Flavor Asymmetry of the Nucleon Sea and

the Connection with the Five-Quark Components

Wen-Chen ChangInstitute of Physics, Academia Sinica

8th Circum-Pan-Pacific Symposium on High Energy Spin PhysicsJune 20-24, 2011 in Cairns, QLD, Australia

Outline

Evidences for the Existence of Sea Quarks

Flavor Asymmetry of Sea QuarksTheoretical Interpretations Intrinsic Sea Quark & Light-cone 5q

ModelCurrent & Future ExperimentsConclusion

2

Deep Inelastic Scattering

3

Q2 :Four-momentum transferx : Bjorken variable (=Q2/2Mn)n : Energy transferM : Nucleon massW : Final state hadronic mass

22 2 2

2 1'

2 2 22 1

[ ( , ) 2 ( , ) * tan ( / 2)]

[ ( , ) / 2 ( , ) / * tan ( / 2)]

Mott

Mott

dW Q W Q

d dE

F x Q F x Q M

• Scaling• Valence quarks• Quark-antiquark pairs

Sum Rules

4J.I. Friedman, Rev. Mod. Phys. Vol. 63, 615 (1991)

Constituent Quark modelAxial vector current matrix

elements:

Scalar density matrix elements:

5

52 , | | , 2 ( )s q p s q q p s s q q q q

| | (3) ( ) ( )( ) ,

(8) ( ) ( ) 2 ( )| |

p qq p F F u F dF q

F F u F d F sp uu dd ss p

sQM exp

5/3 1.26

1 0.6

1/3 0.23

u d 2u d s

(3) / (8)F F

The simplest interpretation of these failures is that the sQM lacks a quark sea.Hence the number counts of the quark flavors does not come out correctly.- Ling-Fong Li and Ta-Pei Cheng, arXiV: hep-ph/9709293

Sum Rules

6J.I. Friedman, Rev. Mod. Phys. Vol. 63, 615 (1991)

Gottfried Sum Rule

7

1

2 20

1

0

[( ( ) ( )) / ]

1 2( ( ) ( ))

3 3

( )1

3 p p

p nG

p p

S F x F x x dx

u x d

i

x dx

f u d

Assume an isotopic quark-antiquark sea,GSR is only sensitive to valance quarks.

Measurement of Gottfried Sum

8

New Muon Collaboration (NMC), Phys. Rev. D50 (1994) R1

SG = 0.235 ± 0.026

( Significantly lower than 1/3 ! )

Explanations for the NMC result

9

• Uncertain extrapolation for 0.0 < x < 0.004

• Charge symmetry violation

• in the proton

,( )n p n pu d d u

( ) ( )u x d x1

0( ( ) ( )) 0.148 0.04d x u x dx

Need independent methods to check the asymmetry, and to measure its x-dependence !

d u

224 1[ ( ) ( ) ( ) ( )]

9 t b t bb t b t

de q x q x q x q x

dx dx x x s

( )1

1 ( ) 1 ( )4 ( )| 1 1

2 2 2( ) ( ) ( ) ( )1

4 ( ) ( )

b t

bpd

t tbx x

ppb t t t

b t

d xd x d xu x

d x d x u x u xu x u x

Drell-Yan Process

10

Acceptance in Fixed-target Experiments

Light Antiquark Flavor Asymmetry

Naïve Assumption:

11

NA51 (Drell-Yan, 1994)

NMC (Gottfried Sum Rule)

NA 51 Drell-Yan confirms

d(x) > u(x)

Light Antiquark Flavor Asymmetry

Naïve Assumption:

12

NA51 (Drell-Yan, 1994)

E866/NuSea (Drell-Yan, 1998)

NMC (Gottfried Sum Rule)

Deep-Inelastic Neutrino Scattering

13

)]()()()([2)(2 xcxsxuxdxxF p

)]()()()([2)(2 xcxsxdxuxxF n

Strange Quark in the Nucleon

14

CCFR, Z. Phys. C 65, 189 (1995)

;N c X c s

)(*5.0)( duss

)()( xsxs

Strange Quark and Antiquark in the Nucleon

15

NuTeV, PRL 99, 192001 (2007)

Strange Quarks from SI Charged-Kaon DIS Production

16

HERMES, Phys. Lett. B 666, 446 (2008)

x(s

+s

)

( ( ) ( )) ( ( ) ( ))s x s x u x d x

HERMES vs. CCFR and CT10

17

Nontrivial QCD VacuumAnimation of the Action Density in 4 Dimensions

18

http://www.physics.adelaide.edu.au/theory/staff/leinweber/VisualQCD/QCDvacuum/welcome.html

Origin of u(x)d(x): Perturbative QCD effect?

