diagnostic power in the “low-resolution”cxc.harvard.edu/cdo/hrxs2015/presentations/... ·...
TRANSCRIPT
Diagnostic Power in the “Low-Resolution” Spectroscopy of Non-Equilibrium Plasmas
Hiroya Yamaguchi (NASA/GSFC, UMD)
Apologies: - I have little experience of “high-resolution spectroscopy” (i.e., grating). - I hate to show lots of “fake data” of future missions (i.e., calorimeter).
So I’m going to ..., present real data from the X-ray CCD Suzaku/XIS with low moderate energy resolution (E/△E ~ 40).
In order to convince you that: - There are a number of powerful diagnostics even in the moderate-energy spectroscopy. - Knowledge of physics behind radiation processes helps get an unexpected, exciting results.
Today’s Talk
5%
35%
60%
Atomic physics(but pretty basic)
Astrophysics
How I love Boston ... sorry, DC/MD guys!
Supernova Remnants (SNRs, especially of Type Ia): to solve progenitor’s evolution and explosion mechanism.
Scientific Motivation
Cliche: “X-ray observation of SNRs is best suitable for studying SNe.”... Is it true?
In principle, TRUE: SNRs allow us to directly measure chemical composition and distribution.In practice, NOT TRUE: Plasma is in non-equilibrium ionization (NEI), making the abundance measurement very difficult.
Thanks to the NEI, - We can understand supernova physics in detail.- We can discover new interesting phenomena.
CIE? NEI? (“Textbook”-ish explanation)
Definition:
Fez+ Fe(z+1)+ + e-
CIE: Ionization rate = Recombination rate (for all ions)NEI: Ionization rate ≠ Recombination rate (for any ion)
Convention among plasma physicists:CIE: Electron temperature = Ionization temperatureNEI: Electron temperature ≠ Ionization temperature
What’s the major difference between CIE and NEI in terms of atomic processes and resulting spectral features?
Ionization & recombination rates depend on electron temperature. ➔ Ion fraction in a CIE plasma is uniquely determined by kTe.
0.01 0.1 1
Ar-like
Fe8+
Ne-like
Fe16+
2+
4+
12+
6+
10+
14+18+
20+ 22+
10
Electron temperature (keV)
1
0.1
K-shell
Neutral Ar-like Fe8+ Ne-like Fe16+ He-like Fe24+ Bare Fe26+
105 K 107 K 108 K106 K
L-shell
M&N-shells
K-shell
L-shell
K-shell
L-shell
K-shell
M-shell
M shellPotential Energy L shell K shell
Ion
fra
ctio
n He-like
Fe24+
Fully-ionized
Fe26+
25+
CIE? NEI? (Atomic-physics point of view)
Abundant in young SNRs
Typical kTe of young SNRs
CIE? NEI? (Atomic-physics point of view)
0.01 0.1 1
Ar-like
Fe8+
Ne-like
Fe16+
2+
4+
12+
6+
10+
14+18+
20+ 22+
10
Electron temperature (keV)
1
0.1
105 K 107 K 108 K106 K
Ion
fra
ctio
n He-like
Fe24+
Fully-ionized
Fe26+
25+
Fe26+
25+12+
10+14+
18+ 20+ 22+
1012
1011
1010
109
Ionization age: net (cm-3 s)
1
0.1
10-2
Ar-like
Fe8+
Ion
fra
ctio
n
He-like
Fe24+
Ne-like
Fe16+
Ionization speed depends on the electron densityTo reach CIE, ne t ~ 1012 cm-3 s ➔ 3×104 (ne / 1 cm-3)-1 yr
Ionization & recombination rates depend on electron temperature. ➔ Ion fraction in a CIE plasma uniquely determined by kTe.
Typical kTe of young SNRs
In an NEI plasma, (i) hot free electrons with the energy higher than the K-shell potential, and (ii) low-ionized ions with a lot of L-shell and M-shell electrons co-exist.
