φ-meson production at forward rapidity in p–Pb collisions at √sNN=5.02 TeV and in pp collisions at √s=2.76 TeV

Contents lists available atScienceDirect

Physics Letters B

www.elsevier.com/locate/physletb

φ -Meson production at forward rapidity in p–Pb collisions at

√ s _NN = ⁵ . 02 TeV and in pp collisions at √

s = ² . 76 TeV

.ALICE Collaboration

a r t i c l e i n f o a b s t ra c t

Articlehistory:

Received16November2015

Receivedinrevisedform8December2016 Accepted9January2017

Availableonline16February2017 Editor:L.Rolandi

Thefirststudyofφ-mesonproductioninp–Pbcollisionsatforwardandbackwardrapidity,atanucleon–

nucleoncentre-of-massenergy√_s

NN=⁵.02 TeV,hasbeenperformedwiththeALICEapparatusatthe LHC.Theφ-mesonshavebeenidentifiedinthedimuondecaychannelinthetransversemomentum(pT) range1<pT<7 GeV/c,bothinthep-going(2.03<y<3.53)andthePb-going(−⁴.46<y<−².96) directions —where ystandsforthe rapidityinthenucleon–nucleon centre-of-mass—the integrated luminosity amounting to5.01±⁰.19 nb⁻¹ and5.81±⁰.20 nb⁻¹,respectively, for thetwo data sam- ples.Differentialcrosssectionsas afunctionoftransversemomentumandrapidityare presented.The forward–backwardratioforφ-mesonproductionismeasuredfor2.96<|^y|<3.53,resultinginaratio

∼⁰.5 withnosignificantpTdependencewithintheuncertainties.The pTdependenceoftheφnuclear modificationfactorRpPbexhibitsanenhancementuptoafactor 1.6atpT=^{3–4 GeV}/cinthePb-going direction.The pTdependenceoftheφ-mesoncrosssectioninppcollisionsat√_s

=².76 TeV,whichis used to determineareference for thep–Pb results,is alsopresented herefor 1<pT<5 GeV/c and 2.5<y<4,fora78±^{3 nb−1integratedluminositysample.}

©^{2017TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense} (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP³.

1. Introduction

Proton–nucleus (p–A) collisions are of special interest in the context of high-energy nuclear physics for two reasons. On one hand, a precise characterisation of particle production processes in p–A collisions is needed as a reference for nucleus–nucleus data. This allows in-medium effects — linked to the forma- tion of a deconfined phase of the QCD matter, the quark–gluon plasma (QGP)[1–3] —tobedisentangledfromtheeffectsalready present in cold nuclear matter. Among them, a sizeable role is played by the transverse momentum broadening of initial-state partons due to multiple scattering inside the nucleus, responsi- blefortheCronineffect [4],which mayleadto an enhancement of intermediate-pT hadron spectra.In addition, p–A collisions at LHC energies provide a wayto probe the partondistributions of the colliding nucleus at small values of Bjorken-x, in a regime where parton densities can reach saturation [5,6]. In particular, thesmallestxvaluescontributingtothewavefunctionofthecol- liding nucleus can be probed by looking at particle production atlarge rapidities,in the p-going direction. Such a measurement canthus extendtowards lower x-valuestheresults ofthelower- energymeasurements by the PHOBOS andBRAHMS experiments atRHIC[7,8].Measurementsofidentifiedparticleproductionmay,

E-mailaddress:[email protected].

inparticular,provideusefulconstraintsforforthcomingtheoretical studiesofthesaturationmechanismatsmallx.

We havealready reportedresults on chargedparticle produc- tioninp–Pbcollisionsatmid-rapidity.Theseresultsfocusedonthe pseudorapidity density[9]andthe pT dependenceofthenuclear modificationfactor[10–12];thelatterwas foundtobeconsistent with unity for pT^{2 GeV}/c. The nuclear modification factor of charged hadrons was also studied by the BRAHMS and PHOBOS Collaborations in d–Au collisions at the nucleon–nucleon centre- of-mass energy √

s_NN=^{200 GeV at RHIC [13,14], as a function} of pseudorapidity, where values smaller than unity were found for

η

1,correspondingtothed-goingdirection.

InthisLetterwereport themeasurementofφ-mesonproduc- tion atforwardrapidityin p–Pbcollisionsat √

sNN=⁵.02 TeV in the transverse momentum (pT) range 1<pT<7 GeV/c, for the centre-of-massrapidity(y)ranges2.03<y<3.53 (p-goingdirec- tion) and−⁴.46<y<−².96 (Pb-goingdirection),inthe dimuon decaychannelwiththeALICE detector.Thismeasurementextends theinvestigationoflight-flavourparticleproductiontoforwardra- pidity.Atthesametime,itrepresentsanessentialbaselineforthe understanding of φ productionin heavy-ion collisions, where an enhancement ofstrange particleyields relative to the onesmea- suredinppcollisionshasbeenproposedlongagoasasignatureof theformationofaQGPphase[15–17]triggeringanintenseexper- imentaleffortalreadyatSPSandRHICenergies [18–24].Itshould http://dx.doi.org/10.1016/j.physletb.2017.01.074

0370-2693/©^{2017TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense}(http://creativecommons.org/licenses/by/4.0/).Fundedby SCOAP³.

benotedthat,despiteitshiddenstrangeness,producingaφ-meson inahadroniccollisionstillimpliesthecreationofas¯spairasitis thecaseforotherstrangehadrons,evenifthehadronisationmech- anismscandifferinreasonofthe differentquark composition.In thiscontext,thep–Pbdatapresentedherewillprovidean impor- tantreferenceforfuturemeasurementsinPb–Pb collisionsinthe LHCRun 2,whichwillbeperformedatacomparableenergy.

