1975 SUMMER STUDY
*^i ft[h, 1 Hff'
LBL/SLAC/SLUQ LEGAL NOTICE This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor the United States Energy Research and Development Administration, nor any of their employees, nor any of their contractors, subcontractor!, or their employees, makes any warranty, expiess or implied, or assumes any legal liability or responsibility for the accuracy, completeness or usefulness of any information, apparatus, product or process disclosed, ir represents that itr use would not infringe pnvitely owned rights.
Prepared for the U. S. Energy Research and Development Admi nitration under Contracts W-7405-F.N(J-4H and K(04-3)-515
Primed in the United States of America Available from National Technical Information Service U. S. Department nf Commerce 5285 Port Royal Road Springfield. Virginia 22I5I Price: Printed Copy $9.00; Microfiche S2.25 LBL-4800 UC-3U. TID MOO. Hi] SLAC-190 PEP-178
PROCEEDINGS of the 1975 PEP Summer Study
July 28-Augusl 20. 1975 nT^^T-T. t«TM • -v„^J H <>. 1- tu 1«ri Sum • ««• nvbi«> •»
Sponsored by sr-ss^„, «»j«-^«>i- „„„ taj.*.... t«t« , Lawrence Berkeley Laboratory Slanlord Linear Accelator Center SLAC/LBL Users Organizaiion
December J 97 5
Lawrence Berkeley Laboratory University of California
Berkeley, California -ii-
Prcfeco
The reports in these Proceedings arc organised in nine sections corresponding to the areas studied by the various working groups. The coordinators and deputy co ordinators for each group gathered and edited the individual contributions in their areas of study. The working groups with their coordinators and deputy coordinators are listed in the Introduction.
Following previous Susmcr Study Proceedings, each report has received a PUP Note manner. A co^lcto list of nil* Notes is included in the (otic*inn introduc tory pages. Orders for individual FTP Notes or the complete Proceedings isiy he placed through
Judy Zclvcr ftiilding -17, toon IC Liwrence Berkeley laboratory Berkeley, a ow>
1 want to thank tlie Sinner Study participants for t?*cir pra=pt reports, tlir coordinators and deputy coordinators for their effective editing, h.ilicr I.-iw).-iki for rhe lively cover, and Boh Barton for his patient fcslp in bringing together tlir sany pieces of these Proceedings.
Pier <\Wu«w Coordinator 19TS PIT Siawr Study -iii-
197S PEP SIMCR STUDY CONTORTS List of Participant* v List of PEP Notes vi Introduction Introduction 1 K. Strauch An fcjqifriDvnter's View ol i'UV (l'JiP<179) 4 .1. M. Patcrson LoKqwrison of Jet and Phase Space Models at l^, - 30 GeV (PEP-UJO) 8
('.. I Vinson ami I'. Uddone
Pnl.iri^^t^pjifccjuiu^iv-a K;wl of the Polarisation Group {1*1x1"-1HI > IS rt. lord. K. tiondo, l:. Martin, t\. Manning, D. Miller, C. Proscott *Uv"~v«K'iii of Ke.iK Interaction Contributions to cV - w'ti" Using Longitudinal r.-lariration of the e* and c' beams (PEP-18J) 19 0. Maiming •1 bean Pipe fcr Polar i:at Urn Iixjier intents (PEP-183) 23 (•- Martin V Systen for Obtaining t/JH£itu*iinal ttero Polarisation at PUP with Vertical Hi pole* Located Outside of the Interaction Region (FEP-184) 26 A- (Jarren, J. K;«dyk Wcafc Interactions Sum-try of the h'eak Interactions Croup (PEP-1S5) 31 A. benvenuti, K. Ford, I). ilitlin, K. fiondo, G. Venning, R. Morse, T. Rhoades, A. Sosson.*, I-. Stevenson. K. Strauch, A. Zollo It*-. Slall Nark II (PEP-ISO) 36 K. T. Ford [iM-hcak Interference in c*c* •* uV Scattering (PEP-187) 39 R. Morse Wcali-Ucctrorugnetii* Interference Effects in eV - Madrons (PEP-188) 43 I). Hitlin, A. Sessoms (icncral User Magnet A Ocncral Users Magnet Design (r8MS9) 46 P. Lobkowics, U. Becker, K, Berkclman, M. A. Green, £. Groves, K. Ifcilbach, J. Kadyk, N. Mistry, A. Sosscms, M. Strovink Use of Discrete Coils in Axial Field Spectrometers (PEP-190) 76 J. U. toy lor, K. A. Kernel Neutral Detectors
The Study of Neutral Particles (PEP-191) 84 K. Bartel, P. Bulos. A. Eisner, G. Hanson, 0. 'Utliti, U. Koetz, R. Kotthaus, P. Luke, H. Marsliak, T. Mast, J. Matthews, C. Peck, K. Strauch, D. Yount
The Crystal Ball at PEP (PEP-192) 88 K. Bartel, F. Bulos, D. Luke, C. Peck, K. Strauch
A Liquid Argon Neutrals Detector (LAND) for PEP (PEP-193) 94 A. Eisner, G. Hanson, D. Hitlin, U. Koetz, M- Marshalc, T. Mast, J. Matthews, C. Peck, D. Yount -iv- Resolving Overlapping Gamas in a Modular Neutrals Detector (PEP-194) 105 F. auios General Detector for Charged and Neutral Particles
Report of the General Purpose Detector Croup (FEP-19S) JOB A. Barbaro-Galtieri, W. Bartel, F. Bulos, R. Cool, G- Hanson, U. Koetz, R. Kotthaus, S. Loken, D. Luke, A. Rothenberg
Considerations for a General Flexible Detector (PEP-196) 116 F. Bulos
The Streamer Chamber as a General Detector (PEP-197) 118 G. Barbiellini, R. Kotthaus, S. Poucher, A. Seidl, P. Villa, D. Yount
A Time Projection Chamber-1975 (PEP-198) 126 D. Nygrcn
Comparison of Time Projection and Drift Cliamber Detectors [PEP-.199) 131 J. A. J. Matthews, A. Rothenberg
Mark [I Magnetic Detector for SPEAR (PEP-200) 138 R. R. Larsen
High Monentum ^ladrons with Particl••• Identification
Detection of High Mcmentun Particles with Identification of the Fin?l State (PEP-201) 139 U. Becker, R. Cashmore, E. Groves, L. Keller, S. Loken, C. Morehouse, S. Poucher, M, Strovink
Use of MicroChannel Electron Multipliers in High Energy Physics (PEP-202) 1S7 P. Lecomte, V. Pertz-Mendez
Photon-Photon Physics
Photon-Photon Physics (PEP-203) 15.9 G. Barbicllini, A. Benvenuti, K. Berkelman, A. Courau, F. Pester, K. -W. Lai, F. Lobkowicz, J. Matthews, N. Mistry, T. Rhoades New Particle Searches
Report of the New Particle Group (PEP-20-1) 176 A. Carroll, B. Cox, A. Eisner, K. W. Lai, F. Lobkowicz, M. Marshak, J. Marx, J. Matthews, N. Mistry, C. Morehouse, J. Poucher, R. Rothenberg, A. Peidl, D. Yount
Experimental Areas
Report of the Experimental Areas Group (PEP-205) 215 A- Carroll, B. Case, D. Coyne, F. Foster, K. Lobkawicr, F. Martin, C. Morehouse, P. Oddone, \ Prescott, L. Keller, G. Manning 1975 PEP Sumner Study
List of Participants
Guido Barbiellini Barrie E, Hughee Robert Horae Lab Nazlonali Wulfrin Bartel John Kadyk Jerry Helaon DESY LBL LBL Dave Nygren Ulrich F. Becker Levis Keller LBL HIT SLAC Pier Oddone" Alberto Benvenuti Ulrica Koet= LBL Univ. of Wisconsin tSESY Charles V, Peck Carl Herkelfflan Kunitaka. Kondo Cal Tech Cornell '.'iiv. Tsububa Univ. Martin Perl Charles D. Buchanan fiainer Kotthaus UCLA Max-Planck-Inst. sue John S. Poucher Fatin Buloa Kvan-Wu Lai Vanderbilt Univ, SLAC Brookhaven National Lab Charles Prescott Alan S. Carroll Alan Litke SLAC Brookhaven national Lab LBL ' 'erry 0. Rhoades Hcbert Cashmere Frederick Lobkowiez •ST Oxford Univ. Univ. of Rochester Barton BIchter* David Cline* Stewart Loken Univ. of WiaconGin LBL sue Allan Rothenberg Rodney L. Cool Dieter Ltikc The Rockefeller Univ. The Rockefeller Univ. SLAC Andrew Seidl Andre" Courau Oeofrey Hannlne Univ. of Michigan LAL Oraay Rutherford HEL Allen Seasons Bradley Cox Karvia L. Karshak Harvard Univ. Fermi lab Univ. of Minnesota Lynn Stevenson Donald Coyne Fred Martin LBL Princeton Univ. SLAC Karl Strauch* Alan M. Eisner Jay Marx Harvard Univ. UC Santa Barbara Yale ITtiiv. Mark Stro-rink William T. Ford Terry Maat LBL Univ. of Pennsylvania LBL Francesco Villa Prank Foster John A. 3. Matthews SLAC Univ. of Lancaster Michigan State Univ. ID iu» Winter Una Galtieri Donald H. Killer CfiHH LBL Horthweatern Univ. David E. Yoiart Eric Groves Narissn B. Miatry Univ. of Hawaii LBL Cornell Uaiv. Adriano Zallo Gail Hanson Charles Morehouse Ub Saziocali dl Fraseati SMC SLAC David Hitlin SLAC PEP Sunmer Study Steering Committee Member •VI- LIST OF PF.P NOTES ANI> RTinRTs TITLE PIP-I M. Mohl, ct. al. \ High Energy Proton - Elect run-Pnsi trim Colliding Bean system Dieter Moh!, Andrew A Model for a High Energy Proton- Sessler fclcctron-Positron CollidinR fleas Systenn; Spear Plus a Protnn Ring M. Stevenson Inelastic E!e<:tron-Proton Scattering Ki/JVl with SPEAR Plus a Proton Ring J'fcP-S S. Bennan, S, Prcll Feasibility Study for a 15 CcV Electron- J. Rees, B. Richter Positron Stoiage Ring PEP-6 Michel Davier, Electromagnetic Backgrounds and Photon 11/*»/"! Al Odian Tagging PLT-7 P.L. Morton, How Much Free Space Can h'c Provide (S«»EAtl - J.R. flees For Expertwenta! Apparatus? 132) PEP-S Dieter Kohl, On the Use of Isibelle in a PEP Andrew Sessler System and Other Related Topics PF.P-9 Stanley Flatte Proposed NAL Photoproduction Experi ments and some Comparisons with PEP Capabilities Stanley Flatte A Question of Duty Cycle Fred Oilman Angular Distributions for Had-ons Produced In PEP Electroproduction Experiments Gerson Goldhaber PEP Kinematics*-Deep Inelastic Scat tering— Twt Exclusive Reactions as Examples John Kadyi, Srookhiven HEDG Meeting of December John Rers 10, 1971 Check on the Equivalent Radiation for PEP PEP-15 Stanley Brodsky Radiative Processes in Electron- Proton Collisions at PEP NUMBER AUTHOR PEP-16 M.L. Stevenson The Kinematics and Possible Dynamics of Inelastic Lepton Scattering in PEP (IS GeV electrons on 70 GeV Protons) PEP-1? fierson Goldheber PEP Kineaatics--AddJiional Remarks on the Reaction of tht Type ep * e'oN* PE?-1» Bieter Kohl, PEP Parameters Andrew Sessler PEP-19 H. Wiedemann Limitations of the Transition Energy in targe ^ - p Colliding Bean Facilities PEP-20 Lloyd Snith On the Calculation of Luminosity for Electron-Proton Colliding Beaa PEP-21 M. A. Allen, High Voltage RF Systems for the PEP J. R. Rees Rings PEP-22 A. Scssler The Self-Destructive Behavior of Stored (ERAN-192) Elttctron Beams: The Disease Patterns, Symptons, and Cures PEP-23 kl Carren PEP Hodel One - A Machine Design Example 6/16/72 PEP-24 M. L. Stevenson Conceptual Design of a Hybrid Detect- 6/15/72 (CRISP 72- or for Electron Physics at ISABELLE 27) and PEP: Solon*/id • Quantaaeter • Hadroaeter (calorimeter) PEP-2S M.A. Allen Further Consideration of the RF System for PEP PEP-26 J.E. Augustin Multiple Couloab Scattering and Gas (SPEAR-147) Breasstrahluag at SPEAR PEP-27 H. Wiedrsixn A Correction to Formulas Computing the Touschek Lifetime in Storage Rings. PEP-28 A. Sessler Strongly Turbulent Collective Motion and the Anomalous Size of Stored Particle Beans PtP-29 R.O. Bangerter Variable Proton Momentum at PEP 3/20/73 PEP-30 E. Hartwig, Noise *n Proton Accelerators 3/73 (i,BL-139S) V.K. Nell, R.K. Cooper PEP-31 R. Bangerter, PEP Lattice Design (LBL-1714) A. Garren, et. al. DUMBER AUTHOR PlP-32 A. Garren, Use of the Electron Ring for T. Elioff Protons in the PEP System PEP-33 Allen. Avery, Proton-Electron-Positron HcMgti Elioff. et. al. Study (LBL-1746 an.1 SLAC-1220) PEP-34 A. Garren PEP Model Five: An Update of (Rev.) PEP Parameters PEP-35 L. Stevenson Kinematics and Possible Dynamics of e • * X e' * Madrons PEP-36 T. Elioff Nctca of PEP "Bull Session" of March 1 - 2, 1973 PEP-37 P. B. Wilson Beam Loading in High Tner^y St•-rage (SPEAR-163) Rings PEP Notes 38 - 72 are in the 1973 PEP Stumer Tu-Jy Report dated August 1973. A. Gerrrn PEP Model Six H. Wiedemann Scaling of FODO-CELL Parameters M. Month, Closed Orbit Beim-Bcam Effect for A Ruggisro Crossing Bean;; Space-charge Effects at Transition Energy: An Attempt to Scale from the CPS to PEP-6 and Other Machines The Excitation of Non-linear Resonances by a Displaced Elliptical Beam Bunch Lengthening and Nidening Effects Due to the Combination of RF Noise and the Prcscncr of Inductive Wall Elements A. RuQgiero Bean-Beam Linit in SPEAR as a Single Resonance Effect F. Sacherer Bunch Lengthening H. Wiedemann Proton Beam Enlargement by Gas Scattering R. Helm Synchrotron Radiation Integrals from PEP-6 NUMBER AUTHOft PEP-arf II. Niedaaann Enlarfaaant of ttia EI«ctron Beta Cross Section in a Storage Ring due to an Oscillating Synchrotron Radiation DanpMg Tina Constant TEP .9 A. Ru|Ki«ro Transversa Bunch-Bunch Instability in PEP (Raiistive Vail) PEP-SO N, King Seme Cowacnts on Homing Points and Resonance Effect! in PEP Lattices PEP-SI G. Rees Aspects of PEP as Compared to EPIC PEP-M A. Runipro, Calculation of Resonance Effects Due L. Saith to a Localized Gaussian Charge Dist ribution PEP-53 ll. Hereward Equilibrium Energy Distribution in • nor.-linear Potential Hell in the Presence of Quantim Fluctuation PEP-54 A. Rufgiero Kead-rail Effect in PEP PFP-SS M. Melvin, Interaction of a Coasting Bean and A. Ruggiero a Bunched Bean wit?) Frequency Slip PEP-56 II. Hercward Soae Possible Causes of Bunch Shape Distortion in SPEAR PEP-57 H. Herevard Possibility of Observing Turbulence in SPEAk PEP-58 H. Wiedemann e - p Luminosity for Different Energies ir ?EP PEP-59 E. Keil Diffusion-like Blow-up in Asynchronous Bunched Bean Collisions PEP-bO R. H. Helm A Negative Momentum-Compacticr. Lattice PEP-bl H.J. Lee. Magnet Insertion Code [MAGIC1 (SPEAH-165) W.W. Lee, et. al. PEP-63 R. Chasman, A. PEP with Crossing Angle Garrcn, et. ai. PEP-63 J. Augustin Longitudinal Beam-Beam Effect in Head-on Collisions NUMBER AUTHOR PEP-64 J. LeDuff Behavior of a Stochastic Non-Linear System Excited by an External Harmonic Force. PEP-6S J. LeDuff Diffusion on a Single Non-Linear Resonance in the Case of e-p Collisions PEP-66 T. Eiioff, Storage Ring Experiments H. Hereward, ct. al. PEP-6^ H. Hereward Influence of the Touschek Effect on Life-time Measurement in SPEAR PEP-68 D. Hohl, The Use of Rf-Knockout to Measure P. Morton Synchrotron Oscillation Frequencies and Energy PEP-69 J. Rees, Preliminary Design of a IS GeV fa. Richter Electron-Positron Variable-Tune Storage Ring PEP-70 M.A. Green Superconducting Dipolis and Quad- rupoies for the PEP Accelerating Storage Proton Ring PEP-71 J. R. LeDuff Properties of Bunch Lengthening Effect Observed on Existing e+e" Storage Rings Incoherent Beam-Beam Effect: A Computer Simulation PcP-75 R. F, Schwitters Comments on Obtaining Longitudinally Polarized beams PEP-76 J. B. McCaslin, Neutron Shielding for PEP R.H, Thomas PEP-77 A. fiarren, Detection of Proton Beam Jet in J. Kadyk 15 x 200 GeV PEP PEP-78 R. Helm, Double Thin-Lens Approximation for 12/73 (SPEAR-169) M. J. Lee Preliminary PEPSI8 Lattice nesign PfcP-T9 B. Richter PEP Parameters 1/74 PEP-8H J. Rees, Tentative KP Stage 1 Bear.t Stny - Clear 1/74 A.V. Lisin Requirenonts and Bean Line Space Allocation NUMdEfl AUTHOR PEP-81 P.J. Channell Transverse Diffusion of Proton 2/14/74 Seams Due to Noise PEP-81 R. H. Thomas Proton Losses from PEP 2/21/74 PEP-83 T.M . Jenkins, Kuon Shielding for PEP 2/21/74 R. H. Thomas PEP-84 G. E. Fischer PEP Inflection 3/1/74 PEP-85 G. A. Loew, A Few Thoughts Regarding Beam Cavity et . al. Mode Excitation in PEP 5/16/74 PEP-86 L. Smith Proton-Electron-Positron - PEP 5/17/74 (LBL-3032) PEP-B? R. Schwitters, A Method for Producing Longitudinal 6/74 [5PEAR-17S) B. Richters Beam Polarization at PEP PEP-88 J. Jurow, Sypchrotroti Radiation Absorbing Surfaces 7/25/74 N. Dean PEP-89 M. Allen Higher Cider Modes in SPEAR II Cavities 7/74 PEP-90 M. Sands Energy Loss to Parasitic Modes of the Accelerating Cavities L, J. Laslett Concerning the Density Distribution and Associated Fields PEP-92 M. Sands Parasitic Cavity Losses in SPEAR-2 PEP-93 I. I. Laslett Examples of Weak-Beam/Strong-Beam fi/5/74 Computations Performed by use of the Program "KEAK8", with Graphic Output PEP-94 L.O. Laslett An Exanple of the Use of Program 8/6/74 "WEAK9" PEP-95 M. Sands A Bench Measurement of the Energy Loss of 8/8/74 a Stored Beam to a Cavity PEP-96 J. Roes The PEP Electron-Positron Ring - PEP 8/12/74 Stage I PEP-97 M.Allen RF Systems for High Energy e~e+ 8/14/74 P. Wilson Storage Rings PEP-98 R. Helm Preliminary Design Considerations for 8/14/74 H. Lee the Stage I PEP Lattice PEP-99 R. Helm, H. Lee, Beam Enlargement by Mismatching the 8/14/74 J. Paterson Energy-Dispersion Function PEP-100 P. Wilson Beam Loading in High-Energy Storage 8/20/74 Rings NUMBER AUTHOR DATE PEP-101 P. Wilson Stored Current Capability of the PEP 9/10/74 R.F, system PEP-102 W. Heme) Comparison of Two Configurations for 9/23/74 Intersection Regions PEP-103 H. Allen Measurement of Higher Order Mode Losses 9/19/74 J.M. Paterson 1n SPEAR II by Shift in Synchrotron Phase P. Wilson and Increase In Net Cavity Power PtP-104 P. Channel I Alternative Theories of the Non-Linear 10/29/74 Negative Mass Instability PEP-105 A.W. Chao Physical Picture of the Electro-magnetic 2/3/75 P.L. Morton Fields Between Two Infinite Conductlna Plates Produced by a Point Charge Moving at the Speed of Light PEP-106 J.H. uenklns Shielding Required for Radiation J.B. CcCaslin Produced by the 15 GeV Stored Electrons R. H. Thomas PEP-107 U. Uenzel PEP Experimental Areas-Winter of '75 2/12/75 PEP-108 LBL-5LAC Joint The PEP Electron-Positron Ring: 2/25/75 Study Group An Update (from SLAC) PEP-109 U.fi. Nelson The Radiation Dose to the Coil 3/6/75 G.J. Warren Linings and the Production of (Health R.L. Ford Nitric Acid and Ozone from PEP Physics) Synchrotron Radiation PEP-110 A.w. Chao Higher Order MultipMe Magnet H.J. Lee Tolerances P.L. Morton PEP-I11 H.J. Lee Control of Closed Orbit Deviation 3/7/75 P.L. Morton Due to Synchrotron Radiation J.R. Bees B. Richter PEP-112 K. Bane 3/25/75 P. Wilson PEP-113 R.T. Avery The PEP Injection System 3/27/75 J.M. Petarson K.L. Brown Pl:P-114 D. BostLC, Vacuum System for the Stan Ford- 3/27/75 et al. LBL Storage Ring (i'F.P) PEP-US M. Allen JJuam Bnergy Loss to Parasitic 3/27/75 J.M. Patersou Modes in SPliAR II J.R. Rees P.U. Wilson 3/28/75 PEP-llfi F. Ha.-tin background Esti mates for PEP 3/28/75 PEP-1 17 H. Weia°mann Implication of Shorter Cells in PEP 4/?4/7S PEP-I1R A.W. Chan Parasitic Loss of a Gaussian Bunch in a Closed Cavity PEP-U9 A.M. Cliau P.L. Morton Differential Energy Loss for a Particle in a Sfiare Pulse of Cliarqe Traveling Between Infinite Conducting Plates .MUMIII.it AUTilOR TITLL UATL PEP-120 J.L. PeUigrini Deflection of fielativistic fleams by 5/30/75 Heans of a Traveling Wave Elecfxodu PEP-I2I E. Keil The Dehavior of Betatrun Oscillation in 5/3B/75 the Vicinity of Half Integral Structure Resonances. Pi:P-122 Roger HcConnell Beam Induced Transit Time Signals at 6/6/75 (SPfAR-IflS) SPEA'; IW-121 P.P.. Wilson 6/9/75 PEP12* T. Jackson High Performance Magnet Power Supply C/1U/75 Optimization PEP-125 J. Paterson Beam Size Control in PEP 7/21/75 J. Rees H. Wiedemann PEP-12<> i). Kcii Hunch Lengthening and Bucket l)is- 5/75 tort ion Due to Cavities i1-.l'-127 R. Avery Injection Transfer Lino Coordinates 7/3J/75 pi:i'-i2H H. Avery PEP Rin'j Coordinates - Configuration E H/1X/75 I'i:IW2!) A.W. OlilO On the Horizontal Shape of An Electron a/13/75 Hunch P'.^-130 A.W. Chao Stationary Solution of the l:okker/ 8/28/7S ' rt-AR-taa Martin Lee Planck liquation for Linearly Coupled Motion in an lilectron Storage Riny A. Chao Changing Particle Distribution for 9/8/75 M. Lee Linearly Coupled Motion in an Electron Storage Ring A. Chao Evaluation of the Field Quality of the 9/9/75 M. Lee Prototype PEP Cell Quadrupole Magnet P. tor ton F. Hall PEP Cooling Water Systems and Under- 10/7/7S D. Robbins ground Piped Utilities Design Criteria Report PEP-137 1974 PEP Sunnar Study (PEP-138-177) Aug. 1974 PEP-138 K. Strauch Introductory Remarks PEP-139 J. Rees The PEP Electron-Positron Ring PEP-140 B. Rienter Event Rates to be Expected at PEP PEP-141 G. Abrams Report of the Study Group for the et.al. Measurement of the Total Cross Section for e e" •* Hadronii PEP-142 H. Lynch, Hon-Hagnetic Detector for Measuring R. Schwltters oy and Charged Multiplicities in e e" PEP-143 G. Feldnan, Precise Measurement of the Total Cross D, Hltlln Section PEP-144 D« Nygren The Time-Projection Chamber - A Hew 4n Detector for Charged Particles PEP-145 B. Shcn Contribution of the Two-Photon Annihilation Process in the Measurement of nT at PEP PEP-146 6. Buschhorn Detection of High-Momentum Hadrons et.al. PEP-147 G. Barbiellinl Heavy Hadrons - Group Report PEP-148 A. Barbaro- Large-Solid-Angle Detector for Charged and Galtieri, et.al. neutral Particles PEP-149 J. Perez-y-Jorba The Detection of Low-Energy Charged Particles Identification Up to 3.5 GeV/c PEF-150 P. Spillantini Study of a Hulti-Hadron Facility for PEP Based on Toroidal Field Magnets PEP-151 Gt Buschhorn, A Streamer Chamber Detector for PEP et.al. PEP-152 P. Yamin Comments on the Simultaneous Measurement of Charged and Neutral Components of Multlhadron Events PEP-153 T. Mast, Sone Design Considerations for a Large J. Nelson Solid Angle Charged Plus Neutrals Detector fo» e+e" Storage Rings PEP-154 E. Bloom Report on Neutral Particle Detectors and et.al. Q.E.D. PEP-155 E. Bloom Properties of Some Photon Detectors Aug, 1974 et.al. PEP-156 Liquid Argon Ganma Ray Detector - Variations of the Willis Chamber PEP-157 D. Buchholz, Report of the Weak Interactions/EM et.al. Final States Group PEP-158 D. Cline, Tests of y-e Universality for Weak Neutral L. Resvanis Currents at PEP PEP-159 V. Camerini, A Compact Magnetic Detector for .- -u" et.al. Asymmetry Measurements and Long' udinal Polarization Utilization at PEP PEP-160 S, Yellln A Suggested Detector PEP-161 K. Strauch Study of the Reaction e e" * p p' with an Iron Solenoid Spectrometer PEP-162 Solenoid Spectrometers for e e~ •+ u u~ and Other Final States PEP-153 P. Limon, Direct Measurement fo Muon Polarization et.al. in e e~ •*• u u" PEP-164 D. Hitlin, Strange Particle Experiments at PEP et.al. PEP-165 J. Klems Parity Violating Momentum Correlations as a Means of Observing Weak Interactions in e+e" -*• Hadrons PEP-166 D. Buchholz, Polarization Group Coordinators' Summary et.al. PbP-167 R. Schwitters Control of Direction of Beam Polar nation PEP-168 K. Wenzel Notes on Longitudinal Beam Polarization PEP-169 W. Toner Resonance Method to Produce a Polarization Asymmetry in Electron-Positron Storage Rings PEP-170 U. Camerini TWo Methods to Measure the e" Polarization at PEP PEP-171 C. Prescott A Pulsed Polarization Monitor for PEP PEP-172 W. Toner A First Look at a Polarimeter for EPIC or PEP PEP-173 D. Buchholz, An Alternate Way of pleasuring Beam Aug. 1974 et.al. Polarization at an e e" Colliding Beam Facility PEP-174 D. Berley, New Particle Searches at PEP et.al. PEP-175 G. Barbiellini Colliding Y Beams: TVo-Photon Processes et.al. and Tagging PEP-176 H. Lynch, Background Sources at PEP et.al. PEP-177 b. Berloy, Report of the Experimenta3 Areas Group ct.al. IMTKODirTIOH K. iHriiiicb The 1975 PHP Summer Study took place In an available be!\>rr Uie r.tart of the 1975 Lepton- atmosphere of great excitement and great expecta Phaton iiyrapoaiu.^. tion. The basic discovery of the polnrized, at leact in the reaction e+t* * u*u~i Chamber. However, because of limitations in time, promising solutions were discussed for the problems only a start was made on the evaluation of the of space limitation and synchrotron radiation performance of the various detector designs for which are present in the polarization scheme of events with Jet structure. Further analyses is Kichter and Schwitters described in the 197'' Pro needed. ceedings. All final large magnetic detector de The PEP design group under John Rees and Tom signs use a solenoidal field produced by a super Elioff worked closely and effectively with the conductor; differences arise mainly Prom the use Summer Study. The participation of colleagues of lumped or continuous coils. A General Ifsers from the CERN, DESY, Praseati, KEK, Orsay and Magnet appears quite feasible and practical; its Rutherford laboratories was most helpful and desirability wiil, to a large extent, depend on much appreciated. We thank the management of the interests of the user community. With one ex these laboratories for making this participation ception, the specifications of the solenoidal field possible. and volume were similaij suggesting that perhaps For the second year in a row, LBL played host two such devices would serve most of the presently to the PEP Summer Study. The LBL staff, under envisioned experiments needing large magnetic the direction of Andy Ressler and Bob Birge, volumes. (The one exception is a design requiring worked most effectively to provide all necessary the use of many micro-channel phototubes which is support and assistance. All participants are not yet an economically viable reality.) Photon greatly indebted to Pier Oddone who more than any detection over a vide energy range remains a other individual was responsible for the fine problem. Hal has most of the desired character organization and smooth working of the Study. istics, but is too expensive to surround a Margit Birge and helpers cheerfully satisfied all magnetic field; argon calorimeters are bulky anil reasonable and unreasonable secretarial requests. have problems with low energy response. Every Jean Lynch was everybody's friend and helper in interaction region needs a tagging system, either her multiple roles as effective housing agent for to study two-pliotcn eventB, or to at least esti families, understanding sorority mother for lonely mate the two-photon background. bachelors, ojok of delicious barbecues and snacks, elegant and charming hostess of a splendid dinner Every general magnetic detector design tries dance at the San Francisco St. Francis Yacht Club. to achieve kit solid angle coverage, reliable i- The hospitality of our Berkeley colleagues was in dentification and accurate momentum determination the best western tradition and greatly appreciated of both charged and neutral particles. The dis by all visitors and their families. coveries of the last twelve months reemphasize the importance of these goeJ.s. Since no one We thank ERDA whose support made the Summer detector has yet been designed which incorporates Study possible. There are many similarities optimum solutions for each goal, designs differ between a Wine Tasting and a Summer Study. Par mainly In their emphasis on different desirable ticipants are exposed to many bouquets and flavors properties. with the expectation that one or two will prove At PEP energies, hadronic events are expected attractive enough to be worth long and continuous to involve about twice as many individual particles enjoyment. Realization of such long term activity as at SPEAR. These particles will probably be requires, of course, interest in participation and emitted in two back-to-back Jets. Such Jets would availability of supplies. I believe that the 1975 produce difficult problems in Individual track PEP Summer Study further encouraged the U.S. High reconstruction and in the use of big Cerenkov Energy Physics community to participate in the PEP counters. To alleviate the problem of track re program. And, I very much hope that our work construction, two-promising central detectors helped some to bring closer the realization of this surrounding the interaction region were discussed: most exciting new tool for probing Into the funda the Streamer Chamber and the Time Projection mental properties of nature. Appendix Smmary Review Sessions of Work of the 1975 PEP Sumner Study Tuesday August 19 Wednesday, August 20: Welcome and Introduction 9:30 A.M. Streamer Chamber Detector 9:30 •• 10:00 A.H, D. Yount An Experimenter's View of PEP 9:45 - 10:30 Time Projection Detector 10:00 • 10:30 J. M. Paterson D. Nygren Polarized Beams at PEP 10:30 - 11:15 Neutral Detectors 10:30 -• 11:15 D. Miller C. Peck Coffee Break 11:15 - 11:30 Coffee Break 11:15 - 11:30 Weak Interaction Effects in Search for New Particles 11:30 • 1:30 P.M. QED Reactions and Hadron J. Marx Production 11:30 - 12:45 W. Ford Lunch (LBL Cafeteria) 12:30 - 1:30 P.M. Lunch (LBL Cafeteria) Two-Photon Physics 1:30 • 2:30 K. Berkelman General Purpose Detector 2:00 - 5:0C P.M. R. Cool Experimental Areas Report Z:30 - 3:30 G. Manning Multi-User Detection System 3:00 - 4:00 F. Lobkowicz Concluding Remarks 3:30 - 3:45 Coffee Break High Momentum Detectors and Particle Identification 4:IS - 5:15 Wijie and Cheese Reception 6:00 (LBL Cafeteria) AN EXPERIMENTER'S VIEW OF PEP 0. H. Paterson Stanford Linear Accelerator Center Abstract LUMINOSITY VS. ENERGY Since the 1974 Summer Study, the PEP design Figure 1 shows the luminosity versts energy has been greatly refined in ways that have very function for the present PEP design. In the region much expanded its physics capability. In this re below 15 GeV, thejjftnrlnosity « E2 and the stored port, I shall briefly discuss the effects of the current I « E, rising from 3 x lo11 particles/ changes on the luminosity vs. energy function, beam bunch at 5 GeV to 10" particles/bunch at 15 GeV. polarization and backgrounds. Experimental facili It should be noted that since last year's Summer ties are discussed elsewhere in these Proceedings. Study, SPEAR II has come Into being with the same rf frequency as FEP and therefore the same short INTRODUCTION bunch length. Already in SPEAK II, currents of 3 * 1011 particles/bunch have been stored at the During the past twelve months, there have been much lower energy of 1.5 iieV. This has greatly two very important changes in the basic PEP magnet reinforced our confidence in the projected PEP per lattice. The first of these is the change from a formance. In this energy region, the wiggler mag 48-cell magnet structure through the curved sec nets will be used to maintain a beam size indepen tions of the ring to a structure with twice as many dent of energy, a necessary feature in obtaining quadrupole focusing elements evenly distributed; this performance. i.e., a 96-cell structure. This structure change does many things to improve PEP, but from the ex 10" perimenter's point-of-view, the most significant is the expansion of the energy range to higher ener , ^ ...... gies. It also leaves open the possibility of fur ther extension to still higher energies in future improvement programs. At the same time this design change was made, six short straight sections of 5-metre length were • 2^ \ \ \ \ • incorporated into the lattice. They are in the centre of each of the curved sextants. The idea S L.mtMby { 1 X was to leave space for the as-yet-unknown beam con trol devices which may be desirable in the future. Total InUaM RF P«Mf: 9MW Shortly thereafter, three of the six straight sec Tx Lt:JlcliiiCoirilrLtnfth tions were occupied by "wiggler" magnets. These NumMiol Bundm:B>3 magnets are used to control the beam size at ener ! i gies less than IS GeV by increasing the synchrotron L.'Se re iu« radiation emitted by the particles. As the wiggler magnets are independent of the main bending magnet system and only locally perturb the orbit, the syn chrotron radiation power emitted by the beam can be controlled at any fixed energy, giving smooth con trol of beam size, energy spread and damping times. A more complete discussion of this system can be found in reports SPEAR-186/PEP-125. Beam size con trol is necessary to optimize the luminosity at Fig. I — Luminosity versus energy, energies less than 15 GeV where one is no longer a rf-powar limited. This control was previously ac As before, the luminosity peaks at 10"cm" sec" complished by the use of different lattice arrange at 15 GeV, where the power lost to synchrotron ra ments at different energies. The new technique is diation and to higher-order-mode losses in the more flexible and has several advantages, including vacuum system is balanced by the planned 9-MW rf shorter injection times and shorter polarization accelerating system. Above this point, the current build-up times at these lower enerqies. fiat can be stored falls very rapidly, and with a fixed magnet lattice, the luminosity would fall as The updated PEP design, which includes the E-10. It is here that the increased flexibility above changes, is being published essentially si of the 96-cell lattice becomes important. By using multaneously with this report (vii., PEP Conceptual the variable-tune method of beam s*ze control to Design Report, LBL-4288/SLAC-189, October 1575). decrease the beam size as the energy increases (in The reader is referred to this report for the de a fixed lattice the beam size increases as E), the tails of what I am about to discuss. rapid fall in the luminosity function can be ame liorated to a &<* E"3 dependence. Figure 1 shows -5- three different cases where, keeping the rf power a ding this procedure. Using the presently-designed c-nstant, the length of the accelerating structure wiggler magnets at full strength, i.e., 20 kG, is increased. The first curve corresponds to the would give too large a beam size for the available p.