DDDiffraction at Tevatron

p p p p

p p p (p) + X

p p p (p) + j j

p p p p + j j

DD

DDColor Singlet Exchange

DD

DDHard Color Singlet Studies

jet

jet

QCD color-singlet signal observed in ~ 1 % opposite-side

events (p )

Publications

DØ: PRL 72, 2332(1994)

CDF: PRL 74, 885 (1995) DØ: PRL 76, 734 (1996) Zeus: Phys Lett B369, 55 (1996) (7%) CDF: PRL 80, 1156 (1998) DØ: PLB 440, 189 (1998) CDF: PRL 81, 5278 (1998) H1: hep-ex/0203011 (sub. to Eur Phys J C)

Newest Results

Color-Singlet fractions at s = 630 & 1800 GeV

Color-Singlet Dependence on: , ET, s (parton-x)

p

DDDØ Detector

EM Calorimeter

L0 Detector

beam

(nl0 = # tiles in L0 detector with signal2.3 < || < 4.3)

Central Gaps

EM Calorimeter ET > 200 MeV || < 1.0

Forward Gaps

EM Calorimeter E > 150 MeV 2.0 < || < 4.1

Had. Calorimeter E > 500 MeV 3.2 < || < 5.2)

Central Drift Chamber

(ntrk = # charged tracks with || < 1.0)

End Calorimeter

Central Calorimeter

(ncal = # cal towers with energy above threshold)

Hadronic Calorimeter

DDMeasurement of fs

Count tracks and EM Calorimeter Towers in || < 1.0

ET >30 GeVfS 1800 = 0.94 0.04stat 0.12sys %

jet

jet

Negative binomial fitto “QCD” multiplicity

fS = color-singlet fraction =(Ndata- Nfit)/Ntotal

(Includes correction for multiple interaction contamination.Sys error dominated by background fitting.)

High-ET sample (ET > 30 GeV, s = 1800 GeV)

DD630 vs. 1800 Multiplicities

Jet ET > 12 GeV, Jet || > 1.9, > 4.0

R1800 = 3.4 1.2

Opposite-Side Data

Same-Side Data

1800 GeV:

630 Gev:

ncal ncal

ncalncal

ntrk

ntrk

ntrk

ntrk

fS 1800 (ET =19.2 GeV) = 0.54 0.06stat 0.16sys %

fS 630 (ET = 16.4 GeV) = 1.85 0.09stat 0.37sys %

630

DDColor Singlet Models

If color-singlet couples preferentially to quarks or gluons, fraction depends on initial quark/gluon densities (parton x)

larger x more quarks

Gluon preference: perturbative two-gluon models have 9/4 color factor for gluons

Naive Two-Gluon model (Bj) BFKL model: LLA BFKL dynamicsPredictions:

fS (ET) falls, fS () falls (2 gluon) / rises (BFKL)

Quark preference: Soft Color model: non-perturbative “rearrangement” prefers

quark initiated processes (easier to neutralize color) Photon and U(1): couple only to quarksPredictions:

fS (ET) & fS () rise

DDModel Fits to Data

Using Herwig 5.9

Soft Color model describes data

s = 1800 GeV

DDSurvival Probability

• Assumed to be independent of parton x (ET , )

• Originally weak s dependence Gotsman, Levin, Maor Phys. Lett B 309 (1993)

• Subsequently recalculated

GLM hep-ph/9804404

• Using soft-color model

(uncertainty from MC stats and model difference)

•

with

2.02.2)1800(

)630(

S

S

1.05.16301800 R

)1800(

)630()Model()Data( 630

18006301800 S

SRR

2.14.3)Data(6301800 R

8.02.2)1800(

)630(

S

S

DDModifications to Theory

BFKL: Cox, Forshaw (manhep99-7) use a non-running s to flatten the falling ET prediction of BFKL (due to higher order corrections at non-zero t)

Soft Color: Gregores subsequently performed a more careful counting of states that produce color singlets to improve prediction.

DDHard Single Diffraction

-4.0 -1.6 3.0 5.2

Measure min multiplicity here

-5.2 -3.0 -1. 1. 3.0 5.2

OR

Gap fractions (central and forward) at s = 630 & 1800 GeV

Single diffractive distribution

Phys Lett B 531 52 (2002)

Measure multiplicity here

DD630 vs. 1800 Multiplicities

s = 1800 GeV:

s = 630 GeV:

DDEvent Characteristics1800 Forward Jets

Solid lines show show HSD candidate eventsDashed lines show non-diffractive events

• Less jets in diffractive events• Jets are narrower and more back-to-back• Diffractive events have less overall radiation• Gap fraction has little dependence on average jet ET

DDComparison to MC

fvisible = gap · fpredicted

gap

*Add diffractive multiplicity from MC to background data distribution

*Fit to find percent of signal events extracted

Find predicted rate POMPYT x 2 / PYTHIA

*Apply same jet cuts as data, jet ET>12GeV

*Full detector simulation

* Model pomeron exchange in POMPYT26 (Bruni & Ingelman)

* based on PYTHIA * define pomeron as beam particle * Use different structure functions

DD

DDPomeron Structure Fits

Demand data fractions composed of linear combination of hard and soft gluons, let overall s normalization and fraction of hard and soft at each energy be free parameters

If we have a s independent normalization then the data prefer :– 1800: hard gluon 0.18±0.05(stat)+0.04-0.03(syst)– 630: hard gluon 0.39±0.04(stat)+0.02-0.01(syst)– Normalization: 0.43±0.03(stat)+0.08-0.06(syst)

with a confidence level of 56%.

