Search for collectivity with azimuthal J/ψ -hadron correlations in high multiplicity p–Pb collisions at √sNN = 5.02 and 8.16 TeV

This is an elec onic ep in o he o iginal a icle. This ep in may di e om he o iginal in pagina ion and ypog aphic de ail. Au ho (s): Ti le: Yea : Ve sion: Please ci e he o iginal e sion: All ma e ial supplied ia JYX is p o ec ed by copy igh and o he in ellec ual p ope y igh s, and duplica ion o sale o all o pa o any o he eposi o y collec ions is no pe mi ed, excep ha ma e ial may be duplica ed by you o you esea ch use o educa ional pu poses in elec onic o p in o m. You mus ob ain pe mission o any o he use. Elec onic o p in copies may no be o e ed, whe he o sale o o he wise o anyone who is no an au ho ised use . Sea ch o collec i i y wi h azimu hal J/ψ -had on co ela ions in high mul iplici y p–Pb collisions a √sNN = 5.02 and 8.16 TeV ALICE Collabo a ion ALICE Collabo a ion. (2018). Sea ch o collec i i y wi h azimu hal J/ψ -had on co ela ions in high mul iplici y p–Pb collisions a √sNN = 5.02 and 8.16 TeV. Physics Le e s B, 780, 7-20. h ps://doi.o g/10.1016/j.physle b.2018.02.039 2018 Physics Le e s B 780 (2018) 7–20 Con en s lis s a ailable a ScienceDi ec Physics Le e s B www.else ie .com/loca e/physle b Sea ch o collec i i y wi h azimu hal J/ψ-had on co ela ions in high mul iplici y p–Pb collisions a √sNN =5.02 and 8.16 TeV .ALICE Collabo a ion a i c l e i n o a b s a c A icle his o y: Recei ed 27 Sep embe 2017 Recei ed in e ised o m 14 Feb ua y 2018 Accep ed 14 Feb ua y 2018 A ailable online xxxx Edi o : M. Dose We p esen a measu emen o azimu hal co ela ions be ween inclusi e J/ψand cha ged had ons in p–Pb collisions eco ded wi h he ALICE de ec o a he CERN LHC. The J/ψa e econs uc ed a o wa d (p- going, 2.03 <y <3.53) and backwa d (Pb-going, −4.46 <y <−2.96) apidi y ia hei μ+μ−decay channel, while he cha ged had ons a e econs uc ed a mid- apidi y (|η| <1.8). The co ela ions a e ex- p essed in e ms o associa ed cha ged-had on yields pe J/ψ igge . A apidi y gap o a leas 1.5 uni s is equi ed be ween he igge J/ψand he associa ed cha ged had ons. Possible co ela ions due o collec- i e e ec s a e assessed by sub ac ing he associa ed pe - igge yields in he low-mul iplici y collisions om hose in he high-mul iplici y collisions. A e he sub ac ion, we obse e a s ong indica ion o e- maining symme ic s uc u es a ϕ≈0and ϕ≈π, simila o hose p e iously ound in wo-pa icle co ela ions a middle and o wa d apidi y. The co esponding second-o de Fou ie coefficien ( 2) in he ans e se momen um in e al be ween 3 and 6 GeV/cis ound o be posi i e wi h a significance o abou 5σ. The ob ained esul s a e simila o he J/ψ 2coefficien s measu ed in Pb–Pb collisions a √sNN =5.02 TeV, sugges ing a common mechanism a he o igin o he J/ψ 2. ©2018 The Au ho . Published by Else ie B.V. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/by/4.0/). Funded by SCOAP3. 1. In oduc ion The measu emen o angula co ela ions be ween pa icles p oduced in had on and nucleus collisions is a powe ul ool o s udy he pa icle p oduc ion mechanisms. Usually he wo-pa icle co ela ion unc ion is exp essed in e ms o di e ences in he azimu hal angle (ϕ) and pseudo apidi y (η) o he emi ed pa icles. In minimum-bias p o on–p o on (pp) collisions, he dom- inan s uc u es in he co ela ion unc ion a e a nea -side peak a (ϕ, η) ≈(0, 0)and an away-side idge loca ed a ϕ≈π and elonga ed in η[1]. The nea -side peak o igina es om je agmen a ion, esonance decays and em oscopic co ela ions. The away-side idge esul s om agmen a ion o ecoil je s. In colli- sions o hea y ions, he wo-pa icle co ela ion unc ion exhibi s addi ional long- ange s uc u es elonga ed in η[2]. These s uc- u es a e usually in e p e ed as signa u es o collec i e pa icle flow p oduced du ing he hyd odynamic e olu ion o he fi eball. They a e analyzed in e ms o he Fou ie coefficien s o he el- a i e angle dis ibu ions. Assuming ac o iza ion, hese coefficien s a e hen ela ed o he Fou ie coefficien s ( n) o he pa icle az- imu hal dis ibu ion ela i e o he common symme y plane o he colliding nuclei’s o e lap a ea. E-mail add ess: alice -publica ions @ce n .ch. The disco e y o a nea -side idge in high-mul iplici y pp [3] and p–Pb [4] collisions has inc eased he in e es in wo-pa icle angula co ela ions in small collision sys ems. These disco e ies we e ollowed by he obse a ion ha he nea -side idge in p–Pb collisions is accompanied by an away-side one [5,6]. Long- ange s uc u es ha e also been epo ed in wo-pa icle co ela ions in d–Au collisions a RHIC [7,8]. Fu he s udies using mul i-pa icle co ela ions ha e p o en ha he obse ed long- ange co ela- ions a e o a collec i e o igin [9–11]. Mo eo e , he ans e se- momen um and pa icle-mass dependencies o he ncoefficien s in p–Pb collisions ha e been ound o be simila o hose measu ed in A–A collisions, sugges ing a common hyd odynamic o igin o he obse ed co ela ions [12,13]. Al e na i e in e p e a ions, including Colo -Glass Condensa e based models [14] and