ApJ,accepted PreprinttypesetusingLATEXstyleemulateapjv.10/09/06 PROPERMOTIONS OF PSRS B1757−24AND B1951+32: IMPLICATIONS FOR AGES AND ASSOCIATIONS B. R. Zeiger1, W. F. Brisken2, S. Chatterjee3,4, W. M. Goss2 ApJ, accepted ABSTRACT Over the last decade, considerable effort has been made to measure the proper motions of the pulsars B1757−24 and B1951+32 in order to establish or refute associations with nearby supernova remnants and to understand better the complicated geometries of their surrounding nebulae. We 8 present proper motion measurements of both pulsars with the Very Large Array, increasing the time 0 0 baselinesofthemeasurementsfrom3.9yrto6.5yrandfrom12.0yrto14.5yr,respectively,compared 2 to previous observations. We confirm the nondetection of proper motion of PSR B1757−24,and our n measurementof(µα,µδ)=(−11±9,−1±15)masyr−1confirmsthattheassociationofPSRB1757−24 with SNR G5.4−1.2 is unlikely for the pulsar characteristic age of 15.5 kyr, although an association a cannot be excluded for a significantly larger age. For PSR B1951+32, we measure a proper motion J 4 of (µα,µδ) =(−28.8±0.9,−14.7±0.9) mas yr−1, reducing the uncertainty in the proper motion by a factor of 2 compared to previous results. After correcting to the local standard of rest, the proper 2 motionindicatesakineticageof∼51kyrforthepulsar,assumingitwasbornnearthegeometriccenter of the supernova remnant. The radio-bright arc of emission along the pulsar proper motion vector ] h shows time-variable structure, but moves with the pulsar at an approximately constant separation p ∼2.5′′, lending weight to its interpretation as a shock structure driven by the pulsar. - o Subjectheadings: pulsars: individual(PSRB1757−24,PSRB1951+32)—ISM:individual(G5.4−1.2, r CTB 80) — stars: neutron — supernova remnants t s a [ 1. INTRODUCTION decelerated by the ISM, making identification of pulsar- Neutron stars (NSs) are born from the core collapse SNR pairs problematic. 2 Thus, cases where a young radio pulsar can be as- v of massive stars, whose deaths are marked by super- sociated with a SNR are particularly important. The 4 nova remnants (SNRs). Measuring the proper motions pulsars B1757−24 and B1951+32 present two such sit- 4 of young NSs allows their trajectoriesto be traced back, uations in which associations might be possible. Both 2 and can thus provide strong tests of proposed NS-SNR areyoungpulsarswithhighspin-downenergy-lossrates, 0 associations. In cases where such associations can be and in both cases, the ram pressure balance between . confirmed, an independent age estimate can be derived 1 the NS relativistic wind and the ISM produces a pul- for both the NS and its associated SNR. 0 sar wind nebula (PWN) with a bow shock structure. 8 However,suchanexerciseis subjectto severalpitfalls. PSRB1757−24appears to be exiting the approximately 0 On one hand, a typical radio pulsar might be detectable : for & 107 yr, while the associated SNR fades into the circularSNRG5.4−1.2,whoseasymmetricbrightness,in v interstellar medium (ISM) in . 105 yr, leaving behind conjunctionwith the PWN producedbythe pulsar,pro- i duces the structure known as “the Duck” (see Fig. 1). X young pulsars with no detectable SNRs. On the other Meanwhile,PSRB1951+32drivesacomplexinteraction hand,therearealsoseveralSNRswithnoassociatedpul- r withinthe SNR CTB 80(G69.0+2.7;see Fig.2), includ- a sars. NotallNSsareradiopulsars,butwhenyoung,they ing a radio-bright structure resembling a bow shock in are all likely to be bright in thermal X-ray emission, as the direction of its motion. illustrated,forexample,bythedetectionofpointsources WepresentVeryLargeArray(VLA)observationsthat intheSNRsKes73(Vasisht & Gotthelf1997)andKes79 improve the time baseline for the proper motion mea- (Seward et al. 2003). The absence of such X-ray detec- surements of these two pulsars. In §2 we detail the ob- tionsinasetofnearby,youngSNRs(Kaplan et al.2004, servations,analysis,and results for PSR B1757−24,and 2006) aggravates the problem. While unusual cooling likewise for PSR B1951+32 in §3. We discuss the impli- scenarios may be required for NSs in these young SNRs, cations of our results in §4. a partial explanation for the “missing” NSs might lie in their birth velocities. Radio pulsars have characteristic 2. PSRB1757−24: OBSERVATIONSANDRESULTS birth