Heavy Flavour Content of the Proton eminar, 8 th October 2008 Paul - PowerPoint PPT Presentation

Heavy Flavour Content of the Proton eminar, 8 th October 2008 Paul Thompson, Birmingham S • Reminder of HERA and kinematics • Why measure proton structure (PDFs)? • Why measure heavy flavours? • Experimental Techniques • charm and beauty cross sections at HERA • Outlook 1

HERA ep collider • Q 2 corresponds to the ± e • Collided protons with 92-07 spatial resolution of probe • E e : 27.5 GeV λ 2 2 5 2 • ~ 1 / Q Q ~ 10 GeV max • E p : 920 GeV − λ 18 ~ 10 m ~ R / 1000 • min proton = • Centre of mass energy : s 320 GeV 2 Deep Inelast ic S cat t ering - DIS

Available Data • In total ~500pb -1 of high energy data collected per experiment • luminosity upgrade in 2001 • detectors adj usted to accommodate focussing magnets • Low energy running to measure F L Many preliminary analyses on full HERA II dat a Working on final publicat ion and combinat ion of result s 3

Deep Inelastic S cattering (DIS ) e ± ( k ' ) Neut ral Current Event e ± ( k ) Q 2 = -q 2 =-(k-k ΄ ) 2 γ , Ζ 0 (xP) p(P) } Proton remnant Current jet • DIS cross section can be described in terms of: • Q 2 : Virtuality of the intermediate boson Q = 2 s x y • x : Bj orken scaling factor -fraction of proton ’ s momentum carried by struck quark • y : Inelasticity -energy fraction transferred from lepton in proton rest 4 frame

Neutral Current Cross S ection F 2 ± σ π 2 2 2 ( ) 2 d e p a y = − ± 2 2 2 Y [ F ( x , Q ) F ( x , Q ) xF ( x , Q )] + 2 L 3 2 4 dxdQ xQ Y + F 2 – dominant contribution to the cross section = ∑ + 2 F e x (q q ) 2 q q S caling violations indicate presence of gluons Data evolution with Q 2 (at fixed x ) described by perturbative QCD 5

QCD Factorisation and Proton PDF DGLAP F 2 ( x,Q 2 ) = Σ f k ( μ ) x C k j (Q,m, α s ( μ )) k 2 and f k are parton density functions – parameterised at Q 0 evolved to high Q 2 using DGLAP equations j perturbative coefficient functions C k 6

PDFs for the LHC proton proton M x 1 P x 2 P momentum fractions x 1 and x 2 determined by mass and rapidity of X DGLAP evolution x dependence of f(x,Q 2 ) determined by fit to data, Q 2 dependence determined by DGLAP equations full NNLO DGLAP now known*, also with small x, QED etc improvements 7 *Moch, Vermaseren, Vogt (2004)

Why do we need PDFs? high precision (S M and BS M) cross section predictions require • precision pdfs: δσ th = δσ pdf + … improved signal and background predictions → easier to spot • new physics deviations • ‘ standard candle ’ processes (e.g. σ (Z) ) to check formalism (factorisation, DGLAP, … ) • measure machine luminosity? • • learning more about pdfs from LHC measurements. e.g. high-E T j ets → gluon? • W + ,W – ,Z 0 → quarks? • forward DY → small x ? • • … 8

How Important Is PDF Precision? • Example 1: σ (M H =120 GeV) @ LHC δ σ pdf ≈ ± 3%, δ σ ptNNL0 ≈ ± 10% δ σ ptNNLL ≈ ± 8% → δ σ theory ≈ ± 9% • Example 2: σ (Z 0 ) @ LHC δ σ pdf ≈ ± 3%, δ σ ptNNL0 ≈ ± 2% → δ σ theory ≈ ± 4% Catani et al, 9 hep-ph/0306211

Production of Heavy Quarks at HERA Predominantly via boson gluon , c fusion Test of perturbative QCD: multi-scale problem ( Q 2 , m b 2 , p t 2 ) , c Directly sensitive to gluon density in the proton (PDFs) 10

cc and F 2 bb F 2 At HERA we can measure the contribution of c and b to the total DIS cross bb and F 2 cc section F 2 bb measurements at high Q 2 important for LHC e.g. bb->H F 2 11

Predictions for Heavy Quark Production Fixed Flavour Number Scheme (FFNS) General Mass VFNS (GM VFNS) Zero Mass Variable Number Scheme (ZM VFNS) CTEQ6.5 uses a General Mass scheme changed from a massless in CTEQ6.1 TW improved their General Mass scheme from MRS MRS T2004 to MS TW2006 12 Thorne, Tung arXiv:0809.0714, P.T. hep-ph/ 0703103

Impact on W, Z @ LHC • Correct heavy flavour treatment affects light partons! • changes in CTEQ 6.1 -> CTEQ 6.6 due to c,b,s treatment • Improved agreement between latest PDFs 13

Heavy Quark contribution to DIS cross section HERA I result: • fraction of total DIS cross section from charm and beauty • large charm fraction(~30% ) • small beauty fraction ~% (lower at low Q 2 ) • mass thresholds visible • reasonable description by pQCD 14

Flavour Tagging - Vertex Detectors H1 (MVD) C entral S ilicon T racker (30 o < θ < 150 o ) • Double layer double sided strips • Precise determination of impact parameter in transverse plane • Resolution of | δ | for hits in both layers; μ 90 m μ ⊕ 33 [ ] m GeV P T Installed for HERA II 15 Installed 1997 (first pub 2004)!

