Stellar laboratories

28 Pages·2016·0.7 MB·English

A&A587,A39(2016) Astronomy
DOI:10.1051/0004-6361/201527324 &
(cid:2)c ESO2016 Astrophysics
Stellar laboratories
iv vii
VI. New Mo – oscillator strengths and the molybdenum abundance
− − (cid:2),(cid:2)(cid:2),(cid:2)(cid:2)(cid:2)
in the hot white dwarfs G191 B2B and RE0503 289
T.Rauch1,P.Quinet2,3,D.Hoyer1,K.Werner1,M.Demleitner4,andJ.W.Kruk5
1 InstituteforAstronomyandAstrophysics,KeplerCenterforAstroandParticlePhysics,EberhardKarlsUniversity,Sand1,
72076Tübingen,Germany
e-mail:[email protected]
2 PhysiqueAtomiqueetAstrophysique,UniversitédeMons–UMONS,7000Mons,Belgium
3 IPNAS,UniversitédeLiège,SartTilman,4000Liège,Belgium
4 AstronomischesRechen-Institut,ZentrumfürAstronomie,RuprechtKarlsUniversity,Mönchhofstraße12-14,69120Heidelberg,
Germany
5 NASAGoddardSpaceFlightCenter,Greenbelt,MD20771,USA
Received8September2015/Accepted13December2015
ABSTRACT
Context.For the spectral analysis of high-resolution and high signal-to-noise (S/N) spectra of hot stars, state-of-the-art non-local
thermodynamicequilibrium(NLTE)modelatmospheresaremandatory.Thesearestronglydependentonthereliabilityoftheatomic
datathatisusedfortheircalculation.
Aims.Toidentifymolybdenumlinesintheultraviolet(UV)spectraoftheDA-typewhitedwarfG191−B2BandtheDO-typewhite
dwarfRE0503−289and,todeterminetheirphotosphericMoabundances,reliableMoiv–viioscillatorstrengthsareused.
Methods.WenewlycalculatedMoiv–viioscillatorstrengthstoconsidertheirradiativeandcollisionalbound-bound transitionsin
detailinourNLTEstellar-atmospheremodelsfortheanalysisofMolinesexhibitedinhigh-resolutionandhighS/NUVobservations
ofRE0503−289.
Results.Weidentified12Movand9MovilinesintheUVspectrumofRE0503−289andmeasuredaphotosphericMoabundance
of1.2−3.0×10−4(massfraction,22500−56400timesthesolarabundance).Inaddition,fromtheAsvandSnivresonancelines,we
measuredmassfractionsofarsenic(0.5−1.3×10−5,about300−1200timessolar)andtin(1.3−3.2×10−4,about14300−35200times
solar).ForG191−B2B,upper limitsweredeterminedfortheabundances ofMo(5.3×10−7,100timessolar)and, inaddition,for
Kr(1.1×10−6,10timessolar)andXe(1.7×10−7,10timessolar).Thearsenicabundancewasdetermined(2.3−5.9 × 10−7,about
21−53timessolar).Anew,registeredGermanAstrophysicalVirtualObservatory(GAVO)service,TOSS,hasbeenconstructedto
provideweightedoscillatorstrengthsandtransitionprobabilities.
Conclusions.Reliablemeasurements andcalculationsofatomicdataareaprerequisiteforstellar-atmospheremodeling.Observed
Mov-vilineprofilesintheUVspectrumofthewhitedwarfRE0503−289werewellreproducedwithournewlycalculatedoscillator
strengths.Forthefirsttime,thisallowedthephotosphericMoabundanceinawhitedwarftobedetermined.
Keywords.atomicdata–line:identification–stars:abundances–stars:individual:G191-B2B–stars:individual:RE0503-289–
virtualobservatorytools
1. Introduction iv-vii. While Werner et al. (2012b)couldmeasure onlythe Kr
andXeabundances,furtherabundancedeterminations(Zn,Ge,
RE0503−289 (WD0501−289, McCook & Sion 1999a,b) is a
Ga,Xe,andBabyRauchetal.2014a,2012,2014b,2015b,a,re-
hot, helium-rich, DO-type white dwarf (WD, effective tem-
spectively)werealwaysinitiatedbynewcalculationsofreliable
perature Teff = 70000K, surface gravity log(g/cm/s2) = 7.5, transitionprobabilities.
Dreizler&Werner1996),thatexhibitslinesofatleasttentrans-
G191−B2B(WD0501+527,McCook&Sion1999a,b)isa
ironelementsinitsfar-ultraviolet(FUV)spectrum(Werneretal.
hot,hydrogen-rich,DA-typewhitedwarfthatwasrecentlyana-
2012b).Theabundanceanalysisofthesespeciesishamperedby
the lack of atomic data for their higher ionization stages, i.e., lyzedbyRauchetal.(2013,Teff = 60000K,logg=7.6).Based
onthismodel,Rauchetal.(2014a,b,2015b)measuredtheabun-
(cid:2) Based on observations with the NASA/ESA Hubble Space dancesofZn,Ba,andGa.
Telescope,obtainedattheSpaceTelescopeScienceInstitute,whichis Molybdenumis anothertrans-ironelement(atomic number
operatedbytheAssociationofUniversitiesforResearchinAstronomy, Z = 42). It was discovered for the first time in a WD (in the
Inc.,underNASAcontractNAS5-26666. spectrumofRE0503−289)byWerneretal.(2012b,fourMovi
(cid:2)(cid:2) Based on observations made with the NASA-CNES-CSA Far
lines). To identifymorelines ofMo and to determineits abun-
UltravioletSpectroscopicExplorer. dance,wecalculatednewtransitionprobabilitiesforMoiv–vii.
(cid:2)(cid:2)(cid:2) TablesA.10–A.13areonlyavailableviatheGermanAstrophysical
Virtual Observatory (GAVO) service TOSS(http://dc.g-vo.org/ In this paper, we first describe the available observa-
TOSS). tions (Sect.2), our stellar-atmospheremodels (Sect.3), and the
ArticlepublishedbyEDPSciences A39,page1of28
A&A587,A39(2016)
computationofthenewtransitionprobabilities(Sect.4).Anew Table1.StatisticsofMoiv-viiatomiclevelsandlinetransitionsfrom
VirtualObservatory(VO)servicethatprovidesaccesstotransi- TablesA.10–A.13,respectively.
tionprobabilitiesispresentedinSect.5.
To use the most elaborated models of G191−B2B and Ion Atomiclevels Lines Superlevels Superlines
RE0503−289 for our Mo abundance analysis, we start with iv
162 2803 7 15
an incorporation and an abundance determination of arsenic v
257 5882 7 22
(Sect.6.1, both stars) and tin (Sect.6.2, RE0503−289). Then, vi
112 988 7 23
weassesstheMophotosphericabundancesinRE0503−289and vii 95 1181 7 16
G191−B2B (Sect.6.3). In Sect.6.5, we determine upper abun-
626 10824 28 76
dancelimitsofkryptonandxenoninG191−B2B.Tounderstand
theabundancepatternsoftrans-ironelements,weinvestigateon
the efficiencyof radiativelevitation acting on the elementsZn,
Ga,Ge,As,Kr,Mo,Sn,Xe,andBainbothstars’atmospheres tothequadraticStarkeffectiscalculatedusingapproximatefor-
(Sect.7).WesummarizeourresultsandconcludeinSect.8. mulaebyCowley(1970,1971).