The differential φ-meson crosssection asa function oftrans- verse momentum is also presented for pp collisions at √

s= 2.76 TeV. This measurement complements the ALICE results on φ-meson production in pp collisions at √

s=^{7 TeV, already re-} portedin[25] and,combinedwiththelatter,isusedtobuildthe ppreferenceforthep–Pbmeasurementspresentedhere.

2. Experimentalsetup

AfulldescriptionoftheALICE detectorcanbefoundin[26,27].

Theresults presentedinthisLetter havebeenobtaineddetecting muon pairs withthemuon spectrometer, covering the pseudora- pidityregion−⁴<

η

lab<−².5.Hereandinthefollowing,thesign of

η

lab is determinedby the choiceof theLHC referencesystem.

The other detectorsrelevantfor theanalysisare the SiliconPixel Detector (SPD)oftheInnerTrackingSystem (ITS),theV0detector andtheZeroDegreeCalorimeters (ZDC).

Theelementsofthemuonspectrometerareahadronabsorber, followedbyasetoftrackingstations,adipolemagnet,anironwall actingasmuonfilterandasetoftriggerstations.Thehadronab- sorberis madeofcarbon,concreteandsteelandisplaced0.9 m away from the interaction point. Its total material budget cor- responds to 10 hadronic interaction lengths. The dipole magnet provides an integrated magnetic field of 3 T·^{m in the vertical} direction. The muon tracking is provided by five tracking sta- tions,eachonecomposed oftwo cathodepad chambers.Thefirst twostationsarelocatedupstream ofthedipole magnet,thethird one inthe middle ofits gap andthe last two downstream ofit.

A 1.2 mthickiron wall,corresponding to7.2 hadronicinteraction lengths, isplaced betweenthe trackingandtrigger detectorsand absorbstheresidualsecondaryhadronsemergingfromthehadron absorber.The hadron absorber together with the iron wall stops muons with total momentum lower than ∼^{4 GeV}/c. The muon trigger detector consists of two stations, each one composed of two planes of resistive plate chambers, installed downstream of themuonfilter.

TheSPDconsistsoftwosiliconpixellayers,coveringthepseu- dorapidity regions |

η

lab|<2.0 and |

η

lab|<1.4 for the inner and outer layer, respectively. It is used for the determination of the primary interaction vertex position. The V0 is composed of two scintillator hodoscopescovering thepseudorapidity regions 2.8<

η

_lab<5.1 and −³.7<

η

_lab<−¹.7. It is used in the definition ofthe minimum bias trigger signal, andallows the offline rejec- tionofbeam-haloandbeam-gasinteractionstobeperformed.The ZDCdetectors,positioned symmetrically at112.5 m fromthe in- teraction point, are used toclean the eventsample by removing beam–beamcollisionsnotoriginatingfromnominalLHCbunches.

3. Dataselectionandsignalextraction

TheanalysispresentedinthisLetterisbasedontwodatasam- ples, collected by ALICE during the 2013 p–Pb and pp runs at

√s_NN=⁵.02 TeV and √

s=².76 TeV, respectively. In thissection wepresentthedetailsofthedataselection,aswellastheproce- durefollowedfortheextractionoftheφ-mesonsignal.

3.1. Dataselection

TheMinimum-Bias (MB)triggerfortheconsidereddatasample is given by the logical AND ofthe signalsin the two V0 detec-

tors [28]. Events containing a muon pair are selected by means ofa specificdimuontrigger,basedonthe detectionoftwo muon candidate tracksinthe triggersystemofthe muon spectrometer, incoincidencewiththeMBcondition.Duetotheintrinsicmomen- tumcutimposedbythedetector,onlymuonswithp_T⁰.5 GeV/c managetoleaveasignalinthetriggerchambers.

BecauseofthedifferentenergyoftheLHCprotonandPbbeams (Ep=^{4 TeV, E}Pb=¹.58 A·^{TeV),in p–Pb collisions the nucleon–}

nucleon centre-of-mass moves in the laboratory with a rapidity y0=⁰.465 in thedirectionoftheprotonbeam. Thedirectionsof theprotonandPbbeamorbitswereinvertedduringthep–Pbdata taking period. Thisallowed the ALICE muon spectrometer to ac- cess two different rapidity regions¹: the region 2.03< y<3.53 wheretheprotonbeamisdirected towardsthemuonspectrome- ter (p-going direction) andtheregion −⁴.46<y<−².96 where thePbbeamisdirectedtowardsthemuonspectrometer(Pb-going direction).Inthefollowing,thesetworapidityranges arealsore- ferred to as“forward” and“backward”, respectively.For pp colli- sionsat√

s=².76 TeV themuonspectrometercoverstherapidity region2.5<y<4.²

Background events not coming from beam–beam interactions are rejected by performing an offline selection, based on the re- quirementthatthe timingsignalsfromtheV0andZDCdetectors arecompatiblewithacollisionoccurringinthefiducialinteraction region|^zvtx|^{10 cm.}

Theintegratedluminosityforthep–Pbdatasampleswas eval- uatedasL_int=^NMB/

σ

MB,whereNMB isthenumberofMBevents corresponding to the analysed triggered events,and

σ

MB the MB trigger cross section. The value of NMB was obtained by averag- ingtheresultsoftwodifferentmethods—onebasedontheratio of trigger rates and the other based on the offline selection of dimuoneventsintheMBdatasample[29]—whiletheMBtrigger crosssections

σ

MB were measuredwithavanderMeerscanand found tobe 2.09±⁰.07 b and2.12±⁰.07 b,respectively,forthe beam configurations corresponding to the forwardandbackward rapidity coverageofthemuon spectrometer[30].Forthe ppdata sample, the integrated luminosity is calculated with the method described in[31],usingasreferencetheMB triggercrosssection

σ

MB=⁴⁷.7±⁰.9 mb,measuredinavanderMeerscan[32].

