TRANSVERSE SPIN EFFECTS COMPASS measurements of TRANSVERSE SPIN EFFECTS Franco Bradamante Trieste University and INFN on behalf of the Collaboration IWHSS 09 International Workshop on Hadron Structure and Spectroscopy March 30–April 1, 2009, Schloss Waldthausen, Mainz

OUTLOOK the COMPASS experiment results from 6LiD data unpolarised azimuthal asymmetries Collins asymmetry Sivers asymmetry other TMD asymmetries first results from NH3 data Collins and Sivers asymmetries conclusions Collins and Sivers mechanisms have been firmly established IWHSS09 F. Bradamante

OUTLOOK the COMPASS experiment results from 6LiD data unpolarised azimuthal asymmetries Collins asymmetry Sivers asymmetry other TMD asymmetries first results from NH3 data Collins and Sivers asymmetries conclusions Collins and Sivers mechanisms have been firmly established IWHSS09 F. Bradamante 3

OUTLOOK the COMPASS experiment results from 6LiD data unpolarised azimuthal asymmetries Collins asymmetry Sivers asymmetry other TMD asymmetries first results from NH3 data Collins and Sivers asymmetries conclusions Collins and Sivers mechanisms have been firmly established IWHSS09 F. Bradamante 4

Unpolarised Target SIDIS Cross-Section A Bacchetta, M Diehl, K Goeke, A Metz, P Mulders, M Schlegel (06) 3 independent azimuthal modulations in h , the hadron azimuthal angle in GNS pQCD contributions expected to be important at pt > 1 GeV/c IWHSS09 F. Bradamante

Unpolarised Target SIDIS Cross-Section Cahn effect + Boer-Mulders DF Boer- Mulders x Collins FF + Cahn effect IWHSS09 F. Bradamante

Unpolarised Target SIDIS Cross-Section Cahn effect kinematical effect due to quark intrinsic momentum Boer- Mulders DF leading order DF quark with spin parallel to the nucleon spin in an unpolarised nucleon Cahn effect + Boer-Mulders DF Boer- Mulders x Collins FF + Cahn effect IWHSS09 F. Bradamante 7

Unpolarised Azimuthal Asymmetries data sample: part of the 2004 data collected with L and T target polarisation with both target orientation configurations to cancel possible polarisation effects event selection: Q2>1 (GeV/c)2 0.1<y<0.9 W>5 GeV/c2 0.2 < z < 0.85 0.1 < pT < 1.5 GeV/c final statistics: IWHSS09 F. Bradamante

Unpolarised Azimuthal Asymmetries to extract the asymmetries the azimuthal distributions have to be corrected by the apparatus acceptance  dedicated MC simulations for L and T target polarisation data acceptance initial azimuthal distribution final azimuthal distribution IWHSS09 F. Bradamante 9

Unpolarised Azimuthal Asymmetries to extract the asymmetries the azimuthal distributions have to be corrected by the apparatus acceptance  dedicated MC simulations for L and T target polarisation data acceptance final azimuthal distribution the final azimuthal distributions are fitted with the function: IWHSS09 F. Bradamante

Unpolarised Azimuthal Asymmetries systematic errors evaluated from: compatibility of results with L and T target polarization (different experimental conditions, different MCs) comparison of results obtained using two MCs with different settings for each data set (LEPTO default, standard COMPASS high pt; ~extreme cases) compatibility of results from subsamples corresponding to different periods different target polarizations different geometrical regions for the scattered muon most important contribution IWHSS09 F. Bradamante

Unpolarised Azimuthal Asymmetries sin modulation error bars: statistical errors bands: systematical errors IWHSS09 F. Bradamante

Unpolarised Azimuthal Asymmetries cos modulation IWHSS09 F. Bradamante

Unpolarised Azimuthal Asymmetries cos modulation comparison with theory IWHSS09 F. Bradamante

