Semi-Inclusive DIS with SoLID on a polarized 3He target Yi Qiang for JP. Chen, H. Gao, X. Qian

Nucleon Structure from Deep Inelastic Scattering Un-polarized Nucleon Structure Function Longitudinal Momentum Distribution Well probed for 50 years over very large kinematic range Longitudinal Polarized Nucleon Structure Functions Since “spin crisis” in 1980s Plotted in fairly large range Transversity: Quark transverse polarization in a transversely polarized nucleon. Equal to g1 when there is no relativistic effect. Chiraly Odd: Can be probed in Semi-Inclusive DIS when convoluted with Collins fragmentation function. New business: recent measurements from HERMES and COMPASS using Hydrogen and Deuteron targets. f1 = g1 = Nucleon Spin Quark Spin h1 =

Longitudinally Polarized Transversely Polarized Leading Twist TMDs Nucleon Spin Quark Spin Quark polarization Un-Polarized Longitudinally Polarized Transversely Polarized Nucleon Polarization U L T h1 = f1 = Boer-Mulder g1 = h1L = Helicity Beside the three parton distribution functions which survive the integration over the quark transverse momentum there are five more transverse momentum dependent distribution functions at leading twist. Here is a table of all the 8 leading twist TMDs, the cartoons on the right hand side of each name tell you the meaning of each TMDs. Again, the virtual photon is supposed to go out of the paper and that’s our longitudinal direction. The red circle and arrow are the quark and it’s spin, and black circle and arrow are the nucleon and it’s spin. The blue arrow shows the momentum direction of the quark. An important information these 5 new TMDs will tell you is the quark oribtal angular motion. They will vanish without orbital angular momentum. In particular, the Sivers function directly tells you the quark orbital angular momentum in a transverse polarized target, Boer-Mulder tells you the correlation of the quark orbital momentum and it’s spin inside a unpolarized nucleon and Pretzelosity, in certain model context, direct measures the relativistic nature of the nucleon. Blue: no kt dependence and T-even Red: T-odd with kt dependence Other: kt dependence and T-even h1T = f 1T = Transversity g1T  = h1T = Sivers Pretzelosity 10/30/2009 Transverse Momentum-dependent Sturecutre of the Nucleon

SIDIS electroproduction of pions Separate TMDs with different angular dependences Sivers angle, effect in distribution function: (fh-fs) = angle of hadron relative to initial quark spin Collins angle, effect in fragmentation function: (fh+fs) = p+(fh-fs’) = angle of hadron relative to final quark spin Pretzelosity angle: 3fh-fs Scattering Plane target angle hadron angle e-e’ plane q No like the pp scattering, you have both Sivers and Collins effects mixed, we can clearly identify the Sivers and Collins effects in the SIDIS pion production based on their different azimuthal distributions. Here is a figure for the SIDIS kinematics. We defined two important azimuthal angles about the Q-vector here, one is the target angle fs, which is the angle of the target spin to the scattering plane and another one is fs, which is the angle of produced hadron to the scattering plane. Here is the projection of the angles if we look into the virtual photon. For the Sivers function, because it is an effect in the distribution function, it has a dependence on the angle between the produced hadron and the initial quark spin. And this angle is fh – fs. For the Collins function, since it’s an effect in the fragmentation process, it will a dependence on the angle between the hadron and the final quark spin. So we got the Collins angle as fh + fs. 10/30/2009 Transverse Momentum-dependent Sturecutre of the Nucleon

Access Leading Twist Parton Distributions through Semi-Inclusive DIS SL, ST: Target Polarization; le: Beam Polarization f1 = f 1T = g1 = g1T  = h1 = h1L = h1T = h1T = Unpolarized Polarized Target Beam and Boer-Mulder Sivers Transversity Pretzelosity Not just transversity and Sivers diftributions, we can actually access all the eight leading twist TMDs in semi-inclusive DIS by using different combinations of beam and target polarizations. During the experiment we will cover in the later half of this talk, we used a transversely polarized 3He target and polarized electron beam. By forming single target spin asymmetry, we can measure Transversity, Sivers and Pretzelosity distribution functions through single target spin asymmetry. The Pretzelosity shows the correletion of the quark transverse spin and it’s momentum inside a transversely polarized nucleon. And in certain model context, it directly measures the relativistic nature of the nucleon. g1T tells us the chance that we find a longitudinally polarized quark inisde a transversely polarized nucleon due to the quark orbital angular momentum. 10/30/2009 Transverse Momentum-dependent Sturecutre of the Nucleon

SIDIS with SoLID Modification based on PVDIS setup Coil Yoke GC HG A C Collimator Polarized 3He Target LC GEMx4 GEMx5 LC: Large angle Calorimeter GC: Gas Cherenkov HG: Heavy Gas Cherenkov A: Aerogel Detector C: Calorimeter TOF

Polarized 3He Target 40 cm polarized 3He target 10 amg: 3x1036 nuclei/s/cm2 with 15 mA beam Successfully run during 6 GeV Hall A transversity experiment 65% polarization with 15 mA beam and 20 min auto spin flip (~ 0.5% relative AFP loss per flip) Faster flipping speed can be achieved by doing field rotation <5s flip time, 1 flip per minute ~0% polarization loss Tread in: real time polarization monitoring Require special treatment for fringe field Shielding Correction coil

Target Collimator Shield high energy electron and photons, which are generated from the target glass wall for the forward angle detector. Target Collimator Large Angle Forward Angle Acceptance Comparison with or without target collimator