TheresultingvaluesofL_intfortheanalysedp–Pbdatasamples are5.01±⁰.19 nb⁻¹ and5.81±⁰.20 nb⁻¹ [29,30]—correspond- ing to∼^{24000 and}∼²⁶000 reconstructed φ→

μμ

decays (see next section) — respectively for the forward and backward ra- pidity regions. For thepp data sample, theintegrated luminosity amounts to 78±^{3 nb−1 for a total number of} ∼^{1400 recon-} structedφ→

μμ

decays.

Track reconstruction in the muon spectrometer is based on a Kalman filter algorithm [25,33,34]. Muon identification is per- formed by requiring the candidate track to match a track seg- mentinthetriggerchambers(triggertracklet).Thisrequestselects muonswithpT,μ⁰.5 GeV/c and,asaconsequence,significantly affects the collected statistics for dimuons with invariant mass ^{1 GeV}/c²andp_T^{1 GeV}/c.Itisalsorequiredthatmuontracks lieinthepseudorapidityinterval−⁴<

η

μ<−².5,where

η

μisde- fined inthelaboratoryframe,inordertoremovethetracksclose to the acceptanceborders of thespectrometer, wherethe accep- tance drops abruptly. Selected tracks are finally required to exit thehadronabsorberataradialdistancefromthebeamaxis,R_abs, intherange17.6<R_abs<89.5 cm:thiscut,forallpracticalpur- posesequivalenttotheoneon

η

μ,explicitlyensurestherejection

1 Thesignofyisdefinedbyassumingtheprotonbeamtohavepositiverapidity.

2 Inthiscasethesignofyisdefinedbyassumingtheprotonbeamenteringthe muonspectrometertohavepositiverapidity.

Fig. 1.Dimuonmassspectrumaftercombinatorialbackgroundsubtraction: pT-integratedppsample(toppanels)and pT-integratedp–Pbsampleinthebackward(centre panels)andforward(bottompanels)rapidityregions,comparedtotheresultofthehadronic-cocktailandtheempirical-functionfits(left- andright-columnpanels,respec- tively).Errorboxesondatapoints(wellvisibleonlyinsomeregionsontheplots)representthesystematicuncertaintyduetothecombinatorialbackgroundsubtraction, whileerrorbarsaccountforthestatisticaluncertainty.Thewidthofthehadronic-cocktailfitresult(redband)combinesthestatisticaluncertaintiesofthefreefitparameters withthesystematicuncertaintiesonthefixedparameters(seetext).(Forinterpretationofthereferencestocolourinthisfigure,thereaderisreferredtothewebversionof thisarticle.)

oftrackscrossingtheregionoftheabsorberwiththehighestden- sitymaterial,wheremultiplescatteringandenergylosseffectsare largeandcanaffectthemassresolution.Muonpairsarebuiltcom- biningtwomuontracksthatsatisfytheabovecuts.

3.2.Signalextraction

TheOpposite-Sign(OS)muonpairsarecomposedofcorrelated anduncorrelatedpairs. The former contain the signal of interest forthe present analysis, while the latter— mainly comingfrom semi-muonicdecays of pions andkaons — form a combinatorial background.Thecontribution ofthecombinatorial backgroundto theOSmassspectrumwasevaluatedusinganeventmixingtech-

nique inwhich uncorrelatedpairs are formed withmuons taken fromdifferentevents.Adetaileddescription ofthetechniquecan befoundin[25].TheratiobetweencorrelatedanduncorrelatedOS dimuonsattheφ-mesonmassis∼0.65 (∼0.40)inp–Pbcollisions at√

sNN=⁵.02 TeV atforward(backward)rapidity,and∼¹.30 in pp collisions at √

s=².76 TeV. A direct comparison of the raw OS mass spectrum and the associated combinatorial background ispresentedin[35],foreachofthe pT intervalsconsideredinthe analysis.

Theinvariant massspectrainppandp–Pbcollisions,obtained after combinatorial background subtraction, are shown in Fig. 1 forthe pT-integratedsamples.Intheleft-columnpanelsofFig. 1, thesignalisdescribedinthelow-massregion(fromthethreshold

up to∼¹.5 GeV/c²) by thesuperpositionofa so-calledhadronic cocktailandtheopencharmandopenbeautyprocesses.Thepro- cesses included in the hadronic cocktail are the two-body and Dalitz decays of the light neutral mesons

η

,

ρ

,

ω

,

η

and φ, which dominate dimuon production for invariant masses below

∼^{1 GeV}/c².Theopen charmandopenbeautycontributionsarise fromcorrelatedsemi-muonicdecaysofcharmandbeautymesons andbaryons.

