Interplay of target mass and threshold corrections in large-x PDF determination

Transcription

1 Interplay of target mass and threshold corrections in large-x PDF determination Alberto Accardi Hampton U. and Jefferson Lab Intersections of BSM Phenomenology and QCD for New Physics Searches INT, Seattle, October 22 nd, 2015 and much more...

2 Plan Intersecons Hadronic, nuclear, parcle physics A global approach across subelds Proton PDFs Proton vs. nuclear targets The CJ15 PDFs almost nalized!! Deuteron correcons and the d/u rao How to use proton targets to study the deuteron Connecon to la#ce QCD Threshold resummaon: the new PDF froner Interplay of Target Mass and Threshold correcons Conclusions 2

3 Intersections

4 Why (n)pdfs? High-energy (large to small x) Beyond the Standard Model searches Precision (Higgs) physics NuTeV weak mixing angle Small-x and gluonic ma3er Accardi Mod.Phys.Lett. A28 (2013) 35 Forte and Watt Ann.Rev.Nucl.Part.Sci. 63 (2013) 291 q-q lumi Hadron structure (large to medium x) E6ects of connement on valence quarks q qbar asymmetries; isospin asymmetry Strangeness, intrinsic charm d/u d/u CJ12 x Nuclear Physics Bound nucleons, EMC e6ect, SRC p+a and A+A collisions at RHIC / LHC Color propagaon in nuclear ma3er JLab E

5 A PDF landscape Pert. order QED correcons Resum maon(s) N3LO NNLO NLO Do we need N 3 LO parton distribuons? Forte et al., PLB 731 (2014) HERAPDF CT NNPDF MMHT JR AB(K)M NNPDF CTEQ-JLab Plenty of NLO NNPDF Theory input (roughly x) On the way to LT NUCL TMC/HT RESUM 1% precision theory uncertaines (scale, as, HQ scheme, masses...) Quark-hadron duality 5

6 A npdf landscape Atomic number ncteq DSSZ EPS HKN Kulagin-Pe# Much room for progress! 1 MMHT AB(K)M / JR CTEQ-JLab LT NUCL TMC/HT RESUM Theory input (roughly x) x > 1, SRC, exoca 6

7 Needs the betrothal of HEP and NUCL A global approach across sub5elds PDFs Nuclear data Global QCD ts pqcd New physics Hadron structure In-medium q & g LQCD pqcd LQCD pqcd HEP data Global QCD ts Nuclear, hadron theory 7

8 Proton PDFs

9 Data coverage LHC data W and Z DY and DIS (on p, d) 9

10 Data coverage LHC data Tevatron Jets CJ12 DY and DIS (on p, d) 10

11 Data coverage LHC data DIS prot. & deut. standard cut resonances CJ12 / 15 CJ12 11

12 Large-x, small-q 2 corrections 1/Q 2n suppressed: Higher-twists (non-pert. parton correlaons) in pracce, 3ed to low-q 2 DIS data Target mass correcons Heavy quark masses hot debates Current jet invariant mass,... Under control (extracted HT include residual power correcons) Accardi et al Alekhin et al Non-suppressed Nuclear correcons or deuteron targets Melnitchouk [WG1]; Thorne [WG1] Threshold resummaon calculaons available Flexible d-quark parametrizaon Need to allow d/u nite, as x 1 (as required in theorecal models) e.g. [used in CJ12, analogous one in CT14] alternave approach Liu [WG1] 12

13 Global fits overview See parallel talks in WG1 Also from RHIC HERAPDF2.0 JLab HERA I+II Tevatron new W,Z LHC di-m Nucl. HT TMC Flex d Res um CT14 MMHT14 NNPDF3.0 [ GJR14 ] ( ) CJ12 * ( CJ15) ( ) ( ) ABM12 ** * NLO only ** No jet data but see no reconstructed W Placakyte 13

