Transcript Düren (ppt 10,1MB)

Results from HERMES
in view of futurepp experiments
Michael Düren
Universität Gießen
—pp-QCD Workshop, ECT* Trento, July 3-7, 2006 —
Outline
 HERMES physics:
Modest aims and rich harvest
 HERMES technology:
polarization and novel techniques
 Physics results:
– Spin structure of the nucleon
– Hard exclusive reactions
– Quarks in nuclei
 Future prospects for pp
 Conclusions
Some transparencies are stolen from
Aschenauer, Hasch, Nowak, Ji, …
M. Düren, Univ. Giessen
2
HERMES physics
Why HERMES?
A historical review
 In 1987 we discovered in
the EMC the “spin puzzle”:
– Violation of the Ellis
Jaffe sum rule
– A large negative
strange sea polarization
Both results were based
on inclusive DIS data
and validity of SU(3)f
Original physics program of HERMES:
 In 1989 we decided to solve the
“spin puzzle” with a semi-inclusive
DIS experiment doing a flavor
decomposition of the quark spin
 Re-measure inclusive polarized
DIS on proton and neutron:
– Ellis-Jaffe sum rule
– Bjorken sum rule
– g 1, g 2
 Measure semi-inclusive polarized
DIS on p and n:
– u(x), d(x), s(x)
4
Why HERMES?
A historical review
Results (in short):
up

Ellis-Jaffe sum rule:
is really broken

Bjorken sum rule:
is fine (fortunately)

g1, g2 are well known nowadays

u(x) is large and positive

d(x) is smaller and negative

s(x) is approx. zero
i.e.
– SU(3)f is broken and
– only ~30% of the spin of
the nucleon is due to the
spin of the quarks
down
M. Düren, Univ. Giessen
u
d
strange
5
Spin of the nucleon
Latest results!
Q( x )  u( x )  u( x )  d ( x )  d ( x )
1

