Grating Observations

Sgr B2 was also well observed with the medium resolution LWS grating mode. This was particularly important for L04 observations to determine the continuum level when the spectral coverage was very narrow. This meant that there were many grating L01 mode pointings across the source often corresponding to individual L04 pointings. These are shown in the right-hand plot of Figure 3.1. The grating observations are detailed in Appendix C. Due to the high flux from Sgr B2 the long wavelength detectors exhibited non-linear behaviour in grating observations. In this case it is necessary to apply the `Strong Source Correction' to recover the true continuum level. This correction was developed by Leeks (2000). The non-linear behaviour seen in grating mode is not present in high resolution observations because transmission through the FPs reduced the incident flux on the detectors.

One grating observation was included as part of the high resolution spectral survey (TDT 28701401) with the same pointing and orientation as the L03 measurements. This was used to determine the overall continuum shape of the source to compare with the FP observations. The basic reduction recipe for L01 observations was carried out and this is described below. The strong source correction and reduction of grating observations is described in more detail by Leeks (2000).

Figure 3.23.2: Grating observation of Sgr B2 after defringing and strong source correction. No extra factors were applied to individual detectors to improve the stitching. The red line is a dust corrected blackbody fit to the data as described in the text.

Each spectral point in the observation was made with a voltage ramp of 0.5 s. The strength of the incoming radiation was measured from the slope of these voltage ramps with time. For Sgr B2 the slope of the voltage ramps were high enough to cause significant debiasing of the detectors, causing a non-linear change in detector responsivity. This caused a curvature in the voltage ramps which reduced the measured slope. In order to reduce the effects of this debiasing in the ramps, the last half of each integration was discarded, effectively reducing the time spent on each point to 0.25 s. This did not remove all of the effects of debiasing and a further correction was applied in the calibration of the relative spectral response (RSRF) of each detector. Mars was used as a calibrator with the standard LWS Interactive Analysis (LIA) version 8 Strong Source Correction routine (see Leeks, 2000). The spectrum also showed large scale fringing across each detector wavelength range. This occurred for extended or off-axis sources due to interference between the reflecting surface of the LWS mirror M2 and its support structure (see Section 1.8.3). The fringes in the spectrum were removed using the standard LIA `Defringe' routine. This routine solves for the period, phase and amplitude of a sine wave fitted to the data. A dust corrected blackbody function was fitted to data using the ISO Spectral Analysis Package (ISAP; Sturm et al., 1998). This produced a modified black body spectrum with equation,

$$F(\lambda)=\Omega (1-\exp{(-\tau)}) B(\lambda,T)\\$$ (C.2)

where the $B(\lambda,T)$ is the standard black body function and $\Omega$ is the solid angle of the emitting region (here constrained to be equal to the approximate LWS beam solid angle, $1.18\times10^{-7}$ sr). The dust opacity is a function of wavelength and the optical depth, $\tau$, is given by,

$$\tau=\tau_{0.55} \left(\frac{0.55}{\lambda}\right)^{\alpha}$$ (C.3)

where $\tau_{0.55}$ and $\alpha$ were parameters in the fit ($\tau_{0.55}$ is the optical depth at 0.55 $\mu$m). The best fit parameters found by the ISAP fitting routine were, $T=32.2$ K, $\tau_{0.55}=2141.1$ and $\alpha=1.45$. Figure 3.2 shows the fit with the data. This fit was used for comparison with the FP data and to remove the large scale continuum shape in some calibration procedures and was not used to determine physical characteristics of the source.

Figure 3.33.3: Relative spectral response function (RSRF) for each detector.