PDA

View Full Version : Are local blazar host galaxies in the green valley too?



Nereid
2010-Jan-24, 04:39 AM
One of several questions triggered by "Galaxy Zoo: The fundamentally different co-evolution of supermassive black holes and their early- and late-type host galaxies (http://arxiv.org/abs/1001.3141)"

"local" = 0.02 < z < 0.05

"blazar" = looking down one of the AGN jets, per the unified model

The paper uses SDSS DR7, and reports analysis of some 942 AGNs (that's a tad under 0.1 local AGNs per square degree, give or take^). They find that green valley galaxies have the highest fraction (proportion) of AGNs, by far (rises to >10%, in some parts of the valley).

Had local blazars been included in their study (they were not), I wonder whether the number of such could have been used to estimate the (average) opening angle of the jets of the (local) AGNs ... and, if the hosts had been classified by zooites (they were not), would they have been able to say something about how AGN jets differ, as a function of host galaxy type?

My first guess would be ... no. Why? Because from ~1,000 AGNs, randomly oriented, you'd expect only ~1^^ to be a blazar (if the opening angle were 5o).

Next.

The selection criteria: "The emission line selection we use in this work does not include unobscured (Type 1) AGN with broad lines" ... does that mean that Type 1 AGN with narrow lines were selected I wonder?

And how does the local (volume) density of unobscured AGN with broad lines compare with ~0.1/square degree (with the same luminosity etc selection criteria)?

Finally, the only SDSS objects selected for possible inclusion in the study were those "with SDSS spectra classified as GALAXY". Of course, if you're going to study host galaxies, you'd better have a galaxy to study! But how many local AGNs are there, in the SDSS DR7 footprint, which do not have "SDSS spectra classified as GALAXY" (with the same luminosity etc selection criteria)?

^ care required: the team's selection criteria are carefully stated, rely upon SDSS to be unbiased and complete (in carefully defined ways), etc.
^^ give or take a factor of 2 or so ...

EDG
2010-Jan-24, 05:27 AM
I just want to say, I have absolutely no clue what the subject line of this thread means :). What is the "green valley"?

Nereid
2010-Jan-24, 02:24 PM
We're all familiar with the HR diagram, right?

It's a colour-magnitude (or colour-luminosity, or spectral class-mass, or ...) diagram, and if you plot the estimates for lots of stars, you find they tend to be in specific parts of the diagram (main sequence, white dwarfs, asymptotic giant branch, etc).

Well, if you do the same thing for galaxies, only you use colour (u-r is popular) and stellar mass (log scale), you find most galaxies lie on the 'red sequence' or the 'blue cloud'; in between is ... (drum roll please) ... the 'green valley'!

In terms of shape ('morphology'), the red sequence is mostly ellipticals and lenticulars, and the blue cloud spirals (though red spirals and blue ellipticals are among the zooites' discoveries).

EDG
2010-Jan-24, 05:06 PM
Ah, I think I follow now. So you tend to find galaxies that are mostly red or blue, but not in the middle?

I guess red ones have lots of red dwarfs and not a lot of young high mass stars (unless they're all red giants, which I guess is possible if they all formed around the same time and had similar mass). And I suppose the bluer ones have lots of active star formation to make the younger, blue stars?

Though given that red dwarfs massively outnumber F and G stars, I don't think we'd be too likely to see a galaxy that was in the middle would we? Even if you somehow removed all the blue stars, the galaxy would still be dominated by the orange and red ones.

Or am I still misunderstanding?

Nereid
2010-Jan-24, 07:01 PM
Ah, I think I follow now. So you tend to find galaxies that are mostly red or blue, but not in the middle?
Sorta.


I guess red ones have lots of red dwarfs and not a lot of young high mass stars (unless they're all red giants, which I guess is possible if they all formed around the same time and had similar mass). And I suppose the bluer ones have lots of active star formation to make the younger, blue stars?

Though given that red dwarfs massively outnumber F and G stars, I don't think we'd be too likely to see a galaxy that was in the middle would we? Even if you somehow removed all the blue stars, the galaxy would still be dominated by the orange and red ones.

Or am I still misunderstanding?
Well, you're mixing up a purely empirical finding (there are regions of the HR diagram with lots of stars, and others with few or none; the colour-mass diagram for galaxies has two regions with lots of galaxies, and the part between them is called 'the green valley') with an explanation of why a region is what it is (e.g. the main sequence is heavily populated because stars spend far more of the lives burning hydrogen in their cores than in any other phase).

