by Walter Lewin
The Italian astronomer Laura Maraschi was visiting MIT in 1976, and one day in February she walked into my office and suggested that the bursts were the result of thermonuclear flashes, huge thermonuclear explosions on the surface of accreting neutron stars. When hydrogen accretes onto a neutron star, gravitational potential energy is converted to such tremendous heat that X-rays are emitted (see previous chapter). But as this accreted matter accumulates on the surface of the neutron star, Maraschi suggested, it might undergo nuclear fusion in a runaway process (like in a hydrogen bomb) and that might cause an X-ray burst. The next explosion might go off a few hours later when enough new nuclear fuel had been accreted to ignite. Maraschi demonstrated with a simple calculation on my blackboard that matter racing at roughly half the speed of light to the surface of a neutron star releases much more energy than what is released during the thermonuclear explosions, and that is what the data showed.
I was impressed—this explanation made sense to me. Thermonuclear explosions fit the bill. The cooling pattern we’d observed during the bursts also made sense if what we were seeing was a massive explosion on a neutron star. And her model explained the interval between bursts well since the amount of matter required for an explosion had to build up over time. At the normal rate of accretion, it should take a few hours to build up a critical mass, which was the kind of interval we found in many burst sources.
I keep a funny kind of radio in my office that always unsettles visitors. It’s got a solar-powered battery inside, and it works only when the battery has enough juice. As the radio sits there soaking up sunlight, it slowly fills up with juice (a lot more slowly in the winter), then every ten minutes or so—sometimes longer if the weather’s rotten—it suddenly starts playing, but only for a couple of seconds, as it quickly exhausts its supply of electricity. You see? The buildup in its battery is just like the buildup of accreted matter on the neutron star: when it gets to the right amount, the explosion goes off, and then fades away.
Then, several weeks after Maraschi’s visit, on March 2, 1976, in the middle of burst fever, we discovered an X-ray source that I named MXB 1730-335 that was producing a few thousand bursts per day. The bursts came like machine-gun fire—many were only 6 seconds apart! I don’t know if I can completely convey just how bizarre this seemed to us. This source (now called the Rapid Burster) was a complete outlier, and it immediately killed Maraschi’s idea. First, there is no way that a sufficient amount of nuclear fuel could build up in six seconds on the surface of a neutron star to produce a thermonuclear explosion. Not only that, but if the bursts were a by-product of accretion, we should see a strong X-ray flux due to accretion alone (release of gravitational potential energy), far exceeding the energy present in the bursts, but that was not the case. So it seemed in early March 1976 that Maraschi’s wonderful thermonuclear model for the bursts was as dead as the proverbial doornail. In our publication on MXB 1730-335, we suggested that the bursts are caused by “spasmodic accretion” onto a neutron star. In other words, what in most X-ray binaries is a steady flow of hot matter from the accretion disk onto the neutron star is very irregular in the case of the Rapid Burster.
When we measured the bursts over time, we found that the bigger the burst, the longer the wait before the next one. The waiting time to the next burst could be as short as six seconds and as long as eight minutes. Lightning does something similar. When there’s a particularly large lightning bolt, the large discharge means that the wait needs to be longer for the electric field to build up its potential to the point that it can discharge again.
Later that year a translation of a 1975 Russian paper about X-ray bursts surfaced out of nowhere; it had been reporting burst detections made in 1971 with the Kosmos 428 satellite. We were stunned; the Russians had discovered X-ray bursts, and they had beaten the West! However, as I heard more and more about these bursts, I became very skeptical. Their bursts behaved so very, very differently from the many bursts that I had detected with SAS-3 that I began to seriously doubt whether the Russian bursts were real. I suspected that they were either man-made or produced near Earth in some odd, bizarre way. The iron curtain made it difficult to pursue this; there was no way to find out. However, I had the good fortune to be invited to attend a high-level conference in the Soviet Union in the summer of 1977. Only twelve Russians and twelve U.S. astrophysicists had been invited. That’s where I met for the first time the world famous scientists Joseph Shklovsky, Roald Sagdeev, Yakov Zel’dovich, and Rashid Sunyaev.
