the timing behaviour of radio pulsars george hobbs australia telescope national facility...

The timing behaviour of radio pulsars

George Hobbs

Australia Telescope National Facility

[email protected]

CSIRO. Gravitational wave detection

Contents

• Radio pulsars• Pulsar timing• A few things that you can do with pulsar timing• Young pulsars - a new (predictive?) model for timing noise

• What has this to do with this conference? - pulsars are compact objects - radio pulsar timing is a powerful technique for studying pulsars - can determine parameters of interest - neutron star masses, rotation rates etc. - can study the pulsar spin-down => implications for internal structure of neutron star.


Let’s start at the beginning

08:35:20.61 -45:10:34.87


Radio pulsars

QuickTime™ and aMicrosoft Video 1 decompressorare needed to see this picture.

Animation: Michael Kramer


Properties of radio pulsars


Must average many thousands of pulses together to obtain stable profile

Must convert to reference frame suitable for the timing model – e.g. solar system barycentre

Must convert to arrival times at infinite frequency

Must convert to conform with terrestrial time standards

Must add extra propagation delays e.g. through the solar system

Pulsar timing: The basics(see Hobbs, Edwards & Manchester 2006, MNRAS)

Obtain pulse arrival times at observatory

Model for pulsar spin down

Form timing residuals – how good is the timing model at predicting the arrival times

Improve timing model


What can we do with the timing model?


Some examples: pulsar velocities

• With long data spans can get accurate pulsar proper motions - with a distance estimate can obtain velocities.

• Mean space velocity ~ 400km/s (Hobbs et al. 2005)


Determine pulsar masses and testing GR

• Champion et al. (2008) Science: PSR J1903+0327, NS mass = 1.74 ± 0.04 Mo (unusually large)

• Double pulsar (B) has mass 1.25 Mo - significantly smaller (Lyne et al. 2004 Sci)


What can we do with the timing residuals?

• Residuals are a measure of unmodelled physics

• Are these residuals from …• The pulsar spin-down• Terrestrial time standards• Pulse propagation through the

interstellar medium• Orbital companions to the

pulsar• Gravitational waves!• Errors in the planetary

ephemeris• …


Spin-down irregularities

No angular signature


Terrestrial time standard irregularities

Monopolar signature


Errors in the planetary ephemerides - e.g. error in the mass of Jupiter

Dipolar signature


What if gravitational waves exist?

Quadrapolar signature


The post-fit planet ‘signal’: The effect of fitting

CSIRO. Measuring the mass of Jupiter using pulsars

Jupiter

Mars

Simulations of 10 years of pulsar residuals with an RMS of 100ns


Current status (Champion et al. 2009, in prep)

• Use data from Parkes, Arecibo, Effelsberg and Nancay radio telescopes

MSun Best Published This work

Mercury 1.66013(7)x10-7 1.660(2)x10-7

Venus 2.4478686(4)x10-6 2.44782(10)x10-6

Mars 3.227151(9)x10-7 3.2277(8)x10-7

Jupiter* 9.547919(8)x10-4 9.547916(4)x10-4

Saturn 2.85885670(8)x10-4 2.858858(14)x10-4

9.54791915(11)x10-4


QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

The timing residuals of young pulsars

• 76-m Lovell Radio Telescope• 366 pulsars with tspan > 10yr• Hobbs, Lyne & Kramer (2004)





Not high time precision experiments


Pulsar timing residuals (fit for F0 and F1)






Difficulties when categorising timing noise

• B1746-20



• B1900+01


Difficulties when categorising timing noise: depends on data span

• PSR B1818-04• Any simple

classification scheme would change with data span.

• Most large-scale analyses of timing noise used ~3 yr of data.




What timing noise is not!

• Not observatory dependent - many pulsars also observed at other observatories - see same timing noise

• Not off-line processing (use ‘tempo2’ and ‘psrtime’)

• Not terrestrial time scales/planetary ephemeris errors - too large

• Not ISM effect - not frequency dependent




Previous models of timing noise

• Random walks in the pulse frequency or its derivatives• Free-precession of the neutron star• Unmodelled planetary companions• Asteroid belts• Magnetospheric effects• Interstellar/interplanetary medium effects• Unmodelled Post-Keplerian orbital parameters• Accretion onto the pulsar’s surface• Large numbers of small glitch events

• These models were based on short data sets• Mainly model random, “noise-like” timing residuals


Significant F2 values (= cubics in timing residuals)

• Glitch events => F2 > 0 (Lyne, Shemar & Graham-Smith 2000)

• All pulsars with c < 105 yr have F2 > 0

• For older pulsars 52% have F2 > 0.