Unpolarised Azimuthal Asymmetries cos2 modulation IWHSS09 F. Bradamante

Unpolarised Azimuthal Asymmetries cos2 modulation comparison with theory IWHSS09 F. Bradamante

Unpolarised Azimuthal Asymmetries summary positive hadrons negative hadrons error bars: statistical errors only IWHSS09 F. Bradamante

OUTLOOK the COMPASS experiment results from 6LiD data unpolarised azimuthal asymmetries Collins asymmetry Sivers asymmetry other TMD asymmetries first results from NH3 data Collins and Sivers asymmetries conclusions Collins and Sivers mechanisms have been firmly established IWHSS09 F. Bradamante 18

Transversity Distribution Function is chiral-odd: observable effects are given only by the product of Tq (x) and an other chiral-odd function can be measured in SIDIS on a transversely polarised target via “quark polarimetry” “Collins” asymmetry “Collins” Fragmentation Function “two-hadron” asymmetry “Interference” Fragmentation Function Λ polarisation Fragmentation Function of q↑Λ all explored in COMPASS I will concentrate on the first IWHSS09 F. Bradamante

Collins Asymmetry Collins effect  azimuthal distribution of the hadrons produced in C “Collins angle” C = h - s’ = h + S - p h,s’,S azimuthal angles of hadron momentum, of the spin of the fragmenting quark and of the nucleon in the GNS ± refer to the opposite orientation of the transverse spin of the nucleon PT is the target polarisation; DNN is the transverse spin transfer coefficient initial  struck quark from the azimuthal distribution of the hadrons one measures the “Collins Asymmetry” IWHSS09 F. Bradamante

Collins Asymmetry using different targets (p, d, n) and identifying the final hadron one can perform flavour separation unique feature of SIDIS has been measured by the HERMES experiment on proton different from zero for charged hadrons of opposite signe for positive and negative charge particles from the azimuthal distribution of the hadrons one measures the “Collins Asymmetry” IWHSS09 F. Bradamante

Collins Asymmetry - Deuteron data charged hadrons 2004: first results from 2002 data [PRL94 (2005) 202002] confirmed in 2006: final results from 2002-2004 data [NPB765 (2007)31] asymmetries compatible with zero within the statistical errors (syst. errors much smaller) the same with leading h with HERMES results, cancellation between u and d quark contributions IWHSS09 F. Bradamante

Collins Asymmetry - Deuteron data identified hadrons 2007: final results from 2002-2004 data [PLB 673 (2009) 127] again, difficult to see a signal … IWHSS09 F. Bradamante

Collins Asymmetry - Fits to Data new results using HERMES (p) and COMPASS (d) pion data, and BELLE data IWHSS09 F. Bradamante

Collins Asymmetry - Fits to Data new results using HERMES (p) and COMPASS (d) pion data, and BELLE data IWHSS09 F. Bradamante

OUTLOOK the COMPASS experiment results from 6LiD data unpolarised azimuthal asymmetries Collins asymmetry Sivers asymmetry other TMD asymmetries first results from NH3 data Collins and Sivers asymmetries conclusions Collins and Sivers mechanisms have been firmly established IWHSS09 F. Bradamante 26

Sivers Asymmetry the “Sivers DF” the most famous of the TMD parton DF appears in SIDIS as a modulation in the “Sivers angle” S = h - S h azimuthal angle of hadron momentum S azimuthal angle of the spin of the nucleon the “Sivers angle” S and the “Collins angle” C are independent  the Collins and Sivers asymmetries can be disentangled and extracted from the same data in SIDIS on a transversely polarised target the “Sivers DF” the most famous of the TMD parton DF J. Collins, Nucl. Phys. B396 (1993) 161 F. Baldrachini, N.S. Craigie, V. Roberto, M. Socolovsky, Fortschritte der Physik 29 (1981) 505 the Sivers asymmetry has been measured by the HERMES experiment on proton different from zero for positive charge particles compatible with zero for negative charge particles IWHSS09 F. Bradamante

