Suzaku Studies of Magnetic Cataclysmic Variables and the Galactic Ridge X-Ray Emission

Suzaku studies of magnetic cataclysmic variables and the Galactic Ridge X-ray Emission

Takayuki YUASA ASTRO-H Project Researcher, JAXA/ISAS

Kazuo Makishima Kazuhiro Nakazawa The University of Tokyo

Publications: Yuasa, Nakazawa, Makishima et al. 2010, A&A, 520, A25+ Yuasa, Ph.D thesis, The University of Tokyo, 2010 Introduction Galactic Ridge X-ray Emission (GRXE) X-ray background emission observed along with the Galactic disk and the Galactic bulge. Total luminosity is ～1038 erg/s in 2-10 keV. The origin of the GRXE has been one of great mysteries in X-ray astrophysics over 40 years (e.g. Cooke et al. 1969). RXTE (3-20 keV) Revnivtsev+06 銀緯 gal. b

gal.銀経 l 0 Spectral feature (e.g. Koyama et al. 1986) • Continuum + Emission lines. Fe line • Coexistence of lines from light (Si/S) and heavy (Fe) elements. -1 Continuum • Multi-temperature plasma?

Questions GRXE by Tenma • Koyama+86 What is the origin? Log count rate (1/s) -2 • What is the emission mechanism? 1 2 5 10 20 keV The origin of the GRXE

Chandra 1 Ms image Recent !ndings from Chandra deep observation

29:32:00.0 • 1 Ms in the Galactic bulge. ‒ • 80% of detected GRXE !ux was resolved into point

sources (Revnivtsev+09). ‒29:33:00.0 • Main candidates of dim point sources :

coronal X-ray sources + accreting WDs. ‒29:34:00.0

soft X-ray band hard X-ray band 17:51:40.0 17:51:30.0 17:51:20.0 • To further improve imaging sensitivity is not feasible. 1 arcmin

‒29:36:00.0 > 400 sources in 5.2arcmin diameter Our approach : Broad-band spectroscopic study ‒29:37:00.0 1. Construct a spectral model of accreting WDs, especially intermediate polars (IPs), and then check its validity using data of nearby sources. - We concentrate on intermediate polars which have hardest spectra among accreting WDs. - WD masses can be estimated as by products in individual source analyses. 2. Use the IP spectral model to spectroscopically decompose the GRXE. Studies of magnetic accreting WDs

Polar Main-sequence and Roche-lobe-!lling low-mass companion

Intermediate Polar accretion column

accretion curtain and column

acc. disk Modeling a spectrum from an accretion column accreting gas Geometry and emission process • Accreting matter freely falls along the B ﬁeld lines. z axis shock front • A shock converts bulk kinematic energy z~100km accretion into internal energy. column GMWDGMWD kTs kTs mH (m>H10(> keV)10 keV) ∝ R∝WDRWD • The heated gas cools in the post-shock region (PSR) WD via optically thin-thermal X-ray emission.

• Equating conservation laws leads to density, kT and ρ proﬁles (MWD=0.7 Msun) temperature, and velocity proﬁles in the PSR. 40 40 ρ (g cm-3) 4E-8 d d GM (ρv)=0, (ρv2 + P )= WD ρ, dz dz − z2 2E-8

dP dv 2 20 20 ) v + γP = (γ 1)Λn T -3 dz dz − −

thin-thermal cooling function (keV) (g cm kT (e.g. Schure+09) ρ ρ 10 10 1E-810-8 • The system reduces to an initial value problem MWD=0.7 Msun kTRWD=0.0112 (keV) Rsun of ODEs (Cropper+99, Suleimanov+05). z0-RWD=0.0165 R WD

kTs=28.2 keV • Example of the M =0.7 M case. 55 5E-9 WD sun 0.001 1%0.01 0.1% (z-RWD)/RWD Height from the WD surface (normalized with RWD=7800km) Construct a total spectrum • APEC was convolved with the emissivity to produce a total spectrum. Thus, the model consists of multi-temperature emission.

