The Binary Companion of Young, Relativistic Pulsar J1906+0746
Joeri van Leeuwen, Laura Kasian, Ingrid H. Stairs, D. R. Lorimer, F. Camilo, S. Chatterjee, I. Cognard, G. Desvignes, P. C. C. Freire, G. H. Janssen, M. Kramer, A. G. Lyne, D. J. Nice, S. M. Ransom, B. W. Stappers, J. M. Weisberg
DD RAFT VERSION J ULY
30, 2018
Preprint typeset using L A TEX style emulateapj v. 12/16/11
THE BINARY COMPANION OF YOUNG, RELATIVISTIC PULSAR J1906+0746 J. VAN L EEUWEN , L. K
ASIAN , I. H. S TAIRS , D. R. L ORIMER , F. C AMILO , S. C
HATTERJEE , I. C OGNARD , G. D ESVIGNES ,P. C. C. F REIRE , G. H. J ANSSEN
1, 10 , M. K
RAMER , A. G. L YNE , D. J. N ICE , S. M. R ANSOM , B. W. S TAPPERS , AND
J. M. W
EISBERG Draft version July 30, 2018
ABSTRACTPSR J1906+0746 is a young pulsar in the relativistic binary with the second-shortest known orbital period,of 3.98 hours. We here present a timing study based on five years of observations, conducted with the 5largest radio telescopes in the world, aimed at determining the companion nature. Through the measurementof three post-Keplerian orbital parameters we find the pulsar mass to be 1.291(11) M (cid:12) , and the companionmass 1.322(11) M (cid:12) respectively. These masses fit well in the observed collection of DNS! s, but are alsocompatible with other white dwarfs around young pulsars such as J1906+0746. Neither radio pulsations nordispersion-inducing outflows that could have further established the companion nature were detected. Wederive an HI-absorption distance, which indicates that an optical confirmation of a white dwarf companion isvery challenging. The pulsar is fading fast due to geodetic precession, limiting future timing improvements.We conclude that young pulsar J1906+0746 is likely part of a double neutron star, or is otherwise orbited by anolder white dwarf, in an exotic system formed through two stages of mass transfer.
Keywords: pulsars: individual (PSR J1906+0746) – stars: neutron – white dwarfs – binaries: close INTRODUCTION
Binaries harboring a neutron star are windows on dynam-ical star systems that have undergone and survived at leastone supernova. By precisely measuring pulsar times of arrival(TOAs), and fitting binary models to these, one can describethe orbital motions of the pulsars and their companions – andhence constrain their masses. In combination with other infor-mation such as the pulsar spin, orbit, and companion nature,these mass estimates can elucidate the binary interaction andmass transfer history.In the vast majority of observed binary pulsar systems, thepulsar is the first-born compact object: there, it is found tohave been spun up by accretion from its companion to a higherspin rate than seen in young pulsars. These spun-up pulsarshave far lower magnetic fields than the general pulsar popula-tion. They thus show very stable rotation and evolve only very ASTRON, the Netherlands Institute for Radio Astronomy, Postbus 2,7990 AA, Dwingeloo, The Netherlands; [email protected] Astronomical Institute “Anton Pannekoek”, University of Amsterdam,Science Park 904, 1098 XH Amsterdam, The Netherlands Department of Physics and Astronomy, University of BritishColumbia, Vancouver, BC V6T 1Z1, Canada Department of Physics, West Virginia University, Morgantown, WV26506, USA Arecibo Observatory, HC3 Box 53995, Arecibo, PR 00612, USA Columbia Astrophysics Laboratory, Columbia University, New York,NY 10027, USA Center for Radiophysics and Space Research, Cornell University,Ithaca, NY 14853, USA Laboratoire de Physique et Chimie de l’Environnement et de l’EspaceLPC2E CNRS-Universit´e d’Orl´eans, 45071 Orl´eans, and Station de ra-dioastronomie de Nanc¸ay, Observatoire de Paris, CNRS/INSU, 18330Nanc¸ay, France Max-Planck-Institut f¨ur Radioastronomie, D-53121 Bonn, Germany Jodrell Bank Centre for Astrophysics, School of Physics and Astron-omy, University of Manchester, Manchester, M13 9PL, UK Department of Physics, Lafayette College, Easton, PA 18042, USA NRAO (National Radio Astronomy Observatory), Charlottesville, VA22903, USA Department of Physics and Astronomy, Carleton College, Northfield,MN 55057, USA slowly – resulting in higher characteristic ages of ∼
10 Gyr.This stability and longevity means such recycled pulsars areobservable for much longer periods of time than the high-fieldfast-evolving young pulsars.The amount of recycling is related to the binary type: low-mass binary pulsar (LMBP) systems host millisecond pulsars(MSPs), which have spin periods of about 1 −
10 ms, and areorbited by a low-mass white dwarfs (WDs). Pulsars with moremassive WD companions or neutron star (NS) companions,generally have longer spin periods (10 −
200 ms), and lowercharacteristic ages (Lorimer 2008). This is consistent with aspin-up picture in which the amount of mass transferred fromthe companion to the pulsar depends largely on the durationof the mass-transfer stage, and hence, on the mass of the pro-genitor of the companion (Alpar et al. 1982).The 144-ms pulsar J1906+0746, discovered in precur-sor PALFA observations in 2004 (Cordes et al. 2006; vanLeeuwen et al. 2006; Lorimer et al. 2006), is one of onlya handful of known double-degenerate relativistic binarieswhere the pulsar is believed to be the younger of the two com-pact objects. Other such pulsars, listed in Table 1, includethe white dwarf binaries J1141 − − − − a r X i v : . [ a s t r o - ph . S R ] N ov VAN L EEUWEN ET AL . Pulsar Period P b Eccentricity Pulsar Companion Companion(ms) (days) Mass ( M (cid:12) ) Mass ( M (cid:12) ) TypeYoung Pulsars in Relativistic Binaries J0737 − . +0 . − . . +0 . − . NSJ1906+0746 . +0 . − . . +0 . − . WD or NSJ1141 − . +0 . − . . +0 . . WDB2303+46 . +0 . − . . +0 . − . WD Recycled Pulsars in Relativistic Double Neutron Star Binaries
J0737 − . +0 . − . . +0 . − . NSJ1756 − . +0 . − . . +0 . − . NSB1913+16 . +0 . − . . +0 . − . NSB2127+11C . +0 . − . . . . NSB1534+12 . +0 . − . . +0 . − . NS Recycled Pulsars in Long-Period ( P b > J1518+4904 . +0 . − . . +0 . − . NS Total Mass ( M (cid:12) ) J1829+2456 . +0 . − . NSJ1753 − − . +0 . − . NS Table 1
Known pulsars in relativistic and/or double neutron star binary systems, ordered by pulsar age, with minor ordering on binary period. (Values from: Krameret al. 2006; this work; Bhat et al. 2008; Thorsett et al. 1993, Kulkarni & van Kerkwijk 1998; Ferdman et al. 2014; Weisberg et al. 2010; Jacoby et al.2006; Fonseca et al. 2014; Janssen et al. 2008; Champion et al. 2005; Keith et al. 2009; Corongiu et al. 2007)
J1141 − ˙ P ∼ × − s/s) andshort period, J1906+0746 is a young pulsar. Its characteris-tic age τ c is roughly 112 kyr, the lowest of all known binarypulsars, and in the 8th percentile of the general pulsar age dis-tribution (Manchester et al. 2005). Young pulsars show rapidspin-down evolution. By definition this early 10 -year stage ismuch more fleeting than the 10 -year detectable life span of recycled pulsars. In binary systems the recycled pulsars arethus common, the young pulsars rare. The recycled pulsarsare generally the older of the two binary components – butthe young pulsars formed more recently than their compact-object companion. Therefore these latter provide a perspec-tive on binary evolution that is different from the typical recy-cling scenario.Pulsar J1906+0746 was intensively monitored with theArecibo, Green Bank, Nanc¸ay, Jodrell Bank and Westerborkradio telescopes, up to 2009, after which the ever-decreasingpulse flux density (Fig. 1) generally prevented significant fur-ther detections. We here present the high-precision follow-uptiming from these telescopes over that initial five year period2005-2009.Through this analysis we are able to significantly improvethe system characterization presented in the Lorimer et al.(2006) discovery paper: we measured two more post-Keplerian orbital parameters (for a total of three) – which,assuming general relativity, provide an over-constrained de-termination of the pulsar and companion masses.In § §
