salvatore fazio bnl university of washington – int seattle, wa – usa november 1-13, 2010...

Salvatore Fazio BNL

University of Washington – INT Seattle, WA – USA

November 1-13, 2010

Simulation of DVCS with an EICusing MILOU

p p

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 2S. Fazio: INT-workshop, Univ. of Washington, Seattle

27.5 GeV electrons/positrons on 920 GeV protons →√s=318 GeV

2 colliding experiments: H1 and ZEUS

Total lumi collected at HERA: 500 pb-1, polarization of electrons/positrons at HERA II

Detectors not originally designed for forward physics, Roman pots added later

From HERA to an EIC collider

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!

STAR

ePHEN

IX

6 pass 2.5 GeV ERL

Coherent

e-cooler

Beam-dump

Polarized e-gun eRHIC

detector

EIC/eRHIC

HERA


Important (as respect of HERA) improvements:• Central Tracking Detector• better em calorimeter resolution• Very forward calorimetry • Rear Trackers! • Roman pots from the early biginning

The EIC detector

ep/N

FORW

ARDREAR


Similarities with HERA detectors:• Hermetic• Asymmetric

22 44 66 88 1100

2.5 m2.5 m

3.5 3.5

mm

1212 1144

90 mm90 mm

5.75 m5.75 m

1616IPIP

Dipole:Dipole:2.5 m, 6 T2.5 m, 6 T=18 mrad=18 mrad

4.5 m4.5 m=18 mrad=18 mrad

=10 mrad=10 mrad

Estimated Estimated **≈ 8 cm≈ 8 cm

=44 mrad

=44 mrad

6.3 cm6.3 cmZDCZDC

pp cc/2.5/2.5

15.7 cm15.7 cm

6 mrad6 mrad11.2 cm11.2 cm

4.5 cm4.5 cmneutronsneutronsppcc/2.5/2.5

ee

Quad Gradient: Quad Gradient: 200 T/m200 T/m

Nov. 9, 2010 S. Fazio: INT-workshop, Univ. of Washington, Seattle 5

0.4

4 m

0.4

4 m

90 m90 m

10 mrad10 mrad 0

.329

m0.3

29

m3.67 mrad3.67 mrad

60 m60 m1010 2020 3030

0.1

88

036

m0.1

88

036

m

18.8

18.8

mm16.8

16.8

mm

6.33 mrad

6.33 mrad4 m4 m

DipoleDipole

© D.Trbojevic© D.Trbojevic

30 GeV e30 GeV e--

325 GeV p325 GeV p

125 GeV/u ions

125 GeV/u ions SpinSpin

rotatorrotator

GPD

GPD

p p

γγ*

Deeply Virtual Compton ScatteringVM (ρ, ω, φ, J/ψ, Υ) DVCS (γ)

Q2Q2 + M2Scale:

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

IPp p

Vγ*



theoretically very cleantheoretically very clean DVCSDVCS ( (): H, E, H, E): H, E, H, E VMVM ( ( H E H E info on quark flavorsinfo on quark flavors PS mesonsPS mesons ( (: H E : H E

~~

~~ ~~

~~

quantum number of final state quantum number of final state selects different GPDs:selects different GPDs:

0 2ud

2ud

ρ0 2ud, 9g/4

ω 2ud, 3g/4

s, g

ρ+ ud

J/ψ g J

qz

1

2x dx H q E q

1

1

t 0

J

qz

1

2q

q L

qz

q

1

2J

qz J

gz

1

2q

q L

qz

q J

gz

Accessing the GPDs

Nov. 9, 2010 8

BHDVCS

p p

ee

Wrong-sign sample: a negative track match to the second candidate

sample: no tracks matching to the second candidate

e sample: a track match to the second candidate

(DVCS+BH)

(BH+ dilepton + J/)

(dilepton + J/)

DVCS @ ZEUS - Strategy

S. Fazio: INT-workshop, Univ. of Washington, Seattle

Q2 dependence for the W slope not clear within the uncertainties!

