Study of the hadron structure using the polarised Drell-Yan process at COMPASS Márcia Quaresma, LIP - Lisbon on behalf of the COMPASS collaboration 7 th July 6, MENU 6 Kyoto COMPASS CERN/FIS-NUC/7/5 Márcia Quaresma (LIP) MENU 6 7 th July 6 / 3
COMPASS experiment at CERN COmmon Muon Proton Apparatus for Structure and Spectroscopy Fixed target experiment at the end of M SPS beam line Around 4 collaborators from 3 countries and institutes Data taking since Márcia Quaresma (LIP) MENU 6 7 th July 6 / 3
COMPASS general purpose spectrometer Two stages spectrometer, wide angular acceptance, ±8 mrad About 35 detector planes Particles identification: Calorimeters, RICH and µ Filters Muon and hadron beams Polarised target (longitudinally and transversely polarised NH 3 and 6 LiD) Márcia Quaresma (LIP) MENU 6 7 th July 6 3 / 3
Polarised Target In transverse mode polarisation reversal every weeks Three target cells - SIDIS case st sub-period 3 cm 5 cm 6 cm 5 cm 3 cm nd sub-period Two target cells - DY case st sub-period 55 cm cm 55 cm nd sub-period NH 3 polarised target Polarisation 8% dilution factor 5% in SIDIS, % in DY Márcia Quaresma (LIP) MENU 6 7 th July 6 4 / 3
Nucleon spin decomposition = Σ + G + L q + L g Presently the contribution from the quarks to the proton spin is Σ (Q = 3(GeV/c ) ) =.3 ±.(stat) ±.(evol) The gluon s contribution seems small but with large uncertainties, more data is needed in a different phase space to constrain it How to address the orbital angular momentum? From D to 3D How are the charges distributed inside the nucleon? The 3D nucleon mapping in the momentum space is described by the Transverse Momentum Dependent (TMD) Parton Distribution Functions (PDFs) Márcia Quaresma (LIP) MENU 6 7 th July 6 5 / 3
Nucleon Structure - TMD PDFs The nucleon structure at leading twist QCD, taking into account the quarks intrinsic transverse momentum, k T, is described by 8 TMD PDFs for each quark flavour Quark longitudinally polarised unpolarised transversely polarised f h unpolarised unpolarised PDF g h L Nucleon longitudinally polarised helicity f T g T h transversely polarised worm-gear T h T Sivers transversity Boer-Mulders worm-gear L pretzelosity k T p l Contributions of f, g and h survive the integration over k T The other 5 distributions describe correlations between the quark s k T, its spin and the nucleon s spin Most of these TMD PDFs can be accessed in COMPASS from PDF γ * SIDIS: PDF FF FF SIDIS and DY lʹ h X π p PDF PDF X γ * DY: PDF PDF X l + l Márcia Quaresma (LIP) MENU 6 7 th July 6 6 / 3
Single polarised DY Ha(Pa) s = (P a + P b ) ū(ka) u(kb) γ (q) l (l) l + (l ) X x a(b) = q /(P a(b) q) x F = x a x b Hb(Pb, S) Q = q = M µµ = sx a x b q T = k T a + k T b DY is an excellent tool to access TMD PDFs at COMPASS: No fragmentation functions involved, but the convolutions of PDFs. All the TMD PDFs are expected to be sizeable in the valence quark region - dominant region when π is used as beam QCD TMD approach valid for Q (M µµ > 4 GeV/c ) p T GeV/c Very low cross-section High intensity beam Closed spectrometer adding a hadron absorber Márcia Quaresma (LIP) MENU 6 7 th July 6 7 / 3
Polarised DY programme π beam at 9 GeV/c two polarised NH 3 target cells Hadron absorber downstream of polarised target Stops hadrons and non interacting beam Degrades resolutions thin lithium foil in the dowstream part of the absorber to stop the slow neutrons produced in the absorber and reduce the radiation level in the first detectors Al and W targets unpolarised DY studies See Takahiro Sawada s talk Márcia Quaresma (LIP) MENU 6 7 th July 6 8 / 3
