Derck, Raman, Alex, Asif, Asif has finished the line-by-line spectral analysis, and I've updated the files on the webserver: http://astro.swarthmore.edu/~cohen/projects/xper/ Much of what's there remains unchanged since we last gave you an update, but let me summarize it anyway: Spectral atlas, with lines identified: http://astro.swarthmore.edu/~cohen/projects/xper/spectrum1.pdf http://astro.swarthmore.edu/~cohen/projects/xper/spectrum2.pdf postscrpt versions are there too. I don't think these have changed since our last update. Spreadsheet summarizing fitting results to each line or line complex (including some we attempted to fit, but due to very low S/N, could not put meaningful constraints on - and which, therefore, we have no figures for). http://astro.swarthmore.edu/~cohen/projects/xper/xper_fits_april06.xls O VIII 18.97 (Lyman alpha) - these have NOT changed since we last updated the directory - progression of fits: Gaussian with fixed centroid, Gaussian with free centroid, wind-profile model: http://astro.swarthmore.edu/~cohen/projects/xper/oviii_1897_fixedgauss.ps http://astro.swarthmore.edu/~cohen/projects/xper/oviii_1897_freegauss.ps http://astro.swarthmore.edu/~cohen/projects/xper/oviii_1897_windprof.ps And visualizations of the 68% and 90% confidence limits on the three interesting parameters of the wind-profile model: http://astro.swarthmore.edu/~cohen/projects/xper/oviii_1897_conf2cntrs_q_vs_t.ps http://astro.swarthmore.edu/~cohen/projects/xper/oviii_1897_conf2cntrs_q_vs_umax.ps http://astro.swarthmore.edu/~cohen/projects/xper/oviii_1897_conf2cntrs_t_vs_umax.ps Summary: presence of some wind attenuation confirmed at >90% level; onset radius of x-ray emission is between 1.5 and 2 Rstar. Fe XVII 17.05, 17.10 - these lines are blended, and thus we fit two wind profile models simultaneously, fixing the parameter values of the two profiles to each other, and fixing the normalization (flux) ratio to the theoretically expected value of 0.8 (note that this implies that densities are not above ~10^14) - we tried fitting Gaussians, but the results were not well constrained; in fact, the wind-profile model results are not that well constrained either, with large uncertainties on the relatively high values of tau_star and u_o (though the 90% confidence limits encompass the values we find for the other, stronger and unblended, lines in the spectrum) - see the spreadsheet: http://astro.swarthmore.edu/~cohen/projects/xper/fexvii_1705_d_windprof.ps Fe XVII 15.26 - S/N low, best-fit wind-profile model is shown, but constraints on model parameters are uninteresting: http://astro.swarthmore.edu/~cohen/projects/xper/FeXVII_1526_windprof.ps Fe XVII 15.01 - these plots have been updated recently - again, we show a progression of fits: Gaussian with fixed centroid, Gaussian with free centroid, and wind-profile model: http://astro.swarthmore.edu/~cohen/projects/xper/fexvii_1501_fixedgauss.ps http://astro.swarthmore.edu/~cohen/projects/xper/fexvii_1501_freegauss.ps http://astro.swarthmore.edu/~cohen/projects/xper/fexvii_1501_windprof.ps Note: the fit domain is truncated on the red side of the line, to avoid blends with weak lines, longward of this. And visualizations of the 68% and 90% confidence limits on the three interesting parameters of the wind-profile model: http://astro.swarthmore.edu/~cohen/projects/xper/confidence_windprof_fexvii_1501_20x20x20_q_tau.ps http://astro.swarthmore.edu/~cohen/projects/xper/confidence_windprof_fexvii_1501_20x20x20_q_umax.ps http://astro.swarthmore.edu/~cohen/projects/xper/confidence_windprof_fexvii_1501_20x20x20_tau_umax.ps Summary: little direct evidence for wind attenuation, but upper limits on tau_star are consistent with results for other high S/N lines (several tenths). R_o a bit lower than generally found, but consistent with 1.5 Rstar at the 90% confidence level. Ne X 12.13 (Lyman alpha) - these have NOT changed since we last updated the directory - progression of fits: Gaussian with fixed centroid, Gaussian with free centroid, wind-profile model: http://astro.swarthmore.edu/~cohen/projects/xper/NeX_1213_fixedgauss.ps http://astro.swarthmore.edu/~cohen/projects/xper/NeX_1213_freegauss.psn http://astro.swarthmore.edu/~cohen/projects/xper/NeX_1213_windprof.ps And visualizations of the 68% and 90% confidence limits on the three interesting parameters of the wind-profile model: http://astro.swarthmore.edu/~cohen/projects/xper/confidence_windprof_contour_NeX_1312_q_tau.ps http://astro.swarthmore.edu/~cohen/projects/xper/confidence_windprof_contour_NeX_1312_q_umax.ps http://astro.swarthmore.edu/~cohen/projects/xper/confidence_windprof_contour_NeX_1312_tau_umax.ps Summary: presence of some wind attenuation confirmed at >90% level; onset radius of x-ray emission is around 1.5 Rstar. *Asif and I will make a set of plots summarizing the best-fit model parameters and their confidence limits for the strongest lines - summarizing values already in the spreadsheet - (like Fig. 6 in my paper on zeta Ori: http://astro.swarthmore.edu/~cohen/Papers/cohen_zori_lprofiles_mnras_2006.pdf) But the executive summary is - just as for zeta Pup and zeta Ori - some evidence of wind attenuation in the x-ray line profiles, but the optical depths we infer are an order of magnitude below what's expected from a smooth wind and the published mass-loss rate. Otherwise, the profiles are consistent with what's expected from a generic wind-shock scenario - broad lines indicating an onset radius for x-ray formation of roughly 1.5 Rstar, with a relatively constant filling factor above that, and some wind attenuation. ____ Now, for the He-like f-i-r complexes: Ne IX at ~13.55: Profile parameters - again, no strong evidence for wind attenuation, but upper limits are consistent with what we see in other lines. R=f/i = 0.15 +/- .05. Note that the umax (Ro) and tau_star constraints are similar to what's seen in other, unblended lines. Figure showing best-fit three component model (wind profile parameters free, but all parameters except the normalizations are tied together): http://astro.swarthmore.edu/~cohen/projects/xper/NeIX_1355_t_windprof.ps Mg XI at ~9.2: Simultaneous fit to the three components, but profile parameters can be constrained - like other lines, tau_star of a few tenths and R_o consistent with 1.5 to 2 Rstar. Best-fit profile model and data: http://astro.swarthmore.edu/~cohen/projects/xper/mgxi_92_fir_windprof.ps Derived value of R=f/i = 0.34 +/- .09 f/i modeling at: http://astro.swarthmore.edu/~cohen/projects/xper/xiPer_MgXI_ftoi.jpg (and .eps) Summary: We've fit a model with a *single* radius of formation, and get a best-fit value of 2.4 Rstar, with 1 sigma errors implying a range of 2.1 to 2.8 Rstar. Note that a more realistic model would include a spatially distributed x-ray emitting plasma. The lower limit to the radius of x-ray emission would be *less* than the best-fit single value (For information on how a *distributed* model of x-rays affects the f/i ratios observed in O stars, see: http://astro.swarthmore.edu/~cohen/projects/xper/leutenegger_f-to-i_ApJ_resubmit.pdf). So, this result is consistent with the Ro value derived from the line profile fitting. Si XIII at ~ 6.7: Simultaneous fits to the complex; S/N isn't so good, so constraints aren't significant. *I still need to model this one too. (But R=f/i ~ 2.0 looks like it may imply formation radius of several Rstar.) ____ Other X-ray data - Asif's made a preliminary search of the literature and also of NASA's X-ray databases, and found (generally short) observations from other instruments. These results are summarized in: http://astro.swarthmore.edu/~cohen/projects/xper/xper-literature.xls There are a few observations from ROSAT and Einstein. Note that the ROSAT data were only published in a conference proceedings (Palsa et al. 1996), with no separate discussion of individual stars. It is very difficult to make comparisons of observations taken with different instruments, as each has a different bandpass (thus what's meant by something as simple as "x-ray luminosity" is different for each instrument). To make such comparisons (to search for variability, as Raman suggested, requires modeling). Of course, separate observations made with the *same* instrument can be more easily compared. We've taken a look at the quick-and-dirty "standard extraction" light curves from the three ROSAT observations. Have a look at these data if you like: http://astro.swarthmore.edu/~cohen/projects/xper/ROSAT_lightcurves/ Note that one of the three observations was made with the boron filter, and thus modeling is also required here to make a comparison with the other, unfiltered, ROSAT observations. *We will continue to work on this and let you know what we find, but I don't expect the results to be conclusive (or all that interesting). ------- So, to summarize - 1. Fits to individual strong lines show consistency with a generic wind-shock scenario, but with effective optical depth reductions of an order of magnitude (conclusions similar to those for zeta Pup and zeta Ori). 2. f/i modeling - so far in detail only for Mg XI - shows formation radius consistent with wind-profile fits - 1.5 to 2 Rstar onset radius (or, if you fit a model that assumes a single formation radius, ~2.5 Rstar). 3. There's some data in the archives from Einstein and ROSAT, but the observations are short, don't show anything that's obviously of interest, in terms of short-term variability, and will be hard to compare with the Chandra x-ray fluxes (because of band-pass mismatches). Asif and I still have to do: a) Make plot summarizing wind profile parameter fits (like Fig. 6 in my recent paper on zeta Ori); this information is already summarized in the spreadsheet, though. b) Do the f/i modeling for Ne IX and Si XII (like what I did for Mg XI); in principle, we could model a spatially distributed source, too. c) Work on the ROSAT and Einstein light curves and, ultimately, a comparison with the Chandra fluxes. ------- Talk to you soon. David