Sivers Asymmetry - Deuteron data charged hadrons 2004: first results from 2002 data [PRL94 (2005) 202002] confirmed in 2006: final results from 2002-2004 data [NPB765 (2007)31] asymmetries compatible with zero within the statistical errors (systematic errors much smaller) cancellation between u and d quark contributions in the dueteron IWHSS09 F. Bradamante

Sivers Asymmetry - Deuteron data identified hadrons 2007: final results from 2002-2004 data [PLB 673 (2009) 127] IWHSS09 F. Bradamante

Sivers Asymmetry - Fits to Data new results using HERMES (p) and COMPASS (d) pion and kaon data IWHSS09 F. Bradamante

OUTLOOK the COMPASS experiment results from 6LiD data unpolarised azimuthal asymmetries Collins asymmetry Sivers asymmetry other TMD asymmetries first results from NH3 data Collins and Sivers asymmetries conclusions Collins and Sivers mechanisms have been firmly established IWHSS09 F. Bradamante 31

SIDIS cross-section 18 structure functions 8 modulations Sivers A Bacchetta, M Diehl, K Goeke, A Metz, P Mulders, M Schlegel (06) Sivers Collins 8 modulations (4 LO) all measured by COMPASS on deuteron IWHSS09 F. Bradamante

again cancellation between the u and d quarks in the deuteron ? one example … beam polarization g1T is chiral-even, T-even, leading twist again cancellation between the u and d quarks in the deuteron ? all these asymmetries are compatible with zero; same result for charged pions and kaons IWHSS09 F. Bradamante

Conclusions from COMPASS Deuteron Data the unpolarized azimuthal asymmetries are different from zero and different for positive and negative hadrons all the transverse spin asymmetries have been measured to be compatible with zero from HERMES, COMPASS, and BELLE data a consistent picture first fits extractions of the transversity and Sivers DFs and Collins FF predictions still, a lot of expectation for the higher energy COMPASS proton data IWHSS09 F. Bradamante

OUTLOOK the COMPASS experiment results from 6LiD data unpolarised azimuthal asymmetries Collins asymmetry Sivers asymmetry other TMD asymmetries first results from NH3 data Collins and Sivers asymmetries conclusions Collins and Sivers mechanisms have been firmly established IWHSS09 F. Bradamante 35

2007 run: transversely polarized NH3 target data taking: May to November equally shared between transverse and longitudinal transverse polarization data taking: Period “Weeks” Target Polarization 1 25 / 26 - + - / + - + 2 27 / 28 3 30 / 31 + - + / - + - 4 39 / 40 5 41 / 42a 6 42b / 43 data used for these results several stability tests have been performed detectors and triggers performances, event reconstruction, K0 reconstruction, distributions of kinematical variables (zvtx, Em’, fm’, xBj Q2,y, W, Ehad, fhadLab , qhadLab , fhadGNS , qhadGNS , pt) ~20% of the total collected data has been used IWHSS09 F. Bradamante

Collins asymmetry – proton data integrated over x and z at small x, the asymmetries are compatible with zero in the valence region the asymmetries are different from zero, of opposite sign for positive and negative hadrons, and have the same strength and sign as HERMES statistical errors only; systematic errors ~ 0.3 sstat IWHSS09 F. Bradamante

Collins asymmetry – proton data comparison with M. Anselmino et al. predictions IWHSS09 F. Bradamante

Sivers asymmetry – proton data statistical errors only; systematic errors ~ 0.5 sstat the measured symmetries are small, compatible with zero IWHSS09 F. Bradamante

Sivers asymmetry – proton data comparison with the most recent predictions from M. Anselmino et al. IWHSS09 F. Bradamante

Sivers asymmetry – proton data comparison with predictions from S.Arnold, A.V.Efremov, K.Goeke, M.Schlegel and P.Schweitzer arXiv:0805.2137 IWHSS09 F. Bradamante