^sent PEP design and now has a luminosity of aperture. However, using the variable tune tech i;H cm-*'sec-1 at 18 GeV. The other curves indicate nique to reduce the beam size as previously described possible Future expansion of the energy range by for operation above 15 GeV, one could achieve the increasing the length of the acceleratina structure performance shown in Fig, 3. This is another exam wthout modifications to the lattice. These steps ple of the improved flexibility and future potential toward a PEP 11, although outside the scope of the of the present PEP design. PEP budget, are inexpensive enough to be undertaken as accelerator improvements. BEAM POLARIZATION During last year's study, there was consider able interest in the potential use of the predicted transverse polarization that builds up in a circu lating electron beam that is emitting synchrotron radiation. However, there was still some uncer tainty as to how fragile the beam polarization wou'd be in routine operation. The very strong non-linear fields associated with operating these high-energv rings with large beam-beam effects could increase the number of depolarizing reso nances to the point where they would make the storage ring so sensitive that effective polarisa tion might be impossible to maintain. Exoerience with SPEAS II in the last year has done 'uch to remove these worries, as a hioh degree of beam 8 10 12 14 16 polarization was achieved in routine operation. E (Gev) With a polarization build-up time of 20 mins, the beams were 85. polarized within an hour and were Fig. 3 -- Polarization build-up time versus energy insensitive to operating configurations. A beam for the case where the wiggler magnets depolarizer, which utilises an oscillating trans are at full strength and the lattice is verse rield, is presently under construction and varied to maintain a constant beam cross will be installed and tested in SPEAR during the section. comino year. This accumulation of experience on polarization in SPEAR will be very useful in olan- nincj for PEP. BACKGROUNDS Figure 2 shows the polarization build-up time Estimates of particle loss rates in the neigh in PEP. The dashed curve corresponds to operation borhood of the interaction regions were presented -! without wiqgler magnets and Tp0j « E \ When one in '..st year's study, They indicated that at the utilises pigolsrs to achieve the-!/'™ I' performance, highest cur.-ents there is an approximately 50' one has the added benefit of a reduction in Tpo) probability of a narticle striking the vacuum cham at the lower energies as Shown in fig. ?, If the ber close to, or in, the interaction reyicn for physics required the use of beam polarization at each bunch passage. Since then, a more complete even lower energies, this can be achieved by exten- Monte-Carlo and tracking calculation has confirmed these estimates (PEP-U6). These calculations show that particles which have undergone bremsstrahlung in the curved sextants have only a very small pro bability Of reaching the interaction region and that the source of background events is the-brems- strahlung (both the photons and the electrons) occurrinq in the 50 metres on either side of the interaction point. This is valid also in SPEAR and has been checked experimentally. It is ubvious that a great effort will be made to keep the pres sure in this region as low as possible. However, one can also attempt to prevent these particles from targetting on the interaction region beam pipe bv including aperture stops in the system. This has been studied for PEP, and Fig. 4 shows a pos sible set of thick, high-7 collimators which should sionificantly reduce the number of particles which can contribute to experimental backgrounds. Figure 4 also shows two new Important fixtures. E -UVI Thpy are: the addition of low-field magnets at the end of the sextants and the movable thin hfipv7 rtiask at the end of the interaction region, these Fii. ? -- Polarization build-up lime versus energy devices are necessary to tdfce care of a ptoblem for the case whL-r«» tin- l Fig. 4 -- Layout of insertion region straight section showing locations of background collimators, synchrotron radiation masks and low-field magnets. more serious in PEP. The problem is that as the beam energy increases, the synchrotron radiation power increases as E1* and the critical photon ener gy as EJ. This means that the number of hard x-rays or soft Y-rays increases rapidly and masking becomes very difficult. In SPEAR at the higher energies, these hard x-rays scattering from the edge of the tantalum masks which masked the inter action region pipe penetrated the vacuum chamber and literally saturated the counters and chambers in the vicinity. To remove this problem, it was necessary to reduce the flux and photon energy of 'ne radiation impinging on this last mask. This was done in SPEAR, and is planned in PEP, by lower ing the bending field at the end of the sextants. These magnets and their synchrotron radiation pat tern are shown in Fig. 4. The photon flux through the interaction region with and without the low- field magnets is shown in Fig. 5. These curves are self-explanatory and clearly show that in designing +l.c interaction region vacuum chambers and experi mental appar^us, one must take cars to avoid the unmasked region close to the bean. Another important background problem lies in beam-gas interactions within the experimental appa ratus. For some experiments, these interactions can simulate the events of interest and require background subtractions. As the gas pressure dis tribution within the apparatus is difficult to moni tor and is a function of the operating conditions of the storage ring, one cannot easily condense the accumulation of data on beam gas background in any quantitative way. With the luminosity region being short (10-15 cm), one can, however, cleanly separate beam qas events from those from e+-e~ annihilation by studying the distribution of events along the beam axis and extrapolating through the luminous region. This, of course, implies tnat the apparatus Fig. 5 -- Synchrotron radiation flux versus photon has an acceptance over, say, ±50 cm, which is well energy for radiation from the main bend understood and has adequate spatial resolution in magnets and from the low-field magnets. the beam direction. When an experimental apparatus becomes large tion should be given to including some pumping ca and vacuum pumping of the vacuum chamber is only pacity within the 20 metres where the experiment is from the pumps in the neighboring quadrupales, the sensitive to the typical e-p scattering events of pressure in the interaction region could be higher than desired. In designing experiments, considera- beam gas interaction. CONCLUSIONS continuing experience with SPE*V* is of great assis tance in refining the design of both the ring and The PEP design has matured during the last the experiments. Certainly the "new" ohysics nf year and fts physics capability has improved both the fast year has only increased our anticipation in overall scope and in detailed execution. The of the excitement of PEP. COMPARISON OF JET AMD PHASE-SI'ACK MODRLE1 AT E =• 30 GeV G. Hanson and P. Oddono I. ifitnouarnoii Figure 3 comiiiires the sphericity distribution :' •• the data and the two modelB. The sphericity iu Data from the SLAU-LBL Magnetic Detector at obtained hy rinding the axis such that the sum of iiPL'AIi has heen compared to a phase-space model ami the squares of the transverse momentum relative to ;i jet model.1-!- Below a eenter-ot'-mass energy to the axis is mitiimizeu. Then the sphericity Is {Ecm' (1f '' i't*V the data does not distinguish be- defined as tv.">«n the two models. Above an Ecm of U GeV, in particular for the high statistics runs at Ecm 1 ~ o. OBSERVED x DISTRIBUTIONS HADRON EVENTS,* 3 PRONGS Data Monte Carto,Umited - Monte Carlo, Phase Space Transverse Momentum Fi<;. 1. Momentum distribution for charged tracks for the SLAC/LBL Magnetic Detector data at E._ = 6.2 GeV and Ecn, = J.k GeV. OSSFRVED SPHERICITY DISTRIBUTIONS HADRON EVENTS,!3 PRONGS • Data — Monte Carlo, Phose Space — Monte Carlo, Limited Transverse Momentum 300 l i > I'll 1 E^ 6.2 GeV (a) 200 /* * ST^ - j •rWi \ L / tt ' f\* S s, l0 v — r l ^fc* •B ° 1 •«cj T(L^ (sp h ) 1 1 1 I 1 I w u o d E^ 7.4 GeV (b) t"o 700 — ~ £ 600 A - UJ V^Mr * > 500 fl ' r4 *• s 400 j " f *4. ^ 300 r * " 200 — / ~ 100 - 1 1 I.I, 0.2 0.4 0.6 0.8 SPHERICITY Fig. S. Sphericity distribution for events from the SUIC/LBL Magnetic Detector at E = 6.2 GeV andE„m = 7.*i GeV. sections that follow ve compare the distributions predicted by the Jet model to those predicted by the phase-space model at Ecm = 30 GeV, and show some distributions which may be helpful in the design of detectors. where the summation is over all pion3, Pjj is the transverse momentum of the i*h particle relative II. PHASE-SPACE AND JET MODELS to a given axis and r is a parameter that can be varied to give the desired mean Pj. The Jet axis In the phase-space model, pions (charged and can be given any desired angular distribution. neutral) are produced with momentum and angular For production through a single virtual photon, distributions according to Lorentz invariant the general form for the angular distribution for phase space. The total multiplicity is given by a unpolarized beams is Poisson distribution whose mean can be specified The fraction of pions vhieh are neutral can also be specified, and tha decays if0 •*• YY are included. -JTT a 1 + a • cos G The Jet model hos all the features of the phase-space model. Jets are simulated by modi with-1 < a < 1. fying phase space, according to a matrix element For the distributions shown below, we have squared (H2) taken an average multiplicity of 15.3 wit'.i a MOMENTUM ALONG AXIS (CEV) MOMENTUM ALONG AXIS 'CE MOMERTUM ALOKC AXIS (CEV) MOMENTUM ALONG AXIS (GEV) MOHEHTUM ALONG AXIS IGEV) OMENTUM ALONC AXIS (CEV) Fig. 3. Some randomly chosen events at Ecm = 30 GeV. On the right hand side are the events generated by the Jet model, and on the left hand side by the phnse-space model. Solid lines are charge pions, and dotted lines are i'--_-tral piona. HP SIMULATION. PH.3P. 30 CKV PEP SIMULATION. JET JOGEV CHARGEP PION HOHCHTUM CHARGE? HOH yOUCHTUU HOHtNTUH(CZV) HOUENTUU(CEV) t>EP SIMULATION. FH S* 3D GEV (•KP SIMULATION. JET JOCEV COS. Or ANGLE SET*. CH. PIONS COS. OF ANCLE BETi CH PIONS 1 0 ENTRIES SPMCIIICITT Fig. h. Distributions for charged tracks generated by the phase-space model and by the Jet model. Solid line in Figs. Ma) :uid l»(b) are from the scaling form suggested by Richter.3 The distributions are all based on 2000 events. PEP SIMULATION. PH.SP. 3D CEV Pi:p SIMULATION. IET SOCEV PHOTON MOMENTUM PHOTON MOMENTUM 1000 \ V 100 \ 100 \ 10 10 1 ; \, 1 1 /w .Y %,.. •OIIENTUM(CEV) a«MVNTUMtCE*) PCP SIMULATION. PH.SP. 10 CEV PCP SIMULATION. JCT 30CEV ANGLE BETWEEN GAMMAS 4NCLC •Kt*ECH CAMMAS ANCLE OP PAIR ANCLE OF PAIR PtP 91NLLATI0N PH SP iZ CEV PfcP SIMLLATION. JCT lOCtV ENERGY VS. ANCLE Of GAMMAS ENERCT VS. ANCLE OF CAHMAS 29103 ENTRIES 305»3 ENTRIES CHERCY or PAIR((CEV) CNERCT OF PAIRUGCV) Fig. 5. Distributions for photons generated "by the phase-space model and by the jet model. All distributions are baaed on 2000 events, except the plots 5(e) and 5(f), which are based on 200 events. PIP SIMULATION. PM SP 10 CtV II- tlCi, tTIOM JIT ICCtV HUlTIPtlClTT At P«0» VMTCX MULTIPLICITY AT MOD VtRU k tK> . i v 10' « • •« i«' 1000 CWtWUS . . t W '6' • » ** '0 ' *500 CKTtffL* (B) I 1 3 'L Vi MUlTIPtiCITV MVLtll'LirtTt I'lf MM( LATlOK PM SP *- Ot* -iNiiaTiON Jft i:c(* MASS sottirfO or CM PI P»IK- MAS* SQUARED or CM PI kiln- •»•••"», COli Hl«'. KOTftll-- MASS SQUARED OF P*IIUCEV • •. ) •Ms* *QLARED OF PAIRICLV" ) PEP SIMULATION. PM SP 10 CC* PIP SIMULATION JIT IflCRV MASS SOUAftCO OF PHOTON PAIR* H4SS 54UAME0 OF PHOWK (>«IR FIR 6. Distributions uoerul for estimating backgrounds ts two body B distributions for the phaae-space and jot nodelo. The distribution! based on TOCO events. fraction of 3° of .51*. The average transverse small size of Pj^ compared to the average Pt leads soaentum relative to the jet axis is 310 MeV, and to a sphericity peaked towards small values for the jet axis has been taken along a fixed di the Jet model. rection, as the distributions shown do not do>end Figure 5 shows some distributions for photons on the angular distribution of the Jet axis. derived from the decay of n°*s. The momentum dis tributions 3hown in Fig3. 5(a) and 5(b) reflect III. COHPARISOHS AT E » 30 GeV the original n° momentum distributions. The angle between photons shown in FI^s. 5(c) and 5(d) showB Figure 3 shown the transverse and longitu the same pronounced peaking for the Jet model as dinal Boaentum with respect to the Jet axis for is shown by the distribution for charged pions. At some typical events. On the left hand aide of small angles a peak due to photons belonging to Figure 3 appear the events generated by the phase- the same ir° appear In the phase-apace model. In space model and on the right hand side events the Jet model, the peak is overwhelmed by the generated by the Jet model. At Ecm - 30 GeV, the difference betveen the two models is dramatic. At number of small ungle pairs. Figures 5(e) and the lower SPEAH energies, statistical distribu 5(f) show a plot of the momentum of the pairs tions with a large number of events were needed to vs. the angle or the pairs for 200 events. Figure 6 shows some distributions of interest distinguish the two models. At Ecm = 30 GeV a coupl? of events will suffice. for calculations of backgrounds to two fcndy mass distributions. Figures 6'a) and 6(b) show the Figures Ma) and Mb) show the momentum dis multiplicity distribution for the models. Figures tribution of charged particles compared to the 6(c) and 6(d) show the effective mass of pairs of scaling form discussed by Blchter in the 197^ Summer Study.3 The phase-space model is much charged pions, and Figures 6(e) and 6(f) show the stepper than the Jet model and is in disagreement effective mass of pairs of photons. For charged vith the scaling form. It is interesting to note particles the Jet model increases the number of 2 *.hat the extrapolated multiplicity and the limited events for small masses (M < 1.7 GeV ) and large transverse momentum, which are the Inputs to the masses (M2 > 'i0 GeV2) compared to the phase- Jet model, automatically lead to the scaling space model. distribution. If the present simple Jet model predicts the actual behavior of particles at L* = 30 TeV, Figures Mc) and Md) show the distribution CB of angles between two charged particles (pairs). then detectors which try to study events in an The distribution for the phase-space model Is exclusive way will need large segmentation, due essentially flat, while the Jet model shows sh"\rp to the increased density of particles near the peaks for angles near 0° and 180°. axis of the Jets. On the other hand. Jets will be Figures Me) and Mr) show the sphericity so easily distinguished that it will be extremely distribution for charged particles. There is interesting to observe if there exists a separate almost no overlap between the two models. The class of non-Jet events. R£FER£NC£S 2. G. Hanson, et al., submitted to Physical 1. 0. Hanson, SLAC Summer Institute Topical Review Letters; SLAC Pub.-lfi55, LBL-li2a7. Conference, July 1975. 3. B. Richter, PEP-lUo, 197^ PEP Summer Study. NIIOKT 01- WE 10LARI2ATIDN GROUP W. I;ord, K. Kondo, I". Martin, II. Manning, U. Miller, C. Prescott 'Ilie physics of intersecting storage rings We have studied tlie use of longitudinal has been thoroughly discussed by V. N. Baier in polarization in the reaction e*e" •* y*u" as a his elegant review.' The mechanism of radiative practical example with high theoretical interest. polarisation and its dependence or- machine charac G. Manning considered modifications of the magnetic teristics is considered in detail, lie reviews a insertion which could reduce synchrotron radiation number of effects which might bi? useful in mea by two or more. In addition, a specific design is surement of single-beam polarizition; in addition, suggested which incorporates the optimized mag he shows that angular correlations for secondaries netic configuration. In this case it is assumed produced in e+e" interactions cm provide a di that no particle detection is necessary near the rect measure of the effective two-beam polarization. interaction vertex and the synchrotron radiation is 'dumped' up- and downstream. W, Ford and The practical utilization of techniques for F. Martin have separately considered vacuum cham measurement of single-beam polarizations at PEP bers in which the synchrotron radiation is absorbed energies was considered in the 1.974 Summer Study. locally so that shielded regions are provided for It was concluded that the measurement of the asym detectors near the interaction vertex. The report of metry in the backward scattering of circularly Ford appears in the Section on Weak Interactions; polarised laser photons could provide a satis the reports of Manning and of Martin appear here. factory fast monitor of single-beam polarization. The critical requirements on detector alignment During the Simmer Study there was discussion and beam stability were discussed in detail. Sev of schemes for rotating the polarization outside eral detectors were suggested which night provide of the experimental areas. One such scheme was satisfactory signal-to-noise in the inevitable worked out after the Summer Study by Al Garren background of synchrotron radiation. and John Kadyk. Their report is included in this section. Most of the synchrotron radiation in During the past year the angular correla their design could be dumped outside the experi tions expected for transversely polarized beams mental areas, greatly simplifying the design of have been observed in a serit-M of elegant experi experiments. ments involving both leptcnic and hadronic final states.2 In SPEAR II polarizations as large as K. Kondo studied the intense local ionization 0.75 are routinely achieved. The observed angular of residual gas in the interaction region due to correlations provide a direct measure of average synchrotron radiation at the insertion. He points two-beam polarization over time intervals which out that appropriate electrode structure could permit statistically significant samples of data. result in effective pumping and better vacuum in The u+u" final state, whose properties are calcu the interaction region. lable from QED, provides a calibration for the two-beam polarization deduced from the more copious It is important at ISR's that potential ex hadronic final states. perimenters have some familiarity with machine operation. To this end we summarize some general The achievement of longitudinal polarization considerations in the production and measurement is especially important in the exploration of of beam polarization. weak-interaction contributions to the e+e" annihil ation process. We reiterate that any system of magnets operating on both the e+ and e" beams can produce only the -*- -*• and -*• •«- states of longitud inal polarization. These are of interest in the I. MECHANISM OF SYNCHROTRON RADIATION study of interactions which do not proceed through the one-photon intermediate state. W. Toner' has In the classical approximation^ 7/8 the considered the resonant depolarization of either total synchrotron radiation is polarized in the or both beams using weak solenoidal fields. With plane of rotation (o); 1/8 is polarized along the either beam depolarized states with net helicity guide field (ir). A = 0 and A = 1 are obtained. Sokolov and Ternov were first to study the The only practical scheme for achieving effect of spin orientation on radiated intensity longitudinal polarization suggested thus far is using exact solutions of the Dirac equation in a that of Richter and Schwitters.4 A symmetric con uniform magnetic field. They showed figuration of magnets is used to Totate transverse a) for longitudinal electron polarization polarization to longitudinal immediately before the radiated intensity is independent of whether the interaction region and rotate it back to its the spin is with or against the direction of motion original state immediately after. Application of b) for polarization along the guide field a the scheme involves severe technical problems small component of the radiated intensity depends since the large magnetic fields required in the on initial and final spin orientations. In particu insertion result in high local intensities of lar synchrotron radiation. field vari;it ions >>n spin mot ion could he stihlied in detail. In this M»ik the r.idi.ituni d:u:iping ten proJiii/i' the spin alignment. Since dumping terms are small it is amven- ient to stihly the spin motion separately. Hie • "TT * start ing |xnnt is tin- equation ot' liarxrrcinii, Michel, .ilkl iVlejiJi."7 W ds Jt .v | (J • a>» It • ad - HH.,1. Here s is the rest-fnum.- polarization and J" e (K*2)/2. '••-'* II.,Jbe ilie guide field on tut? '/losed equilibrium orbit. The spin precession can be The initial electron spin is along [against) the split into three parts field for ; = I(-l). The arrows indicate relative spin orientations in initial and final states. Only the spin-flip term W„ contains; a com -L nic y -0 ponent of radiated intensity dependent on initial spin orientation. At P!:P design energies c taY = :,v (R = 170m, Y = 29,354) the parameter % meT V \ 5 =• 2.936{10-6) so that 12 .cl -III • m) i.H-B0J • ah-I)fi4l W. H . 8.62[10" ) (1.28) {1 ± 0.9884 ) W where JJ, is the precession of the momentum vector The total synchrotron radiation at PliP will be p in the guide fiel:'.; ^, is the precession of 2.6(10 ) W/beam. For initially unpolarized beams the rest-frame polarization with respect to ;i only 28.6 pW of the radiated intensity for either coordinate frame fixed to j>; and u> is the per beam is associated with spin-flip leading to trans turbation, verse polarization. Within this incredibly small component of the synchrotron radiation transitions With 2 along the direction of motion, leading to alignment along (againstj the field for x radially outward, and >' along N™ (for e*) to e*(e") are more probable by a factor 171.4 and the beam gradually becomes polarized. lowest order The polarization buildup time T will be :([(! + 3Y)»r x - a(Y - 1) inversely proportional to the number of"trar.sitions per unit time. With the critical energy 5 Uj. = (3/2) (ficy3/R) as a measure of typical photon energy ^l '- The first two terms represent normal synchrotron motion; Br rotates transverse polarization into longitudinal and with vertical betatron oscilla P "c R2 R2 CR ' RJ tions the component of BQ parallel to the instan taneous motion rotates transverse polarization The detailed calculation by Sokolov andTernov gives into radial. To lowest order the radial betatron oscillations no not depolarize. The third term represents an insert such as a solenoid magnet, 1 2 etc. t - = *J1 ttfc r^r CJ C) X . P 8 R3 Schwitters has developed a general method for solving equations of spin motion and applied For times t » TT the polarization approaches a maximum value of 92.4". it in a study of depolarization effects in several specific storage ring configurations. He concludes there is no net depolarization due to betatron motion providing the spin-precession frequency is II. MACHINE CONSIDERATIONS sufficiently far from interger multiples of the horizontal and vertical betatron tunes. Since the transverse polarization is 'driven' by only an infinitesimal fraction of the total Away from the depolarizing resonances quantum synchrotron radiation, it is important to determine fluctuations in the synchrotron radiation still whether normal machine characteristics might pre lead to a stochastic depolarization. Detailed cal vent the polarization from reaching its theoretical culations show that for normal storage ring pa limit. To do this Baier and collaborators* de rameters the stochastic depolarization time is veloped a formalism within which the effects of long compared to polarization buildup time.1-8 A solenoid couples horizontal and vertical 111. Dirm:ii«s »i I-OLUIIZATIUN oscillations. Consider a plane-parallel beam en tering a region of soleiwidal field B,. The longi A. SinBle-Heam t'olarization tudinal flux within an area irp2 becomes radial flux Bp outside tht- solenoid. Integrating over Baier lias discussed a variety of effects ;i (Russian cylinder .starting outside the solenoid sensitive to lieam polarisation; in principle, any and extending to its center gives of these could he used in its measurement. il J: 1. Tousclick Scattering. Scattering will occur between pans uf particles in a bunch due to rela tive motion. Some fraction of large-angle scat . B /2 . terings will result in loss of particles to the = bunch. Since scattering at large angles is danpt-d for electrons with identical polarization (scat tering at n/J goes to zero in the non-relativistic litis radial field idiie to the return flux.) in linttl the beam lifetime will depend upon the de duces ;m :i:u»uth»l component of momentum p. -o gree of polarization. that particles iniiially parallel to the ;JAis now undergo helical motion in the solenoid with radius ii/1 (i,e. the diameter of the helical trajectory This effect was used in the first measure for each particle is just equal to its original ment of radiative polarization in the VLP!'-2 distance from the solenoid axis.). At the exit of storage ring,* In this case {beam energy 700 MeV) the solenoid a flat beam is rotated through an the relative contribution of polarization-dependent angle terms to beam lifetime was bj. Since the relative contribution decreases with increasing beam energy the Touschek scattering is not useful as an indica tor of beam polarization at PEP energies. 2. Compton Scattering. The cross section The opposite radial flux at the exit reduces the for scattering of circularly polarized photons on remaining Pj, to zoru and the beam particles are polarized electrons may be written again parallel. where t,2 is the degree of circular polarization and C, is the polarization of the electron; $ is the angle between the plane perpendicular to V.A ?, and the plane of scattering. Baier has em phasized that when final state polarizations are -0 unmeasured this is the only correlation allowed by *u: P and T conservation. 3. scatteringoff Polarized Electrons. Scat tering from atomic beams with polarized electrons would yield large asymmetries for measurement of Fig. 1. Longitudinal flux inside solenoid generates single-beam polarisations. At present the intens radial component outside. ities of atomic beams are not sufficient for adequate counting rates. B. Polarization in Beam-Beam Interactions Baier has provided detailed calculations for a number of relevant two (and quasi-two) body final states. Jackson has emphasized, however, that in many cases the angular correlations expected at PEP energies depend only on parity and angular mo mentum conservation for reactions proceeding through a single intermediate photon.9 Fig. 1 End view showing rotation of plane parallel Polarized beams imply well-defined relative beam because of helical motion induced by phases for initial helicity states. With both beams radial fringe field at entrance to solenoid. transversely polarized This coupling of vertical and horizontal l->) C !•> u+> + betatron oscillations is adequately reduced with the addition of compensating coils at each end of the solenoidal insert 50 that net |Bzdz = 0. It 1 ( !->• |- is easy to see that the compensated solenoid also l*o- 1- results in zero net rotation of the polarization to To order y only the |<-> and |-*>(i.e. lowest order. X = ± 1) states couple to the virtual intermedi- .iti- plinton <\. Hie JH.I1.II i.Mt ion *'! Itii phot-.m I Inn i t^ • , S>i .S • . V • • , iIoilH the direct ion of the nuidt- I u-l.i. Consider the reactions inmh ini: iw (••.vwlu •.i.iLir mvsons eV • -V, k*k , h. Ks. .-i. . l iu.it si.itfi with net lie I icm s. I.ili-. hit hi. ' • 1 com r iIn Ilu- Duly lector in the final st.ili i- tin- ivl.it in- ..i ••> tn shut, tJt.it 's = ll I..i inmrientun; the .unpl ititde is then v • p -unl l in.il -.latvs. Consider d,- _' ., it. " c,,!t " Inr .(.itvi with '••, * *,. *', ihi' iMi i'v oiicratnr 111 idiere *' is the angle with respivi h> the livid i •in", A li ' direction. With tvsj>ect to p(| !•, I.M: V'i.*B.o>- Vi,; I I'or complete tninsvcrsc jx>l;iri;:H ion production perjwndicular to the guide field noes to :ero. Consequently, any state for which Jackson has used the hvlicity formalism to generalize to more complicated final states. The "It rotation operators R may he uscvJ to lonn stares ,,1-L of definite J for an arbitrary system of particles with fixed internal configuration. Ihv amplitude is then cannot couple to the intermediate pliotun. I'his is true, for example, when A = n and 11 is anv natural jvirity resonance. In this case onlv states with J l\ x (R.w.K) - 1 U+M D'^RKimi.w.'l S" (K) | '. s> n • : I nrv allowed and JS = I). Mien no conservation law interienes, u„ = it where R represents a rotation of the 'hudy axes', or ii . = 0 may be a consequence^! f well-established \.f\| are initial (C'lall helicities, v|ul is the dynamics, f-'or example, in the 11,11 final state nit initial (filial] helicity along the initial tji-4i predicts that cv * 0 as y"-. In general, how- (.filial) ::-direction; n in the numher of particles ever, the angular correlations j»roi'ide imjwrtant and M. the remaining quantum numbers. new dynamical information. 5. See for example, J. H. Jackson, Classical Rhh'tRGNCliS Kloctrodynamics, Second Uitiou, John Wiley and Sons, Inc., Chapter 11. 1. V. H. Biiicr, Sov. Phys. UspeUii U (1972)095. ii. A. A. Sokolov and I. M. I'ernov, Sov. I'hvs. This comprehensive review article contains Itoklady S (1%4)IJ03. references to much earlier work. 7. V. Hargmann, 1„ Michel, V. Telcgdi, Phys. Rev. 1. Proceedings of the 1975 International Photon- Letters I (.1959) 435. Lepton Conference, Stanford University, to be «. R. I'. Scfiwitters, Nucl-ir Inst, and Methods published. 118 (1974) 531. .i. W. Toner, 1974 PEP Summer Study, p. 39(>. 9. J7"i). Jackson, LIU. Theorv Croup Note. •1. B. Richtcr and R. Schwitters, 197-1 PliP Summer III. M. Jacob and C. C. Wick.'Ann. Phys. (N.Y.) 7 Study, p. 384. (1959). MEASUREMENT OF WKAK INTKUACTION C'ONTRIULTIUNS TO e V - n V USING LONGITUDINAL POLARIZATION OF TI1K e+ AND e" REAMS (i. Manning I. DIFFERENTIAL CROSS SECTION FOR e e~ - jifT FOR POLAH1ZKH REAMS Weinberg - Snlam V-A The cross section for e e — y n~, assuming one phulon exchange and weak neutral current contributions, •"•Vs is given by:' C(4ain28 - l>2 -0,01 "r": 2<* -0. l f«w-t»|(I - P.,P~) (I) :i • V 2 $£» a -fl. 1G -0.03 4 c'(4sin ew - i) - (1 t a,,)sin"tf I\|,P" cua 3^ a -O.lli C(4sin'd - 11 -0.03 5 w + a '1 cos 0 (1 + P*P") + a,(l * cos~tf )(P%-P~) a 0 0 0 a" c 1 cos t> (P Pj- ASYMMETRIES THAT CAN BE MEASURED (0,0J is (he angle of they" with respect io the e" direc A. Charge_Asymmetry or Forward-Backward tion; P~is 'he electron be;im polarization and is defined Asymmetry to be up; P^ is the positron beam polarization and is down; P" and P+ are defined to be positive In the elec tron beam direction. (Note the Schwilters-Hiehter method" of producing - oV(fl) -dffM) longitudinal polarization gives P"£ + P£ = 0. If one ach " d(r(0)*dff(-0> beam is depolarized pt + P^ = * <*. 