ISM composition through X-ray spectroscopy of LMXBs

0.4 MB·English

by  C. Pinto

Astronomy&Astrophysicsmanuscriptno.paper˙lmxb (cid:13)c ESO2013
January9,2013
ISM composition through X-ray spectroscopy of LMXBs
C.Pinto1,2,3,J.S.Kaastra1,4,E.Costantini1,andC.deVries1
1 SRONNetherlandsInstituteforSpaceResearch,Sorbonnelaan2,3584CAUtrecht,TheNetherlands.
2 DepartmentofAstrophysics/IMAPP,RadboudUniversity,POBOX9010,6500GLNijmegen,TheNetherlands.
3 InstituteofAstronomy,MadingleyRoad,CB30HACambridge,UnitedKingdom,e-mail:[email protected].
4 AstronomicalInstitute,UtrechtUniversity,P.O.Box80000,3508TAUtrecht,TheNetherlands.
Received2October2012/Accepted7January2013
3
1 ABSTRACT
0
2 Context.Thediffuseinterstellarmedium(ISM)isanintegralpartoftheevolutionoftheentireGalaxy.Metalsareproducedbystars
andtheirabundancesarethedirecttestimonyofthehistoryofstellarevolution.However,theinterstellardustcompositionisnotwell
n
knownandthetotalabundancesareyettobeaccuratelydetermined.
a
Aims.WeprobeISMdustcomposition,totalabundances,andabundancegradientsthroughthestudyofinterstellarabsorptionfeatures
J
inthehigh-resolutionX-rayspectraofGalacticlow-massX-raybinaries(LMXBs).
8
Methods.Weusehigh-qualitygratingspectraofnineLMXBstakenwithXMM-Newton.WemeasurethecolumndensitiesofO,Ne,

Mg,andFewithanempiricalmodelandestimatetheGalacticabundancegradients.
]
A Results.Thecolumndensitiesoftheneutralgasspeciesareinagreementwiththosefoundintheliterature.Solidsareasignificant
reservoir of metalslike oxygen and iron. Respectively, 15–25% and 65–90% of the total amount of Oi and Fei isfound in dust.
G The dust amount and mixture seem to be consistent along all the lines-of-sight (LOS).Our estimates of abundance gradients and
. predictionsoflocalinterstellarabundancesareinagreementwiththosemeasuredatlongerwavelengths.
h Conclusions.OurworkshowsthatX-rayspectroscopyisaverypowerfulmethodtoprobetheISM.Forinstance,onalargescale
p
theISMappearstobechemicallyhomogeneousshowingsimilargasionizationratiosanddustmixtures.Theagreementbetweenthe
- abundancesoftheISMandthestellarobjectssuggeststhatthelocalGalaxyisalsochemicallyhomogeneous.
o
r Keywords.ISM:abundances–ISM:dust,extinction–ISM:molecules–ISM:structure–X-rays:ISM
t
s
a
[
1. Introduction presenceofdifferentionizationstatesconstrainsthemulti-phase

1 temperaturestructureoftheISM.Asummaryoftheresultsob-
v Theinterstellarmedium(ISM)isoneofthemostimportantcom- tainedby X-rayspectroscopyonthe ISM in the last decadesis
2 ponentofgalaxiesbecauseitinfluencestheirevolution:theISM givenbyPintoetal.(2010).
1
isenrichedwithheavyelementsbystellarevolution,anditpro- Briefly, Schattenburg&Canizares (1986) first constrained
6
vides the source of material for following generation of stars. interstellar oxygen in the X-ray band using the Einstein
1
. Thereareseveralphysicalprocessesthatalterthemetallicityof Observatory, but a thorough study of the ISM was possible
1 theISM.Stellarwindsandsupernovaeexpelpartoftheinterstel- only after the launch of the XMM-Newton and Chandra satel-
0 largasoutoftheGalacticdisk,butgravitygenerallyforcesthe lites, providedwith high-resolutiongratings. Absorption struc-
3
gas to fall back through the process known as ”Galactic foun- ture due to ionized gas and dust was found around the inter-
1
tain”(Shapiro&Field1976).Gasaccretedfromsmallergalax- stellar oxygen K-shell absorption edge in the spectra of sev-
:
v ies,liketheMagellanicClouds,andtheintergalacticmediumin- eralsources(Paerelsetal.2001;deVriesetal.2003;Juettetal.
Xi creasesthereservoiroflowmetallicitygas.TheISMalsoshows 2004;Costantinietal.2005,2012;deVries&Costantini2009;
acomplexstructureconsistingofphasesatdifferentequilibrium Kaastraetal.2009;Pintoetal.2010,2012a,b).
ar temperatures(forareview,seeDraine2011).Thecoldphase is However, in the last decades, the most important discover-
ablendofdust,moleculesandalmostneutralgasbelow104 K. ies on the physics of the ISM phases have been obtained with
Thewarmandhotphasesaremostlygaseous.Thewarmgasis UV,IR,optical,andradiospectroscopy.Dustisusuallystudied
weakly ionized,with a temperatureof ∼ 104 K. The hotgasis in IR, molecules in radio and IR, while the neutral and warm
highly ionized, with temperatures of about 106 K. These three phasesoftheinterstellargasarealsoprobedwithUVdata(see
mainphasesare notentirelyseparatedfromeachother. Forin- Ferrie`re 2001 and references therein). The hot gas is the only
stance,itisthoughtthataconductivecoolinglayer,revealedby ISMphasewhichismoreoftenstudiedinX-rays.However,de-
Ovi, lies between the warm and hot ionized phases (see e.g. spite the very high resolution and signal-to-noise ratio at long
Richter2006). wavelengths, both the chemistry and physics of the interstellar
InthespectraofbackgroundsourcestheISMproducesred- medium are still underdebate. Elementalabundances,dustde-
deningandabsorptionlines.HighresolutionX-rayspectroscopy pletion factors and dust chemistry are not yet well determined
has become a powerful diagnostic tool for constraining the and thus provide an open and interesting research field. The
chemicalandphysicalpropertiesoftheISM.TheK-shelltransi- heavyelementssuchasoxygen(themostabundantone)andiron
tionsofcarbon,nitrogen,oxygen,neon,andmagnesium,andthe areproducedinhigh-massstars.Moleculesanddustcompounds
L-shelltransitionsofironfallinthesoftX-rayenergyband.The are mainly formed in AGB stars and then grow in the diffuse
2 C.Pintoetal.:ISMcompositionthroughX-rayspectroscopyofLMXBs
ISM (see e.g.Mattsson&Andersen2012).Heavy elementsdi- superimposable, we have stacked the spectra according to the
rectlywitnessthehistoryofthepaststellarevolution.Inthelast procedure described by Kaastraetal. (2011). Briefly, we have
yearstherehavebeenmanyattemptstomeasuretheinterstellar reducedthespectrawiththeScienceAnalysisSystem(SAS)ver-
abundances,butinmanycasesonlylowerlimitswereobtained sion11.0,obtainedfluxedspectraforRGS1and2andstacked
duetothedepletionoftheelementsfromthegaseousphasesinto them with the SPEX taskRGS fluxcombine.We cannot stack
dustgrainsandtostronglinesaturationatlongwavelengths.The thespectraofvariablesourcesbecausethismayintroducespuri-
dustdepletionfactorsandespeciallythecompoundsmixtureare ousfeaturesneartheinterstellaredges.Thus,wesimultaneously
not yet well known and thus the total ISM abundances are af- fit their spectra coupling the parameters of the interstellar ab-
fectedbystrongsystematicuncertainties.Forinstance,asignif- sorbers,butadoptingadifferentcontinuumforeachobservation.