14 Deuteron corrections No free neutron use deuteron CJ: Nuclear modeling Connects to underlying nuclear theory Can reject models verify assumpons Cross check with other processes DY(p+d), 3 He, polarized DIS,... Connuous vs. discrete parameters MMHT: parametrize D/(p+n) rao Nuclear uncertainty straighsorward No model bias (beside parametrizaon) Limited nuclear physics output May be missing non-negligible Q 2 dependence Needs one parametrizaon per process quark p or n Nucleus Low-energy factorizaon issues Renorm. of nucl. operators, gauge inv., FSI,... 14

15 Deuteron corrections The 2 complementray strategies agree Can use reliably, extend to lower W 2 MMHT14, arxiv: Nucl / HEP symbiosis: Reduced PDF uncertainty with SLAC/JLab data d-quark suppression CJ10 PRD81 (2010) CJ12 MPLA 28 (2013) W and Z constrain d-quark constrain deuteron correcons Abundant DIS deuteron data precise u/d Tavor separaon 15

16 Use protons to study nuclei (!) Directly reconstructed W: highest sensivity to large x CJ11 nuclear uncertainty Accardi, Mod.Phys.Lett. A28(2013)35 Brady, A.A., Melnitchouk, Owens, JHEP 1206 (2012) 019 p p (p) q q W,Z l l sensive to d at high x Can constrain Deuteron models! New precise DØ data (+BONUS) determine d-quark W; also W l+n and Z (but less sensive) Use DIS on Deuteron to: Constrain nuclear model / parameters Reduce d-quark uncertainty 16

17 The CJ15 parton distributions

18 CJ15 data set 18

19 CJ15 PDFs 19

20 Effect of new data sets on d/u ratio 20

21 Effect of new data sets on d/u ratio 21

22 Effect of new data sets on d/u ratio 22

23 Effect of new data sets on d/u ratio 23

24 Effect of new data sets on d/u ratio 24

25 Hadronic phsyics output 25

26 Effect of new data sets on d/u ratio 26

27 Nuclear corrections At large x, DIS dominated by incoherent ssca3ering from individual nucleons O6shell expansion; parametrize rst order coeucient; x x1 with valence sum rule 27

28 Constrained by D0 data (!) The wrong nuclear correcons creates tension between DIS(D) and W asym The ts then choses the right one Deuteron to nucleon EMC rao D/(p+n) Stable w.r.t. choice of nucleon wave funcon WJC1 disfavored c 2 -wise No evidence for anshadowing OC shell correcon 5rst me in Deuteron! Stable shape Magnitude compensates for wave funcon's missing / excessive strength Physical result or #ng away other physics? Can we get guidance from ladce QCD? 28

29 New lattice QCD technique: PDFs in x-space H.-W. Lin, J.-W. Chen, S.D. Cohen, X. Ji arxiv: CJ12 bands courtesy of J. Guerrero Can we adapt these to study oc-shell PDF modi5caons? 29

30 Threshold resummation: the new PDF frontier

31 Why yet another large-x correction?? New precision data in the (very) near future: Jlab 12 tagged DIS, PVDIS RHIC W, Z W+c at LHCb, Long Range Plan: Direct photons: removes tensions in old, but unused, data Resummaon allows use in global ts 10% reducon in large-x gluon errors N. Sato, Ph.D. thesis 2014 Drell-Yan: Alekhin et al., PRD74 (2006) accidental cancellaon at NLO of large ecects? At NNLO, tension with DIS, vector bosons Resummaon e6ects are large Need to be revisited 31

32 Why yet another large-x correction?? ECects extend down to not so large x For example in DIS Need resummed PDFs for precision physics at LHC Accardi, Anderle, Ringer PRD 91 (2015) To use with resummed partonic cross secons: t-tbar squark and sgluino producon. Resummed u- and d-quarks suppressed at high x Implicaons for npdf extracon? Does it also suppress or enhance the d/u rao? Gluon an-correlated to d, u enhanced at large x are we underpredicng radiave charm at large x? 32

