compass results: transversity

Prague (CZ), August 1st Paolo Pagano (INFN - Trieste) 1

COMPASS results: TRANSVERSITY

Paolo PaganoOn behalf of the COMPASS collaboration


Talk outline

1. The physics case of transverse spin effects2. Probing transversity in SIDIS 3. The COMPASS experiment4. The results on the deuteron

for 1 hadron SSA and interpretationfor hadron pairs SSA and interpretation

5. Conclusions and Outlook


Spin structure functions

3 distribution functions are necessary to describe the spin structure of the nucleon at LO:

All of equal importance !

TTq(x)q(x)

q(x)q(x)

q(xq(x))


Measuring ΔTq(x)

How to access it:SIDIS(e.g. COMPASS and HERMES )

Hard scattering (e.g. RHIC)–Drell-Yan–Single Spin Asym (e.g. p↑p→πX)

Hard scattering (e.g. GSI)–Drell-Yan

FFxqT )(

qq TT

NN NN qq TT

Chiral-odd: requires another chiral-odd partnerChiral-odd: requires another chiral-odd partner

Impossible in DIS Direct Measurement

ppl+l-X epe’hX

Convolution with fragmentation function

epe’X


Transversity

In the last ten years:• Great development in the theory

of transversity;• Remarkable role of ΔTq(x),

notably complementary to Δq(x).

In the last couple of years:• Role of the kT structure functions

clarified (Cahn and Sivers effects, …).

• Tensor charge (1991 – 1992):

in analogy with:

• Soffer inequality (1995):

• Leader sum rule (2004):

in analogy with:

)()(2

1)( xqxqxq TT

)()( xqxqdxg TTT

)()( xqxqdxgA

gqq

Tqq

T Lxqdx,,,

)(2

1

2

1

zz LGS 2

1

Key features of transversity:- Probes relativistic nature of quarks- No gluon analog for spin-1/2 nucleon- Different Q2 evolution and sum rule

than Δq(x)- Sensitive to valence quark polarization


•

)

Transverse Spin Physics at COMPASS

Possible quark polarimeters suggested using SIDIS:

• Azimuthal distribution of single (leading)

hadrons

• Azimuthal dependence of the plane containing

hadron pairs

• Measurement of transverse polarization of spin ½

baryons (e.g. hyperon)

Published RESULTS

2005

PreliminaryRESULTS

2005

Not covered in this talk


Azimuthal asym’s of single hadrons

• Collins effect predicts an azimuthal asymmetry in the quark fragmentation;

• Look at leading hadron (the most sensitive to the parent struck quark);• The larger z = Eh/, the stronger the signal;• An azimuthal asymmetry can also come from the un-polarised quarks; namely from

an azimuthal modulation of quark transverse momentum for a transversely polarized nucleon (Sivers effect)

Left Right

Non vanish Lq

The quark prefers to fragment in one direction

00 hqTT Dq

ChqTD sin ~FF 0

0

sin ~DF

0

0

hq

T

ST

Dq

q


Collins and Sivers asym’s

q

hqq

q

hq

Tq

Siv

q

hqq

q

hqTTq

Coll

Dqe

DqeA

Dqe

DqeA

S

C

2

02

sin

2

02

sin

Asym’s in C and S extracted independently;

Asym’s calculated as function of x, z and pt and for “Leading Hadrons” and for “All Hadrons”

s'hC

shS

Collins angle:

Sivers angle:


Azimuthal dependence of the plane containing hadron pairs • Another quark polarimeter comes from interference in fragmentation

• The struck quark fragments in two hadrons. The fragmentation function has a component dependent on the transverse spin momentum (as for the Collins function):

The hadron pair prefers to fragment in one direction

0 qT Hq ALeft Right

0sin :FF RSqH

A


02

h 2

sin

q

hqq

q qTq

UT Dqe

HqeA RS

Hadron pair azimuthal asymmetry

Calculated as function of x, z and Mh for all pairs of positive and negative hadrons.

s'RRS Azimuthal angle:

A

A. Bacchetta and M. Radici,Proceedings of DIS 2004,hep-ph/0407345

COMPASS Fixed target experiment at CERN SPS240 physicists from 12 countries28 institutes2000 technical run2001 commissioning run2002-04 physics runs (muon program)2005 SPS stop2006-… resume physics


With the muon beam

Gluon Polarization G/G

transverse spin distribution

functions Tq (x)

