sign in sign up
<<
Home , Gamma (satellite), Granat

Gamma-ray Bursts: 15 Years of GRB Afterglows A.J. Castro-Tirado, J. Gorosabel and I.H. Park (eds) EAS Publications Series, 61 (2013) 15–25

EARLY DANISH GRB EXPERIMENTS – AND SOME FOR THE FUTURE?

N. Lund1

Abstract. By 1975 the hunt for GRB counterparts had been on for almost ten years without success. burst instruments of that day provided little or no directional data in themselves. Positions could be extracted only using the time delay technique – potentially accurate but very slow. Triggered by a japanese report of a balloon instrument for GRB studies based on a Rotation Modulation Collimator we at the Danish Space Research Institute started the development of an RMC detector for GRBs, the WATCH wide field monitor. Four WATCH units were flown on the Soviet satellites, and one on ESA’s EURECA satellite. The design and results will be summarized. Now, 35 years later, recent detector developments may allow the construction of WATCH-type instruments able to fit weight, power and data-wise into 1 kg cubesats. This could provide the basis for a true all-sky monitor with 100 percent duty cycle for rare, bright events.

1 Introduction

The Danish Space Research Institute (DSRI) was set up in 1968 to provide a national point for a national participation in the European Space Research Organisation, ESRO. Prior to the formation of DSRI the national efforts in space had been directed towards the ionosphere because of the importance of this at- mospheric region for radio communications with Greenland. The first director of DSRI was Bernard Peters, a well known figure in the post war cosmic ray re- search. Not surprisingly, the first astrophysics project was a cosmic ray isotope experiment, constructed in collaboration with Centre d’Etudes Nucleaire, Saclay in . This experiment was launched in 1979 on the NASA satellite HEAO-3. With a weight of 350 kg the HEAO-3 experiment was very large for its time, and even before the launch it was clear that DSRI had to find a new future research

1 DTU Space, Elektrovej Building 327, 2800 Lyngby,

c EAS, EDP Sciences 2013 DOI: 10.1051/eas/1361002 16 Gamma-ray Bursts: 15 Years of GRB Afterglows theme for its astrophysics group, we could not dream of building anything larger or more complex to improve on the HEAO-3 results. At that time the Cosmic Gamma-Ray Bursts were the big mystery which in- trigued astrophysi-cists all over the world. Why had nobody succeeded in finding any counterparts? How could these events be so luminous in X- and gamma-rays, yet so completely absent at other wave-lengths? The distribution across the sky was peculiar to say the least – no hint of an origin within our Galaxy, yet it seemed obvious that such extraordinary flashes had to have a relatively local origin? A few good positions had been obtained from the early InterPlanetary Network (IPN), but nothing conspicuous was visible within the small IPN error boxes. Figure 1. In the early days it took quite a while to finalize the IPN positions – clock syn- chronization is not trivial! Therefore it was obvious that what was needed was an instrument which by itself could determine the burst position in near real time so ground based follow-up could start within hours or days rather than months. For the follow-up we planned to use classical astronomical search procedures – blink- ing Schmidt-plates, so we expected that a position accuracy better than 1 degree would be adequate – at least it would be far better than anything done in near real time on GRB’s up to that point.

Fig. 1. Error box for GRB 790406 derived by the Interplanetary Network (Evans et al. 1980).

2 Instrument design

For our instrument design we were inspired by a Japanese balloon experiment employing Rotation Modulation Collimators (RMC’s), Nishimura et al. (1976). RMC’s were originally developed by Mertz (1968), actually to be used for electronic read-out of optical Schmidt . However, it was in X-ray astronomy that the RMC technique really made its mark, with experiments on the British ARIEL V and the US SAS-3 satellites. N. Lund: Early Danish GRB Experiments 17

Nishimuras balloon experiment, specifically designed for GRB studies, em- ployed 3 detectors, two RMC detectors with orthogonal grid orientations and one “monitor counter” to provide an unmodulated time history of the burst. Figure 2. In the RMC technique the source positions are derived by an analysis of the “modulation pattern” arising through the rotation of the double grid structure. Figure 3.

Fig. 2. Detector configuration of Fig. 3. Modulation patterns cor- Nishimuras balloon payload. Two or- responding to different off-axis and thogonal RMC detectors and one mon- phase angles. Off-axis: 10◦,20◦,30◦, itor counter. 40◦ and 50◦.

