How black holes stay black

Article metrics

There is probably a massive black hole at the centre of our Galaxy, and perhaps at the centre of most other ordinary galaxies. Their dimness surprises astrophysicists, but a possible explanation has been found in the behavior of the plasma they consume.

In the theory of General Relativity, black holes are concentrations of mass so great that even light cannot escape from them. Paradoxically, in extragalactic astrophysics black holes are considered to be the best way to explain the large luminosities of distant active galactic nuclei, such as quasars. More recently, evidence has been accumulating that large black holes also exist in the centres of many, perhaps most, normal galaxies — but that they cause far less radiation than expected. In a paper to be published in the Astrophysical Journal on 10 January1, Ramesh Narayan and co-workers offer an explanation of why one such candidate black hole, suspected to lie at the centre of our own Galaxy, is so remarkably quiet.

The paradox mentioned above is resolved when one realizes that black holes are completely ‘black’ only within the event horizon, the so-called Schwarzschild radius, which is three million kilometres for a million-solar-mass black hole. Outside this radius, light can escape. Massive black holes in galactic nuclei suck in nearby stars and interstellar gas through their enormous gravitational pull. On its way into the hole, the accreting matter (probably in the form of a rotating disk) heats up through viscous dissipation and so converts gravitational energy into radiation. Calculations show that in such accretion disks about 10% of the initial rest-mass energy is radiated away. This is the most efficient radiation source one can think of, nuclear fusion in stars being a distant second. For this and several other reasons the ‘massive black hole accretion’ model has been the standard explanation for the spectacular energy output of quasars.

During the past decade, astronomers have made observations at ever higher resolution to test this black-hole model in nearby, low-luminosity galaxies — the mass distribution is inferred from velocity measurements of gas and stars. Compact, dark mass concentrations between a million and several billion solar masses exist in about a dozen nearby galaxies2. In the galaxy NGC4258 (ref. 3) and our own Galactic Centre4, we can now say with some confidence that the central dark mass cannot be anything but a black hole. In the case of the Galaxy, the evidence for a 2.6×106-solar-mass black hole is based on measurements of radial velocities and proper motions of stars as close as five light days from the suspected dynamical centre, which coincides with a bright, compact radio source, SgrA* (Fig. 1).

Figure 1: Stellar proper motions in the Galactic Centre.

a, The derived proper motions in the central 6” as vectors (green), with length proportional to the absolute value of the motions, superimposed on an infrared (2 μm) image. b, The proper motion (blue) near the compact radio source SgrA* (cross). c, d, The change in right ascension and declination of two particular stars (marked on the upper images) between 1992 and 1997. The stellar motions systematically increase with decreasing distance from SgrA*, to values as large as 1,500 km s-1 (assuming that the Galactic Centre is 8,000 parsecs away), requiring a compact, dark body of 2.6 million solar masses to exist within five light days of the radio source. Such a massive black hole would be expected to shine brightly in the infrared and X-ray bands, but its dimness may now be explained by the decoupling of electrons and ions in the accreting gas. (Figure adapted from ref. 4.)

When one compares the nuclear luminosities to the accreting black hole model for about ten of the best black hole candidates, a surprising fact emerges: few of them radiate at maximum, quasar-like efficiency. Most nearby black hole candidates are remarkably radiation-deficient (L. Ho, personal communication). The most extreme case is the Galactic Centre; it radiates only 10-8 to 10-7 of the maximum, so-called ‘Eddington’ luminosity (which is fixed by the mass of the central black hole).

One apparently easy way to explain these low luminosities is to postulate that the holes are accreting at very low rates. But the Galactic Centre says otherwise. The central light year is filled with a relatively thick plasma of gas emitted as winds from luminous stars, and from the known wind density one can calculate that the accretion rate onto the central mass has to be at least 10-4 solar masses a year. At 10% conversion efficiency, this would allow the hole to radiate at about 0.2% of the maximum Eddington rate. So unless the current accretion rate onto the central compact source happens to be orders of magnitude smaller than the above value5, the radiation efficiency of SgrA* must be less than a few times 10-6, four to five orders of magnitude smaller than in standard accretion disk models.

One solution to this dilemma was proposed by Melia6,7, who conjectured that the inflowing wind could have very low angular momentum, and so flow in radially. Such purely radial ‘Bondi-Hoyle’ flows have very low radiation efficiencies. Melia then showed that a 10-4-solar-mass yr-1 Bondi-Hoyle flow onto a million-solar-mass black hole would produce something like the observed radio to X-ray spectrum of SgrA*, including its weak infrared emission4. However, even small deviations from the purely radial flow would create a very large amount of infrared radiation, inconsistent with the observations; and such deviations would be likely to occur, at least transiently, through incomplete mixing of the incoming wind gas, or interactions with a pre-existing fossil accretion disk.

