F c measurements at HERA Katerina Lipka New trends in HERA Physics, Ringberg 8
Charm production at HERA: why now? HERA I : PDF central measurement of HERA HERA I F c / F Q = GeV PDF obtained from the fits to inclusive F Inclusive F experimentally very precise Contribution of events with charm to F high Measurement of F c has large uncertainties.4. now: precise PDF crucial importance for the LHC Combined HERA PDF are of unprecedented precision BUT dependent on parameterization of the QCD fit Need a cross check / direct access to the gluon -5-3 x Final state measurements (jets, heavy quarks) extremely important F c @ HERA II on the way to precision measurement K. Lipka Charm production and F c at HERA
Charm production at HERA Dominated by Boson Gluon Fusion (BGF) e γ c c fragmentation hadrons D* gluon directly involved: include in a global PDF fit important cross-check of the g(x g ) charm mass additional hard scale: pqcd calculations possible; multiple scales: calculations complicated Factorization: σ(ep D*X)=Proton Structure Photon Structure Matrix Element Fragmentation to learn something about PDFs: calculate hard ME, measure cross section, understand fragmentation K. Lipka Charm production and F c at HERA 3
Presented in this talk Hard ME: NLO calculations and Monte-Carlo simulations Charm tag methods and extraction of F c : Charm tag via reconstruction of charmed mesons Extrapolation to the full phase space Fragmentation measurement Charm tag via track displacement measurement Results and discussion K. Lipka Charm production and F c at HERA 4
Models of charm production Massive calculation, fixed order QCD calculation, FFNS correct threshold suppression, no collinear divergences, terms ~log(μ/m) no factorization, no conceptual necessity fo FFs, no resummation valid for p t m c, fixed order logarithms ln(p t /m c ) large for p t >>m c Models for charm at HERA: FMNR (Photoproduction), HVQDIS (DIS), Massless calculation (ZM-VFNS) large collinear ln (μ /m c )-terms resummed in evolved PDFs and FFs (LL, NLL), good for large μ ~p t >>m c universality of PDFs and FFs via factorization theorem, global analysis terms (m c /p t ) n neglected in the hard part breaks down @ threshold Not appropriate for charm production at HERA (close to threshold) Generalized mass calculation (GM-VFNS) Hubert s talk Available for charm production in γp at HERA, DIS is on the way K. Lipka Charm production and F c at HERA 5
Monte-Carlo models for data corrections RAPGAP matrix element calculated in LO QCD higher order contributions via parton showers parton evolution in collinear approximation (DGLAP equations) charm is massive in BGF CASCADE gluon density unintegrated in gluon transverse momentum k T only gluons in proton higher order contributions via initial state parton showers based on CCFM equations charm is massive in BGF Hadronization via Lund String model (Jetset) K. Lipka Charm production and F c at HERA 6
Charm tag via D* ± production Kinematics regimes: Liquid Ar Calorimeter π Spaghetti Calorimeter DIS: 5 <Q < GeV Photoproduction Q < GeV e D* + D π s+ K - π + π s+ (+ c.c.) K central tracking detector π s Entries /.5 MeV H Preliminary HERA II + 6 Κ π ± π slow ± N(D*) = 83 ± 8 4 fit Electron reconstructed in SpaCal: Q < GeV LAr: Q > GeV 5 < Q < GeV. < y <.7 η (D*) <.5 p (D*) >.5 GeV T.4.5.6.7 M(Kππ) - M(Kπ) [GeV] K. Lipka Charm production and F c at HERA 7
