Measurement of the cross section for isolated-photon plus jet production in pp collisions at √s = 13 TeV using the ATLAS detector

Article

Reference

Measurement of the cross section for isolated-photon plus jet production in pp collisions at √s = 13 TeV using the ATLAS detector

ATLAS Collaboration AKILLI, Ece (Collab.), et al.

Abstract

The dynamics of isolated-photon production in association with a jet in proton–proton collisions at a centre-of-mass energy of 13 TeV are studied with the ATLAS detector at the LHC using a dataset with an integrated luminosity of 3.2 fb −1 . Photons are required to have transverse energies above 125 GeV. Jets are identified using the anti- kt algorithm with radius parameter R=0.4 and required to have transverse momenta above 100 GeV. Measurements of isolated-photon plus jet cross sections are presented as functions of the leading-photon transverse energy, the leading-jet transverse momentum, the azimuthal angular separation between the photon and the jet, the photon–jet invariant mass and the scattering angle in the photon–jet centre-of-mass system. Tree-level plus parton-shower predictions from Sherpa and Pythia as well as next-to-leading-order QCD predictions from Jetphox and Sherpa are compared to the measurements.

ATLAS Collaboration, AKILLI, Ece (Collab.), et al . Measurement of the cross section for isolated-photon plus jet production in pp collisions at √s = 13 TeV using the ATLAS detector.

Physics Letters. B , 2018

DOI : 10.1016/j.physletb.2018.03.035

Available at:

http://archive-ouverte.unige.ch/unige:103326

Disclaimer: layout of this document may differ from the published version.

1 / 1

Contents lists available atScienceDirect

Physics Letters B

www.elsevier.com/locate/physletb

Measurement of the cross section for isolated-photon plus jet production in pp collisions at √

s = ^{13 TeV using the ATLAS detector}

.The ATLAS Collaboration

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

Articlehistory:

Received30December2017

Receivedinrevisedform27February2018 Accepted13March2018

Availableonline15March2018 Editor: W.-D.Schlatter

The dynamicsofisolated-photonproductioninassociation withajetinproton–proton collisions ata centre-of-massenergyof13 TeVarestudiedwiththeATLASdetectorattheLHCusingadatasetwithan integratedluminosityof3.2 fb⁻¹.Photonsarerequiredtohavetransverseenergiesabove125 GeV.Jets areidentifiedusingtheanti-kt algorithmwithradiusparameterR=⁰.4 andrequiredtohavetransverse momenta above 100 GeV. Measurements ofisolated-photon plus jet cross sections are presented as functionsoftheleading-photontransverseenergy,theleading-jettransversemomentum,theazimuthal angularseparation betweenthe photonand the jet, thephoton–jet invariant massand the scattering angleinthephoton–jet centre-of-masssystem.Tree-levelplusparton-showerpredictionsfromSherpa andPythiaaswellasnext-to-leading-orderQCDpredictionsfromJetphoxandSherpaarecomparedto themeasurements.

©^{2018TheAuthor.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense} (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP³.

1. Introduction

Theproductionofprompt photonsinassociation withatleast onejetinproton–proton(pp) collisionsprovidesatestingground forperturbativeQCD(pQCD).In ppcollisions,allphotonsthatare notsecondariesfromhadrondecaysareconsideredtobe“prompt”.

Themeasurementsofangularcorrelationsbetweenthephotonand thejet canbe usedto probethedynamicsofthe hard-scattering process. The dominant source in pp collisions at the LHC is the qg→^q

γ

process. Thesemeasurements are alsousefulfortuning MonteCarlo(MC)modelsandtestingt-channelquarkexchange [1, 2].Furthermore,precisemeasurementsoftheseprocessesvalidate thegeneratorsusedforbackgroundstudiesinsearchesforphysics beyond the Standard Model which involve photons, such as the search for new phenomena in final states with a photon and a jet [3,4].

The productionof pp→

γ

₊jet+^{X events proceedsvia two} processes: direct, in which the photon originates from the hard process, andfragmentation, in which thephoton arisesfromthe fragmentationofacolouredhightransversemomentum¹ (pT)par-

E-mailaddress:atlas.publications@cern.ch.

1 ATLASusesaright-handed coordinatesystemwith itsoriginat thenominal interactionpoint(IP)inthecentreofthedetectorandthez-axisalongthebeam pipe.Thex-axispointsfromtheIPtothecentreoftheLHCring,andthe y-axis pointsupwards.Cylindricalcoordinates(r,φ)areusedinthetransverseplane,φ beingtheazimuthalanglearoundthez-axis.Thepseudorapidityisdefinedinterms ofthepolarangleθasη= −ln tan(θ/2).Theangulardistanceismeasuredinunits ofR≡

(η)²+(φ)².Therapidityisdefinedasy=0.5ln[(E+pz)/(E−pz)],

ton [5,6].Thedirectandfragmentationcontributionsareonlywell defined at leading order (LO) in QCD; at higher orders this dis- tinctionisnolongerpossible.Thesetwoprocessesexhibitdistinct behaviours in the observables considered here. Precise measure- mentstesttheinterplayofdirectandfragmentationprocesses.

Measurements of prompt-photon production in a final state with accompanying hadrons necessitate an isolation requirement onthephotontoavoidthelargecontributionfromneutral-hadron decaysintophotons.Theproductionofisolatedphotonsinassoci- ation withjetsin ppcollisions at√

s=^{7 and8 TeV was studied} by theATLAS [1,2,7] andCMS [8–10] Collaborations.Theincrease inthecentre-of-massenergyofppcollisionsattheLHCto13 TeV allows theexploration ofthedynamicsofphoton+^{jet production} inanewregime withthegoaloftestingthepQCDpredictionsat higherenergytransfers thanachievedbefore.Itisalsopossibleto investigate whetherthe data in the newenergy regime are well described by the predictions of parton-shower event generators, suchasSherpa[11] andPythia[12].

The dynamics of the underlying processes in 2→^{2 hard} collinear scattering can be investigated using the variable θ^∗, wherecosθ^∗≡^tanh(y/2)andy isthedifference betweenthe rapidities of the two final-state particles. The variable θ^∗ coin- cides withthe scatteringpolarangle inthecentre-of-massframe forcollinearscatteringofmasslessparticles,anditsdistributionis sensitivetothespinoftheexchangedparticle.Forprocessesdom-

whereEistheenergyandpzisthez-componentofthemomentum,andtransverse energyisdefinedasET=Esinθ.

https://doi.org/10.1016/j.physletb.2018.03.035

0370-2693/©^{2018TheAuthor.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense}(http://creativecommons.org/licenses/by/4.0/).Fundedby SCOAP³.

inatedbyt-channelgluonexchange,suchasdijetproductioninpp collisions,thedifferentialcrosssectionbehavesas(1− |^cosθ^∗|)⁻² when|^cosθ^∗|→^{1.Incontrast,processesdominatedby t-channel} quark exchange are expected to exhibit a (1− |^cosθ^∗|)⁻¹ be- haviour when |^cosθ^∗|→^{1. This} fundamental prediction of QCD can be tested in photon plus jet production in pp collisions.

