A non-equilibrium ionization model of the Local Bubble (I)

0.27 MB·English

by  Miguel A. de Avillez

Astronomy&Astrophysicsmanuscriptno.ms (cid:13)c ESO2012
January30,2012
LettertotheEditor
A non-equilibrium ionization model of the Local Bubble (I)
iv v vi
Tracing C , N , and O ions
MiguelA.deAvillez1 andDieterBreitschwerdt2
1 DepartmentofMathematics,UniversityofE´vora,R.RomaoRamalho59,7000E´vora,Portugal
e-mail:[email protected]
2 Zentrumfu¨rAstronomieundAstrophysik,TechnischeUniversita¨tBerlin,Hardenbergstrasse36,D-10623Berlin,Germany
e-mail:[email protected]
2 ReceivedMay2,2011;acceptedMonthDay,2011
1
0 ABSTRACT
2
Aims.Wepresentthefirsthigh-resolutionnon-equilibriumionizationsimulationofthejointevolutionoftheLocalBubble(LB)and
n
LoopIsuperbubblesintheturbulentsupernova-driveninterstellarmedium(ISM).Thetimevariationandspatialdistributionofthe
a Li-likeionsCiv,Nv,andOviinsidetheLBarestudiedindetail.
J
Methods.ThisworkusestheparalleladaptivemeshrefinementcodeEAF-PAMRcoupledtothenewlydevelopedatomicandmolec-

7 ular plasma emission module E(A+M)PEC, featuring the time-dependent calculation of the ionization structure of H through Fe,
2 usingthelatestrevisionofsolarabundances.ThefinestAMRresolutionis1pcwithinagridthatcoversarepresentativepatchofthe
Galacticdisk(withanareaof1kpc2inthemidplane)andhalo(extendingupto10kpcaboveandbelowthemidplane).

] Results.TheevolutionageoftheLBisderivedbythematchbetweenthesimulatedandobservedabsorptionfeaturesoftheLi-like
A ionsCiv,Nv,andOvi.ThemodeledLBcurrentevolutiontimeisbracketedbetween0.5and0.8Myrsincethelastsupernovareheated
G thecavityinordertohaveN(Ovi) <8×1012cm−2,log[N(Civ)/N(Ovi)]<−0.9andlog[N(Nv)/N(Ovi)]<−1insidethesimulated
LBcavity,asfoundinCopernicus,IUE,GHRS-ISTandFUSEobservations.
.
h
Keywords.hydrodynamics–shockwaves–ISM:general–ISM:bubbles–ISM:structure–ISM:kinematicsanddynamics
p
-
o
tr 1. Introduction TheseobservationspointtocolumndensitiesN(Civ) <7×1011,
s N(Nv) < 3 × 1012 cm−2, log[N(Civ)/N(Ovi)] (cid:46) −0.9 and
a
[ The ionization structure of the Local Bubble (LB), like that log[N(Nv)/N(Ovi)] (cid:46) −1.0. In that CIE model, the match be-
of any superbubble, should in principle reflect its evolutionary tween the simulated and observed column densities indicated
1
state for given boundary (and initial) conditions, if we assume thatthelastSNeventoccurred0.5Myrago.
v a multi-supernova/stellar wind origin (e.g, Innes & Hartquist However,significantdeviationsfromglobalCIEconditions
3
1984; Cox & Anderson 1982; Smith & Cox 2001). While are likely to occur (e.g., Kafatos 1973) as suggested by obser-
6
7 the ambient medium at the time of the first supernova (SN) vations of the LB in the EUV (Jelinsky et al. 1995; Hurwitz
5 explosion is unknown, the subsequent time-dependent energy et al. 2005) and X-ray (Sanders et al. 2001; McCammon et al.
. input rate is – quite surprisingly – much more tightly con- 2002;Freyberg&Breitschwerdt2003;Henleyetal.2007)bands
1
strained, although no early-type stars (or relics thereof) have and by non-equilibrium ionization (NEI) one-dimensional sim-
0
been found within the current LB volume. The kinematic anal- ulations of the LB (e.g., Cox & Anderson, 1982; Smith & Cox
2
1 ysis of nearby stars (Ma´ız-Apella´niz 2001), moving groups 2001).