3. In § § F l ux d e n s it y ( m J y ) Figure 1.
The profile evolution of J1906+0746. Shown are the 1998 and2005 Parkes profiles (Lorimer et al. 2006), and yearly Arecibo profiles (thiswork). All subplots are the same scale. The 1998 and 2005 profiles used adifferent recording setup, with somewhat higher noise. OBSERVATIONS AND INITIAL DATA REDUCTION
For the timing follow-up of this pulsar, we have obtainedhigh signal-to-noise data using the Arecibo Telescope and theGreen Bank Telescope (GBT), covering several full orbitswith the latter. High-cadence data from the Nanc¸ay, Jodrell HE Y OUNG , R
ELATIVISTIC B INARY P ULSAR
J1906+0746 3
Telescope Backend Epoch Cadence Interval Central Frequency Bandwidth Channel BW Coherent N
TOAs (MHz) (MHz) (MHz) Dedispersion
Arecibo WAPP ×
100 0.195 N 23250ASP −
32 4 Y 220GBT GASP - - 1440 600 0.781 N -Nanc¸ay BON Table 2
Details of the telescope and backend setup. N
TOAs marks the number of TOAs generated per backend. Spigot data was used for searching, not for timing. Dowd et al. (2000); Demorest (2007), Ferdman (2008); Kaplan et al. (2005); Desvignes (2009); Hobbs et al. (2004); Voˆute et al. (2002); Karuppusamyet al. (2008).
Bank and Westerbork telescopes provided further long-termtiming coverage.Data from the Arecibo telescope (USA) were taken usingtwo backends simultaneously, as detailed in Table 2. ThreeWideband Arecibo Pulsar Processor (WAPP) filterbank ma-chines autocorrelated the two polarization channels. Offline,these were converted to spectra, dedispersed incoherently,summed, and finally folded at the local value of the pulsarperiod. Further data were taken using the Arecibo Signal Pro-cessor (ASP) coherent dedispersion machine, which foldedon-line using the best-known values for the dispersion mea-sure and the local pulse period. Through its coherent dedis-persion capabilities, ASP is complementary to the WAPPswith their larger bandwidths. In parallel to this on-line fold-ing, ASP recorded the 4-MHz wide band around 1420 MHzand stored the baseband data on disk, for off-line investiga-tion of HI absorption toward the pulsar ( § Flux calibration and offline refolding
The WAPP, ASP and GASP data were flux-calibrated usingthe noise diode signal that was injected into the receiver, foreach polarization individually, before each observation. Whengood calibration observations were not available, we normal-ized the flux density in each profile by the root-mean-square(RMS) across the profile for the coherently dedispersed pro-files, while weighting all channels equally for the WAPP fil-terbank data. A continuum source was used to further cali-brate the ASP and GASP data for a significant portion of theepochs, using ASPFitsReader (Ferdman 2008). The WAPPdata were calibrated using
SIGPROC and pre-recorded cal-ibrator data . http://sigproc.sourceforge.net/ For all Arecibo and GBT data, time-averaged pulse pro-files were finally created by adding both polarizations, all fre-quency channels, and five minute integrations. These profileswere remade whenever sufficient new Arecibo and GBT datawere obtained to compute a new ephemeris. As the coherentlydedispersed ASP and GASP data were recorded as a series of30- or 60- second integrated pulse profiles, these were subse-quently realigned to create new 5-minute integrated profiles.The Jodrell Bank and Westerbork/PuMa profiles were pro-duced using their respective custom off-line software, whilethe Nanc¸ay and the Westerbork/PuMaII profiles were reducedusing the
PSRCHIVE software package (Hotan et al. 2004),These data were not flux-calibrated; they were normalized bythe RMS of the noise. HI ABSORPTION IN THE PULSAR SPECTRUM
Lorimer et al. (2006) combined the measured dispersionmeasure (DM) of 218 cm − pc with the NE2001 Galacticelectron-density model (Cordes & Lazio 2002) to estimatethe distance to J1906+0746 of 5.40 kpc. Comparisons ofNE2001 and VLBI distances suggest an error of 20% (Delleret al. 2009b, although selection effects favoring nearby pul-sars may lead to much larger errors in several cases), pro-ducing an overall estimate of . +0 . − . kpc. Refining thatestimate would be beneficial for several lines of follow-upanalysis: it would improve our estimate of the kinematic con-tribution to the observed orbital period decay ˙ P b , and thusstrengthen our general relativity tests ( § § µ Jy in 2012, making it unsuitable for VLBA astrometry. http://psrchive.sourceforge.net/ VAN L EEUWEN ET AL .In the kinematic method, the absorption of pulsar emissionby Doppler-shifted neutral hydrogen (HI) along the line ofsight, can, once combined with a model of the Galactic kine-matics, constrain the distance to the pulsar. In the off-pulsephases, the HI in its line of sight shines in emission; but duringthe on-pulse phases, some of the pulsar’s broadband emissionwill be absorbed by the HI clouds between it and Earth (the“pulsar off” and “pulsar on” simulated spectra in Fig. 2, left).As this absorption occurs at the Doppler-shifted HI frequency,it can be associated with a velocity relative to the observer. AGalactic rotation curve in the line of sight can then model atwhich distances from Earth such a velocity is expected. If themeasured absorption velocities unambiguously map onto thecurve, a lower limit to the pulsar distance can be produced.In some cases, an upper limit on the pulsar distance can alsobe derived: if certain features in the emission spectrum donot have corresponding absorption dips, the pulsar can be as-sumed to lie in front of the emitting region corresponding tothe feature velocity (see Frail & Weisberg 1990 and Weisberget al. 2008 for further discussion).Deriving a kinematic distance to J1906+0746 is especiallyinteresting as its
DM-derived distance is . +0 . − . kpc, nearthe tangent point in this direction, at . kpc. The absence ofabsorption at the highest velocities, for example, would firmlyput J1906+0746 closer to us than this tangent point.Most kinematic HI distances have been determined forbright, slow pulsars (e.g. Frail & Weisberg 1990; Johnstonet al. 2001). Of the 70 pulsars with kinematic distances listedin Verbiest et al. (2012, their Table 1, excluding distances de-rived from associations), J1906+0746 is the 8th fastest spin-ning and, at 0.55 mJy, the dimmest. Combined with the highDM, measurement of HI absorption is challenging. We there-fore used, for the first time in a pulsar HI absorption measure-ment, coherent dedispersion to maximize the signal-to-noiseratio by eliminating smearing between on- and off-pulse bins. Observations and analysis
During four full tracks on J1906+0746 with Arecibo, on2006 June 14, July 11, Oct 12 and November 12, we recordeda total of 7.6 hr of baseband data with ASP ( § DSPSR (van Straten & Bailes 2011), each of the four obser-vations was coherently dedispersed, folded onto 256 phasebins using the ephemeris resulting from the timing campaign(Table 3), and split into the maximum possible number ofchannels of 1024 over 4 MHz, for a velocity resolution of0.83 km/s. The long Fourier transforms needed for coherentdedispersion obviated further windowing functions.The two, main and interpulse, on-regions were defined asthe series of phase bins with signal-to-noise ratio SNR > PSRCHIVE (van Straten et al. 2012). For the on-pulse spectrum, the spectra were averaged while weightingby the square of the pulsar SNR in the concerning on-pulsebin (following Weisberg et al. 2008). For each observation,the channel frequencies were barycentered and converted tovelocities relative to the Local Standard of Rest (LSR).The spectra from the four observations were summed,weighted by the square of the pulsar SNR in each observation,for both the on and off-pulse. The intensity scales were cali-brated by matching the peak of the off-pulse spectrum to thepeak brightness temperature T b as measured in this directionby the VLA Galactic Plane Survey (Stil et al. 2006). Over- I( v ) / I O p ti ca l D e p t h -150 -100 -50 0 50 100 150Velocity (km/s)051015 D i s t a n ce ( kp c ) l=41.6°b=0.1° T b ( K ) T b ( K ) I(v)I(v)pulsar onpulsar off pulsar off method(simulation) data -150 -100 -50 0 50 100Velocity (km/s)
Figure 2.
