in which a method to search
for black holes in the sky
is proposed and pursued
and succeeds (probably)
The Method
Imagine yourself as J. Robert Oppenheimer. It is 1939; you have just convinced yourself that massive stars, when they die, must form black holes (Chapters 5 and 6). Do you now sit down with astronomers and plan a search of the sky for evidence that black holes truly exist? No, not at all. If you are Oppenheimer, then your interests are in fundamental physics; you may offer your ideas to astronomers, but your own attention is now fixed on the atomic nucleus—and on the outbreak of World War 11, which soon will embroil you in the development of the atomic bomb. And what of the astronomers; do they take up your idea? No, not at all. There is a conservatism abroad in the astronomical community, except for that “wild man” Zwicky, pushing his neutron stars (Chapter 5). The worldview that rejected Chandrasekhar’s maximum mass for a white-dwarf star (Chapter 4) still holds sway.
Imagine yourself as John Archibald Wheeler. It is 1962; you are beginning to be convinced, after mighty resistance, that some massive stars must create black holes when they die (Chapters 6 and 7). Do you now sit down with astronomers and plan a search for them? No, not at all. If you are Wheeler, then your interest is riveted on the fiery marriage of general relativity with quantum mechanics, a marriage that may take place at the center of a black hole (Chapter 13). You are preaching to physicists that the endpoint of stellar implosion is a great crisis, from which deep new understanding may emerge. You are not preaching to astronomers that they should search for black holes, or even neutron stars. Of searches for black holes you say nothing; of the more promising idea to search for a neutron star, you echo in your writings the conservative view of the astronomical community: “Such an object will have a diameter of the order of 30 kilometers. . . . it will cool rapidly. . . . There is about as little hope of seeing such a faint object as there is of seeing a planet belonging to another star” (in other words, no hope at all).
Imagine yourself as Yakov Borisovich Zel’dovich. It is 1964; Mikhail Podurets, a member of your old hydrogen bomb design team, has just finished his computer simulations of stellar implosion including the effects of pressure, shock waves, heat, radiation, and mass ejection (Chapter 6). The simulations produced a black hole (or, rather, a computer’s version of one). You are now fully convinced that some massive stars, when they die, must form black holes. Do you next sit down with astronomers and plan a search for them? Yes, by all means. If you are Zel’dovich, then you have little sympathy for Wheeler’s obsession with the endpoint of stellar implosion. The endpoint will be hidden by the hole’s horizon; it will be invisible. By contrast, the horizon itself and the hole’s influence on its surroundings might well be observable; you just need to be clever enough to figure out how. Understanding the observable part of the Universe is your obsession, if you are Zel’dovich; how could you possibly resist the challenge of searching for black holes?
Where should your search begin? Clearly, you should begin in our own Milky Way galaxy—our disk-shaped assemblage of 1012 stars. The other big galaxy nearest to our own, Andromeda, is 2 million light-years away, 20 times farther than the size of the Milky Way; see Figure 8.1. Thus, any star or gas cloud or other object in Andromeda will appear 20 times smaller and 400 times dimmer than a similar one in the Milky Way. Therefore, if black holes are hard to detect in the Milky Way, they will be 400 times harder to detect in Andromeda—and enormously harder still in the 1 billion or so large galaxies beyond Andromeda.
8.1 A sketch of the structure of our Universe.
If searching nearby is so important, then why not search in our own solar system, the realm stretching from the Sun out to the planet Pluto? Might there be a black hole here, among the planets, unnoticed because of its darkness? No, clearly not. The gravitational pull of such a hole would be greater than that of the Sun; it would totally disrupt the orbits of the planets; no such disruption is seen. The nearest hole, therefore, must be far beyond the orbit of Pluto.
How far beyond Pluto? You can make a rough estimate. If black holes are formed by the deaths of massive stars, then the nearest hole is not likely to be much closer than the closest massive star, Sirius, at 8 light-years from Earth; and it almost certainly won’t be closer than the closest of all stars (aside from the Sun), Alpha Centauri, at 4 light-years distance.
How could an astronomer possibly detect a black hole at such a great distance? Could an astronomer just watch the sky for a moving, dark object which blots out the light from stars behind it? No. With its circumference of roughly 50 kilometers and its distance of at least 4 light-years, the hole’s dark disk will subtend an angle no larger than 10−7 arc second. That is roughly the thickness of a human hair as seen from the distance of the Moon, and 10 million times smaller than the resolution of the world’s best telescopes. The moving dark object would be invisibly tiny.