Thehadroniccocktailwassimulatedwithadedicatedgenerator described in [25], tuned to theexisting measurements whenever possible,otherwisebasedonthekinematicdistributionsextracted fromPYTHIA[36].Inparticular,thekinematicdistributionsofthe φ-mesonhave beentuned by meansof an iterativeprocedureto the resultspresented inthis Letterto ensure self-consistency for thisanalysis.Theopencharmandbeautygenerationisbasedona parameterisationofthespectrageneratedwithPYTHIA[33].The detectorresponseforalltheseprocessesisobtainedwithasimu- lationbasedontheGEANT3[37]transportcode.Simulatedevents arethensubjectedtothesamereconstructionandselectionproce- dureasthedata.

Whendescribingthesignalwiththesuperpositionoftheafore- mentioned contributions, four parameters are adjusted in the fit procedureineachofthep_Torrapidityintervalsconsideredinthe analysis:theyieldofthe

η

,

ω

andφ-mesons,andtheoneofthe opencharmandbeautyprocesses,withtherelativebeauty/charm contributionfixed (see later in thisparagraph). Inthis way,each parameter is linked to a process dominating in at least one re- gion ofthe consideredmass spectrum.The remaining degreesof freedomarefixedeitheraccordingtotherelativebranchingratios knownfromliterature[38],orassumingspecifichypothesesonthe crosssection ratios.In particular, theproduction crosssection of the

ρ

-mesonisassumedtobethesameasforthe

ω

assuggested frombothmodelsandppdata[25],whilethe

η

contributionwas derivedfromthe

η

crosssectionby applyingtheratioofthecor- responding cross sections

σ

_η/

σ

_η=0.3 taken from the PYTHIA tunesATLAS-CSCandD6Twhichbestdescribetheavailablelow- mass dimuon measurements at the LHC energies [25]. The open beautynormalisation isfixed to the open charm one via a fitof the p_T- and rapidity-integratedmass spectrain whichthe yields frombothprocessesarefree parameters;whenperformingdiffer- ential studies, the beauty/charm ratio is scaled according to the differentialdistributionsforthetwoprocesses,givenbytheMonte Carlo(MC)simulations.

Foreach pTandrapidityinterval,therawnumberofφ-mesons isdeterminedviaafitprocedurebasedona

χ

²minimisation,per- formedonthesignalobtainedafterthesubtractionofthecombi- natorialbackground,showninFig. 1forthepT-integratedsamples.

Several testshave been performed to evaluate the robustness of the signal extraction andestimate an appropriate systematicun- certaintyforit.Theyincludeinparticular:

– Replacingthe fitbasedonthefull MChadroniccocktail with afitbasedonthesuperpositionofvariousempiricalfunctions.

In this case, illustrated in the right-column panels of Fig. 1, thecontinuum is modelledeitherwithexponential functions orvariable-widthGaussians, whilethe

ρ

+

ω

andφpeaksare describedbyCrystalBallfunctions[39]tunedontheMC.

– Varyingtheratiobetweentheyieldsofopenbeautyandopen charmprocesses.Itwasverifiedthatforperturbationsaslarge as±50% (resultinginareasonablywiderangeofvariationfor theshapeofthetotalcontinuum)nosignificantsystematicef- fectisvisible.

– Varyingtheratiosbetweenthetwo-bodyandDalitzbranching ratiosofthe

η

and

ω

-mesons,aswellasthecrosssection ra- tios

σ

ρ/

σ

ω and

σ

_η/

σ

η,withintheuncertaintiescomingeither

fromthe availablemeasurements orfromthe differencesbe- tweenthePYTHIAtunesconsideredintheanalysisofthepp data.The branching ratio B Rω_→μμ wastaken asthe average (weightedbythecorrespondinguncertainties)oftheavailable measurementsofB Rω→μμandB Rω→^ee [38],assuminglepton universality.

– Varyingtheconsideredfitrange:inparticular,thefitwasper- formedbothincludingandexcludingthemassregionfrom0.4 to 0.65 GeV/c² wherethequalityofthecomparisonbetween thedataandthesumoftheMCsourcesturnsouttobelower.

Thetotalsystematicuncertaintyonthesignalextractionwastaken asthequadraticsumoftheabove sources.The systematicuncer- taintyonthecombinatorialbackgroundisestimatedbycomparing theshapeoftheLike-Signdimuoncontributionscomingfromthe event mixingprocedure andfrom theraw data [25]. Thisuncer- tainty dependsonthe mass,itsrelative contributionbeingmaxi- malinthemasswindow0.5–0.8 GeV/c² andminimalaroundthe φ-mesonpeak,anditisaddedinquadrature,foreachpointofthe massspectrum,to thestatisticaluncertaintyofthesignal:inthis way,thissourceofsystematicsisaccountedforbythe

χ

² minimi- sation procedure, and automatically propagated when evaluating the φ-meson rawsignal fromthefit parameters.The uncertainty associated to the sum of the MC sources (red band in the left- columnplotsofFig. 1)isevaluatedbycombiningtheuncertainties onthenormalisationofeachconsideredprocess.Fortheprocesses whose normalisationisleft free inthe fit,thisuncertaintyis the statistical oneresulting fromthe fitprocedure itself; forthe rest oftheprocesses,wealsopropagatethesystematicuncertaintyon theparameters(branchingratiosorcrosssectionratios)whichfix theirnormalisationstothoseofthefreeprocesses.

4. Results

The results of the φ-meson analysisare presented asfollows.

Wefirstpresentthemeasurementoftheproductioncrosssections, starting withits pT-dependenceinppcollisionsat√

s=².76 TeV, followed by p–Pbcollision resultsasa function of pT andrapid- ity.Then,weshowtheratioofthecrosssectionsmeasuredinthe forward andbackward regions, obtained inthe commonrapidity interval2.96<|^y|<3.53.Finally,themeasurementofthenuclear modification factor RpPb as a function of pT is presented, sepa- ratelyforthep-goingandthePb-goingdirections.