: (Bergho¨fer & Breitschwerdt 2002; BB02), and the whole vol- The present work explores the degree to which non-
v umeof400pcindiametercenteredontheSunfromHipparcos equilibrium ionization (NEI) affects the conclusions of other-
i
X data(Fuchsetal.2006)consistentlypointto14−19starshaving wise similar CIE modeling. We extend our previous CIE work
explodedduringthepast15Myrontheirwaythroughtheregion
r (BA06andAB09)to(i)modeltheLBevolutionhydrodynami-
a nowoccupiedbytheLB. callyandtraceitsionizationstructure(andthatofthesurround-
Collisional ionization equilibrium (CIE) simulations of the ingISM)inatime-dependentfashion,(ii)studythespatialdis-
LBpoweredbythemissingstarsofthesubgroupB1ofPleiades tributionoftheLi-likeionsCiv,Nv,andOviandcomparethem
(BB02) within a turbulent interstellar medium (Breitschwerdt with observations, and (iii) constrain the LB age by matching
& de Avillez 2006 (BA06); de Avillez & Breitschwerdt 2009 simulatedandobservedline-of-sightdistributionsoftheseions.
(AB09)) reproduced the clumpy distribution of Ovi, as well as This letter is organized as follows: Section 2 summarizes the
itscolumndensityanddispersioninsidetheLBcavityobserved new model setup and simulation. In Section 3, the distribution
byCopernicus(Jenkins1978;Shelton&Cox1994)andFUSE ofLi-likeionsandtheircolumndensityratioswithintheLBare
(Oegerle et al. 2005; Savage & Lehner 2006), and the lack of discussed and compared to observations. The discussion of the
CivandNvionsinthebubble,whichisconsistentwiththeob- results(Section4)closesthepaper.
servationswithCopernicus(York1977),IUE(Bruhweileretal. We simulate the effects of the explosions of the stars of a
1980),andGHRS-HST(Bertinetal.1995;Huangetal.1995). moving subgroup, the living members of which now belong to
2 M.A.deAvillezandD.Breitschwerdt:ANEIModeloftheLocalBubble(I)-TracingCiv,NvandOviIons
theUCLandLCCassociations,astheirtrajectorieshavecrossed
theLBvolumetowardstheirpresentposition.
2. Modelandnumericalsetup
WesimulatethesimultaneousevolutionoftheLocalandLoopI
superbubbles. These are the result of the successive explosions
ofmassivestarsfrom,respectively,amovingsubgroupconsist-
ingof17starswithmasses∈[8.2,21.5] M ,whoselivingmem-
(cid:12)
bersnowbelongtotheUCLandLCCassociations,astheirtra-
jectorieshavecrossedtheLBvolumetowardstheirpresentposi-
tion(Fuchsetal.2006),andScoCenconsistingof39starswith
masses∈ [14,31] M (Egger1998).Theirtimeofexplosionis
(cid:12)
calculated by the Fuchs et al. (2006) formula. The simulation
setup is similar to that discussed in BA06 and AB09, but with
threemajordifferencesfromthesepreviousCIEmodels:Firstly,
the Time-dependent evolution of the ionization structure of H,
He, C, N, O, Ne, Mg, Si, S, and Fe ions with Asplund et al.
(2009)abundancesderivedusingtheEboraeAtomic+Molecular
PlasmaEmissionCode(E(A+M)PEC1;Avillezetal.2011).The
physical processes included in E(A+M)PEC are electron im-
pact ionization, inner-shell excitation auto-ionization, radiative
and dielectronic recombination, charge-exchange reactions (re-
combination with Hiand Heiand ionization with Hiiand Heii),
continuum (bremsstrahlung, free-bound, two-photon) and line
(permitted,semi-forbiddenandforbidden)emissionintherange
1Å-610µ,andmolecularlines.TheE(A+M)PECcodealsoin-
cludes inner shell photoionization, ionization of Hiby Lyman
continuum photons produced by the recombination of helium,
and molecular chemistry involving H, C, and O. The cooling
function (calculated at each cell of the grid at all time steps)
includesthese processes,whilethe spectracalculations include
lineandcontinuumemissions.Theseconddifferenceisthatlocal
self-gravityandheat-conduction(Balbus&Mckee1982;Dalton Fig.1.Ovidensitydistributions(cutsthroughtheGalacticmid-
& Balbus 1993) are calculated at every time step; Finally, the plane) in the LB centered at (x = 200,y = 450) pc and Loop I
coarse grid resolution is 8 pc, while the finest AMR resolution
(the adjacent bubble on the right) at 0.5 (top) and 0.8 (bottom)
is 1 pc - the highest so far used for LB evolution calculations.
