Transcript Simulation of DVCS with an EIC using MILOU Salvatore Fazio BNL g* p g p University of Washington – INT Seattle, WA – USA November 1-13, 2010

Simulation of DVCS with an EIC
using MILOU
Salvatore Fazio
BNL
g*
p
g
p
University of Washington – INT
Seattle, WA – USA
November 1-13, 2010
Planing of the talk

The eRHIC accelerator and an EIC detector
compared to HERA

Strategy of a DVCS measurement

Hera results

Extension to EIC

DVCS simulation for an EIC

Summary
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
2
From HERA to an EIC collider
 27.5 GeV electrons/positrons on 920 GeV
EIC/eRHIC
protons →√s=318 GeV
 2 colliding experiments: H1
and ZEUS
Polarized egun
Beamdump
eRHIC
detector
6 pass 2.5 GeV
ERL
 Total lumi collected at HERA: 500 pb-1,
polarization of electrons/positrons at HERA II
HERA
STAR
 20 - 30 GeV electrons on 325 (125)
GeV protons (nuclei). Polarization of
electrons and protons (nuclei)
 Lumi: 1.4 x 1034 cm-2s-1
For exclusive diffraction the concept is
similar to HERA but:
• Dedicated forward instrumentation
• Higher tracker coverage
• Very High lumi!
Detectors not originally designed for
forward physics, Roman pots added
later
S. Fazio: INT-workshop, Univ. of
Nov. 9, 2010
Washington, Seattle
3
The EIC detector
p/N
e
REAR
FORWARD
Estimated
b*≈
8 cm pc/2.5
Similarities with HERA detectors:
• Hermetic
• Asymmetric
4.5 cm
neutrons
11.2 cm
e IP
q=10 mrad
2
Nov. 9, 2010
4
6
8
Dipole:
2.5 m, 6 T
q=18 mrad
10
12
Important (as respect of HERA)
improvements:
• Central Tracking Detector
• better em calorimeter resolution
• Very forward calorimetry
• Rear Trackers!
• Roman pots from the early biginning
14
16
Quad Gradient:
200 T/m
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
4
20
0.44 m
0.329 m
0.188036 m
10
30 GeV e-
30
60 m
90 m
© D.Trbojevic
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
5
Deeply Virtual Compton Scattering
γ*
VM (ρ, ω, φ, J/ψ, Υ)
V
γ*
DVCS (γ)
IP
p
p
p
p
Scale: Q2 + M2
γ
Q2
DVCS properties:
• Similar to VM production, but γ instead of VM in the final state
• No VM wave-function involved
• Important to determine Generalized Parton Distributions sensible to
the correlations in the proton
• GPDs are an ingredient for estimating diffractive cross sections at
LHC
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
GPD
GPD
6
Accessing the GPDs
quantum number of final state
selects different GPDs:
 theoretically very clean
~ ~
DVCS (g): H, E, H, E
 VM (r, w, f): H E
 info on quark flavors
~ ~
PS mesons (p, h): H E
1
1
z
z
 J q + J g   Dq +  Lzq + J gz
2
2 q
q
1
z
J q   Dq +  Lzq
2 q
q
1
1
z
q
q 
Jq 
x
dx
H
+
E
 t0
2  -1

Nov. 9, 2010
)
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
p0
h
2Du+Dd
2Du-Dd
ρ0
ω
f
ρ+
2u+d, 9g/4
2u-d, 3g/4
s, g
J/ψ
g
u-d
7
DVCS @ ZEUS - Strategy
DVCS
BH
e
e
g
g*
p
p
g sample: no tracks matching to the
second candidate
e sample: a track match to the
second candidate
Wrong-sign
sample: a negative track match to the
second candidate
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
(DVCS+BH)
(BH+ dilepton + J/)
(dilepton + J/)
8
DVCS @ HERA
Fit: σ ~ Wδ
Q2 dependence for the W slope
not clear within the uncertainties!
ZEUS: JHEP05(2009)108
H1: Phys.Lett.B659:796-806,2008
t measured indirectly:
2
 2

t ~ PTg + PT2 
e 

Nov. 9, 2010
d
 e -b |t |
dt


by roman pots!
Fit :
No evidence for W dependence of b
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
9
t dependence
The ZEUS result is in agreement with H1
b = 4.5 ± 1.3 ± 0.4
…nevertheless it seems to suggest a lower
trend!
dσ/dt measured for the first time by a
direct measurement of the outgoing
proton 4-momentum using the LPS
spectrometer
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
10
W-dependence: summary
Summary of the W,t-dependence for all VMs + DVCS measured at HERA
Fit: σ ~ Wδ
Fit :
d
 e -b |t |
dt

