The first-known double pulsar discovered at Parkes!
About 100 of the 1500 known pulsars are members of binary systems, but only in six or seven of these is the companion star believed to be a second neutron star. Most other systems, originating from lower-mass stars, have white dwarf or low-mass helium-core companions. Up to now, searches of the double-neutron-star systems to detect the second neutron star as a pulsar have proved fruitless. This is not too surprising, since the detected pulsar has been through a "recycling" process, accreting matter from its evolving companion, which both spins it up to millisecond periods and reduces its magnetic field. As a result of these changes, the first-born pulsar has a very long lifetime, typically 109 years or more, although gravitational decay may further limit the lifetime in short-orbital-period binary systems. The second-born pulsar, on the other hand, is a "normal" pulsar, typically with a surface magnetic field of 1012 Gauss, which spins down and becomes undetectable in 107 years or so. Consequently, there is only a small chance of detecting the second neutron star as a pulsar in any given double-neutron-star system and even in any of the known systems.
An international team working at the Parkes radio telescope has beaten these odds by discovering the first known double-pulsar system! This system, PSR J0737-3039A/B, is outstanding, not only for the detection of the second pulsar, but also for its relativistic orbital motions. It was discovered in the Parkes High-Latitude Pulsar Survey, a collaborative project between groups at Jodrell Bank Observatory (UK), the Universita degli Studi di Bologna and INAF-Osservatorio Astronomico di Cagliari (Italy), Columbia University (USA), National Centre for Radio Astrophysics (India), Arecibo Observatory (USA) and the ATNF, which uses the Parkes multibeam receiver and filterbank receivers to search a region between Galactic longitudes of 220o and 260o and Galactic latitudes of -60o and 60o. The observing parameters were optimized for detection of millisecond pulsars (MSPs) with an integration time of 4 minutes and a sampling interval of 125 microseconds. Data were recorded on Digital Linear Tape and processed mainly using a computer cluster in Bologna using techniques similar to those employed in the Parkes multibeam pulsar survey (Manchester et al. 2001).
Figure1: Velocity curve for the relativistic binary pulsar PSR J0737-3039A. The mean orbital velocity is approximately 0.1% of the speed of light and the orbit eccentricity is 0.088.
PSR 0737-3039A, a 22-ms pulsar, was detected in April, 2003. It was immediately evident that the pulsar was a member of a short-period binary system, which was quickly shown to have an orbital period of only 2.4 hours (Figure 1). Even more interesting was the fact that the mean orbital velocity was very high, about 0.1% of the velocity of light, which in turn implies that the minimum companion mass is about 1.25 solar masses. This and the eccentricity of the orbit meant that the system is very likely to be a double-neutron-star system. With the large orbital velocity, relativistic effects were expected to be large and, with just a few days observation, the relativistic precession of periastron was detected at the predicted value, approximately 17 degrees per year. This is four times the value for the Hulse-Taylor binary pulsar, PSR B1913+16, making PSR J0737-3039A, by far the most relativistic binary pulsar known. Already we have detected four different "post-Newtonian" effects (i.e., effects which require general relativity or a similar gravitational theory for interpretation) including variations in the gravitational redshift during the orbit and, interestingly, a strong Shapiro-delay signal. The Shapiro delay results when the signal path from the pulsar passes close to the companion, resulting in a deflection of the signal path and an extra delay in the time of pulse reception. In the case of PSR 0737-3039A, the extra delay is about 100 microseconds, implying that the orbit is seen nearly edge-on. The derived inclination angle is 87o with an uncertainty of about a degree.
Figure 2: Pulse profile and polarization parameters for PSR J0737-3039A. In the lower part of the figure, the black line is the pulse total intensity, the red line is the linearly-polarized intensity and the green dotted line is the circularly-polarized intensity (V = IL - IR). The position angle of the linearly-polarized component is plotted in the upper part. The zero of position angle is arbitrary.
The pulse profile and its polarization is shown in Figure 2. There are two main pulse components separated by about 60% of the pulse period (as plotted). The time-reversal symmetry of the pulse shape and polarization properties about pulse phase 0.5 strongly suggests that the emitted beam is a very wide cone emitted from magnetic-field lines associated with one magnetic pole. The almost constant position angles through both components are consistent with this idea.
