1. Status and prospects for GPDs studies at COMPASS
Etienne Burtin, CEA/Saclay, DAPNIA/SPhN
on behalf of the COMPASS collaboration
1- Physics impact
2- Experimental issues
3- Recoil detector prototype
GPD2006
June 7, Trento, Italia

2. Generalized Parton Distributions
g,p,r
Factorisation:
Q2 large, -t<1 GeV2
hard
x+x
x-x
soft
GPDs P P’ t
Generalized Parton Distributions
H(x,0,0) = q(x)
measured in DIS
for quarks :
4 functions H(x,x,t)
F(t)
measured in elastic scattering

3. DVCS observables Deep VCS Bethe-Heitler High energy beam
Lower energy => use interference - holography
Cross section
Single Spin Asymmetry
Polarised beam
Beam Charge Asymmetry
+/- charged beam
COMPASS muon beam can do all !

4. DVCS+ Bethe Heitler φ θ μ’ μ *  p BH calculable
The high energy muon beam at COMPASS
allows to play with the
relative contributions DVCS-BH
which depend on
1/y = 2 mp Eℓ xBj /Q2
Higher energy: DVCS>>BH
 DVCS Cross section
Smaller energy: DVCS~BH
Interference term will provide
the DVCS amplitude φ θ μ’ μ *  p

5. Polarized μ+ and μ- beams
Polarized beam: Ep=110 GeV → Em=100 GeV
P(m+) = m/spill (5s)
P(m-) = m/spill (5s)
Same collimator settings :
beam profile unchanged
Switch in 10 mins:
could be done every 8h
Muon section 400m
Hadron decay section 600m
Compass
target
scrapers
Be absorbers
Protons
400 GeV
T6 primary
Be target
Collimators
H V H V
2.108 m/spill
protons/spill

6. Extraction of GPDs in the case of μ+ / μ-
t, ξ~xBj/2 fixed
dσ(μpμp) = dσBH + dσDVCSunpol + Pμ dσDVCSpol
+ eμ aBH Re ADVCS eμ Pμ aBH Im ADVCS
 cos nφ  sin nφ φ θ μ’ μ *  p
Pμ+=-0.8 Pμ-=+0.8

7. Kinematical domain Ix2 Collider : H1 & ZEUS 0.0001<x<0.01
100GeV
Fixed target :
JLAB GeV SSA,BCA?
HERMES 27 GeV SSA,BCA
COMPASS could provide data on :
Cross section (190 GeV)
BCA (100 GeV)
Wide Q2 and xbj ranges
Limitations due to luminosity
new LINAC 4 (SPS injection) in 2010
Radioprotection limits
needs better shielding

8. Sensitivity of BCA to models
Model 1: H(x,ξ,t) ~ q(x) F(t)
Model 2: from Goeke, Polyakov
and Vanderhaeghen
model 1
model 2
H(x,0,t) = q(x) e t <b2>
= q(x) / xα’ t
COMPASS
sensitivity to the different spatial
distribution of partons  when xBj 
Good sensitivity to models
in COMPASS xBj range

9. Projected errors of a possible DVCS experiment
Beam Charge Asymmetry
L = cm-2 s-1
Ebeam = 100 GeV
6 month data taking
25 % global efficiency
6/18 (x,Q²) data samples
3 bins in xBj= 0.05, 0.1, 0.2
6 bins in Q2 from 2 to 7 GeV2
Model 1 : H(x,ξ,t) ~ q(x) F(t)
Model 2 :
H(x,0,t) = q(x) / xα’t
Good constrains for models

10. Hard Exclusive Meson Production (ρ,ω,…,π,η… )
Scaling predictions:
meson p p’
GPDs g*
x + ξ
x - ξ
t =Δ2 L
hard
soft
1/Q6
1/Q4
Collins et al. (PRD ):
1. factorization applies only for g*L
2. σT << σL
vector mesons pseudo-scalar mesons
ρ0 largest production present study
ρ0  π+ π with COMPASS

11. Hard Exclusive Meson Production
It comes for free with the recoil detector and existing COMPASS trackers
Cross section:
Vector meson production (ρ,ω,…)  H & E
Pseudo-scalar production (π,η… )  H & E ~ ~
Hρ0 = 1/2 (2/3 Hu + 1/3 Hd + 3/8 Hg)
Hω = 1/2 (2/3 Hu – 1/3 Hd + 1/8 Hg)
H = /3 Hs - 1/8 Hg
with present COMPASS data
Can be investigated
Single spin asymmetry ~ E/H
for a transverse polarized target

12. 250 physicists 26 institutes
Compass Set-up
at CERN
250 physicists 26 institutes
magnets
muon filter
Calorimeters
160 GeV
pol. m beam
~ 200 detection planes
Silicon, SciFi, Micromegas,
Drift chambers, GEM, Straw
chambers, MWPC
RICH
polarized
target

13. Incoherent exclusive r0 production in COMPASS DATA
Preview of
A. Sandacz’s talk
Mpp
Emiss
pt² m m’ p- p+ N N’ g* r0
2002
Impact on GPD :
sL is dominant at high Q²
(factorisation only valid for sL)

14. Additionnal equipment to the COMPASS setup required for DVCS
all COMPASS trackers:
SciFi, Si, μΩ, Gem, DC, Straw, MWPC
2.5 m Liquid H2 target
to be designed and built μ’
ECAL 1 or 2
  12° 
COMPASS equipment
with additional calorimetry
at large angle p’ μ
Recoil detector to insure exclusivity
to be designed and built