If one could not see the hole’s dark disk as the hole goes in front of a star, might one see the hole’s gravity act like a lens to magnify the star’s light (Figure 8.2)? Might the star appear dim at first, then brighten as the hole moves between Earth and the star, then dim again as the hole moves on? No, this method of search also will fail. The reason it will fail depends on whether the star and the hole are orbiting around each other and thus are close together, or are separated by typical interstellar distances. If they are close together, then the tiny hole will be like a hand-held magnifying glass placed upright on a windowsill on the eighty-ninth floor of the Empire State Building, and then viewed from several kilometers distance. Of course, the tiny magnifying glass has no power to magnify the building’s appearance, and similarly the hole has no effect on the star’s appearance.
8.2 A black hole’s gravity should act like a lens to change the apparent size and shape of a star as seen from Earth. In this figure the hole is precisely on the line between the star and the Earth, so light rays from the star can reach the Earth equally well by going over the top of the hole, or under the bottom, or around the front, or around the back. All the light rays reaching Earth move outward from the star on a diverging cone; as they pass the hole they get bent down toward Earth; they then arrive at Earth on a converging cone. The resulting image of the star on the Earth’s sky is a thin ring. This ring has far larger surface area, and hence far larger total brightness, than the star’s image would have if the black hole were absent. The ring is too small to be resolved by a telescope, but the star’s total brightness can be increased by a factor of 10 or 100 or more.
If the star and the hole are far apart as in Figure 8.2, however, the strength of the focusing can be large, an increase of 10 or 100 or more in stellar brightness. But interstellar distances are so vast that the necessary Earth–hole–star lineup would be an exceedingly rare event, so rare that to search for one would be hopeless. Moreover, even if such a lensing were observed, the light rays from star to Earth would pass the hole at so large a distance (Figure 8.2) that there would be room for an entire star to sit at the hole’s location and act as the lens. An astronomer on Earth thus could not know whether the lens was a black hole or merely an ordinary, but dim, star.
Zel’dovich must have gone through a chain of reasoning much like this as he sought a method to observe black holes. His chain led finally to a method with some promise (Figure 8.3): Suppose that a black hole and a star are in orbit around each other (they form a binary system). When astronomers train their telescopes on this binary, they will see light from only the star; the hole will be invisible. However, the star’s light will give evidence of the hole’s presence: As the star moves around the hole in its orbit, it will travel first toward the Earth and then away. When it is traveling toward us, the Doppler effect should shift the star’s light toward the blue, and when moving away, toward the red. Astronomers can measure such shifts with high precision, since the star’s light, when sent through a spectrograph (a sophisticated form of prism), exhibits sharp spectral lines, and a slight change in the wavelength (color) of such a line stands out clearly. From a measurement of the shift in wavelength, astronomers can infer the velocity of the star toward or away from Earth, and by monitoring the shift as time passes, they can infer how the star’s velocity changes with time. The magnitude of those changes might typically be somewhere between 10 and 100 kilometers per second, and the accuracy of the measurements is typically 0.1 kilometer per second.
What does one learn from such high-precision measurements of the star’s velocity? One learns something about the mass of the hole: The more massive is the hole, the stronger is its gravitational pull on the star, and thus the stronger must be the centrifugal forces by which the star resists getting pulled into the hole. To acquire strong centrifugal forces, the star must move rapidly in its orbit. Thus, large orbital velocity goes hand in hand with large black-hole mass.
To search for a black hole, then, astronomers should look for a star whose spectra show a telltale periodic shift from red to blue to red to blue. Such a shift is an unequivocal sign that the star has a companion. The astronomers should measure the star’s spectra to infer the velocity of the star around its companion, and from that velocity they should infer the companion’s mass. If the companion is very massive and no light is seen from it at all, then the companion might well be a black hole. This was Zel’dovich’s proposal.
8.3 Zel’dovich’s proposed method of searching for a black hole. (a) The hole and a star are in orbit around each other. If the hole is heavier than the star, then its orbit is smaller than the star’s as shown (that is, the hole moves only a little while the star moves a lot). If the hole were lighter than the star, then its orbit would be the larger one (that is, the star would move only a little while the hole moves a lot). When the star is moving away from Earth, as shown, its light is shifted toward the red (toward longer wavelength). (b) The light, upon entering a telescope on Earth, is sent through a spectrograph to form a spectrum. Here are shown two spectra, the top recorded when the star is moving away from Earth, the bottom a half orbit later when the star is moving toward Earth. The wavelengths of the sharp lines in the spectra are shifted relative to each other. (c) By measuring a sequence of such spectra, astronomers can determine how the velocity of the star toward and away from the Earth changes with time, and from that changing velocity, they can determine the mass of the object around which the star orbits. If the mass is larger than about 2 Suns and no light is seen from the object, then the object might be a black hole.
