Review of results on polarized glue from fixed-target DIS experiments G. K. Mallot CERN/PH APS Meeting, St. Louis, MO, April, 2008

APS, 13 April 20082G.K. Mallot

G. Ross 1989 Lepton-Photon 1989 APS, 13 April 2008G.K. Mallot3 possible scenario: ∆G ≈ 6 (Q 2 =10 GeV 2 ) ∆g/g(x) = 1 for x g > 0.1 Let’s measure it!

∆g in fixed-target DIS exp. scaling violations of g 1 – inclusive DIS hadron production via PGF – hadron pairs – single hadrons – charmed hadrons APS, 13 April 2008G.K. Mallot4 EMC, SMC, SLAC, HERMES, COMPASS, JLab SMC, HERMES, COMPASS HERMES, (COMPASS) COMPASS

APS, 13 April 2008G.K. Mallot F 2 (x,Q 2 ) g 1 (x,Q 2 ) 5 proton

World data on xg 1 (x,Q 2 ) APS, 13 April 2008G.K. Mallot6 p d

World data on neutron g 1 APS, 13 April 2008G.K. Mallot7 from p and d from 3 He n

Scaling violations: QCD fits choose scheme MS, AB, jet and Q 0 2 optionally fix ns moments from hyperon decays (a 3, a 8 ) fit PDFs for quark non-singlet and singlet and gluon to g 1 data functional form of PDFs biases error band extra problems in polarized case: – no positivity condition, no momentum sum rule higher twist, … APS, 13 April 2008G.K. Mallot8 NLO DGLAP:

QCD fits many groups performing World Data fits most recent and complete DSSV → De Florian here: AAC, LSS06, COMPASS APS, 13 April 2008 G.K. Mallot Singlet gluon note: errors of AAC account for simultaneous deviation of all parameters AAC

Fit with new deuteron g 1 two solutions with ∆G > 0 and ∆G < 0 uncertainty due to parametrization not included APS, 13 April 2008G.K. Mallot10 Δ GΔ G ΔG = ΔG = ∆G > 0 ∆G < 0 MS Q 0 2 =3 GeV 2 PLB 647 (2007) 8-17

APS, 13 April 2008G.K. Mallot COMPASS deuteron g 1 Q 2 = 3 GeV 2 11

LSS06 / COMPASS06 LSS06 also finds two solutions with small ∆G > 0 and ∆G < 0 includes HT and Clas data How to tell sign of ∆G? APS, 13 April 2008G.K. Mallot12 ∆G > 0 ∆G < 0

Sign of ∆G and low g 1 at small x APS, 13 April 2008G.K. Mallot13 ∆G > 0∆G < 0LSS06 measurable at a polarized ep collider (EIC) xx g1g1 p Leader, DIS2008

Δg/g from hadron production in DIS APS, 13 April 2008G.K. Mallot14

measure calculate, and background by Monte Carlo Principle: Gluon polarization enters via photon-gluon fusion (PGF) APS, 13 April 2008G.K. Mallot Hadron production in DIS via PGF 15 q q

Analysed channels analysed data sets: – high-p T hadron pairs (no ID, pions/kaons) Q 2 > 1 GeV 2 L EPTO Q 2 < 1 GeV 2 or unmeasured P YTHIA – high-p T single hadron small Q 2 or unmeasured P YTHIA – single charmed meson quasi-real photons A ROMA, R APGAP All analyses in LO till now (plus parton showers) APS, 13 April 2008G.K. Mallot16

Why is small Q 2 interesting much larger cross section Hermes cannot measure Q 2 for most events – scattered e out of acceptance – dominated by small Q 2 hard scale given by – high-p T hadrons: – charmed hadron: APS, 13 April 2008G.K. Mallot17

Hermes high-p T hadrons APS, 13 April 2008G.K. Mallot`18 scattered electron: “tagged”  inside acceptance  Q 2 measured  p T wrt γ* “anti-tagged”  outside acceptance  Q 2 small, unmeasured  p T wrt beam 3 samples for p/d, h + /h - : single hadrons  tagged~ 4%  anti-tagged~92% hadron pairs  combined~ 4% rough percentages for events after respective cuts: p T >1 GeV/c e GeV see: N.Bianchi, Pacific SPIN07, Vancouver BC

A LL for single hadrons APS, 13 April 2008G.K. Mallot19 taggeduntagged p d h+h+ h-h- h+h+ h-h- p T (beam) p T Q 2 >0.1 GeV 2, W 2 >4 GeV 2 MC plus asymmetry model with Δg/g(x) = -1, 0, 1 for upper, middle and lower curve A LL

Subprocesses in MC APS, 13 April 2008G.K. Mallot20 VMD (elast.+diffr., soft low-pT): decreasing with p TDIS: increasing (dominating) with p T QCD2->2(q): QCDC/QCD2->2(q): increasing with p T Signal processes are PGF and QCD2>2(g) (resolved photon) anti-tagged, charge combined, deuteron PYTHIA 6.2

Direct Method (I) APS, 13 April 2008G.K. Mallot21 ▲ pairs: from Monte Carlo statistical errors only