33 Brief history of PDFs with large-x resummation 2005: 5rst evaluaon of threshold correcons ecects on u-quark Corcella, Magnea, Phys.Rev. D72 (2005) : combining resummaon with Target Mass Correcons Accardi, Anderle, Ringer, Phys.Rev. D91 (2015) 3, : 5rst 5t of PDFs with threshold correcons NNPDF (Bonvini et al.), JHEP 1509 (2015) 191 Data on DIS only More work to do! But need consistent inclusion of resummaon on top of other large x correcons 33

34 Target mass corretions an overview Kinemacs: Schienbein et al., JPG 35 (2008); Brady, A.A., Hobbs, Melnitchouk, PRD 84 (2011) Light-cone unit vectors Nachtmann variable Naively: 34

35 Target mass corretions an overview Schienbein et al., JPG 35 (2008); Brady, A.A., Hobbs, Melnitchouk, PRD 84 (2011) Idenfy power suppressed (in M 2 /Q 2 = r 2 1 ) but leading twist terms OPE: resummed to all orders Georgi, Politzer, De Rujula 1979 Collinear Factorizaon: up to O(M 4 /Q 4 ) Ellis, Furmanski, Petronzio 1979 THRESHOLD PROBLEM predicts scajering at x B >1! (1) Neglect of partonic kinemac (m N =0 implicitly used here) (2) Non uniform convergence of power series 35

36 Correct partonic kinematics Accardi, Qiu, JHEP (2008) Impose in the handbag diagram phase space close at x B =1 Quesons: How to extend this x to the power suppressed LT terms? Phase space in parton's momentum fracon x is reduced: How to perform threshold resummaon? 36

37 Threshold resummation Perturbave expansion Accardi, Anderle, Ringer PRD 91 (2015) At large x B, sca3ered quark carries large y = x / x : Only so\ gluons (x 0) can be emi3ed Spoils perturbave convergence also if 37

38 Threshold resummation These threshold divergences need to be resummed: Mellin transform Accardi, Anderle, Ringer PRD 91 (2015) Exponenate the divergent pieces in the hard sca3ering C T N, keep nite pieces, Mellin transform back calculable perturbavely 38

39 Threshold resummation with TMC Accardi, Anderle, Ringer PRD 91 (2015) With TMCs, partonic phase space does not extend any longer to y=1 39

40 Threshold resummation with TMC Trick: separate integraon in 2 pieces Accardi, Anderle, Ringer PRD 91 (2015) First one is nite, divergences in second; Mellin transform Exponenate C T 1,N as before, Mellin tramsform back Truncated PDF moments 40

41 Resummed structure functions F2: resummed vs. NLO Points: bin-centered data / NLO theory Lines: NLO+resummaon / NLO theory Resummaon ecects decrease with decreasing Q 2 TMCs harden F2 at small Q 2 Large e6ect in DIS region x > Accardi, Anderle, Ringer, PRD 91 (2015) 41

42 Resummed PDFs Let's do global ts with resummed theory What's the e6ect on the ( resummed ) PDFs? NNPDF3.0 with threshold resummaon E.Nocera' talk Reduced data set cf. xed order NNPDF3.0 Large W 2 > 12.3 GeV 2 cut keeps only x B < 0.7 data Not much data in resummaon region at x > Small e6ect on d, u singularly even less on d/u Need to lower the W cut!! Plot courtesy of E.Nocera 42

43 Let's go non-perturbative SoM-gluon resummaon pqcd treatment of large log(1-x) Pushes perturbave calculaons Catani et al., '90s Collinear Jet funcons Accardi and Qiu, 2008 Can be seen as non-perturbave extension of emission Bridging the rapidity gap Eventually, need to account for so\ interacons So\ factors? Collins, Rogers, Stasto, 2008 (TMD / fully unintegrated factorizaon) 43