Although this method was vastly superior to any previous one, it nevertheless is fraught with many pitfalls, of which I shall discuss just two: First, the weighing of the dark companion is not straightforward. The star’s measured velocity depends not only on the companion’s mass, but also on the mass of the star itself, and on the inclination of the binary’s orbital plane to our line of sight. While the star’s mass and the inclination may be inferred from careful observations, one cannot do so with ease or with good accuracy. As a result, one can readily make large errors (say, a factor of 2 or 3) in one’s estimate of the mass of the dark companion. Second, black holes are not the only kind of dark companions that a star might have. For example, a neutron-star companion would also be dark. To be certain the companion is not a neutron star, one needs to be very confident that it is much heavier than the maximum allowed for a neutron star, about 2 solar masses. Two neutron stars in a tight orbit around each other could also be dark and could weigh as much as 4 Suns. The dark companion might be such a system; or it might be two cold white dwarfs in a tight orbit with total mass as much as 3 Suns. And there are other kinds of stars that, while not completely dark, can be rather massive and abnormally dim. One must look very carefully at the measured spectra to be certain there is no sign of tiny amounts of light from such stars.
Astronomers had worked hard over the preceding decades to observe and catalog binary star systems, so it was not necessary for Zel’dovich to conduct his search directly in the sky; he could search the astronomers’ catalogs instead. However, he had neither the time nor the patience to comb through the catalogs himself, nor did he have the expertise to avoid all the pitfalls. Therefore, as was his custom in such a situation, he commandeered the time and the talents of someone else—in this case, Oktay Guseinov, an astronomy graduate student who already knew much about binary stars. Together, Guseinov and Zel’dovich found five promising black-hole candidates among the many hundreds of well-documented binary systems in the catalogs.
Over the next few years, astronomers paid little attention to these five black-hole candidates. I was rather annoyed at the astronomers’ lack of interest, so in 1968 I enlisted Virginia Trimble, a Caltech astronomer, to help me revise and extend the Zel’dovich-Guseinov list. Trimble, though only months past her Ph.D., had already acquired a formidable knowledge of the lore of astronomy. She knew all the pitfalls we might encounter—those described above and many more—and she could gauge them accurately. By searching through the catalogs ourselves, and by collating all the published data we could find on the most promising binaries, we came up with a new list of eight black-hole candidates. Unfortunately, in all eight cases, Trimble could invent a semi-reasonable non–black-hole explanation for why the companion was so dark. Today, a quarter century later, none of our candidates has survived. It now seems likely that none of them is truly a black hole.
Zel’dovich knew, when he conceived it, that this binary star method of search was a gamble, by no means assured of success. Fortunately, his brainstorming on how to search for black holes produced a second idea—an idea conceived simultaneously and independently, in 1964, by Edwin Salpeter, an astrophysicist at Cornell University in Ithaca, New York.
Suppose that a black hole is traveling through a cloud of gas—or, equivalently, as seen by the hole a gas cloud is traveling past it (Figure 8.4). Then streams of gas, accelerated to near the speed of light by the hole’s gravity, will fly around opposite sides of the hole and come crashing together at the hole’s rear. The crash, in the form of a shock front (a sudden, large increase in density), will convert the gas’s huge energy of infall into heat, causing it to radiate strongly. In effect, then, the black hole will serve as a machine for converting some of the mass of infalling gas into heat and then radiation. This “machine” could be highly efficient, Zel’dovich and Salpeter deduced—far more efficient, for example, than the burning of nuclear fuel.
8.4 The Salpeter–Zel’dovich proposal for how to detect a black hole.
Zel’dovich and his team mulled over this idea for two years, looking at it first from this direction and then that, searching for ways to make it more promising. However, it was but one of dozens of ideas about black holes, neutron stars, supernovae, and the origin of the Universe that they were pursuing, and it got only a little attention. Then, one day in 1966, in an intense discussion, Zel’dovich and Novikov together realized they could combine the binary star idea with the infalling gas idea (Figure 8.5).
Strong winds of gas (mostly hydrogen and helium) blow off the surfaces of some stars. (The Sun emits such a wind, though only a weak one.) Suppose that a black hole and a wind-emitting star are in orbit around each other. The hole will capture some of the wind’s gas, heat it in a shock front, and force it to radiate. At the one-meter-square black-board in Zel’dovich’s Moscow apartment, he and Novikov estimated the temperature of the shocked gas: several million degrees.
8.5 The Zel’dovich-Novikov proposal of how to search for a black hole. A wind, blowing off the surface of a companion star, is captured by the hole’s gravity. The wind’s streams of gas swing around the hole in opposite directions and collide in a sharp shock front, where they are heated to millions of degrees temperature and emit X-rays. Optical telescopes should see the star orbiting around a heavy, dark companion. X-ray telescopes should see X-rays from the companion.
