Localized Recombining Plasma in G166.0+4.3: A Supernova Remnant with an Unusual Morphology

0.95 MB·

by  Hideaki Matsumura

Publ.Astron.Soc.Japan(2014)00(0),1–8 1
doi:10.1093/pasj/xxx000
Localized Recombining Plasma in G166.0+4.3:
A Supernova Remnant with an Unusual
Morphology
7
1
0 Hideaki MATSUMURA1, Hiroyuki UCHIDA 1, Takaaki TANAKA 1, Takeshi Go
2 TSURU 1, Masayoshi NOBUKAWA 2, Kumiko Kawabata NOBUKAWA 3 and

n Makoto ITOU 1
a
J 1DepartmentofPhysics,FacultyofScience,KyotoUniversity,KitashirakawaOiwake-cho,

6
Sakyo-ku,Kyoto-shi,Kyoto606-8502,Japan
1
2Science,CurriculumandInstruction,FacultyofEducation,NaraUniversityofEducation,

] Takabatake-cho,Nara-shi,Nara630-8528,Japan
E
3DepartmentofPhysics,NaraWomen’sUniversity,Kitauoyanishimachi,Nara630-8506,
H
Japan
.
h
p Received;Accepted
-
o
r Abstract
t
s
We observed the Galactic mixed-morphology supernova remnant G166.0+4.3 with Suzaku.
a
[ The X-ray spectrum in the western part of the remnant is well represented by a one-

component ionizing plasma model. The spectrum in the northeastern region can be ex-
1
v plained by two components. One is the Fe-rich component with the electron temperature
64 kTe = 0.87+−00..0023 keV. The other is the recombining plasma component of lighter elements
1 withkTe=0.46±0.03keV,theinitial temperaturekTinit=3keV(fixed)andtheionization pa-
4 rameter net=(6.1+−00..54)×1011 cm−3 s. As the formation process of the recombining plasma,
0 two scenarios, the rarefaction and thermal conduction, areconsidered. The former would not
.
1 be favored since we found the recombining plasma only in the northeastern region whereas
0
thelatterwouldexplaintheoriginoftheRP.Inthelatterscenario,anRPisanticipatedinapart
7
1 of theremnant where blast wavesare in contactwith cooldense gas. Theemission measure
: suggestshigher ambient gas densityin thenortheasternregion. Themorphology oftheradio
v
i shellandaGeVgamma-rayemissionalsosuggestamolecularcloudintheregion.
X
Keywords:X-rays: individual,G166.0+4.3,Recombiningplasma
r
a
1 Introduction theX-raymorphology. MostMMSNRsareactuallyassociated
withmolecularcloudsasindicatedbyCOlineemissionsorOH
Mixed-morphology supernova remnants (MM SNRs) have
(1720MHz)masers(e.g.,Lazendic&Slane2006). Theyalso
center-filled thermal X-ray emissions in a synchrotron radio
exhibit GeV/TeV gamma-ray emissions which suggest shell-
shell (Rho & Petre 1998). Although the formation mecha-
cloudinteractionssincethoseSNRscanemitluminousgamma
nism of the non-standard morphology has not been fully un- raysthroughπ0 decays(e.g.,Ackermannetal.2013;Albertet
derstood, there are two proposed scenarios. One is the evap-
al.2007).
orating cloudlet scenario (White & Long 1991), and the other
is the thermal conduction scenario (Cox et al. 1999; Shelton IntheASCAobservationoftheMMSNRIC443,Kawasaki
et al. 1999). Ineither scenario, interactions between the blast et al. (2002) measured ionization temperatures (T ) based on
z
waveanddensegasplayanimportantroleintheformationof theH-liketoHe-likeKαintensityratiosofSiandS,andcom-
(cid:13)c 2014.AstronomicalSocietyofJapan.
2 PublicationsoftheAstronomicalSocietyofJapan,(2014),Vol.00,No.0
paredthemwithelectrontemperatures(T ).TheyfoundthatT 12keV.Weassume5kpcasthedistancetotheremnantinthis
e z
is significantly higher than T , and suggested that the plasma paper.
