From apj@phy.ohiou.edu Tue Aug 6 11:16:20 2002 Date: Sun, 4 Aug 2002 15:47:38 -0500 From: apj@phy.ohiou.edu To: yaqoob@pha.jhu.edu Cc: ApJ-MS56330@mss.uchicago.edu, apj@phy.ohiou.edu Subject: Your ApJ Submission MS# 56330 Dr. Tahir Yaqoob Department of Physics and Astronomy Johns Hopkins University 3400 North Charles Street Baltimore, MD 21218 USA Dear Dr. Yaqoob: Appended below is the referee's report on your submission to the ApJ entitled "The Kinematics and Physical Conditions of the Ionized Gas in Markarian 509. I. Chandra High-Energy Grating Spectroscopy" ( MS# 56330). The referee has indicated that the results of this paper merit publication, but has noted a number of substantive matters that should be dealt with before the paper is accepted. I would be happy to consider for publication a revised version that addresses these points. When you resubmit the paper through the ApJ ftp site, please include a detailed cover letter containing the (mandatory) listing of the changes you have made to the text and your responses to the report. Processing of your revised manuscript will be expedited if you make your revisions to the manuscript latex file available for downloading from the ApJ Web Peer Review System (http://mss.uchicago.edu/ApJ/). This version includes your previous submission plus any modifications to latex commands necessary for smooth processing by the ApJ electronic system. If you have any questions, feel free to contact me. Best regards, Joseph Shields, Scientific Editor The Astrophysical Journal Phone:740 593-0336 Fax:740 593-0433 apj@phy.ohiou.edu ************************* The Yaqoob et al. paper "The Kinematics and Physical Conditions of the Ionized Gas in Markarian 509. I. Chandra High-Energy Grationg Spectroscopy" presents new data about the warm absorber in this well-studied Seyfert 1 Galaxy. I will highly recommend it for publication after the following comments are addressed. First some general comments. (1) The authors did a very thorough job of considering the many factors which may affect their modeling efforts, including X-ray continuum shape, multiwavelength SED, and velocity profile. I have some concern about the measurement of a Doppler b value, which I discuss below (See comments (13) and (16)). (2) Even though it is likely that a multi-zone or continuous range of ionization states is present, a single zone is apparently adequate to describe the observed X-ray lines. For example, the high-ionization species Si XIV is not present. But what about ionization states between those observed in the UV and those observed in X-ray? In particular, it would be interesting to measure or put constraints on inner-shell transitions from O I-VI, Si I-XII, etc.. which have been seen in other Seyferts in the X-ray band. **(3) What are the more general implications of this study? I come to the conclusion that warm absorbers contain clouds with a wide range of ionization states, and that the UV and X-ray absorbers are associated but not identical. What is implied for the mass flux in the ionized outflow? Are the absorbers associated with the BLR or NLR, or is this not yet clear? The high densities implied by the Pounds et al. result suggest BLR. I am skeptical of this, since the emission lines and associated absn lines are narrow. Kraemer et al. 2002 (your Paper II) suggest an origin in the NLR for a similar reason. Next some specific comments: (4)Intro, para 3, sentence 1: "..models span a wide range in distance...out to the putative (parsec-scale) molecular torus" AND the extended NLR. C.f. Seyferts such as Mrk 3, NGC 4151, NGC 1068, with extended X-ray emitting NLRs. (5) Section 2, sentence 1: "..duration of ~59 ks" In a few places, you give: ~59ks, 57.950 ks, and ~60 ks. Make sure these are consistent. (6) Section 2. Impressively detailed Observations section. Kudos. (7) Section 3. para 3: "The origin of the soft excess is unclear...." I would also add here the possibility that a relativistic O VIII Ly-alpha line contributes at least part of the soft excess. This is important, because the assumed continuum shape may affect your warm absorber column density measurements. (See 21 below.) (8) Section 3, para 4: "The measured edge energies are in excellent agreement with the expected values" I note that the O VII edge energy is marginally inconsistent at the >90% conf level. It concerns me that the O VII edge is contaminated by absorption from Fe M-shell UTAs in some objects, and may also be here. You discuss this later but should also note it here. This can affect your fit ionization parameter if you overestimate the O VII column density. A similar problem with other Fe transitions may cause an overestimate of the O VIII edge depth. You do address the Ne IX, but not the Fe here. This is a crucial point to address. (9) Section 4.1, para2, Figs 4, 5, 6, 7 seem a bit redundant. I would kill figures 4 and 5 and combine figures 6 and 7. One can see the S/N from your errorbars and it is not necessary to include the counts spectra. I will not be adamant about this, since it is a matter of taste. (10) Section 4.2, para 1, Last sentence: "..there may be emission filling in the trough.." OR contamination from other absn lines, OR a truly different absn profile. (11) Section 5, 2nd sentence: "Table 2 gives the abundances we used..." How did you decide on these particular abundances? Reference? (12) Section 5.2, para 1, "...we fixed n_e at 10^8" and "..data were indistinguishable for densities in the range ne=10^5-10^11." The high densities you consider seem to betray a prejudice toward an origin in the BLR. Why not "n_e=10^2-10^11", to include the possibility that the absn is coming from the NLR? **(13) Section 5.2, para 2, 1st sentence: "...all models assumed a velocity turbulence of 1 km/s, and since XSTAR broadens lines by the greater of the turbulent velocity and the thermal velocity, it is the latter which is relevant for these models." This sentence leads me to believe that you are using purely the thermal velocity width for the fit, which is much less than 100 km/s. But then in the next sentence, you correctly recognize that the choice of b is vital for getting the proper equivalent widths and fitting for the column density. I hope this is just a misstatement, since it appears that you _did_ fit with b=100 km/s (Section 5.4)? Please clarify. 