Probing parton densities with γ and γ +Q production. in p A collisions. François Arleo. Workshop on proton-nucleus collisions at the LHC

Transcription

1 Probing parton densities with γ and γ +Q production in p A collisions François Arleo LLR Palaiseau & LAPTh, Annecy Workshop on proton-nucleus collisions at the LHC Trento, May 2013 Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 1 / 25

2 Introduction Prompt photon production in p A collisions Motivations Phenomenology γ+q production in p A & A A collisions Probing PDF in nuclei Probing heavy-quark energy loss in quark-gluon plasma References FA, T. Gousset, Phys. Lett. B660 (2008) 181 FA, K. J. Eskola, H. Paukkunen, C. A. Salgado, JHEP 04 (2011) 055 T. Stavreva et al., JHEP 01 (2011) 152 & JHEP 02 (2013) 072 Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 2 / 25

3 First LHC measurements A robust interpretation of the data requires a quantitative understanding of hard processes in nuclear collisions Probing parton densities in nuclei (npdfs) Essential to predict benchmark predictions in p A and A A collisions Probing energy loss processes Variety of observables to investigate Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 3 / 25

4 Global fits Global fit analyses of nuclear parton densities DIS and Drell-Yan data [ EKS98, HKM, nds, ndsg, ncteq ]...and hadron production at RHIC [ EPS09, DSSZ ] 2 =1.69 GeV 2 ) Pb G (, EPS09NLO nds HKN [ EPS09 Eskola, Paukkunen, Salgado ] Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 4 / 25

5 Global fits Global fit analyses of nuclear parton densities DIS and Drell-Yan data [ EKS98, HKM, nds, ndsg, ncteq ]...and hadron production at RHIC [ EPS09, DSSZ ] Nuclear Modifications for g p/pb (x,q o =1.3 GeV) ncteq decut3 ncteq decut3g3 ncteq decut3g9 1.5 R g Pb x [ ncteq Schienbein et al & ] Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 4 / 25

6 Global fits Global fit analyses of nuclear parton densities DIS and Drell-Yan data [ EKS98, HKM, nds, ndsg, ncteq ]...and hadron production at RHIC [ EPS09, DSSZ ] Question How to probe small-x gluon shadowing at LHC? which observables why prompt photons look promising Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 4 / 25

7 Comparing observables Advantages and limitations Jets high rates, rich phenomenology, forward rapidities large scales Q GeV 2 Heavy-bosons constraints on sea-quark shadowing large scales Q GeV 2 Prompt photons low Q GeV 2, rich phenomenology parton-to-photon fragmentation process Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 5 / 25

8 Comparing observables [ Aurenche et al ] Very good description of isolated / inclusive photon world-data Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 5 / 25

9 Kinematical range Q 2 (GeV 2 ) photons jets heavy bosons NMC Q s x (x,q 2 ) domain covered at the LHC Photons and jets are clearly complementary Photons cover small Q 2 where shadowing should be large Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 6 / 25

10 Dynamics At leading order Compton scattering q( q)g q( q) γ Annihilation process q q g γ At high energy: Compton Annihilation Simple relationship between prompt photon production and parton densities! Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 7 / 25

11 Nuclear production ratio Definition R pa (x ) = 1 A d 3 σ / d 3 σ dy d 2 (p +A γ +X) p dy d 2 (p +p γ +X) p Most naive estimates [ FA Gousset 2008 ] Around mid-rapidity R pa (p,y) 0.5 [ ] R F2 (x e y ) + R G (x e y ) At (very) forward rapidity R pa (p,y) R G (x e y ) At (very) backward rapidity R pa (p,y) R F2 (x e y ) Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 8 / 25

12 Caveats Relationship between photon momenta and parton kinematics spoiled by fragmentation processes and higher-order corrections Photon fragmentation contribution This component very much reduced using isolation criteria E had E max for particles in a cone (η η γ ) 2 +(φ φ γ ) 2 R 2 Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 9 / 25

13 Caveats Relationship between photon momenta and parton kinematics spoiled by fragmentation processes and higher-order corrections Photon fragmentation contribution Next-to-leading order (NLO) corrections 3-body kinematics in the final state Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 9 / 25

14 Checking the approximation Using ndsg in p A collisions at the LHC R ppb (x ) s = 8.8 TeV y = 0 isolated γ NLO w/ ndsg 0.5 ( R F2 + R G ) x Excellent matching between R pa and nuclear density ratios at the LHC < 2 3% at y = 0 5% at y = 2.5 Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 10 / 25

15 Set-up Prompt photon suppression R pa and R AA computed systematically NLO accuracy [ FA Eskola Paukkunen Salgado 2011 ] Using most recent npdf sets available at the time At RHIC and LHC RHIC: d Au and Au Au at s NN = 200 GeV LHC: p Pb at s NN = 8.8 TeV and Pb Pb at s NN = 5.5 TeV Full error analysis uncertainty computed from the 31 EPS09 error sets For an up-to-date analysis: Helenius, Eskola, Paukkunen and talk Wed PM Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 11 / 25