ZEUS: JHEP05(2009)108H1: Phys.Lett.B659:796-806,2008


Fit: σ ~ Wδ

DVCS @ HERA

t ~ PT2 +PTe

2

2

t measured indirectly:

Fit :ddt

e b |t |

No evidence for W dependence of bby roman pots!

Nov. 9, 2010 10

b = 4.5 ± 1.3 ± 0.4

dσ/dt measured for the first time by a direct measurement of the outgoing proton 4-momentum using the LPS spectrometer

The ZEUS result is in agreement with H1

…nevertheless it seems to suggest a lower trend!

t dependence

S. Fazio: INT-workshop, Univ. of Washington, Seattle

W-dependence: summary

Summary of the W,t-dependence for all VMs + DVCS measured at HERA


Fit: σ~ Wδ

Fit :ddt

e b |t |

Exclusive production of VMs can be a golden measurement for an EIC

To successfully measure t indirectly from the electron and photon candidates

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/p⊕ ⊕ T

High resolution em calorimetry (crucial! Remember that one particle is a photon!)

For ZEUS it was σ(E)/E=0.18/√E

Measuring t indirectly with an EIC

CTD acceptance @ ZEUS

DVCS/BH BHAlways measure a track when we can -> better momentum resolution but not only… More acceptance for DVCS!


t ~ PT2 +PTe

22

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!

L = 27.77 pb-1

55 events (DVCS + BH)

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…

Silicon micro-stripsresolution: 0.5% for PL ; 5 MeV for PT


EIC lumi

t vs proton scattering angle


4 GeV el x 50 GeV prot 4 x 100

4 x 250

very strong correlation between t and “recoiling” proton angle Roman pots need to be very well integrated resolution on t!

t=(p4-p2)2 = 2[(mpin.mp

out)-(EinEout - pzinpz

out)]

t=(p3–p1)2 = mρ2-Q2 - 2(Eγ*Eρ-px

γ*pxρ-py

γ*pyρ-pz

γ*pzρ)

1414

Nov. 9, 2010 S. Fazio: INT-workshop, Univ. of Washington, Seattle

MILOU

d3DVCS

dxdQ2dt

2 3s 1 1 y 2 2xR2Q6 e b t F2

2 x,Q2 12

Based on: Frankfurt, Freund and Strikman (FFS) [Phys. Rev. D 67, 036001 (1998). Err. Ibid. D 59 119901 (1999)]

FFS

GPDs-basedBased on: A. Freund and M. McDermott

All ref. in: http://durpdg.dur.ac.uk/hepdata/dvcs.html

• GPDs, evolved at NLO by an indipendent code which provides tables of CFF

- at LO, the CFFs are just a convolution of GPDs:

15

H (,Q2,t) =ei

2

1 x i

Hi x,,Q2,t dx

1

1u,d,s

• Old model: written before data came out! • Used by H1 and ZEUS for their DVCS measurements• The ALLM parametrization for F2 is used

The code MILOU contains two different models for DVCS simulation:Written by E. Perez, L Schoeffel, L. Favart [arXiv:hep-ph/0411389v1]

• provide the real and imaginary parts of Compton form factors (CFFs), used to calculate cross sections for DVCS and DVCS-BH interference.

• The B slope is allowed to be costant or to vary with Q2:

• Proton dissociation (ep → eγY) can be included


MILOU

ddxdyd | t | dd

= 3xB y

16 2Q2 1 2

Ie3

= N l

= T N

2xmN

Q

I BH

2

e6

x 2y 2(1 2)22P1()P2()c0

BH cnBH cos n s1

BH sin n1

2

I DVCS

2

e6

y 2Q2 c0DVCS cn

DVCS cos n snDVCS sin n

n1

2

I 2 e6

xy 32P1()P2()c0

I cnI cos n s1

I sin n1

3

d d t exp B Q2 t ; with : B Q2 ln Q2


• 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)

