The extreme radio emission of PSR B0656+14 -- Is B0656+14 a very nearby rotating radio transient?

0.82 MB·English

by  P. Weltevrede

Proceedingsofthe363.WE-HeraeusSeminaron:“NeutronStars and Pulsars”(Postersand contributedtalks)
Physikzentrum BadHonnef, Germany,May.14−19,2006, eds.W.Becker,H.H.Huang, MPEReport291,pp. 60- 63
The extreme radio emission of PSR B0656+14 —
Is B0656+14 a very nearby Rotating Radio Transient?
P. Weltevrede1, B. W. Stappers2,1, J. M. Rankin3, and G. A. E. Wright4
1 AstronomicalInstitute“AntonPannekoek”,UniversityofAmsterdam,Kruislaan403, 1098SJAmsterdam,TheNetherlands
2 StichtingASTRON,Postbus2, 7990 AA Dwingeloo, TheNetherlands
7 3 Physics Department,405 Cook Physical Science building, Universityof Vermont,Burlington, 05405, USA
0 4 Astronomy Centre, University of Sussex, Falmer, BN19QJ, UK
0
2

n
a Abstract. Wepresentadetailedstudyofthesingleradio 2. Observations
J
pulses of PSR B0656+14.The emission can be character-
8 The results presented in this paper are based on two ob-
ized by two separate populations of pulses: bright pulses
servations made using the 305-meter Arecibo telescope.
have a narrow “spiky” appearance in contrast to the un-
1 Bothobservationshadacentre-frequencyof327MHzand
v derlyingweakerbroadpulses.The shape ofthe pulse pro-
a bandwidth of25MHz. Almost25,000and17,000pulses
9 file requires an unusually long timescale to achieve sta-
withasamplingtimeof0.5125and0.650mswererecorded
8 bility (over 25,000 pulses at 327 MHz) caused by spiky
1 using the Wideband Arecibo Pulsar Processor (WAPP)
emission.Theextremepeak-fluxesofthebrightestofthese
1 for the observation made in 2003 and 2005 respectively.
pulses indicates that PSR B0656+14,were it not so near,
0 The Stokes parameters have been corrected off-line for
could only have been discoveredas anRRAT source.The
7
dispersion,Faradayrotationandvariousinstrumentalpo-
0 strongest bursts represent pulses from the bright end of
larization effects. For more details we refer to Weltevrede
/ an extended smooth pulse-energy distribution, which is
h et al. 2006b and 2006c.
p unlikegiantpulses,giantmicropulsesorthepulsesofnor-
- mal pulsars. Longer observations of the RRATs may re-
o vealthat they, like PSR B0656+14,emit weakeremission 3. Stability of the pulse profile
r
t in addition to the bursts.
s For most pulsars one can obtain a stable pulse profile by
a averaging a few to a few hundred pulses, so our observa-
:
v tions of up to 25,000 pulses could have been expected to
Xi 1. Introduction belongenough.HoweverPSRB0656+14provedtobe far
from typical, as the pulse profile is highly unstable. To
r
a PSR B0656+14 is one of three nearby pulsars in the illustrate this time dependence, the profiles of successive
middle-age range in which pulsed high-energy emission blocksofonethousandpulseswerecalculated(leftpanelof
has been detected (the “The Three Musketeers”). It was Fig. 1). The scintillation bandwidth is much smaller than
included in a recent extensive survey of subpulse modu- the observing bandwidth, so the intensity of the profiles
lation in pulsars in the northern sky at the Westerbork are unaffected by interstellar scintillation. This is not be-
Synthesis Radio Telescope by Weltevrede et al. 2006a. In cause of systematic errors due to polarization calibration
the single pulses analysed for this purpose exceptionally uncertainties (Weltevrede et al. 2006c). A much longer
powerfulandlongitudinally narrowsubpulses reminiscent observationwouldbe requiredto find out if there exists a
of “giant” pulses were found. We therefore set out to ex- time scale for the pulse profile to stabilize.
