An integrated micromachined CMOS spectrometer for biochemical analysis

Spectrometry is a valuable analytical chemistry technique that measures the interactions between electromagnetic radiation and matter. Because these interactions have frequency components that are characteristic in nature, spectrometry can measure analyte composition and concentration. Spectrometers are widely used in the analysis of environmental pollutants, industrial processes, and biomedical assays. However, conventional optical spectrometers are generally complex, bulky, and exacting to assemble. Micromachining offers a method to produce simpler, smaller devices that are easily manufactured. While micromachined spectrometers exist, many designs are not practical to fabricate in volume nor to integrate with the circuitry desired for many applications.; This work covers the design, characterization, and application of a novel micromachined integrated CMOS spectrometer designed for detection of visible light spectra. The design consisted of a grating mounted above a CMOS active pixel imager with 128 pixels. The gratings were developed separately using electron beam lithography on fused silica wafers. The CMOS imager die were fabricated at the Center for Integrated Systems (CIS) at Stanford University using a dual metal poly gate p-well process. To align the grating properly to the underlying imager die, thick photoresist walls were formed on the imager to provide a mechanical alignment guide for placement during bonding. The spectrometers showed a dispersion of 5 nm/pixel, a resolution of 15 urn, a signal to noise ratio of 66 dB, and a detection range of 400-725 nm. This is comparable in performance to other micromachined spectrometers, yet it has the advantage of being integrated and easily manufacturable with standard fabrication processes. In addition, the spectral performance of the device is consistent with the fundamental limits for its compact size. The overall device measured 7.0 x 2.6 x 3.7 mm. Finally the utility of the spectrometer as a real-time process analytical tool was demonstrated by optically monitoring the pH of cell culture media with a repeatability of 0.003 pH units and an accuracy of 0.08 pH units.