The new work of Narayan et al.1 follows a second route. In a hot accretion flow, the gas is ionized to form a plasma. The heavy ions carry most of the mass, and thus of the energy, whereas the electrons produce most of the radiation (through synchrotron, Compton and Bremsstrahlung radiation processes). But, crucially, in a low-density flow the temperatures of the ions and of the electrons may decouple8,10. As a consequence, most of the gravitational energy would be viscously converted into thermal energy of the ions (in a hot ion torus9,10), and not radiated away by the electrons.

Instead, the gravitational energy is carried (‘advected’) with the flow across the event horizon of the black hole11,12. Such a flow leads to a low radiation efficiency even in a highly dissipative accretion disk. The possible application of such a model to the Galactic Centre had already been proposed more than a decade ago13, but Narayan et al. have now shown that a quantitative and self-consistent dynamical model can be constructed that fits most of the SgrA* observations for a 2.6-million-solar-mass black hole accreting at 10-4 solar masses per year. Like Melia's Bondi-Hoyle model, this advection-dominated accretion flow model also explains the spectral shape of the observed emission, with peaks in the millimetre and X-ray bands but very weak emission in the near-infrared and visible.

Turning the argument around, Narayan et al. point out that the success of their model (and the Melia model) leaves little doubt that the dark mass in the Galactic Centre must indeed be a black hole in the strict sense of General Relativity. Only if an event horizon exists can the gravitational energy truly disappear from sight. Otherwise there would have to be a ‘surface’ where the gravitational and thermal energy of the flow was converted to radiation after all. This argument adds to the already strong case for the existence of black holes.

Naturally, there remain some uncertainties. First, one might question how well the current accretion rate is determined by the observations. Apart from an order-of-magnitude uncertainty in the accretion rate estimated above, time variability in the flow may be important. However, the dynamical time for the stellar winds near SgrA* is only about 100 years, and high-energy X-ray observations indicate that SgrA* was not very much more active 100 years ago than it is now. (Hard X-rays are scattered by dense, interstellar clouds. By looking at that scattered radiation from clouds about 100 light years from SgrA*, one can thus infer its X-ray brightness 100 years ago.) Second, the detailed plasma physics of the advection flow and the reality of the two-temperature solution are complex matters that depend on a number of poorly known parameters.

Nevertheless, the new work is the best answer yet to this strange paradox. Why are many massive black holes so black? It is because most of the time they are converting their food very inefficiently into radiation.


  1. 1

    Narayan, R., Mahadevan, R., Grindlay, J. E., Popham, R. G. & Gammie, C. Astrophys. J. 492, 554–568 (1998).

  2. 2

    Kormendy, J. & Richstone, D. Annu. Rev. Astron. Astrophys. 33, 581–624 (1995).

  3. 3

    Miyoshi, al. Nature 373, 127–129 (1995).

  4. 4

    Genzel, R., Eckart, A., Ott, T. & Eisenhauer, F. Mon. Not. R. Astron. Soc. 291, 219–234 (1997).

  5. 5

    Falcke, H., Biermann, P. L., Duschl, W. J. & Mezger, P. G. Astron. Astrophys. 270, 102–106 (1993).

  6. 6

    Melia, F. Astrophys. J. 387, L25-L28 (1992).

  7. 7

    Ozernoy, L. in Back to the Galaxy (eds Holt, S. S. et al.) 40-43 (AIP, New York, 1992).

  8. 8

    Ichimaru, S. Astrophys. J. 214, 840–855 (1977).

  9. 9

    Shapiro, S. L., Lightman, A. P. & Eardley, D. M. Astrophys. J. 204, 187–199 (1976).

  10. 10

    Rees, M. J., Begelman, M. C., Blandford, R. D. & Phinney, E. S. Nature 295, 17–21 (1982).

  11. 11

    Begelman, M. C. Mon. Not. R. Astron. Soc. 184, 53–67 (1978).

  12. 12

    Abramowicz, M., Czerny, B., Lasota, J. P. & Szuskiewicz, E. Astrophys. J. 332, 646–658 (1988).

  13. 13

    Rees, M. i The Galactic Center (eds Riegler, G. R. & Blandford, R. D.) Am. Inst. Phys. Conf. Proc. 83, 166–176 (1982).

Download references

Author information

Rights and permissions

Reprints and Permissions

About this article

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.