velocities ∼ 400-500 km s−1 (Arzoumanian et al. PSR B1757−24 is an energetic young pulsar, with 2002; Hobbs et al. 2005), and they are likely to escape a period P = 125 ms, a spindown energy loss rate their natal remnants once the expansion of the SNR is E˙ = 1036.4 erg s−1, and a characteristic age τ ≡ c 1Center for Astrophysics and Space Astronomy, University of P/2P˙ = 15.5 kyr. As shown in Figure 1, its location Colorado,Boulder,CO80304; [email protected] andthe morphologyofits PWNsuggestthatit is escap- 2National Radio Astronomy Observatory, Socorro, NM 87801; ing the circular SNR G5.4−1.2 (Frail & Kulkarni 1991; wbrisken,[email protected] 3SchoolofPhysics,TheUniversityofSydney,NSW2006,Aus- Manchester et al. 1991). If it was born at the center of tralia;[email protected] the remnant and its characteristic age is close to its real 4JanskyFellow age, that would imply a proper motion ∼ 70 mas yr−1 2 Zeiger et al. and a transverse velocity ∼1500-2000 km s−1, high images. Twoself-calibrationandimagingiterationswere compared to the radio pulsar population. However, performed using all 14 sub-regions to generate the final Gaensler & Frail (2000) placed an upper limit on the images used for the astrometry. motion of the PWN that was inconsistent with such Positions of reference sources and the pulsar were de- a high velocity, and suggested that the true age of termined with a Gaussian fit using the AIPS task JM- the pulsar was much larger than the characteristic age. FIT.Thepropermotionofthepulsarwasmeasuredwith Thorsett et al. (2002) derived comparable limits on the respect to a reference frame defined by six point sources proper motion5 of the pulsar (rather than the PWN) of chosen based on compactness from the 13 imaged sub- (µ ,µ )=(−2.1±7.0,−14±13)masyr−1andsuggested regions. The reference source positions and fluxes from α δ instead that the proximity of B1757−24 and G5.4−1.2 the (ungated) 2004 December 09 epoch are listed in Ta- wasmerely a line-of-sightcoincidence in the crowdedre- ble 2, and our method is described in greater detail in gionneartheGalacticcenter. Argumentsforandagainst McGary et al. (2001). ′′ an association between the pulsar and SNR were sum- The 10 PWN in which the pulsar is embedded is re- marized by Blazek et al. (2006), who improved the up- solvedoutby ourobservations,but it contributesan ad- per limit on the westward proper motion of the PWN ditional uncertainty to the position of the pulsar in the and concluded that both interpretations for the Duck (a ungated data. To account for the ∼50% increase in rms line-of-sight coincidence and a genuine association with noise within this region relative to the background, the a large age) remained viable. pulsar position uncertainties were increased in ungated Based on the dispersion measure, PSR B1757−24 epochs by a factor of 1.5. In the gated data, the PWN is estimated to lie at a distance of 5.2±0.5 kpc contribution to the noise is insignificant and the posi- (Cordes & Lazio 2002), while the distance to G5.27−1.2 tion uncertainties of the reference sources dominate the is greater than 4.3 kpc based on H I absorption proper motion uncertainty. (Frail et al. 1994; Thorsett et al. 2002). Here we adopt 2.2. Results a distance of 5d kpc. 5 The structure of the PWN around the pulsar is un- 2.1. Observations certain and could cause a systematic offset in the mea- surementofthepulsarposition,affectingregistrationbe- VLA observations of PSR B1757−24 spanned 6.4 yr, tween the gated and ungated data sets. Thus, we mea- from1998to2004. The1998June03epochwasobserved sure the proper motion of the pulsar separately for the usingthe BnAconfiguration;allofthe later epochsused gated and ungated data and combine the two indepen- the A configuration. At the frequency of observation, dent results for a final proper motion value. Using un- 1.45 GHz, the A configuration resolution at a declina- tion of −25◦ is ∼1.2′′×2.2′′, elongated north-south. The (g−at1e6d.9d±at1a2,.6w,e−m35e.a8s±ure22a.4p)rmopaesrymr−o1t;iownitohfg(aµtαed,µdδa)t=a, first three epochs employed the pulsar gate on the right wemeasure(µ ,µ )=(−5.9±11.8,6.0±20.3)masyr−1. circular polarizationto allow increased sensitivity in the α δ Combiningtheresultsgivesameasurementof(µ ,µ )= pulsar measurements by only accepting data when the (−11±9,−1±15)masyr−1,consistentwithanoαndeδtec- pulsar was “on.” The 2002 February 21, 2002 May 02, tion. Note that correction to the local standard of rest and 2004 December 09 epochs included the Pie Town (LSR) for the pulsar involvesadjustments at the level of VLBA antenna, more than doubling the maximum east- 1 mas yr−1. Since these are much lower than the uncer- west baseline length. The ungated data from the 2002 tainties, we ignore these corrections for B1757−24. February 21 epoch was not used, as the observation was We note that our results and those of Thorsett et al. not long enough to provide sufficient u-v coverage and (2002) are not independent, since two epochs of data sensitivity needed for clear identification of the pulsar. are common between the two analyses,although the Pie In order to minimize variation in u-v coverage between Town link data were not used in the present work. The epochs and in contrast to Thorsett et al. (2002), we do measurement of Blazek et al. (2006) is completely in- notusethePieTownbaselines. Observationsweremade dependent, since they observe the nebula and not the in the ‘2AD’ correlator mode, offering 15 spectral chan- pulsar. However, only one dimension of motion was nels each of width 1.56 MHz in both circular polariza- tions andallowinga usable fieldof view of∼30′. Details probed by these authors. We derive a 68% confidence upper limit on the westward motion of the pulsar, µ > of these observations are shown in Table 1. −14.9 mas yr−1 (µ > −24.8 mas yr−1 at 95% coαnfi- The data were calibrated using standard procedures α dence), which corresponds to a westward transverse ve- wweitrheinustehdeaAsIflPuSx6dpeancskiatygea.ndSobuarncedspa3sCs c2a8l6ibarnatdor3sCan4d8 locity v⊥ < 360d5 km s−1 (v⊥ < 600d5 km s−1 at 95% confidence). The implications are discussed further be- 1751−253wasusedforphasecalibration. TheAIPStask low, but we note that such velocity limits are consistent UVFIXwasusedtorecalculatetheUVWcoordinatesin- with the velocity distribution of ordinary young radio corporatingaberrationcorrectionsnot performed by the pulsars (Arzoumanian et al. 2002; Hobbs et al. 2005). VLAonlinesystem. Giventhelargefieldofviewandthe sparsedistributionofsourceswithcompactstructure,we 3. PSRB1951+32: OBSERVATIONSANDRESULTS selected14sub-regionsofthefield(containingthepulsar PSR B1951+32, a 39.5 ms pulsar with τ = 107 kyr c and 13 other apparently compact sources) for imaging. anda spin-downenergy-lossrateE˙ =1036.6 ergs−1, lies These fields were jointly deconvolved to produce initial near the southwestern edge of the approximately circu- lar infrared shell of CTB 80 (Fesen et al. 1988). The 5 All uncertainties are 68% confidence intervals, except where pulsar is surrounded by a ∼30′′ diameter asymmetric both68%and95%intervalsareexplicitlyreported. 6 Seehttp://www.aips.nrao.edu/ PWN at the western edge of an 8′×4′ east-west plateau Proper Motions of PSRs B1757−24and B1951+32 3 Fig. 1.— The Duck: the SNR G5.4−1.2, G5.27−0.90, and B1757−24 system. Left: Large-scale structure of the remnant, as observed bytheVLAat327MHzwithanangularresolutionof40′′×23′′ (Broganetal.2006,imagedatacourtesyoftheauthors). Theextended circular structure is SNR G5.4−1.2, while G5.27−0.90 is the “head” of the Duck protruding from the west edge of the remnant. The small white square shows the approximate location of the right-hand panel. Right: Pulsar B1757−24 and its PWN (the western tip of G5.27−0.90), asobservedat1.4GHzin2002. Thelarge-scalestructuresareresolvedoutbytheinterferometer response,leavingonlythe “beak”visibleatanangularresolutionof2′.′2×1′.′2. TABLE 1 ObservationalParametersforB1757−24 Epoch Obs. Poln. Gated Fluxdensity rmsnoise Beam Pos. Tint Code (mJy) (mJybeam−1) (′′) angle (hr) 1998Jun03 AF336 R Y 5.0 0.3 2.3×1.2 16◦ 1.7 L N 1.9 0.1 2.2×1.2 16◦ 1.7 2001Jan01 AB969 R Y 3.4 0.2 2.3×1.2 −1◦ 3.2 L N 1.1 0.1 2.2×1.2 0◦ 3.2 2002Feb21 AB1029 R Y 3.4 0.2 2.0×1.2 5◦ 1.5 2002May02 AB1029 I N 4.5 0.1 2.2×1.2 4◦ 2.3 2004Dec09 AB1139 I N 1.3 0.1 2.3×1.2 0◦ 3.5 Note. — Allobservations areat afrequency of1.45 GHz. Pulsar gating improves the apparent pulsarfluxdensitybyafactor.[(Ton+Toff)/Ton]1/2. surements,whileMigliazzo et al.