e
is overionized. In the observations of IC 443 (Yamaguchi et
al. 2009) and W49B (Ozawa et al. 2009) with Suzaku, they
2 Observationsand data reduction
discovered radiative recombining continua (RRC), which pro-
vide clear evidence that the plasmas are in an extremely re- We performed Suzaku observations of G166.0+4.3. Table 1
combiningstate. SubsequentSuzakuobservations revealedre- summarizes the observation log. In this analysis, we used
combining plasmas (RPs) in other MM SNRs in the Galaxy data from the X-ray Imaging Spectrometer (XIS: Koyama et
(e.g., G359.1−0.5: Ohnishi et al. 2011; W28: Sawada & al. 2007) installed on the focal planes of the X-ray telescopes
Koyama2012;W44:Uchidaetal.2012)aswellasintheLarge (XRT: Serlemitsos et al. 2007). XIS2 and a part of XIS0 are
MagellanicCloud(N49:Uchida,Koyama,Yamaguchi2015). notusedinwhatfollows. Theformerhasnotbeenfunctioning
Twoscenariosaremainlyconsideredfortheformationpro- since2006 November. Thelatter was turned offin 2009 June
cessoftheRPs. Oneistherarefactionscenario(Itoh&Masai (SuzakuXISdocuments1,2).
1989). If a supernova explodes in a dense circumstellar mat- We reduced the data using the HEADAS software version
ter(CSM)aroundthemassiveprogenitor, ejectaandtheCSM 6.19.Weusedthecalibrationdatabasereleasedin2015October
areshock-heatedandquicklyionizedbecauseofthehighden- fordataprocessing. Thetotalexposuretimeofthefiveobser-
sity. When the shock breaks the CSM out to a lower density vations is ∼226 ks after the standard data screening. In the
interstellar medium (ISM),T is decreased by adiabatic cool- processing ofthe XISdata, weremovedcumulative flickering
e
ing. Kawasaki et al. (2002) proposed the other scenario, the pixelsbyreferringtothenoisypixelmaps3providedbytheXIS
thermal conduction scenario. If an SNR shock is interacting team. Wealsodiscardedpixelsadjacenttotheflickeringpixels
withamolecularcloud,thermalconductionoccursbetweenthe (pixelqualityB14=1). NonX-raybackgrounds(NXBs)were
SNRplasmaandthecloud,and,therefore,T drops. Sincethe estimatedbyxisnxbgen(Tawaetal.2008). Theredistribution
e
recombination timescalesofionsaregenerallylongerthanthe matrix files and the ancillary response files were produced by
conductiontimescales,T cannotfollowthedecreaseofT ,and xisrmfgen and xissimarfgen (Ishisaki et al. 2007), respec-
z e
RPscanberealized. tively.
G166.0+4.3isaGalacticSNRwhoseradioshellhasalarge
bipolar structure in the west with a smaller semicircle shell 3 Analysis
in the east (Sharpless 1959; Landecker et al. 1982; Pineault
3.1 Image
et al. 1987). Landecker et al. (1982) estimated the distance
to G166.0+4.3 as 5.0±0.5 kpc by using the Σ-D-z (surface Figure 1 shows a mosaic image of G166.0+4.3 in the energy
brightness-diameter-distanceabovetheGalacticplane)relation- band of 0.5–2.0 keV. We subtracted the NXB from the image
ship. Pineaultetal.(1987)suggestedthatthegasdensityinthe andcorrecteditforthevignetting effectoftheXRT.Contours
eastishigherthanthatinthewestbecauseoftheunusualmor- area325MHzradioimagefromtheWesterbordNorthernSky
phology of the radio shell. G166.0+4.3 was observed in the Survey(WENSS).
X-raybandwithROSAT(Burrows&Guo1994),ASCA(Guo
& Burrows 1997) and XMM-Newton (Bocchino et al. 2009).
3.2 Backgroundsubtraction
Burrows&Guo(1994)foundtheX-raymorphologyisdistinct
Forbackgroundestimation,weusedSuzakuobservationaldata
fromtheradio morphology and reproduced the spectrum with
ofIRAS05262+4432,whichislocatedat(l,b)=(165.◦1,5.◦7).
collisionalionizationequilibrium(CIE:T =T )models.They
z e
estimated the age of the remnant to be ∼2.4×104 yr by ap- TheobservationlogisshowninTable1. Weextractedaspec-
trumfromasource-freeregion.Inordertomodelthespectrum,
plying the evaporating cloudlet model (White & Long 1991)
weadoptedamodelbyMasuietal.(2009). Theyobservedthe
whentheyassumetheestimateddistanceof5kpc.Bocchinoet
direction (l,b) = (230◦, 0◦) with Suzaku in order to study the
al.(2009) classified G166.0+4.3 asanMMSNRand reported
softX-rayemissionfromtheGalactic disk. Their modelcon-
that the spectra can be reproduced by an ionizing plasma (IP:
sistsoffourcomponents: thecosmicX-raybackground(CXB),
T <T )model.