92.) If only one photon exchange contributes: ^UPzPzla3^VPz>a512c036 are zero. 1 2 |(l+P*P^(l-*i1)+ (P*+Pj)a4J d+eos 0) a2 •- DReftrf fi£) (2) • P*P^i(l+a,)cos 2$ + a6Bin 2 = 0, no T violation, and w-e univer = DRe - DRefe*"^) Ch * ^(1+P^p-Kl+g^D) + (P^P>vgAD \ 1-KJOS" 9 / 1 + PZPZ Ml-KIOS-fi/ 2raM! The last approximate expression Is written to show the dependence upon P . The following points can be If time reversal is not violated all the g's are real and made: a_ =0. If^-e universality is assumed, g^ = g^ and a. Use of longitudinal polarized beams allows mea surement of g e. as well aa g£. Table I gives the values of a. - a- for V-A theory b, p£ + P^ = 0 provides no useful information for and the \Velnber5~Salam theory for s = 900 GeV2, where weak interaclions with V or A, It could be of use in C-— sG/4sT2 isa. looking for S, P, or T terms. -20- e. One beam uopolsriced and the other beam polarized d. In expression^) transverse polarization terms gives almoit as big aa effect as when both beams are only appear as PT^T• whereas longitudinal polariza longitudinally polarized to yield P*+ P~ = 2 x 0,92, eo tion terms appear as P* + P" Hence, one can use spin there is no need to struggle to get the difficult polariza rotations of say 60° rather than 90° to obtain measur tion state. able asymmetries, thus greatly reducing Che synchro tron radiation problems. If one beam is unpolarized the P'pPf = 0. Even if this Is not true transverse polar ization effects are small and can be measured since Asymmetry for Longitudinal Polarization they alone produce a tp dependence. Reversal dq [« d + cos Q) * a„ 2 cos 8 4 ']fc*-;l 2 2 J(|l+P*P~)(l+a1){l-Kfos 0Hu32*;os For P~ = P™ =0, no T violations, andjx-e univer Table II gives values of A .and A, for five com sality, the above formula reduces to: binations oE Pj + P~ for V-A ano Weinberg-Snlam theories. P™ and P™ are assumed zero. Dg g {1+COB 0) (l+Dg^)(l+cos~0>+Dg 2 cos e\ A [II. PRODUCTION OF LONGITUDINAL POLARIZATION A system for production of longitudinal polarization was described by .Schwitters and Richter,2 Figure 1 shotVB the general system. Schwitters and Richter had a solution with Bi = to* aiH* Li = L2* Tnis syatem <*oes not result in the minimum production of synchrotron ra The following points can be made: diation. Given an overall system length (L), a gap (G) between magnets and field free region (R) about the in One cai teraction point, the requirement of minimum synchro- SABv- b. One needs to work at small forward and backward Iron power, toother with the requirements that the ang'.^j in 0. spin is rotated through 90° and thai the beam pass c. The comments made in II. A, b, c, and d apply here. 2 Values of Pz V A WSwith sin *) =1/3 Case K K en AL \h AL 1 .92 .92 0.47 C, 0. 998 C2 0.11 C, 0. 9fc.° C3 Fig. 1—Magnet system to produce longitudinal polariza tion. The system is symmetric about the interaction 2 .92 0 0.44 Cj 0.92 C_ 0.11 ct 0.92 C 3 point. 3 0 0 0.19 Cj 0 0. 08 C. 0 4 .92 -.92 0.19 Cj I? 0. OB C. 0 through the interaction point, defines the 4 parameters, 5 -.92 0 0.01 Cj -C.-2 C2 0.05 C, -0.92 C3 L., L„, B., and Br,. Figures 2 and 3 show how these parameters vary for different values of L2. It will be .2 -2c os 0 -0.16(1 n seen that minimum synch rot ronjjower is when L^^Lj; + 00a e) this was pointed out by Wenzel. Some solutions with 0.84(1 +C08 0} -0.32 coa 6 L2 = 2Lj are tabulated in Table Dl. It can be seen that the power increases rapidly with the required central field free space. It can also be seen that the power 0.99(1 tcos 0)- .16 GO] changes very slowly as the gsip between the magnets is 800 w 600 -iZ - 6 \ • iO - 5 200 - 4 - 2 2 -' Fig. 2—Plots of total synchrotron power, magnetic fields, and magnet lengths to achieve 90° spin rotation Fig. 3—plots of total Eynchroiron power, magnetic for a field free region at the inieraction point of + 0. 7 m fields, and magnet lengths to achieve 90° spin rotation with a gap between the magnets of 0.1 m and an overall for a field free region at the interaction point of ± 2.0 m interaction length of ± 9. 7 m. with a gap between the magnets of 0.1 m and an derail interaction length of ± 9. 7 m. IV. A COMBINED DETECTOR AND LONGITUDINAL jr faFts for L,, 21 POLARIZER ind C(>° spin roi.itiun Figs. 4 and 5 show a detector for e e —-M y~ with the polarizing fields built into the detector. The detec 3(1° Spin Rolallon 50° ^ptn Hot jttmt tor is an iron shell with a 5,4 m outer diameter and a U K ^, u K Ct-wer EC, -*., lliwc f Ev j ECfl 1,4 m inner diameter. The ball has a vertical slice cut m ku Mi HV keV fceV kw kcV keV through its center or width of about 20 cm. Coils are 0 il. 1 4.1 7.14 2. 35 -71 106 33 120 71 23 provided to produce a field with approximately circular symmetry about a vertical axis of the sphere. In the Ll.T l). t 4.^5 «.49 ;t.3-> 37(1 126 49 l&l Hi 33 •1.30 9. IS 3, >«3 425 lafi 1-9 91 11. 1 a.s 12. 11 fi. 115 79H 90 351 120 GO •j.-> ll. 1 a.;i 16.50 119* 142 i32 163 95 Inner Region Field Witfi Dnfi Chombe's •1.59 3.5-1 11.4 us Ifil 32 and Showe» Detectors U.G 4.2 S.67 3.19 3i» 129 t:>9 31 : 1.0 •1.0 M.W6 ;i. ii 352 132 31 ii l.-t 3.(1 9.0M 3.03 34 K 135 45 1SI) 30 (J. 7 2. I> 3.5 9.-10 •J. 33 846 HI 44 91 29 3.(1 :i.(i 111. A» 3.13 lf>K •11! 157 104 increased. The critical energy in magnet 1 increases as G is increased. A value of Rj ^ 0. 7 m and G = 1.0m Fig. 4—Sketch of detector for e+e" —•fi'V" with longi is chosen for a detector described below. The table tudinal polarizer built in. The analyzing field of in kG also Includes values for a spin rotation of 60° - this re in the iron Is spread out in the median plane to give a duces the asymmetries by only 17^ but it reduces the field of a. 86 kG over a distance of 4, 0 m. Separate power by 56% and the critical energies by 33l1e. magnets of 4 m length provide a field of 3.11 bG. -22- 3 m vertical height and 20 cm width. It would have sub stantial side walls. Other solutions are also possible. The momentum resolution for muons would be ±15% approximately Independent of momentum. The iron is about 12 absorption lengths. The solid angle is about 0.95 * 4*. e+e~ — u+u~ can be adequately detected with detec tors outside the iron ball, it would be desirable to All the 1.4 m diameter core with chambers and shower de tectors to attempt to detect e+e" — e+e" as an indepen dent normalization check. The practicality of this will depend upon how effectively the synchrotron radiation can be dumped. The weight of the detector would be <* 630 tons. The magnet power requirement would be a Bmall frac tion of a megawatt. V. CONCLUSIONS 1. p* = ± 0.92 with P = 0 Is almost as useful as P+ + P~ = ± 1,84 and it is not worthwhile struggling to get me difficult polarization state. Flg. 5—Schematic view of half of central magnet system 2. It Is not necessary to use 90° spin rotation; 60 ro showing a possible coil system. tation reduces synchrotron power by 56% and leaves the asymmetries 87% of full effect. The transverse polarization effects are proportional lo P-f P^ and hence are small - or zero when one PT Is zero. median plans the iron is extended along the beam direc The effects can be checked as they are only 0 de tion to make a total length of 9.4 m. The fletd is made pendent terms. to lie fl.BR kGover the length of 4.0 m and -8.86 kG 3. A detector for e+e" — JI+/I" can be designed that in~ over the opposite 4.0 m. The field in the 2m thick iron eludes the polarization rotation magnets within It. ball is about 18 kG. Separate coils arc used for the For 90° spin rotation the synchrotron power is median plane section, the section above the plane, and *, 350 kW and =» 156 kW for 60° rotation. It seems the section below, so that the major part of the muon possible to leave room for the synchrotron radiation analyzing field can be kept al 18 kG. independent of the to be kept in vacuum until it is dumped clear of the polarizing fields. Separnte magnets of length L,=4.0m apparatus. are powered to have fields of ± 3.11 kG. 4. Experiments can be made to measure gv and gA to The slot left through the detector allows the syn about ltft if the contribution of the weak interaction chrotron radiation to be kept in vacuum until it reaches is at the level predicted by V-A theory. I^the a suitably shielded dump. A brute force solution would Welnberg-Salam model is correct with sln"^ = 1/3 be to position these dumps welt downstream of the de it should be possible to measure (L, to an accuracy tector. This would require a vacuum chKtnber of about of 10°. W 2. R. Schwilters rad B. Richler. SPEAR-175; PEP Note 87; or 1074 PEP Summer Study (PEP-137), REFERENCES p. 384. •J. W. A. Wenzel. 1074 PEP Summer Sludy (PEP-137). I, Notes by S. Berman at PEP Summer Study. 1974. p. 38B. A BKAM PIPE FOR POLARIZATION EXPERIMENTS* F. Martin 1. the dissipation of the heat produced from absorbing A vacuum chamber Tor the Riehter/Schwitters-type the radiation; polarizer is shown in Fig. 1. Its complicated shape is designed to absorb "all" uf the synchrotron radiation 2. the high-energy densities of the radiation; coming from the polarizing magnetic fields and cover the 3. leakage of the radiation from Compton scattering in energy rnnge from 5 to 15 GeV/c. The basic problems the high-energy tail; center around the following conditions: 4. access to the interaction region for tbe physics. *l U- 0,30 m Fig. l—polarized beam vacuum chamber over the full length of the interaction region. The polarizer occupies the full 20 meters with a The shape of the beam pipe follows the trajectory of tbe two-meter clear area across the interaction point. The 5-GeV beam. The symmetry allows for reversing the parameters are tabulated in the table below. direction of polarization in both beams but one might argue that that is an unnecessary complication. This system has not been optimized to minimize en ergy power but shows what may be a typical worst case. ENERGY DENSITY The vacuum chamber Is designed to absorb all of The radiation is spread out over a length of 175 cm, the radiation close to the Interaction region and inside which Is equivalent to 1.11 kW/cm. Since the radiation the experimental apparatus. At 15 GeV, 93% of the ra is strongly cotllmated along the beam direction, the en* diation falls on the region a-b In the forward direction. ergy densities may run as much as five or six kW per The other 1% passes through the Interaction region and cm2. Consequently, the type and thickneBB of the Initial is lost at the other end of the vacuum pipe and on down Impact surface will require some attention. For now it the storage ring. Around the interaction point is a 30- is assumed that a thin sheet of water-cooled etainleeB cm sphere which sees no direct radiation. steel will be adequate. * Work supported by U.S. Energy Research and Development Administration. Interoction Region Detail 1/2 Scale- Elevotion Synch'otron Radiation Passing Through Fig. 3—Elevation view of the vacuum chamber showing the thin window around the Interaction region. The principal weakness of this system with respect to radiation absorption Is the Compton scattering from the edges of the beam pipe near the interaction point. In an attempt to break up this secondary scattering, two small masks have been placed at a and b, as shown in Fig. 3. An upper estimate of the amount of secondary scattering can be made. Assume the radiation that strikes In a 2 cm zone back from the end of the shield is toe main source of radiation. Then toe radiation in tercepted by the mask is about 2.3% of l. 95 kW. The amount of solid angle covered by this scattering surface is about 25° out of 360°, or 3.5% of 4ir. Therefore, the geometric attenuation reduces the power leaking out to about 0.16 kW per beam. This quantity is further at Fig. 2—Water-cooled radiation absorber inside the tenuated by the mask and vacuum window. Note also vacuum chamber, 41. 73 g/cmr thick. that radiation scattered through relatively large angles is trapped inside the beam pipe. COOLING As is shown in Fig. 2, It is proposed to dissipate SIZE OF THE CHAMBER the heat with circulating water. The total heat in sec tion a-b is roughly 195 k\V; then the volume of water to The vacuum chamber is constructed of stainless remove this heat (assuming 100% collection efficiency) steel both to provide rigidity and to withstand the impact is: of the radiation. The thick walls, as Indicated in the figures, are needed to withstand the atmospheric pres d v/dt/AT = power/c •• 195 kW/4.184 sure. Two crude estimates on the thickness are com 3 3 puted assuming a drum head 1.40 m in diameter and a 47 x io cm /sec/°C 1 cm wide beam 140 cm long with fixed ends. In the where c is the specific heat of water (4.184 J/era ) and former case, the thickness Is 1,25 cm and in the latter, iT is the temperature rise. 1.77 cm, for a maximum deflection of 0. l cm. For our purposes, a 2 cm thick wall is proposed. The total weight of the beam pipe and shield is five metric tons. RADIATION ABSORBER The 1974 PEP Summer Study contained a section on Interaction Region Detail backgrounds 120/11.35 x 0,15 = 1.56 cm . In the second figure, a scheme is shown for a water-and-lead absorber inside the vacuum tank. The "thin" outer wall is needed in order to minimize the en ergy deposition on the outer suriace of the absorber. As shown, the total material, including the outer wall, amounts to 41. 73 g/cmr, and, with an incident angle of 2 Fig. 4—Plan view of the vat-uum chamber with the thin 0.15, this is equivalent to 278 g/cm . window around the interaction region. INTERACTION REGION VISIBILITY Figs. 4 and 5 are views of the beam pipe at the In teraction regit,.:• The three angles shown in Fig, 5 are the half-polar angles for particles coming from e+e" in teractions. Fig. 5—Vertical cross section through the interaction region, showing the outer thin window. The angles, 0, are the clearance angles for particles produced in the interaction. Fift- 5 A SYSTEM FOR OBTAINING LONGITUDINAL BEAM POLARIZATION AT PEP WTTH VERTICAL DIPOLES LOCATED OUTSIDE OF THE INTERACTION REGION A. Carren and J. Kadyk I. INTRODUCTION the IR, Just the region where background should be minimized to assure efficient detector perform Recent experiments at SPEAR anzonatrate that ance. circulating e+ and e~ beams become highly polar It 1B thus important to search for ways to ized along the field direction, to a degree con rotate the electron spins that leave more space sistent with the theoretical maximum, 92.kJ,2i3 for the experimental detectors and reduce syn This property of the beams allows use of new and chrotron radiation in their vicinity. Wenzel was powerful techniques with which to study the weak the first to consider this question. In a pre and electromagnetic interactions. Particularly vious note he discusses two schemes, each using interesting applications of the spin polarization tvo dipoleB, one or both of which were placed become possible if the spins are rotated to the . beyond the IB." The focussing doublets at either longitudinal direction at the interaction point. end of the IR bend both the central orbit and the To maintain full polarization it is necessary dispersion to cross the median plane at the IP. to keep the particles' spin vertical in the The quadrupoles must have very large apertures circular arcs of the machine, rotate it to the (about one meter) and be considerably longer than longitudinal direction at the interaction point the present low-B quadrupoles. Moreover, their and then rotate it back to the original vertical chromatic aberrations result in substantial loss direction in the following arc. Thi3 can be done of luminosity. by utilizing the g-2 precesBion of the electrons In this note we propose a spin rotation in vertical bending magnetB, or in a combination of system (SRS) in which the central orbit i3 bent solenoidal and horizontal bending magnets.5i°" In by vertical dipoles located outside of the Inter our opinion, the most feasible method proposed to action region, and follows the path shown in date for use at PEP is that of Schwitters and Fig. 2. The spin is rotated as before by the Richter.T This method, shown in Fig. 1, uses (g-2) precession in these magnets. Some import ; features of the Bystem are the following: 1 1) There is no reduction in luminosity resulting from adding the SRS. 2) Somewhat less (about 30? less) synchrotron power is radiated than in the , Jt >v Schwltters-Riehter scheme. 3) Low-field vertical bending magnets and masks are included in the SRS ^ * In order to shield the detector from the synchro L tron radiation, at least in part, h) Selection 1 |if f—»-H of either the SRS mode or the conventional oper -.- ' „1 '" * '; ating mode is achieved simply by choice of al ternative settings of magnet currents. 5) The IR is left completely unencumbered for disposition Of the experimental apparatus. Fig. 1. The Schvittere-Rienter Scheme for Produc ing Longitudinal Beam Polarization (PEP-87). The diagram shows the four vertical bending II. DESCRIPTION OP THE SYSTEM magnets (unshaded rectangles) inside the 20 m interaction region. They are used to produce A side view of the spin rotation system (SRS) longitudinal polarization at the interaction is shown in Fig. 2, where the standard PEP lattice point, IP, as indicated. Spin directions are magnets are shown as open rectangles, and those indicated at various points along the of the SRS by shaded/black rectangles for vertical trajectory. dipoles./quadrupoles. Figure 2 showB only half of the system to the left of the IP, the right half being produced by inversion through the IP. four vertical bending magnets in the interaction The beam coming from the left enters the SRS region (IR). The spin rotates relative to the at the vertical dipole BV1, the lattice before momentum by 90° for every 2.305 Tesla-meters of this element being unmodified. There the beam 1B net bending, independent of beam energy. The four deflected upwards 58 rarad, and the following magnets have virtually no effect on the beam elements are centered on the vertically displaced throughout the ring. The principal change is to central orbit. Vertical dipole BV2 deflects the introduce vertical dispersion in the IR, except beam back to the horizontal direction. BV3 and at the interaction point (IP), where it vanishes. the two low field magnets, BVL, bend the beam Hcwever, the magnets fill most of the IR, leaving downwards by '(6 mrad to cross the median plane only a small space (about 3m) for the experimental at the IP. These five dipoles, having a net /Bdl apparatus, and the fields are high (0.82 T of 2.305 Tm, rotate the polarization of the beams compared to 0.3 T for the horizontal bending from the vertical direction in the circular arcs magnets). Consequently, a great deal of synchro to the longitudinal direction at the IP. The tron radiation is emitted (~600 kW) locally within corresponding dipoles on the opposite side of XBI. 7512-M44 Sthoiunc SKM Vm, of Intviion with in Ruution Sjittm Added 2. Schematic Side View of Insertion with Spin Rotation System. nuiRncts of the basic PEP lattice are shown as open rectangles; these are not to be replaced or moved. The new elements, belonging to the spin rotation system, are shown as black (qundrupoles) or shaded (vertical bending magnets). Small arrows indicate spin polarization directions at several points along the beam trajectory. i * QA as U a n H n B n a "'Vr B Bl-fn9Lfn Bv1 1 • • ak^ • BU19yL BVL ai a - i> al md narticil bwn tnt(tops and di XB1. 7512-9845 3. Horizontal and Vertical Beam Envelopes and Dispersions. Beam envelopes are plotted for the horizontal (upper solid line) and vertical (lower solid line) planes, corresponding to 10a (10 times the expected rms width). Also shown arc the dispersed rays, horizontal and vertical, cor responding to energy spread AE/E = ID, the lOo^ value for 15 GeV beam energy. The envelope width as shown includes the dispersion, combined quadratically with the betatron beam width. the IP rotate the spin back to the original verti will also he constrained to be everywhere un cal direction. The vertical low-field magnets, changed from that of the standard configuration. BVL, each deflect the beam 1.6 mrad downwards. The betatron-function and dispersion values are Together with appropriate masks, they are intended given in Table I. to shield the detector system from the radiation emitted in BV3. This is similar to the use of the horizontal low-field magnets BL to shield the IR from radiation emitted in the main lattice bending Table I. Required Values of the Betatron Function magnets. at the Interaction Point And at the Center of QD, The four quadrupoles, QA, QB, QC, QE, lie the First Normal Cell Quadrupole, between the vertical dipoles, and the doublet, Q2*, Q31, is placed above the corresponding GD Center Ii> doublet Q2, Q3 of the standard configuration. Horizontal . Vertical Horizontal Vertical These six new quadrupoles are all centered on the 6 (m) displaced beam line. 12.16 3li.l2 3.30 0.20 a O.OltO -0.052 0 0 III. Beam Optics n (m) 1.238 0 -0.7k 0 A. Requirements on Vertical Dispersion dn/ds 0.0025 0 0 arbitrary In order to maintain full polarization, the spins must be vertical in the circular arcs of the storage ring. In order to maintain this The matching problem was solved hy allowing condition, the particle spins must precess through the strengths of the 12 quadrupoles, QF3 through equal and opposite angles on the two 3ides of the Q3' inclusive, to be varied. The number of IP. Since the precession angle is proportional variables may seem unnecessarily large for the to the rotation of the momentum vector, it follows seven constraints involved, but. this many variables that in order to keep the spins vertical in the were needed to resolve the conflicting require arcs, the dispersion function, nv(s)t oust be an ments of beam and dispersion focussing discussed odd function of s about the IP, If, in addition, above. Even so, the problem is not trivial, and 0 nv(s) is identically zero throughout the rest of two fitting programs were employed to solve it.^i* the ring, systematic spin oscillations outside of The resulting quadrupole strengths and other the SRS will be eliminated. The vertical dis parameters of the system are given in Table II. persion nv(s) has these properties in the method Beam envelopes and dispersion functions cor of Schwitters and Richter, shown in Fig. 1, due responding to ten times the rui.s widths a are to the absence of quadrupole lenses in their shown in Fig. 3. system. In fact, nv(s) in that case is Identical, but reversed in sign, from the vertical displace ment of the central orbit. Unfortunately, this automatic matching of Tabl< It. Kignet the vertical dispersion does not occur in the system presented here, because of the action of the lenses Q2' and Q3'- In the absence of the elements QA, QB, QC and QE, a nearly parallel beam would be focussed at the IP, vertically. However, the dispersed (off-momentum) rays would be parallel when entering BV3, and therefore not parallel after leaving BV3 and entering Q21 . Thus, the dispersed rays would not be focussed at the IP. The four new lenses, QA-QE, are added to focus the dispersed ray as well as the beam envelope at the IP. In addition, the strengths of the other quadrupoles between the normal lattice (left of QD) and the IP are altered from the values of the standard configur ation. B. Beam Hatching and IP Conditions The linear beam properties are determined by the betatron functions and the dispersion. In order to obtain the full design luminosity, and minimize the effects of the SRS on the operation •f the ring, the betatron functions and dispersion at QD and at the IP were constrained to be the same as in the standard configuration (both hori zontally and vertically). Only the slope of the vertical dispersion, nv1 = driv/ds, was allowed to be different (non-aero), but this ia expected to have a negligible effect. Therefore to the left of QD, in the normal lattice, the beam envelope -29- IV. LGuISTICS OF INSTALLATI' U AM) OFFIlA'i IOH system. These, however, must be so designed that third and higher order resonances are not excited. A. Operating Modes A one-fold periodicity is also introduced by energy I033 in the vertical dipoles. Thi3 effect Since the technique of spin rotation has never could be removed with compensating rf cavities been used in a storage ring, it is inevitable that located in the SRS, between BV2 and BV3. fuj-l operation will only occur after a period of testing and development. Therefore, it is de B. Chromatlcity sirable to have operation of the SHS be optional, arid to be able to sviteh over easily to the con Not only will the ring have one-fold peri ventional (non-SRS) operation. A rather natural odicity, but the chromatidties are increased by way to accomplish this is to install two sets of large gradients and beta function values in t l:v-S quadrupole doublets, Q2'-Q3 , along the some quadrupoles of the polarization insertion. vertically displaced beam line as part of the SRS, The maximum beta value is increased from 500 m and Q3-^3 along the conventional beam line as in the standard configuration to SOOm. Again, • part of the standard configuration. This will the consequence of these increases are probably reqiire also two separate beam pipes, which will managable. diverge at BV1 (at 5o" mi-ad) and merge again at the IF (at lib nrad;. With this arrangement, the entire set cf new magnets can be turned on or off C. Apertures to permit or^ration either in the spin rotation mode or in the conventional way. Beam sizes are rather large in some magnets of the SRS. Care will be needed in the design, As presently designed, quadruples Qy, Q3 are fabrication, and testing of these elements to somewhat too high to give sufficient clearance be sure they meet the somewhat tighter tolerance for the new set Q2', Q3' to be directly above, requirements, as shown in Fig. 2. Therefore these elements may have to be staggered with 0?, (J* or designed with D. Synchrotron Radiation smaller vertical dimensions. The new low-field magnets introduced in the 3. Use of the SRS at Different Energies present design should permit the shielding rf the IR from synehroton radiation emitted from Tim design presented here is intended to the vertical bending magnets BV1-EV3. However, produce longitudinal polarisation at 1^ GeV. a detailed study of this question will need to Since the condition for 90° of spin rotation, be made to determine where the shielding masks 2.305 T-m of vertical bending, is independent of should be placed, and how effective such shield energy, a full polarization capability requires ing will be. that the system be movable, with the height of the BV2-BV3 line varying inversely with energy, the V. CONCLUSIONS net /EJdl being fixed. Although occasional changes in beam line of this nature may be This is a "first look" at this particular feasible, it is obviously not desirable to make configuration, not an optimized design. It is in frequent changes. A possible compromise is to tended rather as a demonstration of the feasibil change /Bdl proportionally with beam energy, E, ity of a system in which the large spin so that the beam optica remains invariant. Then rotation magnets are located outside the IR. The the spin rotation is not precisely 90°, and the system can, in principle, be turned on or off to net longitudinal polarisation will be * 0.92k allow conventional operation of PEP or operation cin{™E/ilE ), where E is the design energy for 0 a with the SRS. Operation over a wide energy 95° rotation. Since polarization is never range also appears possible without moving mag complete skyway, one might tolerate values as low nets, with only a relatively small loss in longi as = 0,75 (as long as it is known), allowing E tudinal polarization. The IR is left unen to deviate by from E . For example, this ±^0% 0 cumbered by the SRS, and the sources of synchro would allow beams to be operated through the SRS tron radiation are moved out of the IR. However, over the energy range from about K = fi to lfl GeV a careful study is necessary to determine the without a beam line alteration. best design for reducing or eliminating this background radiation. IV. PRACTICAL CONSIDERATIONS Although we see no fundamental problems REFERENCES associated with use of the SRS, some important considerations come to mind that will need to be 1. Aaimuthal Asymmetry in Inclusive Jladron examined in detail. Production by e+e™ Annihilation, R. F. Schwitters, et al., SLAC Publication-1629 A. Periodicity (to be published, Phys. Rev. Letters, Nov. 17, 1975). J. G. Learned et al., The insertion containing the SRS has a dif (submitted for publication) ferent structure from the others. Since it is 2. R. F. Schwitters, Nucl. Instruments and only perfectly matched to the rest of the ring Methods, 117, 331 (1971*). for the central momentum, the ring has one-fold 3. J. D. Jackson, On Understanding Spin-Flip periodicity and linear stop-bands exist at half- Synchrotron Radiation and the Transverse integral intervals of the tunes vXt \>y for the Polarization of Electrons In Storage Rings, off-momentum orbits. These may be car-rented with LBL-l»232 (Aug. 1975), submitted to Rev. Hod. suitable adjustments in the sextupole correction Phys. k. PEP-178, 1975 FKP Summer Study Proceedings. 7. R. F. Schvitters and B. Rlchter, A Method for Report of the Weak Interactions/EH Final Producing Longitudinal Beam Polarization at States Group, PEP-157. 19T1* PEP Smawr Study PEP, SPEAR-175, PEP-87. Proceeding!, p. 263. 8. W. A. Wenzel, Note on Longitudinal Bean 5. R. F. Schvitten, Comments on Obtaining Polarization, PEP-168. Longitudinally Polarized Beans in e+e~ 9. TRAHSPORT, an Ion Optic Program, LBL version, Storage Rings, PEP ITote 75' Arthur C. Paul, LBL-2697. 6. N, Christ, F. J. K. Farley* and H. C. 10. MAGIC, A Computer Code for Design Studies of Herevard, On the Feasibility of Colliding Insertions and Storage Rings, A. S. King, Beam Experiments with Longitudinally M. J. Lee, W. tf. Lee, SLAC-103. Polarized Electrons. Columbia University report CO-22T1-16 SUMMARY OF THE WEAK INTERACTIONS GROUP •\. Hciu!Vi>iiuLl, H. Ford, D. Hitlia, D. Kondo, C. Manning Mome, T. Rhoadcs, A. Sessoms, L. Stevenson, K. Strauch. A. Zallo We have studied ways cc Isolate effects due to If the beams are unpolarized, the forward-backward a weak neutral current in e e~ collisions at PEP (charge) asymmetry becomes energies. We consider below, and in the subgroup - dateWo (n-fl)_ reports that follow, the reactions e e -*v u~, = 6 do(6)+do (ir-8) e e "e e t and e e -*• hadrons. There is, of course, a substantial overlap between out work and that of the polarization study group, and of the 1974 PEP Summer Study. Our principal conclusions are: (1) 1+ai 1 + cosz with longitudinally polarized e (e ) beams a pro gram of measurements to determine cleanly the weak °3 1+ Z Interaction coupling constants g , g , g , g Is Z Eeasible; (2) subject to model-dependent interpre ' D(s) Re gfg^ilfr tation, the single-hadron inclusive processes pro vide the possibility of observing S,P,T type weak couplings, and an alternative to longitudinally If the beam polarization is rotated through polarized e e beams for the observation of parity- some angle toward the beam direction, by the method violating V, A couplings. of Schwitters and Richter1, we get P * - P (and a reduction of P~) so that all terms in1 the 2cross I. ISOLATING WEAK EFFECTS section formula are suppressed. The interesting situation occurs when one of the beams is depolar A- e e -*u u ized2. Then P +P~* ± P depending on which beam is depolarized, z and 2 l-p P3"p P •<•• for this In his summer study report G. Manning has re z T T 2 corded the expression for the cross section for case the charge asymmetry is ' a + P a e e -m u with arbitrary polarization of the elec 3 , 5 2z l 7 trons: *8 V l+ai+P au 1+z w 1+c s2e (a + P a >(2z/l+z2). of"tl {(u+aiXl+fV)^ «t »'l < ° > 3 5 2 If we record data for both P and - - ( + (a3 U+P*P~)+a5 (P*+P~)) 2 ens 6 j with aL » a2 - DCs) Re g^ gy D B (2) »3 - D(s) «° 4"< • < >»•4 4hT& We can also consider a measurement of the change ai* - D a6 * D(s) Im g 2 ai^d+z ) + as 22 " U+ai) U+*z> + a3 2* 'z z • («i + a52z/(l+z )) P2 (See Manning's report for coordinate definitions, Folding the angular distribution about 90° gives and for numerical values of a *s.). The information provided by this measurement is re dundant with (l)-(J) provided JJ Is not < = DCs) He ** gj PB. (3) g to (5) are summarized in Table I. Numerical differ ences between the model predictions for the magni A test of time-reversal invariance is made by tudes differ slightly from those in Manning's report isolating the a& sin 2$ term: because of the neglect of small (^162) terms In the denominators in the derivation of equations (l)-<4). . _do J large value. The counting rate for ee+up under the a6 P~, (1-2 ) sin 2$ standard assumption3 is v200 per day, or 1.2 x 10u n U+ai>U+z';)-(l+a2) P£(l-zz)cos 2$ for a two-month run. This Is sufficient for measure ments of \l% precision (for rnsults of an analysis of the statistical errors, see Ref. 4). r,/ v r e* i.,.-. (1-z2) sin 2$ ,., A background for the asymmetry and rate measure 3 - D(S> Im 8A *VPT W-PjU-.2) co. 2* .