QCD at the Tevatron Iain Bertram Lancaster University DØ Experiment 6 June 2001 - Bristol

QCD at the Tevatron

Iain Bertram

Lancaster University

DØ Experiment

6 June 2001 - Bristol

6 June 2001 Iain A Bertram 2

Outline

Introduction to Tevatron KinematicsDØ – CDF Detectors

Jet Cross SectionsJet Structure

PhotonsRun II @ DØ Conclusions

Caveat: This is not the entire Tevatron QCD program – that is impossible in one hour. I have chosen the topics I consider

to be most important.


Hadron-hadron collisions

fragmentation

partondistribution

partondistribution

Jet

Underlyingevent

Photon, W, Z etc.

ISRFSR

Hard scattering


Measured Event Variables In a event the following are measured:

massless coscosh2

2/tanln ,, :electron or Jet

21212,1,2

TT

T

EEM

E

Jet, electron 2: ET

2, 2, 2

Jet, electron 1: ET

1, 1, 1

= 0

2tanln

:dityPsuedorapi


The Detectors (Run I 1993-95)

s 1.8 TeV W/Z’s ~40k/3k for mass

Luminosity 21031 ttbar ~10-100 events

Ldt ~100 pb-1 Higgs negligible


DØ Calorimeter: Run I

Full Coverage: || < 4.1Segmentation: =0.10.1 (=0.050.05 EM shower max)

Single Electron Response: 0.15/E + 0.003Single Hadron Response: 0.5/E + 0.004Depth:


Jets at the Tevatron

Mostly Fixed cone-size jets Add up towers

Iterative process Jet quantities:

ET, ,

Correct to Particles Do not include underlying

event. Model underlying event with

Minimum bias event (inelastic scattering)

q

Tim

e

p p

q g

K

“par

ton

jet”

“par

ticle

jet”

“cal

orim

eter

jet”

hadrons

CH

FH

EM

0.7R

towerT

jetT

i

EE

0.7R

towerT

jetT

i

EE


“Typical DØ Dijet Event”

ET,1 = 475 GeV, 1 = -0.69, x1=0.66ET,2 = 472 GeV, 2 = 0.69, x2=0.66

MJJ = 1.18 TeVQ2 = ET,1×ET,2=2.2x105 GeV2


Tevatron x-Q2 Reach

Overlaps and extends reach of HERA


Inclusive Jet Cross Section

TT

jet

TT

TE

LdtE

N

ddE

dEdd

E vs.

1 2

T

T

jet

TT

TE

LdtE

N

ddE

dEdd

E vs.

1 2

bin in the jets of #

Luminosity inst.sizebin

efficiencyselection sizebin

jet

TT

N

L

EE

bin in the jets of #

Luminosity inst.sizebin

efficiencyselection sizebin

jet

TT

N

L

EE

X jet pp X jet pppQCD, PDFs, substructure,..?

d2 /d

ET d

ET

How well do we know proton structure (PDFs)

? Is NLO (s

3) QCD “sufficient” ? Are quarks composite ?


Inclusive Jet Cross Section

CDF: 0.1 < || < 0.7

Good Agreement with NLO QCD

DØ: PRL:82 2451 (1999) , hep-ex/0012046 (acc PRD)

CDF: hep-ph/0102074 (acc PRD)


Theory Uncertainties

NLO pQCD predictions (s

3): - Ellis, et al., Phys. Rev. D, 64, (1990) EKS- Aversa, et al., Phys. Rev. Lett., 65, (1990)- Giele, et al., Phys. Rev. Lett., 73, (1994) JETRAD

Choices (hep-ph/9801285, EPJ C5, 687, 1998): Renormalization Scale

(~10%) PDFs (~20% with ET

dependence) Clustering Alg. (~5% with

ET dependence)


Comparisons with CDF

DØ analyzed 0.1 < || < 0.7 to compare with CDF

Good Agreement

If we calculate 2 between DØ data and fit to CDF and its uncertainties:

2 = 30.8 (0.16)


No indication of an excess above 350 GeV. Good agreement quantitatively

as indicated by 2 test:

Di and Ti data and theory, Cij covariance matrix.