Thedirect-photoncontribution,dominatedbyt-channelquarkex- change,isexpectedtoexhibita(1− |^cosθ^∗|)⁻¹dependencewhen

|^cosθ^∗|→^{1,whereasthatof}fragmentationprocessesispredicted tobethesameasindijetproduction,namely(1− |^cosθ^∗|)⁻².For bothprocesses, there arealso s-channel contributionswhich are, however,non-singularwhen|^cosθ^∗|→^{1.Asa result,ameasure-} mentof the crosssection forprompt-photon plus jet production asa functionof|^cosθ^∗|^{providesa handleontherelativecontri-} butionsofthedirectandfragmentationcomponentsaswellasthe possibilityoftestingthedominanceoft-channelquarkexchange.

The results presented here include a study of the kinematics ofthephotonplus one-jet systemviameasurements ofthecross sectionsasfunctionsoftheleading-photontransverseenergy(Eγ

T) andtheleading-jet transverse momentum(p^jet-lead_T ). The dynam- icsof the photon plus one-jet systemare studied by measuring theazimuthalangularseparationbetweentheleadingphotonand theleading jet (φ^γ−jet),the invariant massofthe leading pho- ton and the leading jet (mγ−^jet) and cosθ^∗. The distribution in mγ−^jet ispredictedtobemonotonicallydecreasinginQCD dueto theabsenceofresonancesthatdecayintoaphotonandajet.The analysisisperformedusing3.2 fb⁻¹ of√

s=^{13 TeV pp collision} datarecordedby ATLAS.Next-to-leading-order(NLO)QCD predic- tions from Jetphox [13,14] and Sherpa as well asthe tree-level predictions ofPythia and Sherpaare compared to the measure- ments.

2. ATLASdetector

The ATLAS detector [15] is a multi-purpose detector with a forward–backward symmetric cylindrical geometry. It consists of an inner trackingdetectorsurrounded by a thinsuperconducting solenoid, electromagneticandhadroniccalorimeters,anda muon spectrometer incorporating three large superconducting toroid magnets. The inner-detector system is immersed in a 2 T axial magneticfield andprovidescharged-particletrackingintherange

|

η

_|<2.5. The high-granularitysilicon pixel detectoris closest to theinteractionregionandprovidesfourmeasurements pertrack;

theinnermostlayer,knownastheinsertableB-layer [16],provides high-resolutionhitsatsmallradiustoimprovethetrackingperfor- mance. The pixel detector is followed by the silicon microstrip tracker, which typically provides four three-dimensional space pointmeasurementspertrack.Thesesilicondetectorsarecomple- mentedby thetransitionradiationtracker,whichenablesradially extended track reconstruction up to |

η

_|=2.0. The calorimeter system covers the range |

η

|<4.9. Within the region |

η

|<3.2, electromagnetic(EM) calorimetry is providedby barreland end- caphigh-granularitylead/liquid-argon(LAr)EMcalorimeters,with anadditionalthinLArpresamplercovering|

η

|<1.8 tocorrectfor energylossinmaterialupstreamofthecalorimeters;for|

η

_|<2.5 theEMcalorimeterisdividedintothreelayers indepth.Hadronic calorimetryisprovidedbyasteel/scintillator-tilecalorimeter,seg- mented into three barrel structures within |

η

_|<1.7, and two copper/LAr hadronic endcap calorimeters, which cover the re- gion 1.5<|

η

|<3.2. The solid-angle coverage is completed out to|

η

_|=4.9 withforwardcopper/LArandtungsten/LArcalorimeter modules,whichareoptimisedforEMandhadronicmeasurements, respectively. Events are selected using a first-level trigger imple- mentedincustom electronics,whichreducesthemaximumevent rateof 40 MHz to a design value of 100 kHz using a subset of

detector information.Software algorithms withaccess tothe full detector information are then used in the high-level trigger to yieldarecordedeventrateofabout1 kHz [17].

3. DatasampleandMonteCarlosimulations

The data used inthis analysiswere collected withthe ATLAS detectorduringthe ppcollisionrunningperiodof2015,whenthe LHC operated with a bunch spacing of 25 ns and at a centre- of-massenergyof 13 TeV.Onlyevents takenduring stablebeam conditions andsatisfying detectorand data-quality requirements, which include the calorimeters and inner tracking detectors be- ing innominaloperation, are considered.The average numberof ppinteractions per bunchcrossing inthedatasetis 13.Thetotal integratedluminosity ofthecollected sampleis 3.16±⁰.07 fb⁻¹. Theuncertaintyintheintegratedluminosityis2.1% andisderived, followingamethodologysimilartothatdetailedinRef. [18],from a calibration of the luminosity scale using x–y beam-separation scansperformedinAugust2015.

Samples of MC events were generated to study the charac- teristics of signal events. The MC programs Pythia 8.186 and Sherpa2.1.1 were usedtogenerate thesimulatedevents.Inboth generators,thepartonicprocesseswere simulatedusingtree-level matrixelements,withtheinclusionofinitial- andfinal-statepar- tonshowers.Fragmentationintohadronswasperformedusingthe Lund stringmodel [19] inthecaseofPythia,andinSherpabya modifiedversion oftheclustermodel [20].TheLONNPDF2.3 [21]

partondistributionfunctions(PDF) setwasused forPythia while the NLO CT10 [22] PDF set was usedfor Sherpatoparameterise the proton structure. Both samples include a simulation of the underlying event (UE). The event-generator parameters were set accordingtotheATLAS2014tuneseries(A14tune)forPythia[23]

andtothetunedevelopedinconjunctionwiththeNLO CT10PDF set for Sherpa. The Pythia simulation of the signal includes LO photon-plus-jetevents fromboth directprocesses (the hard sub- processes qg→^q

γ

and q^q¯ →^g

γ

, called the “hard” component) and photon bremsstrahlung in LO QCD dijet events (called the

