1. XVIII th International Symposium on Spin PhysicsCharlottesville, October 6-11, 2008 GPDs, The Hunt for Quark Orbital Momentum (i) The nucleon spin puzzle (ii) Generalized parton distributions (iii) Deeply virtual Compton scattering (iv) Experimental access to E(Q 2,x, ,t) (v) Perspectives (vi) Experimental strategy (vii) Conclusions Laboratoire de Physique Subatomique et de Cosmologie Grenoble, France Eric Voutier ?

2. Orbital Momentum of gluons (unknown) Orbital Momentum of quarks and antiquarks (first glimpse) Helicity distributions describing the longitudinal polarization of quarks and antiquarks (known) Gluon polarization (first data) Eric Voutier The nucleon spin puzzle 1/1 SPIN2008, Charlottesville, October 6-11, 2008 Transversity distributions describing the transverse polarization of quarks and antiquarks (first experimental attempts) B.L. Bakker, E. Leader, T.L. Trueman, PRD 70 (2004) 114001 Terra Incognita Longitudinal spin sum rule Generalized parton distributions (GPDs) procure the cement of the nucleon spin puzzle, and more generally of the hadron structure.

3. Generalized parton distributions 1/6 Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008  GPDs are the appropriate framework to deal with the partonic structure of hadrons and offer the unprecedented possibility to access the spatial distribution of partons. Parton Imaging M. Burkardt, PRD 62 (2000) 071503 M.Diehl, EPJC 25 (2002) 223 GPDs can be interpreted as a 1/Q resolution distribution in the transverse plane of partons with longitudinal momentum x.  GPDs = GPDs(Q 2,x, ,t) whose perpendicular component of the momentum transfer to the nucleon is Fourier conjugate to the transverse position of partons.  GPDs encode the correlations between partons and contain information about the dynamics of the system like the angular momentum or the distribution of the strong forces experienced by quarks and gluons inside hadrons. X. Ji, PRL 78 (1997) 610 M. Polyakov, PL B555 (2003) 57 A new light on hadron structure

4. The GPDs Framework Generalized parton distributions 2/6 Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 GPDs Coherence between quantum states of different helicity, longitudinal momentum and transverse position.  At leading twist, the partonic structure of the nucleon is described by 4 quark helicity conserving and chiral even GPDs and 4 quark helicity flipping and chiral odd GPDs (+8 gluon GPDs). D. Müller et al., FP 42 (1994) 101 A.V. Radyushkin, PRD 56 (1997) 5524 X. Ji, PRL 78 (1997) 610 P. Hoodbhoy, X. Ji, PRD 58 (1998) 054006 M. Diehl, EPJC 19 (2001) 485 E’s are new and unknown distributions  In the forward limit (t  0,   0), H’s reduce to the forward parton distributions ( density, helicity, transversity ).  E’s, which involve nucleon helicity flip, do not have a DIS equivalent. x Initial longitudinal momentum fraction -2  Transferred longitudinal momentum fraction (skewness) (-t) 1/2 Momentum transfer to the nucleon

5. Generalized parton distributions 3/6 Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 Form Factors GPDs unify in the same universal framework parton distributions, form factors, and the spin of the nucleon. GPDs  The first Mellin moments relate GPDs to Dirac (H q ), Pauli (E q ), axial ( ), and pseudo-scalar ( ) nucleon form factors.  Similar relations relate chiral odd GPDs to tensor form factors.  independence from Lorentz invariance Transverse spin-flavor dipole moment in an unpolarized nucleon Energy Momentum Tensor X. Ji, PRL 78 (1997) 610 M. Polyakov, PLB 555 (2003) 57 M. Burkardt, PRD 72 (2005) 094020  The second Mellin moments relate GPDs to the nucleon dynamics, i.e. parton angular momentum and strong forces distributions. Correlation between quark spin and angular momentum in an unpolarized nucleon

6. Generalized parton distributions 4/6 Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 GPDs ? J.C. Collins, L. Frankfurt, M. Strikman, PRD56 (1997) 2982 X. Ji, J. Osborne, PRD 58 (1998) 094018 J.C. Collins, A. Freund, PRD 59 (1999) 074009 The key requirements for the experimental study of GPDs are luminosity and resolution. GPD(Q 2,x, ,t) Probe tagging (  ) Production of one additional particle (   ) Final state identification Exclusivity Factorization Interaction with elementary partons (Q 2 >>M 2 ) Separation of perturbative and non-perturbative scales (-t<<Q 2 ) Hardness Deep Exclusive Scattering  The factorization theorem allows to express the cross section for deep exclusive processes as a convolution of a known hard scattering kernel with an unkown soft matrix element related to the nucleon structure (GPDs).

