Probing the Proton’s Spin Structure with Hard Scattering, Jets and the STAR Detector at RHIC: Recent Results

W. W. Jacobs for STAR Collaboration IUCF and Dept of Physics, Indiana Univ.

STARSTAR

Introduction STAR longitudinal spin - Recent results

The polarized proton collider RHIC

The STAR detector

Topics: “Topics: “SliceSlice” of STAR Spin Physics Status & ” of STAR Spin Physics Status & OverviewOverview

STAR Transverse spin di-jets - Recent results

STAR EMC Calorimetry! Summary

Look at near future and outlook

Spin Problem & DIS Gluon Helicity Preference Spin Problem & DIS Gluon Helicity Preference ConstraintsConstraints

World data (2005) on g1p = ½

ei2 [qi (x,Q2) + qi (x,Q2)]

All fixed-target data

Only ~30% of proton spin

arises from q and q helicity preferences !

limited info on scaling

violations, on shape or integral of

gluon helicity

preference g(x,Q2).

)GeV1(@

8.1

2

dxg

)GeV5,( 2xgx4.0

g

7.1 g

Only valence

quarks are strongly polarized

gz

qz

pz LLGS

2

1

2

1

ssd

duu

RHIC p+p: Pert QCD Probe of Spin-Dep Partonic RHIC p+p: Pert QCD Probe of Spin-Dep Partonic StructureStructure

pp hX

“soft” parton distribution functions

“soft” frag. function

“hard” dQCD parton-parton

∧

Theory ingredients: pQCD factorization

LO pQCD

+ large parton-level

2-spin sensitivity

• Prefer dominant/”clean” reaction mechanism w/ large “aLL” ( -jet)

• But jet and 0 rates are sufficient to give significant G constrain in initial RHIC polarized p data

Inclusive cross-section (jets, , and ±)

Good agreement with NLO pQCD over many orders of magnitude

PRL 97, 252001 (2006)2003+2004

PLB 637, 161 (2006)

STARSTAR

Hadron inclusive

production

1st RHIC inclusive Jet x-sec

pQCD works!

also direct photon incl. @ PHENIX;

forward incl. @ STAR & incl chg.

hadrons @ BRAHMS

jets

Xpp o

2005 STAR Preliminary

±

Relativistic Heavy Ion Collider 100 GeV beam proton beams

Each bunch filled with a distinct polarization state

Spin Rotators at STAR IR allow for transverse and longitudinal spin orientation

Bunch Xings every 100-200ns

CNI polarimeters + Hydrogen Jet target provide run by run & absolute polarization

HJT calibr to ~ 5% goal in progress

…world’s 1st Collider

p p

pp Run Year FOM=P4L 2002 2003 2004 2005

2006

< Polarization> % 15 30 40-45 45-50 60

Lmax [ 1030 s-1cm-2 ] 2 6 6 16 30

Lint [pb-1 ] at STAR (L/T) 0 / 0.3 0.3 / 0.25 0.4 / 0 3.1 / 0.1 8.5 / 3.4,6.8

BRAHMS

PHENIX

AGS

BOOSTER

Spin Rotators(longitudinal polarization)

Solenoid Partial Siberian Snake

Siberian Snakes

200 MeV PolarimeterAGS Internal Polarimeter

Rf Dipole

RHIC pC PolarimetersAbsolute Polarimeter (H jet)

AGS pC Polarimeters

Strong Helical AGS Snake

Helical Partial Siberian Snake

Spin Rotators(longitudinal polarization)

Spin flipper

Siberian Snakes

STAR

PHOBOS

Pol. H- SourceLINAC

1

2

3

4

jet patch

size:1x1 x

7

8

9

10

11

12

+13,…,18

Allocated Jet Rate to tape: ~15 Hz

1x1 Jet patch ET/GeV

4 8

2006 rate ~ 2.5 Hz, sent to tape without prescaling

Level 0 High Tower & Jet Triggering in STARLevel 0 High Tower & Jet Triggering in STAR

2006 rate ~ 150 Hz, combine with L2 trigger to fit in limited bandwidth

Trigger either on

HT: 1 (of 4800 BEMC or 720 EEMC) tower ET > thresh - Or

JP: 1 (of 12 BEMC or 6 EEMC) hard-wired jet patch ET > thresh.

[note: also a minbias cond]

BARREL EMC

ENDCAP EMC

• Theoretically allowed range in Δg:

• GRSV-STDGRSV-STD: Higher order QCD analysis of polarized DIS

experiments!

