The Past and Future of DrellYan Physics at Fermilab and Structure Functions

1
The Past and Future of Drell-Yan Physics at
Fermilab and Structure Functions

• Drell-Yan and Parton Distributions
• Fermilab E866/NuSea
• Absolute Cross Sections
• Nuclear Measurements-energy loss, nuclear
  structure
• Future Experiments Fermilab E906

2
Proton Constituents Quarks and Gluons

• Naïve Proton
• 3 quarks at some Q0, bound by gluons
• QCD evolution does the rest
• Real Proton Data guides our knowledge
• Distribution of quarks, antiquarks and gluons
• QCD evolution tells us how distribution evolves,
  but not original distributions
• Experimental data provide guide for distributions
• Theoretical assumptions guide (prejudice) our
  expectations of the Parton Distribution (PDF)
  behavior, e.g. Drell-Yan-West relationship

3
How are parton distributions determined?

• Phenomenological fit worlds data to find parton
  distributions
• MRST, Eur. Phys. J C4, 463 (1998)
• CTEQ, Phys. Rev. D55, 1280 (1997)
• GRSV, Phys. Rev. D63, 094005 (2001)
• Quite sophisticated
• NNLO DIS, NLO Drell-Yan
• include estimates of uncertainties in PDFs

• Three sources for measurement
• Deep Inelastic Scattering (DIS)
• W Production Asymmetry
• Drell-Yan

Too many experts are here at DIS2004 for 1 2,
so Ill take 3! (Im not a theorist, so the left
side was not even under consideration)
4
Drell-Yan mm- Production and PDFs

• Detector acceptance chooses range in xtarget and
  xbeam.
• xF xbeam xtarget gt 0
• high-x Valence Beam quarks
• Low-x sea quarks.

Leading Order

• Experiment measures m momenta
• Virtual photon pL and pT

5
Fermilab E866/NuSea Detector
60m x 3m x 3m

• Forward xF, high mass m-pair spectrometer
• Liquid hydrogen and deuterium targets
• Two acceptance defining magnets (SM0, SM12)
• Also used solid W, Be, Fe targets

• Beam dump (4.3m Cu)
• Hadronic absorber (13.4 I0-Cu, C, CH2)
• Momentum analyzing magnet (SM3)
• Three tracking stations
• Muon identifier wall 4th tracking

6
FNAL E866/NuSea Collaboration
Abilene Christian University Donald Isenhower,
Mike Sadler, Rusty Towell, Josh Bush, Josh
Willis, Derek Wise Argonne National
Laboratory Don Geesaman, Sheldon Kaufman, Naomi
Makins, Bryon Mueller, Paul E. Reimer Fermi
National Accelerator Laboratory Chuck Brown, Bill
Cooper Georgia State University Gus Petitt,
Xiao-chun He, Bill Lee Illinois Institute of
Technology Dan Kaplan Los Alamos National
Laboratory Melynda Brooks, Tom Carey, Gerry
Garvey, Dave Lee, Mike Leitch, Pat McGaughey,
Joel Moss, Brent Park, Jen-Chieh Peng, Andrea
Palounek, Walt Sondheim, Neil Thompson
Louisiana State University Paul Kirk, Ying-Chao
Wang, Zhi-Fu Wang New Mexico State
University Mike Beddo, Ting Chang, Gary
Kyle, Vassilios Papavassiliou, J. Seldon, Jason
Webb Oak Ridge National Laboratory Terry Awes,
Paul Stankus, Glenn Young Texas A M
University Carl Gagliardi, Bob Tribble, Eric
Hawker, Maxim Vasiliev Valparaiso University Don
Koetke, Paul Nord
7
The Data Sample

• 3 spectrometer magnet setting which focus
  different muon pair masses into the detector
  low, intermediate and high

8
E866 quark sea distributions

• Select xb gt xt to get first term (detector
  acceptance does this).

• Study ratio of deuterium to hydrogen
• (Actually use full NLO calculation to extract sea
  quark ratio)
• Approx. 360,000 events.

9
Proton Valence Structure d/u as x 1
Theory Exact SU(6) d/u
1/2 Diquark S0 dominance d/u 0
pQCD d/u 3/7

• Data
• Nuclear binding/Fermi Motion effects in
  deuteriumchoice of treatments.
• Proton data is needed.

10
Drell Yan Absolute Cross Sections
11
xtarget NLO comparison (Sea)

• xtarget distribution measures magnitude of
• Data in good agreement with PDFs for x lt 0.15.
  Deuterium starts to fall off above x 0.15
• Sea previously set by HERA small-x data and E605
  Drell-Yan. Present data is much more precise.

12
xbeam NLO comparison (Valence)
xtarget
xbeam

• xbeam distribution measures 4u d as x 1.
• Both MRST and CTEQ overestimate valence
  distributions as x 1 by 15-20.

• Possibly related to d/u ratio as x 1, but
  requires full PDF-style fit.
• Working with CTEQ to incorporate data in global
  fit.
• Radiative corrections have recently been
  calculated.
• Proton-deuterium (cross check) agrees with
  proton-proton data.