The GZ paper (preprint actually) in the OP leans heavily towards the empirical ...

parejkoj
2010-Jan-24, 07:01 PM
Had local blazars been included in their study (they were not), I wonder whether the number of such could have been used to estimate the (average) opening angle of the jets of the (local) AGNs ... and, if the hosts had been classified by zooites (they were not), would they have been able to say something about how AGN jets differ, as a function of host galaxy type?

How many blazars are there below z~0.1? Can't be many. (NED doesn't have a blazar class, which was my first thought for a quick check). And how many of them have host galaxies that are readily distinguishable from the central source? Probably even fewer.

I'm not aware of any SDSS blazars in this volume, but I can check. But your estimates based on opening angle certainly apply here. But you also have to fold in the bolometric luminosity, which for the sources in Schawinski et al. are around 1040 - 1043 erg/s, whereas blazars are closer to 1046 erg/s. There are no known AGN brighter than ~1045 erg/s in the local universe (z ~< 0.1).



The selection criteria: "The emission line selection we use in this work does not include unobscured (Type 1) AGN with broad lines" ... does that mean that Type 1 AGN with narrow lines were selected I wonder?

Are you referring to "narrow line Seyfert-1" sources? Those also have broad lines, just not as broad as a standard broad-line AGN. Just to be sure we're clear on definitions:

Type 1 AGN: have some broad line (~> 1000 km/s FWHM) component to the optical/UV spectrum, due to the region quite close in to the accreting black hole.

Type 2 AGN: no broad lines, but the ratio of narrow lines can only be produced by a luminous source with a hard spectral slope, so there must be an obscured, accreting supermassive black hole, whose emission we can't see directly.

All of the sources in this paper are type 2 AGN. No broad lines at all (or, at least, no obvious ones that the fitting program found).


And how does the local (volume) density of unobscured AGN with broad lines compare with ~0.1/square degree (with the same luminosity etc selection criteria)?

Careful: there are a number of selection criteria going on here, so 0.1/sq. degree isn't exactly the unobscured AGN density in this redshift range.

But as to the answer to your question, that's work in progress... ;)


Finally, the only SDSS objects selected for possible inclusion in the study were those "with SDSS spectra classified as GALAXY". Of course, if you're going to study host galaxies, you'd better have a galaxy to study! But how many local AGNs are there, in the SDSS DR7 footprint, which do not have "SDSS spectra classified as GALAXY" (with the same luminosity etc selection criteria)?

Somewhere in the realm of 400 to 2000. Here's one example (http://cas.sdss.org/dr7/en/tools/explore/obj.asp?id=587731869627777065). Getting the selection functions to match is very hard, though. And some SDSS spectra in this volume are classified as GALAXY, but have broad lines and/or a power-law component that the automatic classification pipeline didn't flag. Remember, there are going to be a few thousand 3-sigma outliers when you've got a million measurements! ;-)

EDG_:

The Wikipedia page on the galaxy color magnitude diagram (http://en.wikipedia.org/wiki/Galaxy_color-magnitude_diagram) might help you. As to the "AGN live in the green valley" assertion, I've attached a figure from Schawinski et al. 2009 (http://adsabs.harvard.edu/abs/2009ApJ...692L..19S) that might help you out. The data plotted there are from roughly the same volume as Schawinski et al. 2010 that Nereid linked in the OP.

Note, that there is some debate as to whether AGN hosts are in the green valley because of dust reddening of otherwise blue galaxies. This is ongoing, fairly cutting edge stuff, and the answer isn't clear yet (in fact, exactly this sort of research makes up the bulk of my postdoctoral research proposal!).

Nereid
2010-Jan-24, 09:08 PM
Thanks parejkoj! :)

I kinda guessed that this general area was one you might be working on, but to learn that you're deep in the green valley was a pleasant surprise.

OK, perhaps "blazar" isn't quite right; more basic questions then:

* how does the incidence of detectable jets (however defined) vary with estimated AGN luminosity?

* specifically, is there some low luminosity cutoff?

* what can be said - if anything - about how the expected 'looking down a jet' observables varies with luminosity (based on studies of jets seen 'from the side')?

"Luminosity" is something I seem to be tripping over: a paper the one in the OP references (http://cdsads.u-strasbg.fr/abs/2009ApJ...692L..19S) finds 21 local (0.01 < z < 0.07) AGN in the SDSS footprint, which have 42.7 < log (L14-195 keV) < 44.5 (erg/s), which I guess is pretty close to bolometric luminosity. However, the luminosities the paper in the OP refers to are (mostly) L[OIII], with the conversion to bolometric covered thus: "Kauffmann & Heckman (2009) argue that the bolometric correction for the extinction corrected [O iii] 5007 luminosity is a factor of 300 – 600, so the bolometric luminosities of the AGN in our sample are between 2.5 and 2.8 dex higher than the [O iii] 5007 luminosities quoted throughout."^

Are you referring to "narrow line Seyfert-1" sources? Those also have broad lines, just not as broad as a standard broad-line AGN. Just to be sure we're clear on definitions:

Type 1 AGN: have some broad line (~> 1000 km/s FWHM) component to the optical/UV spectrum, due to the region quite close in to the accreting black hole.

