From - Fri May 9 19:33:28 2003 Return-Path: Received: from postala.lbl.gov ([128.3.41.61]) by imapd.lbl.gov (Netscape Messaging Server 4.15) with ESMTP id HEHEO500.G7B for ; Tue, 6 May 2003 13:09:41 -0700 Received: from postala.lbl.gov (localhost [127.0.0.1]) by postala.lbl.gov (8.12.9/8.12.9) with ESMTP id h46K9cea026836 for ; Tue, 6 May 2003 13:09:39 -0700 (PDT) Received: from jimbean.lbl.gov (IDENT:Lwaf9MpwJm3y/PtOy8Zl9tzIm1CUb3i2@jimbean.lbl.gov [131.243.48.81]) by postala.lbl.gov (8.12.9/8.12.9) with ESMTP id h46K9crU026832; Tue, 6 May 2003 13:09:38 -0700 (PDT) Received: from localhost (dnkasen@localhost) by jimbean.lbl.gov (8.11.6/8.11.2) with ESMTP id h46K9Zx04684; Tue, 6 May 2003 13:09:35 -0700 Date: Tue, 6 May 2003 13:09:35 -0700 (PDT) From: Dan Kasen To: Gabriele Garavini cc: dnkasen@panisse.lbl.gov, Tony Spadafora , Greg Aldering , Ariel Goobar , Subject: Re: 1999aa draft In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hey Gabrielle, I enjoyed reading your 99aa paper and found it very interesting and (with a little rewriting) definately worth publishing. Below are some comments and suggestions -- probably more than you care to read but hopefully some will be helpful. My main comments concern the synthetic spectra fits and suggest a softening of some of your conclusions on line identifications and an overall more skeptical tone. In some cases your conclusions are difficult to assess because the relevant facts are spread out over many sections and not enough details are given; it would help to have in one place a complete and unified description of all relevant evidence for each of your conclusions (e.g. carbon lines, Nickel lines, high-velocity CaII, etc...). Before submission I would suggest a thorough reorganizing of the writing within individual sections. The most essential thing is to break each section into numerous individual paragraphs, with each paragraph focusing on one issue. If possible the first sentence of each paragraph should introduce or lead into what is going to be said in the rest of it. This will help your passages follow a logical progression and will help out those many readers who only have time to skim. I also noticed a number of ungrammatical or obscure sentences and some typos. I didn't bother to to point these out, though, figuring they will be caught in subsequent revisions by a careful re-reading by you or others. have fun, -dan ABSTRACT: Note that often the terms "direct analysis" and "detailed analysis" are used somewhat as opposites, direct meaning the empirical synow-level analysis you have done, detailed meaning more involved theoretical modeling that includes all relevant physics. It probably won't bother most people, but you may not want to use the description "detailed analysis". You say "carbon is found at high photospheric velocities", but I think you mean just "high-velocities". The photospheric velocity is normal but you find detached CII far above the photosphere. As discussed below, I would not mention the "evidence for highly ionized ions" here as these identifications are only probable and in any case finding highly ionized species is not interesting in and of itself. Also, the SiII is not "detached" until later times when the photosphere recedes below the photosphere. And is this Si layer any more "thin" then in typical SNIa? You mention the thiness many times but the discussion of the evidence for this was too spread out for me to draw a firm conclusion. I didn't see where you found an upper velocity limit to the IME region; perhaps you consider the CII line detached 19,000 km/s as the end of the IME and start of C/O rich region but I wouldn't call this an especially thin IME. INTRODUCTION: The first sentence would make more sense as something like: "Because of the homogeneity of their luminosity, TypeIa supernovae have become excellent tools for distance estimates." i.e the supernova are the tools, not the homogeneity. You may also want to define SNIa, so that you don't have to keep repeating typeIa SN. In the intro and other places you discuss how 99aa fits in with 91t and other "over-luminous" SN, but I didn't find a reference to how bright 99aa itself is (if this known). You might also find it relevant that recent Cepheid measurements to the host of 91t show that its overluminosity is marginal (.3 +/- .3 magnitudes brighter in B, according to Saha et al. 2001ApJ...551..973S) Among the other known 91t-likes, 1995ac appears clearly overluminous while 1997br could be about normal brightness. All three of these observed magnitudes are quite uncertain to due to large uncertainty in the dust extinction, but it does seem that the notion that the spectroscopic peculiarity of 91t-like SNe is clearly associated with over-luminosity can be questioned. 3, MODELS FOR TYPE1a SNe Split the section into 4 or 5 paragraphs and catch typos. I don't understand this sentence and what follows: "To match the observed intermediate mass elements (IME) composition,Branch et al. (1993), the models require the burning front to accelerate and reach the external layer." 4. SPECTRA COMPARISON Describing spectra in detail is very difficult since it is basically trying to put a picture into words. To me this section was an overwhelm of numerous comments that did not hold together. I would suggest: (1) Cut back on the quantity of information by eliminating all comments that are not essential or that can encompassed in a more general description. For example "The IME lines look very similar to the normal SN 94d" covers SiII, SII, CaII, MgII etc.. and there is no need to discuss these individually. You will have a more informative discussion of line ID in the synow section, so there really is no need to describe everything in detail again here. (2) Split each section into paragraphs that are organized logically. For example, one way to group the ideas would be two paragraphs for each epoch spectrum, one discussing how it is alike/different from normal SNIa and one how it is alike/different from 91t. 5. EXPANSION VELOCITIES Split the discussion into paragraphs. Also refer to individual figures as you come to them. You say the analysis here is not "quantitative" but