Subject: Re: xi Per outline From: "David H. Cohen" Date: Thu, 06 Apr 2006 16:37:51 -0400 To: Derck Massa CC: Alex Fullerton , Raman Prinja , dcohen1@swarthmore.edu Derck et al., Thanks for sending the outline. I agree with Alex that it's a good starting point. And also I agree with Alex that the porosity/clumping situation is complicated - in that porosity (large scale clumps, effective opacity reduction from the point of view of x-ray transfer) implies clumping, but not necessarily vice versa. But clumping (on any scale) implies mass loss reductions. You don't come out and say it in the outline, but from talking with you, I'm assuming that you're thinking that xi Per, more than other O stars, will/should show evidence of porosity associated with CIR-type structures. And so, from an x-ray spectral line point of view, may show significant porosity effects, above and beyond mass-loss rate reduction effects. I guess I'd have to say that my preliminary analysis is that the spectrum looks like zeta Pup's and zeta Ori's (and probably iota Ori's and delta Ori's and Cyg OB2 8A's...). That there's no strong evidence for even lower effective opacities than one would expect given how the other O stars look (and so no clear evidence of large-scale structures like CIRs from the x-ray profiles alone). Ditto the f/i ratios (though it's not clear to me if we'd really expect anything different with xi Per than with the other O stars) - hot plasma at a mean radius of ~2.5 Rstar. So, along these lines... I would certainly recommend removing "With the exception of zeta Pup" from the very first bullet point. The main conclusion of my recent paper on zeta Ori is that zeta Pup is not an exception, it is typical. All the O stars - except theta1 Ori C - have x-ray line profiles (and f/i ratios - stay tuned for a new paper momentarily) that are fully consistent with the wind-shock paradigm except for the large effective opacity reduction (factor of 5 to 10...in zeta Pup, zeta Ori, and it looks like now xi Per...and most of the others). I guess, maybe this is disappointing, but I don't see how we are going to be able to connect the Chandra spectrum of xi Per to any specific picture of this star, rotational modulation and whatever it implies about the wind, etc. The only thing I could think of - beside the possibility of an additional porosity effect - is if we had enough signal, we could make separate grating spectra of the high-state and low-state portions of the observation and see if any of the diagnostics differ (though how we'd interpret any differences, I'm not sure). Now, I can tell you that I have just implemented the porosity-based line profile model into the xspec fitting program. I've been running some preliminary fits, looking at joint constraints on tau_star and the porosity length (see Owocki and Cohen 2006 if you need to remind yourself about the porosity length). What I'm finding is that there isn't enough signal in the data to definitively tell them apart - in other words, you can get formally good fits with a smooth wind and very low optical depth (tau_star) or you can get good fits with higher tau_star but big porosity lengths. Now, looking at this tradeoff is interesting. And one approach might be to say, OK, we know M-dot (maybe two versions - standard M-dot and PV-based, lower M-dot), and that dictates tau_star, so how big would the porosity length have to be for this expected tau_star to be consistent with the data. Ultimately, I will be doing some studies of this sort, but xi Per could, in principle, be the first place this is presented (the higher S/N spectrum of zeta Pup will lead to better constraints). Oh, and finally - how would you feel about including Asif on the author list? He's done most of the reduction and analysis. If you'd rather not add two authors (me and Asif), I'd be happy to let Asif have the slot (but I could still be the conduit of x-ray spectral information between him and you, read the manuscript and give you feedback, etc.). OK, shifting gears a bit: We've made some more progress with the spectral analysis, and produced and updated some figures. Here's the summary of where things stand now: New, more cosmetically pleasing two-panel figure of the whole spectrum (also looks like the old one only included one spectral order -whoops! - so these new ones show about twice as many counts): http://astro.swarthmore.edu/~cohen/projects/xper/spectrum1.ps http://astro.swarthmore.edu/~cohen/projects/xper/spectrum2.ps Note that there are pdf files too. A spreadsheet summarizing our fitting results (note that we indicate what lines we tried to fit but were too weak to give meaningful constraints, and also that we fit Gaussians too - and that these are formally good (the result of poor