Predicted ALL sensitivity for different G scenarios

Inclusive ALL measurements (, ±, and jets)

±

jets

Neutral pions (in Jets); Inclusive 0 ALL

/ndf compared to NLO calculations(ignoring systematic errors): GRSV Std: 0.8 GRSV Max: 2.5 GRSV Min: 0.8 GRSV Zero: 0.4 GRSV max scenario disfavored

z pt 0

ptjet

use invariant mass spect. (data from HT trig)

– MC 0 line shape– low inv. mass bkgrd – comb. bkgrd & residual fit

assoc w/ full Jets if within a Jet cone (0.4 in , ) but typ. ~ 5 deg.

2005 data

“geant” jets

“pyt

hia”

jets

Jet Reconstruction in STARJet Reconstruction in STAR

part

on

part

icle

dete

ctor

etcp

e

,,

,

gq,

DataData SimulationSimulation

GEANT

pyth

ia

theo

ry

Search over all possible seeds (pTseed >

0.5 GeV) for stable groupings

• Check midpoints between jet-jet pairs for stable groupings

• Split/merge jets based on Eoverlap

• Add all track/tower 4-momenta

Use cone radius:

• = 0.4 for half-BEMC 2003-5

• = 0.6-0.7 for full B+EEMC 2006

22 coneR

Full jet reconstruction uses “midpoint-cone” algorithm (hep-ex/0005012):

Data well described by MC

Inclusive Jet ALL and x-sec Analysis Issues

Use Simulation (MC) to provide correction to RAW jet yield - trigger and jet inefficiencies - jet resolution & bin migration - undetected particles (n + ) - PYTHIA 6.205 CDF Tune A - GEANT (Geisha)

Verification of DATA/MC agreement essential

The shape of the Fraction of Neutral Energy in the Jet (EMF) is sensitive to the trigger bias as well as contributions from beam background.

C pT PYTHIA(pT )

GEANT(pT )

Minbias• JP2• HT2

STAR Preliminary

On average PARTICLE Jets are reconstructed in the DETECTOR with 20% increase in pT due

to ~25% jet resolution + steep jet pT distribution

Systematic offsets in pT cause dilution of the jet asymmetry which depend on the size of the asymmetry!

Effects of Jet Resolution

Trigger Bias The trigger biases jets toward higher neutral energy. This may change the ratios of qq+qg+gg and therefore change the asymmetries

10 20 30 pT(GeV)

Trigger Bias: JP << HTALL/ALL larger at low pT

Total Systematic:

ALL ALLPARTICLE ALL

TRIGGER

g = g (max)g = -g (min) g = 0GRSV-STD

2005 STAR preliminary

Systematic error band

Measured Jet PT (GeV/c)

Error bars are statistical Systematic band includes 25% scale error from current polarization uncertainties due to online values Results in good agreement with ’03/’04 ALL data* in region of overlap but ~ 4 times more precise and pT range nearly doubled [* PRL 97,252001 (2006)]

Systematic

(x 10 -3)

False Asymmetries

<6.5 (pT dep’t)

Reconstruction + Trigger Bias

2-12 (pT dep’t)

Non-long Polarization

3

Relative Luminosity

2

Backgrounds

<1

2005 Inclusive jet ALL at mid-rapidity

Vogelsang and Stratmann

Significant new constraints on G when compared to predictions derived from one global fit to DIS data

GRSV DIS best fit=0.241 = -0.45 to 0.7

PRD 63, 094005 (2001)

GRSV DIS

Constraint on G

Inclusive Jet (Inclusive Jet ()) Data Data from 2006 -> Greater from 2006 -> Greater Discr’m Power for Discr’m Power for gg

High-statistics (esp. at high pT) inclusive jet and 0 ALL data from 2006 will select among g models, assuming a shape of g(x,Q2).

Need global analysis including these ALL results!

Significant increase in sampled luminosity Polarization typically ~60% acceptance in BEMC

increased by a factor of 2significant increase in figure of merit!

G=G

GRSV-stdG=-G G=0

Projected statistical uncertainties for STAR 2006 inclusive jet ALL

jetjet

also analysis w/ Endcap EMC + vs. - analysis in Barrel

Di-Jet Asymmetry Measurements with Di-Jet Asymmetry Measurements with STAR; Sivers Transverse Spin AsymmetriesSTAR; Sivers Transverse Spin Asymmetries

Needs ISI and/or FSI to evade time-reversal violation

Assuming QCD factorization, subsume ISI/FSI contributions in gauge-invariant kT -dependent parton distribution fcns.