icant fraction of the oxygen which is bound in the solid phase ThiscanbedonewiththeSPEXtasksectors.
isstillunknown:silicatesandcarbonaceouscompoundsarenot The LMXBs GS 1826−238, GX 9+9, Ser X−1,
abundantenoughto coverthe amountof depletedoxygen.Ices 4U 1254−690,and 4U 1636−536showed little or insignificant
maysolvethisproblem,buttheyarehardtodetect(forareview spectralvariability,thuswecouldstacktheirRGS1and2spec-
aboutdepletionfactorsintheISMseeJenkins2009).Therefore, tra, which were taken during all the observations, and produce
our work focuses on an alternative method to determine ISM a single final fluxed spectrum for each LMXB (see Figs. 2 –
columndensities,dustdepletionfactors,molecularcompounds, 4). We adopted a similar approach in Pintoetal. (2010) for
totalabundances(gas+dust),andabundancegradientsthrough GS1826−238.Therewefilteredouttheburstingemissionfrom
X-rayspectroscopy. thespectra butalso reportedthatitwas negligiblecomparedto
In this work we report the detection and modeling of in- thepersistentemission.Thereforewedonotremovetheperiods
terstellar absorption lines and edges in the high-quality spec- ofburstsinanyofourdata.
tra of nine LMXBs: GS 1826−238, GX 9+9, GX 339−4, For Aql X−1, SAX J1808.4−3658, and 4U 1735−444 we
Aql X−1, SAX J1808.4−3658, Ser X−1, 4U 1254−690, 4U have only foundone very goodspectrum. Theirotherobserva-
1636−536, and 4U 1735−444. The observations are taken tions available in the archive have either too short exposure or
with the XMM-NewtonReflection Grating Spectrometer(RGS, much lower flux in such a way that they do not significantly
denHerderetal. 2001), see Table 1 for details. These sources improve the statistics. For each of these sources we simply fit
are well suited for the analysis of the ISM because of their theRGS1and2countspectrasimultaneously.TheLMXBGX
brightness and column densities N ∼ 1 − 5 × 1025m−2 (see 339−4isanexception.Itsspectralcontinuumandslopestrongly
H
Table 3), which are high enough to produce prominent O, Fe, vary,butthethreeobservationswithIDs0148and0605havesu-
andNeabsorptionedgesinthesoftX-rayenergyband.Thispa- perimposablespectra,aswellastheremainingtwoobservations
perprovidesasignificantextensionofthepreviousworkonthe (IDs 0204). Thereforewe split the spectra in these two groups
ISM in the line-of-sight(LOS) towardsGS 1826−238doneby and stack them producing two final fluxed spectra with com-
Pintoetal.(2010).Wehaveusedupdatedatomicandmolecular parable statistics. In Figs. 2 – 4 with show the Ne, Fe, and O
databaseandextendedtheanalysistoseveralLOS(seeFig.1). absorptionedgesoftheGX339−4fluxedspectrumobtainedby
Our analysis focuses on the 7−38 Å first orderspectra of stackingthegroupofobservationswithID0204.
the RGS detectors. We have not used the EPIC spectra of the In Table 1 we report the RGS exposure times and iden-
observationsbecause they have much lower spectralresolution tification number for each observation. We also provide the
and we were not interested in determining the source contin- Galactic coordinates, distance d and hydrogen column density
uumatenergieshigherthantheRGSband.Mostofinterstellar N for each source. Most of the distances have been taken
H
X-ray featuresare indeedfoundin the softX-ray energyband. from Gallowayetal. (2008), which estimated them through
WeperformedthespectralanalysiswithSPEX1version2.03.03 the luminosity peak during the LMXBs burst. The N ranges
H
(Kaastraetal.1996).Wescaledtheelementalabundancestothe refer to the spread between the values measured by the
proto-SolarabundancesofLodders&Palme(2009).Weusethe Leiden/Argentine/BonnSurvey of Galactic Hi and the Dickey
χ2-statisticthroughoutthepaperandadopt1σerrors. & Lockman Hi in the Galaxy, see Kalberlaetal. (2005) and
Thepaperisorganizedasfollows.InSect.2wepresentthe Dickey&Lockman(1990).TheN valuesasprovidedbythese
H
data.InSect. 3 we reportthe spectralfeaturesthatwe analyze. two surveys differ by up to 30% and we will determine them
InSect.4wedescribethemodelsweuseandtheresultsofour fromtheX-rayspectrum.IntheRGSfitswerebinthespectraby
analysis.Thediscussionandthecomparisonwithpreviouswork afactoroftwo,i.e.about1/3FWHM(thefirstorderRGSspec-
aregiveninSect.5.ConclusionsarereportedinSect.6. traprovidearesolutionof0.06−0.07Å).Thisgivestheoptimal
binningfortheRGSandabinsizeofabout0.02Å.
2. Thedata
Forthisworkwehaveused22archivalspectraofninesources 3. Spectralfeatures
takenwithXMM-Newton(seeFig.1andTable1).Thesespectra
In Figs. 2 – 4 we show the neutral absorption edges of neon,
weretakeninperiodsofhigh-fluxstateofthesourcesandeach
iron, and oxygenfor the nine LMXBs, which are the strongest
singlespectrumshowshighstatistics. We haveexcludedin our
spectral features. SAX J1808.4–3658 also shows a prominent
analysisadditionalobservationstakenduringlow-fluxstates.For
Ni1s–2pedgeat31.3Å,butmostofthesourceshaveweakni-
some sourceswe have just one good spectrum,while for other
trogenand magnesiumedges, which meansthat the Mg and N
ones we obtained up to five high-quality spectra. The spectra
abundancesarenotwellconstrained.TheNeiedge(see14.3Å
taken on the same source have been fitted together in order to
inFig.2)isshallowerthanthoseofFeandO,butitisinteresting
increasethestatistics,butwehaveusedadifferentapproachbe-
forthepresenceofadditionalabsorptionlinesduetomildlyion-
tweenstableandvariablesources.Whenthepersistentemission
izedNeii-iiiandheavilyionizedNeixgas.Forafewsourceswe
of a certain source was steady and the spectra were perfectly
are even able to detect a weak Neviii line at about13.7Å. For
1 www.sron.nl/spex somespectraneartheneonedgetwoweakabsorptionfeaturesat
C.Pintoetal.:ISMcompositionthroughX-rayspectroscopyofLMXBs 3
Table1.XMM-Newton/RGSobservationsusedinthispaper.
