Pulsars as tools for fundamental physics and astrophysics

26 Pages·2004·1.52 MB·English

NewAstronomyReviews48(2004)1413–1438
www.elsevier.com/locate/newastrev
Pulsars as tools for fundamental physics and astrophysics
J.M. Cordes a,*, M. Kramer b, T.J.W. Lazio c,1, B.W. Stappers d,
D.C. Backer e, S. Johnston f
aDepartmentofAstronomyandNAIC,CornellUniversity,Ithaca,NY,USA
bUniversityofManchester,JodrellBankObservatory,JodrellBank,UK
cNavalResearchLaboratory,RemoteSensingDivision,Washington,DC,USA
dASTRON,Dwingeloo,TheNetherlands
eDepartmentofAstronomy,UCBerkeley,Berkeley,CA,USA
fSchoolofPhysics,UniversityofSydney,Sydney,NSW2006,Australia
Availableonline19November2004
Abstract
The sheer number of pulsars discovered by the SKA, in combination with the exceptional timing precision it can
provide,willrevolutionizethefieldofpulsarastrophysics.TheSKAwillprovideacompletecensusofpulsarsinboth
the Galaxy and in Galactic globular clusters that can be used to provide a detailed map of the electron density and
magneticfields,thedynamicsofthesystems,andtheirevolutionaryhistories.Thiscompletecensuswillprovideexam-
plesofnearlyeverypossibleoutcomeoftheevolutionofmassivestars,includingthediscoveryofveryexoticsystems
suchaspulsarblack-holesystemsandsub-millisecondpulsars,iftheyexist.Theseexoticsystemswillallowuniquetests
ofthestrongfieldlimitofrelativisticgravityandtheequationofstateatextremedensities.Massesofpulsarsandtheir
binarycompanions–planets,whitedwarfs,otherneutronstars,andblackholes–willbedeterminedto(cid:1)1%forhun-
dredsofobjects.WiththeSKAwecandiscoverandtimehighly-stablemillisecondpulsarsthatcompriseapulsar-tim-
ingarrayforthedetectionoflow-frequencygravitationalwaves.TheSKAwillalsoprovidepartialcensusesofnearby
galaxies throughperiodicity andsingle-pulse detections, yielding important information on the intergalacticmedium.
(cid:1)2004PublishedbyElsevier B.V.
1. Introduction
Neutron stars (NSs) are accessible to observa-
* Correspondingauthor.
tion as pulsars and thus provide our only means
E-mailaddress:[email protected](J.M.Cordes).
1 BasicresearchinradioastronomyattheNRLissupported for probing the most extreme states of matter in
bytheOfficeofNavalResearch. thepresent-dayUniverse,whichinturnwillenable
1387-6473/$-seefrontmatter (cid:1)2004PublishedbyElsevierB.V.
doi:10.1016/j.newar.2004.09.040
1414 J.M.Cordesetal./NewAstronomyReviews48(2004)1413–1438
a vast range of transforming science goals to be force on a surface proton (cid:1)1011 times larger
addressed, among which are: than the gravitational force. Voltage drops
(cid:1)1012 V across the magnetosphere accelerate
(cid:2) Strong-field tests of gravity including the Cos- particles that can radiate across the entire elec-
micCensorshipConjectureandtheno-hairthe- tromagnetic spectrum. This makes some pulsars
orem of BHs. visible in every astronomical window (Thomp-
(cid:2) Detection of a cosmological gravitational wave son, 2000). Most of what is to be learned from
background. pulsars requires observations of their radio emis-
(cid:2) Mapping the complete structure of the Milky sion, often providing a unique source of infor-
Way and revealing properties of the Galactic mation, and otherwise providing information
Center GC. that complements multiwavelength studies, par-
(cid:2) Probing the intergalactic medium in new ways. ticularly at high energies.
(cid:2) Identifying the equation-of-state of superdense The feature that gives pulsars their name –
matter. pulsed radioemission –allowsmostNSstobede-
(cid:2) Quantifyingtherolesofmagneticfieldsandtur- tected at levels well below the sky confusion limit
bulence in core-collapse physics. and also provides the means for using pulsars as
(cid:2) Understandingthesuperfluidinteriorsandrela- physics laboratories. Coherent radio emission is
tivistic magnetospheres of NS. associated with the collimation of the flow of par-
(cid:2) Unravelingtheevolutionaryanddynamicalhis- ticles at the poles of the large-scale magnetic field
tories and properties of all Galactic globular in combination with relativistic beaming. The
clusters. spin-driven sweep of the beam across the line of
(cid:2) Discoveries of extrasolar planets. sight then provides the distinctive pulsation of
the electromagnetic signal.
