Pulsars are Cool. Seriously.
Scott RansomNational Radio Astronomy Observatory /
University of Virginia
Surface temp ~106 K
Magnetic field (Gauss): Millisecond: 108-109
“Normal”: 1011-1013
Magnetar: 1014-1015
Detailed emission mechanisms unknown
Spin rates up to 716 Hz
“Luminosity” up to 10,000x the Sun's!
Surface gravity ~1011 times Earth's
Central densitiesseveral times nuclear
Neutron Stars1.2 - 2 Solar masses
10 - 12 km radii
Surface temp ~106 K
Magnetic field (Gauss): Millisecond: 108-109
“Normal”: 1011-1013
Magnetar: 1014-1015
Detailed emission mechanisms unknown
Spin rates up to 716 Hz
“Luminosity” up to 10,000x the Sun's!
Surface gravity ~1011 times Earth's
Central densitiesseveral times nuclear
Neutron Stars1.2 - 2 Solar masses
10 - 12 km radii
These are exotic objects
TheDiscoveryof Pulsars
PhD student Jocelyn Bell andProf. Antony Hewish
Initially “Little Green Men”Hewish won Nobel Prize in 1974
What are their radio properties?
• Continuum sources
• Typically somewhat to highly linearly polarized
• Steep radio spectra (index of -1 to -3, typical obs freqs 0.3-3 GHz)
• Point sources
• Special ISM effects (freq dependent)
• Highly time variable
• Wide variety of timescales
• Very faint average flux density ~mJy
Confusion?
Timing solns for33 Ter5 MSPs (VLA contours
in green)
None for pulsars!Pulsars separated via time (or spin frequency!) rather
than spatially.
Large beam?Doesn't matter!
Sub-arcsec positions come from pulsar timing.
Gain variations?Who cares?!
Observations are continually“on” and “off” source.
Fundamental Physics with Pulsars
Also many others:
• Plasma physics (e.g. magnetospheres, pulsar eclipses)
• Astrophysics (e.g. stellar masses and evolution)
• Fluid dynamics (e.g. supernovae collapse)
• Magnetohydrodynamics (e.g. pulsar winds)
• Relativistic electrodynamics (e.g. pulsar magnetospheres)
• Atomic physics (e.g. NS atmospheres)
• Solid state physics (e.g. NS crust properties)
Gravitational wave detection (e.g. high precision timing)
Physics at nuclear density (e.g. neutron star interiors)
Strong-field gravity tests (e.g. binary pulsar dynamics)
Basic Physical Information from Pulsars• Rotating dipole magnet in a vacuum (I = 1045 g cm2):
• radiates energy and therefore spins-down (p-dot)
• Surface magnetic field strength (B)
• Spin-down luminosity (E-dot)
• Age (T) and Characteristic Age (c) (braking index: n ~ 3)
P-Pdot DiagramPulsar Hertzsprung- Russell Diagram
HR Diagram:Temp (color) vs Luminosity
P-Pdot DiagramPeriod vs Spindown rate
Pulsar Flavors
Young
Old
High B
Low B
Young(high B, fast spin, very energetic)
CrabNebulaSN1054AD
Pulsar rotates30 times
per second!
Anasazi Indian cave pictogram,Chaco Canyon, NM
The Crab is visible at all energies!
Red = RadioGreen = OpticalBlue = X-ray
Pulsar Flavors
Young
Old
High B
Low B
Young(high B, fast spin, very energetic)
Pulsars move down and right across the diagram as they lose energy (assuming that the magnetic field doesn't change...)
Pulsar Flavors
Normal(average B, slow spin)
Young
Old
High B
Low B
Young(high B, fast spin, very energetic)
Science with “normal” pulsars
Used to:– study the unknown
pulsar emission mechanism
– probe the interstellar medium (scattering, scintillation, rotation measures, electron distribution)
– Measure PSR distances (HI absorption)
Drifting Sub-pulsesBhattacharyya et al 2007
ScintillationWalker et al 2008
Pulsar Flavors
Normal(average B, slow spin)
Young
Old
High B
Low B
Young(high B, fast spin, very energetic)
Eventually they slow down so much that there is not enough spin to generate the electric fields which produce emission.
Their lifetimes are 10-100 Myrs.
