Claude Canizares reviewed the resolving power () of the 3 AXAF-I spectrometers: ACIS ( = 10-60), HETG ( = 100-1000), and LETG ( = 100-1000). The gratings are, of course, designed for point sources. The loss of the XRS (AXAF-S) means a loss of ability for high-resolution spectroscopy on diffuse, extended sources, and a loss of resolution above 5 keV.
The physical content of spectral measurements was also reviewed. Spectra contain information about temperature, abundance, density, equilibrium and motion of the emitting material. ASCA spectra were used to illustrate Doppler shifts (Cas A, SS433) and spectral imaging (Cas A). The AXAF-I gratings will be able to measure spectral broadening down to 100 km/s.
Work is needed to develop procedures and tools for the analysis of AXAF spectra. ASCA data show that present models do not always produce good fits. It will be necessary to incorporate new atomic physics into models and codes, e.g., to predict the strength of the Fe L lines. There is a large volume of atomic data that needs to be checked to ensure that all relevant microphysical processes are incorporated.
Webster Cash asked about background problems and how users might be warned of analysis pitfalls. Answers ranged from `Caveat Observer' to suggestions for ASC instructions to users. Steve Kahn pointed out that journals should have more observational expertise in the refereeing process, and that some major past problems could have been avoided with this approach. Webster Cash suggested that someone in the ASC should work with the journals to ensure that they have a list of good referees.
Steve Kahn also mentioned that model-fitting software with a large number of parameters needs improvement, so that the many local minima can be avoided. Mark Birkinshaw stated that he has had encouraging results using simulated annealing.
For more details see AXAF-I's Spectroscopic Capabilities.