Source ID(a) Length(ks)(b) l,b(c) d(kpc)(d) N (1025m−2)(e)
H
0060740901 28.5
4U1254–690 0405510301 60.6 303.5,−6.4 15.5(f) 2.2–2.9
0405510501 60.7
0105470401 20.6
0303250201 30.9
4U1636–536 332.9,−4.8 5.95(f) 2.6–3.6
0500350301 31.5
0606070201 28.8
4U1735–444 0090340201 20.6 346.1,−7.0 6.5(f) 2.6–3.0
AqlX–1 0406700201 52.6 35.7,−4.1 3.9(f) 2.8–3.4
0150390101 106.3
GS1826–238 9.3,−6.1 6.7(f) 1.7–1.9
0150390301 91.5
0148220201 20.3
0148220301 13.7
GX339–4 0204730201 13.3 338.9,−4.3 8(g) 3.7–5.3
0204730301 13.4
0605610201 32.9
0090340101 10.6
GX9+9 8.5,9.0 7.5(h) 2.0–2.1
0090340601 21.7
SAXJ1808.4–3658 0560180601 63.7 335.4,−8.1 2.77(f) 1.1–1.3
0084020401 21.5
SerX–1 0084020501 21.7 36.1,4.8 7.7(f) 4.0–4.7
0084020601 21.8
(a,b) Identificationnumber and duration of theobservation. (c,d,e) Source Galacticcoordinates indegrees, distance fromthe Sun, and hydrogen
column density (see Kalberlaetal. 2005 and Dickey&Lockman 1990). (f) Distance estimates are taken from Gallowayetal. (2008). (g) For
GX339-4 we adopt the distance as suggested by Zdziarskietal. (2004). (h) The distance for GX9+9 is the mean value between those ones
providedbyChristian&Swank(1997)andZdziarskietal.(2004).
theseedges.We alsoprovidethepositionoftheOviiβ(1s–3p)
andOviiiα(1s–2p)absorptionlinesat18.6and19.0Å,respec-
tively. These two lines also trace the hot ionized gas. Fig. 4
shows the most interesting part of the spectrum: the oxygen
K edge. The Oi 1s–2p line at about 23.5Å is the strongest
ISM spectralfeature,the neutraloxygenisalso responsiblefor
the jump in spectrum between 22.5 and 23.2Å. Additionalbut
not less important 1s–2p absorption lines are produced by Oii
(23.35Å), Oiii (23.1Å), Ovii (21.6Å), and occasionally Ovi
(22.0Å).Dustoxygencompoundsaffectmainlythespectralcur-
vaturebetween22.7−23.0Å(seee.g.Pintoetal.2010,andref-
erences therein). An easy way to check for strong amounts of
dust is to compare the depth of the edge with that of the line.
Dust usually is responsible fora deep edge, while the gas also
providesstrong lines. Forinstance,in SerX-1 the O K edgeis
deeperthanthe1s–2pline,whileinSAXJ1808.4–3658theab-
sorption line is clearly deeper than the edge. This means that
Fig.1. Map of the X-ray sources as projected on the Galactic alongtheLOStowardsSerX-1theamountofoxygenfoundin
plane.TheSunisassumedtobe8kpcfarawayfromtheGalactic dustcompoundsisexpectedtobelarger.Wewillconfirmthisin
Center(GC). Sect. 5. Dust is also expected to providemost of iron (see e.g.
Jenkins2009),whichmeansthatthebulkoftheabsorptionbe-
15.0Åand14.2ÅarevisiblethatarepresumablyduetoFexvii tween17–17.5Åisproducedbyirondustcompounds(seealso
andFexviiiinthehotgasphase. Costantinietal. 2012). Dustis also responsibleformostof the
magnesiumedgeatabout9.5Å.
The iron L2 and L3 edges are located at 17.15 and 17.5Å
(seeFig. 3).Thesourcesshowcleardifferencesin thedepthof
4 C.Pintoetal.:ISMcompositionthroughX-rayspectroscopyofLMXBs
0 0
0
4 0
6
1 3
NeIX NeIX NeIX
NeIII NeIII NeIII
NeI NeII 0 NeI NeII 0
0 2 5 NeI NeII
5 1 2
0
0
0
0
0 1
2
4
4U 1254−690 4U 1636−536
4U 1735−444
0
)1 5
A− 0 3 50
2
2
1 1
−
s
−2 10 0
m 1 0
3
s 0
n 0 0
o 0 2
t 1
o
h 0
p 0 5
( 9 GS 1826−238 2 GX 9+9 Aql X−1
x
u 0
Fl 70 5
0 1
6
1 0
6
0 2
0 6
4 0
1 4
2
0 0
2 5 0
2
1
2
Ser X−1 GX 339−4 SAX J1808.4−3658
0 0
0
0 0
4
1 13.5 14 14.5 15 13.5 14 14.5 15 2 13.5 14 14.5 15
Wavelength (Å) Wavelength (Å) Wavelength (Å)
Fig.2.Dataandbest-fittingmodels:theNeKedge.
4. SpectralmodelingoftheISM row absorption lines which do not depend on the continuum.
Thus,wealsotestedbothPLandacombinationofBB/COMT
TheadvantageofusingLMXBsfortheanalysisoftheISMlies in modeling our RGS spectra (bband comtmodels in SPEX).
in their high-statistics and simple spectrum. All the sources in ThebestfitvaluesofN wereconsistentwithintheerrors,which
H
our sample provided us with well exposed spectra, which are validatesourchoiceofasimplePLcontinuum.
easy to modelby adoptinga simple continuum.In most of the
cases one power-law (PL) continuum gave an acceptable and WehavebuiltanempiricalmodelfortheISMsimilartothat
quickly-convergingfit.WhenasinglePLdidnotprovideagood onesuccessfullyusedbyPintoetal.(2012a)butwithafewuse-
fitthenwewereabletoobtainanacceptablefitbysimplyadding ful changes. Essentially, they used three components modeled
anotherPL. The continuumof LMXBsusuallyoriginatesfrom withtheslabmodelofSPEXtofitthethreemainphasesofthe
their accretion disk and corona in the form of blackbody and interstellargas.Theslabmodelgivesthetransmissionthrougha
comptonizedemission,respectively.However,weareusingonly layerofgaswitharbitraryioniccolumndensities.Herewealso
the soft part of the X-ray spectrum which is relevant for the use twoslabcomponentstomodelthewarm andhotphasesof
ISMabsorptionandisnotbroadenoughtoprovideaccuratecon- the ISM, but we prefer to model the cold neutral gas with the
straintsontheX-rayemissioncomponents.Moreover,thiswork hotcomponentinSPEX.Thehotmodelcalculatesthetransmis-
does not focus on the nature of the continuum intrinsic to the sion of a collisionally-ionizedequilibrium plasma. For a given
sources.Thus,weprefertousePLcomponentsratherthanphys- temperatureandsetofabundances,themodelcalculatestheion-
ical ones like blackbody (BB) or comptonization (COMT). In izationbalanceandthendeterminesalltheioniccolumndensi-
principlethechoiceofthecontinuumcomponentmayaffectthe tiesbyscalingtotheprescribedtotalhydrogencolumndensity.
