The CDS NIS Intensity Calibration.
CDS was calibrated pre-launch using a transfer standard source. Following launch, an in-fight calibration was estimated from the preliminary report on the calibration (Bromage et al. 1996) and the work leading to the publication by Landi et al. (1997) and shown by them in their table 1 and figure 1.
Since then the calibration report has been finished (Lang et al. 1999). This contains, inter alia, a re-evaluation of the laboratory measurements and gives the wavelength dependence of that calibration. In addition the CDS NIS calibration has been compared to that obtained by two calibration rockets, one with the NRL EUV Grating Spectrometer (EGS) (Brekke et al. 1999), and the other with the NASA Goddard Solar EUV Research Telescope and Spectrograph (SERTS).
NIS 2 - First Order
The pre-launch calibration of CDS (Lang et al. 1999) indicated that the absolute responsivity in the NIS 2 spectral waveband was known to ± 30% when the uncertainties in the aperture determination ( ± 8%), the events to photon conversion factor ( ± 23%), the detector operating voltage change between laboratory and in-flight use ( ± 16%) and the calibration (from a synchrotron source) of the secondary standard source used ( ± 7%) are combined. Brekke et al. (1999) report comparisons of in-flight measurements using CDS to simultaneous measurements from a NASA/LASP sounding rocket experiment (EGS) which had been calibrated against a synchrotron (Woods et al. 1994, Woods et al. 1997). These measurements gave a calibration of CDS NIS 2 varying in uncertainty from ± 15% at 584 Å to ± 25% at either end of the NIS 2 waveband. Although the laboratory NIS 2 events to photon conversion was taken overall to be ± 23%, the measurement point at 584 Å was estimated to be correct to ± 9% yielding (4.70 ± 0.42)x10 -4 events/photon while the comparison with the rocket spectra gave (4.75 ± 0.71)x10 -4, in excellent agreement. The calibration adopted for CDS NIS 2 for version 2 of the CDS calibration was that derived by Brekke et al. (1999). Version 4 of the CDS calibration is based on a reanalysis of the same data, with a more realistic treatment of the detector burn-in.
Figure 1 shows the various NIS 2 calibrations, plotted against the reanalyzed data from Brekke et al. (1999).
NIS 1 - First order
For NIS 1, the overall uncertainty for the pre-launch calibration of CDS (Lang et al. 1999) indicated that the absolute responsivity in the NIS 1 spectral waveband was also ± 30%, in this case the events/photon having an uncertainty of ± 25%. The comparison of CDS and the EGS rocket experiment spectra only yielded one comparison point, for Mg IX 368 Å . There the events/photon were (1.86 ± 0.46)x10-4. The laboratory calibration gave (6.77 ± 1.69)x10 -5 which is about a factor three lower. In doing the calibration work for NIS 2 it was possible to observe the source spot image using the largest slit and the very bright He I 584 Å line. However, for NIS 1 no single isolated and bright enough line was available to check the alignment. The difference in laboratory and in-flight calibration is consistent with, and is attributed to, undetected misaligment during the NIS 1 laboratory work. Other sources of disprepancy which would cause a change in NIS 1 but not NIS 2 are discussed in Lang et al. (1999) in their comparison of the laboratory and predicted events/photon conversions. These other sources can also be ruled out in the present case.
The in-flight calibration adopted on 23 Dec 1998 is thus the point obtained for the 368 Å line from comparison with the EGS rocket experiment. This is extended to other wavelengths by adopting the slope of the laboratory measurements. A misalignment should not affect the slope of the laboratory events/photon versus wavelength conversion. The events/photon uncertainty is taken as ± 25% close to 368 Å otherwise ± 45% is assigned.
Analysis of the coordinated measurements with the 1997 flight of the SERTS sounding rocket further refined the NIS-1 calibration. These data, when combined with the 368 Å data from the EGS rocket indicated a return to a calibration with a slope much closer to the original calibration curve, as shown in Figure 2. The error on this calibration is ± 15% over the SERTS range, and ± 25% close to 368 Å.
NIS 2 - Second order.
The laboratory calibration of CDS yielded a sensitivity of (1.32 ± 0.11)x10-5 events/photon for the second-order He II line at 304 Å. The two sounding rocket experiments yielded somewhat higher results of (1.64 ± 0.12)x10-5 for the EGS rocket, and (1.68 ± 0.25)x10-5 for the SERTS rocket, in excellent agreement with each other. The SERTS experiment also yielded a measurement of (1.65 ± 0.25)x10-5 for the Si XI line at 303 Å.
A recalibration of the EGS and SERTS data yields updated values of (1.83 ± 0.19)x10-5 and (1.87 ± 0.28)x10-5 respectively at 304 Å. The combined value of (1.86 ± 0.22)x10-5 has been adopted as part of CDS calibration version 4. The adopted curve is shown in Figure 3. In the CDS software, this second-order curve is referenced as "NIS-4".
Comparison with SUMER
Intercalibration studies of NIS2 with SUMER have been analysed and are reported in Pauluhn et al. (1999). They show that CDS measures generally a 30% higher intensity than SUMER in the He I 584 Å line, while it measures 9% and 17% higher intensities in Mg X 609 Å and Mg X 624 Å, respectively. Both instruments are shown to have a very good temporal correlation and stability, indicating that solar variations dominate over changes in instrumental sensitivity for the intercalibration datasets.
The following table shows the progress of the various calibrations implemented within the software:
The implementation of these calibrations (up to version 2) has been described in an email message from Bill Thompson. It is important to note that any raw data run through the current software will use the LATEST calibration values irrespective of the observation date, but instrumental effects (eg MCP burn in) will always relate to the actual observation date. There has always been the facility to retrieve earlier intensity calibrations (see Bill's note), but that feature was at a fairly low level - ie as a keyword to the database access routines. There is, however, a routine QCALIB (currently in /cs/update) which allows an easy, quick comparison for any particular date.Version Date Notes 1 16-Sep-1996 Original calibration, based on pre-flight absolute measurements and in-flight line ratios. 2 23-Dec-1998 Modified calibration, based on EGS sounding rocket measurements and reanalysis of the pre-flight data. 3 28-Feb-2000 Modified NIS-1 calibration, based on SERTS and EGS sounding rocket measurements. NIS-2 curve unchanged. 4 21-May-2002 Modified NIS-2 calibration, based on reanalysis of EGS sounding rocket measurements. Incorporated second-order NIS-2 curve. Modified GIS calibration curves.
Eg. the command
IDL> qcalib, '1997/10/1'
would produce the following plot.
Initial estimates were that the post-recovery calibration of CDS was the same as before the accident. However, more recent analysis suggests that there was some loss of sensitivity in NIS-1 and second-order NIS-2. More detailed analysis, including the 1998 flight of the EGS sounding rocket, and the 1999 flight of the SERTS sounding rocket, is required before this can be confirmed.
Brekke P, Thompson W T, Woods T N and Eparvier F G 2000, Astrophys
J. 536, 959.