44 Conclusions

45 Conclusions Entering a new precision era in large-x PDFs New data (now and in the future), new #ng approaches Conquering nuclear correcons Time for threshold resummaon High-energy and nuclear physics need to work together! Progress in hadron / nuclear structure Precision PDFs for BSM searches Next steps: the betrothal of PDFs and npdfs! Use la#ce QCD fro nuclear PDF modicaons 45

46 Appendix: strangeness fits

47 Strangeness and strangeness asymmetry In pre-lhc ts, mostly constrained by n+a data CCFR inclusive DIS NuTeV muon pair producon NOMAD and CHORUS 47

48 Strangeness and strangeness asymmetry In pre-lhc ts, mostly constrained by n+a data CCFR inclusive DIS NuTeV muon pair producon NOMAD and CHORUS A 48

49 Strangeness and strangeness asymmetry In pre-lhc ts, mostly constrained by n+a data CCFR inclusive DIS NuTeV muon pair producon NOMAD and CHORUS Nuclear correcons again... A Inial state nuclear wave-funcon mods Partly under control using npdfs But: double counng!! either use in npdf or in PDF ts! Final state propagaon of the charm quark / D meson Not under theorecal / phenomenological control (cf. heavy quark puzzle in A+A at RHIC, LHC) 49

50 Strange tensions n+a dimuons vs. p+p W+c at LHC Alekhin et al., arxiv: FSI? Kaons in e+p at HERMES But.. fragmentaon funcons uncertainty HERMES, PLB 666 (2008)

51 Strangeness and strangeness asymmetry In my opinion: Don't use n+a data in proton PDF analysis!! Use neutrino data only for npdfs combine with W/Z in p+a nuclear strange t proton's strange PDF with, say, LHC/RHIC W(+c), Z Anchor npdfs to proton PDFs, use nuclear data: Detect deviaons from free proton strangeness W and Z from RHIC, LHC t IS nuclear strangeness Dimuons in n+a t charm's FS energy loss (rather than subsuming in proton's s-quark) 51

52 Appendix: Resonance region

53 Confronting the resonance region* But: where can we expect the handbag diagram to be valid? Current and jet separaon in rapidity with LO kinemacs, Berger criterion: [Berger, ANL-HEP-CP , 1987; Mulders, hep-ph/ ] * in collaboration with Simona Malace see arxiv:

54 Appendix: Very large x at Tevatron and LHC

55 W charge asymmetry at Tevatron Brady, Accardi, Melnitchouk, Owens, JHEP 1206 (2012) 019 Directly reconstructed W: highest sensivity to large x From decay lepton W Æ l+n: smearing in x sensive to d at high x Can constrain Nuclear models! Too li3le large-x sensivity in lepton asymmetry: need reconstructed W 55

56 W charge asymmetry at LHC Brady, Accardi, Melnitchouk, Owens, JHEP 1206 (2012) 019 Directly reconstructed W: highest sensivity to large x From decay lepton W Æ l+n: smearing in x CMS LHCb Would be nice to reconstruct W at LHCb Denitely needs more stascs Is it at all possible?? (too many holes in detector?) Systemacs in W reconstrucon? What about RHIC, AFTER@LHC? 56

57 Z rapidity distribution ~ LHCb Brady, Accardi, Melnitchouk, Owens, JHEP 1206 (2012) 019 `` Nucl. Uncert. Direct Z reconstrucon is unambiguous in principle, but: Needs bejer than 5-10% precision at large rapidity Experimentally achievable? At LHCb? RHIC? AFTER@LHC? Was full data set used at Tevatron? 57

58 W+c at LHCb S.Farry and R.Gauld, Benasque workshop, Feb

59 Appendix: Nuclear corrections

60 CJ12 Deuteron corrections No free neutron! Best proxy: Deuteron Parton distribuons (to be 3ed) nuclear wave funcon (AV18, CD-Bonn, WJC1, ) O6-shell nucleon modicaon (model dependent) Theorecal uncertainty Bound vs. free proton+neutron Strong Q 2 dependence at large x! binding o6-shellness Fermi moon 60