2 = (Di-Ti) C-1ij (Dj-Tj)

DØ Comparisons to NLO Theory

|| < 0.5 0.1 <|| < 0.7

CTEQ 3M 25.3 (0.39) 32.7 (0.11)

CTEQ4M 20.1 (0.69) 26.8 (0.31)

CTEQ4HJ 16.8 (0.86) 22.4 (0.56)

MRS(A’) 20.4 (0.67) 28.5 (0.24)

MRST 25.3 (0.39) 29.6 (0.20)


No indication of an excess above 350 GeV. Good agreement quantitatively as indicated by 2 test:

2 = 2stat + S2

CDF Comparisons to NLO Theory

33 bins 0.1 <|| < 0.7

CTEQ4M 63.4

CTEQ4HJ 46.8

MRST 49.5

S2 is a contribution depending on the best fit to the uncertainties


DØ Forward Jets

QCDJETRADQCDJETRAD QCDJETRADQCDJETRAD

ET

d2

dE

T d

(fb

/GeV

)

Phys.Rev.Lett. 86 1707 (2001)


Comparison to Theory

Closed: CTEQ4HJ Open: CTEQ4M Closed: MRTSg Open: MRST


Quantitative Comparison

Discriminates between PDF Now being used by CTEQ and MRST to restrict

gluons to 20% level (see DIS 2001 http://dis2001.bo.infn.it/wg/sfwg.html)

PDF dof ProbCTEQ3M 121.56 1.35 0.01CTEQ4M 92.46 1.03 0.41CTEQ4HJ 59.38 0.66 0.99MRST 113.78 1.26 0.05MRSTgD 155.52 1.73 <0.01MRSTgU 85.09 0.95 0.63


What have we learned?

These results extend significantly the kinematic reach of previous studies and are consistent with pQCD calculations over the large dynamic range accessible (| | < 3).

Once incorporated into revised modern PDFs, these measurements will greatly improve our understanding of the structure of the proton at large x and Q2.

Are gluon distributions at large x enhanced? factor 20 more data in Run II, starting summer 2001,

will extend the reach to higher ET and should make the asymptotic behaviour clearer


KT Jet Algorithm

R=0.7

0

Jet Seeds

Calorimeter ET

Snowmassi

TiT EE

2

22

,2

,, ),min(D

REEd

ijjTiTji

2,JetTcutEy

resolution parameter Jet ET

KT Jet Algorithm:

Cone jet

KT jet (D=1)

jiij

jiij

EEE

ppp

Fixed Cone Algorithm:


Jet Cross Section using KT

KT with D=1.0, equals NLO cross section with Cone R=0.7

Energy difference between KT and cone causes difference in cross section

1-2 GeV Difference caused by Hadronic Showering effects

(parton to particle) Underlying Event Showering

Difference with theory most at low ET.

2=27 (31%)

2=31/24 (31%)

2=27 (31%)

24 d.o.f.


Ratio of Cross Sections

Express Inclusive Jet Cross Section as dimensionless quantity

as a function of

Theory uncertainties could be reduced to 10%

Experimental Uncertainties Cancel

ddEdE

T

T23

2

sE

x TT

2

Naive Parton Model = 1


Ratio of Cross Sections

Phys.Rev.Lett.86,2523 (2001); hep-ex/0012046

Data 10-15% below NLO QCD

No obvious problem:

Interesting!Agreement Probability

(from test) with CTEQ4M, CTEQ4HJ, MRST, MRSTGU:

25-80%


Comparison with CDF

Consistent at high xT, possible discrepancy at low values


Suggested explanations

Different renormalizationscales at the two energies OK, so it’s allowed, but . . .

