1. COMPASS : a Facility to study QCD CO COMMON M MUON and P PROTON A APPARATUS for S STRUCTURE and S SPECTROSCOPY 1 N. d’Hose (CEA Saclay) Studies until now 2011: Studies until now 2011:  Nucleon Spin  Nucleon Spin with high energy polarized  beams + polarized targets: longitudinal spin: gluon and quark helicity distribution transverse spin and transverse momentum dependent distribution  Spectroscopy  Spectroscopy with hadron beams + LH2 (or solid) targets: Search of hybrids and glueballs to better understand quark and gluon confinement

2. Primakoff with π, K beam  Test of Chiral Perturb. Theory DVCS & DVMP with μ beams  Transv. Spatial Distrib. with GPDs SIDIS (with GPD prog.)  Strange PDF and Transv. Mom. Dep. PDFs Drell-Yan with  beams  Transverse Momentum Dependent PDFs New Program (SPSC-P-340) until ~2016 New Program (SPSC-P-340) until ~2016 recommended by SPSC and approved by Research Board recommended by SPSC and approved by Research Board COMPASS-II: a Facility to study QCD CO COMMON M MUON and P PROTON A APPARATUS for S STRUCTURE and S SPECTROSCOPY 1

3. SPS proton beam: 2.6 10 13 /spill of 9.6s each 48s, 400 GeV/c  Secondary hadron beams ( , K, …): 6.10 8 /spill, 50-200 GeV/c  Tertiary muon beam (80% pol): 4.610 8 /spill, 100-200 GeV/c -> Luminosity ~ 10 32 cm -2 s -1 GPD with  + and a 2.5m long LH target -> Luminosity ~ 10 32 cm -2 s -1 GPD with  + and a 2.5m long LH target ~1.2 10 32 cm -2 s -1 DY with  - and a 1.1m long NH 3 target ~1.2 10 32 cm -2 s -1 DY with  - and a 1.1m long NH 3 target LHC high energy beams, broad kinematic range, large angular acceptance SPS CNGS CNGS Gran Sasso 732 kms 732 kms COMPASS 60m

4. Meson Spectroscopy 3

5. Hadron Reactions at COMPASS COMPASS: COMPASS: production mechanisms studied in parallel using proton, pion and kaon beams H2, Ni and Pb targets identification of charged and neutral particles in a large and flat acceptance 10  world statistics  PWA analysis (international task force for COMPASS, JLab, PANDA…) 4

6. Search of the controversial hybrids  1 - (1600) with J PC =1 -+ in the reaction  - p   1 - (1600) p   - p   +  -  - p + Pb (2009 data) 2.3M events + other targets W, Ni

7. Search of the controversial hybrids  1 - (1600) with J PC =1 -+ in the reaction  - p   1 - (1600) p   - p   +  -  - p a1(1260)  2(1670) a2(1320)  1(1600) J PC M  (isobar,  )L Pb 2004 data PRL104 (2010) 241803

8. Search of the controversial hybrids  1 - (1600) with J PC =1 -+ in the reaction  - p   1 - (1600) p   - p   +  -  - p a1(1260)  2(1670) a2(1320)  1(1600) Phase between  1 (1600) and a 1 (1260) J PC M  (isobar,  )L Clear  1 - (1600) signal in  +  -  - for Pb target PRL104 (2010) 241803 Pb 2004 data

9. Nuclear Medium effect Nuclear effect: Production of Spin Projection M=1 states enhanced for larger A Pbfrom Pb (2004) and H 2 (2008) targets Pb H2H2 H2H2 H2H2 H2H2 7

10. QCD at low energy: Primakoff experiments with π, K Primakoff experiments with π, K or inverse Compton Scattering on π, K or inverse Compton Scattering on π, K 8 the point-like cross section is measured with the muon beam the point-like cross section is measured with the muon beam Deviation due to  polarisabilities The chiral perturbation theory (ChPT) predicts the low-energy behavior of the cross section with s varying from threshold (m π 2 ) to a few m π 2 the cross section with s varying from threshold (m π 2 ) to a few m π 2 π (or Κ) γ Q2  0Q2  0 γ s = ( p  +p  ) 2 = (  CM energy) 2 t = ( cos cm -1) (s - m π 2 ) 2s γ