(4) ments is the contribution from order n QED diagrams. This subject is discussed in the report by R. Horse in this summer study. It is found that the asymmetry A harder experiment is the rate as a function of s: Is very nearly independent of s and of $ (even far P-l* 0), and depends upon the energy resolution •g-. The weak and QED contributions to A (*=6 In Horse's «•••>= *[£<•> •£<">] notation) are shown in Fig. 1. The large variation 4s (l+ai)U+22) of A with 41 in the transversely polarized case allows its separation from Ag independent of a • fg" (1+D(s) Re g* 6j) U+s2 quantitative calculation of A^E&. For example, if we form the difference AA(A ) between asymmetries = (1 + D(B) Re 8 in the regions cos 2% > 0 and cos 20 < 0, we find R(B) r f^ sj* ^- IWt.de 0f Effect He». at ISCeV. Parameter. Normal- Weinberg Measured Det'd Polarization f z.tt l.m "V-A" -.08 ^V -•"It* i+r A.. CO) e, 1-? *> - "« SA"S» —• JV - -.06 2 u 2.2i\-7. )7. 2.i(l-Z->Z "" (l+Z2)-2VU-zV 11+7. ) -.2911-7. > n ( "J * Re 8A BA !„„,„« - V aw • -.01 -.025 S-="' l+Z"1 vv a5 '• «• >'"*! LanR.(1 Bean) ..... aw - -.11 aw - -.02 W Lang.il Bean) Bunch-palt -.02S ^Lsym \ - «= «;'•? -" Rate (H> •• • «• «;*«j - 5 -.16 -.01 A Transverse + . -4 0 0 »6 " '" 8»"< , .. n .,.-> Biinch-pnii fev .*.-) »°j * «"•»*«; m o+zr av - -.04 av • -.006 All but the IJSt 1 * Analyzing power increase! slightly with transverse polarization 7 Aaym. ulth P* - +.92U Is twice this size with B. - .£ (0- - I (ara2>) B2 - 1 + ai-S2 2 z '"his quantity is also plotted in Fig. 1(b). Its B3 = (H-2) / tion is clearly useful for this experiment. B5 - 0 Note chat QED does not contribute to the parity- violating effects observed by inverting the longi tudinal polarization- Bt - -2a3 (l+z)/U-t). 10" 0 0.5 cosS Fig. l + - + - ft e -*e c From a talk at thla Sumner study by R. Budny we have the expression for Che fihabhacros s section: H-PV)(I,+(WV)(I-<) Win)'ij Here time-reversal imparlance has been assumed 2 (hence Bs-0). The n 's are defined analogously to P£>" sin 8 K cos 2^+B5 sin 2*) those for ee+uu: 2 2 • (a) unpoia"«d Beotns lb) P/'0.924, P,"=-0.924 - bs^ (ci Pj=0.9Z4. PJ= 0.924 . (dJ P; = 0.924. PT=-0.924 (el p; = 0.924. P}, Pf.O ii t Pj =-0.924. P". Pf-0 - Fig. 3 (a) - r 1 0 1 where D(s) was defined in section A. The pure QED eoi 9 part of the cross section is plotted in Fig. 2 for various beam polarization conditions. Fig. 3 gives Fig. 3 (b) the weak contributions relative to the QED. Al though effects of ^LQSC in the angular distribution formation regarding the coupling constants that can are predicted for some cases, these would be hard be extracted. Essentially it amounts to a complete to observe against the rapidly falling dominant description of the l'ptonic weak neutral currents, distribution. within the context of V- and A- type couplings. One measurement is feasible with the Bhabhas. The change in rate, integrated over the angular C. Inclusive Madron 1c Processes acceptance of the detector, as the polarization P++P -(P or P~=0) is reversed (curve e - curve f The report in this summer study by D. Hit^in in F?g. 5): z and A. Sessoms discusses prospects for Isolating weak effects in single hadron inclusive experiments. (7) Particularly striking is the observation that S, P, ?«>-?v{^ and T type weak couplings give rise to terms linear in P_, in contrast ylth V, A types which show only The analyzing power is given by the integral of tl-e angular factor weighted by the angular distribution; correlations with P-P^' Simple tesrs are therefore ve estimate ^r for |cos B\ <-65- This is a factor possible with a general hadron detector and the of three smaller than the analogous measurement with natural beam polarization. u w~. The event rate is, however, larger by a fac Also considered and found feasible by Hitlin tor of ^3. This experiment has also been listed in and Sessoms is an experiment to measure A° polariza Table I. tions, in lieu of longitudinal beam polarizations, Collecting the data from the orogram of experi to isolate parity-violation- ments listed in Table I we show in Table II the in- Table 1. II. DETECTORS Beam Polar zation Trans ve rse(zero) Longitudinal The summer study reports by tf. Ford and G. Manning discuss arrangements for studying ee+uu and ee with longitudinally polarized beams. Both e u a5 8? detectors are of the "iron ball" type. In the first a 8 a free space of 4 meters is provided between the 3 * A h "3 " *A polarization rotating magnets to allow flexibility to interchange detectors. An iron ball detector c s ) which fits into this tegion is described in some v 4 v detail. The scheme of Manning optimizes the syn 33 " & chrotron radiation by providing only a short (1.4 meter) free space for the central detectors. The •r 4 polarization rotation magnets are extensions of the a Im 8 6 " A *V iron ball, which then can be made thicker to improve *> < punch-through rejection and momentum resolution. -35- The cost is this loss of ability to vary the detector REFERENCES configuration, or to run the analyzing field inde pendently of the polarization rotator (eg., tn uBe 1. R, Schwitters and D. Richter, FEi--B7 ('74PEPSS). transversely polarized beams). Detailed studies of 2. H. Toner, I'LJ'-ie** {'74 PEPSS). the backgrounds due to the synchrotron radiation 3. B. Richter, i'EP-140 ('74 PEPSS), are needed Lo decide uhich approach should be U, P. Wanderer, et ul., PEP-15? <*74 PEPSS), Fig. 4. adopted. [RON BALI., MARK II W. T. Ford The detector described hers is intended to gye,u and gAe,M by 'he method outlined in record events of the reactions e+e~ •*• \i u and Refs. 1 and 2. For some of the measurements the e*e" in sufficient detail to permit extraction e+ or e" beam is longitudinally polarized. of the leptonic weak neutral current parameters ^r:r^\ji=:^^^^^ 1— BSC ' Bum HJV clv.ir ^11 Fig. 1. Plan and elevation views of the experimental area showing the particle trajectories, the location of the polarization rotation magnets and the location of the detector. The muon-clectron detector is shown in Fig. 2. sheets which in turn are glued to the panels. The A toroidal magnet provides both particle identifi wire orientation is across the narrow dimensions cation and momentum analysis for the muons (the of chamber, along constant-z contours (z being the magnetic field of 16 - 18 kG is contained entirely beam axis), to provide good resolution in 9. Half within the iron). The cylindrical central cavity the wire planes are rotnted by 15" to give small- contains tracking chambers for muons and electrons, angle stereo determination of the azimuth. Outside and lead scintillator shower counters to identify the spark chambers are liquid scintillation the electrons over that part of the solid angle counters for triggering and tuiie-of-flight mea that is free of shielding. The tracking devices surements. may be either drift or wire spark chambers, but must provide reasonably good resolution in the The acceptance for muons extends over 2ir in polar angle, since the bending of the magnet is in azimuth and to 15° minimum pul:ir angle, J=. 97$ the 9 direction. Trigger scintillators (not of 4ir. The electron acceptance extends over 2TF shown in the figures) are also to be provided. in azimuth and 50° < 9 < 150°, or i* 4ir. The momentum resolution (for muons) is Ap/p z ,21, Surrounding the magnet are 8 "orange-peal" Limited by multiple scattering. wire spark clumbers, each of two gaps and covering the region :-. 4 7.5 GeV 0.002 10' 1. Weak Interactions Study Group summary in this volume. It should be pointed out that no reliable 2. G. Manning, "Measurement of Weak Interaction estimates of the background levels induced by the Contributions to e+e" •+ u+U" Using Longitudinal synchrotron radiation have been made for this con Polarization of the e+ and e~ Reams", report figuration. It is possible the only safe procedure in this volume. would be to extend the vacuum pipe [and bending 3. R. Schwitters $ B. Richter, PPE-87 ('74 SS magnets) vertically to the maximum envelope of the p. 384). synchrotron radiation, allowing it to pass un 4. G. Barbiellini, N1H 123, 125 (1975). EM-WEAK INTERFERENCE IN e e +u i SCATTERING R. Morse The basic one photon cross-section for e e annihila tion is 2 (1 + cos* 9) - PTPl sin 9 cos 2$ If we allow for the weak neutral current process + + e v *V With restriction that polarization lie in x-z plane, % * <£ (( + (a3(s)(l+pV)+a5(s)(P*+pp) 2 cos 6) v £§ & . (1 + a,B) f° + 2 a3 a cos 6 - (1 + S ) f° W aj " D(s) Re gj gj) where 6 • 3jS + — (2a3 a cos B), where the coefficienta aiand a3 are given in Table 1. a2 - DCs) Re gjj g" The higher order (a3) electromagnetic correctlona z a^jF 2*o also interfere with the baBic one photon process. The effect can be absorbed into 6™ (e,4) defined a, - DCs) Re gf g» s as a modification of the basic single-photon rate. SiroM' e &8 do a - DC.) In 8**8^ «» do• ,- CI + o" + J™) f - <1 + «T) f - fT 6 do i) Rates and Aayametries (a) The »flte ia proportional to R where s - 4» j a cc» e (l + s* (e)) (l + co«2 e) R - f CI + «T) f° d cosed* the Modification of the basic rates js deter o " Note that f is symetric about e - n/2 so that mined entirely by o C6) we can write R - 2 J d cos ell + 6T Cfl)J f° d* where Cb) The ••yiaietry ACe.,) ** Riven by the usuel fomula 0TC8) - \j («T<6) + 0T0l-8)). AC8,«) tV.) - f^-e.e) _ < >•«> fT<8,,) + fTC«-B,») 1 + 5*(8,« We show later that 6 C6) has no * dependence so 1 that the aodlfied rate is slaply where 0* (6,T) • {6 C«.*) - IM,tH Note that the asymmetry depends on the rate. TAS asymmetry due to the WEAK IncerEerence is Aw(0,*) (l+al9) XL) Calculation of the Weak Interfen 1< 5 - a,s and is SII-JWII in Mn. J for s = 900 CeV2 and loo s 1 GeV2,for P*=P~= 0-921 and the V-A interaction. 2a_s The effect of polarization on the $-distribu- tions is quite striking. However, since moat experiments in this region are event limited we look at the it-integrated asymmetry Recall Chat f = 1 + cos28 - IP^Jsin^e cos 2$ so chat at if = 45° there are no polarization W W effects, and at $ = 0° the effect is maximal A (0)= i~ f A (B,$) d* It is also shown in Fig. 2. •• EM ond Weak Corrections lo /I I Photon Exchange Process / 1 1 -20 - me) \ \ r -30 \ \ /- -40 i i J 0 30 60 90 6 (degrees) Pig- 2- Hi) Calculation of the higher order {a3) EM Inter ference Brown et al.*have shown that - for transverse 30 60 polarisation - that e? (degrees) 5™„»" Bl<9> +7^ <«2<«+«3(l,,lP&lco,W Fig. l. ) g_(0) and g. (0) are anti-symmetric with (b) the $ dependence of f compensates the $ respect to E)'(H-H) SO they do not contribute dependence in the E2<6) + g3(6) |PT+P~1 the rate. Further there is no net polariza cos 24 tern tion contribution to the rate because nf the cos 2 eTe ^iTu. EM corrections S|M, S|M EM corrections S|M, S|M S--IOO GeV2 s = 900 GeV2 fclerends E=5.0 •§' O.I E = I5.0 ^=0.01 Brown E = 6.5 «§ = O.OI E=I5.0 f =0.1 Brown E=5.0 1-= 0.1 E=I5.0 <=f=0.2 Brown E=5.0«f =0.2 ~T ~r 'T ^-- 1 1" \ 10 - - o Sv, HO - : 1 1 30 60 50 60 9 (degrees) 8 (degrees) Fig. i.. Fig. 3- than 5jp[. It may be that the Brown et al. it is also statistically weak in that differ formalism has been pushed into the "hard- ences are used. This requires lots of data • photon" region for AE/E=»10-20?£. In Fig. 3 it may be extravagant. and 4 we present Sjp1 and ^ for the values of s and AE/E shown. We also compare with "hard-photon" calculation of Berends et al.2 The agreement is not good - though oJ;n Is e e —-/i/i better than 6™. The behavior remains the Expeded ^-Asymmetry same at E=15 GeV. It seems to indicate that a | - AE/E-102 is a tolerable level. p-- = p = IOO% Polarization (Up + Down>-{ Right + Left) A^10) Conclusion 1.0 1 1 / The grand asymmetry \p+ W4 / the $-integrated asymmetry is ahown in Fig. 2 \ for the values shown. Since the aa EM correc \ tions are nearly ^-independent one can define / AA(6) which is(nearly) independent of the / a3 EM correction (to the degree that they are 4>-independent). 0.5 A (B,« it - 4 '/ / V6> % " i given that the higher order symmetric effects 30 60 90 are small. n(B) which equals A.(6)/2na3 s is 9{ degrees) shown In Fig. 5. The technique*has the strength that the EM corrections need not be well known - Pig. 5. References Brown et al.. Phys.Lett. «B, 403 (1973J S, Berraan, Lecture given at 1974 PEP Summer Berenda et al.. Nuclear Physics B63, 381 (1973) Study (unpublished). WEAK-ELECTROMAGNETIC INTERFERENCE EFFECTS IN e+e% HADRONS D. Hit 1 in - A.. Sessoms Dass and Ross have investigated nossib^e interference terms in e e -+ h(p) + X, where p is the hadron momentum. The effective Lagranglan for A program is described which can ascertain the the neutral weak coupling of the electron field, ¥, complete space-time structure of neutral weak cur to hadrons, has the form rents by studying correlations of inclusive hadron distributions, and polarization,wtth beam polarization. : y J. + f Y,T J, + T Y. x J, V v.v r J, The possibility of observing weak-electromag '>;f netic interference effects is one of the most excit ing experimental prospects of PEP. The simplest approach to the measurement of such effects is in the pure QED reaction e e"-*u u~. Weinberg-Salam The interference terms, and their properties under type models of the weak interaction make predictions Parity, P, and charge conjugation, C, are shown in for the magnitude of such effects, as does "stand Table 1, which is adapted from Dass and Ross. ard" V-A theory. Since neglecting muon form factors and derivative couplings appears justified, it is concluded that a "model independent" analysis Table 1 of the data from such an experiment can be made. In Kinematic structure of the weak electromagnetic order to isolate weak effects, howevet, it is neces interference terms for the process e e -*• hX for the sary either to measure the polarization of the out currents J],...,^. The symbols p, p+.p_»s+1s are going 15 Cev rauoiis, or to produce longitudinal polar the hadron momentum, the momentum of the e and e ization in at least one electron beam. Both of beams, and the transverse polarization of the e these are at best difficult tasks. It is the purpose and e~ beams respectively; 2K-p_-p.t.. Th© last two of this note to point out that it may be practical columns give the behavior under parity inversion to see unambiguous effects of weak-electromagnetic (F) and for charge conjugation because, in tin.- limit m /m •+ 0, S, P and T inter + I i -- p'Kp- "i s_-lp_s ) ference terms are all linearly dependent on the 4 4 transverse polarization of the beams, while V and l\ p-Kl'-<»(«+*sJ) terms arc either independent of transverse pnl.iriza- tion or quadratjcally dependent on. it. Sini-o K, I' p-Kp-(s+-sJ and T couple to e e of the same helicity tln-re ran p'Kp- (K>. (li_K +ti s_J) be no interference ti-rms dependent solely an tlui + + longitudinal polarization. Thus a term involving p- z expected effect at the 3o level. As a corollary, Monte Carlo, Jet Model, P *0.47 a=0.82 ( if no effect is seen, extremely good limits on STP • Dota coupling can be derived. These limits should be 2 substantially better than those achievable in muon- I" eV —/i*/i" P :O.47i0.05 iess neutrino interactions in the forseeable future. As opposed to the V, A case, should a correlation linear in the transverse polarization be qeen, it is not possible to confuse the effect with inter ference with a second order electromagnetic process since those contributions Linearly dependent on transverse polarization are suppressed by ^t'TX /tip ^ 10 relative to weak-first order elec tromagnetic terms at PEP energies. If this correlation is seen, we have then to separate SP couplings from T. This 1B possible on the basis of angular distributions and their be havior when the e or the e beam is transversely polarized. For example, at 0=TT/2 the first four terms in J;,, which are proportional to p*K, vanish. The remaining terras are Identical to the SP correla tions, except that they do not change sign when the ObsL-rved azimuthal asymmetry vs. x at sn5"> e Instead of the e beam is transversely polarized, fVV- tt SPKAR II for h.idr.m evanta, with while the .Ij, J2 terms do change sign. It should J i i'roni'.s. |cos DJi O.f;. be noted that the setting of limits on SP or T neu tral current coupling does not require either longi Indus! v in spectra provide a large nuoiln tudinal polarization of the c e beams or measure of events furr those correlation studies. In what ment of final state helicity, since It is not neces follows we ::\il l assume a luminosity F. Lobkowicz, U. Becker, K. Berkelman, , A. Green, E. Groves, K. Halboch, J. Kadyk, N, MlBtry, A. SeBsoms, M. Strovink I. INTRODUCTION presently used at SPEAH, or its improved version Mark II.1 (2) Highly specialized detectors which Experience from both SPEAH and DORIS has shown can identify uniquely one particular reaction and that in order to be best able to do physics at PEP, analyze it completely. As example, we might as complete information as possible should be ob mention here the iron ball^ and its variouB pro tained for each event. There is a fundamental posed modifications for investigating e+e~ •*• 2u difference between proton storage rings such as or the crystal ball3 for studying the reaction the CERH ISH and ee storage rings such as PEP, e+e- •*• (nv). Such detectors are designed to be which necessitate a different approach for any ideally suited for one particular reaction, but detectors used. At hadronic machines, the inter not easily modified for other reactions. action rates are large (105-106/ sec), and thus (3) Semi-specialized detectors which lie in between the complete analysis of every event is clearly the two previous categories. Such a detector would impossible. Thus, great care has to be taken in have the ability to identify with high reliability the choice of an appropriate trigger to avoid some of the reaction products, but still be versa being swemped by unwanted interaction events. On tile enough to obtain most of the information con the other hand, inclusive or semi-inclusive data tained in the residual reaction products. As may contain as much physical information as some examples, let us mention here the study of the exclusive channels which make up only a small reaction e+e- -+ e+e_ + hudronn (2y processes),^ fraction of the total cross section. Since the investigation of events where at least one large statistical samples are in principle availa particle has a high momentum,5 or the study of the ble, backgrounds can be investigated by modifying new particles.^ the initial trigger requirements. A general user nagnel (GUM) has r.o place in At -'KP the expected interaction rates of < 1 the first two types of ietectors. Thus ttie ques event/minute are sufficiently small that, were it tion which we net out *•-> Investigate is whether it possible to design a perfect detsctor able te is possible to design a. magnet system versatile completely identify nil reaction products of every enough so ^s to be suitably for moat, if not all event, only one such detector would have to be .;emi-specializc i 'Ji-;'.e-jti>r£;, us rpquit'i-d by ii T— built. The necessity far having severni Inter t'erent experiments. ::\x -h 'letecu>rc will have many section regions at ?i-.V, each with its own dif features of a general letti'tor. Tn n vny r-ach GUM ferent er.periment.il setup, is based on the fact i-onfiguration could be '.'unsidered u goneril de that suL-h :i perfect lietector cannot be built. tector where the compromises have been shifted very The muss identification of fast particles via far in optimizing its features for one particular their Cerenkov radiation interferes with the reaction. requirement of detecting neutral products such as photons. I'ions and kaons decay in flight while their ^jotr«ntuBv is being measured, thus, creating II. PHYSICS FOR SEMI-SPECIALIZED PETECT.Jfls; identification difficulties between hadrons and muons. Therefore, the emphasis for any PEP de Given the extremely fast pace with which tector, once a. reasonable annihilation trigger physics is currently progressing nt electron- has been established, lies in identifying unam positron colliding rint;s, it is difficult to pre biguously the reaction of interest from all the dict which of the outstanding questions will have data obtained. been answered by the time PEP starts doing physics, inu how many new questions will have appeared by Vie feel that it is necessary to remind the then, nevertheless, it ia useful to consider which reader of tho3e ruther obvious facts, in order to particular experiments would require a versatile enable him to appreciate the framework within and easily adaptable magnet system. which we have undertaken to study the possibility of designing a general user magnet (GUM) suitable A. Search for Hew Particles 6 for a reasonably large set of different experiments. Basically one can distinguish three types of With the discovery of ii, %* and their decays detectors suitable for PEP. (l) General detectors into new narrow states, it geemB likely that soon which try to achieve a reasonable compromise be a host of new charmed particles will be found. tween ill detection requirements. They typically One can assume that by 1900, when Vt-V will start require a solid angle of nearly Itir, some particle its operations, one has begun I.J study their identification, nn 1 luitfc neutrul and charge de- decay modes. In u q'j'irk-model eaUm'tte, the in teitior.. With sui-h detectors, one can study all clusive charmed particle cross section should be possible reactions, hut may have difficulties i- 'i-20? of the total cross section, comparable to dentlfyiriK specific reactions which form only a o(e+e~ -* u+u~)- Unless the leptouic decay modes small part of the total cross section -- exactly are abnormally low, and if the charmed particles because being a compromise solution they may be are produced in two Jets, the best signature for a unable to uniquely identify reaction products. A charmed particle would be a high energy lepton prime example of such a detector would be Mark I, coming from the de?r_r of its opposite charm partner. Thus, a detector is required which has ag&iii that there is no need for a highly re a n/e rejection of < i0~^ uni/or r/ii rejection strictive trigger. Indeed, any trigger whii'h of the same order of" magnitude, but is wise capable rejects most of the beam-eas interactions:, L-osinic of measuring momenta and, i i" possible, masses of rays and 3tnall angle Bliabha events is sufficient. the hadrons '.rom the decay of the eharmed particle. The apparatus haB to be designed only to allow the selection of the desired interactions from the B. 2Y Processes Tlie reaction e+e~ -* e+e- + hiidrons, corres III. MUM OH GUM ponding to the diagram: Once it has been decided that a versatile magnet facility should be built for FEP, und before it car, he designed in detail, it is neces sary to discutrs how such a facility would be pri marily used. •' aere axe basically two possible extreme modes of operation, with a continuous range in between. (a) In a true multi-user magnet (MUM), the basic facility would be gradually upgraded by indi vidual experimenters who would add their own equip ment. Each new experiment would be added to the previously existing setup, until a whole set of yields the only available information on the experiments is being performed at once on the same process y + y -*• hadrons. As long as the scattered facility. electron of energy E* leaves under a small ai^le (b) A general user magnet (GUM) would consist 6, the (massj^ of the photon of a magnet with a flexible detector inside. This would he modified by each experimental user for 2 q" ; EQE'e his own particular use, without being constrained by any special equipment or modification used in a is quite small and one thus has a colliding photon previous experiment. ThU3, only one experiment facility. On the other hand, since the virtual _ could be performed at any time, except if two ex photon angular distribution falls off only = 1/0 , periments complement each other so that neither q2 values up to 10 (GeV/c)2 are available vith has to compromise its performance. reasonable counting rates. There would he an obvious advantage if the The tvo virtual photons in the interaction sane facility could be used by many -oeriments have typically very different momenta, therefore at once. Given the low interaction rii-es expected the cm, system of the produced hadrons also is at PEP, one cannot afford to throw away any events. moving with large velocities. Any detector cap A possible HUM setup could be arranged by several able of detecting ty processes has to positively experiments, dividing among themselves the avail identify the tvo electrons, and measure their ener able solid angle on the periphery, while all would gies. In addition,it has to have the capability of be using the same central detector, nevertheless, analyzing the product hadxons. Since the hadrons when discussing particular experiments, it is are frequently produced very forward, good mo quickly discovered that the requirements of one mentum resolution at small polar angles is par experiment can only be satisfied by compromising ticularly important. on the other, A facility run in the extreme MUM mode would quickly develop into a general de C. High Momentum Hadr"n Events tector in which the necessary compromises are more dictated by historical accidents One of the weaknesses m' a.V presently con (sequence of experiments installed) instead of by templated general detectors is t^ieir lack of mass physics. To cii.e an example, an experiment de identification for medium to high momentum signed to identify high-momentum particles re hadrons; this is usually given up in order to en quires large Cerenkov counters near the interaction able the detection of the photons coming from nD, region. Another experiment, e.g., one detecting n°, n' , etc, decay. In addit?- to detecting both charged and neutral particles, would want to and identifying high momentum ^urons, it is also install Y-detectors Just outside the magnet. If desirable to be able to study associated particle part of the solid angle is covered by Cerenkov multiplicities, momentum distributions, Jet counters anri part by shower counters, it is very structures, etc. Thus, it is insufficient to likely that neither of the two experiments will be only meauure the high momentum hu'lrt'ii itself; it an unqualified success. is necessary to be able, insofar 'is possible> to Thus, we have concentrated our design con reconstruct the whole event. sideration on a GUM. We try for flexibility in He have given here only a few typical ex- the sense that the basic setup could be easily j.iDles for use of a semi—specialised detector modified for a particular experiment. We do not which would require a magnetic field. Many other attempt to design a basic detector which would be examples could he given. They all have in common suitable for many experiments at the same time. tb« characteristic that one particular feature It should be realized, however, that implicit in serves to uniquely identify the reaction of inter tl* GUM idea ia the problem that experiments plan est with high reliability. Once this feature has ned for the same GUM will perform sequentially and been identified, one wonts to obtain as much ad not simultaneously. This implies that the special ditional information al'out the event as physical physics which forms the basis of the apparatus or financial 'onsiderations allow. We streas design may be partially explored by some general detector group in advance of the GUM proup. It is assumed that those features of the HUM system which would be utilized by ail the user groups would be set up and maintained by a set of A more stringent requirement is given if the goal GUM support people. The systems which would is to investigate the very high rierti'-iituiii region require this support include the U - H/E0 > O.M). Tygu-alljr for tli inUy.nited operation of the superconducting magnet-, the luminosity/ X.it = i0-";i I'm"**-, and n^^'uninc operation and maintenance of the _-enUvtl drift i! * 6, one obtains t'nv a typical experiment chamber array, and finally the software required for the online computer system, the ti-ickfinding n = ibO events, x. > 0/1 and momentum recons-trust ion in the central, detector, and general bookkeeiin^. While we do net anti or about 1% of the total annihilation events. If cipate that these support functions need require it is important to measure the momentum of these a large number of people, a minimum support effort, events to better than 5%, then: such 2S instituted in many high energy physics laboratories, has proven to be essential for ef ficient operation of large facilities. IV. REQUIREMENTS F03 GUM This resolution cannot be achieve.! in our design In specifyin: g the design criteria of a without providing position measurements outside magnet, we have considered typical experiments the magnetic field, to better determin.. the di- such as described in See. II. Our requirement on ivition at the magnet exit. any magnet parameter is given by the most stringent At low momenta, the resolution is limited by reasonable demand of any of the experiments con multiple scattering. The magnet should be suf sidered. While it is impossible to predict exact ficiently strong, and any coil material inter ly all demands^ we have tried to anticipate any posed sufficiently thin, to keep the momentum other future use of the system. resolution Ap/p due to this effeet below 1%, even in the presence of the unavoidable minimum ma terial (beam tub- walls, chambers) of a few per A. Solid An^le cent of a radiation length. As for any detectors to be used at PEP, it is In order to clarify any further discussion, imperative that GUM cover as much of the solid we write down the resolution achievable in a angle as possible. Any coil structure should magnetic field, given the rm3 position resolution, either be sufficiently thin so as not to degrade ax perpendicular to the magnetic field, Az in the charged particle or photon momentum resolution or magnetic field direction, and an evenly distri identification, due to Interactions in the coil, buted amount of x/x . radiation length: or the coil should cover a sufficiently small solid angle so as not to seriously jeopardize the £| = "yi" , 4* physics goals. Any steel necessary for magnetic P 0 due 0 flux return should he either far away or small enough so as not to cover more than about 10? of Unm A particular requirement is posed by any 2Y due to multiple experiment which is interested in large q2 events. scattering Since q" = EoE'0^, and typically the scattered electron energy E' is of the order 1/2 of the beam energy E0, one requires for E0 = 15 GeV an opening angle sin 6 cos 9 =- (3) where S is the distance traversed perpendicular to the magnetic field B, R = z tan 0 and 6 is the angle between the particle momentum and the di which for qmax = 9 GeV corresponds to 0max = 300 mrad * 20°. One wants the magnet to be completely rection of the magnetic field B. 3 is measured in tesla and all distances in meters. If, after the open for 0 < 6 < 9mfUt to permit an external tag ging setup. particle has left the magnetic field, the outgoing direction iB measured over another distance S, then Eq. (l) should be replaced by° B, Momentum Resolution Scaling up flora SPEAR energies, at 15 GeV an ^l£_Ax_£inJ < <1') average charged multiplicity of nch> = T-5 is expected, and assuming that the charged energy is one-half of the total energy, the average mo We accept a different positional resolution Az mentum of a particle would he about 2 GeV/c, How along the magnetic field, because drift chambers ever, on the average, about one particle/event will measure position accurately only in one direction. have a momentum p > 5 GeV.5 If these particle mo Whether one uses a delay-line readout, a small- menta are to be measured to "5% or better, a reason angle stereo arrangement, or Induced pulce readout able requirement, it is necessary to have an in on strips. Az is always much bigger than trinsic momentum resolution at high mementa of ax. Actually, the error Az usually yields a negligible contribution to fip/p. It is important, < 15 kfi. In 'iMition, if supercanduotiiy; coils however, to the UL-i-ufiey of p„ = p cos !;: are used, tyie j P'IV. r.oRnetlo field limits the r-.axi- miiE current density in the coil. This yields a 1 init on nafin-ti.- field of 15-30 kO, depending on —2. = Y(^y + (42.