Mangano proposes an O(3 GeV)non-perturbative shift in jet energy losses out of cone underlying event intrinsic kT

could be under or overcorrecting the data (or even different between theexperiments — DØ?)


Dijet Mass Spectrum

Experiments in excellent agreement.

Different rapidity ranges and very different analysis techniques.

Reasonable agreement with predictions.

Nevents

L M


Dijet Mass Cross Section Ratio

2.4 TeV (95% confidence level)

SystematicUncertainty ~ 8%

Theory uncertainty ~

6% (m) , 3% (PDF)

NLO QCD in good agreement with data

2

sC

sL

< 0.5 ) / 0.5 < 1 ) (s=1800 GeV ) PRL 82, 2457-2462, 1999


BFKL

In hadron-hadron collisions:

Q2

1/x

DGLAP (Dokshitser-Gribov-Lipatov-Altarelli-Parisi)

BFKL (Balitsky-Fadin-Kuraev-Lipatov) evolution

PDF

BFKLmany gluonsall ~ same pT

DGLAP(as in PYTHIA)gluon radiationstrongly ordered

One way to realise this situation is jets widely separated in rapidity: the total energy is then much greater than the jet pT scale, and one can have many gluons of comparable pT emitted between the jets Note: BFKL provides a way to resum the contribution of these gluons; it doesn’t predict how many there are, and there is no BFKL event generator yet


Phys.Rev.Lett. 84, 5722 (2000)

Another attempt to find anobservable which displays BFKLbehaviour:

DØ measurement of 630/1800 GeV cross section ratio at large rapidity separations

use bins such that x and Q2 arethe same in the two datasets(but different )

What’s going on here? data behave qualitatively like BFKL (but also like HERWIG) given that we can’t predict the 630/1800 GeV ratio of inclusive cross sections, how

much can we really infer?

BFKL


Subjet Multiplicity in Quark & Gluon Jets

Jet ETJet ET

Motivation:

• Test of QCD ( Q & G jets are different)

• Separate Q jets from G jets (top, Higgs, W+Jets events)

Measure the subjet multiplicity in quark and gluon jets

Contributions of different initial states to the cross section for fixed Jet ET vary with

1800GeV

100 200 300

gg

qg

qq

Method:

• Select quark enriched & gluon enriched jet sample

• Compare jets at same ( ET , h ) produced at different and assume relative q/g content is known

630GeV

100 200 300

s

s

s


Jet Structure at the Tevatron

Subjets inside jets: perturbative part of fragmentation DØ compares 630 to 1800 GeV data at same ET and , and infers q and g

jet differences

Quark Jets

Gluon Jets

Subjet Multiplicity

jetsjets dN

dM

N

1

04.091.11

1

q

g

M

MR

04.086.1 R

kT algorithmD=0.5, ycut= 10-3

55 < ET(jet) < 100 GeV|jet| < 0.5

DØ Data

HERWIG 5.9

DØ Preliminary

1 2 3 4 5

0.1

0.2

0.3

0.4

0.5


Direct Photon Production

DØ PRL 84 2786 (2000) CDF Preliminary

QCD prediction is NLO Owens et al.Note: ET range probed with photons is lower than with jets


Direct Photon Production

DØ PRL 84 2786 (2000) CDF Preliminary

Low ET excess: Theory QCD NLO Owens et al.


kT Smearing???

Gaussian smearing of the transverse momenta by a few GeV can model the rise of cross section at low ET (hep-ph/9808467)

Motivated by observed pT()


Alternative Viewpoint

Also note that Vogelsang et al. get closer to the observed shape than the Owens prediction with no extra kT (using NLO fragmentation and setting RF)


Photons at s = 630 GeV

CDF data, compared with UA2 data show deficit at high ET compared with the theory. Relation to jets?


DØ Photons at 630 GeV

First measurement at forward rapidities

2 Comparison (7 bins)

CC: 11.4 -- 12%

EC: 4.6 -- 71%


Ratio of Photons at 630 and 1800

Slight excess at low XT Insignificant

statistically Run II won’t help!