Flavor dependent polarized

quark helicity densities q(x)

physics

Diffractive VM-Production

Physics Goals of COMPASS

With hadron beams

Primakoff-Reactions- polarizability of and K

glueballs and hybrids

charmed mesons and baryons- semi-leptonic decays- double-charmed baryons


SM1 dipoleSM1 dipole

SM2 dipole

Polarised Target

HCAL1

Muon-filter1,MW1

Micromegas,DC,SciFi

Gems,SciFi,DCs,straws

MWPC Gems Scifi

trigger-hodoscopes

Silicon

Spectrometer 2002 2004

RICH_1 Gem_11

ECAL2,HCAL2

straws,MWPC,Gems,SciFi

straws Muon-

filter2,MW2

DW45

SciFi

Veto

Beam: 160 GeV µ+

2 . 108 µ/spill

(4.8s/16.2s)

P ~ 76%


The polarized 6LiD-Target

two 60 cm long Target-Containerswith opposite polarizationDuring data taking for transversitydipole field always

Relaxing time > 2000 hrs

superconductive

Solenoid (2.5 T)

3He – 4He Dilution refrigerator (T~50mK)

Dipole (0.5 T)

reversed once a week

1

2

For transversity:

Polarization: 50%

Dilution factor: 0.38


COMPASS data sample• In so far ( 3 years )

– only DIS off 6LiD– only runs with transverse polarized target being analyzed

runs good runsused events

in the analysis (1h)

2002 47545300 SPS

spills1.6 · 106

2003 47942900 SPS

spills~4•(2002)

2004 49694000 SPS

spills~2•(2003)

RICH

E-Calo

• Data sample increased in years 2003/4:– trigger system

upgraded;– DAQ upgraded;– 2004 longer run.

• 2003 and 2004 data production over(analysis in progress)


Kinematical distributions and general cuts

Q2=2.4 (GeV/c)2

Y=0.33

W=9.4

GeV/c2

xBj0.035

hadronshadrons

p

E

muomuonsns

Use of the calorimeters

To access DIS region and get rid of radiative corrections

2002 data

Q2 1 (GeV/c)2

0.1 Y 0.9

W 5

GeV/c2


Cuts for Collins and Sivers SSA

pT0.51

GeV/c z 0.45

2002 useful statistics Lead had’s

Positive hadrons 8.7 ·105

Negative hadrons 7.0 ·105

Total 1.6 ·106

2002 data

pT0.1 GeV/c z 0.25 (0.2 ah)


Extraction of Collins and Sivers asymmetries• Evaluating the

population as a function of Φ, independently in both target cells;

• Extracting 2 asymmetries and calculating the weighted mean.

)sin1()()( jjjjjjaFnN

j = C, S(Φj calculated always with spin = ↑)

F is the muon fluxn the number of target particlesσ the spin averaged cross-sectionaj the product of angular acceptance

and efficiency of the spectrometer;

jj

jjjj

jjjj

NrN

NrN

sin)()(

)()(

SivTS

CollNNTC

APf

ADPf

const)(

)(

jj

jj

a

a


Evaluation of systematics • Stability of the asymmetries:

i) in time;ii) in two halves of the target cells;iii) according to the hadron momentum.

• Evaluated 3 different estimators and compatibility in the results checked.

Look at the paper:

“First Measurement of the Transverse Spin Asymmetries ofthe Deuteron in Semi-Inclusive Deep Inelastic Scattering”(The COMPASS Collaboration)

Phys. Rev. Lett. 94, 202002 (2005)


Collins and Sivers effects (1)

Systematic errors are smaller than the quoted statistical errors.

• Asymmetries as a function of x, z, pt

• Only statistical errors shown

Phys. Rev. Lett. 94, 202002 (2005)


Collins and Sivers effects (2)

Systematic errors are smaller than the quoted statistical errors.

• Asymmetries as a function of x, z, pt


Phys. Rev. Lett. 94, 202002 (2005)


Interpretation of Collins SSA

• Small asymmetries. Two explanations:– Cancellation between proton and neutron;– Too small Collins mechanism.

• If and large as from preliminary measurement by BELLE (see hep-ex/0507063), this is evidence for cancellation in iso-scalar target;

• Also a new phenomenological fit of the data by

Vogelsang and Yuan (see hep-ph/0507266).

00 hqTD


COMPASS (deut.) vs. HERMES (protons) Positive hadrons

• HERMES:– Negative Collins

asymmetries;– Positive Sivers

asymmetries.

HERMES data points for Collins effect flipped by a phase of π to match COMPASS notation !

HERMES data points from: A. Airapetian et al, Phys. Rev. Lett. 94 (2005) 012002[DC53] (hep-ex/0408013)


COMPASS (deut.) vs. HERMES (protons) Negative hadrons

• HERMES:– Large Positive

Collins asymmetries;

– No Sivers effect.

HERMES data points for Collins effect flipped by a phase of π to match COMPASS notation !

HERMES data points from:A. Airapetian et al, Phys. Rev. Lett. 94 (2005) 012002[DC53] (hep-ex/0408013)


Theoretical work on Sivers SSA

• “Extracting the Sivers function from polarized SIDIS data and making predictions”–Phenomelogical model whose parameters are constrained by HERMES proton measurements;– COMPASS 2002 preliminary results for Sivers effect are in agreement with the model.