D* production in DIS (5<Q < GeV ) [nb/gev] d σ / dη dp [nb/gev] d σ / dη dp.5 T.5 T D* in DIS H Preliminary HERA II 5 < Q < GeV. < y <.7.5 < p (D*) <.5 GeV T -.5 - -.5.5.5 η(d*).4.3. D* in DIS. 3.5 < p (D*) < 5.5 GeV T -.5 - -.5.5.5 η(d*).3 < m c <.6 GeV μ = Q + 4m c < μ /μ < 4 f,r α(kartvelishvili) = 3.3 ±.4 [nb/gev] d σ / dη dp [nb/gev] d σ / dη dp T.5.3. T D* in DIS H data (prel.) HVQDIS (MRST4FF3nlo) HVQDIS (CTEQ5f3).5 < p (D*) < 3.5 GeV T -.5 - -.5.5.5 η(d*) D* in DIS. 5.5 < p (D*) < 4. GeV T -.5 - -.5.5.5 η(d*) dσ/dη(d*) (nb) 3.5 3.5.5.5 ZEUS ep e+d*+x - ZEUS (prel.) 6 pb HVQDIS -.5 - -.5.5.5 η(d*) H: FFNs NLO does good job describing the D* kinematics, but underestimates forward region ay low p T (D*) ZEUS: no forward access in the data seen. Data precision will improve K. Lipka Charm production and F c at HERA 8
D* production in DIS (Q > GeV ) dσ/dη [pb] 5 D* production at high Q H Preliminary HERA II [pb] dσ/dp T D* production at high Q H Preliminary HERA II H data (prel.) RAPGAP (CTEQ65m) CASCADE (A) HVQDIS (MRST4FF3nlo) 5 < Q < GeV. < y <.7 p (D*) >.5 GeV T H data (prel.) RAPGAP (CTEQ65m) CASCADE (A) HVQDIS (MRST4FF3nlo) - η (D*) < Q < GeV. < y <.7 η (D*) <.5 5 5 p (D*) [GeV/c] T D* production at high Q : description by the LO Monte-Carlo gets worse NLO FFNs describes data very well K. Lipka Charm production and F c at HERA 9
D* production in DIS: Q slope D* production in DIS ] [pb/gev dσ /dq 3 - - H Preliminary HERA II. < y <.7 η (D*) <.5 p (D*) >.5 GeV T H data (prel.) RAPGAP (CTEQ65m) RAPGAP (CTEQ6l) CASCADE (A) Q [GeV ] 3 Monte-Carlo models don t describe the Q slope of the D* cross section K. Lipka Charm production and F c at HERA
D* production in DIS ] [pb/gev dσ /dq 3 - - D* production in DIS H Preliminary HERA II. < y <.7 η (D*) <.5 p (D*) >.5 GeV T H data (prel.) HVQDIS (MRST4FF3nlo) Q [GeV ] 3 dσ/dq (nb/gev ) - - -3-4 -5 ZEUS ZEUS (prel.) 6 pb - ZEUS BPC (prel.) 98- ZEUS 98- HVQDIS - 3 Q (GeV ) NLO FFNs calculation does good job (surprising, should break down for high Q ) K. Lipka Charm production and F c at HERA
) ( nb / GeV dσ/dq - - -3 More to charmed meson production HERA-II, L=35pb - 5<Q < GeV, p T (D)>3 GeV, Iη(D)I<.6 Lifetime information from the ZEUS Micro Vertex Detector used NLO FFNS describes data well D ZEUS ep e + D + X - ZEUS 65 pb HERA I - ZEUS (prel.) 35 pb HERA II NLO + Fragmentation (nb) D dσ/dη (nb) D ± dσ/dη 3.5.5.5.4..8.6.4. ZEUS ep e + D + X - ZEUS (prel.) 35 pb HERA II NLO + Fragmentation beauty contribution (RAPGAP) -.5 - -.5.5.5 D D + ZEUS η D ± ep e + D + X - ZEUS (prel.) 35 pb HERA II NLO + Fragmentation beauty contribution (RAPGAP) -4 Q (GeV ) 3 -.5 - -.5.5.5 η D ± K. Lipka Charm production and F c at HERA
Extraction of F c from meson cross section cc σ vis (exp) cc F (exp) = F ( theory) σ ( theory) vis Visible cross section: p T (D*)>.5 GeV, η(d*) <.5.<y<.7, 5<Q < GeV Problem: detector sees only 3% of the phase space for c D* strong model dependence due to large extrapolation factors Extrapolation problems: ) Different extrapolation models ) Unknown parameters within a single model: mass of charm quark, scales, fragmentation model experimentally measurable: see next slides K. Lipka Charm production and F c at HERA 3