- tlie confifruriitIon used. If high current density P Vp 1 1 I 1 1 1 If in Addition one wants to limit Ap/p < 1.5? duo Para metes to multiple scattering, one obtains for x/xQ -- 0.03 U : ], 6 = 90°; 360 ~ *<> = 200 urn - B 15 Teste 1.1 t-i ' 11 kO-m . 60 ^^ n?o "p od<*> ' ._ V / - Clearly a minimum magnetic field length of the Track length - 80 order of 60-Bo en is required of any system pro viding the necessary momentum resoluti'- To this 280 Ay *A/ has to be added the beam i Ipe, coil th_^.ness, etc. ' Therefore, the magnetic detector will have a mini mum diameter cf about 3 m. 240 •- V. THE MAGNET 200 -• In this section we discuss various options of magnet configuration, coll manufacture and yoke design. One particular optr'on, a lumped super 160 - conducting solenoid magnet, is investigated in detail. 120 - A. Magnetic Field Configuration There are in principle three basic magnetic 80 o*2>-— - field configurations possible (Figs. 2-5): *-\ (l) A purely transverse field, (2) A toroiaal field, and (3) A solenoidal field. Of course, 40 various split-field type combinations can be built up from these. In a purely transverse field, either the beam p*?e has to be shielded from the 1 1 1 1 .. 1 1 field, or special provisions are necessary to reduce the synchrotron radiation background. In either case, material of thickness of the order of one radiation length has to be introduced around XBI. 7512-9846 the beam pipe. The resulting interactions of Pig. 1. Mags resolution for the iji(3.l), from hadrons, electrons and photons would be intoler able to most, experiments. Toroidal magnetic measurement of it(3.l) •+ y+M~ vs. momentum of fields huve been successfully used at storage t|i, for various z resolutions, Az. rings before (DASP, AD0HE). Other configurations hove been proposed (Octopus, Orange ) but not as yet built. Most of these designs have the ad C. Size and Magnetic Field vantage of vor.5 (< lS fieV1) momentum resolution, but suffer from various defects: As can be seen from Eqs. (l) and {?), the required momentum resolution yields lower limits on both BS^ and BS. S is the minimum distance at {1) In split coil designs, whether they which external detectors of appreciable mass, h.-ive iron ycki-r (IM;T, Orange * or use superconduct such as y detectors, calorimeters or Cerenkov ing coils (Oi'iojnis), less thai ?/3 of the total counters can "be installed. The size and cost of noJ id anale ir accessible for momentum measurement. equipment increases roughly proportional to the In addition, l;»rpe forces occur between the coils area which has to be covered; thus, keeping the of the only proposed superconducting solution. covered solid angle fixed, the cost increases (2) In a continuous coil design (b) Pig. 2. The DASP Double-Arm Spectrometer, as used at T)ESY. Supci conducting coil (t » 13 a W A) Cfyoilai -~- \.J-K ASupport ttructure ai en* XhL7512 9H47 XBL 7512-9848 Fig. 3. The proposed Toroidal Field Super conducting Detector Magnet, "Octopus." Fig. \. Proposed detector, "ORAIKIE." Fig. 5. Proposed continuous coil desicri toroidal field (from PEP-l^O). -52- measured. This would be toierible if the super conducting filaments could be imbedded in alumi num. With the usual copper matrix, the required thickness of -1 radiation length would be too large. Since up to the present there are no com mercially available superconducting wires using an aluminum matrix, a major technological ad vance would be required. Here, also, there are severe mechanical problems associated with the large forces, which could affect superconductor performance. (3) Any toroidal symmetry of the mag netic field necessitates very accurate tracking in the polar angle 6. The most precise tracking devices in use today are drift chambers or streamer chambers. It would be very difficult to adapt a streamer chamber .0 auch a field con figuration, while drift chambers for precise measurement of 9 are ouch more expensive to build than their counterparts that measure $. HISO, more material has to be introduced into the measurement region, since the wires cannot be supported from the end. •) T#I*icoped view. For these reasons we have chosen a solenoidal field for further consideration. This field con figuration has the only disadvantage that compen sating coils of opposite polarity have to be in troduced into the same straight section. We assume that it will be possible to install these compensating coils sufficiently far away so as not to conflict with the experimental setup. Figure 6 shows the Mark I detector at SPEAR as an example of the solenoiflal field detector. B. The Coil A solenoid coil interposes material between the inner magnetic region and outer region where external detectors would be added. Consider a solenoid of radius R = 1 m, magnetic field B = 1.5 T, and length L *= h m. The linear current density in the coil is then j = B/uo = 1.2 x 10 A/m . b) End view. If one wants to limit the power consumption to less than 2 MW a conventional copper coil would XBL 7511-8979 be 30 cm = 21 radiation length thick. Even using aluminum, one cannot build such a coil of Fig. 6. Mark I detector at SPEAR. less than 3 radiation length thickness. On the other hand, recent advances of high current densi ty superconducting materials enable one to build large cryogenic solenoids with moderate {10-20 kG) fields and small (< 1 radiation length) i|= 0.56 sin Q {% GeV"1) thickness. We have therefore narrowed our design study down to a superconducting design of the P following size: for |cos e| < O.89U or 26.6° < Q < 153.V*. Beyond this angular range, the intrinsic resolution worsens rapidly, because less track length 1B Field B = 15 kG available in the radial direction. Figure 7 Radius R = 1 m shows the ideal momentum resolution of such H Length L = U m solenoid, ignoring multiple scattering and the uncertainty in z. In calculating the values given The inside of the 30lenold would be filled with a in the figure, we have assumed that 15 cm 1B lost tracking detector. If we are conservative and on the inside near the beam pipe, and the outer assume a resolution Ax = r&$ = 0.2 mm and further most measured point is at 95 cm radius. The allow that 20 cm of the total radius is lost be actual resolution at high momenta will be somewhat cause of the beam pipe and trigger counters, one worse. If drift chambers are used, the resolution obtains from Eq. (J.) worsens in discrete steps, because with decreasing coil hag X/XQ radiation length, one has to add quadratically at & - 90°: k.2% Jxjx~ p 80 HS O The multiple scattering error will increase as l/^sin 8 at smaller angles, while the magnetic field resolution improves proportionally to sin 6. Thus, for /x/x0 - 0.5 the multiple scattering error dominates the resolution already at ''5°. As described in more detail in Appendix A, it seems possible to build a continuous superconduct ing solenoid of 1 m inner radius, having an over all coll + vacuum thickness of 10-12 cm and a total material thickness of 0.1»5 radiation length. Such a coil thickness is perfectly acceptable for measuring photon energies using a detector out side the coil (e.g., lead glass or liquid argon). As is shown in Fig. 8, pratically all photons of Ey > 150 MeV can be detected outside such a coil XIL71I1M49 Fig. 7. Momentum resolution vs. angle of particle for detector under study (see text). 0.4 X0 angle fewer chambers are available. Furthermore, 2.0 X Eq. (1) assumes that the track is measured over — 0 many points to a corridor of width Ax.* If only 1 1 1 a few points are available, the resolution will o.e a _ be worse. If a continuous device such as a streamer chamber is used, systematic distortions y could also worsen the resolution. To account^Tor b _ these effects we derate the resolution by 20S 0.6 and claim an intrinsic resolution at high momentc r- c of Ap/p2 = 0.TJ& GeV-1, scaling it for other angles as shown in Fig. 7- m 0.4 The resolution at small angles 6 is par < ticularly important for a 2y experiment, where CD the cm. motion of the produced hadrons is ap O - preciable. Furthermore, yy scattering is likely £0.2 to be largely diffractive (via VDM); thus, many of the hadrons will be produced with 0 < 30°. -1-~ I Since the proposed magnet (compare next section) 0,05 0.I 0.5 I.O 3.0 will have an opening core of 19.3D, particles emerging at smaller angles can be measured by ex PHOTON ENERGY (GeV) ternal tagging magnets. Fig. 8. Probability for photons to convert in a Our resolution, while good enough for most continuous coil vs. photon energy, with (a,b, experiments) misses the requirement for measuring c) or withaut (d,e,f) electrons (at least one the very high momentum (x > 0.9) particles. Thus g 2 MeV) coming out of coil, for three coil drift chambers have to be added outside of the thicknesses (radiation lengths) (from Lira coil to measure their momenta. Using F.q. (l1), Galtieri). one has at 6 = 90°: Ap . 3.15% at 15 GeV. 0.3 -^ (b) This resolution is significantly worsened if ap preciable coil material is in the path. If the *s"--« 0.2 N> ^""^ "Corridor" means here sry^Z~^^ t> the path of finite width /ys O.I within which the /// particle trajectory lies. As used here, the width I |~~T::;;;^^ is dominated by systematic uncertainties surveying errors. 0.05 O.I 0.5 l.O 2 PHOTON ENERGY (GeV) If only three equally spaced drift chambers are Fig. 9. Bms error in photon energy contributed used, the resolution given by Eq. (l) is worsened by energy loss of electron pairs from kiinwn by a factor /Jf? = 1.22, Y conversions (from Linn Galtieri). and given an energy resolution for the y-detector of o • 8S vtT even the fluctuations in energy loss give & negligible contribution above thlB energy.* It is also possible to build a lumped coil consisting of six individual superconducting coils In a common cryostat.+ As detailed in Appendix A, each individual coil would have a cross section of 9 x 13 em and present a thick (-2.7 radiation length) material for any particle going through it. However, the material between the coils could be kept very thin: if no special measures are taken, the vacuum vessel walla would represent 0.05 radiation length of aluminum. One could, at some additional expense, provide thin windows between the coils, thus reducing the material to < 0.02 radiation length. The individual coils are separated and sup ported by eight struts which totally cover an azimuthal angle of A$/2ir » 0.1. They constitute - 1.5 radiation length of material and thus reduce the solid angle used for big external drift chambers for momentum determination of very high momentum particles. We have studied the question whether the magnetic field inhomogeneity of a lumped coil system would make momentum reconstruction dif ficult. Figure 10 shows a magnetic field map for the lumped and continuous coils. One can use as a figure a merit two integrals of the magnetic field along a straight line leaving the inter action point at an angle 6. (See Fig. 11) The first integral i^/fS-Sl ds=J Bj_ gives the bending due to the magnetic field in the if-direction for a particle at an angle S. The second integral = /"BJ. s ds = I" [u x B| s ds = -^jp- is a measure of the distance the particle has been bent away from the plane formed by the beam axis and its direction when leaving the target. Note that the "effective length" of the magnet is given fey Leff = V1! As described in PEP- 20^ , even the :r/e rejection does not appreciably suffer if the particles traverse less than one radiation length of material. There has been no study made by this group of the optinum number of lumped coils. In Fig. U we show the two integrals as functions (cf. Figure 7) end is the same for n lumped of 6. The integrals go out either to a radius or a continuous coil. Lumping the coils produces R = 0.95 m or to a distance along the axis ZJMJX the three wiggles in Fig. 12 (solid line), which = 1.90 m, whichever is less. Figure 12 shows the are absent in the continuous coil (dashed lines). ratio of these integrals betveen the real magnet Outside the coil the field falls off rapidly and an ideal solenoid of equal size vhich has (cf. Fig. 13 4 discussion in Appendix B, Sec. 't.ll. everywhere inside a nonstani magnetic field in the Forty cm beyond the outer vacuum vessel, the field z direction. The falloff of ratios for 6 < 25° is is less than 100 G everywhere except near the due to the large opening in the return yoke return yoke. It is difficult at present to decide whether to choose a lumped or a continuous coil if a single GUM is to be built. The lumped coil re I'll 1 1 1 1 1 1 quires less sophistication in the art of fabricat ing superconducting magnets! however, if the pre sently planned tests at L8L are successful, it may well turn out that a continuous coil is cheaper than a lumped coil. If a continuous supercon : 1/ ducting coil, e.g., the Mark II detector, is if chosen for a general detector, then there should 1 be at least one GUM of the lumped type to maxi mally complement the physics which can be done with both facilities. If only one GUM is to be _ // (f«tmfV \ built and developments at SPEAR show that the _..^jl.- ultimate in resolution at very high momenta, or in Ey, 1B not required, a continuous coil should "7 V / be represented. One should note that as far as the inner detector and the steel yoke are con 1* zmJ> " 'We"1*) cerned, the two coil systems are practically interchangeable. C. Performance of Exterior Devices Aside from the measurements obtained from Fig. 11. Integral I1 = / Bj^de and I2 = / Bj_sae, the inner chambers, certain other general measure- giving respectively the net bending and displacement of the particle trajectory in $ direction, vs. particle direction, 6. (a) at coil (b) between coils 21 I 0.5i 0.1 Fig. 12, Ratio of integral I1 or Ig for the lumped coil {"Retl") to that for a uniform field solenoid ("Ideal"). The dashed line XBL75H-S078 shows the same integrals for a continuous Fig. 13. Magnetic field strength vs. radius, solenoid of same length and radluF. showing field decrease outside the coils. merit capabilities will be needed by moat experi ments. Among these are: (a) charged particle identification (up to large x), (b) y-ray measure ita!«i SS; ST* •it,-! m«h u ...>w ./c'l ment (direction and energy), (c) particle (charged and neutral) correlations. Charged particle i- dentification will generally be accomplished using a set of Cerenkov counters, plus the use of time- "" 1ST™' of-flight for low momentum particles. Gamma-ray measurement will require a neutral detection scheme, which crust lie outside the coil If inner ,D i I 1 1: chambers are used for charged particle tracking. "f.s"a"r:s;i"'" Particle correlations will require the neutral detector of (b) and tracking chambers, plus suf s ficient solid angle and resolution to achieve the desired correlation measurements. It is quite i=r "• S3 likely that experimenters will want all three options, but at present (a) and (b) tend to be O.JG mutually exclusive, for practical reasons. The presence of coil material and limi tations of solid angle and space for measuring devices may significantly compromise these goals. The lumped coll and the continuous coil designs pose different problems. The relative advantages of the two coil designs depend upon the choice of the detection system located external to the •"=fr™ :S5 coils. The measurement difficulties encountered with either design result from interactions in ;i-.;I..D•s^xr"'""", -1* •1% the coil, causing a loss of energy resolution, or from failure of the particles to reach the ex ternal detection system. For example, inter actions of y's ln tne continuous coil solenoid will cause measurement of y's to deteriorate, especially at low and moderate energies, while charged particle interactions in the coil will result in loss of some particleB before the i iuy <4ltlu identification system. The lumped coil design sight inprov ~ leaves a large part of the available solid angle MPS*? virtually free of material (- O.03 radiation lengths), but is essentially opaque to particles in the small part obstructed by coils and struts. (We assume in both cases that the y and Cerenkov counters lie outside the coil* since the interior region is fully occupied by tracking chambers.) Table I exhibits some of these effects in detail. The assumptions made in Table 1 should be care fully noted. The relative importance of these effects will probably require a Monte Carlo analysis for each apparatus design. It should be noted, however, that within the limits of the unoertaintitiea in the magnet design, the loss of charged particles due to absorption in the lumped coi-ls is roughly equal to the loss to the identi fication system, due to interaction in a con tinuous coil. In addition to the effects noted in the Table, a loss of particle identification will also occur when a converting y and a charged particle enter the sane cell of the Cerenkov counter Bystem. Figure it gives estimates for the probability of this occurence vs. the number of cells for a continuous coil. (We have assumed that D. The Steel Yoke (a) The steel yoke was designed with the follow ing considerations in mind {describing one half): (i) An opening cone of 20° half-angle should be provided to enable a 2y experiment to measure externally the electron momenta as well as the momenta of any hadrons emitted in the forward cones. (ii) The field lnhoraogeneities should be sufficiently small so as not to impair the momentum resolution for 20° < 6 < 30°. (iii) As little outside interference ^ as possible with the experimental setup for 30D Coil < 6 < 90° is essential. (9x13 cm) (iv) Minimize any steel saturation to maintain field uniformity as much as possible in the central region. (v) The same "jke should be usable for either a continuous or - lumped coil. LP. Clearly some of these requirements, in par ticular (i), (ii) and (iv), work against each other, and thus some reasonable compromise must be reached. XBL 7511-9328 Flux return yokes 20 30 50 No. of cells Fig. 15. Probability that no cell has two charged particles vs. number of cells, for various XBL 7512-9853 total multiplicities, N. Fig. 16. (a) Lumped coil magnet: longitudinal cross section, •0.15 r.l., for example^ in nigh momentum hadron (b) Lumped coil magnet: transverse report. cross section. The yoke is recease.l Just beyond the coil to also reduce the maximum angle available for tag provide more space for equipment outside the coil. ging, nevertheless, one could imagine an experi The large opening of TO cm radius does not in ment with reasonably homogeneous field down to troduce a significant degradation of magnetic e = 15°, while still permitting a parasitic tag field in the region of interest. In Fig. 17 ve ging experiment to operate in the region: 15 mrad show the radial dependence of 5 at 10 cm away from < 6 < 150 mrad. the yoke. In the region of interest 70 cm < r < 95 cm, corresponding to 20.2° < 6 < 26.6°, the E. The Compensating Solenoids^ axial field is near its maximum value. The radial field is mainly affected by the saturation in the Whether a lumped or a continuous coil design steel yoke, which saturates most around r ™ 70 cm is chosen for GUM, compensating coils have to be (tip A) and r = 110 cm (coll radius, beginning of provided to satisfy the condition recess B in return yoke). If it proves to be necessary to reduce Br at large radii, one could achieve this by moving the beginning of the recesB / B dz = 0 . (V-l) B to larger radii. The yoke has octagonal symmetry, with 8 veins This condition la necessary to preserve the de for flux return. The return veins start at a coupling of the vertical and horizontal oscill radius of 3 ra, leaving 1.8 m free space outside ations. However, any solenoid is also a focuB- the coil. The veinB cover an azimuthal angle $ sing device, and thus will contribute a tune * 1.77 radians, leaving slightly more than 0.7 of shift iv: the solid angle free for additional Cerenkov counters. A peak magnetic field inside the iron of 19 k0 is reached at the inner edge of the end )S1&J~ pole piecesi everywhere else the field is limited to less than 18 kG. If greater field homogeneity near the end region is required for Borne particu where B is the field in the compensator, p is the lar experiments, it is easy to insert a steel end- beam momentum, and 8 is the betatron function in cap down to smaller radii. Of couree, this would the straight section. For the main detector, the tune shift is small, because the betatron function has a minimum at the interaction point. However, in order to permit electron tagging for 2Y proces ses down to Qmall angles, one wants to put the compensating coils at the extreme ends of the interaction region. There the betatron function is approximately (V-3) and Eq. (V-2) limits the maximum field allowed. If a tune shift is allowed of A\> < 0,02 per com pensating coil, then one has the condition: (V-l»> P p^ (0.2m) where Lc is the length of each compensator, and 2nax * 10 11 ie the distance to the nearest quad ripole at the end of the straight section. The beam momentum p is to be measured in tesla-netera: 15 GeV/c • 50 T-ra. Since BL (z -L /2)2 . ""T C > 0.32* (=E-)2 x (0.2m) * 56 m. This implies L > 1.5 m. However, barely satis fying Eq. (V-h) would imply that the field in GUM would have to he scaled with the beam energy: B„.,(*°)-E„ «wr>. Fig. 17. Strengtne of Bz and Br components of Thus one will tend to choose a larger length. magnetic field near poletip (z = 190 cm). Another condition is given by the fact that any imperfect alignment of the magnets introduces transverse magnetic fields. If the misalignment errors are "" 0.5 mm, then one haa to require Since {B£/p) = O.OO36 this Implies that: T < 390 m which because of Eq. (V-3) implies that the com pensating magnet center should 1.01 be further from the I.P. than z - /0.2 x 390 = 8.8 m . Since Eq. (V-5) again necessitates scaling B with XIL7H1-MJ1 the bean energy, a reasonable compromise would be Fig. 18. Drift chambers in GUM: longitudinal a set of 3 m long magnets, vith magnetic field of cross section. 10 kG. However, both the main GUM as well a3 the compensating coils should have provision for fine positional adjustment, so that condition (V-5) can be relaxed. One could then use the full magnetic field down to a beam energy of about 10 GeV. VI. THE INNER DETECTOR We propose that an initial detector con sisting of several layers of wire chambers be installed in GUM. This detector would be suf ficiently versatile — assuming the necessary software for track recognition is part of the GUM package ~ to enable initial use of the magnet for experiments. Later on, when special demands are to be met, or the technology has advanced suf Fig. 19. Drift chambers in GUM; transverse ficiently, other user'-supplied inner detectors cross section. could be installed. A. Initial Inner Detector A sample design for the inner drift chambers is shown in Figs. 18-21. We propose six sets of two chambers with delay line readout. Such a system la easier to subdivide in $ than the small- angle stereo system planned, e.g., for Hark II. In addition, the simultaneous readout of a single point ($,z) instead of two correlated angles $ in the small angle stereo system, makes pattern recog nition easier. The chambers are cylinders which are divided into quadrants and supported by four vanes attached to the coil support struts. This allovs any quadrant of any chamber to be removed and replaced with nthvr equipment without affect ing the remaining chambers. The lengths of the chambers are, in meters, 1.20, 2.20, 3.20, 3.80, 3.80, and 3.80. The average radii are, in centi meters, 18, 36, 51*, 72, 82, and 92 respectively. This arrangement gives unambiguous six i^oint tracks in the range 26° J 6 < 15^° and unam biguous It or 5 point tracks in the range 19.2° < 6 < 20° and 15^' < 0 < l60.8°. It allows a 19-2° front-back cone to be free of material. This greatly aids in the studv of two photon processes, as discussed earlier in this report. Figure 20 shows in detail the chamber construction. A cross section for a typical chamber is shown in Figs. 18 and 19. The material of the chambers is approximately 5 x 10-3 radiation Fig. 21. Drift -jhambers; support scheme. -60- lengths (r-1.) thick. The G-10 end plates are error. Ax, from that due to z resolution of -' x 10~ > r.l. thick. The chambers are made of individual chambers. plastic hexed material and aJmninlaed mylar. There are fhree sections of hexcel /'mylar laminate, B- Special Inner Detectors. •.'lie i-V which separates the two layer:- of chamber. !'he \u.< halves of n chamber are shifted re-:itl.e The inner tracking detector, connistinfi of tc one another by one half cell sir.e to .-,•". lvi' the tirM't chambers, is deliberately made in mo'iui .r lei'L-rif-hL siaibiftuity. They arc separated by o nun form to facilitate removal in part or in toto, '-:• -til -.v tl'.e local !;u!,'i'iit imrle of a trac'- in be As has been mentioned previously, llii:: permits r.-'as'ire i. This ('l-fatly aidi; in bittern rertv- part of the solid angle to be covered Irj ouiev experimental requirements (e.g., u.bnn"h»r close 'i'he ieluy Line is at negative high voltage, in, to obtain a clean signal for muon pryil action while t:ie senst* wire is at positive high v-slt.-irfe. over half the sensitive region). It altio pt-rmits :r!u ceLi Liu-e i:: L..1 cm with a full width oV <• BII:I. the entire set of drift chambers to be removed Tliis ••onfiKurat u-.. -illow.s the chambers in operate (perhaps without removing the beam pipe), ard the in a li. kil magno' ;•• fi.-.l.i without use of field installation a totally different inner detector. nh;»puiK wir-*s which voulu add significantly to the Examples of such detectors might be (a) a ;unou:.L of material in the chambers. streamer chamber,10 or (to) a time-projection A possil le support scheme ir, shown in !'IK. chamber.11 Their compatibility with a 'Jeneral .' I. i'he support vanes are in the shidow of the Uiierr. Magnet will be discussed briefly us examples coil -1:1 L return yoke, an 1 no io not occupy -my of potential versatility of inner detection of the solid alible available for detectors. schemes. .. The- res> Int.ion in the asimulhal direction (i) Streamer Chamber.± An installed vlli be ^00 microns or better. The rei-.olut.i--n streamer chamber system wculd appear somewhat as .ilon.: 'he oh amber, A'l, per delay line readout is indicated in Fig. 22. Here the design is from a given -Approximately by: recent study, which uses a set of solid-state imaging chips to readout the visual track inform ;.r. ^- f resolution for fin.-irsK center ation in a filmless manner. Access for the of p-jl:ie, nsec) Hlumlein feed lines {which ire small except near i (propapatior. tine of delay line, the chamber), and CCD readout is through ;?0° end nsec/ein) . cones rr between coils. The chips are inside the rarifinet, since they are not magnetic field sensi tive, and the optical images of the stramer .'.z. 'Hi- rise time of a .'signal, t'ror a senr.e wire, chamber tracks are focussed on the chips by means jn 'j. -lei.ay line of reasonable ien.-"h, "-».'., l'^ of a set or lenses. The non-uniformity of field is ns'-. "-: A* =£. uOU usee, is approximately: not a problem for the chamber operation or track reconstruction (the space-averaged non-uniformity r i" L line - to ' total propagation tine is < 1%). Due to the absence of field-sensitive of the line. equipment, there are no difficulties which can be The resolution for finding the center of su:*h a pulse is approximately: At = i»5 " 2 < rise time. For short line, say 100 nsec, with a propagation time of 1.5 nr.ee/cin, az = 5 mm. For longer lines, i.e., 0.5 nsee/cm, we again obtain the same resolution: Az = 1.8 cm. The obvious improve ments are to make less dispersive line- i.e., decrease the rise times, or measure z m„re often. The former has proven to be technically in- tractible. The latter solution is straight forward, but expensive. In the proposed chambers each delay line is read out from each end, and there are two delay lines looking at each sense wire. We therefore get four independent measure ments of z, and the error is reduced by a factor of two per half chamber. The number of sense wires in the proposed system is 3l€0. Each sense wire has an associated delay line readout at each end. The total number of channels to instrument is then pli30. The disadvantage of our system is the relatively large error in As. This error is particularly troublesome if one wants to re construct the invariant mass of several hadrons. Figure 1 shows the mass resolution of a typical case i^(3-1 GeV) -*• 2u (or 2tr) in function of the ty momentum. The resolution is dominated by A?.. XBL7512-9HS9 It should be noted, however, that the z coordinate of each track is measured independently by each Fig. £2. Example of streamer chamber as central chamber, thereby reducing substantially the net detector In GUM, -61- seen due to possible leakage of stray field. In particles such as charmed hadrci.o ur 'savy the event of a conventional film readout system leptons. These particles are expecttu to decay instead of the newer solid-3tate scheme, the GUM into normal leptons (u or e) with reasonable offers a natural photographic "port" through one branching ratios, and a prot.pt high-energy lepton or both of the open end cones. (say p > 5 GeV/c) may be considered a distinctive The spatial resolution in a streamer chamber signal. For the study of the properties of such is expected to be superior to that of a drift new particles, a suitable system would be similar chamber. This might enable one to use only to the "central detector" proposed for the 2y part of the magnet and still have sufficient experiment.^ An example of such a scheme is shown energy resolution without any external tracking in Fig, 33. The rejection of charged pions is of device. Thus, it would be ideally suited for a the order 10" ^ o in this 3ystem. Contamini-. on high p experiment. Also, the problem of un due to ir°-Dalitz pairs and conversion of n°-decav ambiguously identifying Ks and A decays would be photon; in the beam pi^re can be reduced below this simplified. limit by use of ehambera and dE/dx scintillator near the beam pipe. The "lumped" solenoid coil • (ii) Time Projection Chamber. The of the GUM reduces the probability of contamin time projection chamber (TPC) of B. Nygren offers ation due to charge exchange of it" immediately the potential of both high resolution and direct before the shower counters. The signature for readout of space points, in a large volume de directly produced muons is not optimal in the tector. In addition, it can, in principle, "normal" configuration of drift chambers, where provide ionization information to aid in particle almoBt 2 m of flight path is available for pion identification. As an example, a TPC might be and kaon decays into unions, The hadron rejection ratio is estimated to be "• 8 x 10" o for muona installed in the General Users Magnet, occupying ulJ the central one-meter diameter and two-meter of momentum above 3 GeV/c. However, the flexi length inside a system of coils having Just twice bility in rearranging the inner detector of the these respective dimensions. magnet allows a configuration more suiiable for direct-muon ideu*'i..'i-otion. Thus, a masBive Uniformity of field is important absorber of uranium or nevimet, etc., may be for optimum performance of the TPC. This is be placed surrounding the beam pipe at the inter cause a non-alignment of 5 and t field directions action point over one hemisphere, leaving the leads to a deviation of the drifting electrons other hemisphere open to a track chamber system from the E" direction, leading to a possible dis suitable for observing haaronie or direct- tortion of the track position as read out assuming electron final states. Alternatively, smaller § and 2 are parallel. This is not, however, segments of the aaimuth may be blocked by dense equivalent to a resolution loss, at least not for absorber, leaving free regions in between, for small non-uniformities, where such deviations observation of correlated u-e or u-hadroiiic might well be corrected out by mapping in the states. Details may be found in Ref, 6. Finally, more inhomogeneous regions. Evpn without mapping, the GUM is particularly suited for the study of a non-alignment of fields of aboui 0.3% corre new statuJ produced in Sv reactions, since the sponds „o the intrinsic resolution of the device. This is to be compared with a value of about 0.k% for the maximum non-uniformity (at maximum radius, r = 0.5 m) in the present GUM design, and « 0,5# for the average space point. Thus, with some mapping at large radii (or improvement in magnet deBign), this magnet should prove quite adequate for operation of the TPC, even with the ends open, ao thftt electron tagging may "be done (e.g., in recording the 2v events), VII, SOME POSSIBLE CONFIGURATIONS FOR EXPERIMENTS We have, in close consultation and collabc— ation with many members of other groups at the PEP 1975 Summer Study, studied some particular applications of GUM, A. GUM Used in Hew Particle Detection0 The GUM is well suited for use in a large solid-angle detector system for the study of new Estimates of the ultimate resolution for streamer chambers vary, but a resolution transverse to the streamer direction of perhaps ~100 u seems possi ble. In addition, there is generally much more information per track, due to the "continuous" nature of measurements, so that the momentum reso lution may ultimately be considerably better than can be obtained from drift chambers of equivalent Fig. 23. Side view of charmed particle detector space-point resolution, (See Ref. 10) in GUM (see also PEP-gcJ(). end-caps of the solenoid are free for tagging up C. Uifth-Homentua Detection"* to about 20°.)