7. Generalized parton distributions 5/6 Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 GPDs DVCS Deeply Virtual Compton Scattering Longitudinal response only GPDs DA DVMP Deeply Virtual Meson Production Hard gluon GPDs can be accessed via exclusive reactions in the Bjorken kinematic regime.  The DVCS process is identified via double (e  ) or triple (e  N) coincidences, allowing for small scale detectors and large luminosities.  Additional amplitudes contribute to the reaction process and interfere with the DVCS amplitude.  The identification of the DVMP process requires the detection of the meson disintegration products in large acceptance detectors.  Factorisation applies only to longitudinally polarized virtual photons whose contribution to the electroproduction cross section must be isolated.  Meson production acts as a GPD and a flavor filter.

8. GPDs enter the cross section of hard scattering processes via Compton form factors, that are integrals over the intermediate parton longitudinal momenta. GPDs Q 2 >> M 2 -t << Q 2 6/6 Generalized parton distributions Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008

9. Photon Electroproduction Eric Voutier  The Bethe-Heitler (BH) process where the real photon is emitted either by the incoming or outgoing electron interferes with DVCS.  DVCS & BH are indistinguishable but the BH amplitude is exactly calculable and known at low t.  The relative importance of each process is beam energy and kinematics dependent. Deeply virtual Compton scattering SPIN2008, Charlottesville, October 6-11, 2008 1/5 leptonic plane e -’  p e-e- ** hadronic plane  Out-of-plane angle entering the harmonic development of the reaction amplitude A.V. Belitsky, D. Müller, A. Kirchner, NPB 629 (2002) 323 Polarization observables help to single-out the DVCS amplitude. P 0 = 6 GeV/c

10. Polarized Beam Cross Section Difference Eric Voutier Deeply virtual Compton scattering SPIN2008, Charlottesville, October 6-11, 2008 2/5 The s DVCS,I are bilinear and linear combinations of GPDs at ±  C. Muñoz-Camacho et al., PRL 97 (2006) 262002 Twist-2 Twist-3 Q 2 = 2.3 GeV 2 x B = 0.36 Q 2 = 2.3 GeV 2 x B = 0.36 DVCS / Bethe-Heitler interference amplitude DVCS amplitude Imaginary part  The DVCS process is associated with a non-zero polarized beam signal.

11. Gluons Twist-2 Twist-3 The c DVCS,I are bilinear and linear combinations of Compton form factors. Unpolarized Cross Section Eric Voutier Deeply virtual Compton scattering SPIN2008, Charlottesville, October 6-11, 2008 3/5 C. Muñoz-Camacho et al., PRL 97 (2006) 262002 Q 2 = 2.3 GeV 2 x B = 0.36 Q 2 = 2.3 GeV 2 x B = 0.36  The magnitude of the DVCS process is seen as a deviation from the pure BH cross section. Bethe-Heitler amplitude DVCS amplitude DVCS / Bethe-Heitler interference amplitude Real part

12. E00-110  p-DVCS P.Y. Bertin, C.E. Hyde, R. Ransome, F. Sabatié et al.  Twist-two contributions dominate the DVCS cross section and are Q 2 independent in the measured kinematics range.  Twist-three contributions to the cross section are small and are Q 2 independent within error bars. C. Muñoz-Camacho et al., PRL 97 (2006) 262002 Deeply virtual Compton scattering Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 4/5 Experimental results support the dominance of the handbag diagram at Q 2 as low as 2 GeV 2.