4.1. Productioncrosssectioninppandp–Pbcollisions

The cross section

σ

φ was evaluated for each pT and rapidity intervalas:

σ

φ

(

x

) =

^N

φ→rawμμ

(

x

) [

^A

· ε ](

^x

) ·

^{B R}φ→μμ

·

^Lint

,

wherexstands foranyspecific pT orrapidity intervalconsidered.

ThetotalsystematicuncertaintyonN^raw_φ→μμ(x),aftercombiningthe differentsources described above,rangesbetween3% and8%de- pendingonthecollisionsystemandkinematicrange.Thebranch- ing ratio B Rφ→μμ wastakenfrom[38] asthe average(weighted bythecorrespondinguncertainties)oftheavailablemeasurements of B R_φ→μμ and B R_φ→ee, assuming lepton universality, resulting inafinaluncertaintyofapproximately 1%.Theproductofthegeo- metricalacceptanceAandthereconstructionefficiency

ε

hasbeen evaluated by meansofMC simulations,usingthecocktail predic- tions forthedifferentialinput spectra.The valuesare obtainedas the ratio between the number of dimuons at the output of the reconstruction chain—includingtheeffectof theeventselection

Fig. 2.φ-mesoncrosssectionasafunctionofpTinppcollisionsat√

s=².76 TeV.

Errorbarsandboxesrepresentstatisticalandsystematicuncertainties,respectively.

PredictionsfromPHOJET[42]andthePYTHIAtunesATLAS-CSC[44],D6T[45], Perugia0andPerugia11[43]arealsoshownforcomparison,aswellasthe resultofafitwiththeLevy–TsallisfunctiondefinedbyEq.(1).

criteriaimposed on the data — andthe number of dimuons in- jectedasinput.

Theuncertainty on[A·

ε

] mainly originatesfromthe system- atic uncertainty on the dimuon tracking and trigger efficiencies.

The systematicuncertainty on the tracking efficiency, amounting to6% and 4% forthe backwardand forwardrapidity regions, re- spectively,comesfromtheresidualdifferencesbetweentheresults of the efficiency-determination method based on reconstructed tracks [29,40], applied to both data and MC. For the systematic uncertaintyon the trigger efficiency, we alsorefer to the proce- dure discussed in [29], resulting in an uncertainty of 3.2% and 2.8%,respectively, for the backward andforward rapidity regions consideredintheanalysis.Inordertotestpossibleadditionalsys- tematiceffects related to the hardware trigger pT cut, imposing a non-sharp threshold around 0.5 GeV/c, the analysis was re- peatedwiththeadditionalofflinesharpcutspT,μ>0.5 GeV/c and pT,μ>1 GeV/c on single muons. Foreachof thetwo alternative scenarios, the corresponding measurement of the φ-mesoncross sectionwascomparedtotheonecomingfromthereferenceanal- ysis:the difference betweenthe resultswas found to be smaller thanthequadraticdifferenceofthestatisticaluncertainties,show- ingthatnosignificant biasrelatedtothetriggerthresholdaffects theresults[41].

Thereportedvaluescorrespondtoazero-polarisation scenario forthe 2-bodydecayoftheφ-meson, intheabsenceofevidence supportinglesstrivialassumptions(inparticular,nomeasurement ofφ-mesonpolarisationiscurrentlyavailableattheLHCenergies).

4.1.1. Productioncrosssectioninppcollisions

Theinclusive, pT-differential φ-mesoncrosssection inpp col- lisionsat √

s=².76 TeV isshowninFig. 2. Thedata points,also summarised in Table 1,are compared withthe predictions from PHOJET[42] andPYTHIA [36], wherefor thelatter thePeru- gia0, Perugia11 [43], ATLAS-CSC [44], and D6T [45] tunes areconsidered.An overall goodagreementisfoundbetweenpre- dictionsanddata,withtheexception ofthePerugia0andPe- rugia11 tunes of PYTHIA which underestimate the measured cross section by a factor of two, as already observed for the φ-mesonmeasurementsat√

s=^{7 TeV[25,46].Itisworthtonote} thattheD6Ttuneisnotsuccessfulindescribingthe pT evolution oftheK/

π

ratioatmid-rapidityinppcollisionsat√

s=².76 TeV, asmeasuredbytheCMSCollaboration[47]:thissuggeststhathid- denstrangenessisbetterreproducedthanopenstrangenessinthis

Table 1

pT-differentialproductioncrosssectionfortheφ-mesoninppcollisionsat √ s= 2.76 TeV,for2.5<y<4.Thefirstuncertaintyisstatisticalandthesecondisthe bin-to-binuncorrelatedsystematic.Thebin-to-bincorrelatedrelativesystematicun- certaintyis3.9%.Theχ²/ndf valuesarerelativetothehadronic-cocktailfitandthe [⁰.8,1.2 GeV/c²]^{massregion,wherendf}=^10.

pT(GeV/c) χ²/ndf d²σφ/(dydpT)(mb/(GeV/c)) [1.0,1.5] 1.1 0.423±0.067±0.043 [¹.5,2.0] ^{1.7 0}.182±⁰.025±⁰.018 [².0,2.5] ^{1.1 0}.089±⁰.011±⁰.007 [².5,3.0] ^{1.1 0}.0340±⁰.0056±⁰.0020 [³.0,3.5] ^{0.9 0}.0139±⁰.0032±⁰.0011 [3.5,4.0] 1.1 0.0087±0.0022±0.0006 [4.0,5.0] 1.1 0.0028±0.0012±0.0002

Table 2

Systematicuncertainties(inpercent)contributingtothemea- surement ofthe φ crosssection in pp collisions at √

s= 2.76 TeV.WhentheuncertaintyvaluesdependonthepTin- terval,theirminimumandmaximumvaluesarequoted.

