フェルミガンマ線宇宙望遠鏡で見た銀河系内天体

フェルミガンマ線宇宙望遠鏡
で見た銀河系内天体
田中 孝明（京都大学）
On Behalf of the Fermi LAT Collaboration
Content
• Fermi Large Area Telescope
• Supernova Remnants
• GeV-Bright SNRs
• TeV-Bright SNRs
• Gamma-ray Binaries
• Gamma-ray Flares from Crab Nebula
• Summary
Fermi Large Area Telescope
Pair-production telescope launched in June, 2008
Energy Range: from 20 MeV to > 300 GeV
Angler Resolution: < 1° (68% containment at 1 GeV)
Effective Area: 8000 cm2 (on axis at 1 GeV)
Field of View: 2.4 sr (all-sky coverage in ~ 3 hr)
γ
ACD
e+
[surrounds
4x4 array of
TKR towers]
Tracker
Home Mission Instrument Collaboration Institutions Publications NASA Pictures Internal
e–
The Fermi
Large Area Telescope
Calorimeter
News
The 2012 American Physical Society's W. H. K. Panofsky Prize in Elementary Particle Physics is
Fermi Large Area Telescope
Pair-production telescope launched in June, 2008
Energy Range: from 20 MeV to > 300 GeV
Angler Resolution: < 1° (68% containment at 1 GeV)
Effective Area: 8000 cm2 (on axis at 1 GeV)
Field of View: 2.4 sr (all-sky coverage in ~ 3 hr)
γ
ACD
e+
[surrounds
4x4 array of
TKR towers]
Tracker
Home Mission Instrument Collaboration Institutions Publications NASA Pictures Internal
e–
The Fermi
Large Area Telescope
Calorimeter
News
The 2012 American Physical Society's W. H. K. Panofsky Prize in Elementary Particle Physics is
CGRO EGRET vs Fermi LAT
LAT Specifications & Performance
Quantity
LAT (Minimim Spec.)
EGRET
Energy Range
20 MeV - 300 GeV
20 MeV - 30 GeV
Peak Effective Area 1
> 8000 cm 2
1500 cm 2
Field of View
> 2 sr
0.5 sr
Angular Resolution 2
< 3.5° (100 MeV)
< 0.15° (>10 GeV)
5.8° (100 MeV)
Energy Resolution 3
< 10%
10%
Deadtime per Event
< 100 µs
100 ms
Source Location Determination 4
< 0.5'
15'
Point Source Sensitivity 5
< 6 x 10 -9 cm -2 s -1
~ 10 -7 cm -2 s -1
1
2
After background rejection
Single photon, 68% containment, on-axis
3 1-!, on-axis
4 1-! radius, flux 10 -7 cm -2 s -1 (>100 MeV), high |b|
5 > 100 MeV, at high |b|, for exposure of one-year all sky survey, photon spectral index -2
2FGL catalog contains 1873 sources (cf. 271 sources for 3EG catalog)
超新星残骸
GeV-Bright SNRs
Color: Gamma Rays by Fermi LAT
Contours: Radio Continuum by VLA
W44: Fermi LAT Spectrum
Red vertical bars: 1-σ statistical errors
Black vertical bars: systematic errors
Whipple
HEGRA
Milagro
Abdo+ 2010
Spectral break at a few GeV
π0-decay model can explain the data well
Leptonic scenarios have similar difficulties
Bremsstrahlung: difficult to fit the radio and GeV data at the same time
IC: requires large amount of electrons (~ 1051 erg)
W44: Fermi LAT Spectrum
Red vertical bars: 1-σ statistical errors
Black vertical bars: systematic errors
Whipple
HEGRA
Milagro
Abdo+ 2010
Proton Spectrum: Broken PL
s1 = 1.74, s2 = 3.7, Eb = 8 GeV
Wp = 6 × 1049 (n/100 cm–3)–1 (d/3 kpc)2 erg
Kep = 0.01
Similar Case: W51C
Age: 3.0 × 104 yr, Distance: 6 kpc
Abdo+ 2009
Count Map
(2–10 GeV)
π0-decay
Brems
IC
Contours: ROSAT X-ray (Koo+ 1995)
Dashed magenta ellipse: shocked CO clumps (Koo & Moon 1997)
Green crossed: HII regions (Carpenter & Sanders 1998)
Diamond: CXO J192318.5+143035 (PWN?) (Koo+ 2005)
One of the most luminous gamma-ray sources L = 1 × 1036 (D/6 kpc)2 erg s–1
Spectral steepening
π0-decay model can reasonably explain the data
Leptonic scenarios have difficulties similar to those for W44
H.E.S.S.
