Proton Structure from HERA and the impact for the LHC

Transcription

1 Proton Structure from HERA and the impact for the LHC Katerina Lipka, DESY for the H1 and ZEUS Collaborations Lomonosov Conference on High Energy Physics 13

2 Proton structure: fundamental subject in matter studies baryonic matter nucleons (protons, neutrons) mass M N ~ 1 GeV u u d partons (quarks & qluons) valence quarks (u, d) quantum properties BUT: M u ~.3 M N, M d ~.6 M N Origin of the proton mass: QCD energy parton composition, coupling, masses and momentum distributions 1

3 Proton structure in quark-parton model Point-like constituents (partons) behave incoherently Probability f(x) for a parton f to carry the fraction x of the nucleon momentum is an intrinsic property of the nucleon, i.e. process independent Learn about the nucleon structure via lepton-nucleon scattering Electron-proton scattering in parton picture Electron scatters off a charged constituent (parton) of the proton Identify the charged partons with quarks γ, Z exchange: Neutral Current (NC) W ± exchange: Charged Current (CC)

4 Proton structure in quark-parton model Point-like constituents (partons) behave incoherently Probability f(x) for a parton f to carry the fraction x of the nucleon momentum is an intrinsic property of the nucleon, i.e. process independent Learn about the nucleon structure via lepton-nucleon scattering Electron-proton scattering in parton picture Kinematics 4-momentum transfer Q defines distance scale r at which proton is probed Proton Q r ~1.6 fm Q =-q photon virtuality r hc/q =.[fm]/q[gev] 3

5 Proton structure in quark-parton model Point-like constituents (partons) behave incoherently Probability f(x) for a parton f to carry the fraction x of the nucleon momentum is an intrinsic property of the nucleon, i.e. process independent Learn about the nucleon structure via lepton-nucleon scattering Electron-proton scattering in parton picture Kinematics Infinite proton momentum frame: Q =-q photon virtuality x=-q /p q Bjorken scaling partons do not interact, move parallel to the proton, massless, no transverse momentum parton i carries fraction xi of Pp <xi<1, xi =1 4

6 Proton structure in quark-parton model Point-like constituents (partons) behave incoherently Probability f(x) for a parton f to carry the fraction x of the nucleon momentum is an intrinsic property of the nucleon, i.e. process independent Lepton-proton scattering cross section and proton structure functions Neutral Current measured cross-section γ*, Z d σ e± p dxdq πα xq 4 (1 + (1 y) )F y F L xf 3 structure functions F x f q f + q f parton distributions Q =-q photon virtuality x=-q /p q Bjorken scaling y inelasticity Bjorken scaling: if partons do not interact, q=q(x); F=F(x) 5

7 Proton structure in Quantum ChromoDynamics Quarks do interact via gluon exchange. Probability via splitting functions: g(x) q(x) P q(x) qg P qq P gq P gg q(z) q(z) q(z-x) g(z) g(z) g(z-x) q(z-x) g(x) g(z-x) Parton Distribution Functions: number of partons in the proton, carrying momentum between xp and (x+dx)p, as resolved at Q. Scaling violation: F (x) F (x,q ), q(x) q(x,q ) 6

8 Proton structure in Quantum ChromoDynamics Quarks do interact via gluon exchange. Probability via splitting functions: g(x) q(x) P q(x) qg P qq P gq P gg q(z) q(z) q(z-x) g(z) g(z) g(z-x) q(z-x) g(x) g(z-x) Parton Distribution Functions: number of partons in the proton, carrying momentum between xp and (x+dx)p, as resolved at Q. Scaling violation: F (x) F (x,q ), q(x) q(x,q ) Additional dependence on Q quantitatively described in perturbative QCD via Dokshitzer-Gribov-Lipatov-Altarelli-Parisi (DGLAP) Evolution Equations q(x, Q ) lnq g(x, Q ) lnq 1 x 1 x dz z dz z x x q(z,q )P qq + g(z,q )P qg z z x x q(z,q )P gq + g(z,q )P gg z z Quark and gluon PDF coupled in DGLAP PDFs determined experimentally, mostly using data on Deep Inelastic Scattering 6