it seems to me that it is in the sense that you are determining numbers and comparing them to other numbers. But I don't think that the systematic uncertainties in your fit model will be a serious problem. Your comment: "It is generally believed that the more energetic is the super- nova explosion the more 56 Ni is produced and then more luminous the object should be. This would suggest that over luminous object should have higher velocities but no such correlation has been found" I don't think this will be understood by most readers unless you explain a little more the logic behind it. The last sentence "The major peculiarity is the restricted atmo- sphere region in which SiII is present that makes clearly appear this ion only one week before maximum light." Again I am not clear what your final conclusion is for the extent of the IME is, but I don't think the reason that you don't see SiII early on is because it has a "restricted atmospheric region". More likely it is because, as suggested for 91t, a high temperature ionizes SiII. 6. SYNTHETIC SPECTRA p 10, first col: "in Sobolev LTE approximation and resonance scattering" would be better as "in the Sobolev approximation using LTE occupation numbers and a resonance scattering source function". p 10, second col: the Sobolev approximation does not necessarily imply homologous expansion; these are two separate assumptions. You say "in the present modeling scenario with 3D asymmetric complete NLTE models, SYNOW can...". the models aren't there yet; 1D fairly complete NLTE models exist and some simplified 3D ones are coming, but solving the full 3D NLTE problem is still several years away I think. The Synow fits look great. Most of your line identifications are reasonable and expected and do not need much elaboration. However the unusual identifications most be handled with care; these are: CII, CIII, OIII, SiIV, NiII, NiIII. Although you may be right about some of these, you should make it clear that you understand that they are initially to be treated skeptically. David Branch developed the following terminology in using synow for line identification. If an ion matches only one feature, its identification may be considered "probable" assuming that there is seemingly no other ion that can make that feature. An ion that fits multiple features in the spectrum may be considered a "definite" identification if no other set of ions can make the features. Its somewhat subjective, but I think some of your identifications, such as the carbon lines, should be softened to "probable". And you really shouldn't claim "identification" at all for an ion such as OIII or SiIV when it is only hinted at in a blend with other ions. While such ions may well contribute to the spectrum, you can't demonstrate any hard evidence for this using Synow -- that is just the limitation of a direct, empirical analysis as opposed to a detailed calculations where optical depths are computed self-consistently. The most you could say would be something like: "OIII may make a contribution to the blend at...." One other issue: why do you use a wide variety of v_max and T_exc in your fits? In previous synow papers, these parameters are held constant except for rarely when a change was motivated directly by the quality of the fit. v_max should always be set at the outer edge of the atmosphere (pick, say, v=30,000) unless you think you find evidence of an ion with an upper boundary. Crucially, each section needs to be organized logically into paragraphs. For example, a possible organization would be 1 paragraph discussing the expected line identifications, followed by one paragraph each for each unusual thing noticed (high-velocity CaII -- Carbon lines -- Nickel lines -- whatever else). Carbon lines: Somewhere in the paper (here or the conclusion, or both) a solid paragraph collecting all the information on the probability of carbon line identification and the implication of them is needed. From what I can tell, the strongest indication of carbon is the CII line near 6100. The crucial fact, which you point out, is that this feature cannot be fit with SiII and maybe more discussion and a blow up figure on this would be nice. Its not clear from your discussion if there are other CII lines that you have may have identified -- another line occurs 6900 but it may be too weak to see. Anything relevant should be included. The line you identify as CIII at photospheric velocities is pretty strange. I may be wrong, but I think this same line is seen in 00cx and was considered very mysterious. Look at the paper astro-ph/0302260 by R.C. Thomas et al, 2003. They tentatively suggest the possibility of high velocity Hydrogen Hbeta. This it is not totally implausible because in both 00cx and 99aa you see evidence for detached high velocity CaII at around the same velocities; however I bet you can rule out this possibility in 99aa because the Halpha line would be too strong (in 00cx you can hide Halpha in the SiII feature but 99a doesn't have one). Other possible identifications of the line in 00cx aren't mentioned in that paper but one other I know of is HeII. HeII would be strange because it is basically impossible to thermally ionize Helium at these temperatures -- the presence of HeII would imply a large degree of non-thermal ionization from the decay of radioactive elements (this is known to occur in type1b SN) which may suggest Ni56 out where the Helium line is. One problem with CIII is that this implies there is a fair amount of carbon at lower velocities and then you must ask why CII is only seen detached at higher velocities. At day -11 you might be able to get away with saying that the temperature is so hot near the photosphere that CII is ionized there (I don't think it likely that CII would be completely ionized all the way out to 19,000 kms, but I suppose its something to suggest). But that still doesn't explain why you don't see CII at later times, when the temperature cools off. It may be that at these times all the CII lines are completely hidden in blends, e.g the one near 6000 is swamped by the SiII feature. From the explosion models it is not impossible for there to be carbon