Parsec-Scale Localization of SDSS J1536+0441A

Parsec-Scale Localization of the Quasar SDSS J1536+0441A, a Candidate Binary Black Hole System


The radio-quiet quasar SDSS J1536+0441A shows two broad-line emission systems, recently interpreted as a binary black hole (BBH) system with a subparsec separation; as a double-peaked emitter; or as both types of systems. The NRAO VLBA was used to search for 8.4 GHz emission from SDSS J1536+0441A, focusing on the optical localization region for the broad-line emission, of area 5400 mas (0.15 kpc). One source was detected, with a diameter of less than 1.63 mas (8.5 pc) and a brightness temperature K. New NRAO VLA photometry at 22.5 GHz, and earlier photometry at 8.5 GHz, gives a rising spectral slope of . The slope implies an optically thick synchrotron source, with a radius of about 0.04 pc, and thus K. The implied radio-sphere at rest frame 31.2 GHz has a radius of 800 gravitational radii, just below the size of the broad line region in this object. Observations at higher frequencies can probe whether or not the radio-sphere is as compact as expected from the coronal framework for the radio emission of radio-quiet quasars.

Subject headings:
black hole physics — quasars: individual (SDSS J153636.22+044127.0) — radio continuum: general

1. Motivation

Binary black hole (BBH) systems with subparsec scales are predicted in merging scenarios for galaxy evolution and also factor prominently in predictions for the gravitational wave background (e.g., Colpi & Dotti, 2009). But can such BBH systems be found? A promising search method - identifying candidate BBH systems through their optical broad-line emission properties - recently yielded three such candidates: SDSS J153636.22+044127.0 (Boroson & Lauer, 2009), SDSS J092712.65+294344.0 (Bogdanovic et al., 2009; Dotti et al., 2009), and SDSS J105041.35+345631.3 (Shields et al., 2009).

In this Letter we focus on SDSS J153636.22+044127.0 (SDSS J1536+0441 hereafter), a quasar at a redshift (Boroson & Lauer, 2009). For the assumed flat cosmology4, the inferred 0.1 pc separation subtends 0.02 mas (Boroson & Lauer, 2009). This quasar has also been interpreted as a lone double-peaked emitter (DPE; Gaskell, 2010; Chornock et al., 2010) or as twin DPEs with a subparsec separation (Tang & Grindlay, 2009). The emission lines from a DPE are thought to arise from rotational motion in a relativistic accretion disk. SDSS J1536+0441 is radio quiet and our imaging of it at 8.5 GHz (Wrobel & Laor, 2009) revealed two faint sources, SDSS J1536+0441A and J1536+0441B, separated by 0.97″  (5.1 kpc) with each source being unresolved with a diameter of less than 0.37″  (1.9 kpc). It is now clear that SDSS J1536+0441A is hosted by the radio-quiet quasar (RQQ), while SDSS J1536+0441B is hosted by a companion radio-loud elliptical galaxy (Decarli et al., 2009; Lauer & Boroson, 2009).

In the BBH scenarios for SDSS J1536+0441A (Boroson & Lauer, 2009; Tang & Grindlay, 2009), each of the two broad-line emission systems is itself a quasar. Recent optical spectroscopy finds that the broad-line emission systems have relative positions that localize them to 3 bands of full width 90 mas (470 pc) along a position angle (PA) of 48°  (Chornock et al., 2010) and 60 mas (310 pc) along PA 90°  (T. Boroson, 2010, private communication). The resulting parallelogram localizes SDSS J1536+0441A to an area of 5400 mas (0.15 kpc). Improvements in the localization of the emission from SDSS J1536+0441A would further test its candidacy as a BBH system. In this Letter we use measurement techniques at radio frequencies to investigate the compactness of SDSS J1536+0441A, first by seeking evidence for a synchrotron self-absorbed spectrum and then by direct imaging with mas resolution to localize the emission on parsec scales. Our new imaging is presented in § 2 and its implications are explored in § 3. A summary appears in § 4.

