A 23 GHz Survey of GRB Error Boxes J. N. Hewitt1, C. A. Katz1, S. D. Barthelmy2, W. H. Baumgartner1, T. L. Cline2, B. E. Corey3, G. J. Fishman4, N. Gehrels2, K. C. Hurley5, C. Kouveliotou4, C. A. Meegan4, C. B. Moore1, R. E. Rutledge6, and C. S. Trotter1 1 Department of Physics and Research Laboratory of Electronics, Massachusetts Institute of Technology 2 NASA Goddard Space Flight Center 3 Haystack Observatory 4 NASA Marshall Space Flight Center 5 Space Sciences Laboratory, University of California, Berkeley 6 6 Department of Physics and Center for Space Research, Massachusetts Institute of 9 Technology 9 1 n TheHaystack37-metertelescopewasusedinapilotprojectinMay a J 1995 to observe GRB error boxes at 23 GHz. Seven BATSE error boxesandtwoIPNarcswerescannedbydrivingthebeamofthetele- 0 scoperapidly across theirarea. FortheBATSEerrorboxes,theradio 1 observations took place two to eighteen days after the BATSE detec- 1 tion, and several boxes were observed more than once. Total power v data were recorded continuously as the telescope was driven at a rate 6 of 0.2 degrees/second, yielding Nyquistsampling of thebeam with an 4 integration timeof50milliseconds, correspondingtoatheoreticalrms 0 sensitivityof0.5Jy. Underconditionsofgoodweather,thissensitivity 1 wasachieved. Inapreliminaryanalysisofthedatawedetectonlytwo 0 sources,3C273and0552+398,bothcataloguedsourcesthatareknown 6 tobevariableat23GHz. Neitherhadafluxdensitythatwasunusally 9 high or low at the time of our observations. / h p - o INTRODUCTION r t s The detection of radio counterparts to gamma-ray bursts would provide a : important information on the source emission mechanism, would be likely to v produce more accurate position determinations, and would give a measure of i X the source distance through a measurement of the dispersion delay (Palmer r 1993). There is, therefore, considerable interest in detecting radio counter- a partstogammaraybursts,andseveralsearchesareunderway. TheHaystack 1 telescope can make a unique contribution because of its high frequency ca- pabilities and its ability to slew rapidly. 1Radio astronomy at Haystack Observatory of the Northeast Radio Observatory Consortium (NEROC)is supported bythe National Science Foundation. (cid:13)c 1993 American Institute of Physics 1 2 The Haystack telescope is optimized for operationabove 1 GHz; a recently completed upgrade (Barvainis et al. 1993) has extended operations to fre- quencies as high as 116 GHz. The telescope was originally designed to track objects in low earthorbit,and highslew speeds canbe tolerated. Any source positionabovethehorizoncanbe reachedfromzenithwithin30seconds,and the complementary goals of high resolution and wide field of view may be achieved through rapid scanning of an area of the sky. OBSERVATIONS Rapid continuum mapping of large fields with the Haystack telescope had not been previously attempted, and a two-weekperiod in May 1995was allo- cated to us to develop the observing techniques. At the observing frequency of 23 GHz, the beam is 1.4 arc minutes and the gain is 0.14 K/Jy. The ob- serving frequency was chosen as a compromise between the desire to observe at as high a frequency as possible and a desire to produce a map within a reasonable period of time. The bandwidth was 160 MHz and the nominal system temperature was 140 K; the actual system temperature depended on the elevation and weather conditions (see below). Error boxes were scanned at 0.2◦/second in right ascension, stepping in increments of 0.00625◦ in dec- lination. Continuum data were sampled at 20 Hz, yielding Nyquist sampling inthe rightascensiondirectionandtwiceNyquistsamplinginthe declination direction. Oversampling in declination was necessary because with the rapid scanning the telescope failed to settle to its commanded declination before beginning to scan in right ascension. SevenBATSEerrorboxesandtwoIPNarcswereobserved;theobservations aresummarizedinTable1. Eacherrorboxwasscannedinseveral“slices;”and in most cases each slice was observed more than once. Observing conditions varied considerably with source elevation and as the weather changed. For presentation in Table 1 we selected just one observation of each slice; the observation selected was that with the best sensitivity. Thedataweredisplayedandinspectedinreal-timeduringtheobservations and we believe we would have noted any source with flux density exceeding