Return to the EIT Home Page
1999-May-12 - Calibration and software updated.
1999-June-09 - Bug fix for Emission measure calculation via EIT_TEMP.PRO
1999-November-17 - Calibration and software updated.
2000-January-05 - Relative response normalization and call to EIT_PREP ] updated.
An initial calibration was given by Delaboudiniere et al. (1995) and incorporated into the analysis software as part of the SolarSoftWare package. This is the calibration used in all previous EIT papers. This calibration relied on part by some modelling and as well as measurements of the witness mirrors instead of the flight mirrors. The main calibration data were obtained with the Orsay synchrotron and presented in the Ph.D. thesis of Song (1995). This data set is believed to contain a more realistic calibration. The Song data has been re-analyzed in two separate works, first that of Defise (1999, Ph.D.) and by Dere et al. (1999). The calibration work by Dere is a more detailed examination of the pre-flight calibration. This work is presented in Dere et al. Solar Physics, 2000. Comparisons of Dere's with other calibrations are not yet available.
The online software has been updated on November 17, 1999 to reflect a modified version of the Dere calibration. The modifications include the far wings of the bandpass and a probably more correct treatment of the second order He II for the 304 bandpass. Note, for almost all work, these modifications have a negligible effect on the final output. To graphically view the calibration see Calib.ps. The solid lines in this plot are the Dere calibration, while the dashed are the modifications. This modified calibration is presented in Cook et al., Ap.J. 2000.
In addition to these pre-flight calibration updates, work has been done using the in-flight data. In particular, an in-flight measurement of the photons/DN conversion was made by Moses et al. (1999) which will replace the theoretical conversion used in old calibrations. This will be incorporated with the post-flight calibration updates. A final photometric comparison with the Naval Research Lab EIT CalRoc (Moses et al., 2000) is still pending. It should be noted that these are changes in the photometric calibrations and do not change any conclusions based upon relative changes.
| Sector | Plots | |
|---|---|---|
| Fe IX/X 171 Å | Linear | |
| Fe XII 195 Å | Linear | |
| Fe XV 284 Å | Linear | |
| He II 304 Å | Linear | |
As is well known, the response of the EIT CCD has changed with time. The variation of the instrument throughput is monitored by the total flux in a full field image for each bandpass. A detailed discussion of this and the following brief comments can be found in Moses et al. 1997. The utility of this monitor is determined by the variability of the solar flux in each waveband. The intrinsic solar variability increases with the temperature of the dominant emission line the bandpass. Thus the monitor is very good for the 304 (He II) and the 171 (Fe X,IX) channels where the solar variability is low, while it degrades for the 195 (Fe XII) channel and is useless for the 284 (Fe XV) channel where changes in the instrumental response are masked by solar changes. Since the extremes of the wavelength regime are 171 and 304, this technique allows the full wavelength dependence of the degradation to be monitored.
Since the 304 channel is the most sensitive to degradation, it best illustrates the processes. The initial increase in response over the first month of observations is attributed to an overall outgassing of the instrument. In order to reverse the subsequent decline in response, the CCD was heated (baked out) to 18C on 23 May 1996. Initial recovery was to the highest throughput observed in flight. Successive heat cycles were conducted according to the requirements of the observing schedules and in exploration of the causes of the decline. The degradation process consists of several components which are difficult to separate in detail. The two basic processes contributing to the degradation are 1) the absorption of EUV before it interacts with the CCD by a surface contaminant and 2) the reduction of charge collection efficiency (CCE) in the CCD due to EUV induced device damage.
The response degradation was not uniform. Instead the degradation was patterned in proportion to the EUV exposure. As the average EUV intensity in the coronal and transition region images is highly non-uniform, the calculated flat-field is distorted by strong residuals of local solar activity features, especially outside the disk area. The radiation damage therefore is largely unaffected far from the solar limb, while the degradation is strong (50\,\% in this case) over the solar disc, in the form of a negative imprint of an average solar disk.
Many users of EIT data wish to know the conversion from DN to "real" units, i.e. photons cm-2 sec-1 sr-1 or ergs... This is not a trivial as one would hope. The main complication, is that EIT is not a spectrometer but a broad band (FWHM ~10 Ang) instrument containing various emission lines formed at various temperatures in each bandpass. As the response for each bandpass is not a square well, each line contributes a different amount of the total flux, according to the relative response and strength of the lines. Therefore, there is no unique transformation of DN to real units!!! Instead, it depends upon the Differential Emission Measure (DEM) of the formation region and knowledge of the instrument response.
How do we proceed? There are a couple of ways to move forward. In terms of the instrumental response, one can make assumptions (e.g. square well, dominance of some lines) or one can use the actual response. As for the line formation, we can proceed in two ways. The DEM of a region of interest may be known from another source, e.g. averages given as part of the CHIANTI package or CDS, SERTS, etc, measurements. Then one can use a spectral package (CHIANTI) combined with the DEM bandpass information to produce a synthetic spectrum. This gives you the absolute flux in terms of real units. A different method presently being developed (Cook, Newmark, and Moses) is to use the EIT data itself to compute a rough DEM curve and then the conversion to real units.
The one exception to this is the 304 bandpass. This bandpass is dominated by He II (70-95% on disk depending on type of region) while most of the remainder of the flux is due to the nearby Si XI line (dominant off disk), which therefore has the same instrumental response. Therefore, conversion of DN to irradiance for this bandpass is fairly straightforward.
Cook, J., Newmark, J., and Moses, J., 1999, in prep.
Delaboudiniere, J.P. et al, 1995, Sol. Phys. 162,291.
Defise, J.M., 1999, Ph.D. Thesis.
Dere, K.P., et al., 1999, in preparation.
Moses, J.D., et al, 1997, Sol. Phys. 175, 571,
Moses, J.D., et al., 1999, in preparation.
The first step to obtaining calibrated EIT data is to run EIT_PREP:
IDL> eit_prep,file, hdr, image, /normalize, /filter_norm, /response_norm
or
IDL> read_eit,file, index, data
IDL> eit_prep, index, hdr, image, data = data, /normalize, /filter_norm, /response_norm,
This will give you a dark current subtracted, flat fielded, degridded, exposure time normalized, a first order response normalized image.
****Note, at this time the relative response correction is only good to +/- 10% overall, but can be worse on a pixel by pixel basis.
The next series of steps is complicated because of the various assumptions discussed in the previous section. Interpretation of this is not straightforward and it is recommended that the user consult with an EIT team member!
In order to obtain bandpass information use the routine EIT_PARMS. Please read the header information for all keywords.
IDL> bandpass = eit_parms(waves,band,units=units) ;band=171,195,284, or 304
One can visualize the bandpass by:
IDL> plot_io,waves,bandpass
In order to convert observed DN to real units, now one must choose the pathways described above. To obtain the simplest upper limit one can simply use: (remember the poor assumptions in this)
IDL> peak_calib = 1./interpol(bandpass,waves,wave_of_interest)
IDL> calib_image = image * peak_calib
If one assumes the structure you are looking at is an isothermal plasma then you can use the EIT_FLUX and EIT_TEMP routines discussed in Chapter 3 in order to obtain an emission measure.
If one has a DEM curve, either from CHIANTI or another source put in the same format as CHIANTI, then one can use the CHIANTI package to obtain absolutely calibrated lines by:
IDL>SYNTHETIC,Wmin,Wmax,Fwhm,Pressure,Lambda,Spectrum,List_wvl,List_ident
Return to the EIT Home Page The SDAC home page The SOHO home page