Branch Organization

Major Projects

Annihilation Fountain Images

Recent papers and publications

Web Links

Burst Locations with an ArcSecond Telescope (BLAST)

Introduction Blast Team Members Gamma Ray Bursts Hard X-ray Sky Survey BLAST Mission Concept Precise Location of Cosmological Gamma Ray Bursts

Introduction

Blast Team Members

Co-Investigators:

NRL:

C. Dermer, J. Eric Grove, Paul Hertz, Robert Kinzer, Richard Kroeger, James D. Kurfess, Mike Lovellette, Gerry Share, Mark S. Strickman, Kent Wood

USRA:

Sue Inderhees, Bernard Phlips

Clemson University:

Dieter Hartmann, Mark D. Leising

MSFC:

G. Fishman, C. Meegan

LANL:

E. E. Fenimore

Gamma Ray Bursts

By default, the cosmological model (e.g., Prilutskii & Usov 1975; Paczynski 1986; Paczynski 1995) is favored because burst directions are necessarily isotropic for sources at redshifts z ~ 1 where the Hubble flow flattens the size distribution from isotropy. An extended halo or corona of neutron stars in the Milky Way is, however, also consistent with the GRB statistics if only high-velocity neutron stars burst after some delayed turn-on (e.g., Shklovskii & Mitrofanov 1986; Li & Dermer 1992; Lamb 1995). The halo model has received additional support from the identification of a population of fast radio pulsars (Frail et al. 1994; Lyne & Lorimer 1994) which are not confined by the Galaxy's gravitational field. This leaves an uncertainty in GRB distances by a factor ~ 3 x 10e4 or GRB luminosities by a factor of ~10e9, even supposing that we can exclude a local Solar system origin, which may not be justified from the evidence!

Discovering the nature of GRBs is important because, if coronal, we are directly probing the late evolution of isolated pulsars and the processes which generate brief energetic gamma-ray events in compact objects, and we are indirectly probing the physical processes in supernovae which impart high velocities to neutron stars. If cosmological, we are studying sources with the greatest luminosities and the highest energy densities known. Such sources are intrinsically important as laboratories in an extreme regime of physics. If due to coalescing neutron stars (Paczynski 1986; Eichler et al. 1989; Narayan et al. 1992), cosmological GRBs give a measure of the rate of massive binary stellar formation and merging. Moreover, if the merger scenario is correct, then we are dealing with intense sources of neutrino and gravitational radiation (~10e53 ergs per GRB in each channel) in an environment so compact that neutrino-antineutrino interactions occur. The GRB itself signals the radiative consequences of an electron-positron fireball blast wave expanding into its surrounding medium (e.g., Rees & Meszaros 1992). The identification of counterparts to GRBs is generally recognized as the key to solving the GRB mystery. The slow progress toward this goal is due to the relatively poor imaging capability of GRB detectors. This also contributes to the ambiguity in determining whether GRBs do or do not recur. The rapid response (15-30 second) BACODINE system gives burst locations with uncertainties of ~6 degrees for most GRBs, and uncertainties as small as ~2 degrees for rare bright GRBs that occur about once per month over the whole sky (Barthelmy et al. 1994). The Gamma-Ray Optical Counterpart Search (GROCSE; Akerlof et al. 1994) and the Explosive Transient Camera (ETC; Krimm et al. 1994) take advantage of these real-time coordinates to perform an optical survey, although only to limiting magnitudes of 8 and 10, respectively. The BATSE/COMPTEL/NMSU provides GRB coordinates with uncertainties as small as ~1 degree within hours of bright GRBs falling within the CGRO's Compton Telescope field of view (Harrison et al. 1995).