MyrafterthelastSNintheLB,whichoccurredatevolutiontime
Periodicandoutflowboundaryconditionsaresetalongthever-
13.1 Myr. Both bubbles are surrounded with a thin fragmented
ticalfacesandboththetopandbottom(z=±10kpc)ofthegrid, Ovibearingshell.
respectively.
ity.Here,wediscusstheevolutionoftheLBinthefirstMyrafter
3. Results thelastSNexplosion.
TomodeltheLocalBubble,whichisassumedtobeembeddedin
arealisticbackgroundISM,thesimulationswererunforacon- 3.1. OvidistributionintheLB
siderabletimeinordertoestablishthefullGalacticfountainand
Figure1displaystimesnapshotsofthe Ovidensitydistribution
a global dynamical equilibrium with the density, pressure, and
withintheLBatanevolutiontime(sincethefirstSNintheLB)
temperaturedistributionsconditionssimilartothosedescribedin
oft =13.6,and13.9Myr,thatis∆τ∈{0.5,0.8}Myr,respec-
deAvillez&Breitschwerdt2004).AtanISMevolutiontimeof evol
250Myr,atrackersearchesforamolecularregionofsufficient tively,afterthelastSNeventwithintheLBcavity.TheOvihas
aclumpydistributionresultingfromlocalcoolingandturbulent
masstoformtheScoCenstars(adoptinga5%starformationef-
motions,drivenbyshearflowsinsidetheLBcavity,whosehigh-
ficiency).TheUCLandLCCclusterspathstowardsScoCenare
estvaluesoccurtowardstheboundary.
thensetusingthekinematicparametersderivedbyFuchsetal.
The spatial and temporal variations in Oviare explored by
(2006) with the last SN event in the LB cavity occurring 13.1
calculatingtheOvicolumndensity(N(Ovi))throughLOStaken
Myrafterthefirstexplosion(thatis,afteranISMevolutiontime
from the Sun at (x = 200,y = 450) pc, 100 pc from the inter-
of 263.1 Myr). The Local and Loop I bubbles evolve in time
action region between the Local and Loop I bubbles. The col-
interacting for most of the LB lifetime, and eventually the two
umn densities were calculated with a 5 pc step length out to a
bubblesmergewithLoopImaterialexpandingintotheLBcav-
maximum of 200 pc, spaced at 1 degree intervals to examine
1 The code is written in OpenCL and runs in AMD and NVIDIA thefull200pcradiussphereabouttheassumedSolarposition.
Graphics Processor Units (GPUs). A description of the software and Figure2displaysthemaximumN(Ovi)valuesoverallthelines
ionizationfractions,coolingandemissionspectratablescanbefound of sight at times 0.5 Myr through to 1, 1.5 and 2.0 Myr after
athttp://www.lca.uevora.pt/research.htmlSeealsoreferencestherein. the last SN explosion in the cavity. In addition, the figure also
M.A.deAvillezandD.Breitschwerdt:ANEIModeloftheLocalBubble(I)-TracingCiv,NvandOviIons 3
14 maximumoflogN(Ovi) =13.3(inunitsofcm−2)inthenext0.2
Myr.Fordistances>140pc,thelinesofsightaresamplinggas
notonlyfromLoopIbutalsofromthesurroundinginterstellar
medium, hence the increase in the column density. It is clear
that the modelled column densities fall within and follow the
13
observed(withFUSE)dataforpathlengthsgreaterthan50pc.