Type 2 AGN: no broad lines, but the ratio of narrow lines can only be produced by a luminous source with a hard spectral slope, so there must be an obscured, accreting supermassive black hole, whose emission we can't see directly.

All of the sources in this paper are type 2 AGN. No broad lines at all (or, at least, no obvious ones that the fitting program found).
Here's the full quote (from the preprint):

The emission line selection we use in this work does not include unobscured (Type 1) AGN with broad lines and may be biased against the most highly obscured AGN. If the Unified Model of AGN (Antonucci 1993; Urry & Padovani 1995) is correct and these difference are due to the orientation of the putative obscuring medium in the central engine, and if the obscuration does not cover the narrow-line region, then narrow-line (Type 2) AGN are a fair sub-set of the entire AGN population within the luminosity range covered.
One thing I'm trying to understand is what the incidence of local AGN (in the SDSS footprint) is, that Schawinski et al. did NOT select (with estimated luminosities > 1040 erg/s) ...

There are follow-ons from this, e.g. how many broad-line galaxies there are (among the ~48k sample they used), and how many 'blue-green peas' might there be (the green peas the zooites discovered have redshifts just outside the upper bound - 0.112 vs 0.07 - but it's not clear that slightly less luminous, lower z, objects would turn up in the ~48k GALAXY sample, even though some green peas are AGNs, and the luminosities would likely be > 1040 erg/s).

Now "Somewhere in the realm of 400 to 2000" goes a long way to answering my question (well, one of them)! :)

And that points to possible systematics that Schawinski et al. may not have adequately addressed ...

^ caveat lector: copy/paste loses some important stuff!

parejkoj
2010-Jan-24, 09:53 PM
Oook...

Nereid, I'm not sure I have time to handle most of your questions. I don't work on Blazars directly, so I don't have those numbers immediately at hand. I'll see what I can do, though. Keep in mind that I suspect good answers to some of your questions would end up being decently publishable results!



"Luminosity" is something I seem to be tripping over: a paper the one in the OP references (http://cdsads.u-strasbg.fr/abs/2009ApJ...692L..19S) finds 21 local (0.01 < z < 0.07) AGN in the SDSS footprint, which have 42.7 < log (L14-195 keV) < 44.5 (erg/s), which I guess is pretty close to bolometric luminosity. However, the luminosities the paper in the OP refers to are (mostly) L[OIII], with the conversion to bolometric covered thus: "Kauffmann & Heckman (2009) argue that the bolometric correction for the extinction corrected [O iii] 5007 luminosity is a factor of 300 – 600, so the bolometric luminosities of the AGN in our sample are between 2.5 and 2.8 dex higher than the [O iii] 5007 luminosities quoted throughout."^

What's confusing you here? The Swift BAT sample in the 2009 paper certainly has higher luminosities than the galaxy zoo sample, simply by the fact that they were detected by the BAT! On the other hand, [OIII] is typically used as a proxy for the actual AGN luminosity (which we can't see because of the obscuration) when X-ray data isn't available, but there's quite a bit of scatter, and it can be contaminated by ionization from star formation and other galaxy-scale processes. So converting from an [OIII] luminosity to a bolometric luminosity is a pretty rough estimate of AGN output.



One thing I'm trying to understand is what the incidence of local AGN (in the SDSS footprint) is, that Schawinski et al. did NOT select (with estimated luminosities > 1040 erg/s) ...

Heh... ask me after Aprilish (and/or later this summer), when I've got more time and have started to attack this problem. Expanding this sample is one of my eventual goals. GZ 2 should help with this, because of expanded inclusion criteria.



There are follow-ons from this, e.g. how many broad-line galaxies there are (among the ~48k sample they used), and how many 'blue-green peas' might there be (the green peas the zooites discovered have redshifts just outside the upper bound - 0.112 vs 0.07 - but it's not clear that slightly less luminous, lower z, objects would turn up in the ~48k GALAXY sample, even though some green peas are AGNs, and the luminosities would likely be > 1040 erg/s).

There should be very few to no galaxies with broad lines in the sample used in Schawinski et al. 2010. There are galaxies with broad lines in the redshift range, but they were specifically excluded, because their spectra are hard to fit, and there are a number of other systematics related to SDSS target selection that come into play.