estimate of the hydrogencolumn density,while the otherionic Atlowtemperaturesthismodelmimicsverywelltheneutralin-
columnsarewellconstrainedastheyareestimatedthroughnar- terstellar gas (see SPEX manual and Kaastraetal. 2009). Free
C.Pintoetal.:ISMcompositionthroughX-rayspectroscopyofLMXBs 5
0
5
0
3
2
Fe L2
0 Fe L2 Fe L2
8
30 Fe L3 βOVII αVIII 0 Fe L3 βOVII αVIII 150 Fe L3 βOVII αVIII
5 O 6 O
2 O
0
0
0 0 1
2 4
4U 1254−690 4U 1636−536 4U 1735−444
0
A)−1 15 20 500
2

s−1 60 80 1
2 1 0
− 0
m 0 1
0 6
s 5 1
n
0
o
0 8
t
o 0 4
h 4 1
p 0
( 0 6
ux 0 GS 1826−238 12 GX 9+9 Aql X−1
Fl 3 30
0
2
0
6 5 2
2
0
0
0 2
2
0
4 0
5 8
1 1
Ser X−1 GX 339−4 SAX J1808.4−3658
0
0 0
6
2 1
17 17.5 18 18.5 19 17 17.5 18 18.5 19 1 17 17.5 18 18.5 19
Wavelength (Å) Wavelength (Å) Wavelength (Å)
Fig.3.Dataandbest-fittingmodels:theFeL2andL3edge.
parameters in the hotmodel are the hydrogen column density Yaoetal. 2009 on high-resolution Chandra X-ray spectra of
N ,thetemperatureT,thevelocitydispersionσ ,thesystematic severalLMXBs).We tookintoaccountabsorptionbyinterstel-
H v
velocityv,andtheabundances.However,inX-raysthespectral lar dustwith the SPEXamolcomponent.Theamolmodelcal-
resolutionisnothighenoughtoconstrainthevelocitieswithsuf- culates the transmission of various molecules, for details see
ficientaccuracy,thusweassumethattheinterstellargasisatrest. Pintoetal. (2010), Costantinietal. (2012) and the SPEX man-
Wealsoassumeatemperatureof0.5eV(about5800K),thelow- ual(Sect.3.3).Themodelcurrentlytakesintoaccountthemod-
estvalueavailableintheSPEXhotmodel,becauseinthisway ified edge and line structure around the O and Si K-edge, and
the gas is mostly neutral. In X-rays with the current satellites the Fe K / L-edges, using measured cross sections of various
it is not possible to resolve the narrow lines of the interstellar compoundstakenfromtheliterature.Relevantandabundantsili-
medium.Aspectralfitcaneasilyprovideawrongvalueofveloc- cates,icesandorganicmoleculesarepresentintheourdatabase.
ity dispersion σ and consequently of column density because In fitting the molecular models we limit the range of column
v
they are degenerate. Therefore it is more appropriate adopting densitiestophysicalvalues.Forinstance,wetookcarethatthe
aphysicallyreasonablevaluefortheσ andjustfittingthecol- column density of each weakly-abundant element did not ex-
v
umndensityofeachion.Weadoptanominalvalueof10kms−1 ceedtheproto-Solarvaluepredictedbythebest-fittingN .This
H
for the velocity dispersion of the cold (Oi) and warm (Oii-v) isimportantespeciallyforCaandAl(containedrespectivelyin
gas components, which is similar to that found by Pintoetal. andraditeandhercinite),whoseabundancesaremuchlowerthan
(2012a)usingahigher-resolutionUVspectrum.Forthehotgas oxygen.
(Ovi and higher ionization states) we adopt σ =100kms−1,
v
which is an average value between those suggested in the lit-
erature (see for instance the work by Yao&Wang 2005 and
6 C.Pintoetal.:ISMcompositionthroughX-rayspectroscopyofLMXBs
0 0
2 4
4U 1254−690 OIII OII 4U 1636−536 OIII OII 4U 1735−444 OIII OII
0
15 OVIIα OIV OI 30 OVIIα OIV OI 10 OVIIα OIV OI
0
0 2
1 0
5
0
1
5
)
1
−
A 0
−1 25 GS 1826−238 10 GX 9+9 Aql X−1
s
0
−2 0 4
m 0 8
2
s
n
o 5 0
ot 1 6 0
h 2
p
( 0 0
x 1 4
u
Fl 25
0 Ser X−1 GX 339−4 SAX J1808.4−3658
2 0
1 0
5 5
1 1
0
1 5
0
5 0
1
0 0
21.5 22 22.5 23 23.5 21.5 22 22.5 23 23.5 21.5 22 22.5 23 23.5
Wavelength (Å) Wavelength (Å) Wavelength (Å)
Fig.4.Dataandbest-fittingmodels:theOKedge.
4.1.Thefittingprocedure measuredthe∆χ2oftheabsorbingcomponentsineachedge,we
jointlyfit the edgesorthe spectralrangesthatsharetransitions
Thefirststepisdetermininga preliminarymodelforthewhole fromsameionsorcompounds.Thewavelengthintervalbetween
spectrum.Thisrequiresto fit the spectrumof each source with
15 and 24Å contains the strong Fe and O edges, and the bulk
a simple modelconsisting of one (ormore)power-lawcompo-
ofthe dustsignatures.Thefit ofthis spectralregionprovidesa
nent(s)absorbedbyneutralgas(hotmodelinSPEX).Asalready
startingmodelforthefinalfitextendedtothewholespectrum.