• MWD and Fe abundance are primary parameters of the model. • Updates from previous studies : emission lines + variable metal abundance • Heavier masses result harder spectra, i.e. cutoff energy ∝ shock temp ∝ WD mass • By fitting observed spectra, WD mass can be estimated. • Cutoff energy and Fe line ratio are important factors constraining the mass.

0.45 Close-up view 0.4 10 He-like H-like Fe Kα Fe Kα 1.1 Msun 0.35 1 0.3 -1 MWD s -2 0.9 0.25 1.1 Msun 10-1 vFv 0.7 0.2

10-2 Photons cm 0.15 0.5 MWD=0.3 Msun 0.1 10-3 XIS HXD 0.05 0.3 Msun 10-4 0 1 10 100 6.4 6.6 6.8 7 7.2 Energy (keV) Energy (keV) Counts/sCounts s!1 keV!1 Counts/sCounts s!1 keV!1 !3 !3

3 10 0.01 0.1 10 0.01 0.1 TV Columnbae χ χ • • • 2 2 Spectral ﬁttingwithnear-byIPs(Yuasa+10, A&A) Note：In thepresentstudy6.4-keVFelinewasarbitrarilymodeledusingaGaussian. The IPmodelsuccessfullyreproducedoverallspectraandtheFelinestructure. Single kTmodel(apec) → Felinesandcontinuum The XISclearlyresolvesthreeFelines.HXDdetectssignalsupto40-50keV. v v =1.11 (321)NHP=8% =1.50 (326)NHP=1e-8 5 5 5 Energy (keV) 10 10 10 Energy (keV) Energy (keV) Single-kT (apec) IP PSRmodel + absorption + absorption 20 050 20 50 20 30 40 50

Counts/sCounts s!1 keV!1 Counts/sCounts s!1 keV!1 0 0.1 0.2 0.3 0 0.1 0.2 0.3 from theFelineratio. temperature thanthoseexpected Continuum requireshigher Z M Best ﬁtparameters: Fe WD =0.41(0.38-0.45) Zsun =0.91(0.83-1.00) Msun 6 inconsistent. Neutral Energy (keV) Energy (keV) Energy (keV) He-like 7 H-like 0.1 ) 1 − keV 1 − s 2 − 1 cm 2 0.0 vFv (keV 0.43(0.41-0.45) Msun 0.79(0.75-0.83) Msun 0.93(0.78-1.08) Msun 1.04(0.94-1.16) Msun 3 − 10 0.1 ) 1 − keV 1 − s 2 − 1 cm 2 0.0 vFv (keV

3 0.51(0.47-0.60) Msun 0.80(0.65-1.07) Msun 0.94(0.79-1.05) Msun 1.07(0.92-1.17) Msun − 10 0.1 ) 1 − keV 1 − s 2 − 1 cm 2 0.0 vFv (keV

3 0.63(0.60-0.70) Msun 0.83(0.73-0.96) Msun 0.95(0.73-1.16) Msun 1.09(0.96-1.18) Msun − 10 0.1 ) 1 − keV 1 − s 2 − 1 cm 2 0.0 vFv (keV

3 0.66(0.60-0.71) Msun 0.91(0.83-1.00) Msun 0.95(0.83-1.05) Msun 1.24(0.85-1.30) Msun −

10 5 10 20 50 5 10 20 50 5 10 20 50 5 10 20 50 Energy (keV) Energy (keV) Energy (keV) Energy (keV) Mean M of “isolated WDs” NY_Lup WD 18 (from SDSS by Kepler+07) IGR_J17303 0.593±0.0016 MSun 16 PQ_Gem V709_Cas 14 BG_CMi MU_Cam 12 XY_Ari IGR_J17195 10 YY_Dra TV_Col 8 J2133 TX_Col 6 V1223_Sgr V2400_Oph Mean M of our 17 IPs 4 GK_Per WD 0.88±0.24 MSun AO_Psc 2 FO_Aqe Mean MWD of non-mag CV(RK+03) EX_Hya 0.82 MSun 0 0.2 0.4 0.6 0.8 1 1.2 1.4 WD mass MWD from radial velocity measurements (Msun) References : Hellier97 - XY Ari Penning85 - BG CMi 1.4 Hellier93 - TV Col Haswell+97 - YY Dra Beuermann+04 - V1223 Sgr Beuermann & Reinsch08, Mhlahlo+07 - EX Hya 1.2 XY 1.01.0 V1223 YY 0.8 (RK Cat IP) BG WD EX TV M 0.60.6