final-s a e pa on– pa on sca e ing [15], ha e also been p oposed. Long- ange co e- la ions o o wa d and backwa d muons wi h mid- apidi y had ons ha e also been ound in p–Pb collisions a a cen e -o -mass en- e gy pe nucleon pai √sNN =5.02 TeV [16]. The esul s show ha hese co ela ions pe sis ac oss wide apidi y anges and ex end in o he high muon ans e se-momen um in e al, which is dom- ina ed by decays o hea y fla o s. In pp collisions, he J/ψ esonance is o med mainly om pai s o c and ¯ cqua ks p oduced in ha d sca e ing eac ions du ing he ini ial s age o he collision. The heo e ical models desc ibing he h ps://doi.o g/10.1016/j.physle b.2018.02.039 0370-2693/©2018 The Au ho . Published by Else ie B.V. This is an open access a icle unde he CC BY license (h p://c ea i ecommons.o g/licenses/by/4.0/). Funded by SCOAP3. 8ALICE Collabo a ion / Physics Le e s B 780 (2018) 7–20 J/ψp oduc ion combine calcula ions o he p oduc ion o c¯ cpai s wi hin a pe u ba i e Quan um Ch omodynamics app oach wi h he subsequen non-pe u ba i e o ma ion o he c¯ cbound s a e [17]. In p–Pb collisions, he p oduc ion is a ec ed by he modifi- ca ion o pa on dis ibu ion unc ions inside he nucleus [18]as well as possible ene gy loss and inelas ic sca e ing inside nuclea ma e [19,20]. In A–A collisions, he e a e wo addi ional com- pe ing phenomena ha influence he J/ψp oduc ion. Fi s is he supp essed p oduc ion due o he dissocia ion o he c¯ cpai s in he qua k–gluon plasma [21]. Second is he J/ψenhancemen ia ecombina ion o cha m qua ks he malized in he medium [22, 23]. The ecombina ion is expec ed o become p e alen in cen al collisions a he LHC ene gies. Recen ly, he ALICE Collabo a ion has published a p ecise mea- su emen o he second-o de Fou ie coefficien , 2, o he az- imu hal dis ibu ion o he J/ψp oduc ion in Pb–Pb collisions a √sNN =5.02 TeV [24]. The esul s show significan 2in cen al and semi-cen al collisions. The measu ed J/ψ 2a low and in e - media e ans e se momen um can be quali a i ely desc ibed by a anspo model in which he J/ψazimu hal aniso opy is in- he i ed om ha o ecombined cha m qua ks [25,26]. Howe e , a highe ans e se momen um he da a s ill indica es significan 2while he anspo model p edic s significan ly smalle al- ues coming mos ly om pa h-leng h dependen supp ession in he almond-shaped in e ac ion egion o he colliding nuclei and om non-p omp J/ψp oduced om b-had on decays assuming he - malized b qua ks. Gi en hese esul s in Pb–Pb collisions, i is o in e es o s udy he J/ψ-had on azimu hal co ela ions also in he smalle p–Pb sys em. The ecombina ion o cha m qua ks, i any, should ha e much smalle impac , due o he smalle numbe o ini ially p oduced cha m qua ks wi h espec o Pb–Pb collisions. The small sys em size should no lead o a sizeable pa h-leng h dependen supp ession. Ne e heless, he s udy o he J/ψ-had on azimu hal co ela ions could allow o de e mine whene e J/ψ p oduc ion is a ec ed by he medium possibly c ea ed in hese collisions [27–29]. In his Le e , we p esen esul s o long- ange co ela ions be ween o wa d (p-going, 2.03 <y <3.53) and backwa d (Pb- going, −4.46 <y <−2.96) inclusi e J/ψand mid- apidi y cha ged had ons in p–Pb collisions a √sNN =5.02 and 8.16 TeV. Inclu- si e J/ψ e e s o bo h p omp J/ψ(di ec and decays om highe mass cha monium s a es) and non-p omp J/ψ( eed down om b-had on decays). 2. Expe imen al se up and da a samples A de ailed desc ip ion o he ALICE appa a us can be ound in Re . [30]. Below, we b iefly desc ibe he de ec o sys ems essen ial o he p esen analysis. In he ollowing, ηand ylab will deno e he pseudo apidi y and apidi y in he ALICE labo a o y sys em. The muons a e econ- s uc ed in he muon spec ome e co e ing he ange o −4 <η< −2.5. The spec ome e con ains a on abso be loca ed be ween 0.9 and 5m om he nominal in e ac ion poin . The abso be is ollowed by fi e acking s a ions, each made o wo planes o Ca hode Pad Chambe s. The hi d s a ion is placed inside a dipole magne wi h 3 Tm field in eg al. The acking s a ions a e ollowed by an i on wall wi h a hickness o 7.2 in e ac ion leng hs and wo igge s a ions, each one consis ing o wo planes o Resis i e Pla e Chambe s. The posi ion o he in e ac ion poin is ob ained using he clus- e s econs uc ed in he Silicon Pixel De ec o (SPD) [31,32]. The SPD is loca ed in he cen al ba el o he ALICE appa a us and op- e a ed inside a la ge solenoidal magne p o iding a uni o m 0.5 T magne ic field pa allel o he beam line. The SPD consis s o wo cylind ical laye s which co e |η| <2.0 and |η| <1.4wi h espec o he nominal in e ac ion-poin , o he inne and ou e laye , espec i ely. The associa ed cha ged had ons a mid- apidi y a e econs uc ed ia he so-called SPD ackle s, sho ack segmen s o med om he clus e s in he wo laye s o he SPD and he p i- ma y e ex [32]. The V0 de ec o [33] consis s o wo ings o 32 scin illa o coun e s each, co e ing 2.8 <η<5.1(V0-A) and −3.7 <η<−1.7 (V0-C), espec i ely. I is used o igge ing and e en -mul iplici y es ima ion. The da a samples p esen