“bremsstrahlung” component). The bremsstrahlung component is modelledby final-stateQEDradiationarisingfromcalculationsof all 2→^{2 QCDprocesses. Inthe}particle-level phasespaceof the presentedmeasurements thefractionofthebremsstrahlungcom- ponent decreases from 35% at Eγ

T =^{125 GeV to 15% at Eγ}_T = 1 TeV. The Sherpa samples were generated with LO matrix ele- ments forphoton-plus-jetfinalstateswithup tothree additional partons (2→^{n processes with n from 2 to 5); the matrix el-} ements were merged with the Sherpa parton shower using the ME+PS@LO prescription [24]. The bremsstrahlung component is accounted for in Sherpa through the matrix elements of 2→ⁿ processeswithn≥^{3.In thegeneration oftheSherpasamples,a} requirement on the photon isolation at the matrix-element level wasimposedusingthecriteriondefinedinRef. [25].Thiscriterion, commonlycalledFrixione’scriterion, requiresthetotal transverse energy inside a cone of size V ^{around the generated final-state} photon,excluding thephotonitself,tobe belowa certainthresh- old, E^max_T (V)=

Eγ

T((1−^cosV)/(¹−^cosR))^n,forall V<R^.The parameters forthe threshold were chosen to be R=⁰.3, n=² and =⁰.025.TheSherpapredictionsfromthiscomputationare referredtoasLOSherpa.

All the samplesof generatedevents were passed through the Geant4-based [26] ATLASdetectorandtriggerfullsimulationpro- grams [27]. They are reconstructed and analysed withthe same program chain as the data. Pile-up from additional pp collisions inthe sameand neighbouringbunch crossingswas simulatedby overlayingeachMCeventwithavariablenumberofsimulatedin- elastic ppcollisions generatedusingPythia8.153withtheATLAS

setoftuned parametersforminimumbias events(A2tune) [28].

The MCevents areweighted to reproducethe distributionof the averagenumberofinteractionsperbunchcrossing(

μ

)observed inthedata,referredtoas“pile-upreweighting”.Inthisprocedure, the

μ

valuefromthedataisdividedby afactorof1.16±⁰.07, a rescalingwhichmakesthenumberofreconstructedprimaryver- ticesagreebetterbetweendataandsimulationandreproducesthe visiblecross sectionof inelastic pp collisions asmeasured inthe data [29].

4. Eventselection

Events were recorded using a single-photon trigger, with a transverse energy threshold of 120 GeV and “loose” identifica- tionrequirementsbasedontheshowershapesinthesecondlayer ofthe EM calorimeteras well as onthe energy leaking intothe hadroniccalorimeterfromtheEMcalorimeter [17].Eventsare re- quiredtohaveareconstructedprimaryvertex.Ifmultipleprimary verticesarereconstructed,theonewiththehighestsumofthep²_T oftheassociatedtracksisselectedastheprimaryvertex.

Photon candidates are reconstructed from clusters of energy deposited in the EM calorimeter and classified [30] as uncon- vertedphotons(candidateswithoutamatchingtrackormatching reconstructed conversion vertex in the inner detector) or con- vertedphotons(candidateswithamatchingreconstructedconver- sion vertexor a matchingtrack consistent withoriginating from a photon conversion). The measurement of the photon energyis based on the energy collected in calorimeter cells in an area of size

η

×φ=⁰.075×⁰.175 in the barrel and 0.125×⁰.125 in the endcaps. A dedicated energy calibration [31] is then ap- plied to thecandidates to account forupstream energy loss and bothlateral andlongitudinal leakage.The photonidentificationis basedprimarilyonshowershapesinthecalorimeter [30].Anini- tialselection is derived using the informationfrom the hadronic calorimeter andthe lateral shower shape in the second layer of theEMcalorimeter,wheremostofthephotonenergyiscontained.

Thefinaltightselectionappliesstringentcriteria [30] tothesame variablesusedintheinitialselection,separatelyforconvertedand unconvertedphotoncandidates.Italsoplacesrequirementsonthe showershapeinthefinelysegmentedfirstcalorimeterlayertoen- surethe compatibility of the measured shower profile with that originatingfromasinglephotonimpacting thecalorimeter.When applyingthephotonidentificationcriteriatosimulatedevents,cor- rections are made forsmall differences in the average values of the shower-shape variables between dataand simulation. Events withatleastonephotoncandidatewithcalibrated Eγ

T >125 GeV, wherethe trigger is maximally efficient,and |

η

^γ_|<2.37 are se- lected.Candidatesintheregion1.37<|

η

^γ_|<1.56,whichincludes thetransitionregionbetweenthebarrelandendcapcalorimeters, are not considered. The photon candidate is required to be iso- latedbasedon theamount oftransverse energyinsidea cone of sizeR=⁰.4 inthe

η

–φplanearoundthephotoncandidate,ex- cludingan area ofsize

η

×φ=⁰.125×⁰.175 centred onthe photon. The isolation transverse energy is computed fromtopo- logical clusters of calorimeter cells [32] and is denoted by E^iso_T . Topologicalclustersare built fromneighbouring calorimetercells containingenergysignificantlyabovea noisethresholdthatises- timated from measurements of calorimeter electronic noise and simulatedpile-up noise. The measured value of E^iso_T iscorrected forthe expected leakage of thephoton energyinto the isolation coneas well asfortheestimatedcontributions fromthe UE and pile-up [33,34]. The corrections forpile-up andthe UE are com- putedsimultaneouslyonanevent-by-eventbasisusingthejet-area method [35,36] asfollows:thek_⊥jetalgorithm [37,38] withjetra- diusR=⁰.5 isusedtoreconstructalljets,takingtopologicalclus-

tersofcalorimetercellsasinput;noexplicittransversemomentum threshold is applied. The ambient transverse energy density for the event (

ρ

), from pile-up and the UE, is computed using the medianof thedistributionof theratioofthe jet’s transverseen- ergytoitsarea.Finally,

ρ

ismultipliedbytheareaoftheisolation conetocomputethecorrectiontoE^iso_T .Thecombinedcorrectionis typically2 GeV anddependsweaklyon Eγ

T.Inaddition,forsimu- latedevents,data-drivencorrectionsto E^iso_T are appliedsuch that the peak position in the E^iso_T distribution coincides in data and simulation. Afterall these corrections, E^iso_T is requiredto be less than E^iso_T,cut≡⁴.2·¹⁰⁻³·^EγT +⁴.8GeV [39].Theisolationrequire- mentsignificantlyreducesthemainbackground,whichconsistsof multi-jet events whereone jet typically contains a

π

⁰ or

η

me- son that carries most of the jet energy and is misidentified as a photon because it decays intoan almost collinearphoton pair.