DY and SIDIS cross sections in terms of leading twist asymmetries dσ DY ( + cos (θ) + sin (θ)a cos(φ) UU cos(φ)) + S T [( + cos(θ))a sin(φ S ) sin(φ S ) + sin (θ) (A sin(φ+φ S ) sin(φ + φ S ) + A sin(φ φ S ) sin(φ φ S ))] y S T x y l + x φ φ S q T q θ z H A z H B H A l dσ SIDIS + ε cos(φ h )A cos(φ h) + S UU T [sin(φ h φ S )A sin(φ h φ S ) + ε sin(φ h + φ S )A sin(φ h+φ S ) +ε sin(3φ h φ S )A sin(3φ h φ S ) ] l lʹ S T φ S P ht P h φ h hadron plane lepton plane y x z Márcia Quaresma (LIP) MENU 6 7 th July 6 9 / 3
Leading twist asymmetries in DY and SIDIS - TMD PDFs The same nucleon PDFs can be accessed through the two processes l lʹ π PDF X γ * l γ * p PDF X DY: PDF PDF l + p PDF FF h X SIDIS: PDF FF A cos(φ) UU A sin(φ S ) A sin(φ φ S ) A sin(φ+φ S ) h q,π h q,p Boer-Mulders A cos(φ h) UU h q H h q f q,π f q T,p Sivers A sin(φ h φ S ) h q,π hq,p transversity A sin(φ h+φ S ) h q,π h q T,p pretzelosity A sin(3φ h φ S ) f q T Dh q h q H h q h q T H h q Márcia Quaresma (LIP) MENU 6 7 th July 6 / 3
Sivers function Theoretical prediction of the Sivers sign change f T (x, k T ) DY = f T (x, k T ) SIDIS Crucial test of the QCD TMD approach COMPASS SIDIS proton data from 7 and Comparison between COMPASS and HERMES results Difference points to a possible Q evolution effect p A Siv p A Siv COMPASS π + x<.3. COMPASS π + x>.3 HERMES π + PRL 3 (9).5.5.. + COMPASS K x<.3 + COMPASS K x>.3 + HERMES K PRL 3 (9) (GeV/c) Q DY and SIDIS phase-space overlap at large Q Avoid evolution effects Q vs. x at COMPASS SIDIS ( proton data) Drell-Yan (MC).9.8.7.6.5.4.3.. x.5 z (COMPASS PLB 744 (5) 5).5.5 p h (GeV/c) T -3 - - x. Márcia Quaresma (LIP) MENU 6 7 th July 6 / 3
SIDIS Data MultiDimensional Analysis Sivers asymmetry ) S - ϕ h sin(ϕ A.5 D analysis: 4 Q ranges, each divided in bins of x, z, p Th, W Contribution for evolution studies + h <Q /(GeV/c) <4 h - z >. COMPASS preliminary Proton data -.5 )4<Q sin(ϕ - ϕ A h S.5 -.5 /(GeV/c) <6.5 )6.5<Q sin(ϕ - ϕ A h S.5 /(GeV/c) <6.5.5.5 -.5 -.5 -.5 -.5 ) S - ϕ h sin(ϕ A.5 -.5 - Q /(GeV/c) >6 -.5 -.5 x..4.6.8..4.6.8.5 -.5 z.5.5.5 -.5.5.5 p (GeV/c) T DY/SIDIS overlapping region: Q > 6 (GeV/c) M µµ > 4 GeV/c 6 8 4 6 8 4 W (GeV/c ) Márcia Quaresma (LIP) MENU 6 7 th July 6 / 3
Polarised DY data-taking at COMPASS 9 4 5 π beam at 9 GeV/c C H 4 cells unpolarised NH 3 cells unpolarised NH 3 cells transversely polarised I beam.6 7 π/s I beam 6 7 π/s I beam 8. 7 π/s 3 days data taking 7 days of stable data taking Feasibility of the experiment First unpolarised high mass DY data 4 months of stable data taking First polarised DY data Márcia Quaresma (LIP) MENU 6 7 th July 6 3 / 3
Drell-Yan measurement - 5 run Test production of 3/9 of the data collected in 5 Four mass ranges: < Mµµ < GeV/c < Mµµ <.5 GeV/c.5 < Mµµ < 4 GeV/c J/ψ region 4 < Mµµ < 9 GeV/c High mass range (GeV/c) Q Q > 3 COMPASS ongoing analysis Drell-Yan NH 3 5 data (~3%) 6<Q <8.9.8.7.6 (a.u.) d N/dxdQ 6.5<Q <6 4.<Q <6.5 <Q <4..5.4.3.. 3 x N Márcia Quaresma (LIP) MENU 6 7 th July 6 4 / 3
Dimuon mass distribution dn/dm µµ 6 COMPASS ongoing analysis 5 Drell-Yan NH 3 5 data (~3%) 4 3 Total J/ψ Continuum ψ' 3 4 5 6 7 8 Continuum has several contributions: M µµ (GeV/c ) DY Open Charm (D D decays into correlated muons) Combinatorial Background (π and K decays into uncorrelated muons) small contamination for M µµ > 4 GeV/c (mainly from ψ ) Márcia Quaresma (LIP) MENU 6 7 th July 6 5 / 3
J/ψ and High mass regions dn/dz 35 3 5 5 5 J/ψ region 3 45 COMPASS ongoing analysis 4 Drell-Yan 5 data (~3%).5 < M µµ /(GeV/c ) < 4 3 5 5 5 5 z (cm) dn/dz.6.4..8.6.4. High mass region 3 COMPASS ongoing analysis.8 Drell-Yan 5 data (~3%) 4 < M µµ /(GeV/c ) < 9 3 5 5 5 5 z (cm) All the targets are well separated ( NH 3 cells, Al and W) For the asymmetry studies only the events from the NH 3 target cells are used 3% of statistics from all targets The optimization of selection criteria is still under study Márcia Quaresma (LIP) MENU 6 7 th July 6 6 / 3