Exclusive production of VMs can be a golden measurement for an EIC
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
11
Measuring t indirectly with an EIC
To successfully measure t indirectly from the electron and photon candidates
t ~ PT2g +PT2
2
e
it is important:
Tracker coverage (tracker has higher momentum resolution than Cal!)
Reso of the CTD @ ZEUS: σ(pT)/pT =0.0058pT⊕0.0065⊕ 0.0014/pT

DVCS/BH
CTD acceptance @ ZEUS
BH
Always measure a track when we can -> better momentum resolution but
not only… More acceptance for
DVCS!
High resolution em calorimetry (crucial! Remember that one particle is a photon!)
For ZEUS it was σ(E)/E=0.18/√E
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
12
Direct t measurement at EIC
But… is an indirect measurement of t really an issue for EIC?
We’ll get roman pots in the forward region at EIC!
Silicon micro-strips
resolution: 0.5% for PL ; 5 MeV for PT
L = 27.77 pb-1
55 events (DVCS + BH)
EIC lumi
for eRHIC: 1.4 1034* Ep/325 cm-2s-1
assuming 50% operations efficiency one week corresponds to:
L(1 w)= 0.5 * 604800(s in a week) * (1.4x1034 cm-2s-1) = 4*1039 cm-2 = 4000pb-1
+ Roman Pots
~ 8000 events/week !!
assuming the same acceptance ad LPS (~2%)
Calculations are absolutely not rigorous! But give an idea…
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
13
t vs proton scattering angle
t=(p4-p2)2 = 2[(mpin.mpout)-(EinEout - pzinpzout)]
t=(p3–p1)2 = mρ2-Q2 - 2(Eγ*Eρ-pxγ*pxρ-pyγ*pyρ-pzγ*pzρ)
4 GeV el
x 50 GeV prot
4 x 100
very strong correlation between
t and “recoiling” proton angle
 Roman pots need to be very
well integrated
 resolution on t!
Nov. 9, 2010
4 x 250
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
14
MILOU
Written by E. Perez, L Schoeffel, L. Favart [arXiv:hep-ph/0411389v1]
The code MILOU contains two different models for DVCS simulation:
FFS
d

dxdQ2 dt
3
DVCS
Based on: Frankfurt, Freund and Strikman (FFS)
[Phys. Rev. D 67, 036001 (1998). Err. Ibid. D 59 119901 (1999)]

p 2 3 s 1 + 1 - y )
2
2
2xR Q
6
)e
-b t
F22 x,Q2 )1 + r 2 )
• Old model: written before data came out!
• Used by H1 and ZEUS for their DVCS measurements
• The ALLM parametrization for F2 is used
Based on: A. Freund and M. McDermott
All ref. in: http://durpdg.dur.ac.uk/hepdata/dvcs.html
GPDs-based
• GPDs, evolved at NLO by an indipendent code which provides tables of CFF
- at LO, the CFFs are just a convolution of GPDs:


ei2
H (,Q ,t) =   -1 
  -Hi x, ,Q2,t )dx
1 - x  - i

u,d,s
2
Nov. 9, 2010
1
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
15
MILOU
• provide the real and imaginary parts of Compton form factors (CFFs), used to
calculate cross sections for DVCS and DVCS-BH interference.
d
 3 xB y
I
=
dxdyd| t | dfd 16p 2Q2 1+  2 e 3

I BH
2
I DVCS
f = fN - f l
 = T - f N
  2x
mN
Q
 BH 2 BH

e6
BH

 2 2
c 0 +  c n cosnf ) + s1 sinf )
2 2 2
x y (1+  ) D P1 (f )P2 (f ) 
 
n
1
2
2

e 6  DVCS
DVCS
DVCS
 2 2 c 0
+  c n cosnf ) + sn sinnf )
yQ 

n 1
 I 3 I

e 6
I
c +  c cosnf ) + s1 sinf )
I  3 2
xy D P1 (f )P2 (f )  0 n 1 n

2
• The B slope is allowed to be costant or to vary with Q2:


)
d d t  exp BQ2 )t ; with : BQ2 ) lnQ2 )
• Proton dissociation (ep → eγY) can be included

Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
16
Phase space
• 1.5 < Q2 < 100 GeV2
• 10-4 < x < 0.1
• 0.01 < y < 0.85
• 0 < |t| < 1.5 GeV2
• Radiative corrections: OFF
• t slope: B = 5.00 (costant)
• GPDs evolved at NLO
• ALLM param. used for F2 (FFS model)
100 k event generated for each
config.
30 X 325:
• E_el = 10 GeV
ep -> egp)= 0.186 nb (FFS_ALLM)
ep -> egp)= 0.376 nb (GPDs)
20 X 250:
• E_el = 5 GeV
ep -> egp)= 0.16 nb (FFS_ALLM)
ep -> egp)= 0.32 nb (GPDs)
10 X 100:
• E_el = 0 GeV
ep -> egp)= 8.1*10-2 nb (FFS_ALLM)
ep -> egp)= 0.16 nb (GPDs)
5 X 50:
• E_el = 0 GeV
ep -> egp)= 8.1*10-2 nb (FFS_ALLM)
ep -> egp)= 0.16 nb (GPDs)
eRHIC Luminosity
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
17
qg vs Eg
30 X 325
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
18
qg vs Eg
20 X 250
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
19
qg vs Eg
10 X 100
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
20
qg vs Eg
5 X 50
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
21
t-xsec (ep -> gp)
FFS - 30 X 325
by roman pots!
• 1.5 < Q2 < 100 GeV2
• 10-4 < x < 0.1
• 0.01 < y < 0.8
L = 0.54 fb-1
EIC lumi: 4 fb-1/month @ 30x325
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
• Precision enormously improved
• Roman pots acceptance not yet
included in the simolation
22
t-xsec
GPDs conv. - 30 X 325
• 1.5 < Q2 < 100 GeV2
• 10-4 < x < 0.1
• 0.01 < y < 0.8
Systematics will dominate!!
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
23
t-xsec
20 X 250
10 X 100
5 X 50
 dominated by gluon contributions
 EIC will provide sufficient luminosity to bin in multi-dimensions
 wide x and Q2 range needed to extract GPDs
… we can do a fine binning in Q2 and W!
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
24
Scanning the phase space…
30 X 325
20 X 250
Logarithmic bins:
1 < Q2 < 100 GeV2
10-4 < x < 0.1
10 X 100
Nov. 9, 2010
5 X 50
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
25
Scanning the phase space…
30 X 325
20 X 250
10 X 100
5 X 50
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
26
30x325-t-xsec
Veri precice scan!
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
27
20x250-t-xsec
GPD
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
FFS
28
10x100 5x50 t-xsec
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
29
DVCS: the beam-charge asymmetry
A = ADVCS  +ABH  +AI 
2
2
2
2
Interference term:
Beam charge asymmetry:

DVCS and BH: identical final state → they Interfere
AI  ReADVCS )+ImADVCS )
d + - d AC =
+
-  Re ADVCS )
d + d

The phi angle
At EIC:
d+ - d AC =
d + + d Nov. 9, 2010
Possible if a positron beam
Thanks to a good tracker
coverage
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
30
DVCS on nuclear targets
 How does the nuclear environment modify parton-parton correlations?
 How do nucleon properties change in the nuclear medium?
 DVCS
in coherent region:
new insights into ‘generalized EMC effect’?
(Bethe-Heitler)
 Nuclear GPDs ≠ GPDs of free nucleon
 Enhancement of effect when leaving forward limit?
 caused by transverse motion of partons in
nuclei?
 important role of mesonic degrees of freedom?
 manifest in strong increase of real part of τDVCS
with atomic mass number A?
MC simulation for DVCS on nuclei coming soon thanks to an updated version of MILOU code
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
31
Summary

A lot of experience carried over from HERA

Simulation shows how an EIC forward program can sensibly improve HERA
results and go beyond

Uncertainties will be dominated by systematics

Large potential for diffractive-DIS studies using polarized and unpolarized
protons and nuclei
Outlook:

Simulation of asymmetries

Simulation of DVCS on nuclei

Updating the MILOU code to status of art (Re_Amp, NNLO…)
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
32
Back up
Nov. 9, 2010
S. Fazio: INT-workshop, Univ. of
Washington, Seattle
33