9 DIS at Hadron-Electron-Ring-Accelerator at DESY World-only ep collider Energies and rates of cosmic ray particles E p =46-9 GeV protons E e =7.6 GeV electrons E dn/de (GeV cm - sr -1 s -1 ) Unique machine to study the proton structure e p compatible with 5 TeV γ beam on a fixed p target HERA measurements 7

10 DIS at Hadron-Electron-Ring-Accelerator at DESY World-only ep collider E p =46-9 GeV E e =7.6 GeV protons HERA I : 199- HERA II: 3-7 collider experiments H1 & ZEUS, s max = 318 GeV integrated Luminosity electrons ~.5 fb -1 / experiment HERA switched off June 7, analyses ongoing Central on the way to final precision Tracker H1 and ZEUS combine experimental data accounting for systematic correlations HERA performs the QCD analysis of (semi) inclusive DIS data (HERAPDF) H1 and ZEUS collaborations provide/support the PDF Fitting Tool (HERAFitter) 7

11 Kinematics at HERA as compared to the LHC DIS at HERA: clean lepton probe Kinematics reconstructed from the scattered lepton (or hadronic final state) fi(x) Proton-proton collision at LHC E 1 ŝ = τs = x 1 x s Reconstruction of x1 and x not straightforward E x 1, = M s exp(±y) 8

12 Kinematics at HERA as compared to the LHC DIS at HERA: clean lepton probe HERA covers low, medium x range of the LHC Q evolution via QCD fi(x) LHC TEVATRON Proton-proton collision at LHC E 1 ŝ = τs = x 1 x s DGLAP HERA E x 1, = M s exp(±y) HERA data: backbone for PDF determination 1 9

13 Structure functions in DIS and proton PDFs DIS cross sections provide an access to parton distribution functions in proton γ, Z Exchange: Neutral Current ep e X d σ e± P dxdq LO : F i = πα xq 4 [Y +F Y xf 3 y F L ] (q i (x)+ q i (x)) dominant contribution LO : F 3 i (q i (x) q i (x)) γ/z interference NLO : F L x α S g(x, Q ) contribution from gluon W ± Exchange: Charged Current ep ν X σ e+ p CC x{(ū + c)+(1 y) (d + s)} x{(u + c)+(1 y) ( d + s)} σ e p CC sensitive to individual quark flavours

14 Recent results on Neutral Current Cross Sections JHEP 9:61 (1) precision 1.5% for Q < 5 GeV arxiv:8:6138 xf γz 3 x q val JHEP 9:61 (1) Q =15 GeV F γz q q gluon distribution via scaling violations sensitivity to valence composition 11

15 Recent results on Charged Current Cross Sections HERA II: improvement in precision: e+ (e-) p by a factor of 3 () luminosity wrt. HERA I JHEP 9:61 (1) improved sensitivity to quark flavour composition 1

16 Combined HERA DIS Cross Sections Neutral Current H1 and ZEUS Charged Current H1 and ZEUS r,nc (x,q ) ± HERA I+II NC e + p (prel.) HERA I+II NC e - p (prel.) x =. (x3.) x =.3 (x17.) HERAPDF1.5 e + p HERAPDF1.5 e - p x =.5 (x9.) x =.8 (x5.) x =.13 (x.) x =.18 (x8.) x =.5 (x.4) x =.4 (x.7) x = Q / GeV HERA Inclusive Working Group August r,cc (x,q ) Q = 3 GeV Q = 5 GeV Q = GeV Q = 15 GeV Q = GeV Q = 3 GeV Q = 5 GeV Q = 8 GeV Q = 15 GeV Q = 3 GeV x x HERA Inclusive Working Group August HERA I+II CC e - p (prel.) HERAPDF1.5 Preliminary H1 and ZEUS measurements are combined accounting for correlations Impressive precision up to high Q and high x QCD analysis of combined HERA NC and CC data: HERAPDF 13