Internal densities (cid:1)10 times nuclear density The suitability for using pulsars as clocks de-
have not existed since the Universe was about pends on the regularity by which the spin
1 ms old. The actual state of matter in the core evolves with time. Spin noise in some objects,
of a NS is presently not known. It may consist which evidently reflects activity within the crust
of de-confined quark matter or hyperonic matter and superfluid, is large enough to mask many
produced in a phase transition that occurred of the physical effects of interest that provide
during or shortly after the core collapse of the only subtle timing signatures. However, spin
progenitor star. Intermediate regions of the NS noise itself, including rapid spinups (glitches), is
consist of neutrons and trace protons in, respec- itself valuable information on NS interiors. All
tively, superfluid and super-conducting states pulsars are stable enough so that timing meas-
achieved after the NS cooled to about 109 de- urements can yield fundamental parameters such
grees. The outer regions include an (cid:1)1 km thick as the period vs. epoch, P(t) and its time deriv-
crust composed of iron-like nuclei surrounded by ative P_, along with the dispersion measure DM.
an ocean about 1 cm deep (Shapiro et al., 1983). Objects with the narrowest pulses, the shortest
The magnetic field anchored to the crust and periods, and the most stable rotation rates – mil-
extending to interstellar space is sufficiently lisecond pulsars – yield the greatest opportunities
strong that it elongates the atoms comprising for exploring relativistic gravity. Extrinsic gravi-
the crust. The surface gravity, about 109 times tational effects include perturbations of pulse ar-
that of the Earth(cid:1)s, is the largest of any object rival times from the direct influence of
in the Universe subject to observation, and cor- companion stars on space-time and from evolu-
responds to a gravitational redshift (cid:1)0.3. tion of the orbit in the non-Newtonian gravity.
While NS gravity is strong, radio pulsars and The latter causes binary pulsars and their orbits
probably most NSs are even more extreme elec- to precess whilst their orbit decays owing to loss
tromagnetically. By virtue of spin periods P(cid:1)1 s of energy to gravitational radiation. Besides
and magnetic fields B(cid:1)108–1014 G, the electric being sources of gravitational wave emission
J.M.Cordesetal./NewAstronomyReviews48(2004)1413–1438 1415
(Peters, 1964), pulsars also lend themselves as 2. The cosmic census for pulsars
detectors of long-period gravitational waves that
are cosmological in origin (Detweiler, 1979). Pulsars are of great utility, no matter where we
Thepopulationofisolatedandbinarypulsarsis find them. So far, most known pulsars are Galac-
of great interest because their phase-space distri- tic,residinginornearthediskoftheGalaxyorin
bution and overall numbers reflect the rate at globular clusters. A small number of pulsars is
which NSs are born in Type II supernova explo- known in the Magellanic clouds; however radio
sions, how the explosions themselves produce the pulsars havenot yetbeen detectedin more distant
runawayvelocitiesofNSs,andhowNSsareinflu- galaxies, although a few NS in accreting systems
enced by accretion in those rare binary systems have been seen in M31 via their X-ray emission
that survive the explosions. Unlocking the vast (Kaaret, 2002). In the following we summarize
populationofactiveradiopulsarsthereforeallows whyitisimportanttodiscovermorepulsars,both
us to study the star formation history and evolu- locally and in other galaxies, as well as in particu-
tion of massive stars, aspects of binary evolution, lar regions of the Milky Way.
core collapse physics, as well as the movement of
the high-velocity population of pulsars in the 2.1. Galactic census
Galactic gravitational potential.
With this enormous range of fundamental Why perform a full-Galactic census? The Galac-
physics accessible through pulsar observations, ticCensusisessentialinprovidingthelaboratories
the pulsar field has been extremely fruitful in forawiderangeofpulsarapplicationsthatwillbe
the 37 years since their discovery, as evidenced discussed in more detailed in the following sec-
by the awarding of two Nobel prizes, one to tions. The larger the number of pulsar detections,
the original discovery of pulsars (Hewish et al., the more likely it is to find rare objects that pro-
1968), the other to the discovery of the first videthegreatestopportunitiesasphysicslaborato-
NS–NS binary pulsar (Harrison and Tademaru, ries. These include binary pulsars with black hole
1975) that allowed inference of gravitational companions (see Section 3.1); binary pulsars with
radiation in accord with Einstein(cid:1)s General The- orbital periods of hours or less that can be used
ory of Relativity (Taylor et al., 1979). Nonethe- for fundamental tests of relativistic gravity (see
less, the field has a great deal more to contribute Section 3.1); MSPs that can be used as detectors
to our knowledge of fundamental issues in phys- of cosmological gravitational waves (see Section
ics and astrophysics. 3.3); MSPs spinning faster than 1.5 ms, possibly
Wedescribesomeoftheapplicationsofpulsars as fast as 0.5 ms, that probe the equation-of-state
madepossiblebytheSquareKilometerArraypro- under extreme conditions (see Section 3.2); hyper-
ject in the following sections. In particular, we velocity pulsars withtranslational speedsinexcess
haveidentifiedStrong-FieldTestsofGravityUsing of 103 kms(cid:3)1, which probe both core-collapse
Pulsars and Black Holes as a key science area for physics and the gravitational potential of the
the SKA that is outlined in some detail in Section Milky Way (see Section 3.10); and objects with
3.1 and in particular in Kramer et al. (this unusualspinproperties,suchasthoseshowingdis-
volume). continuities(‘‘glitches’’)andapparentprecessional
To enable this programme of research outlined motions(bothisolatedpulsarsshowing‘‘free’’pre-
below, the SKA must have capabilities that allow cession and binary pulsars showing geodetic pre-
itsenormouscollectingareatobeusedinavariety cession) (see Section 3.2).