( The high-momentum group at thin "'inner :'.tutly B. Two-Photon Studies'' has designed a detector using the GUM as basic deLector. The emphasis of thia group is on The PEP study group on two-photon processes complete reconstruction and identification of has made a design for an experiment using the flUM secondaries accompanying a particle of hlc'i and supplementing it with transverse field spec momentum. A cross sectional sketch of a possible trometers covering the h°-Q0° regions at each end design is shown in Fig. ?5. Since the goal io to of the solenoid. Either the lumped-coil or identify as well as precisely measure all the continuous-coil version of GUM could be used charged particles in the event, a large part of equally well. The design is shown in Fig. 21*. the volume outside the coll is occupied by three The outside rep.ion is covered jy an aerogel layers of Cerenkov counters, operatlnR in dif Cerenkov counter, followed "by a Pb-liquid argon ferent velocity ranges. Just outside the eo_"l, shower counter :md a set of rauo:i chambers. The there are 96 cells of Cerenkov counter Cj oper tagging apparatus consists of a very small-angle ating with l*.l atmospheres ', [isobutane]. Nat array, while electrons with scatterinR angles This is followed by 160 one-atmosphere C0^ 50 flj-mi •; 0 •: . USkG SoJeno-tt) 1 Flu - Ffclurn oneV ) | /^ \(h Ptj-A Snowr^. ml.04 C«ren»o» \ CQ mm Jcckcl J fl 2 "y-* 0 FK11-I 'ding Pipt Ki^. S-. Two-photon experimental setup in GUM (see alao fig. 15 '-if PEI'-yD^). End (ila'c Open TOF 1 IXnPb XBI.7512-9H71 Side view of high-mi 1 h.'viron experiment . •e also PEP-201). yokes, which are also part of the Y measuring The disadvantages of this magnet as a general system. Together, C3 and C^ reBOlve pions from user device are obvious: K's and p's (or p"*s) over 0.82(4ir) sr solid angle (i) All phototube assemblies and other in the momentum range 1,3-15 GeV. Along vith C3, detectors must withstand 3 kO. Microchannel- full n/K/p separation is achieved above 1.3 GeV/c plBte photomultlpliers available at reasonable cost over 0.56(Uir) sr. Time-of-flight measurements would relieve thiB otherwise severe constraint. are obtained over a 3 m flight path by scintill (ii] If the high mass and cost of shower ation counters adjacent to C3. counters (e.g., liquid argon) requires their Gamma rays are measured using (l) drift placement with *1.5 m of the interaction point, chambers behinrl 1 r.l. of lead to determine the Y the reduction in radius will result in poorer direction, and (2) shower counters (0.36 x lm sr) momentum re -luticn. consisting of alternate layers of iron and pro (iil) The magnet may be very expensive. portional tubes in the laminated flux return. If constructed using KTNIMAG techniques, it would Charged particle measurements are made by the cost at least twice as much as the lumped sole inner set of drift chambers. noid described here. However, there is no need In this design, charged particles can be to restrict the superconducting coil thickness to identified and precisely measured over a auite much less than one radiation length or* for that large fraction of the total solid angle. Gamma matter, to rule out a conventional coil. Coll rays also can be detected and measured, though forces are proportional to B^ and therefore very over a smaller solid angle. The potential for small. identification of charged particles and gamma-ray The large volume, low field magnet has at measurement exploits the thin windows between the tractive features aa a general user facility. lumped coils, and In other respects, the geometry Hov«ver, it is less flexible than the magnet and size are very similar to the principal GUM system previously discussed. Perhaps this al design considered previously. ternative should be considered seriously if micro* channel photomultipliers should become available VTII. ALTERNATE DESIGN POSSIBILITIES - ,- in the near future at a cost comparable to LARGE VOLUME LOW B GENERAL USER MAGNET ordinary photomultipller tubes. The 6 m diameter 3 KG magnet proposed by this ACKNOWLEDGMENTS year*B high momentum and particle identification group might also be viewed as a candidate for the We would like to acknowledge the help of general user magnet. Compared to the magnets John D. Taylor in helping to prepare and run the which are the main focus of this report, the program TRIM, on which our field plots are based. large volume device would have the following We also wish to thank Pbillipe Eherhard for a advantages: number of relevant and helpful remarks, end Line. Galtieri for use of her result on y conversions in (1) Out to 2 a radius, no appreciable a solenoid coil. Thanks also to F. Villa and amount Qf material need separate the user's R. Mozley for their advice. detector from the beam pipe. Within thiB radius, the detectors may assume aoy shape. (2) Very precise (typically 2-52) libit II. CcnjortiCE if PnspQtri Maca*tt u,t t>> momentum measurement is possible over the full momentum range if the multiple scattering ma SuptreoMiKlini terial can be kept small over an appreciable fraction of the trajectory. (3) The la-ge extent of the magnetic field permits precise momentum measurement out side relatively massive detectors placed close to the beam pipe. Examples other than Cerenkov detectors are: (i} mucus may be measured outside a small "lead ball" which suppresses Kaon decay and other background, vith a momentum resolution [?U — 5 * greater than achievable with the UBual configurations of magnetized ironi (ii) many con centric shells of thin radiator could be used with drift chambers for electron-pair spectrometry of Boft photons. ThiB could be the most economi cally feasible means of achieving high resolution, both in energy and direction of Y-rays with C«*cka< CviDMit energies < 100 MeV. Gwi ,'; 0.1 p.l. (k) A 3 m flight path may permit veloci ty measurement by relativistic dE/dx rise. This could be combined with trajectory measurement by mm" (Mcuni mewis of large-volume drift chambers, time pro jection detectorB, etc. The ambient magnetic field may be helpful in limiting the range of £- r&ys which confound the dE/dx measurements. (5) Addition of specially shaped iron *nickM» or TV • TMttl nlU bjfmc ei pole faces permit the user to create a particular ileld shape near the interaction point. REFERENCES 1. SLAC r"roposal Si'-c {Hark I Detector). See also PEP-200 , this Proceedings. 2. SLAC Proposal SP-7 (IRON BALL). 3. SUC Proposal SP-2'i (CRYSTAL BALL). I*. Group Report on Two-Photon Physios, I'EP-203. See also PEP-175, 197'i PEP Summer Study. 5. Group Report on High Momentum Particles and Particle Detection, PEP-201. See alr.o PEP- I'te, 197>* PEP Sunmer Study. 6. Group Report on Search for New Particles, PEP-20U. See also PKP-17'i, 197'" PEP Summer Study. 7. PEP-153, T. Hast and J. Nelson, 1971* PEi' Sunmer STudy. 8. The numerical coefficient in this formula represents an average between a setup with only three chambers and one with a very large number of chambers. 9. Private communication, J. M. Patterscn and B. Richter. 10. Group Repo.'t on Streamer Chamber Detectors FEP-197, s.^e also PEP-151, 1971* PEP Summer Study. 11. Group Report on Time Projection Detector, PEP-198. See also FEP-lfcli, 191U PEP Summer Study. APPENDIX A SUPERCONDUCTING DETECTOR MAGNETS - ALTERNATIVES AMD CHOICES The relative advantages of different magnetic field geometries were discussed Sec. V.A. Tt was pointed out that the solenoidal coil ;eems the most attractive from many points of vi-;w, and that geometry will be examined in further detail in this note, with special regard to the engineer ing design eonsiderati •n:;. The questions which are or immediate interest are (l) the size, and (2) whether detectors will be used outside the coil (and certain properties of these detectors). The momentum resolution for Vir.. A-l. 'Hi*.' b>\. current density "cuiiventional" charged particles is ur.ually the single mo.-;, au|'i.-rc 'iviuctine solenoid. important criterion, and this topic has bren discussed in deta . in Sec, IV.li. However, this consideration interacts with both (l) and (:.'), and since a typical experiment is a -'omprondae between various physics objectives and the expense ;>f tli" detection apparatus, it may he important t<> con sider several kinds of magnets. Therefore, we will discuss the engineering problems associated with three kinds of solenoid magnets: (l) the conventional low current density magnet, (;•) the continuous high current density thin coil magnet, and (3) the lumped coil magnet. The conventional solenoid will not permit a significant amount of physics to be performed outside the coil. The continuous thin coil and the lumped coil magnet will permit a considerable amount of interesting physics to be performed outside the coil. TabLe A-l compares the three kinds of coils in a raapnet with a 0.9 *n useful (warm) bore diameter and a useful length of l.fl't m. Longitudinal cross sections of the three kindo of coils are slinwn in Figs. A-l, A-?, and A-'.. The magnetn which :U-T- compared in Table A-.l and at'L: shown i H'lffs. A-l, A-2,nnd A-3 are ever, if large diameters (greater than 2 or 3 meters) and inductions greater than 2 T are re quired for physics reasons, there probably 1B no reasonable alternative to the cryostatically stable conventional solenoid magnet. The major problem which has been encountered in most of the large solenoids built to date has been the cryogenic system. The large bubble chambers are bath cooled. (The superconductor is immersed in a bath of cold helium.) Massive bath cooled solenoidB are difficult to cool down under the bsst of circumstances. The OMEGA magnet represents a positive step forward in large magnet cryogenics. The superconductor, which is hollow like conventional water cooled conductors, carries supercritical helium.A~2 The cooldown of such a system, if properly designed, will be faster than an equivalent bath cooled magnet. The inventory of helium in contact with the magnet is reduced. However, the primary disadvantage of Fig. A-3. Ike lumped high current density sup*r- supercritical helium cooling is the amount of conducting solenoid. refrigeration required to obtain low enough oper ating temperatures. Large conventional bubble chamber solenoidB have been major us rs of superconducting magnet roughly half the size of those being proposed for technology. Large .-agnets of this type will PEP. The proposed PEP magnets, which are about 2 continue to be built. Physics experiments which meters in diameter and U meters long, will have a require some of the particle detection to occur larger percentage of usable solid angle available outside of the magnet winding cannot use con for physics than the magnets.shown in Table A-l. ventional low current density superconducting It is useful to point out ve have more magnets. The other alternatives are discussed in choices available to us than there were Just a the next two sections. few years ago. The relative merits of the three choices are discussed in the sections to come. THE THIN HIGH CURRENT DENSITY SOLENOID MAGNET THE CONVENTIONAL SUPERCONDUCTING SOLENOID MAGNET The thin high current density solenoid mag net has a low radiation thickness over its entire The conventional low current denBity super surface. In PEP, the magnet will permit physics conducting solenoid has been used in high energy to be performed at about $0% of i'-* solid i>ngle physics for the last eight years. Examples of outside the coil. The magnetic induction outside thin kind of magnet includes: the 12-foot bub the coil will be quite low if the central in ble chamber at ANL, the 15-foot bubble chamber at duction is kept below 1.8 T and if the iron return Fermi Lab. the Fluto magnet at K£Y, the LASS yoke is properly designed. When the induction nagnet at SLAC, the BEBC, and OMEGA magnets at outside the coil is low, photomultiplier tubes CEiW and other sensitive electronics may be used in ' , A-l that region. The thin solenoid can be easily In 1965, Steekly showed that if the cur modularized so that individual magnet sections rent in a superconductor could be carried in a can be tested separately. The col.'1 mass is low; copper substrate without heating the super the cooldown from 300° K to k° K is relatively conductor to a temperature above its critical easy. temperature, the superconductor would operate stably. The principle of cryoatatic stability is The thin solenoid is not without its dis used in nearly all of the large high energy advantages: The magnet must operatf at very high physics detector magnets. This type of magnet current densities. As a result, 1-jrge magnetic will not quench. The current will jump from the stresses and high stored energy p*r unit mass can superconductor to the copper. Since the magnet occur. The magnet will quench if any of the will not quench, the solenoid can be made in large superconductors go normal. A practical upper sizes (over 6 meters in diameter) vith fairly central induction limit is Just over 2.0 T. Large large central inductions (up to U T for a large diameter COIIB, which are stress limited, will size magnet). The technology is proven and have a lower central induction. Snail high cur reasonably reliable. rent density solenoids have been buil+> d oper Meaningful physics is nearly impossible out ated successfully in experiments. HOY --jr, large side the magnet coil because, in most cases, its high cv.-rent density solenoids have no- yet been radiation thickness is in excess of two radiation proven. lengths. A eryostatically stable magnet is a low Engineering studies on the thin solenoid . _ current density magnet. Its thick coils are mas show that the concept is technically feasible. sive and difficult to cool down from 300° K to Preliiiinary experimental teats on small magnets U° K, The cryogenic system uBed on large con have been very encouraging. The remainder of this ventional aolenoidB makeB them difficult to section will discuss the following: the design modularize and test before final assembly. A large characteristics of thin solenoid magnetB, the conventional superconducting solenoid is expensive scaling laws for thin magnets, and the LBL test to build and requires a large crew to run. How program for thin magnets. The Desipn Characteristics for Thin Solenoid that the aluminum bore tube will help support the Magnets magnetic stresses in the system, which gives on additional u^rgin of safety. The thin (low radiation thickness} solenoid The thin magnet design proposed for MINIMAG should have tr 5 following characteristics: attempted to combine the cryogenic system arid 1) It must have uniform radiation thickness superconducting magnet into a single integrated (normal to the magnet coil) over the full length system. It is hoped that one can avoid the kinds of the coil. 2) The cryogenic system should "be of cryogenic problems which have been common on designed for ease of cool down and simplicity of the large bubble chamber magnets. Since the construction in the thin region of the magnet. magnet is not cryC3tatically stable, the liquid 3) The coil should be built in modules which can helium inventory in contact with the magnet can be be tested individually. A thin solenoid which reduced. Most of the helium in the system is out meets the above criteria vas designed in con.juc- of direct contact with the magnet. The tubular tion with the MINIMAG experiment.A-1* cooling system also permits a positive well controlled cooldown process. The cooldown of a The MINIMAG solenoid was designed so that "MINIMAG type" thin solenoid should proceed much it could be modularized. The major stress, stered faster than a comparable low current density energy, and quench problems were studied. The magnet. tubular cooling system permits rapid cooldown and positive cooling consistent with *;,e low radiation thickness requirements of the (.-xperiment. The Scaling Laws for Thin Magnet3 thin solenoid consists of four primary parts: the solenoid bore tube, the superc- lucting coil, Thin coils must operate at lower central the refrigeration cooling tube, and the cryostat inductions as the magnet diameter increases. vacuum vessel (See Fig. A-M- There are two primary reasons for this; the stored The solenoid bore tube serves as a winding energy per unit coil mass should be less than form for the coil, but its most important function 25-30 Jg"1. The maximum stress in the conductor is an electrical one. It slows down the quench (the magnet coll is assumed to carry all of the process (the process of going normal and dumping magnetic stress) should be less than 5 x 10° Nm~2 the magnet stored energy), and it serves as a (72,300 psi). Keeping the proceeding limitations thermal sink for most of the magnet stored energy in mind one nay apply the following scaling law during a quench. The superconducting coil is to thin coil constructions: wound with intrinsically stable (twisted fine filamented) superconductor which is operated at r - „2 „3/2 SO? of its critical current or less. The thin solenoid would not have a cryostat in the conventional sense. The inner eryostat vessel is replaced by a coil of aluminum refriger ation tubes which carries flowing two-phase helium. The refrigeration tube forms an integral part of the superconducting magnet. The magnet bore tube, the superconducting windings and refrigeration tube are vacuum impregnated forming :i single structure which is suspended inside the vacuum enclosure. The ends of the thin magnet may be thick from a radiation standpoint. There fore, all of the cryostat support functions, refrigeration feeds and current leads would be in that region. We def]^e the proceeding symbols as follows: The superconducting coil is designed to fl0 is the central Induction (T); D is the coil carry ill of the magnet stresses put into the diameter (ra); L is the coil lenfith between poles system at its peak field. However, we expect (m); fc'_i s the coil stored energy per unit length -1 (J m ); M0 is the permeability of air (u0 = ^n " lO-^). I' is the scaling constant for thin mulcting spacer solenoids; Bmax is the maximum central Induction is tn for the magnet; EjnQX * """timum stored energy per unit length. T, Bmax and ^mix are functions of the coil radiation thickness. In addition, the maximum practical coil diameter Pmax is also a function of radiation thickness. Table A-2 presents estimated values of T, Bmax, En]ax and T'max f*or thin superconducting solenoids of various radiation thicknesses. The four magnets shown in Table A-? which Aluminum spool tiave radiation thickness of O.'ifi, O.'tO, 0. jl', and 0.25 radiation lengths are assumed to hrive aluminum bore tubes and cryostats. The "ultra thin solenoid" is assumed to have a msiRtieni'an Mru-r.ified cross section o the thin <•< bore tube and cryostat. A radiation thickness of iLiriii'"js high current density solenoid • O.lB radiation lengths Is judged to be very close I .• ''.in,: tubo. tn a lower limit for thin coil technology. A -67- magnet which has a radiation thickness of 0,18 thick formvar is applied). Both conductors have radiation lengths will cost 30-50X more than the over 2000 filaments, and they are twisted at the same magnet when it has a radiation thickness of rate of one turn per centimeter. The filament 0.25 radiation lengths. Table A-3 shows a break diameter of both conductors is under 15 ura. Both down of radiation thickness of the magnets shown conductors are modern intinsically stable in Table A-2. conductors. The two different conductors are wound on two 6.35 mi UA thick) 1100 series aluminum allow tubes which are 1030 mm in diameter and 500 nun long (including end flanges - See Fig. A-fi). 1 1 i 1 i r \ \fsCosee, r-6 LTLOLIA CcBpenrntt ur Lcrv pi Casts DBE, r«2-*"* i _U 1 I _L_ as 1.0 15 2D ZS W> 3.5 Magnet coil diameter (m) XBLTS9-4I93 . A-5. Scaling of magnet central induction vs magnet coil diameter: Case A rad. lengths thick Case B 0.40 rad. lengths thick Case C 0,32 rad. lengths thick Case D 0.25 rad. lengths thick Case E 0.18 rad. lengths thick. ,„., •» »"•'• U" .ID .St n.St "Tu.1 — • «,., «„. Figure A-5 shows n plot of the thin magnet central induction B0 vs, the magnet diameter D for magnet with radiation thickness of 0.it8, Q.l*, 0.32, 0.25 and 0.19 radiation lengths. Figure A-5 uses the T, Pflias, Emax. and ^ax given in Table A-2. These values are our "best guess" at this time. They are based on limited experimental data. An experimental program now under way at Lawrence Berkeley Laboratory (LBL) will determine much more accurately the values of r, Bmax, E^^ and ty[)Qx- The LBL Experimental Teat Program for Thin Magnets The experimental program is built around the testing of two 1.03 m diameter MINIMAG type prototype coils- These coils have a radiation thickness of O.L'I' radiation lengths. The two coils will Mse differpnt niobium-titanium super conductors. One eoH will use 1,8 to 1 copper to superconductor ratio conductor; the other coil will -use a 1 to 1 copper to superconductor ratio CBB T? - U83 conductor. Both conductors are 1.0 mm In dia Fig. A-6. The LBL thin continuous coil prototype meter (the bare diameter before a layer 0.05 turn being vound. -68- The bore tube is expected to play an important concept and will provide reasonable experimental role in controlling the superconducting magnet determination of magnet scaling factors. quench. The superconducting coil has a layer of 12.7 os> o.d. (1/2 inch o.d.) aluminum tube wound The DeaiKn Characteristics of Larse Lumped around it. This tube will carry two-phase liquid Solenoid Magneta helium as a coolant for the superconducting magnet (See Fig. A-b). The lumped solenoid will have the following The two superconductors have been teBted at characteristics: high current densities (> 1.2 x 10°- A m~2) and 1. The holes of very thin regions of high magnetic stress (>4x 10° K m~2) in small the magnet or magnet cryoBtat should be aa large oval solenoid tests. The large solenoids will as possible- test the conductor under conditions of high cur 2. The cryogenic ayBtera should be rent density, high magnetic stress, and high designed for ease of cooldown and simplicity of stored energy. We expect to be able to determine construction. experimentally the magnet sealing factors r, 3. The coil should be eaBy to test in 3mBX and EJJJQX- Thus we expect to prove that high modules. current density superconducting coil technology Preliminary designs for a proposed CERN experi- can be applied to relatively large (1-3 m in mentA~T indicate that the proceeding criteria can diameter) magnets with central inductions from be met in t lumped solenoid system. 0.6 to 1.8 T. A lumped solenoid should be easy to modul arize. The modules may or may not be identical THE LUMPED HIGH CURRENT DENSITY SOLENOID MAGNET In physical shape, but they must contain the same number of ampere turns of conductor. The solu The lumped solenoid has a non-uniform radi tions to stress, stored energy and quench problems ation thickness over its length. The supports which are applied in the MINIMAG thin solenoid between coils are also thick from a radiation studies, may be used for the lumped magnet system. standpoint. The radiation thickness is The use of the tubular cooling system, which 1B an 0.01 - 0.06 radiation lengths in the thin important part of the MINIMAG concept and has been regions of the magnet\ the radiation thickness at successfully employed at SIN, car be used to the coils and support members will typically ex advantage in the lumped coil system. ceed two radiation lengths. The regions of very The lumped solenoid magnet consists of four low radiation thictness are useful for certain primary parts: the lumped coil bore tubes, the kinds of physics (i.e.* the accurate measurement superconducting windings, the tubular cooling of the momentum of high momentum charged particles system, and the cryostat vacuum vessels. and low energy gamma rays). Like the thin con The solenoid bore tube serves the sane func tinuous solenoid, the lumped magnet is easily tion in the lumped coil solenoid as it does in the modularized so that individual magnet sections can thin coil solenoid. In both cases, the coil bore be tested separately. The coil mass is not as lew tube controls the magnet qui nch precess. The lumped as the continuous solenoid, but it remains low coils are not as well coupled inductively to the enough so that the cooldown of the magnet from bore tube as the thin coils (9855 coupling is 300° K to h° K is relatively easy. possible in the thin ceil system; 90S coupling •ine lumped solenoid must operate at rela should be possible in the lumped coil magnet). tively high current densities in the conductor. Like the I'kin solenoid, the lumped coil solenoid As a result, the magnet will quench if any of the should be wound with intrinsically stable super superconductors goes normal. The quench process conductors. The conductor is operated at a lower must be understood and dealt with. Since the current density than in the thin continuous coil magnet must be designed with the quench process in case. The primary reason for this is that peak mind, the central induction of the lumped high induction in the conductor can be twice the current density solenoid should be no higher than central induction of the magnet. (In a thin 2.0 to 2.5 T. Large diameter lumped coils will solenoid, there is less than a 5% difference be be stress limited, so they would be operated at tween the peak and central inductions.) The a central induction below 2.0 T. superconductor in the lumped solenoid would be operated at 753? of its critical current, or less. Numerous small magnets have been operated at the current densities proposed for the lumped Like the thin solenoid, the lumped solenoid solenoid (about k x 10° A m~2). Large lurapnd coil should use a tubular cooling system using two type magnets ("2 m in diameter), which is similar phase helium. The refrigeration tubes, the bore to the lumped solenoid coils3 have been bui\t by tube and superconducting coil form an integrated NASAA-5 and by the high energy physics laboratory package. 'fiie lumped coils are assembled to at Stanford University.^"" Both magnets have been gether with cold force carrying members. The operated reliably. The luaped coll concept is, in Btructure, which is mostly holes, is nearly in a sense, more proven than the thin continuous force equilibrium. The support system which solenoid concept. suspends the lumped coil structure to the room Preliminary engineering studies on a lumped temperature outside world is designed to carry coil detector indicates the magnet concept is gravitational forces and those magnetic forces technically feasible. Successful thin coil tests which are generated by aaymetric currents in the are encouraging for the thick lumped magnets as system. well. These magnets are not driven to as high a Two kinds of cryogenic vacuum vessels can be current density as are the thin coils. Success considered for the lumped solenoid system: the ful completion of the lexge thin coil testa will continuous uniform cryostat and the thin window improve our knowledge of the lumped coll system, cryostat. The continuous cryostat is less ex and will verify the technical feasibility of the pensive than the thin window cryostat, but its radiation thickness ,'LS a factor of 3 to 5 greater. In order to minimize radiation thickness, core M. the continuous cryostat would have to be made '- Dore magnesium or magnesium alloy. The thickest par of the cryostat is the cuter vacuum can; this where the symbols are defined 03 follows: E is thickness is needed to resist buckling due to the magnet stored energy (J); D is the magnet coil vacuum loading. Proper mechanical design will diameter (m) at the center of the windings; L is the magnet length between pole tips (m); B is the result in radiation thicknesses between the coils 0 central indi-tion of the magnet (T); M C is the of 0.05 to 0.07 radiation lengths. S total superconu-ctor matrix mass (g), which in If the physics of the experiment requires clude the Hb-Ti and all normal metals in the leBS than 0.05 radiation lengths of material in matrix; Mbore is the bore tube mass; and vQ is the thin regions of the magnet, the thin window the permeability of air u - hir * 10"?. cryostat should be considered. A thin window cryostat can have a radiation thickness as low as If the lumped -oil magnet IB divided into II 0.01 radiation lengths. It can he built with ceil packages, th .iraits for stored energy per either aluminum foil or Mylar windows. Both unit mass are: kinds of windows will be quite fragile. Both will require either plastic foam or a screen to protect the windows from damage. A cryostat without K <- ^10 £ /r(*- ~1\) windows* in which the cryostat vacuum vessel closely surrounds the coils and struts, has been suggested. The cost of such a cryostat plus when U < H < 8 maintenance problems associated with such a 1 s>etem rule it out. In addition, there is little Ksc <12.5 (Jg" ) physics advantage which can be gained by reducing the minimum radiation thickness from 0.01 radi when N > 8 ation lengths to zero. ^^25 (Jg"1) The thin solenoid tests which will soon be conducted at Berkeley will answer many technical when N < !