COMPA Drell-Yan Results and Future Prospect Jen-Chieh Peng University of Illinois at Urbana-Champaign IN Workshop on he Flavor tructure of Nucleon ea Oct. 2-13, 2017 On behalf of COMPA Collaboration (hanks to. Platchkov, W.C. Chang, B. Parsamyan for some slides) 1
he Drell-Yan Process ~1300 citations to date 2
Complimentality between DI and Drell-Yan DI Drell-Yan Ann.Rev.Nucl. Part. ci. 49 (1999) 217 Both DI and Drell-Yan process are tools to probe the quark and antiquark structure in hadrons (factorization, universality) 3 3
Polarized Drell-Yan to Probe ransverse Momentum Dependent Quark Distributions (45 years later!) 4
ransverse Momentum Dependent (MD) Quark Distributions Leading-wist Quark Distributions ( A total of eight distributions) hree survive after K integration he other five are transverse momentum (K ) dependent (MD) ransversity ivers function Boer-Mulders function 5
hree parton distributions describing transverse momentum and/or transverse spin 1) ransversity hree transverse quantities: 1) Nucleon transverse spin N 2) Quark transverse spin q s 3) Quark transverse momentum q k hree different correlations hey can be measured in sem-inclusive DI Correlation between 2) ivers function Correlation between and 3) Boer-Mulders function Correlation between s s q N q and and k k N q q 6
hey can also be measured in Drell-Yan process a) Boer-Mulders functions: - Unpolarized Drell-Yan: b) ivers functions: DY N 1 q q q DY 1 q 1 q - ingle transverse spin asymmetry in polarized Drell-Yan: A f ( x ) f ( x ) c) ransversity distributions: DY 1 q 1 q dσ h ( x ) h ( x )cos(2 φ) - Double transverse spin asymmetry in polarized Drell-Yan: A h( x ) h( x ) Drell-Yan does not require knowledge of the fragmentation functions -odd MDs are predicted to change sign from DI to DY (Boer-Mulders and ivers functions) Remains to be tested experimentally! 7
IDI and single-polarized DY x-sections at twist-2 (LO) LO dσ + dxdydzdp dφ dφ A A A A cos2φh UU sin( φh φs ) U sin( φh + φs ) U h sin( 3 φh φs) U H q h 1 1q q h 1 1q q h 1 1q q h 1 1q sin 2φh q h UL 1L 1q f h h D H H within QCD MD-framework: A h H h h + cos2φh 1 ε AUU cos 2 +... dσ dω LO Boer-Mulders ivers ransversity Pretzelosity F + ( 1 cos θc ) 1 2 U cos2ϕc 1 + D 2 A sin U cos2ϕ θ C C 2 sin2ϕc + L sin θc AL sin 2ϕ C sinϕ A sinϕ sin( 2ϕC ϕ ) + A sin ( 2ϕC ϕ ) + D 2 sin θ C sin( 2ϕC + ϕ ) + A sin ( 2ϕ C + ϕ ) ( o θc ) where D = sin θ / 1+ c s 2 sin θ C 2 2 C q & f q h cos( φh φs ) q h MD A PDFs LL g are 1L Dexpected 1q, A to L be g"conditionally" 1 D universal (IDI DY: sign change) 1q q & h MD PDFs are expected to be "genuinely" universal (IDI DY: no sign change) q 1 1 q 1 1 2 h sin 2φh UL ( FUU, ε FUU, L ) + ε A sin 2 φ + ( φ φ ) 2 cos( φh φ) ( ε ) AL ( φh φ) sin ( φh φ) sin ( 3φ φ ) sin h AU sin sin( φh+ φ) + + ε AU + sin( 3φh φ) + ε AU h + L λ φ h h 2 λ 1 ε A 1 cos φ L h LL ( ) IDI φ 5 April 2017 Bakur Parsamyan 8 A A A A A cos2ϕ U sinϕ sin(2 ϕc ϕ ) sin(2ϕ sin 2ϕ L C C C DY h f q q 1, π 1,p q q 1, π 1,p q 1, π q q 1, π 1,p + ϕ ) q q 1, π 1, p h h h h f h h h q 1 L, p
he Boer-Mulders function A long-standing puzzle in Drell-Yan angular distributions Θ and Φ are the decay polar and azimuthal angles of the μ + in the dilepton rest-frame A general expression for Drell-Yan decay angular distributions: 1 dσ 3 2 ν 2 = 1+ λ cos θ + µ sin 2θ cosφ+ sin θ cos 2φ σ dω 4π 2 Lam-ung relation: 1 λ = 2ν Reflect the spin-1/2 nature of quarks (analog of the Callan-Gross relation in DI) Insensitive to QCD - corrections 9
