AN OSSE SEARCH FOR THE BINARY RADIO PULSAR 1259-63 1 2 2 1 3 P. S. Ray , J. E. Grove , J. D. Kurfess , T. A. Prince , M. P. Ulmer 1 Division of Physics, Mathematics,andAstronomy, Caltech, Pasadena, CA 91125 2 E. O. Hulburt Center for Space Research, Naval Research Laboratory, Washington, DC 20375 3 Northwestern University, Evanston, IL 60208 ABSTRACT We have searched data from the Oriented Scintillation Spectrometer Exper- iment (OSSE) on the Compton Gamma Ray Observatory (GRO) for evidence of low-energy (cid:13)-ray emission from the binary radio pulsar PSR1259-63. This 47ms pulsar is in a long-period, highly eccentric orbit around a Be stellar companion and was observed by OSSE approximately 400 days after periastron. The pe- riod derivative allowed by the published radio ephemeris (Johnston et al. 1992) suggests that the pulsar might be relatively young, and therefore a (cid:13)-ray source. However, the ephemeris is not su(cid:14)ciently accurate to allow the traditional epoch- folding technique over the full OSSE observation. Instead, the OSSE data were analyzed using Fourier transform spectral techniques after applying trial accel- erations to correct for a range of possible orbital accelerations. We searched 48 29 6 (cid:24) accelerations; each FFT was 2 points sampled at 2ms, spanning 10 seconds of observation time. There was no evidence of pulsed emission in the 64{150 keV (cid:0)3 (cid:0)2 (cid:0)1 (cid:0)1 (cid:2) band, with a 99.9% con(cid:12)dence upper limit of 6 10 photonscm s MeV (cid:24) or 40 mCrab pulsars, which suggeststhat the pulsar’s intrinsic period derivative is small and its magnetic (cid:12)eld weak. This work was performed on the Concurrent Supercomputing Consortium’s Intel Touchstone Delta parallel supercomputer as part of a GRO Phase 1 Guest Investigation. 1. BACKGROUND ON PSR1259-63 PSR1259-63 is a recently discovered radio pulsar (Johnston et al. 1992) and is the only radio pulsar known to have a massive, non-degenerate companion. The stellar companion has been optically identi(cid:12)ed as the Be star SS 2883. The pulsar is in a highly eccentric orbit and has gone through only one periastron passagesince its discovery. As a consequence, the published timing solution could not completely discriminate between orbital Doppler shift-induced period changes _ and the intrinsic period derivative (Pint) of the pulsar. The best-(cid:12)t solution gives (cid:0)14 (cid:0)14 _ (cid:2) _ _ (cid:24) Pint = 2 10 , but even Pint = 0 cannot be excluded. An intrinsic Pint 10 implies a young age for the pulsar. With pulse period and distance comparable _ to those of the Crab pulsar, such a large Pint would make it a likely candidate for pulsed (cid:13)-ray emission. Table 1, reproduced from Johnston et al. (1992), lists the (cid:12)tted pulsar and orbital parameters with two di(cid:11)erent assumed values for P_int. The ambiguity in _ Pint will be removed by the time the pulsar reaches periastron again since the orbital period will be determined. Report Documentation Page Form Approved OMB No. 0704-0188 Public reporting burden for the collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to a penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number. 1. REPORT DATE 3. DATES COVERED 1993 2. REPORT TYPE 00-00-1993 to 00-00-1993 4. TITLE AND SUBTITLE 5a. CONTRACT NUMBER An OSSE Search for the Binary Radio Pulsar 1259-63 5b. GRANT NUMBER 5c. PROGRAM ELEMENT NUMBER 6. AUTHOR(S) 5d. PROJECT NUMBER 5e. TASK NUMBER 5f. WORK UNIT NUMBER 7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING ORGANIZATION Naval Research Laboratory,E.O. Hulburt Center for Space REPORT NUMBER Research,4555 Overlook Avenue, SW,Washington,DC,20375 9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSOR/MONITOR’S ACRONYM(S) 11. SPONSOR/MONITOR’S REPORT NUMBER(S) 12. DISTRIBUTION/AVAILABILITY STATEMENT Approved for public release; distribution unlimited 13. SUPPLEMENTARY NOTES 14. ABSTRACT 15. SUBJECT TERMS 16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF 18. NUMBER 19a. NAME OF ABSTRACT OF PAGES RESPONSIBLE PERSON a. REPORT b. ABSTRACT c. THIS PAGE 5 unclassified unclassified unclassified Standard Form 298 (Rev. 8-98) Prescribed by ANSI Std Z39-18 Table 1 Parameters