gamma rays from molecular clouds (part 2)cappa/work_shop/gabici2.pdf · gamma rays from the milky...

Gamma rays from molecular clouds(part 2)

Stefano GabiciAPC, Paris

Overview

Brief introduction on cosmic rays

The supernova remnant hypothesis for the origin of cosmic rays

Why gamma ray astronomy?

Why molecular clouds?

Part 1: Passive Molecular Clouds -> how to use MCs as cosmic ray

barometers

Part 2: Illuminated Molecular Clouds -> how to use MCs to locate CR

sources and constrain CR propagation

|----- Class #2 -----|

(1) Molecular clouds as cosmic ray barometers

Gamma rays from the Milky Way

dnγ(Eγ) =qγ(Eγ ,r )

4πr2d3r =

qγ4π

dr dΩ

The gamma ray emission from the Milky Way is mainly due to p-p interactions

photon flux at Earth

gamma ray emissivity



4πr2d3r =

qγ4π

dr dΩ




dnγ

dΩ=

rmax

0dr

qγ4π

∝ rmax

0dr NCR(ECR, r) ngas(r)



4πr2d3r =

qγ4π

dr dΩ




dnγ

dΩ=

rmax

0dr

qγ4π

∝ rmax

0dr NCR(ECR, r) ngas(r)

measured by EGRET, FERMI...

If we know (we do) the gas distribution in the Milky Way we can use gamma ray observations to constrain spatial variations of the CR intensity

Molecular Clouds as CR barometers: GeVs

<- FERMI sky = CR+gas


<- FERMI sky = CR+gas

Casa

nova

...SG

... e

t al

, 201

0

gas column density ->

CO data from NANTEN telescope

Gamma rays from the Milky WayNANTEN: CO (J=1-0)

-> tracer of H2

LAB HI Survey (Karberla et al 05)

ASSUMPTION: spatial homogeneity of CRs

Distance (kpc)110 1 10

)1 k

pc1

s2(p

hoto

ns c

m

1010

910

810 sum

HIH2

Casa

nova

et

al, 2

010

@1 GeV<--


-> tracer of H2




)1 k

pc1

s2(p

hoto

ns c

m

1010

910

810 sum

HIH2

Casa

nova

et

al, 2

010

@1 GeV<--


-> tracer of H2



2% 84% 14%


)1 k

pc1

s2(p

hoto

ns c

m

1010

910

810 sum

HIH2

Casa

nova

et

al, 2

010

@1 GeV<--


-> tracer of H2



2% 84% 14%

the main contribution to the gamma ray emission comes

from ~ 1 kpc. If the CR intensity is different there, we expect a different flux.

-> “Tomography” with gamma rays!

CR are, on large scales, homogeneously distributed in

the MW (EGRET)


LSR velocity (m/s)160 140 120 100 80 60 40 20 0 20

310×

(K)

BT

0

20

40

60

80

100

H I (b)

Abd

o et

al,

2009

Goul

d Be

lt

loca

l arm

Pers

eus

oute

r ar

m

100o < l < 145o -15o < b < 30o

Recent results from the FERMI collaboration

<- density along line of sight


LSR velocity (m/s)160 140 120 100 80 60 40 20 0 20

310×

(K)

BT

0

20

40

60

80

100

H I (b)

Abd

o et

al,

2009

Goul

d Be

lt

loca

l arm

Pers

eus

oute

r ar

m

100o < l < 145o -15o < b < 30o

Recent results from the FERMI collaboration

<- density along line of sight

R (kpc)0 2 4 6 8 10 12 14 16 18 20

)1 sr1

(sH

Iq

0

5

10

15

20

252710×

radial profile ofgamma ray emissivity

->

Molecular clouds at GeV energies: FERMI

coun

ts/pi

xel

0

5

10

15

20

25

30

35

40

45

50

coun

ts/pi

xel

0

5

10

15

20

25

30

35

40

45

50

° = -30b

° = -20b

° = -10b

° = 0b

° =

190

l

° =

200

l

° =

210

l

° =

220

l

° =

230

l

Mono. R2

Orion A

Orion B

Orion molecular clouds

Oku

mur

a et

al,

2009


coun

ts/pi

xel

0

5

10

15

20

25

30

35

40

45

50

coun

ts/pi

xel

0

5

10

15

20

25

30

35

40

45

50

° = -30b

° = -20b

° = -10b

° = 0b

° =

190

l

° =

200

l

° =

210

l

° =

220

l

° =

230

l

Mono. R2

Orion A

Orion B


Oku

mur

a et

al,

2009

d ~ 400 pc+

CR spectrum from GALPROP(8% less than local CRs)


