Low dimensionality spectral sensing for low cost material discrimination and identification

Abstract

Spectroscopy is a powerful tool in material identification, characterization and discrimination. Unfortunately industrial and laboratory spectrometers are typically very large, costly, and inconvenient. The aim of this thesis is to broaden the awareness and appeal of spectroscopic sensing modalities by exploring specialized, rather than general purpose instruments. Rather than sensing the entire spectrum, these devices work by observing just the particular spectral features needed to perform identification or discrimination. This approach greatly simplifies the instrument reducing the cost, size, power consumption, and analysis complexity by many orders of magnitude. In this work the anatomy of such specialized sensors is explored by way of a thorough discussion of illuminators, current sources, photodetectors, photodiode amplifiers, control systems and part selection. In the following chapters, instruments are designed and fabricated, and their tradeoffs are enumerated and discussed. Finally, these building-blocks are combined to construct several working prototypes which are informally characterized.