Pink Black Holes: Technical Details

[Mt Stromlo Observatory] [Department of Physics] [Astrophysical Theory Centre]

This page is intended for those with a technical background who want more detail than was available in the press releases.


These "pink black holes" are quasars. They were selected in a radio survey: the Parkes Half Jansky survey of the southern sky (Drinkwater et al. 1997, MNRAS 284, 85). This survey consists of 323 sources selected to have flat radio spectra. The optically bright ones have long been studied: our innovation was to energetically pursue the optically faint ones. We obtained accurate radio positions with the VLA and ATCA, and identified all sources with deepish CCD and near-IR array camera imaging. We have spectra for about 80% of the sample, including all but the very faintest (B>23) 5% of the sources.


In 1994-5, we discovered that many of the optically faint Parkes sources were extremely luminous in the near-IR (K-band: 2.1 microns). Radio quiet quasars have essentially uniform B-K colours (around 2.5) while these Parkes sources had B-K ranging from 2.5 out to 8! Their colours were thus clearly much redder than normal quasars. (Webster et al. 1995, Nature 375, 469).

Why the difference? Several rival theories were put forward:

  • Dust: these actually are typical blue quasars obscured by dust somewhere along the line of sight.
  • Synchrotron emission: we know that these compact flat-radio-spectrum sources have relativistic jets. Perhaps the synchrotron emission from these jets extends into the near-IR?
  • Host galaxy contamination: perhaps they are not quasars but high redshift radio galaxies, reddened by the redshifted 4000A break.
Our spectra demonstrated that these sources were not galaxies (apart from a few of the nearest ones, with radio spectra near the steep cut-off) (Masci et al. 1998, MNRAS 301, 975). They have broad emission lines with roughly normal line ratios and equivalent widths. We obtained quasi-simultaneous BVRIJHK imaging of about half the sample, which showed that most had spectra well fit by pure power-laws, with indexes (F(nu) = nu^alpha) 0 > alpha > -2. A small sub-set were much redder, and had spectra turning down in the blue and reddened line ratios: these few are well fit by dust obscuration of blue quasars.

The puzzle is: what are these red power-law sources? They appear pink to the human eye (I combined out BVR imaging weighted such that a sun-like star would appear white). They show broad emission-lines such as MgII and CIII] with roughly normal equivalent widths, so these are not classical BL Lac objects (H-beta, on the other hand, has an equivalent width that anticorrelates nicely with redness). We've measured the near-IR polarisation for a few of them and they are quite strongly polarised out beyond a micron. Many are reasonable X-ray sources as detected by ROSAT (Seibert et al. 1998, MNRAS 301, 261).

The problem is: how do you get synchrotron emission with a slope of -2? This is close to the reddest slope you can get for any plausible electron energy distribution. We can fit it with a cut-off in the electron energy distribution, but this cut-off has to be at 1 micron +- 30% in well over half of the quasars. This would argue for some fundamental physical process locking the cut-off here: we cannot think of any (naievely the cut-off should go as the square of the magnetic field). Even then, the predicted spectrum shouldn't be a pure power-law, due to the Big Blue Bump component coming in as shorter wavelengths (we need something blue to photoionise the BLR). So: a puzzle! Suggestions welcome!

We have a couple of papers almost ready to submit about the data and our modelling of it: watch this space!

Last updated 12th May 1999.
Maintainer: Paul Francis,