17 HERA parton density functions HERAPDF1.5: most precise DIS data, recommended HERA PDF set xf H1 and ZEUS HERA I+II Combined PDF Fit HERAI +HERAII (prel.) xg (.5) xs (.5) -3 HERAPDF1.5 (prel.) exp. uncert. model uncert. parametrization uncert. - Q = GeV xu v xd v -1 1 x HERA Structure Functions Working Group July DGLAP evolution (QCDNUM, arxiv:5.1481) Heavy quarks: massive Variable Flavour Number Scheme Scales: µ r = µ f = Q, starting scale 1.9 GeV Experimentally uncertainty (Δχ =1) Parameterization at starting scale: Model assumptions: Q min =3.5 GeV, α s (M Z )=.1176 m c =1.4 GeV; m b =4.75 GeV; f s (Q )=.31 14

18 HERA parton density functions HERAPDF1.5 NLO and NNLO most precise DIS data, proper treatment of correlation of errors, allows for model studies Available in LHAPDF with eigenvectors for uncertainties and αs variation Describes LHC data very well, used in ATLAS, CMS and LHCb analyses xf H1 and ZEUS HERA I+II PDF Fit xs (.5) NNLO -3 xg (.5) - HERAPDF1.5 NNLO (prel.) exp. uncert. Q = GeV model uncert. parametrization uncert. xd v -1 xu v 1 x HERAPDF Structure Function Working Group March 11 Final combination of HERA inclusive data and QCD analysis on the way (HERAPDF.) 15

19 i Precision QCD: jet production at HERA (Multi)-jet production directly sensitive to the strong coupling constant DIS, 15 GeV < Q < 15 GeV Normalised Inclusive Jet Normalised Dijet Normalised Trijet H1 Preliminary Photoproduction Q ~ Nuclear Physics B 864 (1) 1 Jet / NC NLO c had NLOJet++ and fastnlo QCDNUM CT, s = < Q < GeV (i = ) < Q < 7 GeV (i = 8) 7 < Q < 4 GeV (i = 6) 4 < Q < 7 GeV (i = 4) 7 < Q < 5 GeV 1 (i = ) 5 < Q < 15 GeV (i = ) P T,Jet 7 5 P T Dijet 7 5 P T Trijet [GeV] αs(mz)=.6 (exp) -.35 (theory) αs(mz)=.1163 (exp) (theory) Measure αs running in a single experiment different measurement methods agree < 1% experimental error 16

20 Additional Constraints on PDFs: Charm at HERA stringent test of QCD, direct probe of the gluon Eur. Phys. J. C 73:311 (13), [arxiv: ] e γ c, b cc _ red. H1 VTX H1 D* HERA-I ZEUS D* 98- ZEUS D H1 D* HERA-II ZEUS c µ X ZEUS D* ZEUS D + H1 and ZEUS Q =.5 GeV Q =5 GeV Q =7 GeV p g(x) c, b.5 Q =1 GeV Q =18 GeV Q =3 GeV use different tagging methods: - meson reconstruction, - large mass and long lifetime.5 Q =6 GeV Q = GeV Q = GeV combination: - orthogonal uncertainties - take into account correlations.5 Q =35 GeV Q =65 GeV Q = GeV HERA x 17

21 Additional Constraints on PDFs: Charm at HERA stringent test of QCD, direct probe of the gluon Eur. Phys. J. C 73:311 (13), [arxiv: ] p e γ g(x) c, b c, b use different tagging methods: - meson reconstruction, - large mass and long lifetime _ cc red.6.4 H1 VTX cc _ red H1 D* HERA-II..5.5 H1 D* HERA-I H1 VTX ZEUS c µ X H1 D* HERA-II H1 D* HERA-I ZEUS c µ X ZEUS D* 98- ZEUS D* ZEUS D ZEUS D + H1 and ZEUS Q =.5 GeV Q =5 GeV Q =7 GeV Q =1 GeV Q =18 GeV Q =3 GeV H1 and ZEUS ZEUS D* 98- ZEUS D* HERA ZEUS D ZEUS D + Q =18 Q =6 GeV GeV Q = GeV Q = GeV combination: - orthogonal uncertainties - take into account correlations..5 Q =35 GeV Q =65 GeV Q = GeV HERA x x Precision ~ 5% at medium Q reached 17