Gas at such a temperature does not emit much light. It emits X-rays instead. Thus, Zel’dovich and Novikov realized, among those black holes which orbit around stellar companions, a few (though not most) might shine brightly with X-rays.
To search for black holes, then, one could use a combination of optical telescopes and X-ray telescopes. The black-hole candidates would be binaries in which one object is an optically bright but X-ray-dark star, and the other is an optically dark but X-ray-bright object (the black hole). Since a neutron star could also capture gas from a companion, heat it in shock fronts, and produce X-rays, the weighing of the optically dark but X-ray-bright object would be crucial. One must be sure it is heavier than 2 Suns and thus not a neutron star.
There was but one problem with this search strategy. In 1966, X-ray telescopes were extremely primitive.
The Search
The trouble with X-rays, if you are an astronomer, is that they cannot penetrate the Earth’s atmosphere. (If you are a human, that is a virtue, since X-rays cause cancer and mutations.)
Fortunately, experimental physicists with vision, led by Herbert Friedman of the D.S. Naval Research Laboratory (NRL), had been working since the 1940s to lay the groundwork for space-based X-ray astronomy. Friedman and his colleagues had begun, soon after World War 11, by flying instruments to study the Sun on captured German V-2 rockets. Friedman has described their first flight, on 28 June 1946, which carried in the rocket’s nose a spectrograph for studying the Sun’s far ultraviolet radiation. (Far ultraviolet rays, like X-rays, cannot penetrate the Earth’s atmosphere.) After soaring above the atmosphere briefly and collecting data, “the rocket returned to Earth, nose down, in streamlined flight and buried itself in an enormous crater some 80 feet in diameter and 30 feet deep. Several weeks of digging recovered just a small heap of unidentifiable debris; it was as if the rocket had vaporized on impact.”
From this inauspicious beginning, the inventiveness, persistence, and hard work of Friedman and others brought ultraviolet and X-ray astronomy step by step to fruition. By 1949 Friedman and his colleagues were flying Geiger counters on V-2 rockets to study X-rays from the Sun. By the late 1950s, now flying their counters on American-made Aerobee rockets, Friedman and colleagues were studying ultraviolet radiation not only from the Sun, but also from stars. X-rays, however, were another matter. Each second the Sun dumped 1 million X-rays onto a square centimeter of their Geiger counter, so detecting the Sun with X-rays was relatively easy. Theoretical estimates, however, suggested that the brightest X-ray stars would be 1 billion times fainter than the Sun. To detect so faint a star would require an X-ray detector 10 million times more sensitive than those that Friedman was flying in 1958. Such an improvement was a tall order, but not impossible.
By 1962, the detectors had been improved 10,OOO-fold. With just another factor of a thousand to go, other research groups, impressed by Friedman’s progress, were beginning to compete with him. One, a team led by Riccardo Giacconi, would become a formidable competitor.
In a peculiar way, Zel’dovich may have shared responsibility for Giacconi’s success. In 1961, the Soviet Union unexpectedly abrogated a mutual Soviet/American three-year moratorium on the testing of nuclear weapons, and tested the most powerful bomb ever exploded by humans—a bomb designed by Zel’dovich’s and Sakharov’s teams at the Installation (Chapter 6). In panic, the Americans prepared new bomb tests of their own. These would be the first American tests in the era of Earth-orbiting spacecraft. For the first time it would be possible to measure, from space, the X-rays, gamma rays, and high-energy particles emerging from nuclear explosions. Such measurements would be crucial for monitoring future Soviet bomb tests. To make such measurements on the impending series of American tests, however, would require a crash program. The task of organizing and leading it went to Giacconi, a twenty-eight-year-old experimental physicist at American Science and Engineering (a private Cambridge, Massachusetts, company), who had recently begun to design and fly X-ray detectors like Friedman’s. The U.S. Air Force gave Giacconi all the money he needed, but little time. In less than a year, he augmented his six-person X-ray astronomy team by seventy new people, designed, built, and tested a variety of weapons-blast monitoring instruments, and flew them with a 95 percent success rate in twenty-four rockets and six satellites. This experience molded the core members of his group into a loyal, dedicated, and highly skilled team, ideally primed to beat all competitors in the creation of X-ray astronomy.
Left: Herbert Friedman, with payload from an Aerobee rocket, in 1968. Right: Riccardo Giacconi with the Uhuru X-ray detector, ca. 1970. [Left: courtesy U.S. Naval Research Laboratory; right: courtesy R. Giacconi.]
