A number of industrial processes, especially quality assurance procedures, accept information on relative
quantities of components in mixtures, whenever absolute values for the quantitative analysis are unavailable.
These relative quantities may be determined from infrared intensity ratios even though known standards are
Repeatability [vs precisionhl in quantitative analysis is a critical parameter for meaningful results. In any given
analysis, multiple runs provide "answers" with a certain standard deviation. Obviously, the lower the standard
deviation, the better the precision. In attempting to minimize the standard deviation and thus improve
precision, we need to delineate which contributing factors we have control over (such as sample preparation
techniques, data analysis methodology) and which factors we have little control over (environmental and
instrument noise, for example). For a given set of conditions, the best instrumental precision achievable on an
IR instrument should be determinable. Traditionally, the term "signal-to-noise" (S/N) has been used for a single
spectrum, realizing that S/N improves with an increase in number of scans coadded for generation of that
single spectrum. However, the S/N ratio does not directly reflect the precision achievable for an absorbing
band. We prefer to use the phrase "maximum achievable instrument precision" (MAIP), which is equivalent to
the minimum relative standard deviation for a given peak (either height or area) in spectra.
For a specific analysis, the analyst should have in mind the desired precision. Only if the desired precision is
less than the MA1P will the analysis be feasible. Once the MAIP is established, other experimental procedures
may be modified to improve the analytical precision, if it is below that which is expected (the MAIP).