22 Additional Constraints on PDFs: Charm at HERA Eur. Phys. J. C 73:311 (13), [arxiv: ] Inclusion of charm: reduced uncertainty on gluon, charm and light sea...mostly due to better constrained charm-quark mass 18

23 Additional Constraints on PDFs: Charm at HERA Determination of mc(mc) MS at NLO Study mc -choice in variable flavor number schemes mc(mc) = 1.6 ±.5exp ±.3mod±.par±.αs GeV perform PDF fit varying mc in FFNS coefficients Eur. Phys. J. C 73:311 (13), [arxiv: ] Consistent with the world average mc(mc)wa = 1.75 ±.5 GeV Different schemes prefer different Mc 19

24 Additional Constraints on PDFs: Charm at HERA Prediction for W+ (W-, Z ) production at the LHC Study mc -choice in variable flavor number schemes Eur. Phys. J. C 73:311 (13), [arxiv: ] Uncertainty due to differences in charm treatment in PDFs significantly reduced by using optimal Mc Different schemes prefer different Mc

25 HERAFitter: Open-Source Modular Tool for QCD Analysis Developed at HERA, extended to LHC and theory groups, Study the impact of different data on PDFs and test different theory approaches r,nc (x,q ) x=. H1 and ZEUS x=. HERA I NC e + p ZEUS H1 x=.8 x= /d y T (CTEQ 6.6)/dp /d y /d T /dp d anti-k t jets. R =.4 (.< y <.3) PDF (CL9) Data CTEQ 6.6 HERAPDF 1. ATLAS Preliminary HERAPDF 1.5 GJR 8 x=.8 x=.5 Q / GeV experimental input Tevatron inclusive jet cross sections 3 [pb/gev] jet jet d /dy dp T p [GeV] T jet p [GeV] processes: NC, CC DIS, jets, diffraction, heavy quarks (c,b,t) Drell-Yan, W production theoretical calculations/tools Heavy quark schemes: MSTW, CTEQ, ABM Jets, W, Z production: fastnlo, Applgrid Top production NNLO (Hathor) QCD Evolution DGLAP (QCDNUM) kt factorisation Alternative tools NNPDF reweighting Other models Dipole model + Different error treatment models + Tools for data combination (HERAaverager) D RunII NLOjet++ (HERAPDF1.5) + nonperturbative corr. jet 6 y <.4 (x ) jet 3.4 < y <.8 (x ) jet.8 < y < 1. jet < y < 1.6 (x ) jet < y <. (x ) jet -9. < y <.4 (x ) experiments: HERA, Tevatron, LHC, fixed target HERAFitter xf H1 and ZEUS HERA I+II PDF Fit xg (.5) xs (.5) -3 HERAPDF1.5 NNLO (prel.) exp. uncert. model uncert. parametrization uncert. - Q = GeV xu v xd v -1 1 PDF or DPDF,... αs (MZ),mc,mb,mt, fs,.. Theory predictions Benchmarking Comparison of schemes x HERAPDF Structure Function Working Group March 11 1

26 Summary HERA data are unique and of particular importance for PDF determination HERA II inclusive DIS cross section measurements are published H1+ZEUS combination and analysis towards HERAPDF. ongoing HERA jet and charm data provide additional constrains on αs and charm-quark mass HERA Expertise in QCD analysis preserved in HERAFitter development developed and supported by several experiments and theory groups allows experimentalists to perform QCD analysis and test theory approaches open source program, successfully used by the LHC experiments

27 Back up

28 Determination of parton density functions Structure function factorization: for the exchange of Boson V (γ, Z, W ± ) F V (x, Q )= from measured cross sections i=q, q,g 1 x dz C V,i (x z,q,µ F,µ R, α S ) f i (z,µ F,µ R ) calculable in pqcd PDF x-dependence of PDFs is not calculable in perturbative QCD: parameterize at a starting scale Q : f(x)=axb (1-x) C (1+Dx+Ex ) evolve these PDFs using DGLAP equations to Q >Q construct structure functions from PDFs and coefficient functions: predictions for every data point in (x, Q ) plane χ - fit to the experimental data