0.02
Q dx  0.29  0.03  0.01
~30% up and down
quark contribution
in valence region
S ( x )  s( x )  s( x )
1

0.02
S dx  0.006  0.029  0.007
<3% strange quark
contribution
M. Düren, Univ. Giessen
6
Physics program at HERMES today:
 Flavor decomposition of
quark spin (done)
 Fragmentation process in
vacuum and nuclear medium
 Gluon spin contribution
(first result but large
errors)
 Spin matrix elements of
vector meson production
 Transversity, Collins and
Sivers functions
(first non-zero results)
 ...
 Orbital angular momentum
of quarks (first results,
but model dependent)
 Generalized parton
distributions (first results
on BCA, BSA, TTSA,…)
 Positive Pentaquark signal
In many areas HERMES
contributed as a
pioneering experiment
M. Düren, Univ. Giessen
7
HERMES technology
Polarization and novel techniques
HERMES requires:
 Large polarization of beam
and target; pure target
 Relatively large luminosity
(Much larger than EMC)
 Relatively high beam energy
(Q²>1 GeV²; larger than
Jlab)
 Relatively large acceptance
(Much larger than SLAC)
 Strangeness identification
(kaons)
 Recoil protons
(for exclusive reactions)
Solution:
 High HERA beam polarization
(pushed by HERMES) and ABS
 Storage cell technique
(new at that time)
 HERA fixed target
 Standard open spectrometer
 RICH upgrade in 2000
(well working RICH)
 Recoil upgrade in 2006
(for GPD program)
M. Düren, Univ. Giessen
9
Polarized target
M. Düren, Univ. Giessen
10
The spectrometer
M. Düren, Univ. Giessen
11
Recoil detector
M. Düren, Univ. Giessen
12
Scintillating fiber detector
M. Düren, Univ. Giessen
13
Guide to success:
 Novel technologies (at time of proposal)
 Unique facility (energy, luminosity, precision, …)
 Polarization (that is where one can falsify models)
 Flexibility for upgrades
 Electromagnetic processes
that can be directly related to QCD
My message:
Panda at Fair will also measure completely different
things compared to what is in the proposal today!
M. Düren, Univ. Giessen
14
Outline
 HERMES physics:
Modest aims and rich harvest
 HERMES technology:
polarization and novel techniques
 Physics results:
– Spin structure of the nucleon
– Hard exclusive reactions
– Quarks in nuclei
 Future prospects for pp
 Conclusions
M. Düren, Univ. Giessen
15
Transverse spin distribution of quarks
 The transverse polarization of quarks in a transversely polarized
proton is not identical with the longitudinal polarization of quarks
in a longitudinally polarized proton
(Rotation and boost do not commute)
 There are three leading-twist structure functions:
Quark density
Helicity distribution
Transversity
M. Düren, Univ. Giessen
See talk by Anselmino
16
Transverse spin distribution of quarks
 The azimuthal angular distributions of hadrons from a
transversely polarized target show two effects:
– Collins asymmetry in +s
Target spin
– Sivers asymmetry in -s
M. Düren, Univ. Giessen
17
Transverse spin distribution of quarks
 The azimuthal angular distributions of hadrons from a
transversely polarized target show two effects:
– Collins asymmetry in +s
Product of the chiral-odd
transversity distribution h1(x)
and the chiral-odd
fragmentation function H1(z)
 related to the transverse
spin distribution of quarks
First
results
from Belle
Target spin
– Sivers asymmetry in -s
M. Düren, Univ. Giessen
18
Transverse spin distribution of quarks
 The azimuthal angular distributions of hadrons from a
transversely polarized target show two effects:
– Collins asymmetry in +s
Product of the chiral-odd
transversity distribution h1(x)
and the chiral-odd
fragmentation function H1(z)
 related to the transverse
spin distribution of quarks
First
results
from Belle
Target spin
– Sivers-Asymmetrie in -s:
Product of the T-odd
distribution function f1T(x)
and the ordinary
fragmentation function D1(z)
 related to the orbital
angular momentum of quarks
M. Düren, Univ. Giessen
See talk by Metz
19
Collins asymmetries
 at HERMES (H) significantly
positive/negative for +/ at COMPASS (D)
compatible with zero
20
Sivers asymmetries
 at HERMES (H)
positive/zero for +/ at COMPASS (D)
compatible with zero
The results are unexpected
21
but not inconsistent
Collins and Sivers
Kaon asymmetries
 at HERMES (H)
and
 at COMPASS (D)
are similar
except for the
Sivers K+ asymmetry
M. Düren, Univ. Giessen
22
Perspectives for futurepp experiments:
Drell Yan processes
 Parton distributions accessible similar (complementary) to DIS
 Asymmetries of polarized beam
and/or target experiments give
access to transversity
– if the energy is sufficient
(energy scale ~√(x1x2s))
– if the beam polarizations are
sufficient
 No knowledge of polarized
fragmentation functions needed!
(Contrary to DIS)
M. Düren, Univ. Giessen
qq   *     
pp      X
See talk by Anselmino
23
Hard exclusive reactions
Quantum phase-space „tomography“
of the nucleon
Quantum phase-space
Wigner distribution
 A classical particle is defined by its coordinate and
momentum (x,p): phase-space
 A state of classical identical particle system can be
described by a phase-space distribution f(x,p). The time
evolution of f(x,p) obeys the Boltzmann equation.
 In quantum mechanics, because of the uncertainty principle,
the phase-space distributions seem useless, but…
 Wigner introduced the first phase-space distribution
in quantum mechanics (1932)
 Wigner function:
M. Düren, Univ. Giessen
25
Wigner function
 When integrated over x, one gets the momentum density.
 When integrated over p, one gets the probability density.
 Any dynamical variable can be calculated from it!
The Wigner function contains the
most complete (one-body) info
about a quantum system.
!
 A Wigner operator can be defined that describes quarks in the
nucleon
 The reduced Wigner distribution is related to
Generalized parton distributions (GPDs)
M. Düren, Univ. Giessen
26
What is a GPD?

A proton matrix element which is a hybrid of elastic form factor
and Feynman distribution

Depends on
x: fraction of the longitudinal momentum carried by parton
t=q2: t-channel momentum transfer squared
ξ: skewness parameter
There are 4 important GPDs (among others):
~
~
H q ( x,  , t ), E q ( x,  , t ), H q ( x,  , t ), E q ( x,  , t )
Limiting cases:

t0: Ignoring the impact
parameters leads to ordinary
parton distributions
q( x)  H q ( x,0,0)
~
q( x)  H q ( x,0,0)

Integrating over x: Parton
momentum information is lost,
spatial distributions = form factors
remain
F 1 (t )   H q ( x,  , t ) dx
q
F2 (t )   E q ( x,  , t ) dx
q
27
3-D contours of quark distributions for
various Feynman x values
z
bx
by
Fits to the known form
factors
and parton distributions
M. Düren,
Univ. Giessen
28
with additional theoretical constraints (e.g. polynomiality) and model assumptions
Conclusions: Quarks in the
quantum mechanical phase-space

Elastic form factors  charge distribution (space coordinates)

Parton distributions  momentum distribution of quarks
(momentum space)

Generalized parton distributions (GPDs) are reduced Wigner
functions  correlation in phase-space  e.g. the orbital
momentum of quarks:
L  r p

Angular momentum of quarks can be extracted from GPDs:
Ji sum rule:

Jq 
1
1
2 1


xdx H q ( x,  ,0)  Eq ( x,  ,0)