How small must an error box be to be meaningful? The possibility that cosmological counterparts are in some sense sub-luminous galaxies must be taken into account. Tyson (1988) determined the surface density of stars and galaxies to magnitudes brighter than R ~ 27. At this level he detected about 5 sources per 100 arcsec^2. It is clear that a unique source identification requires position accuracies of a few arcseconds. The BLAST mission will provide 1 arcsec imaging for ~10 GRBs per year, and 5 arcsecond imaging for ~35 GRBs per year. This will permit deep optical and radio follow-up observations within a fraction of a day following the event. If GRBs are associated with neutron stars in an extended galactic halo, they will probably not be detectable even with an accurate position. Indeed empty error boxes, when sufficiently small, become an argument in favor of halo models. If, on the other hand, GRBs are associated with merging compact objects at cosmological distances, then we would expect to find host galaxies inside a fair fraction of the error boxes. While galaxies are present in several small IPN error boxes (e.g., Vrba et al. 1995), not all of them contain bright galaxies (Schaefer 1992; Fenimore et al. 1993). One year of BLAST observations with deep follow-up optical and radio observations should clearly discriminate between galactic halo and cosmological models.

However, the discovery of an enhancement in the rate of GRBs towards M31 over the average rate measured in directions away from M31 is considered the definitive experiment to validate the galactic neutron star halo model of GRBs (e.g., Atteia & Hurley 1986; Liang 1991; Li & Liang 1992, Lamb 1995; Paczynski 1995). At the same time, we must recognize that a negative result is much less meaningful, no matter how sensitive the detector (Lamb 1995). This is because GRB emission may be beamed along the direction of motion of the neutron star, insofar as this property offers a natural solution to the delayed turn-on of GRBs (Li et al. 1994). If the emission is beamed opposite to the direction of motion of the neutron star, then we should not expect to see an enhancement of GRBs no matter how deep the search, because the distant faint bursts the distant side of M31 would be swamped by the numerous bursts from our own galaxy for appropriate GRB luminosity functions.

The situation is better if we assume unbeamed GRB emission, in which case BLAST will be able to examine essentially all of the remaining parameter space assuming that M31's halo is similar to ours. As allowed by Hakkila et al. (1994), sampling distances <170 kpc to the faintest bursts are ruled out at the 90% confidence level for 800 BATSE GRBs, and this minimum sampling distance is growing as BATSE detects more GRBs. Detailed numerical simulations by Podsiadlowski et al. (1995) indicate that a sampling distance > 500 kpc is required to detect the anisotropy from M31 at a statistically significant (>3sigma level). Even if the galactic neutron-star halo is sharp and the roll-over in the size distribution is due entirely to the GRB luminosity function, a sensitivity increase of ~9 is therefore needed in this worst case to detect the neutron star halo from M31 at an assumed distance of 700 kpc. The BLAST M31 experiment, with a sensitivity equal to 10 times the BATSE sensitivity, therefore gives the necessary improvement to test galactic neutron star models.

Hard X-ray Sky Survey

Relatively few sources are known at these energies, as the only survey conducted in the hard X-ray band (HEAO A-4 in 1977-79) was a scanning mission with poor sensitivity and no imaging capabilities. HEAO A-4 detected ~70 sources to a survey limit of ~15 mCrab in the 13-80 keV band (Levine et al. 1984). Balloon and satellite-borne instruments have added perhaps 20-30 additional sources to our catalog of known hard X-ray sources, so that the persistent hard X-ray sky is not known much better today than the HEAO A-4 catalog. The SIGMA instrument on GRANAT imaged small areas, principally the Galactic Center region, to ~30 mCrab in the 35-100 keV band (Paul 1995), and BATSE, using occultation techniques has detected ~20 persistent hard X-ray sources and a comparable number of hard X-ray transients, also with limiting sensitivities of about 20 mCrab. The OSSE instrument on CGRO, much more sensitive than HEAO A-4 above 50 keV, has detected about 30 extragalactic sources and a similar number of galactic sources, but does not have sensitivity below 50 keV (Johnson et al. 1993). The HEXTE instrument on XTE and The hard X-ray instrument on SAX will reach ~1 mCrab in a two day observation over the 15-250 keV range, but their small fields-of-view will limit the catalog of observations to a tiny fraction of the sky.