13
Radiative corrections for total cross section as
function of xbeam (x2) and xtarget (x1)

• Maximum effect is lt4 for intermediate x2
  
• (.4ltx2lt.8)
• gt Will not explain 20 effect observed by E866!

14
How is the sea of the proton formed? E866 and
E906

• pQCD Gluon splitting?

• Meson Cloud? Chiral Solitons? Instantons?

• Models describe well, but
  not pQCD becoming
  dominant?

Soon lattice moment analysis may also weigh in.
15
Monte Carlo Acceptance
Detector Variables
Physics Variables
16
E d3s/dp3 E866 and E772
17
E866 - PRL 84, 3256 (2000) NA3 -
PRL84(2000),3258
Scaling of J/? Suppression?

• Shadowing and initial-state gluon energy loss
  thought to be main reasons for increasing
  suppression at larger xF
• But this effect does not scale with x2 as
  expected for shadowing between E866 at 800 GeV
  NA3 at 200 GeV
• Although does scale with xF which would be
  expected for energy loss
• Remains a puzzle

18
E866 - Correction to Nuclear Dependence for pT
Acceptance

• Acceptance in pT is considerably narrowed at low
  xF
• Use MC acceptance ds/dpT consistent with our
  data to correct for incomplete coverage
• This also is why E772 (which had stronger
  narrowing of pT at small xF) J/Ys got a 0.92

Lesson If you dont have good kinematic
coverage, effects like this are hard to detect,
much less correct.
19
Parton Energy Loss in Nuclei for Drell-Yan
Kopeliovich Model
Johnson, Kopeliovich et al., hep-ph/0105195
Shadowing

• Shadowing when coherence length,
• is larger than nucleon separation
• From E772 E866 Drell-Yan data
• With separation of shadowing
• dE/dz via Mass dependence
• In the color-dipole model
• dE/dz -2.7 .4 .5 GeV/fm
• PRC 65, 025203 (2002)

dE/dx Shadowing
Drell-Yan data from E772 (PRL 64, 2479 (1990))
20
Future Drell-Yan at Fermilab E906

• Fixed-target Drell-Yan with 120 GeV Fermilab Main
  Injector
• sDY?1/vs Larger cross section (more statistics)
• Scheduled to start collecting data in late 2008

21
Drell-Yan and mm- Physics at Fermilab

• Proton structure
• d/u as x 1
• at intermediate x
• Nuclear Measurement
• Parton energy loss
• Nuclear Pion
• The Future E906
• Significant increase in Physics reach over
  previous Drell-Yan experiments.
• Approved in 2001/most likely run in 2008-9
• 1,992k for Magnet
• 993k for Detector improvements

22
Parton Energy Loss

• Colored parton moving in strongly interacting
  media.
• Only initial state interactions are importantno
  final state strong interactions for virtual
  photon and muons.

• Fit shadowing-corrected data to energy loss
  modelsdata consistent with no energy loss
  Vasiliev et al. PRL 83 (1999) 2304.
• BaierdE/dzlt0.046 GeV/fm2 L2A
• Galvin and MilanodE/dzlt0.14/fm
• Brodsky and HoyerdE/dzlt0.44 GeV/fm
• Treatment of parton propagation length and
  shadowing are critical. Johnson et al. PRL 86
  (2001) 4483 ¼ 2.2 GeV/fm.
• Energy loss ? 1/slarger e-loss at 120 GeV.
• Measure energy loss rather than a limit.
• Distinguish between models of energy loss and
  shadowing.
• Important for interpretation of RHIC data

E906 Drell-Yan expected uncertainty and expected
energy loss effect
23
E906Nuclear Structure

• Are there
• pions in the nucleus?
• Antiquark enhancement expected from Nuclear
  Pions.
• Not seen in Fermilab E772 DY.
• E906 will clearly challenge revised predictions.

• Comparison with DIS effects
• Antishadowing not seen in Drell-YanValence only
  effect?better statistical precision neededE906.
• Intermediate-x sea PDFs set by n-DIS on
  ironunknown nuclear effects.

24
E906 Detector
25
E906 Construction Cost and Schedule
Expt. runs
Expt.. Construction
Magnet Design and construction
Expt. Funded
906 Publications
Collider and MINOS running
2009
2007
2006
2005
2008

• Schedule driven by Fermilab long range plan
• Proton EconomicsMINOS will start running in
  Jan. 2005 and run for 3 years (with design
  performance of accelerator complex).
• CKM will, most likely, not be ready until the
  2007/8 time frame.
• E906 construction starting in 2005 will allow for
  completion on an appropriate timescale (ready for
  beam in Jan. 2008).