Good overall agreement with NLO

No significant high ET deficit. No statistics at

high ET (> 40 GeV)

2 Comparison (7 bins)

CC: 6.5 -- 49%

EC: 3.0 -- 89%


Photons: Comments

Situation is complicated and not easily explained. Low ET kT effects?

High ET deficits at 630 GeV

Lack of statistics in both regions

Challenge for both theory and experiment in Run II.


Run II : Coming March 2001s 2.0 TeV W/Z’s >1M/>50k

Luminosity 21032 ttbar >1k events

Ldt 2-30fb-1 Higgs Possible

Upgrades to both detectors: Silicon, Tracking (Drift-CDF, Fibres-DØ) Preshowers, Calorimeter Upgrade – CDF, Muon upgrades,

new trigger electronics (bunch spacing), C++ OO code development


Case Study: Dijet Mass Spectrum

Dramatic increase in high pT cross sectionsLarge gains in statistics


Run II:~100 events ET > 490 GeV~1K events ET > 400 GeV

Run I: 16 Events ET> 410 GeV

Great reach at high x and Q2, the place to look for

new physics!

Looking ahead: Run II


Jet Algorithms Fermilab Run II workshop

Various species of kT

New Cone Algorithm Theoretical desires

“infrared safety is not a joke!” avoid ad hoc parameters like Rsep

Cone algorithm improved by e.g. by modification of seed choices - midpoints seedless algorithm? In development.

Experimental desires sensitivity to noise, pileup, negative energies


Requirements for Progress Quantitative Theoretical Uncertainties

pdf and renormalization scale uncertainties

Full understanding of correlations of experimental uncertainties Statistical Uncertainties in most cases will be negligible

Underlying event Require a better understanding and treatment of the

surrounding event


Status of DØ detector

Run commenced March 2001

Two data taking periods completed of a few weeks.

Detector Commissioning taking place!

Physics data in autumn

First results in early 2002


Firstpp collisions of Run 2 at DØ: April 3, 2001

Luminosity counters timing

Vertex distribution along z of min bias events:

[cm]

Luminosity( coincidence)

Proton halo

~5 1027 cm-2 sec-1

Antiproton halo


Global track finding

36 36 Store Run 119679, Event 232931Level 3 (software trigger) Global Tracking

Track with 5 fiber tracker hits, 5 3D silicon hits

Relative alignmentof silicon and fiber trackersverified to 40 m level


Event with 6 tracks pointing to same vertex

36 36 Store Run 119679, Event 231653Level 3 (software trigger) Silicon-only Tracking


Vertex finding

Primary vertices reconstructed from tracks in silicon:

x

y

z

y

z

x

A. SchwartzmanBuenos Aires


Tracking in Silicon

3D event display showing hits and reconstructed track:

Offline track findingE. BarberisY. Kulik


Silicon Detector

North

South

H1

H4

F1

F11 F12

B1 B4 B6


Tracks in the Silicon + Fiber Tracker

Offline track finding frompp events:

2

1/pT

Since B=0 for this run, real tracks should be found with 1/pT = 0

8 or more points on a track (5 fiber hits, 3 or more Silicon)

D. AdamsFermilab

Data taken in 36 x 36 store (end April)Plot posted 5/18/01


Hits in Central Preshower Detector

Preshower hits

Fiber tracker hits

D. Coppage, Kansas


L1 muon triggered event

xy view, looking North zy view, looking East

xz view, looking down


Calorimeter and forward muon hits in a minimum bias event

Firstpp collisions of Run 2 at DØ


The work of many

people...


Closing Remarks Run II has started: 1 March 2001

BIG Opportunity for QCD

In most cases QCD predictions work exceptionally well.

Exceptions: Low pT processes are problematic

low pT jets and photons

Underlying Event and Event characteristics

Documents

QCD at the Tevatron Iain Bertram Lancaster University DØ Experiment 6 June 2001 - Bristol