of observing modes. Some of these imply usage as The second reason to perform a full Galactic
a huge, effective single dish, while others exploit census is that the large number of pulsars can be
the array aspects of the telescope. As a mantra usedtodelineatetheadvancedstagesofstellarevo-
forpulsarresearch,theSKAmustallowusthefol- lution that lead to supernovae and compact ob-
lowing activities on pulsars: find them, time them jects. In particular, with as large a sample as
and VLBI them. possible, we can determine the branching ratios
1416 J.M.Cordesetal./NewAstronomyReviews48(2004)1413–1438
fortheformationofcanonicalpulsarsandmagne-
tars. We can also estimate the effective birth rates
for MSPs and for those binary pulsars that are
likely to coalesce on time scales short enough to
beofinterestassourcesofperiodic,chirpedgravi-
tational waves. The latter population can be di-
rectly compared to the results of gravitational
wave detectors, which will have been operating
for a sufficiently long time by the time the SKA is
switched on.
The third reason is that a maximal pulsar sam-
ple can be used to probe and map the interstellar
medium (ISM) in a nearly complete way. Measur-
able propagation effects include dispersion, scat-
tering, Faraday rotation, and HI absorption that
provide, respectively, line-of-sight integrals of the
free-electron density n , of the fluctuating electron
e
density, dn , of the product Bn , where B is the
e i e i
LOS component of the interstellar magnetic field, Fig.1. TheP–P_ diagramforradiopulsarsandmagnetars.This
andoftheneutralhydrogendensity.Theresulting isascatterplotofpulseperiodandperiodderivativeforpulsars
measures are DM, SM, RM and N : inthecurrent(April2004)publicATNFcatalogue.Thecircled
HI pointsarepulsarsinbinarysystemswhileobjectsintheupper
rightpartofthediagramhavederivedsurfacemagneticfields
Z D Z D
DM¼ dsn ; SM¼ dsC2; ð1Þ exceeding Bq=4.4·1013 G (see Section 3.4). The ‘‘spin-up’’
e n linedenotestheterminalperiodforobjectsmovingtotheleftin
0 0
thediagramowingtoaccretionthatbecome‘‘recycled’’pulsars
Z D Z D once accretion ceases. The ‘‘graveyard’’ marks the locations
RM¼ dsn B ; N ¼ dsn ; ð2Þ where radio pulsars emit less radiation or turn off entirely,
e k HI HI
0 0 possiblyassociatedwiththecessationofelectron-positron(e±)
pair-production cascades or with saturation of the radio
involving quantities that can also be studied by luminosityatsomefractionofthetotalspin-downenergyloss
othermeans,forinstancebyHIobservationswith rate.
the SKA. The determination of these observables
for a large number of independent line-of-sights magneticfieldstrengthsB(cid:1)1012±1G.Theyare
for pulsars will enable us to construct a complete often thought to be born with periods (cid:1)10 ms,
map of the Galaxy (see Section 3.7). though evidence suggests that some objects are
WithaMilkyWaybirthratethatcurrentlymay born with periods longer than 0.1 s (Kramer
be as large as 10(cid:3)2 year(cid:3)1, about 108 NSs have et al., 2003). In the standard picture of NS for-
been formed over the lifetime of the Galaxy, and mation,allpulsarsstartascanonicalpulsars.In
probably many more because the star-formation theP–P_ diagramofFig.1mostofthesepulsars
was most likely higher in the past. Most NSs are arelocatedatP(cid:1)1sandP_ (cid:1)10(cid:3)15.Youngpul-
inert, their radio emission having shut off long sars are especially important members of this
ago, and up to about half of them will have left class because they are associated with super-
the Galaxy owing to their large space velocities. nova remnants and often show copious num-
Of the known pulsars (see Fig. 1), we can identify bers of glitches.