coun

ts/pi

xel

0

5

10

15

20

25

30

35

40

45

50

coun

ts/pi

xel

0

5

10

15

20

25

30

35

40

45

50

° = -30b

° = -20b

° = -10b

° = 0b

° =

190

l

° =

200

l

° =

210

l

° =

220

l

° =

230

l

Mono. R2

Orion A

Orion B


Oku

mur

a et

al,

2009

Energy (MeV)210 310 410

)-1

MeV

-1 s-2

cm2 (M

eVE

/dNd2

E

-610

-510

-410

Orion A Total

400M 3 10× 4.8) ± 7.5 ±(80.6 /ndf = 16.7/102

Orion A Total

400M 3 10× 2.6) ± 5.2 ±(39.5 /ndf = 9.3/102

Pionic

Electron Bremsstrahlung

d ~ 400 pc+



coun

ts/pi

xel

0

5

10

15

20

25

30

35

40

45

50

coun

ts/pi

xel

0

5

10

15

20

25

30

35

40

45

50

° = -30b

° = -20b

° = -10b

° = 0b

° =

190

l

° =

200

l

° =

210

l

° =

220

l

° =

230

l

Mono. R2

Orion A

Orion B


Oku

mur

a et

al,

2009

Energy (MeV)210 310 410

)-1

MeV

-1 s-2

cm2 (M

eVE

/dNd2

E

-610

-510

-410

Orion A Total

400M 3 10× 4.8) ± 7.5 ±(80.6 /ndf = 16.7/102

Orion A Total

400M 3 10× 2.6) ± 5.2 ±(39.5 /ndf = 9.3/102

Pionic

Electron Bremsstrahlung

d ~ 400 pc+


MA ∼ 8.1× 104M⊙

MB ∼ 4.0× 104M⊙

(2) Illuminated molecular clouds (MC/SNR associations)

Evolution of SuperNova RemnantsSN -> Explosion in a cold, uniform medium

ESN = 1051 E51 ergExplosion energy:

Mass of ejecta: Mej ≈ 1÷ 10 M⊙

Msw -> mass swept up by the shock

Msw << Mejif free expansion phase

Evolution of SuperNova RemnantsSN -> Explosion in a cold, uniform medium

ESN = 1051 E51 ergExplosion energy:

Mass of ejecta: Mej ≈ 1÷ 10 M⊙

Msw -> mass swept up by the shock

Msw << Mejif free expansion phase

ESN =1

2Mej v2 v = 109 E1/2

51

Mej

M⊙

1/2

cm/s

constant velocity

Evolution of SuperNova Remnantsthe free-expansion phase ends when: Mej ≈ Msw


uniform medium with density: gas ≈ 1.7× 10−24 g

4 π

3R3

s gas = Mej Rs ≈ 2

Mej

M⊙

13

pc

shock radius


uniform medium with density: gas ≈ 1.7× 10−24 g

4 π

3R3

s gas = Mej Rs ≈ 2

Mej

M⊙

13

pc

shock radius

duration of the free expansion phase: t ≈ Rs

v≈ 200

Mej

M⊙

− 16

yr

Evolution of SuperNova Remnants

Msw >> Mej -> the shock slows down

we want to find a relation between Rs and t




kbT2 =316

mu21 1 keV

let’s start by considering the shock heating of the gas

τc ∝ T12 106 yr

cooling time much longer than the SNR age!




SNRs emit X-rays

the SNR in this phase conserves the total energy!

kbT2 =316

mu21 1 keV

let’s start by considering the shock heating of the gas

τc ∝ T12 106 yr

cooling time much longer than the SNR age!

