Hyperspectral Imaging with a Step-Scanning Interferometer

Hyperspectral Imaging with a Step-Scanning Interferometer 

The first array detectors to be used for FT-IR hyperspectral imaging were focal-plane arrays (FRA) fabricated from InSb. MCT arrays were incorporated shortly thereafter, and these detectors are usually cold-shielded to improve their performance. The individual elements are bump-bonded onto small spheres of indium (a very soft metal). This method of fabrication sometimes led to poor reliability, but this problem has been largely overcome. Details of the construction of FPAs are given by in the paper by Scribner et al. [6]. Most InSb and MCT arrays are square, with 64 X 64, 128 x 128, or 256 x 256 detector elements (pixels) that are 6.25 X 6.25 pm in size. Thus, the entire detector array is less than 0.5 mm on the side. Note also that when such small detectors are used, their NEP is concomitantly small (see Section 7.1). 

ADC would have necessitated a data acquisition frequency of over 20 MHz 20-MHz ADCs with the high dynamic range needed for FT-IR spectrometry are still not available. In the first imaging FT-IR spectrometers, the data acquisition frequency was reduced to a manageable level through the use of a step-scan interferometer (see Section 5.5). 

In the earlier software releases for hyperspectral imaging, the interferometer mirror was stepped to the next sampling position after reading out all the data from the array of pixels. An improvement in measurement efficiency was achieved by Levin s group [7], who showed that since about half the data acquisition time 

Many of the problems associated with the manufacture of FPAs have now been overcome, and most contemporary instruments no longer suffer from the bad pixel problem. Instruments based on the combination of step-scanning interferometers and FPA detectors are now available from Varian (the Stingray), Thermo Electron (the Continupm XL), and Bruker Optics (the Hyperion 3000). 

---