Abundances of the light elements from UV (HST) and red (ESO) spectra in the very old star HD 84937

0.4 MB·

by  M. Spite

Astronomy&Astrophysicsmanuscriptno.30058_hk2ArX (cid:13)cESO2017
January18,2017
Abundances of the light elements from UV (HST) and red (ESO)
⋆
spectra in the very old star HD84937
M.Spite1,R.Peterson2,A.J.Gallagher1,B.Barbuy3,andF.Spite1
1 GEPI,ObservatoiredeParis,PSLResearchUniversity,CNRS,UniversitéParisDiderot,SorbonneParisCité,PlaceJulesJanssen,
92195,Meudon,France
2 SETIInstitute,189BernardoAve,Suite100,MountainView,CA94043,USA
3 UniversidadedeSãoPaulo,IAG,RuadoMatão1226,CidadeUniversitária,05508-900,SãoPaulo,Brazil
7
1 January18,2017
0
2
ABSTRACT

n
a Inordertoprovideabetterbasisforthestudyofmechanismsofnucleosynthesisofthelightelementsbeyondhydrogenandhelium
J intheoldeststars,theabundancesofC,O,Mg,Si,P,S,K,andCahavebeenderivedfromUV-HSTandvisible-ESOhighresolu-
7 tionspectraintheold,verymetal-poorstarHD84937,atametallicitythatis1/200thatoftheSun’s.Forthishalomain-sequence
1 turnoff star,theabundance determinationof PandS arethefirstpublished determinations. TheLTEprofilesof thelineswerefit-
tedtotheobserved spectra.Whereverpossible, wecorrectedthederivedabundances fornon-LTEeffects.Three-dimensional(3D)
] CO5BOLD model atmospheres have been used to determine the abundances of C and O from molecular CH and OH bands. The
R abundancesoftheselightelementsinHD84937arefoundtoagreewellwiththeabundancesinclassicalmetal-poorstars.OurHD
S 84937carbonabundancedeterminationpointstowardasolar(ormildlyenhanced)valueof[C/Fe].Themodestoverabundanceofthe
alphaelementsofevenatomicnumberZ,typicalofhaloturnoffstars,isconfirmedinthisexample.Theodd-ZelementPisfoundto
.
h besomewhatdeficientinHD84937,at[P/Fe]=–0.32,whichisagainconsistentwiththehandfulofexistingdeterminationsforturnoff
p starsofsuchlowmetallicity.Weshowthattheabundanceofoxygen,deducedfromtheOHbandfrom3Dcomputations,isnotcom-
- patiblewiththeabundancededucedfromtheredoxygentriplet.Thisincompatibilityisexplainedbytheexistenceofachromosphere
o
heatingtheshallowlayersoftheatmospherewheretheOHband,in3Dcomputations,ismainlyformed.Theabundanceratiosare
r
t comparedtothepredictionsofmodelsofgalacticnucleosynthesisandevolution
s
a Keywords. Stars:Abundances–Galaxy:abundances–Galaxy:halo
[

1
v1. Introduction 2013;VandenBergetal.2014;Snedenetal.2016;Amarsietal.
8 2016; Zhaoetal. 2016). The model parameters are very sim-
0TheoldestsurvivingunevolvedstarsintheMilkyWayaremade ilar in all these studies. The adopted effective temperature
6from matter that is not very enriched in elements produced by T ranges from 6300 to 6408K, and the metallicity [Fe/H]1
eff
4previous(i.e.older,moremassive)stars. The elementsin these
from–1.9(Laietal.2007)to–2.3(Snedenetal.2016).Thesur-
0survivingstarsshowlowabundances,buttheobserveddeficien- facegravitylogg rangesfrom4.0to4.1inthesestudies;theonly
1.cies of individualelements are not identical. The measurement exception is a value of logg = 4.5 derived by Laietal. (2007)
0ofthesedeviationsisessentialinordertoconstrainthedifferent fromphotometryratherthanspectra.
7nucleosynthesisprocessesproducingtheseelements.Itisthere- FollowingVandenBergetal.(2014),acomparisonofthepo-
1fore very important to measure the detailed abundances for as
sitionofthestarinthe M vs.logT diagramwithisochrones
v:many elementsas possible, with the highestprecision, in these computedwith differentavgesand meefftallicities yieldsan age of
Xitvheerylioglhdtsetlaerms.eTnthsissupcahpearsiPs,mS,aiannldyKde,vtohteedevtooluthtieoanboufnwdahniccheoinf 12.08 ± 0.14Gyr, adopting Teff = 6408 K and logg = 4.05.
Butthisagedependscriticallyontheadoptedtemperature.The
rtheGalaxyisstillpoorlyknown. error bar takes into account only the parallax uncertainty. If
a
HD84937 is a bright, (V = 8.32, Ducati 2002) very we instead adopt the parameters T = 6300K and logg =4.0,
eff
metal-poor, main-sequence turnoff star, often used as a refer- as in Snedenetal. (2016), then the age of HD84937 becomes
ence for the field metal-poor stars (lately: Lawleretal. 2015), ≈13.8Gyr,veryclosetotheageoftheUniverse:13.799±0.038
and consequently has recently been widely studied from spec- Gyr (PlanckCollaborationetal. 2015). Unfortunately, a new
tra in the visible domain (Gehrenetal. 2006; Laietal. 2007; parallaxforthisstarisnotyetavailableinthefirstGaiarelease
Mashonkinaetal. 2011; Bergemannetal. 2012; Lindetal. 2.
Owing to the weakness of spectral lines at the temperature
⋆ Based on observations made with the NASA/ESAHubbleSpace and metallicity of HD84937, the analysis of very high quality
Telescopeobtained under program GO-14161 at the Space Telescope spectraisnecessarytoestablishaccurateabundancesforthisref-
Science Institute (STScI), which is operated by the Association of erencestar.TakingadvantageofanewUVhighresolutionspec-
Universitiesfor ResearchinAstronomy (AURA)and onobservations
collected at the European Organisation for Astronomical Research in 1 We adopt the classical notation that for element X: A(X) =
the Southern Hemisphere (Archives of programmes 080.D-0347(A), log(N /N )+12 and [X/H]=log(N /N ) −log(N /N ) .