Source Syst. uncertainty onσφ^pp

Uncorrelated

Signal extraction 3–8%

Tracking efficiency 4%

Trigger efficiency 3%

Correlated

Lint 3.8%

B R(φ→) 1%

specificPYTHIAtune.Datapoints werefittedwitha Levy–Tsallis function[48]

1 p_T

dN dp_T

∝

1

+

^mT

−

mφ nT

₋n

,

(1)

where mT=

p²_T+^m2_φ ^{stands for the transverse mass, obtain-} ing the valuesn=¹⁰.2±⁴.8 and T =²⁸⁴±^{72 MeV for the fit} parameters, where the errors reflect the statistical uncertainties only. The cross section integrated over the accessible pT range 1<p_T<5 GeV/c is

σ

φ=⁰.566±⁰.055(stat.)±⁰.044(syst.)mb.

Thesystematicuncertaintiesforthismeasurementaresummarised inTable 2.

4.1.2. Productioncrosssectioninp–Pbcollisions

The φ cross section as a function of pT in p–Pb collisions is showninFig. 3fortheforwardandbackwardrapidityregionscon- sideredin the analysis. The results, also reportedin Table 3,are fittedwiththeLevy–TsallisdistributiondefinedinEq.(1),there- sultingfitparameters beingβ=⁹.6±¹.3 andT =³⁶⁶±^{30 MeV} fortheforwardrapidityregionandβ=¹¹.4±¹.4 andT =³⁸⁴± 24 MeV for thebackward one,wheretheerrorsreflectthestatis- ticaluncertaintiesonly.ThepredictionsfromHIJING(withgluon shadowing) [49] andDPMJET[50] are alsoshown:thesegenera- tors provided a good description ofthe ALICE dN_ch/d

η

lab results atmid-rapidity[9].Averagingovertheavailable pT range,thedis- crepancybetweenthedataandthepredictionsfromHIJINGand DPMJETamounts to ∼^{18% and} ∼^57%, respectively,atbackward rapidity (the Pb-going direction) and ∼^{5% and} ∼⁹.5%, respec- tively,atforwardrapidity (thep-going direction).Inallthecases, thegeneratorsunderestimatethedatapoints.

The φ crosssection inp–Pbcollisions, integratedoverthe ac- cessible pT range, 1<pT<7 GeV/c, is shown as a function of rapidity in Fig. 4. The data points, also summarised in Table 4, exhibit a significant asymmetry between the forward and back- ward rapidity regions.The data pointfrom theφ-mesonanalysis at mid-rapidity in the K⁺K⁻ channel [51], also shown for the

Fig. 3.φ-mesoncrosssectioninp–Pbcollisionsat√

sNN=5.02 TeV asafunctionofpTinthebackward(left)andforward(right)rapidityregions.Errorbars(smallerthan themarkers)andboxesrepresentstatisticalandsystematicuncertainties,respectively.PredictionsbyHIJING[49]andDPMJET[50]arealsoshown,togetherwiththeresult ofafitwiththeLevy–Tsallisfunction (Eq.(1)).

Table 3

Productioncrosssectionfortheφ-mesoninp–Pbcollisionsat√

sNN=5.02 TeV,asafunctionofpT,inthebackwardandforwardrapidityregions.Thefirstuncertaintyis statisticalandthesecondisthebin-to-binuncorrelatedsystematic.Thebin-to-bincorrelatedrelativesystematicuncertaintyis3.6%and3.9%,respectively,forthebackward andtheforwardregions.Theχ²/ndf valuesarerelativetothehadronic-cocktailfitandthe[⁰.8,1.2 GeV/c²]^massregion.

pT(GeV/c) −⁴.46<y<−².96 2.03<y<3.53

χ²/ndf d²σ_φ^pPb/(dydpT)(mb/(GeV/c)) χ²/ndf d²σ_φ^pPb/(dydpT)(mb/(GeV/c))

[1.0,1.5] 0.7 102±8±12 1.5 73.3±5.6±8.0

[1.5,2.0] 1.2 58.6±3.3±5.5 1.9 42.1±2.5±4.3 [2.0,2.5] 2.5 28.3±1.4±2.9 1.7 21.0±1.2±2.0 [².5,3.0] ^{4.2 15}.0±⁰.7±¹.2 3.1 10.07±⁰.77±⁰.97 [³.0,3.5] ^{2.6 7}.66±⁰.40±⁰.70 2.0 6.38±⁰.41±⁰.61 [³.5,4.0] ^{1.9 4}.20±⁰.24±⁰.34 1.2 3.96±⁰.30±⁰.36 [⁴.0,4.5] ^{0.7 2}.15±⁰.17±⁰.16 1.0 1.99±⁰.20±⁰.15 [4.5,5.0] 0.9 1.20±0.11±0.10 0.9 1.06±0.13±0.08 [5.0,6.0] 1.0 0.560±0.052±0.054 1.0 0.570±0.088±0.043 [⁶.0,7.0] ^{1.2 0}.201±⁰.030±⁰.028 0.9 0.199±⁰.045±⁰.016

Fig. 4. φ cross section inp–Pb collisions at √_s

NN=⁵.02 TeV as afunction of rapidity,integratedovertherange1<pT<7 GeV/c.Errorbarsandboxesrepre- sentstatisticalandsystematicuncertainties,respectively.Predictions byHIJING andDPMJETarealsoshown,togetherwiththemid-rapiditydatapointfromthe φ-meson measurementin the K⁺K⁻ channel [51], also evaluatedin the range 1<pT<7 GeV/c.