One of the most interesting aspects of this system is that orbit decay due to emission of gravitational waves is expected to lead to merging of the two neutron stars in about 85 Myr. While this may seem like a long time, it is less than one-third the time required for PSR B1913+16 to merge. These merger events are one of the principal targets for gravitational-wave detectors such as LIGO and VIRGO. Previous estimates of merger rates, largely dominated by PSR B1913+16, gave intervals of order 100 years between events detectable by the present LIGO. Detection of the new system implies a merger rate about seven times greater, meaning that there is a reasonable chance of observing such an event in just a few years of observation with LIGO. The paper announcing the discovery of PSR J0737-3039A, Burgay et al. (2003), discussed both these results and the implications for tests of gravitational theories. Kalogera et al. (2004) give more refined estimates of the predicted rate of LIGO detections.
These results were exciting enough, but this system had another surprise in store. In October, Duncan Lorimer was at Parkes for one of the many pulsar observing sessions, and was testing a pulsar search program on a more-or-less randomly selected data file which happened to be a 20-min observation of PSR J0737-3039A. He was astounded to see a very strong signal at a period of 2.77 seconds and a dispersion measure equal to that of J0737-3039A. Further investigation showed this to have the reversed Doppler variation of PSR J0737-3039A - quite by accident he had discovered the companion to PSR J0737-3039A, making this the first-known double-pulsar system! Incidentally, Marta Burgay was planning a search for pulses from the companion, but Duncan beat her to it. The emission from this second pulsar, PSR J0737-3039B, is strongly modulated with orbital phase, being almost undetectable for all except two roughly 10-minute periods each orbit. As it happened, the initial discovery observation of PSR J0737-3039A was at an orbital phase when the B-pulsar was turned off!
Figure 3: A schematic diagram illustrating the configuration of the PSR J0737-3039A/B double-pulsar system when the A-pulsar is at superior conjunction, that is, on the far side of the orbit as seen from Earth. Both stars orbit about the system barycentre (centre of mass) and, since they have similar masses, the orbits have similar sizes. The line of nodes is the intersection of the orbit plane and the plane of the sky and the apsidal line is the major axis of both orbits. The grey circle around the B-pulsar represents the size of the velocity-of-light cylinder (that for the A-pulsar is 100 times smaller). The lower part of the figure is the system viewed along the line of nodes. Since the orbit inclination is about 87o, the line-of-sight is nearly in the orbital plane and passes through the B-magnetosphere when the B-pulsar is at inferior conjunction. The dotted lines and grey sectors in the upper plot represent the orbit phases when the B-pulsar is strong. (Lyne et al. 2004)
Once the B-pulsar was discovered, it was possible to go back and re-analyse all previous filterbank data, folding at the longer period. This showed that the pulsar was a typical (albeit slow) relatively young "normal" pulsar with a surface-dipole magnetic field of about 1012 G, exactly in accordance with expectations based on the "recycling" model for MSP formation. For the first time it was possible to directly measure the mass ratio of the two stars from the relative amplitudes of the Doppler curves. The results were fully in accord with limits placed on the mass ratio from measurements of the post-Keplerian effects using the A-pulsar alone (Burgay et al. 2003). Figure 3 is a schematic diagram of the system. Not only do we have a highly constrained and highly relativistic double-pulsar system with enormous potential for tests of theories of gravitation, we also have a unique system where we can use the signals from both pulsars to probe the magnetospheric properties of the other. The discovery of the B-pulsar and these implications were announced by Lyne et al. (2004).
Figure 4: Constraints on the masses of the two neutron stars in the PSR 0737-3039A/B system. The grey regions are forbidden by the constraint that the sine of the orbit inclination angle cannot exceed unity; the lower-right region is from the A-pulsar and the upper-left region from the B-pulsar. The solid diagonal lines (marked R) define the limits on the mass ratio from the relative amplitudes of the Doppler curves for the two pulsars. Additional constraints are provided by the detection of relativistic effects in the timing of the A-pulsar, interpreted in the framework of general relativity. The diagonal dashed lines are limits on the sum of the masses based on the observed precession of periastron and the dot-dash lines are limits based on variations in time dilation as the pulsar moves around its somewhat eccentric orbit. The other two constraints, marked r and s, are based on the observed Shapiro delay. The inset shows an expanded plot of the region of intersection of the various constraints.