29 HERA Inclusive DIS Measurements at highest Q ] [pb/gev d /dq NC and CC cross sections at large scales: Inclusive HERA I and II data precision NC: ~1.5%, CC: ~4% Neutral Current Charged Current H1 e p CC + H1 e p CC ZEUS e p CC HERA II + ZEUS e p CC HERA II SM e p CC (HERAPDF 1.5) + SM e p CC (HERAPDF 1.5) y <.9 P e = 3 HERA textbook measurements in electroweak physics HERA H1 e p NC + H1 e p NC ZEUS e 4 p NC HERA II + ZEUS e p NC HERA II SM e p NC (HERAPDF 1.5) + SM e p NC (HERAPDF 1.5) Electroweak unification Q [GeV ] [pb] CC arxiv: JHEP 9:61 (1) EPJC 6 (9) 65 EPJC 7 () 945 EPJC 61 (9) 3 SM: zero cross section for a RH e- or LH e ± HERA Charged Current e p Scattering Q > 4 GeV y <.9 e + p X H1 ZEUS p X [%] e HERAPDF 1.5 H1 ZEUS HERAPDF 1.5 P e Good agreement with the SM 6

30 Combination Procedure Minimized value: µ i measured value at point i δ i statistical, uncorrelated systematic error γ j j correlated systematic error b j shift of correlated systematic error sources m i _ true value (corresponds to min χ ) Measurements performed sometimes in slightly different range of (x,q ) swimming to the common (x,q ) grid via NLO QCD in massive scheme

31 HERAPDF1.7: DIS+ low energy+jets+charm Flexible parametrization heavy flavours: mc=1.5±.15 GeV αs (MZ)=.119 steeper gluon as HERAPDF1.6 xf HERA I+II inclusive, jets, charm PDF Fit xg (.5) xs (.5) -3 HERAPDF1.7 (prel.) exp. uncert. model uncert. parametrization uncert. HERAPDF1.6 (prel.) - Q = GeV xu v xd v -1 1 x HERAPDF Structure Function Working Group June 11 Including the jet and the charm data: decouple the gluon from αs and mc

32 Heavy Quarks and PDF Fits Factorization: F V (x, Q )= i=q, q,g 1 x dz C V,i (x z,q,µ F,µ R, α S ) f i (z,µ F,µ R ) i - number of active flavours in the proton: defines the factorization (HQ) scheme i fixed : Fixed Flavour Number Scheme (FFNS) only light flavours in the proton: i = 3 (4) c- (b-) quarks massive, produced in boson-gluon fusion Q» m HQ : can be less precise, NLO coefficients contain terms ~ ln Q m HQ i variable: Variable Flavour Number Scheme (VFNS) - Zero Mass VFNS: all flavours massless. Breaks down at Q ~m HQ - Generalized Mass VFNS: different implementations provided by PDF groups smooth matching with FFNS for Q m HQ must be assured QCD analysis of the proton structure: treatment of heavy quarks essential

33 Heavy Quark Production at HERA Heavy quarks in ep scattering produced in boson-gluon fusion e γ c (b) Contribution to total DIS cross section: charm: ~ 3% at large Q p g(x) c (b) beauty: at most few % Gluon directly involved: cross-check of g(x) from NC and CC DIS cross sections HQ contributions to the proton structure function F : (e.g. charm) Direct test of HQ schemes in PDF fits

34 HERA Charm Data test PDFs obtained with inclusive DIS HERAPDF is obtained using only inclusive HERA DIS NC and CC data, use VFNS Describes charm cross-sections very good Uncertainty band mostly due to variation of charm quark mass in PDF: 1.35<mc<1.65 GeV red cc. Q =.5 GeV Q = 5 GeV H1 and ZEUS Q = 7 GeV Sensitivity to charm mass in PDF fit increased once HERA charm data used together with inclusive DIS data.5 Q =1 GeV Q =18 GeV Q =3 GeV ) c (M 8 75 HERA-I inclusive H1 and ZEUS Charm + HERA-I inclusive opt M c =1.5 ±.6 GeV.5 Q =6 GeV Q = GeV Q = GeV 7 65 RT standard inclusive+ charm DIS.5 Q =35 GeV Q =65 GeV Q = GeV HERA HERAPDF1.5 6 inclusive DIS only x M c [GeV]