Measurement of charm fragmentation in ep Methods to reconstruct the energy of the parent quark: Jet containing D*; problem: only small region of phase space accessible DIS: inclusive k algorithm applied in the γp frame, E T (D*-jet)>3 GeV significant contribution only from ŝ > GeV² (much above threshold) Hemisphere containing D*: experimental setup similar to e + e -, works at threshold, ŝ 4m² c γp-system cut:η> - take all particles towards γ - project onto the plane to the γ - get Thrust axis - Σ momenta of all particles in the D* hemisphere γdirection K. Lipka Charm production and F c at HERA 4
Charm fragmentation in photoproduction Parent quark approximated by a jet containing D* data: HERAI, L = pb -, Q <GeV ; p T ( D*) > GeV, Iη(D )I<.5 inclusive k algorithm, E T (D*-jet)>9 GeV compared to the NLO FFNS (FMNR) fragmentation model: Kartvelishvili /σdσ/dz z = + E ( E PII ) D* ZEUS ( pb - ) jet ZEUS FMNR C PYT had (Kartvelishvili α=.67+.5 -.3 ) FMNR C PYT had (Kartvelishvili α=.) FMNR C PYT had (Kartvelishvili α=4.) z D D c * ( z) z α ( z) best fit α=.67..4.6.8 z K. Lipka Charm production and F c at HERA 5
Measurement of charm fragmentation in DIS ( E + p Data: H HERA I, L=75 pb - L ) D* both methods used: z jet =, ( E + p) jet Differences between the methods: Hemisphere method should include more final state gluon radiation than jet method Measured distributions of the fragmentation variable should be different Extracted parameters of the non-perturbative fragmentation function should agree z hem = ( E + pl ) D* ( E + p) hem Measure in the common phase space: require presence of a D* jet Extract parameters for n.-p. FF using MC/ HVQDIS. Expect : parameters agree for different methods, different for MCs and HVQDIS Any differences at the threshold (absence of a D*-jet)? K. Lipka Charm production and F c at HERA 6
Fragmentation measurement: D*- Jet sample Hemisphere method RAPGAP MC: Jet method NLO (HVQDIS): Hemisphere method Distributions on hadron level look different (as expected) MC (Rapgap) with standard n.p. FF yield reasonable description of data Extracted n.p. FF parameters from z hem (α=4.5±.6) and z jet (α=4.3±.4) agree HVQDIS Kartvelishvili fragmentation: extracted parameters z hemi and z jet agree K. Lipka Charm production and F c at HERA 7
Fragmentation c D*. No D*- jet sample MC: Extracted parameters (α=.3 ) inconsistent with jet-sample (α=4.5±.6) HVQDIS: no jet (α=6. +. -.8 +.7 -.6 ) inconsistent with jet-sample (α=3.3±.4) Fragmentation at threshold significantly harder than expected from the jet-sample Treatment in extrapolation models: ŝ-dependent fragmentation K. Lipka Charm production and F c at HERA 8
Back to extrapolation of σ(d*) to the F c Extrapolation Models in use: NLO: Riemersma et al: integrated form; HVQDIS: differential form, fixed order massive calculation, Nf=3, FFNS, evolution: DGLAP Parameters: PDFs: MRST4F3, m c =.43 GeV, μ r =μ f =μ= Q +4m c Fragmentation: ŝ<7 GeV : α =6., otherwise α=3.3 CASCADE: massive LO ME + Parton showers, proton structure: gluons only, evolution: CCFM Parameters: PDFs: A, m c =.43 GeV, μ r =μ f =μ= Q +4m c Fragmentation: ŝ <7 GeV ; α = 8., otherwise α = 4.3 K. Lipka Charm production and F c at HERA 9