CONCLUSIONS new COMPASS results for unpolarised hadron asymmetries on deuteron for positive and negative hadrons Collins and Sivers asymmetries on protons Collins asymmetry different from zero the effect is there at COMPASS energies! Sivers asymmetry: smaller, compatible with zero to be understood near future: analysis of the whole 2007 proton data sample longer term: transverse spin physics is one of the items being spelled out for the future COMPASS program the study of transverse spin effects needs further precise measurements and the COMPASS facility is the only place where SIDIS can be measured at high energy IWHSS09 F. Bradamante

CONCLUSIONS one full year of data taking (>2009) the study of transverse spin effects needs further precise measurements and the COMPASS facility is the only place where SIDIS can be measured at high energy IWHSS09 F. Bradamante

THANK YOU !

COMPASS

COMPASS fixed target experiment at the CERN SPS broad physics programme data taking since 2002: muon beam deuteron (6LiD) polarised target 2002 2003 2004 L/T target polarisation 4:1 2006 L target polarisation only proton (NH3) polarised target 2007 L /T target polarisation 1:1 hadron beam LH target 2008 muon beam: 160 GeV/c longitudinal polarisation -80% intensity 2·108 µ+/spill (4.8s/16.2s)

COMPASS high energy beam large angular acceptance broad kinematical range two stages spectrometer Large Angle Spectrometer (SM1) Small Angle Spectrometer (SM2) variety of tracking detectors to cope with different particle flux from θ = 0 to θ ≈ 200 mrad SciFi Silicon Micromegas GEMs Straws SDC MWPC W45 MuonWall SM2 E/HCAL E/HCAL SM1 MuonWall Polarised Target calorimetry, PID RICH RICH detector m beam 47

COMPASS 48 radiator C4F10 threshold: p ~2 GeV/c K ~ 10 GeV/c Polarised Target RICH RICH detector m beam 48

COMPASS Polarised Target m beam 49

The Target System solid state target operated in frozen spin mode 2002-2004: 6LiD (polarised deuteron) dilution factor f = 0.38 polarization PT = 50% two 60 cm long cells with opposite polarisation (systematics) 2006: PTM replaced with the large acceptance COMPASS magnet (180 mrad) 2 target cells  3 target cells 2007: NH3 (polarised protons) dilution factor f = 0.14 polarization PT = 90% during data taking with transverse polarisation, polarisation reversal in the cells after ~ 4-5 days

SIDIS event selection and kinematics 2002-2004 data DIS cuts: Q2>1 (GeV/c)2 0.1<y<0.9 W>5 GeV/c2 hadrons energy deposit in HCALs>Thr. ( ~5 GeV ) pT>0.1 GeV/c z>0.2 (all), z>0.25 (leading)

Sivers Asymmetry IWHSS09 F. Bradamante

Collins Asymmetry IWHSS09 F. Bradamante

unpolarised target SIDIS cross-section Collins and Sivers mechanisms have been firmly established EMC, J.J. Aubert et al., Phys. Lett. 130B (1983) 118 Charged hadrons, Muons of energy 120 and 280 GeV, hydrogen target IWHSS09 F. Bradamante

COMPASS proton data IWHSS09 F. Bradamante

SIDIS kinematics IWHSS09 F. Bradamante

SIDIS kinematics IWHSS09 F. Bradamante

Collins asymmetry for pions and kaons preliminary 2002-2004 data proton (2002-05  DIS07) 2003-2004 data deuteron COMPASS sign convention IWHSS09 F. Bradamante

Sivers asymmetry for pions and kaons d data preliminary 2002-2004 data proton (2002-05  DIS07) 2003-2004 data deuteron COMPASS sign convention IWHSS09 F. Bradamante

Collins asymmetry– deuteron h± data preliminary 2002-2004 naïve interpretation of the data (parton model, valence region) proton data  unfavored Collins FF ~ – favored Collins FF at variance with unpol case u quark dominance (d quark DF ~ unconstrained) deuteron data refs! some (small) effect expected even if  cancellation between Tu (x) and Td (x) access to Td (x) IWHSS09 F. Bradamante 60