X H X H star X H Sun
082.B-0610(A),266.D-5655(A),and073.D-0024(A)). 2 http://www.cosmos.esa.int/web/gaia/dr1
Articlenumber,page1of8
A&Aproofs:manuscriptno.30058_hk2ArX
trum from the HubbleSpaceTelescope(HST), complemented 3. Spectralanalysisandabundancemeasurements
by previous high resolution, high signal-to-noise UVES and
In our analysis we used OSMARCS model atmospheres
HARPS spectra obtained in the visible at the European South-
(Gustafssonetal. 2008) together with the turbospectrum LTE
ern Observatory (Very Large and 3.6m telescopes), we present
synthesis code (Alvarez&Plez 1998; Plez 2012). In this code,
newdeterminationsoftheabundancesofseveralelementswith
the collisional broadening by neutral hydrogen is generally
atomic numbersZ lyingbetween Z=6(carbon)and Z=20(cal-
computedfollowingthetheorydevelopedbyAnstee&O’Mara
cium).NewabundancesofC,Si,P,andSinparticularhavebeen
(1991),Barklemetal.(2000)andBarklem&O’Mara(2000).
determinedfromthe new HST spectrum,which containsmany
For the computations we adopted the model parame-
more useful lines of these elements than the optical and near-
ters and [Fe/H] derived for HD84937 in the same way
infraredspectradoinsuchweak-linedstars.
as in Peterson&Kurucz (2015) and Petersonetal. (2016):
Phosphorus (Z=15) has never been measured in
T =6300K, logg=4.0, and v=1.3kms−1, along with the iron
HD84937 and indeed has been measured in very few metal- eff t
abundancetheyderivedwiththeseparameters,[Fe/H]=−2.25.
poor stars, including only five stars with a metallicity below
ThesevaluesdeducedonlyfromtheUVspectraareinexcellent
[Fe/H] < −2.0dex (Caffauetal. 2011, 2016; Roedereretal.
agreementwith the values derivedfor this star by Lawleretal.
2014). Its behaviour hints at a primary origin in massive stars
(2013)andSnedenetal.(2016)basedonV−Kphotometryand
(Jacobsonetal.2014).
accurateHIPPARCOSparallaxes.
Sulphur(Z=16)hasneverbeenmeasuredinHD84937,butit
Theparametersanddeducedabundanceforeachatomicline
hasbeenmeasuredinasampleofmetal-poorstars(Caffauetal.
considered for this analysis are listed in Table 1. The mean
2005; Spiteetal. 2011; Caffauetal. 2016) from near-infrared
abundances of the different elements are given in Table 2. For
lines (multiplets 1, 6, and 8) and in a carbon-rich metal-poor
the computations of [X/H] we adopted the solar abundances
star(Roedereretal.2016)fromthreeUVlines(multiplet2uv)3.
A(C)=8.5,4 A(Mg)=7.54, A(Si)=7.52, A(P)=5.46, A(S)=7.16,
We present the first determination of the sulphur abundancein
A(K)=5.11, and A(Ca)=6.33, following Caffauetal. (2011) or
HD84937,usingtheUVmultiplet1uvcomparedtotheinfrared
Loddersetal.(2009).
multiplet1.
We also provide 3D corrections for a selected number of
Forthefirsttimethesiliconabundancehasbeendetermined
atomicandmolecularlines.Theywerecomputedfrom3Dsyn-
in HD84937 from the profile of six UV lines and from the Si
theticspectrabasedonaCO5BOLD(Freytagetal.2012)model
lineinthevisible.
atmosphere from the Cosmological Impact of the First STars
(CIFIST; Ludwigetal. 2009) grid, using the latest version of
Linfor3D5(Gallagheretal.2016a).Agridof1Dsyntheticspec-
2. Observationaldata
tra – based on 1D LHD models(Caffau&Ludwig 2007) com-
putedwiththesamemicro-physicsasthe3Dmodels–werefit
The HST spectrum was obtained as a legacy spectrum un-
tothe3Dsyntheticprofilestodeterminethe3Dcorrection(see
der GO-14161, with R. Peterson as PI. As described further
Gallagheretal.2016b).
by Petersonetal. (2016), this spectrum was acquired with the
E230H grating of the STIS echelle spectrograph and the 0.2" The 3D and equivalent 1D LHD models selected from
x 0.09"slit for a resolvingpowerR=114000.Adoptingfive of the CIFIST grid have stellar parameters Teff/logg/[Fe/H] =
thesix E230H“prime”wavelengthsettings, withcentralwave- 6206K/4.0/−2.0,whichiscurrentlythemodelinthegridwith
lengths201.3nm,226.3nm,251.3nm,276.2nm,and301.3nm, stellarparametersclosesttotheadoptedOSMARCSmodelpa-
resultedinananalysiscoverageof188.0nm–314.0nm.Theto- rameters.
talexposuretimeof106.4ks(29.5hours)yieldedS/N>=50per
two-pixel resolution element at all wavelengths, except for the
3.1.Carbonandoxygenabundances
bluestandreddest.
To compare the abundances deduced from the HST spec- In HD84937 the C and O abundances can be derived from
traandthevisible,wealsousedground-basedspectrafromthe atomiclinesandfromthemolecularCHandOHbands.
ESO Very Large Telescope archives(spectrographsUVES and
HARPS).Forwavelengthsshorterthan400nmweusedHARPS
3.1.1. CandOfromatomiclines
spectra.TheyhavearesolvingpowerR=115000andasignal-to-
noiseperpixel(S/N)betterthan150perpixel.Atlongerwave- – Carbon
lengths,inthebluewecombinedtheindividualUVESspectrato Itispossibletodeterminethecarbonabundancefromafitof
a mean UVES spectrum with a resolving power R=65000 and theweakredCilinesnear910nm,theweakUVlinesaround
aS/Nperpixelbetterthan500.Inthered,inordertominimise 199and248nmandtheverystrongUVCilineat193nm.In
theinfluenceofthetelluriclines,andtakingadvantageofthefact thethreecasesweestimate(Table2)thattheerrorin[C/H]
thatfollowingCarney&Latham(1987)anddeBruijne&Eilers is≈0.1dex.
