The Rise of Large Collaborations

1
The Rise of Large Collaborations
P. Grannis APS Panofsky Prize Talk May 1, 2001
Panofsky Prize citation For his distinguished
leadership and vision in the conception, design,
construction, and execution of the DØ experiment
at the Fermilab Tevatron proton-antiproton
collider. His many contributions have been
decisive in all aspects of the experiment.
Had not my colleagues succeeded in building a
superb detector, and produced a world-class set
of measurements, I would not be here today. My
thanks to all my DØ colleagues! Here I comment
on the nature and role of large collaborations in
particle physics, and other fields of science.
2
A brief history of DØ

• After 2 years consideration of small, clever
  proposals for the DØ interaction region, Leon
  Lederman asked Grannis on July 1, 1983 to form a
  new collaboration to complement the existing CDF
  experiment with Stage I approval, sight
  unseen!
• Proposal to Physics Advisory Committee, Nov.
  1983
• First Temple review Nov. 1984 proposed
  experiment was essentially that built
  (cogniscenti will notice some changes though!)

• Collaboration funding building slow in 1984
  88 due to pressures from other major
  initiatives at SLAC (SLD), LEP, FNAL (CDF) (the
  longer you have, the more the tendency to
  optimize ! )
• Roll-in detector Feb. 1992 first collisions
  May 1992 first physics data Sept. 1992
• Tevatron Run I 9/92 1/96
• 100th publication of full collaboration
  early 2001

DØ detector as proposed 1984
3
100th paper author list
List from original 1983 proposal
M. Abolins, M. Adams, L. Ahrens, R. Brock, C.
Brown, D. Buchholz, R. Butz, P. Connolly, B.
Cox, C. Crawford, D. Cutts, R. Dixon, D.
Edmunds, R. Engelmann, H. Fenker, .
Ficenic, D. Finley, P. Franzini, E. Gardella,
B. Gibbard, B. Gobbi, L. Godfrey,
H. Goldman, H. Gordon, P. Grannis,
D. Green, H. Haggerty, M. Harrison, D.
Hedin, J. Hoftun, R. Horstcotte, R.
Johnson, H. Jostlein, S. Kahn, J. Kirz, W.
Kononenko, S. Kunori, R. Lanou, J. Lee-Franzini,
S. Linn, D. Lloyd-Owen, E.
Malamud, P. Martin, M. Marx, P. Mazur, J.
McCarthy, R. McCarthy, M. Month, M.
Murtagh, D. Owen, B. Pifer, B. Pope, S.
Protopopescu, P. Rapp, L. Romero, R.D.
Schamberger, W. Selove, T. Shinkawa, D. Son, S.
Stampke, S. Terada, G.
Theodosiou, P.M. Tuts, R. Van Berg, H.
Weerts, H. Weisberg, D. Weygand, D.H. White,
R. Yamada, P. Yamin, S. Youssef
71 people, 12 institutions
386 authors, 62 institutions
24 red highlighted are in current author list
4
If there is justification for large
collaborations, it had better be the PHYSICS!

• How did the proposed physics program compare with
  reality ? (Talked of 5 pb-1 data accumulation!
  Got 127 pb-1 -- thanks to the Fermilab
  Accelerator Division)
• W/Z boson mass, width, cross-section, study W
  tb ! (explore Electroweak interaction)
• Anomalous trilinear gauge boson couplings
  (WWg) etc.
• Search for new quarks and leptons (did not
  really talk about discovery of top quark in 1983,
  since it seemed clear it would be discovered
  before DØ started! )
• QCD studies jets, W/Z, Drell Yan dileptons,
  photons, aS measurement
• Searches for beyond-the-SM phenomena
  technicolor, leptoquarks, heavier W/Z,
  supersymmetry, compositeness
• Discussed in proposal, but did not do
• Centauro events !
• Quark gluon plasma indications
• But did many things not envisioned QCD color
  coherence diffractive production of jets/W
  magnetic monopole search large extra dimensions
  search J/Y production b production b s
  transitions, electroweak production of top, t
  e universality