➔ Innershell processes take place. Fluorescence emission offers key to diagnosing the plasma condition. (e.g., Kaastra+1993; Palmeri+2003)
Q_ Q`
Electron impact Innershell ionization
Fluorescence
K-shellL-shell
M-shell
Characteristic Atomic Processes in NEI
9.0 9.5 10.0 10.5 11.0 11.5 12.0 12.5 13.0
7.0
7.5
log (net [cm-3 s])
log (T e
[K])
8.5
8.0
6395 6499 6603 6707 6811 6915
6400
6500
6600
6700
0 10 15 20Charge Number
Ener
gy [e
V]
5
Fe K_ Centroid
Neutral Ar-likeNe-like
8 9 10 11log(n t [cm-3 s])e
e.g., centroid energy which depends on the charge number (hence net).
(for kTe = 5 keV)
0
0.05
0.1
0.15
Flux
Rat
io
0 10 15 20Charge Number
5
Fe K`/K_�Flux Ratio
NeutralAr-like Ne-like
(no 3p e-)
8 9 10 11log(n t [cm-3 s])e
Innershell processes offer important diagnostics for NEI plasma.E/△E > 100 is not necessary for the simple diagnostics (centroid, ratio)
0.0 0.028 0.056 0.084 0.112 0.140
9.0 9.5 10.0 10.5 11.0 11.5 12.0 12.5 13.0
7.0
7.5
log (net [cm-3 s])
log (T e
[K])
8.5
8.0
Kβ/Kα flux ratio would be even more interesting (see also Adam’s poster).
(for kTe = 5 keV) CIE
K-shellL-shell
M-shell
K-shellL-shell
Low-netregime
Low-ionization regime: - Drops with ionization. (Fe atoms lose 3p electrons.) - Doesn’t depend on temperature. (ionization dominant)
Characteristic Atomic Processes in NEI
K-shell
High-netregime
High-ionization regime: - Goes up with ionization. (Fe atoms lose 2p electrons too.) - Strongly depends on temperature. (excitation dominant)
Ar-like Ne-like He-like
Fe Kα Emission Diagnostics on SNRs
Yamaguchi+2014a, ApJL, 785, L27
Centroid energies discriminate the SNR progenitor type.
12 16 17 18 19 20 21 22 23 240
N103B
0519-69.0
0509-67.5
G337.2-0.7
W49BCas A
N132D
Sgr A East
N63A
G349.7+0.2
SN 1987AN49
G292.0+1.8
G350.1-0.3
IC 443
Kepler
TychoG344.7-0.1
G352.7-0.1
RCW86
SN 1006
3C 397
G0.61+0.01
L K (1
040 p
hoto
ns s
-1)
6400 6500 6600 6700
110
100
1000
Energy (eV)
Charge number
Galactic Ia SNRsLMC Ia SNRsGalactic CC SNRsLMC CC SNRs
G1.9+0.3
DEM L71
Fe Kα Emission Diagnostics on SNRs
Yamaguchi+2014a, ApJL, 785, L27
Centroid energies discriminate the SNR progenitor type.
Density effect: - Ionization age ∝ ne t - Luminosity ∝ ne NFe
12 16 17 18 19 20 21 22 23 240
N103B
0519-69.0
0509-67.5
G337.2-0.7
W49BCas A
N132D
Sgr A East
N63A
G349.7+0.2
SN 1987AN49
G292.0+1.8
G350.1-0.3
IC 443
Kepler
TychoG344.7-0.1
G352.7-0.1
RCW86
SN 1006
3C 397
G0.61+0.01
L K (1
040 p
hoto
ns s
-1)
6400 6500 6600 6700
110
100
1000
Energy (eV)
Charge number
Galactic Ia SNRsLMC Ia SNRsGalactic CC SNRsLMC CC SNRs
G1.9+0.3
DEM L71
Fe Kα Emission Diagnostics on SNRs
Yamaguchi+2014a, ApJL, 785, L27(Patnaude+2015; Follow-up studies for CC models)
Comparison with hydro models: - Type Ia = uniform ambient - CC = dense CSM
Centroid energies discriminate the SNR progenitor type.