A smallfractionoftheeventscontainmorethanonephotoncandi- datesatisfyingtheselectioncriteria.Insuchevents,thehighest-Eγ T (leading)photonisconsideredforfurtherstudy.

Jetsare reconstructedusingtheanti-k_t algorithm [40,41] with a radius parameter R=⁰.4, using topological clusters as input.

The calorimeter cell energies are measured atthe EM scale,cor- responding to the energy deposited by electromagnetically inter- actingparticles.Thejetfour-momentaarecomputedfromthesum ofthejet-constituentfour-momenta,treatingeachasafour-vector with zero mass. The jets are then further calibrated using the method described in Ref. [42] and these jets are referred to as detector-leveljets.The four-momentumofeach jetisrecalculated topointtotheselectedprimaryvertexoftheeventratherthanthe centreofthedetector.ThecontributionfromtheUEandpile-upis then subtracted on a jet-by-jet basis using the jet-area method.

A jet-energy calibrationisderived fromMC simulations asacor- rectionrelatingthecalorimeterresponsetothetruejetenergy.To determine thesecorrections, thejet reconstruction procedure ap- plied to the topological clusters is alsoapplied to the generated stable particles, which are defined as those witha decaylength of c

τ

>10 mm, excluding muons and neutrinos; these jets are referred toasparticle-leveljets.Inaddition,sequentialjet correc- tions, derived from MC simulated events andusing global prop- erties ofthe jet such astracking information,calorimeterenergy depositsandmuonspectrometerinformation,areapplied [43].Fi- nally, thedetector-leveljetsarefurthercalibratedwithadditional correction factors derived in situfrom a combinationof

γ

₊jet,

Z+^{jet andmulti-jetp}T balancemethods [44,45].²

Jetsreconstructedfromcalorimetersignalsnotoriginatingfrom a ppcollision arerejectedby applyingjet-qualitycriteria [45,46].

These criteriasuppressspurious jetsfrom electronic noisein the calorimeter, cosmic rays andbeam-related backgrounds. Remain- ingjetsarerequiredtohavecalibratedtransversemomentagreater than 60 GeV and rapidity |^yjet|<2.37. Jetsoverlapping withthe candidate photon are not considered if the jet axis lies within a cone of size R=⁰.8 around the photon candidate; this re- quirementpreventsanyoverlapbetweenthephotonisolationcone (R=⁰.4) andthe jet cone (R=⁰.4). Finally, theeventis re- tained if thejet withhighesttransverse energy (leadingjet) has p^jet_T >100 GeV.

Thetotalnumberofdataeventsselectedbyusingtherequire- ments discussed aboveis 895726.Thissample ofeventsis used to measure the cross section as a function of Eγ

T, p^jet-lead_T and φ^γ−jet.Forthemeasurements ofthecrosssectionsasfunctions

2 The effectofthecorrelation betweenthe eventsusedinthe insitu γ+jet analysisandtheeventsselectedherehasanegligibleeffectontheexperimental uncertaintiesassociatedtothemeasurements.

Fig. 1.E^iso_T distributionfortight(blackdots)andnon-tight(dashedhistogram,normalisedaccordingtothefit,seetext)photoncandidatesindatawith|η^γ|<0.6 indifferent E^γ_T regions.TheMCsimulationofthesignalusingPythiaisalsoshown(dottedhistogram).ThesolidhistogramisthesumofthecontributionsoftheMCsimulationofthe signalusingPythiaandthatofthenon-tightphotoncandidatesnormalisedaccordingtothefit.

ofmγ−^jetand|^cosθ^∗|^{,theadditional}constraints|

η

^γ₊y^jet-lead|<

2.37, |^cosθ^∗|<0.83 and mγ−^jet>450 GeV are imposed to re- movethebias [1,2] duetotherapidityandtransverse-momentum requirementsonthephotonandthejet³;thenumberofeventsse- lectedinthedataaftertheseadditionalrequirementsis137738.

5. Backgroundestimationandsignalextraction

Afterthe requirements on photon identification andisolation areappliedto thesampleofevents,thereisstill aresidual back- groundcontribution,whicharisesprimarilyfromjetsidentifiedas photons in multi-jet events. This background contribution is es- timated andsubtracted bin-by-bin using a data-driven technique whichmakes useofthe sametwo-dimensional sidebandmethod employedinapreviousanalysis [47].Inthisapproach,thephoton isclassifiedas:

•“Isolated”,ifE^iso_T <E^iso_T,cut.

•“Non-isolated”,ifE^iso_T >E^iso_T,cut+^{2GeV and Eiso}_T <50 GeV.The non-isolated region is separated by 2 GeV from the isolated regiontoreducethesignalcontamination.Theupperboundis appliedtoavoidhighlynon-isolatedphotons.⁴

•^{“Tight”,ifitsatisfiesthetightphoton}identificationcriteria.

•“Non-tight”, ifit failsat leastone of fourtight requirements ontheshower-shapevariablescomputedfromtheenergyde- positsinthefirstlayeroftheEMcalorimeter,butsatisfiesthe tightrequirementonthetotallateralshowerwidthinthefirst layerandalltheothertightidentificationcriteriainotherlay- ers [30].

Thedistributions in E^iso_T fortightandnon-tight photoncandi- dateswith|

η

^γ_|<0.6 inthedataareshownseparatelyinFig.1for tworegions in Eγ

T.TheMC simulationoftheprompt-photon sig- nalusingPythiaisalsoshown.Afitofthesumofthedistributions ofthePythiasignal photonsandthenon-tight photoncandidates

3 Thefirsttwoconstraintsavoidthebiasinduced byrequirementsonη^γ and y^jet-lead,yieldingslicesofcosθ^∗withthesamelengthalongtheη^γ+^{yjet-leadaxis.}

ThethirdconstraintavoidsthebiasduetotheE^γ_T>125 GeV requirement.

4 Inthisway,thedeterminationofthesignalyielddoesnotdependonthede- scriptionbytheMCgeneratorsofthedistributionofE^iso_T forpromptphotonswith highvaluesofE^iso_T .

tothedistributionofthetightphotoncandidatesisalsoincluded.

A clearsignalofpromptphotonscentredat E^iso_T aboutzeroisob- served.

