7.17 Evaluation of Recovery
109
When well-matched CRMs are not available, it is recommended to perform spiking of a defined amount of analyte (most likely the pure chemical) to the sample. The
measurements are performed for the non-spiked and spiked sample. From the difference of both quantity values, it is possible to estimate the calculated concentration
of the added analyte. In the case of unbiased measurement procedure, the difference
between both values would be equal to the defined amount of added analyte. When
carrying out the recovery test, the following conditions should be maintained: analyte added to the sample should be present in the same chemical form in which it is
present in the real sample, and the total concentration of the analyte after the addition
of the standard should be within linear range of the calibration.
7.18 The Limit of Detection and Quantitation
The dynamic range of a calibration curve refers to the concentration/content of an
analyte that can be determined using a given measurement procedure with acceptable
accuracy and uncertainty. The upper concentration can be evaluated by examination
of the course of linearity, as the lower concentration can be evaluated by examining
the probability of distinguishing the analytical signal from the base line noise. Both
values depend not only on measuring technique but also on the applied measurement
procedure covering the entire analytical process.
In practice, it is convenient to define the scope for which both values are determined. When the primary characteristic of the given analytical technique is of interest,
it is recommended to execute the measurements for blank as well as for matrix-free
standard solution, which enables evaluation of the best instrumental performance of
the technique. In the case the important information is the capabilities of the measurement procedure used for the matrix-rich test samples, then it is recommended to
execute the measurements for the blank solutions that underwent the entire analytical process, as well for real test samples. With regard to the detection/quantification
limits, typically the ‘instrumental’ values are lower than the ‘procedural’ ones.
7.19 The Limit of Detection
The limit of detection (LoD) indicates the smallest amount/concentration of the
analyte that can be detected using a given measurement procedure. The LoD is often
assumed to be threefold the value of the standard deviation for the blank sample or
a sample containing a low concentration of the analyte. The numerical value of the
LoD has the dimension of the concentration/content of the analyte.
There are several methods used for the calculation of LoD; it can be calculated as
a sum of average signal for the blank and three times the standard deviation (blank +
3 s), or can be calculated from the slope of the calibration curve, so the detection
limit is calculated as 3 s/b, where b is the slope of the graph (Table 7.10).
109
When well-matched CRMs are not available, it is recommended to perform spiking of a defined amount of analyte (most likely the pure chemical) to the sample. The
measurements are performed for the non-spiked and spiked sample. From the difference of both quantity values, it is possible to estimate the calculated concentration
of the added analyte. In the case of unbiased measurement procedure, the difference
between both values would be equal to the defined amount of added analyte. When
carrying out the recovery test, the following conditions should be maintained: analyte added to the sample should be present in the same chemical form in which it is
present in the real sample, and the total concentration of the analyte after the addition
of the standard should be within linear range of the calibration.
7.18 The Limit of Detection and Quantitation
The dynamic range of a calibration curve refers to the concentration/content of an
analyte that can be determined using a given measurement procedure with acceptable
accuracy and uncertainty. The upper concentration can be evaluated by examination
of the course of linearity, as the lower concentration can be evaluated by examining
the probability of distinguishing the analytical signal from the base line noise. Both
values depend not only on measuring technique but also on the applied measurement
procedure covering the entire analytical process.
In practice, it is convenient to define the scope for which both values are determined. When the primary characteristic of the given analytical technique is of interest,
it is recommended to execute the measurements for blank as well as for matrix-free
standard solution, which enables evaluation of the best instrumental performance of
the technique. In the case the important information is the capabilities of the measurement procedure used for the matrix-rich test samples, then it is recommended to
execute the measurements for the blank solutions that underwent the entire analytical process, as well for real test samples. With regard to the detection/quantification
limits, typically the ‘instrumental’ values are lower than the ‘procedural’ ones.
7.19 The Limit of Detection
The limit of detection (LoD) indicates the smallest amount/concentration of the
analyte that can be detected using a given measurement procedure. The LoD is often
assumed to be threefold the value of the standard deviation for the blank sample or
a sample containing a low concentration of the analyte. The numerical value of the
LoD has the dimension of the concentration/content of the analyte.
There are several methods used for the calculation of LoD; it can be calculated as
a sum of average signal for the blank and three times the standard deviation (blank +
3 s), or can be calculated from the slope of the calibration curve, so the detection
limit is calculated as 3 s/b, where b is the slope of the graph (Table 7.10).