the only relevant physical quantities are: ESN gasand




the only relevant physical quantities are: ESN gasand




ESN

gas

t2

R5s

we can built a non dimensional quantity ->

Rs ≈ESN

gas

15

t25 Sedov solution

Evolution of SuperNova Remnantsduration of the Sedov phase

tage


tage vs ∝ t−35


tage vs ∝ t−35 T2 ∝ v2s


tage vs ∝ t−35 T2 ∝ v2s τc ∝ T

122


tage vs ∝ t−35 T2 ∝ v2s τc ∝ T

122

tage ∼ τcwhen -> radiative losses become important

tage ∼ 5× 104 yrthis happens at

Rends ≈ 20 pc

vends ≈ 200 km/s

Diffusive Shock Acceleration

ShockUp-stream Down-stream

u1u2

-->

B

B

B

B

B

B

Shock rest frame

Krymskii 1977, Axford et al. 1977, Blandford & Ostriker 1978, Bell 1978

See class by T. Bell


ShockUp-stream Down-stream-->

B

B

B

B

B

B

Up-stream rest frame

u1

u1 − u2

a

b

Ea = Eb


ShockUp-stream Down-stream-->

B

B

B

B

B

B

Down-stream rest frame

u1 − u2

a

b

Ea = Eb

u2

Particle acceleration during the free expansion phase

us

Up-stream

CR

shock

ld = us t

ld =√D t


us

Up-stream

CR

shock

ld = us t

ld =√D t

t =D

u2s

∝ Eequating...

Bohm diffusion -> λ = RL

D ≈ λc → D ∝ E


us

Up-stream

CR

shock

ld = us t

ld =√D t

t =D

u2s

∝ Eequating...

Bohm diffusion -> λ = RL

D ≈ λc → D ∝ E

max energy increases w. time

age max

Particles escape from SNRs

ush

Rsh

THIS IS A SNR

We need to know a bit of shock acceleration theory...

Diffusion length: ldiff ∼D(E)ush

∝ E

Bshush

see e.g. Ptuskin & Zirakashvili, 2003


ush

Rsh

THIS IS A SNR


Diffusion length:

Confinement condition:

ldiff ∼D(E)ush

∝ E

Bshush

D(E)ush(t)

< Rsh(t) → Emax ∼ Bsh ush(t) Rsh(t)



ush

Rsh

THIS IS A SNR


Diffusion length:


ldiff ∼D(E)ush

∝ E

Bshush

D(E)ush(t)


Sedov phase:

Rsh(t) ∝ t2/5

ush(t) ∝ t−3/5



ush

Rsh

THIS IS A SNR


Diffusion length:


ldiff ∼D(E)ush

∝ E

Bshush

D(E)ush(t)


Sedov phase:

Rsh(t) ∝ t2/5

ush(t) ∝ t−3/5

Emax ∝ Bsht−1/5



ush

Rsh

THIS IS A SNR


Diffusion length:


ldiff ∼D(E)ush

∝ E

Bshush

D(E)ush(t)


Sedov phase:

Rsh(t) ∝ t2/5

ush(t) ∝ t−3/5

Emax ∝ Bsht−1/5

Emax decreases with timeParticles with E > Emax escape the SNR

Bsh also depends on

time


Particle escape from SNRsPtuskin & Zirakashvili, 2003

ush [km/s]

t [yr]Emax [GeV]


ush [km/s]

t [yr]Emax [GeV]

200 yrs

~ PeV


ush [km/s]

t [yr]Emax [GeV]

~10 GeV

~105 yrs


ush [km/s]

t [yr]Emax [GeV]

~10 GeV

~105 yrs

particles are released gradually during the Sedov phase: the one with the

highest energy first, and the ones with lower and lower energies at later times

The W28 region in gammas, CO, & radio

W28

HESS - TeV emission


W28

HESS - TeV emission

Aharonian et al, 2008


W28

18h18h03m

-24o

-23o 30

´-2

3o

0

20

40

60

80

100

NANTEN 12CO(J=1-0) 0-10 km/s

l=+6.0o

l=+7.0o

b=-1.0o

b=0.0o

W 28 (Radio Boundary)

W 28A2

G6.1-0.6

6.225-0.569

GRO J1801-2320

HESS J1801-233

HESS J1800-240A B

C

RA J2000.0 (hrs)

Dec

J20

00.0

(deg

)

18h18h03m

-24o

-23o 30

´-2

3o

0

20

40

60

80

100


l=+6.0o

l=+7.0o

b=-1.0o

b=0.0o


W 28A2

G6.1-0.6

6.225-0.569

GRO J1801-2320

HESS J1801-233

HESS J1800-240A B

C

RA J2000.0 (hrs)

Dec

J20

00.0

(deg

)

NANTEN data - CO line

HESS - TeV emission



W28

18h18h03m

-24o

-23o 30

´-2

3o

0

20

40

60

80

100


l=+6.0o

l=+7.0o

b=-1.0o

b=0.0o


W 28A2

G6.1-0.6

6.225-0.569

GRO J1801-2320

HESS J1801-233

HESS J1800-240A B

C

RA J2000.0 (hrs)

Dec

J20

00.0

(deg

)