combinations of π 0 decay (long-dashed curve), bremsstrahlung (dashed curve),
and IC scattering (dotted curve). The sum of the three component is shown as a
solid curve.
H
H
p
H
0
be expected.
Given the interaction with a molecular cloud, we first attribute
the observed gamma-rays to the decay of π 0 mesons produced
in inelastic collisions between accelerated protons (and nuclei)
boundary (a flat elliptical template) as the spatial distribution of
and target gas (Figure 3). The gamma-ray spectrum of π 0 decay
the source gamma-rays. Different spatial distributions such as a
is calculated based on Kamae et al. (2006) using a scaling
flat elliptical template reduced in size (scaled by 0.5) are tested
factor of 1.85
for helium andSTUDY
heavy nuclei
(Mori
2009). Note
to estimate the systematic error. No.
Our1,conservative
estimate is
2009
GAMMA-RAY
OF SNR
W51C
that the scaling factor assumes the local interstellar abundance
!20% in 1–6 GeV and ∼30% above 6 GeV as the systematic
for target material and the observedmomentum
cosmic-ray are
composition.
assumed to be identical f
uncertainty attributable to the unknown shape of the source.
Contributions from bremsstrahlung protons
and inverse
(IC)As we argue bel
andCompton
electrons.
scattering by accelerated electrons are
also the
shown
in Figure
4. DISCUSSION
reflect
character
of 3.
magnetohydrodyna
Electron–ion and electron–electron bremsstrahlung
spectra
are
account for the radio synchrotron index α
The extended gamma-ray emission positionally coincident
computed as in Baring et al. (1999).Koo
The1994),
interstellar
radiation
we adopt
s = 1.5, though a
with SNR W51C has been studied using the Fermi LAT. The
field for IC (see Table 1) is comprised
of
two
diluted
blackbody
s = 1.7) could be reconciled with the
gamma-ray spectrum presented in Figure 3 is not fitted by a
components (infrared and optical) and
the the
cosmic
microwave
within
uncertainty.
The energetic parti
simple power law, exhibiting a remarkable steepening. Here, we
background (CMB). The infrared and
optical
components
are
be uniformly distributed in the volume o
discuss the origin of the extended emission and the underlying
adjusted to reproduce the interstellar
field radius
in the R = 30 pc. T
withradiation
an effective
particle spectra that give rise to the observed spectrum of
GALPROP code (Porter et al. 2008). Cooling effects due to eff
imply a shock velocity of vsh ∼ 400 km
photons.
ionization and synchrotron (or IC) losses, 52which introduce
1.6 × 10 erg cm3 , where ESN and n0 rep
The expanding shock waves driven by the supernova explocooling breaks in particle spectra in addition to pbr , are taken
kinetic
energy
anda the
interstellar density i
sion are expected to be the sites of the acceleration of multi-GeV
into account assuming constant particle
injection
over
period
blast wave
is propagating,
particles. To phenomenologically interpret the spectral curvaT0 ∼ 3 × 104 yr. The synchrotron cooling
becomes
important respectively. The
−3
ture in the LAT+TeV bands, a broken power law is adopted for
in the TeV band for leptonic models.n0 ∼ 0.3 cm . The radio images indicate
the momentum distribution of the radiating electrons/protons
electronsspectrum
are smoothly
distributed in a thic
Figure 4(a) shows the radio+gamma-ray
together
of Koo
& Moon
with the radiation model that uses the
parameters
in4(1997a)
Table 1. suggests the pres
! "−s #
! "2 $−∆s/2
p
p
×−310
M% engulfed
by the blast
We adopt here MH = 2.8 × 104 M%cloud
(n̄H =of10∼1
cm
), which
is
Ne,p (p) = ae,p
,
(1)
1
+
Figure 4.