The japanese design with the monitor counter took into account that GRB’s were known to have unpredictable time structures, and the derivation of the in- strument modulation would be uncertain without an independent measurement of the true light curve. Two orthogonal RMC-units were used because the balloon payload rotated relatively slowly (two revolutions per minute) and many GRB’s would only last for a small fraction of a revolution. We realized that by suitable modifications of the RMC detector we could dis- pense with the monitor counter and achieve with one detector what was done with three in the balloon. The design of our WATCH (Wide Angle for Cosmic Hard X-rays) detector is shown together with the classical RMC in Figure 4. We eliminated the lower shadow grid and replaced it by two interleaved grid detectors with the same pitch as the top shadow grid. Now we can derive the un-modulated time history of a burst by adding the signals from the two detectors, and we can derive the instrument modulation of the signal independent of the signal amplitude from the ratio of the time history from one of the detectors to the un-modulated time history. 18 Gamma-ray Bursts: 15 Years of GRB Afterglows

Fig. 4. Comparison of classical RMC (left) and WATCH design (right). The classical RMC uses two 50% open grids rotating synchronously and a single, large area detec- tor which observes the time pattern of light and shadow as the grids rotate. WATCH uses only one shadow grid, but the co-rotating detector is now more complex with two interleaved grid-detectors.

We also opted for a high spin rate of our detector: 60 revolutions per minute. More details on the design of the original WATCH instruments can be found in (Lund 1981 & Brandt et al. 1990). It should be noted that unlike other GRB-instruments WATCH did not rely on the “burst”-nature of the GRBs to observe and localize them. WATCH performs equally well on persistent X-ray sources, it is a true wide field monitor. We build prototypes of the instrument and flew them in balloons from Spitzbergen in 1979 and 1980. Figure 5. No gamma bursts were observed during these flights but the design was proven and on this basis we got the instrument accepted for flight on ESA’s EURECA (EUropean REtrievable CArrier), a micro- gravity satellite with a planned launch in 1988. The main characteristics of the original WATCH instrument (Fig. 7) and the expected characteristics of a modern version of the instrument can be found in Table 1 of Section 5.

3 The challenger disaster and a new opportunity

The construction of the WATCH flight model was well underway when in January 1986 the Challenger accident put a halt to the US shuttle program. For more than a year it was undecided whether EURECA would ever fly – in the aftermath of the disaster NASA had decided to transfer all future satellite launches back to expendable launchers, the Shuttle would only be used for manned flights – with a few exceptions. N. Lund: Early Danish GRB Experiments 19

In the summer of 1986, in the middle of this limbo I was fortunate to meet Rashid Sunyaev from the at a COSPAR meeting in Toulouse. We both made a presentation in a session devoted to future X-ray satellite missions. Rashid presented “Granat”, a Russian-French mission with a large French gamma-ray instrument, Sigma, and a cluster of Russian X-ray telescopes, ART-P and ART-S. In addition Granat carried two gamma-burst instruments, the French Phebus and the Russian Konus with an associated rapid moving platform, “Tournesol” with X-ray and optical cameras. (Unfortunately the downlink data connection to the entire Russian GRB instrument complex was lost soon after launch – otherwise the GRB afterglows may have been discovered with Granat in 1990 rather than with SAX in 1997.) In his presentation Rashid expressed regret that Granat did not carry an all-sky monitor. Such an instrument had been fore- seen, there was room for it as well as excess payload mass, but the instrument development had been delayed. Immediately after this I presented WATCH – an all-sky monitor which now appeared to be homeless! This was too much of a coincidence to be neglected. After the session Rashid and I met and after a good bottle of French red wine I could go back to Copenhagen with an offer from Rashid to fly four WATCH units on Granat – delivery of the flight units to be executed within a year! Of course we did not succeed to build four flight units adapted for Granat within a year, but the delivery of the flight units began in 1988. A few stones had to be cleared during the adaptation – Granat had room and mass to accommodate four WATCH units, but no data storage and very little power. We had to modify the on-board software to allow the storage of data for four days (the Granat ) inside the 512 kByte RAM memory of each instrument – and we had to negotiate directly with the spacecraft builder, Lavotchkin, to pay for a special WATCH panel to be fitted to the spacecraft. (The Granat spacecraft was the 25th and last copy of the Russian “”-probes handed over to IKI to be used for X- and gamma-ray astronomy on the condition that there would be absolutely no modification on the spacecraft systems! All final decisions can be revoked under suitable temptation!) Granat was launched in December 1989 carrying four WATCH units. Unfor- tunately one of the units did not survive the launch, a thin aluminium foil used as an entrance window for the broke, probably due to air trapped behind the foil. So we had to contend ourselves with a monitor for 75% of the sky rather than 100%. But even the 75% gave us plenty of data between 1990 and 1994 as will be discussed below. By 1989 it had also become clear that NASA would stick to the agreement with ESA: EURECA would be launched and retrieved by the shuttle system. So the EURECA programme began to move forward again. The EURECA WATCH did incorporate a number of improvements relatively to the Granat version, not the least a X-ray entrance window as replacement for the less reliable alu- minium foil. But also a important modification of the scintillator mosaic which re- duced the artefacts present in the Granat RMC correlation images. EURECA was 20 Gamma-ray Bursts: 15 Years of GRB Afterglows launched by the space shuttle Atlantis in July 1992 and retrieved after 11 month in by the Endeavour shuttle in June of 1993.