z e
the local hot bubble (LHB), and two thermal components for
GasenvironmentofG166.0+4.3isclearlydifferentbetween
theGalactichalo(GHcoldandGHhot).
theeastandthewest,accordingtothemorphologyoftheradio
shell. It is one of the best sources to study plasma evolution 1http://www.astro.isas.ac.jp/suzaku/doc/suzakumemo/suzakumemo-2007-
dependingontheenvironment. Inthispaper,wereportonob- 08.pdf
2http://www.astro.isas.ac.jp/suzaku/doc/suzakumemo/suzakumemo-2010-
servationalresultsofG166.0+4.3withSuzaku,whichhashigh
01.pdf
sensitivityandhighenergyresolutionintheenergybandof0.3– 3https://heasarc.gsfc.nasa.gov/docs/suzaku/analysis/xisnxbnew.html
PublicationsoftheAstronomicalSocietyofJapan,(2014),Vol.00,No.0 3
Table1.Observationlog.
Target Obs.ID Obs.date (R.A.,Dec.) EffectiveExposure
G166.0+4.3NE 509022010 2014-Sep-19 (5h27m07.s7,42◦58′52.′′2) 61ks
G166.0+4.3NE 509022020 2014-Sep-22 (5h27m07.s7,42◦58′52.′′2) 61ks
G166.0+4.3NW 509023010 2014-Sep-20 (5h25m46.s8,42◦54′01.′′7) 42ks
G166.0+4.3SE 509024010 2014-Sep-21 (5h26m41.s2,42◦39′06.′′2) 34ks
G166.0+4.3SE 509024020 2015-Mar-13 (5h26m41.s2,42◦39′06.′′2) 27ks
IRAS05262+4432(Background) 703019010 2008-Sep-14 (5h29m56.s0,44◦34′39.′′2) 82ks
Right ascension
5:28:00.0 5:26:00.0 5:24:00.0
0
00. 0.1
0:
3
3:
4
V−1 0.01
n NE Counts s ke−1 10−3 GHcold CXB GHhot
o 0 LHB
i 0.
eclinat 43:00:0 10−244
D W χ 0
−2
−4
0.5 1 2 5 10
0 Energy (keV)
0.
0
30: Fig.2.XIS0+3(black) and XIS1(red) spectra takenfrom the source-free
2: regionoftheIRAS05262+4432datawiththebest-fitmodel.Theblue,green,
4
magenta andorange linesrepresent theCXB,LHB,GHcold and GHhot
models,respectively. Thebottompanelshowsthedataresidualsfromthe
best-fitmodel.
9.8e-4 3.9e-3 1.6e-2 6.3e-2 2.5e-1 1.0
Fig.1.XISimageofG166.0+4.3inthe0.5–2.0keVbandaftertheNXBsub-
tractionandthecorrectionofthevignettingeffect. Thecoordinatesreferto
epochJ2000.0. TheX-raycount rateisnormalized sothatthepeakbe-
comesunity. Weoverlaidcontours ofthe325MHzradioimagefromthe
WENSS.
Table2.Best-fitmodelparametersofthe
backgroundspectra
We fit the background spectrum with the model after the Component Parameter(unit) Value
NXB subtraction using XSPEC version 12.9.0. We fixed the CXB NH(1021cm−2) 3.8(fixed)
Photonindex 1.4(fixed)
electron temperatures of the LHB, GHcold and GHhot at the
Normalization∗ 10.7±0.4
values byMasuietal.(2009). Theabsorption columndensity
(NH) of the CXB was fixed at 3.8×1021 cm−2, the Galactic LHB kNToerm(kaleiVza)tion† 103.1.405±(fi3x.2ed)
valueinthelineofsighttowardIRAS05262+4432 (Dickey& GHcold kTe(keV) 0.658(fixed)
Lockman 1990; Kalberla et al. 2005). The normalizations of Normalization† 2.1±0.2
allthecomponentswereallowedtovary. Figure2andTable2 GHhot kTe(keV) 1.50(fixed)
showthefittingresultandthebest-fitparameters,respectively. Normalization† 3.1±0.5
χ2(d.o.f.) 301.9(234)
∗Theunitisphotonss−1cm−2keV−1sr−1@1keV.
3.3 SNRspectra †Theemissionmeasureintegratedoverthelineofsight,i.e.,
(1/4π)RnenHdlinunitsof1014cm−5sr−1.
Considering the morphology of the radio shell, we extracted
spectrafromthenortheastandwestregions(namedNEandW)
indicatedwiththewhitesolidlinesinFigure1.Figure3shows
4 PublicationsoftheAstronomicalSocietyofJapan,(2014),Vol.00,No.0
(Figure4(ne-i)). Thesespectralfeatures,ifconfirmed,provide
evidencethattheplasmaisinarecombination-dominant state.