BLAST will perform an all sky survey in hard X-rays over a one year period. The wide field-of-view, which covers ~5% of the sky at FWHM, allows two week (10e6 s) pointed observations of each survey field. Although the hard X-ray number-flux relation is poorly known (this is one of the goals of the survey), by scaling from HEAO A-4 and OSSE results we estimate that ~300-700 extragalactic sources and 200-500 galactic sources will be detected. Each source will be positioned to 1 arcmin; essentially all sources will be easily identified with known counterparts. Sources without previously catalogued counterparts may include new black hole candidates and Geminga-like pulsars.

A large fraction of the diffuse X-ray background (DXB) in the BLAST energy band is due to unresolved point sources. BLAST will resolve hundreds of Seyfert galaxies and other extragalactic sources. The direct contribution of compact sources to the DXB will be determined, and extrapolation of the number-flux relation will constrain the presence of any truly diffuse component.

Source Classes for BLAST

• Persistent Black Hole Candidates:
  With the exception of determining the dynamical mass, the best remaining discriminator between black hole and neutron star primaries may be the high energy cutoff. Sources such as Cyg X-1, 1E1740.7-2942, GX339-4, and GRS1758-258 have low state spectra characterized by power laws of ~2 and high energy cutoffs above 200 keV due to the Comptonization of soft photons by hot electrons (Sunyaev & Titarchuk 1980). Neutron star binaries have cutoffs at much lower energies (~50 keV). The BLAST survey will determine the high energy cutoff for all of the persistently bright X-ray binaries in the galaxy; a statistical study of the validity of high energy cutoff as an indicator will be conducted.

  Low Mass X-ray Binaries:

  The ~5 keV thermal emission from most low mass X-ray binaries will be detectable with BLAST. Prior observations (e.g. SIGMA, BATSE, Ginga) have shown that some X-ray bursters and binary pulsars have hard X-ray tails (e.g. Paul 1995). BLAST will provide a complete survey of the hard X-ray tails of all known X-ray binaries for the first time. The large sample of detected X-ray binaries will test the suggestion from HEAO A-4 that spectral index is correlated with luminosity (van Paradijs & van der Klis, 1994).

  Bursters and Pulsars:

  Hard X-ray phase resolved spectroscopy can be conducted on all known X-ray pulsars using the time resolved pointed modes. BLAST will detect pulsed emission from significantly more than the ~20 pulsars detected by BATSE and ~10 pulsars detected by OSSE. Unidentified EGRET sources can be searched sensitively for pulsed signals. The sensitivity of BLAST will increase the sample hard X-ray light curves and test recent theories which attempt to unify the production of pulsed emission from radio to high energy gamma-rays (e.g. Romani et al. 1995).

  Supernova Remnants:

  There are several SNRs, for instance the Vela Remnant, which are expected to show structure in hard X-rays with 1 arcmin resolution. In plerionic remnants, the extent of the synchrotron nebula can be mapped. For shell remnants, a non-thermal component in the expanding shell, perhaps due to synchrotron emission or shock processes, will be searched for.

  Active Galactic Nuclei:

  We expect BLAST will detect several hundred AGNs. This provides the basis for large numbers of population studies including the search for correlations between hard X-ray emission and other properties (IR luminous, radio emission), evolutionary studies relating the BLAST luminosity to redshift, and systematic differences in Seyfert 1, Seyfert 2, quasar, and other populations. All blazars will be studied during the all sky survey; in particular all EGRET blazars will be observed. Extrapolating from OSSE (~65% detected), we expect to detect 80-90% of EGRET blazars.

  Clusters of Galaxies:

  Hard X-ray tails have been predicted in galaxy clusters due to synchrotron emission in the cluster cores. BLAST will provide a sensitive search for such emission.