multiwavelength
modeling
(see quoted
cloud
canThe
act total
as target
somewhat
larger than
the value
above.
energymaterial for relativ
p0
pbrThree different scenarios for the
Table 1). The radio emission (from Moon
&
Koo
1994)
is
explained
by
50
hydrogen ma
of the high-energy protons amountstotal
to W(atomic
× 10molecular)
erg,
p = 5.2 and
synchrotron
radiation,
while
the
gamma-ray
emission
is
modeled
by
different
volume is denoted by M = n̄H mp V . Not
which
is inversely(dashed
proportional
where p0 = 1 GeV c−1 . For simplicity,
the indices
and the
break curve),
combinations
of π 0 decay
(long-dashed
bremsstrahlung
curve), to MH , but insensitive to otherH
be expected.
and IC scattering (dotted curve). The sum of the three component is shown as a
Given the interaction with a molecular clo
solid curve.
Table 1
the observed gamma-rays to the decay of π
Parameters of Multiwavelength Models
in inelastic collisions between accelerated
boundary (a flat elliptical template) as the spatial distribution of
and target gas (Figure 3). The gamma-ray sp
the source gamma-rays. Different
a
Parameters spatial distributions such asEnergetics
is calculated based on Kamae et al. (200
flat elliptical
template
reduced
in size B(scaled by
0.5) are tested
Model
ae /ap
∆s
pbr
n̄H
Wp
We
factor
of 1.85 for helium and heavy nuclei
to estimate the systematic
error.
Our
conservative
estimate
is
−1
−3
50
50
(GeV c )
(µG)
(cm )
(10 erg)
(10 erg)
that the scaling factor assumes the local in
!20% in0.02
1–6 GeV
and ∼30%
above406 GeV as
the systematic
(a) π 0 decay
1.4
15
10
5.2
0.13
for
target material and the observed cosm
uncertainty
attributable
to
the
unknown
shape
of
the
source.
(b) Bremsstrahlung
1.0
1.4
5
15
10
0.54
0.87
Contributions from bremsstrahlung and inv
(c) Inverse Compton
1.0
2.3
20
2
0.1
8.4
11
scattering by accelerated electrons are also
4. DISCUSSION
Electron–ion and electron–electron bremss
Notes. Seed photons for IC include the CMB (kTCMB = 2.3 × 10−4 eV, UCMB = 0.26 eV cm−3 ), infrared
The
extended
gamma-ray
emission
positionally
coincident
computed
as in Baring et al. (1999). The i
(kTIR = 3 × 10−3 eV, UIR = 0.90 eV cm−3 ), and optical (kTopt = 0.25 eV, Uopt = 0.84 eV cm−3 ). The
total
with
SNR
W51C
has
been
studied
using
the
Fermi
LAT.
The
−1
energy content of radiating particles, We,p , is calculated for p > 10 MeV c .
field for IC (see Table 1) is comprised of tw
gamma-ray spectrum presented in Figure 3 is not fitted by a
components (infrared and optical) and the
simple power law, exhibiting a remarkable steepening. Here, we
Similar Case: W51C
Similar Case: IC 443
Spatially extended emission detected with the Fermi LAT
Similar spectral steepening to W51C and W44
π0-decay
Gaensler+ 2006
Abdo+ 2009
Gamma-Ray Production Site
Uchiyama+ 2010
correlated with
e.g. Ohira+ 2010
Emission from the Vicinity of W44
Uchiyama+ 2012
2
Y. Uchiyama
F IG . 1.— (a) Fermi LAT γ-ray count map for 2–100 GeV around SNR W44 in units of counts per pixel (0.◦ 1 × 0.◦ 1) in celestial coordinates
(J2000). Gaussian smoothing with a kernel σ = 0.◦ 3 is applied to the count maps. Green contours represent a 10 GHz radio map of SNR W44
(Handa et al. 1987). 2FGL sources included in the maximum likelihood model are shown as crosses, while those removed from the model are
indicated by diamonds. (b) The difference between the count map in (a) and the best-fit (maximum likelihood) model consisting of the Galactic
diffuse emission, the isotropic model, 2FGL sources (crosses), and SNR W44 represented by the radio map. Excess γ-rays in the vicinity of
W44 are referred to as SRC-1 and SRC-2.