Si
1 Wetriedaone-component RPmodel. However, asatisfactory
S
fitwasnotobtained.
Ne
Counts s keV−1−1 0.010.1 Fe NLi LMg >cfioxme1Fd.p6otoroknateehdVneettbwIaPeihlseoet-drrfieRitnwvPvaeemlsufteooigdufaeontlrdisoW.nthT,,eNhweleHaarr=gbeess0otrr.r8iepcs×ttiiedo1dun0at2hcl1soe,lceuamnmne−dnrg2td,yreiwebndhaseniotrdyenaetioss-
kT andn tareleftfree.ThekT isafreeparameterforthe
10−3 RPemodel,ewhereas it is fixed atin0it.01 keV for the IP model.
Theabundances of Siand S arefreeparameters, and those of
10−40.5 1 2 5 10 Ar and Ca are linked to S. The residuals were not improved
Energy (keV) withtheIPmodel(Figure4(ne-ii)andTable3). Ontheother
Fig.3.XIS0+3spectraofW(black)andNE(blue). Theverticalsolidand hand, we successfully fitted the spectrum with the RP model
dashedblacklinesshowthecenterenergiesoftheKlinesfromHe-likeand (Figure4(ne-iii)andTable3).
H-likeions,respectively. ThearrowsindicatetheenergybandswithFe-L
andNi-Llinecomplexes. We extrapolated the above RP model down to 0.5 keV.
Figure 4 (ne-iv) shows the model and the full-band spectrum
in the energy band of 0.5–10.0 keV. The spectral structure in
XIS0+3spectraofWandNEaftertheNXBsubtraction. One
the 0.7–1.1 keV band cannot be reproduced. With the elec-
canclearlyseeFe-LandNi-Llinecomplexes,andSi-KandS-
Klines.TheFe-LandNi-Llinecomplexesaremoreprominent tron temperature kTe =0.37 keV, the model predicts a peak
inWthanthoseinNE. energyoftheFeandNiLlinecomplexesat∼0.8keV,which
is significantly lower than the observed value of ∼0.95 keV.
Wefittedthespectrawiththemodelconsisting oftheSNR
Figure 5 shows the NE spectrum of XIS0+3 and the plasma
componentandthebackground(§3.2). Theoverallnormaliza-
model in the energy band of 0.5–1.2 keV. The model consists
tionofthebackgroundcomponentwastreatedasafreeparame-
ter.TheabsorptioncolumndensityoftheCXBcomponentwas onlyoftheFeandNicomponentswithNH=1.7×1021cm−2,
fixedat3.6×1021 cm−2 whichistheGalacticvalueintheline kTinit = 3.0 keV and net = 6.0×1011 cm−3s. The elec-
tron temperatures are 0.40 keV and 0.85 keV in Figures 5
ofsighttowardG166.0+4.3.
(a) and (b), respectively. The dominant emission lines of the
We applied a one-component IP model to the W spectrum
Fe and Ni components are FeXVII 2p5 3s1→2p6 (0.725 keV
usingtheVVRNEImodelintheXSPECpackage,wheretheini-
and 0.739 keV) and FeXVII 2p5 3d1→2p6 (0.812 keV and
tialtemperature(kT )isfixedat0.01keVwhereastheelec-
init
trontemperature,ionizationparameter(n t)andnormalization 0.826 keV) with kTe = 0.40 keV, and FeXIX 2p3 3s1→2p4
e
(0.822 keV), FeXVIII 2p4 3d1→2p5 (0.873 keV), FeXX 2p2
arefreeparameters. WeusedtheWisconsin absorptionmodel
3d1→2p3 (0.965 keV) and FeXXI 2p1 3d1→2p2 (1.009 keV)
(Morrison&McCammon1983) fortheabsorption whosecol-
with kT = 0.85 keV. The comparison between the data and
umndensityisafreeparameter.TheabundancesofO,Ne,Mg, e
themodelsindicatesthattheelectrontemperatureoftheFeand
Si, S, and Fe are also free parameters. The Ar and Ca abun-
Niplasmashouldbe∼0.85keVinordertoreproducetheob-
dancesarelinkedtoS,andNiislinkedtoFe. Theabundances
served structure. On the other hand, the Si and S plasma de-
of the other elements are fixed to the solar values (Anders &
mandslowerkT aswesawinthe(ne-iii)fit.
