1

Flavor Separated Quark Polarizationsat HERMES

W. Lorenzon (Michigan)NAPP 2003, Dubvrovnik, Croatia

Collaboration

28 May 2003

• Flavor structure of proton – some phenomenological models

• Flavor tagging and purity analysis

• Experimental results

• Summary and Outlook …

2

Flavor Structure of the Proton

• Constituent Quark Model Pure valence description: proton = 2u + d

• Perturbative Sea sea quark pairs from g qq should be flavor symmetric:

u = d

E866: d > u

3

Flavor Structure of the Proton - II

Non-perturbative models: alternate d.o.f.

Meson Cloud Models Chiral-Quark Soliton Model Statistical Model

Quark sea from cloud of 0 mesons:

d > u

• quark d.o.f. in a pion mean-field• nucleon = chiral soliton• one parameter: dynamically generated quark mass

• expand in 1/Nc:

d > u

d > u

• nucleon = gas of massless partons• few parameters: generate parton distribution functions• input: QCD: chiral structure DIS: u(x) and d(x)

important constraints on flavor asymmetry for polarization of light sea

4

Spin Structure of the Proton

½ = ½ + G + Lq + Lg

“You think you understand something? Now add spin …” -R. Jaffe

• Constituent Quark Model pure valence description of constituent quarks:

u = + 4/3, d = - 1/3

• Relativistic Quark Model relativistic current quarks with light masses: orbital angular momentum is important, and accounts for the deficit of .

• QCD?

0.60 – 0.75

= 1

Lq=½(1-)

5

Polarized Deep-Inelastic Scattering

Note: inclusive DIS is sensitive only to eq2

There are no neutrino DIS measurements on polarized targets (yet!)

inclusive DIS cannot distinguish quark from anti-quark q := (q + q)

21

21

and

1

2

1

2

: :

qq

qq

F e q

g e q

q q q q q q

1/ 2 q

3/ 2 q

One measures double spin asymmetries:

In QCD Parton Model:

1/ 2 3/ 21

1/ 2 3/ 2

A

22 220

2 11 2 2 2

1

( , )g ( , )( , )

( , ) ( , )

gqq

qq

e q x Qx QA x Q

F x Q e q x Q

6

Comparison of Data for g1(x)

•From NLO-QCD analysis of inclusive DIS measurements… SMC, PRD 58 (1998) 112002

(in AB scheme)

s = -0.02 to –0.15 (model dependent) slight neg. sea-quark polarization?

What about q? (inclusive DIS gives no answer)

= 0.38

G= 1.0 +1.9 –0.6

7

Anti-quark Spin in the Proton

Meson Cloud Models Chiral-Quark Soliton Model Statistical Model Li, Cheng, hep-ph/9709293 Goeke at al, NPA 680 (2000) 397 Bourelly et al, EPJ C23, (2002) 487

Light sea quarks polarized:

with … but …

qvalence > 0 qsea < 0, but … “higher-order” cloud ofvector mesons can gene- rate a small polarization.

q = 0

u – d >

0

d < 0, u > 0

u(x) – d(x) > d(x) –

u(x)

u(x) – d(x) ~ d(x) –

u(x)

8

Quark Polarization from Semi-Inclusive DIS

In semi-inclusive DIS a hadron h is detected in coincidence with the scattered lepton

Goal: Flavor Separation

of quark and anti-quark helicity distributions

Technique: Flavor Tagging z=Eh/

The flavor content of the final state hadrons is related to the struck quark through the agency of the fragmentation function Dqh(z,Q2). In LO

QCD: 2

, .

2

,..

2( ( ,, ) )h hq

q u u

hq

d dq x D zQe

dz dzQ

target fragmentationregion

current fragmentationregion

9

Is HERMES Fragmentation “Universal”?

Compare pion multiplicities: HERMES vs EMC (both at Q2=25 GeV2)

Assumptions for flavor tagging (in LO QCD):• factorization (DIS, fragmentation)• fragmentation functions exhibit scaling

Good agreement, despite order of magnitude difference in energy

10

Purity Analysis of Asymmetries

• Perform a simultaneous global analysis of all A1h(x,Q2)’s

• The photon-nucleon asymmetry is:

• The hadron quark purity Pqh(x,Q2) is the probability that a

quark q was struck in an event e+N e’+h+X

• Need at least six independent asymmetry sets A1h(x) to

determine six unknown helicity distributions• Purity equations are generally under determined add symmetry assumptions (HERMES 2002)

max

min

max

min

max

min

max

min

2 2 2

'

2 2

2 2

2

2

''

2 2

1

2

2

2

( , )

( ,

'( , ) ( , )

( , )

( , ))( , )

( ,

() ,

)( )

,

z h

q qq z

z h

q qq z

z h

q qq z

z hq q

h

q

z

e

e D

q x Q dzD z Q

q z Q dz

e D z Q dz

x Q

q x Q

e q x Q dzD

A x Q c

cxq

z qQQ

Δu(x), u(x), Δd(x), d(x), Δs(x), s(x)

Δs(x) s(x)

s(x) s(x)

FL contribution to unpolarizaed PDF’s:

2

2

1 ( , )

1

R x Qc

11

Generation of Purities

• Use Monte Carlo model of DIS process (LEPTO), fragmentation process (LUND model in JETSET) and detector• Systematic uncertainties from

– Variation of fragmentation parameters– Use of alternative PDF set GRV98LO vs. CTEQ5L

12

Purities

Syst. uncertainties from PDF sets (GRV98LO, CTEQ5L) and LUND parameters

13

Semi-Inclusive Asymmetries

•

• is an all-sea object• Covered range: 0.023 x 0.6 0.2 z 0.7 xF 0.1

• Asymmetries are independent of z in this range

1 ( ) 0KA x

( )K us

1996-97 (Cerenkov)

1998-2000 (RICH)

14

Latest q Results from HERMES

First 5-flavor fit to q(x) ( assumed)

Results show:

• u-quark positively polarized

• d-quark negatively

polarized• light sea quark polarization small

• s-quark polarization somewhat positive (1)

( ) / ( ) ( ) / ( )s x s x s x s x

15

(A)symmetry of Polarized Light Quark Sea

• no strong breaking of flavor asymmetry in light sea

• data disfavor QSM of Dressler et al.

• statistical model consistent with BUT…

• meson cloud model seems to describe data best

0u d

16

Isoscalar Extraction of s

• A1K provide largest sensitivity to s: and

– but: large systematic uncertainties on PsK

• Alternative approach:

– use only total kaon flux K+ + K- on deuterium target (isoscalar)

– s’ = s + s is also isoscalar quantity

• Can measure s using inclusive and semi-inclusive

– use fragmentation functions from e+e- collider experiments,

directly compute purities PsK (no MC model needed)

• Provides independent check on strange polarization result

• Two component analysis also favors positive s

1, ( )dA x 1,K KdA

2( , )K Kq qD z Q

( )K us ( )K us

17

Summary and Outlook

• First 5-flavor separation from HERMES

– helicity distributions confirm inclusive DIS results:

– light sea in unpolarized:

– no indication for

• The Next Round of Experiments

– COMPASS: complementary results

at higher energy

– RHIC: polarized W production

• sensitivity

• largely improved precision

• HERMES Run 2 with transverse target

– focus on transversity

0, 0u d

0u d

( ) 0s s

( ) vs ( ), ( )u x d x s x