several subclasses. 2. Modestly recycled pulsars: are objects in bina-
ries that survived a first SN explosion and sub-
1. Canonicalpulsars:Thesepulsars,likethosefirst sequently accreted matter that spun-up the
discovered, have present-day spinperiods rang- pulsarandreducedtheeffectivedipolarcompo-
ing from tens of milliseconds to 8 s and surface nent of the magnetic field. Accretion is termi-
J.M.Cordesetal./NewAstronomyReviews48(2004)1413–1438 1417
nated in these objects by a second supernova
15
explosion that may or may not disrupt the bin-
ary.Thosethatsurviveareseentodayasrelativ-
istic NS–NS binaries. Evolutionarily, it is 10
possible that some surviving binaries include
black-hole companions. In the P–P_ diagram of 5
Fig.1thesepulsarsaretypicallylocatedaround
P(cid:1)30 ms and P_ (cid:1)10(cid:3)18. kpc) 0
3. Millisecond pulsars (MSPs): objectsin binaries Y (
thatsurvivethefirstSNexplosionandinwhich
–5
the companion object eventually evolves into a
white dwarf. The long, preceding accretion
–10
phasespinsthepulsaruptomillisecondperiods
while attenuating the (apparent) dipolar field
componentto108–109G.Theconsequentsmall –15
spin-downratesseemtounderlythehightiming –15 –10 –5 0 5 10 15
precision of these objects and imply spin-down X (kpc)
time scales that exceed a Hubble time in some Fig. 2. Simulated Galactic pulsar population discovered in a
cases.IntheP–P_ diagramofFig.1thesepulsars SKA survey of the entire sky. The (cid:1)20,000 pulsars shown
are typically located around P(cid:1)5 ms and togetherwiththespiralarmsstructure.TheGalacticCenteris
P_ (cid:1)10(cid:3)20. Evolutionary scenarios that produce locatedattheoriginwhiletheSunisat(0.0,8.5)kpc.
recycled pulsars and MSPs are discussed in
Bhattacharya and den Heuvel (1991). 1989), so the fiducial birth rate implies (cid:1)2·104
4. Strong-magnetic-field pulsars: Recently discov- detectable pulsars in the Galaxy. Non-canonical
ered radio pulsars have inferred fields J1014 classesofpulsarsaddtothesenumbersonlynegli-
G (Camilo et al., 2000; McLaughlin et al., gibly because their effective birth rates are smaller
2003), rivalling those inferred for ‘‘magnetar’’ by a factor 10(cid:3)4–10(cid:3)3.
objects identified through their X-and-c radia-
tion that seems to derive from non-rotational 2.2. Galactic center
sources of energy. The relationship between
magnetars and these high-field radio emitting TheGalacticCenter(GC)isanespeciallytanta-
pulsars, whose radiation derives solely from lising but exceedingly difficult region to search for
spin energy, is not yet known. In the P–P_ dia- pulsars. In many respects, the GC is similar to an
gram of Fig. 1,these radio pulsars are typically especially large globular cluster with regard to the
located around P(cid:1)5 s and P_ (cid:1)10(cid:3)13. density of stars (cf. Section 2.3). It differs in that
molecular material and, hence, star formation are
Weenvisionafull-Galacticcensusofradiopul- both much more prominent in the GC than in
sars that aims to detect at least half of the active globulars. Additionally, the (cid:1)3·106M black
x
radio pulsars that are beamed toward Earth (Fig. hole (Scho¨del et al., 2002) that underlies the com-
2). The typical lifetime of canonical pulsars (cid:1)10 pactradiosource,SgrA*,perturbsspacetimesig-
Myr, where we define lifetime as the duration of nificantlyandisfedepisodicallybyinspiralinggas
the radio-emitting phase that, for objects with and stars.
B(cid:1)1012 G, is the time needed for a rapidly rotat- Why find pulsars in the GC? Any pulsars de-
ing object to reach the ‘‘death band’’ on the right- tected in or beyond the GC that are viewed
hand side of the P–P_ diagram. At long periods through the region are potentially unique probes
short-wardofthedeathband,wherepulsarsspend of the gas, magnetic field, and space-time of the
most of their detectable lifetimes, the beaming GC region. First of all, pulsars in the GC can be
fraction (cid:1)20% (e.g., Emmering and Chevalier, used to probe the magnetoionic medium along
1418 J.M.Cordesetal./NewAstronomyReviews48(2004)1413–1438
the line-of-sight, including possible detection of To combat pulse broadening, observations at
the inner scale for electron density turbulence higher frequencies are therefore needed that ex-
(see Section 3.8). Second, the initial mass func- ploit the strong m(cid:3)4 dependence. For example,
tion and overall stellar evolution in the GC is 10 GHz observations yield (cid:1)0.11 s broadening,
likely to be very different from the rest of the small enough to allow detection of longer period
Galaxy. The number and ages of pulsars and pulsars. However, pulsar spectra are power-law
their binary membership will provide clues about in form, often with steep dependences (cid:2)m(cid:3)a with
these areas (see Sections 3.8 and 3.10). Finally, a ranging from 0 to 3 and Æaæ(cid:4)1.5. While ongo-
the large stellar density offers the possibility of ing searches may yield detections of a few pul-
finding pulsars with stellar black hole compan- sars in the GC, some of which may provide
ions, allowing unprecedented tests of gravita- important probes of the region, the sensitivity
tional theories in the strong field limit and the of the SKA is clearly needed at high frequencies
study of black hole properties. Similar studies to detect a meaningful sample of pulsars in the
will be possible for Sgr A*, the supermassive GC, as demonstrated in Fig. 3. The figure shows
black hole in the center, if pulsars in compact the radio luminosity (defined as L =flux den-
p
orbits around the black hole are found (see Sec- sity·distance2) at 1.4 GHz plotted against spin
tion 3.1). Pulse timing measurements may pro- period along with detection threshold lines for
vide the possibility of measuring the spin of the Green Bank Telescope (GBT) and for the
the GC(cid:1)s central black hole. SKA.