2. New Imaging

The CnB configuration of the NRAO VLA5 (Thompson et al., 1980) was used under proposal code AL739 to observe SDSS J1536+0441A and J1536+0441B for 2 hours near transit on UT 2009 May 27. We followed the strategies described by Wrobel & Laor (2009), except that for our new observations the center frequency was 22.5 GHz, the switching time was 120 s, and the amplitude scale was set to an accuracy of about 5%. The elliptical Gaussian resolution, 0.96″  times 0.81″  at PA -89°, was sufficient to obtain photometry for each source. We will report elsewhere on SDSS J1536+0441B. For SDSS J1536+0441A, the flux density was  mJy and comparison with the 8.5 GHz value (Wrobel & Laor, 2009) implies a rising spectrum with index (). The 2009 December 31 release of the NRAO AIPS software was used for calibration and imaging.

The NRAO VLBA (Napier et al., 1994) was used for 5 hours under proposal code BL168 on 2009 October 14 UT to search for 8.4 GHz emission from SDSS J1536+0441A. Phase-referenced observations were made in the nodding style, using 32 MHz per circular polarization. A 120 s scan of SDSS J1536+0441A was preceded and followed by a 60 s scan of the reference source J1539+0430, favorably located at a switching angle of 0.7°  from SDSS J1536+0441A. The 2009 December 31 release of the NRAO AIPS software was used for calibration and imaging. We followed the calibration strategies described by Wrobel & Ho (2006) and the large-area imaging strategies described by Wrobel et al. (2005).

We conducted a large-area VLBA search in a region set by the 3 VLA astrometric accuracy (Wrobel & Laor, 2009) and of area 280,000 mas. Using natural weighting of the visibility data, this search achieved an elliptical Gaussian resolution of 2.57 mas times 1.04 mas at PA -5.2°, a geometric-mean resolution of 1.63 mas, and a root-mean-square sensitivity of 0.056 mJy beam in Stokes . Given the large number of resolution elements searched, the conservative search strategies developed by Wrobel et al. (2005) were followed. Specifically, a 6 detection threshold was adopted and only one source emerged above this threshold. This VLBA detection, shown in Figure 1, is unresolved with a diameter of less than 1.63 mas and has a flux density of mJy. The ratio of the flux density of the VLBA detection to that measured with the VLA (Wrobel & Laor, 2009) is 0.750.11. The position of the VLBA detection is and , with a conservative 1 error of 2 mas per coordinate estimated from analysis of the check source, J1544+0407, observed at a switching angle of 1.5°  from the reference source.

We also conducted a slightly deeper (5 ) small-area search for an additional point-like source near the VLBA detection and within the 5400 mas localization region described in § 1 for the broad-line emission systems. Figure 2 shows that if an additional source is present within that parallelogram-shaped region, it is fainter than the VLBA detection by a factor of 2.6 or more. This upper limit was corrected for a 20% coherence loss, estimated from the ratio, about 1.2, of the VLBA detection’s integrated-to-peak flux densities. This ratio is applicable to a switching angle of 0.7°  and is also consistent with the higher ratio, about 1.6, measured for the check source which had a switching angle of 1.5°.

3. Implications

3.1. Size of the Radio-Sphere

From the new VLA photometry, the RQQ SDSS J1536+0441A has a rising spectrum with index at rest frequencies of tens of gigahertz. (This is consistent with our prior suggestion that the overall spectrum of A and B was flat or rising between 1.4 GHz [White et al. 1997] and 8.5 GHz [Wrobel & Laor 2009].) SDSS J1536+0441A thus resembles the 45%-50% of RQQ that show flat or rising integrated spectra at similar frequencies (Barvainis et al., 1996; Ulvestad et al., 2005). This spectrum suggests that SDSS J1536+0441A is compact enough to be synchrotron self-absorbed, as expected in the coronal framework for RQQ (Laor & Behar, 2008). SDSS J1536+0441A has a bolometric luminosity of  ergs s (Boroson & Lauer, 2009) and a 22.5 GHz luminosity density of  ergs s Hz. Applying equation (22) of Laor & Behar (2008), for a homogeneous synchrotron source with equipartition between magnetic and photon energy densities, implies a radio-sphere of radius about 0.04 pc at a rest frequency of 31.2 GHz, and thus K. This is just below the Readhead (1994) limit of K, expected for equipartition between the electron energy density and the magnetic energy density within the radio-sphere. The 0.04 pc radius corresponds to about 800 gravitational radii for a   BH, a mass thought to be applicable to SDSS J1536+0441A (Lauer & Boroson, 2009; Tang & Grindlay, 2009). The radius of the radio-sphere implies that the 31.2 GHz emission arises just within the optical broad-line emission region (BLR) discussed in § 3.2.