Simulated (left) and measured (right) HI absorption data. Left: il-lustration of the subtraction and scaling methods used to define the absorptionspectrum I(v)/I . A smoothed version of the measured “pulsar-off” spectrumis shown as the bottom solid curve. The top, gray, dashed curve is basedon the measured “pulsar-on” spectrum but the increase in intensity over thepulsar-off spectrum is amplified 100 × for clarity, and shows the expected on-pulse spectrum if there were no absorption and I ( v ) = I . In the middle,solid, gray-filled curve, which then defines the absorption spectrum I ( v ) /I ,absorption of the two features at 13 and 35 km/s is simulated.Right: The measured HI spectrum. In the top panel, the off-pulse HI emis-sion spectrum is shown. The 35 K criterion ( § I ( v ) /I .The optical depth, derived from e − τ = I ( v ) /I , is noted with a dashed line at τ =0.3. In the bottom panel, the Galactic rotation curve for this line of sightis plotted. The features in the absorption spectrum span the velocity range upto the tangent point. The derived lower and upper bounds to the distance areeach marked with a vertical dashed line. all “pulsar on” and “pulsar off” spectra were thus produced.Their difference I ( v ) , illustrated in Fig. 2, can be attributedto the pulsar minus the absorption. By dividing by I , thebroadband unabsorbed strength of the pulsar signal, the rel-ative absorption spectrum for J1906+0746 was produced, asshown in the right-middle panel of Fig. 2. It shows severaldeep absorption features. Interpretation
Any absorption features deeper than an optical depth τ of0.3 is considered significant (Frail & Weisberg 1990). Four ofthese appear, peaking at 4, 13, 35 and 63 km/s. To determinethe distances to which these velocities correspond, we con-structed a Galactic rotation curve (the same as Verbiest et al.2012) using a distance from the Galactic Center to the Sunof R =8.5 kpc and a flat rotation of Θ =220 km/s (Fich et al.1989). HE Y OUNG , R
ELATIVISTIC B INARY P ULSAR
J1906+0746 5From this curve (Fig. 2, bottom panel), we find that thehighest-velocity emission component, at 63 km/s, is emittednear the tangent point. This means the distance lower limit isat or beyond that tangent point at 6.4 kpc. By propagating theestimated spread in velocities of ± T b =35 K is thought to be suffi-ciently bright to allow for an absorption measurement (Weis-berg et al. 1979). The upper limit to the distance is thusprovided by the farthest-out such peak that is not accompa-nied by absorption. The first peak to meet these criteria isat −
32 km/s. This means the hard distance upper limit de-rived using this kinematic distance is 15.9 ± we com-bine our upper and lower limits with the discovery mean fluxdensity of 0.55 mJy (Lorimer et al. 2006), and obtain an over-all distance estimate of . +2 . − . kpc. That estimate is largerthan the previous, DM-derived distance of . +0 . − . kpc. PROFILE EVOLUTION
The profile of J1906+0746 changes drastically over atimescale of years. The prominent interpulse in the 2004 dis-covery pulse profile was absent from the 1998 Parkes archiveprofile (Fig. 1). On the shortest time scales, our observationsrange from single-pulse data to 10, 30, 60 and 120-second in-tegrations. From visual inspection we have detected no modechanges on either of these timescales, or in the hours to weeksbetween observations.We attribute the profile evolution to geodetic precession ofthe pulsar’s spin angular momentum vector about its total or-bital angular momentum vector. In general relativity (GR) theprecession rate (e.g. Barker & O’Connell 1975) is: Ω geod = T / (cid:12) (cid:18) πP b (cid:19) / − e m (4 m + 3 m )2 ( m + m ) / (1)where T (cid:12) = GM (cid:12) /c = 4 . µs is the solar massexpressed in time units; m and m are the pulsar and com-panion masses, respectively, in solar masses; P b is the orbitalperiod, and e is the eccentricity. Using the timing solution andmasses presented later, in Table 3, we find a predicted geode-tic precession period of ∼
165 years, which equals a rate of . degrees per year. Over the 2005 − ◦ – since the 1998 archival Parkes observation,qualitatively consistent with the significantly different pulseshape then. http://psrpop.phys.wvu.edu/LKbias Main Pulse
Phase AABB M J D Phase0.700.24
Interpulse
Figure 3.
Illustration of the alignment based on Gaussian components, forGASP and ASP profiles. Shown are the main pulse (left) and interpulse (right,flux scale increased 7 times). The vertical line at phase 0.25 marks the loca-tion of the fiducial point. The vertical line at phase 0.75 can be used to tracethe changing location of the interpulse, over time. The red lines represent theGaussian component A or B that was used for the alignment. The blue linesare the other Gaussian components. For MJD 54390 the data (black line) andtotal profile model (dashed green line) are also shown.
The secular profile changes observed in J1906+0746 offeran exciting opportunity to study geodetic precession; but thechanging profile shape poses a problem in determining pre-cise pulse times-of-arrival (TOAs). To ensure that the fiducialpoints of all profiles are consistent, and limit introduction offurther timing noise, we used a series of Gaussian standardprofiles developed from the well-modelled epochs of ASP,GASP and WAPP data, and next aligned these Gaussian tem-plates as described below.Using
BFIT (Kramer et al. 1994) we fit sets of up to 3(as necessary) Gaussians to both the pulse and interpulse inthe summed profile of each epoch (illustrated for MJD 54390in Fig. 3). This approach was previously used to determinethe geometry of the PSR B1913+16 system (Kramer 1998).We next identified a stable component, and used this compo-nent as our timing fiducial point. We found that the smoothestalignment was achieved by keeping the phase of the initiallytallest component (“Component A”) constant. Fig. 3 showsthis approach for a subset of the ASP and GASP summedprofiles. After modified Julian day (MJD) 54700 we could nolonger reliably identify Component A, and so instead alignedthe tallest component for those epochs (Component B), andintroduced a fiducial point phase shift based on the transitionprofile in which components A and B were both identifiable(see Fig. 3). The full collection of Gaussian-modelled pro-files is shown in Fig. 4. This method of aligning the profilesproduced fairly monotonic behaviour in the phase of the in-terpulse, as can be seen by the gradual widening between theinterpulse peaks and the fixed, vertical dashed line in Fig. 4.After this alignments, these profiles were used as the stan-
VAN L EEUWEN ET AL . Main Pulse
Phase M J D Phase0.700.22
Interpulse
Figure 4.