(2012), HD84937 shows no detectable velocity variability, we The red Ci lines are located within a H O telluric band.In
2
used three different mean spectra each obtained by combining thisregionwehadseveralspectraobservedatdifferentdates
individual spectra with almost the same geocentric radial ve- (2002,2003,2004),andforeachmeanspectrumtherelative
locity. Therefore, the positions of the telluric lines relative to positionofthetelluriclineswasdifferent.Thuswechoseto
the stellar lines are almost the same. These three mean spec- measure each Ci line from only those spectra in which tel-
trawereobtainedconsecutivelyonasinglenighteveryyearbe- luriccontaminationwasminimal.The906.25linewasmea-
tween 2002 and 2004. The resulting mean red spectra reach a suredonlyonthemean2003spectrum,the907.83linecould
resolving power of about 45000 and a S/N per pixel equal to be measured on the 2002 and 2003 mean spectra, and the
310,510,and350,respectively,intheregionoftheredCilines.
4 withtheclassicalnotation:A(X)=log (N /N )+12
10 X H
3 WiththemultipletnumberingofMoore(1945) 5 http://www.aip.de/Members/msteffen/linfor3d
Articlenumber,page2of8
XX1etal.:AbundancesofC,O,Mg,Si,P,S,K,andCainHD84937
Table 1. Line-by-line 1D LTE abundances A(X) from atomic lines. Table2.Meanabundances ofthedifferentelements. InCols.2and3
The adopted non-LTE corrections to be applied are given in the last aregiventheabundancesofthedifferentelementsderivedfromthe1D
column.ForS,thenon-LTEcorrectionincludesthe3Dcorrection.Air computations.Weadoptedthesolarabundancesgiveninsection3and
wavelengths are given for wavelengths > 200nm and vacuum values [Fe/H]=–2.25.Inthelastcolumnisgiven[X/Fe]correctedforNLTE
below. and3Dwhenpossible.
Element [X/H] [X/Fe] σ [X/Fe]
Z λ(nm) Ex.Pot. loggf A(X)LTE NLTEcorr 1D 1D corrected
6 C(Ci) LTE LTE (adopted)
193.0905 1.26 -0.13 6.30 <0.01 C(CH) –1.86 +0.39 – +0.11
199.2012 1.26 -5.74 6.47 <0.01 C(Ci)(UV) –2.09 +0.16 0.15 +0.17
199.3620 1.26 -3.72 6.60 0.01 C(Ci)(red) –2.06 +0.19 0.04 +0.10
247.8561 2.68 -0.96 6.28 0.01
906.2470 7.48 -0.455 6.41 -0.07 O(OH) –1.47 +0.78
907.8280 7.48 -0.581 6.43 -0.08 O(Oi)(red) –1.48 +0.77 – +0.65
908.8509 7.49 -0.430 6.43 -0.08
911.1799 7.49 -0.297 6.49 -0.10 Mg –1.99 +0.26 0.07 +0.30
8 O
777.1944 9.15 0.369 7.28 -0.12 Si –1.90 +0.35 0.06 +0.38
777.4166 9.15 0.223 7.28 -0.12
777.5388 9.15 0.002 7.28 -0.12 P –2.57 –0.32 0.07 –0.32
12 Mg
279.5528 0.00 0.100 5.55 0.00 S(UV) –1.83 +0.42 0.03 +0.42
280.2705 0.00 -0.210 5.55 0.00 S(red) –1.75 +0.50 0.08 +0.20
457.1096 0.00 -5.623 5.61 —
470.2991 4.35 -0.440 5.60 0.04 K(red) –1.68 +0.57 0.03 +0.37
517.2684 2.71 -0.380 5.59 0.06
518.3604 2.72 -0.158 5.54 0.02 Ca –1.78 +0.47 0.08 +0.53
552.8405 4.34 -0.341 5.49 0.01
14 Si
197.7598 0.00 -1.309 5.66 <0.01
198.0618 0.01 -1.437 5.69 <0.01
198.3233 0.01 -1.192 5.65 <0.01
208.2021 0.78 -1.900 5.63 -0.01
208.4462 0.78 -1.880 5.62 -0.01
212.2990 0.78 -1.840 5.57 0.00
212.4122 0.78 0.149 5.40 <0.01
390.5523 1.91 -1.041 5.70 0.05
15 P
213.5469 1.41 -1.240 2.98 —
213.6182 1.41 -0.111 2.94 —
215.2939 1.41 -0.357 2.81 —
253.3986 2.32 -1.114 2.86 —
255.4911 2.33 -1.231 2.85 —
16 S
190.0287 0.00 -3.709 5.30 —
191.4697 0.05 -4.261 5.31 —
216.8884 1.14 -3.962 5.30 —
921.2863 6.53 0.420 5.33 -0.3
922.8093 6.53 0.260 5.49 -0.3
Fig.1.Examplesofthebestfits(redthickline)oftwoatomiclinesof
923.7538 6.53 0.040 5.40 -0.3
Ciinthered.Syntheticprofiles(bluethinlines)havebeencomputed
19 K
withA(C)=6.1,6.4,and6.7.
766.4899 0.00 0.149 3.46 -0.2
769.8964 0.00 -0.154 3.40 -0.2
20 Ca 911.18lineonlyonthe2004spectrum.Anexampleofthefit
428.3011 1.89 -0.220 4.55 0.07 isgiveninFig.1.InFig.2wepresentthefitofthewingsof
431.8652 1.90 -0.210 4.49 0.06 thestrongCilineat193.1nm.
442.5437 1.88 -0.360 4.50 0.07 Fabbianetal.(2006,2009b)computedanon-LTEcorrection
443.5679 1.89 -0.520 4.52 0.07 forHD84937of–0.2dexforthehighexcitationCiredlines.
445.4779 1.90 0.260 4.50 0.04 Zhaoetal.(2016)andMashonkina(privatecommunication)
526.5556 2.52 -0.260 4.63 0.13 computedacorrectioncloseto–0.1dexfortheredCilines
534.9465 2.71 -0.310 4.50 0.12 and≤0.01dexfortheUVCilines;weadoptedthesecorrec-
558.1965 2.52 -0.710 4.67 0.09 tions (Table 1). These new measurementspoint to a [C/Fe]
558.8749 2.52 0.210 4.62 0.08 ratioinHD84937closetosolar(Table2).