The spectral index of this RQQ is clearly too shallow to be a homogeneous source, and likely implies a superposition of emission from an inhomogeneous source, as commonly adopted for flat spectrum radio-loud systems (e.g., Phinney, 1985). Given the newly measured flat spectral index for SDSS J1536+0441A, its ratio of radio to X-ray (Arzoumananian et al., 2009) luminosities drops from (Wrobel & Laor, 2009) to , putting it close to the average ratio characterizing lower-luminosity active galactic nuclei, a ratio expected in the coronal framework for the radio emission from RQQ (Laor & Behar, 2008).

Some RQQs are time variable (Barvainis et al., 2005) so our inference of a rising spectrum for SDSS J1536+0441A is weakened by using non simultaneous photometry. However, for the source to have an optically thin spectral index steeper than , the flux density at either 8.5 or 22.5 GHz would need to vary by a factor greater than two between the 100 days separating the measurements. Causality arguments would then imply a size upper limit of light days, or 0.06 pc. In this case, both the steep spectrum and the high variability brightness temperature, K, would exclude a thermal free-free origin for the radio emission.

From the new VLBA imaging (Figs. 1 and 2), a single source was detected at 8.4 GHz within the 470 pc by 310 pc localization region for the broad-line emission (Chornock et al. 2010; T. Boroson, 2010, private communication). The VLBA detection has a geometric-mean diameter of less than 8.5 pc and a rest-frame brightness temperature, modified for an elliptical Gaussian, of K. The ratio of the VLA and VLBA flux densities near 8 GHz is broadly consistent with unity. This suggests that the VLBA recovers all of the VLA signal but time variability remains a concern. For now, we tentatively assign the VLA-derived spectral index, , to the VLBA detection. Then the isotropic power at a rest frequency of 8.4 GHz is  W Hz. This VLBA detection has a power and limit at the low end of the values reported for other RQQs detected with the VLBA (Blundell et al., 1996; Blundell & Beasley, 1998; Ulvestad et al., 2005). The crude estimate made above for a synchrotron-self-absorbed size is also consistent with the VLBA detection.

The implications of the above findings are examined below, first within the context of BBH scenarios for SDSS J1536+0441A (Boroson & Lauer, 2009; Tang & Grindlay, 2009) and then within the context of it being a lone DPE (Gaskell, 2010; Chornock et al., 2010).

3.2. Binary Black Hole Scenarios

Our VLBA findings are consistent with the projected separation of the two quasars being less than 8.5 pc. The VLBA localization area for SDSS J1536+0441A is 82 pc, improving over the emission-line localization area by a factor of about 1800. The VLBA detection appears to have a rising, synchrotron self-absorbed spectrum, which bodes well for imaging it at higher resolutions to improve the localization further. Optical spectroscopic monitoring of SDSS J1536+0441A implies an orbital period longer than about 200 years (Lauer & Boroson, 2009). Unfortunately, such a long period means that VLBA monitoring could not usefully constrain the astrometric wobble of SDSS J1536+0441A.

The twin DPE model of (Tang & Grindlay, 2009) reproduces the observed H line profile by emission from a disk extending from 7000 down to 800 gravitational radii. Thus, the size of the radio-sphere at 31.2 GHz is just below the inner boundary of the BLR. Since the size of an optically thick radio-sphere scales as (e.g. equation 22 of Laor & Behar 2008), observations at mm wavelengths will allow us to probe the radio-sphere on smaller scales, potentially down to the optically emitting region at a few 10s of gravitational radii. If the source remains optically thick, then sub-mm observations can probe down to the X-ray emitting region at a few gravitational radii.

The VLBA detection of the RQQ SDSS J1536+0441A represent the first parsec-scale localization of a candidate BBH system identified through its broad-line properties. This VLBA detection also demonstrates that parsec-scale localizations of candidate BBH systems need not be restricted to radio loud objects like the radio galaxy 0402+379 with its 7-pc separation (Rodriguez et al., 2006).

Our VLBA findings cannot exclude additional sources with K in the same field of view. Such a value is atypically low compared to other RQQs detected with the VLBA (Blundell et al., 1996; Blundell & Beasley, 1998; Ulvestad et al., 2005). But those studies targeted RQQ stronger than several millijansky and were thus biased toward detecting higher brightness temperatures. Several RQQ were not detected in the survey of Blundell & Beasley (1998), with brightness temperature limits similar to the present study. This suggests that the effects of source resolution could also contribute to non detections of RQQs.