The resulting fits to the pulse and interpulse of J1906+0746, basedon data taken with the WAPPs (red: 1170 MHz, green: 1370 MHz, blue:1570 MHz) ASP (magenta) and GASP (cyan). The vertical axis spans fromJuly 2005 to August 2009. The interpulse flux is magnified by a factor often relative to the main pulse. The vertical line for the main pulse (left) il-lustrates the chosen alignment. A vertical line in the interpulse panel (right)illuminates the interpulse phase shift. dards for high-precision timing.
Profile flux variations
Throughout the period covered by this study, the profileevolution of J1906+0746 was accompanied by a steady de-crease in the pulsar mean flux density (Fig. 1). Independentlyprocessed, well calibrated, ASP and GASP data producedmean flux density estimates that are consistent, falling from0.8 mJy in 2006 to 0.2 mJy in 2009 (Fig 4.6 in Kasian 2012). TIMING
Times of arrival
Times of arrival (TOAs) were created for the profiles re-sulting from the iterative data reduction process described in § § NançayASPGASPJodrell WAPPs Westerbork R e s i du a l ( m s ) - - R e s i du a l s ( P ) - Figure 5.
Residuals plotted versus MJD and Year for the entire data set.Shown are the residuals after using the timing solution that included a 10thorder polynomial in time. Also fitted were offsets between the various ob-servatories, between pre- and post-54100 Jodrell data, and between PuMaand PuMaII. Data from ASP are magenta, GASP is yellow, Jodrell is cyan,Nanc¸ay is red, the combined WAPP data are shown in green, and Westerborkis blue.
Timing solution for J1906+0746
The complete set of 28,000 TOAs was fit for the parametersdescribing the state and evolution of both the individual pul-sar and the binary system. This initial fitting was performedusing the
TEMPO2 (Hobbs et al. 2006) package. Using thedata from Arecibo, GBT, Jodrell, Nanc¸ay and Westerbork, weproduced a phase-connected solution over the entire period,effectively accounting for every one of the 10 rotations overthe 2005 − FITWAVES (Hobbs et al. 2004), but better noise removalwas achieved by modeling the pulsar rotation frequency as a10th order polynomial in time, the highest degree of complex-ity currently implemented in
TEMPO2 . The residuals of thattiming solution are shown in Fig. 5.Even this last solution, however, shows large variations (of420 µ s rms) due to remaining timing noise or unmodeled pro-file variations. To determine the system parameters with thehighest precision, we thus chose to include only the timingdata that used the evolving timing profiles, from Arecibo,GBT and Nanc¸ay ( § § TEMPO code allows up to a20th-order polynomial in frequency, and we tested fitting eachof these orders while simultaneously fitting the offsets, DMvalues, position and spin and binary parameters. The fit thatallowed for a 4th-order frequency polynomial produced thelowest reduced- χ and was adopted as our preferred timingsolution.The final, best timing solution is presented in Table 3, and http://tempo.sourceforge.net/ HE Y OUNG , R
ELATIVISTIC B INARY P ULSAR
J1906+0746 7
NançayASPGASP WAPPs
With OffsetsWithout Offsets
MJD R e s i du a l ( µ s ) R e s i du a l ( µ s ) -2×10 Figure 6.
Residuals are plotted versus MJD for the best fit. The toppanel shows the residuals for the solution that includes jumps between the118 individual epochs. Those jumps remove the long-term timing noisetrend that is clearly visible in the bottom panel. There we show the samephase-connected timing solution but without any jumps between epochs.Green residuals represent WAPP Arecibo data, magenta represents ASPArecibo, yellow is GASP Green Bank, and red is Nanc¸ay data. the resulting residuals (17 µ s rms) are shown in the top panelof Fig. 6. We re-weighted our data so that the reduced χ ofthe fit was equal to 1 for each data set, and overall. Followingcommon use, all reported uncertainties are twice the valuesproduced by TEMPO after this re-weighting.
Measurement of post-Keplerian parameters
Our timing solution presented in Table 3 includes severalpost-Keplerian parameters. We describe these in some moredetail below.
Measuring γ , the gravitational redshift We tested how our measurement of γ depends on the cur-rent stage of the periastron precession cycle, which affects theviewing geometry. We simulated TOAs over various pointsspanning one whole cycle, as was previously shown for PSRB1913+16 and B1534+12, in Damour & Taylor (1992, Fig.5, top line). We find that our set of TOAs, collected while ω moved from ∼ ◦ to ◦ , corresponds to relatively low the-oretical fractional uncertainties, and we are moving towardseven better measurability as ω increases (Kasian 2012). Measuring ˙ P b , the orbital period derivative The observed value of ˙ P b needs to be corrected for twoeffects, before it can be compared to the value predictedby general relativity (GR), and used to constrain the systemmasses. These are, first, the different Galactic accelerationfelt by the pulsar and by Earth; and second, the Shklovskiieffect (Shklovskii 1970), which incorporates the Doppler ef-fect caused by the pulsar proper motion into the measured ˙ P b value. We have no measurement of proper motion for thispulsar, but we can estimate the Galactic contribution and cal-culate the limit on the Shklovskii contribution and the systemproper motion. If that limit is in line with the measured propermotions of similar systems, ˙ P b can be used to constrain thebinary system. Following Damour & Taylor (1991) and Nice& Taylor (1995) we write the observed orbital period decay ˙ P obs b as: ˙ P obs b = ˙ P int b + ˙ P Gal b + ˙ P Shk b (2)where the Galactic contribution ˙ P Gal and Skhlovskii term ˙ P Shk b add to the intrinsic decay ˙ P int b . If we assume that ˙ P int b is equal to ˙ P GR b , the value determined by GR, we can rewriteEq. 2 to isolate the Shklovskii contribution, scaled to the bi-nary orbit: (cid:32) ˙ P b P b (cid:33) Shk = (cid:32) ˙ P b P b (cid:33) obs − (cid:32) ˙ P b P b (cid:33) GR − (cid:32) ˙ P b P b (cid:33) Gal (3)On the right-hand side of Eq. 3, each of the terms can beestimated or calculated. First, our best fit value of the orbitaldecay ˙ P obs b is − × − (Table 3).Second, the value predicted by GR and computed fromfitting the DDGR model (Damour & Deruelle 1986; Taylor& Weisberg 1989) to our data ˙ P GR b is − × − (Table 3).The third term, the Galactic contribution, can be written as ( ˙ P b /P b ) Gal = (cid:126)a · ˆ n/c , where (cid:126)a is the differential accelera-tion in the field of the galaxy and ˆ n is the unit vector alongour line of sight to the pulsar. The components parallel andperpendicular to the Galactic plane that make up this termcan be calculated (Nice & Taylor 1995, Eqs. 3 − Θ = 220 km/s and R = 8 . kpc, plus the pulsar co-ordinates and HI-absorption distance from Table 3, we find ( (cid:126)a · ˆ n/c ) (cid:107) = (6 . ± . × − s − and ( (cid:126)a · ˆ n/c ) ⊥ =(4 . ± . × − s − . These translate to a Galactic correc-tion ˙ P Gal b = × − . Given the large distance un-certainty, this value does not significantly change when usingthe more recent Reid et al. (2014) Galactic kinematics. Com-bined, the three terms limit the Shklovskii contribution to beessentially zero, ˙ P Shk b = P b µ d HI /c < . × − (at 95%confidence level), where µ and d HI are the total proper motionand the distance of the pulsar respectively. The error rangeson P b and d HI allow for a proper motion of < mas/year ortransverse velocity of v = µ d HI < km/s (95% CL). Thiseasily encompasses the range of published system velocitiesfor other relativistic pulsars (e.g. Hobbs et al. 2005).Our results thus imply that the orbital period decay ˙ P b forJ1906+0746 is consistent with the value predicted by generalrelativity. Mass measurements
