In more detail:

ADVANCED CALORIMETRY COLLABORATION
Cerenkov Compensation MC Studies for High Resolution NLC Calorimetry


   Principal Investigators:
   David R. Winn, Fairfield
   Yasar Onel, Iowa

-------------------------------------------------------
   Prof. Erhan Gulmez + 1 grad. student
   BOGAZICI UNIVERSITY
   physics dept, Bebek, 80815
   ISTANBUL, TURKEY
   tel: 90-212-263 15 40
   fax 287 24 66
   e-mail gulmez@boun.edu.tr

   Prof. D.Winn + 2 Res. Asst. +1 Mech. Eng./Draft. (V.Podrasky) + 1 Prog.
   (C.Sanzeni)
   FAIRFIELD UNIVERSITY
   Dept. Physics P Bannow 118
   N. Benson Rd
   Fairfield, CT 06430-5195
   Tel: 203-254-4000 x2359
   FAX: 203-254-4277
   winn@fair1.fairfield.edu

   Professor Ramazan Sever + 1 faculty + 1 grad. student
   Middle East Technical University
   Physics dept, 06531, Ankara, Turkey
   tel : 90-312-210 32 92
   fax : 90-312-210 12 81

   Professor Y. Onel ( PI) , Professor E. Norbeck
   J.P.Merlo,  A.Mestvirisvili (post-doc ),
   U.Akgun, A.S. Ayan, F. Duru (grad.students)
   I.Schmidt ( Mechanical Engineer), M.Miller ( electronics engineer)
   Jon Olson ( undergrad. scholar)

   UNIVERSITY OF IOWA
   Physics Dept, Van Allen Hall
   Iowa City, Ia 52242, USA
   tel: 319-335-1853
   fax : 319-335-1753
   e-mail: yasar-onel@uiowa.edu

   Professor Aldo Penzo + 1 grad. student + 1 tech.
   UNIVERSITY OF TRIESTE
   INFN and physics dept. Trieste, Italy
   tel : 39-40 375 6249/6261
   Aldo.Penzo@ts.infn.it

   Introduction:
----------------------
   We propose to study a novel idea to employ a dual readout calorimeter,
   simultaneously measuring the Cerenkov light with ionization on
   hadron-initiated showers on an event-by-event basis to compensate
   calorimeters, and to achieve precision energy resolution.

   Briefly, the idea is that as a shower fluctuates more into charged pions
rather than neutral pions, that a Cerenkov signal generated in a
transparent absorber/active medium, which arises mainly from the e-m
component of the shower, is reduced in a correlated fashion with the
ionization signal,thereby enabling a correction of the energy given by an
ionization signal.

   Preliminary infinite media GEANT simulations have indicated that the
correction can in principle enable an energy resolution substantially
better than existing calorimeters [1], which rely instead on suppressing
the e-m signal relative to the hadronic signal.

   Technical Proposal:
------------------------
   If a hadron shower were to fluctuate entirely into neutral pions (i.e. an
   extreme charge exchange for example), ionization and Cerenkov signals both
   can achieve excellent resolutions if sufficiently well sampled (NaI and
   Pb-glass calorimeters can have excellent resolutions on electrons, for
   example). However, as the hadron shower fluctuates into charged pions
   and neutrons (etc.), both signals or measures of the energy, the
   ionization and a Cerenkov
   signals, become degraded. In general, with a single calorimeter signal, it
   is not possible to know how much the signal is degraded or reduced
   compared with the initial hadron emergy. However, in preliminary studies,
   the Cerenkov signal appears to degrade at a much larger rate as a
   function of Fpi, the fraction of charged pions, compared to ionization
   signals (both scintillation light from LScint, BaF2, and NaI, and a
   drifted ionization signal collected from LArgon
   were studied). If the Cerenkov and the ionization signals are highly
   correlated, then measuring both will determine how large the fluctuation is
   on any event, which can be then used to correct the energy.

   These preliminary homogeneous calorimeter GEANT studies done some years ago
   indicate that an achievable energy resolution may allow a stochastic term less
   than 20%/SqrtE, perhaps as low as 15%/SqrtE, with a constant term tuned less
   than 1% on a hadron calorimeter. We propose to make an exgtensive MC study of
   designs which could be more easily be used in practice.