Characterize by Sivers effect:

Sensitive to parton orbital angular momentum.

0processobserved

partonTprotonproton kps

pproton

sproton

kTparton ?

kTparton ?

Do quarks and/or gluons have transverse motion preferences in a proton polarized transverse to its momentum?

Do we observe q Sivers consistent

w/ HERMES, after inclusion of proper

pQCD-calculable ISI/FSI gauge link factors

for pp jets? Tests limited TMD

“universality”.

First direct measurement of gluon Sivers effects.

Expect Collins and Boer-Mulders asy’m contributions

small: no charge-sign bias from EMC-only

L2 trigger jet reconstruction & small role of q-q scattering

at probed pT

2006 p + p run at RHIC; 3 wks transverse spin @ STAR

Motivation for pp Motivation for pp Di-Jet Di-Jet MeasurementMeasurement

zx

y

Colliding beams

proton spin

parton kTx

HERMES transverse spin SIDIS asymmetries u and d quark Sivers

functions of opposite sign, different magnitude.

Sivers effect in pp spin-dependent sideways boost to di-jets, suggested

by Boer & Vogelsang (PRD 69, 094025 (2004))

Both beams polarized, x +z x z can

distinguish high-x vs. low-

x (primarily

gluon) Sivers effects.

spin

bisector

Jet 1

Jet 2

Reco cos(bisector) measures

sign of net kTx for event

=+2

= -1

TPC

EMC Barrel

EMC Endcap

BBC East BBC West

Yellow (-z) beamBlue (+z)

STAR EMC-Based (Level 0 + 2) Di-Jet Trigger in STAR EMC-Based (Level 0 + 2) Di-Jet Trigger in 20062006

Endcapessential for hi-x vs.lo-x Siversdistinction

Signed azimuthal opening angle zz xx /ln21

2006 p+p run, 1.1 pb1

2.6M di-jet triggered events

2 localized clusters = 0.6 0.6, with ET

EMC > 3.5 GeV, | | > 60; ET wt’ed centroid

Broad 1,2 coverage Full, symmetric 1,2 coverage

Fast MC Simulations Illustrate Di-Jet Sivers Fast MC Simulations Illustrate Di-Jet Sivers EffectsEffects 2-parton events,

transverse plane

match full jet reco. pT distribution

Gaussian + exp’l tail kT distribution fits distribution

random kTx,y (rms = 1.27 GeV/c) for each parton

Sivers spin-dep. kTx offset shift, L-R di-jet

bisector asymetry

1-spin effects vary linearly with kTx offset

f = 0.85 dilution corrected in data

Error-weighted average of 16 independent AN(>) values for |cos(bisector)| slices

effective beam pol’n for each slice = Pbeam |cos(bisector)|

rotation samples kT

y, parity-violating sp•kT

correlation

STAR data - bothjets rotated by 90

Null Tests

STAR Results Integrated Over PseudorapiditySTAR Results Integrated Over Pseudorapidity

Sivers asymmetries consistent with zero with stat. unc. = 0.002

Fast MC sensitivity to Sivers kTx offset few MeV/c 0.002 (kT

x)21/2

Systematic uncertainties smaller than statistics

All null tests, including forbidden 2-spin asym. cos(bisector), consistent with zero, as are physics asymmetries for all polarization fill patterns

Note: P_beam from online CNI analysis, with 20% calibr. uncertainty

AN+z AN

-z 2-spin AN+z AN

-z 2-spin

STAR data

What Did We Expect? Constraints from SIDIS What Did We Expect? Constraints from SIDIS ResultsResultsFits to HERMES SIDIS Sivers asym constrain u and d quark Sivers functions, for

use in pp dijet + X predictions.

Bomhof, Mulders, Vogelsang & Yuan, hep-ph/0701277

5 < pTparton < 10 GeV/cInitial State Inter’ns only (à la Drell-Yan)Trento sign conv. (opposite Madison)

ISI only

FSI only

ISI+FSI

VY 2 SIDIS Sivers fit

Theory of Transverse SSA Developing Very Theory of Transverse SSA Developing Very Rapidly!Rapidly!Bacchetta et al. [PRD 72, 034030 (2005)] deduce gauge

link struct for pp jets, hadrons:

AN (ISI+FSI) 0.5 AN (ISI)