2012), Galactic interstellar diffuse emission, and an isotropic
component (extragalactic and residual particle background).
The Galactic diffuse emission is modeled using the standard
ring-hybrid model, gal_2yearp7v6_v0.fits, with its normalization being left free. We use a tabulated spectrum written in
iso_p7v6source.txt as the isotropic diffuse emission. The
sky model. Significant excess γ-rays are seen in the vicinity
of W44; the features are referred to here as SRC-1 and SRC2. The statistical significance is found to be ∼ 9σ for SRC-1
and ∼ 10σ for SRC-2.
The residual count map depends weakly on the choice of the
spatial template that describes γ-rays from W44. Our simula-
ld in the filaments as nm ! 7 × 103 cm−3 and Bm ! 0.8
owing the prescription in Uchiyama et al. (2010).
mplicity we describe the energy distributions of CR
s and protons in the filaments as a cut-off power law
entum: ne,p (p) = ke,p p−1.74 exp(−p/pc), where the inhosen to match the radio spectral index of α ! 0.37
etti et al. 2007). The ratio of radio and γ-ray fluxes
e /k p = 0.05. The spectral break in the LAT specreproduced by pc = 10 GeV c−1 . As shown in Fig. 2,
ctrum below a few GeV is dominated by the decays
mesons produced in the dense filaments12 , while the
IG . the
3.— LAT
Fermispectrum
LAT residual
count map highlighting
thethe
γ-ray
partFof
is contributed
largely by
emission from the surroundings
of density
SNR W44.
bremsstrahlung.
The energy
of Magenta
the CR contours
propresent the synchrotron 2radio map of SNR
W44.
Green
contours
3
−1
−3
ounts
to u(Jp=!
4×
10 (Msh
/5 × 10over
Mvelocity
eV
cm
!)
show CO
1→
0) emission
integrated
from
30
to, 65
esc
ould
bewith
explained
reacceleration
of Galactic
CRs
km s−1
respect to by
the Local
Standard of Rest
(Dame et al.
2001),
tracing the molecular
complexcloud
that surrounds
SNR W44.
pre-existing
in a cloud
molecular
(Uchiyama
et al.The
−1
contours start
20 Kspecify
km s with
interval ofsource
10 K km
However,
wefrom
do not
the an
dominant
ofs−1 .
Some molecular
clouds notsince
associated
with W44
are also seen
along
ay-emitting
particles,
freshly
accelerated
CRs
the Galactic plane (dashed line) in the CO map.
so enter the filaments diffusively from the intercloud
m. Provided that u p corresponds to the mean CR denhe shell of the remnant,4.the
total kinetic energy of CRs
DISCUSSION
50
3
−1
50periphery of
0.4We
× 10
/5
×
10
M
)
erg.
have (M
discovered
GeV
γ-ray
sources
on
the
sh
!
esc
2
Nesc (p) structures
dominated by filamentary
synchrotron
−(r−Rescof
)2 /R
−(r+Rescradiation
)2 /R2d
d
e radio emission
− e is thought ,to (2)
n(p, r,t) = al.3/2
(Castelletti et4π
2007).
The
Rd Resc r
arise from radiatively-compressed gas behind fast dissociawhere
tive shocks driven into molecular
clouds that are engulfed by
%
the blastwaveR(Reach
≡al.2 2005).
DISMAssuming
(p)[t − tesctypical
(p)]. magnetic (3)
d (p,t)et
fields of molecular clouds, the GeV γ-ray flux relative to the
The
diffusion
coefficient
theenough
interstellar
medium
radio
flux is expected
to be of
high
to account
for theisγ-often
parameterized
as SNR W44, irrespective of the origin of the
ray emission from
high-energy particles (Uchiyama
& et al. 2010).'
Emission from the Vicinity of W44
Kinetic energy of the
escaped cosmic rays (W )
can be estimated
W
=
(0.3–3)
×
10
erg
SNR W44. It has long been known that a complex of giant
cloudsfrom
(GMCs)
surrounds SNRofW44;
2.molecular
Gamma-rays
the Surroundings
W44the spatial
extent is as large as 100 pc and the total mass of the complex
CR
distributions
in6M
the
surroundings
ofSeta
SNR
W44
amounts
to ∼ 1 × 10
et al. 1986;
et al.