mentioned,one or two power-law componentsare sufficient to
reproduce all the spectral continua. This fit provides prelimi-
nary values for the normalization and slope of the power-law 4.2.Step-by-stepanalysisforGX339-4
as well as the hydrogencolumndensity and the abundancesof
theneutralgas.Thesecondstepisdeterminingthecontributions InthissectionweshowthedetailedanalysisofthetwoGX339-
of each absorbing component.Some compoundsand ions pro- 4 stacked spectra between 15 and 24Å. This will also indicate
duce relevant absorption features only near certain absorption the relevance of each absorber in the final model. We choose
edges.Forinstance,ionssuchasOii-viandmoleculeslikewa- to show this source because of its highest statistics and large
terice, andraditeand pyroxenemostlyaffectthe spectralrange column density that maximize the significance of the spectral
near the O K-edge. Therefore, we first fit each edge with the features. Moreover,the results obtainedforGX 339-4are sim-
preliminarymodelandaddinsequenceionsand,inasecondin- ilar to those of most sources in our sample. We first jointly
stance,molecularcompoundslookingforsignificantchangesin fit the Fe L and O K edges (15− 24Å) with a simple model
the chi2.InthecaseoftheoxygenK-edge,significantchanges consisting of a power-law absorbed by neutral gas. Free pa-
intheχ2aregivenbyionslikeOii-iv,Ovii-viii,andcompounds rametersare the normalizationand slope of the power-law and
likewaterice,pyroxene,magnetite,andhematite.Oncewehave the Fe and O abundances of the neutral gas. This fit provides
C.Pintoetal.:ISMcompositionthroughX-rayspectroscopyofLMXBs 7
Table 2. GX 339-4 spectral fits within 15 − 24Å and relative
Fe L2
∆χ2. Fe L3 Neutral gas model
30
Neutral + ionized gas model
Parameter ∆χ2 Parameter ∆χ2 Parameter ∆χ2 A)−1 OVIIβ Gas + dust (complete) model
Oii 14 MetallicFe 101 CO 1 s −1
Oiii 163 H2Oice 168 FeO(OH) 1 m −2 20 OVIIIα
Oiv 20 MgSiO3 4(a) N2O 0 s OIII
n
Ov 1 Fe O 60 Fe O 0 o OII
3 4 2 3 ot
OOvviii 3661 Ca3FFe2eSAi3l2OO124 32 FeSO4 0 ux (ph 10 dust OVIIα OIV OI
Oviii 35 Mg Fe SiO 2 Fl
1.6 0.4 4
The∆χ2 valuesrefertothesequentialchangesobtainedbyaddingthe
parametersfromtoptobottomandfromlefttoright:Oii,Oiii,...,Oviii, dust
0
Metallic Fe, ..., FeSO4. (a) The ∆χ2 given by pyroxene MgSiO3 is 16 18 20 22 24
small because its features are similar to those produced by H2O ice Wavelength (Å)
(seealsoFig.7).
Fig.5.SpectralfitsoftheFeLandOKedgesforGX339-4(for
clarityonlyonestackedspectrumisshown).Formoredetailsee
χ2 = 1692/454 = 3.73,whichisclearlypoorandunacceptable Sect.4.2andTable2.
ν
(seealsothebluelineinFig.5).Adifferentcontinuumdoesnot
produce any change. Therefore,we add ions in sequence from 1
OiitoOviii,seeTable2.This(neutral+ionized)gasmodelwith
all these ions givesχ2 = 1092/447 = 2.44and a big improve-
mentto the fit. Howeνver,the fit is notyet acceptableand there 0.8 O III
are residuals at 17.5 and 22.9Å (see red line). Moreover, the
n
ptiuornelginaessmaonddeelddgoeessfnoortbroepthroidrounceanthdeorxaytigoebne.tDwueestnitshdeeafibnsioterply- ssio0.6 O IV
needed.WecompletetheISMmodelbyaddinginsequencedif- mi O VII
s
ferentmoleculeswiththeorderassuggestedbypreliminary fits an0.4 ANDRADITE O III−to−VII
toeachedge,whichismentionedabove:metalliciron,waterice, Tr MPYARGONXEETNITEE OO III gas
pyroxenesilicates,uptoironsulfate(seealsoTable2).Thefinal ICE ISM
χ2 dropsdowntoabout1.7asshownbythefinalmodel(purple 0.2
ν
line in Fig. 5). Althoughthis fit is notyet formallyacceptable,
the finalχ2 (significantlyhigherthan one)is mainlydue to the
ν
highstatistics.Similarχ2 valuesareobtainedforfeaturelessre- 0
ν 21 22 23 24
gionsatsimilarstatisticalqualityandindicatethefinalsystem-
Wavelength (Å)
aticuncertaintiesoftheinstrumentmodel.InFig.6weshowthe
transmissionof theISM best-fitting modelnearthe Oi edgeof Fig.6.TransmissionoftheISMneartheOK-edgeinthespec-
GX339-4.Wealsoreporttherelativecontributionsprovidedby trumofGX339-4.Themodelreferstothebest-fittingmodelas
Oi(redsolidline),Oii(purpledottedline),Oiii-iv(blackdashed giveninTable3.Thesolidredandblacklinesprovidethetrans-
line),as well as byspecific compounds.Molecularcompounds missionoftheneutralgasandtheentireISM,respectively(see
dominatethespectralrangewithin22.8–23.2Å andaffectthe alsoSect.4.2).
ratiobetweenthedepthsoftheOiedgeandresonanceline.
Onceweobtainagoodfitforthe15−24Åspectralrange,we
genfoundindust.WewilldiscusstheresultsinSect.5.Briefly,
fitthemodeltotheentirespectrum.Additionalfreeparameters
we have found a general agreement in the spectra of the dif-
in the model are the neon and magnesium abundances of the
ferent sources. Ions like Ov-vi and molecular compounds like
coldgas(hotmodelinSPEX)andthecolumndensitiesofNeii-
hematite(Fe O ),andCOarehardtodetect.Instead,Ovii-viii,
iii, Neviii-x,Mgxi-xii,andFexvii-xviii(whichareprovidedby 2 3
H Oices,andcompoundsofCa(andradite),andAl(Hercinite)
theslabmodel). 2
are detected along any LOS. The cold neutral gas provides on
averageOi∼ 1−3×1022m−2,whichmeansthatmostoxygen
4.3.Results is found in the neutral phase. We have summed all the contri-
butionstothemolecularoxygenaswellasalltheioniccolumn
This procedureused forGX 339-4is identically applied to the densitiesofthemildlyionized(Oii-v)andheavilyionized(Ovi-
othereighttargetsandtheresultsofourbest-fittingmodelsob- viii) gas, and comparethem in Table 4. As we have mentioned
tainedonalltheX-raysourcesareshowninTable3andplotted above,mostoftheoxygenisprovidedbytheneutralgasphase.
inFigs.2–4.Themodelfitsverywellboththeabsorptionedges Dustalsoseemstobeubiquitous.Thewarmandhotionizedgas
and lines. The reducedχ2 forthe ninesourcesare between 1.4 phases generally contribute less to oxygen, but there are some
and 1.7. We show the absolute abundances and hydrogen col- cases in which they constitute significant portions of the total
umn density for the cold gas, the column densities for all ions columndensity.SAXJ1808.4–3658isanexceptionalcasewith
and molecules that we were able to detect. In the cases of no detectable Ovi and prominent Ovii absorption lines (see also
detectionwereportthe2σupperlimits.Wealsoreportthefor- Fig.4,bottom-rightpanel).