Grevesse1989). TheenergybandaroundtheneutralSiK-edge e
(1.73–1.78 keV) is ignored because of the known calibration The result indicates that the plasma has different kT be-
e
uncertainty. The fitting left large residuals around 0.82 keV tween Fe (+Ni) and the other elements. Kamitsukasa et al.
and1.23 keV duetothe lackofFe-Llines inthemodel (e.g., (2016)decomposedtheejectaofSNRG272.2−3.2intocompo-
Yamaguchietal.2011). Therefore,weaddedtwonarrowlines nentswithdifferentelements,consideringthateachelementhas
at0.82keVand1.23keV.TheresultsareshowninFigure4(w- differentradialdistributions. Followingtheiridea,weadopted
i)andTable3. Wefoundthatthespectrumiswellreproduced atwo-componentVVRNEImodel: oneforFeandNi,andthe
byaone-componentIPmodelwiththeelectrontemperatureof other forlighter elements. Thespectracanbewell fittedwith
0.83keV. the model comprising a CIE for Fe and Ni and an RP for the
We applied the same IP model as W to the NE spectrum. other elements. Figure 4 (ne-v) shows the spectrum with the
Significant residuals are found at ∼2.0 keV and ∼ 2.6 keV best-fitmodel. Although wetriedIPandRPfortheFeandNi
which correspond to Si Lyα (2.0 keV) and the edge of the component,theionizationparametern tcannotbeconstrained
e
RRCofHe-likeSi(2.67keV)+SLyα(2.63keV),respectively well.Therefore,wehereassumedtheCIEmodelforsimplicity.
PublicationsoftheAstronomicalSocietyofJapan,(2014),Vol.00,No.0 5
(w-i) (ne-i)
1 IP + 2gauss model 1 IP + 2gauss model
(0.5-10.0 keV) (0.5-10.0 keV)
V−1 0.1 V−1 0.1
Counts s ke−1 01.00−13 Counts s ke−1 01.00−13
10−4 10−4
4 4
2 2
χ 0 χ 0
−2 −2
−4 −4
0.5 1 2 5 10 0.5 1 2 5 10
Energy (keV) Energy (keV)
(ne-ii) (ne-iii)
1 IP model (1.6-10 keV) 1 RP model (1.6-10 keV)
V−1 0.1 V−1 0.1
Counts s ke−1 01.00−13 Counts s ke−1 01.00−13
10−4 10−4
4 4
2 2
χ 0 χ 0
−2 −2
−4 −4
0.5 1 1.6 2 5 10 0.5 1 1.6 2 5 10
Energy (keV) Energy (keV)
(ne-iv) (ne-v)
1 RP + 2gauss model 1 CIE (only Fe & Ni) + RP (other elements) + 2gauss model
(0.5-10.0 keV) (0.5-10 keV)
Counts s keV−1−1 01.000.−113 Counts s keV−1−1 01.000.−113
10−4 10−4
4 4
2 2
χ 0 χ 0
−2 −2
−4 −4
0.5 1 2 5 10 0.5 1 2 5 10
Energy (keV) Energy (keV)
Fig.4.XIS0+3(black)andXIS1(red)spectraofWorNE.Eachspectrumisfittedwiththeplasmamodel(solidlines)andthebackgroundmodel(dottedlines).
6 PublicationsoftheAstronomicalSocietyofJapan,(2014),Vol.00,No.0
Table3.Fitparametersofsourceregions.