at lower velocities, but people will find it strange. As you mention, the 3D deflagration models leave clumps of unburned C-O, although if the deflagration transitions into a detonation, these clumps will likely be burned away. I don't think people will believe that there is a ton of carbon down there -- if, for example, it was 50% C/O in the IME layer, then its likely that there wouldn't be enough Si, S and Ca to make the other lines look right. You can claim CIII as a "probable" identification of this line; but don't be overly confident about it and note the possible problems and alternatives if any. Always state conclusions as if-then statements: "If the identification of CIII is correct, this implies the presence of original unburnt material in the IME region of the atmosphere". In the end you can always admit that the "these question can only be fully addressed using detailed radiative transfer models" OIII,SiV: These lines come from highly ionized states and to my knowledge have not been identified in SN before. Although these ions may well contribute to the spectrum, you can't claim clear identification of them since they are blended in with other ions. Because these lines do not dramatically improve your fits or affect your conclusions, I would consider removing them from the paper (they may only arouse suspicion). Or if you do mention them, do it with some qualification -- "OIII may contribute to the spectrum at... but its presence is not definite " and don't use them to strongly justify your conclusions as you do when you say "the identification of CII, CIII, and OIII favor the presence of of an unburned C+O rich composition." To comfort skeptical readers you may want to mention if the optical depths you use are consistent with what is expected in LTE based on the plots in Hatano et al. Nickel lines: You haven't clearly shown what the evidence is for Ni lines -- much more detail is needed, as these lines are somewhat unusual (I think they've only been seen in 91t-likes). The most obvious NiII line observed 91t is around 5500 I think -- is it there in 99aa? From what I can tell, the evidence for nickel in 99aa is weak. Of course a non-detection is interesting in its own right, since it distinguishes 99aa from 91t. It is strange that you use nickel for the day -1 spectrum, but none for the day -11. The radioactive nickel decays to cobalt pretty fast so this be another reason to claim that the nickel you are talking about is stable. But it still is strange if you don't see the 5500 feature early on, when blending is low. A discussion of this would be welcomed. Mazalli and Ruiz-Lapuente claim that 91t had an "inverted" composition structure with Ni on the outside, IME in the middle and then a Ni core as you might see in a late detonation. That was challenged by Hatano 2002 and I think your analysis challenges it also. It may be interesting to look at those papers carefully in talking about Ni lines. Hi-Velocity CaII A solid paragraph somewhere (here or the conclusion) summarizing the evidence for high-velocity CaII would be interesting. It is your choice how much detail to go into, but you might mention (1) the double minimum visible in the HV IR triplet is a strong indication that your identification is correct (2) The two minima of the observed HV IR triplet appear to be about equal strength (although it may be hard to tell given the blending from OI and the continuum slope). Meanwhile -- as you see in your synthetic spectrum -- the blue IR triplet line is about twice as strong as the red. The equal minima depth can be reproduced in 3D (as discussed in my paper) if you have, say, an optically thick clump that only partially covers the photosphere. The 99aa feature resembles my plot of a clump viewed from a line of sight where it just barely covers the photosphere. I would also find it interesting to know how well a high-velocity component can be determined just from the H&K -- I mean, do you really need the HV Ca or can reproduce the same thing with a blend of Si or by using a shallower density profile? But the observation of a double-dipped IR triplet makes me convinced that your identification of HV CaII is right. Note high velocity CaII has also been seen in 00cx (Thomas et al 2003). One final thing about Section 6.5 +40 days: You might want to mention that a +40 days the spectrum is basically nebular and synow's assumption of a sharp photosphere is not expected to hold. However, as already found, synow still does a pretty good job and is still worth using. 7. SUMMARY You say here that: " As shown in Figs 12 and 13, FeIII lines do not appear to be formed at very high velocities, (< 12000 km/s). Thus, SN 1999aa does not necessarily require a more powerful explosion and a higher production of 56 N i". However previously you have said that there is no observed correlation between high velocity features and high explosion energy (section 5). Also the FeIII lines need not come from completely burned material since a good deal of iron still exists in the IME layer. Maybe the point you want to make is that, based on the SiII velocities, the IME layer seems to extend down to 12,000 kms, which means the Ni core does not appear to be overly big. But I think the issue of whether 99aa had a higher Ni56 production depends more on its absolute magnitude and light curve, and you haven't discussed either here. 8. CONCLUSIONS You say here that: "The iron peak core extends further out and dominates the spectra at larger radii than in normals." But in section 7 you seemed to be saying that the extent of the iron peak region was consistent with a typical w7 model. And elsewhere you say that the question of how far out the Ni core extends (whether 10,000 or 12,000) cannot be determined! I think the conclusion and summary need to be rewritten. Discussion of one point follows another that has nothing to do with it in no obvious logical progression. In the conclusion, you seem to be repeating alot of what you just said in the summary. Maybe the two can be combined into one section and broken into paragraphs, each of which states the evidence and discusses the implications of one individual finding.