3.3. Lone Double-Peaked Emitter Scenario

The lone DPE scenario for the RQQ SDSS J1536+0441A (Gaskell, 2010; Chornock et al., 2010) requires the presence of only one quasar. The line profiles for SDSS J1536+0441A do make it an unusual DPE however (Chornock et al., 2010; Lauer & Boroson, 2009). As a class, DPEs are rare, constituting only 4% of the spectroscopically-selected sample at of Strateva et al. (2003). Yet 76% of those DPEs are radio quiet like SDSS J1536+0441A. Using 1.4 GHz detections from White et al. (1997), Strateva et al. (2003) tabulate 1.4 GHz luminosities for 18 RQ DPEs and report typical values of several times  ergs s. SDSS J1536+0441AB is not detected by White et al. (1997) and the 1.4 GHz luminosity is less than  ergs s, a limit consistent with detected RQ DPEs in Strateva et al. (2003). Moreover, except for SDSS J1536+0441A, no RQQs detected with the VLBA (Blundell et al., 1996; Blundell & Beasley, 1998; Ulvestad et al., 2005) are known to exhibit DPEs, so it is impossible to say whether or not the VLBA detection of SDSS J1536+0441A is in any way unusual. In this regard, VLBA imaging of the DPE sample of Strateva et al. (2003) would be a useful undertaking. As the VLBA detection of SDSS J1536+0441A demonstrates, such imaging is feasible for both radio-quiet and radio-loud DPEs.

4. Summary

Our VLBA search for 8.4 GHz emission from the RQQ SDSS J1536+0441A found only one source within the localization region, of area 0.15 kpc, for the broad-line emission. The VLBA detection has a diameter of less than 8.5 pc and a K. This detection of SDSS J1536+0441A represents the first parsec-scale localization of a candidate BBH system identified through its broad-line properties. The VLA photometry at a rest frequency of 31.2 GHz yields a rising spectrum, consistent with synchrotron self-absorption, which implies a radius of about 0.04 pc for the radio-sphere, and K. The observed compact flat spectrum radio sphere is consistent with the trait predicted in the coronal framework for RQQs. The radio-sphere at 31.2 GHz happens to be just inside the estimated inner boundary for the BLR, of 800 gravitational radii, in this object.

It would be useful to investigate the spectrum at higher frequencies, corresponding to the mm and sub-mm range, to determine where the spectral slope steepens as the source becomes optically thin. This will establish the size of the most compact synchrotron-emitting region, and potentially allow a direct exploration of relativistic electrons in the accretion disk corona. We plan such investigations using the Expanded VLA (Perley et al., 2009) and the Atacama Large Millimeter/submillimeter Array (Wootten & Thompson, 2009).

Concerning the BBH scenarios for SDSS J1536+0441A, the VLBA detection is consistent with a quasar separation of less than 8.5 pc. No additional sources with K or more are found in within the localization region for the broad-line emission.

Concerning the lone DPE scenario for SDSS J1536+0441A, its emission line profiles do make it an unusual DPE. But as no other radio-quiet DPEs have been imaged on parsec scales, it is impossible to say whether or not the VLBA detection of SDSS J1536+0441A is in any way unusual.

We acknowledge prompt and helpful feedback from the referee, and useful discussions with Craig Walker and Greg Taylor. This research was supported by THE ISRAEL SCIENCE FOUNDATION (grant #407/08), and by a grant from the Norman and Helen Asher Space Research Institute. Facilities: VLA VLBA
Figure 1.— VLBA image of Stokes emission from SDSS J1536+0441A at a frequency of 8.4 GHz and spanning 20 mas (104 pc). The rms noise is 0.056 mJy beam (1 ) and the hatched ellipse shows the Gaussian beam dimensions at FWHM. Geometric-mean beam width is 1.63 mas (8.5 pc) at FWHM. Contours are at 6, 4, 2, 2, 4, 6, 8, 10, 12, … 14 times 1 . Negative contours are dashed and positive ones are solid. Image peak is 0.79 mJy beam. Linear gray scale spans 0.24 mJy beam to 0.79 mJy beam.
Figure 2.— Stokes emission at 8.4 GHz centered on the VLBA detection of SDSS J1536+0441A and spanning 200 mas (1.04 kpc). The 1 rms noise, beam dimensions, image peak and gray scale are as in Fig. 1. Contours are at 5 and 5 times 1 . The square shows the field of view for Fig. 1. The parallelogram encloses the localization region, of area 5400 mas (0.15 kpc), for the broad-line emission systems. Any additional point-like source within the parallelogram has an observed peak below 5 .