30 X 325:• E_el = 10 GeVep -> ep)= 0.186 nb (FFS_ALLM)ep -> ep)= 0.376 nb (GPDs)

20 X 250:• E_el = 5 GeVep -> ep)= 0.16 nb (FFS_ALLM)ep -> ep)= 0.32 nb (GPDs)

10 X 100:• E_el = 0 GeVep -> ep)= 8.1*10-2 nb (FFS_ALLM)ep -> ep)= 0.16 nb (GPDs)

17

5 X 50:• E_el = 0 GeVep -> ep)= 8.1*10-2 nb (FFS_ALLM)ep -> ep)= 0.16 nb (GPDs)

Phase space

eRHIC Luminosity

100 k event generated for each config.


vs E

30 X 325


vs E

20 X 250


vs E

10 X 100


vs E

5 X 50


t-xsec (ep -> p)FFS - 30 X 325

by roman pots!

L = 0.54 fb-1

EIC lumi: 4 fb-1/month @ 30x325 • Precision enormously improved• Roman pots acceptance not yet

included in the simolation

• 1.5 < Q2 < 100 GeV2

• 10-4 < x < 0.1• 0.01 < y < 0.8


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!!


20 X 250 10 X 100 5 X 50

t-xsec

dominated by gluon contributionsdominated by gluon contributions EIC will provide sufficient luminosity to bin in multi-dimensionsEIC will provide sufficient luminosity to bin in multi-dimensions wide x and Qwide x and Q22 range needed to extract GPDs range needed to extract GPDs

… … we can do a fine binning in Q2 and W!we can do a fine binning in Q2 and W!


Scanning the phase space…

Logarithmic bins:

1 < Q2 < 100 GeV2

10-4 < x < 0.1

y = 0.85

y = 0.

01

y = 0.85

y = 0.85

y = 0.85

y = 0.

01

20 X 250

10 X 100

30 X 325

5 X 50


Scanning the phase space…

y = 0.

01

y = 0.

85

y = 0.

01

y = 0.

85

y = 0.

85

y = 0.

85

30 X 325 20 X 250

10 X 100 5 X 50


30x325-t-xsec

Veri precice scan!

y = 0.

85

y = 0.

01


20x250-t-xsec

GPD FFS

y = 0.

01

y = 0.

85

y = 0.

01

y = 0.

85


10x100 5x50 t-xsec

y = 0.

85y =

0.85

The phi angle

AC =d+ d

d+ +d

DVCS: the beam-charge asymmetry

At EIC:Possible if a positron beam Thanks to a good tracker coverage


AC =d d

d+ +d Re ADVCS

AI Re ADVCS +Im ADVCS Interference term:

Beam charge asymmetry:

A 2 = ADVCS 2 + ABH 2 + AI 2 DVCS and BH: identical final state → they Interfere

31

How does the nuclear environment modify parton-parton correlations? How does the nuclear environment modify parton-parton correlations? How do nucleon properties change in the nuclear medium?How do nucleon properties change in the nuclear medium?

(Bethe-Heitler)(Bethe-Heitler)

Nuclear GPDs ≠ GPDs of free nucleonNuclear GPDs ≠ GPDs of free nucleon Enhancement of effect when leaving forward limit? Enhancement of effect when leaving forward limit?

caused by transverse motion of partons in caused by transverse motion of partons in nuclei?nuclei?

important role of mesonic degrees of freedom?important role of mesonic degrees of freedom? manifest in strong increase of real part of τmanifest in strong increase of real part of τDVCS DVCS

with atomic mass number A? with atomic mass number A?

DVCS in coherent region: DVCS in coherent region: new insights into ‘generalized EMC effect’?new insights into ‘generalized EMC effect’?


DVCS on nuclear targets

MC simulation for DVCS on nuclei coming soon thanks to an updated version of MILOU code

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…)


Back up


salvatore fazio bnl university of washington – int seattle, wa – usa november 1-13, 2010...

Documents

seattle slide

e vm h e info

candidate e sample

h e ps mesons

e quad gradient

eic dvcs simulation

simulation of dvcs

hera strategy