The new constraints on gravitational theories provided by the detection of the B-pulsar are dramatically illustrated in the so-called "mass-mass" diagram (Figure 4). The constraints on sin i (the so-called mass-function limits) already limit the masses to the white region in the upper right of the diagram. With the detection of the companion star as a pulsar, we have an entirely new constraint on the mass ratio (R). This constraint is important because it is largely independent of gravitational theories. Together with the constraint on the sum of the masses provided by the observed precession of periastron (interpreted within Einstein's general theory of relativity), we immediately determine the masses of the two stars to high precision: mA = 1.337 +/- 0.004 Msun and mB = 1.251 +/- 0.004 Msun. Constraints based on detection of other relativistic effects are fully consistent with these masses, providing new verifications of Einstein's theory. At present these limits are not very tight, but they will improve rapidly with time. We also expect additional relativistic effects to become detectable. For example, orbit decay due to emission of gravitational radiation from the system should be detectable within a year or so. With its highly relativistic properties, we also expect to be able to detect some previously unobserved effects. Geodetic precession of the rotation axes of the two stars is expected to have periods of about 75 years for A and 71 years for B. Not only is this likely to lead to changes in the observed pulse profiles, it will also very likely make possible detection of aberration effects due to the pulsar rotation. Higher-order relativistic terms, for example, variations in the observed orbit precession rate, are also likely to be detectable. There is no doubt that this system will result in manifold constraints on possible theories of gravitation. So far, Einstein is King. Will this still be true in five or ten years from now? A big question indeed!
Figure 5: Orbital variations in the intensity and pulse shape of the B-pulsar. The vertical axis represents about one-quarter of the pulse period for observations at each of three frequencies as marked. The horizontal axis is orbital longitude (true anomaly) and the vertical line represents inferior conjunction of the B-pulsar, that is, when B is on the near side of the orbit and aligned with the A-pulsar. (Lyne et al. 2004)
Figure 5 illustrates the orbital variations in the emission from the B-pulsar. The pulsed emission is only strong for two intervals, each of about 10-minutes duration, when the pulsar is near inferior conjunction. Not only does the pulse intensity vary, but the pulse shape also changes! Especially near the start of the first burst, the trailing component is strong and there is a weaker leading component which fades during the burst. In the second burst, the two components are more equal in amplitude and more widely separated. These changes, which are unprecedented in the history of pulsar astronomy, repeat every orbital period and are largely achromatic, putting strong constraints on possible interpretations. A clue to the likely mechanism is shown in Figure 3. The relativistic wind from pulsar A is far more energetic than that from B, and penetrates far into the B-pulsar magnetosphere. Because of the changing aspect of the B-pulsar magnetosphere as seen from A, the degree of penetration will depend on both the B rotational-phase and the orbital phase. It is almost certain that this is greatly affecting the B-pulsar emission, but understanding the details of this process is going to require a lot more investigation.
As if all this was not enough, we have also detected an eclipse of the A-pulsar emission as it passes behind the B-pulsar. The eclipse is very short, lasting only about 30 seconds, is total within the uncertainties and again is largely achromatic. This interaction gives us an unprecedented probe of a pulsar magnetosphere. For example, we might expect changes in the polarization of the A-pulsar as the signal propagates through the B-magnetosphere. Because of the short timescales involved, such measurements are an observational challenge, one we are ready to take up!
To sum up, this system is a wonderful and unique laboratory for the investigation of a variety of astrophysical phenomena. The next few years are going to be interesting indeed!
We thank the Parkes Observatory staff who, as always, have provided outstanding support for our observations. We made extensive use of the PSRCHIVE pulsar analysis system developed by Aidan Hotan and Willem van Straten and we thank them for their efforts.
References
Burgay, M., D'Amico, N., Possenti, A., Manchester, R.N., Lyne, A.G., Joshi, B.C., McLaughlin, M.A., Kramer, M., Sarkissian, J. M. Camilo, F., Kalogera, V., Kim, C, & Lorimer, D.R., 2003, Nature, 426, 531-533
Kalogera, V., Kim, C., Lorimer, D.R., Burgay, M., D'Amico, N., Possenti, A., Manchester, R.N., Lyne, A.G., Joshi, B.C., McLaughlin, M.A., Kramer, M., Sarkissian, J. M. & Camilo, F., 2004, ApJ Lett., 601, 179
Lyne, A.G., Burgay, M., Kramer, M., Possenti, A, Manchester, R.N., Camilo, F., McLaughlin, M.A., Lorimer, D.R., D'Amico, N., Joshi, B.C., Reynolds, J.E. & Freire, P.C.C., 2004, Science, 303, 1153
Manchester, R.N., Lyne, A.G., Camilo, F., Bell, J.F., Kaspi, V.M., D'Amico, N., McKay, N.P.F., Crawford, F., Stairs, I.H., Possenti, A., Kramer, M. & Sheppard, D.C., 2001, MNRAS, 328, 17-35
Dick Manchester, on behalf of the Parkes High-Latitude Pulsar Survey team:
Marta Burgay, Fernando Camilo, Nichi D'Amico, Paulo Freire, Bhal Chandra Joshi, Michael Kramer, Duncan Lorimer, Andrew Lyne, Maura McLaughlin, Dick Manchester, Andrea Possenti, John Reynolds and John Sarkissian.
(Dick.Manchester@csiro.au)