J/ψ region Much larger statistics wrt the high mass region, 7k pairs but with several contributions Valence quarks region x N.9.8.7.6.5.4.3.. COMPASS ongoing analysis Drell-Yan NH 3 5 data (~3%).5 < M µµ /(GeV/c ) < 4...3.4.5.6.7.8.9 x π.9.8.7.6.5.4.3.. N/dx π dx N (a.u.) d dn/dq T 3 8 6 4 8 6 4 COMPASS ongoing analysis Drell-Yan NH 3 5 data (~3%).5 < M µµ /(GeV/c ) < 4 dn/dx F 4 8 6 4 3 COMPASS ongoing analysis Drell-Yan NH 3 5 data (~3%).5 < M µµ /(GeV/c ) < 4 3 4 5 q (GeV/c ) T...4.6.8 x F Márcia Quaresma (LIP) MENU 6 7 th July 6 7 / 3
Sivers asymmetry prediction for J/ψ Recent publication from Anselmino et al, arxiv:67.75v (4 July) Significant Sivers asymmetry in COMPASS kinematic coverage x F. and q T. GeV/c Márcia Quaresma (LIP) MENU 6 7 th July 6 8 / 3
High mass region: 4 < M µµ < 9 GeV/c Valence quarks region q T GeV/c x F.3 x N.9.8.7.6.5.4.3.. COMPASS ongoing analysis Drell-Yan NH 3 5 data (~3%) 4 < M µµ /(GeV/c ) < 9...3.4.5.6.7.8.9 x π.9.8.7.6.5.4.3.. N/dx π dx N (a.u.) d dn/dq T 45 4 35 3 COMPASS ongoing analysis Drell-Yan NH 3 5 data (~3%) 4 < M µµ /(GeV/c ) < 9 dn/dx F 5 COMPASS ongoing analysis Drell-Yan NH 3 5 data (~3%) 4 < M µµ /(GeV/c ) < 9 5 5 5 5 5 3 4 5 q (GeV/c ) T...4.6.8 x F Márcia Quaresma (LIP) MENU 6 7 th July 6 9 / 3
Asymmetries uncertainties in high mass range A cos(φ) UU ) CS cos(φ A U.. h q,π h q,p (Boer-Mulders) COMPASS Drell-Yan NH3 5 data (~3%) Estimated uncertainties (ongoing analysis) 4 < M µµ /(GeV/c ) < 9 A sin(φ φ S ) ) S φ CS sin(φ A T.5 h q,π hq,p (transversity) COMPASS Drell-Yan NH3 5 data (~3%) Estimated uncertainties (ongoing analysis) 4 < M µµ /(GeV/c ) < 9 ) S sin(φ A T.. A sin(φ S ).5 x N f q,π f q T,p (Sivers) COMPASS Drell-Yan NH3 5 data (~3%) Estimated uncertainties (ongoing analysis) 4 < M µµ /(GeV/c ) < 9.5 A sin(φ+φ S ) ) S φ + CS sin(φ A T.5 x N h q,π h q T,p (pretzelosity) COMPASS Drell-Yan NH3 5 data (~3%) Estimated uncertainties (ongoing analysis) 4 < M µµ /(GeV/c ) < 9.5.5 x N x N Márcia Quaresma (LIP) MENU 6 7 th July 6 / 3
Future polarised DY programmes Worldwide plans to study TMD PDFs via the polarised DY process Facility type timeline COMPASS fixed target, π p 5 + 8 Fermilab (SeaQuest) fixed target, p p > 6 RHIC (STAR*, PHENIX) collider, p p > 6 J-PARC fixed target, p p > 8 FAIR (Panda) collider, p p > 8 JINR (NICA) collider, p p > 8 *STAR results on TSA in p + p W ± /Z (PRL 6, 33(6)) corresponds to a very high Q 8 (9 ) (GeV/c ) and favour theoretical models that include a change of sign for the Sivers function, if TMD evolution effects are small. Márcia Quaresma (LIP) MENU 6 7 th July 6 / 3
Projected Sivers asymmetry uncertainty 5 data + 8 assuming 4 days S sinφ A T..5 COMPASS DY (5 + 4 days) P. Sun & F. Yuan, PRD 88, 4 (3) M. Echevarria et al, PRD 89, 743 (4)..5.5..4...4.6.8 x F.3% statistical uncertainty Estimated systematic smaller than statistical Márcia Quaresma (LIP) MENU 6 7 th July 6 / 3
Final remarks The first ever polarised DY measurement using π beam and proton polarised target was done by COMPASS in 5. These data are being analysed. Unique opportunity to extract the TMD PDFs from both SIDIS and DY in the same experiment. DY results will be fundamental to test the Sivers sign change prediction. The possibility to have a second year of polarised DY data taking in 8 is about to be approved. First COMPASS DY polarised data results on the asymmetries will be ready soon Márcia Quaresma (LIP) MENU 6 7 th July 6 3 / 3