Tracking with GEM detectors 5 planes reconfigured from PVDIS GEM detectors Total surface ~ 18 m2 Need to build the first plane Electronics can be shared

Rates and Random Noise on GEM Background rates simulated using GEANT3, with all physics processes (Moller/Mott ... etc). 50 ns ADC gate 1 0.1 0.01 Chamber Rate (kHz/mm2) Radius (cm) LF2 L1 LF3 LF4 F5 F6 L1 LF2 LF3 11 51 40 LF4 F5 F6 49 36 40

“Progressive” Tracking Algorithm The tracking is based on “progressive” method. Starts from a seed, normally a hit in the first chamber. Then it looks for match in the second chamber based on designed momentum and angular coverage. With the two hits, narrow down the search range. Then we look for a smaller region in chamber-3 and same procedure until we goes to the last layer of chamber. After the coarse tracking, a global fitting is implemented to judge the tracks so that we can order tracks by c2/DOF. Then the “similar tracks” will be merged (de-ghosting).

Tracking Performance Assume 200mm spatial resolution and 98% hit efficiency Require 3/4 chambers firing for large angle and 4/5 for small angle Performance with real tracks: Large angle side: (Tracking efficiency > 99%) Multi-tracks: 0.32% No-tracks: 0.49% Small angle side: (Tracking efficiency > 99%) Multi-tracks: 0.28% No-tracks: 0.31 % Performance with pure background (false track): Large angle side: 0.30% Small angle side: 0.29% Speed: (preliminary) 100 Hz with single CPU, can be improved.

Particle Identification Large angle side: 14.5o – 22o Electron only Momentum: 3.5 – 6.0 GeV/c p/e < 1.5 Shashlyk calorimeter Forward angle side: 6.6o – 12o Electron and pion Momentum: 0.9 – 7.0 GeV/c Calorimeter: Pre-shower/Shower splitting Cherenkov and TOF detectors for hadron identification

Hadron Identification Momentum range: 0.9 – 7.0 GeV/c Configuration for only pion identification P (GeV) Gas Cherenkov: CO2@1 atm n = 1.000585, 200 cm N.P.E. ~ 17.4 Heavy Gas Cherenkov: C4F10@1.5 atm n = 1.0021, 30 cm N.P.E ~ 9.3 Aerogel Cherenkov: n = 1.015, 10 cm N.P.E ~ 13.6

Hadron Identification (continued) Configuration for pion and kaon Require TOF to identify kaon below 3.5 GeV/c P (GeV) Gas Cherenkov: CO2@1.3 atm n = 1.00045, 200 cm N.P.E. ~ 13.2 Heavy Gas Cherenkov: C4F10@3 atm n = 1.00419, 30 cm N.P.E ~ 18.5 Aerogel Cherenkov: n = 1.012, 15 cm N.P.E ~ 16.4

Time-of-Flight K/p separation up to 3.5 GeV assume 8 meter pathlength: sT < 150 ps (4s separation) Can also help to suppress photon events Scintillator MRPC Multi-Resistive Plate Chamber s < 90ps No PMT, can live with field Rates > 25 kHz/cm2 TK-Tp Tp-TK

Table of Rates 15 uA beam and 40 cm target@10 amg w/ collimator Models for rate estimation have been adjusted based on real rates measured in E06-010 Only momentum cut: 3.5 GeV/c on large angle.

Trigger Setup Divide the azimuthal angle into 10 sectors. Level 1 trigger: Electron Trigger Large side: Calorimeter (11 kHz/10 = 1 kHz), OR Small side: coincidence between Pre-Shower/Shower Signal and Gas Cherenkov (140 kHz/10 = 14 kHz) Total rate of 10 sectors: 10 x (1+14) = 150 kHz Level 2 trigger: Electron Pion Coincidence Trigger Pion Trigger: Coincidence signal between Aerogel counter and Pre-Shower VETO (5MHz/10 = 500kHz) Coincidence of electron and pion triggers from different sectors: 45 total combinations. Assuming 50 ns coincidence window, total rate of all 45 combinations is: 45 x 500kHz x 50ns x 15 kHz + 3 kHz = 30 kHz Level 3 trigger: Online track cleaning up.

Data transfer rate to tape Size of single event is about 3kB. If we use Level 2 trigger, the total data transfer rate is about 3kB x 20 kHz = 60 MBps = 480 Mbps. Already pretty reasonable. With Level 3 trigger, probably will be able to further reduce the recording rate by a factor of 2.

Kinematic Coverage Precision 3-D (x, pT and z) mapping of Collins, Sivers and pretzelosity. Coverage with 11GeV beam plotted here. xB: 0.1 ~ 0.6 PT: 0 ~ 1.3 GeV/c W: 2.3 ~ 4 GeV z: 0.3 ~ 0.7 Mm: 1.6~ 3.3 GeV

Huge advantage from Angular Coverage 2p fS coverage and very large fh coverage: Impact of fh coverage on uncertainties of extracting different SSAs: Sivers, Collins and Pretzelosity. An example of fh coverage

Preliminary Beam Time Request

Projected Data Total 500(300) bins in x, Pt and z for 11(8.8) GeV beam. z ranges from 0.3 ~ 0.7, only a sub-range of 11 GeV shown here.

Coordinate with PVDIS collaboration Work together with PVDIS Yoke design Pre-Shower/Shower design Single Gas Cherenkov setup for both experiments? Try to optimize the way of switching over: Reallocation of baffles, GEMs and various detector Electronics, DAQ

Backup Slides

Multi-Resistive Plate Chamber Time resolution: < 90 ps Efficiency: > 95 % Gain: ~ 105