D* cross sections vs NLO/CASCADE ep e D* X ] ep ed*x Q 4 d σ/dq dy [nb GeV ] Q = 7 GeV GeV 8GeV 3GeV 65GeV GeV [nb GeV σ/dydq d 4 Q Q = 7 GeV 3 GeV GeV 65 GeV 8 GeV GeV GeV - - - - 44 GeV H Preliminary y Data HERA-II HVQDIS (MRST4FF3) total theory uncertainty fragmentation uncertainty - GeV - - 44 GeV - y H Preliminary Data HERA II CASCADE (A) Lowest y (highest x) overestimated by NLO, underestimated by CASCADE K. Lipka Charm production and F c at HERA
Extrapolation factors NLO/CASCADE Extrapolation Factor Ratio CASCADE/HVQDIS 3 Q = 7 GeV GeV 8 GeV 3 GeV 65 GeV GeV GeV 44 GeV -5-3 -5-3 - x Extrapolation factors (σ tot /σ vis ) differ in NLO vs CASCADE: 3%-% (low x) -% (high x) Differences in the models: LO+PS vs NLO Evolution Hadronization Possible reason: Hadronization More studies have to be done K. Lipka Charm production and F c at HERA
Results: F c from D* measurement F cc_ in the NLO DGLAP scheme (HVQDIS) cc F.6 Q = 7 GeV GeV 8 GeV.4 cc F.6.4 cc F in the CCFM scheme (CASCADE) Q = 7 GeV GeV 8 GeV...4 3 GeV 65 GeV GeV.4 3 GeV 65 GeV GeV...4. GeV 44 GeV -5-3 -5-3 - H Preliminary Data HERA-II CTEQ5F3 MRST4FF3 total theory uncertainty PDF variation.4. -5 GeV -3-5 44 GeV -3 x - H Preliminary D* HERA II CASCADE (A) Graph Graph theory Graph uncertainty:.3 < m c <.6 GeV.5μ < μ < μ r x Experimental errors will further decrease we are on the way to the final precision K. Lipka Charm production and F c at HERA
Charm/beauty via other tag methods μ p t rel B + B - Jet δ Muons 3 Charm comes alone with beauty: - More experimental details in Massimo s talk Large mass : transverse momentum to Jet axis: muon p t rel Large lifetime: impact parameter δ, Significance S=δ/σ(δ) Semileptonic decays (e,μ) Different systematic uncertainties wrt. D-meson measurements ZEUS (prel.) HERA II 5 pb - MC sum c b uds Tracks 7 6 cc bb Measurement of F and F H Data (Prel.) H Preliminary Total MC uds c b Jet 5 4 -. -.5 -. -.5.5..5. δ / cm K. Lipka Charm production p rel T (GeV) and F c at HERA 3 3
Charm in semileptonic events ZEUS ZEUS dσ/dη μ (pb) 8 ZEUS (prel.) HERA II 5 pb - Charm Beauty HVQDIS ep eqq X e μ X _ cc F.6.5.4 Q =3 GeV _ cc F.6.5.4 Q =3 GeV 6.3.3 4.... -.5 - -.5.5.5 Reasonable description by the NLO FFNS η μ _ cc F.6.5.4.3.. -3 - - x Q = GeV -3 - - x Charm -3 - - x ZEUS (prel.) HERA II, 5 pb - (μ) ZEUS HERA I, 8 pb - (D*) H HERA I, 57 pb - (VTX) ZEUS-S-FF m c =.5, m b =4.75 GeV PDF uncertainty m c =.3, m b =4.5 GeV m c =.7, m b =5. GeV K. Lipka Charm production and F c at HERA 4
Charm cross sections via lifetime tag vs D* σ cc _.4 H CHARM CROSS SECTION IN DIS Q = 6.5 GeV Q = GeV Q = GeV..4 Q = 35 GeV Q = 6 GeV Q = GeV..4. Q = GeV Q = 4 GeV -4-3 - x H HERA I H HERA II (Prel.) H D HERA II (Prel.) H Preliminary -4-3 - x -4-3 - x K. Lipka Charm production and F c at HERA 5
Charm cross sections vs VFNS (TR) σ cc _.4 H CHARM CROSS SECTION IN DIS Q = 6.5 GeV Q = GeV Q = GeV..4 Q = 35 GeV Q = 6 GeV Q = GeV..4. Q = GeV H Preliminary -4-3 - Q = 4 GeV -4-3 - x H HERA I H HERA II (Prel.) CTEQ6.6 MSTW8 (Prel.) CCFM x -4-3 - K. Lipka Charm production and F c at HERA 6 x
_ F cc 4 i 5 4 3 F c from lifetime tag vs MSTW NNLO H F cc _ (x,q ) x=., i= x=.3, i=9 x=.5, i=8 x=.8, i=7 x=.3, i=6 Recall Robert s talk on Monday: H Preliminary NNLO better fits the measured F c Only Lifetime data of H are shown Data will get better! x=., i=5 x=.3, i=4 x=.5 i=3 - H HERA II (Prel.) MSTW8 (Prel.) MSTW8 NNLO (Prel.) x=.8 i= x=.3 i= x=. i= 3 K. Lipka Charm production Q / GeV and F c at HERA 7