13. Beam Spin Asymmetry Deeply virtual Compton scattering Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 5/5 F.-X. Girod et al., PRL 100 (2008) 162002 @ CLAS (5.77 GeV) GPDs Twist-2 JML / Regge GPDs Twist-3  Calculations based on hadronic degrees of freedom, within a Regge approach, are in fair agreement with data up to 2.3 GeV 2.  GPD based calculations reproduce reasonably well the main features of the data. The angular dependence of the asymmetries is compatible with expectations from leading-twist dominance. F.-X. Girod talk in GPD parallel session

14. F. Ellinghaus,W.-D. Nowak, A.V. Vinnikov, Z. Ye EPJ C46 (2006) 729 Z. Ye, Doctorat Thesis, Universität Hamburg (2006)  DVCS Transverse Target Asymmetry Experimental access to E(Q 2,x, ,t) Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 1/6 Twist-2 A.V. Belitsky, D. Müller, A. Kirchner, NPB 629 (2002) 323 M. Diehl, S. Sapeta, EPJC 41 (2005) 515 B. Zihlmann talk in GPD parallel session

15. 2/6 Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 Experimental access to E(Q 2,x, ,t) DVMP Transverse Target Asymmetry Transverse Target Asymmetry xBxB 2J u +J d Transverse Target Asymmetry xBxB 2J u -J d  The s-channel helicity conservation (SCHC) allows to extract the longitudinal/transverse ratio from the angular distribution of the decay products of the vector mesons, avoiding a Rosenbluth separation.  The longitudinal polarization of the vector meson is obtained from the angular distribution of the decay products. Transversally polarized proton targets are of primary importance for the determination of E(Q 2,x, ,t). K. Goeke, M. Polyakov, M. Vanderhaeghen, PPNP 47 (2001) 401 S.V. Goloskokov, P. Kroll, arXiv/hep-ph:0809.4126

16. 3/6 SPIN2008, Charlottesville, October 6-11, 2008 Eric Voutier Experimental access to E(Q 2,x, ,t) DVMP Longitudinal Cross Section S.A. Morrow et al., arXiv/hep-ex:0807.3834 @ CLAS (5.74 GeV)  Standard GPD calculations fail to reproduce data in the valence region, while successfull at large W.  Data can be interpreted in terms of hadronic degrees of freedom, following a Regge approach.  Data can be interpreted in terms of partonic degrees of freedom via strongly modified GPDs, via a D-term like adjusted component. VGG / GPDs GK / GPDs JML / Regge VGG++ / GPDs The question is asked whether current GPD parametrizations must be revisited or DVMP cannot be interpreted GPD wise in the valence region ?

17. Neutron Target Suppressed because F 1 (t) is small Suppressed because of cancellation between u and d quarks From polarized beam cross section difference Neutron targets allow to access the least known and constrained GPD that appears in the nucleon spin sum rule. Neutron targets provide new linear combinations of GPDs Neutron and proton targets are sensitive to the u quark flavor and, following isospin symmetry, appear then complementary. Eric Voutier Experimental access to E(Q 2,x, ,t) SPIN2008, Charlottesville, October 6-11, 2008 4/6

18. E03-106  n-DVCS P.Y. Bertin, C.E. Hyde, F. Sabatié, E. Voutier et al. S. Ahmad et al., PR D75 (2007) 094003 M. Vanderhaeghen et al., PR D60 (1999) 094017 M. Mazouz et al., PRL 99 (2007) 242501  The twist-2 effective harmonic coefficients of the neutron are small, compatible with zero.  The measured t-dependence can be used to constrain the parametrization of the GPD E, within a particular model. Impulse approximation Q 2 = 1.9 GeV 2 x B = 0.36 M. Mazouz et al., PRL 99 (2007) 242501 Experimental access to E(Q 2,x, ,t) Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 5/6 Twist-2

19. 6/6 Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 u M. Mazouz et al., PRL 99 (2007) 242501 Measurements off neutron are sensitive to J d (u quark in the neutron) Measurements off proton are sensitive to J u (u quark in the proton) A. Airapetian et al., arXiv/hep-ex:0802.2499 Model Dependent Quark Angular Momenta d A.W. Thomas, arXiv/hep-ph:0803.2775 Neutrons are a mandatory step in the hunt for the quark orbital momentum. Experimental access to E(Q 2,x, ,t)