ed he e we e collec ed du ing he 2013 and 2016 p–Pb LHC uns. The collision ene gy was √sNN = 5.02 and 8.16 TeV o he 2013 and 2016 da a samples, espec- i ely. Pa o he 5.02 TeV da a we e collec ed du ing he 2016 p–Pb un. Da a wi h bo h beam configu a ions, namely Pb–nucleus momen um (deno ed as Pb–p collisions) o p o on momen um (de- no ed as p–Pb collisions) o ien ed owa ds he muon spec ome e , ha e been analyzed. The asymme ic beam ene gies, imposed by he wo-in-one LHC magne design, esul ed in collisions whose nucleon–nucleon cen e -o -mass e e ence sys em is shi ed in a- pidi y by 0.465 in he di ec ion o he p o on beam wi h espec o he ALICE labo a o y sys em. The da a we e aken wi h a ig- ge ha equi ed coincidence o minimum-bias (MB) and dimuon igge s. The MB igge was p o ided by he V0 de ec o eques - ing a signal in bo h V0-A and V0-C ings. I s efficiency is ound o be abou 98% [34]. The dimuon igge equi ed a leas a pai o opposi e-sign ack segmen s in he muon igge sys em, each wi h a ans e se momen um (pT) abo e he h eshold o he on- line igge algo i hm. This h eshold was se o p o ide 50% effi- ciency o muon acks wi h pT=0.5GeV/c. The collec ed da a samples o p–Pb and Pb–p collisions a 5.02 TeV (8.16 TeV) co espond o in eg a ed luminosi ies o 8.1 and 5.8 (8.7 and 12.9) nb−1, espec i ely. The maximum in e ac- ion pile-up p obabili y anged up o 3% and 8% du ing 2013 and 2016 da a aking, espec i ely. 3. E en , ack and dimuon selec ion The beam-induced backg ound is ejec ed by equi ing ha he iming signals om bo h ings o he V0 de ec o a e compa i- ble wi h pa icles coming om collision e en s. E en s con aining mul iple collisions (pile-up) a e ejec ed by equi ing one single in e ac ion e ex econs uc ed in he SPD and by exploi ing he co ela ion be ween he numbe o clus e s in he wo laye s o he SPD and he numbe o he econs uc ed SPD ackle s. The longi udinal posi ion o he econs uc ed p ima y e ex (z x) is equi ed o be wi hin ±10 cm om he nominal in e - ac ion poin . The econs uc ed SPD ackle s a e selec ed by ap- plying a z x-dependen pseudo apidi y cu . The cu is adjus ed o exclude he con ibu ion om he edges o he SPD whe e he de ec o accep ance is low. Fo example, we selec ackle s wi hin −1.8 <η<0.5, −1.3 <η<1.3 and −0.5 <η<1.8 o e en s wi h z x =10, 0 and −10 cm, espec i ely. The con ibu- ion om ake and seconda y ackle s is educed by applying a || <5m ad cu on he di e ence be ween he azimu hal an- gles o he clus e s in he wo laye s o he SPD wi h espec o he p ima y e ex. Wi h his cu , he mean pTo he selec ed cha ged had ons is ound o be app oxima ely 0.75 GeV/c[16]. The acks econs uc ed in he muon spec ome e a e equi ed o eme ge a a adial ans e se posi ion be ween 17.6 and 89.5 cm om he end o he on abso be in o de o a oid egions wi h highe ma e ial budge . The acks econs uc ed in he acking chambe s a e iden ified as muons by equi ing hei ma ching wi h co esponding ack segmen s in he igge chambe s. Back- g ound acks a e emo ed wi h a selec ion on he p oduc o ALICE Collabo a ion / Physics Le e s B 780 (2018) 7–20 9 Fig. 1. The Mμμ dis ibu ion in he 3 <pμμ T<6GeV/cin e al fi ed wi h a combina ion o a CB2 unc ion o he signal and a VWG unc ion o he backg ound, o high-mul iplici y (le panel) and low-mul iplici y ( igh panel) p–Pb collisions a √sNN =8.16 TeV. he o al ack momen um and he dis ance o closes app oach o he p ima y e ex in he ans e se plane [35]. The selec ed dimuons a e defined as pai s o opposi e-sign muon acks ha - ing −4 <yμμ lab <−2.5, ans e se momen um pμμ Tbe ween 0 and 12 GeV/cand in a ian mass Mμμ be ween 1 and 5GeV/c2. Only e en s wi h a leas one dimuon sa is ying hese selec ion c i e ia a e conside ed. The da a samples a e spli in o mul iplici y classes based on he o al cha ge deposi ed in he wo ings (V0-A and V0-C) o he V0 de ec o (V0M) [34]. The high-mul iplici y (low-mul iplici y) e en class is defined as 0–20% (40–100%) o he MB igge e en sam- ple. 4. Analysis The Mμμ dis ibu ion in each e en -mul iplici y class and pμμ T bin is fi wi h he combina ion o an ex ended C ys al Ball (CB2) unc ion o he J/ψsignal and a Va iable-Wid h Gaussian (VWG) unc ion o he backg ound [36]. The ail pa ame e s o he CB2 unc ion we e fixed o he alues used in [37,38]. The J/ψpeak posi ion and wid h we e ob ained om he fi in he 0–100% e en class and fixed o hese alues in he o he e en -mul iplici y classes. Examples o he Mμμ fi in he 0–20% and he 40–100% e en classes in he 3 <pμμ T<6GeV/cin e al a e shown in Fig. 1. The angula co ela ions be ween J/ψand cha ged had ons a e ob ained om he associa ed-pa icle (SPD ackle s) yields pe dimuon igge . The yields a e defined as Yi(z x,Mμμ,pμμ T,ϕ,η) =1 Ni ig(z x,Mμμ,pμμ T) d2Ni assoc(z x,Mμμ,pμμ T) dϕdη =1 Ni ig(z x,Mμμ,pμμ T) SEi(z x,Mμμ,pμμ T,ϕ,η) MEi(z x,Mμμ,pμμ T,ϕ,η),(1) whe e Ni ig(z x, Mμμ, pμμ T)is he numbe o dimuons, Ni assoc(z x, Mμμ, pμμ T)is he numbe o associa ed SPD ackle s co ec ed o accep ance and combina o ial e ec s (as shown in he second line o he equa ion and