11. Pion Polarisabilities and Chiral predictions 9 The pion: fundamental role for QCD at low-energy Goldstone boson (spontaneous breaking of chiral symmetry) Goldstone boson (spontaneous breaking of chiral symmetry) lightest quark-gluon bound state system lightest quark-gluon bound state system  understanding its internal structure is a fundamental challenge  understanding its internal structure is a fundamental challenge The polarisabilities give the deformation of the pion shape by an EM field   > 0 S=0 diamagnetic contr.   <0 2-loop ChPT prediction:   +   = (0.2  0.1) 10 -4 fm 3   -   = (5.7  1.0) 10 -4 fm 3 Previous experiments:   -   varies from 4 to 14.10 -4 fm 3 accuracy accuracy  0.025  0.025  0.66  0.66 Goal of Goal ofCOMPASS in 120 days

12. PxPx  p   ’X Deep Inelastic Scattering  Q²x B x p γ* γ* Parton Distribution q ( x ) x boost x P z y QCD at high energy: QCD at high energy: Deep Inelastic Scattering Deep Inelastic Scattering ’’ 10 Quark with momentum x in a nucleon Quark with spin parallel to the nucleon spin in a longitudinally polarized nucleon Quark with spin parallel to the nucleon spin In a transversely polarized nucleon Chiral odd, accessible in SIDIS and DY 3 twist-2 PDFs or f 1 q (x) or g 1 q (x) Unpolarized PDF Helicity PDF Transversity PDF or h 1 q (x)

13. Helicity quark distributions From polarized semi-inclusive DIS analysis, Using charged pions and kaons which tag quark flavour on both proton and deuteron polarized targets  Full quark flavour decomposition (at LO) Longitudinal spin asymmetry We need unpolarized PDF (MRST04) and parameterization of FF from DSS (DeFlorian, Sassot, Stratmann, PRD75 (2007) 114010) quark Fragmentation Function (FF) 11

14. COMPASS PLB693 (2010) 227 full flavour separation at LO down to x=0.004 o HERMES PRD71 (2005) -DSSV predictions at NLO (DeFlorian, Sassot, Stratmann, Vogelsang PRD80 (2009) 034030) LO Helicity quark distributions Good agreement with global fits to g 1 inclusive data The flavor asymmetry of the sea helicity distributions is compatible with 0 (while there is a considerable wide asymmetry in the unpolarized case)  s ~ 0 but the DSSV fit suggests negative values at smaller x 12

15. Transverse spin or momentum PDF 13 Sivers Collins Collins fragmentation function, depends on spin Usual quark fragmentation function at LO: SIDIS with transversely polarized target Measure simultaneously two azimuthal asymmetries (  h,  S ) Sivers: Nucleon spin & quark transverse momentum Collins: Outgoing hadron direction & quark transverse spin  p  p h +/- Transversity PDF or h 1 q (x) = f 1T  (x) One Transverse Momentum PDF beyond the collinear approximation, quark transverse momentum (k T ) effects

16. Transversity from Collins asymmetry Fit to A d COLL (x) from COMPASS and A p COLL (x) from HERMES and Collins fragmentation function from BELLE (e + e - scatt.) in good agreement with new proton data 14 COMPASS proton 2007

17. Sivers TMD from Sivers asymmetry Positive hadrons Negative hadrons Comparison with predictions from Anselmino et al., based on fit of Hermes-p and Compass–d data Extraction of Sivers fct (HERMES p and COMPASS d) x Present proton data not in fit - COMPASS signal < HERMES signal -Possible W dependence -proton 2010 data to be analysed = f 1T  (x) 15 COMPASS proton 2007

18. After SIDIS, Drell-Yan to study TMDs π - p    +  - X Drell –Yan π - p    +  - X with intense pion beam (up to 10 9  /spill) with the transversely polarised NH 3 target with the COMPASS spectrometer equipped with an absorber Cross sections: In SIDIS: convolution of a TMD with a fragmentation function In SIDIS: convolution of a TMD with a fragmentation function In DY: convolution of 2 TMDs ‘’’’ In DY: convolution of 2 TMDs ‘’’’  complementary information and universality test  complementary information and universality test 16 - ‘

19. Why DY is very favourable at COMPASS? 17 dominated by the annihilation of a valence anti-quark from the pion and a valence quark from the polarised proton This will be the only such experimental measurement for at least 5-10 years 1 rst moment of Sivers function for u quark large acceptance of COMPASS in the valence quark region for p and π where SSA are expected to be larger  Access to TMDs for incoming pion  target nucleon (TMDs as Transversity, Sivers, Boer-Mulders, pretzelosity)

20. Experimental check of the change of sign of Experimental check of the change of sign of TMDs confronting Drell-Yan and SIDIS results TMDs confronting Drell-Yan and SIDIS results The T-odd character of the Boer-Mulders and Sivers function implies that these functions are process dependent Boer-Mulders Sivers In order not to be forced to vanish by time-reversal invariance the SSA requires an interaction phase generated by a rescattering of the struck parton in the field of the hadron remnant 18 FSI ISI Need experimental verification Test of consistency Test of consistency of the approach of the approach COMPASS + HERMES have already measured non zero Sivers SSA in SIDIS