mulaforeachcompoundandthetotalamountofironandoxy-
Table3.RGSfitswiththecompleteISMmodel.Thedustandthethreegasphasesarerepresentedbythefourmainblocks. 8
Parameter 4U1254–690 4U1636–536 4U1735–444 AqlX–1 GS1826–238 GX339–4 GX9+9 SAXJ1808.4–3658 SerX-1
N (1025m−2) 2.69±0.03 3.58±0.07 3.28±0.08 5.21±0.05 4.14±0.06 5.1±0.4 2.15±0.05 1.40±0.03 5.0±0.3
H
O/H(a) 0.92±0.04 0.85±0.10 0.93±0.03 0.88±0.02 0.92±0.04 1.06±0.05 0.99±0.03 0.91±0.03 0.9±0.1
Ne/H(a) 0.91±0.03 1.31±0.05 1.00±0.05 0.95±0.07 1.23±0.05 1.05±0.05 1.35±0.07 1.37±0.05 1.20±0.05
Mg/H(a) 1.7±0.3 <0.8 2.0±0.3 1.0±0.2 1.4±0.2 1.2 ± 0.3 0.4±0.5 <0.2 0.9±0.5
1.0 1.0 0.8 0.6 0.8 0.3 0.1
Fe/H(a) 0.16±0.05 0.18±0.04 0.25±0.10 0.11±0.05 0.22±0.06 0.4±0.1 0.19±0.05 <0.1 0.50±0.05
0.1
Oii(b) 0.33±0.15 1.4±0.7 0.6±0.2 0.2±0.1 0.6±0.1 1.4±0.5 0.3±0.1 1.67±0.05 2.0±0.2
Oiii(b) 0.08±0.08 <0.1 0.2±0.1 <0.3 <0.3 <0.8 <0.05 0.34±0.05 <0.4
0.04
Oiv(b) 0.5±0.2 0.3±0.2 0.4±0.2 0.4±0.1 0.6±0.2 0.5±0.1 0.4±0.2 0.31±0.03 0.7±0.5
C
Ov(b) <0.04 <0.02 <0.1 <0.03 <0.06 <0.01 <0.002 0.006±0.002 0.3±0.2 .
P
Neii(b) 0.11±0.06 1.2±0.3 0.4±0.2 0.7±0.2 1.4±0.3 1.2±0.5 0.4±0.2 1.3±0.1 0.3±0.2 in
to
Neiii(b) 0.19±0.08 0.9±0.2 0.5±0.2 0.6±0.2 0.6±0.3 0.9±0.2 0.2±0.1 1.1±0.1 0.8±0.4 e
t
Ovi(b) <0.03 <0.06 <0.01 <0.06 <0.07 <0.04 <0.004 0.02±0.01 <0.05 al.:
Ovii(b) 0.10±0.05 0.20±0.07 0.2±0.1 0.15±0.06 0.15±0.06 0.4±0.1 0.04±0.02 2.2±0.5 0.24±0.05 IS
M
Oviii(b) 0.09±0.04 0.19±0.05 0.3±0.1 0.10±0.04 0.10±0.03 0.3±0.1 <0.005 0.11±0.02 0.20±0.06
c
Neviii(b) <0.03 0.07±0.04 <0.05 <0.01 <0.01 0.06±0.04 <0.004 <0.003 <0.03 om
p
Neix(b) 0.05±0.03 0.09±0.04 0.2±0.1 0.07±0.05 0.06±0.02 0.16±0.04 0.03±0.02 0.05±0.02 0.14±0.05 o
s
Nex(b) <0.04 <0.03 <0.02 <0.1 <0.02 <0.02 <0.006 <0.008 <0.04 itio
n
Mgxi(b) 0.16±0.05 0.08±0.04 0.4±0.3 <0.05 <0.1 0.10±0.05 0.06±0.03 0.18±0.09 <0.15 th
Mgxii(b) <0.01 <0.01 <0.01 <0.02 <0.02 <0.01 <0.005 <0.01 <0.03 ro
u
Fexvii(b) <0.01 0.02±0.01 0.02±0.01 0.02±0.01 0.02±0.01 0.02±0.01 0.008±0.004 0.02±0.01 <0.01 gh
X
Fexviii(b) <0.02 <0.01 <0.01 0.02±0.01 <0.01 <0.01 <0.003 <0.04 <0.01 -
ra
WatericeH O (b) 2±1 0.9±0.5 3±2 3±1 3±2 4.4±2.5 2±1 0.3±0.7 <6.6 y
2 0.1 s
AndraditeCa3Fe2Si3O12 (b) 0.02±0.01 0.02±0.01 0.02±0.01 0.03±0.01 0.03±0.01 0.03±0.01 0.012±0.004 0.008±0.004 0.03±0.01 pec
HerciniteFeAl2O4 (b) 0.04±0.02 0.05±0.02 0.04±0.02 <0.025 0.06±0.02 0.07±0.02 0.03±0.01 0.015±0.007 0.07±0.02 tros
LepidocrociteFeO(OH) (b) 0.11±0.06 0.4±0.2 <0.2 0.6±0.3 0.4±0.2 0.03±0.02 0.2±0.1 0.15±0.08 0.15±0.10 co
p
PyroxeneMgSiO (b) <1.2 1.2±0.5 <1.1 <1.2 <1.1 0.3±0.7 <1.0 <0.4 1.7±0.5 y
3 0.1 o
CarbonmonoxideCO (b) <0.11 0.3±0.2 <0.2 <0.1 <0.3 0.07±0.05 <0.05 0.6±0.1 <0.2 fL
M
MetallicFe (b) 0.3±0.3 0.3±0.3 0.3±0.2 <0.7 <1.2 0.6±0.1 0.3±0.3 0.13±0.05 <0.7
0.1 0.1 0.1 0.1 X
MagnetiteFe O (b) 0.03±0.02 0.2±0.1 <0.04 0.3±0.2 0.3±0.1 0.2±0.1 <0.02 <0.005 0.2±0.1 B
3 4 s
OlivineMg Fe SiO (b) <0.04 0.06±0.03 <0.1 <0.04 <0.08 0.07±0.03 0.2±0.1 <0.03 <0.1
1.6 0.4 4
LaughinggasN O (b) 0.4±0.3 <0.04 <0.1 <0.03 <0.2 <0.03 <0.03 0.03±0.02 <0.04
2
HematiteFe O (b) <0.04 <0.01 <0.1 <0.02 <0.2 <0.003 <0.02 <0.01 <0.02
2 3
IronsulfateFeSO (b) <0.04 <0.01 <0.05 <0.03 <0.01 <0.004 <0.04 <0.03 <0.02
4
Totaloxygenindust (b) 3±1 7±2 4±1 6±1 6±1 7±2 4±1 1.5±0.5 7±2
Totalironindust (b) 0.6±0.2 1.3±0.5 0.5±0.3 1.7±0.5 1.5±0.6 1.4±0.5 0.7±0.2 0.3±0.1 0.8±0.4
(a)Abundancesratiosareinthelinearproto-SolarabundanceunitsofLodders&Palme(2009).