4WATCHresults

The original inspiration for WATCH was the enigma of the gamma-ray bursts. And GRBs were detected and localized. But as will be illustrated it was as an X-ray all-sky monitor WATCH made its most significant discoveries.

4.1 Gamma-ray bursts

A total of 47 bursts were localized by Granat WATCH between January 1990 and September 1994 (Sazonov et al. 1997). For the first time GRBs were imaged in real time, and for the brightest of the bursts the source position was even derived on-board within seconds. But the precision of the localization, 0.2 to 1 degree error radius – although far better than anything previously obtained – was in- adequate for the facilities available on the ground in those days. Only Schmidt telescopes could cover the error boxes, and their photographic recording system was lacking in sensitivity. But probably the primary obstacle was the slowness of the communication systems. Data was only retrieved from Granat once every four days, so most GRB detections were “out of date” already by the time the data was retrieved – this can be said today! Data analysis on the ground station in would add another few hours, and then any messages would have to go to Copenhagen via slow telephone systems under tight security control. From Copenhagen things would go easier, but the rescheduling of a Schmidt telescope observation plan to make room for the seventeenth fruitless attempt to find a GRB counterpart would at that time not make many astronomers jump in excitement to their feet. 12 GRBs were localized during EURECAs 11 month lifetime (Brandt et al. 1995). The communications with EURECA WATCH was easier, and the orbital period was shorter, only 90 minutes. But the ground data transfer and analysis would still add a couple of hours. And the scheduling bottleneck at the ground observatories remained. It must be remembered that in 1992 nobody could know what to expect and what to look for. One important result which came out of the WATCH GRB positions was that it allowed an independent test of the positions coming from the Interplanetary Network and later the BATSE experiment, as illustrated in Figure 5 for GRB 921022. WATCH also detected some unusual bursts, the light curve of GRB 921022 (Fig. 6), is a fascinating example of a single burst exhibiting simultaneously the characteristics of a short and a long GRB. The initial spike in this burst was resolved in the raw detector data with a FWHM duration of about 10 ms, whereas the second pulse evolves over a time scale about 2000 times longer than that of the spike. N. Lund: Early Danish GRB Experiments 21

Fig. 5. IPN (U), BATSE (B) and WATCH Fig. 6. Top: full light curve, 0.9 s/bin. (W) localizations of GRB 901022. Bottom: initial spike, 0.0035 s/bin.

4.2 X-ray Novae

WATCH made real headlines (even in the newspaper Izvesti )R with the dis- covery and localization of the X-ray transient Nova Musca 1991 (Lund & Brandt 1991). This transient was later observed in detail with the Sigma telescope on Granat and in a brief episode a transient line was observed near 500 keV (Gilfanov et al. 1991). Another X-ray nova highlight is GRO J0411+22 (Nova Persei 1992). This transient was discovered by the BATSE experiment on GRO during a period where Granat was off. But as soon as the Granat observations were resumed the source was detected and localized accurately. Armed with the WATCH posi- tion our conference host, Alberto, went off from Evpatoria to the nearby Crimean Astrophysical Observatory and there he managed against all odds, with a small telescope – but some very helpful local astronomers – to identify the optical coun- terpart! (Castro-Tirado et al. 1992a).