(2002)directlymeasure TABLE 2 apropermotionµ=25±4masyr−1 ata positionangle ReferenceSourcesUsed in theProper 252◦ ±7◦ after correcting for the effects of differential Motion Fit of B1757−24 Galactic rotation. Their implied transverse velocity of R.A. Decl. Fluxdensity θsep the pulsar is V⊥ =(240±40)d2 km s−1. (J2000) (J2000) (mJybeam−1) (′) In the direction of the proper motion vector measured by Migliazzo et al. (2002), Moon et al. (2004) observe 180127.40 −250738.1 8.7(2) 17.3 a cometary X-ray synchrotron nebula and an Hα bow 180158.57 −245548.5 6.6(2) 14.0 180047.31 −244245.0 2.6(1) 9.2 shock. The X-ray emission peaks at the pulsar loca- 180015.71 −250022.0 3.6(1) 13.4 tion and is confined within the Hα structure (Hester 180128.11 −243423.7 1.2(2) 18.2 2000) that is clearly defined at an angular separation of 180041.14 −244204.4 4.9(1) 10.3 ∼7′′ from the pulsar. The X-rays are produced by syn- Note. —Unitsofrightascensionarehours,min- chrotronemission from the pulsar wind, confined within utes, and seconds, and units of declination are de- a bow shock produced by the ram pressure of the wind grees,arcminutes,andarcseconds. interacting with the SNR wall. Inradioimages,B1951+32appearstobesituatedjust of emission (Castelletti et al. 2003). CTB 80, observed inside a limb-brightened bubble ∼30′′ in diameter (see at 1.4 GHz to have three arms that cover 1.8 deg2 and Fig. 2, right). A 5′′ portion of the bubble nearest to the convergenearB1951+32(Castelletti et al.2003), hasan pulsar is substantially brighter than any other portion expandingHIshellthatyieldsadynamicalageestimate of the shell. In MERLIN observations at 1.6 GHz with ′′ of 77 kyr for the pulsar-SNR system (Koo et al. 1990). 0.15 resolution,Golden et al.(2005)findthattheradio- Fromitsdispersionmeasure,thedistancetoB1951+32is bright arc shows compact structure, resembling a radio estimated as 3.1±0.2 kpc (Cordes & Lazio 2002), while bowshock. Thestructureis∼2.5′′fromthepulsar,which the distance to CTB 80 is estimated as 2 kpc from H I puts it at the head of the cometary X-ray nebula, but absorption (Strom & Stappers 2000). Here we adopt a enclosed within the Hα bow shock which is ∼7′′ away distance of 2d kpc. Kulkarni et al. (1988) estimated a from the pulsar (Moon et al. 2004). 2 speedof300kms−1forthepulsarfromscintillationmea- 4 Zeiger et al. Fig. 2.—CTB80(G69.0+2.7)andthePWNproducedbyPSRB1951+32. Left: Infraredratioimage(60µm/100µm) constructedfrom theIRASarchiveshowsthenearlycompleteshellofCTB80. Overplottedlinesindicatetheprojectedpathofthepulsar,projectedbackin timewiththe1σ uncertaintyonthemeasuredpropermotion,correctedtotheLSR.Circlesindicatetheestimatedbirthlocationsforage estimates ofτp ∼51 kyr(the age ofclosest approach to theSNR geometric center, as shownwiththe white cross)andthe characteristic (spin-down) age τc = 107 kyr. The projected path does not pass close to the expansion center obtained by Kooetal. (1990), indicated by the white triangle. Right: VLA 1.4 GHz image of B1951+32 and its PWN, with an angular resolution of 1′.′0×0′.′9 and rms noise 0.03mJybeam−1. Thearrowrepresents300yroftravelalongthepropermotionvector(see§3.2). Thebowshockisvisibleasthebright arctothesouth-westofthepulsar. Notethedifferenceinscalebetween thetwopanels. 3.1. Observations MeasurementsofthepropermotionofPSRB1951+32 were made using five epochs of archival VLA data from 1989.04to2003.50. ObservationdetailsareshowninTa- ble3. AllobservationsweremadewiththeVLAAarray except the 2000November 25 epoch, which included the Pie Town antenna. To maintain consistent u-v coverage in each epoch, the baselines to Pie Town were removed before data were reduced. Observations were made with 15 channels of 1.56 MHz each. Data were calibrated with VLA calibrator sources 3C 286 (for flux density andbandpasscalibration)andJ1925+211(forphasecal- ibration), following the method outlined in §2.1. Nine- teen sources with compact structure were found in the Fig. 3.