Instant Notes: Analytical Chemistry

Section B – Assessment of data

B5 QUALITY CONTROL AND

CHEMOMETRICS

Control charts The purpose of a control chartis to monitor data from an ongoing series of
quantitative measurements so that the occurrence of determinate (systematic)
errors (bias), or any changes in the indeterminate (random) errors affecting the
precision of replicates can be detected and remedial action taken. The predomi-
nant use of control charts is for quality control (QC) in manufacturing industries
where a product or intermediate is sampled and analyzed continually in a
process stream or periodically from batches. They may also be used in analytical
laboratories, such as those involved in clinical or environmental work, to
monitor the condition of reagents, standards and instrument components,
which may deteriorate over time.
Shewart chartsconsist of a y-axis calibrated either in the mass, concentration
or range of replicated results of an analyte or in multiples of the estimated stan-
dard deviation, s, of the analytical method employed, and the sample number
along the x-axis. An averagesor X-chart, the most common type, is pre-
prepared with a series of five parallel horizontal lines, the centre one being posi-
tioned along the y-axis to correspond to the true, accepted or target valuefor the
analyte (Fig. 1). The other four lines are positioned in pairs on either side of the
target value line and act as critical levels that, when exceeded, indicate probable
instability in the system. The inner pair are defined as warning levelsand the
outer pair as action levels.
An averages chart is used to monitor the level of an analyte either as single
values or as means of Nreplicates to check for determinate errors (bias) in the
results. Chart criteria and decisions are based on the following:

● plotted values have a Gaussian or normal distribution (Topic B2);

● warning linesare positioned to correspond to a selected probability level

Key Notes

Graphical representations of quantitative data from an ongoing series of

measurements can be used to monitor the stability of a system for quality

control (QC) purposes.

Schemes have been devised to compare results for the analysis of

homogeneous samples or standard materials from groups of analytical

laboratories to test specified methods of analysis.

Advanced chemometric procedures can be applied to the statistical

analysis of large amounts of data and numerous experimental variables

so as to extract the maximum amount of useful information.

Related topic Quality in analytical laboratories (A6)

Control charts

Collaborative testing

Multivariate

statistics