Pauli blocking guu is more suppressed than gdd in the

proton since p=uud (Field and Feynman 1977)

pQCD calculation (Ross, Sachrajda 1979)

Bag model calculation (Signal, Thomas, Schreiber 1991)

Chiral quark-soliton model (Pobylitsa et al.

1999) Instanton model (Dorokhov, Kochelev 1993) Statistical model (Bourrely et al. 1995; Bhalerao

1996) Balance model (Zhang, Ma 2001)

19

The valence quarks affect the gluon splitting.

Origin of u(x)d(x): Non-perturbative QCD effect?

Meson cloud in the nucleons (Thomas 1983, Kumano 1991):

Sullivan process in DIS.

Chiral quark model (Eichten et al. 1992; Wakamatsu

1992): Goldstone bosons couple to valence quarks.

20

The pion cloud is a source of antiquarks in the protons and it lead to d>u.

2:3:4:: , 0 qq

0:1:2:: , 0 Np

n

Meson Cloud Model (Signal and Thomas, 1987)

Chiral Field (Burkardt and Warr , 1992)

Baryon-Meson Fluctuation (Brodsky and Ma , 1996)

Perturbative evolution (Catani et al., 2004)

21

s(x)=s(x)?

( ) ( ) at large s x s x x

( ) ( ) at large s x s x x

( ) ( ) at large s x s x x

( ) ( ) at large s x s x x

Spin and Flavor are Connected

22

J.C. Peng, Eur. Phys. J. A 18, 395–399 (2003)

1

0[ ( ) ( )]I u x d x dx

HERMES (PRD71, 012003 (2005))

COMPASS (NPB 198, 116, (2010))

DSSV2008 (PRL 101, 072001 (2008))

23

0.32 2

0.023

( ) 0.048 0.057 0.028 at Q =2.5 GeVu d dx

Is = ?u d

0.32 2

0.004

( ) 0.052 0.035 0.013 at Q 3 GeVu d dx

12 2

0

( ) 0.117 0.036 at Q 10 GeVu d dx

Light quark sea helicity densities are flavor symmetric.

Origin of Sea Quarks

Extrinsic Intrinsic

Gluon splitting in leading twist Gluon fusion & light quark scattering (higher-twist)

Perturbative radiation Non-perturbative dynamics

CP invariant Possible CP non-invariant

Fast fluctuation With a longer lifetime

Of small x Of large x (valence like)

Strong Q2 dependent Small Q2 dependent

24

It is generally agreed that the observed flavor asymmetry mostly resulted from the intrinsic sea quarks.

For further investigation, it will be good to separate their contributions.

: Flavor Non-singlet Quantity

25

( ) ( )d x u x

• is a flavor-non-singlet (FNS) quantity.

• Extrinsic sea quarks vanish at 1st order in s .

• Non-perturbative models are able to describe the trend.

• Greater deviation is seen at large-x valence region.

• No model predicts

( ) ( )d x u x

( ) ( )d x u x

Intrinsic Sea & Flavor Non-singlet Variables

Select a non-perturbative model with a minimal set of parameters.

Construct the x distribution of flavor non-singlet quantities: , , at the initial scale.

After a QCD evolution with the splitting function PNS to the experimental Q2 scale, make a comparison with the data.

26

d u d u s s

“Intrinsic” Charm in Light-Cone 5q Model

3 5| | | .....q qp P uud P uudQQ

Dominant Fock state configurations have the minimal invariant mass, i.e. the ones with equal-rapidity constituents.

The large charm mass gives the c quark a larger x than the other comoving light partons, more valence-like.

27

In the 1980’s Brodsky et al. (BHPS) suggested the existence of “intrinsic” charm (PLB 93,451; PRD 23, 2745).

The intrinsic charm in | can contribute to the charm production at large x .Fuudcc

25 52 2

1 5 51 1

( ,..., ) (1 ) / [ ]ii p

i i i

mP x x N x m

x

ISR

Experimental Evidences of IC

28

Still No Conclusive Evidence…..

CTEQ Global AnalysisPRD 75, 054029

arXiv:hep-ph/9706252

“Intrinsic” Sea 5q Component

29

2 In principle, the probability of 5q state ~1/M .

So the probability is larger for | of light Q.

We consider the flavor asymmetry of sea quark

as the experimental evidences for the intrinsic

|

Q

uudQQ

uu

, | , | 5-quark states.duu uudd d uudss

“Intrinsic” Sea 5q Component

30

In the limit of a large mass for quark Q (charm):

)]/1ln()1(2)101)(1[(2

1)( 555

2555

255

~

5 xxxxxxxNxP

| uudcc mc=1.5, ms=0.5, mu, md=0.3 GeV is obtained numerically.

5 ( ; )uudQQQ

P P x uudQQ dx

( ; )Q

P x uudQQ

Data of d(x)-u(x) vs. Light-Cone 5-q Model

The shapes of the x distributions of d(x) and u(x) are the same in the 5-q model and thus their difference.