Indirect Method (II) APS, 13 April 2008G.K. Mallot22 Fit parameters of polynomial function to A ║ using MC h + +h - data

MC tuning/systematics APS, 13 April 2008G.K. Mallot23 varied: MC parameters, PDFs, low p T asymmetry

Original HERMES measurement proton data only, high-p T pairs missing generator processes (Pythia 5.7 vs 6.3) APS, 13 April 2008G.K. Mallot24 N.Bianchi, Pacific SPIN07, Vancouver BC Phys. Rev. Lett. 84 (2000) 2584

Results for ∆g/g APS, 13 April 2008G.K. Mallot25  g/g(x,  2 ) = ± 0.034(stat) ± 0.010(syst-exp) (syst-model)  g/g(x,  2 ) = ± 0.034(stat) ± 0.011(syst-exp) (syst-model) Method I Method II

+ Resolved  Q 2 < 1 GeV 2 q,g q,gq,g ΔG/G from high-p T pairs APS, 13 April 2008G.K. Mallot26 Photon Gluon Fusion ~ 30% Leading OrderQCD Compton all Q 2 gqq use also inclusive A LL

Ratios for hadron pairs with for Q 2 < 1 High-p T pairs, low Q 2 APS, 13 April 2008G.K. Mallot 27 Resolved photons signal

Resolved photons APS, 13 April 2008G.K. Mallot28 More than 50%, however assuming a min and max scenario, shows little difference. Probing photon at large x, where photon PDF rather well determined  u/u Glück, Reya, Sieg

APS, 13 April 2008G.K. Mallot Data versus MC, low Q 2 excellent to good agreement for all kinematics variables pTpT 29

APS, 13 April 2008G.K. Mallot Systematics: e.g. k T tuning, low Q 2 systematic error: – determined using 15 independent MC simulations – exploring the parameter space in k T of nucleon and photon fragmentation functions parton shower on/off, renormalization scale nucleon photon 30

Δg/g from high-p T hadron pairs, low Q 2 APS, 13 April 2008G.K. Mallot31  2002 – 2004 data: high-p T, Q 2 < 1 GeV 2  G/G = ± (stat) ± (syst) = 0.085,  2 = 3 GeV 2

High-p T pairs, Q 2 >1 GeV 2 tune MC (Lepto) to reproduce data estimate relative fractions R on event-by-event basis using NN trained with MC events (kin. variables) use in NN output as event weight looser cuts possible – Q 2 > 1 GeV 2 – p T1,2 > 0.7 GeV – x F > 0, inv. mass 2h > 1.5 GeV 2 for details see M. Stolarski DIS2008 APS, 13 April 2008G.K. Mallot32

Monte Carlo APS, 13 April 2008G.K. Mallot33 Leading hadron Sub-leading hadron

NN description APS, 13 April 2008G.K. Mallot34

NN output APS, 13 April 2008G.K. Mallot35 inclusivehadron pairs

APS, 13 April 2008G.K. Mallot Can we learn about x g ? All processesPGF events (MC) here Lepto and Q2>1 36

Δg/g from high-p T hadron pairs APS, 13 April 2008G.K. Mallot37  2002 – 2004 data: High p T, Q 2 > 1 GeV/c 2   G/G = 0.08 ± 0.10 (stat) ± 0.05(syst)  = 0.082, (range: – 0.123)  2 ~ 3 (GeV/c)2

cleanest process (“golden channel”) – no or little physics background (LO, QCDC) observe asymmetry in D meson production – strongly statistics limited – only one D meson via D →π K (BR ~ 4%) – combinatorial background large – drastically reduced when looking to D* decay in coincidence with slow pion for details see F. Robinet DIS2008 Δg/g from open charm APS, 13 April 2008G.K. Mallot38

D mass spectra APS, 13 April 2008G.K. Mallot39 nD 0 = nD* = 8675 D 0 → K π π 0 D 0 → K π

 G/G from open charm Analysis uses both a LL and S/(S+B) weighting a LL obtained from Neural Network trained on MC (AROMA) using: Q 2, x bj, y, p T, z D S/(S+B) given by a parameterization in fP μ a LL, p K, θ K, z D, cosθ*, p T, RICH likelihoods weighting brings significant improvement in statistics due to large variations of a LL and S/(S+B) in phase- space APS, 13 April 2008G.K. Mallot40

a LL variation APS, 13 April G.K. Mallot

S/S+B weighting APS, 13 April 2008G.K. Mallot42 5 bins in  = S/(S+B) D 0 -untagged D * -tagged

 G/G from open charm APS, 13 April 2008G.K. Mallot43  2002 – 2006 data D 0 + D*   G/G = ± 0.27 (stat) ± 0.11 (syst)  ~ 0.11, ~ 13 (GeV/c) 2

Summary of results APS, 13 April 2008G.K. Mallot44 new high-p T point new charm point

DIS-FT  G/G measurements APS, 13 April 2008G.K. Mallot45 GRSV,  G std, 0.6 min, 0.2 QCD Fits |  G| ~ 0.3 max, 2.5

Summary all results point to small Δg/g DIS results are very precise NLO analysis urgently needed use results in global fits APS, 13 April 2008G.K. Mallot46