2. Observations 4. Atomicstructureandradiativedatacalculation
G191−B2B. We used the spectra obtained with the Far New calculations of oscillator strengths for a large number
Ultraviolet Spectroscopic Explorer (FUSE, 910Å < λ < of transitions of molybdenum ions that are considered in the
1190Å, resolving power R = λ/Δλ ≈ 20000, for details present work were carried out using the pseudo-relativistic
see Rauch et al. 2013) and the Hubble Space Telescope/Space Hartree-Fock (HFR) method (Cowan 1981), including core-
Telescope Imaging Spectrograph (HST/STIS, 1145 Å < λ < polarizationcorrections(see,e.g.,Quinetetal.1999,2002).
3145Å,resolutionof≈3kms−1,seeRauchetal.20131). For Moiv, the configuration interaction was considered
among the configurations 4d3, 4d25s, 4d26s, 4d25d, 4d26d,
4d4f2,4d5s2,4d5p2,4d5d2,4d5s5d,4d5p4f,4d5p5f,and4d4f5f
RE0503−289. We analyzed its FUSE (described in detail by for the even parity and 4d25p, 4d26p, 4d24f, 4d25f, 4d5s5p,
Werner et al. 2012b) and HST/STIS observations (1144Å < 4d5s4f, 4d5s5f, 4d5p5d, 4d4f5d, and 4d5d5f for the odd par-
λ < 3073Å). The latter was co-added from two observa- ity. The core-polarization parameters were the dipole polariz-
tions with grating E140M (exposure times 2493s and 3001s, ability of a Movii ionic core reported by Fraga et al. (1976),
1144−1709Å, R ≈ 45800), and two observations with grat- i.e., α = 1.82au, and the cut-off radius corresponding to the
d
ing E230M (1338s, 1690−2366Å and 1338s, 2277−3073Å, HFR mean value (cid:4)r(cid:5) of the outermost core orbital (4p), i.e.,
R ≈ 30000).These STIS observationsare retrievablefrom the r = 1.20au. Using the experimentalenergylevelsreportedby
c
BarbaraA.MikulskiArchiveforSpaceTelescopes(MAST). Sugar & Musgrove (1988)and Cabeza et al. (1989), the radial
integrals(averageenergy,Slater,spin-orbitandeffectiveinterac-
tionparameters)of4d3,4d25s,4d26s,4d25d,and4d25pconfig-
3. Modelatmospheresandatomicdata urationswereadjustedbyawell-establishedleast-squaresfitting
process that minimizes the differences between computed and
WeemployedtheTübingenNLTE2ModelAtmospherePackage
experimentalenergies.
(TMAP3,Werneretal.2003,2012a)tocalculateplane-parallel, v
ForMo ,theconfigurationsretainedintheHFRmodelwere
chemically homogeneous model-atmospheres in hydrostatic
4d2,4d5s,4d6s,4d5d,4d6d,4d5g,5s2,5p2,5d2,4f2,5s5d,5s6s,
and radiative equilibrium. Model atoms were taken from the
5p4f,5p5f,4p54d24f,4p54d25f,and4p54d25pfortheevenpar-
Tübingen Model Atom Database (TMAD4, Rauch & Deetjen
ity and 4d5p, 4d6p, 4d4f, 4d5f, 4d6f, 4d7f, 4d8f, 4d9f, 5s5p,
2003)thathasbeenconstructedaspartoftheTübingencontribu-
5p5d,5s4f,5s5f,4f5d,4p54d3, 4p54d25s, and4p54d25dforthe
tiontotheGermanAstrophysicalVirtualObservatory(GAVO5).
oddparity.Inthision,thesemi-empiricalprocesswasperformed
For our Mo model atoms, we follow Rauch et al. (2015b) and
to optimize the radial integrals corresponding to 4d2, 4d5s,
usedastatisticalapproachtocalculateso-calledsuperlevelsand
4d6s,4d5d,4d5g,5s2,5p2,5s5d,5s6s,4d5p,4d6p,4d4f,4d5f,
superlineswithourIronOpacityandInterface(IrOnIc6,Rauch
4d6f, 4d7f, 4d8f, 4d9f, 5s5p, 5s4f, and 4p54d3 configurations,
&Deetjen2003).WetransferredournewModataintoKurucz-
whichusetheexperimentalenergylevelsreportedbyReader&
formattedfiles7thatweretheningestedandprocessedbyIrOnIc.
Tauheed(2015).Core-polarizationeffectswereestimatedusing
ThestatisticsofourMomodelatomissummarizedinTable1. viii
a dipole polarizability value corresponding to a Mo ionic
ForMoandallotherspecies,leveldissolution(pressureion-
core taken from Fraga et al. (1976), i.e., α = 1.48au, and a
ization)followingHummer&Mihalas(1988)andHubenyetal. d
cut-offradiusequalto1.20au.
(1994)isaccountedfor.BroadeningforallMolinesthataredue vi
In the case of Mo , the 4d, 5d, 6d, 7d, 8d, 5s, 6s, 7s, 8s,
1 Available at http://www.stsci.edu/hst/observatory/cdbs/ 5g,6g,7g,8g,7i,8i,4p54d5p,4p54d4f,and4p54d5fevencon-
calspec.html figurationsandthe5p,6p,7p,8p,9p,10p,11p,4f,5f,6f,7f,8f,
2 Non-localthermodynamicequilibrium. 9f,6h,7h,8h,8k,4p54d2,4p54d5s,and4p54d5doddconfigura-
3 http://astro.uni-tuebingen.de/~TMAP tionswereexplicitlyincludedinthe HFRmodelwith thesame
v
4 http://astro.uni-tuebingen.de/~TMAD core-polarizationparametersasthoseconsideredforMo .The
5 http://www.g-vo.org semi-empiricaloptimizationprocesswascarriedouttofitthera-
6 http://astro.uni-tuebingen.de/~TIRO dialparametersinthend(n=4−8),ns(n=5−8),ng(n=5−8),
7 GFxxyy.GAM, GFxxyy.LIN,andGFxxyy.POSfileswithxx=ele- ni (n = 7−8), np (n = 5−11), nf (n = 4−9), nh (n = 6−8),
mentnumber,yy=elementcharge,http://kurucz.harvard.edu/ 8k, 4p54d2, and 4p54d5sconfigurationsusing the experimental
atoms.html energylevelspublishedbyReader(2010).