21. Results from test measurements in 2009 19 2009 with an absorber up to 1.5 10 8  /spill (without radioprotection pb) Expected: 3600±600 J/  and 110±22 DY Measured: 3170±70 J/  and 84±10 DY 4 < M  +  - < 9 GeV 2 < M  +  - < 2.5 GeV M  +  - M  +  - GeV 1- safe domain for Drell-Yan Large Combinatorial Background S/B ~ 1 Open charm decays (from D 0 ) smaller than 15% of signal 2.9 < M  +  - < 3.2 GeV 4 < M  +  - < 9 GeV 2-reasonable domain for Drell-Yan 2 < M  +  - < 2.5 GeV 3 domains of study 3- domain for J/  mechanism 2.9 < M  +  - < 3.2 GeV if q-qbar dominates over g-g fusion Drell-Yan can be considered

22. 20 Predictions for Drell-Yan at COMPASS Sivers function in the safe dimuon mass region 4 < M  +  - < 9 GeV 2 years of data 190 GeV pion beam 6.10 8 π - /spill (of 9.6s) 1.1 m transv. pol. NH 3 target Lumi=1.2 10 32 cm -2 s -1 Blue line with grey band: Anselmino et al., PRD79 (2009) Black solid and dashed: Efremov et al., PLB612 (2005) Black dot-dashed: Collins et al., PRD73 (2006) Squares: Bianconi et al., PRD73 (2006) Green short-dashed: Bacchetta et al., PRD78 (2008) The change of sign can already be seen in 1 year

23. Deeply Virtual Compton Scattering Generalized Partons Distrib. H( x, , t )  p   ’p’  P’ GPDs ** Q²Q² x+  x-  p  t x P y z bb x boost ( P x, b  ) from inclusive reactions from inclusive reactions to exclusive reactions to exclusive reactions Observation of the Nucleon Structure in 1 dimension in 1+2 dimensions in 1 dimension in 1+2 dimensions 21 PxPx  p   ’X Deep Inelastic Scattering  Q²x B x p γ* γ* Distrib. de Partons q ( x ) x boost x P z y ’’

24. 22 From Wigner phase-space-distributions (Ji, PRL 2003, Belitsky, Ji, Yuan PRD 2004) We can build « mother-distributions » (Meissner, Metz, Schlegel, JHEP 0908:056 2009) and derive kk Exploring the 3-dimensional phase-space structure of the nucleon structure of the nucleon GPD: Generalised Parton Distribution (position in the transverse plane) GPD: Generalised Parton Distribution (position in the transverse plane) TMD:Transverse Momentum Distribution (momentum in the transv. plane) TMD:Transverse Momentum Distribution (momentum in the transv. plane) PDF GPD TMD New research fields

25.   Allow for a unified description of form factors and parton distributions transverse imaging (nucleon tomography)  Allow for transverse imaging (nucleon tomography) total angular momentum and give access to total angular momentum (through E) Generalized Partons Distributions (H,E,…) Tomographic parton images of the nucleon Impact parameter b  Longitudinal momentum fraction x  23 x  0.003 x  0.03 x  0.3

26. CERN High energy muon beam 100 - 190 GeV μ + and μ - available 80% Polarisation with opposite polarization  Will explore the intermediate x Bj region  Uncovered region between ZEUS+H1 and HERMES+Jlab before new colliders may be available presently only 2 actors: Jlab and COMPASS  Transverse structure at x~10 -2 essential input for phenemenology of high- energy pp collision (LHC) B What makes COMPASS unique for GPDs? 24

27. Experimental requirements for DVCS μ’μ’ p’ μ  Phase 1 Phase 1 ~ 2.5 m Liquid Hydrogen Target ~ 4 m Recoil Proton Detector (RPD) ~ 4 m Recoil Proton Detector (RPD)  ECALs upgraded + ECAL0 before SM1 + ECAL0 before SM1 ECAL2 ECAL1 μ p  μ ’ p  SM1 SM2 25

28.  Contributions of DVCS and BH at E  =160 GeV θ μ’μ’ μ **  p Deep VCS Bethe-Heitler BH dominates study of Interference DVCS dominates BH dominates study of Interference DVCS dominates excellent  Re T DVCS study of d  DVCS /dt reference yield or Im T DVCS  Transverse Imaging 26 d   |T DVCS | 2 + |T BH | 2 + Interference Term Monte-Carlo Simulation for COMPASS set-up with only ECAL1+2  Missing DVCS acceptance without ECAL0

