An overview of the spectral resolution capacity of each flight device is given in section 1.3 of the latest (Oct. 1998 rev2.2) ACIS calibration report. It presents an average spectral resolution for each node in each flight device obtained from XRCF phase I telemetry data. It also points to available spectral resolution subassembly data sets.
In the following we will have a closer look at those data with the emphasis on spatial variations across the node. For the time being this includes primarily analysis of the ACIS subassembly data sets, which are taken with the single devices and its associated electronic units. These data differ to some extend from the XRCF phase I data, which are telemetry from the integrated ACIS flight unit. At a later time we will, however, include the XRCF phase I data in this analysis. From the CTI analysis we know that additional noise is introduced when charge is clocked through the device. In the case of front-illuminated devices (FI) CTI is sufficiently small not to expect significant variations of the spectral resolving power across a single node as can be seen in the example of w182c4r in S2. Back-illuminated devices show higher CTI and a spectral resolution that degrades with increasing distance to the readout node, as can be seen in the example of w134c4r in S3.
Description of the Data
Extensive calibration measurements were performed at the MIT Center for Space Research Those measurements were designed to determine of the energy-to-pulse-height relationship, spectral resolution, and other dependencies of the response function as a function of instrument parameters such as temperature, detector clock levels, and readout mode (Bautz et al. 1997). A detailed description of the various measurements can be found in the (Oct. 1998 rev2.2) ACIS calibration report. Each flight CCD was illuminated at 15 X-ray energies. Unfortunately the X-ray sources were not entirely monochromatic and suffered from a variety of impurities and the data have to be properly filtered by pulse height.
Table 1 shows a list of the X-ray sources with the corresponding line energies, total accumulated counts per energy and device quadrant, and the approximate exposure time. The measurements are exclusively in flat field mode and were designed to determine the energy scale with an error less than 0.1%. Although most measurements were not designed to test spatial dependencies to a high accuracy, the statistics allows us in most cases to determine the centroid of a line in a 64x64 detection cell with better than 1% accuracy. Count rates did not vary throughout each measurement and were low enough that only single photons were counted.
Table 1: List of X-ray sources with average counts par frame and total number of counts available (in brackets). The device w182c4r represents measurements of front-illuminated CCDs, w134c4r and w140c4r are the two back-illuminated CCDs.
Front-illuminated Devices (FI) [IN PREPARATION]
Back-illuminated Devices (BI)
Because of some extended treatment during the manufacturing process back-illuminated devices show a higher affinity towards CTI effects. This became already a fact during Lincoln Lab screening of the old flexed devices, which indicated about a factor 100 higher charge loss per pixel transfer than observed in FI devices. It also appears that the two BI devices in the ACIS-S array themselves behave quite differently. The average FWHM versus energy in both devices, as specified in the ACIS IPI report, differ from 90 eV to 220 eV for w134c4r (S3) to 150 eV to 270 eV for w140c4r (S1) between 277 eV and 8029 eV respectively. In the following we therefore treat both BI CCDs as different devices.
The shift of pulse heights affects the spectral resolution of the device. However, besides the expected global shift in the FWHM value due to the CTI shift, a significant increase in noise can be observed away from the read out nodes. In the following we will present the measured change in spectral resolution across each node in devices w134c4r (S3) and w140c4r (S1) at the 15 energies available during subassembly tests. Later we will add similar results obtained from XRCF phase I telemetry data. While the figures exclusively show the variation of FWHM, the accompanying tables also show the shift of the centroids.
w134c4r in S3 (BI)
In backside devices the energy scale is not as linear as it appears in front illuminated devices. Figure 61 (Figure 1.5 in the ACIS calibration report) shows the linear energy scale residuals for the back-illuminated devices. Depending on energy, these residuals show extensive deviations with up to +/-30 eV near oxygen, which more than 30% of the spectral resolution in w134c4r, which averages to about 100 eV at O-K in the phase I data analysis. Table 62 and figure 62 show the spectral resolution of w134c4r averaged over the whole node for XRCF phase I data (from table 1.7 of the ACIS Calibration Report). The fits in figure 62, unlike fits to front side devices, are best fit polynomials with no physical meaning.
Table F, Table G, Table H, Table G, and Table J show the variations of spectral resolution as observed during subassembly measurements. The results show, that a single average resolution for each single node by no means appropriately describes the device behavior. The variations between near and far from read-out are quite substantial, as they differ from node to node within a single energy. The difference between extreme values of FWHM within a single node also seems to increase with energy. It also can be observed that above 1 keV the spectral resolution in the middle of each node is quite in agreement with the average FWHM measured for the phase I data. Below 1 keV the subassembly data show much higher resolution throughout the nodes, which may reflect the difference in noise produced by the different electronic units.
Table F: 3-dim figures and tables of spatial variations of FWHM and line centroids across each node for line measurements C-k (277 eV), O-K (525 eV), and F-K (677 eV).
Table G: 3-dim figures and tables of spatial variations of FWHM and line centroids across each node for line measurements Al-K (1487 eV), Si-K (1740 eV), and P-K (2014 eV).
Table H: 3-dim figures and tables of spatial variations of FWHM and line centroids across each node for line measurements Ti-K (4511 eV), V-K (4952 eV), and Mn-K (5899 eV).
Table I: 3-dim figures and tables of spatial variations of FWHM and line centroids across each node for line measurements Fe-K (6403 eV), Co-K (6930 eV), and Ni-K (7478 eV).
Table J: 3-dim figures and tables of spatial variations of FWHM and line centroids across each node for line measurements Cu-K (8048 eV), Zn-K (8639 eV), and Ge-K (9252 eV).
w140c4r in S1 (BI)
In backside devices the energy scale is not as linear as it appears in front illuminated devices. Figure 61 (Figure 1.5 in the ACIS calibration report) shows the linear energy scale residuals for the back-illuminated devices. Depending on energy, these residuals show extensive deviations with up to +/-30 eV near oxygen, which more than 20% of the spectral resolution in w140c4r, which averages to about 135 eV at O-K in the phase I data analysis. Table 123 and figure 123 show the spectral resolution of w40c4r averaged over the whole node for XRCF phase I data (from table 1.7 of the ACIS Calibration Report). The fits in figure 123, unlike fits to front side devices, are best fit polynomials with no physical meaning.
Spatial variation analysis in preparation.......