A39,page2of28
T.Rauchetal.:Stellarlaboratories:newMoiv–viioscillatorstrengths
vii
Finally, forMo , the HFR multiconfigurationexpansions
included the 4p6, 4p55p, 4p56p, 4p54f, 4p55f, 4p56f, 4s4p64d,
4s4p65d, 4s4p66d, 4s4p65s, 4s4p66s, 4p44d2, 4p44d5s, and 0
4p45s2 evenconfigurationsandthe4p54d,4p55d,4p56d,4p55s, er
d
4p56s, 4p57s, 4p58s, 4p59s, 4p510s, 4p55g, 4p56g, 4s4p65p, ea
4s4p66p,4s4p64f,4s4p65f,4s4p66f,4p44d5p,and4p44d4fodd gfR-2
configurations.Here,becausesomeconfigurationswithopen4s og
and 4p orbitals were explicitly includedin the physicalmodel, l
-4
the core-polarization effects were estimated by considering a
Moxv ionic core with the corresponding dipole polarizability Mo V Mo VI
value taken from Johnson et al. (1983), i.e., α = 0.058au,
and a cut-off radius equal to 0.41 a.u, which cdorresponds to -4 -2 0 -4 -2 0
loggf
the HFR mean value (cid:4)r(cid:5) of the 3d subshell. The fitting pro-
cess was then carried out with the experimental energy levels Fig.1. Comparison of our weighted oscillator strengths to those of
v
classifiedbySugar&Musgrove(1988)andShiraietal.(2000) Reader & Tauheed (2015) for Mo (left) and of Reader (2010) for
vi
toadjusttheradialparametersthatcharacterizethe4p6,4p55p, Mo (right). The larger, red symbols refer to the lines identified in
4p54f,4p55f,4s4p64d,4p44d2,4p54d,4p55d,4p5ns(n=5−10), RE0503−289(TableA.9).
4p55g, and 4s4p65p configurations. The parameters adopted
in our computations are summarized in Tables A.1−A.4 and FUSE FUSE
computedandavailable experimentalenergiesare comparedin 2 STIS STIS
TablesA.5−A.8,forMoiv–vii,respectively. )er
d
Tables A.10–A.13 give the weighted HFR oscillator ea
strengths (loggf) and transition probabilities (gA, in s−1) for gfR 0
Moiv–vii, respectively, and the numerical values (in cm−1) of gf/
lower and upper energy levels and the corresponding wave- (
g-2
lengths (in Å). In the last column of each table, we also give o
l
the absolute value of the cancellation factor CF, as defined by
Mo V Mo VI
Cowan(1981).Wenotethatverylowvaluesofthisfactor(typ-
ically<0.05)indicatestrongcancellationeffectsinthe calcula- 2.5 3.0 2.5 3.0
tion of line strengths. In these cases, the correspondinggf and log λ/Ao
gA values could be very inaccurate and, therefore, need to be
Fig.2.RatioofourweightedoscillatorstrengthsandthoseofReader&
consideredwithsomecare.However,veryfewofthetransitions v vi
Tauheed(2015)forMo (left)andofReader(2010)forMo (right).
that appear in Tables A.10–A.13 are affected. These tables are ThewavelengthrangesofourFUSEandHST/STISspectraaremarked.
provided via the newly developed GAVO Tübingen Oscillator The larger, red symbols refer to the lines identified in RE0503−289
StrengthsServiceTOSS8thatisbrieflydescribedinSect.5. (TableA.9).
Oscillator strengths were published by Reader & Tauheed
v
(2015)for923linesofMo andbyReader(2010)for245lines
vi
of Mo .Wecomparedloggf valuesandwavelengthsofthose whichallowsconventionalqueriesbywavelength,element,ion-
lineswhosepositionsagreewiththoseofourlineswithinΔλ ≤ isationstage,etc.,exportingtovarioustabularformatsandalso
0.02Å(theseare921linesof Mov,178linesofMovi)(Figs.1 directlyintouserprogramsviaSAMP9.
and 2). In general, we find a rather good agreement, although TheSLAPinterfacecanbeusedfromspecializedprograms
ournewloggf-valuesseemtobe,onaverage,smallerthanthose (“SLAP clients”) like VOSpec (Osuna et al. 2005);this is nor-
previouslypublished.Thiscanbeexplainedbythefactthatour mallytransparenttotheuser;inaserverselector,theservicewill
calculationsexplicitlyincludealargersetofinteractingconfig- typicallyappearunderitsshortnameTOSSSLAP.
urations, in particular with an open 4p subshell, as well as a The TAP interface allows database queries against the line
pseudo-potentialmodeling of the remaining core-valence elec- database, including comparisonswith user-provideddata (“up-
troniccorrelations.FortheMovandMovilinesthatwereiden- loads”). In the TAP dialogs of applications like TOPCAT, the
tifiedinRE0503−289(TableA.9,Fig.12)andwereusedforthe user should look for “GAVO DC” or manually enter the ac-
abundancedetermination,theloggf-valuesarealmostidentical. cessURL10. Asthe discussionof TAP’spossibilities isbeyond
the scopeof this paper,see on-lineexamplesforTOSSservice
usage11.
5. TheGAVOserviceTOSS
In the frameworkof the GAVO project,we developedthe new, 6. Photosphericabundances
registeredVO service TOSS.It is designedto provideeasy ac-
To improve the simulation of the background opacity, we in-
cesstocalculatedoscillatorstrengthsandtransitionprobabilities
cludedAsinourcalculationsforG191−B2BandRE0503−289
ofanykindinVO-compliantformat.
and determined its abundance from these models (Sect.6.1).
LinedataisstoredintermsoftheSpectralLineDataModel Then,weincludedSnforRE0503−289(Sect.6.2)becauseour
(Osunaetal.2010)andisaccessiblethroughawebbrowserin-
terface, via the Simple Line Access Protocol SLAP (Salgado 9 The SimpleApplication Messaging Protocol isa VO-defined stan-
et al. 2010), and via the Table Access Protocol TAP (Dowler dard protocol facilitating seamless and fast data exchange on user
et al. 2010). The browser-based interface offers a web form, desktops.
10 http://dc.g-vo.org/tap
8 http://dc.g-vo.org/TOSS 11 http://dc.g-vo.org/toss/q/q/examples
A39,page3of28
A&A587,A39(2016)
n 0 VII VIII n 0 VII VIII
o o
acti VVI acti V VI
r r
n f IV n f IV
o o
ati -5 ati -5
z z
ni ni
o o
i i
g g
o o
l-10 l-10
-5 0 -5 0
log (m / g/cm2) log (m / g/cm2)
Fig.3. Arsenic ionization fractions in our RE0503−289 model. m is Fig.4.AsinFig.3,forG191−B2B.
the column mass, measured from the outer boundary of the model
atmosphere.
