Methodologies and Industrial Applications of Infrared Spectral Imaging
E. Neil Lewis
Spectral Dimensions, Inc., 3416 Olandwood Court, Olney, MD 20832
There is a growing demand and need for high speed, parallel chemical analysis in a number of industries to support the ever increasing speed of product development, production and quality control. In addition, because of the complexity and intrinsic chemical heterogeneity of many manufactured products, analytical metrics that provide spatial as well as spectral information are critical.
Current technologies employing simple ‘point and shoot’ near-infrared or mid-infrared spectroscopic approaches or mapping techniques in which spectrometers rapidly collect a sequential series of spectra are, in many cases, inadequate. The methods are time-consuming and inefficient, and obtaining the data to fully characterize a sample utilizing mapping techniques can take many days to complete. New developments in high-speed infrared focal-plane array (FPA) detectors and solid-state tunable filter technology promise to revolutionize these applications by providing a fundamental departure from existing sequential modalities to a highly parallel approach. Using infrared FPAs and tunable filters it is possible to collect tens of thousands of spatially resolved spectra simultaneously. This technology can deliver spectral performance per channel that is equivalent to many single point spectrometers with several orders of magnitude increase in the speed of the analytical process.

Figure 1 shows a neat-infrared spectral image and of a finished pharmaceutical product. The data was set contains 76,800 spectra recorded over the spectral range of 1000-1700 nm. Each pixel contains a separate NIR spectrum corresponding to approximately 40x40 microns of the sample surface. Several spectra are shown in the panel on the right corresponding to spatial regions containing both the active ingredient and excipients in the pharmaceutical matrix.