1<p_T<7 GeV/c p_T range, fits well into the trend defined by thetwoseriesofpoints inthebackwardandforwardrapidityre- gions. This observationcomplements the previous measurements oflight-flavourparticleproduction(chargedunidentifiedparticles) reported in p–Pb by ALICE at the LHC at mid-rapidity [9], and ind–Au byPHOBOS atRHIC ranging frommidto forwardrapid- ity[14].Thecomparisonbetweenthedataandthepredictionsby

Table 4

Productioncrosssectionfor theφ-mesoninp–Pbcollisionsat√

sNN=5.02 TeV, asafunctionofrapidity,integratedovertherange1<pT<7 GeV/c.Thefirstun- certaintyisstatisticalandthesecondisthebin-to-binuncorrelatedsystematic.The bin-to-bincorrelatedrelativesystematicuncertaintyis3.6%and3.9%,respectively, forthebackwardandtheforwardregions.Theχ²/ndf valuesarerelativetothe hadronic-cocktailfitandthe[⁰.8,1.2 GeV/c²]^massregion.

y χ²/ndf dσ_φ^pPb/dy(mb) y χ²/ndf dσ_φ^pPb/dy(mb) [−⁴.46,−⁴.25] ^{0.9 89}±¹⁰±⁹ [².03,2.35] ^{2.6 104}±¹¹±⁶ [−⁴.25,−⁴.05] ^{1.8 89}±⁶±⁷ [².35,2.55] ^{1.5 102}±⁷±⁵ [−4.05,−3.85] 0.9 103±5±8 [2.55,2.75] 2.0 96±5±6 [−3.85,−3.65] 2.9 117±6±9 [2.75,2.95] 1.6 86±4±5 [−3.65,−3.45] 1.2 128±7±9 [2.95,3.15] 2.3 68±4±4 [−³.45,−³.25] ^{3.6 133}±⁹±⁹ [³.15,3.35] ^{1.0 66}±⁵±⁵ [−³.25,−².96] ^{1.2 136}±¹⁴±¹¹ [³.35,3.53] ^{1.2 45}±⁸±⁶

HIJINGandDPMJET,illustratedinFig. 4,clearly showshowthe models—whichsuccessfullydescribedchargedparticleproduction at mid-rapidity in the same collision system [9] — fail to prop- erly reproduce the shape and the normalisation of the observed rapiditydependenceoftheφcrosssection.Still,theHIJINGpre- diction qualitativelyreproducestheforward–backwardasymmetry observed in the data,aswell as —ignoring the normalisation — the shapeofthe y-dependenceinthe backwardregion.DPMJET, onthecontrary,failstoreproduceevenqualitativelytheobserved forward–backwardasymmetry.

4.2. Forward–backwardratioinp–Pbcollisions

To establisha more direct comparison ofthe cross section in thep-goingandPb-goingdirections,

σ

_φ^pPbwasextractedasafunc-

Fig. 5.Forward–backward ratiofor the φ-meson inp–Pb collisions at √ sNN= 5.02 TeV asafunctionofpT,intherapidityrange2.96<|y|<3.53 commonto thetworapidityregionsconsideredintheanalysis.Errorbarsandboxesrepresent statisticalandsystematicuncertainties,respectively.Theblueboxontheleftrepre- sentsthebin-to-bincorrelatedsystematicuncertainty,seeTable 7.Predictionsfrom HIJINGandDPMJETarealsoshownforcomparison.(Forinterpretationoftheref- erencestocolourinthisfigurelegend,thereaderisreferredtothewebversionof thisarticle.)

tion of pT in the common |^y| ^{range 2}.96<|^y|<3.53. The pT interval1.0<pT<1.5 GeV/c was discardedinthismeasurement because of the poor statistics available in this limited rapidity range,resultinginanuncertaintylargerthan50%.

The ratio between the forward and backward cross section, R_FB, is shown asa function of p_T inFig. 5. The data points ex- hibit no significant pT dependence within the experimental un- certainties.Predictions by HIJINGand DPMJETare also shown, withHIJINGslightlyoverestimatingthedatapointsandDPMJET clearly failing to reproduce the observed values, staying above R_FB=^{1 in the whole p}T rangeconsidered here. Thisobservation isconsistent withtheobservations inFig. 4, wherethe forward–

backwardasymmetryoftheφ-mesonyieldwasbetterreproduced byHIJINGthanbyDPMJET.