14) Section 5.3, para 2: "We had to apply an additional (but non-physical) correction of -270 km/s... XSTAR v2.1.d contains wavelengths of SOME atomic transitions which disagree with wavelengths published elsewhere." I have to say that this does not inspire great confidence in your fits!!! Are you confident that your correction does a decent job of matching the wavelengths of _ALL_ the important observed lines? "It is the published values which we used to construct the velocity spectra in Fig. 8 and Fig. 9". Why don't you use these values for your model fits? Accurate wavelengths are really important for fitting the line strengths. 15) Section 5.4, 1st para: "the strongest absorption line features are NVII... ..and Mg XI 1s2.." I don't see the Mg XI line in Fig 3. It would be good to include it there. **16) Section 5.4, 1st para (curves of growth): Clever idea to use lines from multiple ions to constrain a Doppler b-value. However, you need to state that this depends on the assumption of a particular set of abundances (eg solar). Also, you use a somewhat circular argument: points in Fig 10 use the "predicted column densities". Predicted from where? From the model which you fit with b=100 km/s? A more correct argument might go: All Ly-alpha transitions from H-like and He-like ions have the same equivalent width (to within uncertainty). This suggests that all of these lines are on the saturated part of the COG. The measured equivalent width of ~1 milliAngstrom is consistent with b=100 km/s, assuming solar abundances. The best way to determine an independent b-value is doing a curve-of-growth analysis on at least 2 lines from the same ion. Can you do this for O VII 1s-2p through 1s-5p?? 17) Section 5.4, 1st para (velocity profiles): Fig 8. and Fig. 9. I don't understand why you don't simply plot the fit model against the data? Why do you go through this complicated excercise of re-calculating the model with Gaussians of the appropriate b and EW? It leads me to believe that the original model was somehow inadequate, along the lines of comments 13 and 14 above. 18) Section 5.4, para 3, 1st sentence: "...the best-fitting XSTAR model is moderately good at the O VII edge but should be somewhat lower in flux blueward of the O VII edge." I don't follow. Do you mean "good at O VIII", but "poor at O VII," or pretty good at "O VII", but not quite deep enough at the threshold energy? Please restructure this sentence. 19) Section 5.4, para 4, "We see that more sophisticated modeling with better data now gives a smaller O VIII edge." After giving your new, improved edge optical depths, I would add the caveat that they still may be overestimated due to the presence of Fe inner shell UTAs. (See comment 8 above). 20) Section 5.5, para 2, 1st sentence: " C IV Lya" should be "C VI Lya". 21) Section 5.5, last sentence: "..the evidence for relativistically broadened O VIII... is not compelling". While it may not be compelling, neither is it ruled out. I would be careful here and suggest that further observations are required. Also, how much did your O VII column change when you added in the O VIII rel. line? I would guess that it decreased. 22) Section 6, para 2 "...a broad absorption trough, likely to be associated with an Fe M-shell UTA" Which ionization state of Fe? **23) Section 6, para 3 "...the emitting gas did not respond to the continuum in ~180 days implies....d>0.15pc". Let me get this straight. The continuum increased by a factor of 2.8. The intercombination line EW did not change significantly. Therefore the flux of the line could have increased along with the continuum, and therefore MAY HAVE responded. This would invalidate your size lower limit. If the line flux was constant, the EW should have dropped by a factor of 2.8 to EW=0.5 eV; which is also consistent with your observation. I would say your observation is insufficient to establish whether or not the emission line flux varied! **24) Section 6, para 3 "We verified the conclusion of Pounds et al.(2001), that the line-emitter must have ne>10^11." You are saying that XSTAR model spectra require ne>10^11 in order to explain the line ratios observed by Pounds et al. So you haven't actually verified the Pounds et al. observation; you just concur that if they measured these lines properly, it supports a high density? I would eliminate this statement since you aren't actually providing any new data on the line ratios. Furthermore, I have looked up the Pounds et al. reference. They find: O VII r: EW=0.5+/-0.5 eV i: EW=1.5_/-0.7 eV f: EW=0.6+/-0.4 eV So the i line is the only one detected, and at best at 2.1 sigma. Are you willing to support such a detection by "verifying" it with your own modeling efforts? R=f/i=0.4+/-0.3 (R<1.0 at 2-sigma level) is the relevant density sensitive diagnostic. From Porquet and Dubau 2001, Fig. 8, I read ne>10^11), which is as Pounds and Reeves stated. Also see PD01, Fig. 11, it looks like ne=10^12 is necessary to produce the observed f/i ratio as well. I belabor this point because I think ne>10^11 is rather unlikely, but very interesting if true. This density would put the absorber in the BLR at 0.1 pc, which seems inconsistent with its narrow width (few hundred km/s). I will not be surprised if longer observations of Mrk 509 do not confirm the Pounds et al. result. 25) Section 8, para 3 (point2), line 7: floating period "O VIII ." 26) Conclusions, point 4: I would eliminate point about emitting material being at least 0.15 pc from center (see comment 23 above). I hope that you find the above comments useful and not overwhelming. Overall, it is a great paper. Pay special attention to asterisked comments 3, 13, 16, 23, and 24. *************************