plore the full nature of PSR B0656+14’spulse behaviour
in the radio band (Weltevrede et al. 2006c).This pulsar’s
4. The spiky emission
extreme bursts are far from typical of older better-known
pulsars, but are similar to those detected in the recently Atypicalpulsesequenceofthispulsarisshownintheleft
discoveredpopulationofburstingneutronstars.TheseRo- panelofFig.2.Onecanseethatthefrequentoutburstsof
tatingRAdioTransients(RRATs;McLaughlinetal.2006) radio emission are much narrower than the width of the
typically emit detectable radio emission for less than one pulse profile. The emission also has burst-like behaviour
second per day, causing standard periodicity searches to in the sense that the radio outbursts tend to cluster in
fail in detecting the rotationperiod. The intermittent ex- groups of a few pulse periods. Furthermore, this cluster-
treme bursts we have detected in PSR B0656+14 led us ingsometimesseemstobeweaklymodulatedwithaquasi-
to argue that this pulsar, were it not so near, could itself periodicity of about20 pulse periods (see for instance the
have appeared as an RRAT (Weltevrede et al. 2006b). bursts around pulse numbers 55, 75, 95, 115 and 135).
P. Weltevredeet al.: The extremeradio emission of PSR B0656+14 61
Fig.1. The pulse profiles obtained by averaging successive blocks of one thousand pulses each of the 2005 Arecibo
observation. The left panel shows all the emission and the middle and right panel show the spiky and weak emission
separately. The 1-sigma error bars are plotted in the top left corner.
Thebrightestburstsarealsoshowntohavequasi-periodic Thereasonwhythepulseprofileisunstableisthepres-
structures with a ∼11 ms and a ∼1-ms timescale (Wel- ence of the spiky emission. These spikes have a very un-
tevrede et al.2006c).Besides these bursts there aremany evenlongitude distributionandthereforemanypulses are
pulses (and largefractionsofthe pulse window)thatcon- required to obtain a steady profile. This is demonstrated
tain no signal above the noise level. We will use the term in Fig. 1, where the profiles of successive blocks of one
spiky to refer to these bursts of radio emission. thousand pulses each are plotted separately for the spiky
andweakemission.Onecanclearlyseethatitisthespiky
Although the pulse sequence of the left panel of Fig. emission that is highly unstable, whereas the weak emis-
2 is dominated by the very apparent spiky emission, this sion converges rapidly to a stable profile.
is accompanied by an almost indiscernible background of
weak emission. To separate these two components of the
5. The radio bursts of PSR B0656+14
emission, we have applied an intensity threshold to the
data. The intensities of the time samples in the pulse The brightest measured pulse is 116 hEi (where hEi is
stack of the weak emission are truncated if they exceed the average integrated pulse energy). This is exceptional
this threshold. The time samples in the pulse stack of the for regular radio pulsars and based on the energy of
spikyemissioncontainsonlysamples withintensityinex- thesepulsesalone,PSRB0656+14wouldfitintotheclass
cess of this threshold. When the pulse stack of the weak of pulsars that emit so-called giant pulses. Nevertheless,
emissionisaddedtothe pulsestackofthe spikyemission, there are important differences between giant pulses and
one retrieves exactly the original pulse stack. the bright bursts of PSR B0656+14. The bursts of PSR
B0656+14have timescales that are much longer than the
We have set the threshold intensity such that 99% of nano-second timescale observed for giant pulses, do not
thenoisesamplesarebelowthisthresholdvalue.Notonly showa power-lawenergy-distribution,arenotconfinedto
dothenoisefluctuationsmakeitimpossibletocompletely a narrow pulse window and are not associated with an
separate the weak and spiky emission, it is also very well X-ray component. This suggests differing emission mech-
possible that the energy distributions of the two compo- anismsfortheclassicalgiantpulsesandtheburstsofPSR
nents overlap. In Fig. 2 one can see the pulse stacks ob- B0656+14.Alsothepossiblecorrelationbetweenemission
tained after separation of the spiky and weak emission. of giant pulses and high magnetic field strengths at the
The integrated power of this sequence of weak pulses is light cylinder clearly fails for PSR B0656+14. However,
about 3 times greater than that of the sequence of the giantpulseshavebeenclaimedinother(slow)pulsarsthat
spiky emission. This shows that a significant fraction of also easily fail this test and for millisecond pulsars a high
the pulsar’s emission lies at or below the noise level. magnetic field strengths at the light cylinder seems to be
62 P. Weltevredeet al.: The extremeradio emission of PSR B0656+14
Fig.2. A typical sequence of successive pulses (left panel). The same pulses are shown in the middle and right panel,
but there the emission is separated into the spiky and weak emission respectively.
