Deepsearch Meeting Notes, 2000 January 26

Mark Phillips and Mario Hamuy are visiting this week. Peter is introducing Mark Phillips, a supernova pundit with whom Peter's been working since 1993 (and with whom Peter's published way more papers than he has with anybody in our group). Peter, with Mark and Mario are looking at the spectra of the Calan-Tololo supernovae, which haven't been published. Mario wants to get them out there, and Peter says that they are looking at a number of interesting supernova science things. Mark and Mario will each show us a few things.

Carl talks for a few minutes.

Mark Phillips and the LCO IR Supernova Survey

Mark says that he's a firm believer that where we should be spending our energy right now is on nearby supernovae. He thinks that there's a lot to be learned, especially if we're trying to learn about evolutionary effects (metallicity, mass, etc.). Mark mentions that he changed observatories a year and a half ago, and he's been doing some different stuff.

He tells us about the LCO IR Supernova Survey. (Las Campanes, near-infrared.) It's an attempt to characterize the JHK lightcurves of Type Ia (and other) supernovae. He's working with Mario, Chris Smith, Nick Suntzeff, Lou Strolger, Jose Maza, and others. They're using mostly the 1m telescope at LCO with an IR camera; they're also using the 2.5m. They've also started getting some IR data at CTIO with the YALO telescope; optical stuff is being done almost entirely at CTIO.

Alex Kim asks if it's a search, or followup. Mark says it's a combination. Lou Strolger's doing a search at Kitt Peak with the 0.9m (which we searched with them the first time around). Maza is also doing a plate search. But, they observe anything else that comes up. They aren't doing any IR searching, just IR followup.

The principal goals are to characterize JHK lightcurves; very little of this has been done in the past. They want to derive bolometric lightcurves in combination with UBVRI lightcurves. Finally, they want to improve the precision in distance measurements from Type I and Type II supernovae.

Badde-Wesselink Distances to SNe II is what Mario will be doing for his thesis. JHK has the advantags that extinction uncertainty is less than half that from optical photometry. Near IR spectra of Type II's have fewer strong absorption/emission features, making it much easier to determine the photometric temperature from JHK colors. Departures from a blackblody are less severe in the near-IR, translating to a flux dilution factor (zeta) cloer to unity and more constant with pahse than in optical. The factor is very close to 1, and virtually constant with time, Mark says. 1-2 so far have been close enough that you might also get Cephied distances. The have 3 more which are more distant. (Plus 10 without IR from the Calan-Tololo survey.) They hope to get a few more.

For Ia distances, JHK offeres two major advantage. One is again that dust absorption will be negligible except for the most heavily extinguished objects. (Most Ia's have fairly small reddening; in the IR, it's truly dinky in almost all cases.) The second is that the dependence of peak absolute magnitudes in the near-IR on the light curve decline rate (i.e. stretch?) may be small enough to be ignored for most SNe Ia. Something to look for, anyway.

(Brief aside about spectrophotometry in the IR.)

So far, they've had two campaigns in 1999. They got about 7 Ia's (of which 4-5 are good), 1 Ic, 5 II's. They had a Ic (1999ex) which was in the same galaxy as a Ia (1999ee), and it went off while they were following the Ia. So they should get a good rise measurement in the IR for the Ic.

(Aside about how much unreduced data everybody has.)

Mark shows some data, BFRIJHK data of 1999ac. This was a 91T like event, though a peculiar one (early spectra doesn't have Silicon, but does have Calcium). Peter says in that sense it's closer to a 90N supernova. (These guys, and their intimate knowledge of famous supernovae....) The lightcurves don't look particularly peculiar, however. This one was caught well before maximum, and is the first time anybody's really seen the rise in the infrared (12-13 days before max). Saul suggests that they could get a bolometric rise lightcurve.

Carl asks what the secondary peak in the redder lightcurves comes from. Mark says that he thinks its due to a change in the opacity in the infrared. It becomes more opaque, and the retreat of the photosphere slows down. Peter grunts agreement.

Mark shows a second example, closer to the limit of what you can do with a 40" telescope (1999aw). It's a little rattier looking, and wasn't caught nearly as early. He remarks how slow the decline rate of this is. Delta m-15 is 0.76, he says. (What stretch does this correspond to?) This is basically a hostless Ia; no obviously galaxy there. At z=0.038. Greg says you can just barely see the host on the POSS. Lou stacked the good seeing data, and said he could just see it in R.

Gerson says that we have on with a stretch of 1.55; 92bi (did I get that right), the first one found. That one did have a host. Saul points out that the stretch measurement on that supernova is not very well constrained.

Mark notes that in this, and in the previous, the H band lightcurve is really quite flat from maximum through the first 10 days. That's one reason H might be a nice place to observe. Of course, you need to see how this holds up, which of course is one of the reasons that they're doing this survey.