0.4 EX

No systematic uncertainty0.20.2 strongly aﬀects the result. The model seems0.2 valid0.2 for 0.4the presently0.6 available0.8 data.1.01.0 1.2 1.41.4 This work MWD(Msun) →Apply the model to the broad-bandM GRXEWD (! spectrum.is work) Spectral analysis of the GRXE Suzaku GRXE observations • 4.4 Ms exposure with 92 pointings in the Galactic Center region as of 2010. • Selected "elds “Region 1” and ”Region 2” → Total exposure = 1Ms.

XIS mosaic image (4.4Ms total) Region1 590ks HXD/PIN full FOV

Region2 420ks 3.5° 5° Correlation of GRXE and NIR • Revnivtsev+06 reported nice correlation between the surface brightness of the GRXE and the NIR diﬀuse emission using RXTE Galactic plane scan data. • Since NIR surface brightness traces the stellar density, the GRXE origin was suggested to have connection with it (i.e. stellar origin). • The present Suzaku data also con"rms this correlation.

0.08 Region 1 0.07

0.06

0.05

0.04

(This work)(This 0.03 Region 2

GRXE HXD countGRXE rate 0.02

0.01

0 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 NIR COBE/DIRBE (Dwek+95) Close-up view of the XIS spectrum

• Lines from lighter (S/Ar/Ca) elements coexist with those from Fe (recon"rmation of Ebisawa+08 with much higher statistics). • This strongly indicates contributions from thermal emissions with diﬀerent temperatures. • At least two distinct plasma temperatures are necessary to reproduce the spectrum.

Power law + Gaussian lines (phenomenological)

0.1 S XV Kα Ar XVII Kα Fe I Kα Fe XXV Kα

1 Fe XXVI Kα − S XVI Kα Ar XVIII Kα keV 1 1 − Ca XIX Kα 0.0 Counts s 3 −

10 2 3 4 5 6 7 8 9 Energy (keV)

S XV Kα ~ 25 eV Ar XVII Kα ~ 40 eV Fe I Kα ~ 75 eV S XVI Kα ~ 35 eV Ar XVIII Kα ~ 15 eV Fe XXV Kα ~ 320 eV Ca XIX Kα ~ 25 eV Fe XXVI Kα ~ 80 eV Fitting the GRXE in the hard X-ray band • A power-law model gives a soft index, Γ=2.8±0.2, and extrapolation of this to lower energies contradicts with the XIS spectrum. • Single-temperature thermal model gave the best "t at kT=15.7 (13.7-18.4) keV.

• The IP spectral model well reproduced the data with MWD=0.66 (0.59-0.75) MSun. - This could be interpreted as a representative WD mass of IPs in the Galaxy. (c.f. ～0.5 MSun by Krivonos+07 with INTEGRAL data) 3 −

10 IPIP model Model (IP_PSR) ×

5 Region1 Region1 M WD- M=0.66(0.59-0.75)WD=0.66!"#$%&"#'$( Msun Msun 2 χ v=1.37(30)2 1 - χ =1.37(30)

− ν 3 − keV Region2Region2 1 − 10 Count rate 4 Counts s − 10 × 2 2 χ 0 2 − 20 3030 4040 50 EnergyEnergy (keV) (keV) Broad-band decomposition Fit with the IP PSR model + single-kT thermal model. (1) MWD was ﬁxed at the HXD result. Region 1 1 Single-kT thermal (apec) nH = 3.6±0.3 x 1022 cm-2

kT = 1.52±0.04 keV ZFe = 0.86±0.05 Zsun

1 2 χ v = 1.30 (753) 0.