Density effect: - Ionization age ∝ ne t - Luminosity ∝ ne NFe
Ionization state (Fe-K centroid) constrains SNR’s environment.
12 16 17 18 19 20 21 22 23 240
N103B
0519-69.0
0509-67.5
G337.2-0.7
W49BCas A
N132D
Sgr A East
N63A
G349.7+0.2
SN 1987AN49
G292.0+1.8
G350.1-0.3
IC 443
Kepler
TychoG344.7-0.1
G352.7-0.1
RCW86
SN 1006
3C 397
G0.61+0.01
L K (1
040 p
hoto
ns s
-1)
6400 6500 6600 6700
110
100
1000
Energy (eV)
Charge number
Galactic Ia SNRsLMC Ia SNRsGalactic CC SNRsLMC CC SNRs
G1.9+0.3
DEM L71
SNR 3C 397
Fe Kα Emission Diagnostics on SNRs
Yamaguchi+2014a, ApJL, 785, L27(Patnaude+2015; Follow-up studies for CC models)
Comparison with hydro models: - Type Ia = uniform ambient - CC = dense CSM
Centroid energies discriminate the SNR progenitor type.
Density effect: - Ionization age ∝ ne t - Luminosity ∝ ne NFe
Ionization state (Fe-K centroid) constrains SNR’s environment.
Most evolved
SNR 3C 397
Reverse Shock
Forward Shock
Explosion Center
Young SNR Evolved SNR
X-rays from entire ejecta
cool ejecta
0 0.5 1
0.0
10.1
1
Ni
Fe
Mn
Mn
Cr
Ar
SiO
Lagrangian Mass (M )
Mass F
raction
0.01
0.1
1
Energy [keV]5 6 7 8 9
Cou
nts
s-1 k
eV-1
10-4
10-3
Cr Mn
Fe
Ni
Suzaku XIS
Yamaguchi+2015, ApJL, 801, L31
The Fe Kα diagnostic suggests 3C 397 is the most evolved Type Ia SNR.
Allows us to investigate if electron captures indeed take place in the Type Ia SN core.
Electron capture reactions (p + e- ➔ n + νe) increase abundances of n-rich elements.
Detected strong emission from Mn and Ni, also due to fluorescence; New AtomDB!
SN Ia nucleosynthesis model
(radius)
Tycho’s SNRConclusions: - Identified physical condition and location of immediate postshock plasma - (Actually,) this was robust evidence for collisionless electron heating
Tycho Brahe (1546−1601)
12 16 17 18 19 20 21 22 23 240
N103B
0519-69.0
0509-67.5
G337.2-0.7
W49BCas A
N132D
Sgr A East
N63A
G349.7+0.2
SN 1987AN49
G292.0+1.8
G350.1-0.3
IC 443
Kepler
TychoG344.7-0.1
G352.7-0.1
RCW86
SN 1006
3C 397
G0.61+0.01
L K (1
040 p
hoto
ns s
-1)
6400 6500 6600 6700
110
100
1000
Energy (eV)
Charge number
Galactic Ia SNRsLMC Ia SNRsGalactic CC SNRsLMC CC SNRs
G1.9+0.3
DEM L71
Ne-likeYamaguchi+2014b, ApJ, 780, 136
An early Suzaku observation achieved the first detection of Cr and Mn from a Type Ia SNR (Tamagawa+2009). ➔ Weak-line search became popular.
