The neutral gas extent of galaxies as derived from weak intervening CaII absorbers

0.92 MB·English

by  P. Richter

Astronomy&Astrophysicsmanuscriptno.ms15566 (cid:13)c ESO2011
January6,2011
The neutral gas extent of galaxies as derived from weak
ii ⋆
intervening Ca absorbers
P.Richter1,F.Krause1,C.Fechner1,J.C.Charlton2,andM.T.Murphy3
1 Institutfu¨rPhysikundAstronomie,Universita¨tPotsdam,Karl-Liebknecht-Str.24/25,14476Golm,Germany
e-mail:[email protected]
2 DepartmentofAstronomyandAstrophysics,PennsylvaniaStateUniversity,UniversityPark,PA16802,USA
3 CentreforAstrophysics&Supercomputing,SwinburneUniversityofTechnology,Hawthorn,Victoria3122,Australia
1 Received11August2010/Accepted03December2010
1
0 ABSTRACT
2
WepresentasystematicstudyofweakinterveningCaiiabsorbersatlowredshift(z < 0.5),basedontheanalysisofarchivalhigh-
n
resolution(R ≥ 45,000)opticalspectraof304quasarsandactivegalacticnucleiobservedwithVLT/UVES.Alongatotalredshift
a path of ∆z ≈ 100 we detected 23 intervening Caii absorbers in both the Caii H & K lines, with rest frame equivalent widths
J
5 Wabr,3ia93s4-c=or1re5c−ted79n9ummbÅeranddencsoiltuymonfwdeenaskitiinestelrovgenNin(CgaCiai)ii=a1b1so.2rb5e−rs1o3f.0d4N(o/bdtzai=ne0d.1b1y7fi±tti0n.g04V4oiagtt-hpzrofiil=ec0o.m35pofonrenatbss).orWbeersobwtaiitnh
log N(Caii) ≥ 11.65 (W ≥ 32 mÅ). This is ∼ 2.6 times the value obtained for damped Lymaanbsα absorbers (DLAs) at low
O] redshift.AllCaiiabsorberr,3s93i4noursampleshowassociatedabsorptionbyotherlowionssuchasMgiiandFeii;45percentofthem
haveassociatedNaiabsorption.FromionizationmodellingweconcludethatinterveningCaiiabsorptionwithlog N(Caii) ≥ 11.5
C arisesinDLAs,sub-DLAsandLyman-limitsystems(LLS)atHicolumndensitiesoflogN(Hi) ≥ 17.4.Usingsupplementary Hi
. informationfornineoftheabsorberswefindthattheCaii/HiratiodecreasesstronglywithincreasingHicolumndensity,indicating
h
acolumn-density-dependentdustdepletionofCa.TheobservedcolumndensitydistributionfunctionofCaiiabsorptioncomponents
p
followsarelativelysteeppowerlaw, f(N)∝N−β,withaslopeof−β=−1.68,whichagainpointstowardsanenhanceddustdepletion
-
o in high column density systems. The relativelylarge cross section of these absorbers together with the frequent detection of Caii
r absorptioninhigh-velocityclouds(HVCs)inthehalooftheMilkyWaysuggeststhataconsiderablefractionoftheinterveningCaii
t systemstrace(partly) neutral gasstructures inthehalosand circumgalacticenvironment of galaxies(i.e.,theyareHVC analogs).
s
a BasedontherecentlymeasureddetectionrateofCaiiabsorptionintheMilkyWayHVCsweestimatethatthemean(projected)Caii
[ cLo≥ve0ri.n0g5Lfr⋆acwtieocnaolcfuglaaltaextiheastatnhdetchheairragcateseriosutischraaldoisalisex(cid:10)tfec,nCtaIoI(cid:11)f=(pa0r.t3l3y.)UnesuintrgatlhgisasvaclluoeudasndwcitohnlsoidgeNri(nHgia)ll≥ga1l7a.x4ieasrowuintdhlluomw-inreodssithiiefst
2 galaxiesisR ≈55kpc.
HVC
v
1 Keywords.Keywordsshouldbegiven
0
2
2
1. Introduction determined what consequences the accretion mode has for the
.
8 star-formation activity of a galaxy. Part of this lack of under-
Theanalysisofinterveningabsorptionlinesinthespectraofdis-
0 standing(particularlyatlowredshift)isrelatedtoobservational
0 tantquasars(QSO) hasbecomean extremelypowerfulmethod limitations.Since, byfar,mostofthe iontransitionsofinterest
1 to study the distribution and physical properties of the inter- for QSO absorption spectroscopy are located in the ultraviolet
: galactic medium (IGM) and its relation to galaxies. Over the
v (UV),QSOabsorptionlineobservationsatlowredshiftarelim-
lastcoupleofdecades,QSOabsorptionspectroscopyofvarious
i ited by the number of bright extragalactic backgroundsources
X metalionssuchasMgiiandCiv,togetherwithgalaxyimaging,
forwhichhigh-resolutionUV spectroscopywithcurrentspace-
hasbeenusedextensivelytoconstrainthenatureofthevarious
r basedspectrographs(i.e.,HST/STISorHST/COS)canbecarried
a processes (infall, outflow, merging) that govern the matter ex-
out.At higherredshift,moreandmore UV linesare redshifted
changebetweengalaxiesandtheIGMaspartofthehierarchical
intotheopticalregime,allowingustoobtainamuchlargernum-
evolutionofgalaxies(e.g.,Bergeron& Boisse´ 1991;Steidelet
berofhigh-qualityQSO absorptionspectrawithlarge,ground-
al.1992;Churchilletal.1999;Steideletal.2002).
basedtelescopessuchastheESOVeryLargeTelescope(VLT).
Because of the physicaland spatial complexityof the IGM
Because galaxies at high redshift are dim, however, the char-
andthevarioushydrodynamicalprocessesthatareinvolved,our
acterization of the IGM-galaxy connection at high z using di-
knowledge about the exact role of the IGM-galaxy connection
rect observationsis a challenging task (Adelberger et al.2003;
fortheevolutionofgalaxiesatlow andhighredshiftis stillin-
Crighton et al.2010). Fortunately, hydrodynamicalsimulations
complete. In particular, it is not known whether the gas infall
of the IGM have turned out to be an extremely valuable tool
ontogalaxiesoccursintheformofcondensedneutralgasclouds
to assess the properties of intergalactic matter and its relation
(“cold” mode) or in the form of ionized gas (“warm” or “hot”
to galaxies at high and low redshifts (e.g., Fangano, Ferrara &
mode; see, e.g., Bland Hawthorn 2009). It also remains to be
Richter2007;Kaufmannetal.2009;Kacprzaketal.2010a).
⋆ Basedonobservations collectedattheEuropean Organisationfor Mostoftherecentabsorptionlinestudiesofinterveningsys-
AstronomicalResearchintheSouthernHemisphere,Chile tems have concentrated on the analysis of ionic transitions of
2 P.Richter,F.Krause,C.Fechner,J.C.Charlton,andM.T.Murphy:InterveningCaiiabsorbers
hydrogen and heavy elements in the UV (e.g, Hi, Mgii, Civ,
Ovi,andmanyothers).Thetwomostimportantopticalabsorp-
tion lines for IGM studies, the Caii H & K lines near 4000Å,
arelesscommonlyused,althoughtheyhavealsoplayedanim-
portantroleinourunderstandingofthediffusematterinspace.
