Discussion of two complementary detectors

Dear Colleagues of the World Wide Study,

The text below has been written by Tsunehiko Omori, Ron Settles,
and Jim Brau, with important criticism as devil's advocate from Joel Butler.
What we have attempted to do, is to understand and to express the value
of having two complementary detectors operating at the start of the ILC
collisions. We invite comments on this statement from the community.

We are collecting historical examples to back up this statement. A
first attempt at this list is included in section (6). This section
still needs editing. If you have other examples to offer, or comments
on these, please send them.

During the ILC Workshop at Snowmass, there will be a 'town meeting'
on Thursday, August 18, when the argument will be discussed. If you
would like to make a five (5) minute presentation during that session,
regarding this issue, please send me an email request requesting to be
put on the formal agenda.

This is a very important issue, and timely, and we urge you to spend
some time considering it. It would particularly useful if you could
reply with feedback by August 1.

Sincerely,

Jim, David, and Hitoshi




Arguments for Two Complementary Detectors
------------------------------------------
DRAFT 1.1 - 8 July, 2005
------------------------------------------

Introduction
-------------------
The ILC experimental program should ideally include at least two high energy detectors at two separate interaction regions (IRs), designed for the full energy and luminosity reach of the collider, with operation of both scheduled to begin with start of ma
chine collisions. Scientific reasons for two detectors/IRs are summarized in the five points that follow. Counter-arguments can be made, especially if there is a limit to the finances, and these are addessed in a sixth point. Since two detectors/IRs wi
ll enhance the output of the ILC facility and should attract more international participation and funding, this option should be included from the beginning. Since we have to know what the different options mean financially, a summary of all of the detec
tor options for the ILC and a recommendation for the next step are presented in a seventh and final point.

(1) Cross-check and Scientific Redundancy
-------------------------------------------------
Each of the two experiments would provide a cross-check on the discoveries and on the physics measurements of the other, a critical element of past practice. Discoveries require confirmation; precision measurements require redundancy. Two collabora
tions would develop complementary analyses with detectors characterized by independent data sets and differing systematic errors. Two would ensure the most accurate assessment of new physics found by discoveries or by precision measurements.
Confirmation and redundancy have been a key guarantee for progress in high-energy physics, as demonstrated at past fixed-target and collider facilities. For decades the ILC will be at the cutting edge of the unknown where confirmation and redundancy
are imperative to a rapid, thorough understanding of the data and physics. Historically, cross-checks have been indispensable in all branches of science, a principle understood broadly.

(2) Complementarity, future collider options
----------------------------------------------------
Ideally, given the unknowns of the experimental environment at
future colliders, the program must be prepared with two detector
philosophies in order to provide complementary sensitivity to physics,
backgrounds and fake effects. There is no unique optimal design for an
ILC detector, because it is not known what will be discovered, and what
physics will prove to be the most important. For this reason,
the LEP/SLC detectors were designed with different
strengths and weaknesses, arising from different assumptions on
physics and technical advantages; their complementarity
broadened the coverage. It is important for the ILC detectors to provide
similar breadth in detector response.
A second IR will allow future implementation of the gamma-gamma
collider option if the science case becomes compelling, without
disrupting the continuation of high energy e+e- studies. It could also
accomodate the larger crossing angle needed if a future upgrade to
much higher energies were based on a CLIC-like concept.

(3) Competition
-----------------------
The competition between two IRs will drive the scientific productivity of both experiments, as has been demonstrated frequently. This important force in the scientific enterprise results in a more effective utilization of the program.

(4) Efficiency, Reliability, Insurance
-----------------------------------------------
The efficiency of operation will be higher, since the maintenance of one detector can be carried out while the other is accumulating data. Furthermore, unexpected problems with one detector will not stop the operations of the collider; the risk asso
ciated with the large concentration of hardware at the detector IR implies that a major failure could disable the program for a long time period without the backup of a second detector.

(5) Sociology, Scientific Opportunity
---------------------------------------------
A research facility for decades of exploration is being planned,
meaning this facility will provide the opportunities for more than a
generation of physicists. It is obvious that two detectors are better
than one, doubling the possibilities for meaningful contributions to
the experimental program, and accomodating the research interests of
twice as many physicists. With two detectors employing complementary
technical solutions, the development and training opportunities,
including those for young scientists and engineers, will be enhanced.

(6) Historical examples where having multiple experiments was important
---------------------------------
(i) TPC and Mark II magnet coils shorted -- other PEP experiments were able to take data while the coils were repaired
(ii) When Jade's wires broke Petra was able to continue running
(iii) High backgrounds in H1 at HERA are not troubling ZEUS
(iv) Recent CDF tracker problems are backed up by D0
(v) BaBar and Belle have greatly exceeded expectations; would this have happened if there had been just one
(vi) LSND has an anomalous observation; experiment repeated with MINIBOONE
(vii) Hi-Q^2 jet distributions of CDF and D0 needed both experiments
(viii) Crystal Ball saw Zeta -> repeat experiment
(ix) charm and B lifetimes have needed multiple experiments to sort out observations
(x) Ab has a history requiring multiple experiments
(xi) Leptoquark signals at ZEUS and H1 were compared and found to be inconsistent
(xii) Pentaquark signals cannot be judged by one experiment alone.
(xiii) Epsilon-prime.
(xiv) At LEP, Aleph found a 4-jet mass peak at 105 GeV/c^2. It was not confirmed by the other experiments, an important check.
(xv) Kl -> mu+ mu-
(xvi) QED violation (Pipkin)


(7) Counter points of view
---------------------------------
---Counter arguments.
(i) In the absence of sufficient funding for two detectors/IRs, a single
IR can provide a cross-check by "repeating a collider run".
(ii) Another often-cited possibility is to organize two independent analysis chains within the same detector collaboration in order to promote the competition and redundancy.
(iii) With proper organization the visibility for young physicists, and
the opportunities to make significant contributions can be enhanced within
one collaboration.
(iv) Reduced efficiency: the two-IR solution would mean that the tuning-for-two would require more effort than tuning-for-one would have, and the more complicated operation will yield less total luminosity than for one IR.

---Comments on the counter arguments.
(i) This is true (see next point).
(ii) This technique has been used in the past but is not be as effective as having two different detectors. There are many historical examples of effects not being resolved by parallel analyses in one experiment in high-energy physics: the Aleph 4-jet
105 GeV mass peak, the split-A2, leptoquarks, the zeta...
(iii) Visibility for young physicists and opportunities for significant
contributions occur naturally if there are two detectors; they may or may
not if there is only one.
(iv) Clearly this may be true, but the "insurance" addressed under item (4) above says the argument may cut the other way. This is the price one has to pay for a more attractive and robust scientific environment.

(8) Options
------------------
Two detectors will cost more than one, but not double since the detector-optimization process would be different. Many developments for the two IRs could be in common, e.g., bunch-to-bunch feedback, slow-control, DAQ architecture, magnetic-field-map
ping gear, etc.

Since cost is a major issue and the financial basis for the ILC is not yet known, reliable estimates for the different options beyond the baseline are needed. In the following list the symbols mean: IIR=Instrumented IR and NIR=Non-Instrumented IR.
The options are in order of our preference:
-2IIRs/2detectors would entail a cost increase of A%,
-1IIR/2detectors(push-pull)+1NIR of B%,
-1IIR/2detectors(push-pull) of C%,
-1IIR/1detector+1NIR of D%, relative to
-1IIR/1detector = the base-line program.

Recommendation for the next step:
We ask the Snowmass machine+detector costing committee to provide estimates of A, B, C, and D