}}={\frac {\left(y+3.2s-b\right)}{a}}}$$

There are a number of concepts derived from the detection limit that are commonly used. These include the instrument detection limit (IDL), the method detection limit (MDL), the practical quantitation limit (PQL), and the limit of quantitation (LOQ). Even when the same terminology is used, there can be differences in the LOD according to nuances of what definition is used and what type of noise contributes to the measurement and calibration.

The figure below illustrates the relationship between the blank, the limit of detection (LOD), and the limit of quantitation (LOQ) by showing the probability density function for normally distributed measurements at the blank, at the LOD defined as 3 × standard deviation of the blank, and at the LOQ defined as 10 × standard deviation of the blank. (The identical spread along Abscissa of these two functions is problematic.) For a signal at the LOD, the alpha error (probability of false positive) is small (1%). However, the beta error (probability of a false negative) is 50% for a sample that has a concentration at the LOD (red line). This means a sample could contain an impurity at the LOD, but there is a 50% chance that a measurement would give a result less than the LOD. At the LOQ (blue line), there is minimal chance of a false negative.