29.  θ μ’μ’ μ **  p Deeply Virtual Compton Scattering d  DVCS / d t ~ exp(-B|t|) Phase 1: DVCS experiment to study the transverse imaging with  + ,  -  beam + unpolarized 2.5m long LH2 (proton) target S CS,U  d  (  +  ) + d  (  -  )  Using S CS,U and integration over  and BH subtraction  dσ (μp  μp  ) = dσ BH + dσ DVCS unpol + P μ dσ DVCS pol + e μ a BH Re A DVCS + e μ P μ a BH Im A DVCS 27

30. DVCS: Transverse imaging at COMPASS d  DVCS /dt ~ exp(-B|t|) 2 years of data 160 GeV muon beam 2.5m LH 2 target  global = 10%  global = 10% 28 without any model we can extract B(x B ) B(x B ) = ½ r  is the transverse size of the nucleon

31. DVCS: Transverse imaging at COMPASS d  DVCS /dt ~ exp(-B|t|) 14 Transverse size of the nucleon 0.65  0.02 fm H1 PLB659(2008) 1. 0.5 ? xBxB COMPASS without any model we can extract B(x B ) B(x B ) = ½ r  is the transverse size of the nucleon

32. DVCS: Transverse imaging at COMPASS d  DVCS /dt ~ exp(-B|t|) α ’ B(x B ) = b 0 + 2 α ’ ln(x 0 /x B ) ansatz at small x B inspired by Regge Phenomenology: with the projected uncertainties we can determine : α ’ α ’ slope of Regge traject  B with an accuracy of 0.1 GeV -2 α ’  α ’ with an accuracy  2.5  α ’  0.26 if α ’  0.26 with ECAL1+2 α ’  0.125 if α ’  0.125 with ECAL0+1+2

33. Phase 1: DVCS experiment to constrain GPD H with  + ,  -  beam + unpolarized 2.5m long LH2 (proton) target and  Im( F 1 H) and  Re( F 1 H) S CS,U  d  (  +  ) + d  (  -  )  D CS,U  d  (  +  ) - d  (  -  )  Deeply Virtual Compton Scattering  Re H ( ,t) = P  dx H (x, ,t) /(x-  )  Im H ( ,t) = H (x= , ,t)   x B / (2-x B ) 29 dσ (μp  μp  ) = dσ BH + dσ DVCS unpol + P μ dσ DVCS pol + e μ a BH Re A DVCS + e μ P μ a BH Im A DVCS

34. Beam Charge and Spin Difference D CS,U Beam Charge and Spin Difference (using D CS,U ) Comparison to different models 2 years of data 160 GeV muon beam 2.5m LH 2 target  global = 10% Systematic error bands assuming a 3% charge-dependent effect between  + and  - (control with inclusive evts, BH…)  θ μ’μ’ μ **  p 30  ’=0.8  ’=0.05

35. 2008-9 tests: observation of BH and DVCS events 31 10 BH events expected  ~30 « pure » DVCS events with ~10  0 contamination During the hadron program with 1m long recoil proton detector (RPD) and 40cm long LH2 target and the 2 existing ECAL1 and ECAL2 2008: observation of exclusive single photon production, ε global = 0.13 +/- 0.05  confirmed ε global = 0.1 as assumed for simulations 2009: observation of BH and DVCS events

36. GPDs investigated with Hard Exclusive Photon and Meson Production  the t-slope of the DVCS cross section ……………… LH 2 target + RPD ……phase 1  transverse distribution of partons  the Beam Charge and Spin Sum and Difference and Asymm…………….phase 1  Re T DVCS and Im T DVCS for GPD H determination  the Transverse Target Spin Asymm……… polarised NH 3 target + RPD ……phase 2  GPD E and angular momentum of partons (future addendum) Summary for GPD @ COMPASS 32

37. Only selected topics among the large harvest of (present and incoming) results on Spin and Spectroscopy The main topics of COMPASS-II are:  GPD with DVCS and DVMP  TMD with DY and SIDIS  precise unpolarised PDF measurements  Chiral perturbation theory - soft QCD + updated programs on spectroscopy, DY, GPD… Jlab 12 GeV, huge luminosity, very promising facility Jlab 12 GeV, huge luminosity, very promising facility and for the next 10 years, before ENC, EIC, eRHIC, LHeC, and for the next 10 years, before ENC, EIC, eRHIC, LHeC, COMPASS@CERN can be a major player in QCD physics COMPASS@CERN can be a major player in QCD physics using its unique 200 GeV hadron and polarised muon beams using its unique 200 GeV hadron and polarised muon beams Conclusions 33

