A Spectrometer That Spans Octaves

Hank Hogan

Singing the musical scales provides a hint as to why users of grating-based spectrometers must be careful: After beginning with do, the scale ends up back at the same note but an octave higher.

In spectrometers, diffraction orders also repeat, and that prevents the measurement of octave-spanning spectra unless bandpass filters are used.

In an all-reflective Fourier transform spectrometer developed at Colorado State University, a beam strikes a split mirror, creating a spatial interferogram. Moving the slit samples the interferogram, with the detector measuring it. Because it uses reflective optics and does not use gratings, the device works across a wide wavelength range. Courtesy of Philip Schlup, Colorado State University.

Now researchers David G. Winters, Philip Schlup and Randy A. Bartels, all of Colorado State University in Fort Collins, have developed an all-reflective Fourier transform spectrometer that works from the near-UV at 400 nm to the mid-IR at 10 μm — a span of about 4.6 octaves.

Compared with previous Fourier transform spectrometers, this version has a simpler architecture, a fact that is evident in the ease with which the group implemented the idea. “When we set up the experiment, we had it working almost immediately. The only real technical issue was writing the data acquisition software,” Bartels said.

At the heart of the spectrometer is a split mirror. Spatially coherent light, as with that from a broadband laserlike source, enters the device and strikes the two halves of the mirror, which are set at an angle to each other. The resulting spatial interferogram is sampled by moving a box containing a slit and a detector. Once the apparatus is set up, the user need only change out the detector to measure a new spectral range. The slit may be swapped out, with larger slits used for longer wavelengths to boost the signal. However, it is not necessary to do this.

For their prototype device, the researchers wanted performance to be as achromatic as possible. They accomplished this by using all-reflective optics and by changing out the scanning slit. For the visible range, they used a 5-μm-wide slit, whereas for longer wavelengths, they used a 25-μm one, both from National Aperture Inc. of Salem, N.H. In addition, they used a silicon photodiode from Thorlabs Inc. of Newton, N.J., in the visible and near-IR and a liquid nitrogen cooled HgCdTe detector from IR Associates of Stuart, Fla., for the mid-IR.

They calibrated the spectrometer using a Ti:sapphire oscillator from Kapteyn-Murnane Laboratories Inc. of Boulder, Colo., operating it at 773.4 nm. They compared the performance of the device in the visible and near-UV with that of a conventional grating-based spectrometer from Ocean Optics Inc. of Dunedin, Fla. They found that the two agreed very well. For the mid-IR, they used a CO₂ laser, demonstrating that the Fourier transform spectrometer worked at 10.8 μm.

As for future research directions, the group is building a glancing incidence configuration of the spectrometer that will work in the extreme-UV (~15 nm). Bartels noted that this approach offers some advantages over other spectrometers. “Such a configuration would allow spectral measurements without the use of lossy extreme-UV/soft x-ray gratings.”

Optics Express, Feb. 5, 2007, pp. 1361-1368.