F-VIPGI: a new adapted version of VIPGI for FORS2 spectroscopy. Application to a sample of 16 X-ray selected galaxy clusters at 0.6 < z < 1.2

1.2 MB·English

by  Alessandro Nastasi

Astronomy&Astrophysicsmanuscriptno.aa19862 (cid:13)c ESO2013
January14,2013
F-VIPGI: a new adapted version of VIPGI for FORS2 spectroscopy
Application to a sample of 16 X-ray selected galaxy clusters at 0.6 ≤ z ≤ 1.2
1 2 1 1 4 3 2 2
A.Nastasi ,M.Scodeggio ,R.Fassbender ,H.Bo¨hringer ,D.Pierini ,M.Verdugo B.Garilli ,andP.Franzetti
1 Max-Planck-Institutfu¨rextraterrestrischePhysik(MPE),Giessenbachstrasse1,85748Garching,Germany
e-mail:[email protected]
2 IASF-INAF-ViaBassini15,I-20133,Milano,Italy
3 Institutfu¨rAstrophysikUniversita¨tWien,Tu¨rkenschanzstr.17,1180Vienna,Austria
3 4 Freelancescientist
1
0 Received...;accepted...
2
ABSTRACT
n
a
Aims.Thegoalofthispaperistwofold.Firstly,wepresentF-VIPGI,anewversionoftheVIMOSInteractivePipelineandGraphical
J
Interface(VIPGI)adaptedtohandleFORS2spectroscopicdatatakenwiththestandardinstrumentconfiguration.Secondly,weinves-
0 tigatethespectro-photometricpropertiesofasampleofgalaxiesresidingindistantX-rayselectedgalaxyclusters,theopticalspectra
1 ofwhichwerereducedwiththisnewpipeline.
Methods.WeprovidebasictechnicalinformationabouttheinnovationsofthenewsoftwareandreferthereadertotheoriginalVIPGI
] paperforadetaileddescriptionofthecorefunctionsandperformances.Asademonstrationofthecapabilitiesofthenewpipeline,
M
wethenshow resultsobtainedfor16distant(0.65 ≤ z ≤ 1.25) X-rayluminousgalaxyclustersselectedwithintheXMM-Newton
I DistantClusterProject.Weperformedaspectralindicesanalysisoftheextractedopticalspectraoftheirmembers,basedonwhich
. wecreatedalibraryofcompositehighsignal-to-noiseratiospectra.Wethencomparedtheaveragespectraofthepassivegalaxiesof
h
oursamplewiththosecomputedforthesameclassofobjectsthatresideinthefieldatsimilarhighredshiftandingroupsinthelocal
p
Universe.Finally,Wecomputedthe“photometric”propertiesofourtemplatesandcomparedthemwiththoseoftheComaCluster
-
o galaxies,whichwetookasrepresentativeofthelocalclusterpopulation.
r Results.Wedemonstrate thecapabilities of F-VIPGI,whose strength isan increased efficiency and a simultaneous shortening of
t FORS2spectroscopic data reduction time by a factor of ∼10 w.r.t.the standard IRAF procedures. We then discuss the quality of
s
a thefinal stacked optical spectra andprovide theminelectronic form⋆ ashigh-quality spectral templates, representative of passive
[ andstar-forminggalaxiesresidingindistantgalaxyclusters.Bycomparingthespectro-photometricpropertiesofourtemplateswith
thelocalanddistantgalaxypopulationresidingindifferentenvironments,wefindthatpassivegalaxiesinclustersappeartobewell
1 evolvedalreadyatz∼0.8andevenmoresothanthefieldgalaxiesatsimilarredshift.Eventhoughthesefindingswouldpointtoward
v
asignificant acceleration of galaxy evolutionindensest environments, wecannot exclude theimportance of the massasthemain
7 evolutionarydrivingelementeither.Thelattereffectmayindeedbejustifiedbythesimilarityofourcompositepassivespectrumwith
2 theluminousredgalaxiestemplateatintermediateredshift.
3
2 Keywords. Instrumentation:spectrographs–Methods:dataanalysis–Techniques:spectroscopic–galaxyclusters:general,redshifts
.
1
0
3 1. Introduction ters by means of their red galaxy population (e.g., SpARCS,
1 Wilsonetal.2006),theirSunyaev−Zel’dovich(SZ)effectsigna-
Galaxyclustersarethesignatureoftheprimordialdensity fluc-
: ture(e.g.,SPT,Williamsonetal.(2011);ACT,Menanteauetal.
v tuations that have grown via hierarchical accretion since the
(2010)) or the diffuse X-ray emission originating from the hot
Xi epoch of recombination. Because their abundance at different intraclustermedium.Thelastapproach,inparticular,hasproved
epochs is extremely sensitive to the matter content and accel-
r very powerful to the above aim as shown e.g. by the XMM-
a eration of the universe, clusters are sensitive probes for test- Newton Distant Cluster Project (XDCP, Bo¨hringeretal. 2005;
ing different cosmological models. In addition, they are cos-
Fassbenderetal. 2011a). This is a serendipitous X-ray survey
mic laboratoriesin whichcomplexprocessesthatshapegalaxy
specifically designedforfindingandstudyingdistantX-ray lu-
evolution can be studied in great detail. Nowadays many ef-
minousgalaxyclustersatz≥0.8andithascompiledthelargest
forts are invested into the observational challenge of provid-
sampleofsuch systemstodate. Fora comprehensiveoverview
ingsizable samplesofgalaxyclustersathighredshift(z >0.8)
ofthesurveyandanextensivediscussiononitsstrategyandre-
to trace the evolution of the cluster population and its mat-
sultswereferthereadertoFassbenderetal.(2011a).