1998).
! (Dame
beIndetermined
(1)emission
how CRs
intoofthe
Figure 3, the by
γ-ray
fromare
thereleased
surroundings
W44
is compared
CO diffusion
(J = 1 → 0)
map (Dame
al. in2001)
medium,
andwith
(2)athe
coefficient
ofetthe
integrated over
velocity
of 30
km s−1We
appropriate
r medium,
DISMa(p)
(e.g.,range
Gabici
et to
al.652009).
contheprotons
GMC complex.
The regions
of excess
overlap
lyforCR
since leptonic
emissions
areγ-rays
unimporwith
surrounding
GMC complex.
an
e-ptheratio
of ∼ 0.01.
Also, the non-detection of
The γ-ray emission in the vicinity of W44 can be ascribed
tron
radiotoemission
the GMC
complex
formally
possible from
imperfection
of the
maps ofindicates
gas column
ctron
bremsstrahlung
is notofthe
γ-ray
production
densities
used in the model
themain
Galactic
interstellar
diffuse
δ
p
2 −1
DISM (p) = 10 D28
cm
s ,
(4)
10 GeV c−1
with constants of D28 ∼ 1 and δ ∼ 0.6 based on the GCR
propagation model (Ptuskin et al. 2006). The diffusion coefficient in the close vicinity of a SNR may be different from
the Galactic average, for example, because of Alfvén waves
generated by CRs themselves (Fujita et al. 2010). We allow
D28 to vary, but fix the index to be δ = 0.6 for simplicity.
Given that the γ-ray emission by the escaping CRs is visible against SNR W44 in the LAT image, we expect the size
of the CR halo, RCR (p) = Rd (p,tage ) + Resc (p), is much larger
than the size of the remnant R = 12.5 pc. Since we find
RCR (100 GeV c−1 ) ! 2R for D28 = 0.1, we set D28 ≥ 0.1. Note
however that the constraint depends on the exact choice of
tage ; D28 ≥ 0.05 is obtained for tage = 20, 000 yr.
To calculate the γ-ray emission from the surroundings of
W44, we also need to specify the mass distribution. The total
F IG .of
4.—
the γ-rayilluminated
emission from by
the molecular
cloud CRs
mass
theModeling
cloud of
complex
the runaway
complex that surrounds W44. Data points
are from Fig. 2, but SRC5
is 1estimated
as
M
∼
5
×
10
M
by adding
sixwith
molecMC
! spectrum
and SRC-2 are co-added.
The SRC-1+2
(red up
points
ular
clouds
(Clouds
1–6) in
Seta
et al. γ-rays
(1998).
mass is
statistical
errors)
is attributable
to the
π 0 -decay
from The
the cloud
complex illuminated
by the CRs
that have escaped
(blueof the
assumed
to be uniformly
distributed
withinfrom
theW44
radius
curves). Three
casescomplex,
of diffuse coefficient,
D28pc.
= 0.1,Given
1, 3, are the
shown.
molecular
cloud
RMC = 50
spheriA gray curve indicates the γ-ray spectrum produced by the sea of
calGCRs
symmetry
of the model, we combine the γ-ray spectra of
in the same CR-illuminated clouds.
SRC-1 and SRC-2.
