Spitzer Documentation & Tools
IRS Instrument Handbook


9.3  Calculated Values for Merged Spectra

9.3.1             SNR Estimates  (keywords: SNR_SL2, SNR_SL1, SNR_LL2, SNR_LL1)

An SNR estimate is given for each order in the merged spectrum.  It is computed as the mean of all “flux density” values divided by the mean of all “error” values.  These values are only in the product headers, not in the Catalog.

9.3.2             Spectrophotometry (keywords: IRAC8, IRAS12, IRS16, IRS22, MIPS24, IRAS25)

Source photometry was synthesized from the merged spectra in 6 standard filters:

IRAC 8, IRAS 12, IRS blue peak-up (PU 16), IRS red peak-up (PU 22), MIPS 24, and IRAS 25 (Fig. 9.1).  The merged IRS spectra were convolved with the band-pass transmission curves to compute the flux at the effective central wavelength in each filter (Section 9.5).

The uncertainty values for the IRS Enhanced spectrophotometric fluxes were computed by propagating the uncertainties from the merged spectra.  An additional uncertainty factor of 1.1-1.8% was added in quadrature with the pipeline uncertainties to account for additional sources of photometric noise, such as slit-loss or flat field noise (Section 9.5).  This ‘repeatability’ uncertainty factor was measured for each order from the fractional standard deviation of the ensemble of HR 7341 primary standard star flux measurements.

IRS Enhanced spectrophotometry was checked against actual IRAC 8, PU 16, PU 22, and MIPS 24 observations of IRS calibration stars (Fig. 9.2). These calibration stars were observed many times by IRS, so median flux values and their 1σ errors are shown for each band .  The synthetic IRS photometry for these stars matches the IRAC 8, PU16, PU 21, and MIPS 24 photometry to better than 5%.




Figure 9.2 - Comparison of median synthetic photometry of IRS flux standard stars with actual IRAC 8, MIPS 24, PU 16, and PU 22 measurements of these same stars. These stars were observed many times by Spitzer IRS. The vertical error bars represent the standard deviation of the mean over all visits for the IRS spectrophotometry.  The slopes of the best-fit (dotted) lines give an indication of how well the synthetic photometry compares to the integrated light measurements of the same stars. Unity slopes are indicated by the solid lines.




9.3.3             Source Colors (keywords: mips24_irac8, mips24_irs16, irs16_irac8, iras25_iras12)

Four source colors (MIPS 24/IRAC 8, MIPS 24/IRS 16, IRS 16/IRAC 8, and IRAS 25/ IRAS 12) are computed from the ratios of synthetic fluxes in five of the photometric bands. The distribution of MIPS 24 versus IRAC 8 fluxes is shown in Figure 9.3.  The locus of stars with Rayleigh-Jeans spectra stands out with a MIPS 24/IRAC 8 flux density ratio of ~0.2. Star-forming galaxies typically have MIPS 24/IRAC 8 colors in the range 1-10.



Figure 9.3 - MIPS 24 versus IRAC 8 synthetic IRS spectrophotometry. A narrow locus of stars with blue (Rayleigh-Jeans) MIPS 24/IRAC 8 colors of ~0.2 and a broader band of galaxies with red colors in the range of 1-10 are both apparent.  IRS calibration stars (blue diamonds) and galaxies and AGNs (red diamonds) are shown for reference.


9.3.4             Extracted Source Positions (keywords: RA, DEC, RA_SL, DEC_SL, RA_LL, DEC_LL)

Source positions in each slit were computed from the location of the source along the slit determined from the spectral trace (‘ridge’) measured by the SSC pipeline, compared to the slit center position, as recorded in the FITS header for each exposure (see Section 5.7.2 in the IRS Instrument Handbook).  This computation interpolated between the “_FOV” and “_SLT” coordinates given in the header of the bksub.tbl files. If SL is available, the reported right ascension and declination (RA, DEC) is for the average of all SL exposures and nods. If SL is not available, then LL is used. The position angles (PA_SL and PA_LL) are the same as PA_FOV in the bksub.tbl files.

A plot of extracted minus requested positions for calibration targets (Fig. 9.4) gives an indication of positional uncertainty, for observations using high-accuracy blue peak-up. Location in this plot will depend upon the accuracy of the requested position and peak-up accuracy, as well as the intrinsic pointing accuracy of the telescope.  Circles drawn in Figure 9.4 show the 68%, 95%, and 99% error circles for positional accuracy. The radii of these circles are 0.36”, 0.62”, and 1.35”, respectively, after deleting 12 deviant points from Campaign 7 (3 of which are plotted; 9 fall off the plot), when there was some difficulty pointing the telescope.  For all other campaigns, the 1σ uncertainty on the extracted IRS source positions is 0.36”.



