IRS: Best Observing Practices

Additional important observation planning advice and information is available throughout the IRS SOM chapter. For example, we recommend that observers pay careful attention to all material that is presented in bold font or (especially) material that is enclosed in text boxes. The latter usually summarizes in one or two sentences an important aspect of IRS observation planning that is described at length in the surrounding text (see SOM index under IRS: boxed item for a list of page numbers). In addition, the following list of highlights includes references to SOM sections. The basic characteristics of the spectrograph are summarized in the IRS Pocket Guide, available from the SSC website. They are also available in SOM Table 7.1, Table 7.2, Table 7.5, and Figure 7.6. Assessing this information in light of your science goals should be the first step in determining whether to propose to use the IRS.

Updates 24 Jan 06:

• Tips and Techniques for Properly Designing IRS Spectral Maps
• Report on ultradeep IRS spectroscopy of faint sources (AORs to go with report)

Peak-Up Accuracy

Observers should be mindful of their required photometric precision, which will translate into peak-up accuracy. In general, high-accuracy peak-ups are not much more expensive than others and produce better results (spectroscopic repeatability of 1 sigma=2 to 5%), but medium accuracy peak-ups are appropriate for working with the wide slits (LL and LH). Low accuracy peak-up should not be used. See section 7.2.4.

Peak-Up Acquisition Targets

The observer should ensure that he/she has picked up an isolated point source (no objects within 60" × 80") within 30' of the target by checking the MSX, 2MASS, and/or DSS images. The source should have a flux density within 5 mJy and 150 mJy in the blue or 15 mJy and 340 mJy in the red, with the exact minimum value depending on the local background level. This value should be entered correctly in Spot, as the peak-up source will be rejected if the observed flux is discrepant from the predicted flux by a factor larger than 5. The observer should take care to enter the epoch (the year for which the position was obtained), not to be confused with the equinox (which determines the actual coordinate system used) of the coordinates and the stellar proper motion. The later should be given in arcsec/year. Note that peak-up stars from the 2MASS catalog do not include proper motion information, and that proper motions in the PCRS are in milliarcseconds per year. See section 7.2.4.4.

Redundancy

The observer should ensure that he/she has at least two cycles for each observation this will provide greater redundancy for cosmic ray hits and rogue pixel identification. For example, rather than having 1 cycle x 480s, the observer should choose 4 x 120s.

Spectroscopic Saturation

A-to-D saturation limits for the low resolution modules are ~5 Jy (10 Jy) at 10 microns (25 microns). For the high resolution modules, the limits are 60 Jy (70 Jy) at 12 microns (25 microns). This means that in the shortest possible observation (6 sec), at least one of the samples will be saturated. Given the characteristics of the SSC pipeline, saturation is not necessarily fatal although it will result in less signal-tonoise than expected. The observer should decide whether or not the loss of signal-to-noise in certain regions of the spectrum is compensated by a better exposure elsewhere. See Section 7.2.1.2.

Imaging Saturation

For the shortest exposure times, in low background, point sources brighter than 0.19 Jy at 16 microns will saturate the blue peakup image. Point sources brighter than 0.42 Jy at 22 microns will saturate the red one. Extended sources brighter than 450 MJy/sr for both the blue and red will saturate the detector in the shortest available ramps. As in spectroscopy, source saturation is not necessarily fatal, as the image flux can usually be recovered. See Section 7.2.3.4.5.1.

Sensitivity

FIn order to estimate the feasibility of his/her observations, the observer should use the sensitivity calculator available on the SSC website. Observers should be aware, however, that this calculator does not include any noise introduced by the extraction. It assumes perfect extractions with perfectly centered sources unaffected by rogues or fringing. These are very idealized conditions and observers should design their observations with generous signal-to-noise margins. See Section 7.2.2.

SL Observations

For the SL observations, the observer should check that there are no bright sources in the PU field of view. Since the same array is used for the PU and SL spectra, saturation on the PU FOV may introduce artifacts into your spectrum. See Section 7.2.5.1.

Sky Observations, high resolution

For the LH module, the SSC strongly recommends that the observer plan for sky background observations close in time to the primary observation for sources fainter than 2 Jy at 25 microns. Sky subtraction will significantly reduce the impact of the rogue pixels in the signal-to-noise. For the SH module, sky subtraction is recommended if accurate flux calibration is needed. The sky observations for the high resolution modules should be observations of nearby empty fields, as extracting sky measurements from within the target observations is almost impossible with the high-res slits. See Section 7.2.5.2.

Sky Observations, low resolution

The sky can be subtracted using the two nod positions or observations in which the target is in the other slit in the detector. The latter are possible because the observer automatically gets "off-slit" background measurements for (for example) order 2 when observing in order 1. If using off-slit sky observations, the observer should make sure that the off-slit regions are not contaminated by a bright source or extended nebulosity in the field of view. Also, the off-slit sky is only useful if the two orders have the same exposure time. Off-slit sky measurements may also be less accurate than nod-sky measurements, if separated by long times or if there is a large gradient in sky background. Observers interested in extended sources will need to assess whether or not these procedures are acceptable for their goals. If not, dedicated sky observations of empty fields will likely be necessary. See Section 7.2.5.2.

Spectrophotometric Calibration

The SSC provides the observer with photometrically calibrated products. The absolute photometric accuracy of point source spectra is limited by the accuracy of the models used for calibration and currently stands at ~10%. Representative slope deviations from the beginning to the end of an order with respect to point source models are 7% for the high resolution modules and 3% for the low resolution ones. Residual features in the flats produce systematic residual artifacts of 2 to 3%, which may affect detection of faint spectral lines. See Section 7.3.4.6.

High-Res resolution

Unresolved spectral lines are only Nyquist-sampled at the longest wavelength ranges in each module. If using the high resolution modules, the observer should understand that this will result in errors in the measured line widths. Observers should be careful when interpreting lines with R>~400 as resolved. See Section 7.1.6.

Go back to IRS page