At present, none of the known pulsars is within The maximum detectable distance is
or beyond the GC. The primary reason is that an
(cid:1) L (cid:2)1=2
especially intense scattering region lies between D ¼ p N1=4;
us and the GC at close proximity ((cid:1)100 pc) to max Smin1 hp
the GC (Lazio and Cordes, 1998). The large scat- ffiffiffiffiffiffiffiffiffi
where S ¼mS = 2BT is the single-harmonic
tering has been known since shortly after Sgr A* min1 sys
threshold with m, the threshold signal-to-noise ra-
itself was discovered 30 years ago and it has been
tio,andN isthenumberofharmonicsdetectedin
h
probed through angular diameter measurements
the search power spectrum. The minimum lumi-
of OH/IR masers in the GC region and through
nosity for m=10 is thus
surveys for background AGNs. The implication
for pulsars is that, at a standard search frequency (cid:1) D (cid:2)2(cid:1)N (cid:2)(cid:3)1=2
of1.4GHz,apulseemanatingfromthelocationof Lpmin ¼87:5 mJykpc2 1 Mmapxc 16h :
Sgr A* would be broadened to (cid:1)300 s owing to
multi-path propagation. 2 Forshortperiodpulsars,thenumberofharmonics
Thepulsebroadeningtimes scaleswiththeob- detected will be less than the 16 number assumed
d
served angular diameter h as s ¼W Dh2=2c, in this equation.
d d s d
whereDisthesourcedistanceandW isageomet- Also shown in Fig. 3 are detection curves for
s
rical weighting factor that takes into account the the minimum detectable period P vs. L (1.4
min p
location of the scattering region relative to the GHz). These are shown for pulsars at the Galac-
source and observer. For the GC, the weighting tic Center that are detectable with the Green
factor is very large because the region is much Bank Telescope (GBT) at 9 GHz and with the
nearer the GC than the observer. For scattering SKA at 9 and 15 GHz. These curves were calcu-
by fluctuations in the electron density, h (cid:2)m(cid:3)2 lated for pulsars near the location of Sgr A* and
d
andthusthebroadeningtimeisaverystrongfunc- take into account pulse broadening from scatter-
tion of frequency, s (cid:2)m(cid:3)4. ing and the spectral dependence of the pulsar
d
flux density using a conservative value for spec-
tral index, a=2. The detectability was thus cal-
2 PulsarslocatedclosertousthanSgrA*butstillbehindthe
culated for surveys at these frequencies but has
screenarescatteredlessthanthiswhilepulsarsbeyondtheGC
viewedthroughthescreenaremorescattered. been scaled to the equivalent sensitivity at 1.4
J.M.Cordesetal./NewAstronomyReviews48(2004)1413–1438 1419
2.3. Globular clusters
GlobularClustersholdvastreservoirsofMSPs,
which are formed at a rate per unit mass which is
at least an order of magnitude higher than in the
Galactic disk. The reason for this overabundance,
which may be even higher than presently seen due
totheescapeofhighvelocityobjects,isthoughtto
be the formation of binaries via two- and three-
body encounters in the high density environments
of Globular Clusters.
Why perform a full census of globular clusters?
The greatly increased likelihood that many of
these MSPs in globular clusters will have under-
gone some form of dynamical interaction means
that the chances of finding exotic binaries, such
as the long sought-after MSP-black hole system,
are perhaps highest in globular clusters (see Sec-
tion3.1).Furthermorethepulsarsineachglobular
Fig. 3. Detectability of pulsars in the Galactic Center and in
other galaxies is shown in this plot of ‘‘pseudo luminosity’’ cluster will provide us with exceptional probes of
L =S D2 vs. spin period for pulsars in the current (April the history and evolution as well as its present
p 1400
2004)publicATNFcatalogue;S1400istheperiod-averagedflux properties, including the dynamics, gas content,
densityat1.4GHz.Givendutycyclesforpulsarsrangingfrom
accurate distances and proper motion (see Section
0.01to0.7,thepeakfluxdensitycanbealargemultipleofthe
3.9). Moreover, an open question is whetherglob-
period-averagedfluxdensity.Solidanddashedlinesmarkthe
minimumdetectablespinperiodP forpulsarsattheGalactic ular clusters contain massive black holes in their
min
CenterthataredetectablewiththeGBTat9GHzandtheSKA centers (Merritt et al., 2004; Gu¨rkan et al., 2004),
at9and15GHz.Scatteringisassumedtobeinafilledregion or possibly even binary black holes (Colpi et al.,
centeredonSgr A*with1/ecylindricalradiusof0.15kpc,as
2003). Pulsar timing with the precision achievable
includedintheNE2001model.Forpulsarsatthe location of
withthe SKA will provide us with atoolto reveal
Sgr A*, the dispersion and scattering measures are
DM =1577pccm(cid:3)3andSM =107kpcm(cid:3)20/3.Thecurves whether such systems are present. This can be
GC GC
are based on harmonic sums of the Fourier power spectra of achieved either by probing the inner most regions
dedispersed time series. They correspond to surveys at these of the cluster to reveal velocities or accelerations
frequenciesbuthaveallbeenscaledtotheequivalentsensitivity
expected due to the presence of a black hole, or
at1.4GHzassumingthatthepulsarfluxdensitiesscaleas(cid:2)m(cid:3)2.