   Historically, the E1A neutrino calorimeter, a pure liquid scintillator
   ionization hadron calorimeter, achieved a stochastic term of 11%/SqrtE(GeV),
   showing the remarkable effect of large (i.e. 1/SqrtN) signals, but with a
   constant term of 9% [8]. On the other hand, the SPACAL lead-fiber calorimeter
   achieved a hadron energy resolution of ~35%/SqrtE, with a constant term of
   about 1%, as limited by the packing fraction of 20%. A compensated Cu-SciFi
   calorimeter constructed for SSC and the scintillator tile-Cu absorber
   calorimeters for ATLAS achieve resolution terms of about 60%-50%/SqrtE, largely
   due to the low compensated packing fraction of about 2-3%. If the packing
   fractions in these practical devices were to be increased to about 25%-30%, the
   stochastic term could be reduced by ~x3, provided that the sampling fraction F
   and the sampling thickness d are such that the sampling fluctuations are less
   than the sampled energy statistics [5] [i.e. s/E = (d/F)^0.5 x (E)^-0.5, where
   d is the sampling thickness and F is the sampling fraction.] However, the
   constant term would increase to about 7%-8%. Thus it is worth considering if a
   "2nd" measurement could be used to adjust the constant term downwards, while
   allowing a large signal for a small stachastic term. Using typical SPACAL data
   for Fpi [6], [7] the pion fluctuation fraction, and estimating the contribution
   from nuclear breakup by Wigmans [7], measuring the energy of the e-m component
   to about +/- 30%/SqrtE should allow the adjustment of the constant term to ~1%.

   The very first absorption calorimeters used homogeneous media Cerenkov light,
   in order to measure electromagnetic shower energy. Modern Pb-glass and
   especially water (Super-K) calorimeters achieve excellent resolutions on
   electrons (<2%/SqrtE P 9,000 p.e./GeV). However, on hadrons, both Pb-glass
   walls [2] and swimming pool calorimeters [3] have achieved a hadronic energy
   resolution of ~35%/SqrtE, but with a constant term of ~10%.

   Recent results by the CMS Forward Calorimeter Group (in which the proposers are
   participants) have shown that sampling Cerenkov calorimeters consisting of
   quartz fibers embedded in Cu serve as an adequate forward calorimeter[4]; the
   results indicate that the signal response is approximately given by:
   (1 p.e./F)(NA/0.2)^1.5, where F is the fiber packing fraction in percent, and
   NA is the fiber numerical aperture. At F~1%, at NA=0.2 and 0.4 mm diameter
   fibers, the Cu-fiber calorimeter achieves an energy resolution of about
   100%/SqrtE(GeV) on electrons, with a constant term <0.1%. With a F~25% packing
   fraction of NA~0.6 200 micron core clear fibers (n~1.6), one would therefore
   expect an electromagnetic energy resolution of better than ~10%/SqrtE. This
   would be sufficient to measure Fpi, the fluctuations in the shower, to about
   +/- 30%/SqrtE, which would in principle allow a constant term of 1-2%. Using
   similar scaling for a packing/sampling fraction of the ionization medium
   embedded in Cu, at say, F~25% for the ionization medium and d~0.5 mm thick
   sampling, one might obtain s/E ~15%-18% (as scaled from either the ATLAS
   (calorimeter), with a constant term near 1%.


   If successful in R&D, the main uses in LC calorimeters would be to:

   (1) High Resolution E-M Calorimeter Compensation for Jet Energy Resolution

   To correct for jet energy from hadrons interacting in high resolution e-m
   calorimeters. At present, the use of an extremely non-compensated but very high
   resolution e-m calorimeter in front of a compensated hadron calorimeter results
   in relatively poor jet energy resolution, as in the CMS calorimeter system,
   where a PbWO3 front end with superb em resolution results in a jet resolution
   degraded to ~100%-120%/SqrtE, mainly from jet energy deposited in an
   uncompensated, e/h~2, ~1-2Lint em calorimeter. For example, in a lead tungstate
   or cerium fluoride calorimeter in the front of NLC experiments, 2
   photo-readouts would be provided, with optical filters which accept either the
   scintillation light or the Cerenkov light generated in the crytal. Or with a Si
   or \Largon e-m calorimeter, additional cerenkov sampling via fibers or plates
   would be provided.

   (2) Intrinsic Hadron Calorimeter Energy Resolution

   Increase hadronic and jet energy calorimeter energy resolution sufficiently so
   that Zo identification and other precision dM/M and missing transverse energy
   measurements by jets becomes more feasible P i.e. so that at least the
   intrinsic particle energy resolution is such that the calorimeter contribution
   to the jet-jet mass width is below the intrinsic Z or narrow Higgs widths P
   this may require s/E ~ 25%/SqrtE (together of course with requirements on
   increased transverse segmentation and adaptive global jet-cone algorithms which
   are not part of this study).