Fortheestimationofthebackgroundcontaminationinthesig- nalregionatwo-dimensionalplaneisformedby E_T^isoandabinary variable(“tight”vs. “non-tight”photoncandidate)sincethesetwo variables are expected to be largely uncorrelatedfor background events.Inthisplane,fourregionsare defined:the“signal”region (A), containing tight and isolated photon candidates; the “non- isolated”backgroundcontrolregion(B),containingtightandnon- isolatedphotoncandidates;the“non-tight”backgroundcontrolre- gion(C),containingisolatedandnon-tightphotoncandidates;the background control region containing non-isolated and non-tight photoncandidates(D).

Thesignal yieldN^sig_A inregion A isestimatedby usingthere- lation

N^sig_A

=

^NA

−

^Rbg

· (

N_B

−

^fBN^sig_A

) · (

NC

−

^fCN^sig_A

)

(

ND

−

fDN^sig_A

) ,

(1) whereNK,with K=^A,B,C,D,isthenumberofeventsinregion K andR^bg=^Nbg_A ·^Nbg_D/(N^bg_B ·^N_C^bg)istheso-calledbackgroundcor- relationandistakenasR^bg=^{1 forthenominalresults; Nbg}K with K =^A,B,C,D is the unknown number of background events in each region. Equation (1) takes into account the expected num- ber of signal events in each of the three backgroundcontrol re- gions (N^sig_K ) via the signal leakage fractions, fK=^Nsig_K /N^sig_A with K=^B,C,D,whichareestimatedusingtheMCsimulationsofthe signal.

The only assumption underlying Eq. (1) is that the isolation andidentificationvariablesareuncorrelatedforbackgroundevents, thus R^bg=^{1.Thisassumption isverified indata typicallywithin}

±^{10% invalidationregions,5 whicharedominatedby}background.

Deviations of R^bg from unity in the validation regions, after ac- countingforsignal leakage usingeither thePythia orLOSherpa simulations,are propagated throughEq. (1) andtakenassystem- atic uncertainties.The signal purity,definedas N^sig_A /NA,isabove 90% in all bins of the measured distributions. The use ofPythia

5 ThevalidationregionsaredefinedwithinthecontrolregionsBandDbysplit- tingeachofthemintotwosubregions.

Table 1

Summaryoftherequirementsatparticlelevelthatdefinethefiducialphase-space regionofthemeasurements.

Requirements on photons

E^γ_T>125 GeV,|η^γ|<2.37 (excluding 1.37<|η^γ|<1.56) E^iso_T <4.2·10⁻³·E^γ_T+10 GeV

Requirements on jets

anti-ktalgorithm withR=⁰.4

the leading jet within|y^jet|<2.37 andR^γ−jet>0.8 is selected p^jet-lead_T >100 GeV

UE subtraction usingk⊥algorithm withR=0.5 (cf. Section4) Additional requirements for dσ/dm^γ−jetand dσ/d|cosθ^∗|

|η^γ+^yjet-lead|<2.37,|^cosθ^∗|<0.83 andm^γ−jet>450 GeV

orLOSherpatoextract thesignal leakage fractionsleadto simi- larsignalpurities;thedifferenceinthesignalpurityistakenasa systematicuncertainty.

Thebackgroundfromelectronsmisidentifiedasphotons,mainly producedinDrell–YanZ^(∗)/

γ

^∗_→e⁺e⁻andW^(∗)→^e

ν

processes, isalsostudied.Suchmisidentifiedelectronsarelargelysuppressed by the photon selection. This backgroundis estimated using MC samplesoffullysimulatedeventsandfoundtobenegligibleinthe phase-spaceregionoftheanalysispresentedhere.Theseprocesses weresimulatedusingtheSherpa2.2.1generator.Matrix-elements werecalculatedforuptotwoadditionalpartonsatNLOandupto fourpartonsatLO [48,49].

6. Fiducialphasespaceandunfolding 6.1. Fiducialphasespace

The crosssections are unfolded toa phase-space region close to theapplied eventselection. The fiducialphase-space regionis defined atthe particle level. A summary of the requirements at particle level that define the fiducial phase-space region of the measurementsisgiveninTable1.Thecrosssectionsasfunctions of|^cosθ^∗| ^{andmγ−jet aremeasured inafiducial}phase-spacere- gion withadditional requirements,asdetailed inthe last row of Table 1. The particle-level isolation requirement on the photon is builtby summing the transverse energyof all stable particles (seeSection4), exceptformuonsandneutrinos,inaconeofsize R=⁰.4 aroundthephotondirectionafterthecontributionfrom theUE issubtracted.The sameunderlying-event subtractionpro- cedure used atthe reconstruction level isapplied at theparticle level.The particle-levelrequirement on E^iso_T isoptimised to best matchtheacceptanceatreconstructionlevelusingthePythiaand LO Sherpa MC samples by comparing the calorimeter isolation transverse energy withthe particle-level isolation transverse en- ergyonanevent-by-eventbasis.Theparticle-levelrequirementon E^iso_T thus optimised is E^iso_T (particle)< 4.2·¹⁰⁻³·^Eγ_T +^{10 GeV;}

thesamerequirementisobtainedwhetherPythiaorLOSherpais used.Particle-level jetsarereconstructed usingtheanti-kt jet al- gorithm withradiusparameter R=⁰.4 and are builtfromstable particles,excludingmuonsandneutrinos.Atparticlelevel,thepar- ticles associatedwiththe overlaid ppcollisions (pile-up)are not considered.

6.2. Unfolding

The distributions ofthe background-subtracted signal yield as functionsofEγ

T, p^jet-lead_T ,φ^γ−jet,mγ−^jetand|^cosθ^∗|^areusedto measurethecorrespondingdifferentialcrosssectionsforisolated- photon plusjet production.The distributions are unfolded tothe

particle level using MC samples of events via a bin-by-bin tech- nique which corrects for resolution effects andthe efficiency of thephotonandjetreconstructionthroughtheformula

d

σ

dO

(

i

) =

^N

sig

A

(

i

)

C^MC

(

i

)

L

O

(

i

) ,

where d

σ

/dO(i) is the crosssection as a function ofobservable O in bin i, N^sig_A(i) isthe number of background-subtracted data events inbin i, C^MC(i) is thecorrection factor inbin i, L^{is the} integrated luminosity andO(i) is the width of bin i. The cor- rection factors are computed from the MC samples as C^MC(i)= N^MC_part(i)/N^MC_reco(i),where N^MC_part(i)is thenumberofeventsthat sat- isfy the kinematic constraints of the phase-space region at the particle level, and N_reco^MC(i) is the number of events which meet alltheselectioncriteriaatthereconstructionlevel.