Giacconi’s seasoned team took its first astronomical step with a search for X-rays from the Moon, using a detector patterned after Friedman’s, and like Friedman, flying it on an Aerobee rocket. Their rocket, launched from White Sands, New Mexico, at one minute before midnight on 18 June 1962, climbed quickly to an altitude of 230 kilometers, then fell back to Earth. For 350 seconds it was high enough above the Earth’s atmosphere to detect the Moon’s X-rays. The data, telemetered back to the ground, were puzzling; the X-rays were far stronger than expected. When examined more closely, the data were even more surprising. The X-rays seemed to be coming not from the Moon, but from the constellation Scorpius (Figure 8.6b). For two months, Giacconi and his team members (Herbert Gursky, Frank Paolini, and Bruno Rossi) sought errors in their data and apparatus. When none could be found, they announced their discovery: The first X-ray star ever detected, 5000 times brighter than theoretical astrophysicists had predicted Ten months later, Friedman’s team confirmed the discovery, and the star was given the name Sco X-1 (1 for “the brightest,” X for “X-ray source,” Sco for “in the constellation Scorpius”).
8.6 The improving technology and performance of X-ray astronomy’s tools, 1962–1978. (a) Schematic design of the Geiger counter used by Giacconi’s team in their 1962 discovery of the first X-ray star. (b) The data from that Geiger counter, showing that the star was not at the location of the Moon; note the very poor angular resolution (large error box), 90 degrees. (c) The 1970 Uhuru X-ray detector: A vastly improved Geiger counter sits inside the box, and in front of the counter one sees venetian-blind slats that prevent the counter from detecting an X-ray unless it arrives nearly perpendicular to the counter’s window. (d) Uhuru’s measurements of X-rays from the black-hole candidate Cygnus X-t. (e) Schematic diagram and (I) photograph of the mirrors that focus X-rays in the 1978 X-ray telescope Einstein. (g, h) Photographs made by the Einstein telescope of two black-hole candidates, Cygnus X-1 and SS-433. [Individual drawings and pictures courtesy R. Giacconi.]
How had the theorists gone wrong? How had they underestimated by a factor of 5000 the strengths of cosmic X-rays? They had presumed, wrongly, that the X-ray sky would be dominated by objects already known in the optical sky—objects like the Moon, planets, and ordinary stars that are poor emitters of X-rays. However, Sco X-1 and other X-ray stars soon to be discovered were not a type of object anyone had ever seen before. They were neutron stars and black holes, capturing gas from normal-star companions and heating it to high temperatures in the manner soon to be proposed by Zel’dovich and Novikov (Figure 8.5 above). To deduce that this was indeed the nature of the observed X-ray stars, however, would require another decade of hand-in-hand hard work by experimenters like Friedman and Giacconi and theorists like Zel’dovich and Novikov.
Giacconi’s 1962 detector was exceedingly simple (Figure 8.6a): an electrified chamber of gas, with a thin window in its top face. When an X-ray passed through the window into the chamber, it knocked electrons off some of the gas’s atoms; and those electrons were pulled by an electric field onto a wire, where they created an electric current that announced the X-ray’s arrival. (Such chambers are sometimes called Geiger counters and sometimes proportional counters.) The rocket carrying the chamber was spinning at two rotations per second and its nose slowly swung around from pointing up to pointing down. These motions caused the chamber’s window to sweep out a wide swath of sky, pointing first in one direction and then another. When pointed toward the constellation Scorpius, the chamber recorded many X-ray counts. When pointed elsewhere, it recorded few. However, because X-rays could enter the chamber from a wide range of directions, the chamber’s estimate of the location of Sco X-1 on the sky was highly uncertain. It could report only a best-guess location, and a surrounding 90-degree-wide error box indicating how far wrong the best guess was likely to be (see Figure 8.6b).
To discover that Sco X-1 and other X-ray stars soon to be found were in fact neutron stars and black holes in binary systems would require error boxes (uncertainties in position on the sky) a few minutes of arc in size or smaller. That was a very tall order: a 1000-fold improvement in angular accuracy.
The needed improvement, and much more, came step by step over the next sixteen years, with several teams (Friedman’s, Giacconi’s, and others) competing at each step of the way. A succession of rocket flights by one team after another with continually improving detectors was followed, in December 1970, by the launch of Uhuru, the first X-ray satellite (Figure 8.6c). Built by Giacconi’s team, Uhuru contained a gas-filled, X-ray counting chamber one hundred times larger than the one they flew on their 1962 rocket. In front of the chamber’s window were slats, like venetian blinds, to prevent the chamber from seeing X-rays from any direction except a few degrees around the perpendicular (Figure 8.6d). Uhuru, which discovered and cataloged 339 X-ray stars, was followed by several other similar but special-purpose X-ray satellites, built by American, British, and Dutch scientists. Then in 1978 Giacconi’s team flew a grand successor to Uhuru: Einstein, the world’s first true X-ray telescope. Because X-rays penetrate right through any object that they strike perpendicularly, even a mirror, the Einstein telescope used a set of nested mirrors along which the X-rays slide, like a tobogan sliding down an icy slope (Figures 8.6e,f). These mirrors focused the X-rays to make images of the X-ray sky 1 arc second in size-images as accurate as those made by the world’s best optical telescopes (Figures 8.6g,h).