ter components back to the first half of the universe life-
Irrespectiveoftheinitialapproachusedtodetectdistantclus-
time, corresponding to lookback times of 7 − 10 Gyr. Many
ters,thefinalmandatorystep oftheconfirmationprocessisthe
surveys have been designed to efficiently detect distant clus-
redshiftassessmentofthesystembymeansofspectroscopicob-
⋆ The library of spectra is available at the CDS via anony- servationsofitsgalaxypopulation.Tomaximizethenumberof
mous ftp to cdsarc.u-strasbg.fr 130.79.128.5 or via galaxiesatz>0.8whoseredshiftcanbesuccessfullymeasured,
http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/ and at the link pro- onehastodesignspectroscopiccampaignstoobservethespec-
videdinAppendixA(Sect.A.2). tralfeatureswiththehighestsignal-to-noiseratio(S/N).Because
1
A.Nastasietal.:F-VIPGI:anewadaptedversionofVIPGIforFORS2spectroscopy
galaxy populations in clusters are dominated by red, passive needtoprocessthehugeamountofspectroscopicdataexpected
galaxies,the most suitable featureto thisaim is the continuum from surveys such as the VIMOS-VLT Deep Survey (VVDS,
breakat4000Å(D4000)andtheassociatedabsorptioncalcium LeFe`vreetal. 2004) or zCOSMOS (Lillyetal. 2007), each of
lines(CaH/K).Forz ∼0.8−1thisspectralregionisredshifted whichproducedatotalof∼50,000spectraofgalaxiesinavery
toλ ≈7200−8000Å,awavelengthwindowreadilyaccessible wideredshiftrange(0≤z≤5).
formanyground-basedtelescopes.Inparticular,oneofthe cur- The core of VIPGI is a library of C-written routines, the
rentlymostefficientspectrographsabletocoverthiswavelength VIMOSDataReductionSoftware(DRS,Scodeggioetal.2001),
rangeandtosimultaneouslyguaranteeaveryhighandstableef- currently used by ESO for the VIMOS online reduction. All
ficiencyofitsCCDuptoλ∼11000ÅistheFOcalReducerand these fundamental functions can interact with each other and
lowdispersionSpectrograph(FORS2,Sect.3.1)mountedonthe with the user thanks to the adopted standard Python Tkinter
UT1ofVLT.Becauseofitsperformancesatlongerwavelengths, graphicalinterface.Thischoiceallowsone to obtaina pipeline
wherethe powerand efficiencyof the C codecomputationand
thisinstrumentiswidelyusedforimagingandspectroscopy of
distantclustergalaxiesandishencetheinstrumentofchoicefor thepossibilityofa continuousqualitycheckbythe useronthe
the spectroscopicfollow-upof the distantcandidatesystems in intermediatereductionresultsarebothpresent.Namely,theuser
XDCP. canconstantlymonitorthequalityoftheongingreductionstep
bystepwithVIPGI,repeatinganintermediatereductionstepif
AtthetimeofwritingXDCPprovidesthelargestsampleof
necessarywithoutrestartingtheentireprocedure.
X-ray selected distant galaxy clusters, with 30 confirmed clus-
VIPGIwas also designedto optimizethe storageof a large
ters at z > 0.9 and a final aim of more than 50 clusters at
amountofreduceddata,with aclearandeasily understandable
z > 0.8 (30 at z > 1) to allow statistically meaningful evo-
filingstrategy.MoretechnicaldetailsonVIPGIareprovidedin
lution studies of the cluster population in at least three mass-
theoriginalpaperofScodeggioetal.(2005).
and redshift bins. One of the main disadvantagesconnected to
these expectations is the significant work related to reducing
thelargeamountofdataproducedbythespectroscopicfollow-
3. TheFORS2-VIMOSInteractivePipelineand
up campaigns. Each XDCP targetis observedwith the FORS2
multi-objectspectroscopy(MOS)mode,enablinganaverageof GraphicalInterface(F-VIPGI)
50slits permask. Accordingto the surveyexpectationsquoted
Although VIPGI was specifically designed for the VIMOS in-
above,thespectroscopiccampaignsshouldfinallyyieldseveral
strument, its capabilities are general enough to make it poten-
thousandslits,resultinginasimilaramountofspectratobere-
tially usablewith anyotherMOSspectrograph.Thismotivated
ducedandextracted(seeFig.2,top).Thisprocessisverytime-
us to consider VIPGI as a potentially useful tool for XDCP,
consumingifcarriedoutwiththetraditionalIRAFpackages,and
whoseclustercandidateshavebeen(orareplannedtobe)spec-
the pipelines provided by ESO for automating the procedures
troscopicallyfollowed-uptosafelyconfirmtheirnatureofgrav-
wouldnotprovidethenecessaryaccuracyforanefficientextrac-
itationallybound,distantsystems. Theinstrumentused forthis
tion of the spectra of such distant galaxies, which are mostly
purposeistheFOcalReducerandlowdispersionSpectrograph
faint (I > 20)1. These reasons motivated us to develop a new
(FORS2)mountedonVLT.TheXDCPspectroscopiccampaigns
dedicatedpipelineforaquickandefficientreductionofspectro-
havetargeted∼70clustercandidatessince2005,producingato-
scopic FORS2 data. Because of the many similarities between
talof∼3,500singlespectratobeextractedandreduced.Withthe
FORS2andVIMOSdata,wechosetobuildthenewpipelineon
same original spirit of VIPGI we therefore adapted the power
VIPGI(Scodeggioetal.2005)andnameditF-VIPGI.
and efficiency of VIPGI to FORS2 data as well. The result is
Thispaperisstructuredasfollows:inSect.2themainchar-
F-VIPGI, a new pipeline that allows us to shortenthe time for
acteristicsofVIPGIarebrieflydescribedtointroducetheinno-
reducingFORS2spectroscopicdatabyafactorof∼10w.r.t.the
vationsofF-VIPGI,whichareextensivelydiscussedinSect.3.
standardIRAFprocedures.
Wethenshowanapplicationofthenewpipelinetoasampleof
distant clusters (Sect.4) that results in a new libray of spectro-
scopictemplates,whiletechnicaldetailsandpropertiesaredis- 3.1.TheFORS2instrument
cussed in Sect.5 and a conclusivesummaryis givenin Sect.6.