1 0
5 8
6 4
7. 9.
new HST/STIS observation provides access to Sniv lines that 98 102
are necessary for an abundance analysis. Based on these ex-
tended models, we considered Mo and determined its abun-
dancesforRE0503−289(Sect.6.3)andG191−B2B(Sect.6.4).
x
u
6In.1.thGe19F1U-BS2EBoabnsderRvaEti0o5n03o-f28R9E:A05rs0e3n−ic28(Z9,=W33er)ner et al. ative fl 987.651 029.480
v el 1
(2012b) discovered As λλ987.65,1029.48,1051.6,1056.7Å. r
v
For As , lifetimes were measured with the beam-foil tech-
nique (Pinnington et al. 1981). The multiplet f-value (0.78±
v
0.06) of the As resonance transition was determined with
the help of arbitrarily normalized decay curve (ANDC) analy-
ses, which were confirmed by calculations of Fischer (1977),
10%
Migdalek & Baylis (1979), and Curtis & Theodosiou (1989). 1Ao
Morton (2000) lists both components in his compilation,
λ987.65Å (4s2S −4p2P◦ , f = 0.528) and λ1029.48Å Δλ / Ao
1/2 3/2
(e4tsa2lS.1(/220−145p)2tPo◦1/d2e,tefrm=ine0.a2r5s3en).icTmheasseswfrearcetiounsse.dTbhyeyCfhoauynedr FRiEg.055.0S3e−c2ti8o9ns(booftotoumrFrUowS)EaorbosuenrdvaAtisonvsλo9f8G71.6951−ÅB(2leBft(ctooplurmown))aanndd
6.3 × 10−8(6timessolar)and1.6 × 10−5(1450timessolar)in Asvλ1029.48Å (right column). The thick red and thin green lines
G191−B2BandRE0503−289,respectively. show a comparison with theoretical spectra of two models, with and
Asvλλ1051.6,1056.7Å belong to the 4d2D−4f2F◦ mul- withoutAs,respectively.
tiplet (λλ1050.67,1051.64,1055.60,1056.58) but no oscillator
strengthshavebeencalculatedsofar.
atmWosephinecrelucdaeldcuthlaetTioMnsA.DAsAvsiivis-vthiieimdoomdeinlaantotmioinnionutrhmeolidneel-- ents5 4f 2Fo
forming region vin both stars (Figs.3 and 4). We reproduced effici4 5s2S
tFhUeSoEbsseprevcetdraAosfG1λ9λ19−8B7.26B5,a1n0d2R9E.4085Å03li−n2e8p9roafitmleasswsefrlalcintiothnes e co3 44sd22SD
orefs3p.e7ct×ive1l0y−(7F(i2g9.5ti)m.Oesursoalbaurn)daanndc8e.s3d×ev1ia0t−e6b(y7f6a0cttoimrseosfs6olaanr)d, artur2
0.5for G191−B2BandRE0503−289,respectively,fromthose dep1 4p2Po
ofChayeretal.(2015),whomadeanLTEassumptionbecause
oflackofatomicdata.Bothstarsareoutofthevaliditydomain 0
forLTEmodeling(e.g.,Rauch2012),whichiscorroboratedby -5 0
aninspectionofthedeparturecoefficients(ratioofNLTEtoLTE log (m / g/cm2)
v
foocrcGup1a9t1io−nBn2uBmabnedrsR)oEf0t5h0e3fi−v2e8lo9wtheasttdAesviatleevferolsminuonuitrym(Foidge.l6s Fig.6.DeparturecoefficientsofthefivelowestAsvlevelsinthemodel
forG191−B2B.
and7, respectively). Therefore, the results of an LTE analysis
maybeafflictedbyalargesystematicerror.
Figure8 shows a comparison of theoretical profiles theabundancehadtobereducedto1.0×10−7tomatchtheobser-
of Asvλλ987.65,1029.48Å that are calculated from our vation.ForRE0503−289,theyaretooweakand,thus,ahigher
NLTEmodelandanLTEmodel(calculatedbyTMADbasedon abundanceof1.7×10−5isneededtoreproducetheobservation.
the temperature and density stratification of the NLTE model, This may explain, in part, the deviation of our As abundances
LTEoccupationnumbersoftheatomiclevelsareenforcedbyan fromthoseofChayeretal.(2015).
artificialincreaseofallcollisionalratesbyafactorof1020).In However, a precise abundance analysis by means of
thecaseofG191−B2B,theLTEprofilesaretoostrongand,thus, NLTEmodel-atmospheretechniquesrequiresreliableoscillator
A39,page4of28
T.Rauchetal.:Stellarlaboratories:newMoiv–viioscillatorstrengths
7 5
s 3 2
efficient3 4s2S e flux NiV 1314.5 NiVCoV CoVGaIVNiV 1437.5
e co2 ativ
ur el
rt r
a1
p
de 44pf 22FPoo 10% 1Ao
5s2S
0 4d2D Δλ / Ao
-5 0
log (m / g/cm2) Fig.9.SectionsofourHST/STISobservationsofRE0503−289around
iv iv
Sn λ1314.537Å(left)andSn λ1427.525Å(right).Thethick,red
Fig.7.AsforFig.6,forRE0503−289. andthin,greenlinesshowacomparisonwiththeoreticalspectraoftwo
modelswithandwithoutSn,respectively.
1 0
65 48 n 0 VIII
987. 029. ctio VII
1 a
n fr VI
o V
ati -5
z
x ni IV
u o
ative fl 987.651 029.480 log i-10 -5 0
el 1
r log (m / g/cm2)
Fig.10.MolybdenumionizationfractionsinourRE0503−289model.
6.3.RE0503-289:Molybdenum
10%
1Ao Our RE0503−289 model (Teff =70000K, log g = 7.5) in-
Δλ / Ao cludes opacities of H, He, C, N, O, Si, P, S, Ca, Sc, Ti, V, Cr,
Mn,Fe,Co,Ni,Zn,Ga,Ge,As,Kr,Mo,Xe,andBa.Figure10
vi+vii
Fig.8.AsforFig.5,thethickredanddashedbluelinesshowacompar- shows the Mo ionization fractions in this model. Mo
isonwiththeoreticalspectraofNLTEandLTEmodels,respectively. are the dominatingionizationstages in the line-formingregion
(−4.0 (cid:2) logm (cid:2) 0.5). The element abundances are given in
Table2. In general, their uncertainty is about 0.2dex. This in-
strengths,notonlyforthelinesemployedintheabundancede- cludestheerrorpropagationduetotheerrorrangesofTeff (cf.,
e.g.,Vennes&Lanz2001),logg,andthebackgroundopacity.
termination,butforthecompletemodelatomthatisconsidered
Figure11 shows the relative line strengths (normalized to
inthemodel-atmosphereandspectral-energy-distributioncalcu- vi
that of Mo λ1038.642Å) of the Mo lines in the synthetic
lations. Once these oscillator strengths become available for a
large number of Asv-vii lines, it will be possible to construct spectrumofRE0503−289.We notethatthese strengthsreduce
ifthespectrumisconvolvedwithaGaussiantomatchtheinstru-
muchmorecompleteAsmodelionsandanewdeterminationof
ments’resolutions(Sect.2).
theAsabundancewillbemoreprecise.