by using the pulsars in the cluster as an in situ
Abandwidthof800MHzwasassumedinallcases.Integration
timesof1.5hand104swereusedfortheGBTandtheSKA, gravitational wave detector sensitive to the pres-
respectively.Inaddition,theSKAcurvesassumethatthefull ence of binary black holes (see Section 3.3).
sensitivityoftheSKAisavailable.Dottedhorizontallinesshow At present there are 76 pulsars known in 23
theminimumvaluesforL neededfordetectionat1.5GHzin
p globular clusters. Not all clusters are equally rich,
periodicitysearchesofpulsarsatdistancesof1and5Mpc.The
and it is not clear what determines the number of
dottedlinesassumedetectabilityisindependentofspinperiod,
whichwillnotbethecaseifthereissignificantscatteringalong active pulsars in a given cluster. Based on our
thelineofsight. knowledgeoftheluminosityfunctionoftheMSPs
intheGalacticdisk combined withthecontinuum
radio emission and the numbers of MSP-like
GHz. The calculations assume that the full sensi- X-ray sources associated with globular clusters,
tivity of the SKA is available. If only a fraction the total population is likely to be at least two or-
f can be used in a blind survey using a ‘‘core’’ ders of magnitude more than this.
c
array, then the curves will move upward by an Theremainingpulsarsintheseglobularclusters
amount logf . aredifficulttofindwithpresentinstrumentsdueto
c
1420 J.M.Cordesetal./NewAstronomyReviews48(2004)1413–1438
a combination of the intrinsic low luminosity of vidual giant pulses. Additional successes will fol-
MSPsandthetypicallylargedistancestoglobular low from targeted surveys of individual
clusters.Furthermore,mostpulsarsareincompact supernovaremnantsinthenearestgalaxies.InSec-
binaryorbits.Giventhepresentlylongintegration tion4.1,wediscusstherequirementsonsensitivity
timesneededtodetectthepulsars,thepulsesignal and field of view (FOV).
isoftensmearedoutduetotheDopplereffect.Re- Fig. 3includes detectionlinesforsurveys at1.4
cent developments of sophisticated techniques to GHz for pulsars at 1 Mpc and 5 Mpc assuming
correct for this smearing have resulted in an in- that the full SKA sensitivity is available. Full sen-
crease in the number of systems known, but blind sitivitywouldapplytotargetedsearchesof,forin-
searches are often too computationally expensive stance, supernova remnants in nearby galaxies.
andthevastmajorityofsystemsarestilltooweak However, (cid:1)full-FOV blind surveys will provide
for detection. only a fraction f ) corresponding to a core array.
c
Targeted SKA surveys of all Galactic globular With full sensitivity, a reasonable fraction of the
clusters will enable us to detect all of the pulsars luminosity function can be sampled out to 1 Mpc
contained therein which are beamed in our direc- and a few pulsars can be detected to 5 Mpc.
tion.Thiscompletecensuswilluncoverlargenum- Giant pulses from the Crab pulsar serve as a
bers of pulsars with which to study their usefulprototypeforestimatingdetectionofstrong
formation,evolution,spinparameters,binarynat- pulses from nearby galaxies. The strongest pulse
ure and emission properties. observed at 0.43 GHz in one hour has S/
N =104 – even with the system noise domi-
max
2.4. External galaxies nated by the Crab Nebula. For objects in other
galaxies, the system noise is dominated by non-
Pulsars beyond the disk of the Milky Way are nebular contributions, implying that the S/N in
knownonlyinglobularclustersandintheMagell- thiscasewouldhaveincreasedbyafactorofabout
anicCloudsowingtotheirintrinsicfaintness.With 300. We can estimate the maximum distance of
the SKA, galaxies in the local group, including detection at a specified signal-to-noise ratio, (S/
M31 and M33, are within reach using periodicity N) as
det
searches while giant pulses like those seen from
the Crab pulsar can be detected from galaxies D (cid:4) 1:6 Mpc (cid:1)fcASKA(cid:2)1=2; ð3Þ
out to at least 5 Mpc. max pffiðffiSffiffiffi=ffiffiNffiffiffiffiÞffiffidffiffieffitffi=ffiffiffi5ffiffi AArecibo
What is the importance of detecting pulsars in
other Galaxies? Pulsars likely to be detected will wheref A istheSKA(cid:1)scollectingareathatcan
c SKA
be young pulsars with high luminosities that can be used for a giant pulse survey and A is
Arecibo
becorrelatedwithcatalogsofsupernovaremnants Arecibo(cid:1)s effective area at 0.43 GHz. The usable
and will yield estimates of the star-formation rate area for the SKA in a blind survey will be limited
and the branching ratio for supernovae to form by what fraction f of the antennas are directly
c
spin-driven pulsars as opposed to magnetars and connected to a central correlator/beamformer.