I gave a talk on—you guessed it—X-ray bursts, and I got to meet the authors of the Russian burst paper. They generously showed me data of many bursts, way more than they had published in 1975. It was immediately obvious to me that all this was nonsense, but I did not tell them that, at least not at first. I first went to see their boss, Roald Sagdeev, who at the time was the director of the Space Research Institute of the USSR Academy of Sciences in Moscow. I told him that I wanted to discuss something rather delicate with him. He suggested we not do that in his office (bugs were all over the place), so we went outside. I gave him my reasons why their bursts were not what they thought they were—he immediately understood. I told him that I was afraid that my telling the world about this might get these guys into deep trouble under the Soviet regime. He assured me that that would not be the case, and he encouraged me to meet with them and tell them exactly what I had told him. So I did, and that was the last we ever heard of the Russian X-ray bursts. I’d like to add that we are still friends!
You may be curious to know what caused these Russian bursts. At the time I had no idea, but now I know; they were man-made, and guess who made them—the Russians! I’ll solve this mystery in a bit.
For now let’s return to the real X-ray bursts, which we were still trying to figure out. When the X-rays of the bursts plow into the accretion disk (or into the donor star) of an X-ray binary, the disk and the star get hotter and light up briefly in the optical part of the spectrum. Since the X-rays would first have to travel to the disk and donor star, we expected that any optical flash from the disk would reach us seconds after the X-ray burst. So we went hunting for coordinated X-ray and optical bursts. My former graduate student Jeff McClintock and his co-workers had made the first two optical identifications of burst sources (MXB 1636-53 and MXB 1735-44) in 1977. These two sources became our targets.
You see how science works? If a model is correct, then it ought to have observable consequences. In the summer of 1977 my colleague and friend Jeffrey Hoffman and I organized a worldwide simultaneous X-ray, radio, optical, and infrared “burst watch.”
This was an amazing adventure all by itself. We had to convince astronomers at forty-four observatories in fourteen countries to devote precious observing time during the most favorable hours (known as “dark time,” when the Moon is absent) staring at one faint star—that might do nothing. That they were willing to participate shows you just how significant astronomers considered the mystery of X-ray bursts. Over thirty-five days, with SAS-3, we detected 120 X-ray bursts from the burst source MXB 1636-53 but absolutely no bursts were observed with the telescopes on the ground. What a disappointment!
You might imagine that we had to apologize to our colleagues around the world, but the truth is that none saw it as a problem. This is what science is all about.
So we tried again the following year using only large ground-based telescopes. Jeff Hoffman had left for Houston to become an astronaut, but my graduate student Lynn Cominsky and the Dutch astronomer Jan van Paradijs (who had come to MIT in September 1977) joined me in the 1978 burst watch.* This time we selected MXB 1735–44. On the night of June 2, 1978, we succeeded! Josh Grindlay and his co-workers (including McClintock) detected an optical burst with the 1.5-meter telescope at Cerro Tololo in Chile a few seconds after we, at MIT, detected an X-ray burst with SAS-3. We made it to the front page of Nature, which was quite an honor. This work further supported our conviction that X-ray bursts come from X-ray binaries.
What w
as very puzzling to us was why all burst sources except one produce only a handful of bursts in a day and why the Rapid Burster was so very different. The answer lay with the most wonderful—and most bewildering—discovery of my career.
The Rapid Burster is what we call a transient. Cen X-2 is also a transient (see chapter 11). However, the Rapid Burster is what we call a recurrent transient. In the 1970s it became burst-active about every six months, but only for several weeks, after which it would go off the air.