   (3) Background Rejection

   The Cerenkov signal in CMS prototype copper-quartz-fiber forward calorimeter
   for 375 GeV single pions has been shown to rise in <1 ns and to fully develop
   in less than 5 ns (0%->95% of the signal on the end of a cable). The superb
   timing available has been shown in MC to allow beam-gas and beam-halo muon
   rejection, and to associate signals with the beam crossing and with other
   calorimeter cells to a high enough precision to play a useful role in
   determining interesting events from the multiple events in an LHC crossing
   (very different from NLC of course). However, the rate capability (small PMT
   have been run near 1 GHz for LHC tests) and timing of a well-designed Cerenkov
   fiber or plate component may play a crucial role in the environment of the NLC
   interaction region where a calorimeter may still receive a considerable load of
   uninteresting signals & potential pile-up from beam-associated backgrounds and
   high instantaneous rates (albeit for short times, say ~10Us of ns per
   crossing). Multiple measurements of the same hadron/calorimeter shower allow
   consistency checks for event-associated upsets (for example P a splash through
   a PMT or a FET). Therefore a simultaneous Cerenkov-signal readout of an
   ionization calorimeter may be interesting on these grounds alone.

   Proposal

   We therefore propose:

   FY 03 and 04:

   Cerenkov Compensation MC Studies:

   Study Cerenkov Compensation schemes using GEANT and NLC simulation tools:

   (a) MC "Calibration": Tune existing codes and reproduce the reported
   resolutions and response of existing calorimeters: the ATLAS scintillator
   plate-WLS fiber calorimeter, the CMS Forward Cerenkov Fiber calorimeter, and of
   at least one tested/published drifted-ion sampling calorimeter (Si or LArgon),
   and of at least on homogeneous crystal calorimeter. These will include full
   propagation of individual signal photons or electrons (for example, as captured
   on the WLS fibers, and realistic photodetectors, including both APDs and PMTs).


   (b) WLS Fiber-Scintillator + Clear Fiber Geometry: MC Study an
   ATLAS-style/Gildmeister [6] Scintillating Tile/fiber Cu absorber Calorimeter
   geometry with high scintillator packing fractions, up to 40% of scintillator,
   and up to 40% of clear Cerenkov radiator fibers. (A very brief study will also
   be made using WLS fibers on clear C-radiators, but this is anticipated to
   fail.)

   (c) Plate Geometry: MC Study of a classic plate absorber geometry: Cu absorber
   plate + [scintillator, LArgon, or Silicon] plate + Cerenkov plate. The Cerenkov
   plate would be read-out using an APD array

   (d) All Fiber Geometry: MC study of Cu-absorber + Scintillating Fiber + Clear
   Fiber Calorimeter.

   (e) Homogeneous Calorimeter Geometry: MC study of the simultaneous Cerenkov
   readout of e-m crystal calorimeter (lead tungstate), using filters and 2
   photodetectors, and of collecting drifted ions and Cerenkov light in LXe. The
   authors have shown in detail that Cerenkov light and ionization light can be
   measured independently and simultaneously in LScintillator using filters
   (somewhat counterintuitively, the Cerenkov light is measured by using a
   low-pass filter P i.e. the long-wavelength Cerenkov light P despite the lower
   yield P because of the shifting properties of the fluors in the scintillator).

   REFERENCES:

   [1] D.R.Winn and W.A. Worstell, "Compensating Hadron Calorimeters with Cerenkov
   Light", IEEE Trans. Nuclear Science Vol. NS-36 , No. 1, 334 (1989).
   [2] Bitsadze et al., NIM (1988)
   [3] B.C.Brown et al., Fermilab Report 1987 (also published in IEEE
   Trans.Nuc.Sci. 1989-90)
   [4] N. Akchurin et al., NIM A399, 202 (1997)
   [5] O.Ganel et al., Proc.7th Conf.C alor.in HEP, 141 (1997) Tucson
   D.Acosta et al., NIM A316, 184 (1992)
   [6] O.Gildmeister et al., Proc.2nd Conf.Calor.in HEP (1991) Capri
   ATLAS Tile Calorimeter TDR CERN/LHCC/96-42
   [7] R.Wigmans, Proc.7th Conf.Calor.in HEP 182 (1997) Tucson
   [8] A.Benvenuti et al., NIM 125 447 (1975)

-----------------------------------------------------------
Tel: 319-335 1853    (office)
Fax: 319-335-1753

Please send your replies to my generic e-mail address

Yasar-Onel@uiowa.edu
------------------------------------------------------------