Pulsars in rays

1
Pulsars in ?-rays MAGIC
Marcos López Moya INFN/Padua

• IFAE, July 17th, 2008

2
Outline I ? -ray sky pulsars II Models of
? -ray emission Estimation of detection
sensitivityIII Software/Hardware
developmentsIV Previous pulsar observations
with CTsV MAGIC detection of Crab pulsar!
3
The ?-ray sky and pulsars
4
Gamma(?)-ray Astronomy

• Relatively new discipline (Egt10MeV), in between
  Particle physics and Astrophysics

• Started in 1912 when Victor Hess discovered the
  cosmic rays
• nuclei (p,He,..), e ?, ?
• ?
• Charged cosmic rays not point to the source
• ? Only ?s can be used for astronomy
• Reveals the non-thermal Universe

5
Detection Techniques
Basic fact ?-rays absorbed in atmosphere

• Satellites
• Direct detection
• Small Effective Area lt1m2
• Few background
• Energylt30GeV

• Ground Detectors
• Indirect detection
• Huge Effective Area 105m2
• Enormous background
• Energygt100 GeV

6
Sources of ?-rays
SNRs
Pulsars
AGNs
Binary systems
Dark Matter
GRBs
7
Pulsars

• Pulsars are highly magnetized and rapidly
  rotating neutron stars
• Formation of a neutron star

Supernova explosion
Star
Collapse
Supernove remnant
Neutron star

• Typical mass 1.4 Msun, and radius ?10 km
• Extreme internal density and huge magnetic fields

? Unique lab for nuclear and particle physics
8
?-ray pulsars

• Present situation
• More than 1800 radio pulsars are known today.

• They can be grouped in canonical and ms

normal pulsars
millisecond pulsars (100 become known)
9
?-ray pulsars

• Present situation
• More than 1700 radio pulsars are known today.
• They can be grouped in canonical and ms

• Only 7 (3) detected in ?-rays, with EGRET

• 7 ?-ray pulsars
• 3 candiates

10
EGRET Pulsars Light curves

• Typically 2 peaks and interpulse emission.

• Crab is the only pulsar which presents the same
  behavior at all wavelengths !

11
Candidates to EGRET pulsars
12
Multi-wavelength spectra of EGRET pulsars

• Maximum of emission in the hard X- and ?-ray range

Cherenkov Telescopes

• Spectra are very different above 1 GeV
• High energy spectral cutoffs

13
Spectral tails of EGRET pulsars
No evidence of cutoff lt30 GeV for PSR B1706 PSR
B1951
14
Spectral tails of EGRET pulsars
Evidence of cutoff for Geminga and Crab
15
EGRET pulsars at GeV energies

• Crab pulsar

above 100 MeV Two peaks, broad structures
two100 GeV photons!
above 5 GeV First peak seems to disaperar
16
EGRET pulsars at GeV energies
Pulse Profiles above 100 MeV two peaks, broad
structures
P1
P2
Pulse Profiles above 5 GeV single pulses with
duty cycle ?5 or less.
P2
P1
17
Candidates to new ?-ray pulsars (I)

• Millisecond pulsars
• About 100 are known (several in binary systems)
• Very old systems up to Gyears
• Millisecond pulsars must be spun up by external
  mechanisms, e.g. accretion from the companion
  star.

• Very low magnetic field
• B?5x108 GeV
• High energy photons can escape pair-production

? Very good candidates for Cherenkov telescopes
18
Candidates to new ?-ray pulsars (II)

• Unidentified EGRET sources
• Pulsars are the only class of sources seen by
  EGRET in the galactic plane

? Many U.E.S in the galactic plane must be
pulsars !
19
Candidates to new ?-ray pulsars (III)

• Unidentified EGRET sources

• About 1/3 of U.E.S have hard pulsar like
  spectra.

• There are many (?30) coincidences between U.E.S
  and young pulsars.