5
Physics highlights top quark
DØ and CDF discover top quark in March 1995. DØ
mass determination 172.1 /- 7.1 GeV using l
jets and even ll 2 jets ET where
underconstrained kinematically
4 precision with only 40 events!
Top quark is the last matter particle expected in
the SM is astoundingly heavy its mass is
scale of EW symmetry breaking. This seems
provocative !!
Cross section measured to 30, using ll, l, and
even the 6 jet final state where background is
106 X signal.
6
Physics highlights Electroweak bosons
W mass measured in en decay
80.482 /- 0.091 GeV
( 1 per mil in one channel)
WWg, WWZ, ZZg, Zgg gauge couplings measured. As
with (g-2), couplings probe new physics beyond
the SM (1995 demonstration
that SU(2)xU(1) couplings required)
Combination of top quark and W mass constrain the
Higgs mass through its contributions to loop
diagrams. Light Higgs preferred in SM (fit
with Susy somewhat better).
7
Physics highlights QCD
Cross sections for jets are the modern analog of
Rutherford scattering measures proton
structure, tests QCD and seeks new level of
constituents. Jet XS agrees w/ QCD for pT up to
½ beam energy, and rapidity 3, and gives new
constraints on high-x gluon distribution.
Highest energy qq scatter
b-quark production exceeds QCD prediction by
factor X2
Measurements of W,Z production cross sections
agree with QCD and probe non-perturbative
effects. New studies of diffractive
production of jets helps illuminate high Q2
colorless-exchange processes.
8
Physics highlights Searches for New Phenomena
No substructure of leptons/ quarks at scale 3.3
- 6 TeV from Drell Yan qq ee-
bknd
data
cosq
cosq
M(ee)
M(ee)
Effective Planck scale limit for extra
dimensions. qq/gg ee-/gg
1.4 TeV (nextra2)
1.0 TeV (nextra4-7)
cosq
M(ee)
Limits on scalar leptoquarks to 225 GeV
Signal bknd
No Supersymmetric equal mass squark gluinos
(mSUGRA) M Dirac spin 1/2 monopole limit of M 870 GeV
9
Are large collaborations effective?
0.02 papers/author/year
750K/publication Detector
Cost
20K/person/year 6.5/lb (good steak)
DØ 350 authors 9 yrs to 1st data, 13 yrs to
end of Run I, 18 yrs to 100 papers.
Experiment cost 75M. (typical for large
contemporary collider experiments) 3
contemporary fixed target experiments
(KTeV, g-2, E706 combined) (10 yrs
duration) 220 members, 32 publications My
thesis experiment 10 members, 1 publication
( 3 yrs duration)
0.015 papers/author/year
0.03 paper/author/year
Number of papers/participant/year is not so
different for large collider experiments,
contemporary fixed purpose experiments, and
experiments of 30 years ago. But
large collider experiments can do physics of a
complexity not possible with smaller experiments
due to nearly 4p coverage for tracks, energy
flow, particle ID 106 electronics channels .
e.g. tt production with 6 final state objects
(lepton, n, and 4 jets) invokes all aspects of a
full 4p detector. The large collider detectors
have proven to be capable of a wide range of
measurements not originally envisioned.
10
Physics justification of large collaborations
Large costly detectors/experiments have the reach
to study largest questions before us not only
in HEP, but also in Biology (genome research,
xFELs), Astronomy (HST,MAXIMA, CHANDRA, GLAST,
), Nuclear Physics (RHIC, CEBAF), Materials
Science (Light sources, SNS, ) etc. Size in
HEP has been essential in understanding the basis
for the SM (LEP, SLC, Tevatron expts) and
without them we would not have ns, W/Z bosons,
top quark, direct and indirect Higgs indications,
confirmation of the gauge structure

of SM. We wont discover Susy, large extra
dimensions, or strong WW scattering without large
detectors. We wont unravel the mystery of
neutrino masses, oscillations, proton decay or
GUTs and wont illuminate CP violation in the B
sector without them. The pressing questions of
our field are no longer amenable to small limited
scope experiments.
11
Education of young scientists
DØ Graduate students 136 completed Ph.Ds
123 presently in progress. Of those completed
(about 60 of students took HEP postdocs
initially) NOW 34 HEP postdocs 15
Academic 10 National Labs 32 Technical
industry 8 Financial industry 1
Government DØ Postdocs 162 completed postdocs
67 presently in progress.Of
those completed (some duplication of postdocs
and students) NOW 31 Academic 31 Technical
Industry 27 National Labs 6 Financial,
legal 3 Government 2 Other
20 of the 44 graduate students at the time of the
top discovery paper
12
The price of being large

• Organization of large experiments becomes
  complex. Line authority in project
  organization governance documents rules
  for authorship Speakers bureau Decision
  processes are complex and involve many
  constituencies. Extensive review by Labs,
  funding agencies.
• Individual success in large collaborations
  requires communications skills (often of use in
  subsequent jobs!)
• Size of the enterprise means that most
  individuals specialize in particular aspects --
  software, electronics, detector building (but
  not so different in smaller collab.)
• Few see the experiment from inception of design
  to final physics.