Uchiyama+
2012
Effectively,
model has filaments
three adjustable
parameters,
D28 ,
The dense our
radio-emitting
are indeed
the most
28
TeV-Bright SNRs
H.E.S.S. Images
RX J1713.7−3946
The Fermi LAT collaboration recently published the results (Abdo+ 2011)
Spatially extended source at the location of the SNR
The extent determined by a maximum likelihood fit is consistent with that of the
SNR observed in other wavelengths
– 19 –
Fermi LAT count maps (> 3 GeV)
1FGL J1714.5-3830c
Source B
1FGL J1716.9-3830c
400
Source C
Source A
1FGL J1705.5-4034c
1FGL J1707.9-4110c
1FGL J1702.4-4147c
1FGL J1707.1-4158c
262.0
300
100
1FGL J1716.9-3830c
c)
200
1FGL J1714.5-3830c
200
150
100
Source B
Source C
50
0
Source A
1FGL J1705.5-4034c
262.0
Before background subtraction
1FGL J1717.9-3729c
1FGL J1718.2-3825
200
260.0
258.0
256.0
Right Ascension (J2000, deg)
1FGL J1720.7-3707c
.0
500
1FGL J1717.9-3729c
-42.0
1FGL J1718.2-3825
b)
600
Declination (deg)
-40.0
-38.0
a)
1FGL J1720.7-3707c
1FGL J1717.9-3729c
-42.0
Declination (deg)
-40.0
-38.0
1FGL J1720.7-3707c
-50
1FGL J1707.9-4110c
1FGL J1702.4-4147c
-100
1FGL J1707.1-4158c
-150
260.0
258.0
256.0
Right Ascension (J2000, deg)
-200
After background (contributions from diffuse
backgrounds + other sources) subtraction
RX J1713.7−3946
Fermi LAT spectrum: Very hard with Γ = 1.5 ± 0.1 (stat) ± 0.1 (sys)
Leptonic Models
Hadronic Models
The Fermi LAT + H.E.S.S. spectrum can be fit well with leptonic models
How to reconcile with the large magnetic field?
If interpreted with hadronic models, extremely efficient particle acceleration is required to fit the data
(proton index must be sp ~ 1.5 to fit the Fermi LAT spectrum)
RX J0852.0−4622
Fermi LAT count maps
0.6
(b)
(a)
0.8
0.6
-46.0
0.4
-44.0
Dec [deg]
-44.0
Dec [deg]
Tanaka+ 2011
> 10 GeV
> 1 GeV
0.5
0.4
-46.0
0.3
0.2
-48.0
-48.0
0.2
0.1
138.0
136.0
134.0
132.0
RA [deg]
130.0
128.0
0
138.0
136.0
134.0
132.0
130.0
128.0
0
RA [deg]
H.E.S.S. contours
Spatially extended source at the location of the SNR RX J0852.0–4622
The emission clearly detected in the high energy region (Hereafter we show results with events > 5 GeV)
Using a uniform disk as a spatial template, we obtain a radius of 1.12 (+0.07, –0.06) deg,
which is consistent with the extent observed in radio, X-rays, and TeV gamma rays
Model
sp
p0p
[TeV c−1 ]
Hadronic
Leptonic
1.8
—
50
—
(1.7 ± 0.2 ± 0.4) × 10−11
86.1
−11
(2.8 ± 0.4 ± 0.6) × 10
102.4
−11
p
B
n
0e± 0.8 ± 0.6) × 10
(2.9
40.9
−1
−3
c and] second
[µG]
[cm statis]
Note. —[TeV
The first
errors denote
tical and systematic errors, respectively.
8.49
35.3
se147
Wp
[erg]
We
[erg]
5.2 × 1050
—
3.9 × 1046
1.1 × 1048
RX J0852.0−4622
1.8
2.15
Note. — Wp and We are total kinetic
10
25
100
12
0.1
0.1
Table 2
Parameters
for the Models
energy
of particles
integrated
Tanaka+ 2011
above 10 MeV.
Model
sp
p0p
[TeV c−1 ]
se
p0e
[TeV c−1 ]
B
[µG]
n
[cm−3 ]
Wp
[erg]
We
[erg]
Hadronic
Leptonic
1.8
—
50
—
1.8
2.15
10
25
100
12
0.1
0.1
5.2 × 1050
—
3.9 × 1046
1.1 × 1048
Note. — Wp and We are total kinetic energy of particles integrated above 10 MeV.
sync
π0 decays
IC
sync
IC
Figure 2.
Fermi LAT spectral energy distribution (SED) in 1–
AT300spectral
energy
(SED)
in 1–
GeV with the
H.E.S.S. distribution
SED by Aharonian
et al. (2007b)
plotted
Figure 3. Broadband SED of RX J0852.0−4622 with (a) the hadronic
together.SED
For the
LAT points,
the (2007b)
vertical redplotted
lines and the
.E.S.S.
byFermi
Aharonian
et al.
model 3.
andBroadband
(b) the leptonic
model.
