When you have eliminated all other possibilities, Sherlock Holmes instructed, whatever remains, however improbable, must be the answer.
In the mysterious case of the Soft Gamma Repeaters, or SGRs, the answer appears to be a magnetar, a neutron star with a super-strong magnetic field a thousand trillion times stronger than Earth's.
Indeed the magnetic field actually slows the star's rotation and causes starquakes that pump enough energy into the surrounding gases to generate bursts of soft gamma radiation. These led to discovery of the first SGR in 1979. For almost two decades, scientists speculated about the source, and eventually proposed a new class of highly magnetized stars - magnetars.
"The importance of this discovery goes beyond just adding a new oddity to the list of star types," said Dr. Chryssa Kouveliotou, the lead scientist on the discovery. "It ties together two rare, very peculiar classes of stars we have been puzzling over, and puts the evolution of neutron stars and even galaxies in a new light." It may also swell the population of our galaxy to include a few hundred million undiscovered magnetars.
In the May 21 issue of Nature magazine, Kouveliotou and her colleagues describe how they eliminated the other possibilities to make the "Discovery of an X-ray pulsar with a superstrong magnetic field in the Soft Gamma Repeater called SGR 1806-20."
Kouveliotou works for the Universities Space Research Association (USRA) at NASA's Marshall Space Flight Center. Working with her were Dr. Jan van Paradijs of of the University of Amsterdam and the University of Alabama in Huntsville, Dr. Stefan Dieters of the University of Alabama in Huntsville, both working at NASA's Marshall Space Flight Center (MSFC) in Huntsville Ala., and Dr. Tod Strohmayer of NASA's Goddard Space Flight Center in Greenbelt, Md.
Their find strongly supports the magnetar theory offered in 1992 by astrophysicists Dr. Robert Duncan of the University of Texas at Austin and Dr. Christopher Thompson of the University of North Carolina at Chapel Hill. It's a dynamic model that has the neutron star going through a violent afterlife lasting about 10,000 years. Many colleagues discounted the concept, saying that internal pressures and other factors would keep a star from generating such an intense field.
"We were just trying to understand the origin of the magnetic fields of radio pulsars, which are the ordinary, familiar type of neutron stars," Duncan said.
Neutron stars are left when a massive star expends itself in a supernova. Most become pulsars when rotation of the neutron star's magnetic field produces a repeating, clocklike signal in radio, light, even X-rays (right) and gamma rays. One mystery is why a large number of supernovas create magnificent nebulas yet leave no pulsar at the center. This anomaly poses a problem for theorists trying to calculate the rates of star births and deaths and, eventually, the ages of galaxies and the universe.
"The pulsar problem seemed rather subtle because the known pulsar magnetic fields, although enormous on terrestrial scales, are actually very weak compared to what is possible in forming neutron stars," Duncan continued. "It was much easier to find solutions with much stronger fields. Finally we began to wonder: Well, if much more strongly-magnetized stars did form, what would they look like? Only then did we make the connection with SGRs."
SGRs were not recognized as separate class until 1986 even though all three had been seen in 1979.
The telltale outburst
On March 5, 1979, gamma ray detectors on nine spacecraft across our solar system recorded an intense radiation spike. It was just 2/10th of a second long - with as much energy as the sun releases in 1,000 years - followed by a 200-second emission that showed a clear 8-second pulsation period (most SGR bursts release as much energy as the sun releases in one year). The position tied the burst to a supernova remnant known as N49 in the Large Magellanic Cloud.
Immediately, scientists recognized something odd. N49's youth - it's only a few thousand years old - contrasted with its 8-second spin, typical of a much older neutron star.
Something was putting the brakes on it.
The mystery expanded in 1986 when astrophysicists, meeting in Toulouse, France, realized that they had two more objects like this. Each emitted low-energy gamma rays. And each emitted repeated bursts (most gamma ray bursts are one-time events).
Thus, they were dubbed Soft Gamma Repeaters, or SGR. The object associated with N49 was designated SGR 0526-66 (the numbers indicate the position in the sky). The others are SGR 1806-20 at 14 kiloparsecs - one of the most active - and 1900+14, both in the Milky Way.
Theories abounded, but no one could be certain of the cause. So, it remained what Holmes would call "quite a three-pipe problem" until it became a three-satellite problem 10 years later.
In November 1996, the Burst and Transient Source Experiment aboard the Compton Gamma Ray Observatory detected SGR 1806-20 flaring up again. Kouveliotou used time she was allotted on Rossi X-ray Timing Explorer to take a closer look at the X-ray activities that followed.
During Nov. 5-18, 1996, RXTE captured several hours worth of data as bursts came in a "bunching" mode that had not been seen before. Following her observations, RXTE kept watching SGR 1806-20 to provide data for Strohmayer who was allotted time to observe during SGR 1806-20's quiescent phase. The result was a complementary data set that led to collaboration.
"Combining our data gave us both the capability to make a more sensitive search as well as provide a way to verify each others analysis of the data," said Strohmayer.
Reading the evidence
As the name implies, RXTE carries instruments that read data quickly. Where most telescopes really take time exposures, RXTE Proportional Counter Array acts like a fast electronic counter which, combined with its size, was highly effective in searching for a pattern in the X-rays.