16 Isospin effect Due to QED coupling σ(ug uγ)/σ(dg dγ) = eu 2 /e2 d = 4 When valence quarks dominate (large x 1) σ(nn γ X) < σ(pn γ X) < σ(pp γ X) σ(aa γ X) < σ(pa γ X) < σ(pp γ X) Consequence R pa (x ) < 1 at large x large transverse momentum: 2p / s NN 1 backward rapidity: e y 1 Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 12 / 25

17 RHIC At mid-rapidity (y = 0) R dau y <0.35 s 1/2 =200 GeV proton PDF EPS09 nds HKN p T [GeV] Interplay between anti-shadowing and EMC effect Large differences between the various npdf sets Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 13 / 25

18 RHIC At forward rapidity (y = 3) R dau y=3 s 1/2 =200 GeV proton PDF EPS09 NLO EPS09 NLO p T [GeV] Important isospin effect from the deuteron projectile Interplay between shadowing and anti-shadowing Effects rather similar in the pion channel Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 13 / 25

19 LHC At y = 0 R ppb y <0.5 s 1/2 =8.8 TeV proton PDF EPS09 nds HKN p T [GeV] Interplay between shadowing and anti-shadowing (like at RHIC at forward rapidity) Future data should discriminate between the various predictions Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 14 / 25

20 Counting rates in p A collisions d /dp T dy (pb/gev) s 1/2 =200 GeV y=0 s 1/2 =8.8 TeV y= p T (GeV) RHIC-II (assuming L int = 0.7 pb 1 ) N > 100/GeV for p 35 GeV LHC (assuming L int = 0.1 pb 1 ) N > 100/GeV for p 100 GeV Statistical accuracy in a year much better than the present spread of theoretical predictions Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 15 / 25

21 Counting rates in p A collisions d /dp T dy (pb/gev) s 1/2 =200 GeV y=0 s 1/2 =8.8 TeV y= p T (GeV) RHIC-II (assuming L int = 0.7 pb 1 ) N > 100/GeV for p 35 GeV LHC (assuming L int = 0.1 pb 1 ) N > 100/GeV for p 100 GeV Another interesting observable : photon plus heavy-quark jet production Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 15 / 25

22 γ +Q production in p A collisions γ +Q production from p p to A A collisions dominated by the LO Compton scattering process Qg Qγ Offers a unique access to heavy-quark nuclear PDF gluon nuclear PDF through QCD evolution Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 16 / 25

23 γ +Q production in p A collisions Caveats NLO corrections 2 3 hard-scattering subprocesses - O(αα 2 s) g +g Q + Q +γ Q +Q Q +Q +γ g +Q g +Q +γ Q + Q Q + Q +γ Q +q q +Q +γ q + q Q + Q +γ Q + q Q + q +γ Photon fragmentation contribution Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 16 / 25

24 Comparison between theory and Tevatron data Measurements by DØ Collaboration [ ] compared to NLO theory [ Stavreva Owens ] Excellent agreement between data and theory in the γ +b channel γ +c data above theory at large p Tγ hint for intrinsic charm? Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 17 / 25

25 Predictions in p A collisions NLO calculations assuming that charm PDF is radiatively generated (c(x,µ = m c ) = 0) dc(x, Q) dt = α s dy 2π y [c(x/y)p Q Q(y)+g(x/y)P g Q (y)] in the variable flavour number scheme Probing charm npdf allows for constraining the gluon sector R g Pb ncteq EPS09 HKN07 R c Pb /Rg Pb ncteq EPS09 HKN07 R c Pb R uv Pb u R c Pb /Ruv Pb x x Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 18 / 25

26 Predictions in p A collisions NLO calculations assuming that charm PDF is radiatively generated (c(x,µ = m c ) = 0) dc(x, Q) dt = α s dy 2π y [c(x/y)p Q Q(y)+g(x/y)P g Q (y)] in the variable flavour number scheme Predictions at RHIC and LHC RHIC: d Au collisions at s NN = 200 GeV LHC: p Pb collisions at s NN = 8.8 TeV Calculation of quenching factors R γq pa = σ(pa γ Q X) A σ(pp γ Q X) using different npdf sets then available: ncteq, EPS09, HKN07 Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 18 / 25

27 Spectra and rates Absolute cross sections dσ/dp Tγ (pb/gev) d+au-> γ +c +X S = 200 GeV, ncteq NLO LO+gg->QQγ gq->gqγ qq->qqγ qq->qqγ ; QQ->QQγ dσ/dp Tγ (pb/gev) p+pb-> γ +c +X S = 8800 GeV, ncteq NLO LO+gg->QQγ gq->gqγ qq->qqγ qq->qqγ ; QQ->QQγ p Tγ (GeV) p Tγ (GeV) Cross section dominated by Compton scattering Important NLO corrections Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 19 / 25