In the FUSE and HST/STIS observations, we identified
v vi
12 Mo and nine Mo lines (Fig.12). Their strengths were
6.2.RE0503-289:Tin(Z=50) wellreproducedata Mo abundanceof 1.88×10−4 (massfrac-
tion), which is 35400 times the solar value (Grevesse et al.
Rauch et al. (2013)measuredthe Sn abundancein G191−B2B 2015). Many more weak Mov and Movi lines are visible in
(3.53 × 10−7, 37 times solar). They used the lines of the res- thesyntheticspectrumbuttheyfadeinthenoiseofthepresently
onancedoublet,namely,Snivλ1314.537Å(5s2S1/2−5p2P◦3/2, avaivlaibile observations. The search for the strongest Moiv and
f = 0.637) and λ1437.525Å(5s2S −5p2P◦ , f = 0.240). Mo lines (Fig.11) was not successful. This was not unex-
1/2 1/2 pected, given the predicted weakness of the lines and the S/N
Their f-valueswere calculatedbyMorton(2000)based onen- vii
of the data. Figure13 shows the two strongest Mo lines of
ergy levels provided by Moore (1958). These lines are visible
vii
in our HST/STIS spectrum of RE0503−289 as well. To de- ourmodel.WenotethatBa λ1255.520Ådominatestheob-
termine the Sn abundance, we used the same model atom like servedabsorptionaroundλ1255.5Å(Fig.13).
Rauchetal.(2013)andreproducedtheobservedlineprofilesof The identification of Mo lines in the wavelength region
bothlines(Fig.9)wellatamassfractionof2.04×10−4 (about λ >∼ 1700Å was strongly hampered by the lower signal-to-
22500timessolar).ThisSnabundanceanalysiswillalsobeim- noise (S/N) and resolution (only a fourth of the exposure time
proved, as soon as reliable transition probabilities for Sniv-vi and 66% of the resolving power of the spectrum, compared to
arepublished. theregionatλ <∼ 1700Å,see Sect.2).Anexampleisshownin
A39,page5of28
A&A587,A39(2016)
VI
o
M
1
h
gt
n
e
str
e
v
ati 0
el
r
MoIV
MoV
MoVI
MoVII
1000 1100 1200 1300 1400 1500 1600 1700
λ/Ao
Fig.11.RelativestrengthsofMolinescalculatedfromourstellar-atmospheremodelofRE0503−289.Topgraph:Moiv–viilines,thefourmost
vi
prominent are Mo lines that were identified by Werner et al. (2012b) are marked. Graphs 2 to 5 (from top to bottom): lines of individual
Moiv–viiions(intensitiesreducedbyafactorof0.22comparedtothetopgraph),respectively.
Table2.PhotosphericabundancesofRE0503−289. 6.4.G191-B2B:Molybdenum
Mass Number Wlogega=dd7e.6d)fMoroGi1n9to1−oBu2rBa,twmhoiscphhceorensmidoerdseHl ,(THeeff,C=,6N0,0O0,0AKl,,
Element [X]
Fraction Si,P,S,Ca,Sc,Ti,V,Cr,Mn,Fe,Co,Ni,Zn,Ga,Ge,As,Mo,
He 9.74×10−1 9.92×10−1 0.592 Sn,andBa.TheabundancesaregiveninTable3.Movi+viiare
C 2.23×10−2 7.56×10−3 0.974 the dominating ionization fractions in the line-forming region
N 1.73×10−4 5.04×10−5 −0.602 (Fig.15).
O 1.97×10−3 5.01×10−4 −0.464
Si 1.61×10−4 2.33×10−5 −0.617 WeperformedasearchforMolinesintheobservedspectra
P 1.15×10−6 1.51×10−7 −0.705 of G191−B2B,analogousto thatin Sect.6.3.However,we did
S 3.96×10−5 5.04×10−6 −0.892 notidentifyany.Figure16showsacomparisonofoursynthetic
IG 1.00×10−6 9.19×10−8 −1.787 spectra to the observations. Since the HST/STIS spectrum of
Fe <1.30×10−5 <9.50×10−7 <−1.967 G191−B2B(Sect.2)isofexcellentquality,Moviλ1469.168Å
Ni 7.26×10−5 5.04×10−6 0.028 gives a stringent upper abundance limit of 5.3×10−7 by mass
Zn 1.13×10−4 7.05×10−6 1.814 (about100timessolar,Grevesseetal.2015).
Ga 3.45×10−5 2.02×10−6 2.810
Ge 1.59×10−4 8.90×10−6 2.845
As 8.27×10−6 4.50×10−7 2.879 6.5.G191-B2B:Krypton(Z=36)andXenon(Z=54)
Kr 5.05×10−5 2.46×10−6 2.666
Mo 1.88×10−4 8.00×10−6 4.549 For G191−B2B, we have determined a relatively high upper
Sn 2.04×10−4 7.00×10−6 4.351 abundance limit (about 100 times solar) for Mo (Sect.6.4).
Xe 6.30×10−5 1.96×10−6 3.577 This is well in agreementwith a factor of about100−1000be-
Ba 3.58×10−4 1.06×10−5 4.301 tween the trans-iron element abundances in RE0503−289 and
G191−B2B.LikeMo,Kr,andXeexhibitprominentlinesinthe
Notes.IGisagenericmodelatom(Rauch&Deetjen2003)comprising UV spectra of RE0503−289, but not in those of G191−B2B.
Ca,Sc,Ti,V,Cr,Mn,andCo.[X]denoteslog(fraction/solarfraction)
To investigate on their abundances, we individually included
ofspeciesX.
Kr and Xe in our G191−B2B models and calculated theoret-
ical profiles for all Kr and Xe lines that were identified in
Fig.14. Movλ1718.088Åappearsat comparablestrengths of RE0503−289(Werneretal.2012b;Rauchetal.2015a).Anup-
other lines that were identified (Fig.12). However, it is within perKrabundancelimitof10timessolar(1.09×10−6 bymass)
thenoiseleveloftheobservationandhastobejudgeduncertain. is determined from lines of Krvi-vii simultaneously (Fig.17).