Having obtained reasonable estimates of the advance of pe-riastron ˙ ω , the gravitational redshift/time dilation parameter γ , and the orbital decay ˙ P b for J1906+0746, we use thesethree parameters to place constraints on the masses of the pul-sar ( m ) and companion ( m ).If we use the dependence of the post-Keplerian parame-ters on the masses, as defined in general relativity (see, e.g. VAN L EEUWEN ET AL . Measured Parameter DD Value DDGR Value
Right ascension, α (J2000.0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19:06:48.86(4) 19:06:48.86(4)Declination, δ (J2000.0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 07:46:25.9(7) 07:46:25.9(7)Spin Period, P (s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.14407315538(5) 0.14407315538(5)Pulse Frequency, ν (s − ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.940918295(2) 6.940918295(2)First derivative of Pulse Frequency, ˙ ν (s − ) ( × − ) . . . . . . . . . . . − − ¨ ν (s − ) ( × − ) . . . . . . . . . 5.0(7) 4.9(7)Third derivative of Pulse Frequency (s − ) ( × − ) . . . . . . . . . . . . . − − − ) ( × − ) . . . . . . . . . . . . − − DM (cm − pc) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217.7508(4) 217.7508(4)Ephemeris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DE405Clock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TT(BIPM)Orbital Period, P b (days) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.16599304683(11) 0.16599304686(11)Projected Semimajor Axis, x (lt s) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4199620(18) 1.4199506(18)Orbital Eccentricity, e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.0853028(6) 0.0852996(6)Epoch of Periastron, T (MJD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54288.9298810(2) 54288.9298808(2)Longitude of Periastron, ω (degrees) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76.3320(6) 76.3317(6)Rate of Periastron Advance, ˙ ω (degrees/yr) . . . . . . . . . . . . . . . . . . . . . . . 7.5841(5)Time Dilation and Gravitational Redshift Parameter, γ . . . . . . . . . . . . 0.000470(5)Orbital Period Derivative, ˙ P b ( × − ) . . . . . . . . . . . . . . . . . . . . . . . . . − ˙ P b ( × − ) . . . . . . . . . . . . . . . . . . 0.03(3)Total Mass, M total ( M (cid:12) ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6134(3)Companion Mass, m ( M (cid:12) ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.322(11) Derived Parameter DD Value DDGR Value
Pulsar Mass, m ( M (cid:12) ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.291(11)Rate of Periastron Advance, ˙ ω (degrees/yr) . . . . . . . . . . . . . . . . . . . . . . . 7.5844(5)Time Dilation and Gravitational Redshift Parameter, γ . . . . . . . . . . . . 0.000470(5)Orbital Period Derivative, ˙ P b ( × − ) . . . . . . . . . . . . . . . . . . . . . . . . . − i (degrees) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.7(5)Galactic Latitude, l (degrees) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.5982Galactic Longitude, b (degrees) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.1470Mass Function, f mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.1115662(2) 0.1115636(4)Characteristic Age τ c = P/ P (kyr) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112.6 112.6Surface Magnetic Field, B S = 3 . × ( P ˙ P ) / ( G) . . . . . 1.73 1.73DM-derived Distance to Pulsar, d DM (kpc) . . . . . . . . . . . . . . . . . . . . . . . . +0 . − . . +0 . − . HI-derived Distance to Pulsar, d HI (kpc) . . . . . . . . . . . . . . . . . . . . . . . . . . +2 . − . Table 3
Timing parameters for J1906+0746. The columns contain the parameters measured and derived using, on the left, the DD (Damour & Deruelle 1986)model-independent timing model; on the right, the DDGR (Taylor 1987) timing model, which assumes general relativity to be the correct theory of gravity. TheDD model measures ˙ ω , γ , and ˙ P b , which can each be used to put constraints on the masses of the pulsar and companion. The DDGR model measures the totalmass M total and the companion mass m directly, and the post-Keplerian parameters can be derived from the values of the masses. The errors on these DDGRparameters i , ˙ ω , γ and ˙ P b were derived through a Markov chain Monte Carlo analysis based on the errors of M total and m . Other errors reported here are TEMPO σ values. The DM-derived distance to the pulsar was estimated using the NE2001 model (Cordes & Lazio 2002). Taylor & Weisberg 1989), each parameter constrains the al-lowed ( m , m ) pairs. The intersection of the allowed regionsin m - m parameter space represents the most likely valuesof the pulsar and companion masses. The mass-mass diagramfor our timing solution of J1906+0746 is shown in Fig. 7.The three measured post-Keplerian parameters provide con-sistent values of the pulsar and companion masses. Using theDDGR (Damour & Deruelle 1986, Taylor & Weisberg 1989)binary model, we finally conclude that the pulsar mass m = M (cid:12) and the companion mass m = M (cid:12) .That pair of masses is marked in Fig. 7, and indeed falls withinthe overlap of the constraints from the post-Keplerian param-eters.These masses differ from the initial estimates reportedin Kasian (2008), at m p = 1 . M (cid:12) and m c =1 . M (cid:12) for the pulsar and companion respectively.This change can be explained by the longer data span usedhere, and by the improved method of eliminating the stronglong-term timing noise. DISPERSION MEASURE VARIATIONS
We next investigate changes in dispersion measure, eitheras long-term evolution, or as trends that could recur everyorbit. To estimate such variations within our data span, weused the WAPP data, with its superior wide bandwidth and512-channel spectral information.
Secular DM variation
For each epoch of WAPP data with 3 WAPPS, we used
TEMPO to fit for the DM using only TOAs from that day. Weaccomplished this by using the “DMX” model within
TEMPO ,fitting the individual-day DM values simultaneously with therest of the timing solution. This can bring to light intrinsicDM variations, but it can also absorb other frequency depen-dent effects. Note that the ASP data at these epochs were stilfit with surrounding offsets, because they did not in generalagree with the WAPP DM values. This is presumably dueto the fact that the ASP profiles were obtained with coherentdedispersion but also folded in real-time with some unavoid-able ephemeris smearing. Generally, the observed variationsare larger than the error bars. There is, however, no long- HE Y OUNG , R
ELATIVISTIC B INARY P ULSAR
J1906+0746 9
Figure 7.