(b)Ionicandmolecularcolumndensitiesareinunitsof1021m−2(seealsoSect.4).
C.Pintoetal.:ISMcompositionthroughX-rayspectroscopyofLMXBs 9
4.4.Limitations
Therearesomefactorsthatmaylimitandaffecttheuniqueness
ofourbest-fittingmodels.Firstofall,thefeaturesproducedby
molecules in a certain absorption edge are similar and degen-
erate. Near the Fe L-edge the RGS absorption lines blend and
may not give unique solutions. Chandra spectra are provided
with higherresolution,buttheygiveaccurateresultsonly fora
fewsourcesevenbrighterthanours.Therefore,inourstandard
modeling we have simultaneously fitted all the molecules and
all the absorption edges and lines in order to obtain the best-
fittingmixtureandtobreakasmuchaspossiblethedegeneracy.
Unfortunately,forsomecompounds,suchasmagnetiteandpy-
roxene, we only have the transmission near the oxygen edge,
whilefortheotheredgeswe simplyusethe pureatomiccross-
section without absorption lines (for details consult the amol
model in the SPEX manual). Fortunately, these systematic ef-
fects do not strongly alter the depletion factors, especially for
oxygen. The Mg K-edge is very weak and it does not make Fig.7.Waterice /pyroxeneanti-correlation.Thepyroxenecol-
any difference whether we model it with atomic or molecular umn densities refer to the spectral fits of the Fe L and O K
cross-sections.Anotherproblemis the degeneracybetween the edges for GX 339-4 obtained with fixed amount of water ice.
absorption features produced by water ice and pyroxene (see The red values show the χ2 values of each fit. For more detail
Fig. 6 and 7 and Costantinietal. 2012, Appendix A), which seeSect.4.2and4.4.
arethemoleculesthatbestfittheOK-edge.Thisisalsoshown
by the large uncertaintieson their column densities in Table 3.
(which have higher effective area and resolution than RGS at
Therefore,inordertotesttherobustnessofourresultsandtoes-
timatethesystematiceffects, wehaveperformedadditionalfits shortwavelength)in combinationwith sourceswith higherhy-
withdifferentcombinationsofmoleculesbyexcludingsomeof drogencolumndensities.
them. One model includes only the molecules that do not lack Our estimates of column densities and abundancesprovide
interstellarpropertiesasintegratedalongtheLOSandonalarge
any relevant cross-section, such as water ice, Metallic Fe, CO,
scale.Thisisreasonableforthediffusegas,whichisalmostuni-
andHematite(alltheothermoleculessuchaspyroxeneandmag-
netiteareexcluded).Inanothertestwereplacewatericewithpy- formly distributed between 3–15kpc from the Galactic Center
and whose warm phase may reach a few kpc height from the
roxeneinordertomeasuresystematicchangesintheMgabun-
Galactic place (see e.g. Ferrie`re 2001). This scale-height is in-
danceof the cold gas.Severaltests like these providedus with
systematicchangesinthecolumndensitiesofthemoleculesand deed comparable to the average altitude of our sources (see
Table1).However,thisassumptionmaynotbecorrectfordust
the ions as well as in the abundancesof the cold gas. We have
andmolecules,whichfollowtheGalacticspiralsandareusually
summedallthesystematicandstatisticerrorsandreportedthem
togetherwiththebest-fittingvaluesinTable3.Interestingly,the foundatloweraltitudes.Thismeansthatdustsamplingtowards
our targets may be closer to the Sun rather than in the middle
totalamountofoxygenandironindustdoesnotstronglydepend
ofthepath.Inthecoldphase,ironandmagnesiumarestrongly
onthechosendustmixture.Therefore,theirdepletionfactorsare
nothighlyeffected,butthedetailedchemicalstructureis. depletedintodust.Therefore,theaveragelocationoftheFeand
Mgabsorberscanbefoundatshorterdistancesthanthoserefer-
Onanotherhand,themoleculardatabaseisnotcomplete.We
ringtooxygenandneon.However,inSect.5.2wewillshowthat
have tested all the molecules available in the SPEX database,
alongourlow-latitudeLOSthedifferentscale-heightsofgasand
which are about 40 (carbon oxides, ices, iron and magnesium
dustwillnotaffecttheresults.
silicates, and more complex molecules). Although these com-
Another problem in our grating spectra is the presence of
poundsareamongthemostabundantintheISM,wecouldstill
several bad pixels or dead columns that affect the estimates of
miss important contribution from other species like forsterite
someparameters.Oneofthemostrelevantislocatedat22.75Å,
(Mg SiO ,seee.g.Jones2000).OurISMteamatSRONiscur-
2 4 just near the Oiv resonance line (see Figs. 4 and 6). The Oiv
rentlytakingsynchrotronandelectroscopicmeasurementsofthe
column densities of our targets are all determined just through
cross-sectionsneartheO,Mg,andSiK-edges,andboththeFe
this line and their values may be inaccurate.New observations
K/L edges of several silicate compounds with a high spectral
taken with the multi-pointingRGS mode will surely provide a
resolutiontoo.Theselaboratorydatawillprovideaccurateesti-
solutiontothisproblem.
matesofmolecularcolumndensities.
The results obtained on the various absorption edges have
different weights. The oxygen edge is the deepest and the col-
5. Discussion
umndensitiesofoxygencompoundsareconstrainedbetterthan
foriron. The Mg edge is veryweak forhydrogencolumnden- Our analysis shows that the ISM has a multi-phase structure
sities like those of our sources and the estimates of magne- characterizedbycoldneutralgas(anddust),mildlyionizedgas
sium abundances and depletion factors are much more uncer- and heavily ionized gas. This complex structure is found to-
tain. This does not apply to neon as it is supposed to be only wardsanyofthestudiedsources(seeTable3)andisconsistent
inagaseousphase.Thisworkmostlyfocusesonmeasurements withourpreviouswork(seee.g.Kaastraetal.2009;Pintoetal.
ofinterstellaroxygen,iron,andneoncolumndensities.A deep 2010, 2012a; Costantinietal. 2012). Signatures of dust are al-
analysis of magnesium and silicon would require the use of wayspresentwithmaincontributionsbyicesandsilicates. The
both the XMM-Newton RGS and the Chandra HETG gratings oxygenedgeandlinesprovidethe mostaccurateresultsdueto
10 C.Pintoetal.:ISMcompositionthroughX-rayspectroscopyofLMXBs
their large depth, but the results are even more robust because 5.2.IstheISMchemicallyhomogeneous?