— Proper motion of PSR B1951+32 and the reference field of view and were included in the imaging and self- sources usedforthe fit. Referencesource motions withrespect to eachotherareconsistentwithnoisewithandscatteredaroundzero. calibrationprocess. Eightpointlikesources(seeTable4) Thepulsarliesinthebottomrightcornerwithapropermotionof were used in the determination of the proper motion in (µα,µδ)=(−28.8±0.9,−14.7±0.9)masyr−1. Inthisfigurethe the manner described in §2.1. derivedmotionofthepulsarisrelativetotheframedefinedbyall The pulsar B1951+32 lies in a complicated region of eightreferencesources. the CTB 80 SNR and is embedded in its PWN, making A significant correction is required to obtain the mo- the position fit sensitive to observation parameters such tion of the pulsar in its LSR so that it can be related to as observing frequency and u-v coverage at each epoch. the remnant. We use a flat Galactic rotation curve with Toaccountfor′t′histofirstorder,alinearbrightnessgra- velocity220kms−1 andassumeasolarradiusof8.5kpc. dient in a 1.75 square around the pulsar was fit simul- Furthercorrectionforthesolarpeculiarmotionisapplied taneously with the pulsar position. Trial fits performed (Dehnen & Binney 1998). The magnitude of the correc- without this procedure resulted in substantially greater tiondependsonthedistance;anadditional0.6masyr−1 scatter in the position measurements. Additional trials has been added in quadrature to the proper motion demonstrated that the fit was not very sensitive to the uncertainty to accommodate the magnitude range of exact size and position of the region fitted. the correction over the 1–3 kpc distance range consid- 3.2. Results ered. The LSR-corrected proper motion is found to be (µ′ ,µ′)=(−26.9±1.1,−10.5±1.1)masyr−1,foratotal The proper motion of B1951+32 was measured as α δ (µ ,µ )=(−28.8±0.9,−14.7±0.9)mas yr−1. Figure 3 proper motion µ′ =29±1 mas yr−1 at a position angle α δ 249±2◦ east of north. demonstrates the significance of the proper motion rel- ative to that of the eight presumably stationary sources 3.3. Shock Separation usedtodefine the frameinwhichthe propermotionwas measured. For each of the reference sources, we derive To maximize sensitivity to the extended structure of a proper motion in the frame defined by the seven other theshock,allu-vdatawithineachepochwereaddedwith sources. As shown in Figure 3, these proper motion val- the AIPS task DBCON. Before reimaging, the resulting ues are consistent with zero at ∼1 σ, indicating that a data were convolved with a circular beam with a 1.2′′ self-consistent reference frame has been established. diameter,thelongestsemi-majoraxisofanybeaminany Proper Motions of PSRs B1757−24and B1951+32 5 TABLE 3 ObservationalParametersforB1951+32 Epoch Obs. IF Pol. Freq. Fluxdensity rmsnoise Beam Position Tint Code (MHz) (mJy) (mJybm−1) (′′) Angle (hr) 1989Jan13 AS357 1 RR 1385 1.01 0.07 1.1×1.1 50◦ 5.9 2 LL 1652 0.62 0.09 1.0×0.9 81◦ 5.9 1991Jul18 AF214 1 RR 1385 1.95 0.08 1.3×1.2 −47◦ 5.8 2 LL 1652 1.32 0.08 1.0×0.9 −74◦ 5.8 1993Jan08 AF235 1 RR 1385 1.75 0.06 1.2×1.2 −64◦ 5.9 2 LL 1652 1.79 0.05 1.1×1.0 −58◦ 5.9 2000Nov25 AG602 1 RR,LL 1385 0.63 0.05 1.3×1.2 −69◦ 5.2 2 RR,LL 1516 0.91 0.07 1.1×1.0 −75◦ 5.2 2003Jun03 AG650 1 RR,LL 1665 1.05 0.03 1.0×0.9 −70◦ 6.2 TABLE 4 ReferenceSourcesUsed in theProper Motion Fit of B1951+32 R.A. Decl. Fluxdensity θsep (J2000) (J2000) (mJybeam−1) (′) 195323.70 +325335.3 4.7(3) 5.4 195316.37 +324846.5 21.7(3) 5.5 195317.92 +324911.4 6.8(3) 5.4 195313.42 +330121.8 97.8(3) 9.3 195245.25 +330330.0 6.2(3) 11.2 195215.80 +324935.7 839.3(8) 9.4 195230.01 +325526.2 5.4(3) 6.5 195213.32 +325922.1 61.7(3) 11.6 Note. —Unitsofrightascensionarehours,min- utes, and seconds, and units of declination are de- grees,arcminutes,andarcseconds. ofthe concatenateddatasets,toensureequalsensitivity to large-scale structure across all epochs. Full hydrodynamical modeling of the shock is beyond the scope of the present work. Instead, we simply ex- tractthe flux density in slices alongthe observedproper ◦ motion vector in the LSR (position angle =249 east of north) in order to estimate the angular separation be- tween the pulsar and the shock structure at each epoch. As shown in Figure 4, the pulsar position is well defined in these slices, but the transverse structure of the shock changeswithepochandtheseparationofthepulsarfrom the shock peak (∆θ ) varies between 2.4′′ and 3.1′′. Fig. 4.