Wednesday

May 16, 2012

A generic Fourier Transform spectrometer works by splitting a collimated light beam in two using a beamsplitter cube. Each of the two light beams travels to a retroreflector some distance away, then both beams return and are recombined using another beamsplitter cube (usually referred to as the beam combiner cube). By varying the difference in the optical path length between both beams, an interference pattern is created, consisting of intensity variations as a function of delay. The Fourier Transform of this interference pattern produces the spectrum of the light being observed.

Fourier Transform spectrometers are the most precise instruments available for spectrographic studies. An FTS can obtain resolutions many times greater than conventional spectrographs with better sensitivity. On the other hand, a typical FTS has poor throughput, making studies on faint objects such as stars impractical. There are very few FTSs in use for astronomical studies. One example of a successful astronomy FTS is the McMath-Pierce FTS facility at Kitt Peak National Observatory. This FTS, however, is used primarily for observing the sun and labratory sources.

We intend to improve the sensitivity of our FTS by making three critical modifications to the classical design.

• A dispersive element, or a diffraction grating, is placed in the optical train after beam recombination. This chops the spectra into several narrow-band spectral channels without introducing noise. When these spectral sections are combined into the final spectrum, noise outside their spectral band are omitted. This effectively raises our sensitivity by a factor of the number of spectral channels, N.
• A laser metrology system with nanometer resolution overlaps the starlight beam greatly reducing systematics caused by vibrations and thermal fluctuations.
• Because of our dispersive element breaks the spectra into N channels, the interferogram can be sampled N times fewer. This allows us to integrate on the star at each delay position N times longer than a conventional FTS, and greatly reduces the errors caused by photon noise.

The principal advantage of using an FTS is that the wavelength scale in the spectrum is known absolutely. For example, if one were to shine white light through a prisim and image the resulting spectrum using a CCD, it would not be possible to know the wavelengths of spectral features, such as absorption lines, precisely. One technique is to pass the beam through an absorption source with known wavelength lines. Marcy and Butler, two of the leading extrasolar planet finders, use an iodine absorption cell to calibrate their echelle spectrometer. The disadvantage is that this technique has systematic errors that prevent radial velocity precision beyond a few m/s. With our FTS, we hope to achieve a precision better than 1 m/s.