The green peas would look blue at low redshift. Remember, most of their observed light is coming from very strong [OIII], which at their redshifts made them look greenish in the SDSS gri filters. Selection function!

They appear to be most closely related to luminous blue compact dwarf galaxies, which contain few, if any, AGN.



And that points to possible systematics that Schawinski et al. may not have adequately addressed ...


Oh, believe me: Kevin knows about this. But the conclusions in the paper aren't about "all" AGN, but rather about "these AGN over here that we've defined thusly, and hope are generally representative of all AGN." And the latter seems to be true, base on some other evidence; it appears that type 1 and type 2 objects are in similar galaxies, as would be expected if the unification theory is correct, but there's a lot of testing left to be done before that is completely clear. I'll be able to be more specific about all this in a few months, hopefully. *gonna graduate... gonna graduate...*

parejkoj
2010-Jan-24, 10:54 PM
After a bit of reflect, I can point you toward some results on radio emission and AGN.

Try looking at the various SDSS/FIRST papers by Phil Best et al., such as this one (http://adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2005MNRAS.362...25B&link_type=ABSTRACT). I don't think they'll answer your questions directly, but they'd be a place to start.

In general, radio loud AGN lie in large, red elliptical galaxies. I'm not sure about a galaxy morphology : radio morphology relation, however.

Nereid
2010-Jan-25, 09:01 PM
Thanks parejkoj.
Oook...

Nereid, I'm not sure I have time to handle most of your questions. I don't work on Blazars directly, so I don't have those numbers immediately at hand. I'll see what I can do, though.
I see you posted subsequently, and I did some literature searching of my own, so I'll comment in another post.


Keep in mind that I suspect good answers to some of your questions would end up being decently publishable results!

That's good to know ... I felt this was somewhat unchartered territory, but wanted to be sure I hadn't overlooked something ...



"Luminosity" is something I seem to be tripping over: a paper the one in the OP references finds 21 local (0.01 < z < 0.07) AGN in the SDSS footprint, which have 42.7 < log (L14-195 keV) < 44.5 (erg/s), which I guess is pretty close to bolometric luminosity. However, the luminosities the paper in the OP refers to are (mostly) L[OIII], with the conversion to bolometric covered thus: "Kauffmann & Heckman (2009) argue that the bolometric correction for the extinction corrected [O iii] 5007 luminosity is a factor of 300 – 600, so the bolometric luminosities of the AGN in our sample are between 2.5 and 2.8 dex higher than the [O iii] 5007 luminosities quoted throughout."^
What's confusing you here? The Swift BAT sample in the 2009 paper certainly has higher luminosities than the galaxy zoo sample, simply by the fact that they were detected by the BAT! On the other hand, [OIII] is typically used as a proxy for the actual AGN luminosity (which we can't see because of the obscuration) when X-ray data isn't available, but there's quite a bit of scatter, and it can be contaminated by ionization from star formation and other galaxy-scale processes. So converting from an [OIII] luminosity to a bolometric luminosity is a pretty rough estimate of AGN output.

To ~0.1 to 0.3 dex, bolometric luminosity^ is just the integrated SED, assuming isotropy and a distance, right?

Which is fine, except that the jets tell you you don't have isotropy (more in a later post).

In most cases, if you've got some high S/N x-ray data, you're not going to be too wrong in estimating bolometric luminosity (outside the x-ray region, how much can the SED vary, right?).

Time for me to read Kauffmann & Heckman (2009) ...



One thing I'm trying to understand is what the incidence of local AGN (in the SDSS footprint) is, that Schawinski et al. did NOT select (with estimated luminosities > 1040 erg/s) ...
Heh... ask me after Aprilish (and/or later this summer), when I've got more time and have started to attack this problem. Expanding this sample is one of my eventual goals. GZ 2 should help with this, because of expanded inclusion criteria.

Cool! :)



There are follow-ons from this, e.g. how many broad-line galaxies there are (among the ~48k sample they used), and how many 'blue-green peas' might there be (the green peas the zooites discovered have redshifts just outside the upper bound - 0.112 vs 0.07 - but it's not clear that slightly less luminous, lower z, objects would turn up in the ~48k GALAXY sample, even though some green peas are AGNs, and the luminosities would likely be > 1040 erg/s).
There should be very few to no galaxies with broad lines in the sample used in Schawinski et al. 2010. There are galaxies with broad lines in the redshift range, but they were specifically excluded, because their spectra are hard to fit, and there are a number of other systematics related to SDSS target selection that come into play.