4.3 GRS 1915+105

After successfully identifying the optical counterpart of the Persei nova Alberto returned to Evpatoria, and immediately noted another transient: GRS 1915+105! (Castro-Tirado et al. 1992b). Off again to the observatory, but this time: no luck! GRS 1915+105 was not visible in the optical. But apart from that, GRS 1915+105 is probably the most significant discovery from the WATCH instruments. It was soon found to be a radio source and it was the first Galactic source observed to exhibit the characteristics of a (Mirabel et al. 1993). 22 Gamma-ray Bursts: 15 Years of GRB Afterglows

4.4 Solar flare observations One of the WATCH units on Granat was always observing in the direction of the . Normally the Sun was completely invisible (the lower energy threshold of WATCH was about 6 keV). But when a flare erupts the Sun becomes bright – sometimes even extremely bright. It so happened that Granat was launched only one month after the demise of the Solar Max Satellite – and the next Solar observatory was only launched in 1994. So the WATCH data came as a very welcome bridging data set between these two Solar observatories (Crosby et al. 1998). 5 A WATCH concept for the future?

It is now more than 30 years since the WATCH concept was developed and it may be reasonable to question whether the instrument today can be of more than historic interest? But I think it is, and I shall describe why. First of all, there is a continuing interest in keeping an eye on the transient X-ray sky. In the 50 years of X-ray astronomy nature has entertained us with ever new forms of variable X-ray sources, and it would not be wise to suppose that by now we have seen all. In particular because the understanding of even the well known transients is still incomplete – not the least in the case of gamma-ray bursts.

Fig. 7. Original WATCH detector. But why should we revert to the old-fashioned technique of the rotation mod- ulation collimator when so much more advanced and powerful techniques like the coded mask telescopes have been developed? I will argue that we should revive the RMC technique considering three facts: • The RMC, technically and data-wise, is much simpler than the coded mask. • The RMC can be installed on a simple, spinning satellite platform. • A wonderful new detector, exactly matching the RMC, has been developed. N. Lund: Early Danish GRB Experiments 23

Therefore, if your wish is to produce a (cluster of) low cost satellite(s) to keep an eye on the X-ray sky, the RMC is the best choice. I have already mentioned that WATCH on both Granat and EURECA was working on a very meagre telemetry budget, 37 bits/s per WATCH unit for Granat and 120 bits/s for EURECA. I should hope that by today we can do better even on a 1 kg Cubesat, but if we need to survive on 100 bits/s we know it can be done. The for a spinning satellite is much simpler than for a 3-axis stabilized satellite, both regarding the sensors and the actuators. Considering the on-board software we can now do much better than we did in 1989, we will be able to make a better selection of the important data and localize on-board a much larger fraction than we did 20 years ago. And today the ground based astronomers know what to look for, and they are aware of the importance of being fast! But the real crux is the emergence of the Silicon Drift Detectors, the SDDs (Vacchi et al. 1991; Rashevsky et al. 2002). The SDDs are now vigorously being developed and space qualified as X-ray detectors for the LOFT mission – one of the candidates for the M3-slot in the ESA programme (Feroci et al. 2012). Their key advantages for WATCH is the low weight and the low energy threshold, 2 keV, giving the new instrument access to a much richer sky – and provide better opportunities for analysis of emission temperatures and radiation transfer parameters of the sources. The SDD is an ideal match for an RMC. Through the string of point-like anodes (Fig. 8) which assures the very low read-out noise of the SDD this detector is born with one dimensional position sensitivity. This comes “for free”, and this exactly what is needed for an RMC.

Fig. 8. The electrical structure and working principle of the Silicon Drift Detector. Note the string of small anode pads on the left edge providing 1-D position sensitivity.

Table 1 compares the original WATCH parameters with the parameters for a “CubeWATCH”, a minimal proof of concept instrument matched to a 10 × 10 × 10 cm3 cubesat. Initial studies have been carried out at DTU Space which have demonstrated that the required attitude control stability can be achieved with 24 Gamma-ray Bursts: 15 Years of GRB Afterglows

Table 1. Main characteristics of original and modern WATCH instruments.

