Constraining progenitors of SNRs using X-ray

morphology and spectra

Hiroya YamaguchiHarvard-Smithsonian Center for

Astrophysics

Classification of SN Progenitors

Optical obs of SNeClassification is relatively straightforward

- Spectrum (historically well established) - Luminosity (56Ni yield)

X-ray obs of SNRsClassification (Ia/CC) is (was) controversial in many SNRs

- Similar X-ray luminosity

- Morphology? SNRs can be spatially resolved, strong advantage of X-ray

- Spectrum?

SNe Ia: nuclear reaction energy ~ 1051 erg

SNe CC: gravitational energy ~ 1053 erg 99% neutrino + 1% kinetic (~ 1051 erg)

Ia (SD)

Ia (DD) CC (1987A)

=> transformed to thermal energy (X-ray luminosity)

Type Ia

CC

Elli

pti

cit

y

Mirror asymmetricity

Morphology of SNRsCC SNRs are more asymmetric than Ia SNRs (Lopez+09;11)

G344.7-0.1 found to be Type Ia (HY+12)

Chandra images of Galactic/Magellanic SNRs

Reflects nature of explosion and/or environment?

SNR E0102-72 (CC) 0104-72.3 (Ia candidate)

Doesn’t work for SMC SNRs… (Lopez+12)

0104-72.3

E0102-72

X-Ray Spectra of SNRsAdvantage- Optically thin (self absorption is almost negligible, but see Miyata+08)- K-shell emission from He- & H-like atoms (kTe ~ hn ~ 0.1–10 keV, comparable to K-shell potential), so physics is simple

Si

SArCa Fe

Ni

Mg

Ne

Artificial features(a sort of bgd)

Simple Quiz

CC (W49B)Ia (SN1006)

YOU LOSE

m9(^Д^)Suzaku spectrum of Tycho (Hayato+10)

X-Ray Spectra of SNRs

Si

SArCa Fe

Ni

Mg

Ne

Artificial features(a sort of bgd)

W49B (CC)

Large foreground extinction makes O/Ne/Mg emission in W49B weak

Absorption for different column density (NH [cm-

2])

SN1006

W49B

Note: although we use NH to describe the column, what we measure in X-rays is the column of metals

Yet, weakness of Fe emission in SN 1006 (Ia SNR) is puzzling

=> Understanding of NEI is essential

Non Equilibrium in Ionization (NEI)Pre-shocked metals in ISM/ejecta are almost neutral (unionized)

Shock-heated electrons gradually ionize atoms by collision, but ionization proceeds very slowly compared to heating

Fe ion population in NEI plasma for kTe = 5 keV

net (cm-3 s)

Ion f

ract

ion Fe24+

Fe25+

Fe26+

Fe16+lowly ionized

highly ionized

Fe24+

Fe25

+

Fe26+Fe16

+

Electron temperature kTe (keV)

CIE

net : “ionization age” ne : electron density

t : elapsed time since gas was heated

Non Equilibrium in Ionization (NEI)Fe ion population in NEI plasma for kTe = 5 keV

net (cm-3 s)

Ion f

ract

ion Fe24+

Fe25+

Fe26+

Fe16+lowly ionized

highly ionized

net : “ionization age” ne : electron density

t : elapsed time since gas was heated

Timescale to reach CIE for ISM t ~ 3 x 104 (ne/1 cm-3)-1 yr

As for ejecta…Time when the masses of swept-up ISM and ejecta becomes comparable

Ionization state for the ejecta becomes almost “frozen” after an SNR evolved.

Ionization age for the ejecta strongly depends on the initial CSM density rather than its age.

Non Equilibrium in Ionization (NEI)

Full X-ray band

Magnified spectra in the 6-7 keV band (Fe K emission)

Fe-L blend

Fe-K

Observed spectrum (Convolved by Suzaku response)

net = 5x109 1x1010 5x1010 1x1011 3x1011

Ar-like Ne-like C-like Be-likeHe-like

Model spectra of Fe emission [kTe = 5 keV]

How does ionization age affect a spectrum? How can we measure ionization age?

H-like

6.42 keV 6.44 keV 6.60 keV 6.64 keV 6.67 keV

6.0 7.0

0.5 10

Si

SArCa Fe

Ni

Mg

Ne

Artificial features(a sort of bgd)

Ozawa+2009HY+2008, Uchida+, in prep.

SN1006 (Type Ia SNR) W49B (CC SNR)

SN1006: Searching for Fe emission

Fe?

BeppoSAX MECS

spectrum

Chandra image

- Prototypical Type Ia SNR, but emission from Fe has never been detected.

