Swift blinded by the death of a star

It was one of the biggest star explosions ever recorded to the point that the data-analysis software processing the measurements from the satellite ignored it as if it was an anomaly.

 

When collapsing a star emits a burst of gamma and X rays. Credit: DR.

The actual explosion took place 5 billion years ago, even before our Sun and its series of planets were formed. A massive star caved in on itself and transformed into a black hole, sending a violent flow of Gamma and X-ray radiation through the universe which only reached NASA’s Swift observatory, orbiting the Earth, on 21 June 2010.

Launched by NASA in November 2004, Swift is an observatory specialised in detecting and analysing gamma bursts which usually indicate that events of a rare violence have occurred. It then observes them by means of three telescopes in different light wavelength frequencies (gamma, X-ray and ultraviolet).

The Swift observatory undergoes final preparation at the Kennedy Space Center prior to its launch in November 2004. Credit: NASA KSC.

“The intensity of these X-rays was unexpected and unprecedented” said Neil Gehrels, head scientist for the Swift mission at NASA’s Goddard Space Flight Center. “Just when we were beginning to think that we had seen everything that gamma-ray bursts could throw at us, this burst came along to challenge our assumptions about how powerful their X-ray emissions can be”.

Surprise on his return from holidays
In fact the burst was so powerful that the camera was blinded and the software that automatically analyses the measurements from the satellite threw in the towel, assessing that it was dealing with an anomaly, such that the event initially went unnoticed. It took several days before Phil Evans, a postdoctoral research assistant from Leicester University in Great Britain, just back from his holidays, understood what had happened as he reviewed the data from the satellite.

During such an event, the gamma radiation arrives first. Once it has been detected, the detectors are pointed in the direction indicated so that the X-ray radiation which then follows can be measured. Phil Evans wrote the software which converts the Swift data into light intensity curves. When examining the recordings, he noticed that there were no X-ray signals corresponding to the gamma event that had been recorded and dubbed GRB 100621A (standing for Gamma Ray Burst 2010-06-21-A). Taking a closer look, he identified a “gap” in the data. It was then that he realised what had happened and reconstructed the light curve using the peripheral areas of the image in accordance with a method that is commonly used when the signal itself is “overexposed”. The burst lasted one minute, but for 0.2 seconds, there was a peak at 143,000 X-ray photons per second which exceeded the capacities of the camera.

The software was overwhelmed
“So many photons were bombarding the detector each second that it just couldn't count them quickly enough”, explained Phil Evans. “It was like trying to use a rain gauge and a bucket to measure the flow rate of a tsunami”.

The GRB 100621A source as seen by the XRT (X-rays) and UVOT (ultraviolet) telescopes onboard the Swift observatory. Credit: NASA, Swift, Stefan Immler.

Phil Evans’ software is usually able to make a correction on its own in the event that the camera is overexposed, but in this instance, the values obtained were so far off the scale that it assessed them as erroneous and quite simply did not communicate the results.

This burst was 140 times more powerful than the most powerful continuous X-ray source known in the sky (excepting the Sun, of course), a “nearby” neutron star situated 10,000 light years away. The most powerful X-ray burst detected to date was GRB 080319B which was seven times more powerful seen from the Earth (so powerful that it could have been seen by the naked eye), but also twice as close. Adjusted to the same distance, GRB 100621A was 1.5 times more powerful.

A climate precedent
This filtering of data which prevent researchers from becoming snowed under with measurement errors, has once again shown its limitations.

At the beginning of the 1980’s, a similar mishap occurred with the SME (Solar Mesosphere Explorer) satellite which was measuring the concentration of ozone in the atmosphere. Although some theoretical studies were announcing a risk of the ozone layer diminishing above the Antarctica, it was not detecting anything abnormal. It was only when a British team published some particularly alarming results of a survey using balloons in 1985 that the SME results were reviewed and new acceptable data thresholds were set. Amazed scientists then became aware not only of the existence of a “hole” in the ozone layer but also of its size. SME had always been able to see it, but its measurements, deemed “absurd” by the quality control algorithm, had been ruled out during the data processing.