—SlicesthroughPSRB1951+32anditsbowshock. The Peak fluxdensity is plotted as afunction of the offset fromthe derived At each epoch, given the map rms noise σ, we define pulsarposition(asindicatedbythesolidline),alongthemeasured the shock thickness as the range where the emission is propermotionvector. Successiveepochsareverticallydisplacedby within 1 σ of the shock peak. The near and far edges of 1mJy(exceptfor2003,whichisdisplacedby1.25mJy)forclarity, the shock, defined in this manner, are listed in Table 5 butnote that the timeinterval between successiveobservations is not uniform. The map rms noise σ at each epoch is indicated by and indicated in Figure 4. theerrorbaratthefarleft. Thepeakoftheshockismarked,and Overthe14.5yrspanofobservations,thepulsarmoves thewidthoftheshockwithin1 σ oftheshockpeakisalsoshown by ∼ 0.47′′, but the distance between the pulsar and bythehorizontalerrorbar. Themeasuredseparationisconsistent the near edge of the shock is ∼2.1′′±0.2′′ and the sep- withbeingconstantover∼14.4yr,whilethepulsarmoves∼0.47′′in aration to the far edge is ∼3.6′′±0.6′′. Since the sep- thattimeperiod,asshownbythehorizontalarrow. aration between the pulsar and the radio-bright fea- remnantlikelytobe relatedtothe pulsar,butthatasso- ture does not show a secular decreasing trend, we con- ciation has serious problems (see Blazek et al. [2006]for cur with previous inferences (Hester & Kulkarni 1988; arecentdiscussion). Toremaintenable,suchanassocia- Chatterjee & Cordes 2002) that the feature must be a tionrequiresapulsaragemuchlargerthanthecharacter- shock driven by the pulsar wind as the pulsar travels istic age τ = 15.5 kyr, an explosion site much closer to through the ambient medium. c thecurrentlocationofthepulsar,orsomecombinationof 4. DISCUSSION thetwo. Forexample,G5.4−1.2couldbeexpandinginto aregionwithadensitygradient(e.g.,Gvaramadze2004). 4.1. The Association between SNR G5.4−1.2 and Thebrighteningoftheremnantonthe westernlimb, the PSR B1757−24 side nearer to the Galactic plane, may be evidence for The identity of the remnant associated by birth with such an asymmetric expansion. However,even in an op- the pulsar B1757−24 is still a mystery. Based on mor- timistic scenario, the birth site could be no closer than phology and location, SNR G5.4−1.2 is the only known the westernmost limb of G5.4−1.2. In such a case, for a 6 Zeiger et al. ′ tionassuminga1 uncertaintyinthereportedexpansion TABLE 5 center. A geometric center at coordinates 19h54m50s, AngularSeparationbetween ◦ ′ ′′ 33 0030 (J2000.0)was estimated by superposing a cir- PSRB1951+32andtheRadio BowShock cleontheremnantshellseenintheleftpanelofFigure2. Thepulsar’sclosestapproachtothispointimpliesanage Epoch ∆θPeak ∆θNear ∆θFar tp ∼51kyrwithaminimumseparationof74′′,consistent (′′) (′′) (′′) with the 56′′ positional uncertainty based solely on the 1989 2.4 1.9 3.2 propermotionuncertainty. Ourmeasurementofthepul- 1991 3.1 2.0 3.5 sar motion is therefore not consistent with the ∼77 kyr 1993 3.1 2.4 3.6 age and the expansion center estimated by Koo et al. 2000 2.5 1.9 4.3 (1990)fortheCTB80SNR,butitisconsistentwiththe 2003 2.6 2.2 3.0 pulsar age determined by Migliazzo et al. (2002). We Note. — We list the angular dis- note that the true explosion center may be masked by tancefromthepulsarpeaktothepeak asymmetric expansion due, for example, to density gra- oftheshockandtherangewithin1σ dientsandstructureintheISM,andneitherestimatefor oftheshockpeakateachepoch. The the age may be correct. However,the true age is almost positionfituncertainties areinsignifi- cantcomparedtothethicknessofthe certainlyless thanτc ∼107kyr,since the pulsar reaches shockitself. the far edge of the SNR when its measured proper mo- tion is projected backwards over that time interval, as true age ≤τ , a westwardproper motion &18 mas yr−1 c shown in Figure 2. isrequiredforanassociation,butallthreemeasurements As B1951+32 moves through its ambient medium, its reportedinTable6,wherewesummarizeourresults,im- relativisticwindoutflowisconfinedbyrampressure,and ply westward motions no greater than 15 mas yr−1. It asseeninsome other PWNe, synchrotronemissionfrom thus appears that either the true age is larger than τc, theconfinedpulsarwindproducesacometaryX-rayneb- probablysignificantlyso,orthe associationofthe pulsar ula. On the other side of the contact discontinuity, the with G5.4−1.2 must be excluded. shocked ISM produces Hα emission (see, e.g., the illus- We note that the characteristic ages of some tration by Gaensler et al. 2004). In such a picture, the young pulsars with well-determined kinetic ages are radio-brightarc would be produced by the confined pul- overestimates, for example, J1811−1925 (Kaspi et al. sar