Withanionizationpotentialof11.9eVCaiiservesasatracerof
warmandcold,neutralgasingalaxiesandtheirgaseoushalos.
After all, the Caii H & K lines were the first absorption lines
ever detected in interstellar gas (Hartmann 1904). These lines
were also the first tracers of extraplanar gas clouds in the halo
oftheMilkyWay(Mu¨nch1952)andwereamongthefirstmetal
absorptionlinesdetectedintheIGMin earlyQSO spectra(see
Blades1988).
Bowen (1991), Bowen et al.(1991), and, more recently,
Wild, Hewett & Pettini (2006, 2007), Zych et al.(2007, 2009)
and Nestor et al.(2008) have investigated intervening Caii ab-
sorberstowardsQSOsatrelativelylowspectralresolution.From
the early studies by Bowen et al. (using a small sample of
medium resolution optical QSO spectra) it was concluded that
Fig.2. Examplefora Voigt-profilefit (redsolid line) with four
interveningCaiiabsorptionpredominantlyoccurswithinthein-
absorptioncomponentsoftheCaii absorberatz = 0.23782to-
ner20kpcofgalaxies,thusspeakingagainstmanygalaxieshav-
wardsJ095456+174331.
ing extended Caii absorbing neutral gas halos with large fill-
ingfactors.Fromtheanalysisofseveralthousandlow-resolution
spectrafromtheSloanDigitalSkySurvey(SDSS)Wild,Hewett currentradiotelescopesmaybecrucialforourunderstandingof
& Pettini (2006) concluded that strong Caii systems with rest gasaccretionratesofgalaxiesatlowredshift(coldaccretionvs.
frameequivalentwidths> 350mÅariseonlyfromtheinnerre- warm accretion;see Bland Hawthorn2008).Because interven-
gions of galaxies and thus represent a subclass of the damped ing Caii absorbersshouldtrace neutralgasatcolumndensities
Lymanαabsorbers(DLAs).ThesehavethehighestHicolumn logN(Hi)> 17itisexpectedthattheyarerelativelyrare.Their
densities of all interveningQSO absorbers(log N(Hi) ≥ 20.3) numberdensitiesshouldbelowerthanthevalueofdN/dz≈0.6
andarebelievedtorepresentgas-richgalaxiesandprotogalactic atz=0.3,derivedfortheso-called“strong”Mgiisystems(sys-
structures. tems with rest frame equivalent widths ≥ 0.3 Å in the Mgii
λ2976line;seeNestor,Turnshek&Rao2005),buthigherthan
However, because of their limited spectral resolution and
that of DLAs (dN/dz = 0.045 at z = 0; Zwaan et al.2005).
their typically low signal-to-noise ratios (S/N), previous stud-
ies would have missed a population of weak, intervening Caii Therefore,a largenumberof QSO sightlinesmustbe surveyed
inordertostudythepropertiesofinterveningCaiisystemsona
absorberswithrestframeequivalentwidths≤ 100mÅ.Theex-
istence of these weak Caii systems was recently indicated by statisticallysecurebasis.
high-resolutionobservationsofweakCaiiabsorptioninthehalo Weherepresentthefirstsystematicstudyofweakinterven-
ing Caii absorbers in the redshift range z = 0−0.5, based on
andnearbyintergalacticenvironmentoftheMilkyWay(Richter
a very large sample of more than 300 optical, high-resolution
et al.2005, 2009; Ben Bekhti et al.2008; Wakker et al.2007,
2008).ItwasfoundthatCaiiabsorptioniscommoninthemany (R > 40,000) QSO absorption line spectra obtained with the
VLT. Our paper is organized as follows. In Sect.2 we present
neutralgasstructuresoutsidetheMilkyWaydisk(i.e.,intheso-
called “high velocity clouds”, HVCs), even at low Hi gas col- theobservations,thedatahandling,andtheanalysismethod.The
umn densities, N(Hi) ∼ 1017 − 1019 cm−2. If the Milky Way mainresultsofoursurveyarepresentedinSect.3.InSect.4we
environmentis typicalfor low-redshiftgalaxies,weak Caii ab- discussthephysicalpropertiesandtheoriginoftheweakinter-
veningCaiisystems.Wecompareourresultswithpreviousstud-
sorptionshouldariseinboththeneutralgasdisksofgalaxiesand
iesofinterveningCaiiabsorbersinSect.5.Finally,wepresenta
intheirextended,patchyneutralgashalos(i.e.,inHVCanalogs),
summaryofourstudyinSect.6.
buthighsensitivityisrequiredtodetectthehalocomponent.In
termsofthecommonQSOabsorptionlineclassificationthisim-
plies that Caii absorption should arise in DLAs as well as in 2. Observations,datahandling,andanalysis
a certain fraction of sub-DLAs (19.3 <log N(Hi) ≤ 20.3) and
method
Lymanlimitsystems(LLS;17.2 <log N(Hi) ≤ 19.3),depend-
ing on the Ca abundanceand dust depletion in the gas and the 2.1.Spectraselectionandreduction
localionizationconditions.
ForourCaiisurveywemadeuseoftheESOdataarchive1 and
ThecommondetectionofweakCaiiabsorptionintheMilky
retrieved all publically available (as of Sept. 2008)absorption-
Way’sextendedneutralgashalothereforesuggeststhattheoc-
line data for low- and high-redshift QSOs observed with the
currence of Caii absorption in the gaseous outskirts of low-
Ultraviolet and Visual Echelle Spectrograph (UVES) on the
redshiftgalaxiesshouldbereassessedwithhighresolution,high
VLT. This enormousdata archive (Spectral Quasar Absorption
S/N absorption line data. Such a study may be very useful to
Database, SQUAD; PI: M.T. Murphy) provides high-quality
constrain the absorption cross section of neutral gas in the lo- spectraldatafor∼400quasarsandactivegalacticnuclei(AGN).
cal Universe. Most importantly, it can be used to measure the
Most of these spectra were taken in the UVES standard con-
sizesandmassesofneutralgasgalaxyhalos.Thelatterpointis
figuration using the 1” slit, providing a spectral resolution of
particularlyinteresting,asknowledgeabouttheneutralgasdis-
tributionaroundgalaxiesbelowthe Hi21cmdetectionlimitof 1 http://archive.eso.org
P.Richter,F.Krause,C.Fechner,J.C.Charlton,andM.T.Murphy:InterveningCaiiabsorbers 3
Fig.1. UVESspectrumofthequasarJ215501−092224inthewavelengthrangebetween4240and4300Å.WeakinterveningCaii
absorptionisclearlydetectedatz=0.08091inbothCaiilines,asindicatedbythetickmarks.