28 Spectra and rates Expected rates at RHIC (PHENIX kinematics) Assuming L year = 0.74 pb 1 Photon + charm N for p Tγ > 7 GeV N 10 2 for p Tγ > 20 GeV Photon + bottom < 100 events for p Tγ > 17 GeV Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 19 / 25

29 Spectra and rates Expected rates at LHC (ALICE kinematics) Assuming L year = 0.1 pb 1 σ ppb γ+q N ppb γ+q γ +c PHOS pb 2270 γ +b PHOS 3300 pb 330 γ + c EMCal pb γ +b EMCal pb 2270 Large for γ +c and γ +b events at LHC Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 19 / 25

30 Constraining the gluon npdf LHC p+pb->γ+c+x / p+p->γ+c+x S = 8800GeV EPS09 HKN07 ncteq Nuclear Modifications for g p/pb (x,q=50 GeV) EPS09 HKN07 ncteq R σ γ+c R g Pb p Tγ (GeV) follows RPb g very closely Almost no overlap between EPS09 and HKN07, and ncteq decut3 R γ+c ppb Measurements with sufficiently small error bars should disentangle the various npdf sets 0.75 x 0.01 Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 20 / 25

31 Constraining the gluon npdf RHIC d+au->γ+c+x / p+p->γ+c+x S = 200 GeV ncteq ncteq - no nuclear corrections to d EPS09 HKN Nuclear Modifications for g p/au (x, Q = x S / 2 ~ p T ) ncteq EPS09 HKN07 R dau γ+c R g Au p Tγ (GeV) Same observation than at LHC : R γ+c dau RAu g Kinematic region probed at RHIC (x = ) complementary to that at LHC (x = ) x Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 20 / 25

32 Probing (massive) parton energy loss in QGP Energy loss of massive partons Heavy quark mass acts as a collinear cutoff for medium-induced gluon radiation, just like in vacuum (dead cone) [ Doskhitzer Kharzeev 2001 ] Clear hierarchy expected ( E ) g > E q > E c > E b Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 21 / 25

33 Probing (massive) parton energy loss in QGP Energy loss of massive partons Heavy quark mass acts as a collinear cutoff for medium-induced gluon radiation, just like in vacuum (dead cone) [ Doskhitzer Kharzeev 2001 ] Clear hierarchy expected ( E ) g > E q > E c > E b [ ALICE ] Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 21 / 25

34 Probing (massive) parton energy loss in QGP Energy loss of massive partons Heavy quark mass acts as a collinear cutoff for medium-induced gluon radiation, just like in vacuum (dead cone) [ Doskhitzer Kharzeev 2001 ] Clear hierarchy expected ( E ) g > E q > E c > E b γ +Q unique tool to probe Q energy loss in the plasma γ q γ c γ b Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 21 / 25

35 Analysis in A A collisions Calculations performed at NLO accuracy at s = 5.5 TeV Heavy quark energy loss ǫ Q estimated on an event-by-event basis from the quenching weight (probability distribution) obtained perturbatively [ Armesto Dainese Salgado Wiedemann 2005 ] Various observables investigated Photon jet energy asymmetry A J A J = E T1 E T2 E T1 +E T2, φ > π/2 Momentum imbalance z γq z γq = p Tγ. p TQ p 2 Tγ Photon jet pair momentum q q = p Tγ +p TQ Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 22 / 25

36 Transverse momentum distributions [ Stavreva FA Schienbein 2013 ] Comparing photon and jet p spectra No (or little) modifications of the photon p Stronger effects on the heavy quark jet spectrum Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 23 / 25

37 Transverse momentum distributions [ Stavreva FA Schienbein 2013 ] Nuclear production ratios R AA Sensitive to the amount of energy loss assumed in the calculation Needs to be compared to γ+ inclusive jet production Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 23 / 25

38 Pair momentum distribution Why q distribution q ǫ Q at LO accuracy if the photon is produced directly The shift in the q distributions should reflect the strength of heavy quark energy loss in the medium Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 24 / 25

39 Summary (γ and) γ +Q in nuclear collisions extremely useful p A collisions Sensitive probe of the gluon and heavy quark PDF in nuclei NLO calculations performed using various npdf sets Reasonably large rates expected at LHC Could include more recent npdf sets e.g. DSSZ & EPS09s A A collisions Access to the mass hierarchy of parton energy loss in QCD plasma Preliminary study at the LHC Francois Arleo (LLR & LAPTh) γ and γ+q production in p A collisions Trento Workshop pa@lhc 25 / 25