Decay angular distributions in pion-induced Drell-Yan 1 dσ 3 2 ν 2 = 1+ λ cos θ + µ sin 2θ cosφ+ sin θ cos 2φ σ dω 4π 2 NA10 π - +W Z. Phys. 37 (1988) 545 Dashed curves are from pqcd calculations ν 0 and ν increases with p 10
h 1 h 1 h 1 Boer-Mulders function h 1 represents a correlation between quark's k and transverse spin in an unpolarized hadron (analogous to Collins function) is a time-reversal odd, chiral-odd MD parton distribution h 1 h 1 can lead to an azimuthal dependence with ν f f 1 1 α M M 2 h ( xk, ) = c e f( x) 2 C H α k 1 H 2 2 1 π k + MC ν = 16κ QM 2 2 C 1 2 2 2 ( Q + 4 MC) κ 1 =0.47, M C =2.3 GeV Boer, PRD 60 (1999) 014012 ν>0 implies valence BM functions for pion and nucleon have same signs 11
Boer-Mulders function from cos2φ Distribution in Drell-Yan DY 2 1 cos sin 2 cos dσ + λ θ + µ θ φ ν + 2 θ ν h ( x ) h ( x ) 1 q 1 2 sin cos 2 q φ Lingyan Zhu et al., PRL 99 (2007) 082301; PRL 102 (2009) 182001 Extraction of Boer-Mulders function h 1 : ν(π - W µ + µ - X)~ [valence h 1 (π)] * [valence h 1 (p)] ν(pd µ+µ-x)~ [valence h 1 (p)] * [sea h 1 (p)] First extraction of BM functions in Drell-Yan has been obtained 12
Extraction of ivers function from IDI data orino group, Anselmino et al., Eur. Phys. J. A39 (2009) 89 (2013) u and d- quark ivers functions have opposite signs ea-quark ivers functions are non-zero (from K + data) 13
IDI As (ivers) dσ ( FUU, + ε FUU, L ) 1 +... dxdydzdp dφ dφ 2 h sin( φh φ) AU sin ( φh φ ) sin( φh+ φ ) + ε AU sin ( φh + φ ) sin( 3φh φ ) + + ε AU sin ( 3φh φ ) sinφ + 2ε( 1+ ε) AU sinφ sin( 2φh φ ) 2ε( 1 ε) AU sin ( 2φh φ) + + 2 cos( φh φ ) ( 1 ε ) AL cos( φh φ ) cosφ + λ + 2ε( 1 ε) AL cosφ cos( 2φh φ ) + 2ε( 1 ε) AL cos( 2φh φ) Measured on proton and deuteron Gluon ivers paper: submitted to PLB CERN-EP/2017-003, hep-ex/1701.02453 A sin( φh φs) q h U 1 1q f D PLB 744 (2015) 250 [ ] A twist-2 COMPA proton ivers effect at COMPA is slightly smaller w.r.t HERME results (Q 2 is different by a factor of ~2-3) Q 2 -evolution? Intriguing result!. M. Aybat, A. Prokudin,. C. Rogers PRL 108 (2012) 242003 M. Anselmino, M. Boglione,. Melis PRD 86 (2012) 014028 5 April 2017 Bakur Parsamyan 14 z x
Outstanding questions on ivers function Does ivers function change sign between DI and Drell-Yan? ign and magnitude of the sea-quark ivers functions? Q 2 -evolution of the ivers function? 15
ivers function on the lattice Musch, Haegler, Engelhardt, Negele & chaeffer, PRD 85 (2012) 094510 IDI Drell-Yan As the vertical gauge link goes from (IDI) to - (Drell-Yan), the sign of ivers function changes 16
COMPA Collaboration (Common Muon and Proton Apparatus for tructure and pectroscopy) 24 institutions from 13 countries nearly 250 physicists Fixed-target experiment at P north area Physics programs: Nucleon spin and partonic structures Hadron spectroscopy 17
CERN accelerator complex COMPA 18
COMPA is an old experiment - already two phases COMPA I : 2002 2011 Longitudinally polarized DI and IDI ransversely polarized IDI Muons Beam Polarized target Hadron spectroscopy and chiral dynamics Hadrons COMPA II : 2012 2018 Primakoff studies (2012) Hadrons DVC and Meson-production (2016, 2017) Polarized Drell-Yan process (2015, 2018) Muons Hadrons Beam 19
CERN M2 beam line Beams at COMPA: muons: µ +, µ, P protons hadrons: h + (p, π +, K + ), h (π, K p) Abbon et al., NIM A779 (2015) 69-115 electrons: e 1 or 2 spills of 5 s every 33-48 sec Unique combination of positive and negative beams of different species COMPA arget 20
Beam particle identification Differential Cerenkov counters (CEDARs) two CEDARs CEDAR principle he two CEDARs are located upstream of the COMPA experimental hall 21
