W+ jets and heavy flavour production

W + jets and heavy flavourproductionAll-D0 meeting, July 23rd 2004John CampbellArgonne National LaboratoryW + jets and heavy flavour production – p.1

OutlineThe MCFM Monte Carlo⋆ overview of the program⋆ implementation of vector boson + jets processesNLO predictions for W + 2 jets and W b¯b⋆ basic cross-sections⋆ theoretical similarities between the two processes⋆ effect of NLO on distributionsSingle-tagged eventsSummaryW + jets and heavy flavour production – p.2

¡¢£¤ ¥ ¤¦Heavy flavour as a backgroundEvents containing jets that are heavy-quark tagged are importantfor understanding both old and new physics:⋆ Top decays t → W + b⋆ Much new physics couples preferentially to massive quarks,for instance a light Higgs with m H < 140 GeV decaying to b¯bW + jets and heavy flavour production – p.3

¤¥ ¦¡¡¡¢£ ¤ ©©§¨§ Top processes−→2 jets, both b’s−→2 jets, only one is a b−→≥ 2 jets, two are b’sW + jets and heavy flavour production – p.4

Overview of MCFMW + jets and heavy flavour production – p.5

MCFM overviewJC and R.K. EllisParton level cross-sections predicted to NLO in α Sp¯p → W ± /Zp¯p → W ± + Zp¯p → W ± + γp¯p → W ± + g ⋆ (→ b¯b)p¯p → W ± /Z + 1 jetp¯p(gg) → Hp¯p(V V ) → H + 2 jetsp¯p → W + + W −p¯p → Z + Zp¯p → W ± /Z + Hp¯p → Zb¯bp¯p → W ± /Z + 2 jetsp¯p(gg) → H + 1 jet⊖ low particle multiplicity (no showering)⊖ no hadronization⊖ hard to model detector effects⊕ less sensitivity to µ R , µ F⊕ rates are better normalized⊕ fully differential distributionsW + jets and heavy flavour production – p.6

MCFM InformationVersion 3.4.5 available at:http://mcfm.fnal.govImprovements over previous releases:⋆ more processes⋆ better user interface⋆ support for PDFLIB, Les Houches PDF accord(−→ PDF uncertainties)⋆ ntuples as well as histograms⋆ unweighted events⋆ Pythia/Les Houches generator interface (LO)⋆ ‘Behind-the-scenes’ efficiencyComing attractions:⋆ more processes (Z + b, single top, . . .)⋆ separate variation of factorization and renormalization scalesW + jets and heavy flavour production – p.7

Vector boson + jets in MCFMMany diagrams, sensitive to all parton PDF’sNLO corrections are separated into two classes:REAL extra parton radiation, e.g. (W/Z + 2 jets)soft gluoncollinearquarksVIRTUAL loop diagrams:W + jets and heavy flavour production – p.8

Vector boson + heavy flavour in MCFMIn lowest order b-quark pairs are produced in association with W ’sby gluon splitting alone:Beyond LO, the b-quark is treated as a massless particle in MCFM⋆ a finite cross-section requires a cut on the b-quark p T⋆ this means that this calculation is not suitable for estimatingthe rate with only a single b tagW + jets and heavy flavour production – p.9

Mass effectsLowest order comparison of dijet mass and leading jet p Tdistributions for W b¯b⋆ m b = 4.75 GeV (lowest order only)⋆ m b = 0 (can be calculated to NLO)Overall the cross section decreases by approximately 10% whenincluding the mass. Kinematic distributions are not much affectedaway from regions of low p T (b).W + jets and heavy flavour production – p.10

W/Z+ jet cross-sectionsThe W/Z + 2 jet cross-section has only recently been calculatedat NLO and should provide an interesting test of QCD (cf. manyRun I studies using the W/Z + 1 jet calculation in DYRAD)For instance, the theoretical prediction for the number of eventscontaining 2 jets divided by the number containing only 1 is greatlyimproved.W + jets and heavy flavour production – p.12

Effect of NLO corrections onW b¯b and W + 2 jet ratesW + jets and heavy flavour production – p.13

Jet p T dependenceIncreasing the minimum jet p T reduces the 3 jet contributioncompared to the 2 jet one, so the behaviour of the inclusivecross-section improves.p T (jet) > 30 GeVp T (jet) > 7.5 GeVW + jets and heavy flavour production – p.15

Scale dependence of K-factorsStrong scale dependence.The W b¯b K-factor varies greatly with the minimum jet p T , whereasthe W + 2 jets one does not.dotted: W + 2 jetssolid: W + b¯bScale dependence has a similar shape for both processes.W + jets and heavy flavour production – p.16

K-factor ratioImportant for CDF’s “Method 2”. Essentially, is a lowest orderestimate of (W b¯b/W + 2 jets) reproduced at NLO?A qualified ”yes” - it is for scale choices around 50 GeV or greaterand p T cuts of about 15 GeV or greater.As the jet p T cut is lowered, the ratio gets worse (increases).W + jets and heavy flavour production – p.17

Heavy flavour fractionAt NLO, this fraction is stable across a wide range of scales.dotted: LOsolid: NLO exclusiveFor a p T cut of 15 GeV and µ ∼ M W , we have:[σ(W b¯b)σ(W + 2 jets)]LO= 1.16%,[]σ(W b¯b)σ(W + 2 jets)NLO= 1.23%W + jets and heavy flavour production – p.18

Effect of NLO corrections onW b¯b and W + 2 jet distributionsW + jets and heavy flavour production – p.19