a poor indicator of the rate of emission of giant pulses thispulsariscapableofproducingintensesporadicbursts
(Knight et al. 2006). ofradioemissionevenatearlyphasesoftheprofile.There
are only two bursts with a peak-flux above the noise level
The bursts of PSR B0656+14 are even more extreme
detected at the longitude of the peak of this pulse out of
whenweconsidertheirpeak-fluxes.Thehighestmeasured
the total of almost 25,000 pulses (see right panel of Fig.
peak-flux of a burst is 420 times the average peak-flux
3). This implies either that these two bursts belong to an
of the pulsed emission, which is an order of magnitude
extremely long tail of the distribution, or that there is no
brighter than the giant micropulses observed for the Vela
emissionatthislongitudeotherthansuchsporadicbursts.
pulsar (Johnston et al. 2001) and PSR B1706–44 (John-
ston & Romani 2002). Giant micropulses are not neces-
sarily extreme in the sense of having a large integrated
energy (as required for giant pulses), but their peak-flux
6. The RRAT connection
densities are very large. Not only are the bursts of PSR
B0656+14 much brighter (both in peak-flux and inte-
TheobservationalfactsarethatPSRB0656+14occasion-
grated energy) than those found for giant micropulses,
allyemitsextremelybrightburstsofradioemissionwhich
they are also not confined in pulse longitude and they
areshortinduration.Althoughtheseburstsappearstobe
do not show a power-law energy-distribution as the giant
different than that of the giant (micro)pulses, it seems to
pulses and micropulses do.
besimilartothosefoundfortheRRATs.Weltevredeetal.
At the leading edge of the profile we detected a burst 2006bshowthatthe luminositiesofthe burstsoftherela-
with an integrated pulse-energy of 12.5 hEi. What makes tivelynearbyPSRB0656+14(288pc;Briskenetal.2003)
this pulse so specialis thatit has a peak-flux that is 2000 isverytypicalfortheknownRRATsources.Althoughthe
times that ofthe averageemissionatthat pulse longitude slope of the top end of the peak-flux distribution of PSR
(left panel of Fig. 3). Its dispersiontrack exactly matches B0656+14 is in the range of the giant pulses (between
what is expected for this pulsar (middle panel of Fig. 3), −2 and −3), it is better described by a lognormal than
proving that this radio burst is produced by the pulsar. by a power-lawdistribution. This again suggests that the
Notice that the effect of interstellar scintillation is also brightburstsofPSRB0656+14aredifferentfromtheclas-
clearly visible (different frequency channels have differ- sical giant pulses. The top end of the RRAT distribution
ent intensities) and that the dispersion track is the same with the highest number of detections seems to be harder
for the two pulses in the centre of the profile. This burst (with a slope −1), but for the other RRATs this is as yet
demonstrates that the emission mechanism operating in unclear. For instance, the tail of the distribution of PSR
P. Weltevredeet al.: The extremeradio emission of PSR B0656+14 63
10000
30
25 1000
malized intensity 1250 Counts 100
Nor 10 10
5
1
0
-5 0 5 10 15 20 25 30 35
100 150 200 250 Fi / 〈Fp〉
Pulse longitude (deg)
Fig.3.Thebrightradioburstdetectedattheleadingedgeofthepulseprofileinthe2003observation.Left:Theburst
(solid line) compared with the average pulse profile (dashed line). Middle: The same burst, but now with frequency
resolution. The data in this plot is not de-dispersed and its dispersion track matches exactly what is expected for the
known dispersion measure (DM) of this pulsar (dashed line). Right: The longitude-resolved energy-distribution at
the longitude of the peak of the strong pulse (solid line) and the off-pulse distribution (dashed line). The peak-fluxes
(F ) are normalized to the average peak-flux of the profile (hF i).