COMPA : a large, fixed-target, versatile setup COMPA apparatus Built for detecting several particles in the final state wo spectrometers: mall-angle and Large-angle large and flat acceptance Polarized target Energy: 100 225 GeV Intensity: up to 10 9 /spill Large acceptance, PID detectors everal particles in the final state Large (1.2 m) polarized target Minor changes to the setup switch between various physics programs ame spectrometer used for IDI and Drell-Yan measurements 22
COMPA polarized target Polarized target with two 55 cm long cells (for DY); three cells for IDI uperconducting magnets: solenoid + dipole arget filled with ammonia (NH 3 ) solid beads; also available: 6 LiD Polarization in longitudinal mode, data is taken in transverse mode Polarization is periodically reversed 23
Hadron Absorber & Nuclear argets Absorber: 236 cm long, made of Al 2 O 3. Radiation lengths (multiple scattering for µ): x/x 0 = 33.53 Hadronic interaction lengths (stopping power for π): x/λ int = 7.25 7 cm Al target 120 cm W beam dump Al W 24
Hadron Absorber & Nuclear argets 25
COMPA-II ransversely Polarized Drell-Yan Program 26
Dimuon Vertex Distributions (2015 rans.-pol. Drell-Yan Runs) ungsten beam plug NH 3 Al COMPA, PRL 119 (2017) 112002 27
COMPA DY: high mass range Final sample: 35 000 dimuons in HM 5 April 2017 Bakur Parsamyan 28
Kinematic Acceptance (2015 rans.-pol. Drell-Yan Runs) COMPA, PRL 119 (2017) 112002 29
LO 2 dσ αem = 4 2 σ d qdω Fq ransverse pin Asymmetries in rans.-pol. Drell-Yan: ivers LO U ( LO cos 2ϕ 1 D A cos2 ) U ϕ A Density f ivers f 1 π 1 + 2 + [sin θ ] sinϕ LO 2 ) A sinϕ + D 2 sin(2 ϕ ϕ ) sin(2 ) [sin ] + A ϕ ϕ ϕ + ϕ θ sinφs ( sin(2ϕ ϕ ) sin( + A ) p COMPA, PRL 119 (2017) 112002 30
ransverse pin Asymmetries in rans.-pol. Drell-Yan: ransversity LO 2 dσ αem = 4 2 σ d qdω Fq LO U ( LO cos 2ϕ 1 D A cos2 ) U ϕ A sin( 2 φ φs ) BM h ransversity h + 2 + [sin θ ] sinϕ LO 2 ) A sinϕ + D 2 sin(2 ϕ ϕ ) sin(2 ) [sin ] + A ϕ ϕ ϕ + ϕ θ 1 ( sin(2ϕ ϕ ) sin( + A ) π 1 p COMPA, PRL 119 (2017) 112002 31
ransverse pin Asymmetries in rans.-pol. Drell-Yan: Pretzelosity LO 2 dσ αem = 4 2 σ d qdω Fq LO U ( LO cos 2ϕ 1 D A cos2 ) U ϕ A BM h Pretzelosity h sin(2 φ+ φs ) 1 π 1 + 2 + [sin θ ] sinϕ LO 2 ) A sinϕ + D 2 sin(2 ϕ ϕ ) sin(2 ) [sin ] + A ϕ ϕ ϕ + ϕ θ ( sin(2ϕ ϕ ) sin( + A ) p COMPA, PRL 119 (2017) 112002 32
dσ ( FUU, + ε FUU, L ) 1 +... dxdydzdp dφ dφ 2 h IDI and DY As at COMPA ( φh φ) sin ( φh φ) ( φ φ ) sin( φh φ) AU sin sin( φh+ φ ) + ε AU + sin( 3φh φ) + + ε AU sin 3 h sinφ + 2ε( 1+ ε) AU sinφ sin( 2φh φ) + 2ε ( 1 ε ) AU sin ( 2φh φ) + 2 ( θc ) 1 FU 1+ cos 1 +... sinϕ A sinϕ sin( 2ϕC ϕ ) A sin ( 2ϕC ϕ ) + D 2 sin θ C sin2 ( ϕc + ϕ ) A sin ( 2ϕC ϕ ) + + + sin( ϕc ϕ ) A sin ( ϕc ϕ ) + D [ sin 2θC ] sin( ϕc + ϕ ) + A sn i ( ϕc + ϕ ) COMPA, PRL 119 (2017) 112002 COMPA, PLB 770 (2017) 138 dσ dω LO 33
ivers Asymmetry in Drell-Yan: Hint of ign Change! sign change 0 < q < 1 GeV no sign change s A ϕ = 0.060 ± 0.057( stat.) ± 0.040( sys.) sin COMPA, PRL 119 (2017) 112002 34
Can the existing Drell-Yan data already test the predicted sign-change of B-M function? 1) From IDI data, one deduces that the proton B-M functions are negative for both u and d quarks: ( p) < 0 ; h ( p) < 0, DI, DI h 1, u 1, d 35
Can the existing Drell-Yan data already test the predicted sign-change of B-M function? 1) From IDI data, one deduces that the proton B-M functions are negative for both u and d quarks: h ( p) < 0 ; h ( p) < 0, DI, DI 1, u 1, d 2) From NA10 pion Drell-Yan data, one deduces that the product of the pion valence quark B-M function and the proton valence quark B-M function is positive. Using u-quark dominance, we have: herefore, either or h ( p) * h ( π ) > 0, DY, DY 1, u 1, u a) h ( p) > 0; h ( π ) > 0, DY, DY 1, u 1, u, DY, DY 1, u 1, u ( sign change) b) h ( p) < 0; h ( π ) < 0 ( no sign change) 36