• Young pulsars 3EG associations usually do not
  show spectral cutoff

Good candidates for Cherenkov telescopes
20
Models of ?-ray emission in pulsars
21
Models of ?-ray emission in Pulsars (I)
Pulsar Magnetosphere Goldreich Julian (1969)

• Rotation generates huge induced electric field,
  which overcomes gravity
• ? charges are pulled from star
• ? plasma fills magnetosphere

• The light cylinder
• divides the magnetosphere into
• open field lines crossing L.C.
• closed field lines confined inside

• Two main emission models
• Polar cap
• Outer gap

22
Models of ?-ray emission in Pulsars (II)
Polar Cap Model Sturrock (1971) Ruderman
Sutherland (1975) Harding (1981) Daugherty
Harding (1982)

• Acceleration of electrons
• Cooling mechanisms
• Curvature radiation
• Synchrotron, I.C. of X-rays
• ?-rays interact with magnetic field, via Magnetic
  pair production

Super-exponential cutoff
Polar Cap model predicts super-exponential cutoff
in high energy ?-ray spectra !
23
Simulations of pulsar emission at GeV (I)
Goal Understand light curves and predict the
spectral tail of pulsars in energy range of CTS

• Lets follow the polar cap model

• The algorithm
• Given a pulsar P, B and ?
• Primary electrons are accelerated
• Simulate curvature radiation
• Interaction photons-magnetic field
• Calculate direction of emitted ? in sky
• Output
• ? Light curve are computed for different viewing
  angles
• ? Spectra are obtained

24
Simulations of pulsar emission at GeV (I)

• Understanding light curves
• Example Simulation of a ms pulsar with ?60º

• Light curves depends on
• pulsar geometry, hence on P (polar cap size ?
  P-1/2)
• Observation effects
• Different observers can see completely different
  light curves for the same pulsars
• 2 and 1 peak light curves are explained in this
  scenario

25
Simulations of pulsar emission at GeV (II)

• Predicting VHE Crab pulsar spectrum
• Simulation parameters B3.81012 G, P33ms
• Free parameters ?init, hinit
• Try to reproduce the EGRET spectrum

Eo 7 GeV

• EGRET spectrum reproduced with ?init2107,
  hinit3RNS
• Spectrum can be fit to a power law x
  super-exponential cutoff

26
Spectral cutoffs predicted by polar cap model
27
Models of ?-ray emission in Pulsars (III)

• ?-ray emission near LC
• Assume formation of vacuum gap in outer
  magnetosphere
• Charges accelerated in gap, escaping through L.C
  
• ? ?-rays via Curv. rad.
• B not strong enough for pair-production. But,
  curvature photons interact with non-thermal
  X-rays (or IC photons with IR)

Outer Gap model Cheng, Ho Ruderman (1986)
Romani (1996)
Out gap model predicts softer exponential cutoff
in the high energy ?-ray spectra !
Electrons may up scatter IR photons to TeV
Gamma-rays !
28
Models of ?-ray emission in Pulsars (IV)
Where do ?-rays come from? Outer gap or polar cap?

• Discrimination between models
• Different models predict different spectral
  cutoff.
• Measuring the spectral tail is possible to
  distinguish between models.

MAGIC SumTrig
29
Capabilities of MAGIC for detecting ?ray pulsars
30
MAGIC Observation time for EGRET pulsars (I)
Goal Estimate observation times needed by MAGIC
to detect EGRET pulsars at given significance
level

• Strategy
• Observation times given by

x ? significance

• We need to estimate expected ? rate
• Used following model for pulsar spectrum at GeV,
  assuming super-exponential cutoff

The values for K,?, E0, b, were obtained fitting
the spectra of the EGRET pulsars above 1 GeV
31
MAGIC Observation time for EGRET pulsars (II)

• Results
• Expected pulsed rates and required observation
  times, for a detection at 5?