Modelfunction Parameter(unit) (w-i) (ne-i) (ne-ii) (ne-iii) (ne-v)
Absorption NH(1021cm−2) 0.8±0.1 0.3±0.1 0.8(fixed) 0.8(fixed) 1.7±0.1
VVRNEI1 kTe(keV) 0.83±0.01 0.90+−00..0012 1.36+−00..5217 0.37±0.04 0.46±0.03
kTinit(keV) 0.01(fixed) 0.01(fixed) 0.01(fixed) ≥2.98 3.0(fixed)
H(solar) 1(fixed) 1(fixed) 1(fixed) 1(fixed) 1(fixed)
He(solar) 1(fixed) 1(fixed) 1(fixed) 1(fixed) 1(fixed)
C(solar) 1(fixed) 1(fixed) 1(fixed) 1(fixed) 1(fixed)
O(solar) 0.4+0.2 0.2+0.2 1(fixed) 1(fixed) 0.02+0.03
−0.1 −0.1 −0.01
Ne(solar) 0.6+0.2 0.6+0.2 1(fixed) 1(fixed) 0.4±0.1
−0.1 −0.1
Mg(solar) 1.2±0.2 0.6±0.1 1(fixed) 1(fixed) 0.6±0.1
Si(solar) 0.7+0.1 0.6±0.1 1.4+0.3 2.3±0.3 1.0±0.1
−0.2 −0.1
S=Ar=Ca(solar) 1.2±0.3 1.2±0.2 1.8±0.4 3.3+0.7 1.5±0.2
−0.6
Fe=Ni(solar) 1.2±0.2 0.4±0.1 1(fixed) 1(fixed) ————-
net(1011cm−3s) 4.0+−00..75 3.1±0.5 1.6+−10..77 6.2±0.6 6.1+−00..54
VEM(1057cm−3)† 2.7±0.2 2.9+0.2 1.2+0.3 6.6+1.6 10.3+0.8
−0.3 −0.2 −1.3 −1.0
VVRNEI2 kTe(keV) ————- ————- ————- ————- 0.87+−00..0023
Fe=Ni(solar) ————- ————- ————- ————- 0.14±0.01
VEM(1057cm−3)† ————- ————- ————- ————- =VVRNEI1
Gaussian Centroid(keV) 0.82(fixed) 0.82(fixed) ————- ————- 0.82(fixed)
Normalization‡ 4.6+1.3 2.1+0.8 ————- ————- 6.7+1.5
−0.7 −0.7 −1.3
Gaussian Centroid(keV) 1.23(fixed) 1.23(fixed) ————- ————- 1.23(fixed)
Normalization‡ 3.4+1.3 0.8±0.1 ————- ————- 0.7±0.2
−0.2
χ2(d.o.f.) 339.2(247) 443.9(251) 232.7(150) 172.1(150) 347.6(250)
†VolumeemissionmeasureVEM=RnenHdV ,wherene,nH,andV aretheelectronandhydrogendensities,andtheemittingvolume,
respectively.
‡Theunitisphotonss−1cm−2sr−1.
4 Discussion
(a)
2.0 4.1 Spatialstructure
Counts s keV−1−1 101...055 kTe = 0.40 keV Tcohmepsopneecntrtasl;athneallyoswistsehmopwesratthuartetRhePNcoEmpploasnmenatcaonndtathineshtiwgoh
0 temperature Fe-rich component. This suggests that the distri-
-2 Photons cms keV−1−1 0000....01234 bIi1nnu.3ttoi0horendkseeSrVoi-t)fKo.tWhibneeavnetdwsditvoi(gi1dca.eot7edm5–ttphh2oee.n1mde0nisbtktsyreiVabtrhu)eetiaodcnniodf,frerwterheseepnotmFnfeadr-doiLnemgbXeauIannScddhei(mr0olya.t7hgin0eerg–s.
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
(b) Energy (keV) continuum image for which we selected the line-free band of
2.0 1.50–1.70 keV. We show the images in Figure 6, which indi-
Counts s keV−1−1 101...055 kTe = 0.85 keV cisatmesortehactoSmipdaicstt.ribInutseosmueniSfoNrRmsl,yrweghaerrdelaesssthoefFtheedtyisptreibouftitohne
explosion,distributionsofFeareobservedtobecentered(e.g.,
0
V−1 Uchidaetal.2009;Hayatoetal.2010). Theobserveddistribu-
-2 ms ke−1 0.1 tions areconsistent with the widely believed picture of super-
Photons c0.005 ncloovsaerntuocltehoescyennthteersiosftthhaetphreoagveienriteolre.mentstendtobeproduced
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
Energy (keV) For a detailed analysis of the Fe distribution, we extracted
spectra from the three regions (i)–(iii) in NE indicated by the
Fig.5.(a) NE spectrum of XIS0+3 (crosses) and the plasma model with
(top)andwithout(bottom)responsematrixesmultiplied(solidlines)inthe whitelinesinthebottompanelofFigure6.Wefittedthespectra
energybandof0.5–1.2keV.TheplasmamodelconsistsonlyofFeandNi of each region with the same model as Figure 4 (ne-v). We
components withNH =1.7×1021 cm−2, kTe =0.40keV, kTinit = fixedkT oftheRPcomponent tothebest-fitvalue, 0.46keV.
3.0keVandnet=6.0×1011cm−3s. (b)Sameas(a)butwithkTe= e
0.85keV. Figure7showstheresultantabundanceratiosofFetoSi. The
ratiointheinnerregionissignificantlyhigherandsuggeststhe
PublicationsoftheAstronomicalSocietyofJapan,(2014),Vol.00,No.0 7
Region
(i) (ii) (iii)
Si ) 0.25
Right ascension Si /
5:28:00.0 5:26:00.0 5:24:00.0 Fe )(/ 0.2
e /
0.0 (F
0:0 0.15
3
3:
4
0.1
NE
n 0.05
o 0 0 6 12 18
ati 00. Angular distance from the NE rim (arcmin)
n 0:
ecli 43:0 Fig.7.AbundanceratiosofFetoSiintheregions(i)–(iii)asafunctionof
D
angulardistancesfromtheNErimtowardthecenteroftheremnant.