blackholes. Extragalactic pulsars will also provide For f =1, A /A (cid:4)20, the standard 1-h
c SKA Arecibo
information about the magnetoionic media along pulse seen at Arecibo could be detected out to
the line of sight through determination of DM, 7.3 Mpc.
SM, and RM. Unambiguous study of the interga-
lactic medium in the local group requires removal
of contributions to these measures from the fore- 3. Fundamental physics & astrophysics
ground gas in the Galaxy and gas in the host gal-
axy. The more pulsars detected in a galaxy, the Having discovered a large sample of pulsars in
more robust this removal will be. the census of the Milky Way,the Galactic Center,
Extragalactic pulsars can be found through Globular Cluster and external galaxies, a vast
blind surveys for both periodic sources and indi- range of fundamental problems in physics and
J.M.Cordesetal./NewAstronomyReviews48(2004)1413–1438 1421
astrophysics can be studied. We highlight some of bemeasured.Forpointmasseswithnegligiblespin
those in the following. contributions, the PK parameters in each theory
should only be functions of the a priori unknown
3.1. Tests of theories of gravity NS masses and the well measurable Keplerian
parameters. With the two masses as the only free
The theory of General Relativity (GR) has to parameters, the measurement of three or more
date passed all observational tests with flying col- PK parameters over-constrains the system, and
ours. Nevertheless, one of the most fundamental therebyprovidesatestgroundfortheoriesofgrav-
questions remaining is whether Einstein(cid:1)s theory is ity (Damour and Taylor, 1992). In a theory that
the last word in our understanding of gravity or describes the binary system correctly, the PK
not. Solar system tests of GR are made under parameters produce theory-dependent lines in a
weak-fieldconditions,illuminatedbythesmallor- mass-mass diagram, which compares the masses
bital velocities and gravitational self-energies in- ofthetwoneutronstars,that allintersectinasin-
volved (e.g., the self-energies for the Sun and the gle point.
Earth,expressedinunitsoftheirrest-massenergy, The best example for such tests is currently gi-
are e (cid:1)(cid:3)10(cid:3)6 and e (cid:1)(cid:3)10(cid:3)10, whilst for a ven by the double pulsar system PSR
Sun Earth
pulsar and black hole (BH) we find e (cid:1)(cid:3)0.2 J0737(cid:3)3039 where five PK parameters are availa-
PSR
and e =(cid:3)0.5). Therefore, none of these tests blefortests,inadditiontothetheoryindependent
BH
of gravitational theories can be considered to be mass ratio of the two NSs (Lyne et al., 2004). For
complete without probing the strong-field limit, the relativistic binary systems that will be discov-
which is done best and with amazing precision ered and timed with the SKA, one would be able
using binary radio pulsars (Damour and Espos- to routinely measure five or more PK parameters,
ito-Fare`se, 1998). However, even the existing bin- severely over-constraining the system. In particu-
ary pulsar tests only begin to approach the lar, more double pulsar systems will be found,
strong-field regime and the discovery of more ex- and PK parameters that are currently impossible
treme binary systems is required. Indeed, the (or extremely difficult) to measure, such as those
SKA is the only instrument which promises to arisingfromaberrationeffectsandgeodeticpreces-
probe the strong-field limit via the discovery and sion, would becomeaccessible with theSKA, pro-
timing ofsuchsystems,inparticular thatofapul- viding also full 3-D information about the
sar orbiting a BH. The promises provided by the orientation of these binary systems.