More than 100 hours

• Crab and PSR B195132 in principle detectable in
  lt 30 h if cutoffs are above gt30 and 40 GeV
  respectively

32
Candidates among radio pulsars (I)
Goal Estimate how many radio pulsars could be
seen in ?rays by MAGIC

• Method
• Assume all radio pulsars are ?-ray emitters
• Estimate minimum required ?-luminosity to be
  detect the pulsar
• ? detectable pulsars will be those satisfying

• Problem How to estimate the ?-ray luminosity of
  a given radio
• pulsar, knowing only P, Pdot, d
  ?
• Luminosity is given by

Estimate dN/dE
33
Candidates among radio pulsars (II)

• ?-luminosity required for 5? in 30 h vs
  Sping-down-power, assuming E030GeV

• All pulsars with ?gt1 are undetectable ? 84 de
  studied radio pulsars
• For realistic value of ?1, we end with 22
  pulsars.
• Excluding those with Z.Agt20º, we end with 5
  candidates.

34
Timing Analysis
35
Timing analysis (I)

• Goal Find the periodic signal of the pulsar,
  hidden in
• the noise

• The timing analysis involves 4 steps
• Barycenter correction
• Obtain the Light curve
• Application of Uniformity test
• Upper limits calculation

36
Timing analysis (I)

• Barycenter correction
• Remove the effect of the earth movement on the
  arrival times tUTC.
• Transform the measured arrival times to the Solar
  System Barycenter

37
Timing analysis (II)

• Ligth curve
• If F is the known rotational frequency of the
  pulsar at time T0, the number of revolutions in
  dtt-T0 is
• Integrating, and taking the fractional part, we
  get the rotational phase ?
• where t is the barycenter time

Taylor
38
Timing analysis (III)

• Ephemeris are usually taken from radio
  observations but affected by irregularities in
  pulsar rotation
• Timing noise
  Glitches

Crab is glitching once every 3-6 years !
Need to have contemporaneous ephemeris We use
monthly ephemeris by Jodrell Bank
39
Timing analysis (III)

• Uniformity tests
• If no periodicity ? events will be uniformly
  distributed

• ?² - test
• Applied to k-bins histograms
• Powerful test for narrow pulses
• Probability follows a ?2 of k-1 d.o.f

• Zm2-test
• Independent of k
• Based on trigonometric moments of pulse profile
• mnumber of harmonics
• Sensitive to narrower pulses as m increases

• H test
• Find optimal m
• Powerful against a large variety of light curves

40
Previous pulsar observations from ground at TeV
energies
41
Results fromSolar Plants
CELESTE
42
CELESTE Crab observations

• No significance pulsed signal found.
• Obtained conservative upper limits

David A. Smith, 2002, CENBG
43
CELESTE Crab pulsar limit
44
Results from HESS
45
HESS results

• HESS searched for emission gt100 GeV from 7 young
  pulsars (4 were seen by EGRET)
• No pulsed signal found ? Upper limits

46
HESS results

• HESS searched for emission gt100 GeV from 7 young
  pulsars (4 were seen by EGRET)
• No pulsed signal found ? Upper limits

• U.l. implies that

• constrain IC component predicted by outer gaps

47
HESS results

• Only the Pulsar Wind Nebulae are visible at TeV

48
Results from MAGIC
49
Optical observations of Crab

• MAGIC PMTs designed to detect fast Cherenkov
  pulses ?2ns
• ? Need to be adapted to low frequency
  observations
• A PMT was modified to be set at the camera center
  for optical observations
• Electronic Chain
• Pre-Amplifier
• Signal transmission
• DAQ.
• Cpix signal is split in 2
• To 16 bits ADC, rate 2-20 kHz
• MAGIC FADC

50
Optical observations of Crab

• Observed part of the Crab campaign in optical and
  ? simultaneously ? Made Crab result robust !