desc ibed below), ϕand η=yμμ lab −η ackle a e he azimu hal angle and (pseudo) apidi y di e ence be ween he igge dimuon and he associa ed SPD ackle . The yields a e calcula ed sepa a ely in each e en -mul iplici y class (index i) and 1 cm-wide z x in e al. The dis ibu ion SEi(z x,Mμμ,pμμ T,ϕ,η)=d2Ni same(z x,Mμμ,pμμ T) dϕdη is he yield o associa ed SPD ackle s om he same e en . The dis ibu ion MEi(z x,Mμμ,pμμ T,ϕ,η) =αi(z x,Mμμ,pμμ T)d2Ni mixed(z x,Mμμ,pμμ T) dϕdη is cons uc ed using he e en -mixing echnique, i.e. combining dimuons om one e en wi h SPD ackle s om o he e en s selec ed in he same e en -mul iplici y class and z x in e al. I se es bo h o co ec o de ec o accep ance and efficiency and o ake in o accoun he combina o ial backg ound. The no maliza ion ac o αi(z x, Mμμ, pμμ T)is defined as 1/(d2Ni mixed(z x, Mμμ, pμμ T)/dϕdη)in he η egion co esponding o he maximal accep ance [16]. Wi hin each e en -mul iplici y class and bin o Mμμ, pμμ T, ϕ and η, he yields Yia e aged o e z x a e ob ained by fi ing he dis ibu ion YiN ig(z x)iMEi(z x) o he dis ibu ion SEi(z x). A Poisson likelihood fi is used in o de o p ope ly deal wi h he cases o low numbe o ackle s. Then, he a e age yields a e p o- jec ed on he ϕaxis in he ange o 1.5 <|η| <5using he me hod desc ibed in [16]. In o de o ex ac he yields pe J/ψ igge , he yields pe dimuon igge in each e en -mul iplici y class, pμμ Tand ϕbins a e fi as a unc ion o Mμμ using he ollowing supe posi ion Yi(Mμμ)=S S+BYi J/ψ +B S+BYi B(Mμμ), (2) whe e Sand Ba e he numbe o J/ψand he backg ound dimuons in each bin o Mμμ ob ained om he in a ian mass fi (using a CB2 unc ion o he J/ψsignal and a VWG unc ion o he backg ound) desc ibed abo e, YJ/ψ is he associa ed yield co esponding o he J/ψ igge and YB(Mμμ)is a second-o de polynomial unc ion aimed o desc ibe he associa ed yields co - esponding o he backg ound. The fi ange is chosen be ween 1.5 and 4.5 GeV/c2. Examples o fi s in high-mul iplici y and low- mul iplici y e en classes a e shown in Fig. 2. Fig. 3shows he ob ained associa ed ackle yields pe J/ψ ig- ge o p–Pb and Pb–p collisions a √sNN =5.02 and 8.16 TeV. 10 ALICE Collabo a ion / Physics Le e s B 780 (2018) 7–20 Fig. 2. Example o associa ed ackle yields pe dimuon igge in he 3 <pμμ T<6GeV/cin e al o high-mul iplici y (le panel) and low-mul iplici y ( igh panel) p–Pb collisions a √sNN =8.16 TeV. The esul o he fi wi h he unc ion om Eq. (2)is ep esen ed wi h he blue solid line. The dashed ed line co esponds o he associa ed ackle yields pe backg ound dimuon. (Fo in e p e a ion o he colo s in he figu e(s), he eade is e e ed o he web e sion o his a icle.) As expec ed, in low-mul iplici y collisions we obse e a signifi- can co ela ion s uc u e on he away side (Fig. 3, op panels), p esumably o igina ing om he agmen a ion o ecoil je s. In high-mul iplici y collisions (Fig. 3, middle panels), a possible en- hancemen on bo h nea (ϕ≈0) and away (ϕ≈π) side can be spo ed on op o he away-side s uc u e. In o de o isola e possible co ela ions due o collec i e e ec s be ween he J/ψand he associa ed ackle s, we apply he same sub ac ion me hod as in p e ious measu emen s [5,6,12,16], namely sub ac ing he YJ/ψ yields in low-mul iplici y collisions om hose in high-mul iplici y collisions (Fig. 3, bo om panels). The sub ac ion me hod elies on he assump ions ha he je co ela ions on he away side e- main unmodified as a unc ion o he e en mul iplici y and ha he e a e no significan co ela ions due o collec i e e ec s in low-mul iplici y collisions (see discussion in Sec ion 6). In o de o quan i y he emaining co ela ion s uc u es, he sub ac ed yields Ysub J/ψ (ϕ)a e fi wi h a0+2a1cosϕ+2a2cos2ϕ.(3) The second-o de Fou ie coefficien V2{J/ψ − ackle ,sub}o he azimu hal co ela ion be ween he J/ψand he associa ed cha ged had ons is finally calcula ed as a2/bhigh 0. The denomina o bhigh 0= a0+blow 0co esponds o he combina o ial baseline o he high- mul iplici y collisions, whe e he pa ame e blow 0is he combina- o ial baseline o he low-mul iplici y collisions ob ained a he minimum o he pe - igge yields, namely in ϕ<π/6. The pa- ame e blow 0is he no maliza ion ac o used in Fig. 3. The pa am- e e a1, which desc ibes he s eng h o he emaining away-side co ela ion s uc u e, is ound o be compa ible wi h ze o in p ac- ically all pJ/ψ Tin e als, in bo h p–Pb and Pb–p collisions a bo h 5.02 and 8.16 TeV. As an al e na i e ex ac ion me hod, he calcula ion o blow 0, he sub ac ion o low-mul iplici y om high-mul iplici y collision yields and he fi o Eq. (3)is done in each bin o Mμμ sepa a ely. Then he V2{J/ψ − ackle ,sub}coefficien is ex ac ed by fi ing V2{μμ − ackle ,sub}(Mμμ)wi h a supe posi ion simila o he one defined in Eq. (2) V2{μμ − ackle ,sub}(Mμμ) =S S+BV2{J/ψ − ackle ,sub} +B S+BVB 2{μμ − ackle ,sub}(Mμμ), (4) whe e he VB 2{μμ − ackle ,sub}(Mμμ)is he second-o de Fou ie coefficien o he azimu hal co ela ion be ween he back- g ound dimuons and associa ed ackle s. The backg ound co- efficien VB 2{μμ − ackle ,sub}(Mμμ)is