Nevertheless, many individuals take a series of
diverse tasks over the history of the
experiment. Actual research projects (build a
trigger, do a physics analysis, design m ID code
) are typically 3 4 persons, as in smaller
science research.
13
International collaboration
Large high energy physics collaborations are a
driver for developing more effective
relationships with scientists across the
globe. Improvement of understanding of the
frameworks in various nations. In a
collaboration, one needs to understand the
special forces at play in each participating
nation. Helps developing nations to build ST
infrastructure at home, and promotes technically
competent people for high level positions.

developing (in science) nations in DØ
Brazil, Colombia, Ecuador, Korea, Mexico,
(Vietnam?) Low funding, but very talented
people.
14
Large collaboration dynamics

• HEP collaborations are strange organizations !
• Multi-instutional, multinational
  (DØ has now 68 institutions/ 17 nations)
• No line authority spokespersons/ group heads
  have no real authority (hires/salary/ reward
  mechanisms)
• Research program, detector choices, etc. are
  based upon collective decision to collaborate
  towards a common goal.

• Its a miracle that it works at all !
• Indeed some individualists are repelled there
  is some degree of peer pressure to conform to
  group decisions.
• Competition with other experiments is intense.
  (If the competitor has a 1 better measurement,
  spurred to do better, and the innovations born
  are often significant.)
• For many, the collaboration is the primary
  loyalty above that to ones university, Lab.

15
Large collaboration publications
The big author lists what to do about
them? Pressure from APS, European Physical
Society, University promotion and tenure
committees to reduce the size. Only put on
the list those who did the work DØ and other
collaborations have resisted this
Take our recent paper on inclusive production of
jets at high transverse momentum and all
angles One student postdoc did the analysis
group of about 5 very closely tied to the
particular analysis through closely related
studies. The work builds on that of perhaps 30
physicists who developed the jet energy scale,
resolutions, errors, underlying event corrections
etc. The analysis benefited from 20 or so who
made strong critical contributions to the
analysis, interpretation. A group of 30 more
contributed to the algorithms for jets,
underlying energy from Main Ring beam, vertex
determination, event selection,etc. Then there
are the critical contributions (another 30?) to
the building of the calorimeters, the triggers,
the online analysis, beam monitoring, data bases,
luminosity determination, offline CPU farms, etc.
And what about the 50 or so more who did
necessary experiment management, HV power,
controls, accelerator interfaces, databases,
computer farms, Monte Carlo generation, etc.
? It is essentially impossible to delineate
contribution it is a continuum. Trying to do
so would destroy the fabric of the collaborative
spirit. DØ and others list the full set of
collaborators as authors.
16
How to reward physicists in large collaborations
The large experiments comprise as many
individuals as are in many entire academic fields
worldwide for example DØ comparable to the
world set of 19th century Russian historians (and
a lot bigger than many fields). Typically, in
cases of hires or promotions, one asks experts in
the field who have not directly collaborated with
an individual. For those working in large
experiments, this is thought to be a problem
cant ask collaborators. But those from other
experiments typically have no way to judge the
individual they cannot peer into the internal
workings of a collaboration. One must therefore
rely, even more strongly than usual, on the
written evaluations of members of a collaboration
who have not themselves worked closely with a
candidate. They will have seen the person in
action internal talks, internal notes, oral
evaluations and know who has really done a good
job. There is no evidence that such evaluations
are any more biassed than the outside
commentator. There is just as much pressure in
this case to offer the sort of sound advice that
one wants and needs at ones own institution.
In my experience, letters from collaborators
are often more critical and insightful than
outsiders.
17
Defects of large collaborations

• Large collaborations, even more than smaller
  HEP experiments, tend to foster
  compartmentalization individuals gravitate to a
  speciality.
• It is hard for individuals students in
  particular to see an experiment from inception
  to publication (though this is becoming the norm
  with smaller HEP experiments as well !)
• The large general experiments tend to freeze
  out smaller dedicated experiments. Since they
  can do many things at least moderately well and
  have high costs, they tend to saturate the
  resources available. But there are topics that
  large experiments cannot attack speculative new
  physics can elude the big detectors.
• Once started, the large collaborations are hard
  to stop (DØ started with proposals in 1981 will
  continue until at least 2006 to try to discover
  Higgs ). Keeping them vital and responsive to
  new needs is challenging.

18
Strengths of large collaborations

• We are in an era where we know many of the most
  crucial questions, and to answer these, we need
  large scale experiments.
• The large experiments have given us many high
  profile results in the past 20 years, and these
  have percolated to the attention of the general
  public.
• The large experiments, properly accounted, are
  less expensive than a set of smaller experiments.
• The flexibility of large detectors is
  impressive many topics studied that are not
  foreseen.
• Large collaborative efforts have helped drive
  technological advances for all of science large
  scale electronics, new detection techniques,
  large data set organization, the WEB,
  multivariate analysis techniques ...

19
Conclusions
The nature of the questions we have to ask
assures us that large experimental collaborations
are here to stay. HEP experiments have led the
way, but are not unique. The large experiments
have been remarkably successful in advancing
physics. Seen from the inside, work is not so
different from small experiments. Giving
adequate recognition to young physicists is a
problem. Continued effort is needed to keep the
environment in large experiments healthy and
conducive to innovative ideas.