15. Recoil detector design
Goals: Detect protons of MeV/c
t resolution => sTOF = 200 ps
exclusivity => Hermetic detector
Design :
2 concentric barrels of 24 scintillators counters read at both sides
European funding (127 k€) through a JRA for studies and
construction of a prototype ( Bonn, Mainz, Saclay, Warsaw)

16. Physical Background to DVCS
Source : Pythia 6.1 generated DIS events
Apply DVCS-like cuts
one m’,g,p in DVCS range
no other charged & neutral in active volumes
detector requirements:
24° coverage for neutral
50 MeV calorimeter threshold
40° for charged particles
in this case
DVCS is dominant

17. Timing and triggering issues
Slow protons
TOF~60ns
Fast protons
TOF=3ns
PMT
PMT
B: 0.7 MHz/counter
PMT
PMT
A: 2 MHz/counter
Time window
kinematics (60ns)
light propagation (up to 40 ns)
safety margin (30 ns)
=> all signal in a 128 ns window
Light attenuation
- ADC(PMT) = Edep e-x/l
Analog trigger
- hard to implement
- background… (next slides )
Digital trigger
- could work at 50 MHz
- t>trigger latency…
One solution : use the inclusive trigger
and sample the signals
DAQ rate : 100 kHz

18. Part I : Simulation of the Recoil Detector
Goals :
- evaluate front-end requirements
- tune reconstruction algorithms
Dedicated Géant 3.21 program
LH2 target with envelope
inner scint. 4 mm thick
outer scint. 5 cm thick
light attenuation
all processes turned on
custom beam halo generation
PMT waveform generation

19. Effect of d-rays
With simulation of d-rays

20. Digitization : Waveform generation
For each Geant step in active volume :
record DE,z,t
propagate step to photocathode exit by applying:
decay constant of the scintillator
speed of light in the medium
attenuation length
light yield + quantum efficiency
fill the histogram of time of production of gelectrons (0.1 ns bins)
Add contribution from additional muons according to intensity
(2 108 m in a 5s long spill)
Smear with a gaussian
of s=3ns to mimic the
time response of the PMT
now 1 ns bins (1Gs/s) up
down
-30ns ns

21. PMT signals : only 1m in the set-up
Red is DVCS proton
Blue is background
upstream
PMT
downstream
INNER
OUTER

22. PMT signals : 2 108 m/spill (5s)
recording the
waveform of all
signals and segmentation are mandatory
Hints for analysis:
facing counters
m m’ vertex (z,t)
DEINNER vs DEOUTER
DEOUTER vs b

23. Simple waveform analysis
For each PMT signal (up to 5 “hits”):
perform leading edge discrimination
correct for time-walk effect
extract pulse height and integrate charge (35ns window)
For each up/down pair (up to 5*5=25 ”points”):
calculate time & position of crossing point using time information
unfold attenuation length to extract energy loss
request that position is within scintillator volume +/- 10cm

24. Criteria for proton candidatesBi
Ai+1 Ai
Have points in corresponding A and B counters
For each pair of “points”
Energy loss correlation
Energy loss vs bmeas correlation b
DEA
DEB
( no additional beam
muons in this plot –
just for pedagogy )

25. Coincidence with the scattered muon
Use reconstructed muon vertex time
to constraint proton candidates
Use vertex position to evaluate
the effective signal

26. Performances : number of events with proton identified Efficiency =
number of “triggers”
Efficiency =
trigger = one event with at least one good combination of A and B with hits
identified proton = f of proton candidate matches f of generated proton
Seff for 1000 events
m/5s spill

27. Recoil Detector Prototype
4 m
Mechanical design for the prototype
30° sector , Scale 1
of what could be
the DVCS recoil detector

28. tests on the m beam in fall-2006Bi
Ai+1 Ai CH
Target
Inner
Layer
Outer Layer
Goals for this 2-month data taking period:
Validate detection with sTOF ~ 250 ps
Scintillators + Light guides + PMT + electronics
Measure background with equivalent CH2 target ( ~1MHz/detector)
use 1GHz sampler of the signals with a stand-alone DAQ
and standard multihit TDCs and ADCs

29. next step : write the proposal !
Conclusions
This initiative is now an “Express of Interest” : SPSC-EOI-005
Towards a GPD experiment using COMPASS…
- COMPASS is complementary to other experiments
- has good sensitivity to GPD models through BCA
- has good Q² range for 0.03<xbj<0.2
Experimental challenges
- measure TOF with good resolution in high background conditions
- 1 GHz sampling and recording of the PMT signals
- increase beam intensity …
- recoil detector prototype test this fall
next step : write the proposal !

30. Spare

31. Roadmap for GPDs at COMPASS
2005: Expression of interest SPSC-EOI-005
2006: Test of recoil detector prototype
Proposal
: construction of
recoil detector
LH2 target
ECAL0
≥ 2010: Study of GPDs at COMPASS
In parallel present COMPASS studies with polarised target
Complete analysis of ρ production
Other channels: , 2π …
GPD E/H investigation with the transverse polarized target

32. Proton luminosity upgrade at CERN
Gatignon
compass

33. Front-End Electronics
COMPASS DAQ : 100 kHz – no dead time
Data flow from recoil detector : 2.5 GByte/s
100 kHz, 100 channels, 128 samples, 10 bits
=> 25 % increase in event size
discussion about the DC282 module
10 bits resolution (16 bits used for coding)
350ns dead time
memory depth MBytes
400 MBytes/s PCI transfer