FORS2 is the visual and near-UV FOcal Reducer and low-
Throughoutthis manuscript we assume a concordance ΛCDM
cosmology,with H = 70 km s−1 Mpc−1, Ω = 0.7, Ω = 0.3 dispersionSpectrographmountedontheUT1unit(Antu)ofthe
0 Λ m Very LargeTelescope (VLT) (Appenzelleretal. 1998). The in-
andw=−1.
strumentcoversthewavelengthrangefrom330nmto1100nm
withanimagescaleof0.25′′/pixel(or0.125′′/pixelifthehigh-
resolutioncollimatorisused)inthestandardreadoutmode(2x2
2. TheVIMOSInteractivePipelineandGraphical
binning)andafieldofview(FoV)of6.83′×6.83′.Thedetector
Interface(VIPGI)
consistsoftwoMIT/LLCCID-20chips,with4096×204815µm
VIPGI is a semi-automatic data reduction pipeline released in pixels,eachcharacterizedbyanexcellentsensitivitytowardthe
2005 (Scodeggioetal. 2005) written to process and archive red part of the spectrum (up to λ ∼ 11000Å) and the almost
the data obtained with the VIsible MultiObject Spectrograph totalabsence of fringingpatterncontamination.FORS2 can be
(VIMOS) mountedat the MelipalUnit Telescope(UT3) of the used in many modes, includingmulti-objectspectroscopywith
VLTinaquickandefficientfashion.Itisabletohandlethedata exchangable masks, long-slit spectroscopy, imaging, spectro-
takenwithallthethreeavailableVIMOSmodes:MOS,imaging, polarimetryandhigh-timeresolutionimagingandspectroscopy.
andintegralfieldunit(IFU)spectroscopy.Thecreationofsucha F-VIPGI was designed to work with all FORS2 data taken
new,VIMOS-dedicated,reductionpipelinewasmotivatedbythe with the standard spectroscopic instrument configuration and
straightslits,definedasthoseslitswherethepixel-to-wavelength
1 ApparentVegaI-bandmagnitudeofaz>0.5L⋆passivelyevolving relationis“constant”alongtheslitlength,andthereforethesky
galaxywithformationredshiftz =5andsolarmetallicity. linesareperfectlyalignedalongtheCCDrows(orcolumns,de-
form
2
A.Nastasietal.:F-VIPGI:anewadaptedversionofVIPGIforFORS2spectroscopy
Table1.PropertiesofthegrismformingtheFORS2standardinstrumentconfigurationforwhichF-VIPGIisusable.Inparenthesis
are the wavelength rangesactually used by F-VIPGIto providethe best-qualityresults. The listed valuesof resolution λ/∆λ are
computedatthecentralwavelengthandfora1′′slit.
Grismname Centralwavelength Wavelengthrange Dispersion Resolution Orderseparation
[nm] [nm] [Å/mm] [Å/pixel] λ/∆λ filter
GRIS 600B+22 465 330-621 50 1.50 780 none
GRIS 300V+10 590 330(350)-660(925) 112 3.36 440 none
GRIS 300V+10 590 445(450)-865(850) 112 3.36 440 GG435+81
GRIS 300I+11 860 580(600)-1100(1050) 108 3.24 660 none
GRIS 300I+11 860 600-1100(1050) 108 3.24 660 OG590+32
GRIS 150I+27 720 330(370)-650(980) 230 6.90 260 none
GRIS 150I+27 720 445(430)-880(990) 230 6.90 260 GG435+81
GRIS 150I+27 720 600(590)-1100(1050) 230 6.90 260 OG590+32
pendingon the originalorientationof the data). A summaryof 3.3.F-VIPGIcalibrationfiles
thepropertiesoftheFORS2 standardgrismsequipmentispro-
As forVIPGI, the user hasto initially providespecific calibra-
videdinTable1.
tionfilesalsoforF-VIPGI.Thesearethenusedbythepipeline
tolocatethepositionandlengthoftheslitsontheCCD,tocom-
putethecoefficientoftheinterpolatingpolynomyalforthewave-
3.2.ConversionofFORS2filesintotheVIMOSformat
lengthdispersion,to correctforbadCCD pixels,and to finally
extract the single 1-D spectra. The calibration files are the fol-
The functionalities of F-VIPGI were enabled not by creating
lowing:
new routinesspecifically codedforthe FORS2 data formatbut
insteadbymanipulatingtheFORS2 datathemselvestoconvert • grism table: it contains all the main spectroscopic infor-
them into the VIMOS format. Despite the differences between mation of the grism, i.e. central wavelength, wavelength
FORS2andVIMOSCCDsarchitectures(withfoursquaredsen- range, and resolution. The emission sky lines usable for
sorsfortheformerandtworectangularonesforthelatter), our an additional refinement of the wavelength calibration (see
idea was to make the software able to identify chips 1 and 2 Sect.3.4) are defined here, too. It is defined for each grism
(hereafterQ1andQ2,respectively)ofFORS2asthefirsttwoin- configuration.
dividualquadrantsofVIMOS.Inthiswayonecanusethestan- • CCD table:itmapsallbadpixels/columnsoftheCCD that
dardVIPGIrecipeswiththe“adapted”FORS2fitsfileswithout haveto be correctedforby the pipeline.Two differentfiles
affectingthequalityandreliabilityofthedatareductionresults. areneededforQ1andQ2,buttheseremainthesameforany
ThepartsofFORS2 data thatare manipulatedby F-VIPGIare FORS2observation.
theimageandtheheader: • line catalog: it contains the list of all arc lines used for
thewavelengthcalibration(Sect.3.4).ForFORS2onlythree
filesareneeded,dependingonthekindoflampused(HeAr,
HgCdHeArorHgCdHeNeAr).