i p
B0656+14seems to be consistent with the distribution of weak emission (with a high occurrence rate over the full
the RRAT with the second highest number of detections. width of the pulse). PSR B0656+14 intermittently emits
Normal periodicity searches failed to detect the pulses that areextremelybrightcomparedto normalpul-
RRATs,whichplacesanupper limitonthe averagepeak- sars and with pulse energies well above ten times the av-
flux density of weak pulses among the detected bursts of erage pulse-energy these pulses formally qualify as giant
about1:200(McLaughlinetal.2006).Becausethebright- pulses. Nevertheless these pulses differ from giant pulses
est burst of PSR B0656+14exceeds the underlying peak- and giant micropulses in important ways. Many of the
flux by a much greater factor,PSR B0656+14could have exceptional properties of PSR B0656+14 have led us to
beenidentifiedasanRRAT,wereitnotsonearby.Wereit point out that this pulsar,were it not so near,could have
located twelve times farther away (thus farther than five been discovered as an RRAT.
of the RRATs), we estimate that only one burst per hour Our identification of PSR B0656+14with RRATs im-
wouldbedetectable(theRRATs haveburstratesranging plies that at least some RRATs could be sources which
fromoneburstevery4minutestooneevery3hours).The emit pulses continuously, but over an extremely wide
typicalburstduration(about5ms)ofPSRB0656+14also range of energies. This is in contrast to a picture of infre-
matches that of the RRATs (between 2 and 30 ms). quent powerful pulses with otherwise no emission. There-
Were PSR B0656+14 twelve times more distant, the fore,ifitindeedturnsoutthatPSRB0656+14(despiteits
RMS of the noise would increase by a factor 144 relative relatively short period) is a true prototype for an RRAT,
to the strength of the pulsar signal. When we artificially we can expect future studies to demonstrate that RRATs
addgaussian-distributednoiseatthisleveltotheobserva- emit much weaker pulses among their occasional bright
tion, we find no sign of the pulsar’s (2.6-Hz) rotation fre- bursts. We would also predict that their integrated pro-
quencyin35-minutesegmentsofthedatayetthebrightest fileswillbefoundtobefarbroaderthanthewidthsofthe
pulseiseasilydetectedwith18σ(Weltevredeetal.2006b). individualbursts,andwillneedmanythousandsofbursts
For telescopes with a lower sensitivity than Arecibo (e.g. to stabilize.
Parkes)thenevenifPSRB0656+14werequiteabitcloser
itwouldnotrevealit’speriodicityinasimilarlylongobser-
References
vation.Only inthe spectrumofthe whole 1.8-hourobser-
vationthe periodicity ofa twelvetimes more distantPSR Brisken, W. F., Thorsett, S. E., Golden, A., & Goss, W. M.,
B0656+14 would be marginally detectable for Arecibo. 2003, ApJ, 593, L89
This means that a distant PSR B0656+14 could only be Johnston, S.& Romani, R.,2002, MNRAS.,332, 109
Johnston,S.,vanStraten,W.,Kramer,M.,&Bailes,M.,2001,
found as a RRAT in a survey using Arecibo, unless the
ApJ., 549, L101
pointings were unusually long.
Knight, H. S., Bailes, M., Manchester, R. N., Ord, S. M., &
Jacoby, B. A., 2006, ApJ,640, 941
7. Implications and discussion McLaughlin, M. A., Lyne, A. G., Lorimer, D. R., et al. 2006,
Nature,439, 817
The emission of PSR B0656+14 can be characterized by Weltevrede, P., Edwards, R. T., & Stappers, B. W., 2006a,
spiky (with low occurrence rate within each pulse) and A&A,445, 243
64 P. Weltevredeet al.: The extremeradio emission of PSR B0656+14
Weltevrede, P., Stappers, B. W., Rankin, J. M., & Wright,
G. A.E. 2006b, Astrophys.J., 645, L149
Weltevrede, P., Wright, G. A. E., Stappers, B. W., Rankin,
J. M., & 2006c, accepted by A&A(astro-ph/0608023).