wind at the location of the contact discontinuity. As 2001b), J0538+2817 (Ng et al. 2007), and B1951+32 we have shown here, the structure of the bright arc is (Migliazzo et al. 2002, our §4.2). Thus, the true age of dynamic, changing from epoch to epoch, but overall, it B1757−24 might be no greater than 15.5 kyr, render- moves with the pulsar at an approximately steady sep- ing the associationunlikely. But Vela, for example, may aration ∼2.5′′, corresponding to a standoff distance of have a true age greater than τc (Lyne et al. 1996), and ∼0.024d pc. 2 Blazek et al. (2006) argue that a growing surface mag- netic field, which decouples the true age from the char- 4.3. Two Similar Pulsars in Very Different acteristicage,couldaccountforsuchanagediscrepancy Environments in the Duck while preserving the association. Given a Currently, over 1600 radio pulsars have measured val- long enough time baseline, the pulsar proper motion (or ues for P and P˙, as listed in the ATNF pulsar cata- the motion of the nebula) will surely be detectable, al- log7 (Manchester et al. 2005). In this large ensemble, lowing a firm conclusion about this vexing association. B1757−24 and B1951+32 are among the fifty most en- ergetic pulsars, with E˙ = 1036.4 and 1036.6 erg s−1, re- 4.2. The Age and Bow Shock of B1951+32 spectively. They are also among the hundred youngest For B1951+32, we measure a proper motion of known pulsars, with characteristic ages τ = 15.5 and (µα,µδ) = (−28.8 ± 0.9,−14.7 ± 0.9) mas yr−1, con- 107 kyr, respectively. On one hand, theccharacteristic firming and improving on the results of Migliazzo et al. ageofB1951+32islikelytobeanoverestimate,sinceits (2002). Migliazzo et al. (2002) filter the u-v plane to current spin period is not much greater than the typical ′′ remove all spatial scales greater than 4 and to leave birth spin periods of pulsars (Faucher-Gigu`ere & Kaspi onlycompactsources. Weuseadifferentapproachtore- 2006), andits kinematic aget ∼51kyr,assumingbirth p movethe complexity ofthe PWN,subtracting the linear near the center of CTB 80. On the other hand, the up- slope of the PWN from the region around the pulsar in per limits on its westward proper motion suggest that the image domain. The inclusion of a longer time base- B1757−24maybesubstantiallyolderthanitscharacter- line and a largerfield of reference sources increases both istic age. Although the true ages of the pulsars differ in the precisionand the accuracyof this measurementover relationto their respective characteristicages, these two that of Migliazzo et al. (2002). The new measurement pulsars are quite similar in energetics and age. corresponds to an LSR-corrected transverse velocity of Bothpulsars also exhibit cometary PWNe. B1757−24 (274±12) d2 km s−1. At a position angle of 249◦, the is associated with a ∼ 10′′ elongated nebula visible proper motion vector is well matched to the long sym- at both radio (as in Fig. 1) and X-ray wavelengths metry axis of the X-ray PWN (Moon et al. 2004). (Kaspi et al.2001a). B1951+32showsacometaryPWN Themeasurementofthepropermotionallowsarevised in X-rays (Moon et al. 2004), but its radio structure re- estimate for the age of the pulsar. The LSR-corrected sembles a bow shock and an extended, limb-brightened proper motion of the pulsar is shown in Figure 2. In bubble, and it is enclosed by a complex nebula visible projection, its closest approach to the expansion center estimated by Koo et al. (1990) is ∼330′′, a 5.5 σ devia- 7 Seehttp://www.atnf.csiro.au/research/pulsar/psrcat Proper Motions of PSRs B1757−24and B1951+32 7 TABLE 6 Summaryof ProperMotion Measurements B1757−24 B1951+32 Thiswork Ref.1 Ref.2 Thiswork Ref.3 Target B1757−24 B1757−24 G5.27−0.9 B1951+32 B1951+32 µα (masyr−1) −11±9 −2.1±7.0 >−7.9 −28.8±0.9 −29±2 µδ (masyr−1) −1±15 −14±13 ... −14.7±0.9 −8.7±1.3 TimeBaseline(yr) 6.5 3.9 12.0 14.5 11.9 Frequency (GHz) 1.4 1.4 8.5 1.4 1.4 RAofPulsar 180100.016±0.008 195258.206±0.001 DECofPulsar −245127.5±0.2 325240.51±0.01 EpochofRA,DEC 2004Dec09 2003Jun03 References. — (1) Thorsettetal. (2002); (2) Blazeketal. (2006); (3) Migliazzoetal.(2002) Note. — All provided uncertainties and limits are are 68% confidence intervals. Blazeketal.