pa ame e ized wi h a second-o de polynomial unc ion. This pa ame e iza ion is cho- sen since i ep oduces he dimuon 2(Mμμ)cons uc ed om he measu ed muon 2coefficien [16] assuming ha he domi- nan pa o he backg ound is combina o ial. An example o he V2{μμ − ackle ,sub}(Mμμ)fi is shown in Fig. 4. Following he p ocedu e used in Re s. [5,12,16], he V2{J/ψ − ackle ,sub}coefficien is ac o ized in o a p oduc o J/ψand cha ged-had on 2coefficien s. Thus, he J/ψsecond-o de Fou ie azimu hal coefficien J/ψ 2{2,sub}is ob ained as J/ψ 2{2,sub}=V2{J/ψ − ackle ,sub}/ ackle 2{2,sub},(5) whe e he ackle 2{2,sub}is he ackle second-o de Fou ie az- imu hal coefficien ob ained by pe o ming he analysis consid- e ing SPD ackle s as bo h igge and associa ed pa icles. The ob ained alues o ackle 2{2,sub}a e be ween 0.067 and 0.069 depending on he beam configu a ion and collision ene gy, wi h 1–2% ela i e s a is ical unce ain y and 5–6.5% ela i e sys ema ic unce ain y. 5. Sys ema ic unce ain ies The combined s a is ical and sys ema ic unce ain ies o he measu ed ackle 2{2,sub}coefficien o each beam configu a ion and collision ene gy a e aken as global sys ema ic unce ain ies o he co esponding J/ψ 2{2,sub}coefficien s. All he o he sys ema ic unce ain ies o he J/ψ 2{2,sub}coeffi- cien s a e ob ained o each da a sample and pTin e al sepa a ely. The ollowing sou ces a e conside ed. A possible inaccu a e co ec ion o he SPD accep ance is as- sessed by a ying he z x ange be ween ±8 and ±12 cm. Sys- ema ic unce ain ies a e assigned only in he cases o a significan change o he esul s. The significance is defined acco ding o he p ocedu e desc ibed in Re . [39]. The sys ema ic e ec ela ed o he unce ain y o he shape o he dimuon backg ound yields YB(Mμμ)is es ima ed by pe o m- ing he fi wi h Eq. (2)using a linea unc ion o he backg ound e m and a ying he fi ange. The sys ema ic e ec coming om he unce ain y o he signal- o-backg ound a io S/Bis checked ALICE Collabo a ion / Physics Le e s B 780 (2018) 7–20 11 Fig. 3. Associa ed ackle yields pe J/ψ igge in 3 <pJ/ψ T<6GeV/cin p–Pb and Pb–p collisions a √sNN =5.02 TeV (le panels) and 8.16 TeV ( igh panels). The op and he middle panels co espond o he low-mul iplici y and he high-mul iplici y e en classes, espec i ely. The bo om panels show he yields a e he sub ac ion o he low-mul iplici y collision yields om he high-mul iplici y collision ones. The solid line ep esen he fi o he da a as desc ibed in he ex . The dashed, do -dashed and do ed lines co espond o he indi idual e ms o he fi unc ion defined in Eq. (3). All he yields a e no malized o he alue in ϕ<π/6in he low-mul iplici y (40–100%) e en class. Only he s a is ical unce ain ies a e shown. (Fo in e p e a ion o he colo s in he figu e(s), he eade is e e ed o he web e sion o his a icle.) by employing a ious in a ian mass fi unc ions, bo h o he backg ound and o he J/ψsignal. The maximal di e ence o he esul s ob ained wi h he abo e checks wi h espec o he de aul app oach is aken as he co esponding sys ema ic unce ain y. The unce ain y a ising om he employed analysis app oach is ob ained as he di e ence be ween he wo ex ac ion me hods desc ibed in Sec ion 4. As desc ibed in Sec ion 4, by de aul he mixed-e en dis ibu- ion ME(ϕ, η)is no malized o uni y in he η egion co e- sponding o he maximal accep ance. As an al e na i e app oach, no malizing he in eg al o ME(ϕ, η) o uni y is used. No sig- nifican e ec on he ob ained esul s is obse ed and hus no sys ema ic unce ain y is assigned. The used e en -mixing echnique can in oduce sys ema ic bi- ases. The e en mul iplici y dis ibu ion o he selec ed dimuons (1 <Mμμ <5GeV/c2) di e s om ha o he J/ψsignal. Since he cha ged-had on spec a and he cha ged-had on densi y as a unc ion o ηchange wi h e en mul iplici y [34], he non-uni o m (bo h in he azimu hal and longi udinal di ec ions) SPD accep ance can in oduce a bias. The co esponding sys ema ic unce ain y is e alua ed by doing he e en mixing in fine e en -mul iplici y bins. 12 ALICE Collabo a ion / Physics Le e s B 780 (2018) 7–20 Table 1 Summa y o absolu e sys ema ic unce ain ies o he J/ψ 2{2,sub}coefficien s. The unce ain ies a y wi hin he indica ed anges depending on pJ/ψ T. The alues no p eceded by a sign ep esen double-sided unce ain ies. Sou ce o sys ema ics √sNN =5.02 TeV √sNN =8.16 TeV p–Pb Pb–p p–Pb Pb–p Accep anceco ec ion 0 o0.019 0 o0.057 0 o0.011 0 o0.007 Backg ound shape 0.007 o 0.013 0.015 o 0.056 0.011 o 0.013 0.003 o 0.012 Ex ac ion me hod 0.003 o 0.015 0.010 o 0.040 0.002 o 0.011 0.008 o 0.018 E en mixing 0.003 o 0.015 0.004 o 0.025 0.002 o 0.008 0.004 o 0.012 Residual away-side je co ela ion –−0.030 o 0 −0.018 o 0 – To al +0.009 o +0.024 +0.024 o +0.084 +0.013 o +0.019 +0.015 o +0.021 −0.009 o −0.024 −0.024 o −0.090 −0.015 o −0.026 −0.015 o −0.021 Fig. 4. Example o he fi om Eq. (4)in he 3 <pμμ T<6GeV/cin e al o p–Pb collisions a √sNN =8.16 TeV. The dashed line co esponds o he VB 2{μμ − ackle ,sub}(Mμμ). The non-uni o m accep ance o he muon spec ome e