• Timing noise in young pulsars caused by glitch recovery.

• Timing noise in older pulsars caused by something else!

• Globular cluster pulsar


Periodicities: B1540-06

• Significant 4.38yr periodicity• If planet then Earth-mass. However,

significant residuals remain in the timing after fitting for a planet







• Time between successive peaks range from 3.4yr to 6.6yr

• Radius of curvature smaller at local maxima than at minima







• Time between peaks ranges between 7 and 10 years. No significant individual periodicities.







• Significant periodicity at 2.9yr (however time between peaks varies by ~10%).

• Local maxima have smaller curvature than minima







• Significant periodicities - main periodicity at 500d.

• 3 components to the slow-down• Modelled by Stairs et al. as free-

precession






Periodicities: B2148+63

• Significant periodicity at 3.2yr, 7.1yr and 2.1yr.

• Larger radius of curvature at maxima than at minima






PSR B1931+24

• PSR B1931+24 has recently been reported to undergo “extreme nulling” events (Kramer et al. 2006)

• Normal pulsar for 5 to 10 days• Switches off for up to 35 days• The pulsar spin-down rate changes

by ~50% between the on and off states (pulsar spinning down faster when “on”)


Directly looking at F1 values

• PSR J2043-2740• First pulsar we looked at:• Has 2 F1 values• Has correlated pulse shape

changesQuickTime™ and a

TIFF (Uncompressed) decompressorare needed to see this picture.


Modelling B1828-11

• Implication: B1828-11 is not undergoing free-precession!• Undergoes mode 1, 2, 3, 2 ….




More 1828-11 simulations


PSR J1107-5907

• Recently discovered pulsar with three pulse profiles:• 1) a very strong profile (brightness rivals that of Vela)• 2) weak profile• 3) completely undetectable

• => some pulsars exhibit 3 “magnetospheric modes” - have not yet checked to look for correlated slow-down rates.


The implications of the model

• Have a link between various time-dependent phenomena in pulsars: long-term moding/extreme nulling/intermittency/free-precession/timing noise

• Timing noise linked to magnetospheric changes• Quasi-random nature of the mode switches => a random walk in F1 => large

scale cubics can exist in the timing residuals

• Note: have no understanding of the process creating multiple spin-down rates, but it seems that large changes in spin-down rate => large pulse shape changes.

• 50% change in spin-down rate B1931-24 (large shape changes)• ~% change in spin-down rate B1828-11 (moderate shape changes)• Fraction of a % change in spin-down rate B1540-06 (small shape changes?)


Glitches


An aside: slow glitches

• Zou et al. (2004) reported a new phenomenon known as “slow glitches”

• No difference between “slow glitches” and timing noise!



• B1822-09 (vertical lines are slow-glitches according to Shabanova (2007)


Glitches

• Sudden speedup in rotation period, relaxing back in days to years

• The pulse structure is not notably affected by a glitch => phenomena internal to the neutron star

• Current model is that superfluid vortices in the neutron star ‘pin’ to the surface/crust. Catastrophic unpinning leads to a glitch event.

• 285 glitches published in 101 objects.• 65% of the glitching pulsars have only glitched once• PSR J1740-3015 has glitched 33 times


Glitches

• Melatos, Peralta & Wyithe (2008, ApJ) suggest that glitch events follow an avalanche model.

• Waiting time between glitches is consistent with a Poissionian process.

• … we’re writing a new paper containing more pulsar glitches … • What would you like us to present? Clearly, the glitches are

telling us something about the interior of the pulsar … but how do we extract the information?


Conclusion

• You can do lots of physics/astronomy with radio pulsar timing observations

• Most millisecond pulsars are very stable rotators• The spin-down of the youngest pulsars is dominated by glitch

recovery• The spin-down of most pulsars is dominated by a quasi-

periodic phenomenon.

• This is probably telling us something about the interior of the neutron star!

the timing behaviour of radio pulsars george hobbs australia telescope national facility...

Documents

timing noisewhat

years of pulsar residuals

timing noiseb174620b1900

pulsar timingyoung pulsars

largedouble pulsar b

accurate pulsar proper

ns mass

nancay radio telescopes9