20. E08-021  -DVCS H. Avakian, V. Burkert, M. Guidal, R. Kaiser, F. Sabatié et al. Perspectives 1/6 Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 G. Thorn talk in Polarized Sources & Targets parallel session Advantages  The HDice frozen spin target figures a large spin relaxation time (~1 year) at operational temperature (0.5 K) and field (0.9 T). High T -> high cooling power compensates beam heating Low B -> ideal for transverse polarization @ CLAS Pure target -> small physics background & faster data taking Low Z -> small bremsstrahlung background (e - ) Drawbacks  The many delicate processes between production and operation make the complete system quite complex.  Operation with an electron beam is promising but has not yet been established. The technical feasability of electron experiments with the HDice target has to be demonstrated. HD is condensed into the target cell with ~2000 50 µm Al wires soldered to a copper cooling ring HD (77%) Al(16%) C 2 ClF 3 (7%) HDice Lab @ JLab A. Sandorfi et al., Proc. of the Workshop on Exclusive Reactions, Edts. P. Stoler and A. Radyushkin, World Scientific (2008) 425.

21. Eric Voutier E07-007 & E08-025  p-DVCS & n-DVCS P.Y. Bertin, C.E. Hyde, C. Muñoz Camacho, J. Roche et al. A. Camsonne, C.E. Hyde, M. Mazouz et al. -0.36 Within the VGG model, the neutron pure DVCS amplitude is sensitive to E. A.V. Belitsky, D. Müller, A. Kirchner, NP B629 (2002) 323 Perspectives SPIN2008, Charlottesville, October 6-11, 2008 2/6  The pure DVCS contribution to the cross section is separated from the interference contribution, in the twist- 2 approximation, taking advantage of the beam energy dependence of the kinematic factors. Twist-2

22. E05-114  -DVCS A. Biselli, L. Elouadrihri, K.Joo, S. Nicolai et al. Perspectives 3/6 Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 A.V. Belitsky, D. Müller, A. Kirchner, NP B629 (2002) 323 A. Biselli and H. Egiyan talks in GPD parallel session Access to another Compton form factor, that is another linear combination of GPDs Gluons Twist-2 Twist-3 Access to the imaginary part of H from the target spin asymmetry. Access to the real part from double polarization observable.

23. Perspectives 4/6 Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 R. Kaiser talk in GPD parallel session The GPDs World xBxB One may expect to get a reasonable mapping of the GPDs in the intermediate and valence regions by 2020.  H1, ZEUS, HERMES, JLab 6 GeV are providing the first but limited data sets.  The energy upgrade of the CEBAF accelerator allows access to the high x B region which requires large luminosity.  The DVCS project at COMPASS will explore intermediate x B (0.01-0.10) with a reasonable overlap with the JLab 12 GeV kinematic domain.

24. Perspectives Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 GPDs @ JLab 12 GeV  Several experiments will measure DVCS (Hall A & B) and DVMP (Hall B) with (un)polarized beam and target, extending the experimental techniques and methods developed at 6 GeV.  An eventual transversally polarized target program in Hall B is attached to the technical success of the HDice target.  The development of neutron detection capabilities in the central detector region (Hall B) and of polarized neutron targets sustaining high beam currents (Hall A) are investigated. D(e,e’  n)p 5/6 D(e,e’  n)p

25. Perspectives Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 6/6 GPDs @ COMPASS μ  p’ μ’μ’  The GPDs program is part of the COMPASS Phase II (2010-2015) proposal to be submitted to CERN in 2008-2009. Step 1 (~2011) to constrain H d  (  + ,  ) + d  (  - ,  )  Im(F 1 H) sin  d  (  + ,  ) - d  (  - ,  )  Re(F 1 H) cos  Step 2 (~2013) to constrain E d  ( ,  S ) - d  ( ,  S +π)  Im(F 2 H – F 1 E) sin (  -  S ) cos   The first step of this program requires a 4 m long recoil proton detector (RPD) together with a 2.5 m long LH 2 target. An additional electromagnetic calorimeter will enlarge the kinematical coverage at large x B.  The second step requires either a transversally polarized NH 3 target inserted in the RPD or a new SciFi (?) RPD inserted in the existing NH 3 target system.