coupled o sizeable co ela ions be ween he dimuons and SPD ackle s can bias azimu hally he sample o SPD ackle s used o e en mixing. In o de o check o possible e ec s on ou measu emen , he e en mixing is pe o med in in e als o azimu hal angle o he selec ed dimuons. We obse e no significan sys ema ic e ec as he ob ained esul s show negligible de ia ions wi h espec o he esul s using he de aul e en -mixing echnique. The e ec o a possible esidual nea -side peak is checked by a ying he apidi y gap be ween he igge dimuons and asso- cia ed cha ged-had ons om 1.0 o 2.0 uni s. We obse e no in- dica ion o inc easing 2wi h educed gap and hus conside he de aul gap o 1.5 uni s sufficien o elimina e any significan esid- ual nea -side peak con ibu ion. As shown in Sec ion 4, he ecoil-je away-side co ela ion s uc u e in he high-mul iplici y e en class is g ea ly diminished a e he sub ac ion o he low-mul iplici y e en class. By de- aul , any emaining away-side s uc u e is supposed o be aken in o accoun by he cosϕ e m in Eq. (3). In o de o check o esidual e ec s we p oceed in he ollowing way. Fi s , he co - ela ion unc ion in he low-mul iplici y e en class is fi wi h a Gaussian unc ion cen e ed a ϕ=π. Then, he co ela ion unc- ion in he high-mul iplici y e en class is fi wi h he unc ion om Eq. (3), whe e he cosϕ e m is eplaced by a Gaussian unc ion wi h a wid h fixed o he alue ob ained om he fi in he low-mul iplici y collisions. No clea signa u e o sys ema ic change o he esul s is seen, excep some hin s o a possible e ec in he highes pJ/ψ Tin e al. Conse a i ely, we assign sys ema ic unce ain y as he di e ence wi h espec o he de aul analy- sis app oach. Since he ypical alues o he Gaussian wid h a e a ound 1 ad, one-sided (nega i e) sys ema ic unce ain y is as- signed. In Table 1we p esen a summa y o he assigned sys ema ic unce ain ies o he J/ψ 2{2,sub}coefficien s. No sizeable co ela- ions be ween he pJ/ψ Tin e als a e obse ed and he e o e in he ollowing he unce ain ies a e conside ed unco ela ed. Ou measu emen is o inclusi e J/ψ. The ac ion o J/ψ om decays o b-had ons eaches up o abou 15% a pJ/ψ T≈6GeV/cin p–Pb collisions a √sNN =5.02 [40] and 8.16 TeV [41]. The e o e he eed-down con ibu ion is unlikely o influence significan ly ou esul s. In p inciple, a possible s ong mul iplici y dependence o he eed-down ac ion can po en ially a ec he sub ac ion app oach. Howe e , no e idence o such a s ong dependence is obse ed in pp collisions [42]. As addi ional c oss-checks he analysis is done using al e na- i e e en -mul iplici y es ima o s, a ying he ackle ||cu , applying a cu on he asymme y o ans e se momen um o he wo muon acks, emo ing he pile-up cu s and excluding he SPD egions wi h non-uni o m accep ance in pseudo apidi y. The co e- sponding esul s a e ound o be compa ible wi h hose ob ained wi h he de aul analysis app oach and he e o e no u he sys- ema ic unce ain ies a e assigned. 6. Resul s In Fig. 5we epo he measu ed J/ψ 2{2,sub}coefficien s as a unc ion o pJ/ψ T o p–Pb and Pb–p collisions a √sNN =5.02 and 8.16 TeV. Up o pJ/ψ To 3 GeV/c, no significan de ia ion om ze o is obse ed o ei he p–Pb o Pb–p collisions a he wo colli- sion ene gies. On he con a y, in he pJ/ψ Tin e al be ween 3 and 6 GeV/c, he J/ψ 2{2,sub}is ound o be posi i e al hough wi h la ge unce ain ies. As also shown in Fig. 5, he J/ψ 2coefficien s in 2.5 <y <4in cen al Pb–Pb collisions a √sNN =5.02 TeV each maximal alues in he same pJ/ψ Tin e al [24]. Two me hods a e employed in o de o ob ain he p obabili y ha he J/ψ 2{2,sub}is ze o in he 3 <pJ/ψ T<6 GeV/cin e al. In he fi s me hod, he J/ψ 2{2,sub} alues in he wo pJ/ψ Tin e als (3 <pJ/ψ T<4GeV/cand 4 <pJ/ψ T<6GeV/c) a e combined in o a weigh ed a e age o each apidi y and collision ene gy. The ob- ained p obabili ies a e 0.13% and 0.13% (7.8% and 0.23%) o p–Pb and Pb–p collisions, espec i ely, a √sNN =8.16 TeV (5.02 TeV). Combining all eigh J/ψ 2{2,sub} alues yields a o al p obabil- i y o 1.7 ×10−7. This co esponds o a 5.1σsignificance o he measu ed posi i e J/ψ 2{2,sub}coefficien . The second me hod is Fishe ’s combined p obabili y es [43]. Wi h his me hod one ob- ains p obabili ies o 0.14% and 0.23% (10.3% and 0.41%) o p–Pb and Pb–p collisions a √sNN =8.16 TeV (5.02 TeV), espec i ely. ALICE Collabo a ion / Physics Le e s B 780 (2018) 7–20 13 Fig. 5. J/ψ 2{2,sub}in bins o pJ/ψ T o p–Pb, 2.03 <y <3.53 (le panels), and Pb–p, −4.46 <y <−2.96 ( igh panels), collisions a √sNN =5.02 TeV ( op panels) and 8.16 TeV (bo om panels). The esul s a e compa ed o he J/ψ 2{EP}coefficien s measu ed in cen al Pb–Pb collisions a √sNN =5.02 TeV in o wa d apidi y (2.5 <y <4) using e en plane (EP) based me hods [24]. The s a is ical and unco ela ed sys ema ic unce ain ies a e ep esen ed by lines and boxes, espec i ely. The quo ed global sys ema ic unce ain ies co espond o he combined s a is ical and sys ema ic unce ain ies o he measu ed ackle 2{2,sub}coefficien . The o al p obabili y is 1.4 ×10−6which co esponds o a 4.7σ significance. In