signal-to-noise)): http://astro.swarthmore.edu/~cohen/projects/xper/xper_fits_april06.xls Now, individual lines/complexes (some of these have changed only cosmetically since the last batch we showed you; note: the confidence limit plots - I wouldn't bother showing any in the paper - just quote formal uncertainties in a table, but I'm including them for your reading pleasure; also I haven't provided the figures showing the Gaussian fits, but I have them if you want them): Ne X Ly-alpha at 12.13: http://astro.swarthmore.edu/~cohen/projects/xper/NeX_1213_windprof.ps 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 O VIII Ly-alpha at 18.97: http://astro.swarthmore.edu/~cohen/projects/xper/oviii_1897_windprof.ps 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 Fe XVII at 15.01: http://astro.swarthmore.edu/~cohen/projects/xper/fexvii_1501_windprof.ps http://astro.swarthmore.edu/~cohen/projects/xper/confidence_windprof_fexvii_1501_30x30x30_q_tau.ps http://astro.swarthmore.edu/~cohen/projects/xper/confidence_windprof_fexvii_1501_30x30x30_q_umax.ps http://astro.swarthmore.edu/~cohen/projects/xper/confidence_windprof_fexvii_1501_30x30x30_tau_umax.ps Fe XVII at 17.05, 17.10 (note that they're blended and have to be fit simultaneously; we fixed the relative normalizations at the theoretically expected ratio and let q, tau_star, and u_max be free parameters but forced them to be identical for the two different lines - the best-fit tau_star is kind of high, but with big uncertainties - see the spreadsheet; this is likely because the lines are blended): http://astro.swarthmore.edu/~cohen/projects/xper/fexvii_1705_d_windprof.ps No confidence region plots yet. Fe XVII at 15.26 (weak; parameters can't really be constrained in a meaningful way): http://astro.swarthmore.edu/~cohen/projects/xper/FeXVII_1526_windprof.ps Mg XI f-i-r: http://astro.swarthmore.edu/~cohen/projects/xper/coadded_poisson_MgXI_fir.ps (we're going to remake this plot, with the three lab wavelengths indicated; plus we still need to put formal error bars on the ratio; however, here's some modeling of the f/i ratio): http://astro.swarthmore.edu/~cohen/projects/xper/xiPer_MgXI_ftoi.jpg Have to put an error range on this. As you can see from the spreadsheet, we've also got a measurement of Ne IX f/i, but have yet to model it. Would you like me to put together a latex-formatted table of the fit results? David Derck Massa wrote: > Hi Guys > > The attachment is a rough outline of how I see the xi Per paper going. If no one has a major objection to the overall form or theme, then I'll start flushing it out. > Basically, I see the paper as having two objectives: > > 1. To imply that the X-rays in O stars are associated with large scale structures in the wind (a soft sell, since we certainly don't have proof). > 2. To set up future marketing, so that follow on observations for variability will seem important > > I've attached both a pdf and the latex file, in case you want to fiddle with it. > Derck > > ------------------------------------------------------------------------ > > %\documentclass{aastex} % default manuscript style % > \documentclass[preprint]{aastex} % one column, single-spaced % > %\documentclass[preprint2]{aastex} % two column, single-spaced % > % % > %==============================================================================% > \citestyle{apj} % for BIBTeX % > \renewcommand\ion[2]{#1\,{\sc{\romannumeral #2}}} % Better ion command % > %!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!% > % % > %==============================================================================% > % Define manuscript-dependent marcros % > %------------------------------------------------------------------------------% > \newcommand{\copernicus}{{\it Copernicus}} > \newcommand{\ebv}{$E$(\bv)} > \newcommand{\funits}{erg/s/cm$^2$/{\AA}} > \newcommand{\fuse}{{\it FUSE}} > \newcommand{\grav}{$\log\,g$} > \newcommand{\ha}{H$\alpha$} > \newcommand{\hst}{{\it HST}} > \newcommand{\iue}{{\it IUE}} > \newcommand{\kms}{km\,s$^{-1}$} > \newcommand{\lsun}{$\log L / L_\sun$} > \newcommand{\mdot}{$\dot{M}$} > \newcommand{\mdotlab}{\mdot~[$\times 10^{-6}$ \msunpyr]} > \newcommand{\mdotq}{$\dot{M} q$} > \newcommand{\mdotqi}{$\dot{M} q_i$} > \newcommand{\msun}{${\rm M}_\sun$} > \newcommand{\msunpyr}{${\rm M}_\sun / {\rm yr}$} > \newcommand{\rsun}{$R / R_\sun$} > \newcommand\sk[2]{Sk\,{$#1{^\circ}#2$}} > \newcommand{\teff}{$T_{\rm eff}$} > \newcommand{\trad}{$\tau_{rad}$} > \newcommand{\vinf}{${v}_\infty$} > \newcommand{\vsini}{$v \sin i$} > \begin{document} > > \section{Introduction} > \noindent \underline{Background on X-rays in O stars} > \begin{itemize} > \item With the exception of $\zeta$ Pup, X-ray lines do not show expected profiles \item Analyses imply that the wind opacity is much smaller than expected. This