Models of the Nucleon & Parton Distribution Functions

Transcription

1 11th CTEQ Summer School on QCD Analysis and Phenomenology Madison, Wisconsin, June 22-30, 2004 Models of the Nucleon & Parton Distribution Functions Wally Melnitchouk Jefferson Lab

2 Outline Introduction Phenomenological PDFs - large x - valence quarks, nuclear EMC effect - small x - sea quarks, nuclear shadowing Connection with low energy models PDFs from lattice QCD - moments - x dependence Outlook

3 1. Introduction

4 Looking for quarks in the nucleon is like looking for the Mafia in Sicily - everybody knows they re there, but it s hard to find the evidence! Anonymous

5

6 Why is (accurate) knowledge of PDFs important? PDFs provide basic information on structure of bound states in QCD - DIS paved way for development of QCD Integrals of PDFs (moments) test fundamental sum rules (Adler, GLS, Bjorken,...) Provide input into nuclear physics (relativistic heavy ion collisions) and astrophysics calculations Needed to understand backgrounds in searches for new physics in high-energy colliders

7 F (x, Q 2 ) = i Factorization of structure functions dz C i (x/z, Q 2 /µ 2, α s (µ 2 )) f i (z, α s (µ 2 )) Wilson coefficient PDF PDFs embody nonperturbative (long-range) structure of nucleon - cannot calculate from pqcd Calculated from low-energy models - evolved to larger Q via DGLAP Computed on the lattice - moments only, reconstruct PDF

8 2. Phenomenological PDFs

9 Phenomenological PDFs PDFs extracted in global pqcd (NLO) analyses of data from e.m. & neutrino DIS, Drell-Yan & W-boson production in hadronic collisions... Parameterized using some functional form, e.g. xf(x, µ 2 ) = A 0 x A 1 (1 x) A 2 e A 3x (1 + e A 4 x) A 5 determined over several orders of magnitude in x and Q < x < 1 1 < Q 2 < 10 8 GeV 2

10 Recent global PDFs (CTEQ6) Lai et al., Eur. Phys. J. C12 (2000) 375

11 Comparison with data Lai et al., Eur. Phys. J. C12 (2000) 375

12 Lai et al., Eur. Phys. J. C12 (2000) 375 The χ 2 per data point for these data sets are 1.11 (378/339) for BCDMS and 1.52 (305/201) for NMC. The fit to the BCDMS data is clearly excellent, both by inspection of Fig. 4a and by the 2

13 small x - sea quarks and gluons dominate structure of hadron or structure of probe? Botje, Eur. Phys. J. C 14 (2000) 285

14 Most direct connection between PDFs and models of the nucleon is through valence quarks x f (x,q) uv dv! u d! s g/15 Q 2 = 25 GeV 2 large x - valence quarks dominate PDFs at large x are small x small uncertainties can have large relative effects

15 Botje, Eur. Phys. J. C 14 (2000) 285 Uncertainty due to nuclear effects in neutron

16 Nuclear effects Need proton and neutron to resolve u and d flavors No free neutron targets (neutron half-life ~ 12 mins) - use deuteron as effective neutron target However F2 d F p 2 + F 2 n!! Nuclear effects obscure neutron structure information - nuclear binding + Fermi motion at large x - nuclear shadowing at small x - antishadowing? pion cloud? at intermediate x

17 Nuclear EMC effect F A 2 (x, Q 2 ) AF N 2 (x, Q 2 ) Original EMC data Later SLAC data Aubert et al., Phys. Lett. B 123, 123 (1983) Gomez et al., Phys. Rev. D 49, 4348 (1994) Figure 6. Cross-section ratios compared with deuterium for SLAC data from Gomez et al (1994) ( ) and Stein et al (1975) updated by Rock and Bosted (2001) ( ) for Be, Al, Fe and Au. Data

18 EMC effect in d at large x quark q off-shell nucleon k p k+q k p q Nuclear impulse approximation incoherent scattering from individual nucleons in deuteron P deuteron P spectator nucleon F2 d (x, Q 2 ) = dy f N/d (y) F2 N (x/y, Q 2 ) + δ (off) F2 d (x, Q 2 ) nucleon momentum distribution off-shell correction

19 Nuclear physics in the deuteron Nucleon momentum distribution in deuteron relativistic dnn vertex function f N/d (y) = 1 4 M d y p 2 max dp 2 E p p 0 Ψ d ( p 2 ) 2 momentum fraction of deuteron carried by nucleon

20 Nuclear physics in the deuteron Nucleon momentum distribution in deuteron relativistic dnn vertex function f N/d (y) = 1 4 M d y p 2 max dp 2 E p p 0 Ψ d ( p 2 ) 2

21 Nuclear physics in the deuteron Nucleon momentum distribution in deuteron relativistic dnn vertex function f N/d (y) = 1 4 M d y p 2 max dp 2 E p p 0 Ψ d ( p 2 ) 2

22 Nuclear physics in the deuteron Nucleon momentum distribution in deuteron relativistic dnn vertex function f N/d (y) = 1 4 M d y p 2 max dp 2 E p p 0 Ψ d ( p 2 ) 2 Wave function dependence only at large y sensitive to large p components of wave function not very well known

23 Nuclear physics in the deuteron Nucleon momentum distribution in deuteron relativistic dnn vertex function f N/d (y) = 1 4 M d y p 2 max dp 2 E p p 0 Ψ d ( p 2 ) 2 Nucleon off-shell correction δ (off) F d 2 kinematical: OK dynamical:??