559.0114 2.52 -0.710 4.67 0.09 – Oxygen
560.1277 2.52 -0.690 4.69 0.09
585.7451 2.93 0.230 4.50 0.08 Articlenumber,page3of8
610.2723 1.88 -0.790 4.51 0.05
612.2217 1.89 -0.320 4.52 0.00
616.2173 1.90 -0.090 4.54 0.00
643.9075 2.52 0.470 4.42 -0.02
A&Aproofs:manuscriptno.30058_hk2ArX
Fig.2.ObservedprofileofthestrongCilineinHD84937.Thebluethin
linesrepresentthesyntheticprofilescomputedwithA(C)=6.0and6.9.
Thethickredlinerepresentsthebestfit(1Dcomputations)A(C)=6.3.
Fig.3.Comparisonbetweentheobservedandsyntheticprofilesofthe
redoxygentripletinHD84937.Thesyntheticprofiles(bluethinlines)
oftheoxygentriplethavebeencomputedwithA(O)=6.9and7.5.The
thickredlinerepresentsthebestfitforA(O)=7.28.
We derived the oxygen abundance in HD84937 from the
oxygen triplet at 777nm (Fig. 3). The non-LTE correction
(−0.12dex)wasadoptedfromthecalculationsofZhaoetal. Fig. 4. Observed profile of the CH and OH bands in HD84937. For
theCHbandsthe1DETLsyntheticprofiles(bluethinlines)havebeen
(2016)andleadsto[O/H]=7.16±0.1.Theresultingvalue
computed with A(C)= 0.0, 6.4, and 6.7. The absorption by the wing
[O/Fe] = +0.65 is in good agreement with the value ob-
ofHγhasbeenincludedinthecomputation oftheG-band. Thethick
tained from the forbidden [Oi] line in the EMP giants by
redlinerepresentsthebestfit:A(C)=6.60fortheUVCHfeatureand
Spiteetal. (2005), and is in a reasonable agreement, in- A(C)=6.64for the G-band. The synthetic profile of the OH band has
side the measurement errors, with the value obtained by been computed withA(O)=7.0 and 7.6. Thebest fit isobtained with
Zhaoetal. (2016) from the red Oi lines (part of the differ- A(O)=7.29
enceindeedcomesfromtheadopted[Fe/H]value).
–3DComputations
Wecomputed3Dcorrectionstothe1Dabundancesdetermined
3.1.2. CandOfromCHandOHmolecularbands for the G-band, and 14 OH lines in the wavelength range
312.24−312.83nm.Wedefinea3Dcorrection,∆ ,withthefor-
3D
–1DComputations malismgivenbyGallagheretal.(2016b),i.e.,∆ = A(X) −
3D 3D
-We computed the CH carbon abundance by fitting the profile
A(X) , wherethe 1D synthesisis computedfromLHD model
1D
oftheCHfeatureat314.3nmandoftheG-bandbetween430.8
atmospheres.Wecomputed3DprofilesofCHandOHusingthe
and 431.4nm. Line lists for 12CH and 13CH (Masseronetal.
absolute carbon and oxygen abundances derived from the 1D
2014)wereincludedinthesynthesis.Theresultispresentedin non-LTE computations of the red Ci and Oi atomic lines and
Fig. 4. For the UV feature the best fit correspondsto [C/H] = thus with a C/O ratio of ≈ 0.15. (Unlike the 1D computations,
−1.90dex, i.e., [C/Fe] = +0.35dex, and for the G-band to the3DcomputationsaresensitivetotheC/Oratio)6.
[C/H] = −1.86dex,i.e., [C/Fe] = +0.39dex. The CH bandin
For the G-band we found a 3D correction of –0.28dex.
HD84937isveryweakandthusthecorrespondingabundanceis The resulting C abundance, [C/H]= –2.14 or [C/Fe]= +0.11,
verysensitive to the positionof the continuumand in this case
is in very good agreement with the abundance deduced from
theerroron[C/H]isestimatedto≈0.15dex.
the atomic lines (Table 2). However, the OH lines were far
-AfitoftheOHbandbetween312and314nmleadsto[O/H]=– stronger in 3D than in 1D. In fact, owing to the low C/O ratio
1.47(Fig.4)and[O/Fe]=+0.78.
in the atmosphere of a typical metal-poor star like HD84937,
Thereis,inHD84937,arathergoodagreementbetweenthe the OH lines are mainly formed in the most external regions
abundances deduced from the atomic lines (Table 2) and the of the atmosphere, which are about 300K cooler in the 3D
molecularbandsfrom1Dcomputations.However,inmetal-poor model than in the 1D model (see Fig. 1 in Gallagheretal.
stars the molecularlines are sensitive to convectiveeffects, ne- 2016b, for a metallicity of –2). In this case, the 3D correction
glected in the 1D computations,and we have to take them into
account. 6 X/Y=N(X)/N(Y)=10[A(X)∗−A(Y)∗]
Articlenumber,page4of8
XX1etal.:AbundancesofC,O,Mg,Si,P,S,K,andCainHD84937
fortunately,inHD84937aninterstellarlinewitharadialvelocity
veryclose to the radial velocityof the star perturbsthis profile
on the red side, and the red peak of the emission is not clearly
visible.
We also computed the 3D correction for two Ci lines: the
verystronglineat193.09nm(Fig.2)andthelineat199.36nm.
Inbothcaseswefoundacorrectionlowerthanthemeasurement
error(respectively+0.06and-0.07).
3.2.Abundancesofclassicalαelements:magnesiumand
calcium
Foramorehomogeneouscomparison,inthevisiblewederived
the magnesium and calcium abundances from the equivalent
widths of a set of lines already used in extremely metal-poor
(EMP)main-sequenceturnoffstarsbyBonifacioetal.(2009).