¯b mass cutSuch a cut would be helpful, if it could be experimentally enforced:⋆ It improves the massless approximation⋆ It reduces this background compared to, for example, t¯tproduction, since here the b’s like to lie at low invariant mass.LO (m b = 0)LO (m b = 4.75)NLO (m b = 0)W + jets and heavy flavour production – p.20

Shape comparisonsShapes of distributions are important in order to either:⋆ make kinematic cuts to reduce the W + jet backgrounds; or,⋆ to fit components of signal and background.In particular, are the shapes of relevant distributions similar in theb-tagged and untagged samples? Is this only true at LO?W + jets and heavy flavour production – p.21

Kinematic distributionsNLO behaviour may provide clues to processes with more jets (→relevant for t¯t), especially for more inclusive variables such as∑ET (jet) and H T = ∑ event E T .W b¯b shape is relatively unchanged at NLO, compared toW + 2 jets.W + jets and heavy flavour production – p.22

NLO predictionsAt NLO, there is a change of shape in the H T distribution.Lowest orderNLO inclusiveLowest order+jetNLO exclusiveThis change is not entirely due to the extra W + 3 jet eventsallowed in the inclusive sample.The p T distribution of the hardest jet shows no change in shape.W + jets and heavy flavour production – p.23

Extra jet contributionIn the NLO inclusive result, the contribution to the H T distributionfrom W + 3 jet events is negligible at small H T and dominant atlarge H T .Similar ratio for W b¯bj to W b¯b.Extra jet contribution to the jet p T distribution is never dominantover this range.W + jets and heavy flavour production – p.24

PDF uncertaintiesTotal cross-section uncertainty:W b¯b → 2.5% , W + 2j → 1.5%.Uncertainty in the (W b¯b/W + 2 jet) ratio:W + jets and heavy flavour production – p.26

Single-tagged heavy flavourW + jets and heavy flavour production – p.27

¡¡¡¢¡¢ ¦¤§©¨¥£ £¤Heavy flavour fraction revisitedOften the presence of two b-quarks in the final state is actuallyonly inferred from a single b-tagIn this case, there is another way of computing the theoreticalcross-section. For instance, in the case of Z+ heavy flavour:Requires knowledge of b-quark pdf’s, but compare to:Expansion in α s ln(M Z /m b )and NLO calculation difficultW + jets and heavy flavour production – p.28

¡¢¡£¢£¡¡¢¢Z + b at NLO - Run IIJC, K. Ellis, F. Maltoni and S. Willenbrock, hep-ph/0312024p jetT > 15 GeV, |ηjet | < 2σ(Z + one b tag) = 20 pbFakes from Z+ jet eventsare significantPrediction for ratio ofZ + b to untagged Z + jet isgb → Zb44 %0.02 ± 0.004q¯q → Z(b¯b)22 %34 %Z + 1 jet (fake rate of 1%)W + jets and heavy flavour production – p.29

Experimental resultBased on 189 pb −1 of data from Run IIEvents / 10 GeV1510DØ Run II PreliminaryDataBkgd+MCBkgdPreliminary ratio of crosssections:σ(Z+b)σ(Z+j)= 0.024 ± 0.0750 50 100 150 200(GeV)E Tcompatible with the NLOpredictionZ + b process in the nextversion of MCFM will allow amuch better comparison withthe analysisW + jets and heavy flavour production – p.30

LHC expectationsp jetT > 15 GeV, |ηjet | < 2.5σ(Z + one b tag) = 1 nbFakes from Z+ jet eventsare much less significantand q¯q contributionis tinyZ + 1 jet (fake)13 %4 %q¯q → Z(b¯b)This should allow afairly clean measurementof heavy quark PDF’s(currently, only derivedperturbatively)83 %gb → ZbW + jets and heavy flavour production – p.31

£¡¢§¤ ¥¦¨¨ © b-PDF usessingle-top qb → qtWsingle-top gb → tW(charged) Higgs+binclusive HiggsW + jets and heavy flavour production – p.32

Calculational directionsW + 3,4 jet cross-sections at NLO⋆ New technology needed: probably not ready for Run IINagy and Soper, hep-ph/0308127Giele and Glover, hep-ph/0402152Inclusion of b mass effects in W b¯b and Zb¯b⋆ Technology available: some efforts are underway . . . c.f. Hb¯bW. Beenakker et al., hep-ph/0211352S. Dawson et al., hep-ph/0311216Merging of existing NLO calculations with a parton shower⋆ Possible: MC@NLO has yet to be applied to W/Z+ jetsFurther study of recent ideas regarding parton showers (mostpromising in the short term)⋆ Matrix elements corrections - CKKW, . . .⋆ How much of the effects of NLO are taken into account bycombining matrix element calculations with parton showers?F. Krauss et al.W + jets and heavy flavour production – p.33

OutlookThe W + jets channel (including heavy quarks) is very importantfor many studies in Run II.Unfortunately, for events with many jets we are limited to LOpredictions for rates and distributions.However, there should be lots to learn from the NLO correctionsthat we know about. The highest multiplicity that is currentlyavailable is production of W b¯b and W + 2 jets.Implications for Run II analyses.⋆ Results suggest that some relevant observables do not sufferfrom large NLO effects and we can proceed with moreconfidence in analyses based on LO tools.⋆ However, beware of variables that change shape at NLO (H T ).⋆ These statements are heavily dependent on scale choices.Further comparisons with parton shower approaches and data isthe way forward.W + jets and heavy flavour production – p.34