GPDs provide a unified theoretical framework for various
experimental processes
M. Düren, Univ. Giessen
29
Hard exclusive reactions at
Kinematic coverage
M. Düren, Univ. Giessen
31
Experimental Access to GPDs
 QCD handbag diagram
Deeply virtual Compton
scattering (DVCS)
Hard exclusive meson
production (HEMP)
M. Düren, Univ. Giessen
32
Deeply virtual Compton scattering (DVCS)
 DVCS is the cleanest way to access GPDs: *N N
Factorization theorem is proven!
Handbag diagram separates

hard scattering process
(QED & QCD) and

non-pertubative structure of
the nucleon (GPDs)
x+: longitudinal momentum fraction of the quark
x
B
2: exchanged longitudinal momentum fraction   1
2 1  xB / 2
t :squared momentum transfer
GPDs = probability amplitude for N to emit a parton (x+)
and for N’ to absorb it (x-)
M. Düren, Univ. Giessen
33
DVCS and BH Interference (epe‘p)
FF
Bethe-Heitler
(BH)
Kinematics:
dσ | τ DVCS |2  | τ BH |2
 (τ τ
*
BH DVCS
τ
DVCS
τ )
*
DVCS BH
DVCS-BH interference I gives
non-zero azimuthal asymmetry

x[-1,1]
xB/(2- xB)

t=(q-q‘)2
Q2=-q2
Use BH as a vehicle to
study DVCS.
34
Laser and nucleon holography
(Belitsky/Mueller)
M. Düren, Univ. Giessen
35
Azimuthal asymmetries in
beam spin and beam charge
Fourier decomposition of interference term:
2
 I 3 I

I   c0   cn cos( n )    snI sin( n ) 
n 1
n 1


charge
spin
Access to real and imaginary part of helicity conserving amplitude M1,1
(GPDs enter in linear combinations in amplitudes)
• beam spin asymmetry (BSA)


dσep  dσep  s1I sin   sin   Im M 1,1
• beam charge asymmetry (BCA)
 
 
1,1
dσ e p  dσ e- p  c1I cos M.Düren,
cosUniv.
 Re
M
Giessen
HERMES is the only
experiment which
measures BCA
36
Results: BSA and BCA on proton


1
N ( )  N ( )


ALU ( ) 
| PB | N ( )  N ( )
BSA:
Significant sin() dependence


dσep  dσep  s1I sin   sin   Im M 1,1
N  ( )  N  ( )
AC ( )  
N ( )  N  ( )
BCA:
Significant cos() dependence
 
 
dσ e p  dσ e- p  c1I cos  cos  Re M 1,1
37
BCA and BSA on proton
HERMES results
and projected error-bars of the recoil running
M. Düren, Univ. Giessen
38
DVCS: Transverse Target Spin Asymmetry
sin(  s )cos(  )
AUT
~
Im( F2 H  F1 E )
cos(  s )sin(  )
AUT
~
Im( F2 H  F1E )
first model dependent extraction of Ju possible
Code: VGG
Jd assumed to be zero M. Düren, Univ. Giessen
39
HERMES can constrain the total angular
momentum of up and down quarks (Ju+Jd)!
same statistics with electron beam on tape
independent data set to constrain (Ju+Jd)
M. Düren, Univ. Giessen
40
Generalized EMC-Effect in nuclear DVCS
 Beam spin asymmetries in nuclear DVCS (Neon)
 More to come: 2H,4He,14N,20Ne,82-86Kr,129-134Xe:
A-dependence of coherent DVCS processes
to study quarks in nuclei
M. Düren, Univ. Giessen
41
First results for Hard Exclusive Meson Production
(HEMP) from HERMES: Pseudoscalar Mesons
σ L  Sσ L  S||σ LT  A
sinφ
UL
sinφ
~ ~
σ | ST | sinφ  E  H
 Pion form factor
HEMP cross section
ep  e’n +
 q
GPD Model:
42
Vanderhaeghen, Guichon, Guidal
First results for HEMP from HERMES:
Vector Mesons (r0,,w)
-Clean signal without background subtraction;
-DIS-background well described by MC;
HEMP target spin
asymmetry
Univ.
e’pGiessen
r0
M. ep
Düren,
43
M. Düren, Univ. Giessen
44
Quarks in nuclei
Study fragmentation
in nuclei

 understanding of the space-time
evolution of the hadron formation
process
K
 N h (z) 


N
e A

R hM (z) 
 N h (z) 


N
e D

p
M. Düren, Univ. Giessen
47
2-hadron production in nuclei
 disentangle absorption
and energy loss models
leading hadron: z1 0.5
subleading hadron: z2  z1
 N 2 (z 2 ) 


 N1  A
R 2h (z 2 ) 
 N 2 (z 2 ) 