Mass-mass diagram for the ephemeris presented in Table 3. Thelines represent the values of m and m allowed by the three measured DDpost-Keplerian parameters: ˙ ω (the solid line), γ (the dotted lines) and ˙ P b (thedashed lines). The dot indicates the best-fit value for m and m . term behaviour that could be included in the timing analysis(Fig. 8). The short-term variations observed in the DM valuesare most likely induced by other TOA changes that system-ically depend on frequency; these could be profile evolutionthat varies with frequency, potentially amplified when scintil-lation affects the relative contributions of different parts of theband. Orbital DM variation
The dispersion measure may vary with orbital phase, ifthe pulsar emission travels through varying plasma densi-ties throughout the binary orbit. Such variations have beenseen in several pulsars with non-degenerate companions (e.g.,B1957+20, Fruchter et al. 1995; and J2051 − MJD D M Figure 8.
Dispersion measure variations versus time, for all WAPP-dataepochs, fit simultaneously with the rest of the timing model.
Figure 9.
DM versus orbital phase for all epochs where good WAPP datawere available for J1906+0746, folded on the Kasian (2012) ephemeris. Thebottom plot shows the DM over orbital phase for all of the epochs together. compact companion, we investigated the behaviour of the DMof J1906+0746 over orbital phase.For each WAPP-data epoch we divided TOAs over 16 binsacross the ∼ TEMPO -doubled errorbars, there are no compelling overall trends in DM over thecourse of an orbit, and no evidence for a more extended com-panion. ORBITAL ABERRATION
We have so-far attributed the observed profile changes to geodetic precession , an effect seen only in strong gravita-tional fields. It is, however, also possible that the specialrelativistic effect of aberration contributes to the observedprofile changes. Aberration on an orbital timescale arisesfrom the relativistic velocities with which the pulsar and com-panion travel in their orbit (Damour & Taylor 1992; Rafikov& Lai 2006), and was measured in the double neutron starB1534+12 (Stairs et al. 2004; Fonseca et al. 2014). As thevelocity at periastron of J1906+0746, v p ∼ . c , is simi-lar to that of B1534+12 (Eq. 8.35; Lorimer & Kramer 2005),one might expect to detect aberration. A longitudinal delaywill shift the pulse profile in phase, while keeping the shapeof the pulse intact; meanwhile, a latitudinal delay shifts theobserved emission angle with respect to the pulsar spin axis(Rafikov & Lai 2006). That latter change in our line of sightcould produce measurable profile changes over the course ofa binary orbit.If detected, we can combine the profile changes due to or-bital aberration with the secular changes due to geodetic pre-cession to put limits on the geometry of the system (Stairset al. 2004). Fitting the polarimetry data for J1906+0746 to0 VAN L EEUWEN ET AL .the classical rotating vector model (RVM) resulted in the fol-lowing measured angles (Desvignes 2009): the current anglebetween the spin and magnetic axes α = 80 ◦ +4 − ; the geodeticprecession phase Φ SO = 109 ◦ +51 − ; and the precession coneopening angle (or misalignment angle) δ = 110 ◦ +21 − . A de-tection of orbital aberration could constrain further angles andwould allow for a measurement of the geodetic precession pe-riod independent of the profile beam model. Observations
The detection of profile changes on an orbital timescale,against a backdrop of steady profile evolution from geode-tic precession, requires several complete orbits of coveragecollected over a relatively short time span. We thus col-lected GBT GASP data during two separate, high-cadencecampaigns - over four days between October 4 and 12, 2006,and over 14 days between March 9 and 23, 2008. Thesecampaigns (cf. §
2) are each much shorter than the ∼
165 yeargeodetic precession period. The fully steerable GBT allowedtracks of the pulsar for at least one full orbit per day, in bothcampaigns.
Method
To measure pulse profile changes versus orbit, separatelyfor each campaign, we used the following approach, adaptedfrom Ferdman et al. (2008). We produced time-averagedpulse profiles per five minute interval, using the ephemerisderived by Kasian (2012). From these, we first created an av-erage profile for the entire campaign. Each five-minute pro-file was next binned by true anomaly, over 8 orbital phases.After scaling all profiles to a uniform height, the bins weresummed. We then investigated the differences, shown in Fig.10, between the total profile and each binned profile.For both the interpulse and the main pulse, we subtracted alinear baseline over their respective data window (the horizon-tal range in the subplots of Fig. 10). We then looked for seriesof 5 subsequent time bins that were offset from the mean bymore than 1 standard deviation.
Results
Only one such instance was found, indicated with the graycircle in Fig. 10, but no related changes were detected at otherorbital phases. Thus no variations on orbital timescales weredetected with this method. The difference profiles, shown inFig. 10, do not show significant other changes over the courseof an orbit. From this we conclude that J1906+0746 has asmall aberration amplitude; and that the long-term profile evo-lution can be used to constrain the emission beam. NATURE OF THE COMPANION
From our timing campaign we find a companion mass m c =1.322(11) M (cid:12) and a pulsar mass m p =1.291(11) M (cid:12) . Asevident from Table 1, these masses well fit the observedcollection of double neutron stars (DNSs), and the standardmodel for DNS evolution, in which the recycled companion ismore massive than - or at least comparable to - the young neu-tron star. The mass of the companion is, however, also similarto that of the massive WD in the relativistic binary with youngpulsar B2303+46 (Table 1). Thus, the masses alone cannotrule out the companion is a WD.For some binary pulsars it is possible to observe the WDcompanion optically; however, as discussed in Lorimer et al. (2006), this is not a viable option for J1906+0746. As detailedin the next section, we would expect a WD companion to havean age of at least ∼ b = . ◦ ) the modeled extinction A V out to d HI =7.4 kpc can range from 4.1–8.4 (model A–S in Amˆores &L´epine 2005). That suggests a V magnitude of at least 29, toofaint to reasonably detect. Optical studies of the companioncan therefore not confirm or rule out its nature.If J1906+0746 is part of a double neutron star binary, it maybe possible to detect the companion as a second radio pulsar– as seen in the double pulsar system, J0737 − SIGPROC time series, dedispersed at the DM of the knownpulsar, was transformed into the companion rest frame for arange of possible projected semimajor axes. We then searchedfor periodicities in these data, and folded the transformed timeseries at the candidate periods. No convincing pulsar signalswere found. If the companion is a pulsar, it is either beamedaway from Earth; or too dim. The minimum flux density S min that we could have detected was: S min = β σ min [ T sys + T sky ( l, b )] G (cid:112) n p t obs ∆ ν (cid:114) W e P − W e (4)(Dicke 1946; Dewey et al. 1985) where σ min = 8 is thethreshold detection signal-to-noise ratio (SNR); n p = 2 isthe number of summed polarizations, ∆ ν is the bandwidth(600 MHz with the Spigot; 3 ×
100 MHz for the WAPPs), β ∼ . is the quantization factor for 3-level quantization (Lorimer& Kramer 2005), G is the antenna gain (2.0 K/Jy for the L-band receiver at the GBT and 10 K/Jy for the L-wide receiverat Arecibo ), T sys is the system temperature (20 K and 25 Kfor the GBT and Arecibo, respectively), T sky ( l, b ) is the tem-perature of the sky at the location of the source (Haslam et al.1982), W e is the effective pulse width of the pulsar, and P is the pulse period. For an integration time of t obs ∼ S min (cid:39) µ Jy. For our ∼ S min (cid:39) µ Jy. At the HI-absorption dis-tance of 7.4 kpc, any companion pulsar beamed toward usthus has a pseudoluminosity S d < . Com-paring this with the L-band pseudoluminosities of the recy-cled pulsars in known double neutron stars (1.9 mJy kpc for J0737 − for B1534+12, and 45 mJykpc for B1913+16 – Burgay et al. 2006; Deller et al. 2009a;Kramer et al. 1998; Taylor & Cordes 1993) we conclude thatour search would have detected 2 out of 3 of these at the HIdistance of J1906+0746, and thus had sufficient sensitivity todetect the average known recycled pulsar in a DNS, if its beamintersected Earth.If the opening angle between the spin axis of this neutronstar companion and the angular momentum of the orbit islarge enough, the putative recycled pulsar will become visi-ble within a geodetic precession timescale (Eq. 1). Continuedfollow up and search for pulsations may thus prove the com-panion is a neutron star.Without a direct optical or radio detection at the mo- http://naic.edu/˜astro/RXstatus/Lwide/Lwide.shtml HE Y OUNG , R
ELATIVISTIC B INARY P ULSAR
J1906+0746 11
Figure 10.