Table1.QSOnames,redshiftsandionsof23interveningCaiiabsorbersatz=0−0.5detectedinourQSOsample
QSO Alt.Name z Biasa DetectedIonsb UndetectedIonsb
abs
J121509+330955 Ton1480 0.00396 yes Caii,Nai
J133007-205616 PKS1327-206 0.01831 no Caii,Nai
J215501-092224 PHL1811 0.08091 no Caii Nai
J044117-431343 HE0439-4319 0.10114 yes Caii,Nai
J095456+174331 PKS0952+179 0.23782 yes Caii,Mgii,Feii
J012517-001828 PKS0122-005 0.23864 no Caii,Mgii
J235731-112539 PKS2354-117 0.24763 yes Caii,Nai,Mgii
J000344-232355 HE0001-2340 0.27051 no Caii,Mgii,Feii Nai
J142249-272756 PKS1419-272 0.27563 yes Caii,Mgii,Feii Nai
J042707-130253 PKS0424-131 0.28929 no Caii,Mgii,Feii
J113007-144927 PKS1127-145 0.31273 yes Caii,Mgii,Feii
J102837-010027 B1026-0045B 0.32427 no Caii,Mgii,Feii
J231359-370446 PKS2311-373 0.33980 no Caii,Mgii,Feii
J110325-264515 PG1101-264 0.35896 no Caii,Nai,Mgii,Feii
J094253-110426 HE0940-1050 0.39098 no Caii,Mgii,Feii
J121140+103002 B1209+1046 0.39293 no Caii,Mgii,Feii
J123200-022404 PKS1229-021 0.39498 yes Caii,Mgii,Feii
J050112-015914 PKS0458-020 0.40310 no Caii,Mgii Nai
J224752-123719 PKS2245-128 0.40968 no Caii,Mgii,Feii
J220743-534633 PKS2204-540 0.43720 yes Caii,Mgii,Feii Nai
J044117-431343 HE0439-4319 0.44075 no Caii,Mgii Nai
J144653+011356 B1444+0126 0.44402 no Caii,Mgii,Feii
J045608-215909 HE0454-2203 0.47439 no Caii,Feii
aBiasflagindicateswhethertheobservationsweretargetedobservations(seeSect.3.2)
bOnlyabsorptionbyCaii,Mgii,Feii,andNaiisconsideredinthisstudy
R ∼ 45,000 (corresponding to a velocity resolution of ∼ 6.6 in the wavelength range between 3934 and 5955 Å. From all
kms−1 FWHM). Only a few spectra have been observed at 397availablequasarspectraintheSQUADsampleweselected
slightlyhigherspectralresolution,uptoR ∼ 60,000.Thespec- the304spectrathata)fullyorpartlycovertheabovementioned
tral coverage as well as the signal-to-noise ratio (S/N) varies wavelength range, and b) have an average S/N of ≥ 12 per ∼
substantiallyamongthespectra,reflectingthevariousscientific 6.6kms−1 wideresolutionelementforR = 45,000;seeabove).
goalsoftheoriginalproposals. NotethatwerequiretheoccurrenceofbothCaiiabsorptionlines
inthespectrumforapositivedetection(seealsoSect.2.2).The
AlldatawerereducedusingamodifiedversionoftheUVES
minimumcolumndensitythatcanbedetectedat4σsignificance
reductionpipeline.Themodificationswereimplementedtoim-
for an unresolved,optically thin absorptionline at wavelength,
provethefluxextractionandwavelengthcalibrationforthedif-
λ , with oscillator strength, f, in a spectrum with a given S/N
ferent echelle orders. Different orders then were combined us- 0
per resolution elementand a spectralresolution R = λ/∆λ can
ingthecustom-writtencodeUVES poplerwithinversevariance
be calculated (e.g., Tumlinson et al.2002; Richter et al.2001)
weightingandacosmicrayrejectionalgorithm,toformasingle
via
spectrumwith a dispersionof2.5 kms−1 per pixel.Finally,the
spectra were continuum-normalizedwith an automated contin-
uumfittingalgorithm.
BecauseweaimtoanalyseCaiiH&K(λλ3934.77,3969.59) 4
N ≥1.13×1020cm−2 . (1)
absorption in the redshift range z = 0−0.5, we are interested R(S/N)f (λ /Å)
0
4 P.Richter,F.Krause,C.Fechner,J.C.Charlton,andM.T.Murphy:InterveningCaiiabsorbers
clearlydetectedandthatthetwolinesshowasimilarpatternof
absorptioncomponentsandsimilarlineshapes.
Forapositivedetectionatredshiftz≥0.2wefurtherrequire
that the Caii absorptionis accompaniedby absorptionin Mgii
or Feii. For all possible Caii systems atz ≥ 0.2a Mgii and/or
astrongFeiitransitioniscovered.Thedetectionofoneofthese
ions at the same redshift as Caii proves the presence of an in-
terveningCaiisystem,whereasthesignificantnon-detectionof
Mgii and Feii indicates a false detection. The mandatory co-
existence of Mgii and Feii with Caii is justified given relative
abundancesoftheseions,theirionizationpotentials,andtheos-
cillatorstrengthsof theabovelisted transitions.Caii absorbing
cloudswithoutsignificantMgiiandFeiiabsorptionarenotex-
pectedtoexist.
The situation is different for redshifts z < 0.2, where we
havenocoverageofassociatedMgiiandFeiiabsorptioninour
opticalspectra.Generally,thesignificantnon-detectionofNaiin
aCaiicandidatesystemdoesnotindicateafalseCaiidetection.
Naiexistsonlyinrelativelydenseneutralgasandthusispresent
Fig.3. Relative redshift coverage for Caii absorption for 270 onlyinasubsetoftheCaiisystems.Fortunately,inourparticular
QSOsigthlines,alongwhichinterveningCaiiabsorptioncanbe survey,allbutoneCaiicandidatesystemsatz < 0.2forwhich
detectedatcolumndensitieslogN(Caii)>11.65. Naiwascoveredintheopticalspectrum,didhappentohaveNai
detected.Furthermore,theonlyCaiicandidatesystematz<0.2
for which Nai was not detected (the system towards J215501-
In our case we have λ = 3969.59 Åand f = 0.3145 for 092224),couldbeconfirmedbyexistingUVabsorptionlinedata
0
the weaker of the two Caii lines, so that the minimum S/N fromHST/STIS(Jenkinsetal.2003).
of 12 in our sample corresponds to a column density thresh-
old of N(Caii) = 6.7 × 1011 cm−2 (log N(Caii) = 11.83) for Finally,asachecktoourprocedure,andinordertoidentify
weaker Caii absorbers, we reversed our search and looked for
R=45,000.Thiscolumndensitylimitcorrespondstoanequiv-
Caiiabsorptionin(previouslyidentified)strongMgIIabsorbers
alentwidthlimitof50mÅ inthestrongerCaiiλ3934line.
atz = 0.2−0.5.Inthisway,weidentifiedfourveryweakCaII
Forall304selectedQSOspectrawethenscannedthespec-
absorbers,with N(Caii)< 11.5,whichwehadmissedinourdi-
tral region of interest (3934−5955 Å) and excludedfrom fur- rect search for Caii. We also founda numberof Caii absorber
theranalysisthoseregionsthatarecontaminatedbyLyαforest candidates,forwhichonlyoneofthetwoCaiilinesisseen,e.g.,
linesandotherspectralfeaturesthatwouldcausesevereblend- in systems so weak that Caiiλ3969 was not formally detected,
ing problems. As a result, we define for each sightline a char- orinwhichonememberoftheCaiidoubletwasblendedorwas
acteristicredshiftpathdz ≤ 0.5(typicallycomposedofsev-
CaII notcoveredinthespectrum(seeJonesetal.2010foranexample
eralredshiftchunks)alongwhichinterveningCaiicouldbede-
case).Thesecandidatesystemsarenotconsideredanyfurtherin
tected. The mean redshift path per sightline in our sample is
our analysis, but they are listed in the Appendix in Table A.3.
hdzCaIIi=0.33;thetotalredshiftpathis∆z=100.60. We moreoverexcluded all local Caii absorptionfeatures at ra-
Atablelistingallselected304spectrawiththeQSOnames, dial velocities 0−500 kms−1, as these features are caused by
theaverageS/N,anddz isprovidedintheAppendix(Tables
CaII neutralgasin the Milky Way disk and halo (see Ben Bekhtiet
A.4andA.5).
al.2008).