The green peas would look blue at low redshift. Remember, most of their observed light is coming from very strong [OIII], which at their redshifts made them look greenish in the SDSS gri filters. Selection function!

They appear to be most closely related to luminous blue compact dwarf galaxies, which contain few, if any, AGN.

What I am trying to get a handle on is limits on AGNs in 0.02 < z < 0.05, in the SDSS footprint, that were not studied (Schawinski et al. 2010 looks at ~1,000), AND which have bolometric luminosities >~10<sup>42</sup> erg/s.

Type 1 AGN were not included.

What the green peas point to is a class of object that would also be not included (narrow line yes, but maybe not a GALAXY, r < 17 AB mag, Mz,Petro < −19.5 AB mag). Sure, most green peas are not AGN (at least, not the ones reported in the GZ paper) ... but some are, and I can't tell if a local population of 'weak AGN peas' (whether narrow line Seyfert I's or not), with bolometric luminosities of ~10<sup>42</sup> erg/s or more, would have been completely missed (assuming such things actually exist).

And this is just a couple of possibles ...



And that points to possible systematics that Schawinski et al. may not have adequately addressed ...
Oh, believe me: Kevin knows about this. But the conclusions in the paper aren't about "all" AGN, but rather about "these AGN over here that we've defined thusly, and hope are generally representative of all AGN."
Yep.


And the latter seems to be true, base on some other evidence; it appears that type 1 and type 2 objects are in similar galaxies, as would be expected if the unification theory is correct, but there's a lot of testing left to be done before that is completely clear.
And this is what will be interesting to find out! :)

... but 400 < 942 < 2000

so it would seem there's lots of room for some most interesting discoveries!

BTW, one nice morsel in Schawinski et al. 2010 is this: "While it is certainly the case that some X-ray detected AGN exhibit nuclear LINER spectra (e.g.,
Gonz´alez-Mart´ın et al. 2009), the presence of LINER-like line ratios over the large physical footprint of the SDSS spectroscopic fiber does not imply the presence of a low luminosity AGN. LINERs are thus potentially highly diverse and the majority of them are not low-luminosity AGN." (emphasis in original)


I'll be able to be more specific about all this in a few months, hopefully. *gonna graduate... gonna graduate...*
Got my thumbs pressed!

^ at least in the local universe; at significant z, you have to make assumptions about the LCDM model parameters

Nereid
2010-Jan-25, 09:14 PM
After a bit of reflect, I can point you toward some results on radio emission and AGN.

Try looking at the various SDSS/FIRST papers by Phil Best et al., such as this one (http://adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2005MNRAS.362...25B&link_type=ABSTRACT). I don't think they'll answer your questions directly, but they'd be a place to start.

In general, radio loud AGN lie in large, red elliptical galaxies. I'm not sure about a galaxy morphology : radio morphology relation, however.
What's the best, closest example?

Perhaps M87.

It would not be included in Schawinski et al. 2010 - while it's in the SDSS footprint (I think), its z is too low.

If we were looking straight down the jet, would M87 be a blazar? I think so (but haven't yet read enough in the literature to be sure).

Now here comes the key question: what's the M87 AGN's bolometric luminosity? I was (quite) surprised to learn that it's likely only ~1041 erg/s (see here (http://www.iop.org/EJ/abstract/0004-637X/582/1/133), for example)!

So, given the unified AGN model, why do the AGN in Schawinski et al. 2010 not have rather obvious jets (at least the ones whose log(L[OIII]) > 40.5, say)? Are jets somewhat optional?

ngc3314
2010-Jan-25, 11:08 PM
What's the best, closest example?
So, given the unified AGN model, why do the AGN in Schawinski et al. 2010 not have rather obvious jets (at least the ones whose log(L[OIII]) > 40.5, say)? Are jets somewhat optional?

Pretty clearly, at least the large-scale powerful ones. Lots of Seyfert nuclei have at most very tiny weak jets (hundreds of parsecs) and some have no discernible resolved jets. Historically, disk galaxies are very seldom fond to host powerful jets (exceptions: 0313,192, maybe 3C 120), with borad ideas floating around that it has to do in part with the denser and colder ISM disrupting the jets' collimation so their energy gets dumped into a large volume. (I don't think this can be the whole story because the hosts still don't have enough radio emission for that). Other ideas have had to do with different accretion modes in which most of the energy output is either radiative or kinetic; this remains pretty speculative as far as I can tell.

The unified jet model is not the only unified picture, being independent of the Sy 12 unification idea driven by circumnuclear obscuration. When we see both ionization cones and radio jets, the radio jet always projects within the ionization cone, so there is a connection between obscuration and ejection axes when both are present.