ï IP PSR model

1 MWD = 0.66 Msun(ﬁxed) keV 0.0 ï CXB ï 10 Foreground ï Counts s 10 2 0 r 2 ï 2 5 10 20 50 EnergyEnergy (keV)(keV) Broad-band decomposition Fit with the IP PSR model + single-kT thermal model. (2) MWD was allowed to freely vary. Region 1 ! Single-kT thermal (apec) nH = 3.2±0.4 x 1022 cm-2

kT = 1.44±0.07 keV ZFe = 0.73±0.07 Zsun ! 2 χ v = 1.28(752) "#

ï IP PSR model

! MWD = 0.48±0.05 Msun *+,- "#" ï

ï CXB !" Foreground ï $%&'()*) !" / " r / ï / . !" /" ." Energy (keV) low-kT component ～1-1.5 keV →0',123*4+,-5 typical to coronal sources (active binary stars) high-kT component～>10 keV → well reproduce by the IP model The X-ray observatory ASTRO-H (2014-) X-ray micro-calorimeter ~5 eV resolution (FWHM) at Fe K emission lines. → Resolves the "ne structures. → Direct measurements of plasma density, velocity, Einstein redshift. → Does the GRXE have red-shifted neutral Fe K line? (if emitted from re!ection from the WD surface, a shift of ~50 km/s~1eV) Hard X-ray Imager with a collection mirror High statistics and imaging in the hard X-ray band.

TV Col (simulatedSimulation for AH/SXS 100ks 5eV of opt the blck filt) IP TV Col observationtvcol hxi 100ks for 100 ks SXS Hard X-ray Imager

He-like H-like 0.1 ) 1 1 − Fe Kα Fe Kα − keV keV 1 1 − − s 1 2 − (Photons cm 2 0.5 2 normalized counts s keV 0.02 0.05

6.5 6.6 6.7 6.8 6.9 7 0.01 10 20 50 6.6 Energy6.8 (keV) 7 keV 10 Energy20 (keV) 50 keV Summary of results Hinode/JAXA • An X-ray spectral model for IPs was constructed, and its validity was conﬁrmed by applying it to Suzaku spectra of 17 nearby IPs. WD mass estimates were obtained as byproducts.

• Broad-band GRXE spectra (2-50 keV) were extracted from 1-Ms Suzaku data. • Hard X-ray spectrum was well reproduced with the IP spectral model. • The temperature of the low-kT component is consistent with typical temperature of coronal sources such as active binary stars. Conclusion

The present study precisely measured the GRXE spectrum over 1-50 keV, and decomposed it into to distinctive components which have physical counterparts such as active binary stars and accreting WD binaries.

Being complementary to the imaging result, this result also supports the point source scenario as the origin of the GRXE.

LogN-LogS of required point sources will be available in Q&A session. Number density of hard X-ray sources • The GRXE in the hard X-ray band could be explained by magnetic CVs. • Surface number density : α L − N(>L)=N0 logN-logS of hard sources L0 2 • Integrated luminosity : 10 30 -1 Lmin=2×10 ergerg s Revnivtsev+09 L min dN(>L) ~20 arcmin-2 S = L dL 10 dL L =3.8×1031eergrg s-1 Lmax min

) ) 1 -2 -2 - α = 1.0 - 1.5 (e.g. Muno+09) -1 - S = 2.34×1035 erg/s/HXD-FOV 10 ) (arcmin (GRXE ﬂux measured by the HXD) ) (arcmin L L -2 α=1.5 (> 34 -1 (> 10 N - Lmax = 1.3×10 erg s N !is workwork (from absence of bright point 10-3 Chandra Galactic center sources inside the FOV) (Muno et al. 2009) - L min 10-4 α=1.0 case 1 : 2×1030 erg s-1 (Revnivtsev+09) 30 31 32 33 34 case 2 : 3.8×1031 erg s-1 (Muno+09) 10 10 10 10 10 Luminminositity (erg s-1 in the 12-50 keV)