of PSR1259-63 h m s (cid:6) s Right acension (J2000) 13 02 47:72 0:03 (cid:14) 0 00 00 (cid:0) (cid:6) Declination (J2000) 63 5008:5 0:2 (cid:0)3 (cid:6) Dispersion measure (pccm ) 146:7 0:2 Distance (kpc) 2.3 Pulse period (ms) 47.76164 47.76219 (cid:0)14 (cid:2) Period derivative 0.0 2 10 (cid:6) (cid:6) Orbital period (d) 1133 24 2150 100 (cid:6) (cid:6) Epoch of ascending node (MJD) 48043 2 48027 3 (cid:6) (cid:6) Epoch of periastron (MJD) 48120 2 48117 3 (cid:6) (cid:6) asini (ls) 3480 1900 3450 1000 (cid:6) (cid:6) Longitude of periastron (deg) 164 9 158 6 (cid:6) (cid:6) Eccentricity 0:976 0:025 0:967 0:017 Mass function (M(cid:12)) 35 10 (from Johnston et al. 1992) 2. DETAILS OF THE OBSERVATION The characteristics and performance of the OSSE instrument have been de- scribed by Johnson et al. (1993). The instrument consists of four identical (cid:14) (cid:14) (cid:2) large-area NaI(Tl){CsI(Na) phoswich detector systems, each with a 3:8 11:4 (FWHM) (cid:12)eld-of-view de(cid:12)ned by a tungsten collimator. The pulsar was observed serendipitously in OSSE’s NMus91/Galactic plane pointing in September 1991, spanning MJD 48518{48532, about 400 days after periastron andimmediately followingthe epochstudied by Johnston et al. (1992). (cid:14) (cid:24) The pulsar was about 8.2 o(cid:11)-axis, which reduced the detector response to 25% of the on-axis value. The OSSE observation strategy is implemented by alternately pointing each detector at source and background positions on a time scale (131 seconds) that is short with respect to typical Earth orbital background variations. The di(cid:11)erence spectrum derived from these source and background pointings for the PSR1259- 63 (cid:12)eld shows a signi(cid:12)cant soft-spectrum excess, presumably due to the sum of di(cid:11)use emission from the plane and several point sources. Temporal signatures can be used to resolve the ambiguity over the source of this excess. In addition to the spectral data, time-tagged data (resolution = 0:125 ms) 5 (cid:24) (cid:2) were collected in the 64{150 keV band, with a total livetime of 3:3 10 detector-seconds. These data are the subject of the search described here. 3. SEARCH ALGORITHM In order to search for pulsed (cid:13)-ray emission from PSR1259-63, we generated a binned time series from the OSSE data to which we applied our standard radio pulsar search algorithms (see Anderson et al. 1990). This was accomplished by (cid:12)rst converting the spacecraft photon arrival times to Solar System barycentric arrival times using the known position of the pulsar. These arrival times were 29 then used to generate a 2 point time series consisting of 2 ms bins. This yields a Nyquistfrequencyof250Hz,sothepowerspectrumcontainsthe(cid:12)rst11harmonics of the 47 ms pulsar without aliasing. This search utilized large 1-dimensional Fast Fourier Transforms (FFTs) to 29 search for periodic signals in the data. Because the frequency bins in a 2 point time series are very narrow, a signal must have a constant frequency to within a very narrow tolerance to keep from being smeared over multiple bins in the spectral domain with a concomitant loss in sensitivity. For a pulse period of 6 (cid:24) 50 ms, the largest period derivative that can be searched over a 10 s interval (cid:0)16 _ (cid:24) _ is P 10 . We correct for an assumed P by applying a quadratic stretch to the time series. A quadratic stretch is an adequate model of either linear spin-down of the pulsar or Doppler shifts over a short portion of a binary orbit. Since the published ephemeris was not accurate enough for us to know exactly what trial \acceleration" to use, we estimated an envelope of possible total period derivatives (intrinsic and orbital) from the trend of P with time from Figure 2 of Johnston et al. (1992). We then selected 48 trial accelerations spanning the (cid:0)14 (cid:0)13 (cid:2) _ (cid:2) range 1:8 10 < P < 1:5 10 . The accelerations were spaced such that the di(cid:11)erenceincorrectedarrivaltimeswas< 10msanywhereinthetimeseries. Thus the maximum residual pulse smearing was < 1=4 period, and up to 4 harmonics could have been signi(cid:12)cant. The complete search was performed in about 1.5 hours on 512 nodes of the