If the hard to soft ratio is constrained the data prefer:– Hard gluon: 0.30±0.04(stat)+0.01-0.01(syst)– Norm. 1800: 0.38±0.03(stat)+0.03-0.02(syst)– Norm. 630: 0.50±0.04(stat)+0.02-0.02(syst)

with a confidence level of 1.9%.

To significantly constrain quark fraction requires additional experimental measurements.

CDF determined 56% hard gluon, 44% quark but this does not describe our data without significant soft gluon or 100% quark at 630

DD

Where is the momentum fraction lost by the proton

*Can use calorimeter only to measure

*Weights particles in well-measured region

*Can define for all events

i

yT

s

eE i

i

* calculation works well

* not dependent on structure function or center-of-mass energy

*Collins

(hep-ph/9705393)

true = calc · 2.2± 0.3

Calculation

DD

distribution for forward and central jets using (0,0) bin

p p=

0.2 for s = 630 GeV

i

yT

s

eE i

i

s = 1800 GeV forward

central

s = 630 GeV forward

central

Single Diffractive Distributions

DD

DDCDF Diffractive W

CDF used asymmetry to extractdiffractive component of the W signal

1)TOPOLOGY-lepton favors the hemisphere

opposite the rapidity gap -compare multiplicity for region onsame side of lepton vs opposite side

2)CHARGE-proton(uud) pomeron(qq) gives

twice as many W+ as W-

-W+ production is associatedwith gaps in p direction (and W- with p)

Drawback: Asymmetry approach reduces statistical power of data

DDCDF Diffractive W

CDF {PRL 78 2698 (1997)} measured RW = 1.15 ± 0.55%where RW = Ratio of diffractive/non-diffractive W a significance of 3.8

DDDØ Central W Multiplicity

Peak at (0,0) indicates diffractive W-bosonSignal: 68 of 8724 events in (0,0) bin

ncal

L0nL0

ncal

-1.1 0 1.1 3.0 5.2

Minimum side

ncal

nL0

DØ Preliminary

DDW Event Characteristics

Standard W Events Diffractive W Candidates

ET=37.12

ET=35.16

ET=36.08

MT=70.64 MT=70.71

Electron ET

Neutrino ET

Transverse Mass

DØ Preliminary

ET=35.27

DDW/ZW/Z Data ResultsData Results

Sample Diffractive Probability Background All Fluctuates to Data Central W (1.08 + 0.19 - 0.17)% 1 x 10-14 7.7Forward W (0.64 + 0.18 - 0.16)% 6 x 10-8 5.3All W (0.89 + 0.19 – 0.17)% 3 x 10-14 7.7Z (1.44 + 0.61 - 0.53)% 5 x 10-6 4.4

*Observed clear Diffractive W and Diffractive Z signals*Measured Diffractive W/All W and Diffractive Z/All Z

DØ Preliminary

DDRate Comparison

Correct MC for gap efficiency 20-30% for quark and hard gluon (soft gluon fractions <0.02%)

FINAL GAP FRACTION

Sample Data Quark Hard Gluon Cen W (1.08 + 0.21 - 0.19)% (4.1 0.8)% (0.15 0.02)%For W (0.64 + 0.19 - 0.16)% (7.2 1.3)% (0.25 0.04)% Z (1.44 + 0.62 - 0.54)% (3.8 0.7)% (0.16 0.02)%

NOTE: Observe well-known normalization problem for all structure functions, also different dependence on for data and MC, as in dijet case

DØ Preliminary

DD

Demand gap on one side, measure multiplicity on opposite sideDemand gap on one side, measure multiplicity on opposite side

Gap Region 2.5<||<5.2

Double Gaps at 1800 GeV|Jet | < 1.0, ET>15 GeV

DØ Preliminary

DD

Demand gap on one side, measure multiplicity on opposite sideDemand gap on one side, measure multiplicity on opposite side

Gap Region 2.5<||<5.2

Double Gaps at 630 GeV|Jet | < 1.0, ET>12 GeV

DØ Preliminary

DDTevatron Summary

Measurements at two CM energies with same detector adds critical new information for examining pomeron puzzle– The higher rates at 630 were not widely predicted

before the measurements

Pioneering work in Central Rapidity Gaps– Modifications to theory based on data have occured

Hard Single Diffraction – Many processes observed and studied including

Diffractive W (diffraction is not low pT process)

– Discrepancy in shape and normalization of

distribution confirms factorization breakdown

– Results imply a non-pomeron based model should be considered

Hard Double Pomeron– Observation of interesting new process

– Confirms factorization breakdown

DDConclusion

Diffraction is a complex subject! Much new experimental and theoretical

work with promise of new more precise data to aid understanding

Many open questions:

Are the hard and soft Pomeron distinct objects?

How to combine perturbative and non-perturbative concepts?

How to extract unique parton distribution densities from HERA and Tevatron data?

How to understand gap survival probability?

Is there a particle-like pomeron?