"I found a candidate periodic signal at 7.5 seconds in Chryssa's data, but we needed to have confirmation that the signal was also in Tod's dataset before we would be convinced it was real," remarked Dieters. The study was complicated by the need to carefully remove the data segments which had SGR bursts in them. "When you're looking for such a weak pulsar signal, the bursts could totally mask the modulations."
Then they looked through older data gathered by Japan's Advanced Satellite for Cosmology and Astrophysics (ASCA) in 1993. It had observed SGR 1806-20 while it was not bursting (right; links to 371x450-pixel, 155KB JPG) and was instrumental in establishing that the SGR was associated with a supernova remnant.
"When you know what period you are looking for it gives you a great advantage in sensitivity," remarked Strohmayer. "Finding the pulsed signal in the RXTE data allowed us to go back and also find it in the ASCA data, this removed the last shred of doubt that the pulsed signal could possibly be from another, previously unknown object in the RXTE field of view."
Between the ASCA and RXTE observations, SGR 1806-20 had slowed by 8/1,000th of a second. The difference would be miniscule except that it happened in less than four years to an object with more mass than our sun.
To test her finding, Kouveliotou asked Jeff Kommers of the Massachusetts Institute of Technology, another colleague, to check the data. Using a different approach, he came up with 7.5-seconds.
"That was the clincher," Kouveliotou said. "Two good data sets and two different methods of analysis gave the same answer."
Having established that SGR 1806-20 is associated with a pulsar and is slowing, rapidly, the team asked what might fit that profile.
In science, proving what something is often involves proving what it is not. Scientists suspected that SGRs are magnetars, but first they had to eliminate objects other than pulsars as the sources, and then eliminate possibilities other than magnetars as the answer.
The first possibility was simple accretion where material from another star is scooped up by the pulsar, or the magnetar.
Radio telescope observations by team member Dale Frail at the National Radio Astronomy Observatory helped rule out the accretion model. He showed that SGR 1806-20 coincides with a supernova remnant, SNR G10.0-0.3, whose radio broadcasts suggest a compact shape. It is also may be orbiting a nearby massive blue star every 10 years.
But 1806's own stellar wind is too powerful to let material fall inward, so it can't be an accreting pulsar.
Narrowing the field
That leaves a single suspect.
"We found that the pulsar was slowing down at a rate that suggested a magnetic field strength of about 800 trillion Gauss, a field strength similar to that for so called magnetars predicted by previous theoretical work," said Kouveliotou.
By comparison, Earth's magnetic field is a mere 0.6 Gauss at the poles, and the best we can sustain in laboratories on the ground is 1 million Gauss - and that's in a small volume. Normal radio pulsars reach about 1 trillion to 5 trillion Gauss, strong but still short of a magnetar.
"If the field really is this strong," notes Kouveliotou, "then magnetism itself can keep the star hot - about 10 million degrees C (18 million deg. F) at the surface - and power the X-rays coming from its rotating surface."
"At the surface of the star a chunk of magnetizable metal like iron would feel a force equal to 150 million times the Earth's gravitational pull on it," added Strohmayer
At this intensity, the magnetic field's movements wrinkle the crust of the neutron star and cause starquakes that are the source of the soft gamma-ray bursts.
Neutron stars are the only stars with a solid surface, a 1-km (0.6 mile) deep crust covering a thick fluid of neutrons over a superfluid - or possibly solid - core of subatomic particles (left; links to 1296x781-pixel, 289KB JPG. Credit: NASA/Marshall Space Flight Center).
"In ordinary neutron stars the crust is stable, but in magnetars, the crust is stressed by unbearable forces as the colossal magnetic field drifts through it," said Duncan. "This deforms the crust and sometimes cracks it." Violent seismic waves then shake the star's surface, generating Alfven waves - the electromagnetic equivalent of a Slinky toy - which energize clouds of particles above the surface of the star.
It also drags the star down, slowing it to about a 10-second period in just 10,000 years, about the age and speed of SGR 1806-20.
Eventually, the magnetars may become yet another oddity.
"We know six Anomalous X-ray Pulsars (AXPs) that are different from the bulk of the X-ray pulsars," said van Paradijs, who won the 1997 Rossi Prize for identifying the first optical counterpart for a gamma ray burst. "In terms of colors, the X-ray colors of the anomalous pulsars were very red compared to what you might call the blue normal pulsars." Their rotational periods also slow faster than other stars.
"Third, their pulse periods were close together," said van Paradijs. "All of them were like 6 to 10 seconds, which is very different from what you find with normal X-ray pulsars, which have pulse periods as short as less than a tenth of a second and as long as half an hour."
Since they slow down rapidly, then only a handful would be active for us to observe.
"So that even though there may be many of them, most of them are inactive, dead, so to say, lying in the graveyard," van Paradijs said. If the 10 or so SGRs and AXPs are magnetars, each less than 10,000 years old, then they probably form about once every thousand years.
"I think that conservatively 1 million magnetars have formed in our galaxy, and perhaps as many as 30 to 100 million," Duncan added.
Many of the supernova remnants that lack pulsars actually have them in the form of invisible, dead pulsars that exploded as supernovas, sputtered as SGRs concealing magnetars, then faded through the AXP stage to become invisible. Some may be made visible with more sensitive instruments like NASA/Marshall's Advanced X-ray Astrophysics Facility slated for launch in December 1997.
"Our plans are to go back to that source with RXTE, and hopefully with AXAF, and make more sensitive observations," Kouveliotou. "This is just starting. I'm looking forward to a hot debate on the subject."