-2m] With increasing time, there is a slight increase in the amount
) [cVImax BWSBLL222000010608 olifneOsvaitattimsmesal1le.5rMdisytraanncdes2(.0dM<y5r0afptecr;tohreanlagsetSanNdedvaernktignrethene
O 12 ∆t = 0.5 Myr cavity).
g N( ∆tSSNN = 0.6 Myr For the first million years of evolution of the simulated
lo ∆tSN = 0.7 Myr bubbles, the majority of the LOS (> 75%) samples gas with
∆tSN = 0.8 Myr log(N(Ovi))<13,withthisnumberreaching99%and95%lev-
∆t = 0.9 Myr
SN els at 0.5 Myr and 0.7 Myr after the last explosion (top panel
11 ∆t = 1.0 Myr
SN of Figure 3) and decreasing to some 80% in the next 0.3 Myr
∆t = 1.5 Myr
SN (bottompanelofFigure3).
∆t = 2.0 Myr
SN
0 20 40 60 80 100 120 140 160 180 200
d [pc]
3.2. DistributionofLi-likeionratios
Fig.2.Maximum(top)modelOvicolumndensitiesina200pc
Figure 4 displays the histograms of the LOS ratios of
radius sphere centered at the Sun at times 0.5 through 1.0 Myr
log[N(Civ)/N(Ovi)] and log[N(Nv)/N(Ovi)] in the LB be-
after the last SN event in the LB, overlaid by FUSE data from
tween 0.5 Myr and 1.0 Myr after the last SN event. The
Savage & Lehner (2006; SL2006), Welsh & Lallement (2008;
fraction of LOS sampling gas with column density ratio
WL2008) and Barstow et al. (2010; BB2010) for observations
log[N(Civ)/N(Ovi)]<−0.9increaseswithtimetowards87%at
againswhitedwarfsintheupto200pcfromtheSun.Thehor-
0.6Myrand93%at0.7Myr(toppanel,leftcolumnofFigure4)
izontalaxisindicatestheintegrationdistanceforthemodeland
then decreases to less than 80% in the next 0.1 Myr of the LB
thestardistancefromtheobservationaldata.
evolution (bottom panel, left column of the figure). The small
fractionofLOSwithlog[N(Civ)/N(Ovi)] > −0.9isconsistent
40
∆tSN = 0.5 Myr with the detection by Welsh et al. (2010) of interstellar Civ in
∆tSN = 0.6 Myr theLBwithN(Civ) (cid:39)1012cm−2towardsastarlocatedat74pc
30 ∆tSN = 0.7 Myr from the Sun. For log[N(Nv)/N(Ovi)] < −1.0, the fraction of
%] LOSsamplinggasuptothisthresholdalsoincreasesto89%and
[S 92%levelsat0.6and0.7Myr,respectively,afterthelastexplo-
O
NL20 sionintheLB,followedbyasteepdecreasetolessthan70%in
/S
O thenextfewhundredthousandyears(rightcolumnofFigure4).
NL
d ThisbehaviourofthecolumndensityratiosinthesimulatedLB
10 confirmsthepresenceofaverysmallamountofCivandNvions
andtheiruncorrelateddistributionwithOvi.
0
40
∆t = 0.8 Myr
SN 4. Discussionandconclusions
∆t = 0.9 Myr
SN
30 ∆tSN = 1.0 Myr Theworkdescribedherefocusesonthedetailednumericalevo-
%] lution of a multi-supernova origin of the LB, following the
[OS studies by BA06 and AB09. A major difference from previ-
/NSL20 ous studies by other authors is that we (i) follow in a Eulerian
NLO fashion the joint evolution of the Local and Loop I bubbles
d in the turbulent ISM, (ii) use a three-dimensional high resolu-
10 tion supernova-driven ISM model to simulate a realistic ambi-
ent medium, (iii) calculate the time-dependent evolution of the
completeionizationstructure,andthereforealsothecorrespond-
0
11.6 11.8 12 12.2 12.4 12.6 12.8 13 13.2 13.4 ing time-dependent cooling function at each cell of the compu-
log N(OVI) [cm-2] tational grid, and (iv) include the latest revision of solar abun-
Fig.3. Histograms of the N(Ovi)model predictions for LOS dances.
