A panchromatic review of thermal and non-thermal AGNs

1
A panchromatic review of thermal and non-thermal
AGNs

• Robert R. J. Antonucci
• Department of Physics
• University of California, Santa Barbara
• Santa Barbara, CA 93106-9530
• ski_at_physics.ucsb.edu

2
SUMMARY

• Analyze the literature advocating the existence
  of "True Seyfert 2s," that is, Seyfert 2s that
  have no visible or hidden broad lines,
• Carefully explain nine errors frequently made in
  polarimetry papers along with six
  non-polarimetric arguments.
• I will argue that I don't find any of the
  published cases convincing (though there are
  probably some papers on this topic with which I'm
  not familiar.)
• For more details see review A panchromatic
  review of thermal and non-thermal AGNs on
  astro-ph.

3
OUTLINE

• Very Common Errors and Omissions in Papers
  Advocating True Seyfert 2s.
• I. Polarimetric evidence 9 comments II. Other
  evidence 4 comments
• Note I will NOT discuss individual objects or
  papers, not even in the question period (though
  you can speak up), because I don't remember the
  facts for each one, and I have to think slowly
  and critically to respond to assertions by
  others!

4
Disclaimers

• Some of what I will say is probably wrong!
• I'm critiquing. Unlike Fox News, I do not
  strive to be fair and balanced.
• Not all of my arguments apply to all objects or
  all papers!!
• Some of the points argue the plausibility of
  identification as common NLXGs, and some of big
  blue bumps/BLRs hidden by a nuclear torus. I'm
  not trying to present a single scenario for all.
• I don't really give a damn about whether or not
  there are True 2s!
• NOTE A True Seyfert 2 is defined here to be an
  AGN with broad line EW (relative to the AGN
  continuum) much less that that of normal
  Seyfert 1s.

5
Polarimetric Arguments

• 1) Hidden BLRs revealed laterHidden broad lines
  are often detected at later epochs, or with
  better SNR. Even in their low states, I think the
  points below greatly weaken their
  candidacy.Note One should NEVER say certain
  objects lack a hidden BLR, only that none has
  been detected so far. There is no evidence of
  bimodality some with and some without hidden
  BLRs!

6

• 2) Lack of upper limitsFew if any upper limits
  have been published for broad lines of
  candidates, in either flux or polarized
  flux!!!If someone gives does give EW upper
  limits, they must be relative to the AGN
  continuum, which is hardly detected in many cases.

7

• 3) Polarization is NOT due to nuclear scattering
  in most well-studied Seyferts!!
• Scrupulous removal of interstellar polarization
  in the Milky Way is necessary.
• Polarization is very often due to dust
  transmission in the host.
• Polarization may be due to large-scale dust
  scattering in the host.

8

• 4) Upper limits depend on assumed line
  widthsFor example, Stockton et al. claimed to
  put tight limits on BLs in Cygnus A in total
  flux, but they turned out to be wrong because the
  authors assumed much too narrow a line width.

9

• 5) "Secure non detections" of broad lines
• Several objects have non-detections of BLs in
  polarized light that are called "secure," though
  no upper limits are given.
• I'm not even sure that a non-detection can ever
  be secure.

10

• 6) Spectropolarimetry at Hb is inadequate
• Hb is often too weak to be detected, sometimes
  due to reddening. Early example 3C 234.
  Undetected polarized Hb means almost nothing to
  me.

11

• 7) Near-IR polarimetry often required
• Often kpc-scale obscuration of the reflecting
  region blocks even Ha in polarized light... e.g.,
  Cen A and Cyg A, for which the optical percentage
  polarization is ? 1 yet the near-IR percentage
  polarization is ? 20!!You have to check for
  this.

12

• 8) Dilution of polarization by a starlightOld
  stars must be carefully removed. But alsoThe
  nuclear regions of most well-studied Seyfert 2s
  are dominated by starburst light!Only the broad
  lines themselves tell the scattering
  polarization.The percentage polarization is the
  polarized flux of the line divided by their total
  flux, yet generally the broad lines aren't
  detected in total flux in Seyfert 2s, by
  definition!True test does the polarized flux
  look like a Seyfert 1?

13

• 9) It takes luck to find hidden broad line
  regions!From Tran et al. (2011) Could the
  True 2s be the hidden counterparts to the
  NLS1...? This can be ruled out by the simple
  observation that no emission lines of any kind,
  broad or narrow, are seen in the polarized flux
  spectra.But you need to be lucky to have a
  suitable mirror!

14
Non-polarimetric evidence for True Seyfert 2s.

• 1) Independent analysis by Stern Laor
  (2012)Stern and Laor (2012) have shown proxy
  equivalent width limits by normalizing the BLs to
  the Hard X-ray...What fraction of Seyfert 1s in
  their sample have a value for this that's less
  than their upper limit then NGC 3147?Is it very
  small? No, it's 50!!!And even these limits
  apply only to lines matching their velocity
  prescription, so they are really lower limits to
  upper limits!

15

• 2) Near-IR emission linesCandidates are almost
  always overwhelmedby dust emission and
  starlight
• EW limits are only intelligible if given
  relativeto the AGN continuum, which is usually
  undetected, especially in the near-IR.

16

• 3) The mid-IR calorimeterAn excellent test for
  a hidden AGN is the mid-IR warm calorimeter,
  which indicates the luminosity of any visible or
  hidden AGN approximately but (in my opinion)
  robustly.But in the (heterogeneous) sample of
  Haas et al. (2007) the mid-IR bumps are just as
  strong relative to O III l5007 in the putative
  True 2s as in their other objects!!!

17

• 4) Are the True 2s even Seyfert 2s at all?
• Some candidates have optical spectra inadequate
  for classification. Though some liners do have
  visible or hidden BLRs, liners with no BLRs
  aren't surprising.
• Classification is crucially dependent on precise
  starlight subtraction. The best- observed NGC
  3147 spectrum (Ho et al. 1985) shows no Hb
  emission line!!
• It can only be seen with precise starlight
  subtraction.
• Thus O III l5007 / Hb (and N II l6584 / Ha)
  is biased upward because the Balmer emission
  lines sit in stellar absorption lines!
• This is a HUGE effect at low luminosities such as
  those of True 2s.

18

• Additional Point Isotropic selection is key for
  polarization and other surveysNGC1068 and
  NGC4151 have the same UV excess, but we're only
  seeing 1 of the nuclear UV in the former case,
  so that AGN comes from five magnitudes higher on
  the luminosity function!!Hidden BLR objects
  tend to have systematically larger optical flux
  and luminosity, and much better contrast against
  the host galaxy starlight. That's why their
  hidden BLRs were detected in the first
  place.Keel et al. (1994) is a fundamental paper
  because it was selected independently of
  orientation. Thus the Type 1s and the Type 2s
  can be compared legitimately.
• Here is a related and shocking widespread
  practice Conclusions are routinely drawn from
  luminosity-luminosity plots
• Very influential example that set this field back
  many years
• Rawlings Saunders 1991 Nature paper.
• Plot numbers of bars versus number of bookstores
  in all cities and towns do drinkers like to
  read or does reading drive you to drink?

19
Conclusion

• The conclusion is obvious.