Moviλ1718.238Åisweakerbutabetterobservationwouldal- In the case of Xe, the intersystem lines Xeviiλ995.51Å
lowanidentification. (5s21S−5s5p3P◦)andXeviiλ1077.12Å(5s5p1P◦–5p21D)are
A39,page6of28
T.Rauchetal.:Stellarlaboratories:newMoiv–viioscillatorstrengths
FeV 1011.889 NiVI GaV1148.502 ZnV 1590.414 XeVII FeV995.806 ZnV 1182.142 ZnV 1479.168
STIS
0 0 0 77 1 0 93 5
3 9 5 27 4 4 82 3
5 6 0 22 5 6 90 4
GaVI 1101.GaV1101.ZnV XeVI ZnV 1186.ZnV1186.1186. ZnV 1653.CIV GeV 1038.ZnVGaV NiV 1263.1264.ZnV NiV 1595.
x
u
e fl 72 88 54 61 15 83 82 23
v 6 9 9 0 2 3 1 5
elati 1125. 1125. GaV 1186.NiVI1187. 1661. 1661. ZnV1047. GaVOIV NIIINIVNIII 1270. NiV
r
5 9 8 2 2 0
9 2 9 6 4 3
9 2 8 6 1 0
V 37. 38.V 86. 68. V 82. V 27.
n 1 1n 5 6 n 1 n 4
Z 1 1Z 1 1 Z 1 Z 1
FUSE
10% Mo V Mo VI
1Ao
Δλ / Ao
v vi
Fig.12.Mo lines(leftpanel,markedwiththeirwavelengthsfromTableA.11inÅ,green)andMo lines(rightpanel,markedblue,wavelengths
fromTableA.12)intheFUSE(forlinesatλ < 1188Å)andHST/STIS(λ > 1188Å)observationsofRE0503−289. Thesyntheticspectraare
convolvedwithaGaussian(fullwidthathalfmaximum=FWHM=0.06Å)tosimulatetheinstruments’resolutions.Otheridentifiedphotospheric
linesaremarkedinblack.Thethickredandthingreenlinesshow acomparison withtheoreticalspectraoftwomodelswithandwithout Mo,
respectively.Theverticalbarindicates10%ofthecontinuumflux.
2 3 8 8
8 6 8 3
ux NiV 1255.1 ZnVBaVIIZnVNiVI NiV 1341.1 1718.0 1718.2 NIV
e fl x
v u
relati ative fl
10% 1Ao rel
Δλ / Ao
vii 10%
Fig.13.SameasFig.12,fortwoMo lines. 1Ao
Δλ / Ao
verystrong(Fig.17)andrequireanupperXeabundancelimitof
Fig.14.SectionofourHST/STISobservationsofRE0503−289around
solartofadeinthenoiseoftheobservation.However,thismay iv
vii N λ1718.55Å.Thethickredandthingreenlinesshowacomparison
be strongly underestimated because of the rudimentary Xe
with theoretical spectra of two models with and without Mo, respec-
model atom presently provided by TMAD. In that, only two v vi
vii tively.Mo λ1718.088ÅandMo λ1718.238Åaremarked.
Xe lineswithreliableoscillatorstrengthsareknown,namely,
vii
0.245 for Xe λ995.51Å (Kernahan et al. 1980) and 0.810
vii
for Xe λ1077.12Å (Biémont et al. 2007). Since, for the
calculationofaccurateNLTEoccupationnumbersoftheatomic but will be investigated in detail immediately after new Xeiv-
vii
levels of an specific model ion, reliable transition probabilities transition probabilities become available. However, from
are mandatory for the complete ion, the Xevii upper limit is the Xevi lines alone, we achieve an upper limit of 1.7×10−7
regardedasuncertain.Thisissueisoutofthescopeofthispaper (10timesthesolarvalue,Fig.17).
A39,page7of28
A&A587,A39(2016)
n 0 VIII Table3.SameasTable2,forG191−B2B.
ctio VII
ation fra -5 VVI Element MassFractionNumber [X]
z H 9.99×10−1 9.99×10−1 0.132
oni IV He <1.98×10−5 <5.00×10−6 <−4.099
g i C 6.31×10−6 5.30×10−7 −2.574
lo-10 N 2.08×10−6 1.50×10−7 −2.522
-5 0 O 1.90×10−5 1.20×10−6 −2.479
log (m / g/cm2) Al 1.12×10−5 4.20×10−7 −0.675
Si 5.29×10−5 1.90×10−6 −1.099
Fig.15. Same as Fig.10 forvi+G1v9ii1−B2B. For comparison, the P 1.54×10−6 5.00×10−8 −0.579
dashed lines show the Mo ionization fractions in the S 5.72×10−6 1.80×10−7 −1.733
RE0503−289model. IG 1.78×10−6 4.00×10−8 −1.538
Fe 6.50×10−4 1.17×10−5 −0.269
7. Impactofdiffusionontrans-ironelements
Ni 3.84×10−5 6.60×10−7 −0.249
At almost the same logg, RE0503−289 has a significantly Zn 3.50×10−6 5.40×10−8 0.304
lhoigghge=r7T.e5ffvcso.mTpeffar=ed60to00th0aKt,olfogGg19=17−.6B,2rBesp(Tecefftiv=el7y0).0T0h0uKs,, GGAaes 233...527641×××111000−−−667 345...750000×××111000−−−889 111...615853051
the much stronger enrichment of the trans-iron elements in
RE0503−289(Fig.18)maybetheresultofamoreefficientra- Kr <1.09×10−6 <1.31×10−0 <1.000
Mo <5.33×10−7 <5.60×10−9 <2.000
diativelevitation.Therefore,weusedtheNGRT12code(Dreizler Sn 3.53×10−7 3.00×10−9 1.589
& Wolff 1999; Schuh et al. 2002) to calculate diffusion mod- Xe <1.67×10−7 <1.28×10−9 <1.000
els for both stars, using exactly the same model atoms for H, Ba 4.00×10−6 2.94×10−8 2.350
He, C, N, O, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Zn, Ga, Ge,
As, Kr, Mo, Sn, Xe, and Ba, which were used for our chem-
ically homogeneousTMAP models. For RE0503−289,H was The Mov/Movi ionization equilibrium is well reproduced
formallyincludedinthecalculation,butitsabundancewasfixed (Fig.12).Wedeterminedaphotosphericabundanceoflog Mo=
to 1.0 × 10−20. Therefore, its contribution to the background −3.73 ± 0.2 (mass fraction 1.2−3.0 × 10−4, 22500−56400
opacity is negligible. Disregarding the fixed H abundance in
times the solar abundance). In addition, we determined the ar-
RE0503−289,thediffusionmodelsdifferonlyinTeff andlogg. senic and tin abundances and derived log As = −5.08 ± 0.2
TheTMADmodelatomsforAsandSnarepresentlyrather (0.5−1.3 × 10−5, about 300−1200 times solar) and log Sn =
rudimentary, especially as only a very few oscillator strengths −3.69 ± 0.2 (1.3−3.2 × 10−4, about 14300−35200 times so-
areknown.Restrictingtheradiativelevitationcalculationtoin-
lar).ThesehighlysupersolarAs,Mo,andSnabundancesagree
clude transitions with known oscillator strengths thus leads to
well with the high abundances of other trans-iron elements in
an unrealistically small effect. We follow Rauch et al. (2013) RE0503−289(Fig.18).