35 HERA Charm Data vs QCD Analysis in FFNS QCD Predictions at NLO (~αs ) and NNLO (~αs 3 ) describe data very well Running mass of charm quark is used in coefficient functions in QCD analysis H1 and ZEUS Determine mc(mc) in MS at NLO cc _ red..5 Q =.5 GeV Q =5 GeV Q =7 GeV Q =1 GeV Q =18 GeV Q =3 GeV 7 68 H1 and ZEUS Charm + HERA-I inclusive FF (ABM) mc(mc) m= c (m1.6 )=1.6 ±.5 GeV c ±.5exp ±.3mod±.par±.αs GeV.5 Q =6 GeV Q = GeV Q = GeV perform PDF fit varying mc in FFNS coefficients.5 Q =35 GeV Q =65 GeV HERA Q = GeV ABM9NNLO MS ABM9NLO MS x m c (m ) [GeV] c Consistent with the world average mc(mc)wa = 1.75 ±.5 GeV

36 cc _ red. HERA Charm Data vs QCD Analysis in VFNS Data are confronted to predictions using Variable-Flavour Number Scheme at NLO (αs) and NNLO (αs ) H1 and ZEUS Q =.5 GeV Q =5 GeV Q =7 GeV cc _ red. H1 and ZEUS Q =.5 GeV Q =5 GeV Q =7 GeV.5 Q =1 GeV Q =18 GeV Q =3 GeV.5 Q =1 GeV Q =18 GeV Q =3 GeV.5 Q =6 GeV Q = GeV Q = GeV.5 Q =6 GeV Q = GeV Q = GeV.5 Q =35 GeV Q =65 GeV Q = GeV HERA.5 Q =35 GeV Q =65 GeV Q = GeV HERA MSTW8NNLO MSWT8NNLO (opt.) CTNLO CTNNLO (prel.) x x Predictions using heavy quark coefficients at higher order describe data better at lower Q

37 Charm mass in Variable Flavor Number Scheme Study charm mass choice in PDF using different VFNS implementations using HERAFitter ) c (M 75 Charm + HERA-I inclusive H1 and ZEUS different implementation of VFNS use mc pole in the HQ coefficients 7 RT standard RT optimised ACOT-full S-ACOT- ZM-VFNS opt M C matching between Nflavor to Nflavor+1, (choosing an interpolation approach and different methods for truncation of the perturbative series) definition of mc(pole) gets as uncertain as matching conditions: mc pole Mc 65 6 parameter Mc is implicitly used in predictions for the LHC processes using VFNS PDFs (CTEQ, MSTW, NNPDF, HERAPDF) M c [GeV] Different schemes prefer different Mc

38 Effect of charm mass in VFNS PDF on σ(w, Z) at NLO NLO prediction for W+ (W-, Z ) production at the LHC: dependence on charm mass in PDF Larger M c charm σ W

39 Effect of charm mass in VFNS PDF on σ(w, Z) at NLO NLO prediction for W+ (W-, Z ) production at the LHC: dependence on charm mass in PDF Only one HQ scheme Mc variation in PDF 1.3 < Mc < 1.5 GeV 3% uncertainty on W prediction Larger M c charm σ W

40 Effect of charm mass in VFNS PDF on σ(w, Z) at NLO NLO prediction for W+ (W-, Z ) production at the LHC: dependence on charm mass in PDF Only Several one HQ schemes Mc variation in PDF 1.3 < Mc < 1.5 GeV 3% uncertainty on W prediction Using different HQ schemes: + 7% uncertainty Larger M c charm σ W

41 Data sensitivity to different heavy quark treatments in PDFs NLO prediction for W+ (W-, Z ) production at the LHC: dependence on charm mass in PDF [nb] - W 44 4 Use optimal Mc in each scheme Charm + HERA-I inclusive RT standard RT optimised ACOT-full S-ACOT- ZM-VFNS opt M C < % H1 and ZEUS ) c (M 75 7 Charm + HERA-I inclusive RT standard RT optimised ACOT-full S-ACOT- ZM-VFNS opt M C H1 and ZEUS s = 7 TeV M c [GeV] [GeV] Uncertainty due to differences in charm treatment in PDFs significantly reduced by using optimal Mc in each HQ scheme in PDF 6 M c

