Change page style: 

Telluric, flux and wavelength calibration

As at near-IR wavelengths, telluric standard star observations are required for mid-infrared spectroscopic observations to cancel telluric (atmospheric) absorption features in the data. The following is a guide to assist in selecting the most appropriate telluric standard stars. Two telluric standard observations must be included in the phase II definition for each observation lasting more than ~30 minutes, one airmass-optimized for measurement before the science target, and one for after the target. The Gemini staff member at the telescope will select one of these for baseline calibration, which will not be charged to the science program. If the PI desires both calibration stars or any additional calibration stars to be observed, these additional calibrations will be charged to the science program and should be accounted for in the initial time request. Please see the definition of baseline calibrations for more information and the observing tool libraries for example observations. Additional standard observations for spectrophotometric flux calibration, although not part of the baseline calibration set, are also discussed below, as is wavelength calibration.




Phase II science programs for spectroscopic observations must specify telluric standards, usually from the following lists:



Early K-type Cohen standards or B-G stars are recommended as telluric standards for lowN or lowQ spectra. B-G stars or asteroids should be used for medium resolution spectra, and in nearly all cases asteroids should be used for echelle spectra.

Mid-IR spectral features in standard stars: which spectral type is best?

At low resolution B, A, F, and G stars have smooth spectra in the 7-25um region, and therefore may be used to correct for atmospheric absorption features. In late K and M stars the fundamental vibration-rotation band of SiO significantly depresses the spectrum at 7.5-10um and therefore affects ratios (see eg Cohen & Davies 1995, MNRAS, 276, 715, Fig. 3). Thus late K and M stars should be avoided. A list of B, A, F, and G stars that are likely to be bright enough for observations in the mid-infrared is available. They are excellent for the removal of telluric features, but their mid-infrared fluxes are in general not well known and thus are not useful for accurate flux calibration of spectra.

Many of the Cohen bright spectrophotometric standards are early K dwarfs and have the added advantage of having accurate IRAS mid-infrared flux densities, making them potentially decent spectrophotometric calibrators (i.e. providing telluric line removal and flux calibration). At low resolution the fundamental vibration-rotation band of SiO significantly depresses the spectrum at 7.5-10 microns in stars later than K0III - K2III, and therefore affects ratioing (see eg Cohen & Davies 1995, MNRAS, 276, 715, Fig. 3). Nonetheless, provided the Cohen model template spectra are used to remove this band (this is taken care of by the Gemini IRAF spectral reduction task msabsflux, which can also be called from within msreduce), any of the Cohen stars can be used for lowN band spectra (see Cohen et al. 1999 AJ, 117, 1864). The low res Q-band spectra of the Cohen stars are smooth and thus they also can be used to calibrate lowQ spectra. However, programs aimed at detecting or measuring the detailed shapes of weak spectral features on strong continuum sources in the N or Q bands should use Cohen stars with caution, except for a few of the primary and secondary standards such as alpha CMa which have good quality ISO spectra.

At medium and high resolution the Cohen stars are not suitable as telluric standards because their spectra contain a multitude of photospheric absorption lines.

When using the early-type stars as telluric standards, a check should be made (using the Michelle or T-ReCS ITC or TEXES sensitivities, for example) that the star is bright enough to produce a high signal-to-noise spectrum in a relatively short observation. N-band magnitudes of 1.5 or less are sufficient for lowN spectral calibration with a 2 minute on-source observation, while for lowQ the magnitude should be 0 or less to produce the same result. The same applies to the Cohen stars, but these are generally brighter than the early-type stars at N-band and Q-band. Asteroids are often a better choice for telluric standards at Q-band because they are somewhat brighter than the majority of the stellar calibrators.

Airmass considerations - cancelling telluric features

Phase II Science Programs must specify two telluric calibration stars, one that would match the airmass of the target if observed before it and one that would match it if observed after. The PI should factor into their choice of calibration stars the amount of time to be spent observing the science target. If, for example, the science target requires two hours, ideally the first standard would have an RA 1 hr earlier than and a Dec equal to that of the science target and the second would have an RA one hour later than the target and a similar Dec. The elevation plot facility in the observing tool can be used to check that the chosen standard stars are appropriately matched in airmass to the science target.

Absolute flux calibration

If absolute flux calibration is required the program must include obtaining a spectrum of a Cohen bright spectrophotometric standard. Several of the stars in the list of very bright mid-infrared standards in the Southern Hemisphere that are commonly used by the T-ReCS Team are also useful spectrophotometric calibrators. Time used to observe the spectrophotometric standard will be charged to the observer's program if a separate telluric standard (B-G star or asteroid) is also needed.

Wavelength calibration

Spectra can be wavelength calibrated using sky emission and/or absorption lines observed in standard stars and/or the raw science data. Low resolution atmospheric transmission spectra are available (10um [PDF] and 20um [PDF]), as are ASCII files of atmospheric transmission at spectral resolutions up to 25,000. Tools to calculate Mauna Kea transmission spectra and produce line identifications at echelle resolution are also available.

The line identification process is described in some detail on the spectroscopy data reduction page.