Usingtheaboveoutlineddetectioncriteriaafinallistofver-
2.2.Absorberidentificationandlinefitting
ified Caii absorberswas produced.As an example we show in
Thenextstep inouranalysiswas thesearchforCaii H&Kab- Fig.1theUVESspectrumofthequasarJ215501−092224inthe
sorption at z = 0−0.5 along the 304 selected QSO sightlines. wavelengthrangebetween4240and4300Å.Weakintervening
First,weusedanautomatedline-finderalgorithmtoidentifyab- Caii absorptionis clearly detectedat z = 0.08091in both Caii
sorptionfeatureswhosewavelengthswouldcorrespondtocom- lines(indicatedbythetickmarks).Forthefurtherspectralanaly-
binedCaiiH&Kabsorptionininterveningabsorbersatz≤ 0.5. sisoftheabsorptionlineswemadeuseoftheFITLYMANpackage
In this way we created a candidate list of possible Caii sys- implementedintheESO-MIDASanalysissoftware(Fontana&
tems. Secondly, two of our team (Richter & Krause) indepen- Ballester 1995).Thisroutineuses a χ2 minimizationalgorithm
dently analysed all 304 spectra by eye and identified possible to derive column densities (N) and Doppler parameters(b) via
Caii systems. All three candidate lists then were merged into multi-component Voigt profile fitting, also taking into account
one.Foreachofthecandidatesystemswethencheckedforas- thespectralresolutionoftheinstrument.Anexampleforamul-
sociated absorption in Mgii λλ2796,2803, Feii λλ2586,2600, ticomponentVoigtprofilefit of an interveningCaii absorberis
andNaiλλ5891,5897toverifyorexcludethepresenceofinter- shown in Fig.2. Total Caii column densities for each system
veninggasabsorptionattheredshiftindicatedbythecandidate havebeendeterminedbysummingoverthethecolumndensities
Caiiabsorption.Alargenumberoffalsedetectionsisexpected, of the individual subcomponents. The total rest frame equiva-
caused by interveningabsorptionlines that coincidentlymimic lentwidths(W)andvelocitywidthsoftheabsorption(∆v)were
r
the wavelength pattern expected for a redshifted Caii doublet. derived for Caii λ3934, as well as for associated Mgii λ2796
ForapositivedetectionwethusrequirethatbothCaiilinesare absorption,byadirectpixelintegration.
P.Richter,F.Krause,C.Fechner,J.C.Charlton,andM.T.Murphy:InterveningCaiiabsorbers 5
Table2.Columndensities,absorptioncomponents,velocitywidthsandequivalentwidthsforthe23interveningCaiiabsorbers
QSO z logN(Caii) na ∆v W ∆v W logN(Nai)
abs c CaII r,CaII,3934 MgII r,MgII,2796
(Nin[cm−2]) [kms−1] [mÅ] [kms−1] [mÅ] (Nin[cm−2])
J121509+330955 0.00396 12.31±0.01 3 149 160±9 ... ... 11.64±0.04
J133007-205616 0.01831 13.04±0.06 9 378 799±46 ... ... 13.28±0.05
J215501-092224 0.08091 11.76±0.02 1 38 44±5 ... ... ≤10.49
J044117-431343 0.10114 12.60±0.05 7 130 292±6 ... ... 12.25±0.03
J095456+174331 0.23782 12.22±0.07 4 105 119±7 173 1101±12 ... ...
J012517-001828 0.23864 11.48±0.06 1 22 34±4 154 328±30 ... ...
J235731-112539 0.24763 12.70±0.03 3 53 261±16 ≥470 ≥2244 12.41±0.02
J000344-232355 0.27051 11.66±0.02 2 42 33±2 176 850±7 ≤10.30
J142249-272756 0.27563 12.07±0.03 1 29 80±6 302 1260±50 ≤10.90
J042707-130253 0.28929 12.03±0.03 2 122 67±2 158 368±13 ... ...
J113007-144927 0.31273 12.71±0.01 8 230 376±7 378 1789±12 ... ...
J102837-010027 0.32427 12.43±0.02 2 143 224±14 165 681±20 ... ...
J231359-370446 0.33980 12.81±0.04 3 41 166±5 149 976±16 ... ...
J110325-264515 0.35896 11.26±0.02 1 31 15±2 148 532±15 11.92±0.01
J094253-110426 0.39098 11.89±0.02 2 31 54±2 ≥178 ≥900 ... ...
J121140+103002 0.39293 11.38±0.06 1 30 18±3 355 1114±15 ... ...
J123200-022404 0.39498 12.39±0.05 5 111 194±6 297 2018±10 ... ...
J050112-015914 0.40310 12.26±0.03 3 103 132±17 ... ≥715 ≤11.20
J224752-123719 0.40968 12.23±0.02 2 101 140±11 448 1013±21 ... ...
J220743-534633 0.43720 11.98±0.05 1 26 65±8 68 318±15 ≤11.30
J044117-431343 0.44075 11.42±0.07 3 60 24±3 72 317±11 ≤10.80
J144653+011356 0.44402 11.25±0.06 1 42 21±6 188 219±10 ... ...
J045608-215909 0.47439 12.18±0.07 4 102 122±5 ... ... ... ...
an isthenumberofCaiiabsorptioncomponentspersystem;seeTablesA.1andA.2
c
3. Results tionbiasseeSect.3.2).Theredshiftdistributionoftheabsorbers
(Fig.4,rightpanel)isnonuniform,showinganenhancementof
3.1.FrequencyandcolumndensitiesofCaiiabsorbers Caii systemsin the rangez = 0.2−0.4.Thisisbecauseof the
inhomogeneous redshift coverage of our QSO sample and not
Basedontheanalysismethoddescribedintheprevioussection
becauseofan intrinsicredshiftevolutionof the absorberpopu-
we found23interveningCaii absorbersalongthe selected 304
lation(seeSect.3.2).Themedianabsorptionredshiftinourtotal
QSO sightlines. The redshift of the absorbers ranges between
sample (23 systems) is z = 0.32 (0.35 in the bias-corrected
z=0.00396andz=0.47439.Notethatseveraloftheseabsorp- abs
sample).
tion systems represent DLAs and LLSs that have been identi-
fied and studiedpreviously,based on the analysisof ionsother
thanCaii(e.g.,Mgii).Consequently,itisexpectedthatsomeof 3.2.NumberdensityofCaiiabsorbers
ourQSOsightlinesthatexhibitinterveningCaiiabsorptionhave
been explicitly selected (based on lower resolution spectra) to To calculate the number density of intervening Caii absorbers
study the absorption characteristics of DLAs and LLSs at low perunitredshifz,dN/dz,itisnecessarytoconsiderindetailthe
z, and thus represent targeted observations. Therefore, one im- completeness of our Caii survey and the selection bias in our
portant concern about the detection rate of intervening Caii in datasample.
ourdataistheselectionbiasintheQSOsamplethatweareus- The weakest Caii system in our survey has log N(Caii) =
ing. This issue will be further discussed in Sect.3.2, where we 11.25, but only 103 of the 304 spectra have a sufficiently high
estimatethenumberdensityofCaiiabsorbersatlowredshift. S/N to detect Caii at this column density level (see Eq.1).