The J0852.0−4622
radio data points are
in- (a) the hadronic
blackLAT
caps represent
statistical
and systematic
errors,
respectively.
Figure
SED
of
RX
with
ermi
points,
the
vertical
red
lines
and
the
tegrated fluxes of the SNR determined based on the 64-m Parkes
The dotted line indicates the best-fit power law obtained from the
model
and yields
(b)datathebyΓleptonic
TheThe
radio
data
points are int maximum
statistical
and Power-law
systematic
errors,
respectively.
fit
toGeV
theband.
Fermi
LAT spectrum
= 1.85
± 0.05
(stat)
± blue
0.17
radio telescope
Duncan
& model.
Green
(2000).
line(sys)
likelihood
fit for the entire
1–300
The buttershows thefluxes
X-ray flux
by Aharonian
al. (2007b),
who
an-the 64-m
fly shape
shows
the 68%
confidence
region.
The dashed
curve
is the
tegrated
ofestimated
the SNR
determined
based
on0.1
50 et
–3) Parkes
icates
the
best-fit
power
law
obtained
from
the
The
hadronic
model
requires
a
large
amount
of
protons
(5
×
10
erg
for
n
=
cm
0
alyzed
ASCA
data.
The
solid,
dashed,
and
dot-dashed
curves
repreπ -decay
spectrum
by 1–300
BerezhkoGeV
et al. (2009).
radio
telescopefrom
data
The blue line
d fit
for the
entire
band. The butter0 by Duncan & Green (2000).
sent
contributions
π
decays,
inverse
Compton
scattering,
and
How to reconcile
the
weak
magnetic
fieldshows
with
X-ray
filaments
inThe
the
case
of
the
leptonic
model
the
X-ray
flux
estimated
by
Aharonian
et
al.
(2007b),
who an68%
confidence
region.
The
dashed
curve
is
the
synchrotron
radiation,
respectively.
parameters
adopted
for
the
ton scattering, we considered the CMB as well as the interstelmodel calculation
are summarized
in Table
2.
ASCA data.
The solid,
dashed,
and dot-dashed curves repreradiation field
in infrared
bylarBerezhko
et al.
(2009).and optical bands. The spectrum alyzed
ガンマ線連星
Gamma-ray Binaries
Gamma-ray emitting X-ray binaries
First discovered by air Cherenkov telescopes
LS 5039
Aharonian+ (2005)
O star + ?
H.E.S.S. detected
Periodicity (3.9 days)
LS I +61° 303
Albert+ (2005)
Be star + ?
MAGIC & VERITAS detected
Periodicity (26.5 days)
Possible Scenarios
Mirabel (2006)
Particles accelerated at a jet from the compact object
or
at a shock generated by interaction between stellar wind and pulsar wind
LS 5039 in X and TeV
Folded Lightcurves
Suzaku XIS
X-ray
HESS
TeV Gamma
Aharonian+ 2006, Takahashi+ 2007
Geometry of LS 5039
= 0.058
外合 (Superior Conjunction)
逆コンプトン散乱の確率：大
ガンマ線吸収の確率：大
電子
= 0.0
近星点 (Periastron)
ガンマ線
O型星
可視・紫外光
= 0.5
遠星点 (Apastron)
電子
= 0.716
内合 (Inferior Conjunction)
逆コンプトン散乱の確率：小
ガンマ線吸収の確率：小
ガンマ線
観測者
田中 他 天文月報より
デルでは、放射は毎周期、似たものになることが予
という極めて短いタイムスケールの構造が安定して
は、さらに観測を重ねることが必要であろう。
LS 5039 by Fermi LAT
= 0.058
外合 (Superior Conjunction)
逆コンプトン散乱の確率：大
ガンマ線吸収の確率：大
電子
= 0.0
近星点 (Periastron)
ガンマ線
O型星
可視・紫外光