42 Heavy Quarks in PDFs and W/Z at LHC Prediction of W ± cross LHC: dominant uncertainty due to PDF s=14 TeV NLO HERAPDF1. Prediction using m c =1.4 GeV Error band: PDF uncertainty Experimental error Parametrization variation Model assumptions including mc variation rapidity mc variation in PDF: significant uncertainty on W@LHC in central region

43 (HERA)PDF and top quark at the LHC Top quark at CMS: cross approx. NNLO Q =4mt HERAPDF1.5NNLO _ t t (pb) 6 4 approx. NNLO (Langenfeld et al.) HERAPDF15NNLO only PDF uncertainties (68% CL): experimental parametrization model variations: Q > 5 GeV thanks A. Cooper-Sarkar CMS (Prel.) CMS-PAS-TOP-11-5 (11) m t pole (GeV) Dominant uncertainty: variation of Q min imposed on data used in the fit top quark production at the LHC has potential to constrain the high-x gluon

44 PDF fits using HERA jet data: fixed αs xu v xs H1 and ZEUS HERA I+II PDF Fit with Jets Q = GeV - Q = GeV x -1 1 x xd v xg Q = GeV HERAPDF1.6 (prel.) Exp. uncert. Model uncert. Param uncert. HERAPDF1.5f (prel.) - Q = GeV x -1 1 x HERAPDF Structure Function Working Group March 11 Inclusive DIS data: combined HERAI+HERAII Jet data: H1 high Q, EPJ C65 () low Q, EPJ C67 () ZEUS incl. jets PLB547 () incl.+jets NP B765 (7) PDF Fit: - flexible parametrisation - αs(mz) fixed Inclusion of jet data into the PDF fit using fixed αs does not have large impact 13

45 PDF fits with free αs (MZ) xf H1 and ZEUS HERA I+II PDF Fit no jet data xg (.5) xs (.5) -3 HERAPDF1.5f (prel.) free s (M ) Z exp. uncert. model uncert. parametrization uncert. - Q = GeV xu v xd v -1 1 x HERAPDF Structure Function Working Group March 11 14

46 PDF fits with free αs (MZ) xf H1 and ZEUS HERA I+II PDF Fit no jet data xg (.5) xs (.5) -3 HERAPDF1.5f (prel.) free s (M ) Z exp. uncert. model uncert. parametrization uncert. - Q = GeV xu v xd v -1 1 x HERAPDF Structure Function Working Group March 11 xf H1 and ZEUS HERA I+II PDF Fit with Jets + jet data xg (.5) xs (.5) -3 HERAPDF1.6 (prel.) free s (M ) Z exp. uncert. model uncert. parametrization uncert. - Q = GeV xu v xd v -1 1 x HERAPDF Structure Function Working Group March 11 Inclusion of jet data into the PDF fit decouples the gluon and αs (MZ)

47 αs (MZ) from PDF fits including HERA jet data min - Scan of the αs (MZ) in the PDF fit 15 5 H1 and ZEUS (prel.) no jet data + jet data HERAPDF1.5f HERAPDF S (M ) Z HERAPDF Structure Function Working Group March 11 HERAPDF1.6 Preliminary H1 high Q H1 low Q k T Eur. Phys. J. C67, 1 () PDF and αs (MZ) determined in the common fit: Eur. Phys. J. C65, 363 () norm. k T multijets multijets ZEUS incl. k T jets Phys. Lett. B 547, 164 () ZEUS incl. k T jets 98- Phys. Lett. B 649, 1 (7) World average S. Bethke, Eur. Phys. J. C64, 689 (9) H1 and ZEUS (prel.) αs (MZ)=. ±.13exp ±.7model/param ±.1had+.45scale exp. uncert. th. uncert s (M ) Z HERAPDF Structure Function Working Group March 11 From including the Jet data in the PDF fit: determine gluon and αs (MZ)

48 HERAFitter: Open-Source Modular Tool for QCD Analysis Open source code, available at Version.3. released in March 13. Joined effort of experimentalists, theorists and tool-developers Successfully used by the LHC experiments