The pulse and interpulse average profile of J1906+0746 (top row), and the difference between these total averages, and the average for 8 orbitalphase bins (bottom rows). Data were taken over four days in October 2006 (left) and three days in March 2008 (right). The interpulse flux has been magnified8-fold relative to the main pulse. The top row shows the average profiles per campaign. The subpanels below shows the difference of the profile at orbital phasecompared to the average. ment, the nature of the companion remains best investigatedby comparing the J1906+0746 system masses to the col-lection of known DNS and relativistic WD binaries withprecise mass estimates (Table 1). The companion mass m c = 1 . M (cid:12) is likely higher than that of the mostmassive known similar WD, the . +0 . − . M (cid:12) companion toPSR B2303+46. The companion mass reported does, how-ever, fall well within the mass range of the recycled stars inknown DNSs. IMPLICATIONS AND CONCLUSION
We presented an updated timing solution for J1906+0746that allows for the measurement of three post-Keplerian pa-rameters, ˙ ω , γ and ˙ P b . We measured pulsar and companionmasses of m p = M (cid:12) and m c = M (cid:12) , re-spectively, compatible with a neutron-star or possibly a white-dwarf companion.If the binary companion to this young unrecycled pulsar is aWD, it must have formed first. Such systems are observation-ally rarer by an order of magnitude than NS-WD systems inwhich the NS is recycled. They also require a different masstransfer history than the average binary pulsar.The existence of young pulsars in binaries around WDs waspredicted by Dewey et al. (1985) and Tutukov & Yungelson(1993), and was subsequently confirmed by the detections ofthe 12-day binary B2303+46 (Thorsett et al. 1993; later iden-tified as a NS-WD system by observation of the WD compan-ion; van Kerkwijk & Kulkarni 1999) and the relativistic binaryJ1141 − can explain this class is out-lined in Tauris & Sennels (2000): the binary progenitor in-volves a primary star with mass between and M (cid:12) anda secondary with initial mass between and M (cid:12) . The pri-mary evolves and overflows its Roche lobe, and the secondaryaccretes a substantial amount of mass during this phase, whichlasts ∼ − ˙ P b (Kramer et al. 2006;Weisberg et al. 2010), only very little dipolar gravitational-wave emission is predicted by alternative theories of gravity.Thus, for a pulsar-WD system, the closer the agreement ofthe observed ˙ P b with the value predicted by general relativ-ity, the stronger are the constraints on any alternative theoriesthat predict extra gravitational wave emission. Currently, thebest limits on such alternatives come from the measurementof the orbital decay of the millisecond pulsar - WD systemPSR J1738+0333 (Freire et al. 2012). The main limitationin the precision of this test is the precision (and accuracy)of the masses of the components of that system, which wasderived from optical spectroscopy and is limited, to some ex-tent, by uncertainties in the atmospheric models used (An-toniadis et al. 2012). This is not so much a limitation forPSR J1906+0746, where the masses are known to very goodrelative precision from the measurement of ˙ ω and γ . There-fore, if the companion of PSR J1906+0746 is indeed a WD,then this system could in principle provide a test of alterna-tive theories of gravity that is much superior to any currenttest. However, for that to happen, three conditions must befulfilled: First, the companion should be confirmed to be aWD, as Antoniadis et al. (2011) did for the companion ofPSR J1141 − ˙ P b ( § VAN L EEUWEN ET AL .expected to be very faint at optical wavelengths ( § § τ c = 112 . kyrs old (Table 3). Therefore the current eccen-tricity e = 0 . , the lowest of any of the knownDNSs, must reflect the state of the orbit after the second su-pernova. This implies a small supernova kick to the newbornyoung pulsar J1906+0746, well within our upper limit on thesystem velocity.For the general population of relativistic binaries there is aselection effect favoring the detection of such low-eccentricitysystems: high eccentricities greatly increase the emission ofgravitational radiation, and those systems quickly coalesceafter their orbits have decayed (Chaurasia & Bailes 2005).But for the detection of young systems such as J1906+0746that selection effect has not yet developed. Only binarieswith eccentricities e > . can expect to merge withinJ1906+0746’s age of τ c = 112 . kyrs (Peters 1964).The low eccentricity and system velocity, combined withthe relatively low mass of J1906+0746, suggest it was formedin an electron-capture, O-Ne-Mg supernova (van den Heuvel2007). In such a case, the spin axis of the recycled pulsaris more likely to still be aligned with the orbital angular mo-mentum, in which case it will show little geodetic precession( § does precessinto view.In conclusion, we currently cannot confirm with certaintyor rule out that the companion of J1906+0746 is a neutronstar; and given the fast decline in pulse flux due to geodeticprecession, we will likely not improve on our timing solutionuntil the pulsar precesses back into view. Pulsar J1906+0746is likely in a binary containing a double neutron star; or it isorbited by a white dwarf, in a system formed by through anexotic binary interaction involving two stages of mass trans-fer. ACKNOWLEDGMENTS