The distributions of the signal yields as functions of Eγ T, p^jet-lead_T , φ^γ−jet, mγ−^jet and |^cosθ^∗| ^{in data after background} subtraction are well described by the LO Sherpa MC simula- tions, butsome differences are observed when compared to the Pythia MC simulations, in particular in the tail of the p^jet-lead_T distribution.Abetterdescriptionofthedatadistributionsasfunc- tions of Eγ

T, p^jet-lead_T , φ^γ−jet, mγ−^jet and |^cosθ^∗| ^{by Pythia is} obtained by increasing/decreasing the relative contribution from direct processes with respect to bremsstrahlung processes [1,2];

theresultingPythiasimulationsarereferredtoasPythia-adjusted simulations.

Theunfoldedcrosssectionsaremeasuredusingthesignalleak- agefractionsfromPythia-adjustedsimulationssincetheseinclude an unbiased sample ofnon-isolated photons,⁶ andthe correction factors, C^MC, fromLOSherpasince thesesimulations give some- whatbetter agreementwiththedatadistributionsasfunctionsof Eγ

T, p^jet-lead_T , φ^γ−jet,mγ−^jet and|^cosθ^∗|^{.The correction factors} varybetween1.08 and1.21 dependingontheobservable.There- sults of the bin-by-bin unfolding procedure are checked withan iterativeBayesianunfoldingmethod [50] basedonLOSherpasim- ulations,givingconsistentresults.

7. Uncertaintiesinthecross-sectionmeasurements

Photonenergyscaleandresolution.Adetailedassessmentofthe uncertainties in the photon energy scale and resolution is made using the methodreported inRef. [47]. The photon energy scale uncertainties come mostly from calibration studies using 8 TeV data [31],withadditionalsystematicuncertaintiestotakeintoac- count thedifferencesbetweenthe 2012and2015configurations.

The uncertainties are split into independent components to ac- count forcorrelations oftheuncertainties betweendifferent bins of the measured cross sections.The individual sources of uncer- tainty arevaried by ±¹

σ

inthe MC simulations andpropagated through the analysis separately to maintain the full information aboutthecorrelationsoftheuncertainties betweendifferentbins ofthe measured crosssections.Theimpact ofthephoton energy resolutionuncertaintyismuchsmallerthanthatofthephotonen- ergy scaleuncertainty. Theresulting uncertaintyinthemeasured crosssectionsisobtainedbyaddinginquadratureall theindivid- ualcomponentsandincreasesfrom1% atEγ

T =^{125 GeV to4}.5% at Eγ

T∼¹.5 TeV.

6 IntheLOSherpasamples,theapplicationoftheFrixione’scriteriontothepho- tonisolationatmatrix-elementlevelpreventstheradiatedphotonfrombeingclose toaparton. InthePythiasamples,the bremsstrahlungcomponentissimulated withapartonshowerapproachand,asaresult,theradiatedphotoncanbeclose toaparton.

Jetenergyscaleandresolution.Adetailedassessmentofthe un- certainties in the jet energy scale and resolution is made us- ing the method reported in Ref. [42]. The individual sources of uncertainty [42] are varied by ±¹

σ

in the MC simulations and propagated through the analysis separately, to maintain the full information about the correlations of the uncertainties between different bins of the measured cross sections. The resulting un- certaintyinthemeasured crosssectionsisobtainedbyaddingin quadraturealltheindividualcomponentsandincreasesfrom1.9%

atp^jet-lead_T =^{100 GeV to7}.5% at p^jet-lead_T ∼^{1 TeV.}

Parton-showerandhadronisationmodeldependence.Theeffects duetotheparton-showerandhadronisationmodelsonthesignal purityanddetector-to-particle-levelcorrection factorsare studied separately.The effects on the signal purity are estimatedas the differencesobserved between the nominal results andthose ob- tained using either the (non-adjusted) Pythia or LO Sherpa MC samplesforthedeterminationofthesignalleakage fractions.The differencebetweenthe nominalresultsandthose obtainedusing thePythia-adjustedMCsamplesforthedeterminationoftheun- foldingcorrectionfactorsistakenasasystematicuncertainty.The resultinguncertaintiesinthemeasuredcrosssectionsaretypically smallerthan 2%.

Photonidentification efficiency. The uncertainty in the photon identificationefficiencyisestimatedfromtheeffectofdifferences betweenshower-shape variabledistributions in data andsimula- tion.FromthestudiespresentedinRefs. [30,51],thisprocedureis foundtoprovideaconservativeestimateoftheuncertainties.The resultinguncertaintyinthemeasuredcrosssectionsisintherange 1–2%.Theeffectsonthemeasuredcrosssectionsduetotheuncer- taintyinthephotonreconstructionefficiency,whichareevaluated byrepeatingthefullanalysisusingadifferentdetectorsimulation withincreasedmaterialinfrontofthecalorimeter,arefoundtobe negligible.

Photonisolationmodelling. The differences between the nomi- nal results and those obtained without applying the data-driven correctionstoE^iso_T insimulatedeventsaretakenassystematicun- certainties inthe measurements dueto the modellingof E_T^iso in theMCsimulation.Theresultinguncertaintyinthemeasuredcross sectionsislessthan 1.1%.

Definitionofthebackgroundcontrolregions. The estimation of thebackground contamination inthe signal region isaffected by the choice of background control regions. The control regions B andD aredefinedby thelowerandupperlimitson E^iso_T andthe choiceofinvertedphotonidentificationvariablesusedintheselec- tionofnon-tightphotons.Tostudythedependenceonthespecific choices,thesedefinitionsarevariedovera widerange.The lower limit on E^iso_T inregions B and D isvaried by ±^{1 GeV, which is} largerthananydifference betweendataandsimulations andstill provides a sufficient sample to perform the data-driven subtrac- tion.Theupperlimiton E^iso_T inregions B andD isremoved.The resultinguncertaintyinthemeasured crosssectionsisnegligible.

Likewise, thechoice of invertedphoton identificationvariables is varied. The analysis is repeated using differentsets of variables:

tighter(looser)identificationcriteriaaredefinedbyapplyingtight requirementstoanextended(restricted)setofshower-shapevari- ables in the first calorimeter layer [30,51]. The resulting uncer- taintyinthemeasuredcrosssectionsissmallerthan1.3%.

Photonidentificationandisolationcorrelationinthebackground.

The photon isolation and identification variables used to define

theplaneinthetwo-dimensionalsidebandmethodtosubtractthe backgroundareassumedtobeindependentforbackgroundevents (R^bg=^{1 in Eq. (1)).Any} correlation between thesevariables af- fectstheestimationofthepurityofthesignalsampleandleadsto systematicuncertaintiesinthebackground-subtraction procedure.