1s2 2s2 2p6
3s2 3p6 3d6 4s2
Kα↑
↓Kβ
K-shellL-shell
M&N-shells
Neutral Fe
1s2 2s2 2p6
Kα↑
Ne-like Fe16+
K-shellL-shell
No bound 3p electronresponsible for Kβ
10ï�
0.01
0.1
Cou
nts
s-1 k
eV-1
Energy [keV]
Cr Mn
Fe K_
Fe K`Ni
95 6 7 8
0
0.05
0.1
0.15
Flux
Rat
io
0 10 15 20Charge Number
5
Fe K`/K_�Flux Ratio
NeutralAr-like Ne-like
(no 3p e-)
8 9 10 11log(n t [cm-3 s])e
Kα centroid = 6435 (± 1) eVKβ centroid = 7104 (± 10) eVKβ/Kα flux ratio = 5.5 (± 0.5) %
Anomalously Strong Fe Kβ Emission
Detection of Fe Kβ is more surprising.
10ï�
0.01
0.1
Cou
nts
s-1 k
eV-1
Energy [keV]
Fe K_
Fe K`
9
Red: Fe (K`/K_ ~ 0.01)Green: Fe (K`/K_ ~ 0.15)
~16+
~8+
5 6 7 8
7000
7400
7800
0
0.05
0.1
0.15
Ener
gy [e
V]Fl
ux R
atio
0 10 15 20Charge Number
5
Fe K` Centroid
Fe K`/K_�Flux Ratio
Neutral
Neutral
Ar-like
Ar-like
Ne-like
Ne-like(no 3p e-)
8 9 10 11log(n t [cm-3 s])e1s2 2s2 2p6
3s2 3p6 3d6 4s2
Kα↑
↓Kβ
K-shellL-shell
M&N-shells
Neutral Fe
1s2 2s2 2p6
Kα↑
Ne-like Fe16+
K-shellL-shell
No bound 3p electronresponsible for Kβ
Anomalously Strong Fe Kβ Emission
Detection of Fe Kβ is more surprising.
An early Suzaku observation achieved the first detection of Cr and Mn from a Type Ia SNR (Tamagawa+2009). ➔ Weak-line search became popular.
3 arcmin
Contour: Kα, Color map: Kβ
0.5
01.
52.
01.
0
Radius [arcmin]30 1 2
K_ × 0.1
K`
[10-6
pho
tons
cm
-2 s
-1 a
rcm
in-2]
Surfa
ce B
right
ness
RS radius
- Fe Kβ emission reveals the location of the reverse shock front.- Furthermore, it actually indicates efficient electron heating at the RS.
Reverse Shock
Unshocked FeLow-ionized Fe
Highly-ionized Fe
Reverse Shock
Forward Shock
Explosion Center
Low-ionized plasma at the immediate post-reverse-shock region.
Origin of the Strong Kβ Emission
Evidence for Collisionless Electron Heating
151050Mean Charge of Fe ejecta
101
0.1
Elec
tron
Tem
pera
ture
[keV
]
` = `min ~ 10-5
` = 0.003` = 0.01
` = 0.03
RS front
7.47.2 7.6 7.8 8.0SNR Radius [1018 cm]
( = me /mFe)Temperature increasesdue to Coulomb collisions
β = Te /TFe at RS front
Predicted line luminosity for each charge number
151050Charge Number
200.
010.
1
` = 0.01
Collisionless heating caseK`
K_
Line
Lum
inos
ity [1
040 p
hoto
ns s
-1] 10
1151050
Charge Number20
0.01
0.1
Line
Lum
inos
ity [1
040 p
hoto
ns s
-1] 10
1
` = `min ~ 10-5
Coulomb-only case
K`
K_No emission from low-ionized Fe
Comparison with hydrodynamical simulations indicates instantaneous electron heating to a temperature ~1000 times higher.
Strong K-shell emission from very low ionized ejecta is NOT easy to produce, because these Fe ions must coexist with free electrons energetic enough to ionize K-shell electrons.
Efficient collisionless heating at the SNR RS! See Yamaguchi+2014b, ApJ, 780, 136 for details.