About a year and a half after we discovered the Rapid Burster, we noticed something about its burst profiles that transformed this mystery source into a Rosetta Stone of X-ray bursters. In the fall of 1977, when the Rapid Burster was active again, my undergraduate student Herman Marshall looked very closely at the X-ray burst profiles and discovered a different kind of burst among the very rapid bursts, one that came far less frequently, about every three or four hours. These special bursts, as we called them at first, exhibited the same black body–like cooling profile that characterized all the bursts from the many other burst sources. In other words, perhaps what we were calling special bursts—we soon called them Type I bursts, and gave the rapid bursts the designation Type II—weren’t so special at all. The Type II bursts were clearly the result of spasmodic accretion—there was never any doubt about that—but maybe the common Type I bursts were due to thermonuclear flashes after all. I’ll tell you shortly how we figured that out—just bear with me.
In the fall of 1978 my colleague Paul Joss at MIT had made some careful calculations about the nature of thermonuclear flashes on the surface of neutron stars. He concluded that the accumulated hydrogen first quietly fuses to helium, but that the helium, once it reaches a critical mass, pressure, and temperature, can then violently explode and produce a thermonuclear flash (thus a Type I burst). This led to a prediction that the X-ray energy released in the steady accretion should be roughly a hundred times larger than the energy released in the thermonuclear bursts. In other words, the available gravitational potential energy was roughly a hundred times larger than the available nuclear energy.
X-ray bursts from the Rapid Burster detected with SAS-3 in the fall of 1977. The height of the line represents the number of detected X-rays in about one second, while the horizontal axis represents time. each panel shows about 300 seconds of data. The rapidly repetitive Type II bursts are numbered sequentially. One “Special Burst” is visible in each panel; they have different numbers. They are the Type I bursts (thermonuclear flashes). This figure is from Hoffman, Marshall, and Lewin, nature, 16 Feb. 1978.
We measured the total amount of energy emitted in X-rays from the Rapid Burster during the five-and-a-half days of our fall 1977 observations, and we found that about 120 times more energy was emitted in the Type II bursts than in the “special” Type I bursts. That was the clincher! At that point we knew that the Rapid Burster was an X-ray binary and that Type I bursts were the result of thermonuclear flashes on the surface of an accreting neutron star and that the Type II bursts were the result of the release of gravitational potential energy of the matter flowing from the donor star to the neutron star. There simply was no doubt about this anymore; from that time on, we knew that all Type I burst sources were neutron star X-ray binaries. At the same time we knew conclusively that black holes could not be the source of the bursts. Black holes have no surface, so they cannot produce thermonuclear flashes.
Even though it was already crystal clear to most of us by 1978 that burst sources were accreting neutron star binaries, Grindlay at Harvard continued to insist that the bursts were in fact produced by massive black holes. He even published a paper in 1978 in which he tried to explain how the bursts are produced by very massive black holes. I told you scientists can get emotionally attached to their theories. The Real Paper in Cambridge ran a long story, “Harvard and MIT at the Brink,” featuring pictures of Grindlay and me.
Evidence for the binary nature of burst sources came in 1981 when my Danish friend Holger Pederson, Jan van Paradijs, and I discovered the 3.8-hour orbital period of the burst source MXB 1636–53. Yet, it was not until 1984 that Grindlay finally conceded.
So it was the weirdest X-ray source, the Rapid Burster, that helped confirm the theory of normal (Type I) X-ray bursts, which had been mystifying in their own right. The irony is that for all it explained, the Rapid Burster has remained mostly a mystery. Not so much for observers, but for theoreticians it remains an embarrassment. The best we could do, and in some ways the best we’ve ever been able to do, is come up with the explanation of “spasmodic accretion”—I know, it sounds like something you could catch on an exotic vacation. And the truth is, it’s words, not physics. Somehow, the matter headed for the neutron star is temporarily held up in the disk before a blob or a ring of matter is released from the disk and spurts down to the surface of the star, releasing gravitational potential energy in bursts. We call this release a disk instability, but that too is just words; no one has a clue why and how it works.
Frankly, we also do not understand what the mechanism is behind the recurrent transient behavior of X-ray sources. Why do they turn on and off and on and off? We just don’t know. Once in 1977 we started to pick up bursts simultaneously in all of SAS-3’s detectors. This was bizarre, since they were viewing the sky in totally different directions. The only reasonable explanation we could come up with was that very-high-energy gamma rays were penetrating the entire spacecraft (something that X-rays cannot do) and leaving signals behind. Since all detectors “fired” at the same time, we had no clue what direction these gamma rays were coming from. After we had observed a few dozen of these episodes over a period of several months, they stopped. But thirteen months later they started up again. No one at MIT had a clue.