Need to evolve the 5-q model prediction from the initial scale to the experimental scale at Q2=54 GeV2.

31

5 5 0.118uudd d uuduuP P

W.C. Chang and J.C. Peng, arXiv: 1102.5631

Data of x(s(x)+s(x)) vs. Light-Cone 5-q Model

The x(s(x)+s(x)) are from HERMES kaon SIDIS data at <Q2>=2.5 GeV2.

Assume data at x>0.1 are originated from the intrinsic |uudss> 5-quark state.

32

5 0.024uudssP

W.C. Chang and J.C. Peng, arXiv: 1105.2381

Data of x(d(x)+u(x)-s(x)-s(x)) vs. Light-Cone 5-q Model

The d(x)+u(x) from CTEQ 6.6.

The s(x)+s(x) from HERMES kaon SIDIS data at <Q2>=2.5 GeV2.

Assume Probabilities of 5-q

states associated with the light sea quarks are extracted.

33

5

5

0.240

0.122

uudd d

uuduu

P

P

W.C. Chang and J.C. Peng, arXiv: 1105.2381,1102.5631

5 0.024uudssP

Comparison of 5q Probabilities

P(uu)

P(dd)

P(ss)

P(cc)

Reference

0.198

0.148

0.093

0.011

Bag model; Donoghue and Golowich, PRD15, 3421 (1977)

0.003

Light-cone 5q model;Hoffmann and Moore, ZPC 20, 71 (1983)

0.250

0.250

0.050

0.009

Meson cloud model; Navarra et al., PRD 54, 842 (1996)

0.10 -0.15

Constituent 5q model;Riska and Zou, PLB 636, 265 (2006)

0.122

0.240

0.024

Light-cone 5q model;Chang and Peng, this work (2011)

34

The Light-Cone 5-q Model It is surprising that many FNS quantities

can be reasonably described by such a naïve model with very few parameters (mass of quarks and the initial scale).

For completeness, this model should be extended to take into account: Anti-symmetric wave function Chiral symmetry breaking effect Spin structure Higher configuration of Fock states

35

FNAL E906/SeaQuest Experiment

36

Fermilab E866/NuSea Data in 1996-1997 1H, 2H, and nuclear targets 800 GeV proton beam

Fermilab E906/SeaQuest Data taking planned in 2010 1H, 2H, and nuclear targets 120 GeV proton Beam

Cross section scales as 1/s – 7x that of 800 GeV beam

Backgrounds, primarily from J/ decays scale as s– 7x Luminosity for same detector

rate as 800 GeV beam

50x statistics!!

Fixed Target

Beam lines

Tevatron 800 GeV

Main Injector

120 GeV

224 1[ ( ) ( ) ( ) ( )]

9 t b t bb t b t

de q x q x q x q x

dx dx x x s

d/u From Drell-Yan Scattering

Ratio of Drell-Yan cross sections

(in leading order—E866 data analysis confirmed in NLO)

Global NLO PDF fits which include E866 cross section ratios agree with E866 results

Fermilab E906/Drell-Yan will extend these measurements and reduce statistical uncertainty.

E906 expects systematic uncertainty to remain at approx. 1% in cross section ratio.

37

Longitudinal and Transverse View of E906 Experimental Area

38

Run Schedule

39

Charged Asymmetry of W at RHIC

40Yang, Peng, and Groe-Perdekam, Phys. Lett. B 680, 231 (2009)

p+p at sqrt(s)=500 GeV

41Kensuke’s talk on Monday

J. Mans :: CMS EWK Measurements 42

20 GeV PT Results

● Caveats

● Very preliminary, not part of publication on the topic

● Only muons (no electrons)

● Uncertified systematic errors

Future Experiments

COMPASS Polarized -induced DY experiment at CERN: spin structure of sea quark.

MINERνA at FNAL: x-dependence of nuclear effects for sea and valance quarks.

JLAB-12 GeV: transverse spatial distribution of partons.

(Polarized) DY experiment at J-PARC: d/u at very large-x region.

EIC at RHIC: sea quark distributions and their spin dependence.

43

Conclusion

Using DIS, Drell-Yan and SIDIS processes, the structure of sea quarks in the nucleon are explored. A large asymmetry between d and u was

found at intermediate-x regions. No large asymmetry was observed

between s and s.

44

Conclusion

The observed large flavor asymmetry mostly resulted from the non-perturbative effects.

The measured x distributions of (d-u), (s+s) and (u+d-s-s) could be reasonably described by the light-cone 5q model. The probabilities of the intrinsic 5q states of light sea quarks are extracted.

45

Conclusion

The sea quarks are connected with the non-perturbative feature of QCD. They could be the key to understand the confinement!

46