39. Pion Polarisabilities measurement 5 Polarisability effect with increasing s at backward or forward angle  The 3 components (   -   ), (   +   ), (  2 -  2 ) can be measured point-like case for ChPT prediction

40. The Drell-Yan process in π - p 15 Lepton plane Collins-Soper frame (of virtual photon)  lepton plane wrt hadron plane target rest frame  S target transverse spin vector /virtual photon Hadron plane  Access to TMDs for incoming pion  target nucleon TMD as Transversity, Sivers, Boer-Mulders, pretzelosity

41. π-π- 190 GeV DY and COMPASS set-up 17 Key elements for a small cross section investigation at high luminosity 1.Absorber (lesson from 2007-8 tests) to reduce secondary particle flux 2.COMPASS Polarised Target 3.Tracking system and beam telescope 4.Muon trigger (LAS of particular importance - 60% of the DY acceptance) 5.RICH1, Calorimetry – also important to reduce the background

42. x F =x  -x p  Results from test measurements in 2009 19 = 1 GeV/c  COMPASS sensitive to TMDs

43. 21 Predictions for Drell-Yan at COMPASS x F =x  -x p  Statistical accuracy from 0.01 to 0.02 in two years -0.2 < x F < 0.85

44. 21 Predictions for Drell-Yan at COMPASS x F =x  -x p  Statistical accuracy 1  2% in two years x F =x  -x p 

45. 22 Summary for Drell-Yan @ COMPASS Summary for Drell-Yan @ COMPASS First ever polarised Drell-Yan experiment sensitive to TMDs Pure valence u dominance because of the  - beam Key measurements: TMDs universality SIDIS  DY change of sign check from SIDIS to DY for Sivers and Boer-Mulders study of J/  production mechanism After a series of beam tests, the feasibility is proven In a phase 2 (future addendum), measurement with anti-proton beams will be envisaged

46. PxPx  p   ’X Deep Inelastic Scattering  Q²x B x p γ* γ* Parton Distribution q ( x ) x boost x P z y QCD at high energy: QCD at high energy: Deep Inelastic Scattering Deep Inelastic Scattering ’’ 11 While unpolarised light quark PDF well constrained, strange quark distributions are not so well known COMPASS with SIDIS

47. Semi-Inclusive Deep Inelastic Scattering 8 Semi-Inclusive DIS measurements during GPD program with a pure proton target with RICH detector and Calorimeters Charge separation and identification K +, K -, K 0,  +,  -,  0,  … Major progress as compared to previous experiments to improve Unpolarised PDF Hadron multiplicities at LO PDF quark Fragmentation Function depend on x depend on z (fraction of energy of the outgoing hadron) Final goal: extensive measurements (x,z,…) to provide input to NLO global analysis for PDF and FF

48. Projection for 1 week with 2.5m LH 2 target  high statistics Projection of errors for s(x) 9 Short term goal: LO analysis from COMPASS data alone integrated over z

49. Semi-Inclusive Deep Inelastic Scattering 11 Semi-Inclusive DIS measurements during GPD program with a pure proton target with RICH detector and Calorimeters Cahn effect  info on Boer-Mulders TMD  Collins FF + Cahn effect measurement of A cos  h and A cos 2  h + knowledge of Collins FF  Boer-Mulders TMD determination Unpol. cross section:

50. 14 Projection of errors for A cos  h and A cos 2  h 12 Projection proton (LH2) 1 week COMPASS 2004 deuteron 4 weeks + flavor separation using RICH particle identification  Input for global analysis

51. hh Semi-Inclusive Deep Inelastic Scattering 10 Asymmetries in the azimuthal angle  h of the outgoing hadron around the virtual photon can reveal quark transverse spin and quark transverse momentum (k T ) effects beyond the collinear approximation At leading twist, not only f 1 q (x, k T ), g 1L q (x, k T ), h 1 q (x, k T ) but also 5 otherTransverse Momentum Dependent PDF (TMD (x, k T )) which do not survive after integration on k T 2 examples of TMDs The Boer-Mulders function correlates the quark k T and the quark transverse spin (unpol N) The Sivers function correlates the quark k T and the nucleon spin (transv. Pol. N)

52. gluon polarization Direct measurement (at LO untill now ): global QCD analysis : Use Q2 evolution of spin Dependent gluon and singlet Quark distribution g 1 (x,Q 2 ) COMPASS NLO fit of g 1 data: 2 solutions with |  G|=0.2 – 0.3 14

53. gluon polarization Longitudinal spin asymmetry 13