18h18h03m

-24o

-23o 30

´-2

3o

0

20

40

60

80

100


l=+6.0o

l=+7.0o

b=-1.0o

b=0.0o


W 28A2

G6.1-0.6

6.225-0.569

GRO J1801-2320

HESS J1801-233

HESS J1800-240A B

C

RA J2000.0 (hrs)

Dec

J20

00.0

(deg

)


HESS - TeV emission

Radio (Dubner et al 2000)



W28

18h18h03m

-24o

-23o 30

´-2

3o

0

20

40

60

80

100


l=+6.0o

l=+7.0o

b=-1.0o

b=0.0o


W 28A2

G6.1-0.6

6.225-0.569

GRO J1801-2320

HESS J1801-233

HESS J1800-240A B

C

RA J2000.0 (hrs)

Dec

J20

00.0

(deg

)

18h18h03m

-24o

-23o 30

´-2

3o

0

20

40

60

80

100


l=+6.0o

l=+7.0o

b=-1.0o

b=0.0o


W 28A2

G6.1-0.6

6.225-0.569

GRO J1801-2320

HESS J1801-233

HESS J1800-240A B

C

RA J2000.0 (hrs)

Dec

J20

00.0

(deg

)


HESS - TeV emission

Radio (Dubner et al 2000)


FERMI and AGILE - GeV emission

AGILE -> Giuliani et al, 2010

–23:00:00

–24:00:00

18:04:00 18:00:00

Right Ascension (J2000)

Dec

linat

ion

(J20

00)

0.1 0.2 0.3 0.4 0.5 0.6 0.7

HESS J1800-240

Source N

Source S

A B C

(a)

[counts/pixel]FERMI -> Abdo et al, 2010

W28: a cartoon

SNR shell

radio emission

-> GeV electrons

-> ongoing (re)acceleration?

W28: a cartoon

SNR shell

radio emission

-> GeV electrons


Molecular clouds

W28: a cartoon

SNR shell

radio emission

-> GeV electrons


Molecular clouds

gamma rays + gas

-> CR protons?

γ

γ

W28: a cartoon

SNR shell

radio emission

-> GeV electrons


Molecular clouds

gamma rays + gas

-> CR protons?

-> coming from the SNR?

γ

γ

CR

CR

W28: a cartoon

SNR shell

radio emission

-> GeV electrons


Molecular clouds

gamma rays + gas

-> CR protons?


-> projection effects?

γ

γ

~12 pc

~ 20 pcCR

CR

?

W28: a cartoon

SNR shell

radio emission

-> GeV electrons


Molecular clouds

gamma rays + gas

-> CR protons?



+ masers

γ

γ

~12 pc

~ 20 pcCR

CR

?

W28: a cartoon

SNR shell

radio emission

-> GeV electrons


Molecular clouds

gamma rays + gas

-> CR protons?



+ masers

-> most of the protons

already escaped (> GeV?)

γ

γ

~12 pc

~ 20 pcCR

CR

Escaping particles assumptions

-> point like explosion

-> (isotropic) diffusion coefficient independent on position

-> particles with energy E escape after a time: tout ∝ E−δ





this is just a parametrization





higher energy particles are released first

when the SNR is smaller









ld ≈ (D t)1/2

t = tage − tout

diffusion length

? model dependent...








ld ≈ (D t)1/2

t = tage − tout

diffusion length


≈ (D tage)1/2

high energy particles








ld ≈ (D t)1/2

t = tage − tout

diffusion length


≈ (D tage)1/2

high energy particles

lower energies -> point-like approx not so good + model dependent (tout ~ tage)

Escaping particles: CR spectrumtage -> ld = 10, 30, 100 pc

ld ≈ (D tage)1/2


const ∝ ηCR ESN

l3d

ld ≈ (D tage)1/2


const ∝ ηCR ESN

l3d

∝ e−

Rld

2

ld ≈ (D tage)1/2


const ∝ ηCR ESN

l3d

∝ e−

Rld

2

γ

ld ≈ (D tage)1/2


const ∝ ηCR ESN

l3d

∝ e−

Rld

2

γ

ld ≈ (D tage)1/2

0.1 ηCR 1

tage ≈ 3× 104 ÷ 105yr

ESN 2÷ 4× 1050erg

estimate D?

MNcl ≈ 5× 104M⊙

if we want SNRs to be the sources of CRs...