Pulsar Flavors
Normal(average B, slow spin)
Millisecond(low B, very fast, very old, very stable spin, best for basic physics tests)
Young
Old
High B
Low B
Young(high B, fast spin, very energetic)
1982:
Enigmatic bright, steep-spectrum, polarized, and scintillating radio source... Using Arecibo:
1.558ms pulsar (640 Hz)!
Courtesy Bob Rood
21x faster than Crab!~half an octave above “Concert A”!
(6 pulses)
Millisecond Pulsars: via “Recycling”
Supernova produces a neutron star
Red Giant transfersmatter to neutron star
Millisecond Pulsaremerges with a white
dwarf companion
Picture credits: Bill Saxton, NRAO/AUI/NSF
Alpar et al 1982Radhakrishnan & Srinivasan 1984
Pulsar Flavors
Normal(average B, slow spin)
Millisecond(low B, very fast, very old, very stable spin, best for basic physics tests)
Young
Old
High B
Low B
Young(high B, fast spin, very energetic)
Recyc
ling!
The Primary Pulsar TelescopesArecibo
Jodrell BankParkes
GBT
New All-Sky Pulsar Surveys● All major radio telescopes are
conducting all-sky pulsar surveys● We know of only about 5% of the
total pulsars in the Galaxy!● Generate lots of data (~50MB/s!):
– 1000s of hrs, 1000s of channels, 15000 kHz sampling: gives more than a Petabyte!
● Requires huge amounts of high performance computing
– Processing 2 min of GBT data requires 2 days on a fast CPU!
– Millions of false positives
Green BankTelescope
DispersionLower frequency radio waves are delayed with respect to higher frequency radio waves by the ionized interstellar medium
t DM-2 (DM = Dispersion Measure)
Coherent Dedispersion
exactly removes this effect, but
is verycomputationally
difficult
High Freq
Low Freq
Scattering and Pulse Broadening
Multipath propagation causes frequency dependent pulse broadening.
-4.4
Searching for New Pulsars• Pulsars are:
• Very weak radio sources
• Binary pulsars show Doppler effects
• Often distant (therefore weaker and high DM)
• Predominantly found in the Galactic Plane (ISM effects)
• Solutions:• Use large telescopes and sensitive receivers
• Use longer integration times
• Use advanced algorithms to adaptively remove interference
• Use advanced algorithms to optimize sensitivity to weak binary MSPs (the hardest PSRs to detect)
Sensitivity (A /Ttot) (tint BW)1/2 Computations Fspin
3 tint
2
Basic Radio Pulsar Search Recipe
Step (% of CPU Time)1. Interference identification and removal (1%)2. De-dispersion of the raw data (5%)3. Normal FFT search (slow pulsars) (15%)4. Acceleration search (binary MSPs) (60%)5. Single-pulse search (15%)6. Sifting of candidates (<1%)7. Folding of candidates (3%)
Processing a single ~2-min “pointing” takes ~2 days!
Big surveys have ~105 pointings, therefore 5+ CPU centuries!
Ter5 A(4th harm)
Ter5 N(3rd harm)
Single Pulse Searches• Some pulsars have highly variable pulse amplitudes or
shut off completely (i.e. nulling) RRATs• Look for dispersed individual pulses (e.g. McLaughlin &
Cordes, 2003, ApJ, 596, 982)
New PALFA Pulsar J1904+07
Year
Numbers have:quadrupled in last 10 yrsdoubled in last ~3 years
Why?Rise in computing capability, sensitive new radio surveys, Fermi!
New Millisecond Pulsars
Currently ~70 new Radio/gamma-ray MSPs because of Fermi!
Courtesy: Paul Ray
~10% of them look like they will be “good timers”
Millisecond Pulsars are Very Precise Clocks
PSR J1737+0747At 12:40PM PST February 17 2015:
P = 4.570136528819804 ms+/- 0.000000000000001 ms
The last digit changes by 1 every 2 minutes!
This extreme precision is what allows us touse pulsars as tools to do unique physics!
This digit changes by 1 every ~4000 years!
Observation 1
Pulses
Time
Obs 2
Pulsar Timing:Unambiguously account for every rotation of a pulsar over years
Obs 3Model(prediction)
Pulse Measurements(TOAs: Times of Arrival)
Predict each pulse to ~200 ns over 2 yrs!
Measurement - Model = Timing Residuals
Time in days
Single day at telescope
Does it work?
PSR J1231-1411~3yrs of Fermi gamma-ray data
~3000 photons (~3/day)
~560 binary orbits
~24 billion rotations of MSP
Perfectly lined up from radio pulsar timing
2 Pulse Rotations
Demorest et al. 2010, Nature
...get a spectacular answer!