Quasar names,redshiftsanddetected(andundetected)ions EighteenCaii absorbershavelog N(Caii) > 11.65(thesewere
forthe23CaiiabsorptionsystemsarelistedinTable1.Detailed found by the direct search for interveningCaii absorption;see
results from the Voigt-profile fitting of the absorbers and their Sect.2.2)andfor270outof304sightlinestheselectedspectral
subcomponentsarelistedinTablesA.1andA.2intheAppendix. regionsaresensitiveabovethiscolumndensitylevel.Forcom-
VelocityplotsofCaiiandotherdetectedionsforall23absorbers pleteness reasons we therefore consider only the total redshift
are shown in the Appendix, in Figs.A.1 to A.6. Total column path covered by these 270 sightlines, together with the 18 ab-
densities, absorption components, velocity widths, and equiva- sorbers that have log N(Caii) > 11.65, for the estimate of the
lent widths for the 23 Caii absorbers and for their associated Caiiabsorbernumberdensity.
MgiiabsorptionaregiveninTable2. The relative redshift coverage, f (z), for Caii absorption
z
ThetotallogarithmicCaiicolumndensitiesoftheabsorbers alongtheselected270sightlines(fz(z)=Pdz(z)/270)isshown
rangebetweenlogN(Caii) = 11.25and13.04.Thedistribution in Fig.3. The distribution reflects blending issues and low S/N
of the (total) logarithmic Caii column densities of the 23 ab- for many high-redshift sightlines in the range zCaII < 0.25, as
sorbersaswellasthedistributionoftheirredshiftsareshownin wellasindividualfeaturescausedbytheintegratedwavelength
Fig.4.TheCaiicolumndensitydistributionhasapeaknearlog coverageoftheoriginalUVESobservationsinoursample.The
N(Caii) ∼ 12.2, which is also the median value. If we remove totalredshiftpathfortheselected270sightlinesis∆z=89.15.
theeightsystemsthatareflaggedas“biased”inTable1,theme- To quantify the selection bias in our data we inspected the
dianvalueforlogNisreducedto∼11.9(fordetailsontheselec- originalproposalabstractsintheUVESarchiveforthesightlines
6 P.Richter,F.Krause,C.Fechner,J.C.Charlton,andM.T.Murphy:InterveningCaiiabsorbers
Fig.4.Distributionof(total)logarithmicCaiicolumndensitiesofthe 23absorbersdetectedin ourtotalQSO sample (leftpanel)
andtheirredshiftdistribution(rightpanel).
Fig.5. Distribution of Caii λ3934 rest frame equivalent widths for the 23 Caii absorbers (left panel) and their velocity width
distribution(rightpanel).
along which interveningCaii was found.As it turnsout, 8 out above a column density limit of log N = 11.65 arise exclu-
of the 18 QSO spectra with intervening Caii absorption at log sivelyinstrongMgiisystems, butnotinweakMgiiabsorbers.
N(Caii) > 11.65 represent data from targeted observations of Therefore, the number densities of strong Mgii absorbers and
DLAsandMgiiabsorbersatlowredshift(asindicatedinTable CaiisystemswithlogN(Caii)>11.65canbedirectlycompared
1, Col.4). This implies that our sample contains ∼ 40 percent toeachother.FromourdataweestimatedN/dz≈1.0forstrong
moreCaiisystemsperunitredshiftthaninafullyrandomQSO Mgiisystemsathzi=0.3.ThestudybyNestor,Turnshek&Rao
sample. (2005)basedonSDSSdata,however,indicatesalowervalueof
dN/dz ≈ 0.6 for this redshift. These numbers further suggest
Another way to check for a possible selection bias in our a substantial, ∼ 40 percentoverabundanceof strong Mgii sys-
data is to compare the frequency of Mgii absorption systems tems, and thus Caii systems with log N(Caii) > 11.65 in our
in our sample with that of securelyunbiasedabsorbersearches
QSOsample,owingtoaselectionbias.
(e.g., from spectra of the Sloan Digital Sky Survey, SDSS).
InterveningMgiiabsorberscommonlyaredividedinto“weak” Without any bias correction our detection of 18 Caii ab-
systems (for rest-frame equivalent widths in the λ2976 line, sorbers with log N(Caii) > 11.65 along a total redshift path
W ≤ 300 mÅ) and “strong” systems (W ≥ 300 mÅ). of ∆z = 89.15 implies dN/dz(Caii) = 0.202± 0.048 for this
2796 2796
From a Mgii-selected absorber search of our data we find column density range. Considering the above estimate for the
that strong Mgii systems outnumber Caii systems with log selectionbias(40percent)andremovingtheeightrelevantsight-
N(Caii) > 11.65 by a factor of ∼ 5. Moreover, Caii systems linesfromthestatisticswederiveabias-correctednumberden-
P.Richter,F.Krause,C.Fechner,J.C.Charlton,andM.T.Murphy:InterveningCaiiabsorbers 7
Fig.6.DistributionofDopplerparameters(bvalues)ofthe69Caiiabsorptioncomponentsinour23Caiisystems(leftpanel),and
numberofabsorptioncomponentspersystem(rightpanel),derivedfromVoigt-profilefitting.
sity of intervening Caii systems with log N(Caii) > 11.65 of
dN/dz(Caii) = 0.117±0.044.Theerrorincludesa 20percent
uncertaintyestimateforthebiascorrection.Forcomparison,the
number density of DLAs at low redshift is dN/dz(DLA) ≈
0.045 (Zwaan et al.2005), thus a factor of 2.6 lower. This is
an interesting and important result: weak intervening Caii ab-
sorbers at low redshift outnumber DLAs by a factor of two to
three.
Ifwetransformdzintothecomovingabsorptionpathlength
dXvia
dX =(1+z)2[Ω +Ω (1+z)3]−0.5dz, (2)
Λ m
the data indicate dN/dX = 0.075 for hz i = 0.35 (bias
abs
corrected).Thoughoutthispaperwe use a flat ΛCDM cosmol-
ogywithH = 73kms−1Mpc−1,Ω = 0.238,andΩ = 0.762
0 m Λ
(Spergeletal.2007).