W
0
0.
0
0: center-filleddistributionofFe.
3
2:
4
4.2 AmbientgasdensityandSNRage
0.03 0.06 0.13 0.25 0.50 1.0
In our spectral analysis, the volume emission measures are
Right ascension 2.7×1057 cm−3 and 1.0×1058 cm−3 inW and NE,respec-
5:28:00.0 5:26:00.0 5:24:00.0 tively. The volumes areobtained to be4.1×1059 f cm3 and
1.8×1059 f cm3 for the W and NE plasmas, respectively,
0
00. wheref isthe filling factor, assumingthe morphologies tobe
0:
3 prisms whose bases arethe Wand NE regions and depths are
3:
4 40pcand20pcforWandNE,respectively. Withthevolume
emission measures, we obtain the electron densities ne,W =
n (i) (ii) (iii) 0.3(f/0.1)−1/2 cm−3andne,NE=0.9(f/0.1)−1/2 cm−3for
o 0 WandNE,respectively. Theysuggest thatthe ambientgas is
nati 0:00. denser in NE than that in W. This is supported by the radio
cli 3:0 morphology,thesmallershellintheeasternside. Anotherclue
e 4
D
of denser gasin NE isthe GeVgamma-rayemission detected
in the eastern part of the remnant by FermiLAT observations
(Araya2013).Ifπ0decaysarethemainproductionprocessfor
0
0. thegammarays,densegas,whichactsastargetsforaccelerated
0
0: protonsforπ0production,wouldbeinNE.
3
2:
4
In the spectral analysis, we obtained ne,W t = 4.0×
1011 cm−3 s and ne,NE t = 6.1×1011 cm−3 s for the W
and NE plasmas, respectively. With ne,W and ne,NE calcu-
0.03 0.06 0.13 0.25 0.50 1.0 lated above based on the emission measures, the timescales
are estimated to be t = 1.4×105 f1/2 yr for W and t =
Fig.6.XISimagesofG166.0+4.3in1.75–2.10keVband(Top: Si-Kband)
andthe0.70–1.30keVband(Bottom:Fe-LandNi-Lband)aftersubtraction 7.1×104 f1/2 yrforNE. Theionization timescale fortheW
oftheNXBandcorrectionofthevignettingeffect. Theimagesaredivided plasma,whichisanIP,isexpectedtobealmostthesameasthe
bythecorrespondingunderlyingcontinuumimageforwhichweselectedthe age of the remnant, and, therefore, we estimate the age to be
line-freebandof1.50–1.70keV.Thescaleisnormalizedsothatthepeak
becomesunity.ThecontoursarearadioimageoftheWENSSat325MHz. tage=4×104 (f/0.1)1/2 yr. With ne,NE t, wecanestimate
the elapsedtimesince recombination startedtodominate over
ionizationtobet =2×104(f/0.1)1/2 yr(<t ).
rec age
8 PublicationsoftheAstronomicalSocietyofJapan,(2014),Vol.00,No.0
4.3 Originofrecombiningplasma WENSSteam.ThisworkissupportedbyJSPSScientificResearchgrant
numbers15J01842(H.M.),15H02090,26610047and25109004(T.G.T.),
The rarefaction (Itoh & Masai 1989) and thermal conduction
26800102(H.U.),25109004(T.T.),and16J00548(K.K.N.).