Gauge links more robust for SSA wt’ed by pT or |sin |, due to kT - factorization breakdown (Collins & Qiu, arXiv:0705.2141)

models (2) of Sivers fcn. x-dep

W. Vogelsang and F. Yuan, PRD 72, 054028 (’05)

Jet 1 rapidity Jet pT (GeV/c)

no hadronization no gluon Sivers functions

xxxud

xxxuuVY

T

T

186.1)(/

181.0)(/:1)2/1(

)2/1(

xxxdd

xxxuuVY

T

T

176.2)(/

175.0)(/:2)2/1(

)2/1(

STAR Di-Jet Sivers Results vs. Jet Pseudorapidity STAR Di-Jet Sivers Results vs. Jet Pseudorapidity Sum ISum I

Tra

ns

Sp

in A

sym

Emphasizes (50%+ ) quark Sivers

Spin

Polquark

Unpol

qluon+z

+z

y

x

Blue beam Yellow beam

.,),(

),(

),(

),()(

1)(/1)(

)(|cos|

etcN

N

N

Nr

withrr

APf

bj

bj

bj

bjbz

bzbz

zNbz

Extract analyzing powers averaged over and b w/ fit to asym’s in |cos b| using cross ratio:

Typical xT ~ 0.05 - 0.10;

1+2 range 0.01 < xBj < 0.4

STAR Di-Jet Sivers Results vs. Jet Pseudorapidity STAR Di-Jet Sivers Results vs. Jet Pseudorapidity SumSum

STAR AN all consistent with zero both net high-x parton and low-x gluon Sivers effects ~10x smaller in pp di-jets than SIDIS quark Sivers asym.!

Blue beam Yellow beam All calcs. for STAR acceptance

Reverse calc. AN signs for Madison convention

Scale Bomhof calcs by 1/|sin | 3.0 to get AN of unit max. magnitude

u vs d and FSI vs ISI cancellations sizable SSA in inclusive fwd. h prod’n and SIDIS (weighted SSA) compatible with small weighted di-jet SSA -- test via LCP flavor select

Near Term future, and RHIC run 8 & 9Near Term future, and RHIC run 8 & 9

Inclusive channels for longitudinal spin prog suffer from broad integration over x model-dep. G extraction as well as other systematic issues

With improved beam & detector performance, focus will now shift to jet-jet and -jet coincidences for event-by-event constraints on colliding parton x1,2 .

0 10 3020pT (GeV)

xgluon

500200

GeV101

102

Inclusive 0

N.B. x-range sampled depends on g(x,Q2) ! -- M. Stratmann

.2

;2

1tanh*cos

;2

;2

21

212121

pp

TT

TT

s

pxwith

eex

xeex

x

Sivers -> Ongoing analysis incorporates TPC tracks for full jet reconstruction allowance for cuts on jet pT , u vs. d filtering via leading hadron charge sign, etc. (w/o cuts consistent w/ EMC only)

Presently in detailed planning stages for RHIC run 8 (accelerator cool-down ~1 Nov. 2007) … as per the Beam Use Request (BUR) there will be a significant amount of polarized pp running (divided among longitudinal and transverse spin orientations) … as we also work on run 6 analysis!

Similarly we expect a signifcant amount of polarized pp beam in run 9

2008-12: Coincidence Measurements to Map 2008-12: Coincidence Measurements to Map g(x) Fullyg(x) Fully

For example, simulations (L. Bland) of STAR capabilities for - jet coincidences give rough indication of g discriminating power for various models of input gluon polarization.

Simplified LO analysis used for simulations here to illustrate sensitivity

- jet and di-jet measurements @ s = 200 & 500 GeV, will map g(x 0.01-0.3,high Q2), when included in NLO treatments of entire spin structure database.

SummarySummary NLO pQCD describes hadron x-sections at RHIC for inclusive jets,and ±) allowing spin program to access G directly

Longitudinal spin: 2005 inclusive jet data provide significant new constraint on G when compared to predictions derived from oneglobal fit to DIS data (GRSV-max scenario ruled out w/ ‘03/’04 data);2006 data should provide sizable add’l constraint (as will global fits!) STAR longitudinal spin program entering phase of correlation anddirect measurements, while continuing to expand the pT reach of the incl. channels; 2008-09, focus on ALL for di-jet and +jet production Transverse spin: spin asym’s for pp di-jet production Sivers

asym’s consistent w/ zero, whether dominated by valence or sea partons … data will constrain unified theoretical accounts of SSA in hard trans spin pQCD, and connection to parton orbital momentum.

Present pQCD calcs. reconcile small observed asym’s with larger effects seen in SIDIS (& pp forward hadron), via cancelling ISI vs. FSI and u vs. d contr.

p+p in pQCD regime viable complement to DIS more data coming!