HERA measurement of F c HERA F cc F cc 9 8 7 6 5 4 3 - x =.3 ( 4 ) x =.5 ( 4 9 ) x =.7 ( 4 8 ) x =.3 ( 4 7 ) x =.8 ( 4 6 ) NLO QCD: CTEQ5F3 MRST4FF3 H HERA I (D*) H HERA I (VTX) ZEUS HERA I (D*) ZEUS HERA I (D +, D, D s + ) ZEUS (prel.) HERA II: D* D + μ x =.3 ( 4 5 ) x =.35 ( 4 4 ) x =.5 ( 4 3 ) x =.6 ( 4 ) x =.8 ( 4 ) x =. ( 4 ) x =. ( 4 9 ) x =.5 ( 4 8 ) H (prel.) HERA II: D* VTX x =. ( 4 7 ) x =.3 ( 4 6 ) x =.4 ( 4 5 ) x =.6 ( 4 4 ) x =.8 ( 4 3 ) x =. ( 4 ) x =. ( 4 ) x =.3 ( 4 ) Plenty of measurements Nice agreement between methods Experimental precision of several measurements will further improve Different methods will be combined: orthogonal errors! H and ZEUS results will be combined Sensitivity to the models (PDFs) Need proper (precise) theory 3 K. Lipka Charm Q production (GeV ) and F c at HERA 8
Conclusions Experimentally charm measurements at HERA get very precise: We will get soon D* measurement at H on the way to 5% precision measurement D* measurements at ZEUS full HERA-II on the way Combination of different measurement methods Combination of H and ZEUS: cross calibration Enlarge the visible phase space (H) Extrapolation to the full phase space model dependent Theory: massive NLO pqcd describes the data quite well Model uncertainties larger than experimental errors We still need: GMVFNS for DIS : proper charm treatment in the global fit NNLO K. Lipka Charm production and F c at HERA 9
Outlook: Precision will improve Results will be combined F cc.4. Q = GeV H HERA I: D* VTX ZEUS HERA II: D* D +, D, D s + H (prel.) HERA II: D* VTX ZEUS (prel.) HERA II: D* D + μ HERA F cc 4 GeV 7 GeV NLO QCD: CTEQ5F3 MRST4FF3 GeV 8 GeV 3 GeV HERA I F c / F Q = GeV.4..4 6 GeV 3 GeV 5 GeV..4. -5-3 x -5-3 -5-3 -5-3 - x
Backup
t D* in photoproduction (Q < GeV ) dσ(ep ed*x)/dη[nb] 4 3 H Preliminary HERA II D* in Photoproduction H data (prel.) FFNS (CTEQ5F3) <.7 GeV.3 < m c.5 < μ / m c +p < r,f t GMVFNS (CTEQ6.5) [nb/gev] dσ(ep ed*x)/dp H Preliminary HERA II D* in Photoproduction H data (prel.) FFNS (CTEQ5F3) <.7 GeV.3 < m c.5 < μ / m c +p < r,f t GMVFNS (CTEQ6.5) -.5 - -.5.5.5 η(d*) - - Q < GeV < W γ P < 85 GeV p (D*) >.8 GeV t η(d*) <.5 4 6 8 p (D*)[GeV] t [nb/gev] dσ(ep ed*x)/dw D* in Photoproduction Measurement in the visible range:.8 H Preliminary H data (prel.) HERA II FFNS (CTEQ5F3).3 < m c <.7 GeV P.6.5 < μ / m T (D*)>.8 GeV, Iη(D*)I<.5 c +p < r,f t GMVFNS (CTEQ6.5) p γ.4. 5 5 W γ p [GeV] Q < GeV,.<y<.8 Good agreement with both NLO Large model uncertainties due to variation of the scales K. Lipka Charm production and F c at HERA 8
D* in photoproduction dη (D*)[nb/GeV] d σ/dp t 3 H Preliminary HERA II D* in Photoproduction.8 < p (D*) <.5 GeV t dη (D*)[nb/GeV] 8 6 d σ/dp t H Preliminary HERA II D* in Photoproduction.5 < p (D*) < 4.5 GeV t 4 dη (D*)[nb/GeV] d σ/dp t -.5 - -.5.5.5 η(d*).5.4.3.. H Preliminary HERA II D* in Photoproduction 4.5 < p (D*) <.5 GeV t -.5 - -.5.5.5 η(d*) H data (prel.) FFNS (CTEQ5F3).3 < m c <.7 GeV.5 < μ / m c +p < r,f t GMVFNS (CTEQ6.5) -.5 - -.5.5.5 η(d*) GM VFNS underestimates the forward region at high p T K. Lipka Charm production and F c at HERA 9