(Kawasakietal.2002) scenarios have beenconsidered forthe
formation process of RPs. In either case, the environment of
theremnantisakeytotheRPproductionprocess. Wediscov- References
eredanRPonlyinapartofG166.0+4.3, theouter partof the
Ackermann,M.,etal.2013,Science,339,807
NEregion, wheretheambient gas isdenser. IntheWregion, Albert,J.,etal.2007,ApJL,664,L87
thespectrumisreproducedbytheIPmodel. Thedifferentgas Anders,E.,&Grevesse,N.1989,Geochim.Cosmochim.Acta,53,197
environmentsbetweentheWandNEregionsmightbeactually Araya,M.2013,MNRAS,434,2202
responsibleforthedifferenceoftheplasmastates. Bocchino,F.,Miceli,M.,&Troja,E.2009,A&A,498,139
Burrows,D.N.,&Guo,Z.1994,ApJL,421,L19
In the rarefaction scenario, the electron temperature is de-
Cox,D.P.,etal.1999,ApJ,524,179
creasedbyadiabaticcoolingwhentheblastwavebreaksoutof
Dickey,J.M.,&Lockman,F.J.1990,ARA&A,28,215
dense CSM into rarefied ISM. The lower the ISM density is,
Guo,Z.,&Burrows,D.N.1997,ApJL,480,L51
themoreeffectivetheadiabaticcoolingmustbe(Shimizuetal. Hayato,A.,etal.2010,ApJ,725,894
2012). Therefore, inthis scenario, wecannaturally expect an Ishisaki,Y.,etal.2007,PASJ,59,113
RPintheWregionwhoseambientgasdensitywouldbelower. Itoh,H.,&Masai,K.1989,MNRAS,236,885
Ourresult,wherewefoundanRPonlyintheNEregion,isnot Kalberla,P.M.W.,etal.2005,A&A,440,775
Kamitsukasa,F.,etal.2016,PASJ,68,S7
simplyexplainedbytherarefactionsenario.
Kawasaki,M.T.,etal.2002,ApJ,572,897
Inthethermalconductionscenario,anRPisanticipatedina
Koyama,K.,etal.2007,PASJ,59,23
partoftheremnantwhereblastwavesareincontactwithcool
Landecker, T. L., Pineault, S., Routledge, D., & Vaneldik, J. F. 1982,
dense gas. In the caseof G166.0+4.3, the RP was discovered ApJL,261,L41
indeedinthedenserNEregion. Inourresults,thetemperature Lazendic,J.S.,&Slane,P.O.2006,ApJ,647,350
oftheinnerFe-richplasmaishigher(kT =0.87keV)thanthat Masui,K.,etal.2009,PASJ,61,S115
e
oftheouterRP(kT =0.46keV). Inthisscenario,theplasma Morrison,R.,&McCammon,D.1983,ApJ,270,119
e
Ohnishi,T.,etal.2011,PASJ,63,527
iscooledfromtheoutsidelayers. Therefore,itispossiblethat
Ozawa, M., Koyama, K., Yamaguchi, H., Masai, K., &Tamagawa, T.
theplasmaiscooledonlyintheouterpartoftheNEregion.
2009,ApJL,706,L71
Pineault,S.,Landecker,T.L.,&Routledge,D.1987,ApJ,315,580
Rho,J.,&Petre,R.1998,ApJL,503,L167
5 Summary
Sawada,M.,&Koyama,K.2012,PASJ,64,81
We analyzed the Suzaku XIS data of SNR G166.0+4.3. The Serlemitsos,P.J.,etal.2007,PASJ,59,S9
resultsweobtainedaresummarizedasfollows. Sharpless,S.1959,ApJS,4,257
Shelton,R.L.,etal.1999,ApJ,524,192
1. The plasma in the NE region consists of two components.
Shimizu,T.,Masai,K.,&Koyama,K.2012,PASJ,64,24
One is the Fe-rich component with kTe =0.87+−00..0023 keV. Tawa,N.,etal.2008,PASJ,60,S11
TheotheristheRPcomponentoftheotherlighterelements Uchida,H.,Tsunemi,H.,Katsuda,S.,Kimura,M.,&Kosugi,H.2009,
with kT = 0.46±0.03 keV, kT = 3 keV(fixed) and PASJ,61,301
e init
n t=(6.1+0.5)×1011cm−3s. Uchida,H.,etal.2012,PASJ,64,141
e −0.4
Uchida,H.,Koyama,K.,&Yamaguchi,H.2015,ApJ,808,77
2. The spectrum in the W region is well fitted with a one-
White,R.L.,&Long,K.S.1991,ApJ,373,543
componentIPmodelwithkT =0.83±0.01keV,kT =
e init
Yamaguchi,H.,etal.2009,ApJL,705,L6
0.01keV(fixed)andn t=(4.0+0.7)×1011cm−3s.
e −0.5 Yamaguchi,H.,Koyama,K.,&Uchida,H.2011,PASJ,63,S837
3. The Fe-rich plasma is concentrated near the center of the
remnant whereas the other elements are distributed in the
outerregion.
4. Ourresultsofthespectralanalysisarenotsimplyexplained
by the rarefaction scenario, whereas the origin of the RP
wouldbeexplainedbythethermalconductionwiththecool
densegasinthenortheasternpartoftheremnant.
Acknowledgments
The authors are grateful to Prof. Katsuji Koyama for helpful advice.
WedeeplyappreciatealltheSuzakuteammembers. Wealsothankthe