• image: the original FORS2 matrix of 4096 × 2048 pixels
• grismPAFfile:itisaparameterfileinthestandardESOdata
is transposed to obtain the same vertical orientation of the
interface controlterminology.Basically, it is just a text file
VIMOSframes.Thewavelengthdispersionorientationfrom
where informationcan be collected in a formatthat is very
lower (blue) to upper (red) side of the image is recovered
similartotheFITSkeywordsformatandisthenusedtoup-
in the same way. We also stress that this process does not
dateFITSkeywordsintheheaderofFITSfiles.Specifically,
affect the quality of the data because only the positions of
this file contains all information related to the wavelength
thepixelsarerearranged,whereastheirinformation(counts)
dispersion,opticaldistortion,andcurvatureofeachCCDand
areleftuntouched.
eachgrism.
• header: the header structure of FORS2 files differs signif-
• SPHOTtable:itisafitstablecontainingmagnitudevs.wave-
icantly from those of VIMOS. In particular, while in each
lengthofthestandardstarthatisusedforthespectrophoto-
VIMOSfiletheheadercontainsonlytheinformationonthe
metriccalibrationofthedata.Ithastobenamedexactlyas
slitsofthechiptowhichthefilebelongs,inFORS2eachfile
the target name reported in the header keyword “ESO OBS
hastheinformationofalltheslitsofthemaskonitsheader
TARG NAME”.
that are associated with Q1 and Q2. During the conversion
process F-VIPGI therefore changes these settings and re- All the above calibration files except for the SPHOT table
movesalltheinformationabouttheslits thatdonotbelong are provided together with the installation package and should
to the chip of the frame itself. Only the six reference slits, notbemodifiedbytheuser.
commontobothchips,arepreservedalongthisprocess.We
highlightthatthesechangesaremadeforallFORS2frames
excepttheBIAS,astheydonotcontainanyslitinformation 3.4.Wavelengthcalibration
intheirheaders.
One of the most powerful aspects of VIPGI, and hence of F-
VIPGI, is that the user can perform sensitive calibration steps
viavisualtoolsthatfacilitateexecutingandconstantlychecking
A specific and slightly different treatment is reserved to all the procedures. An example of this interactivity within the
standard-starframes(STD),asdiscussedinSect.3.7. reduction process is shown in Fig.1. Here the pipeline shows
3
A.Nastasietal.:F-VIPGI:anewadaptedversionofVIPGIforFORS2spectroscopy
Fig.1.AnexampleofaHeAr
arc line catalog displayed on
a raw lamp frame (in back-
ground)withinF-VIPGI.The
green regions mark the ex-
pected positions of the arc
lines according to the infor-
mationextractedfromthefits
header whiletheredones on
the right represent the line
catalog used for the calibra-
tion.Theredpartofthespec-
trumisintheupperpartofthe
image.
theexpectedpositionsofthearclinesofeachslit(greenregions
overlaid on the raw lamp frame) togetherwith the line catalog
usedforthecalibration(redregionsontherightside).The user
caniterativelyrefinethewavelengthcalibrationacrosstheentire
chip by shifting each green line until the region line patterns
perfectlymatchtheunderlyinglampframe.Thepositionsofthe
line centroids are then computedautomatically by the pipeline
by applyinga 2.5 sigma iterative rejectionto removelines that
aredeviatetoofarfromthefit.
This method produces a very accurate calibration with a
median uncertainty on the wavelength calibration that roughly
corresponds to one fifth of a pixel. As an example, for grism
300I+11,whosespectrahavealineardispersionof3.24Å/pixel,
the typical uncertainty on the wavelength calibration has an
< rms > = (0.30 ± 0.05)Å, as shown in the bottom panel of
λ
Fig.2.Anotherrefinementofthecalibrationcanalsobeobtained
byusingthepositionofsomebrightrelativelyisolatedskyemis-
sionlineswhosewavelengthsareknownanddefinedinthegrism Fig.2. Top: Redshift distribution of 1543 FORS2 spectra pro-
table(seeSect.3.3).Fordatatakenwithgrism300I+11weused duced by XDCP and extracted with F-VIPGI. The solid line
marks the secure assigned redshifts (≥75% confidence on the
eight sky lines between λ = 6300Å and λ = 10120Å, which
assignedvalue)whilethedottedlinereferstounsafeones(con-
produced a final refined wavelength calibration with a typical
fidence ≤50%).The secure assessments represent ∼72%of the
< rms > < 1Å. This quantity can be converted into a corre-
λ entiresampleat0<z≤3andincreasesto∼78%includingalso
spondingredshiftuncertaintyof δz ∼ 1−2·10−4 byreferring
λ stars.Bottom:Distributionofrmswavelengthcalibrationscom-
to λ = 8600Å. After the calibration process, all data are pre-
c putedforthecorresponding1230reducedslitsobservedwiththe
liminarily reduced in the standard way by subtracting the bias
grism300I+11.Themedianvalueis<rms> =(0.30±0.05)Å.
frames,correctingfortheCCDbadpixelsand,finally,applying λ
theflat-fieldingcorrection.
science spectrum. If this area is large enough (at least 2′′ left
3.5.Skylinesubtractionandatmosphericabsorption betweenthespectrumandtheslitedges,onbothsides),thesky
corrections linescanbeaccuratelymodeledandefficientlysubtractedfrom
thefinalspectrum,asshownintherightpanelsofFig.3and4.
A significant source of noise along the spectroscopic reduc-
Afterskysubtraction,weneedtocorrectallthespectraalsofor
tionprocessisintroducedbythestrongabsorption(telluric)and
theprominentskyabsorptionfeatures(telluriclines)thatcould
emission sky features produced by the O and H O molecules
2 2
and the OH− radical at λ > 6000Å (see Fig.3, left). F-VIPGI
along the wavelength axis. Themean signal level and rms(sigma) of
isabletoremovethetwoundesiredskycontributionsina very
thisslitprofilearethencomputedwitharobustiterativeprocedureby
efficientway,asshownin Fig.4.Theskyemission spectrumis
usingthebiweigthestimatordescribedbyBeersetal.(1990).Finally,
computed for each slit in the free regions2 on the sides of the allgroupsofatleastthreepixelsthatareNsigmaabovethemeanlevel
areconsideredasobjectspectrawhileeverythingelseisconsideredfree
2 Theseregionsareidentifiedautomaticallybythepipelinewiththe skyregion.Intheaboveprocessthesigmathreshold(N)canbesetby
followingprocedure:firsta“slitprofile”istracedbycollapsingthedata theuser.
4
A.Nastasietal.:F-VIPGI:anewadaptedversionofVIPGIforFORS2spectroscopy
Fig.3.Left:Spectralabsorption(topside)andemission(bottomside)featuresintroducedbythenightskyintheλ>6000Åspectral
region. Right: 2D emission spectra of the sky extracted for all slits of a FORS2 mask. The verticalalignmentof the sky lines is
indicativeofaverygoodwavelengthcalibrationovertheentiremask.
Fig.4. Example of the good quality achievable through the removal of absorption (left) and emission (right) sky features with
F-VIPGI.Leftpanel:Atmosphericcorrectionforaspectrumofapassivegalaxyat z ∼ 0.9observedwithgrism300I.Top:Flux-
calibratedspectrum(inblack)withouttheatmosphericcorrectionapplied.Itisevidentthatthetelluriclinesatλ∼7600Åstrongly
affecttheS/Nofthe4000-ÅbreakandtheCaII(H,K)lines.Bottom:Thesamespectrumafterapplyingtheatmosphericabsorption
correction.NowtheCaIIlinesarecompletelyrecovered.Inred(solidanddottedcurves)thepositionsofthemostprominenttelluric
linesareshown.Theblackdashedlinesmarkthepositionofthemostimportantspectralfeaturesinthedisplayedwavelengthrange.
Thebest-fittingtemplateofbothspectra(anLRGoneatz=0.930)isoverlaidinblueinbothpanelsforreference.
Right panel: Subtraction of the emission sky lines from two spectra with different redshift (spectrum1 at z = 1.086, spectrum2
atz = 0.282).Foreach objectthe same partof the spectrumbeforeand afterthe sky subtractionisshown atthe top andbottom,
respectively.Inaddition,thestrongestspectroscopicfeaturesofeachobjectaremarkedbytheredboxesandarelabeledonthetop
side. Althoughsome residualsofthe skyremovalare visible in thebottompanels,thepipelineefficientlysubtractsthe undesired
skyfeaturesandrecoversthesciencefeatures,whicharesometimescompletelyoutshonebytheatmosphericemission.
otherwise be interpretedas absorption lines in the object spec- atmosphericcorrectionbycombiningthespectrawhosetelluric
tra.Thesespuriousfeaturescanberemovedprovidedthatasuf- linesaremoreevidentandisolatedandfinallyappliessuchacor-
ficientnumberofwellexposedfluxcalibratedspectraareavail- rectionto allslitsofthemask.Anexampleofthegoodquality
able. Evenif the best results are obtained when the spectra se- achievablefortelluriclineabsorptioncorrectionisshowninthe
lectedforthecorrectionspanawideredshiftrange(becausethe leftpanelofFig.4.
intrinsicfeaturesareefficientlyremovedinthiswayandonlythe
skyonesareenhanced),agoodtelluriclineabsorptioncorrection
canalsobegainedbyusingclustergalaxies,whoseredshiftval-
uesarerelativelyclosetoeachother.Ifso,thepipelinefindsthe
5
A.Nastasietal.:F-VIPGI:anewadaptedversionofVIPGIforFORS2spectroscopy
Fig.5.Imageofthespectrumofastandardstartakenwiththe
FORS2300Iinstrumentsetupinthefirstchip(Q1).Inredare
shownthevirtualslitsof30′′lengththatF-VIPGIcreatesinthe
headerofthescientificandcalibrationframesofthestandard
starstomimictheVIMOSsetting.InFORS2observationsthe
standardstarspectraarealwaysimagedinQ1andarealways
enclosedinthefirst(#1)virtualslit.
Fig.6. Example of the spectrophotometric calibration for a FORS2 observation with grism 300I and sorting filter OG590. Left
panel:Extracted1-Dspectrumofthestandardstar,notyetcorrectedfortheCCDresponsefunction.Rightpanel,top:Theexpected
spectrumofthestandardstarprovidedbytheSPHOTtable(continuousline)andthefittedone(redpoints)accordingtothespectral
resolution of the observation. The sensitivity function (bottom panel) is then computed as the ratio between the extracted 1-D
spectrumandthefittedstandardstarspectrumconvertedintoergcm−2s−1Å−1.Theusercanedittheoriginallycomputedsensitivity
function(redpoints)andchoosethefinalone(continuousline)toapplytothesinglespectraforcalibratingthemintoflux.
3.6.Thefinalproduct of the standard star frames (and into the correlated calibration
files) that mimic the presence of 11 slits in total (6 in Q1 and
Once the preliminary reductionand the wavelength calibration 5inQ2)3 with30′′ lengthand1′′ width.Itisimportanttonote
are applied to each frame, these are calibrated in flux (see
thatinFORS2observationsthestandardstarspectraarealways
Sect.3.7) and finally combined to obtain a stack of 2D sky-
centered on Q1 at the position given by the header keywords
subtractedspectra.Singlespectraarethenextractedfromthe2D
CRPIX1 and CRPIX2. The virtual slits are therefore created in
stackedframe,usinga Horneoptimalextraction(Horne1986),
the way to have the first slit (#1 of Fig.5) always centered on
andare saved in localfoldersin fits format.Foreach spectrum
(CRPIX1,CRPIX2)and,so,alwayscontainthestarspectrum.An
F-VIPGIalsoproducestheassociatednoisevectorgeneratedby
exampleofthepartofthe virtualmaskrelativetoQ1isshown
theextractionresidual.
inFig.5.
Once the standard-star spectrum is reduced and extracted,
3.7.Spectrophotometriccalibration thepipelinecomputesthesensitivityfunctionbycomparingthe
observedstellarspectrumwiththeexpectedonecontainedinthe
AnimportantdifferenceoftheFORS2 observationstothoseof
SPHOTtable(Sect.3.3)asshowninFig.6.Theextractedcurve
VIMOS concerns the methods used by the two instruments to
is finally saved in a calibration table that can be applied to the
acquirethespectraofspectro-photometricstandardstars(STD).
1-D spectra for their flux calibration and for the correction of
This step is mandatory to calibrate each spectrum in flux and
the CCD responsefunction.Since standardstars in FORS2 are
tocorrectitscontinuumshapefortheCCDsensitivityfunction.
observedonlyonQ1,forQ2 data thesame calibrationtableas
StandardstarsinVIMOSareobservedinallfourquadrantsand
computedforthe firstchip hasto be used. An exampleof how
with a specifically designed mask composed of eight slits per
thesensitivityfunctioniscomputedforastandardstarobserved
chip, with fixed position and length (10′′). In FORS2, instead,
withthe300IgrismisshowninFig.6.
theSTDframeisacquiredonlyonQ1andbymeansofasingle
long-slitwith5′′width.
Toovercomethisdiscrepancyandenabletheuseofthestan-
dardVIPGIroutinesforspectrophotometriccalibrationalsofor 3 Thisisthemaximum number of 30′′ length slitsthat can be uni-
FORS2 data, F-VIPGI writes some keywords into the header formlyplacedontheareaofthetwochips.
6
A.Nastasietal.:F-VIPGI:anewadaptedversionofVIPGIforFORS2spectroscopy
Fig.7. Distribution of the absolute values of the differences Fig.8.Redshiftdistributionofthe selected187galaxies,mem-
between the redshift measurements |∆z| for 14 members of bersofthe16XDCPclustersreportedinTable2.
XMMUJ1230.3+1339atz=0.975frommultiple,independent,
spectroscopic observations. The median value of < |∆z| > =
veys),whichcontaintemplatesrepresentativeofthepopulation
5.1·10−4ismarkedbytheverticaldashedline. offieldgalaxiesatdifferentredshiftsuptoz∼1.5.Inadditionto
identifyingredshiftsofdistantgalaxies,the providedtemplates
3.8.Redshiftaccuracytest canalsosignificantlyimprovetheefficiencyandthereliabilityof
thephotometricredshiftassessmentofdistantclusterswithcom-
The XDCP massive clusterXMMUJ1230.3+1339at z = 0.975
plexstarformationhistories(see,e.g.,Guennouetal.(2010)and
(Fassbenderetal. 2011b; Lerchsteretal. 2011) has been tar-
Pierinietal.(2012)foradiscussion)andtheycanbeusedtopre-
getedbyfourdifferentFORS2spectroscopicobservationsinor-
dictrest-frameU−Bcolorsofclustergalaxiesathighredshifts.
dertofullycharacterizeitsextendedanddynamicallyunrelaxed
galaxypopulation.This resultedin 65 confirmedspectroscopic
members,15ofwhichhadmultiple(double)observations.This 4.1.Thespectroscopicsample
enabledustousethesetargetsforaself-consistencytestonthe
F-VIPGIredshiftmeasurements,basedondifferentindependent For our study we selected all the XDCP clusters with at least
spectroscopicobservationsofthesameobject. threespectroscopicallyconfirmedmemberswith anS/N > 2 in
Tothisaimweselected14outofthe15duplicatedobserva- the rest frame wavelength range of 4000 - 4300 Å in their re-
tions, considering only those couples where the final extracted ducedspectra.Eachspectrumwasalsovisuallyinspectedtover-
dataweregoodenoughtoprovideasaferedshiftassessmentfor ifythatnocontaminationsfromcosmicraysorbadskysubtrac-
both cases. The median of the differences (in absolute value) tion were present. For the cluster member selection, we firstly
between the redshift measurements is < |∆z| >= 5.1 · 10−4, estimated the redshiftcluster(zcl) as the medianof the redshift
with a spread given by the semi-interquartile range (SIQ) of peakfoundintheproximityoftheX-rayemissioncenter.After
< |∆z| > = 2.5·10−4.Thisvaluecanbetranslatedintoacor- that, we selected the cluster members as those galaxies with a
SIQ
respondingrest-framevelocityuncertaintyofδv,rest ≈77kms−1 rest-framevelocityoffset<3000kms−1 fromzcl,corresponding
foragalaxyatz=0.975. to a redshift window cut of ∆z < 0.01×(1 + zcl). The list of
Thedistributionof|∆z|isshowninthehistogramofFig.7. the final targets selected for the next spectroscopic analysis is
reportedinTable2.ForeachtargetasequentialID,thespectro-
scopicredshifts,thenumberofconfirmedmembersusedinthe
4. Anapplicationtoasampleofdistantgalaxy subsequentanalysis,andtheliteraturereference(ifexisting)are
clusters reported.Atotalof187clustermembershavebeenidentifiedin
thisway,witharedshiftdistributionintherange0.65≤z≤1.25
InthissectionweshowanapplicationofF-VIPGItotheanaly- asshowninFig.8.
sisofspectroscopicdataforasampleof16distantXDCPgalaxy
clustersintheredshiftrange(0.65≤z≤1.25)observedwiththe
grism300I+11(Table1).Themaingoalofthissectionistopro- 4.2.Spectralindicesanalysisresults
vide the communitywith a new set of spectroscopic templates
Wegroupedtheabovegalaxiesintofivedifferentspectralclasses
thatare intendedto representthe galaxypopulationresiding in accordingtothemeasured4valuesoftheequivalentwidths(EW)
distant galaxy clusters. For this reason, as shown in Sect.5.1,
thepresentlibraryiscompletelydifferentfromtheexistingones 4 Allspectralindicesweremeasuredinanautomatedwaybymeans
(e.g.,theonesconstructedfromSDSS, zCOSMOSorK20sur- ofaPythonscriptdevelopedbyourgroupthatusesthetrapezoidalrule
7
A.Nastasietal.:F-VIPGI:anewadaptedversionofVIPGIforFORS2spectroscopy
Table2.Listofthe16XDCPclustersusedforourspectroscopicanalysis.
ID z N References
spec members
cl1 0.7690 13 (1)
cl2 1.2310 8 (1)
cl3 0.9410 16 (1)
cl4 0.9590 6 C20in(2)
cl5 0.8290 8 (1)
cl6 0.7850 5 (1)
cl7 0.6770 12 (1)
cl8 0.8830 4 (1)
cl9 0.8940 19 (1)
cl10 0.7890 13 (1)
cl11 1.1220 7 C11in(2)
cl12 1.0740 3 (1)
cl13 0.8270 9 (1)
cl14 0.7460 7 (1)
cl15 0.9750 52 XMMUJ1230.3+1339in(3)and(4)
cl16 1.2030 7 (1)
References.(1)Tobepublished;(2)Fassbenderetal.(2011a);(3)Fassbenderetal.(2011b);(4)Lerchsteretal.(2011)
Table3.ClassificationcriteriatakenfromPoggiantietal.(2009)usedinthisstudyforgroupingtheclustergalaxiesintofivepectral
typesaccordingtothe equivalentwidthsvaluesof their[OII]λ3727andHδ lines. Thefollowingconventionis used: EW < 0 for
emissionlines,EW >0forabsorptionlines.Thenumberofgalaxiesfoundineachclassisgiveninthelastcolumn.
Spectroscopicclass EW([OII])value EW(Hδ )value Number
A
[Å] [Å] ofgalaxies
Passive >-5 <3 108
Post-starburst >-5 ≥3 21
Quiescentstar-forming ]-25,-5] <4 23
Dustystarburst ]-25,-5] ≥4 14
Starburst ≤-25 any 21
of their [OII] and Hδ lines, as summarized in Table3. These ativenumberofgalaxiesinthedifferentclassescanbeinferred
two spectralfeaturesare, in fact, reliable indicatorsof ongoing fromthelastcolumnofTable3.
and recent star formation activities within timescales of ∼ 107 We adopted the bandpasses defined in Baloghetal. (1999)
and ∼ 109yr for [OII] and Hδ, respectively. The amplitude of forEW([OII])andD4000(hereafterD 4000)whileforEW(Hδ)
n
their 4000-Å break (D4000) and the equivalent widths of Hβ, we used the definition of Hδ given by Worthey&Ottaviani
A
[OIII]λ4959and[OIII]λ5007werealsocomputedbutwerenot (1997).Thegalaxieswereclassifiedintofivespectralclassesfol-
used for the spectral classification. The errors on the indices lowingthesamemethodasdescribedinPoggiantietal.(2009),
were estimated by means of the noise vector provided by F- which is based only on the strength of [OII] and Hδ lines, as
A
VIPGI for each extracted spectrum at the end of the reduc- summarizedinTable3.
tionprocess(seeSect.3.6).Specifically,foreachspectrum1000
Fig.9 shows the distribution of the 187 galaxies in the
Monte-Carlorealizationsweregeneratedusingtherelatednoise
EW([OII])-EW(Hδ )plane.Theadoptedlociforthefivespec-
vector as variance. The uncertainties on the indices were then A
tral classes are marked by the dashed lines and a color coding
computed as the rms of the results obtained by repeating the
wasusedforabettervisualizationofthegroups.
spectralindicesanalysisontheseriesofsimulatedspectra.
We stress that the relative fraction of galaxies in the differ- TheDn4000indexwasthenusedtoqualitativelytesttheva-
lidity of our classification. The top panel of Fig.10 shows the
entspectraltypesis substantiallybiased bythemethodused in
distribution of our galaxy sample in the EW(Hδ ) vs D 4000
XDCP to select the targets forthe spectroscopicfollow-up. As A n
spacewiththecolorcodeadoptedinFig.10foridentifyingthe
discussed in Fassbenderetal. (2011a), to maximize the proba-
differentspectralclasses.Thedistributionwefindisremarkably
bilityoftargetingactualclustermembers,FORS2masksarecre-
similartothatreportedinFig.1ofGallazzi&Bell(2009)fora
atedinafashionthattheslitsarepreferentiallyplacedoncolor-
setofmodelgalaxies.Namely,weobservethatconsistentlywith
selected galaxies close to the expected red-sequence color and
theresultsoftheseauthors,thestarburst(dustyandnot)galaxies
withinthedetectedX-rayemission.Thisimpliesthatthemajor-
tend to be concentratedon the upperleft side of the plot, indi-
ityoftheobservedtargetsisactuallyexpectedtobepassiveand,
cating a relatively young (and still forming) stellar population.
hence,thatnostatisticallysignificantconclusionsabouttherel-
The passive and quiescent star-forming galaxies instead reside
to perform the integration. The code functioning was tested on some mostlyinthecenterandonthelowerrightside,suggestingthat
spectra by comparing its results with those provided by the standard theyexperiencedthelastepisodeofmajorstar-formationactiv-
IRAFpackages. ity more than 3 Gyr earlier. Finally, the post-starbursts are all
8
A.Nastasietal.:F-VIPGI:anewadaptedversionofVIPGIforFORS2spectroscopy
Fig.9.Distributionofthe187galaxiesinEW([OII])-EW(Hδ )
A
plane. The dashed lines and the different colors mark the po-
sitions of the five spectral classes, with the same criteria as in
Poggiantietal. (2009). The color code is the following: blue:
starburstgalaxies;cyan:quiescentstar-forming;magenta:dusty-
starburst;red:passive;green:post-starburst.
locatedabovethesequencedrawnbythepassiveandquiescent
star-forming galaxies, a sign that their starburst activity ended
withintheprevious2Gyr.
InthebottompanelofFig.10thedistributionintheD 4000
n
- EW([OII]) diagram is shown. It is evident also here that for
the adopted classification method the different spectral classes
are confinedto specific regionsof the plots. In particularif we
considertherelationadoptedbyFranzettietal.(2007),
D 4000+EW([OII])/15>0.7
n
(markedbythedottedlineintheplot)todividepassivelyevolv-
ing (early-type) galaxies from star-forming (late-type) objects
foraspectroscopicsampleofVVDSgalaxiesat0.45≤ z≤1.2, Fig.10. Dn4000 versus EW([OII]) (top panel) and EW(HδA)
all passive and post-starburst galaxies appear to be efficiently (bottom panel) for the selected targets. The dotted line in the
rightpanelrepresentsthedivisionbetweenearly-type/late-type
separatedfromthestarburstones.However,theplotalsoshows
galaxiesadoptedinFranzettietal.(2007).Thecolorcodeisthe
that the majority of quiescent star-forming and dusty-starburst
galaxiesareclassifiedasearly-typeinthisway.Thiseffectwas sameasinFig.9.
already predicted by the same authors, who claimed that some
contaminationintheearly-typeregionwasexpectedduetoearly
In Fig.11 the distribution of these objects in the
spirals and that this is a general feature observed for color- or
log(EW[OII]λ3727/EW(Hβ)) vs log(EW[OIII]λ5007/EW(Hβ))
spectroscopicclassificationschemes.
plane is shown.Only one object(markedin red) appearsto lie
To identify possible active galactic nuclei (AGNs, e.g.
inthe“AGNregion”oftheplotandwasconsequentlyexcluded
Seyfert2) erroneouslyidentified as starburst galaxies, we used
fromthefinalstackingofthestarburstspectra.
the diagnostic diagram of Bongiornoetal. (2010), that was
originally defined in Lamareilleetal. (2004) (and references
therein), where the ratios between the EW of [OII]λ3727,
5. Anewlibraryofspectroscopictemplates
Hβ, and [OIII]λ5007 lines are used to separate star-forming
from AGN-photoionized spectra. The constraint of having the Before proceeding with the final stacking procedure, we re-
[OIII]λ5007line within the observed wavelength range 6000Å stricted the usable portion of each spectrum to the observed
<λ <9800Åmakesthisanalysisfeasibleonlyforsevenstar- wavelengthrangeof6100-9280Å.Thiscutwasmadetoavoid
obs
burstgalaxieslyingatredshiftz<0.95. the higly sky-contaminated region at λ ≥ 9300Å (Fig.3, left)
9
A.Nastasietal.:F-VIPGI:anewadaptedversionofVIPGIforFORS2spectroscopy
Table 4. Spectral properties of the five averaged spectra of Fig.13. The reported S/N values refer to the wavelength range
4000Å<λ<4300Å.
Spectroscopicclass S/N EW([OII]) EW(Hδ ) D 4000 EW(H ) EW([OIII]λ5007)
A n β
[Å] [Å] [Å] [Å]
Passive 9 1.431±1.191 0.148±0.981 1.810±0.030 1.829±0.606 0.958±0.611
Post-starburst 6 -0.178±1.739 4.910±1.395 1.543±0.034 2.753±1.164 0.540±1.448
Quiescentstar-forming 7 -7.946±1.791 0.593±1.231 1.550±0.031 1.664±0.887 -1.184±0.985
Dustystarburst 7.5 -7.719±1.605 5.827±1.123 1.412±0.030 -2.357±1.849 -5.842±0.693
Starburst 5 -34.943±2.929 1.368±1.887 1.254±0.036 -8.816±1.797 -13.516±2.256
1
0.8
0.6
0.4
0.2
Fig.11.Diagnostic diagramused byBongiornoetal. (2010) to
0
distinguish between pure star-forming and AGN-contaminated 3000 4000 5000
galaxies applied to those galaxies of our sample classified as
starburst(seeTable3)andlyingatredshift0.6<z<0.95tohave
their [OIII]λ5007 line falling within the observed wavelength Fig.12.Normalizedrest-framewavelengthcoverageforthe186
range.Thesolidanddashedcurvesshowthedemarcationandits galaxiesusedinthestackingprocedure.Thegeneralresultforall
±0.15dex uncertainty, respectively, between pure star-forming galaxiesisrepresentedbythethicksolidcurve,whilethethinner
galaxies (bottom region) and AGN (top region), as defined in lines indicate the different contributions from the five spectral
Lamareilleetal. (2004). The red point marks the starburst ob- classes definedin Sect.4.2. Thesame colorcodeas in Fig.9 is
jectpossiblyhostinganAGNthatwasthereforeexcludedfrom used.
thefinalstackingprocedure.
corner.Themostimportantspectralfeaturesrecognizableinthe
while simultaneously preserving the observability of spectral
featureslikeHβand[OIII]forgalaxiesuptoz∼0.85. rest-framewavelengthrange∆λ=2700-5300Åarelabeledon
The totalrest-framewavelengthrangecoveredbythe com- thetopsideoftheplot,andtheirexpectedpositionsaremarked
bydashedverticallines.
posite spectra is 2700- 5300Å, with a normalized wavelength
coverage for all 186 considered galaxies shown in Fig.12. In Thespectralpropertiesoftheaveragespectraarequantified
the same figure we also show the different contributions from in Table4 together with their S/N computed in the wavelength
thefivespectralclassesdefinedinSect.4.2.Werecallthathere range4000Å<λ<4300Å.
the possible AGN-contaminatedstarburst object of Fig.11 was WenotethatthevaluesofEW(Hδ )and,especially,D 4000
A n
excludedfromouranalysis. obtainedforthe“passive”templatearesimilartothoseobserved
Forthefinalstackingprocesseachspectrumwasrescaledto forthe mostmassive and passive galaxiesin the localuniverse
themeanfluxcomputedinthewavelengthrange4000Å< λ < bySDSS-DR4(D 4000| ≈1.9;EW(Hδ )| ≈-1.5)thatwere
n z=0 A z=0
4300Å and then combined with the other objects of the same reportede.g.byGallazzietal.(2005).
spectralclassusingthemedianasoperator. In the next two sections (5.1 and 5.2) we compare the
The resulting five composite spectra are shown in Fig.13. spectro-photometric properties of our composite spectra with
Foreachspectrumtherelativespectralclassandthenumber of those of galaxiesresiding in differentredshiftrangesand envi-
galaxiesusedinthestackingprocessarereportedinthetop left ronments.
10