3.3.Caiiequivalentandvelocitywidths
The Caii rest-frame equivalent widths for the stronger λ3934
transition,Wr,3934,rangefrom15to799mÅ inoursample(see Fig.7. Column density distribution function, f(N), for Caii
Table 2). The equivalent width distribution is shown in Fig.5, components, as derived from our bias-corrected Caii absorber
left panel. Most absorbers (18 out of 23) have Wr < 200 mÅ; sample. Above our completeness limit (log N = 11.65; dotted
notethatthesesystemswouldbeinvisibleinspectraldatawith line),thedatafitbesttoapowerlawoftheform f(N) =CN−β
lowspectralresolutionand/orlowS/N(e.g.,inmostSDSSspec- withslopeβ=1.68±0.20(solidline).
tra). The median rest-frame equivalentwidth in the λ3934 line
is 118 mÅ (76 mÅin the unbiased sample). If we divide our
absorber sample into “strong” absorbers with W ≥ 300 mÅ absorption profiles. The absorption width is defined as the to-
r
and “weak” absorbers with W < 300 mÅ (similar to the di- talvelocityrangeinwhichCaiiλ3934absorption(includingall
r
visions for Mgii), our sample contains 21 weak absorbers, but velocity subcomponents)is detected in a system. The Caii ab-
onlytwostrongsystems.OurstudythusindicatesthatweakCaii sorption widths range from 22 to 378 kms−1, but only two of
absorberswith32≤W <300mÅ2areeighttimesmorenu- the 23 systems show large widths with ∆v > 200 kms−1. For
r,3934 ∆v < 200 kms−1 the distribution of the 21 systems is bimodal
merousthanstrongabsorbers.
with 10 narrow absorbers with ∆v < 50 kms−1, and 9 some-
Figure 5, right panel, shows the distribution of the absorp-
tion width of the Caii systems, ∆v , as derivedfromthe Caii whatbroadersystemswithvelocitywidthsbetween100and150
abs kms−1. The median value for the velocity width of Caii is 60
2 32 mÅcorresond toour completeness level of log N = 11.65 for kms−1 (40 kms−1 for the unbiased sample). While the typical
opticallythinCaiiabsorption inanunresolved, single-component ab- velocity spread of interveningCaii absorbers is < 150 kms−1,
sorptionline. theassociatedMgiicoversanabsorptionrangethattypicallyis
8 P.Richter,F.Krause,C.Fechner,J.C.Charlton,andM.T.Murphy:InterveningCaiiabsorbers
severaltimes largerthanthat forCaii. Thiswill be furtherdis- Caiiabsorptioncomponentsismildlysteeperthanthatofinter-
cussedinSect.3.6. veningHiandMgii.Notethatasimilarslope(β=1.6±0.3)has
beenfoundfortheCaiiCDDFinHVCsinthehalooftheMilky
Way(BenBekhtietal.2008).
3.4.Dopplerparametersandsubcomponentstructure
The relatively steep slope of the Caii CDDF most likely is
The distribution of the Dopplerparametersof all individual69 related to the depletion of Ca into dust grains. Dust depletion
absorbersubcomponentsisshownintheleftpanelofFig.6.The is particularly important for high column density systems (see
distribution peaks at b = 4 kms−1 with a broad tail that ex- Sects.3.6and4.2).Thiseffectleadsto a reductionof highcol-
tends to b = 20 kms−1. The median value is b = 6 kms−1. umndensityCaiisystemsandthusisexpectedtocauseasteep-
In absorption spectroscopy of extendedgaseous structures it is eningof f(N)comparedtoundepletedelements.Notethatifwe
usually assumed that the total (=measured) Doppler parameter fittheCDDFtoall69absorptioncomponentsinthetotaldataset
of an absorber is composed of a thermal component (bth) and (i.e.,includingthe biased absorbersand their components),the
a non-thermal component (bturb), so that b2 = b2th +b2turb. The slopeoftheCDDFsteepenssubstantiallywithβ = 2.12±0.17.
thermal component of b depends on the gas temperature, T, Obviously, the biased systems predominantly add components
and the mass (m) or atomic weight (A) of the particle, where with relatively low Caii column densities to the total sample,
bth = (2kT/m)1/2 ≈ 0.129(T[K]/A)1/2kms−1.Thenon-thermal resulting in a steeper CDDF. This most likely is related to an
component includes macroscopic motions in the gas, such as enhanceddustdepletioninthebiasedsystems.
turbulence and flows. Assuming pure thermal broadening(i.e.,
b = 0 kms−1), Caii Doppler parameters of (5,10,15,20)
turb
kms−1thereforecorrespondtologarithmicgastemperatures,log 3.6.AssociatedMgii,FeiiandNaiabsorption
(T/K), of (4.8,5.4,5.7,6.0).Since the temperature of the Caii
AllCaiiabsorbersinoursampleforwhichinformationonMgii
absorbing gas is likely to be much lower than that (a charac-
teristic temperature range is T = 102 − 104 K; see Richter et and Feii is available (Mgii: 18/23, Feii: 15/23; see Table 1
and Figs.A1-A5) do show associated absorption in these ions,
al.2005; Ben Bekhti et al.2008), this would suggest that the
measured Caii b values are dominatedby turbulentmotions in following our original selection criteria (Sect.2.2). The equiv-
alent widths of the associated Mgii λλ2796,2803 and Feii
thegas.Alternatively,theobservednon-thermallinewidthsmay
λλ2586,2600 absorption lines are typically much larger than
bepartlyduetounresolvedsubcomponentstructureinthelines,
those of the Caii H&K absorption.This is expected,since Mg
which is plausible in view of the limited spectral resolution of
thedata(FWHM≈ 6.6kms−1 formostofthespectra).Ineither is 17 times and Fe is 14 times more abundantthan Ca (assum-
case, the measured Doppler parameters of the Caii absorption ingrelativesolarabundancesfortheseelements:log(Ca/H)⊙ =
components unfortunately cannot be used to constrain the gas −5.69 ± 0.04; log (Na/H)⊙ = −5.83 ± 0.04; log (Mg/H)⊙ =
temperatureintheabsorbers. −4.47±0.09;log(Fe/H)⊙ =−4.55±0.05;Asplund,Grevesse&
Sauval2005)andtheoscillatorstrengthsoftheabovementioned
Figure 6, right panel, shows the number of absorption
components that we resolved in each Caii system. Seventeen MgiiandFeiitransitionsarerelativelystrong.MgiiandFeiiab-
out of the 23 Caii absorbers (i.e., 74 percent) have three or sorptionthat is associated with Caii systems showsa velocity-
fewer absorptioncomponentswithin a velocityrangeof ≤ 150 componentstructurethattypicallyisfarmorecomplexthanthat
kms−1,andsevensystemsaresingle-componentCaiiabsorbers. ofCaii.ThemedianvelocitywidthofMgiiis∆v=175kms−1,
The mean velocity separation between Caii absorption sub- thussubstantiallylargerthanthemedianvelocitywidthofCaii
components is only ∼ 25 kms−1, which suggests that the in- (60 kms−1; see above). This behaviour indicates that Mgii ab-
sorption traces more extended gaseous structures that have a
dividualsubcomponentsare physically connected (e.g., as part
larger velocity extent, while Caii absorption is detectable only
of a coherentgas structure). Note that the velocity width of an
in certain, spatially more confined regions within these struc-
absorber(seeabove)iscorrelatedwiththethenumberofabsorp-
tures.ThisisbecauseMgii-inviewofitslargerabundanceand
tioncomponents(seeTable3).