of the simultaneousmodelingof all the absorptionedges. This
was crucial especially for fitting dust and molecules which af- On small scales the ISM is notexpectedto be highlyhomoge-
fectbothO,Fe,andMgedges.Aninterestingandsimplecheck neous because there are significant physical and chemical dif-
ofdifferentamountsofoxygenindustisprovidedbythecom- ferencesbetweentheinterstellarenvironmentslikeHiiregions,
parisonoftheOK edgeinFig.4.AsweanticipatedinSect.3, PNenebulae,darkclouds,etc. However,thediffusemediumin
theOi1s–2pabsorptionlineisdeeperthantheoxygenedgefor ourGalaxymaybe homogeneouson largescales. Somechem-
SAXJ1808.4−3658,whileforSerX–1andGX339–4theyare ical and physical properties may be possibly consistent when
comparable. We attribute this to oxygen depleted from the gas comparingdistantregionswhichunderwentasimilarevolution.
phaseintodustgrains,whichisconfirmedbythemeasurement Thisispossiblefordustasmostofitisexpectedtogrowinthe
of a largeramountof oxygenin solids in the LOS towards the ISM rather than in stars (see e.g. Mattsson&Andersen 2012).
twolattersources(seeTable3). Itisthusworthtocompareitschemicalpropertiesasintegrated
alongthedifferentLOS.Firstofall,wenoticethatthemolecular
compositionissimilarinanydirection.Aluminatesandcalcium
5.1.Thenatureofthegasphases silicates arefoundtoprovidemostof AlandCa, whicharein-
deed expected to be highly depleted into dust. Water ice may
In our standard model we have used a collisionally-ionized
be present, but in an uncertain amount due to the degeneracy
model in SPEX because at low temperatures this mimics
betweenitsfeaturesandthoseofpyroxene.Jenkins(2009)sug-
well the cold interstellar gas. However, at these tempera-
gested water ice as a possible principal resource of oxygen in
tures photo-ionization could also provide a significant contri-
dust,buthealsoshowedthatatlow-intermediateN itishardto
bution. Therefore, we tried an alternative model in which a H
detect.Cross-sectionsoftheseandofadditionalcompoundsare
photo-ionized component (xabsin SPEX) substitutes the cold
currentlybeingmeasuredwitharesolutionhigherthanthecur-
collisionally-ionizedhotcomponent.Thefitsareequivalent,the
rentoneandmayprovidemoreaccurateresults(seeSect.4.4).
gas is almost neutral with a very low ionization parameter ξ.
To understandthe roleofdust, we havesummedallthe contri-
Abundancesandχ2areconsistentwithourstandardmodel.The
butions to oxygen from solids and molecules and compared it
maindifferenceisthatthefitwiththephoto-ionizedmodeltakes
withthehydrogencolumndensitythatwehaveobtainedinour
more time to converge and the absolute abundances, i.e. rela-
bestfits(seeFig.8).Weshowthetrendsofboththetotalcolumn
tive to hydrogen,have largeruncertainties(see also Pintoetal.
densityofdustyoxygenanditsfractionwithrespecttothetotal
2012a).Therefore,weprefertokeepourstandardmodelevenif
amountofneutraloxygen(gas+dust)asfunctionoftheN .The
photo-ionizationisnotruledout. H
oxygendust column correlates well with the hydrogencolumn
We have also consideredan alternative, physicalmodel for
the warm mildly-ionized gas that produces most of the Oii-iv density(slope= (1.3±0.2)×10−4)andthefractionofoxygen
andNeii-iiiabsorptionlines.Insteadofaslababsorberwehave boundin dustis constant(19±5%) along any LOS. These re-
sults suggestthaton averagethe coldISM phase is chemically
testedaphoto-ionizedxabscomponent,whichwassuccessfully
homogeneous.Wenoticethatthetotalamountofoxygenindust
used by Pintoetal. (2012a) in the deep UV and X-ray spectra
is solidly measured as it does not strongly depend on the dust
oftheAGNMrk509.Thefitwaspoorbecausea singlephoto-
mixture. For instance, we tested different molecular combina-
ionized component is not able to fit all the large column den-
tions by removing some of the main compounds from the fit,
sities as measuredby ourempiricalstandardmodeland shown
butintheendthesumofalltheoxygenmolecularcolumnden-
inTable3.Weneedalow-ionizationcomponentwhichprovides
thebulkofOiiandNeiitogetherwithanintermediate-ionization sities providethe same results in the fits. The same applies for
iron, whose dust fraction does not depend on the N and the
component for the other ions. Interestingly, a good fit is ob- H
LOS, but it has a larger uncertainty and spreads between 65–
tained when using the same abundance pattern as provided by
90%.ThisoccursbecausetheironLedgeisshallowerthanthe
thecoldneutralgas,whileproto-Solarabundancesgiveabadfit.
oxygenK edge.Aswementionedin Sect.4.4,thescale-height
However,athoroughstudyofthewarmgasrequireslongerex-
oftheinterstellardustissmallerthanforthegas,thus,duetothe
posuresandcomplementaryUVdata,whichisbeyondthescope
altitudesofthetargets,thedustisexpectedatshorterdistances
ofthispaper.
fromthe Sun. This meansthat the analysis of the dustin these
Theheavilyionizedgasisusuallyfoundtobeincollisional
different LOS probes properties of the ISM as measured on a
equilibrium.However,itisthoughtthatthishotgasintheGalaxy
largescale,butprobablysmallerthanwiththegas.Asadditional
is not characterized by a single phase, but it is constituted of
two main phases. Most Ovii-viii should arise from a very hot checkwehavecomputedthedust-to-gasratioforoursourcesas
theratiobetweenthetotalamountofoxygenindustandthehy-
(∼ 2 × 106K) collisionally-ionized phase, while the bulk of
Ovi is thought to be embedded in a cooling phase with tem- drogencolumndensity.The dust-to-gasratiois constant(slope
=0.0±0.2×10−4)atanyvalueofGalacticlatitudeasshownin
peratures1−5×105K (seee.g.Richter2006).Wehavetested
Fig. 9. This result and the trend seen in Fig. 8 suggest that the
bothsingle-phaseandtwo-phasemodelsbysubstitutingtheslab
componentresponsiblefortheOvi-viiiabsorptionwithoneand distributionofthedustshouldnotdeviatefromthegas.Thismay
be due to the low latitudes of our sources, which was also one
twocollisionally-ionizedhotcomponentsinSPEX.Thefitsare
ofthereasonsforustochoosethesetargetsasrepresentativesof
comparable due to the weakness of their absorption lines. We
wouldneedUVOvidatainordertoputstrongconstraints,but theGalactic(inner)disk.
most of our targets have not been observed with COS/HST or
FUSE. However, Table 3 shows a spread larger than 10 in the
5.3.ISMcolumndensities
Ovii / Oviii column ratio, which for a single gas component
would provide a very large scatter in temperature between the Ourempiricalmodelprovidesuswithestimatesofcolumnden-
nine LOS (from 1.2 to 2.6 million K). We think that combina- sities for several molecular and ionic species (see Table 3). In
tionsoftwostable,butdifferenthot(coolingandcoronal)phases somecaseswe couldonlyobtainupperlimits, butwe preferto
isamorelikelydescription. showthemastheymaybeusefulforfuturework.Inordertofur-