M. Anselmino et al.

hep-ph/057181

(including HERMES

Prel. Results shown at DIS05)


Prediciton of Sivers asymmetries on a proton target

M. Anselmino et al.

hep-ph/057181


Expected accuracy for Collins asymmetries

AColl statistical errrors:Positive hadrons

(to the left):

2002 - Red;

2002 – 2004 Blue;

2006 Blue.

Negative hadrons

(to the right):

2002 - Red;

2002 – 2004 Blue;

2006 Blue.

• Accuracy from total statistics accumulated with LiD (deuteron) target;

• Projections from 30 days of data taking (included in COMPASS schedule for 2006) with NH3 (proton) target :

Taking into account:

– Variation of statistical errors:

– taking into account the variation of:

)(34.1)( 63 LiDNH AA

fPT


Kinematics from the analysis on hadron pairs

2002-2003:

2.8 106

combinations

Select all combinations of positive (h1) and negative (h2)

9.0

1.0,

1.0,

21

21

21

zzz

xx

zz

FF


Extraction of asymmetries for hadron pairs• Evaluating the

population as a function of ΦRS, independently in both target cells;

• Extracting 2 asymmetries and calculating the weighted mean.

)sin1()()( RSRSRS aFnN

ΦRS calculated always with spin = ↑; F is the muon fluxn the number of target particlesσ the spin averaged cross-sectiona the product of angular acceptance and efficiency of the spectrometer;

RS

RSRS

RSRS

NrN

NrN

sin)()(

)()(

RSUTNNT ADPf sin

const)(

)(

jj

jj

a

a


Effects on hadron pairs

• Asymmetries as a function of x, z, Mh


Systematic errors are smaller than the quoted statistical errors. Presented for the first time @ DIS2005


Theoretical work on hadron pair (asym vs. invariant mass)

R. L. Jaffe, X. Jin and J. Tang, Phys. Rev. Lett. 80, 1166 (1998)

Radici, Jakob, Bianconi, PRD 65, 074031

H∢ (

z,M

2

+-)∼

sin 0

sin 1

sin

(0-

1)

AM

h

sin


Theoretical work on hadron pairs (asym vs. xbj and z)

Axsi

n

(

z)

∢

Radici, Jakob, Bianconi, PRD 65, 074031


Summary and Outlook

• Collins and Sivers SSA shown from 2002 data published by PRL:– The asymmetries are small and compatible with zero; – Making use of all the deuteron data, accuracy will

improve by a factor ~ 3. • SSA from hadron pairs calculated from 2002 + 2003 statistics:

– The asymmetries are small and compatible with zero;– Using 2004 data, accuracy improves by a factor 1.4.

• Stay tuned because:– New channels in hadron pair production are under study.– PID is being included in all the analyses;– New results on Lambda transverse polarization are coming.

• Data (of comparable statistics) will be collected on a transversely polarized proton target (NH3) in 2006.


Thank you!


Save slides


Estimators

Weaker assumptions on acceptance effects

less sensitive to distorsions set by Cahn effect

)sin41(cost)(

)(

)(

)( 2

,

,

,

,jj

jdj

jdj

juj

juj

N

N

N

N

jj

jjjjjjjj

jjjjjjjj

NNNN

NNNN

sin )()()()(

)()()()(3

Reconstruction time independent in the solid angle


Why a proton run?

Flavour decomposition of Tq:

1. proton target, combining + and -

2. deuterium target:

dudu1 DDDDD dudu2 DDDDD

21

21TTTTpTp

p11 DD

ΔDΔD

du4d4u

dΔu4ΔdΔu4ΔDPfA

21

21vTvTpTp

p21 DD

ΔDΔD

du4d4u

dΔu4ΔDPfA

21

21TTTTdTd

d11 DD

ΔDΔD

dudu

dΔuΔdΔuΔDPfA

21

21vTvTdTd

d21 DD

ΔDΔD

d)ud5(u

)dΔu3(ΔDPfA

Flavour decomposition of Tq:

1. proton target, combining + and -

2. deuterium target:

dudu1 DDDDD dudu2 DDDDD

21

21TTTTpTp

p11 DD

ΔDΔD

du4d4u

dΔu4ΔdΔu4ΔDPfA

21

21vTvTpTp

p21 DD

ΔDΔD

du4d4u

dΔu4ΔDPfA

21

21TTTTdTd

d11 DD

ΔDΔD

dudu

dΔuΔdΔuΔDPfA

21

21vTvTdTd

d21 DD

ΔDΔD

d)ud5(u

)dΔu3(ΔDPfA

vTvT

T

T

du

dd

uu

,

)(

)(

To measure the shape of

compass results: transversity

Documents