higherionizationpotential-isamuchmoresensitivetracerfor
diffuse neutral gas than Caii, and it also traces diffuse ionized
3.5.Columndensitydistributionfunction(CDDF) gas (see Sect.4.1). Therefore, the velocity pattern of Mgii ab-
sorptionreflectstheoveralldistributionofdenseanddiffusegas
In Fig.7 we show the column density distribution function
intheabsorber,whileCaiiabsorptionarisesonlyintheregions
(CDDF)ofthe37Caiiabsorptioncomponentsfoundintheun-
thathavethehighestneutralgascolumndensities.
biased absorber sample. Following Churchill et al.(2003), the
CDDF can be written as f(N) = m/∆N, where m denotes the AllbutoneoftheCaiiabsorbersinoursample,forwhichin-
numberofabsorbersinthecolumndensitybin∆N. Integration formationonMgiiisavailable,areassociatedwithstrongMgii
of f(N)overthetotalcolumndensityrangeofinterestthende- absorbers. However, from our Mgii selected absorber sample,
liversthetotalnumberofabsorbersinthatrange.TheCDDFof wefindthatonlyeveryfifthstrongMgiisystematz≤0.5shows
low and high ion absorbers in intervening absorption line sys- Caii absorption above log N(Caii) = 11.65 (see Sect.3.2).
temsusuallyfollowsapowerlawintheform Themedianrest-frameequivalentwidthoftheMgii λ2976ab-
sorption that is associated with a Caii system is 700 mÅ, thus
f(N)=CN−β, (3) about six times larger than the median Caii equivalent width
in the λ3934 line. As shown in Fig.8 (left panel), the Mgii
where β varies in the range between 1 and 2 for differ- λ2976 equivalent width correlates with the Caii λ3934 equiv-
ent ions (e.g., β ≈ 1.5 for Hi and Mgii; Kim, Christiani & alentwidth,whilethevelocitywidthsoftheCaiiandMgiiab-
D’Odorico2001; Churchill et al.2003). Fitting f(N) to the 37 sorptiondonotshowsuchacorrelation(Fig.8,rightpanel).The
Caii absorption components we find β = 1.68± 0.20 and log formeraspectindicatesthattheCaiiabsorptionfollowsonlythe
C = 9.05±2.42forallabsorbersthatare aboveourcomplete- strongestMgiiabsorptioncomponents,whichdominatethetotal
nesslimit(logN(Caii)=11.65).Thus,theCDDFofintervening Mgii equivalentwidth.Thisis in linewith theobservationthat
P.Richter,F.Krause,C.Fechner,J.C.Charlton,andM.T.Murphy:InterveningCaiiabsorbers 9
umndensityrangenorthedistributionofimpactparametersfor
theabsorberslistedinTable3arestatisticallyrepresentativefor
intervening Caii absorbers. The data for five out of the seven
systemsarenottakenfromarandomabsorbersample,butcome
fromtargetedobservationsofDLAsandsub-DLASatlowred-
shift. Thus, we expect that Table 3 is strongly biased towards
high column density systems and low impact parameters.Still,
thesupplementaryinformationlistedinTable3providesimpor-
tantinformationonthephysicalconditionsinthesesystems,as
willbediscussedbelow.
InFig.9we showtherelationbetweenlog N(Caii) andlog
Fig.8.ComparisonbeweenCaiiandMgiiequivalentandveloc- N(Hi)forthesesevensystems(leftpanel),aswellastherelation
itywidthsintheinterveningCaiiabsorbers. betweenthe log(Caii/Hi) = log (N(Caii)/N(Hi)) scaled to the
solar(Ca/H)ratioandlogN(Hi)(rightpanel).Despitethescat-
terthereisaclearcorrelationbetweentheCaiicolumndensity
andtheHicolumndensity,whichcanbefittedas
thereisnoobviousvelocityoffsetbetweenthemainCaiiandthe
mainMgiiabsorptioncomponents.Thelatteraspectshowsthat logN(CaII)=8.415+0.193[logN(HI)], (4)
weakerMgii satellite componentsareeithernottracedby Caii
(as they may arise in predominantlyionized gas rather than in in Fig.9 (left panel) indicated as solid line. The slope of
neutralgas)ortheyaretooweaktodetectinCaiiabsorption. this relation agrees remarkably well with the slope found for
For11ofthe23absorbers,informationonNaiisavailable, the Caii/Hi ratio in Milky Way disk and halo clouds (Wakker
andfiveofthesesystemsareindeeddetectedinNaiabsorption. & Mathis 2000),but the Milky Way data points are systemati-
TheNaicolumndensitiesinthesystemswhereNaiisdetected callyoffsetineitherthexorydirection.Mostlikely,thisoffset
range between log N(Nai) = 11.64 and 13.28 (see Table 2). is in y and is caused by an above average dust-to-gas ratio in
The upper limits for log N(Nai) for the remaining six systems theMilkyWay,comparedtointerveningabsorbers.Thiswould
varybetween10.30and 11.20.Unlikethe associated Mgii and causeahigherdepletionofCaintodustgrains,sothatthemea-
Feii absorption, the observed Nai component structure in the suredgasphaseCacolumndensitiesintheMilkyWayabsorbers
five systems that show Nai absorption follows that of the Caii would be systematically lower. The observedrelation indicates
absorption very closely (see Figs.A1-A5). This indicates that thatitneedsanincreaseofaboutfiveordersofmagnitudeinthe
Caii and Nai absorption arises in the same physical regions. totalneutralgascolumndensity(from1017to1022cm−2)tolift
Neutralsodiumhasanionizationpotentialofonly5.1eV(com- the gas phase Caii column density by one order of magnitude
pared to 11.9 eV for Caii; e.g., Morton 2003) and thus is be- (from1011.7 to1012.7 cm−2).Asaconsequence,theCaii/Hira-
lievedtoserveastracerforcoldneutralgasinsidegalaxies(e.g., tiodecreasesstronglywithincreasingHicolumndensity(Fig.8,
Crawfordetal.1992).However,Naiabsorptionisoccasionally rightpanel).Thistightrelationinourdatacanbeapproximated
found also in HVCs in the halo of the Milky Way (Richter et bythefit
al.2005; Ben Bekhti et al.2008)and in the winds of starburst-
log(CaII/HI)=8.374−0.807[logN(HI)], (5)
inggalaxies(Martinetal.2005).Wewillfurtherinvestigatethe
Nai/Caii relation in our absorbers in Sect.4.1, where we also
as shown as solid line in the right panel of Fig.9. The ob-
discusstheionizationanddustpropertiesinthegas.
served decline of Caii/Hi over three orders of magnitude as a
functionof N(Hi) cannotbeprimarilycausedbyaCa/H abun-
3.7.Supplementaryinformationontheabsorbers dance gradientamongthe absorbers,butis instead a clear sign
ofa column-density-dependentdustdepletionof Ca in thegas.