24 An illustration of possible effects... WM, Schreiber, Thomas, Phys. Rev. D49 (1994) 1183 Note: ratio depends on input F n 2

25 1.2 d N 2 2 F / F convolution with AKL off!shell with MST off!shell MST (conv) MST (conv+off) x 100% error assigned AKL: Alekhin, Kulagin, Liuti, Phys. Rev. D 69 (2004) MST: WM, Schreiber, Thomas, Phys. Lett. B 335 (1994) 11

26 WM, Thomas, Phys. Lett. B 377 (1996) 11 with binding & off-shell SU(6) helicity retention Fermi motion only scalar diquarks Without EMC effect in d underestimated at large x! F n 2

27 At large x (sea quarks, gluons suppressed) F p u d F n u d d u 4 F n 2 /F p 2 4F n 2 /F p 2 1

28 νp, νp Does d/u 0 as x 1?

29 pqcd : d/u constant as x 1

30 pqcd + SU(6) wfn : d/u 1 5 as x 1 Farrar, Jackson, Phys. Rev. Lett. 35 (1975) 1416

31 WM, Peng, Phys. Lett. B400 (1997) 220 d u d u x2 exp( (1 x) 2 )

32 Yang, Bodek, Phys. Rev. Lett. 82 (1999) 2467 d u d u x (1 + x)

33 Cleaner methods of determining d/u e p ν( ν)x ν( ν) p l X p p( p) W ± X e L ( e R ) p e X e p e π ± X e 3 He( 3 H) e X need high luminosity low statistics need large lepton rapidity low count rate need z ~ 1, factorization tritium target

34 Cleaner methods of determining d/u target d recoil p e d e p X backward slow p neutron nearly on-shell minimize rescattering JLab Hall B experiment ( BoNuS )

35 Nuclear shadowing Interference of multiple scattering amplitudes For deuteron:!*!*!* X!* N N N N N N d N (a) d d (b) d Nuclear impulse approximation Double scattering e.g. Piller, Weise, Phys. Rep. 330 (2000) 1

36 Nuclear shadowing Interference of multiple scattering amplitudes For deuteron:!*!*!* X!* N N N N N N d N (a) d d (b) d F d 2 = F p 2 + F n 2 + δ (shad) F d 2 e.g. Piller, Weise, Phys. Rep. 330 (2000) 1

37 Space-time view of shadowing see e.g. Piller, Weise, Phys. Rep. 330 (2000) 1 "! * q q propagation length of hadronic fluctuation of mass (in lab frame) λ 1 E = µ 2ν µ 2 + Q 2 1 2xM, µ2 Q 2 d if propagation length exceeds average distance between nucleons λ > d 2 fm coherent multiple scattering can occur x < 0.05

38 Shadowing in deuterium Q 2 = 4 GeV 2 vector meson dominance (higher twist)

39 Shadowing in deuterium Pomeron exchange

40 Anti-shadowing in deuterium meson (pion) exchange WM, Thomas, Phys. Rev. D47 (1993) 3783

41 VMD important even at moderate Q 2

42 Shadowing in nuclei Perturbative or nonperturbative origin of Q 2 dependence? Q 2 = GeV 2 WM, Thomas, Phys. Rev. C52 (1995) 3373 NMC, Nucl. Phys. B481 (1996) 23

43 WM, Thomas, Phys. Rev. C 52 (1995) see also Badelek, Kwiecinski (1992), Nikolaev, Zoller (1992) Comparison with data

44 Effect on neutron structure function at small x F n 2 (F n 2 ) exp = 1 δf d 2 F d 2 ( 1 + (F n 2 /F p 2 ) exp (F n 2 /F p 2 ) exp ) where experimental n/p ratio is defined as F n 2 F p 2 F 2 d exp F p 2 1

45 Effect on neutron structure function at small x 1-2% enhancement at x ~ 0.01

46 Gottfried sum rule Integrated difference of p and n structure functions S G = 1 0 = dx F p 2 (x) F 2 n (x) x 1 Experiment: S G = ± NMC, Phys. Rev. D 50 (1994) 1 dx (ū(x) d(x)) d(x) ū(x) flavor asymmetric sea!

47 Saturation of Gottfried sum rule S G (x, 1) = 1 x dx F p 2 (x ) F n 2 (x ) x

48 Saturation of Gottfried sum rule correction to S G (0, 1) 0.02 ~ 10% decrease due to shadowing

49 Fermilab E866 Drell-Yan experiment q q annihilation in hadron-hadron collisions q q γ µ + µ beam target d 2 σ = 4πα2 dx b dx t 9Q 2 q Drell, Yan, Phys. Rev. Lett. 25 (1970) 316 e 2 q (q(x b ) q(x t ) + q(x b )q(x t )) For x b x t σ pd 2σ pp 1 2 ( 1 + d(x ) t ) ū(x t ) Ellis, Stirling, Phys. Lett. B256 (1991) 258

50 Towell et al., Phys. Rev. D 64 (2001) dx ( d(x) ū(x)) = ± 0.012

51 Why is d ū? Pauli blocking Since proton has more valence u than d easier to create than uū d d Field, Feynman, Phys. Rev. D15 (1977) 2590 Explicit calculations of antisymmetrization effects in and g d d ū > d g uū asymmetry tiny Ross, Sachrajda, Nucl. Phys. B149 (1979) 497 Steffens, Thomas, Phys. Rev. 55 (1997) 900

52 uū d d Steffens, Thomas, Phys. Rev. 55 (1997) 900

53

54 Why is d ū? Pion cloud some of the time the proton looks like a neutron & π + π + (Heisenberg Uncertainty Principle) p n p p π + n p at the quark level uud (udd)( du) uud Sullivan, Phys. Rev. D5 (1972) 1732 d > ū! Thomas, Phys. Lett. 126B (1983) 97 see Nucleon Models...

55 Tomorrow s lecture: 3. Connection with low energy models 4. PDFs from the lattice

56