FortheopticalMgilinesinTable1,weadoptedthenon-LTE
correctioncomputedbyZhaoetal.(2016)andwefoundamild
overabundanceofMg(Table2:[Mg/Fe] =+0.30dex),asdid
LTE
Gehrenetal.(2006)andZhaoetal.(2016).Thisvalueisalsoin
goodagreementwiththevaluewededucefromthewingsofthe
very strong Mgii h and k lines around 280nm ([Mg/Fe] =
LTE
+0.26dex).Thenon-LTEcorrectionhasnotbeencomputedfor
Fig.5.CentreofthethehandkMgiilinesintheHST/STISUVspectra
theseUVlines.However,followingAbia&Mashonkina(2004),
ofHD84937andHD140283.Theprofilesofthehandklineshavebeen
sinceMgiiisthemajorityspecies,departuresfromLTEareex-
overplottedtoshowtheemissionatthecentreofbothlinesindicating
pectedto be caused mainlyby radiativeb-b transitions, and no
theexistenceofachromosphere.InHD84937aninterstellarline,witha
radialvelocityclosetotheradialvelocityofthestar,perturbstheprofile processseems to affect the Mgii ground-statepopulation.As a
ontheredsideofthecore. consequence,thenon-LTEcorrectionfortheresonancehandk
Mgiilinesshouldbenegligible.
Our LTE calcium abundances in Table 1 have been corrected
reaches−0.72dex.Applyingthis3DcorrectiontotheOHband fornon-LTEeffectsfollowingSpiteetal.(2012).Wehavethree
would lead to a value of [O/H]=–2.19([O/Fe]=+0.06),a value linesincommonwith Zhaoetal.(2016),andthenon-LTEcor-
hardlycompatiblewiththeobservationoftheatomiclines.For rections for these lines do not differ by more than 0.02dex.
HD84937 the adopted 3D model seems to be too cool in the The Ca abundance found in HD84937 is in very good agree-
most outer layers. This could be due to the omission of a hot ment with the mean abundances obtained, using the same set
solar-likechromosphereinour3Dmodel. of lines, for EMP turnoff stars with −3.6 < [Fe/H] < −2.5
(Bonifacioetal.2009;Spiteetal.2012).
From HST spectra of the Mgii lines near 280nm,
Peterson&Schrijver (1997) showed that solar-like chromo-
spheres routinely occur among turnoff metal-poor stars despite 3.3.Siliconabundance
theirrelativelylowdegreeofactivity.Fromsevensuch starsof
a wide metallicity range down to 1/300 solar whose radial ve- Thesiliconabundancecouldbededucedfromsevensiliconlines
locities largely shifted the line coresaway from the interstellar intheUVandanotheroneat390nm.Thislastlinewasusedto
absorption,thisworkdemonstratedthatdouble-peakedemission determinetheSi abundancein the sampleofEMPturnoffstars
cores were always present, and that the blue peak was usually studiedinBonifacioetal.(2009).Shietal.(2009)computedfor
strongerthantheredasisalsotheusualcaseforthequietSun. thislinea non-LTEcorrectionof+0.05dex.Thenon-LTEcor-
Theyconcludedthat“whilethesedatadonotruleoutmagnetic rection for the ultraviolet lines were computed by Mashonkina
fieldstheysupportanacousticoriginofchromosphericemission (privatecommunication)andarelessthan0.1dex(Table1).Our
andshowthatrelativelyinactivesolar-typestarsofallageshave derivedoverabundanceof Si is then [Si/H] = −1.87±0.10 or
chromosphereswhosecharacteristicsarelargelyindependentof [Si/Fe]=+0.38.Itisinterestingtonotethatthisvalueishigher
metallicity”. Subsequent work, based on observations of Hei than the mean silicon overabundance derived in EMP dwarfs
1083nmabsorptionfeatures,haveconfirmedtheseresults(e.g., from the 390nm line alone: [Si/Fe] ≈ +0.09± 0.14. But this
Takeda&Takada-Hidai2011).Moreworkonthissubject,based mean value has not been corrected for non-LTE effects and, in
on the profiles of the Mgii lines in HST/STIS spectra of sev- theseEMPturnoffstarswith[Fe/H]≤ −3,thenon-LTEcorrec-
eralmetal-poorstars,isongoing(Peterson&Tarbellinprepara- tioncouldreach0.25dex(Shietal.2009).Asaconsequencethe
tion).InFig. 5we presentthe profileof theMgii h andklines silicon abundancefoundin the UV in HD84937is compatible
in our HST/STIS spectrum of HD84937 and in the G0-14161 with the results of Bonifacioetal. (2009). As for Mg and Ca,
E230Hspectrumofanotherclassicalextremelymetal-poorstar, our[Si/H] value is veryclose to the valuefoundby Zhaoetal.
HD140283(Petersonetal.2016).InHD140283thetwopeaks (2016), and the small difference in [Si/Fe] is attributed to the
–typicalofasolar-likechromosphere–are clearlyvisible;un- adopted[Fe/H]values.
Articlenumber,page5of8
A&Aproofs:manuscriptno.30058_hk2ArX
3.4.Phosphorusabundance
Thephosphorusabundanceisderivedfromfiveultravioletlines
of the multiplets 4(uv), 8(uv), and 9(uv), leading to [P/Fe] =
−0.32.Thestatisticalerrorreaches0.07dex.SomeotherPlines
are visible in the spectrum, but these lines are so severely
blended that they cannot be used to determine the P abun-
dance with precision. In Fig. 6 we show the fit of the ob-
served lines with the computed spectra. The value of [P/Fe] in
HD84937confirmsthe low value of [P/Fe] in metal-poorstars
(Fig.7).
Recently R. Kurucz has published new loggf values for the P
lines in the UV and in the red 7. These new values are about
0.1dex higher than the old ones. Consequently, their adoption
herewouldreducethe[P/Fe]valuesofTables1and2byabout
0.1dex.
To date, no non-LTE calculations have been performed for
Pi features;followingJacobsonetal.(2014)theyareexpected
tobesmallinthistypeofstar.
3.5.Sulphurabundance
InHD84937,thesulphurabundancecouldbederivedfromthe
profileoftwoUVfeaturesat190.03and191.47nmandfromthe
profiles of the red multiplet 1 (Table 2). This multiplet is con-
taminated by a telluric band of H O. We measured the S lines
2
only on spectra wherethey were notblendedby a telluric line.
The921.2linecouldbemeasuredonthe2002,2003,and2004
spectra,buttheothertwolines(922.8and923.7m)onlyonthe
2002 and 2003 spectra. We obtained good agreement between
the abundances deduced from the two different sets of lines:
[S/Fe]=+0.42dexfromtheUVlinesand[S/Fe]=+0.50fromthe
redlines.