26
FNAL E906 Collaboration
Abilene Christian University Donald Isenhower,
Mike Sadler, Rusty Towell Argonne National
Laboratory John Arrington, Don Geesaman, Roy
Holt, Hal Jackson, Paul E. Reimer, David
Potterveld University of Colorado Ed
Kinney Fermi National Accelerator
Laboratory Chuck Brown Co-Spokespersons
University of Illinois Jen-Chieh Peng Los Alamos
National Laboratory Gerry Garvey, Mike Leitch,
Pat McGaughey, Joel Moss Rutgers University Ron
Gilman, Charles Glashausser, Xiaodong Jaing, Ron
Ransome Texas A M University Carl Gagliardi,
Bob Tribble, Maxim Vasiliev Valparaiso
University Don Koetke
27
SummaryDrell-Yan Cross Sections and Proton
Parton Distributions

• Fermilab E866 has measured the Drell-Yan Cross
  section with 800 GeV pp and pd interactions.

• Data are in good agreement with previous
  Drell-Yan data.
• Proton structure
• d and u as x1 current PDFs seem to
  overestimate valence distributions.
• d-bar and u-bar at intermediate x current PDFs
  in agreement with data

• The Future E906
• Significant increase in Physics reach over
  previous Drell-Yan experiments.
• Approved in 2001/most likely run in 2008

28
FNAL E866/NuSea Total Cross-section Measurments

• FNAL E866/NuSea has measured
• First extensive Drell-Yan double-differential
  cross sections in pp collisions
• Most precise Drell-Yan cross sections in pd (or
  pA) collisions
• Triply-differential cross sections in both pp
  and pd collisions over a broad kinematic range
• dbar/ubar ratios, nuclear dependence, and other
  wonderful things!

• These results
  demonstrate
• Recent NLO PDF fits provide a reasonable
  description of antiquark distributions for 0.02 lt
  x lt 0.25
• Recent NLO PDF fits may overestimate the valence
  quark distributions at large x
• Radiative effects cannot account for x-dependence
  seen
• We need to run Fermilab E906
• (cool experiment name suggestions are
  welcome) -)

29
Data vs. PDF comparison
K0px sdata/sNLO

• Reasonable overall agreement with recent
  Next-to-Leading Order parton distribution fits.
• 6.5 global normalization uncertainty not
  included in fit.

30
JHF Letter of Intent
Physics of High-Mass Dimuon Production at the 50
GeV Proton Synchrotron J.C. Peng et al.

• What will be measured?
• Unambiguous establishment of scaling violations
  in Drell-Yan process
• Light anti-quark asymmetry via pp and pd
• Nuclear effects via Drell -Yan
• Partonic energy loss in nuclei
• Quarkonium production
• If polarized proton source
• 1. Transversity - correlation between quark
  momentum and its perpendular spin component
• 2. Sea quark polarizations

31
Physics Issues in heavy-quark production in
nuclei - Shadowing

• Shadowing of gluons ? depletion of the small x
  gluons
• Very low momentum fraction partons have large
  size, overlap with neighbors, and fuse to thus
  enhancing higher population at higher momenta at
  the expense of lower momenta
• Or, coherent scattering resulting in destructive
  interference for coherence lengths longer than
  the typical intra-nucleon distance

32
Peng et al, PLB 344 (1995) 1-5.
DY
?
J/Y

• J/y suppression an effective signature of
  Quark-gluon plasma (QGP) formation?
• Color screening in a QGP would destroy
  pairs before they can hadronize into charmonium
• But ordinary nuclear effects also absorb or
  modify J/ys
• We need a comprehensive understanding of
  charmonium production in nuclei
• Competing effects may be identified in p-A
  collisions by their strong kinematic
  dependencies, together with complementary studies
  of Drell-Yan scattering and open-charm production

33
Energy loss of gluons in nuclei
Energy loss of incident parton shifts effective
xF and produces nuclear suppression which
increases with xF
Color-dipole model Kopeliovich, Tarasov,
Hufner Nucl.Phys. A696 (2001) 669-714 (hep-ph/0104
256)
34
PT Broadening at 800 GeV
E772 E866 p-A at 800 GeV
Upsilons
Drell-Yan
a(pT) shape is independent of xF same for NA3
at a lower energy (curves are
with A slightly
different for each)
J/Y Y
35
Nuclear modification of parton level structure
dynamics
Drell-Yan
Drell-Yan Process
Ratio(W/Be)
1.0
0.9
0.8
E866 R(W/Be)
NMC DIS
E772 R(W/D)
0.7

• Modification of parton momentum distributions of
  nucleons embedded in nuclei
• e.g. shadowing depletion of low-momentum
  partons. Process dependent?
• Nuclear effects on parton dynamics
• energy loss of partons as they propagate through
  nuclei
• and (associated?) multiple scattering effects
• Production of heavy vector mesons, e.g. J/y, y ',
  ?
• production color singlet or octet ( )
  and color neutralization timescale
• hadronization time
• Coherence length for cc fluctuations
• absorption on nucleons or co-movers
• feed-down from higher mass resonances, e.g. hc

36
Upsilon Polarization E866/NuSea, Phys. Rev.
Lett. 86, 2529 (2001)
Y2S3S
DY
Y1S
Y1S
Y2S3S
Y2S3S
DY
Y1S

• Y2S3S has maximal polarization,
• like Drell-Yan
• Y1S has very small polarization