MAGIC is the only telescope doing simultaneous
observations ?/optical
51
First Crab observations

• Data taken in Oct-Dec 2005.
• 16 hours of optimal quality
• A hint of a signal found
• 2.9? in phase with EGRET
• Derived upper limits
• Eolt27 GeV (exp. case)
• Eolt60 GeV (super-exp case)

J. Albert et al., Astrophys. J. 674,1037 (2008)
52
PSR B195132

• 31 hours taken in July-Sept. 2006
• No signal found
• Constrained cutoff energy to Eolt32 GeV

J. Albert et al., Astrophys. J. 669,1143 (2007)
53
Summary first MAGIC pulsar campaign

• No pulsed signal found from any of the observed
  pulsars
• But obtained the lowest upper limit so far
• Conclusion
• Even the low energy threshold of MAGIC (50-60
  GeV) was not enough for catching pulsars
• Solution Develop a new trigger threshold concept
  ? The MAGIC SumTrigger

54
MAGIC Crab pulsar detection
55
The difficulty of triggering pulsars
SumTrigger
A?(E)
FLUX or COLLECTION AREA
NO
PULSED SPECTRUM (EGRET)
YES
ENERGY
56
A new trigger concept

• Idea
• clusters of several pixels
• sum up analog signals from individual pixels
• discriminate on the summed signal
• Build at MPI (Munich) in summer 2007
• Advantages
• Small signals contribute to the trigger signal ?
  Lower threshol
• Improved signal/noise ratio.

57
A new trigger concept

• Improvements
• Size distribution peak shitfs to
• lower energies
• Higher coll. area at low energies
• New trigger rate 1kHz
• Sum- std trigger data taken in parallel

Trigger threshold decreased in a factor 2 50?25
GeV
58
Crab pulsar detection !!!

• Data sample
• Observations with new trigger between Oct07 and
  Feb08
• 24 h. of optimal data at low zenith angle
• Analysis
• 3 different analysis show clear signal gt 5?
• 2 independent timing software
• barycenter correctionfoldingperiodicity tests
• 4 different image cleaning algorithms

59
Crab pulsar detection !!!
Astronomical Telegram 1491 (28th April 2008)

• Results
• Clear signal at 6.4? in phase with EGRET light
  curve
• Observed 8.5k pulsed ?-ray events

P1 clearly visible, conversely to EGRET ?First
Surprise !
60
Additional periodicity tests
Fine frequency scan

• Combination of ephemerides from 5 periods allows
  frequency scan

x 10-6
-0.2 -0.1 0 0.1
0. 2
Relative frequency
Coarse frequency scan
x 10-6
-2 -1 0 1
2
Relative frequency
61
Crab Light Curve in different energy bands
P2 becomes dominant !
Preliminary
62
Estimation of the cutoff
FLUX or COLLECTION AREA

• Method
• Extrapolate EGRET spectrum with cutoff and fold
  with Aeff
• Calculate the expected excess and compare with
  measurement

63
Estimation of the cutoff
Preliminary
Exponential cutoff 16.4 - 1.5stat
- 4.5syst GeV Superexponential cutoff 20.5
- 1.5stat - 5.0syst GeV
64
Conclusions (I)

• Mechanism responsible for the pulsed emission of
  neutron stars unknown !

Polar cap, outer gap, both, other, ?

• We can try to find it by measuring the spectral
  cutoffs expected at tens of GeV

different models ? different cutoffs

• But, the number of known ?-ray pulsars is still
  very few and all of them discovered by CGRO/EGRET
  with very poor sensitivity at GeV

Try to detect them with CTs, with much higher
collection areas than satellites, BUT, higher
energy thresholds
65
Conclusions (II)

• All ground based ?-rays detectors have tried to
  detect pulsars, but none have succeeded so far..

• MAGIC built to achieve the lowest possible energy
  threshold
• First MAGIC pulsar observation campaigns already
  show a hint of Crab pulsar. New threshold built
  to confirm it.
• First detection of Crab pulsar with a CT!
• Both peaks visible!
• Cutoff higher than expected!

A new road open to new discoveries
66
Outlook The coming future
Exciting physics is coming

• AGILE successfully in space since more than 1
  year

Full coverage of the ??-sky from 100 MeV to 10 TeV

• GLAST just launched
• ?More than 100 new ?ray pulsars could be
  discovered !

• MAGIC II nearly ready
• first light in Sept. 2008

• HESSII under construction

67
Thank you for your attention !
The end.
68
(No Transcript)
69
BACKUP
70
Outline
Gamma-ray sky and Pulsars
Part I

• Detection techniques
• Pulsars in ?-rays (EGRET pulsars candidates)

Models of ? -ray emission and Predictions
Part II

• Models
• Sensitivity of MAGIC for detecting pulsars

Results of pulsars observations from ground
Part III

• Previous results
• MAGIC Crab pulsar detection!