North

South

Fγ

Mcl

North

≈

Fγ

Mcl

South

Escaping particles: W28

North

South

Fγ

Mcl

North

≈

Fγ

Mcl

South

ld both clouds within a distance ld


North

South

Fγ

Mcl

North

≈

Fγ

Mcl

South


Fγ ∝ fCR Mcl

d2


North

South

Fγ

Mcl

North

≈

Fγ

Mcl

South


Fγ ∝ fCR Mcl

d2∝ ηCR ESN

(D tage)3/2

Mcl

d2


North

South

Fγ

Mcl

North

≈

Fγ

Mcl

South


Fγ ∝ fCR Mcl

d2∝ ηCR ESN

(D tage)3/2

Mcl

d2


D ∝ηCRESN Mcl

Fγd2

2/3

t−1ageand we get... ->

W28: numbers

Measured quantities:

apparent size: ~ 45’-50’, distance ~ 2 kpc -> Rs ~ 12 pc ratio OIII/Hβ -> vs = 80 km/s

Assumptions:

mass of ejecta -> Mej ~ 1.4 Msun

ambient density -> n ~ 5 cm-3

Derived quantities (Cioffi et al 1989 model):

explosion energy -> ESN ~ 0.4 x 1051 erg initial velocity -> v0 ~ 5500 km/s SNR age -> 4.4 x 104 yr (radiative phase)

W28: TeV emission

North

South

A B

~ 5 104 Msun

~ 6 104 Msun

~ 4 104 Msun

Gabici et al, 2010

W28: TeV emission

North

South

A B

~ 5 104 Msun

~ 6 104 Msun

~ 4 104 Msun

North

South A

South B

virtually the same curves for:

(η,χ) = (10%, 0.028); (30%, 0.06); (50%, 0.085)

acceleration efficiency

diffusion coefficient in units of the galactic one

Gabici et al, 2010

W28: TeV emission

North

South

A B

~ 5 104 Msun

~ 6 104 Msun

~ 4 104 Msun

North

South A

South B


(η,χ) = (10%, 0.028); (30%, 0.06); (50%, 0.085)



diffusion

coe

ffici

ent is

supp

ressed

?

Gabici et al, 2010

W28: TeV emission

North

South

A B

~ 5 104 Msun

~ 6 104 Msun

~ 4 104 Msun

North

South A

South B


(η,χ) = (10%, 0.028); (30%, 0.06); (50%, 0.085)



size of “CR bubble” @3 TeV

ld ≈ 70 pc → χ = 0.028

≈ 95 pc → χ = 0.06

≈ 115 pc → χ = 0.085diffusion

coe

ffici

ent is

supp

ressed

?

Gabici et al, 2010

W28: GeV emission

Energy (GeV)110 1 10 210 310

) 1 s2

dF/d

E (e

V cm

2 E

110

1

10

210(b) W28 South (Source S)

Energy (GeV)110 1 10 210 310

) 1 s2

dF/d

E (e

V cm

2 E

110

1

10

(a) HESS J1800 240A

Energy (GeV)110 1 10 210 310

) 1 s2

dF/d

E (e

V cm

2 E110

1

10

210(a) W28 North (Source N)

FERM

I ->

Abd

o et

al,

2010

if the cloud is within the “CR bubble” [d < ld(E)]

if cloud A and B are at the same distance -> same spectrum

AB

cloud A might not be physically associated with the

W28 system

W28: GeV emission

W28: GeV emission

~ TeV

W28: GeV emission

~ TeV

~ GeV

W28: GeV emission

d = 12 pc d = 32 pc d = 65 pc

η = 30%

χ = 0.06

North South B South A

W28: GeV emission

d = 12 pc d = 32 pc d = 65 pc

η = 30%

χ = 0.06this peak is removed as background


W28: GeV emission

d = 12 pc d = 32 pc d = 65 pc

η = 30%

χ = 0.06this peak is removed as background

some problems here...+ some of the approximations are not very good at low energies

+ subtraction of the background?+ other contributions? (Bremsstrahlung)


Take-home message

Current observations of MCs in gamma rays can be used to probe the CR

intensity at different locations within the Galaxy (i.e. Gal centre ridge, orion cloud,

W28) -> indirect identification of CR sources

For MC located close to the (supposed) CR accelerator, estimates of the local CR

diffusion coefficient can be made.

Future observations (Fermi, HESS/MAGIC/VERITAS, but most of all CTA) will

increase the number of detected clouds and make the results presented here more

(or less?) solid

gamma rays from molecular clouds (part 2)cappa/work_shop/gabici2.pdf · gamma rays from the milky...

Documents