Arangein R^bgissettocoverthedeviationsfromunitymeasured inthevalidationregions aftersubtracting thesignalleakage with eitherPythia-adjustedorLOSherpaMCsamples.Theresultingun- certaintyinallmeasuredcrosssectionsislessthan 2%.

Pile-up. The uncertainty is estimated by changing the nominal rescaling factor of 1.16 to 1.09 or 1.23 and re-evaluating the reweighting factors. The resulting uncertainty in the measured crosssectionsistypicallylessthan0.5%.

Unfoldingprocedure.The uncertaintyis estimatedbycomparing the nominal results with those obtained by unfolding with LO Sherpa MC samples reweigthed to match the data distributions.

Theresulting uncertaintyinthemeasuredcrosssectionsisnegli- gible.

The total systematic uncertainty is computed by adding in quadrature the uncertainties from the sources listed above, the statistical uncertainty ofthe MC samples, the uncertainty inthe triggerefficiency(1%)andtheuncertaintyintheintegratedlumi- nosity, which isfullycorrelated betweenall bins ofall the mea- suredcrosssections.Therearelargecorrelationsinthesystematic uncertaintiesacrossbinsofoneobservable,particularlyintheun- certainties due to the photon and jet energy scales, which are dominant. The total systematicuncertainty, excluding that inthe luminosity,islessthan 5% for Eγ

T, 4% for φ^γ−jet, 6% formγ−jet and 4% for|^cosθ^∗| ^{andincreases from4% at pjet-lead}T =^{100 GeV} to10% at p^jet-lead_T ∼¹.5 TeV.Fig.2showsthetotalsystematicun- certainty for each measured cross section, excluding that in the luminosity;thedominantcomponentsareshownseparatelyinthis Figure.Thesystematicuncertaintydominatesthetotalexperimen- tal uncertaintyfor Eγ

T 700 GeV andmγ−jet¹.5 TeV,whereas for higher Eγ

T and mγ−jet values, the statistical uncertainty of the data limits the precision of the measurements. For p^jet-lead_T , φ^γ−jet and|^cosθ^∗|^{,thesystematic}uncertaintydominatesinthe wholemeasuredrange.

8. Theoreticalpredictions

The NLO pQCD predictions presented in this Letter are com- puted using two programs, namely Jetphox 1.3.1_2 and Sherpa 2.2.2.TheJetphoxprogramincludesafullNLOpQCDcalculationof both thedirectandfragmentationcontributions tothe crosssec- tion for the pp→

γ

₊jet+^{X process. The numberof massless} quark flavours is set to five. The renormalisation scale

μ

R, fac- torisation scale

μ

F and fragmentationscale

μ

f are chosen to be

μ

R=

μ

F=

μ

f=^EγT [14].Thecalculationsareperformedusingthe MMHT2014 [52] PDFset and the BFG set II of parton-to-photon fragmentation functions [53], both at NLO. The strong coupling constantissetto

α

s(m_Z)=⁰.120.Thecalculationsareperformed using a parton-level isolation criterion which requires the total transverseenergyfromthepartonsinsideaconeofsizeR=⁰.4 aroundthephotondirectiontobebelow4.2·¹⁰⁻³·^Eγ_T +^10GeV.

The NLO pQCD predictions from Jetphox are at the parton level whilethemeasurements are atthe particlelevel.Thus,there can bedifferencesbetweenthetwolevelsconcerning thephotoniso- lationaswellasthephotonandjetfour-momenta.Sincethedata are correctedforpile-upandUE effectsandthedistributions are unfoldedtoaphase-spacedefinitioninwhichtherequirementon

Fig. 2.Totalrelativesystematicuncertainty(solidlines),excludingthatintheluminositymeasurement,asafunctionofE^γ_T,p^jet-lead_T ,φ^γ−jet,m^γ−jetand|^cosθ^∗|^.Thethree dominantcontributionsarealsoincludedseparately:thejetenergyscale(dashedlines),thephotonenergyscale(dottedlines)andthephotonidentification(dot-dashed lines).Theshadedbanddisplaystherelativestatisticaluncertainty;forthelastpointinE^γ_T (m^γ−jet)therelativestatisticaluncertaintyis32% (30%).

E^iso_T at particle level is applied after subtractionof the UE, it is expectedthatparton-to-hadroncorrectionstotheNLO pQCDpre- dictionsaresmall.CorrectionfactorstotheJetphoxpredictionsare estimatedby computing the ratio ofthe particle-level cross sec- tionforaPythiasample withUE effectstotheparton-levelcross sectionwithoutUEeffects.Thesefactorsareclosetounitywithin

±^{5% forthe}observablesstudied,exceptforp^jet-lead_T ^{600 GeV;in} thisregion,whichisdominatedbythebremsstrahlungcomponent, thefactorscandifferbyupto30% fromunitysincehadronisation ofanearbypartoncansignificantlychangetheparticle-levelisola- tioncomparedtotheparton-levelisolation.

The Sherpa 2.2.2 program consistently combines parton-level calculationsof

γ

+(1, 2)−^{jet eventsatNLOand}

γ

+(3, 4)−^jet events at LO [48,49] supplemented with a parton shower [54]

whileavoidingdouble-countingeffects [55].Arequirementonthe photonisolationatthematrix-elementlevelisimposedusingFrix- ione’scriterion withR=⁰.1,n=^{2 and} =⁰.1.Dynamicfactori- sationandrenormalisationscalesareadoptedaswellasadynam- icalmerging scale with Q¯_cut=20 GeV [56]. The strongcoupling constantissetto

α

s(mZ)=⁰.118.Fragmentationintohadronsand simulationoftheUEareperformedusingthesamemodelsasfor theLOSherpasamples.Thenext-to-next-to-leading-order(NNLO) NNPDF3.0PDFset [57] isusedinconjunctionwiththecorrespond- ing Sherpatuning. All the NLO Sherpapredictions are based on theparticle-levelobservablesfromthiscomputationafterapplying therequirementslistedinTable1.

8.1.Uncertaintiesinthepredictions

TheuncertaintyintheNLOpQCDpredictionsfromJetphoxdue to terms beyond NLO is estimated by repeating the calculations using values of

μ

R,

μ

F and

μ

f scaled by the factors 0.5 and 2.