he calcula ion o he abo e p obabili ies, bo h s a- is ical and sys ema ic unce ain ies o he measu ed alues a e aken in o accoun . The global sys ema ic unce ain y is no aken in o accoun as i is i ele an in he case o he ze o hypo hesis. The analysis me hod p esen ed in his Le e elies on he as- sump ion ha he e a e no significan co ela ions due o collec i e e ec s in he low-mul iplici y e en class. In case o a p esence o such co ela ions, he measu ed V2{J/ψ − ackle ,sub}is equal o V2{J/ψ − ackle ,high}−blow 0 bhigh 0 V2{J/ψ − ackle ,low},(6) whe e V2{J/ψ − ackle ,high}and V2{J/ψ − ackle ,low}a e he second-o de Fou ie coefficien s o he azimu hal co ela ion be- ween he J/ψand he associa ed cha ged had ons in he high- mul iplici y and he low-mul iplici y collisions, espec i ely, and blow 0/bhigh 0≈1/3 is he a io o he combina o ial baseline in he low-mul iplici y and high-mul iplici y collisions (see Fig. 3). As is demons a ed in Re . [44], he assump ion o no significan collec- i e co ela ions in he low-mul iplici y collisions is ce ainly ques- ionable o ligh -fla o had ons. Ou da a indica es he same, as we obse e a s a is ically significan inc ease o he measu ed al- ues o ackle 2{2,sub}when sub ac ing a lowe e en -mul iplici y, e.g. 60–100%, class. Ul ima ely, he alue o he ackle 2coefficien is ound o be abou 17% highe in case no sub ac ion is applied. The e o e, eplacing he sub ac ed ackle 2{2,sub}coefficien in Eq. (5)by he non-sub ac ed coefficien would mean ha he J/ψ 2 coefficien s a e up o 17% lowe wi h espec o he measu ed J/ψ 2{2,sub}coefficien s. Howe e , assuming ha he J/ψ 2coeffi- cien s ollow he same end as a unc ion o e en mul iplici y as he ackle 2coefficien , hey would be up o 17% highe wi h e- spec o he measu ed J/ψ 2{2,sub}coefficien s. Sub ac ing lowe e en -mul iplici y classes in he measu emen o he J/ψ 2{2,sub} coefficien does no imp o e he p ecision o ou measu emen , be- cause o he limi ed amoun o J/ψsignal in he low-mul iplici y collisions. The nuclea modifica ion ac o o J/ψin p–Pb and Pb–p colli- sions [37,38]as well as he cha ged-pa icle 2coefficien [45–47] in pp collisions show no significan √sNN dependence. As seen in Fig. 5, he measu ed J/ψ 2{2,sub}coefficien s a √sNN =5.02 and 8.16 TeV also appea o be consis en wi h each o he . The la ges absolu e di e ence be ween he esul s a he wo collision en- e gies is obse ed in Pb–p collisions in he 3 <pJ/ψ T<6GeV/c in e al. The significance o his di e ence is a he low (below 1.5σ), because o he la ge unce ain ies o he measu emen a √sNN =5.02 TeV. Hence, he da a o he wo collision ene gies a e combined as a weigh ed a e age aking in o accoun bo h s a- is ical and sys ema ic unce ain ies. In Fig. 6, we p esen hese combined esul s o p–Pb and Pb–p collisions oge he wi h mea- su emen s and model calcula ions o Pb–Pb collisions a √sNN = 5.02 TeV [25]. In Pb–Pb collisions, he posi i e J/ψ 2coefficien s a pJ/ψ Tbe- low 3–4 GeV/ca e belie ed o o igina e om he ecombina ion o cha m qua ks he malized in he medium and a e desc ibed ai ly well by he anspo model [25](see Fig. 6). In p–Pb collisions, he amoun o p oduced cha m qua ks is small and he e o e he con ibu ion om ecombina ion should be negligible. Ou mea- su ed alues a pJ/ψ T<3GeV/ca e compa ible wi h ze o, in line wi h his expec a ion. The e is one publica ion [28] which sugges s ha e en in p–Pb collisions a sizeable con ibu ion om ecom- bina ion could occu due o canonical enhancemen e ec s. The unce ain ies o ou esul s do no allow o confi m o o ule ou his scena io. In Pb–Pb collisions, he measu ed J/ψ 2coefficien s exceed sub- s an ially he heo e ical p edic ions a pJ/ψ T>4GeV/c, whe e he 14 ALICE Collabo a ion / Physics Le e s B 780 (2018) 7–20 Fig. 6. Combined J/ψ 2{2,sub}coefficien s in p–Pb and Pb–p collisions compa ed o he esul s in cen al and semi-cen al Pb–Pb collisions a √sNN =5.02 TeV [24]and he anspo model calcula ions o semi-cen al Pb–Pb collisions a √sNN =5.02 TeV [25]. The solid line co esponds o he con ibu ion om pa h- leng h dependen supp ession inside he medium. The band shows he esul ing J/ψ 2including also he ecombina ion o he malized cha m qua ks and he eed- down om b-had on decays assuming he maliza ion o b qua ks. main con ibu ion o J/ψ 2is expec ed o come om pa h-leng h dependen supp ession inside he medium [25](see Fig. 6). In p–Pb collisions, he medium, i any, has a much smalle size [48] and hence e y li le, i any, pa h-leng h dependen e ec s a e expec ed. In p inciple, he eed-down om decays o b-had ons can gi e a posi i e J/ψ 2a high ans e se momen um in case o a posi i e b qua k 2. Howe e , he la e would ha e o each un easonably high alues gi en he magni ude o he measu ed J/ψ 2{2,sub}and he small eed-down ac ion. Despi e hese con- side a ions, he measu ed posi i e J/ψ 2coefficien s would imply ha he J/ψpa icipa es in he collec i e beha io o he p–Pb col- lision sys em. 7. Summa y We p esen ed a measu emen o he angula co ela ions be ween o wa d and backwa d J/ψand mid- apidi y