samplingonlygasinsidethecavity,i.e.,LOSlengths< 100pc, Threeconditionsdeterminethepresentevolutiontime(tevol)
between0.5and1.0MyrafterthelastexplosionintheLBcavity. oftheLB:theN(Ovi) thresholdof8×1012 cm−2 andtheratios
log[n(Civ)/n(Ovi) < −0.9 and log[N(Nv)/N(Ovi)] < −1.0. A
comparisonbetweentheLOSmeasurementsatdifferenttimesin
showsFUSEdataagainstwhitedwarfs(Savage&Lehner2006 thesimulatedLBindicatesthatt canbeconstrainedbetween
evol
(SL2006); Welsh & Lallement 2008 (WL2008); Barstow et al. 0.5Myrand0.8MyrsincethelastSNexplosion.Theothertimes
2010(BB2010)). areexcludedbecausetheydonotsatisfysimultaneouslythethree
The maximum N(Ovi)increases with both path length and threshold conditions, therefore, contradicting Copernicus, IUE,
time,beinglowerthan1013cm−2insidetheLBuptoaradiusof and GHRS-HST observations. It should be kept in mind, how-
100pcinthefirst0.5-0.6MyrsincethelastSNandgrowingtoa ever,thattheageoftheLB,solelybasedonthesimulatedOvi,
4 M.A.deAvillezandD.Breitschwerdt:ANEIModeloftheLocalBubble(I)-TracingCiv,NvandOviIons
50 50
∆t = 0.5 Myr ∆t = 0.5 Myr
SN SN
∆t = 0.6 Myr ∆t = 0.6 Myr
40 ∆tSN = 0.7 Myr 40 ∆tSN = 0.7 Myr
SN SN
%] %]
[S30 [S30
O O
NL NL
/S /S
O20 O20
NL NL
d d
10 10
0 0
50 50
∆t = 0.8 Myr ∆t = 0.8 Myr
SN SN
∆t = 0.9 Myr ∆t = 0.9 Myr
40 ∆tSN = 1.0 Myr 40 ∆tSN = 1.0 Myr
SN SN
%] %]
[S30 [S30
O O
NL NL
/S /S
O20 O20
NL NL
d d
10 10
0 0
-2 -1.8 -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 -2 -1.8 -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2
log [N(CIV)/N(OVI)] log [N(NV)/N(OVI)]
Fig.4.HistogramoftheLOSmeasurementsofthecolumndensityratiosN(Civ)/N(Ovi)(leftpanel)andN(Nv)/N(Ovi)(rightpanel)
insidethesimulatedLB(l <100pc)for0.5-1.0MyrafterthelastSNexplosion.Theverticaldottedlinemarksthecriticalcolumn
LOS
densityratiosof−0.9(left)and−1.0(right),respectively(fordetailsseetext).
isbetween0.5Myrand1.1MyrsincethelastSNevent.Thede- References
rivedLBpresentagecorrespondstoadelayofatmost0.3Myr
(0.6 Myr if we only consider the Ovidistribution) compared to Asplund,M.,Grevesse,N.,&Sauval,A.J.,&Scott,P.,2009,ARA&A47,481
Barstow,M.A.,Boyce,D.D.,Welsh,B.Y.,Lallement,R.,Barstow,J.K.,Forbes,
the0.5MyrobtainedintheCIEsimulationsofBA06andAB09. A.E.,&Preval,S.,2010,ApJ,723,1762
Breitschwerdt,D.,Feige,J.,Dettbarn,C.,Schulreich,M.,deAvillez,M.A.,&
The comparison of Li-like ions between observations and
“Fuchs,B.,2012,Astronomy&Astrophysics,inpreparation
NEIsimulationspinsdownforthefirsttimetheageoftheLBre- Breitschwerdt,D.,&deAvillez,M.A.,2006,A&A452,L1(BA06)
liablywell,whichwasimpossibleinpreviousCIEsimulations. Bergho¨fer,T.,&Breitschwerdt,D.,2002,A&A390,299
Inparticular,modellingthedensitiesofdifferentionswithdiffer- Bertin,P.,Vidal-Madjar,A.,Lallement,R.,Ferlet,R.,&Lemoine,M.,1995,
A&A302,889
ent ionization and recombination timescales gives abundance-
Bruhweiler,F.C.,Kondo,Y.,&McCluskey,G.E.,1980,ApJ237,19
independent results. Moreover, since spectral resolution in our
Cox,D.P.,&Anderson,P.R.,1982,ApJ253,268
modelsisnotanissue,wewillbeableinthenearfuturetoalso deAvillez,M.A.,&Breitschwerdt,D.,2004,A&A425,899
comparetheoreticallinewidthstoobservations. deAvillez,M.A.,&Breitschwerdt,D.,2009,ApJ697,L158(AB09)
de Avillez, M.A., Spitoni, E., & Breitschwerdt, D., 2011, in ”Advances in
The small column densities observed over a wide range of ComputationalAstrophysics:Methods,ToolsandOutcomes”,R.Dolcetta,
temperaturesinthesimulationsaretheresultofdelayedrecom- M.Limongi,A.Tornambe(Editors),ASPConf.Series.,inPress.