Concurrent Supercomputing Consortium’s (CSCC) Intel TouchstoneDelta Super- computer at Caltech. The unstretched, barycentered time series was loaded onto theDelta’s90GBConcurrent FileSystempackedat1byteper2msbin, using1/2 GB of storage. Since the Delta has more than 4 GB of usable memory available after the OS and programs have been loaded, we could store both the unstretched time series and one trial stretched time series. So, the time series need only be read o(cid:11) disk once at the beginning of the run. The analysis then proceeds by making a new stretched copy of the time series in memory, performing a parallel FFT in place, generating a power spectrum and searching for signi(cid:12)cant peaks in the region of the spectrum where the pulsar should appear. Harmonic folds of 2, 4, and 8 harmonics were also used to search for the pulsar signal. This process is repeated for each acceleration trial. 12 (cid:2) The 48 trials consisted of about 4 10 (cid:13)oating point operations performed at an average rate of about 1 GFLOP/s by the Delta. This analysis would have been considerably more cumbersome and much slower on a machine that had less than 2 GB of usable main memory, since the running time would be dominated by the tremendous amount of disk access needed to do out-of-core FFTs. Figure 1: Plot of delay in samples vs. sample number for the 48 acceleration trials used in the search. 4. SENSITIVITY AND RESULT OF THE SEARCH Sensitivity was assessed by Monte Carlo simulation. To the actual data stream, we added counts from a simulated pulsar with pulse period of 50 ms and pulse FWHM = 15% at various intensities. The pulsed (cid:13)ux corresponding to the 99.9% con(cid:12)dence upper limit was then simply interpolated from the observed powers. No statistically signi(cid:12)cant powers were seen in any of the 48 acceleration (cid:24) trials. The resulting upper limit is 40 mCrab pulsar (cid:13)ux units (see Table 2). For comparison, the sensitivity of an epoch-folding search with known frequency (cid:24) and phase would be 20 mCrab pulsar (cid:13)ux units, for a pulse FWHM = 25%. Table 2 Search Sensitivity 2 Nfreq Naccel (cid:1)E (keV) Aeff (cm ) Flux limit (99.9% conf) (cid:0)3 (cid:0)2 (cid:0)1 (cid:0)1 (cid:2) (cid:2) 48 75000 48 64{150 400 < 6 10 phcm s MeV (cid:24) 40 mCrab pulsars 5. CONCLUSIONS The known radio pulsars with signi(cid:12)cant low-energy (cid:13)-ray emission (Crab, Vela, PSR1509-58) are young, with reasonably short periods and large intrin- sic period derivatives. By virtue of the similarity of the period and distance of PSR1259-63tothe Crab, and the possibilitythat its period derivative isalsocom- parable to that of the Crab, PSR1259-63 is a likely candidate for (cid:13)-ray emission. < We have placed a limit on the 64{150 keV (cid:13)ux from PSR1259-63 of F (cid:24) 40 (cid:24) mCrab pulsar (cid:13)ux units, which corresponds to a luminosity of 5% of the Crab pulsarinthesameband. Thissuggeststhatthepulsar’sintrinsicperiodderivative, andtherefore itsmagnetic(cid:12)eld, aresubstantially smallerthanthe Crab’s. Further _ radio observations will produce a de(cid:12)nitive value of Pint by the time of the next periastron. This observational limit lends support to the suggestion that this pulsar may 4 (cid:24) not be a young pulsar within 10 years of its birth, but an older pulsar whose short period is due to spin-up by accreting matter from its companion at each periastron passage. Since it is widely believed that neutron star magnetic (cid:12)elds decay during accretion, PSR1259-63 would then be expected to have a weak magnetic (cid:12)eld and no signi(cid:12)cant (cid:13)-ray emission. If this scenario is the case, then the pulsar must be experiencing substantial accretionepisodes,andatthenextperiastronpassageonewouldexpectthematter being accreted to eclipse the pulsar’s radioemission, while making it an X-ray and (cid:13)-ray source. So, searches for transient hard X-ray emission near the time of the next periastron passage, including a searches for pulsed emission, are warranted. REFERENCES Anderson, S.B., et al. 1990, Nature, 346, 42. Johnson, W.N., et al. 1993, accepted for publ. in Ap. J. Suppl. Johnston, S., et al. 1992, Ap. J. Lett., 387, L37.