cha ged had ons in p–Pb and Pb–p collisions a √sNN =5.02 and 8.16 TeV. The da a indica e pe sis ing long- ange co ela ion s uc u es a ϕ≈0 and ϕ≈π, eminiscen o he double idge p e iously ound in cha ged-pa icle co ela ions a mid- and o wa d a- pidi y. The co esponding J/ψ 2{2,sub}coefficien s in 3 <pJ/ψ T< 6GeV/ca e ound o be posi i e wi h a o al significance o 4.7σ o 5.1σ. The ob ained alues, albei wi h la ge unce ain- ies, a e compa able wi h hose measu ed in Pb–Pb collisions a √sNN =5.02 TeV in o wa d apidi y. Al hough he unde lying mechanism is no unde s ood, he compa able magni ude o he J/ψ 2coefficien s a high ans e se momen um in p–Pb and Pb–Pb collisions indica es ha his mechanism could be simila in bo h collision sys ems. Acknowledgemen s The ALICE Collabo a ion would like o hank all i s enginee s and echnicians o hei in aluable con ibu ions o he cons uc- ion o he expe imen and he CERN accele a o eams o he ou - s anding pe o mance o he LHC complex. The ALICE Collabo a ion g a e ully acknowledges he esou ces and suppo p o ided by all G id cen es and he Wo ldwide LHC Compu ing G id (WLCG) collabo a ion. The ALICE Collabo a ion acknowledges he ollow- ing unding agencies o hei suppo in building and unning he ALICE de ec o : A.I. Alikhanyan Na ional Science Labo a o y (Ye e an Physics Ins i u e) Founda ion (ANSL), S a e Commi ee o Science and Wo ld Fede a ion o Scien is s (WFS), A menia; Aus ian Academy o Sciences and Na ionals i ung ü Fo schung, Technologie und En wicklung, Aus ia; Minis y o Communica- ions and High Technologies, Na ional Nuclea Resea ch Cen e , Aze baijan; Conselho Nacional de Desen ol imen o Cien ífico e Tecnológico (CNPq), Uni e sidade Fede al do Rio G ande do Sul (UFRGS), Financiado a de Es udos e P oje os (Finep) and Fun- dação de Ampa o à Pesquisa do Es ado de São Paulo (FAPESP), B azil; Minis y o Science & Technology o China (MSTC), Na- ional Na u al Science Founda ion o China (NSFC) and Minis y o Educa ion o China (MOEC), China; Minis y o Science, Edu- ca ion and Spo s and C oa ian Science Founda ion, C oa ia; Min- is y o Educa ion, You h and Spo s o he Czech Republic, Czech Republic; The Danish Council o Independen Resea ch – Na u- al Sciences, he Ca lsbe g Founda ion and Danish Na ional Re- sea ch Founda ion (DNRF), Denma k; Helsinki Ins i u e o Physics (HIP), Finland; Commissa ia à l’Ene gie A omique (CEA) and Ins i- u Na ional de Physique Nucléai e e de Physique des Pa icules (IN2P3) and Cen e Na ional de la Reche che Scien ifique (CNRS), F ance; Bundesminis e ium ü Bildung, Wissenscha , Fo schung und Technologie (BMBF) and GSI Helmhol zzen um ü Schwe- ionen o schung GmbH, Ge many; Gene al Sec e a ia o Resea ch and Technology, Minis y o Educa ion, Resea ch and Religions, G eece; Na ional Resea ch, De elopmen and Inno a ion Office, Hunga y; Depa men o A omic Ene gy, Go e nmen o India (DAE), Depa men o Science and Technology, Go e nmen o India (DST), Uni e si y G an s Commission, Go e nmen o India (UGC) and Council o Scien ific and Indus ial Resea ch (CSIR), India; In- donesian Ins i u e o Science, Indonesia; Cen o Fe mi – Museo S o ico della Fisica e Cen o S udi e Rice che En ico Fe mi and Is i- u o Nazionale di Fisica Nuclea e (INFN), I aly; Ins i u e o Inno a- i e Science and Technology, Nagasaki Ins i u e o Applied Science (IIST), Japan Socie y o he P omo ion o Science (JSPS) KAKENHI and Japanese Minis y o Educa ion, Cul u e, Spo s, Science and Technology (MEXT), Japan; Consejo Nacional de Ciencia (CONACYT) y Tecnología, h ough Fondo de Coope ación In e nacional en Cien- cia y Tecnología (FONCICYT) and Di ección Gene al de Asun os del Pe sonal Academico (DGAPA), Mexico; Nede landse O ganisa ie oo We enschappelijk Onde zoek (NWO), Ne he lands; The Re- sea ch Council o No way, No way; Commission on Science and Technology o Sus ainable De elopmen in he Sou h (COMSATS), Pakis an; Pon ificia Uni e sidad Ca ólica del Pe ú, Pe u; Minis y o Science and Highe Educa ion and Na ional Science Cen e, Poland; Ko ea Ins i u e o Science and Technology In o ma ion and Na ional Resea ch Founda ion o Ko ea (NRF), Republic o Ko ea; Minis y o Educa ion and Scien ific Resea ch, Ins i u e o A omic Physics and Romanian Na ional Agency o Science, Technology and Inno a- ion, Romania; Join Ins i u e o Nuclea Resea ch (JINR), Minis y o Educa ion and Science o he Russian Fede a ion and Na ional Resea ch Cen e Ku cha o Ins i u e, Russia; Minis y o Educa- ion, Science, Resea ch and Spo o he Slo ak Republic, Slo akia; Na ional Resea ch Founda ion o Sou h A ica, Sou h A ica; Cen- o de Aplicaciones Tecnológicas y Desa ollo Nuclea (CEADEN), Cubaene gía, Cuba, Minis e io de Ciencia e Inno acion and Cen o de In es igaciones Ene gé icas, Medioambien ales y Tecnológicas (CIEMAT), Spain; Swedish Resea ch Council (VR) and Knu & Alice Wallenbe g Founda ion (KAW), Sweden; Eu opean O ganiza ion o Nuclea Resea ch, Swi ze land; Na ional Science and Technology De elopmen Agency (NSDTA), Su ana ee Uni e si y o Technol-