From Giacconi’s rocket to the Einstein telescope in just sixteen years (1962 to 1978), a 300,000-fold improvement of angular accuracy had been achieved, and in the process our understanding of the Universe had been revolutionized: The X-rays had revealed neutron stars, black-hole candidates, hot diffuse gas that bathes galaxies when they reside in huge clusters, hot gas in the remnants of supernovae and in the coronas (outer atmospheres) of some types of stars, and particles with ultra-high energies in the nuclei of galaxies and quasars.
Of the several black-hole candidates discovered by X-ray detectors and X-ray telescopes, Cygnus X-1 (Cyg X-1 for short) was one of the most believable. In 1974, soon after it became a good candidate, Stephen Hawking and I made a bet; he wagered that it is not a black hole, I that it is.
Carolee Winstein, whom I married a decade after the bet was made, was mortified by the stakes (Penthouse magazine for me if I win; Private Eye magazine for Stephen if he wins). So were my siblings and mother. But they didn’t need to worry that I would actually win the Penthouse subscription (or so I thought in the 1980s); our information about the nature of Cyg X-1 was improving only very slowly. By 1990, in my view, we could be only 95 percent confident it was a black hole, still not confident enough for Stephen to concede. Evidently Stephen read the evidence differently. Late one night in June 1990, while I was in Moscow working on research with Soviet colleagues, Stephen and an entourage of family, nurses, and friends broke into my office at Caltech, found the framed bet, and wrote a concessionary note on it with validation by Stephen’s thumbprint.
Right: The bet between Stephen Hawking and me as to whether Cygnus X-1 is a black hole. Left: Hawking lecturing at the University of Southern California in June 1990, just two hours before breaking into my office and signing off on our bet. [Hawking photo courtesy Irene Fertik, University of Southern California.]
The evidence that Cyg X-l contains a black hole is of just the sort that Zel’dovich and Novikov envisioned when they proposed the method of search: Cyg X-l is a binary made of an optically bright and X-ray-dark star orbiting around an X-ray-bright and optically dark companion, and the companion has been weighed to make sure it is too heavy to be a neutron star and thus is probably a black hole.
The evidence that this is the nature of eyg X-1 was not developed easily. It required a cooperative, massive, worldwide effort carried out in the 1960s and 1970s by hundreds of experimental physicists, theoretical astrophysicists, and observational astronomers.
The experimental physicists were people like Herbert Friedman, Stuart Bowyer, Edward Byram, and Talbot Chubb, who discovered eyg X-1 in a rocket flight in 1964; Harvey Tananbaum, Edwin Kellog, Herbert Gursky, Stephen Murray, Ethan Schrier, and Riccardo Giacconi, who used Uhuru in 1971 to produce a 2-arc-minute-sized error box for the position of Cyg X-1 (Figure 8.7); and many others who discovered and studied violent, chaotic fluctuations of the X-rays and their energies—fluctuations that are what one would expect from hot, turbulent gas around a black hole.
8.7 Left: A negative print of a photograph taken with the 5-meter (200-inch) optical telescope at Palomar Mountain by Jerome Kristian in 1971. The black rectangle outlines the error box in which Uhuru’s 1971 data say that Cygnus X-1 lies. The white x marks the location of a radio flare, measured by radio telescopes, which coincided with a sudden change in the X-rays from Cyg X-1. The x coincides with the optical star HOE 226868, and thus identifies it as a binary companion of Cyg X-1. In 1978 the X-ray telescope Einstein confirmed this identification; see Figure 8.6g. Right: Artist’s conception of Cyg X-1 and HDE 226868, based on all the optical and X-ray data. [Left: photo courtesy Dr. Jerome Kristian, Carnegie Observatories; right: painting by Victor J. Kelley, courtesy the National Geographic Society.]
The observational astronomers contributing to the worldwide effort were people like Robert Hjellming, Cam Wade, Luc Braes, and George Miley, who discovered in 1971 a flare of radio waves in Uhuru’s Cyg X-1 error box simultaneous with a huge, Uhuru-measured change in Cyg X-1’s X-rays, and thereby pinned down the location of Cyg X-1 to within 1 second of arc (Figures 8.6d and 8.7); Louise Webster, Paul Murdin, and Charles Bolton, who discovered with optical telescopes that an optical star, HDE 226868, at the location of the radio flare is orbiting around a massive, optically dark but X-ray-bright companion (Cyg X-1); and a hundred or so other optical astronomers who made painstaking measurements of HDE 226868 and other stars in its vicinity, measurements crucial to avoiding severe pitfalls in estimating the mass of Cyg X-1.