*, questions which the lumped solenoid poses. The cooling system is essentially the same for the two systems. In some ways, the lumped soleno.'d design is conservative compared to the continuous thin In general, the aluminum bore tube should b:. de signed so that K < 10 Jg"1. Magnetic stress solenoid design. Therefore, a successful L meter bore only becomes a problem when Kgc is greater than diameter thin solenoid test will advance the cause 1 of the lumped solenoid as well. 20 Jg" . The only further restriction on the design i3 that the superconductor operating cur rent be less than T55S of the superconductor criti Sealing Laws for Lumped Magnets cal current at the peaK induction in the coil. The resulting scaling is as follows: 1) For There are two primary criteria which deter a given central induction the solid angle lost mine the size of the lumped coils in a lumped due to the coils and their support is a constant. coil magnet system. They are: stored energy per 2) The solid angle lost goes as l/D if one allows unit coil mass and magnetic stress. To the first B to go down as D~V Both of the preceeding order, the stored energy per unit mass and Bag- 0 statements assume that the thickness of the super netic stress gc together. As a result, the conducting coil package and its cryostat remain scaling law given here will be based on stored constant as D changes. Scaling probably does not energy per unit coil mass alone. The coil in this apply when D is greater then 3 meters. case includes the superconductor and the bore tube. Each is treated separately. The scaling laws given here for lumped coil systems are approximate. Better scaling laws will The current density in the superconductor result from the series of experiments now goinp (the Nb-Ti alone) is a function of the local on at LBL. temperature, induction and stress. The current density in the overall matrix (copper, otner normal metals and superconductors) is a func+ion SUMMAHY of the stored energy. The stored energy for a lumped coil magnet with return path and a Two basic types of detectors can be con sidered for use on PEP and other intersecting central induction BQ < 1.5 T is to the first order: storage ring devices. Both will utilise solenoid- al fields with flux paths parallel to the di rection of the beams. One type, the conventional superconducting solenoid, permits physics to be performed only inside the coil in the magnetic field. The second type based on high current density technology permits physics t.) be performed The stored energy per unit superconductor matrix both inside and outside the magnet coil. mass is: Two kinds of high current density magnets can be considered. One is based on a continuous thin coil of uniform radiation thickness; the second is based on lumped coils which have regions of high radiation thickness and regions of near The stored energy per unit aluminum bore tube zero radiation thickness. Tor many experiments mass is: the high current density magnets could result in -70- substantial cost savings in both capital and oper A-1*. M. Alston-Garnjost, A. Harbara-Galtieri, ating funds. The central induction of such mag 0. Dahl, P. Eberhard, R. Kennoy, A. Litke, nets is 2 T or less, and the maximum practical G. Lynch, T. Mast, D. Miller, J. Nelson, diameter is 2 to 3 meters. S. Parker, H. Pripstein, R. ROBS, B. Sadoulet, Each of the two types high current density F. Solmitz, H. Stevenson, and D. Young, "A magnets has advantages for some kinds of experi Proposal to Simultaneously Measure Charged ments. If high central inductions and/or large Particles and Gamma Rays over a Large Solid field volumes are required, there is probably no Angle at SPEAR II," Lawrence Berkeley Labor practical alternative at this time to the con atory Proposal, October 15, 1971*- ventional low current density superconducting solenoid. Since the performance parameters of A-5. W. L. Pope, G. F. Smoot, L. H. Smith, and the high current density magnets are not fully L. H. Smith, and C. E. Taylor, "Super understood, experimental work now under way at conducting Magnet and Cryostut for Space LBL will establish the practical limits of oper Application," in Advances in Cryogenic ation for magnets using high current density in Engineering, Vol. 20, p. U7. trinsically stable superconductors. A-6. "Superconductivity Accelerates Cancer Research," in Cryogenics and Industrial REFERENCES Gases. August 1975. P- 15. A-7. A Proposal to Study High P, Physics with a A-l. Z. J. J. Stekly and J. L. Zar, "Stable Large Aperture Hadron Spectrometer at the Superconducting Coils", IEET Transactions» CERN ISR, British-American-Scandinavian ISR NS-12 (3), P. 367 (1965). collaboration; CERW/ISRC 75-18, 12 Hay, 1975, A-2. M. Morpurgo, "A Superconducting Solenoid CERN/ISRC 75-21, 25 June, 1975. Cooled by Forced Circulation of Super critical Helium", CERN 69-25 (1969). A-8. CERN Courier, "Villtgen Superconducting Muon A-3. H. A. Green "The Large Superconducting Channel Begin Operation,"{A Description of Solenoid for the MINIMAG Experiment", to the Swiss Institute for Nuclear Research — be published in the Advance in Cryopenie Supercritical Helium-cooled Solenoid) 15 (2): Engineering, LBL-367T, July 1975. 36 (1975). APPENDIX B DESIGN CONSIDERATIONS FOR A LUMPED SOLENOrD* Klaus Halbach , Lawrence Berkeley Laboratory University of California Berkeley, California 94702 1- Introduction steel everywhere to the riqht of the dashed line in Fig. Bl) and then investigate one by one the This note is primarily intended to demonstrate effects of the modifications that fwve to be which characteristics of a lumped solenoid (with a applied to the Idealized magnet 1n order to obtain steel return path that may also be lumped) are the real magnet. important for good performance. This understanding is clearly necessary to make judgments about trade 3. Useful Tools and Concepts for offs between qualities of the magnetic field and the Discussion of Magnetic FfeTds other desirable attributes of a particular magnet geometry. We also want to show the use of a method It 1s often quite easy to obtain a good picture ology and some "thinking tools" that can help to of a magnetic fiald distribution by applying direct obtain the good qualitative understanding of mag ly the magnetos£atic equations, usually in integra netic fields that is necessary for good magnet de ted form (f H»ds = I, f S-da = 0). However under sign, is indispensable for hand calculations, and some circumstances this procedure does not lead should be the basis for magnetic field computer easily to a good qualitative feeling for the fields, runs. While it 1s clear that it is not possible to particularly when one wants to know the fields close say anything basically new along these lines, it to a coll with steel in the vicinity. If the fields seems worthwhile to formulate and state these con have cylindrical symmetry with no component in the cepts clearly, since they are not nearly as well azimuthal direction, or 1f the fields are of a two known as they ought to be. dimensional nature, the analog model of current flow 1n a two dimensional conductive sheet often gives To accomplish the stated objective, we first the needed insight much more easily. Although this describe the design problem and the general method analog model has been used to physically mode'; 2D used to solve it. We then introduce two analog magnets (B-lathis use Is somewhat complicated in the models that often help to get a good qualitative cylindrical case, and the model is used here only understanding of magnetic fields. In addition, we as a conceptual aid. Similar comments apply to a derive and use a simple formula that can be useful slightly different analog model that we use to for making quantitative estimates of some properties visualize magnet fields produced by magnetic charges. of magnetic fields. Finally, we apply these methods I.i both cases the same labels z, r are used for the to the design of the lumped solenoid. Cartesian coordinates on the 2D conductive sheet and for the coordinates of the cylindrically symmetric magnets. 2. The Oesign Problem and the Methodology Used to Solvent An expansion of 2D dipole fields into exponen Fig. B-l shows a cross-section of the magnet tials will be done in Sect. 3.3. Even though this under discussion, with the center line (CL) re may seem to be a rather specialized formula, its presenting an axis of cylindrical symmetry. The derivation 1s reproduced here not only because it vertical line at the left represents a symmetry is useful for the magnet under discussion, but is, plane. The cross-hatched area indicates the steel In the "Uithors opinion, the single most important that 1s used to conduct the magnetic flux around and useful formula (after the magnetostatic equa the lumped solenoid of radius rj, the part at the tions) for the understanding and design of magnets. top of the drawing representing either a cylindri cal shell or a cross-section through a "vane" 1n 3.1. The Orthogonal Analog Model (QAM) case one chooses to "lump" that part of the steel structure. Inspection of the magnetostatic equations in cylindrical geometry, and of the equations governing The design goal is the production of a reason the current flow in a two dimensional conductive ably uniform field over a volume that 1s not too medium, shows the following relations: (see rows strongly restricted in either the axial or radial 1 and 2 of Trtle 1, and also Fig. B-2. direction. Furthermore the magnetic field should be small in a region outside the coils that begins 1) If one makes the resistivity p(r,z) of the not too far from the coils. sheet proportional to rv'r.z) (column 1), and To understand and assess these problems we 2) If one injects a current density J3(r,z) start with a discussion of the properties of an Into the top surface of the conductive sheet that idealized magnet (infinite permeability n, T2 ~ », is proportional to the exciting current density* j${r,z), in the magnet (column 2), Then: 3) Scalar equi-potential lines 1n the model correspond to surfaces of constant rA (i.e., field -72- surfaces) in the magnet (column 3), and the vectors Hx, Hy as shown in Fig. 4. It follows directly that H and r'B are obtained by rotating the 2D current both HX and Hy are expandable into Fourier series density vector j and electric field vector E in the with period g that should converge quite rapidly model by 90° (columns 6 and 7). since both the functions and at least their first two derivatives with respect to y are continuous. Vtw table of equivalents for two dimensional Using the complex representation for these Fourier magnetic fields is obtained by removing all factors series, we then form the combination H* = Hx - illy r r>om Table B-l. and obtain, by combining the coefficients into an: Mile this treatment of the permeability u can be advantageous, it is often more convenient to = 2 n1y/g H* \ - i i'v - ? ane " . (1) treat finite permeability effects differently: First u is assumed to be infinite in order to We have chosen the combination H. - iH because the obtain an understanding of the fields in the vacuum magnetostatic equations for the vacuunrVield compon region. From this one learns how magnetic flux ents H and-Hyare the same equations as the Caucny- enters the steel, leading in turn to a qualitative RiemanR conditions for the real and imaginary part understanding of the field lines in the steel, and of an analytical function of the complex variable at the steel-vacuum interface, whenever a field z = x + iy. Since we have an explicit expression line in the steel at that -nterface forms a non-zero for the function H* for the case x = 0, we obtain angle i with the normal co the interface, ; non-zero an expression for H* for x * 0 by replacing in Eqn. tangential field component |Ht|- sin a |B|/u exists (1) iy by z = x + iy. Because H - 0 in the mid- at that interface. Since that field component is plane, the coefficients an must Be purely imaginary. continuous through the interface, it is this compon Since the deviations from a uniform field will get ent that modifies the field in the vacuum region. smaller as one moves from the left side (x=0) or the Usually, the field produced by this tangential field right side (x = xj) of the magnet toward its center, component llt is most easily understood ifnne replaces the terms with n <: 0 fjn contribute significantly Hf by a current sheet of strength I" = |lltl on tne only on the left fide of the magnet, and those with steel surface. It is easy to see that the current n > 0 only on the right side. I* is therefore con in this sheet points out of (into) the paper plane venient to rewrite Eqn. (1) in the following way i.f the steel lies to the left (right) of the vector St. The fields produced by these current sheets are in turn often most easily obtained by applying again 2 !i,|H the 0AM. H*= i(h„• ? h i." ™/* + I he !"»). 0 n=l "n nil n (2) 3.2. The Direct Analog Model (PAH) The h in this formula real and h represents In discussing magnetic fields produced by mag p are 0 netic charges (See Sect. 4.3), it is clear that the central field value. scalar potentials can be used to describe the properties of the fields. In cylindrical geometry, It is clear from Eqn. (2) that deviations from the radial dependencies become clearer if one uses a uniform field due to the truncation of the poles again a two-dimensional conductive sheet model with or by the manner in which the magnet is excited de the appropriate conductivity and current injection. cay very rapidly if the pole edges are shaped in The equivalences between the quantities in the DAM such a way that h_i = hi = 0. That is, of course, and a cylindrical magnet are given in rows 2 and 3 what one does when one shims a pole. A wery sophis of Table B-l, with o representing the conductivity ticated pole contour would also make h_2 and h2 zero, or very small compared to h » but trying to of the sheet, and q the magnetic charge density in 0 the magnet. cancel higher orders would not be worthwhile. It should also be pointer out that the dependence of Hy on y can be used to mturure the low order hn 3.3. Expansion of 2D-Dipole Fields into Exponentials with relative ease by usiiij appropriately made stacks of measuring coils. Fig. B-3 shows the cross-section of a magnet assumed to be sufficiently long in the direction perpendicular to the paper plane so that the fields If the excitation of the magnet, or the trunca can be considered two-dimensional. We assume fur tion of the poles, violates mid-plane symmetry, one ther that the magnet has mid-plane symmetry and will also have midplane-antisymmetric field?. For that ii = ••> in the steel. As a consequence, the such fields Hy is an odd function of y and Hx an fields are perpendicular to the steel surfaces and even function of y. A consideration similar to the the midplane. one made above gives for midplane-antisymmetric fields: To find an expression for the fields that is suitable for our purposes and is valid in the region H* = l bne*(2n+l)z/g {3) bounded on top and bottom by the flat part of the pole, we first look at, and graphically represent, how Hy depgnds on y for x = 0 (see Fig. 4): In this formula, the bn are real, and one can write Since div H =• 3Hx/3x + %/ay = 0, and Hx = 0 far H* of course in the same fashion as Eqn. (2). Com y = 0 and y = ± g/2, it follows that 3Hy/3y = 0 for parison of Eqn.'s (2) and (3) shows that the most y = 0, y = * g/2. Mid-plane symmetry requires that slowly decaying terms in Eqn, (3) decay only half as Hy is an even function of y, and Hx is an odd fast as the slowest terms in Eqn. (2). It should function of y, leading qualitatively to functions finally be pointed out that Eqn's (2) and (3) can -73- b'.' applied to manv magnets other than dfpoles (e.g., 1n that region will be of the same order, and their strong focussing dipoles, quadrupoles, etc.) by effe.-t on the field at location **i will be cut down conformally mapping them into dlpoles. 2n r r by another factor e~ < 2- l)/9# We conclude from 4. Magnet Design Consideration?; this that the effect at r = r; of moving the steel shtM to a finite value r2 is of order 4.1. Properties of the Ideal. Lumped Solenoid H e-4n(r2-n)/gf ie fn most cases negligible, provided (rg-r^j/g > 1/4. The ideal lumped solenoid is (see Fig. B-l characterized by )t = ~; by a steel-vacuum interface If we now "lump" the return flux shell in the along the dashed line on the right side of Fig. 1; azimuthal direction Into "vanes", the details of and by a cylindrical shell at r2 = » to conduct the the field modification are more complicated. If flux entering along the dashed line around the we assume first that the endplates still extend to solenoid. r = o°, the periodicity 1n z is not affected by lumping the steel in the azimuthal direction, We For symmetry reasons, the distance between the therefore still expect field modifications at r = left boundary of the problem and the center of the 47,ft r H of the order H0e" '2~ l)''9, but there will be adjacent coil is g/2, and we assume the same dis some dependence on azimuth. If we then reduce the tance between the dashed line and the center of the radius of the endplates to r = r2, there will be coil adjacent to it. This means that we are dealing some additional field modifications. Although the with a periodic system or fundamental period length periodicity in z is destroyed by this process, these g, so that we need to treat only one cell, indicated by the dotted lines. Because of symmetry the magne additional fields will also be extremely small, tic field must be perpendicular to these lines. The since the pre-existing fields at the endplates for line going through the center of the coil is also a r "•> T2 are very small for a reasonable number (^ 6) line of symmetry, with the field perpendicular to it of vanes. While it is not too difficult to pursue as well. To get z qualitative fealing for the th*s argument in more detail, that treatment would fields in the cell, we use the 0AM (see Sect. 3a). go beyond the scope of this report. In it the dotted lines represent insulators, the space between them has a resistivity proportional to If, for a given distance g between coils, it is the distance r from the center line. The current necessary to obtain a decay of the Fields that is injected over the cross-section of the coil flows in 2 r r p faster than e" ' ( - lJ/9 outside the coils, one an easily imaginable pattern to the axis of the can accomplish this by putting vanes close to the system, r - 0. While the current density in the coil and separating them in the azimuthal direction 0AM is quite non-uniform in the immediate vicinity by a distance D somewhat smaller than g/2. Under of the coil, it will obviously smooth out very quickly as one moves toward the axis. A small these circumstances, eqn. 3 of Sect. 3.3 will fraction of the injected current will flow in the approximately describe the decay of fields between vicinity of the symmetry line radially outward, vanes, giving there a most slowly decaying term then turn toward the axial direction and, close to proportional to e_7Tflr' , with Ar representing the the dotted line, flow radially inward. It is quite distance from the inside edge of the vanes. In apparent from this picture that the fields get weak this case, the proximity of the coils could cause very fast as one moves radially away from the coils. local saturation problems in the vanes that would If one were dealing with two-dimensional fields have to be studied in more detail '.ian can be done Instead of cylindrical fields, the expansion of 2D here. fields Into exponentials (Sect. 3.3} could be directly applied and would yield the result that the dominant term in the description of the field 4.3. Effects Caused by Removal of Steel from Region 1 for r > n, and of the inhomogeneity for r < r-\t is When removing the steel from region 1 (see proportional to e*27rlr-ri"9. A more detailed 2D Figure 1), i.e., the space between the dashed line model calculation shows that the constant of pro and the actually desired steel contour, the field portionality for the dominant term equals the aver to the left of the dashed line will obviously be age field value H0 on axis. The fact that we are reduced over some distance. While this field re dealing with cylindrical geometry will obviously duction can be partially compensated by a coil close modify the details of the decays, hut not their basic to steel surface a, the use of that coil does not character. Using again the 0AM to see qualitatively accurately sipulate the field modification caused the difference between the decay 1n cylindrical by the removal of the steel. Instead, we use the geometry compared to 2D geometry, the fact that the following procedure: He first put a magnetic sur resistivity 1n the model is proportional to r leads face charge onto the steel-vacuum interface indica us to conclude that the field outside the coil ted by the dashed line to the left of region 1. If decays a little faster than indicated above, where that surface charge equals the local surface magne as the field Inhomogeneity inside the coil will tic field value B everywhere, removal of the steel decay a little slower. does not change any magnetic fields. We therefore obtain the magnetic field modification everywhere 4.2. The Effects Caused by Lumped Return Yokes by removing those magnetic charges or, equlvalently, by adding the same charges with opposite polarity If we bring the return flux shell from Infinity all along the dashed line to the left of region 1. To see the resulting fields, we use the DAM (see to a finite value r?, we bring steel into a reyion Sect. 3.2 and Table 1), with the vacuum region where the pre-existing field is of the order represented by a 2D sheet with conductivity propor- 27l(r2 n,/g H0e ~ . The resulting field modifications -74- tional to r, the steel modeled with perfectly con same polarity as the solenoidal currents, and the ducting electrodes, and the added magnetic charges current in the sheets at surfaces B and Y has the represented by current injection from the third opposite polarity, lie apply now the 0AM to see the dimension. Visualizing the resulting current den resulting fields and find the following: Saturation sity and electric field in the 2D sheet, it becomes effects on surface a will increase the field in evident that along the dashed line, the axial field region 1 and its vicinity, while saturation effects component will be very close to 1/2 of the pre m surfaces B and Y will reduce the field over a existing value, and that there will be an increase region that starts near the left edge of region 1 of the fields all along the inside boundary of the and has an axial size of the order of the outer endplate. It should be noted that the increase of radius of surface B. the conductivity with r in the DAM indicates that the field modification far from the axis is smaller The fields produced by saturation on surfaces for cylindrical geometry than it would be in 2D 6 and Y will also have a considerable range in the geometry. If the hole that is opened up is compar radial direction. However, the increase of resis able to the coil radius r], the fields outside the tivity with r in the QAM indicates that the radial coils will clearly increase substantially in the range of these fields is smaller than it would be end region. in 2D geometry. 4.4. Effects Caused by Removal of To keep the proper perspective, one has to Steel from Region 2 remember that the saturation-produced fields are excited by exceedingly weak current sheets unless To obtain a qualitative understanding of the one drives the steel extremely hard. field modifications resulting from removal of steel from region 2, we use the same technique as in REFERENCE section 4.3, and should, in fact, apply it simultan eously in che following way: first place appropriate magnetic charges all along the dashed line, remove B-l. A. Asuer, P. Bossard, Ch. Iselin, the steel, and then place charges with opposite Proc. International Symposium on Magnet polarity on the dashed line and find the resulting Technology, Stanford, 1965. fields with the DAM. Since in region 2 the new steel surface is quite close to the old one, most of the injection current in the DAM will flow to the electrode surface representing the new steel surface. Combining that with the fact that the pre Table B-l. Equivalences Between Quantities In existing field behaves like e-':™r'9, one can con clude that the removal of steel from region 2 has Cylindrical Magnet and Two Analog Models only a small effect on the field distribution. 1 2 3 j 4 |S| 6 , 7 Breaking the endplate Into radial spokes is a ! more drastic modification. Its effect can be vis 1 OAH p ualized also with the magnetic charge - scalar pot -h V - ' e3xj e3xt ential surface representation, and it is qualita tively clear that the effects are small provided 2 Cyl. Magn ' m rA rq V , i) ; rB one uses a reasonable number of spokes (;> 6) and \ does not start the spokes closer to the coil than - -} \l'. f . J g/2. 3 0AM o - 3 4.5. Effect of Saturation of Steel DEFINITIONS: By providing sufficient amounts of steel, sat p = resistivity in the 2D sheet uration of the vanes and the outer part of the end plate can be reduced to any desired level. In the o = conductivity in the 2D sheet region where the magnetic flux enters the steel ne j, = current density injected into the top of cannot add steel at will. Consequently saturation J the 2D sheet will occur there above a certain level of excitation, and we want to discuss only the field modification e, = unit vector perpendicular to 2D sheet plane caused by saturation of this part of the steel V = scalar potential in 2D sheet structure. To do so, we use the method outlined in section 3.1. j, t = current density and electric field 1n 2D sheet It is clear that the flux entering the center part of the endplate (surface 6) will produce field j. = excitation current density in magnet lines in the region bounded by surfaces a,B»Y» that q = magnetic charge density in magnet are in the same general direction as surfaces a and Y- From this follows that the tangential field com u = permeability of steel ponent at surfaces a end y will be considerably A = vector potential in magnet larger than iis value at surface 6- Representing these tangential field components by current sheets, fi\$ = magnetic fields and looking at the polarities of the current in the solenoids, the fields, and the current sheets, it r = distance from axis is clear that the current sheet at surface a has the 1 *r-^ f 1 ~"Tmn ' /,<• CI • If / f f 1 (••H Fig. B-l. Lumped Solenoid Magnet Hg. B-2. Continuous Solenoid in Steel Can a) fi-Field in Magnet b) j-Field in 0AM \ KBL75I2J)!J0 Fig. B-3. Dipole Cross Section Fig. B-4, Dipole Field Components H , H vs. y. USE OF HIHCRLT): COILS IN AXIAL Fian spucrwMm-RS John U. Taylor and W. A. WenscI Lawrence Berkeley laboratory University of California Berkeley, California 94720 Abstract 3, The support structures for the mapnet and cryostataremore complicated than for the con A study of possible configurations for dis tinuous solenoid. crete coil axial field superconducting magnets has been carried out, with emphasis on parameters that This report is a study of the magnetic and affect magnetic field strength and uniformity as aperture characteristics of discrete coil spectrom well as the fraction of open aperture which can be eters. We are not concerned with detector design; made available for external detectors. It is found rather we will examine particularly the parameters that field uniformity is sufficiently good for most that affect the strength and uniformity of the mag purposes if the coil separation is less than the netic field and the amount of open aperture be coil radius. The ratio of field strength to tween coils. fractional los. of open aperture is independent of size for suitably large magnets, but is sensitive to the shape of the coll cross-section. GENERAL DESIGN CONSIDERATIONS General procedures for the design of super INTRODUCTION conducting coils with materials of the type con sidered for use here are discussed in a series of There are some important advantages in the reports by M. A. Green.^ For each coil intrinsi use of discrete rather than continuous coil units cally stable superconductor with a relativelj- high in an axial field spectrometer for a colliding superconductor to copper ratio would be wound on beam facility.* Assuming that a Wgc fraction of and epoxy-bonded to an aluminum frame. Closed the aperture can be left clear, i.e., with at most circuit refrigeration would employ two-phase helium a few percent of a radiation length of material, in cooling tubes welded to the frame. The con these advantages are: ductor would be tightly coupled electro-magnetic ally to the aluminum, which, in the event of a L. Secondaries passing through the open aperture quench would act as a high conductivity shorted are more reliably identified because there is turn to dissipate most of the stored energy of the no absorption in the coils. This is most im magnet in a time of some seconds. This process also portant for charged hadrons. would distribute widely the thermal energy, causiig the quench to spread. Because of these effects, 1. Secondaries passing through the open aperture large local voltage buildup in the superconducting are more accurately measured by calorimetric coil would be avoided. methods because there is no absorption in the coils. This is important for hadrons, y's and Figure 1 shows schematically possible magnet e's. configurations of the type considered. The coils are held together by axial support members with 4S-degrec symmetry. The thin vacuum windows are . Seconardies passing through the open aperture supported by a periodic support structure lined up may be more accurately measured in momentum be with the coils and the axial coil support members. cause the resolution obtainable with precision The coiling tubes arc not shown. Two kinds of coil detectors outside the magnet is not limited by cross-sections, rectangu'..." and elliptical, are coulomb scattering. considered. For reasons of simplicity and cost we will want Lo use modular elementary units; but it is obvious that the use of several units of dif Backgrounds from interactions in the coils are ferent cross-sections can reduce considerably the absent in the open aperture. This circumstance loss of aperture. Examples arc shown in Fig. 1(b) can be used to prepare a clpaner event trigger. and (ci. Disadvantages of the open magnet structure MAGNETIC FIELD 1. Because the field is ncn-uniform, algorithms We have studied the field configuration for for the reconstruction of events are more com a half-cell with the geometry shown in Fig. 2. We plicated than for the continuous coil. have used TRIM3 to find the distribution of the field for nine combinations of dimensions using 2. For a given average field the discrete coil solenoid has a larger maximum field than does the continuous solenoid. Therefore, the amount 1.1 0.4 of superconductor needed is larger for the dis 0.0 crete coil solenoid. 1.0 ''L^L. • Axis of symmetry • — Plane of symmetry XBL7512-97B8 I XBL7512 97B9 Fig. 1. Schematic of possible coil configurations, Fig. 2. Dimensions of half-cell studied using a) Single coil module; b) Compound coil TRIM. R and Rfe are the radii of coil and structure, n is the luanber or radially dis iron return yoke, respectively. 2W and 2H are placed subunits. c) Alternative coils at the axial and radial widths of the coils, re different polar angles. 2 is the separation spectively. S is the coil area. I is the between coil packages. current per coil. The permeability of the iron is taken to br infinite. Z is the dis tance from the ce iter of the coil to the center of its virtual image in the iron In each case the coil dimensions were poletip and is the effective coil separation in a multicoil system. J-g- 0.05. -^ R, Uf ,Z,W, and H are defined in Fig. 2. The permeability of die iron was set equal to <*. For parameters sensitive to the details of coil shape and size we have used approximate analytic expressions. Field Uniformity in an axially symmetric field, assumed for the purposes of this study, the important field XBL 751Z9T30 quantity for measuring particle transverse mo Fig. 3. Projected orbit. In terms of the disper mentum over radial range r is B-,r2/2 = r A^, sion d at the axis, the transverse momentum where A,, is the vector potential. Figure 3 shows pj_ is given in terms of the vector potential. how this quantity can be used to infer the particle transverse momentum in terms of the slope of the trajectory at r and the dispersion at the axis. In general for axially symmetric fields (2) where p^ is the momentin in GeV/c and A* is in In terms of the transverse momeiitrjn ?.t is ob T-m. For particles passing through the axis vious from Fig, 3 that (3) It) Figure 4 shows the extent of the variation in 0 n\f with Z at various radii. The effect of the steel return yoke is seen to be small for fields T inside the coils, but it is relatively important where S is coil area and S is the area used in outside. The non-zero radial width of the cryostat Q obtaining eQ from 'HUM. We have neglected the ef prevents detectors from being loc.'ited very near fect of the iron in the small term ^e. Figure the coils, hence the results shown in Fig. A should 5 shows e and Ac as functions of 2/R and S /S. be reasonably independent of coil size and shape. Q 0 Fig. 4. Amplitude (mis) of the variation of rA.. a) in side coil and b) at surface of iron return yoke. Plotted data assumes rAtoax*rAmin Amplitude frms) •^"W+ rW Stored Energy XBl 7512 9782 For a thin continuous solenoid the stored energy Fig. 5. Stored energy e = CQ + Ac. eQ is the per cm length is given by volume integral of the energy density for a rectangular coil of area SQ = 0.01R2. Ac Bl R2 is the incremental change m stored energy li Cergs) = -\~ (emu) (5; For a coil of area S. Ae = (Z/^R)Hn (S../S). 