We thank Jeroen Stil for advice on interpreting the VGPSdata, Bryan Gaensler and Avinash Deshpande for discussionson the kinematic absorption method, and Lindley Lentati forcomparing timing algorithms.The Arecibo Observatory is operated by SRI Interna-tional under a cooperative agreement with the National Sci-ence Foundation (AST-1100968), and in alliance with AnaG. M´endez-Universidad Metropolitana, and the UniversitiesSpace Research Association.The National Radio Astronomy Observatory is a facility ofthe National Science Foundation operated under cooperativeagreement by Associated Universities, Inc.The Westerbork Synthesis Radio Telescope is operated byASTRON with support from the Netherlands Foundation forScientific Research NWO.The Nanc¸ay radio Observatory is operated by the ParisObservatory, associated to the French Centre National de laRecherche Scientifique (CNRS).This work was supported by European Commission GrantFP7-PEOPLE-2007-4-3-IRG-224838 (JvL) and by U.S. Na-tional Science Foundation Grants AST-0647820 (DJN) andAST-0807556/AST-1312843 (JMW). LK acknowledges the support of a doctoral Canada Graduate Scholarship. Pulsarresearch at UBC is supported by an NSERC Discovery Grantand by the Canada Foundation for Innovation.REFERENCES
Alpar, M. A., Cheng, A. F., Ruderman, M. A., & Shaham, J. 1982, Nature,300, 728Amˆores, E. B., & L´epine, J. R. D. 2005, AJ, 130, 659Antoniadis, J., Bassa, C. G., Wex, N., Kramer, M., & Napiwotzki, R. 2011,MNRAS, 412, 580Antoniadis, J., van Kerkwijk, M. H., Koester, D., et al. 2012, MNRAS, 423,3316Barker, B. M., & O’Connell, R. F. 1975, ApJ, 199, L25Bhat, N. D. R., Bailes, M., & Verbiest, J. P. W. 2008, Phys. Rev. D, 77,124017Breton, R. P., Kaspi, V. M., Kramer, M., et al. 2008, Science, 321, 104Burgay, M., Joshi, B. C., D’Amico, N., et al. 2006, MNRAS, 368, 283Champion, D. J., Lorimer, D. R., McLaughlin, M. A., et al. 2005, MNRAS,363, 929Chatterjee, S., Brisken, W. F., Vlemmings, W. H. T., et al. 2009, ApJ, 698,250Chaurasia, H. K., & Bailes, M. 2005, ApJ, 632, 1054Clifton, T. R., Frail, D. A., Kulkarni, S. R., & Weisberg, J. M. 1988, ApJ,333, 332Cordes, J. M., & Lazio, T. J. W. 2002, preprint (arXiv:astro-ph/0207156)Cordes, J. M., Freire, P. C. C., Lorimer, D. R., et al. 2006, ApJ, 637, 446Corongiu, A., Kramer, M., Stappers, B. W., et al. 2007, A&A, 462, 703Damour, T., & Deruelle, N. 1986, Ann. Inst. H. Poincar´e (PhysiqueTh´eorique), 44, 263Damour, T., & Esposito-Farese, G. 1996, Phys. Rev. D, 54, 1474Damour, T., & Ruffini, R. 1974, Academie des Sciences Paris ComptesRendus Ser. Scie. Math., 279, 971Damour, T., & Taylor, J. H. 1991, ApJ, 366, 501—. 1992, Phys. Rev. D, 45, 1840Deller, A. T., Bailes, M., & Tingay, S. J. 2009a, Science, 323, 1327Deller, A. T., Tingay, S. J., Bailes, M., & Reynolds, J. E. 2009b, ApJ, 701,1243Deller, A. T., Brisken, W. F., Phillips, C. J., et al. 2011, PASP, 123, 275Demorest, P. B. 2007, PhD thesis, University of California, BerkeleyDesvignes, G. 2009, PhD thesis, l’Universit´e d’Orl´eansDewey, R. J., Taylor, J. H., Weisberg, J. M., & Stokes, G. H. 1985, ApJ, 294,L25Dicke, R. H. 1946, Rev. Sci. Instrum., 17, 268Dowd, A., Sisk, W., & Hagen, J. 2000, in Pulsar Astronomy - 2000 andBeyond, IAU Colloquium 177, ed. M. Kramer, N. Wex, & R. Wielebinski(San Francisco: Astronomical Society of the Pacific), 275Ferdman, R. D. 2008, PhD thesis, University of British ColumbiaFerdman, R. D., Stairs, I. H., Kramer, M., et al. 2008, in American Instituteof Physics Conference Series, Vol. 983, 40 Years of Pulsars: MillisecondPulsars, Magnetars and More, ed. C. Bassa, Z. Wang, A. Cumming, &V. M. Kaspi, 474Ferdman, R. D., Stairs, I. H., Kramer, M., et al. 2014, MNRAS, 443, 2183Fich, M., Blitz, L., & Stark, A. A. 1989, ApJ, 342, 272Fonseca, E., Stairs, I. H., & Thorsett, S. E. 2014, ApJ, 787, 82Frail, D. A., & Weisberg, J. M. 1990, AJ, 100, 743Freire, P. C. C., Wex, N., Esposito-Far`ese, G., et al. 2012, MNRAS, 423,3328Fruchter, A. S., Bookbinder, J., & Bailyn, C. 1995, ApJ, 443, L21Haslam, C. G. T., Stoffel, H., Salter, C. J., & Wilson, W. E. 1982, A&AS,47, 1Hobbs, G., Lorimer, D. R., Lyne, A. G., & Kramer, M. 2005, MNRAS, 360,974Hobbs, G., Lyne, A. G., Kramer, M., Martin, C. E., & Jordan, C. 2004,MNRAS, 353, 1311Hobbs, G. B., Edwards, R. T., & Manchester, R. N. 2006, MNRAS, 369, 655Hotan, A. W., Bailes, M., & Ord, S. M. 2005, ApJ, 624, 906Hotan, A. W., van Straten, W., & Manchester, R. N. 2004, PASA, 21, 302Jacoby, B. A., Cameron, P. B., Jenet, F. A., et al. 2006, ApJ, 644, L113Janssen, G. H., Stappers, B. W., Kramer, M., et al. 2008, A&A, 490, 753Johnston, S., Koribalski, B., Weisberg, J. M., & Wilson, W. 2001, MNRAS,322, 715Kaplan, D. L., Escoffier, R. P., Lacasse, R. J., et al. 2005, PASP, 117, 643Karuppusamy, R., Stappers, B., & van Straten, W. 2008, PASP, 120, 191 HE Y OUNG , R
ELATIVISTIC B INARY P ULSAR
J1906+0746 13
Kasian, L. 2008, in American Institute of Physics Conference Series, Vol.983, 40 Years of Pulsars: Millisecond Pulsars, Magnetars and More, ed.C. Bassa, Z. Wang, A. Cumming, & V. M. Kaspi, 485Kasian, L. 2012, PhD thesis, University of British Columbia, https://circle.ubc.ca/handle/2429/41515
Kaspi, V. M., Lackey, J. R., Mattox, J., et al. 2000, ApJ, 528, 445Keith, M. J., Kramer, M., Lyne, A. G., et al. 2009, MNRAS, 393, 623Kramer, M. 1998, ApJ, 509, 856Kramer, M., Wielebinski, R., Jessner, A., Gil, J. A., & Seiradakis, J. H.1994, A&AS, 107, 515Kramer, M., Xilouris, K. M., Lorimer, D. R., et al. 1998, ApJ, 501, 270Kramer, M., Stairs, I. H., Manchester, R. N., et al. 2006, Science, 314, 97Kulkarni, S. R., & van Kerkwijk, M. H. 1998, ApJ, 507, L49Lorimer, D. R. 2008, Living Reviews in Relativity,http://relativity.livingreviews.org/Articles/lrr-2008-8/Lorimer, D. R., & Kramer, M. 2005, Handbook of Pulsar Astronomy(Cambridge University Press)Lorimer, D. R., Stairs, I. H., Freire, P. C., et al. 2006, ApJ, 640, 428Lyne, A., Hobbs, G., Kramer, M., Stairs, I., & Stappers, B. 2010, Science,329, 408Lyne, A. G., Burgay, M., Kramer, M., et al. 2004, Science, 303, 1153Manchester, R. N., Hobbs, G. B., Teoh, A., & Hobbs, M. 2005, AJ, 129,1993Nice, D. J., & Taylor, J. H. 1995, ApJ, 441, 429Peters, P. C. 1964, Physical Review, 136, 1224Rafikov, R. R., & Lai, D. 2006, Phys. Rev. D, 73, 063003Reid, M. J., Menten, K. M., Brunthaler, A., et al. 2014, ArXiv e-prints, arXiv:1401.5377 [astro-ph.GA]arXiv:1401.5377 [astro-ph.GA]