Sivers asymmetry naïve interpretation of the data (parton model, valence region) proton data asymmetry for p+ > 0, asymmetry for p- ≈ 0  Sivers DF for d-quark ≈ - 2 Sivers DF for u-quark 2 deuteron data preliminary 2002-2004 refs! the measured asymmetries compatible with zero suggest IWHSS09 F. Bradamante 61

Extraction of the asymmetries the Collins and Sivers asymmetries can be disentangled and extracted from the same data in SIDIS on a transversely polarised target 2002-2004 data NPB765(2007)31 to extract the Collins and Sivers asymmetries the “ratio product” quantities measured in each angular bin -reduces acceptance variation effects - at first order, spin independent effects cancel out are fitted separately with the functions check: fit of with (“2D fit”) J. Collins, Nucl. Phys. B396 (1993) 161 F. Baldrachini, N.S. Craigie, V. Roberto, M. Socolovsky, Fortschritte der Physik 29 (1981) 505 confirmed by 8 parameter fit IWHSS09 F. Bradamante 62

Collins Fragmentation Function measurable in e+e– annihilation from the correlation between the azimuthal angles of p’s first attempts to measured it from e+e– annihilation using LEP data in the last years BELLE measurements the asymmetries are different from zero and the Collins FF can be measured spin2004 ref? IWHSS09 F. Bradamante 63

L polarimetry systematic errors not larger than statistical errors RICH ID not used yet; someother improvement in selection still foreseen IWHSS09 F. Bradamante

two-hadron asymmetries IWHSS09 F. Bradamante 65

two-hadron asymmetries 2002-2004 data all +/- pairs x preliminary 2002-2004 data all +/- pairs deuteron target deuteron COMPASS kin also the two hadron asymmetry ~ 0 HERMES: positive similar results with identified hadrons ordering in pTh, z A. Bacchetta, M. Radici PRD74(2006)114007 model for DiFF IWHSS09 F. Bradamante 66

two-hadron asymmetries identified hadrons on the deuteron target p+p– K+p– IWHSS09 F. Bradamante 67

two-hadron asymmetries IWHSS09 F. Bradamante

Transversity Distribution Function three quark Distribution Functions are necessary to describe the structure of the nucleon at leading order q(x) f1q (x) q=uv, vv, qsea unpolarised DF quark with momentum xP in a nucleon well known – unpolarised DIS Dq(x) g1q(x) helicity DF quark with spin parallel to the nucleon spin in a longitudinally polarised nucleon known – polarised DIS DTq(x) = q↑↑(x) - q↑↓(x) h1q(x), dq(x), dTq(x) transversity DF quark with spin parallel to the nucleon spin in a transversely polarised nucleon largely unknown ALL 3 OF EQUAL IMPORTANCE IWHSS09 F. Bradamante

unknown for several years … Collins asymmetry asimmetry in the azimuthal distribution of the hadrons produced in on transversely polarised nucleons Collins effect (J. Collins, 93) a quark moving “horizontally” and polarized “upwards” would emit the leading meson preferentially on the “left” side of the jet  the fragmentation function has a spin dependent part “Collins Fragmentation Function” the chiral-odd partner of the transversity DF in SIDIS unknown for several years … measurable in e+e– annihilation from the correlation between the azimuthal angles of p’s first attempts using LEP data; in the last years BELLE measurements: the asymmetries are different from zero and the Collins FF can be measured IWHSS09 F. Bradamante

Transversity Distribution Function DTq(x) contribution of the quarks to the transverse spin of the nucleon properties: is chiral-odd: decouples from incl DIS because helicity of quark must flip probes the relativistic nature of quark dynamics no contribution from the gluons  simple Q2 evolution positivity (Soffer) bound first moments: tensor charge sum rule for transverse spin in parton model Bakker, Leader, Trueman, PRD 70 (04) IWHSS09 F. Bradamante