61 Nuclear corrections for p+d DY Same nuclear model for DY cross secons Ehlers, AA, Brady, Melnitchouk, PRD90 (2014) Same as in DIS (in Bj. limit) O6-shell model extended to sea quarks and gluons Spectral funcon in suitable spectator model Pion-cloud e6ects also studied Kamano, Lee, PRD86 (2012) 61

62 Nuclear corrections... Ehlers, AA, Brady, Melnitchouk, PRD90 (2014) Red band: combined wave fn. & o6-shell model uncertainty OC-shell correcons help makes dbar-ubar stay posive 62

63 Future DY reaches into large-x Ehlers, AA, Brady, Melnitchouk, PRD90 (2014) E906/Sea Quest: o6-shell e6ects even more important J-PARC: can cross-check nuclear smearing vs. DIS 63

64 Sea quarks

65 Charge symmetry breaking E866 lepton pairs: Maybe even negave (a theory challenge...) E906 / SeaQuest Will focus on large x LHC W/Z producon: Access to x ~ 0.01 range But Theory correcons needed for few % level accuracy 65

66 Appendix: Large-x data

67 New Large-x data: a partial list DIS data minimally sensive to nuclear correcons DIS with slow spectator proton (BONUS) Quasi-free neutrons 3 He/ 3 H raos (Marathon) Jlab Data on free (an)protons, sensive to d e+p: parity-violang DIS HERA (e + vs. e ), EIC, LHeC n+p, n+p : ShiP, ELBNF Near Detector, MINERvA p+p, p+p at large posive rapidity W charge asymmetry, Z rapidity distribuon Drell-Yan data Dimuons: E906, J-PARC (?) p+d at large negave rapidity dileptons; W, Z Sensive to nuclear correcons, cross-checks e+d Tevatron: CDF, D0 LHCb(?) RHIC!! AFTER@LHC RHIC?? AFTER@LHC 67

68 JLab 6 GeV: Quasi-free neutrons for today Spectator proton tagging Nuclear correcons minimized experimentally n X BONUS coll.,,. Tkachenko et al. arxiv: D p 68

69 JLab 12 proton, deuteron structure functions CJ cut: Jlab12 experiment E JLab 12 GeV More than double Q 2 range Similar precision as JLab 6 GeV (largely improve cf. SLAC) DIS region Resonance region 69

70 JLab 12: Quasi-free neutrons for tomorrow Nuclear correcons largely cancel: JLab E n X Spectator tagging 3He/3H cross sec. rao D p JLab E BONUS 12 MARATHON n X n X 70

71 JLab 12: Parity-Violating DIS Longitudinally polarized electrons PV asymmetry Jlab12 experiment E d/u projected CJ12 71

72 Appendix: old and new experiments examples

73 At the EIC Neutral current DIS MEIC s = 31 GeV (ca. 2010) Pseudo data using CTEQ6X ts, L=230 (35) f -1 [Accardi, Ent, Keppel, 2010] 73

74 At the EIC Charged current DIS plot for polarized sca3ering, similar for unpolarized Not opmized at large-x: likely to add a bin around x = 0.85 [Aschenauer et al, 2013] 74

75 . Lipka, DIS 13 WG1 summary Intersections of BSM and QCD, INT, Oct

76 W lepton asymmetry at LHC 76

77 Constraints on strangeness: W,Z, W+c K. Lipka, DIS 13 WG1 summary 77

78 Constraints on strangeness: K ± at the EIC Aschenauer, Stratmann, in

79 Intrinsic charm at the EIC The ulmate test of the intrinsic charm mechanism is possible in charm SIDIS at the EIC with modest luminosies Guzzi, Nadolsky, Olness, Sec. 1.9 in