BLAST Mission Concept

BLAST can most easily be understood as two independent imaging systems in which a coded aperture telescope operating at energies >50 keV provides arcmin source localization. Such systems have been flown successfully several times from satellites (e.g. SIGMA; Paul et al, 1991) and balloons. The aperture for lower energy X-rays (<50 keV) is defined by a Phase Modulation Grid mosaic which provides complementary arcsec information. This concept has been amply validated by the recent Yohkoh mission. The detectors view the sky through the grids and mask, placed in series. The detector array consists of position-sensitive NaI(Tl) scintillation detectors which provide a position accuracy of ~2 mm. The combined grid/mask system is the most novel design feature. All of its capabilities are exercised in the positioning of gamma-ray bursts. The first mask is a modified uniformly redundant array (MURA; Gottesman & Fenimore, 1989) coded aperture with ~6 mm pixels which is opaque to the highest energy measured, 150 keV, and forms the "coded aperture telescope". The mask is positioned ~2 m above the detector plane. Since the other mask arrays are designed to be transparent above ~50 keV, the BLAST system effectively consists only of the URA mask and the detectors above that energy. The position-sensitive NaI detector array records shadows of the URA cast by the imaged field. Image reconstruction consists of standard URA inversion methods. The URA mask localizes bursts to 1 arcmin or better over the full field of view of the BLAST instrument.

Figure 1. Phase Modulation Grid imaging concept. A broad scale interference pattern (shown in the lower portion) is created by two Phase Modulation Grid planes separated by a large distance. The pitch or spacing of the upper grids differs slightly from the pitch of the lower plane. The phase of the pattern changes dramatically with a small change in incident angle of the radiation.

The remaining aperture masks are a mosaic of matched pairs of fine grids of the Phase Modulation Grids and comprise the low energy grid telescope. One of the pair is placed at the detector plane and the other is positioned with the coded aperture mask. Together, each pair of fine grids provide supplementary information, in one dimension, at finer angular scales - down to an arcsec - and are made so thin that they effectively absorb only photons below ~50 keV. The upper and lower grids of a grid pair are nearly identical, consisting of parallel ribs of width 250 mm on ~500 mm centers. The spacing or pitch of the ribs in the upper grid differs slightly from that in its associated lower grid so that across a mask segment, the number of ribs in the upper grid array differs by one from that in the lower array. This arrangement causes a point source to cast a broad shadow or fringe on the detector plane such that a 1 arcsec shift in the location of a source in a plane perpendicular to the grid openings results in a translation of the fringe pattern by ~1 cm. (See Figure 1.) Determining the phase of this fringe in one coordinate on the detector plane gives the location of the source on the sky in one direction. Centroiding on a bright gamma-ray burst will reconstruct two orthogonal strips on the sky whose widths are of the order of 1 arcsec. There are strips of this width in both the x and y directions repeating every 50 arcsecs (corresponding to the 500 mm rib spacing) but for all but the faintest bursts, only one of the intersections of these strips lies in the error region given by the high-energy MURA mask described above. Combining information from the coded aperture measurements at high-energy and the Phase Modulation grids at low-energy localizes the source to a single region of order 1 arcsec. As discussed in the later sections, this positioning accuracy is dependent on burst detection significance such that ~10 bursts per year would be positioned to 1 arcsec and ~35 per year to 10 arcsec.

Table 1.

Summary of BLAST Mission

Orbit		Altitude	550 km
		Inclination	28.5 degrees
		Lifetime	3 years

S/C pointing	Absolute	1 arcmin
		Stability	1 arcsec/sec
		Knowledge	1 arcsec

Power	        Instrument	220 W
	        S/C		160 W

Weight	        Instrument	970 kg
	        S/C		760 kg

CMD/TLM	        Telemetry	160 MByte/day
		TLM dumps	2 times per day
		Command		1 command load/day

Sensitivity	Burst sensitivity (1 sec)	0.03 ph/cm^2-s
		Positioning (1 arcsec)		10 GRB/yr @1 ph/cm^2-s
		Sky Survey (106 sec)		0.2 mCrab @ 10 keV

Field of View	~50 degrees
 

Figure 2. LogN/logP distribution for 799 BATSE bursts (Pendleton et al. 1995) and extrapolation to BLAST sensitivity of 0.03 photon/cm^2-s.

Precise Location of Cosmological Gamma Ray Bursts

Figure 3. Number of GRB/yr expected as a function of positioning accuracy. The hatched area at large positioning resolution represents the uncertainty in extrapolating BATSE logN/logP.

References