The three scales are either varied simultaneously, individually or byfixingoneandvaryingtheothertwo.Inallcases,thecondition 0.5≤

μ

A/

μ

B≤^{2 is imposed,where A},B=^R, F, f.The final un- certaintyistakenasthelargestdeviationfromthenominalvalue amongthe14 possiblevariations. Inthe caseoftheNLO Sherpa prediction,whichdoesnotincludethefragmentationcontribution,

μ

Rand

μ

F arevaried asaboveandthelargestdeviationfromthe nominalvalueamongthe6possiblevariationsistakenastheun- certainty.

TheuncertaintyintheNLOpQCDpredictionsfromJetphoxdue tothechoice ofprotonPDFsis estimatedby repeatingthe calcu- lationsusingthe50 sets fromtheMMHT2014erroranalysis [52]

andapplying theHessian method [58] for evaluationof the PDF uncertainties.InthecaseofNLOSherpa,itisestimatedusing100 replicasfromtheNNPDF3.0analysis [57].

The uncertainty in the NLO pQCD predictions from Jetphox (NLO Sherpa) due to the uncertainty in

α

s is estimated by re- peatingthe calculationsusingtwo additionalsets ofprotonPDFs fromtheMMHT2014(NNPDF3.0)analysis,forwhichdifferentval- uesof

α

s atmZ were assumedin thefits, namely 0.118 (0.117) and0.122 (0.119);inthisway,thecorrelationbetween

α

sandthe PDFsispreserved.

Theuncertaintyintheparton-to-hadroncorrectionisestimated bycomparingthevaluesobtainedusingdifferenttunesofPythia: theATLAS setoftuned parametersfortheunderlyingevent(tune AU2) [28] with the CTEQ6L1 PDF set [59], the A14 tune with theLO NNPDF2.3PDF set aswell as thetunes inwhich the pa- rametersettings ofthe latter relatedto the modellingof the UE arevaried [23].Larger differencesareobtainedfromthecompari- sonofthetwo centraltunes thanfromthevariations aroundthe A14tune. The nominal correction is takenas theaverage of the correctionsusing the two central tunes, while the uncertaintyis estimatedashalfofthedifferencebetweenthetwocentraltunes.

The dominant theoretical uncertainty is that arising from the terms beyond NLO and, inthe case of Jetphox (NLO Sherpa), is

≈10% (15–25%)forEγ

T,mγ−^jetand|^cosθ^∗|^{andincreasesfrom5%}

(15%)at p^jet-lead_T =^{130 GeV to30% (30%)forpjet-lead}_T =¹.5 TeV.In the caseof theNLO Sherpapredictionfor d

σ

/dφ^γ−jet,theun- certaintyincreasesfrom10% atφ^γ−jet∼

π

to40% atφ^γ−jet∼

π

/2.Theuncertainty inthepredictions ofJetphox(NLO Sherpa) arising from that in the PDFs is ^{2% (3%) for all} observables.

The uncertainty arising from the value of

α

s(m_Z) is below 2%

(5%). Theuncertaintyintheparton-to-hadroncorrection isinthe range 1–3% except for p^jet-lead_T ^{600 GeV, where it increases to} 20% at p^jet-lead_T =¹.5 TeV; thisuncertaintyisincluded intheJet- phox predictions, but not in the caseof NLO Sherpa since it is a particle-level Monte Carlo generator.⁷ The total theoretical un- certainty is obtained by adding in quadrature the individual un- certaintieslistedabove and,inthecaseofJetphox(NLOSherpa), is 10–15% (15–25%) except for p^jet-lead_T , where it is in the range 10–40% (15–30%); in the caseof the NLO Sherpa prediction for d

σ

/dφ^γ−jet,thetotaluncertaintyis10–40%.

9. Results

The measurements presented here apply to isolated prompt photonswithE^iso_T < 4.2·¹⁰⁻³·^EγT +^{10GeV atparticleleveland} jetsof hadrons. The measured fiducial cross section forisolated- photonplusone-jetproductioninthephase-spaceregiongivenin Table 1is

σ

meas=³⁰⁰±¹⁰(exp.) ±⁶ (lumi.)pb, where “exp.”

denotes the sum in quadrature of the statistical and systematic uncertaintiesexcludingthatduetotheluminosityand“lumi.”de- notes the uncertainty due to that in the integrated luminosity.

ThefiducialcrosssectionspredictedbyNLO QCDJetphox(multi- leg NLO QCD plus parton-shower Sherpa) using the MMHT2014 (NNPDF3.0)PDFsetare

σ

Jetphox

=

²⁹¹⁺₋²⁵₂₁

(

scale

)

⁺₋²₃

(

PDF

)

⁺₋⁴₅

( α

s

) ±

⁶

(

non-perturb.

)

pb and

σ

NLOSherpa

=

³¹⁹⁺₋⁵⁴₄₅

(

scale

) ±

³

(

PDF

)

⁺₋¹⁰₁₁

( α

s

)

pb

,

whichareconsistentwiththemeasurementwithinthetheoretical uncertainties.

Fig.3showstheisolated-photonplusjetcrosssectionsasfunc- tions of Eγ

T, p^jet-lead_T ,φ^γ−jet,mγ₋jet and|^cosθ^∗|^.Themeasured d

σ

/dEγ

T decreases by almost six orders of magnitude over the Eγ

T range studied. Values of Eγ

T up to 1.5 TeV are measured.

The measured d

σ

/dp^jet-lead_T decreases by more than four orders ofmagnitudefrom p^jet-lead_T =^{100 GeV uptothehighestmeasured} value, p^jet-lead_T ≈¹.5 TeV.Themeasurementofd

σ

/dφ^γ−jet isre- strictedto φ^γ−jet>

π

/2 toavoidthe phase-spaceregiondomi- natedbyphotonproductioninassociationwithamulti-jetsystem.

The measured d

σ

/dφ^γ−jet increases asφ^γ−jet increases.The measuredd

σ

/dmγ−^jetdecreasesbymorethanfourordersofmag- nitudeup to thehighestmeasured value,mγ−jet=³.25 TeV.The measuredd

σ

/d|^cosθ^∗|^increasesas|^cosθ^∗|^increases.

The tree-level predictions of the Pythia and LO Sherpa MC models are compared to the measurements inFig. 3. These pre- dictionsare normalised to themeasured integratedfiducial cross

7 Anuncertaintyrelatedtothemodelling ofthehadronisationprocessshould alsobeassignedtotheNLOSherpapredictions,butnotuneotherthanthedefault oneisavailable.Itisexpectedthattheuncertaintyshouldbeofsimilarsizeasthat evaluatedusingPythia.