Relationship between shock velocity (Vs ) and downstream temperature (Ti ):
kTi = (3/16) mi Vs2 Tycho: Vs = 5000 km/s ➔ kTe ≈ 0.03 keV
12 16 17 18 19 20 21 22 23 240
N103B
0519-69.0
0509-67.5
G337.2-0.7
W49BCas A
N132D
Sgr A East
N63A
G349.7+0.2
SN 1987AN49
G292.0+1.8
G350.1-0.3
IC 443
Kepler
TychoG344.7-0.1
G352.7-0.1
RCW86
SN 1006
3C 397
G0.61+0.01
L K (1
040 p
hoto
ns s
-1)
6400 6500 6600 6700
110
100
1000
Energy (eV)
Charge number
Galactic Ia SNRsLMC Ia SNRsGalactic CC SNRsLMC CC SNRs
G1.9+0.3
DEM L71
He-like
IC 443Conclusions: - Detected a signature of “recombining plasma” (opposite of typical NEI) - Initiated many discoveries of such SNRs
Yamaguchi+2009, ApJL, 705, L6Ohnishi+2014, ApJ, 784, 74
Energy (keV)4 6 8 10
0.1
0.01
10-3
10-4
Cou
nts
s-1 k
eV-1
Fe XXV
K-shell
Recombination into the
excited level and cascade
L-shell
M-shell
K-shell
L-shell
M-shell
K-shell RRC
(hi = IK + Ee)
Free electron (Ee)
He-like line
(hi = IK - IL)
(L-shell RRC)
Direct recombination
into the ground level ➔ Direct recombination into ground level should also be observed.
IC 443 is an evolved SNR with a low electron temperature (~ 0.6 keV)
105 K
0.01 0.1 1
Ar-like
Fe8+
Ne-like
Fe16+
2+
4+
12+
6+
10+
14+18+
20+ 22+
10
Electron temperature (keV)
1
0.1
105 K 107 K 108 K106 K
Ion
fra
ctio
n He-like
Fe24+
Fully-ionized
Fe26+
25+
Why Surprising?
0.6
Anyway, K-shell electrons cannot be excited by a sub-keV free electron.
X-ray energy: hν = IK + Ee
Spectrum ∝ exp(− )hν − IK
kTe
CIE plasma
Emission originates from H-like ions?
Energy (keV)6.56.4 6.6 6.7 6.8 6.9
zw
Energy (keV)4 6 8 10
0.1
0.01
10-3
10-4
Cou
nts
s-1 k
eV-1
Fe XXV
Fe RRC
IK = 8.83 keV
Interpretation: the plasma was much hotter in the past and cooled rapidly.kTe(past) must be > 10 keV, not ~ 2 keV.
Another Type of NEI (Recombining)Radiative recombination continuum (RRC) was detected, confirming the H-like-ion origin of the Fe XXV line. cf. CIE: Fe XXV mainly from He-like ions.
Another kind of “co-existence”: (i) Low-energy electrons that can barely excite L-shell electrons.(ii) Highly-charged ions with a K-shell vacancy.
ASTRO-H should confirm enhanced forbidden emission (high G-ratio).
Conclusions and Prospects- SNRs are indeed best suitable for studying supernova physics, owing to the NEI (innershell process, recombination, etc.).- SNRs offer a good “laboratory” for fundamental atomic physics.- There are a number of simple, but powerful diagnostics (e.g., centroid, Kβ/Kα ratio, RRC/line) where the grating/calorimeter resolution (E/△E ~ 1000) is not necessarily required.
- There MUST be still new diagnostics in future high-resolution spectral data, not only famous ones, like G-ratio, R-ratio, etc. - Don’t rely on (only) performance of hardware and software. - Understand the physics behind radiation processes. - “Press-‘fit’ analysis” might overlook a key spectral feature.
- Ambitious for ground experiments of innershell processes.