- Only one possible detection reported by BeppoSAX

- XMM-Newton failed to detect

Vink+00

Suzaku spectrum(HY+08)

Detected! but weak despite of its Type Ia origin

Fe-K centroid ~ 6420eV (< Ne-like) … Corresponding net is ~ 1 x 109 cm-3 s

Fe24+

Fe25+

Fe26+

Fe16+

SN1006: Multiple net Components in Si

Approx with 2-net componentsfor Si and S ejecta

net1 ～ 1×1010 cm-3 snet2 ～ 1×109 cm-3 s

cf. Fe: net ～ 1×109 cm-3 sSi ion fraction @1keV

Si12+

Si13+

Si6+ Si8+

C~O-like He-like

MgSi

broad feature

S

Reverse shock heats from outer region

Outer ejecta = highly ionizedInner ejecta = lowly ionized

SN1006: Fullband Spectrum & Abundances

ISM (w/ solar abundance)Outer ejecta (net ~ 1010 cm-3 s)Inner ejecta (net ~ 109)Non-thermal (synchrotron)

Fe

Derived abundance ratios compared to the W7 model of Nomoto+84

Outer ejecta

Inner ejecta

Suggests stratified composition with Fe toward the SNR center, which results in the lowly-ionized (thus weak) Fe emission

HY+08

Ejecta Stratification in Type Ia SN/SNRsXMM image of Tycho

Radius (arcmin)

Radial profile

Fe

Si

Color: Si-KContour: Fe-K

Decourchelle+01

Mazzali+07IME

56NiEncl

ose

d m

ass

SN 2003du(Tanaka+10)

See also Badenes+06

Si

SArCa Fe

Ni

Mg

Ne

Artificial features(a sort of bgd)

Ozawa+2009HY+2008, Uchida+, in prep.

SN1006 (Type Ia SNR) W49B (CC SNR)

W49B: Peculiar Ionization State

- RRC can be enhanced only when the plasma is recombining (e.g., photo-ionized plasma)

Similar recombining SNRs - IC443 (HY+09) - SNR 0506-68 (Broersen+11) - other 3 & a few candidates

“Recombining NEI” in SNRs is not unique => Need to define “recombination age”

Cr Mn

He-like Fe Ka

Ni + Fe Kb

Fe-K RRC

H-like Fe

Ozawa+09

Ejecta is highly ionized to be He-like

Radiative recombination continuumFe25+ + e- → Fe24+ + hn

… indicates presence of a large fraction

of H-like FeMeasured kTe ~ 1.5 keV

Temperature (keV)

Fe24+

Fe25+

Fe26+Fe16+

Fe ion population in a CIE plasma

W49B: Possible Progenitor

blast wave

2nd reverse shock

reverse shockBlast wave breakout into ISM

BW speed becomes faster and expand adiabatically, resulting in rapid cooling with “frozen” ionization state

Shimizu+12Explosion in dense CSM

- Numerical (Shimizu+12)- Analytical, more progenitor- oriented (Moriya 12)

RSG case (vw ~ 10 km/s) WR case (vw ~ 1000 km/s)

Type II-P or IIn could be a progenitor of a recombining SNR (Moriya 12)

Fe-K diagnosticsExtreme cases have been shown

SN1006: Type Ia SNR, Fe lowly-ionized due to a low ambient density and ejecta stratification with Fe more concentrated toward the center

W49B: CC SNR, Fe over-ionized (recombining), possibly due to interaction with high-density CSM … and inhomogeneous ejecta structure?

Red: SiBlue: FeGreen: continuum

Other SNRs?

Fe-K diagnostics

Type Ia

CC

- Type Ia and CC SNRs are clearly separated (CC more ionized)

- Luminosity of both groups are distributed in the similar range.

(HY+, in prep.)

net = 5x109 1x1010 5x1010 1x1011 3x1011

Can be explained by ionization (and temperture, density effects) --- Measuring ionization state is essential for measuring element abundances!!

Fe-K diagnostics

Type Ia

CC

(HY+, in prep.)

Ionization ages expected if the SNRs have evolved in uniform ISM with typical density

Hachisu+01

Badenes+07

If the SD scenario is the case, a large, low-density cavity is expected around the progenitor

No evidence of an “accretion wind” and a resultant cavity but for a few Type Ia SNRs

Evidence of cavity/CSM in Ia SNRs

RCW86 (Williams+11)

Unique Ia SNR where the presence of a surrounding cavity is suggested

Kepler (Reynolds+07)

N103B (Lewis+03)

Summary- X-ray observation of SNRs is one of the best methods to study stellar/explosive nucleosynthesis. (optically-thin, K-shell emission)

- Understanding of non-equilibrium in ionization is, however, essential for accurate measurement of element abundances.

- Fe emission in Type Ia SNRs is commonly weak due to low-density ambient and stratified chemical composition.

- In CC SNRs, on the other hand, Fe is highly ionized, sometime overionized, possibly due to initial CSM interaction.

- No evidence of a large cavity expected from an “accretion wind” around Type Ia SNRs, except for RCW86, constraining progenitor system??