  1. affiliation: National Radio Astronomy Observatory, P.O. Box O, Socorro, NM 87801;
  2. affiliation: Physics Department, Technion, Haifa 32000, Israel;
  3. slugcomment: Accepted by ApJL on 2010 March 31
  4.  km s Mpc and , implying a luminosity distance of 2.1 Gpc, an angular size distance of 1.1 Gpc and a scale of 5.2 pc per mas.
  5. Operated by the National Radio Astronomy Observatory, which is a facility of the National Science Foundation, operated under cooperative agreement by Associated Universities, Inc.


  1. Arzoumananian, Z., Lowenstein, M., Mushotzky, R. F., & Gendreau, K. C. 2009, ATel, 1931
  2. Barvainis, R., Lonsdale, C., & Antonucci, R. 1996, AJ, 111, 1431
  3. Barvainis, R., Lehar, J., Birkinshaw, M., Falcke, H., & Blundell, K. M. 2005, ApJ, 618, 108
  4. Blundell, K. M., & Beasley, A. J. 1998, MNRAS, 299, 165
  5. Blundell, K. M., Beasley, A. J., Lacy, M., & Garrington, S. T. 1996, ApJ, 468, L91
  6. Bogdanovic, T., Eracleous, M., & Sigurdsson, S. 2009, ApJ, 697, 288
  7. Boroson, T.A., & Lauer, T. R. 2009, Nature, 458, 53
  8. Chornock, R., et al. 2010, ApJ, 709, L39
  9. Colpi, M., & Dotti, M., arXiv:0906.4339v1
  10. Decarli, R., et al. 2009, ApJ, 703, L76
  11. Dotti, M., Montuori, C., Decarli, R., Volonteri, M., Colpi, M., & Haardt, F. 2009, MNRAS, 398, L73
  12. Gaskell, M. 2010, Nature, 463, E1
  13. Kukula, M. J., Dunlop, J. S., Hughes, D. H., & Rawlings, S. 1998, MNRAS, 297, 366
  14. Laor, A., & Behar, E. 2008, MNRAS, 390, 847
  15. Lauer, T. R. & Boroson, T.A., 2009, ApJ, 703, 930
  16. Napier, P. J., Bagri, D. S., Clark, B. G., Rogers, A. A. E., Romney, J. D., Thompson, A. R., & Walker, R. C. 1994, Proc. IEEE, 82, 658
  17. Perley, R., Napier, P., Jackson, J., Butler, B., Clark, B., Hayward, R., Durand, S., Revnell, M., McKinnon, M. 2009, Proc. IEEE, 97, 1448
  18. Phinney, E. S. 1985, Astrophysics of Active Galaxies and Quasi-Stellar Objects, 453
  19. Readhead, A. C. S. 1994, ApJ, 426, 51
  20. Rodriguez, C., Taylor, G. B., Zavala, R. T., Peck, A. B., Pollack, L. K., & Romani, R. W. 2006, ApJ, 646, 49
  21. Shields, G. A., et al. 2009, ApJ, 707, 936
  22. Strateva, I. V., et al. 2003, AJ, 126, 1720
  23. Tang, S., & Grindlay, J. 2009, ApJ, 704, 1189
  24. Thompson, A. R., Clark, B. G., Wade, C. M., Napier, P. J. 1980, ApJS, 44, 151
  25. Ulvestad, J. S., Antonucci, R. R., & Barvainis, R. 2005, ApJ, 621, 123
  26. White, R. L., Becker, R. H., Helfand, D. J., & Gregg, M. D. 1997, ApJ, 475, 479 (
  27. Wootten, A., & Thompson, A. R. 2009, Proc. IEEE, 97, 1463
  28. Wrobel, J. M., Taylor, G. B., Rector, T. A., Myers, S. T., & Fassnacht, C. D. 2005, AJ, 130, 923
  29. Wrobel, J. M., & Ho, L. C. 2006, ApJ, 646, L95
  30. Wrobel, J. M., & Laor, A. 2009 ApJ, 699, L22
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description