The measured difference between the semi-major and semi-minor axes is:
2.8 +/- 0.2 mm!
Ask the right question...
Highly circular orbit has a radius of ~3.4 million km(~5 x Solar radius or ~9 x Earth-Moon distance)
Demorest et al. 2010, Nature
The Binary Pulsar: B1913+16 First binary pulsar discovered at Arecibo Observatory by
Hulse and Taylor in 1974
NS-NS BinaryPpsr = 59.03 ms
Porb = 7.752 hrs
a sin(i)/c = 2.342 lt-s
e = 0.6171
ω = 4.2 deg/yr
Mc = 1.3874(7) M⊙
Mp = 1.4411(7) M⊙
Besides the normal 5 “Keplerian” parameters (Porb, e, asin(i)/c, T0, ω), General Relativity gives:
where: T⊙ GM⊙/c3 = 4.925490947 μs, M = m1 + m2, and s sin(i)
Post-Keplerian Orbital Parameters
(Orbital Precession)
(Grav redshift + time dilation)
(Shapiro delay: “range” and “shape”)
These are only functions of:- the (precisely!) known Keplerian orbital parameters P
b, e, asin(i)
- the mass of the pulsar m1 and the mass of the companion m
2
Besides the normal 5 “Keplerian” parameters (Porb, e, asin(i)/c, T0, ω), General Relativity gives:
where: T⊙ GM⊙/c3 = 4.925490947 μs, M = m1 + m2, and s sin(i)
Post-Keplerian Orbital Parameters
(Orbital Precession)
(Grav redshift + time dilation)
(Shapiro delay: “range” and “shape”)
These are only functions of:- the (precisely!) known Keplerian orbital parameters P
b, e, asin(i)
- the mass of the pulsar m1 and the mass of the companion m
2
Need eccentric orbit andtime for precession
Need compact orbit and a lot of patience
Need high precision,Inclination, and m
2
The Binary Pulsar: B1913+16Three Relativistic Observables: ω, γ, Porb
From Weisberg &Taylor, 2003
Indirect detection of Gravitational Radiation
In 1993, Russell Hulse and Joseph Taylor were awarded the Nobel Prize for their workon PSR B1913+16!
The Double Pulsar: J0737-3039 Faster spin, more compact orbit,
edge on system, 6 relativistic observables, 2 pulsars!
Overall, much better than Hulse-Taylor binary PSR.
Currently GR tests to ~0.01%!
Measured vsPredicted Relativistic
Shapiro Delay
Kramer et al., 2006, Science, 314, 97
Shapiro Delay
NRAO / Bill Saxton
Irwin Shapiro 1964Shapiro et al. 1968, 1971
J1614-2230: Incredible Shapiro Delay Signal
Demorest et al. 2010, Nature, 467, 1081D see Ozel et al. 2010, ApJL, 724, 1990
Mwd = 0.500(6) M⊙
Mpsr = 1.97(4) M⊙!Inclination = 89.17(2) deg!
Full Shapiro Signal
No General Relativity
Full Relativistic Solution
An MSP in a Triple Stellar SystemRecently with GBT:a stellar triple system!
Direct Gravitational Wave Detection (Pulsar Timing Array)● Looking for nHz freq gravitational waves from super massive black hole binaries
● Need good MSPs:● Significance scales with
the number of MSPs being timed
● Must time 20+ pulsars for 10+ years at precision of ~100 nanosec!
For more information, see nanograv.org
Australia Europe North America
Bill Saxton (NRAO/AUI)
Where do these GWs come from?
Coalescing Super-Massive Black Holes• Basically all galaxies have them• Masses of 106 – 109 Solar Masses• Galaxy mergers lead to black hole mergers• When BHs within 1pc, GWs are main energy loss• For “nearby” very massive binaries, we can get
10s of nano-second timing residuals
Potentially measurable with a single MSP, but much better using an array of MSPs.
What about the future?• We only know of about 2,000 out of ~50,000+
pulsars in the Galaxy!
• Many of them will be “Holy Grails”• Sub-MSP, PSR-Black Hole systems, MSP-MSP binary
• Several new huge telescopes...
We need them because we are sensitivity limited!
FAST (500m, China))MeerKAT (64 dishes, SA)
Summary
Pulsars are Cool. Seriously.