It is well known that non-LTE effects influence the profiles of
the red multiplet 1. Following Takedaetal. (2005), the non-
LTE correction is about –0.2 dex for the model parameters of
HD84937. However, following Spiteetal. (2011), based on a
new model atom of sulphur by Korotin (2008, 2009a,b) and
Fig. 6. Comparison of the observed (crosses) and synthetic (curves)
newphotoionisationrates,theNLTEcorrectionforthemodelof spectrainHD84937intheregionofthefourPIlinesanalysedinthis
HD84937reaches–0.4dex.Ontheotherhand,ithasbeencom- work.Theredthicklinemarksthebest-fitsyntheticspectrum,obtained
puted that in the case of this multipletthe 3D correctionis not withthevalueofA(P)indicatedinthefigure.Thethinbluelinesshow
negligible and reaches +0.1dex (Spiteetal. 2011). This small syntheticspectracomputedwithA(P)=2.6andA(P)=3.2.Asisusually
butpositive correctionhasbeen addedto the NLTE correction. done,theloggf valuesoftheblendinglineshavebeenadjustedtogive
Finally,wefindfromtheredmultiplet1,[S/Fe]=+0.20dex. amorecorrectrepresentationofthedifferentfeatures.
As far as we are aware, non-LTE corrections of the UV
features have never been computed, but they are probably as
4. Discussionandconclusion
small asforC and Si. We estimate a sulphuroverabundancein
HD84937of[S/Fe]≈0.3dex,ingoodagreementwiththemean
Our analysis of high quality spectra for HD84937,notablythe
sulphuroverabundancefoundinmidmetal-poorstars(François
HST ultraviolet spectra as well as ground-based optical/near-
1986)andinEMPstars(Spiteetal.2011).
infraredspectra,hasallowedustorefinetheabundancesofsev-
eral light elements between C and Ca in this star, and provide
thefirstmeasurementsofitsP,S,andKabundancesandamore
3.6.Potassiumabundance
firmly established Si abundance. Accurate abundances of such
The abundance of K has been deduced from the red lines keyelementsarevitalinreferencestarslikeHD84937inorder
at 766.4899nm and 769.8964nm. Following Takedaetal. tobetterinterpretchemicalevolutionmodelsandsupernovaeen-
(2002), the non-LTE correction for the model parameters of richmentintheGalactichalo.
HD84937reaches–0.3dex,butZhaoetal.(2016)hascomputed The presence of emission-core reversal at the centre of the
a correction of –0.2dex and we adopted this latter value. As a very strong Mgii resonance lines confirms the existence of a
consequencewefindforthisstararatio[K/Fe]=0.37±0.10,in chromospherearoundHD84937(Peterson&Schrijver1997).
excellentagreementwith the value foundfor the EMPstars by The abundancesof C and O deducedfromthe atomic lines
Takedaetal.(2009)andAndrievskyetal.(2010). areingoodagreementwiththeabundancesobtainedintheEMP
stars(Bonifacioetal.2009;Spiteetal.2005).
7 http://kurucz.harvard.edu/linelists.htmlseefileGFALL. TheCabundancededucedfromtheCHmolecularbandafterthe
Articlenumber,page6of8
XX1etal.:AbundancesofC,O,Mg,Si,P,S,K,andCainHD84937
arehighlystrengthened.However,current3Dmodelsformetal-
poorturnoffstarsdonotincludetheriseintemperatureexhibited
byasolar-likechromosphere.Thesemodelshavebeenbuiltonly
fortheSun,butsimilarmodelswouldbeusefulinordertoanal-
ysemolecularfeaturesinclassicalmetal-poorturnoffstars.
TheclassicalαelementsMg,Si,andCahaveaboutthesame
overabundance in HD84937 as observed in the EMP turnoff
starsstudiedinBonifacioetal.(2009).HD84937isalsomildly
enhancedinK,asaretheothermetal-poorstars.
–The ratio [P/Fe] is found to be slightly below the solar
value, as is typically seen in very metal-poor stars (Fig. 7)
Fig. 7. [P/Fe] vs. [Fe/H] is shown for the five stars analysed by
and more generally for the light odd-Z elements like Na and
Roedereretal.(2014)andRoedereretal.(2016)with [Fe/H] < −2.0
Al. Molaroetal. (2001) also measured [P/Fe] in a dust free
(blue dots), plus our result for HD84937 (blue star symbol). The
Ly-α system that has the same metallicity as HD84937, and
theoretical predictions of Timmesetal. (1995) (red dot-dashed line),
Goswami&Prantzos(2000)(pinkdottedline),Kobayashietal.(2011) they found [P/Fe]=–0.27, about the same ratio as observed in
(bluethickline),andCescuttietal.(2012)(greendashedline)arealso HD84937(Table2).
represented. In Fig. 7 we show the P measurements of Roedereretal.
(2014, 2016). The two most metal-poor stars in the figure
areBD+44◦493 ([Fe/H]=−3.88)andG64-12,([Fe/H]=−3.28).
They are both carbon-enhanced metal-poor stars (CEMP) fol-
lowing Roedereretal. (2016) and Placcoetal. (2016), but in
thesestarstheneutron-captureelementsarenotenhanced.They
are “CEMP-no” stars and thus the composition of their atmo-
sphere is supposed to reflect the abundance of the cloud that
formed the star. This cloud would have been enriched by a
faintsupernova(SN)providingcarbonandthelighterelements,
but the elements heavier than Mg are not affected (see, e.g.,
Bonifacioetal. 2015). As a consequence, the P abundance in
thesestarsshouldbecomparabletothePabundanceinclassical
Fig. 8. [S/Fe] vs. [Fe/H] for metal-poor stars with [Fe/H] < −2.0 metal-poorstars(i.e.,metal-poorstarswithoutCenhancement)
(blue squares), plus our result for HD84937 (blue star symbol). The andweincludetheminthefigure.
theoretical predictions of Timmesetal. (1995) (red dot-dashed line), According to Woosley&Weaver (1995) and West&Heger
Goswami&Prantzos(2000)(pinkdottedline),Kobayashietal.(2011) (2013), the isotope 31P is produced mostly in SNII in both O
(bluethickline),andMatteucci(2016)(greendashedline)arealsorep- and Ne burning shells, and its ejection is not very sensitive to
resented. theexplosionmechanism.Takingintoaccountcore-collapsesu-
pernovae and hypernovae, Kobayashietal. (2006, 2011) pro-
ducedmodelsofchemicalevolutionforthehalo.Weoverplotted
theirbestmodeltotheveryfewavailablePdatapoints(Fig.7).