= 0.5
遠星点 (Apastron)
電子
= 0.716
内合 (Inferior Conjunction)
逆コンプトン散乱の確率：小
ガンマ線吸収の確率：小
ガンマ線
観測者
Anti-correlation between GeV and TeV flux as predicted
Abdo+ (2007)
LS 5039 Gamma-ray Spectrum
E2 F(E) [ erg cm-2 s-1 ]
Hadasch+ (2012)
Fermi, 30 months
H.E.S.S., SUPC, 2004/05
-10
10
H.E.S.S., INFC, 2004/05
10-11
10-12
10-13
3
10
104
5
10
6
10
107
Energy [MeV]
Cutoff energy of 2.2 GeV is too low compared to that expected
from γγ absorption (> 30 GeV)
1FGL J1018.6−5856
Periodic source (P = 16.6 days) found in 1FGL catalog
Follow-up observations confirmed this source is a X-ray binary
Power (photons cm−2 s−1)2
5.0×10−16
−16
4.0×10
3.0×10−16
Rate (ph cm−2 s−1 ×10−7)
Ackermann+ (2012)
5.6
5.4
5.2
5.0
4.8
0.0
0.5
1.0
1.5
2.0
Phase
2.0×10−16
1.0×10−16
0.0 −3
10
10 −2
10 −1
Frequency (days−1)
10 0
Radio Flux Density
(mJy)
0.05 Swift XRT (0.3–10 keV)
0.04
0.03
0.02
0.01
0.00
7
6
5
4
3
2
1 ATCA 5.5 GHz & 9 GHz
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
X−ray Count Rate
(counts s−1)
X-ray & Radio Lightcurve
Orbital Phase
Suzaku observation during last summer (PI: Tanaka)
Fig. 4. X-ray (upper panel) and radio (lower panel) observations of 1FGL J1018.6 5856 folded
on the orbital period. The X-ray data are from the Swift XRT and cover the energy range 0.3 to
10 keV. For the X-ray observations the different colors indicate data taken from different 16.58
day orbital cycles. For the radio data the green diamonds indicate 9 GHz and red circles 5.5
GHz data. The radio data are from the ATCA.
1FGL J1018.6−5856
かに星雲からの
ガンマ線フレア
Fermi LAT Images
Flux > 100 MeV [ 10-7 cm-2 s-1 ]
The April 2011 Flare
1
2
3 4
5
6 7 8
9
10
11
Buehler+(2012)
200
~ 9 min time bins
100
55660
55662
55664
55666
MJD
55668
55670
55672
55674
✛ Synchrotron nebula brightened by a factor of ~30
✛ Flux doubling time : 4–8 hours
✛ No change in pulsar flux and phase
Crab SED
ȷɳȯɭɈɭɳ
ԓȳɳɗɈɳ‫ܨ‬෾
ɀȹɈǘȖǻ
ஈૈ೰‫ޝ‬
Spectral Evolution
102
10
3
102
3
Energy [MeV]
102
10
10
3
102
10
3
Buehler+(2012)
10
10
-9
-10
0
F [ ergs cm-2 s-1 ]
1
0
0
2
3
4
10-9
Γ = 1.26 ± 0.11
εc = 361 ± 26 MeV
10-10
5
10
10
6
10
10
7
10
10
8
10-9
Lγ ~ 1036 erg/s
~ 1% of Lsd
10-10
9
10
11
Why Puzzling?
✚ Compactness
Doubling time t ~ 4-8 hours ➜ Emission region < ct ~ 3x10-4 pc
(Inner ring ~ 0.1 pc)
Large luminosity (~ 1% of spindown power) from a compact region
✚ Spectrum
Γ = 1.26 ± 0.11: Flare energy is carried by the highest energy
electrons εc = 361 ± 26 MeV: Appears to violate the radiation
reaction limit
synchrotron cooling time = gyroradius/c
➜ Cutoff of synchrotron spectrum must be:
2
εc < (9/4αF) mec = 160 MeV
Summary
•
•
Fermi LAT は順調に全天観測を進めている
いくつかの超新星残骸のスペクトルは中性パイ粒
子の崩壊でうまく説明できる
•
•
今後の低エネルギー帯域のデータ解析に期待
TeV ガンマ線源の連星系の他、新たなガンマ線連
星からも GeV ガンマ線を検出
•
•
かに星雲からフレアを観測
このフレアは既存の理論では説明が困難