4.3.Nuclearmodificationfactorinp–Pbcollisions

Theφ-mesonnuclearmodificationfactorRpPbisdefinedasthe ratiobetweentheproductioncrosssection

σ

_φ^pPb(p_T)inp–Pbcolli- sionsandthecrosssection

σ

_φ^pp(pT)inppcollisions—evaluatedat

√s=⁵.02 TeV asdescribedinthefollowing—scaledby APb:

R_pPb

(

p_T

) = σ

_φ^pPb

(

p_T

)

σ

_φ^pp

(

pT

) ·

A_Pb

,

(2) where A_Pb isthe nuclearmassnumber forthePbnucleus. Since fortheppcrosssection

σ

_φ^pp at√

s=⁵.02 TeV nodirectmeasure- mentis currentlyavailable, it was evaluated by interpolatingthe measurements in the rapidity interval 2.5<y<4 at √

s=².76 (see Section 4.1.1) and 7 TeV [25]. For each pT interval, the √

s dependenceof the differential cross section d²

σ

_φ^pp/(dydp_T) was describedwithapowerlaw

σ

^pp(√

s)=^C·(√

s)^α,whereC and

α

aredetermined usingthe dataat2.76and7 TeV. Alternativepa- rameterisationswere alsoconsidered[52],namelyalinearandan exponential function, andthe mean ofthe results obtainedwith thethreefunctionswastaken.Sincetheppmeasurementsarelim- ited to 1<p_T<5 GeV/c, the cross section at √

s_NN=⁵.02 TeV was extrapolated towards higher pT by means of a Levy–Tsallis function,which describesthe calculateddifferential cross section in the p_T range covered by the measurements. The uncertainty

Table 5

Differentialcrosssectionfortheφ-mesoninppcollisionsat√

s=⁵.02 TeV inthe backwardandforwardrapidityregionsofinterestfortheanalysisofthep–Pbdata, asobtainedinterpolatingtheavailablemeasurementsat√

s=2.76 and7 TeV.Total uncertainties,combiningstatisticalandsystematicsources,arereported.

pT(GeV/c) d²σφ/dydpT(mb/(GeV/c))

−⁴.46<y<−².96 2.03<y<3.53 [¹.0,1.5] ⁰.491±⁰.067 0.656±⁰.090 [¹.5,2.0] ⁰.223±⁰.015 0.297±⁰.020 [2.0,2.5] 0.0995±0.0071 0.1328±0.0095 [2.5,3.0] 0.0467±0.0032 0.0623±0.0043 [3.0,3.5] 0.0234±0.0015 0.0312±0.0020 [³.5,4.0] ⁰.0125±⁰.0011 0.0167±⁰.0015 [⁴.0,4.5] ⁰.00706±⁰.00094 0.0094±⁰.0012 [⁴.5,5.0] ⁰.00419±⁰.00082 0.0056±⁰.0011 [⁵.0,6.0] ⁰.00213±⁰.00060 0.00284±⁰.00081 [6.0,7.0] 0.00093±0.00039 0.00124±0.00052 on the interpolated cross sections arises fromthe choice of the function usedfor theinterpolation, fromtheuncertainties in the measurementsat2.76and7 TeV,and—forpT>5 GeV/c —from the extrapolation based on the Levy–Tsallis fit. They range from about7% for p_T=^{1 GeV}/c to20% for p_T=^{5 GeV}/c, andexceed 30% for pT>5 GeV/c, representing the major source of system- aticuncertainty forthemeasurement ofthenuclear modification factor.The interpolatedcrosssection,whichrefers totherapidity range2.5<y<4,wasfinallyscaledtotheforwardandbackward rapiditywindows2.03<y<3.53 and−⁴.46<y<−².96,consid- eredfortheanalysisofthep–Pbdata.Therelativescalingfactors ffwd=¹.135±⁰.031 and fbkw=⁰.850±⁰.028 wereevaluatedas an average from simulations with PHOJET and the Perugia0, Perugia11, ATLAS-CSCand D6T PYTHIAtunes. In doing so, we also retainedthe PYTHIAtunes which were observed tofail in describing the pp data (see Section 4.1.1): the reason is that thedisagreement betweenmodels anddata concernsinthiscase theabsolutenormalisationmorethan theshape ofthekinematic distributions, which is the only relevant feature in the evalua- tion of the ffwd and fbkw factors. The uncertainties (amounting toabout 3%)correspondtothedifferencesbetweentheconsidered MC predictions.ThenumericalvaluesarereportedinTable 5.

The nuclear modification factor R_pPb as a function of pT is shownin thetwo panelsofFig. 6 forthe backward andforward rapidity regions considered inthe analysis. The numerical values are also quoted in Table 6. For each p_T interval, the systematic uncertaintydetailedinTable 7 resultsfromthe quadraticsumof theuncertaintyontheφcrosssectioninp–Pbandtheoneofthe pp reference.ArisingtrendofRpPbwhengoingfrompT=^{1 GeV}/c to p_T≈^{3–4 GeV}/c canbeobservedbothatbackwardandforward rapidity.The values of R_pPb inthe two rapidity ranges, however, are significantly different. In particular, at backward rapidity we observeanenhancementoftheφcrosssectionwithrespecttothe scaledppreferencepeakedaround pT=^{3–4 GeV}/c.Thisenhance- ment,absentintheforwardrapidityregion,reachesafactorofup to ∼¹.6 and could be associated eitherto an initial-state effect (includingapossibleCronin-likeenhancement[4,53])ortoafinal state effect relatedto radial flowin p–Pbas proposed forrecent ALICEmeasurementsatmid-rapidity[12].Discriminatingbetween thesetwo effects requires moredetailed investigations, including differential analyses as a function of globalevent propertieslike collisioncentrality.

Concerning the behaviour at high pT, we observe that the φ-meson R_pPb is compatiblewith unity for p_T^{4 GeV}/c in the p-going direction, similar to what was observed for the R_pPb of chargedparticle production atmid-rapidity [10,12].The observa- tions in thePb-going directiondo not allowa clear trendofthe R_pPb factorat high p_T to be established.A possiblesaturation at