A Small and Cheap Spectrometer

Hank Hogan

For spectrometers, it is better to be small and cheap than big and bulky — provided that one’s optical performance targets are met. Now a team of researchers from Delft University of Technology and from TNO Institute of Science and Technology, both in Delft, the Netherlands, have developed one that measures just 3 × 3 × 11 mm. This small instrument has produced some big results, achieving 3-nm resolution over a spectral range of 450 to 750 nm with an optical throughput of ∼9 percent.

The size of a coin, the spectrometer shown here comprises two transparent rectangles mounted over an image sensor. The fiber tip in the top-left corner feeds in the light. Courtesy of Semen Grabarnik, Delft University of Technology, and STW, the Dutch Technical Foundation in the Netherlands.

Gleb Vdovin, project co-leader and an associate professor of electronic instrumentation at the university, noted that the design trade-offs led to the relatively low throughput but that the device was, nonetheless, useful. “The practical field of application is still quite large, as [the] signal-to-noise ratio can be improved by frame integration, especially if we take into account [the] potentially very low price of fabrication.”

The group, which included graduate student Semen Grabarnik, constructed the microspectrometer from a single aluminum-coated glass wafer. The researchers patterned the aluminum to create two gratings on one 3 × 11-mm section and a mirrored input slit and intermediate mirror on another similarly sized section. They diced the wafer to free the sections, spaced them 3 mm apart and set the two pieces atop a CCD board camera from Videology Imaging Solutions of Greenville, R.I., that they used as a photodetector.

The spectrum of a neon lamp, as produced by the miniature spectrometer, is shown. Reprinted with permission of Optics Express.

The manufacturing and alignment steps, according to Vdovin, are not complicated and are standard for a batch process. There also is the possibility of mounting the device directly atop a sensor chip, which could lead to a very low cost product.

During operation, light entering one end of the spectrometer bounces off a mirror on the bottom glass piece and is reflected off a planar grating on the top plate. The diffracted light strikes a mirror on the bottom plate, which sends it to a second grating on the top glass wafer. From there, the now spectrally separated light travels to the photodetector.

The group tested the spectrometer on a neon lamp, capturing the output from 550 to 750 nm, and determined that the resolution was better than 3 nm and that the ability to locate the relative position of the spectral lines was better than 1 nm.

Reinoud F. Wolffenbuttel, a professor at the university and group co-leader, has been working on developing microspectrometers for more than a decade, with this version funded primarily by STW, the Dutch Technical Foundation in the Netherlands. The goal is to increase the resolution by several times that of the dual-grating device without further sacrificing the light throughput.

Wolffenbuttel noted that the structure of the current microspectrometer is straightforward but that the results are important for future projects. “The measurements validate our modeling, which is now used for the design of nonplanar gratings.”

Optics Express, March 19, 2007, pp. 3581-3588.