Identifying a substance by its NIR spectrum can be achieved by a functional group analysis library search, but identifying small differences between spectra has - until now - been difficult
FT-NIR is a useful technique for materials checking during the manufacturing process, particularly at the raw materials inspection stage.
Identifying a substance by its NIR spectrum can be achieved by a functional group analysis library search, but identifying small differences between spectra is difficult.
The result may suggest a series of possible compounds rather than a single substance, which is insufficiently precise for QA/QC requirements.
The Compare function of PerkinElmer's Spectrum software is designed to overcome these problems.
A spectrum can be compared against another to determine how closely they match, or against a number of spectra to identify the most similar spectrum.
Reproducibility of sampling, varying baselines, or non-uniform noise distribution may cause problems in calculating the similarity of spectra.
Mathematical filters are used to reduce the contribution where the information tends to be unreliable.
These filters maximise spectral differences, while minimising variations due to instrumental, sampling, and atmospheric conditions.
A technical note from PerkinElmer demonstrates the use of Spectrum Compare to estimate the similarity of a group of materials used in the pharmaceutical industry for tablet coatings.
A spectral library of five different materials (four batches of each) was built.
Results are described and illustrated both with and without the use of mathematical filters.
Results were consistently improved by using Compare filters to enhance sensitivity due to chemical differences, and reduce the sensitivity to common sampling and instrument problems.
Using extended wavelength NIR and Compare filters allows sampling in amber vials and facilitates sampling in original containers.