71
The MAGIC Telescope

• 17 m diameter dish
• Ultra light carbon
• fibre frame
• Active mirror control
• 577 pixels,
• FOV 3.5º-3.8º
• Optical signal transport
• Fast pulse sampling 300MHz

72
?/hadron separation (I)

• Different kind primary particles
• ? different showers
• ? different images

• ?/hadron separation based on image parameter
  distributions
• ?-images are smaller and point to camera center
• Hadron showers are broader are randomly oriented

73
?/hadron separation (II)

• After applying ?/hadron cuts based on image
  shape, exploit shower direction

74
The ?ray sky

• Satellites give a nice crowded picture of
  energies up to 10 GeV.

271 sources 170 unid.
Unexplored Gap 30-250 GeV

• Ground-based experiments show very few sources.
• In spite they have much better sensitivity
  than EGRET !

?10 sources (increasing fast).
75
The Big Four
17 m
VERITAS(USA England)20074 telescopes10
meters Ø
MAGIC(Germany, Italy Spain)Winter 20041
telescope 17 meters Ø
HESS(Germany France)Summer 20024
telescopes 11 meters Ø
CANGAROO III(Australia Japan)Spring 20044
telescopes 10 meters Ø
76
The ?ray sky

• Satellites give a nice crowded picture of
  energies up to 10 GeV.

271 sources 170 unid.

• Ground-based experiments show very few sources
  with energies gt250 GeV.
• In spite they have much better sensitivity
  than EGRET !

?40 sources increasing very fast !
77
CELESTE Crab signal should be maximum on dry
nights
H relative humidity
Driest Hlt34
Drier H
Effect at the correct phase!
Dry Hlt57
wettest Hgt57
78
CELESTE PSR B15132
2000
2000 2001
2001
These two bins very close to EGRET phase
separation, BUT wrong absolute phase. Too small
to claim detection.
79
Multi-wavelength pulse profiles of Crab pulsar
Energy
80
Observation time for EGRET pulsars (I)
Goal Estimate the observation times needed by
MAGIC to detect EGRET pulsars at a given
significance level

• Strategy
• We have to assume a given spectral shape at GeV
  energies.
• Use the following model, which assume a
  super-exponential cutoff
• The values for K,?, E0, b, were obtained fitting
  the spectra of the EGRET pulsars above 1 GeV
  (Nel, de Jager 95)
• Then, multipliying by the collection area, we
  will get the rates
• and finally, the observation times will be given
  by

81
Observation time for EGRET pulsars (I)

• The big advantage of pulsars, is that we can base
  the detection on timing analysis, instead of DC
  excess.
• For that one has to use a given periodicity test,
  like the Z²m. Its xpected value, for a given
  signal is
• where
• is approx. the DC excess, and ? depends on the
  pulse profile
• If period pulse profile are known a priori,
  timing analysis enhance significance of the
  detection
• Example x3? excess in spatial analysis,
  and for a narrow single peak (?5.8 for a 5
  FWHM), one gets Z72, i.e, P810-8 (9?)
• But If P is unknown, we must multiply by number
  of trials M?T?f, which reduces the
  significance.
• Example for T6h, ?f30Hz, M6.5106. Then
  for x3? now P0.5

82
Predicted flux distributions of ?-ray pulsars
J13576435 J17401000 J17472958 B192910 J112459
16 J22296114 J02056449 B065614 B150958 B17064
4 J06311036
1043 pulsars included
Crab Geminga
Vela
83
Number of predicted pulsars to be discovered by
GLAST
as DC excess
GLAST should detect 750 pulsars as point sources.
This includes only 120 known radio pulsars.
84
Number of predicted pulsars to be discovered by
GLAST
in periodicity search
GLAST could detect ?100 pulsars in periodicity
searches
This includes ALL of the unidentified EGRET
sources.