 N1  D
M. Düren, Univ. Giessen
48
Hard exclusive reactions and GPDs in
futurepp experiments
Perspectives for futurepp experiments
 For exclusive processes,
„crossing“ allows to measure
the same matrix elements
with completely different
experiments and probes
 Instead of having an initial
electromagnetic process, a
final state electromagnetic
process is selected
Example:
Compton scattering
Annihilation
M. Düren, Univ. Giessen
 p  p
pp  
50
Perspectives for futurepp experiments
 Experimental difficulty:
the total cross section of a
hadronic beam is typically
much larger than the one of
the exclusive electromagnetic process
 The experiment needs a huge
background suppression
factor as the majority of
events have a hadronic final
state
Example:
Compton scattering
Annihilation
M. Düren, Univ. Giessen
 p  p
pp  
51
Turning Matter into Light:
pp annihilation at large energies and angles
What is the dominant mechanism?

Large pT in this exclusive reaction
defines hard scale: parton picture

Emission of a single photon from one
(quasi-free) parton is suppressed

Emission of two photons from the
same parton is allowed

At intermediate energies (s~10
GeV) the hard interaction of one
parton can be separated from the
soft part which is parameterized by
General Distribution Amplitudes:
handbag diagram
See talk by Kroll
Quasi-free
partons
cannot emit
single photons
(suppressed)
Handbag
diagram: one
quark emits
both photons
(allowed)
Soft part of
annihilation
M. Düren, Univ. Giessen
52
Further processes of interest:
Reactions with the QCD handbag diagram
hard
gluon
, r, , ...
Hard exclusive
meson
production
(large pT)
Q2 small: like Crossedchannel wide angle Compton
scattering (large pT)
Q2 large: additional
degrees of freedom
e-, 
e+, 
M. Düren, Univ. Giessen
53
Further processes of interest:
Another „distribution amplitude“
•No handbag diagram
•Here the photons and the pion are produced in forward direction!
•Measure „Transition distribution amplitudes“
pp   *  explores the pion cloud
pp   * r explores the r cloud
pp   *  explores the photon cloud
M. Düren, Univ. Giessen
(Study next to lowest
Fock state of the proton)
See talk by
54
B. Pire and L. Szymanowski
Turning light into matter.....
 Compare pp  
with the time inverted
process   pp
Results e.g. from Belle:
 Asymmetric e+e- collider
 High luminosity machine
 Available at e+e- machines:
quasi real photons
s  10.6 GeV
L  1034 /cm 2s
e  e   e  e  * *  e  e pp
M. Düren, Univ. Giessen
55
Recent results from Belle
Energy dependence
  pp
Angular dependence
M. Düren, Univ. (GPD
Giessen curve from Kroll/Schäfer)
56
Processes of interest:
hard
gluon
-process is not only background but also signal!
, r, , ...
Hard exclusive
meson
production
(large pT)
•Much larger cross section
(compared to ) makes
it easier to access!
•3- final state
M. Düren, Univ. Giessen
2
E760 results
57
(curve from Kroll/Schäfer)
More physics potential at PANDA:
Time-like proton Formfactors at large q2
 wide kinematic range
accessible at PANDA
(up to q²~20 GeV²)
q2>0
 high statistical
precision possible
 separate GM and GE
q2<0
M. Düren, Univ. Giessen
58
A few words on polarization:
Target polarization
 PAX wants to use a polarized H,D storage cell
target (like HERMES)
 PANDA could (in principle) use a polarized 3He
storage cell target (like HERMES in 1995) where
the polarized 3He is injected from outside the
solenoid into the cell
 A more attractive option is to run HESR in collider
mode (to obtain high enough energy especially for
Drell Yan) with both beams polarized
 The PANDA detector would be suitable for the
asymmetric collider mode, so (part of) the PAX
and ASSIA program could be done in an extended
PANDA detector (“PANDAX”)
M. Düren, Univ. Giessen
59
A few words on polarization:
Polarization of the final state
The polarization of the final state can be determined
in some special cases:
 Using parity violation of weak decays (e.g. of ´s)
by observing angular distributions of decay
products
 Angular distribution of decay products without
parity violation (e.g. decay of vector mesons or of
virtual photons)
 New idea: angular distribution of the nuclear
interaction of proton in calorimeter can be used to
determine the transverse proton spin (from the
azimuthal distribution of the shower profile in the
calorimeter)
M. Düren, Univ. Giessen
60
Some spin physics possible
with out polarized target!
M. Düren, Univ. Giessen
61
Conclusions
 HERMES has done and still does unique and pioneering
measurements in the fields of
– Spin structure of the nucleon
– Exclusive reactions and GPDs
– Hadronization studies in nuclei
– …
 Part of the program can be completed by the measurement
of complementary reactions inpp annihilation experiments
at PANDA, PAX and ASSIA
 An upgrade of HESR to a polarizedpp-collider seems ideal
to follow up the HERMES program.
M. Düren, Univ. Giessen
62