The theoretical astrophysicists contributing to the effort included people like Zel’dovich and Novikov, who proposed the method of search; Bohdan Paczynski, Yoram Avni, and John Bahcall, who developed complex but reliable ways to circumvent the mass-estimate pitfalls; Geoffrey Burbidge and Kevin Prendergast, who realized that the hot, X-ray-emitting gas should form a disk around the hole; and Nikolai Shakura, Rashid Sunyaev, James Pringle, Martin Rees, Jerry Ostriker, and many others, who developed detailed theoretical models of the X-ray-emitting gas and its disk, for comparison with the X-ray observations.
By 1974 this massive effort had led, with roughly 80 percent confidence, to the picture of Cyg X-1 and its companion star HDE 226868 that is shown in an artist’s sketch in the right half of Figure 8.7. It was just the kind of picture that Zel’dovich and Novikov had envisioned, but with far greater detail: The black hole at the center of Cyg X-1 has a mass definitely greater than 3 Suns, probably greater than 7 Suns, and most likely about 16; its optically bright but X-ray-dark companion HDE 226868 has a mass probably greater than 20 Suns and most likely about 33, and it is roughly 20 times larger in radius than the Sun; the distance from the star’s surface to the hole is about 20 solar radii (14 million kilometers); and the binary is about 6000 light-years from Earth. Cyg X-l is the second brightest object in the X-ray sky; HDE 226868, while very bright in comparison with most stars seen by a large telescope, is nevertheless too dim to be seen by the naked eye.
In the two decades since 1974, our confidence in this picture of Cyg X-l has increased from roughly 80 percent to, say, 95 percent. (These are my personal estimates.) Our confidence is not 100 percent because, despite enormous efforts, no unequivocal signature of a black hole has yet been found in Cyg X-l. No signal, in X-rays or light, cries out at astronomers saying unmistakably, “I come from a black hole.” It is still possible to devise other, non-black-hole explanations for all the observations, though those explanations are so contorted that few astronomers take them seriously.
By contrast, some neutron stars, called pulsars, produce an unequivocal “I am a neutron star” cry: Their X-rays, or in some cases radio waves, come in sharp pulses that are very precisely timed. The timing is as precise, in some cases, as the ticking of our best atomic clocks. Those pulses can only be explained as due to beams of radiation shining off a neutron star’s surface and swinging past Earth as the star rotates—the analogue of a rotating light beacon at an airport or in a lighthouse. Why is this the only possible explanation? Such precise timing can come only from the rotation of a massive object with massive inertia and thus massive resistance to erratic forces that would make the timing erratic; of all the massive objects ever conceived by the minds of astrophysicists, only neutron stars and black holes can spin at the enormous rates (hundreds of rotations per second) of some pulsars; and only neutron stars, not black holes, can produce rotating beams, because black holes cannot have “hair.” (Any source of such a beam, attached to the hole’s horizon, would be an example of the type of “hair” that a black hole cannot hang on to.1
An unequivocal black-hole signature, analogous to a pulsar’s pulses, has been sought by astronomers in Cyg X-1 for twenty years, to no avail. An example of such a signature (suggested in 1972 by Rashid Sunyaev, a member of Zel’dovich’s team) is pulsar-like pulses of radiation produced by a swinging beam that originates in a coherent lump of gas orbiting around the hole. If the lump were close to the hole and held itself together for many orbits until it finally began to plunge into the horizon, then the details of its gradually shifting interval between pulses might provide a clear and unambiguous “I am a black hole” signature. Unfortunately, such a signature has never been seen. There seem to be several reasons: (1) The hot, X-ray-emitting gas moves around the black hole so turbulently and chaotically that coherent lumps may hold themselves together for only one or a few orbits, not many. (2) If a few lumps do manage to hold themselves together for a long time and produce a black-hole signature, the turbulent X-rays from the rest of the turbulent gas evidently bury their signature. (3) If Cyg X-1 is indeed a black hole, then mathematical simulations show that most of the X-rays should come from far outside its horizon—from circumferences roughly 10 times critical or more, where there is much more volume from which X-rays can be emitted than near the horizon. At such large distances from the hole, the gravitational predictions of general relativity and Newton’s theory of gravity are approximately the same, so if there were pulses from orbiting lumps, they would not carry a strongly definitive black-hole signature.
For reasons similar to these, astronomers might never find any kind of definitive black-hole signature in any electromagnetic waves produced from the vicinity of a black hole. Fortunately, the prospects are excellent for a completely different kind of black-hole signature: one carried by gravitational radiation. To this we shall return in Chapter 10.