Original WATCH CubeWATCH Energy range 6to80keV 2to20keV Energy resolution 25% @ 60 keV 1% at 6 keV Field of view 3steradian 3steradian Localization error 0.3◦ 0.1◦ Point spread function 5◦ 2◦ Sensitive area (through mask) 45 cm2 25 cm2 Detector NaI(Tl)/CsI(Na) Silicon Drift scintillator mosaic RMC spin rate 60 rpm 60 rpm Burst-trigger 6to10keVNaI 2to5keV energy bands 10 to 60 keV NaI and CsI 5to20keV Time resolution 4 ms (1/256 spin period) 4 ms (1/256 spin period) Burst data 4 s (spin periods) 1 s (spin period) Count rates (2 ch.) 256 s (spin periods) 256 s (spin periods) 2 channel data 4096 s (spin periods) 2048 s (spin periods) 16 channel data Background count rate 400 cps ∼1000 cps Data rate 37 bits/s (Granat) 200 bits/s 120 bits/s (EURECA) Weight 11 kg 0.4 kg Power 12 W 0.3 W Dimensions 27 × 28 × 29 cm3 9.5 × 9.5 × 3cm3 existing sensors and actuators for a fast spinning cubesat. A project is underway to develop a special SDD matched to the specific cubesat requirements.

6 Discussion

The WATCH wide field monitor was originally developed to identify the sources of the cosmic gamma ray bursts. Although the instrument performed according to expectations on two space missions the primary goal was not achieved. With present day knowledge about the GRB optical afterglows we can see that the limited precision of the burst localizations and the limited possibilities at the time for very rapid follow-up from ground based telescopes with adequate sensitivity conspired to make a successful identification of a gamma-burst host from WATCH data very unlikely. However WATCH performed very well as an all-sky monitor, and particularly for the Granat mission this allowed to make important discoveries using the pointed instruments. The usefulness of permanently keeping the sky under surveillance to detect new or rare phenomena has been demonstrated time and again in X-ray astronomy. N. Lund: Early Danish GRB Experiments 25

The RMC technology offers a relatively simple way to realize this goal and the recent development of X-ray sensitive Silicon Drift Detectors opens the possibility to build ultra light versions of WATCH (<0.5 kg) with a sensitivity fully matching the original units. If realized this would allow to launch an all-sky monitoring swarm of minisatellites for a moderate cost.

The original development of WATCH was supported by the Danish Natural Sciences Research Council and by the Danish Space Board. The author gratefully acknowledges the support from the University of Granada to participate in the conference.

References

Brandt, S., Lund, N., & Rao, R.A., 1990, ICRC, 4, 145B Brandt, S., Lund, N., & Castro-Tirado, A.J., 1995, AdSpR, 16, 43 Castro-Tirado, A.J., Brandt, S., Lund, N., et al., 1992a, IAUC 5587 Castro-Tirado, A.J., Pavlenko, P., Salyapikov, A., et al., 1992b, IAUC 5588 Crosby, N., Vilmer, N., Lund, N., & Sunyaev, R., 1998, A&A, 334, 299 Evans, W.D,. Klebesadel, R.W., Laros, J.G., et al., 1980, ApJ, 237, L7 Feroci, M., Stella, L., van der Klis, M., et al., 2012, ExpA, 34, 415 Gilfanov, M., Sunyaev, R., Churazov, E., et al., 1991, SovAL, 17, 437 Lund, N., 1981, Ap&SS, 75, 145 Lund, N., 1986, ed. Culhane J.L., SPIE Proc., 597, 95 Lund, N., & Brandt, S., 1991, IAUC 5161 Mertz, L., 1967, Modern Optics (New York, Brooklyn Polytechnic Press), 787 Mirabel, I.F., Rodriguez, L.F., Marti, J., et al., 1993, IAUC 5773 Nishimura, J., Fujii, M., Tawara, Y., et al., 1978, Nature, 272, 337 Rashevsky, A., Bonvicini, V., Burger, P., et al., 2002, NIM, A485, 54 Sazonov, S.Y., Sunyaev, R.A., Terekhov, O.V., et al., 1998, A&AS, 129, 1 Vacchi, A., Castoldi, A., Chinnici, S., et al., 1991, NIM, A306, 187