and add all allowed line transitions in our As and Sn model
G191−B2BdoesnotexhibitKr,Mo,andXelinesinitsUV
atoms, using default f-values of 1. Therefore, the results of
spectrum. We investigated the strongest lines in the modeland
our diffusion models for these elements should be regarded as foundupperlimitsfortheabundancesofKr(1.1×10−6,10times
preliminary. solar), Mo (5.3× 10−7, 100 times solar), and Xe (1.7 × 10−7,
Figure19 shows the calculated depth-dependentabundance
10 times solar). Whether radiativelevitationyields abundances
profilesforZn,Ga,Ge,As,Kr,Mo,Sn,Xe,andBa.Alltheseare
of these elements that are consistent with observations should
stronglyoverabundantintheline-formingregions(Fig.19).The
be demonstrated with advancedline-profile calculations in dif-
predicted abundance profiles suggest abundance enhancements
fusiveequilibrium,asdepictedinFig.19.Inaddition,wedeter-
inRE0503−289relativetoG191−B2B.Thisisqualitativelyin
minedthearsenicabundanceandderivedlog As = −6.43±0.2
agreement with the abundance patterns in Fig.18, which were (2.3−5.9 × 10−7,about21−53timessolar).
determinedfromourstaticTMAPmodels.However,whetherit
ivsanpcoesdsiblilnee-toprorefialcehca2l−cu3ldaetixonsshoinulddiffbuesdiveemeoqnusitlriabtreidumw.itThhaedse- MoTivh–eviciomwpasutaatiopnreroefquriesliitaeblfeortrathnesitiidoenntpifirocbataiboinlitioefs Mfoor
lines and the subsequent abundance determination. The hith-
would provide stringent constraints for suggested weak stellar
erto known abundance pattern of RE0503−289 (Fig.18) indi-
winds(Chayeretal.2005)andthinconvectivezonesatthestel-
cates that other yet unidentified species should be detectable.
lar surface (Chayer et al. 2015), which might impact the inter-
Therefore,thepreciseevaluationoftheirlaboratoryspectra,i.e.,
playofradiativelevitationandgravitationalsettling.
the measurement of line wavelengths and strengths, as well as
the determinationof levelenergiesand the subsequentcalcula-
8. Resultsandconclusions tionoftransitionprobabilitiesareabsolutelyessential.
The example of the arsenic abundance determination
With our NLTE model-atmosphere package TMAP, we calcu-
(Sect.6.1) had demonstrated that state-of-the-artNLTE stellar-
lated new models for the DO-type white dwarf RE0503−289
atmospheremodelsaremandatoryfortheprecisespectralanal-
withmolybdenuminaddition.TheMomodelatomswerecon-
ysisofhotstars.
structed with newly calculated Moiv–vii oscillator strengths.
v vi
Wehaveunambiguouslyidentified12Mo andnineMo lines
Acknowledgements. T.R. and D H. are supported by the German Aerospace
in the observed high-resolution UV spectra of RE0503−289. Center (DLR, grants 05OR1402 and 50OR1501, respectively). The GAVO
projecthadbeensupportedbytheFederalMinistryofEducationandResearch
12 NewGenerationRadiativeTransport. (BMBF)atTübingen(05AC6VTB,05AC11VTB)andisfundedatHeidelberg
A39,page8of28
T.Rauchetal.:Stellarlaboratories:newMoiv–viioscillatorstrengths
011.889 186.050 186.227186.277 590.414 995.811 182.143 595.435
1 1 11 1 1 1
FUSE
0 0 1 1 0 3
3 9 6 4 4 4
5 6 0 5 6 1
1. 1. 7. 3. 8. 2.
0 0 8 5 3 8
1 1 1 6 0 1
1 1 1 1 1 1
x
u
e fl 2 8 5 3 4 8 STIS
v 0 9 1 8 8 6
ati 8.5 6.8 1.2 1.3 7.1 9.1
el 14 58 66 66 04 47
r 1 1 1 1 1 1
2
6
6
8.
6
6
1
10% Mo V Mo VI
1Ao
Δλ / Ao
Fig.16.LikeFig.12forG191−B2B.ThethickredandthinbluespectraarecalculatedfrommodelswithphotosphericMoabundancesof5.3×10−6
and5.3×10−7(massfractions,about1000and100timesthesolarvalue),respectively.
4 6 1
3 4 9 6 3 5
9 3 3 3 1 5
9. 7. 1. 8. 9. 7.
1 2 3 2 2 6
9 9 9 9 9 9
944.05 956.015 965.093 996.23 017.270 080.080
1 1
e flux 970.084 980.421 1002.746 1091.634 1101.947 1136.412
v
ati
el 1 5 1
r 14 77 23 54 46 91
1. 5. 5. 9. 1. 8.
1 1 4 7 8 9
0 0 0 1 1 2
1 1 1 1 1 1
9
1052.95 1061.06 KrVII918.440 XeVII995.511 XeVII1077.12
10% Kr Xe
1Ao
Δλ / Ao
vi vii vi vii
Fig.17.TheoreticallineprofilesofKr(leftpanel,14Kr linesand1Kr line)andXe(rightpanel,12Xe and2Xe lines)comparedto
theobservationsofG191−B2B.Thespectraarecalculatedfrommodelswithphotosphericabundances(massfractions)ofKr:1.1×10−5 (thick
red,100timessolar)and1.1×10−6(thinblue,10timessolar)andofXe:1.7×10−6(dashedblack,100timessolar),1.7×10−7(thickred,10times
solar),and1.7×10−8(thinblue,solar).
(05AC11VH3). Financial support from the Belgian FRS-FNRS is also ac- STScIisoperatedbytheAssociationofUniversitiesforResearchinAstronomy,
knowledged.P.Q.isresearchdirectorofthisorganization.Wethankourreferee, Inc.,underNASAcontractNAS5-26555.SupportforMASTfornon-HSTdatais
StéphaneVennes,forconstructivecriticism.Someofthedatapresentedinthis providedbytheNASAOfficeofSpaceScienceviagrantNNX09AF08Gandby
paperwereobtainedfromtheMikulskiArchiveforSpaceTelescopes(MAST). othergrantsandcontracts.ThisresearchhasmadeuseofNASA’sAstrophysics
A39,page9of28
A&A587,A39(2016)
He C O Si S TiCrFeNiZnGe Kr Mo Sn XeBa DataSystemandtheSIMBADdatabase,operatedatCDS,Strasbourg,France.
