Differences between ATOMDB Version 1.2.0 and MEKAL/SPEX
We compare element-by-element line spectra from ATOMDB v1.2.0 and SPEX V1.1.0. (Although newer versions of SPEX may be available, the v1.1.0 models are similar to the MEKAL model currently available in XSPEC and Sherpa.) These figures provide valuable information for X-ray spectral analysis near 1 keV, and illustrate the major differences (click to see in PDF).
| Oxygen | Neon | Silicon | Iron | Nickel |
 |
 |
 |
 |
 |
We also show comparisons of Iron emission lines in Energy (keV) units here:
There are some noticeable differences between ATOMDB and MEKAL/SPEX, even at moderate (CCD) resolution (R ~ 30-50)
- ATOMDB does not yet include fluorescence lines, whereas SPEX does. The most significant implication is that the ATOMDB/APEC models do not include the Fe K 6.4 keV (1.9 Angstrom) line (see other caveats).
-
In Oxygen, the most significant differences occur in the O VIII Ly alpha/ O VIII Ly beta ratio, as shown by Smith et al. (2001). Shown here is a comparison of Oxygen Line Spectra at 40 Million K from ATOMDB and SPEX/Mekal:
- Other H-like ions show a similar difference in Ly alpha/Ly beta ratio. Silicon also shows large differences in He-like emissivities at low temperatures, presumably caused by differences in the ionization balance. SPEX assumes Arnaud and Rothenflug (1985), while ATOMDB v1.3.2 uses Mazzotta et al. (1998). Here is a comparison of Silicon Line Spectra at 2.5 Million K:
- Fe L-shell data show significant differences. Some effects can be attributed to the differences in ionization balance (see (3) above). Other effects can be attributed to scaling laws used in SPEX, which are not used in ATOMDB (see, e.g. Gu et al. 1999). For example, the large differences at low temperatures are from Fe XVII only, and must be related to the scaling laws. Differences at higher temperature appear to be a combination of scaling laws and ionization balance effects for Fe XXI to XXII. The ATOMDB uses collision strengths produced by HULLAC and have been tested relative to the original calculations. Here are two comparisons of Iron Line Spectra, at log T=7.1 and 7.4:
Examples of noticeable differences between ATOMDB and MEKAL/SPEX at high resolution
- The DR satellite lines appearing near the resonance lines show significant spectral structure not captured by assumptions introduced in SPEX (e.g. near O VII 21.7 Angstroms at 3 million K and Fe XVII 15.0 Angstroms at 2 million K). Proper treatment of the DR lines is important for distinguishing between line and continuum emission (see Brickhouse et al. 1995 for details). Shown here is a comparison of DR Satellite Line Spectra:
- Fine structure lines are split up in the ATOMDB, but not always in the SPEX list. For example, the O VIII H-alpha lines near 102.5 Angstroms should be treated carefully when trying to measure the density diagnostic Fe XXI 102.22 Angstrom line:
- Wavelength differences are also apparent. The ATOMDB has included the measurements of Brown et al. (1998), and Brown et al. (1999, Internal Report LLNL.JC-136647B). H- and He-like wavelengths in the ATOMDB are from relativistic quantum mechanics calculations and include fine structure.
- SPEX includes inner shell excitation lines which are not currently included in the ATOMDB (see other CAVEATS), for example, the Fe line at 17.2 Angstroms; however, this line strength in SPEX is much stronger than in Raymond-Smith models and needs investigation. Note also wavelength differences noted in (3) above. Shown here is a Fe XVII Comparison:
- The ATOMDB includes more lines at high principal quantum number than SPEX (see Brickhouse et al. 2000). This is apparent near 10 Angstroms at 6 million K). However, ATOMDB models are still not necessarily complete enough in detail to achieve good spectral fits at high resolution, depending on the spectrum to fit. We do not believe ATOMDB is complete enough yet at soft X-ray wavelengths to perform global fits to LETG spectra. Shown here is a Comparison of High-n Lines in Iron between ATOMDB and SPEX:
