Constraining the thermal history of the Warm-Hot Intergalactic Medium

0.19 MB·English

by  L. Zappacosta

Astronomy&Astrophysicsmanuscriptno.zappacosta February2,2008
(DOI:willbeinsertedbyhandlater)
Constraining the thermal history of the Warm–Hot Intergalactic
Medium
L.Zappacosta1,R.Maiolino2,A.Finoguenov3,F.Mannucci4,R.Gilli2,
A.Ferrara5
5 1 DipartimentodiAstronomiaeScienzadelloSpazio,LargoE.Fermi2,I-50125Firenze,Italy,
0 2 OsservatorioAstrofisicodiArcetriLargoE.Fermi5,I-50125Firenze,Italy
0 3 Max-Planck-Institutfu¨rextraterrestrischePhysik,Giessenbachstraße,D-85748Garching,Germany
2 4 IstitutodiRadioastronomia,SezionediFirenze-CNRLargoE.Fermi5,I-50125FirenzeItaly
n 5 SISSA/InternationalSchoolforAdvancedStudiesViaBeirut,434014Trieste,Italy
a
J Received;accepted

9
1 Abstract. We have identified a large-scale structure traced by galaxies at z=0.8, within the Lockman Hole, by means of

multi-objectspectroscopicobservations. ByusingdeepXMMimageswehaveinvestigatedthesoftX-rayemissionfromthe
1
Warm-Hot IntergalacticMedium (WHIM) expected tobe associated withthislarge-scale structureand we set atight upper
v
limittoitsfluxintheverysoft0.2–0.4keVband.Thenon-detectionrequirestheWHIMattheseredshiftstobecoolerthan
2
0.1keV.CombinedwiththeWHIMemissiondetectionsatlowerredshift,ourresultindicatesthattheWHIMtemperatureis
0
4 rapidlydecreasingwithredshift,asexpectedinpopularcosmologicalmodels.
1
0 Keywords. Large-scalestructureofUniverse-X-rays:diffusebackground
5
0
/
h 1. Introduction and Milky Way Halo are strong. Nevertheless, several inde-
p pendent detections of WHIM emission were obtained by de-
- Bothobservationsandcosmologicalmodelssuggestthatmost tailed analysis of soft X-ray maps (ROSAT and XMM) in re-
o
of the baryonic matter is located in the intergalactic medium
r gions characterized by galaxy overdensities and by the spec-
t (IGM). Cosmological models have also shown that the evo-
s tral analysis of clusters of galaxies and of their surroundings
a lution of the baryonic matter is driven by progressive gravi- (Wangetal. 1997; Sołtanetal. 2002; Zappacostaetal. 2002,
: tational heating in the potential field of dark matter filaments
v 2004;Kaastraetal.2003;Finoguenovetal.2003).
i (Cen&Ostriker1999;Dave´etal.2001).Inparticular,athigh
X AllWHIMdetectionsdiscussedabovehavebeenobtained
redshifts (z >2–3) the baryonic gas is relatively cold (T <
r 105 K), and is identifiedwith Lyα absorbersalongthe line of for gas in the local universe or at low redshift. The most dis-
a tant WHIM emission detected so far was obtained at z∼0.45
sightofquasars,whileatlowerredshifts(z<1)theshocksdue
byZappacostaetal.(2002).Athigherredshiftthedetectionis
totheinfallofthegasonthedarkmatterfilaments(tracedby
moredifficultbecauseof bothtechnicalandphysicalreasons.
theregionsofhighgalaxynumberdensities)graduallyheatthe
gastotemperaturesintherangeT∼105−107K. Indeed,thelackofbright,highredshiftquasarspreventsthede-
tectionofWHIMfeaturesinabsorption,whilethethermalcut-
TheidentificationofsuchWarm-HotIntergalacticMedium
offisredshiftedtolowerenergiesmakingmoredifficulttode-
(WHIM)inthelocaluniversehasreceivedgrowinginterestin
tecttheWHIMemissioneveninthesoftX-rays.Anadditional
thelastfewyears.Thedetectionofabsorptionlinesfromhighly
issue is that, according to cosmological models, the WHIM
ionized species (OVI, OVII, OVIII), both in the UV and in
should be cooler at higher redshift, implying a lower ioniza-
the softX-rays,hasallowedto unambiguoslyidentifyWHIM
tionstateofthegas(i.e.loweropticaldepthofhighionization
along the line of sight of a few bright quasars (Nicastroetal.
absorbers) and a thermal cutoff further moved to lower ener-
2002; Mathuretal. 2003). The detection of emission due to
gies.Yet,itwouldbemostusefultoobtainsomeconstraintson
WHIM is quite challenging. Indeed, the WHIM is expected
the WHIM properties at high redshift since, when compared
to emit weak and diffuse radiation mostly in the softest X-
withtheWHIMdetectionsinthelocaluniverse,itwouldpro-
ray bands (<∼ 1keV), where both the Galactic absorption and
videcontraintsonthecosmologicalmodelsoftheevolutionof
the foreground emission from the Local Hot Bubble (LHB)
baryons.
Send offprint requests to: L. Zappacosta, e-mail: To pursue the latter goal we started a detailed inves-
[email protected] tigation of the Lockman Hole, which is one of the fields
2 Zappacostaetal.:ConstrainingthethermalhistoryoftheWHIM
36:00 0.782
0.782
0.805
0.780
57:30:00
0.788
0.792
0.784
0.8 (phot)
24:00
0.807
30 10:54:00 30 53:00 52:30
Fig.1. Spatial distributionof objects reportedin the literature
inthenarrowredshiftrange0.78<z<0.81withintheLockman
Hole (see Table 1). Each object is labelled with its redshift.
Theobjectmarkedwithadiamondhasaphotometricredshift.
Boxesindicatetheregionscoveredbyourspectroscopicsurvey Fig.2. Upper panel: redshift distribution of the objects for
with10masks. which we have obtained high quality (shaded histogram) and
lowquality(hollowhistogram)spectroscopicredshiftmeasure-
ments. Lower panel: zoom of the redshift distribution around
where the Galactic absorption is minimum (N ∼ 5.6 ×
H thethreemainconcentrationsofobjects.
1019cm−2) and where deep X-ray observationshave been ob-
tained(Hasingeretal. 1998a, 2001;Hasinger2003). We have
searched for superstructures by analyzing the redshift distri-
isstronglysuggestiveoftheexistenceofalarge-scalestructure
bution of sources already identified in this field and found
atthisredshiftinthisregionoftheLockmanHole.
8 sources in a narrow redshift range at about z∼0.8, located
Toconfirmthistentativeindicationwehaveobtainedmulti-
within a regionofabout20arcmin.We obtainedmulti-object
object spectroscopy of 215 galaxies in the same area. We
spectroscopyinthesameareaand,asdiscussedinSect.2,we
usedthemulti-objectspectroscopicmode(MOS)oftheoptical
haveconfirmedthepresenceofasuperstructureatz∼0.8.Then
spectrometerDOLORES,attheTelescopioNazionaleGalileo
wehaveanalyzedanXMMmapoftheLHinthesoftestband
(TNG),withtheLR–Rgrating,whichcoversthe4470–10360
and,asdiscussedinSect.3,wehaveobtainedtightconstraints
Årangeataresolutionof11Å.Thisspectralrangeallowsthe
onthepossiblediffuseemissionduetoWHIMassociatedwith identification of Hβ+[OIII] and [OII] at z∼0.8. The observa-
the large-scale structure. In Sect. 4 we discuss these observa-
tionswereperformedduringfournightsinMarch2003.
tionalconstraintsontheevolutionoftheWHIMwithredshift.
We selected two samples of galaxies within the subarea
of interest of the Lockman Hole. In particular, we selected
2. Opticalobservations:detectionofalarge-scale a shallow sample, made of galaxies in the magnitude range
structureatz∼0.8 20.0 ≤ R < 21.3, and a deep sample, containing galaxies in
themagnituderange21.3 ≤ R < 22.Theshallowsamplewas
As mentioned in the Introduction, an analysis of the redshift observedwithsevenmasksandwithanintegrationof1.5hours
distributionofthesourcespreviouslyidentifiedintheLockman foreachmask,whilethedeepsamplewasobservedwiththree
Hole(mostofwhichareopticalcounterpartsofX-raysources masksandwithanintegrationof3hourspermask.Mostofthe
detectedbyROSATandXMM)hasrevealedtheexistenceof8 masks(7ofthem)werelocatedalongtheN–Sdirectiontraced
sourcesinthenarrowredshiftrange0.780–0.807(∼74Mpc1). bythe4sourcesatz=0.78,whilethreemaskswerelocatedto
Thesesources(listedintable1)aremarkedwithacircleinFig. thewestandtotheeasttocheckpossibleextensionsofthepu-
1 and are located within a region of about 20 arcmin. In the tative large scale structure. The location of the variousmasks
sameregionwe havefoundanotherobject(markedwith adi- isshownbytheboxesinFig.1.
amond)withaphotometricallyestimatedredshiftof0.8±0.1.
The MOS spectra were reduced following the standard
Note that 4 of such objects (distributed along the N–S direc-
threads(darkandbiassubtraction,flatfielding,wavelengthcal-
tion)haveredshiftsinthenarrowerinterval0.780–0.784.This
ibration).Thedeepsamplewasobservedusingaditheringof6
1 InthewholepaperweassumeacosmologywithΩ =0.3,Ω = arcsec along the slit axis to enable a better subtractionof sky
m Λ
0.7andH =70kms−1Mpc−1 andallthedistanceswillbeexpressed lines and an easiest detection of the weak emission lines. We
0
inthecomovingrestframe. observed215 sources in total. We measuredredshifts for 103
Zappacostaetal.:ConstrainingthethermalhistoryoftheWHIM 3
Table1.Objectsinthenarrowredshiftrange0.78<z<0.81locatedwithintheLockmanHolereportedintheliterature.
Object RA(J2000) DEC(J2000) Type z Reference
RDS117Q 10h53m48.8s +57◦30′34′′ AGN 0.780 Lehmannetal.(2001)
[HGG98]5 10h53m44.9s +57◦35′15′′ Galaxy 0.782 Hasingeretal.(1998b)
[HGG98]8 10h53m42.3s +57◦35′41′′ Galaxy 0.782 Hasingeretal.(1998b)
RXJ105335.1+572542 10h53m35.1s +57◦25′42′′ QSO 0.784 Mainierietal.(2002)
RXJ105303.9+572925 10h53m03.9s +57◦29′25′′ QSO 0.788 Mainierietal.(2002)
[MBH2002]41 10h53m05.4s +57◦28′10′′ X–raysource 0.792 Mainierietal.(2002)
[FFH2002]105 10h53m15.80s +57◦24′50.0′′ X–raysource 0.8(phot) Faddaetal.(2002)
LOCK-6cmJ105304+573055 10h53m04.83s +57◦30′55.9′′ Radiosource 0.805 Ciliegietal.(2003)
RXJ105225.3+572246 10h52m25.3s +57◦22′46′′ AGN 0.807 Mainierietal.(2002)
objects(see AppendixA for thecatalog)with typicalrandom perstructure outlined by our survey and the archival objects
errors of ±0.002, as inferred by the uncertainty on the gaus- cover a region of ∼ 7.5Mpc (at the mean redshift z=0.791).
sianfittotheemissionlines2.For47sourcestheredshiftcould Thiscouldbeconsideredas a lowerlimitto the dimensionof
be determined unambiguously thanks to the detection of two the structure because its size is limited by the extent of our
or more emission lines. We assigned unambiguouslythe red- observations. However, we note that other spectroscopic sur-
shiftalsoincaseofdetectionofonlyonestrongemissionline veyscoveringtheentireXMMandROSAT-HRIfieldshavenot
identified as [OII] line at high redshift 3 (in Appendix A the found additional objects in the narrow redshift range outside
qualityoftheseredshiftsismarkedas“high”).Forother56ob- theareaconsideredbyus,suggestingthatthelarge-scalestruc-
jectsthe redshiftdeterminationwasless secure,based online ture isnotfurtherextendedatleast inthe XMM andROSAT-
detections with low signal to noise ratio (in Appendix A the HRIfields.
qualityoftheseredshiftsismarkedas“low”).For112sources
the spectrum was too weak and without bright lines, and we
3. X-raydata:constraintsontheWHIMemission
couldnotrecoveranyinformationontheirredshifts.
The redshift distribution is shown in the top panel of Once demonstrated that large-scale structures at high redshift
Fig. 2, where a prominent peak (6σ significant with respect exist within the Lockman Hole, we have then investigated
to the mean histogram level) is seen at the expected red- whether the associated WHIM emission could be detected.
shift of z=0.78, demonstrating beyond any doubt the pres- Such investigation requires sensitive maps at soft energies
ence of a large-scale structure at this redshift in this region. E<0.5 keV, both because of the low temperatures typical of
As for the archivalobjects, thisspike showstwo distinct sub- WHIMandbecauseofasignificantDopplershiftoftheemis-
concentrationsoneatz∼0.78andoneatz∼0.807.Theformer sion to lower energies. Several deep X-ray observations have
is further splitted into a main peak at z∼ 0.784 and a nearby been carried outin the LockmanHole area, mostimportantly
spikeatz∼ 0.776,asshowninthebottompanel.Fig.3shows withROSATandXMM-Newton(Hasingeretal.1998a,2001).
the distribution projected on the sky of the sources for which
ROSAT maps would have the required field of view and
theredshiftcouldbedetermined,andinparticularforthosein
sensitivityatlowenergiestoproperlyconstrainthepresenceof
the redshift spikes. Although this redshift survey is not com- diffusewarmgasathighredshifts.However,theregionwhich
plete,ourdatasuggestthatthesestructuresatslightlydifferent
we are investigating is also characterized by a high density
redshiftstend also to be distributed in differentregionsof the
of X-ray point sources (a fraction of which belonging to the
sky.Inparticular,galaxiesontheredtailofthemainspike,and large-scalestructureatz=0.78)andbyafewclustersofgalax-
specificallyatz∼ 0.776arelocatedinthesouthernpartofthe
ies(Hasingeretal.1998b).Afterthesubtractionoftheinstru-
field, while galaxies in the main spike at z∼ 0.784 are pref-
mental backgroundand the point source removalthrough the
erentially distributed in the northern part. The distribution of
proceduresdescribedin Zappacostaetal. (2002), we find dif-
thegalaxiesin thefartherspike atz∼ 0.807overlapswiththe
fuseemissioncoincidentwiththesuperstructure.However,the
previoustwo,butthepresenceofanothersourceat0.807(the
relatively extended wings of the ROSAT PSF from the point
onelocatedmosttothewestinFig.3),previouslyidentifiedby sources probably contribute significantly to the diffuse emis-
Lehmannetal. (2001), suggests thatthis substructureextends
sion. Indeed, this apparently extended emission elongated in
towardsthewest.Wehavetentativelyencircledthethreemain
the direction N–S is also found in the harder maps (R45 and
substructureswiththreeellipsesinFig.3.Onthewholethesu-
R67).Thespectralshapeofsuchextendedemissionisthesame
asforthepointsources,associatingthisemissionwithresidual
2 Theonlysourceofpossiblesystematicerrorsisspectralcalibra-
wings of point sources as well as a possible detection of un-
tion.However,suchsystematicerrorsaresmallerthantheestimated
resolved AGNs associated with the large-scale structure (e.g.
randomerrors.
3 AlternativelyitwouldbetheHαoflocalobjects(z∼0),butinthe Gillietal.2003).Theobservedhardnessoftheemissionisalso
lattercasetheobjectsshouldbeextendedandbright,unliketheones atvariancewithwhatfoundinotherfieldswherediffuseemis-
inoursamples. sionhasamuchsofterspectrum(e.g.Zappacostaetal.2002).
4 Zappacostaetal.:ConstrainingthethermalhistoryoftheWHIM
0.7840
36:00
0.782
0.782
0.7840
0.7848
0.7833
0.805
0.780
57:30:00
0.7756
0.8061 0.788
0.7932
0.7871
0.792
0.8069
0.7848
0.8077
0.7825
0.784
0.8069
0.8092 0.8 (phot)
24:00 0.7764 0.7779 0.7772
0.807
0.7764
18:00
30 10:54:00 30 53:00 52:30
Fig.3. Spatial distribution of the various spectroscopically identified sources and in particular of those at the redshift corre-
spondingtothespikesatz≈0.8.Crossesmarkthepositionofallthegalaxieswithredshiftsmeasuredbyus(thethickonesare
thesourcesbelongingtothesuperstructure).Thecirclesandthediamondarethearchivalobjectsatthesameredshiftofthespike
(as in Fig. 1). For all the galaxiesbelongingto the superstructurewe havereportedtheir measuredredshift. The largeellipses
tentativelyindicatetheregionswhichmaybemostlypopulatedbythethreesubspikes.
The presence of such N-S unresolved AGN emission further XMM(Hasingeretal.2001;Hasinger2003).Inparticular,the
suggeststhepresenceofanoverdenseregionofgalaxiesinthis 100 ksec observation obtained by Hasingeretal. (2001) was
areaoftheLockmanHole. performedwiththe“thin”filter,whichallowsthedetectionof
photons down to 0.2 keV. Additional 800 ksec of integration
Chandrahasamuchbetterangularresolutionwhichallows
(Hasinger2003)wereobtainedwiththe“medium”filter,which
theremovalofthecontributionfrompointsourceswithamuch
absorbsphotonswithE<0.5keV.Thelatterobservationcannot
higheraccuracy.However,itssensitivitydropsdrasticallyaten-
beusedtoconstraintheWHIMemissionbecauseofitsenergy
ergiesE<0.5keV, preventingusto studythe levelof soft dif-
cutoffat0.5keV,howeveritcanbeefficientlyusedtosubtract
fuse X-ray emission. The small field of view of ACIS-S (the
spuriouscontributionto the diffuse emission by hardsources,
Chandrachipwhichhashighersensitivityinthesoftbandthan
asweshalldiscusslateron.Inthefollowingwewillfocuson
ACIS-I)isalsoproblematictodetectextendedemission.
theanalisysoftheXMMdatatakenwiththe“thin”filter.
XMM has the appropriate compromise between angu-
lar resolution, good sensitivity in the softest X-ray band The major difficulty in using the soft energy band is the
at E<0.5 keV, and extension of the field of view. The presence of the electronic noise that dominates the emission
Lockman Hole has been subject of deep observations with (Lumbetal. 2002; Read&Ponman 2003). The spectrum of
Zappacostaetal.:ConstrainingthethermalhistoryoftheWHIM 5
theelectronicnoiseisverystable,anditsstatisticalnoisecon-
42:00
sists in a number of small-amplitude events occurring during
every frame exposure. There are no fluctuations in a number
of events, so the removal of these has no major influence on
36:00
thedetectionstatisticsfortheX-rayemission.Electronicnoise
hasasimilarspatialdistributionforthesamedetectorread-out
mode,butits spectrumvariesasa functionofframetime and 57:30:00
issubjecttoanenergyoffsetonthe10eVscaleforeachindi-
vidualpixel.Thisenergyoffsetappliesto allevents,resulting
in a decreaseof theenergyresolutionforanextendedsource. 24:00
Investigationsof the electronic noise by the EPIC calibration
team at the Max-Planck Institut fu¨r Extraterrestrische Physik
18:00
showedthatitispossible,byusingtheshapeoftheelectronic
noise,toactuallydeterminetheenergyoffsetineachpixelfor
eachobservationandthenefficientlyandaccuratelyremovethe 10:55:00 54:00 53:00 52:00 00
electronicnoisefromtheeventlists.Thesoftwareandprocess-
Fig.4.XMM-NewtonEPIC-pnwaveletreconstructedimagein
ingrecipesaremadeavailabletoageneraluserasanewtask,
the0.2–0.4keVband.Crosses,smallcirclesandthediamond
epreject,withinSAS6.0release.InthisPaperwedescribethe
arethegalaxiesbelongingtothesuperstructuredetectedinop-
resultsobtainedfor a single LockmanHole pointingas a part
ticalasinFig.3.Bigcirclesaretheregionswherewemeasured
ofsoftwaretestingstagebyKonradDennerl(seeDennerletal.
theresidualsoftX-rayflux.Thethinlineshowsthepositionof
(2004)andeprejecttaskdescription).Forfurtherdetailsofthe
theEPIC-pncamera.
noiseremovalandadiscussionofassociateduncertaintiessee
epnrejecttaskdescription.WeselectedthelongestEPIC-pnob-
servation, made with the “thin” filter and detector full frame
mode (for a description of EPIC-pn see Stru¨deretal. 2001),
which yields 35 ksec of useful exposure.For purposesof our
move instrumental scattering effect, two clusters located near
analysis,asubtractionofout-of-timeevents(OOTE)isnotnec-
the superstructure, as well as the contribution by unresolved
essary,giventhepositionofbrightsourceswithrespecttothe
AGNs(whichhaveamuchharderspectrumthantheWHIM).
superstructureand an orientationof CCDs in the selected ob-
Wethenusedthe0.2–0.4keVbandtoextractthecountsfrom
servation. The sensitivity reached by this observation is good
theobservationwiththethinfilter.Wealsoselectedtwoaddi-
enoughtosetrelativelytightconstraintsonthepresenceofdif-
tional regions, closer to the center of the pointing and at the
fusesoftX-rayemissionintheregionoftheLockmanHole.
edge of the CCD to estimate both a level of the foreground
The image in the softest available band 0.2–0.4 keV was
andbackgroundemissionclosetothepositionofourmeasure-
extractedandprocessedthroughavariablewaveletfilterdetec-
ment and to estimate the possible contributionof the induced
tionalgorithm(Vikhlininetal.1998)toidentifybothpointand
background,asduetothesoftprotons.Todothat,wemakea
extendedsources(seeZappacostaetal.2002, fordetails).We
reasonableassumption,basedonourbestknowledge,thatthe
set the wavelet peak detection threshold to 4σ, and followed
sky components and X-ray foreground components (such as
theextensionofthedetectedfluxdownto1.7σ.Weperformed
ten iterations at wavelet kernel scales ranging from 4′′ to 4′. the LocalHotBubble, LHB) haveflatdistributionon thesky,
atleastwithintheXMMfieldofview,andthusarevignettedby
We started from the smallest scale and removed the detected
thetelescope,whiletheinducedbackgroundcomponentshave
sources prior to proceeding with the next larger scale. “Point
aflatdistributionoftheirintensityoverthedetector(e.g.Lumb
sources” were identified as sources detected with kernels of
size4′′ (centralpartofthefieldofview)and8′′ (outerregion, etal.2002).
wherethePSFislarger).Theresultingwaveletmap(thesumof After subtraction of these in-field estimated background
allwaveletorders)isshowninFig.4.Severalpointsourcesare components,theresidualsoftX-rayfluxisF0.2−0.4keV = 1.5±
detectedaswellasafewknownclusters.However,wedonot 1.4×10−16ergs−1cm−2arcmin−2,thatincreasesby∼30%after
findanyevidencefordiffuseextendedemissiondowntoalimit thecorrectionforHIabsorption.Thisisaverymarginaldetec-
of2σofconfidencefromthehighorderwaveletmaps,particu- tion and could be explained by statistical fluctuations around
larlyintheregionwherethelarge-scalestructureatz=0.78has thezerovalue.Therefore,weconsidersuchafluxasanupper
beendetected. limittothesoftdiffuseX-rayemissioninthisregion.Itshould
Toplacelimitsonthepresenceofasoftdiffusecomponent benotedthattheLockmanHoleistheclearestwindowforthis
associatedwiththeopticalstructure,identifiedabove,wehave kindofX–raystudies,beingtheregionwiththelowestGalactic
selectedanumberofcirclesalongthesuperstructuretracedby hydrogencolumndensity.Thismeansthatasuperstructure,as
galaxies (see in Fig. 4 the big circles). We masked out from the one we have found in optical, should clearly show X–ray
finalanalysisallthe partswhereX-rayemission hasbeende- diffuseemissionduetothecollapsedgasinthedarkmatterpo-
tectedinthe0.5–2keVandinthe2–7.5keVbandsinthefull tential welltraced by the galaxies, unlessthe gastemperature
∼ 800 ksec XMM-Newton exposure on the Lockman Hole, is verylow. This issue will be discussed more in detail in the
obtained with the “medium” filter; in this way we could re- nextsection.
6 Zappacostaetal.:ConstrainingthethermalhistoryoftheWHIM
4. TemperatureoftheWHIMatz∼0.8 offinour0.2–0.4keVband,buttheconsequent K–correction
stillchangestheobservedfluxsignificantly,andmorespecifi-
Inthissectionwediscusstheupperlimitobtainedforthedif-
callybyafactorof1.8.Summarizingalloftheseeffects,inab-
fuseX-rayemissionandestimatewhetheritiscompatiblewith
senceoftemperatureevolutionthesurfacebrightnessexpected
the predictionsof the cosmologicalmodelsand/orconstraints
forWHIMassociatedwithasuperstructureatz=0.8relativeto
onthelattercanbeinferred.
asuperstructureatz=0.1is
Inthelocaluniverse(z <1)variousdetectionsofemission
btoyaWboHuItM∼0i.n2dikceaVte(tWemanpgereattuarle.s19ra9n7g;iZngapfproacmos∼ta1ekteaVl. 2d0o0w2n, SS00..22−−00..44kkeeVV((00..81)) = nn00..18!2 11++00..18!−4 (1.8)−1=3.4
2004;Sołtanetal.2002;Kaastraetal.2003;Finoguenovetal.
2003).Thelowertemperaturelimitcouldbeascribedtoobser- combinedwiththesurfacebrightnessoftheWHIMobservedin
vational issues, since the bulk of the emission from gas with theSculptorsupercluster,weobtainanexpectedsurfacebright-
the temperatures below 0.1 keV is absorbed even in regions nessfortheWHIMatz=0.8of
otefrilsotwicGofaltahcetiWcNHHIM.M(oCreenov&erO,lsotwrikOexry1g9e9n9;abFuinnodgaunecneocvhaertaacl-. S0.2−0.4keV(0.8)≈3.1×10−15ergs−1cm−2arcmin−2
2003) precludes WHIM detection through the OVII emission whichisroughlyanorderofmagnitudemore(morethanafac-
lines. The situation will change, however, with advent of mi- torof8)thantheupperlimitobtainedbyusforthesuperstruc-
crocalorimeters with ’large grasp’. The alternative technique tureintheLockmanHole.
of using the OVII and OVIII X-ray absorption lines allowed One possibility could be that filaments are more rare at high
the detection of Local WHIM at much lower temperatures redshift,andthatnoneispresentinourfield.However,cosmo-
(Nicastroetal.2002).Forahomogeneouscomparison,herewe logical simulations predict that filamentary structures should
focusonthepropertiesoftheWHIMdetectedinemission. alreadybeformedbyz∼1,andthatsuper-structures(tracedby
We take as a reference for the local universe the WHIM overdensityof galaxies) are the locations where filaments are
emission from the Sculptor supercluster (z=0.1) reported in most likely to be present. Since we are clearly investigating
Zappacostaetal. (2004). In this region the median tempera- one of such super-structures (as demonstrated by the galaxy
ture detected for the WHIM is about T∼0.4 keV (extending overdensity), the filaments expected by cosmological models
up to 0.5 keV and also to <0.3 keV). The minimum average atsuchredshiftsaremuchmorelikelytobepresentinourre-
WHIMemission(i.e.assumingthemaximumpossiblesubtrac- gion than in the field. Another, more likely possibility is that
tionoftheLHBcontribution)is∼ 240×10−6ctss−1arcmin−2 baryonicfilamentsareindeedpresentinthislargescalestruc-
in the 0.14–0.28 keV ROSAT band, corresponding to a flux ture, but that their temperature at z=0.8 is significantly lower
F0.2−0.4keV = 9×10−16ergs−1cm−2arcmin−2. We assume that thaninthelocalUniverse(z≤0.1),makingthemundetectable.
these values are roughly representative of the emission by Indeed a lower temperature moves the thermal cutoff of the
WHIM in the superstructures of the local universe, however spectrumtolowerenergiesandspecificallybelowtheband0.2-
wewillshowthatourconclusionsarenotcriticallydependent 0.4keVobservedbyus.Wehavecalculatedthatthemaximum
onsuchassumptions.Ifwemovesuchamediumtoaredshift allowedtemperatureatz=0.8,whichwouldmaketheobserved
of0.8,inabsenceofevolutionitsdensitywillincreasepropor- fluxconsistentwithourupperlimitinthe0.2–0.4keVband,is
tionally to (1+z)3 due to the Hubble flow, andmore precisely about0.07–0.1keV.Itisimportanttonotethatthistemperature
byafactor[(1+0.8)/(1+0.1)]3=4.4.However,wealsohaveto limitisweaklysensitiveonthedensityandemissivityassump-
includetheintrinsicdensityevolutionexpectedfortheWHIM. tions discussed above. Indeed, the importantresult is that the
From Dave´ etal. (2001, therein Fig. 4) we can estimate that expected flux is so much higher than our upper limit that the
theWHIMathighredshift(z∼1)shouldbedenserbyafactor onlywayoutistorequirethetemperaturetobelowenoughto
of ∼1.5 than locally4. When combined with the effect of the movethethermalcutoffbelowthe0.2–0.4keVband.
Hubbleflow,thedensity increasesbya factorn /n = 6.6.
0.8 0.1
Thethermalemissionincreasesproportionallyton2,therefore
5. ThethermalhistoryoftheWHIM
from z=0.1 to z=0.8 the WHIM emissivity is expected to in-
creasebyafactorǫ0.8/ǫ0.1 = (n0.8/n0.1)2 = 43.6.Sinceweare AsummaryoftheconstraintsontheobservedWHIMtemper-
observing(orconstraining)thesurfacebrightness,wehavealso atures as a function of redshiftis given in Fig. 5. We include
toaccountforacosmologicaldimmingproportionalto(1+z)4, resultsbyotherworkslistedinTable2(bothemissionandab-
i.e. [(1+0.8)/(1+0.1)]4 =7.2. Assuming that the temperature sorption detections). The points give either the median value
doesnotchange(ourworkinghypothesis)thespectralshapere- or the best fitting value. As expected, imaging and statistical
mainsunchanged,buttheredshiftmovesthethermalspectrum measurements,whichtendtodetectthebrightestWHIMemis-
tolowerenergies,certainlynotenoughtopushthethermalcut- sions,obtainhighertemperaturevalues.Thisisaconsequence
oftheWHIMemissivityperunitenergybeingproportionalto
4 These predictions refer to the bulk of the WHIM that virtually,
the temperature5. On the contrary,absorptionsystems sample
havingonaveragelowerdensitiesthantheWHIMinsuperclusteren-
vironment,shouldnotbedetectablewiththepresentdayinstruments. 5 Thethermalemissivityperunitenergyisrelatedtothetempera-
ItispossiblethatthedenserWHIMphaseinthesuperclusterenviron- tureT andthedensitynasT−0.5n2(atenergiesbelowthethermalcut-
mentevolveinadifferentway.Wearenotawareoftheoreticalworks off).OntheotherhandDave´etal.(2001)showedthattheWHIMtem-
exploringtheevolutionofthisdenseWHIMphase. peratureanddensityaredirectlyproportional,andthereforeǫ ∝T1.5.
Zappacostaetal.:ConstrainingthethermalhistoryoftheWHIM 7
Fig.5. Observed WHIM temperatures as a function of the redshift taken from the reference reported in Table 2. The solid
linesarethepredictionsofequation4inCen&Ostriker(1999)andDave´etal.(2001).Theshadedregionshowstheintervalof
temperatureswhichincludes50%oftheWHIMasreportedinFig.5ofDave´ etal.(2001).
random regions of the WHIM, including the very low tem- It is possible to go beyond this qualitative statement and
perature regions. Soft excess around clusters also tend to be investigate the quantitative agreement with the model predic-
biased against high temperatures,because the latter would be tions.Cen&Ostriker(1999)presentedintheirpaperasimple
confusedwiththeclusteremissionitself.MoreovertheWHIM argumentthatcanreproducerepresentativevaluesforthetem-
originofclustersoftexcesseshaverecentlybeenpartlybrought perature of the WHIM (see also Fig. 5 in Dave´etal. 2001).
intoquestionbycosmologicalsimulations(Mittazetal.2004; Theyinferthepostshocktemperatureofacosmicgascollaps-
Chengetal.2004). inginsideaslightlynonlinearlarge-scalestructureofsizeLas
A homogeneous comparison should be limited to detec- T ∝ c2s ≈ K(Lt−H1(z))2, where cs is the gas sound speed and
tions(andupperlimits)madewiththesametechnique.Inthis tH(z) the Hubbletime at the redshiftof interest and K is con-
paperwehavesearchedforWHIMemission,andthereforeour stant(seeequation4inCen&Ostriker1999).Suchsimplees-
upper limit must be compared with the other detections and timateisshowntocorrectlyreproducetheresultsfromnumer-
upperlimits obtainedwith imagingor statistical detectionsof icalsimulations(Cen&Ostriker1999;Dave´etal. 2001).The
WHIMemission(fullsymbolsinFig. 5).Suchmeasurements sameapproachwasusedbyDave´ etal.(2001)whoobtainsim-
indicate that the observed WHIM temperature decreases with ilarresults.InFig.5weshowthetrendsoftheWHIMtempera-
redshift(Fig. 5),justaspredictedbycosmologicalmodels.The tureasafunctionofredshiftobtainedbyCen&Ostriker(1999)
numberofobservationaldataissmallandwithlargeerrorbars, andDave´etal.(2001)byusingtheargumentdiscussedabove
however the plot in Fig. 5 is the first attempt of constraining (the shaded area indicates the interval of temperatures which
theevolutionoftheWHIMtemperaturewithredshift,withthe includes ∼50% of the WHIM in the Dave´etal. (2001) distri-
currentlyavailabledata. butions). Although the uncertaintiesof the theoreticalmodels
8 Zappacostaetal.:ConstrainingthethermalhistoryoftheWHIM
Table2.WHIMtemperatures(sortedbyredsfhift)measuredsofarwithdifferentmethodsbyvariousauthors.
Object z ObjectType Method kT(keV) Reference
PKS2155-304 0a QSO SpectralAbsorption ∼0.043 Nicastroetal.(2002)
Mkn421 0.0116a QSO SpectralAbsorption 0.06÷0.22 Nicastro(2003)
Coma 0.0231 ClusterofGalaxies Softexcess 0.22±0.01b Finoguenovetal.(2003)
Coma 0.0231 ClusterofGalaxies Softexcess 0.187±0.011 Kaastraetal.(2003)
Abell2052 0.035 ClusterofGalaxies Softexcess 0.208±0.007 Kaastraetal.(2003)
MKW3s 0.045 ClusterofGalaxies Softexcess 0.23±0.012 Kaastraetal.(2003)
Sersic15903 0.058 ClusterofGalaxies Softexcess 0.18±0.006 Kaastraetal.(2003)
Abell1795 0.0631 ClusterofGalaxies Softexcess 0.2±0.008 Kaastraetal.(2003)
AbellClusters 0.068+0.102c ClustersofGalaxies Statistical <1 Sołtanetal.(2002)
−0.023
SculptorScl 0.105±0.01 Supercluster Statistical 0.4+0.1c Zappacostaetal.(2004)
−0.2
Abell2125 0.247 ClusterofGalaxies Imaging 0.85+0.45d Wangetal.(1997)
−0.39
Warwickfield 0.45±0.15 Field Imaging 0.3±0.15c Zappacostaetal.(2002)
LockmanHole 0.791±0.016 Superstructure Imaging <0.1 thispaper
a redshiftoftheabsorptionsystem
b 1σerror-bars
c medianvalue
d 90%levelofconfidence
are large, both predictionsfit reasonablywell the data points, tiallysupported by the ItalianMinistryof Research (MIUR) and by
and in particular the temperatures inferred from the imaging the Italian Institute of Astrophysics (INAF). We thank G. Hasinger,
andstatisticalmethods. H. Boehringer, X. Barcons and A. Fabian for letting us to use the
Lockman Hole image to remove point source contamination and
Konrad Dennerl for his assistance withXMM data analysis. AF ac-
knowledges support fromBMBF/DLRunder grant 50OR0207 and
6. Conclusions MPG.
We have identified a large-scale structure of galaxies in the
Lockman Hole at redshift z ∼ 0.79± 0.015 by means of an
opticalspectroscopicsurvey.Thesuperstructureextendsovera References
regionof more than 7.5 Mpc (in projection)and is structured
Cen,R.&Ostriker,J.P.1999,ApJ,514,1
inthreesub-concentrationsatmedianredshiftsof0.776,0.784
Cheng, L.-M., Borgani, S., Tozzi, P., et al. 2004, in astro-
and0.806.InthissuperstructuretheWHIMpredictedbycos-
ph/0409707
mological models should have already formed. By analysing
Ciliegi,P.,Zamorani,G.,Hasinger,G.,etal.2003,A&A,398,
ROSATandXMMpointingswecouldsetatightupperlimiton
901
the WHIM emission associated with the superstructure.From
Dave´,R.,Cen,R.,Ostriker,J.P.,etal.2001,ApJ,552,473
this flux limit we could estimate an upper limit of ∼0.1 keV
Dennerl,K.,Aschenbach,B.,Briel,U.G.,etal.2004,inastro-
on the WHIM temperature at z∼0.8. The combination of this
ph/0407637
tightupperlimitwithotherpreviousWHIMtemperaturemea-
Fadda,D.,Flores,H.,Hasinger,G.,etal.2002,A&A,383,838
surements(atlowerredshifts)stronglysuggeststhattheWHIM
Finoguenov,A.,Briel,U.G.,&Henry,J.P.2003,A&A,410,
temperature must be rapidly decreasing with redshift, as ex-
777
pectedbythecosmologicalmodels.Theagreementofthered-
Gilli,R.,Cimatti,A.,Daddi,E.,etal.2003,ApJ,592,721
shift distribution of the observed WHIM temperatures with
Hasinger,G.2003,inastro-ph/0310804
thecosmologicalpredictions(Cen&Ostriker1999;Dave´etal.
Hasinger, G., Altieri, B., Arnaud,M., et al. 2001,A&A, 365,
2001) is reasonably good even from a quantitative point of
L45
view. However furtherwork is requiredto improvethe statis-
Hasinger,G.,Burg,R.,Giacconi,R.,etal.1998a,A&A,329,
tics on the WHIM temperaturemeasurements(or constraints)
482
athighredshift.
Hasinger, G., Giacconi, R., Gunn, J. E., et al. 1998b, A&A,
340,L27
Acknowledgements. Thepaperisbasedonobservationsobtainedwith
Kaastra, J. S., Lieu, R., Tamura, T., Paerels, F. B. S., & den
XMM-Newton,anESAsciencemissionwithinstrumentsandcontri-
Herder,J.W.2003,A&A,397,445
butionsdirectlyfundedbyESAMemberStatesandtheUSA(NASA).
Lehmann,I.,Hasinger,G.,Schmidt,M.,etal.2001,A&A,371,
TheXMM-NewtonprojectissupportedbytheBundesministeriumfu¨r
BildungundForschung/DeutschesZentrumfu¨rLuft-undRaumfahrt 833
(BMFT/DLR),theMax-PlanckSocietyandtheHeidenhain-Stiftung, Lumb,D.H.,Warwick,R.S.,Page,M.,&DeLuca,A.2002,
and also by PPARC, CEA, CNES, and ASI. This work was par- A&A,389,93
Zappacostaetal.:ConstrainingthethermalhistoryoftheWHIM 9
Mainieri,V.,Bergeron,J.,Hasinger,G.,etal.2002,A&A,393, AppendixA: CatalogofObjects
425
Mathur,S.,Weinberg,D.H.,&Chen,X.2003,ApJ,582,82 Id RA(J2000) DEC(J2000) z Quality
Mittaz,J.,Lieu,R.,Cen,R.,&Bonamente,M.2004,inastro- 1 10h53m41.77s +57◦21′29.56′′ 0.281 low
ph/0409661
2 10h53m45.41s +57◦21′40.96′′ 0.085 high
Nicastro,F.2003,inIAUSymposium
Nicastro,F.,Zezas,A.,Drake,J.,etal.2002,ApJ,573,157 3 10h53m43.71s +57◦22′01.01′′ 0.608 low
Read,A.M.&Ponman,T.J.2003,A&A,409,395 4 10h53m28.84s +57◦22′05.78′′ 0.776 low
Sołtan, A. M., Freyberg, M. J., & Hasinger, G. 2002, A&A, 5 10h53m36.44s +57◦22′17.45′′ 0.482 low
395,475
6 10h53m33.12s +57◦22′34.99′′ 0.735 high
Stru¨der,L.,Briel,U.,Dennerl,K.,etal.2001,A&A,365,L18
Vikhlinin,A.,McNamara,B.R.,Forman,W.,etal.1998,ApJ, 7 10h53m30.17s +57◦22′41.48′′ 0.863 low
502,558 8 10h53m30.86s +57◦22′48.10′′ 0.634 low
Wang, Q. D., Connolly, A., & Brunner, R. 1997, ApJ, 487, 9 10h53m19.95s +57◦22′51.48′′ 0.528 low
L13+
10 10h53m34.39s +57◦23′06.16′′ 0.405 low
Zappacosta, L., Maiolino, R., Mannucci, F., Gilli, R., &
Schuecker,P.2004,MNRASaccepted,astro-ph/0402575 11 10h53m38.38s +57◦23′06.04′′ 0.419 high
Zappacosta,L.,Mannucci,F.,Maiolino,R.,etal.2002,A&A, 12 10h53m44.53s +57◦23′26.43′′ 1.017 high
394,7 13 10h53m37.79s +57◦23′28.09′′ 0.436 low
14 10h53m41.86s +57◦23′53.08′′ 0.892 low
15 10h53m45.87s +57◦24′10.48′′ 0.777 high
16 10h53m38.60s +57◦24′08.85′′ 0.484 high
17 10h53m47.71s +57◦27′15.48′′ 0.785 high
18 10h53m43.83s +57◦26′59.82′′ 0.729 low
19 10h53m54.44s +57◦26′11.21′′ 0.644 low
20 10h53m40.40s +57◦26′36.23′′ 0.821 low
21 10h53m45.22s +57◦26′25.45′′ 0.808 high
22 10h53m26.84s +57◦26′30.70′′ 0.354 high
23 10h53m45.09s +57◦25′56.43′′ 0.783 high
24 10h53m23.87s +57◦27′29.10′′ 0.807 low
25 10h53m44.91s +57◦26′03.48′′ 0.619 high
26 10h53m39.80s +57◦25′05.50′′ 0.807 high
27 10h53m39.41s +57◦25′30.62′′ 0.335 high
28 10h53m44.89s +57◦27′27.74′′ 0.594 high
29 10h53m32.13s +57◦25′07.74′′ 0.479 high
30 10h53m44.77s +57◦24′45.67′′ 0.917 low
31 10h53m27.98s +57◦27′53.20′′ 0.146 low
32 10h53m42.73s +57◦27′51.53′′ 0.684 high
33 10h53m48.68s +57◦28′33.26′′ 0.279 low
34 10h53m58.52s +57◦28′37.64′′ 0.751 low
35 10h53m30.46s +57◦28′35.05′′ 0.531 low
36 10h53m30.22s +57◦28′55.08′′ 1.011 low
37 10h53m48.84s +57◦29′17.08′′ 0.889 high
38 10h53m56.29s +57◦29′25.65′′ 0.580 high
39 10h53m50.98s +57◦29′33.65′′ 0.776 low
40 10h53m40.91s +57◦29′42.15′′ 0.919 low
41 10h53m44.08s +57◦30′17.99′′ 0.490 low
42 10h53m26.96s +57◦31′10.83′′ 0.377 low
43 10h53m56.19s +57◦31′14.36′′ 0.701 low
10 Zappacostaetal.:ConstrainingthethermalhistoryoftheWHIM
Id RA(J2000) DEC(J2000) z Quality Id RA(J2000) DEC(J2000) z Quality
44 10h53m32.85s +57◦31′52.09′′ 0.422 high 89 10h54m11.02s +57◦31′42.79′′ 0.396 high
45 10h53m48.02s +57◦32′05.64′′ 0.783 low 90 10h54m22.84s +57◦31′12.17′′ 0.343 high
46 10h53m40.31s +57◦32′09.61′′ 0.692 low 91 10h54m05.16s +57◦31′20.91′′ 0.977 low
47 10h53m42.17s +57◦32′25.60′′ 0.321 high 92 10h54m05.16s +57◦31′22.56′′ 1.062 low
48 10h53m49.23s +57◦35′39.75′′ 0.231 low 93 10h54m10.99s +57◦30′56.14′′ 0.226 low
49 10h53m49.31s +57◦34′28.74′′ 0.391 low 94 10h54m30.51s +57◦29′57.57′′ 0.619 high
50 10h53m40.12s +57◦36′19.28′′ 0.784 high 95 10h54m01.59s +57◦28′13.58′′ 0.890 low
51 10h53m33.68s +57◦36′10.40′′ 0.500 high 96 10h54m03.72s +57◦29′15.01′′ 0.806 high
52 10h53m37.67s +57◦36′04.33′′ 0.405 low 97 10h54m23.93s +57◦29′10.58′′ 0.410 low
53 10h53m40.23s +57◦35′50.81′′ 0.933 low 98 10h54m16.76s +57◦29′03.93′′ 0.589 low
54 10h53m44.73s +57◦36′08.73′′ 0.591 low 99 10h54m03.14s +57◦28′57.45′′ 0.793 low
55 10h54m01.68s +57◦35′34.23′′ 0.256 low 100 10h54m22.93s +57◦28′42.62′′ 0.787 high
56 10h53m41.03s +57◦35′12.77′′ 0.493 low 101 10h54m29.85s +57◦28′17.79′′ 0.711 high
57 10h53m44.76s +57◦35′15.24′′ 0.093 low 102 10h54m21.11s +57◦30′14.17′′ 0.965 low
58 10h53m42.62s +57◦34′03.58′′ 0.784 high 103 10h54m33.97s +57◦29′52.14′′ 0.276 high
59 10h53m41.48s +57◦34′28.55′′ 0.256 low
60 10h53m46.09s +57◦34′11.09′′ 1.084 low
61 10h53m42.87s +57◦34′03.80′′ 0.784 high
62 10h53m36.94s +57◦33′57.55′′ 0.420 high
63 10h53m36.32s +57◦33′13.74′′ 0.706 high
64 10h53m36.32s +57◦33′12.91′′ 0.784 high
65 10h53m45.19s +57◦33′42.05′′ 0.701 high
66 10h54m03.72s +57◦33′29.03′′ 0.662 low
67 10h53m37.33s +57◦32′52.90′′ 0.562 low
68 10h53m38.25s +57◦32′49.74′′ 0.785 high
69 10h53m29.13s +57◦36′41.92′′ 0.427 high
70 10h52m50.23s +57◦28′25.09′′ 0.536 low
71 10h52m49.71s +57◦28′41.56′′ 0.342 high
72 10h52m39.67s +57◦28′57.53′′ 0.669 low
73 10h52m48.20s +57◦29′54.99′′ 0.291 low
74 10h52m41.79s +57◦30′53.26′′ 0.206 high
75 10h52m42.14s +57◦31′23.02′′ 0.121 high
76 10h52m55.15s +57◦31′47.06′′ 0.343 low
77 10h52m53.80s +57◦33′11.11′′ 0.669 high
78 10h52m45.50s +57◦28′05.49′′ 0.606 high
79 10h52m41.35s +57◦27′21.95′′ 0.552 high
80 10h52m49.02s +57◦26′39.76′′ 0.322 low
81 10h52m42.18s +57◦26′02.02′′ 0.343 high
82 10h54m20.31s +57◦24′02.86′′ 0.778 low
83 10h54m29.59s +57◦24′11.82′′ 0.776 high
84 10h54m23.22s +57◦24′46.59′′ 0.809 low
85 10h54m17.46s +57◦25′02.44′′ 0.138 high
86 10h54m27.92s +57◦26′14.55′′ 0.578 high
87 10h54m14.04s +57◦27′24.94′′ 0.913 low
88 10h54m29.28s +57◦27′44.63′′ 0.306 low