H N Al P MnCo GaAs
0 RE0503-289 TheTOSSservice(http://dc.g-vo.org/TOSS)thatprovidesweightedoscil-
latorstrengthsandtransitionprobabilitieswasconstructedaspartoftheactivities
G191-B2B
oftheGermanAstrophysicalVirtualObservatory.
n
o
cti References
a
r
ss f-5 Asplund,M.,Grevesse,N.,Sauval,A.J.,&Scott,P.2009,ARA&A,47,481
a Biémont,É.,Clar,M.,Fivet,V.,etal.2007,Eur.Phys.J.D,44,23
m
Cabeza,M.I.,Iglesias,L.,Rico,F.R.,&Kaufman,V.1989,Phys.Scr,40,457
g Chayer,P.,Vennes,S.,Dupuis,J.,&Kruk,J.W.2005,ApJ,630,L169
o
l Chayer,P.,Dupuis,J.,&Kruk,J.W.2015,in19thEuropeanWorkshoponWhite
Dwarfs,eds.P.Dufour,P.Bergeron,&G.Fontaine,ASPConf.Ser.,493,3
Cowan,R.D.1981,Thetheoryofatomicstructureandspectra(Berkeley,CA:
UniversityofCaliforniaPress)
Cowley,C.R.1970,Thetheoryofstellarspectra(NewYork:Gordon&Breach)
4 Cowley,C.R.1971,TheObservatory,91,139
Curtis,L.J.,&Theodosiou,C.E.1989,Phys.Rev.A,39,605
2
Dowler, P., Rixon, G., & Tody, D. 2010, Table Access Protocol Version 1.0,
X] 0 IVOARecommendation
[ Dreizler,S.,&Werner,K.1996,A&A,314,217
-2 Dreizler,S.,&Wolff,B.1999,A&A,348,189
-4 Fischer,C.F.1977,J.Phys.B,10,1241
Fraga,S.,Karwowski,J.,&Saxena,K.M.S.1976,HandbookofAtomicData
(Amsterdam:Elsevier)
10 20 30 40 50
Grevesse,N.,Scott,P.,Asplund,M.,&Sauval,A.J.2015,A&A,573,A27
atomic number Hubeny,I.,Hummer,D.G.,&Lanz,T.1994,A&A,282,151
Hummer,D.G.,&Mihalas,D.1988,ApJ,331,794
Fig.18. Solar abundances (Asplund et al. 2009; Scott et al. 2015b,a;
Johnson,W.R.,Kolb,D.,&Huang,K.-N.1983,AtomicDataandNuclearData
Grevesseetal.2015,thickline;thedashedgreenlinesconnecttheel- Tables,28,333
ementswithevenandwithoddatomicnumber)comparedwiththede- Kernahan,J.A.,Pinnington,E.H.,O’Neill,J.A.,Bahr,J.L.,&Donnelly,K.E.
terminedphotosphericabundancesofG191−B2B(bluecircles, Rauch 1980,J.Opt.Soc.Am.,70,1126
etal.2013)andRE0503−289(redsquares, Dreizler&Werner1996; McCook,G.P.,&Sion,E.M.1999a,ApJS,121,1
Werner et al. 2012b; Rauch et al. 2013, 2014a,b, 2015a,b, and this McCook,G.P.,&Sion,E.M.1999b,VizieROnlineDataCatalog: III/210
work). Top panel: abundances given as logarithmic mass fractions. Migdalek,J.,&Baylis,W.E.1979,J.Phys.B,12,1113
Moore, C. E. 1958, Atomic Energy Levels as Derived from the Analysis of
Arrowsindicateupperlimits.Bottompanel:abundanceratiostorespec-
tivesolarvalues,[X]denoteslog(fraction/solarfraction)ofspeciesX. Optical Spectra Molybdenum through Lanthanum and Hafnium through
Actinium(NIST)
Thedashedgreenlineindicatessolarabundances.
Morton,D.C.2000,ApJS,130,403
Osuna,P.,Barbarisi,I.,Salgado,J.,&Arviset,C.2005,inAstronomicalData
RE0503-289 AnalysisSoftwareandSystemsXIV,eds.P.Shopbell,M.Britton,&R.Ebert,
G191-B2B ASPConf.Ser.,347,198
-2 Zn Osuna, P., Guainazzi, M., Salgado, J., Dubernet, M.-L., & Roueff, E. 2010,
SimpleSpectralLinesDataModel,IVOARecommendation2Dec
-7 Pinnington,E.H.,Bahr,J.L.,Kernahan,J.A.,&Irwin,D.J.G.1981,J.Phys.B,
-2 Ga 14,1291
Quinet,P.,Palmeri,P.,Biémont,E.,etal.1999,MNRAS,307,934
-7 Quinet,P.,Palmeri,P.,Biémont,E.,etal.2002,J.AlloysComp.,344,255
-2 Ge Rauch,T.2012,ArXive-prints[arXiv:1210.7636]
Rauch, T., & Deetjen, J. L. 2003, in Stellar Atmosphere Modeling, eds.
-7 I.Hubeny,D.Mihalas,&K.Werner,ASPConf.Ser.,288,103
Rauch,T.,Werner,K.,Biémont,É.,Quinet,P.,&Kruk,J.W.2012,A&A,546,
-2 As
n A55
o
ss fracti ---772 Kr RRRRaaaauuuucccchhhh,,,,TTTT....,,,,HWWWoeeeyrrrnnnereee,rrr,,,DKKK.,...,,,QQBQuouuinhiinneliteen,tt,,,P.PRP,...G,,,&&a&llKKaKrrrdruuuokkk,,,,MJJJ....WW,W&...222R000a111i44n3abe,,,rAiAA,&&&MAAA.,2,,505561660546,a,,A,AAA1410&106A,577,
ma A88
g -2 Mo Rauch,T.,Werner,K.,Quinet,P.,&Kruk,J.W.2015b,A&A,577,A6
o Reader,J.2010,J.Phys.B,43,074024
l -7
Reader,J.,&Tauheed,A.2015,J.PhysB,48,144001
-2 Sn Salgado,J.,Osuna,P.,Guainazzi,M.,etal.2010,SimpleLineAccessProtocol,
IVOAhttp://www.ivoa.net/documents/SLAP/
-7 Schuh,S.L.,Dreizler,S.,&Wolff,B.2002,A&A,382,164
-2 Xe Scott, P.,Asplund,M.,Grevesse, N.,Bergemann, M.,&Sauval, A.J.2015a,
A&A,573,A26
-7 Scott,P.,Grevesse,N.,Asplund,M.,etal.2015b,A&A,573,A25
-2 Ba Shirai, T., Sugar, J., Musgrove, A., & Wiese, W. L. 2000, Spectral Data for
HighlyIonizedAtoms:Ti,V,Cr,Mn,Fe,Co,Ni,Cu,Kr,andMo(AIP)
-7 Sugar,J.,&Musgrove,A.1988,J.Phys.Chem.Ref.Data,17,155
Vennes,S.,&Lanz,T.2001,ApJ,553,399
-5 0
log (m/g/cm2) Werner, K., Deetjen, J. L., Dreizler, S., et al. 2003, in Stellar Atmosphere
Modeling,eds.I.Hubeny,D.Mihalas,&K.Werner,ASPConf.Ser.,288,31
Fig.19. Abundance profiles in our diffusion models for G191−B2B Werner,K.,Dreizler, S.,&Rauch,T.2012a,TMAP:TübingenNLTEModel-
(thinblue)andRE0503−289(thickred).Thedashed,horizontallines AtmospherePackage(AstrophysicsSourceCodeLibrary)
Werner,K.,Rauch,T.,Ringat,E.,&Kruk,J.W.2012b,ApJ,753,L7
indicatesolarabundancevalues.TheformationregionsofUVlinesin
bothmodelsareindicatedatthetop.
A39,page10of28