Mario asks how good the absolute photometry is; if you measure standards on different nights, etc. Mark says he hasn't tried it yet, so he doesn't know. He doesn't expect it will be as good as it is in the optical, but he doesn't think it will be bad.

Supernova Correlations with Host Environments (M. Hamuy)

Mario says that he once had some spare time, and tried to study the relationship between environment and type Ia supernovae. The easiest things you can measure for galaxies are colors and luminosities. He shows a plot of (B-V) versus delta-m-15. There is a feature that blue galaxies produce preferentially slow declining supernovae. This result is similar to what they got before for a Hubble type vs. delta-m-15 plot a number of years ago. There used to be a rule that elliptical galaxies didn't produce slow decliners, for which there is now an exception. But, the distribution for ellpticals is definitely shifted towars higher values of delta-m-15.

Mario says that they think that this plot is telling us that the age of the stellar population is an important parameter in the luminosity of the supernova.

He also plots the luminosities of galaxies (in a large aperture) versus the decline rate. He's combined Calan-tololo, CFA, and nearby supernovae, a total of some 40 supernovae. Slow decliners tend to happen in fainter galaxies- though this is based largely on two very bright cases. But, even take those out and there's a weak trend for spiral galaxies. Ellipticals don't show as much of a trend. Mario says that he first thought that this was a metallicity effect. Today, though, he plotted the color of galaxies versus their luminosity, and saw a correlation that bluer galaxies tend to be dimmer. So, this could be not a metallicity effect, but just the same age effect he mentioned previously. Disentangling the age and metallicity effects will be difficult in this data.

Mark asks if we've seen this sort of thing; I don't think we've looked carefully, and nobody's going to remember anything. Peter and Greg show that there's nothing in our data for color versus stretch. Mark says that he worries for these sorts of plots that the supernovae come from differing search techniques, and may have all sorts of different selection biases and such. For instance, with hostless objects, they might have been ignored as unlikely candidates back in the early days of photometric searching.

Peter says that it almost looked like the high redshift supernovae had a population that was missing in the nearby supernoave (one corner on the color vs. stretch plot).

Mario plotted the residuals on the Hubble diagram as a function of color and metallicity, and didn't see anything to worry about.

One way to disentangle age and metallicity, Mario says, is that there are lines in host galaxies which depend more on either age or metallicity (according to models). One is a Magnesium feature, an absorption feature in elliptical galaxies. He plots that vs H-Beta, which has a grid of models for single-burst stellar systems with different ages and metallicities. He says that this is a promising route to disentangle these things.

Greg says that this spring we have time to get host galaxy spectra from our nearby campaign last spring, with which we might be able to attempt this sort of thing. Greg says that the hope is to get spectra of exactly where the supernova went off; at CTIO, this will be done by dialing in the same offsets. Mario ponits out that these lines may not be as strong at the supernova positions rather than integrated over the whole galaxy. Greg says that the primary goal is to get a template spectrum to subtract from the supernova spectra. For the ellipticals, Greg doesn't think it will be a problem, based on how cloe the supernovae were to the cores of their hosts.

Mark says that in order to do this right, it will take a whole lot of observing. Looking for a weak line like H-beta on a region of the galaxy that's dim, and where you're just trying to get something good enough to subtract from a much-brighter supernova, is going to be a big challenge.

Of course, these models are for single-burst systems. For spiral galaxies, it's harder. Since the idea is that star formation is ongoing, you have no idea if a metallicity observed now is the same as what it was when the supernova progenitor formed.

On this plot, Mario has 8 points, 7 from Calan-Tololo, one from literature (he thinks it's 91bg). The errors in H-Beta are large, because you need good S/N to get a good measurement. It's difficult to get a precise age (as H-beta is mainly sensitive to age). So, they won't be able to draw strong conclusions from this; it's more showing what can be done. (One also has to worry about H-beta being biased by emission.)

The Mg line is stronger, and hence can be measured more precisely. Thus, it's likely that they can get metallicities more preciesly than ages, but of course both are needed.

Units: Mg is equiavlent width expressed in magnituds. H-beta is equivalent width in angstroms. (Mario says that this is the Lick system.)

Greg asks how long he would have to expose on the 4m to get this data. He says it's 10-15m on a 4m. Greg says that this is encouraging, since we're planning on going a lot longer.

Rob Talks about SSH

Everybody should always use SSH. Telnet, FTP, rlogin, and the like are insecure, because they all send your password unencrypted over the network. Anybody on any network link anywhere in between you and where you're going can evesdrop on your session. This means stealing your password. (It also means stealing everything you type. Evesdropping on E-mail is just as easly.)

Rob shows an example where he telnetted from cha-am to rivoli. On Picante, another computer on the same network segment as rivoli, he stole the password from the telnet session. It really is that easy.

Some pointers to SSH:

Unix and Linux

Macs & Windows/DOS