26. Experimental strategy Eric Voutier SPIN2008, Charlottesville, October 6-11, 2008 1/1 Q 2 = 2.3 GeV 2 x B = 0.36 Compton Form factors M. Guidal, arXiv/hep-ph:0807.2355  Within a fitting code relying on a leading order and leading twist description of the DVCS amplitude, it was shown that a reliable and model-independent extraction of the real and imaginary parts of each Compton form factor requires a minimal number of observables.  Cross section ( , single and double polarized cross section differences (  ij ) should be measured; asymmetries are providing additional constraints but are not enough selective by themselves.  Dispersion relation constraints are expected to improve this picture by reducing the number of required observables and/or improving the accuracy of the fitting procedure. Current  and  z0 data suggest that VGG predictions underestimate slightly the imaginary part of H and miss the real part. Imaginary partsReal parts It remains a theoretical concern to extract GPDs from CFFs. A.V. Belitsky, D. Müller, arXiv/hep-ph:0809.2890

27. Conclusions 1/2 Eric Voutier Summary GPDs offer the unique opportunity to access the contribution of the quark angular momentum to the spin of the nucleon. In this effort, neutron and transversally polarized proton targets are mandatory steps. SPIN2008, Charlottesville, October 6-11, 2008 Special thanks to François-Xavier Girod, Michel Guidal, Nicole d’Hose, Malek Mazouz, Andy Sandorfi Together with the evolution of lattice QCD calculation capabilities, the road towards a comprehensive and quantitative understanding of the nucleon structure is OPEN!!  The exploration of the GPD world has been starting at DESY & JLab and will continue at JLab 12 GeV and COMPASS with high precision and exclusive experiments.  Recent phenomenological & theoretical developments tend to indicate that this world should be efficiently mapped with the measurement of a limited set of observables.

29. Eric Voutier Perspectives SPIN2008, Charlottesville, October 6-11, 2008 Hall B HDice Lab HDice Lab @ JLab  HDice Lab design and infrastructure – Nov’08  Building construction – (Feb’09, Jun’09)  Cryogenic infrastructure – (Apr’09, Sep’09)  Polarized target tests – (Sep’09, Nov’09)  Design & construction of a new In-Beam Cryostat for CLAS – (May’08, Jul’10)  Set of polarized targets for  experiments – (Mar’10, Aug’10)  Installation in Hall B – (Jul’10, Sep’10)   run – (Sep’10, Apr’11)  Electron beam tests – (Apr’11)  e run – (Nov’11, Dec’11) Courtesy of A. Sandorfi

30. Eric Voutier  Experimental data are taken for LH 2 and LD 2 targets at 4.82 GeV/c and 6.00 GeV/c.  The LD 2 -LH 2 yields binned in (k, M x 2, , t) are fitted with Monte-Carlo simulated yields taking into account experimental and radiative effects. Unpolarized Cross Section Projected systematics Perspectives SPIN2008, Charlottesville, October 6-11, 2008 P R O J E C T E D D A T A

31. Eric Voutier n-DVCS @ CLAS 12 INFN Frascati, INFN Genova, IPN Orsay, LPSC Grenoble, University of Glasgow … … European initiative towards the central detector …  Neutron detection at laboratory angles larger than 40° would allow an efficient and fully exclusive study of the n-DVCS process up to the valence region.  The possibility of developing neutron detection capabilities in the central region of CLAS12 is forseen. A. El Alaoui (LPSC Grenoble) (2008) SPIN2008, Charlottesville, October 6-11, 2008 Perspectives ToF scintillators Vertex detector Neutron detector 5 T @ target D(e,e’  n)p

32.  Perspectives Eric Voutier Polarized Neutron Target Longitudinal Target Spin Asymmetry Transverse Target Spin Asymmetry The twist-2 target spin asymmetries are derived below in the case of a polarized neutron, assuming that the Dirac form factor and the non spin-flip polarisation dependent GPD are 0. The most sensitive coefficient to E appears to originate from the pure DVCS amplitude while the kinematical factors enhance H in the other coefficients. A. Belitsky, D. Müller, A. Kirchner, NP B629 (2002) 323 SPIN2008, Charlottesville, October 6-11, 2008

33. Eric Voutier Forward Calorimeter Forward Cerenkov (LTCC) Beamline Instrumentation Preshower Calorimeter Forward Drift Chambers Inner Cerenkov (HTCC) Superconducting Torus Magnet Central Detector Inner Calorimeter Forward Time-of-Flight Detectors CLAS12 SPIN2008, Charlottesville, October 6-11, 2008 Perspectives