The golden age of theoretical black-hole research (Chapter 7) coincided with the observational search for black holes and the discovery of Cyg X-1 and deciphering of its nature. Thus, one might have expected the youths who dominated the golden age (Penrose, Hawking, Novikov, Carter, Israel, Price, Teukolsky, Press, and others) to play key roles in the black-hole search. Not so, except for Novikov. The talents and knowledge that those youths had developed, and the remarkable discoveries they were making about black-hole spin, pulsation, and hairlessness, were irrelevant to the search and to deciphering Cyg X-1. It might have been different if Cyg X-1 had had an unequivocal black-hole signature. But there was none.
These youths and other theoretical physicists like them are sometimes called relativists, because they spend so much time working with the laws of general relativity. The theorists who did contribute to the search (Zel’dovich, Paczynski, Sunyaev, Rees, and others) were a very different breed called astrophysicists. For the search, these astrophysicists needed to master only a tiny amount of general relativity—just enough to be confident that curved spacetime was quite irrelevant, and that a Newtonian description of gravity would be quite sufficient for modeling an object like Cyg X-1. However, they needed enormous amounts of other knowledge, knowledge that is part of the standard tool kit of an astrophysicist. They needed a mastery of extensive astronomical lore about binary star systems, and about the structures and evolutions and spectra of the companion stars of black-hole candidates, and about the reddening of starlight by interstellar dust—a key tool in determining the distance to Cyg X-1. They also needed to be experts on such issues as the flow of hot gas, shock waves formed when streams of hot gas collide, turbulence in the gas, frictional forces in the gas caused by turbulence and by chaotic magnetic fields, violent breaking and reconnection of magnetic field lines, the formation of X-rays in hot gas, the propagation of X-rays through the gas, and much much more. Few people could be masters of all this and, simultaneously, be masters of the intricate mathematics of curved spacetime. Human limitations forced a split in the community of researchers. Either you specialized in the theoretical physics of black holes, in deducing from general relativity the properties that black holes ought to have, or you specialized in the astrophysics of binary systems and hot gas falling onto black holes and radiation produced by the gas. You were either a relativist or an astrophysicist
Some of us tried to be both, with only modest success. Zel’dovich, the consummate astrophysicist, had occasional new insights about the fundamentals of black holes. I, as a somewhat talented relativist, tried to build general relativistic models of flowing gas near the black hole in Cyg X-1 But Zel’dovich didn’t understand relativity deeply, and I didn’t understand the astronomical lore very well. The barrier to cross over was enormous. Of all the researchers I knew in the golden age, only Novikov and Chandrasekhar had one foot firmly planted in astrophysics and the other in relativity.
Experimental physicists like Giacconi, who designed and flew X-ray detectors and satellites, faced a similar barrier. But there was a difference. Relativists were not needed in the search for black holes, whereas experimental physicists were essential. The observational astronomers and the astrophysicists, with their mastery of the tools for understanding binaries, gas flow, and X-ray propagation, could do nothing until the experimental physicists gave them detailed X-ray data. The experimental physicists often tried to decipher what their own data said about the gas flow and the possible black hole producing it, before turning the data over to the astronomers and astrophysicists, but with only modest success. The astronomers and astrophysicists thanked them very kindly, took the data, and then interpreted them in their own, more sophisticated and reliable ways.
This dependence of the astronomers and astrophysicists on the experimental physicists is but one of many interdependencies that were crucial to success in the search for black holes. Success, in fact, was a product of joint, mutually interdependent efforts by six different communities of people. Each community played an essential role. Relativists deduced, using the laws of general relativity, that black holes must exist. Astrophysicists proposed the method of search and gave crucial guidance at several steps along the way. Observational astronomers identified HDE 226868, the companion of Cyg X-l; they used periodically shifting spectral lines from it to weigh Cyg X-l; and they made extensive other observations to firm up their estimate of its weight. Experimental physicists created the instruments and techniques that made possible the search for X-ray stars, and they carried out the search that identified Cyg X-l. Engineers and managers at NASA created the rockets and spacecraft that carried the X-ray detectors into Earth orbit. And, not least in importance, American taxpayers provided the funds, several hundreds of millions of dollars, for the rockets, spacecraft, X-ray detectors and X-ray telescopes, and the salaries of the engineers, managers, and scientists who worked with them.
Thanks to this remarkable teamwork, we now, in the 1990s, are almost 100 percent sure that black holes exist not only in Cyg X-l, but also in a number of other binaries in our galaxy.
1. Chapter 7. The electric field hair of a charged black hole is evenly distributed around the hole’s spin axis and thus cannot produce a concentrated beam.