(2006) donot measure µδ, and their68% µα has been inferredfromtheir 5 σ limit; results of Migliazzoetal. (2002) have been adjusted to remove the differen- tialGalacticrotationappliedinthepublishedresult. Propermotionsandpositionsare reportedinJ2000.0coordinatesanddonotincludeLSRcorrections. Unitsofrightascen- sion are hours, minutes, and seconds, and units of declination are degrees, arcminutes, andarcseconds. in Hα (Hester 2000), with extended lobes and a bow riences for Undergraduatesprogramof the National Sci- shock structure. Hester & Kulkarni (1988) suggest that ence Foundation and the Department of Scientific and thepulsarisinteractingwiththewallofitsevolvedSNR, Academic Affairs of the National Radio Astronomy Ob- and we find that the radio shock structure changes with servatory (NRAO) for funding this research. S.C. ac- time while moving with the pulsar. On the other hand, knowledgessupportfromthe UniversityofSydney Post- B1757−24 appears to be outside any parent remnant. doctoral Fellowship program, and he was previously a If indeed it is associated with G5.4−1.2, it would have Jansky Fellow of the NRAO. The VLA is a facility of interacted with the wall of the remnant in the past (as the NRAO, funded by the NSF and operated under co- B1951+32is at present), possibly re-energising the shell operativeagreementbyAssociatedUniversities,Inc. We andleavingbehindthestructureweseeastheDuck. Al- thank Dale Frail for helpful conversationsregarding this ternatively,theparentremnantofB1757−24isnolonger research,andCrystalBroganfor providingthe 327MHz visible, and we are witnesses to a cosmic coincidence. image of G5.4−1.2. Facilities: VLA B.R.Z.acknowledgessupportfromtheResearchExpe- REFERENCES Arzoumanian,Z.,Chernoff,D.F.,&Cordes,J.M.2002,ApJ,568, Kaplan, D. L., Gaensler, B. M., Kulkarni, S. R., & Slane, P. O. 289 2006,ApJS,163,344 Blazek,J.A.,Gaensler,B.M.,Chatterjee,S.,vanderSwaluw,E., Kaspi,V.M.,Gotthelf,E.V.,Gaensler,B.M.,&Lyutikov2001a, Camilo,F.,&Stappers,B.W.2006,ApJ,652,1523 M.,ApJ,562,L163 Brogan, C. L., Gelfand, J. D., Gaensler, B. M., Kassim, N. E., & Kaspi, V. M., Roberts, M. E., Vasisht, G., Gotthelf, E. V., Lazio,T.J.W.2006,ApJ,639,L25 Pivovaroff,M.,&Kawai,N.2001b,ApJ,560,371 Castelletti, G., Dubner, G., Golap, K., Goss, W. M., Vela´zquez, Koo,B.-C.,Reach,W.T.,Heiles,C.,Fesen,R.A.,&Shull,J.M. P.F.,Holdaway,M.,&Rao,A.P.2003,AJ,126,2114 1990,ApJ,364,178 Chatterjee,S.,&Cordes,J.M.2002,ApJ,575,407 Kulkarni, S. R., Clifton, T. C., Backer, D. C., Foster, R. S., & Cordes, J. M., & Lazio, T. J. W. 2002, ArXiv e-print, Fruchter,A.S.1988, Nature,331,50 astro-ph/0207156 Lyne, A. G., Pritchard, R. S., Graham-Smith, F., & Camilo, F. Dehnen,W.,&Binney,J.1998,MNRAS,294,429 1996,Nature,381,497 Faucher-Gigu`ere,C.-A.&Kaspi,V.M.2006, ApJ,643332 Manchester,R.N.,Hobbs,G.B.,Teoh,A.&Hobbs,M.,2005,AJ, Fesen,R.A.,Saken,J.M.,&Shull,J.M.1988,Nature,334,229 129,1993 Frail,D.A.,Kassim,N.E.,&Weiler,K.W.1994,AJ,107,1120 Manchester, R. N., Kaspi, V. M., Johnston, S., Lyne, A. G., & Frail,D.A.,&Kulkarni,S.R.1991,Nature,352,785 D’Amico,N.1991,MNRAS,253,7P Gaensler,B.M.,&Frail,D.A.2000,Nature,406,158 McGary, R. S., Brisken, W. F., Fruchter, A. S., Goss, W. M., & Gaensler, B. M., van der Swaluw, E., Camilo, F., Kaspi, V. M., Thorsett,S.E.2001,AJ,121,1192 Baganoff, F. K., Yusef-Zadeh, F., & Manchester, R. N. 2004, Migliazzo,J.M.,Gaensler,B.M.,Backer,D.C.,Stappers,B.W., ApJ,616,383 vanderSwaluw,E.,&Strom,R.G.2002,ApJ,567,L141 Golden, A., Bourke, S., Clyne, G., Butler, R. F., Shearer, A., Moon,D.-S.,Lee, J.-J.,Eikenberry,S.S.,Koo,B.-C.,Chatterjee, Muxlow,T.W.B.,&Brisken,W.F.2005,ApJ,635,L153 S.,Kaplan,D.L.,Hester,J.J.,Cordes,J.M.,Gallant,Y.A.,& Gvaramadze, V.V.2004,A&A,415,1073 Koch-Miramond,L.2004,ApJ,610,L33 Hester,J.J.,&Kulkarni,S.R.1988,ApJ,331,L121 Ng, C.-Y., Romani, R. W., Brisken, W. F., Chatterjee, S., & Hester,J.2000,BAAS,32,1542 Kramer,M.2007, ApJ,654,487 Hobbs, G., Lorimer, D. R., Lyne, A. G., & Kramer, M. 2005, Seward, F. D., Slane, P. O., Smith, R. K. & Sun, M. 2003, ApJ, MNRAS,360,974 584,414 Kaplan, D. L., Frail, D. A., Gaensler, B. M., Gotthelf, E. V., Strom, R. G., & Stappers, B. W. 2000, in ASP Conf. Ser. 202: Kulkarni, S. R., Slane, P. O., & Nechita, A. 2004, ApJS, 153, IAU Colloq. 177: Pulsar Astronomy - 2000 and Beyond, ed. 269 M.Kramer,N.Wex, &R.Wielebinski,509 8 Zeiger et al. Thorsett, S. E., Brisken, W. F., & Goss, W. M. 2002, ApJ, 573, L111 Vasisht,G.,&Gotthelf,E.V.1997,ApJ,486,L129