With the help of one of my undergraduate students, Christiane Tellefson, I started to catalog these bursts, and we even classified them as bursts A, B, and C, depending on their profiles. I stored them all in a file that I labeled SHIT BURSTS.
I remember giving a presentation to some people from NASA (who would visit us yearly), telling them our latest exciting news on X-ray bursts and showing them some of these bizarre bursts. I explained my reluctance to publish; they just didn’t look kosher to me. However, they encouraged me not to delay publishing. So Christiane and I started to write a paper.
Then one day, completely out of the blue, I received a call from my former student Bob Scarlett, who was doing classified research at the Los Alamos National Laboratory. He asked me not to publish these weird bursts. I wanted an explanation, but he was not allowed to tell me why. He asked me to give him some of the times that these bursts had occurred, which I did. Two days later he called again and this time he urged me not to publish for reasons of national security. I nearly fell off my chair. I immediately called my friend France Córdova, who had once worked with me at MIT but who at that time was also working in Los Alamos. I told her about my conversations with Bob and hoped that she could cast some light on what was going on. She must have discussed it with Bob, because a few days later she too called and urged me not to publish. To put my mind at rest, she assured me that these bursts were of zero astronomical interest. To make a long story short, I did not publish.
Many years later I learned what had happened: the “shit bursts” had been produced by several Russian satellites that were powered by nuclear electrical generators, which contain extremely strong radioactive sources. Whenever SAS-3 came near any of the Russian satellites, they would shower our detectors with gamma rays emitted by the radioactive sources. Now, remember those weird bursts detected by the Russians back in 1971? I’m now quite certain these were also caused by the Russians’ own satellites… what irony!
This period of my life, beginning in the late 1970s and going through 1995, was incredibly intense. X-ray astronomy was the cutting edge of observational astrophysics then. My involvement with X-ray bursts pushed me to the pinnacle of my scientific career. I probably gave a dozen colloquia yearly a
ll over the world, in Eastern and Western Europe, Australia, Asia, Latin America, the Middle East, and throughout the United States. I got invited to give talks at many international astrophysics conferences and was the chief editor of three books on X-ray astronomy, the last one, Compact Stellar X-ray Sources, in 2006. It was a heady, wonderful time.
And yet, despite the amazing advances we made, the Rapid Burster has resisted all attempts to unlock its deepest mysteries. Someone will figure it out some day, I’m sure. And they in turn will be confronted with something equally perplexing. That’s what I love about physics. And why I keep a poster-size reproduction of the Rapid Burster’s burst profiles prominently displayed in my MIT office. Whether it’s in the Large Hadron Collider or at the farthest reaches of the Hubble Ultra Deep Field, physicists are getting more and more data, and coming up with more and more ingenious theories. The one thing I know is whatever they find, and propose, and theorize, they’ll uncover yet more mysteries. In physics, more answers lead to even more questions.
CHAPTER 15
Ways of Seeing
Most high school and college students hate taking physics because it is usually taught as a complicated set of mathematical formulas. That is not the approach I use at MIT, and it is not the approach I use in this book. I present physics as a way of seeing our world, revealing territories that would otherwise be hidden to us—from the tiniest subatomic particles to the vastness of our universe. Physics allows us to see the invisible forces at play all around us, from gravity to electromagnetism, and to be on the alert not only for where and when we’ll find rainbows, but also halos, fogbows, and glories, and maybe even glassbows.
Each pioneering physicist changed the way we look at the world. After Newton, we could understand and predict the movements of the entire solar system, and we had the mathematics—calculus—to do so. After Newton, no one could claim that sunlight was not made up of colors, or that rainbows came from anything but sunlight refracting and reflecting in raindrops. After Maxwell, electricity and magnetism were forever linked: it was even hard for me to separate them into different chapters in this book.