Figure 9.4 - Distribution of IRS position measurements for IRS calibration sources, relative to requested position. Circles with radii of 0.36”, 0.62”, and 1.35” enclose 68%, 95%, and 99% of the measurements, respectively. The three X’s mark observations taken during Campaign 7, when observations were affected by known telescope pointing issues. (Nine additional deviant observations from this campaign fall off of the plot.)

9.3.5             Spatial Profile Widths (keywords:  WIDTH_SL2, WIDTH_SL1, WIDTH_LL2, WIDTH_LL1)

Spatial profiles are generated for each background-subtracted spectrum as part of the standard SSC pipeline. The computation of these profiles is described in section 5.7.1 of the IRS Instrument Handbook.  The FWHM (“) of these profiles are reported in the Catalog for each observed spectral order.  The FWHM is found by first finding the maximum flux in the profile. Starting at this maximum point, the profile is searched on the left and right side for the first flux value less than half of the maximum flux.  The adopted width is the distance between the left point and the right point.  A profile FWHM value is not considered reliable if the SNR (“SNR_*” keyword) is less than 5.0.  In this case the WIDTH value is set to -99.0.

The distribution of profile widths is shown in Table 9.3 and Figure 9.5.  The distribution of profile widths for IRS calibration stars  has 95.4 percentile widths in  SL2, SL1, LL2, and LL1 of 2.56”, 2.68”, 6.97”,  and 8.34”, respectively (Table 9.3).  Source profile widths may be compared to these values to determine whether a source may potentially be extended.   However, these fiducial values are strictly only valid for a source with a stellar (Rayleigh-Jeans) spectrum.  For sources with redder spectral slopes, we have separated blue and red sources based on a MIPS 24/IRAC 8  flux ratio of 1.0, which roughly divides stars from galaxies in Figure 9.3.  The 95.4 percentile widths for blue sources range from 2.88”-22.62”, going from SL2 to LL1 (Table 9.3).   The 95.4 percentile widths for red sources range from 5.54”-27.69”.  Red sources generally have larger profile widths in a given order because their spatial profiles are weighted more heavily towards the red end of the spectrum, which is broader in our diffraction-limited data, and because (red) galaxies are more likely to be spatially resolved than (blue) stars.

Table 9.3 - Distribution of spatial profile widths:


68.2% FWHM(“)

95.4% FWHM (“)

99.7% FWHM (“)

SL2 Cal Stars




SL2 Blue




SL2 Red




SL2 All




SL1 Cal Stars




SL1 Blue




SL1 Red




SL1 All




LL2 Cal Stars




LL2 Blue




LL2 Red




LL2 All




LL1 Cal Stars




LL1 Blue




LL1 Red




LL1 All







Figure 9.5 - Distributions of IRS spatial profile widths for the SL2, SL1, LL2, and LL1 orders. The profile widths generally increase going from short to long wavelengths, reflecting the spatial resolution of the telescope and instrument. Distributions are shown for blue sources (blue lines) and red sources (red lines), separated at a MIPS 24/IRAC 8 flux density ratio ratio of 1 (Fig. 9.3).  Calibration stars (not shown) follow a distribution similar  to that of the blue sources. Galaxies tend to be red sources with MIPS/IRAC8>1.

9.3.6             Order Ratios (keywords: RATIO_SL2_SL1, RATIO_Sl1_LL2, RATIO_LL1_LL2)

Each of these values is the ratio of the mean flux density at the edges of two neighboring orders.  If a ratio differs significantly from unity, it could indicate any of the following. (1) The observed position falls away from the intended source, which misses or is offset perpendicular to one of the slits. (2) The observations are pointed at different sources. (3) The source is extended. A RATIO_SL1_LL2 of less than unity may indicate an extended source, where the larger LL slit admits more flux than the smaller SL slit. (4) There are strong spectral features near the order edges. 

Order ratios are calculated as follows:





The edges of orders are in the ranges shown in Table 9.4.


Table 9.4 - Orders and wavelengths in a Merged Spectrum file:



“LO” edge (micron)

“HI” edge (micron)




7.25 - 7.56   



7.57 - 8.15

13.60 - 14.28



14.29 - 15.05    

19.80 - 20.66



20.68 – 22.30