implies that either \begin{itemize} > \item \mdot's are lower than thought, and/or \item the winds are porous due to clumping or {\em large scale} structures. > \end{itemize} > \item Lack of detectable X-ray variability from most O stars favors clumping. \end{itemize} > > \noindent \underline{Background on $\xi$~Per} > \begin{itemize} > \item A normal, single mid-O star, with a high (but not extraordinary) $v \sin i$ > \item UV wind lines and optical lines influenced by the wind both show a distinctive, well-documented, persistent 2.1d period (deJong et al.) > \item Consensus is that the period in the variability results from rotational modulation of large scale wind structures. \item Consequently, this star presents an excellent case to examine for X-ray variability in order to establish a link between X-rays and large scale wind structures. \end{itemize} > > \section{The Data} > > \noindent \underline{TBD -- should be standard} > > \section{Analysis} > \noindent \underline{The time averaged profiles} > \begin{itemize} > \item They appear similar to most mid-late O stars. > \item Profile analysis implies low optical depth and, hence, porosity. > \item Need quantitative comparisons from final numbers > \end{itemize} > > \clearpage > \noindent \underline{The time series} > \begin{itemize} > \item Details \begin{itemize} > \item Background limited > \item Advantages of studying the zeroth order spectrum > \item Characterization of the background > \end{itemize} > \item KS test for variability in the zeroth order spectrum > \item Results from a linear regression of the temporal data \begin{itemize} > \item Why least squares is appropriate for the data > \item 20--30\% change over the 2 days of observing run! > \item The probability that the observed trend is real > \item The probability that the observed trend is linear -- i.e., monotonic > \end{itemize} > \item Variability of other aspects of the spectrum, e.g., a super-line, the ``pseudo-continuum'', etc. (TBD -- preliminary analysis done, but may not add much. If so, may be best to leave it out.) > \item H$\alpha$ observations obtained with the hope of phasing the observations -- (Alex?) > \end{itemize} > > \section{Discussion} > \noindent \underline{X-ray profiles} > \begin{itemize} > \item How they compare to other stars (give numbers -- a table?) > \item Implications from the implied $\tau ({\rm wind})$ and fir analysis. \item Bottom line -- porosity > \end{itemize} > > \noindent \underline{X-ray variability} > \begin{itemize} > \item Did we observe a partial period? See the H$\alpha$ variability observed by Morel et al. (2004) -- its shape is very similar! So cyclic behavior is {\em plausible}. \item Compare level of variability to other single, normal O stars -- all part of a Table? \item $\zeta$ Oph is the only other one (true?) -- its {\it ASCA}\/ data also varies for 20\% (2001, A\&A, 378, L21). > \end{itemize} > > \section{Conclusions} > > \noindent \underline{} > \begin{itemize} > \item MUST verify whether the variability is periodic. However, the alternative is that the X-ray flux from the star dropped by nearly a third over 2 days as part of some sort of a secular variation -- even stranger! > \item Profiles indicate much of the X-ray flux we observe may come from deep within the wind, or even the far side of the star. > \item IF the variability is cyclic, then the amplitude is {\em very} large. In fact, if the X-ray variability originates in CIRs, such a large amplitude may indicate that ALL of the X-rays are formed by the CIRs > \item Then why don't the X-ray fluxes of all mid-late O stars vary? > \begin{itemize} > \item It may be that the CIRS must be viewed very near to equator-on to observe the variability (see, Dessart's model) \item This would be consistent with $\zeta$~Oph and $\xi$~Per being the two X-ray variable stars. > \item Does this imply that we would expect the optical depths of the X-rays from the winds of $\zeta$~Oph and $\xi$~Per's to be somewhat larger than in comparable stars? > \item If the X-rays originate in the CIRs, then why don't the UV observations of variable sources reveal distinctive modulations in the ionization states of the winds. \end{itemize} > % \end{itemize} > > % \noindent \underline{} > % \begin{itemize} > \item Many unanswered questions > \item Reiterate that it is essential to verify whether the variability is periodic. If it is, it could signify a fundamental change our view of stellar winds. \end{itemize} > > \end{document} > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > \noindent \underline{} > \begin{itemize} > \item \item \item \end{itemize} > > \noindent \underline{} > \begin{itemize} > \item \item \item \end{itemize} > > \noindent \underline{} > \begin{itemize} > \item \item \item \end{itemize} > >