For several of our Caii absorbers, supplementary information This effect is well known from observations of Caii and other
on the properties of these systems (based on previous studies) ionsintheinterstellargasintheMilkyWay.Indeed,theCaii/Hi
is available in the literature. Because intervening Caii absorp- vs. N(Hi) relationfromMilky Way disk andhalo observations
tionmostlikelyarisesintheinnerandouterregionsofgalaxies, (Wakker&Mathis2000;dashedline)reflectsexactlythetrend
wearemostlyinterestedininformationthatindicatesapossible that is seen in our intervening Caii absorbers (albeit the small
absorber-galaxy connection. Table 3 lists relevant information relativeoffset;seeabove).AswewilldiscussinSect.4.2,addi-
for seven outof the 23 QSO sightlines, such as the type of the tional evidence for dust depletion in the Caii absorbers comes
galaxy that probably hosts the absorbing cloud in its environ- fromtheobservedNai/Caiiratios.
ment(Col.5),theproperabsorber-galaxydistance(“impactpa-
rameter”;Col.6),andthegalaxyluminosity(Col.7).References
aregiveninthelastcolumn(seealsoZwaanetal.2005).Theto- 4. NatureandoriginofinterveningCaiiabsorbers
talneutralgascolumndensityfortheassociatedHiLyαabsorp- 4.1.CaiiabsorbersasHVCanalogs
tion is also listed (Col.3), allowing us to distinguish between
damped Lyman α absorbers (DLAs; log N(Hi) > 20.3), sub- UsingEq.(4),togetherwiththeobservedCaiicolumndensities
damped Lyman α absorbers (sub-DLAs; 19.2 ≤log N(Hi) ≤ (and their uncertainties),one can calculate that3−5 of the 10
20.3)andLyman-limitsystems(LLS;17.2≤logN(Hi)≤19.2), unbiasedCaiiabsorberswouldhaveHicolumndensitiesinthe
followingtheusualHiLyαabsorberclassification.Fouroutof DLArange(logN(Hi) ≥ 20.3),sothatanindirectestimatefor
thesevenCaiiabsorbersareDLAs,twoaresub-DLAs,andthere theDLAnumberdensityatz≤0.5inourdataisdN/dz(DLA)=
is one LLS. The derivedimpact parameters are typically small 0.035−0.058.ThisagreeswellwithdirectestimatesfortheDLA
(d < 8 kpcforfoursightlinesoutof seven,forwhichinforma- numberdensityatlowredshift(dN/dz(DLA)=0.045;Zwaanet
tion on d is available). Note, however, that neither the Hi col- al.2005).WithdN/dz = 0.117(Sect.3.2),Caii absorberswith
10 P.Richter,F.Krause,C.Fechner,J.C.Charlton,andM.T.Murphy:InterveningCaiiabsorbers
Table3.Supplementaryinformationforselectedabsorbersinoursample
QSO zabs logN(Hi)a AbsorberType GalaxyType d L/L⊙ Ref.
(N in[cm−2]) [kpc]
J121509+330955 0.00396 20.34 DLA EarlyType 7.9 0.44 a
J215501-092224 0.08091 17.98 LLS Spiral 34.0 0.50 b
J044117-431343 0.10114 20.00 sub-DLA Disk 6.8 1.00 c,d,e
J095456+174331 0.23782 21.32 DLA DwarfLSB ≤3.9 0.02 f
J113007-144927 0.31273 21.71 DLA Group(tidaldebris) 17−241 ... f,d,g,h
J123200-022404 0.39498 20.75 DLA Irr.LSB 6.6 0.17 i
J045608-215909 0.47439 19.50 sub-DLA ... ... ... j
a fromHiLyαabsorption
References: (a) Miller et al.(1999);(b) Jenkins et al.(2003);(c) Petitjean et al.(1996);(d) Chen & Lanzetta (2003);(e) Chen et
al.(2005); (f) Rao et al.(2003);(g) Lane et al.(1998);(h) Kacprzak, Murphy& Churchill (2010b);(i) Le Brun et al.(1997);(j)
Turnshek&Rao(2002)
Fig.9. Leftpanel:relation betweenCaii and Hi columndensitiesforseven Caii absorptionsystems. The solid line indicatesthe
bestfitthroughthedatapoints,thedashedlineindicatestherelationfoundforCaiiandHiintheMilkyWaybyWakker&Mathis
(2000).Rightpanel: relationbetween log (Caii/Hi), scaled to the solar (Ca/H) ratio, and log N(Hi) for the same absorbers.The
solidlineagainindicatesthebestfitthroughthedatapoints,whilethedashedlineshowsthetrendfoundintheMilkyWay(Wakker
&Mathis2000).
log N(Caii) ≥ 11.65 outnumberDLAs at low z by a factor of HVCs at columndensities that are verysimilar to the onesde-
2−3.ThisimpliesthatCaiiabsorbershaveahigherabsorption rivedbyusfortheinterveningCaiiabsorberpopulation(Richter
crosssectionthanDLAs,i.e.,theytraceneutralgasthatspansa et al.2005;Ben Bekhtiet al.2008;Richter et al.2009;Wakker
largercolumndensityrangeandthatisspatiallymoreextended etal.2007,2008).
thanDLAs.SinceDLAsatlowzarebelievedtoariseintheinner Based on these arguments,we are led to suggest that more
regionsofgalaxies(i.e.,intheirinterstellarmedia),weconclude thanhalfoftheCaiiabsorbersinoursampletraceneutraland
that Caii absorbers, which trace gas below the DLA limit (log partlyionizedgascloudsin thehalosandcircumgalacticenvi-
N(Hi)< 20.3),arisepredominantlyintheouter,moreextended ronmentofgalaxiesandthusrepresentdistantHVCanalogs.
regionsofthesegalaxies,i.e.,intheirgaseoushalos.
Strong observational support for this scenario comes from 4.2.TheNai/Caiiratioanddustdepletion
thedistributionoftheHi21cmemissionintheMilkyWaydisk
andhalo.The21cmemissionmeasuredinthediskfromthepo- FortheelevenCaiiabsorbers,forwhichinformationonNaiab-
sition of the sun indicates a neutral gas column density range sorptionisavailable(Sect.3.6),wefindawiderangeinNai/Caii
thatwouldcausethedisktoappearasaDLA ifobservedfrom ratios (or upper limits) from log (Nai/Caii) = −1.36 to log
outside. In contrast, the majority of the neutral gas clouds in (Nai/Caii)=+0.66,aslistedinTable4.IntheMilkyWay,these
the Milky Way halo with vertical distances > 1 kpc from the ratios are typical of the diffuse, warm neutral medium (WNM;
Galacticplane(thiswouldincludeprobablyallHVCsandsome T = 102−104K;n ≤10cm−3),whereCaiiandNaioftenare
H
of the IVCs; see Wakker 2001;Richter 2006) have Hi column notthedominantionizationstates,butserveastracespecies(see,
densitieslogN(Hi) < 20.3andwouldthereforebeclassifiedas e.g., Crawford1992;Welty, Morton& Hobbs1996).For com-
sub-DLAsandLLSsifseenasinterveningabsorbers.Moreover, parison,theNai/CaiiratiosthataretypicallyfoundintheMilky
CaiiabsorptionisfrequentlydetectedintheGalacticIVCsand Way in the more dense, cold neutral medium (CNM; T < 102