BACKUPS

2005 Inclusive Jet ALL

GRSV curves*G = GG = -GG = 0G = STD

1

PY PB

N parallel RN antiparallel

N parallel RN antiparallel

Phys.Rev.Lett 97 252001 (2006)

2005 ALL is consistent

with previous

2003/2004 results.

PYTHIA+GEANT full jet reconstruction vs. parton-level resolution:

EMC-Only Information OK For 1EMC-Only Information OK For 1stst Dijet Sivers Dijet Sivers AsymmetryAsymmetry

Jet finder

•TPC+EMC

• jet cone radius 0.6

Full offline di-jet reconstruction for ~2% of all runs shows triggered jet pT spectrum:

Typical xT ~ 0.05 - 0.10;

1+2 range 0.01 < xBj < 0.4

and angle resolution loss @ L2 OK:

[()=5.0, ()=0.10] full reco. jet vs. parton angles

Net L2-to-parton (jet) = 6.3, (di-jet) = 9.0

(full reco) – (L2) [deg]

[()=3.9, ()=5.8] L2 vs. full jet << observed() 20, mostly from kT

Distinguishing Sivers from Collins AsymmetriesDistinguishing Sivers from Collins AsymmetriesIn SIDIS, can distinguish transverse motion preferences in PDF’s (Sivers) vs. in fragmentation fcns. (Collins) via asym. dependence on 2 azimuthal angles:

HERMES results both non-zero, but + vs. – difference suggests Sivers functions opposite for u and d quarks.

Collins Sivers

Theory of Transverse SSA Developing Very Theory of Transverse SSA Developing Very Rapidly!Rapidly!

Bacchetta, Bomhof, Mulders & Pijlman [PRD 72, 034030 (2005)] deduce gauge link structure for pp jets, hadrons:

AN (ISI+FSI) 0.5 AN (ISI)

Gauge links more robust for SSA weighted by pT or |sin |, due to kT - factorization breakdown (Collins & Qiu, arXiv:0705.2141)

ISI only

FSI only

ISI+FSI

Bomhof, Mulders, Vogelsang & Yuan, hep-ph/0701277

VY 2 SIDIS Sivers fit

Bomhof, Mulders, Vogelsang & Yuan, hep-ph/0701277

u quark

d quark

u+d

Sivers fcns. from twist-3 qg correl’n fits to pp forward hadron

Ji, Qiu, Vogelsang & Yuan [PRL 97, 082002 (2006)] show strong overlap between Sivers effects & twist-3 quark-gluon (Qiu-Sterman) correlations:

twist-3 fits to AN(p+p fwd. h) can constrain Sivers fcn. moment relevant to weighted di-jet SSA

Kouvaris et al. [PRD 74, 114013 (2006)] fits give nearly complete u vs. d cancellation in weighted di-jet SSA

Future: W Production @ 500 GeV Future: W Production @ 500 GeV u/ u and u/ u and d/d to Illuminate Origin of the Nucleon’s qq d/d to Illuminate Origin of the Nucleon’s qq

SeaSea

¯̄ ¯̄¯̄ ¯̄¯̄

2 asyms. 2 charges pol’n of valence q, sea q separately for u,d.

Detect W± via isolated high-pT daughter e± or ± , no away-side jet

¯

Projected uncertainties for quark and antiquark polarizations

Many non-perturbative models of nucleon structure predict sign & magni-tude difference between u and d polar-izations in nucleon sea, not yet seen.

Probe via single-spin parity-violating asym. AL for p + p W + X with respect to helicity flip of each beam.

Detector at RHICSTARSTAR

= -ln[tan(/2)]

TPC

||<1.4

Charged particle momentum

BEMC

||<1.0

Neutral Energy

High pT Trigger

EEMC

1<<2

Neutral Energy

High pT Trigger

BBC

3.4<<5

MinBias Trigger

Relative Lumi

(also ZDC)

EndCapEMC

BBCEast

Barrel EMC

BBC West

TPC

Yellow beamBlue beam

STASTARR

FMS, EEMC and BEMC provides nearly complete EM coverage from -1 η +4

STAR Calorimeter Coverage

proton spin

proton momentum

proton spin

proton momentum +

Blue beam asymmetry Yellow beam asymmetry2006 p+p run STAR measurement of Sivers transverse single-spin asymmetry for di-jets -- shows smaller effects than predicted for observable sensitive to orbital components of parton motion in proton.