Can the existing Drell-Yan data already test the predicted sign-change of B-M function? 1) From IDI data, one deduces that the proton B-M functions are negative for both u and d quarks: h ( p) < 0 ; h ( p) < 0, DI, DI 1, u 1, d 2) From NA10 pion Drell-Yan data, one deduces that the product of the pion valence quark B-M function and the proton valence quark B-M function is positive. Using u-quark dominance, we have: herefore, either h ( p) * h ( π ) > 0, DY, DY 1, u 1, u a) h ( p) > 0; h ( π ), DY, DY 1, u 1, u, DY, DY 1, u 1, u > 0 ( sign change) or b) h ( p) < 0; h ( π ) < 0 ( no sign change) 3) he crucial measurement is to determine the sign of the pion B-M function in polarized π p D-Y, since the is sensitive to the sign of h ( π )., DY 1, u sin (2 φ C φ ) modulation 37
COMPA DY: Charmonia mass range M. Anselmino et al. arxiv:1607.00275 5 April 2017 Bakur Parsamyan 38
COMPA-II Programs 2014-2018: Commissioning of polarized Drell-Yan experiment started in mid-october 2014. 2015: Polarized Drell-Yan program. 2016-2017: DVC program. 2018: Polarized Drell-Yan program (improved statistics errors of ivers asymmetries are expected). 2020-2024 (under planning) : Polarized 6 LiD target: flavor separation of MD As. Long LH 2 and nuclei targets: un-polarized pioninduced DY. 39
Di-muon sample some results from 2015 Drell-Yan data from 2015: 4.3 8.5 GeV/c 2 x N vs x π distribution Mass spectrum NH3 Al W Vertex distribution Many dimuon events for unpolarized Drell-Yan with pion beam! 40
Experiments at CERN and Fermilab Exp P (GeV) targets Number of D-Y events WA11 175 Be 500 (semi-exclusive) WA39 40 W (H 2 ) 3839 (all beam, M > 2 GeV) NA3 150, 200, 280 Pt (H 2 ) 21600, 4970, 20000 (535, 121, 741) NA10 140, 194, 286 W (D 2 ) ~84400, ~150000, ~45900 (3200, --, 7800) E331/E444 225 C, Cu, W 500 E326 225 W E615 80, 252 W 4060, ~50000 41
42
43
44
Exp P (GeV) targets D-Y events WA39 40 W (H 2 ) NA3 150, 200 Pt 688, 90 Exp P (GeV) targets D-Y events WA39 40 W (H 2 ) NA3 200 Pt 170 Exp P (GeV) targets D-Y events WA39 40 W (H 2 ) NA3 150, 200 Pt 275, 32 E537 125 W, Cu, Be 380 45
46
CERN strategy : support physics other than LHC Physics Beyond Colliders initiative meetings ept. 6/7, 2016; March 1 st 2017 Initiated by CERN DG as part of the CERN s future Exploit the full potential of CERN s accelerator complex; complement the LHC programme et working groups with deliverables for each project that include: evaluation of the physics case detector building and optimizations CERN s uniqueness Physics QCD subgroup (among others) : COMPA++, LHC Fixed-target, other experiments... Conventional beams subgroup comprises studies for RF-separated beams fro COMPA Info at: http://pbc.web.cern.ch// Deliverables due end 2018 In time for the European trategy for Particle Physics update in 2019/2020 47
48
ummary COMPA has successfully collected first ever transversely polarized Drell-Yan data in 2015: ivers asymmetry is found to be above zero at about 1 sigma. DY ivers asymmetry is consistent with the predicted sign change! A second year of polarized DY run in 2018 will provide more stringent test of ivers nonuniversality. New unpolarized DY data on nuclear targets with pion beam are being analyzed Future DY experiments with RF-separated kaon and antiproton beams are under consideration 49