Alsoplottedinthisfigurearethemetallicity-dependentmodelof
Timmesetal.(1995),themodelofGoswami&Prantzos(2000)
using yields from Woosley&Weaver (1995) at variablemetal-
licities,andmodel4ofCescuttietal.(2012).Thebestfitseems
tobeobtainedwiththeKobayashietal.(2011)model.However
intheinterpretationoftheabundancetrendswehavetobevery
prudent.We note that NLTE computationshave not been done
for this element and that following Caffauetal. (2016) for the
disk stars there is a shiftof 0.23dex between the UV-HST and
Fig. 9. [K/Fe] vs. [Fe/H] for metal-poor stars with [Fe/H] < −2.0 the near-infrared measurements of the P abundance (see their
(blue squares), plus our result for HD84937 (blue star symbol). The
Fig. 5). This disagreementcould be the result of differentnon-
theoretical predictions of Timmesetal. (1995) (red dot-dashed line),
LTEcorrectionsfortheUVandnear-infraredlinesbutcouldbe
Goswami&Prantzos(2000)(pinkdottedline),Kobayashietal.(2011)
alsotheresultofsystematicerrorsinthegf values.
(bluethickline),andMatteucci(2016)(greendashedline)arealsorep-
resented. –Sulphur is an α element mainly produced in massive su-
pernovae or hypernovae. In Fig. 8 we compare the observed
[S/Fe] ratios to the abundance ratios predicted by the mod-
3D-1Dcorrectionisinverygoodagreementwiththeabundance elsofGoswami&Prantzos(2000),Kobayashietal.(2011)and
deducedfromtheatomiclines(Table2). Matteucci (2016). The best fit is obtained with the model of
The oxygen abundance deduced from the OH molecular band Matteucci(2016).
computed with 1D models agree with the abundancesdeduced –In the very metal-poor stars, potassium (an odd element:
fromtheatomiclines;however,when3Dcomputationsaredone, Z = 19) is enhanced relative to iron (Fig. 9) as are the (even)
theabundanceofoxygendeducedfromtheOH bandisincom- α elements. The models struggle to represent the abundance
patiblewiththeabundancededucedfromtheatomiclines.In3D of K in the Galactic halo. The best fit (Fig. 9) is achieved by
computations,atvariancewith1Dcomputations,theOHbandis the model of Goswami&Prantzos (2000); however, these au-
formedpredominantlyinshallowatmosphericlayers,andsince thorspointoutthatthisgoodrepresentationcouldbefortuitous
3D models are cooler at the surface than 1D models, OH lines and might be due to a factor of two reduction of the Fe yields
Articlenumber,page7of8
A&Aproofs:manuscriptno.30058_hk2ArX
combinedwiththeadoptedIMF(Kroupaetal.1993).Following Korotin,S.A.2008,OdessaAstron.Pub.,21,42
Kobayashietal.(2011)theunderproductionofKinthetheoret- Korotin,S.A.,2009a,AstronomicheskiiZhurnal,86,702
ical massive supernovae/hypernovaecould be the consequence Korotin,S.A.2009b,AstronomyReports,53,651
Kroupa,P.,Tout,C.A.,&Gilmore,G.1993,MNRAS,262,545
ofaneglectoftheνprocess.Whenthecoreofmassivestarscol-
Lodders,K.,Palme,H.,&Gail,H.-P.2009,LandoltBörnstein,NewSeries,As-
lapses, a large flux of neutrinos is emitted and these neutrinos tronomyandAstrophysics,SpringerVerlag,Berlin
interact with heavy elements through neutral-current reactions Lai,D.K.,Johnson,J.A.,Bolte,M.,&Lucatello,S.2007,ApJ,667,1185
producinginparticularFandK. Lawler,J.E.,Guzman,A.,Wood,M.P.,Sneden,C.,&Cowan,J.J.2013,ApJS,
205,11
The HST has permitted an extension of the pattern of the
Lawler,J.E.,Sneden,C.,&Cowan,J.J.2015,ApJS,220,13
light elements produced by the explosion of the progenitor of
Lind,K.,Melendez,J.,Asplund,M.,Collet,R.,&Magic,Z.2013,A&A,554,
a very old, metal-poor, unevolved star whose composition is A96
highlyrepresentativeofthe earlieststagesofpriormassivestar Ludwig,H.-G.,Caffau,E.,Steffen,M.,etal.2009,Mem.Soc.Astron.Italiana,
nucleosynthesis.Theseelementsareakeytoabetterunderstand- 80,711
Mashonkina, L.,Gehren, T.,Shi,J.-R.,Korn,A.J.,&Grupp,F.2011,A&A,
ing of the nucleosynthetic processes that have formed the el-
528,A87
ements in the early massive supernovae/hypernovae.It is very Masseron,T.,Plez,B.,VanEck,S.,etal.2014,A&A,571,A47
importantthat futureaccess to the UV regionof the spectra be Matteucci,F.2016,JournalofPhysicsConferenceSeries,703,012004
preserved. This region is a very powerful tool for the study of Molaro, P., Levshakov, S. A., D’Odorico, S., Bonifacio, P., & Centurión, M.
2001,ApJ,549,90
theearlyphasesoftheGalacticevolution.
Moore,C.E.1945,ContributionsfromthePrincetonUniversityObservatory,vol.