bination, which also manifests itself in X-ray emission at low Egger,R.1998,InIAUColloq.166,LectureNotesinPhysics506,287
Freyberg,M.,&Breitschwerdt,D.,2003,AN324,162
temperatures,aswewilldiscussinthenextfewpapersofthese-
Fuchs,B.,Breitschwerdt,D.,deAvillez,M.A.,Dettbarn,C.,&Flynn,C.,2006,
ries.Finally,wenotethatwehaveperformedcalculationsofthe MNRAS373,993
distributionof60Feinthesolarneighbourhoodandcomparedthe Henley,D.B.,Shelton,R.L.,&Kuntz,K.D.,2007,ApJ661,304
sequenceofexplosionswiththemeasuredironpeakinthedeep Huang,J.-S.,Songaila,A.,Cowie,L.L.,&Jenkins,E.B.,1995,ApJ450,163
Hurwitz,M.,Sasseen,T.P.,&Sirk,M.M.,2005,ApJ623,911
sea ferromanganese crust (Knie et al. 2004), including the new
Innes,D.E.,&Hartquist,T.W.,1984,MNRAS209,7
half lifetime of Rugel et al. (2009), and found excellent agree-
Jelinsky,P.,Vallerga,J.V.,&Edelstein,J.,1995,ApJ441,653
mentinthetimesequences(Breitschwerdtetal.2012).Further Jenkins,E.B.,1978,ApJ219,845
developmentsofthemodelcomprisingtheinclusionofmagnetic Kafatos,M.,1973,ApJ182,433
fieldandcosmicrayswillalsobeexploredinpaperstofollow. Knie,K.,Korschinek,G.,Faestermann,T.,Dorfi,E.A.,Rugel,G.,Wallner,A.,
2004,Phys.Rev.Lett.,93,171103
Ma´ız-Apella´niz,J.,2001,ApJ560,L83
McCammon,D.,Almy,R.,Apodaca,E.,Bergmann-Tiest,W.,etal.,2002,ApJ
576,188
Acknowledgements. We would like to thank the referee, Don Cox, for his Rugel,G.Faestermann,T.,Knie,K.,Korschinek,G.,Poutivtsev,M.,Schumann,
detailed reports, constructive criticism, and valuable suggestions that allowed D.,Kivel, N.,Gnther-Leopold, I.,Weinreich,R., Wohlmuther,M., 2009,
us to improve this letter significantly. The simulations were carried out at Phys.Rev.Lett.,103,072502
theMilipeiaSupercomputer(CenterforComputationalPhysics,Universityof Sanders,W.T.,Edgar,R.J.,Kraushaar,W.L.,McCammon,D.,&Morgenthaler,
Coimbra) and at the ISM-Cluster of the Computational Astrophysics Group J.P.,2001,ApJ554,694
(Dept.ofMathematics,UniversityofE´vora).Thisresearchissupportedbythe Savage,B.D.,&Lehner,N.,2006,ApJS162,134
FCT(Portugal)projectPTDC/CTE-AST/70877/2006. Schmutzler,T.,&Tscharnuter,W.M.,1993,A&A273,318
M.A.deAvillezandD.Breitschwerdt:ANEIModeloftheLocalBubble(I)-TracingCiv,NvandOviIons 5
Shelton,R.,&Cox,D.P.,1994,ApJ434,599
Smith,R.K.,&Cox,D.P.,2001,ApJS134,283
Welsh,B.Y.,&Lallement,R.,2008,A&A490,707
Welsh,B.Y.,Wheatley,J.,Siegmund,O.H.W.,&Lallement,R.,2010,ApJ712,
L199
York,D.G.,1977,ApJ213,43