ten times the rms sensitivity listed in Table 1. This detection limit must be correctedfortheweather-dependentopacityat23GHz;forthedatapresented here the correction is typically about 15%. Asanexampleofthe qualityofthe dataproducedinthesurveywepresent in Figure 1 our map of part of the error box of BATSE #3598. This map was produced by fitting a linear baseline to five-second segments of data and subtracting the fit from the central three seconds of data. The measured pixel-to-pixel rms was 0.49 Jy, consistent with the expected value of 0.51 Jy (see Table 1). Regions in which the fit was poor are evident in the map as horizontal “streaks” over a portion of a single row; occasional unsampled pixels can also be seen. Sources (here, 3C273) appear as bright regions with 3 TABLE 1. Journal of Observations Date and Date of Theoretical rms Time of Radio Sensitivity Size of Burst (UT) Observations (UT) (Jy) Map IPN3509 1 95/04/16 95/05/24 0.52 4.4◦ 0.6◦ ◦× ◦ IPN3509 2 13:26:59.76 95/05/22 0.52 4.3 0.8 ◦× ◦ IPN3509 3 95/05/22 0.52 3.7 1.0 IPN3509 4 95/05/22 0.53 2.6◦×1.0◦ × ◦ ◦ IPN3512 1 95/04/18 95/05/21 0.58 1.0 1.5 × IPN3512 2 23:16:35.63 95/05/21 0.58 IPN3512 3 95/05/21 0.58 IPN3512 4 95/05/21 0.57 IPN3512 5 95/05/21 0.59 IPN3512 6 95/05/22 0.51 IPN3512 7 95/05/22 0.53 IPN3512 8 95/05/22 0.56 IPN3512 9 Not observed IPN3512 10 95/05/22 0.58 IPN3512 11 95/05/21 0.58 IPN3512 10 95/05/21 0.58 ◦ ◦ BATS3552 1 95/05/06 95/05/27 0.51 5.7 1.3 × BATS3552 2 23:25:12.75 95/05/27 0.50 BATS3552 3 95/05/28 0.52 BATS3552 4 95/05/28 0.51 BATS3552 5 95/05/28 0.50 BATS3567 1 95/05/09 95/05/24 0.79 6.4◦ 1.1◦ × BATS3567 2 23:16:05.93 95/05/24 0.62 BATS3567 3 95/05/24 0.58 BATS3567 4 95/05/23 0.53 BATS3567 5 95/05/24 0.57 BATS3567 6 95/05/23 0.49 ◦ ◦ BATS3588 1 95/05/21 95/05/23 0.68 4.7 0.9 × BATS3588 2 06:59:17.29 95/05/23 0.65 BATS3588 3 95/05/23 0.68 BATS3588 4 95/05/23 0.65 BATS3588 5 95/05/23 0.67 BATS3588 6 95/05/23 0.65 ◦ ◦ BATS3593 1 95/05/22 95/05/26 0.52 3.4 0.85 × BATS3593 2 23:41:23.18 95/05/26 0.50 BATS3593 3 95/05/31 0.53 BATS3593 4 95/05/30 0.56 4 TABLE 1. continued BATS3594 1 95/05/23 95/05/27 0.53 6.6◦ 1.1◦ × BATS3594 2 05:39:29.07 95/05/27 0.54 BATS3594 3 95/05/27 0.50 BATS3594 4 95/05/28 0.65 BATS3594 5 95/05/27 0.50 BATS3594 6 95/05/28 0.52 BATS3598 1 95/05/24 95/05/26 0.55 4.2◦ 1.1◦ × BATS3598 2 04:08:19.12 95/05/26 0.51 BATS3598 3 95/05/27 0.52 BATS3598 4 95/05/27 0.55 ThetheoreticalrmssensitivityisgivenbyTsys/(√∆ντG)whereTsysisthemeasured system temperature at the time of observation, ∆ν =160 MHz, τ = 50 msec, and G=0.14 K/Jy. the extent of the beam. Further data analysis is in progress. RESULTS Inthe preliminaryanalysisofthe dataduringtheobservations,only3C273 wasdetected;0552+398wasnotedwhenweexaminedourmapsforpreviously known, catalogued sources. No other source was detected in the maps sum- marizedinTable1. 3C273isintheerrorboxofBATSE#3598(seeFigure1); 0552+398 is in the error box of BATSE#3594. The data analysis is still in progress;careful examination of all radio maps and averagingof the data for boxes scanned more than once may yield further detections. Itispossiblethatradio-loudAGNarethesourcesofthegammaraybursts. However, both 3C273 and 0552+398 are already known to be variable and neither had a flux density that was unusually high or low during the time of our observations. Therefore,there is no compelling reasonto associate either with a GRB. FUTURE PROSPECTS Now that the continuum mapping technique with the Haystack telescope has been demonstrated, further observations of GRB error boxes have been proposed. During this pastsummer,the telescopeservosystemwasreplaced. With the improved pointing control more rapid scanning of boxes should be possible, making feasible observations at higher frequencies. A new receiver 5 1.5 1.0 Kilo ARC SEC 00..50 -0.5 -1.0 -1.0 -0.5 0.0DEGREE0S.5 1.0 1.5 FIG. 1. Contour plot of map of part of the error box of BATSE#3598. The contoursare-3,3,4, 5,6,7,8, 9,10,11, 12, 13,14, 15,16, 17, 18,19, and20times the root-mean-square value of the map of 0.49 Jy. The source visible in the upper right corner is 3C273. will provide lower receiver noise at 23 GHz and allow us to observe farther from the 22 GHz water line, further increasing the sensitivity. REFERENCES 1. Barvainis, R. E., Ball, J. A., Ingalls, R. P., and Salah, J. E., P.A.S.P. 105, 1334 (1993). 2. Palmer, D. M., Ap.J. (Lett.), 417, L25 (1993).