Acknowledgements. WearegratefultoLyudmilaMashonkinaforaveryuseful 20,p.1
andcompetent reportandforthecommunication ofseveral unpublished non- Peterson,R.C.,&Schrijver,C.J.,1997,ApJ,480,L47
LTEcorrections.WealsothankPiercarloBonifaciowhocomputedtheATLAS9 Peterson,R.C.,&Kurucz,R.L.2015,ApJS,216,1
modelandWIDTHprofiles forustocheck theCilines. Thisworkwas sup- Peterson,R.C.,Kurucz,R.L.,&Ayres,T.R.2016,inpreparation
portedbythe“ProgrammeNationaldePhysiqueStellaire”andthe“Programme Placco,V.M.,Beers,T.C.,Reggiani,H.,Meléndez,J.2016,
NationaldeCosmologieetGalaxies”(CNRS-INSU).PartialsupportforR.Pe- PlanckCollaboration,Adam,R.,Ade,P.A.R.,etal.2015,arXiv:1502.01586
tersonwasprovidedbyNASAunderHSTGO-14161.R.Petersonthanksher PlezB.,2012,http://2012ascl.soft05004P
GO-14161coinvestigatorsTomAyresforhisassistanceinacquiringandreduc- Roederer,I.U.,Jacobson,H.R.,Thanathibodee,T.,Frebel,A.,&Toller,E.2014,
ingtheHSTspectraandRobertKuruczforhisadviceinanalysingthem.A.J. ApJ,797,69
GallagheracknowledgesthefundingbyFONDATIONMERACandthematch- Roederer,I.U.,Placco,V.M.,Beers,T.C. 2016,ApJ,824,L19
ingfundgrantedbytheScientificCouncilofObservatoiredeParis.B.Barbuy Shi,J.R.,Gehren,T.,Mashonkina,L.,&Zhao,G.2009,A&A,503,533
waspartiallysupportedbyFAPESP,CNPq,andCAPES. Sneden,C.,Cowan,J.J.,Kobayashi,C.,etal.2016,ApJ,817,53
Spite,M.,Cayrel,R.,Plez,B.,etal.2005,A&A,430,655
Spite,M.,Caffau,E.,Andrievsky,S.M.,etal.2011,A&A,528,A9
Spite,M.,Andrievsky,S.M.,Spite,F.,etal.2012,A&A,541,A143
References Takeda,Y.,Zhao,G.,Chen,Y.-Q.,Qiu,H.-M.,&Takada-Hidai,M.2002,PASJ,
54,275
Abia,C.,&Mashonkina,L.2004,MNRAS,350,1127 Takeda,Y.,Hashimoto,O.,Taguchi,H.,etal.2005,PASJ,57,751
AlvarezR.,&PlezB.,1998,A&A330,1109 Takeda,Y.,Kaneko,H.,Matsumoto,N.,etal.2009,PASJ,61,563
Amarsi,A.M.,Lind,K.,Asplund,M.,Barklem,P.S.,&Collet,R.2016,MN- Takeda,Y.,&Takada-Hidai,M.2011,PASJ,63,547
RAS,inpress(ArXiv:1608.06390) Timmes,F.X.,Woosley,S.E.,&Weaver,T.A.1995,ApJS,98,617
Andrievsky,S.M.,Spite,M.,Korotin,S.A.,etal.2010,A&A,509,A88 VandenBerg,D.A.,Bond,H.E.,Nelan,E.P.,etal.2014,ApJ,792,110
Anstee,S.D.,&O’Mara,B.J.1991,MNRAS,253,549 West,C.,&Heger,A.2013,ApJ,774,75
Barklem,P.S.,Piskunov,N.,&O’Mara,B.J.2000,A&AS,142,467 Woosley,S.E.,&Weaver,T.A.1995,ApJS,101,181
Barklem,P.S.,&O’Mara,B.J.2000,MNRAS,311,535 Zhao,G.,Mashonkina,L.,Yan,H.L.,etal.2016,arXiv:1610.00193
Bergemann,M.,Lind,K.,Collet,R.,Magic,Z.,&Asplund,M.2012,MNRAS,
427,27
Bonifacio,P.,Spite,M.,Cayrel,R.,etal.2009,A&A,501,519
Bonifacio,P.,Caffau,E.,Spite,M.,etal.2015,A&A,579,A28
Caffau,E.,Bonifacio,P.,Faraggiana,R.,etal.2005,A&A,441,533
Caffau,E.,&Ludwig,H.-G.2007,A&A,467,L11
Caffau,E.,Bonifacio,P.,Faraggiana,R.,&Steffen,M.2011,A&A,532,A98
Caffau, E., Ludwig, H.-G., Steffen, M., Freytag, B., & Bonifacio, P. 2011,
Sol.Phys.,268,255
Caffau,E.,Andrievsky,S.,Korotin,S.,etal.2016,A&A,585,A16
Carney,B.W.,&Latham,D.W.1987,AJ,93,116
Cescutti,G.,Matteucci,F.,Caffau,E.,&François,P.2012,A&A,540,A33
deBruijne,J.H.J.,&Eilers,A.-C.2012,VizieROnlineDataCatalog,354,
Ducati,J.R.2002,VizieROnlineDataCatalog,2237,
Fabbian,D.,Asplund,M.,Carlsson,M.,&Kiselman,D.2006,A&A,458,899
Fabbian,D.,Asplund,M.,Barklem,P.S.,Carlsson,M.,&Kiselman,D.2009a,
A&A,500,1221
Fabbian, D., Nissen, P. E., Asplund, M., Pettini, M., & Akerman, C. 2009b,
A&A,500,1143
François,P.1986,A&A,165,183
Freytag,B.,Steffen,M.,Ludwig,H.-G.,etal.2012,JournalofComputational
Physics,231,919
Gallagher,A.J.,Caffau,E.,Bonifacio,P.,etal.2016,A&A,593,A48
Gallagher,A.J.,Steffen,M.,Caffau,E.,etal.2016,arXiv:1610.04427
Gehren,T.,Shi,J.R.,Zhang,H.W.,Zhao,G.,&Korn,A.J.2006,A&A,451,
1065
Goswami,A.,&Prantzos,N.2000,A&A,359,191
GustafssonB.,EdvardssonB.,ErikssonK.,Graae-Jørgensen U.,NordlundÅ.,
PlezB.,2008,A&A,486,951
Jacobson,H.R.,Thanathibodee,T.,Frebel,A.,etal.2014,ApJ,796,L24
Kobayashi,C.,Umeda,H.,Nomoto,K.,Tominaga,N.,&Ohkubo,T.2006,ApJ,
653,1145
Kobayashi,C.,Karakas,A.I.,&Umeda,H.2011,MNRAS,414,3231
Articlenumber,page8of8