Tagging Heavy Quarks ( b ) Beauty quarks rarely produced, use properties of beauty hadrons: • semileptonic decays( μ , e ) • mass rel relative to - transverse momentum p t j et axis • lifetime (vertex detectors) - reconstrucion of a secondary vertex - impact parameter δ 16

igned Impact Parameter δ S igned impact parameter δ ignificance = δ/σ(δ) S , S Charm and beauty asymmetric (positive) due to lifetime Light flavours mostly symmetric (resolution dominates) S imilarly for secondary vertices (>=2 tracks), decay length L and decay length significance =L/ σ (L) 17

Tagging Heavy Quarks ( c ) resonances D*, D + , D 0 ,… Full HERA II statistics (~350pb -1 ) resonances and decay length tagging using vertex detectors 18

D* Cross S ection H1 prelim-08-072 H1 prelim-08-074 • good description by NLO calculation ) in wide Q 2 (HVQDIS range • Also at large Q 2 , where massive approach not expected to be appropriate 19

D* Cross S ection D * • differential cross sections of several D mesons measured • reasonably described by NLO QCD (HVQDIS ) • double differential in x and Q 2 cc allows extraction of F 2 H1 prelim-08-072 D 0 D + 20 ZEUS -prel-07-034 ZEUS -prel-07-009

21 D* Fragmentation

D* Fragmentation • RAPGAP MC: p T,j et > 3 GeV, parameters consistent with e + e - • no j et sample (low photon gluon COM) needs harder frag. 22 • S imilar story for NLO QCD DES Y-08-080 (Juraj Bracinik)

Charm and Beauty Cross S ection miss ||μ , p T rel and • combine p T impact parameter distributions • use 3D fit to decompose into beauty, charm and light flavour • Q 2 > 20 GeV 2 , 0.01< y < 0.7, P T μ > 1.5 GeV, -1.6 < η μ < 2.3 • c and b cross sections ZEUS -prel-08-007 23 described by NLO QCD(HVQDIS )

Charm and Beauty Cross S ection • beauty tends to be above NLO QCD at low Q 2 • may be measured double differentially in x, Q 2 and cc , F 2 bb extrapolated to full phase space to compare F 2 24

H1 Inclusive Analysis H1 prelim-08-173 • Publication on HERA I data (54 pb -1 ) in 2004 & 2005 • H1 CS T rebuilt to account for HERA II beamline • Preliminary analysis on full HERA II data (190pb -1 ) this summer (H1prelim-08-173) • Inclusive analysis: use all tracks with hits in silicon detector ( p t > 0.3 GeV ) • Precise determination of impact parameter in transverse plane • Divide events into 1 track, 2 track and >= 3 track samples 25

S igned Impact Parameter (H1) H1 prelim-08-173 lifetime resolution Charm and beauty asymmetric due to lifetime, Light flavours mostly symmetric MC describes resolution! 26

S ignifcance 2 nd highest significance for N track =2 S ignificance for N track =1 27

Neural Network • Improve c,b separation power (especially at low Q 2 ): use neural network for >= 3 track events • Choose inputs which are different for c and b, and largely physics model independent L , track p t , 2 nd highest track p t , number of • Inputs: S 1 , S 2 , S 3 , S CS T tracks, number of tracks associated to secondary vertex • Network trained with b as “ signal” c as “ background” . Light flavours will be subtracted out due to their symmetry (see later) 28

29 Neural Network Inputs

30 Neural Network Input

31 Neural Network Input (Neg. subtracted) + -

Neural Network Output • S ign given by S 1 . S ubtract – ’ ve from +’ ve to reduce systematic error due to resolution and light contribution 32

Extracting Flavour Fractions These distributions are fitted for ρ c , ρ b in each x,Q 2 bin with ρ uds constrained by total number of DIS events ρ c .N c gen f c = ρ c .N c gen + ρ b N b gen + ρ uds .N uds gen 33

Inclusive b cross section (H1) • HERA I agrees with HERA I • HERA II reaches lower Q 2 (NN) • HERA I and HERA II data combined for improved precision 34

Inclusive b Cross S ection (HERA) • comparison of different methods [acceptance] − Inclusive (H1VTX) [>90% ] − μ p t 03/ 04 μ ) rel (ZEUS [20-35% ] − μ p t rel + δ (ZEUS 05 μ ) [25-50% ] • ZEUS tend to be higher than H1 • generally described by NLO QCD (FFNS , GM-VFNS ) 35

Improvements in Theory tatus summer 2007 (e - data) S • MRS T04 factor 2 larger than CTEQ at Q 2 =12 GeV 2 • Chance to distinguish models with full HERA II data • S ince then MS TW08 was released which is in much better agreement with CTEQ (and data)! 36