discovery of such particular systems are discussed Most importantly, current PK parameters are
in detail in a companion key science article (Kra- only measured to the lowest post-Newtonian
mer et al., this volume), where we detail how pul- approximation. The timing precision achievable
sars can be used to test the ‘‘Cosmic Censorship withtheSKAmeansthathigher-ordercorrections
Conjecture’’ and the ‘‘No-hair’’ theorem for the are likely to become important, demanding the
description of BHs. Here, we concentrate on gen- development of a timing formula that is accurate
eral aspects of tests of gravitational theories using to at least second post-Newtonian order. For in-
pulsars timed with the SKA. stance, corrections would need to account for the
Since NSs are very compact massive objects, assumptions currently made in the computation
doubleneutronstar(DNS)binariescanbeconsid- oftheShapiro delay (seeSection3.2)thatgravita-
ered as ideal point sources. Finding and timing tional potentials are static and weak everywhere
DNSs in tight binary orbits – ideally close to coa- (Wex, 1995; Kopeikin and Scha¨fer, 1999; Kopei-
lescence and thus emitting strong gravitational kin, 2003). In addition, other effects such as those
waves–providestringenttestsoftheoriesofgrav- related to light bending and its consequences will
ityinthestrong-gravitational-fieldlimit.Testscan become important where an additional signal
beperformedwhenanumberofrelativisticcorrec- would be superposed on the Shapiro delay as a
tions tothe Keplerian description of the orbit, the typically much weaker signal, which arises due to
so-called ‘‘post-Keplerian’’ (PK) parameters, can a modulation of the pulsars(cid:1) rotational phase by
1422 J.M.Cordesetal./NewAstronomyReviews48(2004)1413–1438
the effect of gravitational deflection of the light in performgenuinelynewtests.Theseprospectsarise
the field of the pulsar(cid:1)s companion (Doroshenko from the assumption that the timing precision can
and Kopeikin, 1995). beimprovedbytwooreventhreeordersofmagni-
Like the DNS binaries, most NS-white dwarf tude over the current standards. However, as we
(NS-WD) binaries can also be considered as pairs discuss in some detail in Section 4.3, this requires
of point masses. Significant measurements of PK not only an increase in raw sensitivity but also
parameters can only be obtained in a few cases the ability to correct for propagation effects from
so far, since relativistic effects for NS-WDs are intervening plasmas and sufficient polarization
generally much smaller than for DNS binaries purity and (quasi-) simultaneous multi-frequency
(e.g., van Straten et al., 2001; Bailes et al., 2003). observations. Moreover, the intrinsic phase-jitter
However, with the SKA, PK parameters can be ofpulsarswillpreventthetimingprecisionofsome
determinedformanymoresystems.Suchmeasure- MSPs from reaching the theoretical limit given by
ments canbe similarly valuable since NS-WD sys- radiometernoise,sothattheapplicationofcorrec-
temstestdifferentaspectsofgravitationaltheories. tion schemes needs to be considered on a case-by-
For instance, as shown recently (Esposito-Farese, case basis. However, the current timing precision
2004), the tests provided by the PSR-WD system ofanumberofMSPsappearstobelimitedbytel-
J1141(cid:3)6545 are more constraining for the class escope sensitivity, as opposed to systematic effects
of tensor-scalar theories than those tests provided (e.g., Straten, 2003), suggesting that much im-
by the double pulsar. The reason is that unlike proved timing precision can be achieved for a
GR, some alternative theories of gravity, such as few objects and that new, unique tests of relativis-
tensor-scalar theories, predict effects that depend tic gravity will be enabled through greater tele-
strongly on the difference between the gravita- scope sensitivity.
tional self-energy of two orbiting bodies, which is Itisimportanttonotethatmanyoftheseexper-
large in NS-WD systems. Using such systems, imentsrequirecorrectionsofthemeasuredparam-
oneisabletoprobepossibleviolationofconserva- eter values for kinematic effects. For instance, the
tion of momentum, equivalence principles and ex- precision of GR tests achievable with PSR
pectedLorentz-andpositionalinvariances(Stairs, B1913+16 is limited by the accuracy to which the
2003). A manifestation of such effects would, for pulsar(cid:1)s motion and acceleration in the gravita-
instance,bemeasurableinthedetectionofgravita- tional potential of theGalaxy isknown (Weisberg
tional dipole radiation where as GR only predicts andTaylor,2003).Observedvaluesforparameters
quadrupole emission. liketheorbitaldecayrate,P_ =P ,orchangesinthe
b b
The computer power available when the SKA semi-major axis, x_=x, are affected by a kinematic
comes online will enable us to do much more DopplertermgivenbyDamourandTaylor(1992)
sophisticated searches in parameter space, includ-
D_ 1 V2
ing full searches in acceleration due to binary mo- (cid:3) ¼ K~ (cid:5)ð~a (cid:3)~a Þþ T; ð4Þ
D c 0 PSR SSB cd
tion, than possible today, and the SKA sensitivity
allows much shorter integration times, so that where K~0 is a vector from the Solar System Bary-
searchesforcompactbinarypulsarswillnolonger centre (SSB) towards the binary pulsar,
be limited. Hence, the combination of SKA sensi- ~aPSR and~aSSB are the Galactic accelerations at
tivityandcomputingpowermeansthatthediscov- the location of the binary system and the SSB.
ery rate for relativistic binaries is certain to The last term including the transverse velocity,
increasebeyondthenumberofatleasta100com- VT, and distance, d, to the pulsar is known as sec-
pact binary systems that we can expect from an ularaccelerationor‘‘Shklovskiiterm’’(Shklovskii,
extrapolation of the present numbers. 1970).Correctingforthistermtoderivetheintrin-
Insummary,testsofgravityaffordablewiththe sic values obviously requires precise astrometric
SKA will not simply be a continuation of the pre- information like proper motion and distances
sent tests at higher precision levels, but the better which can be derived using the VLBI capabilities
sensitivity and timing precision will enable us to of the SKA. In order to get precise distances for