0 The curves arc accurate only for Z -•> 21V. where the average a.xial magnetic field B is given Maximum fjold on_Coil_ (6) The muxhiam current density in the super conductor is limited by the maximum field on the The average axial field in the case of discrete coil, which For syrnmetiic coils occurs at the inside coils is give/i by center surface. The field here, relative to th? central field is given by The stored energy li for discrete coils of the same (10) nidius \v written in terms of the ratio (8j where we neglect the effect of the iron, which is c0 shown in the cases considered by TRIM to be 5 10 per cent. Under the conditions where C[), which was evaluated using TRIM, depends pr i- marily on the ratio Z/R, and Ae depends on coil size and shape. We have estimated Ae analytically, « 1. [,fj « under the conditions that l * — - 2 F • i tan"1/!'! INT " JET J5 using the field distribution for a circular coil n Z21 cross-section. This gives where 1= is a form factoi for coil shape which de This is shown in Fig. 7. Obviously, a rel pends only on the ratio W/H. See Tigs. 6 and 7. atively smaller field on the coil or a higher For rectangles central field can be obtained if the coil shape is extended cither axiully or radially. (U) For ellipses in general we have not calcu "/nil "3T [-ml lated F. The W/H dependence will be similar to that for rectangles, however, and may be normalized to the case for circular cross-section (F = 1) shown in Fig. 7. SUPERCONDUCTOR For purposes of this study the use of an intrinsically stabilized material like Supercon 2&2E9, with copper to s/c ratio about 1:1, con taining about 3000 strands of superconductor, is assumed. The critical B-i characteristic2 is shown in Fig. 8. We assume also that the magnet operating point will lie on the dashed line, derived by reducing by a factor of Z both field and current from the short sample characteristic. This conservative position allows us in what fol lows to neglect the space lost to insulation and to make the slightly optimistic assumption that "*'s XBL 7&1Z-9793 the superconducting material is distributed uni formly within the coil area. 6. Maximim field on rectangular coil (inside center) relative to central field for dif ferent aspect ratios of coil cross-section. In the region of interest the magnet B-i characteristic may be approximated by a straight (See Fig. 2). The field ratio m=m0 + A m. line, i.e., Bm - Bc - J i (13) where B = 5.25*10G and -3 is the slope in emu. The load line slope is given by D0 4TnnSf = (emu) (14) where f is the fraction of the cross-sectional 1.0 -i 1 r— P*~—i r- area of the coil occupied by superconductor. J^ ~*»0" Circular cross sec 1-f is the fraction of aluminum. Solving (13) and (14) we find for the operating current 0.9 - (emu). (15) OB V A - O.V Rectangular coil \ ~ 0.6 - 05 i i i i I.I XBL 7B13-97BB Fig. 8. Current-field characteristics. The solid Fig. 7. Form factor for coil. curve is from Ref. 2 for SUPERCON 252E9. The dashed characteristic is taken to be the locus of: operating points foT a magnet. The load line is of the form B/i = 4irSf /Z, for which the parameters are defined 9n the text. The centra] field B.. is given by where Z is in an and S in an". DECAY TIME CONSTANT It is essential that immediately following a local quench the decay time of the magnet current be long enough that the energy dissipated in the coil forms has time to raise the coil temperature elsewhere, causing a general quench. The decay time constant is given by . Inductance _. 2E . Resistance Power = 2TTR o S(l-f) (H) XBL 16119796 where t is in ns and resistivity p is in ohm-an. !). Heat capacity and final energy density in Solving for S coil. Effective heat capacity "for coil with superconductor fraction f is Vf = (1.5 + i")C... The temperature rise i* given in terms of energy density ll. f) In practice the resistivity depends sensitively on purity, fabrication and heat treatment. / In a quench essentially ;•:: the stored en ergy 8oes eventually into heating the supercon [•or ellipses, K = [ J ffj + [} ductor and coii frame. In terms of the (volume) heat capacity and the initial and final tcmper- tures For rectangles, K «/fj +/^R » { r HT - Stored energy _ Q e |ergs) ,.„•, C dT C19 n is the number of subunits into which a I eff - Coil volume " T^S M" > compound rectangular coil is divided (see Fig. 1(b)). Figure IJ shows K as a function of various angular variables, including solid angle, rapidity where B„ is in Gauss. and L = cost) + cos" fj Ceff °bv'ously depends on f. C , shown in Fig- 9, is a composite for aluminum and copper Using only Fig. 10 we would tend to emphasize (used for the superconductor), as is the integral, small values of Iv'/H for large solid angle, ;ind defined by assuming that the temperature"dependence large values of IV/M for a detector s :isitive to a of the heat capacity is the same for copper and wide range in L and i\. With the compound coil or aluminum so that with two or more different coil shape* (see Fig. 1(c)) optimization over a wider range for dt « is BnRZe possible. Energy density = (l.S + f) :TolfS~' (20) CHOICE OF MAGNLT PARAMFILHS There are many experimentally sensitive fac LOSS OP APERTURE tors that need to be taken into account in the design of the magnet. The important ones considered In terms of the various dimensions defined in here are: Fig. 2, the fractional aperture loss in 9, defined 1. Field uniformity as the opacity 0, is given by 2. Field strength or momentum resolution 3. Opacity. 0 = ^| K (21) W Figure 4 shows that field uniformity depends where K is a function of the ratio n- and the sensitively and almost exclusively on the ratio poJar angle 0. Z/R. Uniformity deteriorates rapidly for 2/R i I. Fran (11), (16). and (21) we sec that Mag netic field and opacity have sbnilar dependences on Z and S, so that a trade-off between nomcitun resolution and open aperture is necessary. To study this in acre detail we define a figure of writ to be the ratio of the magnetic field to the opacity, -1.0 K /£ OERIT) (*-*") (23) Using (li) 2 «I (24) The last two terns are negligible. The second and third terms arc optimized (i.e., minimized) for S • Sm, where , 3Z (ZS) Substituting (25) in (24) wc obtain in general Fig. 10. Form factor K for loss of aperture. Fcr rectangles (straight lines in figure) and ellipses (curved lines) sec Eq. 22 in text, (26) n is defined in Fig. 1 for rectangles. Solid ^[v&K^j lines arc for n • 1, dashed lines arc for n • 2, and dotted lines are for n • •*>. This function is shown in Fig. 11. I ' I ' I I I l I l 1 I l I I I I Fig. 11. Ratio of opacity to magnetic field strength vs coil separation. Parameters are defined in the text. 1 m-tensitive to thy r;ilK» S/S .tlul to the si.:e of advantage of the tit:.vjiMt ivn* m i !>i; rln- -xignet for :!•-!' > 50 en. IW lost ot aperiiire tin* .irea 01 the coil .'.in he \aiu-.! '••< .• •ittrol tV i*- very sensitive, however, to I- ind *t. i-«r M *l'i • li'M}',!! Meld over .1 -itM.lhle t\Hi>;v. V'ti that t:ir «t- see tlun the loss of .ijterrure i< ,il \nuiltet" constraint is pion.ii.i !>« deuMty allowable lit t!u- coil. 1 .jii.it -i;: he rewritten i or ; • •> seconds, . * 10 oh>T*.n !' •'" the nvn temperature value), ue fiiwl Kp - InO en. I^r a niniran of the ratio I'-'h,. the najtnel it field, as determined Iron 116), '-'M, .-~l and T / eit rc,U I * !•• i:/f(t • M./cKli C " where h, m Joules/os' . t> AIPSS) allowable enerj;v density. figure 1- shows 1\, for different \.t»«e" .»?' U. tins constraint is snil.ir :o that luiiler this ttiiunutt condition. Ihesi- fields ah' in l .'(•>, except thai it provide a \ r 2 n o-t r- / It 0 0 iN!***' ^"^^•*—£ ^""^"•,- B| - „ H,F? 100 m 1 1 1 1 rig. 12. Magnetic field obtained under conditions lhat maximise the ratio of field strength to opacity. The parameters are defined in tlie text. -83- fur present purposes h' is soKwhat arbitrary, we 1=01- R » Rt, this gives a thickness of about cne- note thai for W • 200, T • - 100 degrees centigrade. half en or one-third radiation length. And for 2/H - 0.4, K - 2, this gives In this report we have ignored many en gineering considerations. Kn have not considered -,r- •: 0.1 Tcsla** (31) the loss of aperture from the axial coil supports. It appears that this loss will be comparable with similar to the range established in (26) by the that from the coils themselves. The determining magnetic properties of the superconductor. factors are primarily structural. Other factors affecting the choice of raagnct parameters arc de- lu Mamary, reasonable field uniformity re lector valine, time of flight, diameter of bfiam quires that the coil separation not exceed the coil pipe, cost and problems of fabrication. Kc have radius. Kith currently available superconducting considered none of these. Material, the ratio of aperture loss to field strength is asymptotically Minimized at about 101 Kc are indebted for a number of essential dis per TVs la for coil separations greater than .ibout cussions to P. H. tiberhard, M. A. Green, and SO cm. The conditions on field uniformity and li. S. Groves. aperture Joss per Testa arc* therefore, well satisfied for magnets in the range of interest of Mil' spectrometers. The characteristics related to RlirtRKNCES momentum revolution and aperture improve rapidly with magnet s»;o, because (in the absence of 1. A Proposal to Study High l*i Physics with a Large cojiloah scattering) resolution involves a factor of R* tines the magnetic field. Coil shape is an Aperture lladrori Spectrometer at the ChRN ISR, important par ^..~tcr in the optimization of ;i given British-America;!-Scandinavian ISK collaboration; magnet configuration. O-RN/tSRC 75-18, U Mav, 1975, G1HN/ISRC 7&-21 :s June, iy:s. Assuming that coils with the properties im J. Michael A. Green, Cryogenic Engineering Con plied here could be built, the total amount of ference, Kingston, Ontario, Canada, July 22-25, material required would he small, From (25) and 197S. LBL-!ig?7. Also M. A. Crecn, Mechanical W"), for exa«plct we find that, under the mini - Engineering Note, Code ABU1G3, Serial M4853, on condition, the average radial thickness T of 28 May, 1975. coil material is given by 3. J. S, Colonias, TRIM: A Magnetostatic Computer X - S 'I - (V20 (1 • R /cR). (32) Program for the CUC 6600. UCRL-18459, August tf ( 29, 196B. PEP 191 •mi . -pfi v . -, •; W. I'm". :. . i-il.-. , A. M i- -, :. .:. I., b. HiM in, '.!. K -i . . i.. t.i.; Unm-, I). l,iIK« , M. M-ir.hik, T. Maut J. [.htth.-wi'., C. Iv-k, K. atriu-h, D. Yount iimnm&rv w.'tLi t - posed at SPEAR.'"- I'IIC1! nr- -ipparatus muiit -ontain sufficient matter if ubsDrb most of The range of physios problems fur vhl'ih a the energy of the final state particles, both (etoctor emphasising neutrals Ir. n":". :;ul table is chaig^d and neutral, must cover essentially iiscussed. The prinnry ,joals are U.-- all neutral:-. **~ solid angle, and must have high efficiency for •f-'.T:! section, a (e*e" * neutral:;!, :he .'burncter- detecting the presence of charged particles to i.:ation of the neutral energy In mu!:. i-hadronlc isolate the all neutral events. :'ince all y de • •vents, the search Cor monoenervoti • jh.'tons, and tectors (except pair spectrometers) involve a con- ^ood sensitivity in the difficult iv^ion »r low aidernblff amount of matter, it is natural tc in energy photons. corporate calorimetry and high resolution photon Those features of multl-hadroni.- events detection in the sane apparatus. Considering the which are most relevant to a neutral ietector have importance of the value of a, it seer.s very de been calculated us Inn u Jet rr.oiel wi*h parameter:; sirable to measure it by both magnetic and calo- extrapolated from SPEAB energies. These distri rimetrle methods since they have different biases. butions are presented and dlscussei. Crtloriinetry Is the only practical method we know of i\n' measuring oQ and this component of c, may I. WHAT TC STUDY contain reaononces or threuhold steps which are hard to observe in a, itself. Much leas is known about the properties of -.he neutral particles emitted in high energy re Ft has been found? at SPEAR that about half actions thuii about the properties cf the charged of ths total energy available when two or more particles. The reason Is obvious: neutral parti charged particles are detected is emitted as cles are harder to detect and their properties are neutral energy. What is the nature and what PT<; nore difficult to measure. The neutral energy can the pr<~pertie;i of this neutral energy' The;e are appear in the form of photons (decay and single"), many detailed questions whose answers are expected neutrons, neutral kuons, and neutrinos. In e+e- to be very interesting. [and >>) reactions, most of the emitted neutral (a) What fraction of the energy is tmitted energy is believed to be electromagnetic und as n0l5, as n°'s, "... single Y'S, US neutrons, as primarily from -r° and 1° decays; tittle in 'is ye I other neutral particles? How J^ ^hes-? Tractions knovn about the overall importance L>I" the recently vary with energy? Are there ro;;r.tiances yr iiscovered single photon emission process. Thus, tiii.".::!:'-1'* steps? ietectors emphasising the measurement <;" photons (l>) WJmt inc the inclusive spectra <>( >'s, ire :i»*>del for the stuiiy of the n-u.ure of the " 'n, rir"s, :ind other neutral part lei t-s? How -lo neutral energy, "eutron and V.'* •t"» " •' ]'•:: 1;: 'ie- they conjure with the charged pai-ticles; -ir-.tV- "or -J "•• -f the Jefii!-! i r.-.-.-' ; ,-.i i-.n.n , (cj Are there other monoer.'-ri;e: ic photons in -.:: 1 niiy be Lmp-.rT-int at the hitii-: ' ! i': "iK-reie.-,. alditi':n to those recently disc vorei?" In Vhnt Is ur present know!--!,"'- >:' 'lie neutral ctise'tl" iecays involving sort- tium .me y, what is • •t.'-rgy emitted in e+e- • haun •. ., -n: t vh'i* theic ' in-.- .-jrier -jf emission-; r. :;tion:i •tri:'.-V First, th'M-e i:- • •-•...•i.t.i-ii ly n- fhir.K Kn-'Wri f-Kporirnentally aWiV • :.• -out.rLI-n- •; .11 tr. (..*.- - hadron) ff ali n-".-:-ii final - •:• iteu, .-.i!!(.- [!•'.;;,.nt measurer-•:.' " '. rely ' n .. . w.-.-t.-t-v-i-part i.-le trig,--T. ;;. -.11" ::••"> :-'!\ • .nU-Llmti Ji:, -1^. is helli-v.-il sm.iti T.IIISI- * he obvious reactions e e- ' |f' *• '|'." l|'.'i '- I,-'...I .ire forbidden hv i-hnr^c conltij'..it ion •ivnnin-rrv to I i rit order in the elei'trniruittm-t !• . oii|»l IHR. How ever, other rc.ict ions SHHI .is e <--,*- \>f' >• i|-' , n t n f pr' , en-., are allowed, .-iml ri-membcrtni; th.it Ihr ufe- .'innllii Lit Ion process h.is shown unexpected i.,iiy:: ]>•-• involving identi fical i.'n of -li'm-e-i hehivior before, we believe It is lmt.,.rtant to nvthe h-.dr n:i in the finnl state, tln.-ir cnrrclat ion with dlreit experiment.) I tests for these possibilities. tin- n-iitraln, and the compl'-tc re^onnLrti'M ion of ,\ direct .tppro.ich to the me.i'iiiri-raenl nf both .:•.(-•,.• '-:-isr:PG of fi::-il stater,. And, <>r ^•••urse, : .ind o Is the rnlorlmetr'.c method uslnn n tot.il t-veryli'-iy is looking hard for nr-w, unexpected phe- .•i»Ttjv trij?t!°r- Th!s wan omph.ts i/nl hv Keldnvin ti"ii!fn:i, and w are excited by Lh<.- prosp'.-cl-s of a ind HUlinl for I'KP, .ind .in la.,-.! .-mei.i .it i.»i of It j. ti.,;(fi- with sensitivity in an unexplored >-egion. (y r.iiiKle- phn' ,n:; are meant, j'iiot'iii: ••!•! i t.t'*i! in • r iiiitionc oel-wijon hadronic nl.'it.ic:. II. HOW TO STUDY In this aection, we give the results of an extra polation of this model to /a « 30 GeV, and com It la very much easier to write down the parison vlth the pure phase space isotropic model. interesting features of the neutral energy than The general features of the two models are to devise a practical apparatus which has all of similar. Only pion.: were assumed In the final the deairable features. In practice, compromises state. In the phase apace model, pions (charged must be made involving performance, complexity, and neutral) are produced with momentum and and cost so as to optimize the study of particular angular distributions according to Lorentz in areas. The 197^ PEP Summer Study report contains variant phase space. The total multiplicity is a very useful summary of the properties of photon given by a Poisson distribution whose mean is ad detectors^ and of the design considerations justed to give an average of about 7 charged plons inv"".ved. and T neutral pions. These numbers are extrapola Since everyone expects most of the neutral tions In a + b In* of the e*e" data for SH • £-k energy to be electromagnetic, the neutral detection to t.k GeV.• The Jet model has the same features as schemes we have considered first concentrate on the phase space model, except that an invariant electromagnetic energy detection, with neutron and matrix element squared (M^) is Imposed with K? detection added on in one case, and magnetic detection in the other. We recognize that a gener al detector will surely provide some neutrals de tection, but we ,-xpect practical considerations to limit Its energy resolution. Our aim was to pro vide a oor.^ specialized capability complementary where the summation is over all pions, P 4 is the to a general detector. Since the 1971* Summer Study transverse momentum of the Jth pion relative to a there has be^n o detailed design for SPEAR2 of an given axis and r is a parameter which was adjusted all neutral detector baaed on Hal (Ti> technology to give lj: 'JltModtl • Jot Modil Li; - • Phott Spoce Modal • PhoM Spoct Modtl Li Normalized to *omt numbir Normoliztd to samt .lumtitr L. of photons ol »*'t _l_ -U- _l_ _]_ _1_ _1^ :ln 1 i 60 90 120 ISO 30 GO 90 120 ISO ENERGY (GeV) , MOMENTUM (GeV/c) XUL IV) 'J301 FIK. 2 Flu. i ver sine, eaeh photon 'lis— *-'er -i tiuii-ncglijjlble portio the ilatyihutiori n!' unKles ANGLES BETWEEN PAIRS OF PHOTONS J] s- ./s=30 GeV [ '-. \ ,-•-. / Angle Subtended by 90% of Energy in lr — = Jet Model -I = Phosa Space Model (• H Land Normalized to same number of photon pairs ; 54 72 90 108 126 162 180 ANGLE (DEGREES) XUL /5987CM Fla. 4 for au enhnticer.v.in ui nraall anplec Out to high energy i0l3. tin- phase space nod©1, yle'dn a nearly isotropic distribution, aa one would e>pect. The -l " r -1 1 transverse mc^ieiiljra damping about tlio J5t axle, however, ciunirests Itself very strongly ITJ these ungulur correlations. In fact, the amnl'i angle spike from r^'s in the high enerty tail Is not - even visible, being buried under tue p;reat multi plicity of snail angles between photons in the .let. The figure also 3hnws the angular nue fif an area - \ vhk-h fnntninn about 90% or the energy ->f n single photon !'nr the two neutrals detectors reported on / FWHM • 16 G«V \ ID the suncer sty-iy. Clearly, the .let structure !.•: a serious natter in neutrals detectors. About hull' the pairs "n nne hemisphere lie within a cone ul' hali'-nrwle 11°, whieh apreej well with the V crude e-tinate of 1 I 0 6 12 18 24 30 EBM„M (GeV/ j_. - (multiplicity) * «". Jet Model-Charged/ Neutral Energy Distribution Thus, ir the Jet nodel with transverse monentu limited to about iOO MeV ia applicable at PEP, MM. 7J9 8L>< and there is every rea3on to believe It is, one Ftp. 5 can expect about ~i or i* photons In about 1? of U. This in approximately the size of a single shower in both the crystal hall and LAKP (Liquid simple, direct measurement of their angular distri Argon Neutrals Detector). A typical event would bution and electromagnetic component. Fir ally, then contain two such neutral energy blobs dia- the goals of measuring oQ and ct by calorimetry cel.ricnlly opposed, each blob containing ;?, and of delineating the general features of the perhaps 3, charged particles in the neutral energy are not disluroad by this structure crystal ball. In the LAUD detector which we of the multihadronic states. studied, the magnet would sweep tue charged parti cles aside and disperse than over apace, signifi cantly reducing the charged-neutral overlaps. REFERENCES Of interest in oalorinetry are the neutral 1. G. Peldman and D. Hitlin, PEP Note 1^3 (197l< and charged energy distributions. These energy PEP Sunr,er Study). distrititions in the Jet model are shown in Fig. 5. 2. Stanford Linear Accelerator Center Proposal Of course, the equal partition of energy between SP-2li. cnarged and neutral pions Just reflects the param 3. See S. Schwitters, 1975 Pheton-Lepton Confer eters fed into the model, although the well-known ence, for the most recent data, energy crisis supports it. I*. G. Feldman and B. Wiik, 1975 Photon-Leptcn In conclusion, we note that the expected Conference. Jet structure at PEP energies will maRo the energy 5. E. Bloom et al., PEP Mote 155 (197A PEF Svunner aeasureaent of single photons in multihadronic Study Report). states very difficult. 6. T. Hast and J. Nelson, PEP »'ote 153 U971* PEP neither or t>e neutrals detectors Summer Study Report) we considered ia able to solve this particular 7. W. J. Willis and V. RadeKi!, Hue. Instr. and experimental problem. On the other L^rd, if nature Methods 120. 221 (l97u, does Indeed provide us with auch a stro-.ig focusing S. G. Hanson, 1975 SLAC Summer Institute, and effect, the neutrals detectors will pro'ide a R. Schwitters, 1975 Photon-Lepton Symposium. BALI. .V. IKI T..H.- I. ii-oi.«rti ;t.- r.-.i::'i -itlons * > *ho .Tf'AH ven- L.i, m1 • li- Jryj"-.. r:.LI :-.• i-iircd tiy i*,c hi»:hor e-ntsyioa -t' MliT.-H:i PtT ar- liac ;:•:-.-:. £Um-.- the t.'i iro:: mult i j.l I .• i • ;.• is ox]--'It-1 !-• ":::e a:' '. r: :', ltn-ir avera*-.- i-[.o«,-y niuaf ;•:.;«.•. t; 'ho oti.er irtnl, i!' the halrvm; -ii— Signal proi'K'fi in .'i'*;;, the i..v ••ii-:>*y part .•:* their | -..". i.wf decay tir. jSpeoti-iT. ;s I..'. l.faviiy :e::lele1. This ir.pl it-:: Ii.-ji.ieA'utai i>- t-iHur. ;>. tha* roJifica* ions for \Af;h eii.-r*:;/ particles i-h.-ul no', ie'.yriori" v l.>w energy perforaanoe. An ex Jolll Angle [Tphen- = J.90 • ternal Iron „•»!.-rimetor for measuring the hi*:h Coverage' t-.ptiere = J."> energy hadronn, i-harged an! neutral. Is ."our: k-:v-1 ,1-vA ,'artJ _ To ir.provn the angular rei u 1 at i „•« on >'::, an ac'.iv Thi.'Kness • f :ih-" internal inverter has l-oen studied, estimate;: have b^er. nado. . t" its exj-ejteJ rei-fjrmaiu-o, an i It fil'.x' :-A. U».:f difficulties requiring further study have boot. " i h-t ir-iil.- i:.'..-!- outline!. .i.-t ;,.ti U-IH,">. ThE k . SAIL r.iturvj:. !ol.ir'.;.£ The Crystal Ball is a neutrals detector ie- Segmentation I ^htipe - * riui.euiu' prisns slgned for SFEAP. Its major component is a spheri j col id angle,.;•»• - .01'.") ar .• 155 cal shell cf Nalf'FO, 16 r.l. thick, with inside ji'ractj'-n of aphoro 'zv* radius jf ?5,.' en, covering about QQ% of the sphere. The irtericr cavity is filled with track chambers .00139 • i'S* to determine the number and directions of charged hall" angle cf .vi.o with sace particles, and the conical benn pipe openings are olid anftU' = -.2° = 73 ca coverei lovn tc about 10" with an array of hexa gonal sodium iodide counters, 20 r.l. long. A summary of the relevant parameters of this device binet ic energy jf ran^e -'•'J MeV pior.u is given in Table I and the performance charac particle | ~'UQ MeV kaons teristics in the SPEAR energy range are given in reT. 1. \j}2'> "•'eV pro*.rms Momentum of rang. l~?50 MeV/o plena In the ^i-'EAR enerry ranpe i Nevcrtheless, It v a * hadrons proceeds via system. As noted above, a separation of one photon parton pair production, then those ;.0,i produced annihilation from >> -reactions where the de backward in the parton's frave give rise to •*"» graded electrons arc not detected can be achieved of low energy. Tnus we qualitatively obtain t»-.- by measuring the total oncrgy of a reaction ever ]ot model result that as s Increases, the dynamic a solid *n>jlo close to J.n. •'he anou.it of rat n sir.- range of photon energies inepcaaesj the relative '.ncrgy then allows an identification of af-reaetio»s. importance of the low energy photons m neutral Further, infomi-'.ion 00 the direction of the clo hadron reconstruction is reduced, but thoy cannot ning nosentu-i, although expected to oe poor, cay be ignored, furthemore, the possibilities of also prove useful. This argument specifically ap Eurpriuos involving low energy photons persuades plies to the total cross-Beetle.*) eeaaureaeat. us th.it this energy dopain nust be well covered. Another foe* In that the ball Is essentially a v The major thru - of our study has therefore been ivtector (It is ftbout 1 n&dronic abscrptlen l»ngth thu design of an ontern*I Iron calorimeter to pick thick.). Ileuirons and Kg*are detected only through up tho unorgy of high momentum hadrons and an nuclear int...•/actions, and so an external caloricetei internal active converter to Icprovc the angular would increase the aaoant of energy collected froa resolution or. the photons ulth an li*-tl» etiftrjry these neutral particles us well as froc the charged resolution degradition M possible. particles which do not stop In th* ball. We find that an iron calorimeter is quite Two possible configurations hjve been c.n- feasible but massive (about ^00 tone], even given srdcredj one specifically adapted to the crystal the relative compactness of tho ball. We had ball and thu other using the modules designed by hoped that an internal converter could be added the getvtral detector group and described by thez. " without serious degradation of tho energy reso rron considerations of interaction lengths per lution, but wo fino that Otis uoet not seen pos radiation length, ease of handling, and cost, iron sible because adequate segmentation cannot be pro was chasm in both. vided. h calorlnetcr based upon scintillation count ers was considered first. From the work of Engler, EXTERNAL CAL0RIME7EB ct al. * , we deduce the following] Let us begir- by considering the measurement of ot with the unadorned ball at »T = JO GeV, (Lj Five absorption lengths are needed. Since ignoring any charged hadron energy it detects. the Kal provides one absorption length, the calori first, we note that it is a superb calorimeter for meter requires at least 7^ en of Fe. the purely alcctronagnetic energy in an event and (.2) To achieve a E*HHM resolution of 60S at an efficient counter of charged particles. Thus d.'j GeV, one must sanple every 1 en of iron. Two in the absence of bean-gas and 2*i event?, no ad en between plates has been provided for scintil ditional apparatus in needed to identify the in lator. The geometry of the calorimeter is shown teresting annihilation)?, and the observed energy in Fig. 1 and Scin: 15* of iron 570 tans Fe 15 , « is about O.05 nb *t «T = 30. Thisis 16J of ot (e e -*- hadrons) if R » J. Thus we conclude 80 tons ttiat the unembellished ball could probably be used to measure tf-t , but the corrections for undetected 630 tens Total If events and detected 2y events would be uncom fortably large. Clearly, before using the simple ball in this mode, much more careful background and signal loss estimates would have tu be done Assuming that the light from the scintillators than what we have done in this summer study. We fills X.'j cm of the 2 cm gaps between plates, the,, now turn to the situation with an external calori total area of livjht emitting surface is about 27*0*" meter. which corresponds to the photocathode area of 2.£ l Mil. :vi H:II.I, End Cap replaced with iron and the last ana largest module removed. The geometry is shown in fig. 3> lfc uses 12 of the "standard" liquid argon ionization chamber modules plus four special enJ cap modules (not shown). The details of this design are given in the report of the general detectot group and will not be repeated here. --1 i INTERNAL CONVERTERS ( !) The angular resolution of the spherical part of the crystal ball is determined by the size of the prism segments. Each of these covers a solid angle equivalent to a cone with half-angle of if. 2°, Thus a photon's direction can be determined to about this angle but not to much better than it. In fact, detailed Monte Carlo studies' have shov»i. that for an isolated Y, one can interpolate in side a prism because of the transverse size of a shower and obtain an angular resolution of about 3° at loo MeV and 2° at 2 GeV. In the ball the uncertainty in reconstructing the mass of a n° is dominated by the angular resolution for momenta above about a hundred MeV and in this region, Table II. Calorimeter Performance. in s ^ pm a where p is the momentum of the particle of mass m and &{l is the angle uncertainty on one photon. Wire Plane Kin u mom. to n punch through This uncertainty formula assumes that one calculates [lumber reach plane probability the invariant mass of a y-pulr with the formula &ElKp{l-co3 0)» where ET 2 are '*-ne energies of the (MeV/c) [%) two photons and e, the angle between them. However, Lecause of the size of a shower, each photor. covers a cone of half angle fie, where 60 > A(!. When 600 p > m/60, the cones from the two y,s overlap, rnd 950 other methods than the simple kinematic formula Riven above must be used to calculate the invariant 1300 mass of the pair. Figure h shows these relations {including the small effect of the energy resolu tion in Mai) for representative crystal ball param eters. Although these uncertainty hands are broad compared to the masses of the TJ° and n° themselves Another possible arrangement for the calori {which are the usual comparison quantities), it is meter is based upon the modules considered by the more relevant to compare them Lo the width of the general detector group with the Pb front end part fs fraction of 4ir used by I photon f-0.46% f-0.83% f«4.2% 5 10 IS 20 25 30 Number of Gammas \BI.759-83G1 A more profitable approach is to introduce converters in the cavity of the ball. An elaborate system is not possible because of Space limitations, and we settled on the following requirements: (a) Angular resolution of about 10 mr. This is five tinea better than the unmodified ball and Fig. would hopefully {set below) provide a count of the number of Y'S contained in an unresolved energy distribution of invariant mass between pairs of blob observed in the ball. Since the individual random photons with the same total energy. This energies of two nearby T'S cannot be disintangled, distribution ranges from 0 to E2, where E=E}+E2, and the average direction of the total energy has and, for reference, Fig, h showB this range. a 50 mr uncertainty, one cannot calculate a mass Detailed Monte Carlo studies of the problem of in the manner assumed earlier, but must resort to identifying irDfs and n°'s in multi-hadronic events statistical fitting to distinguish ff's from acci have teen done for the SPEAR energy range, and the dental overlaps. results are given in Ref. 1. (b) energy resolution compatible to thai of One approach to improving the angular reso the main detector. This forces one to an Nal lution in to simply increase the number of seg active converter Co preserve the energy resolution ments in the sphere and increase its radius R, In in the 100 to 200 MeV region. Table III shows the addition to improving the angular resolution, this resolution of the system if a lead glass active option provides other benefits, viz., reduction in converter with the probability that two unrelated T*S have over lapping showers and finer binning of the trans f- = .05/«e («fcv) verse spread of a shower. This later is poten tially useful in reconstructing the invariant mass were used. of a pair of overlapping Y's by Walker's second (c) conversion efficiency of about 70$. moment method or extensions of it.6 the active converter aa a function of the total Buil-OJItldt BouinWry number of gammas incident. It is moat relevant to consider N as the number of photons in an un resolved energy blob In the ball so that fl in the 2-6 range is suggested by the Jet model. A poBBible geometry for the active converter is shown in figs 7 and 8. The changes from the SPEAR version of the ball are: (a) reduction of the radial thickness of the ball by 2X,,, this being replaced with the active converter. The inside radius then becomes 2o cm. (b) beveling of the inner edges as shoim to allow for the cylindrical geometry of the converter. Only 1TS6 of the sphere is affected by this modi fication. •rift Chamber 8*0m Pip* Table III. Active Converter Properties Fig. ? 1Xo tMeV) The active converter is segmented into 2U 100 200 UOQ 100Q parts, 12 modules around the circumference, split at the center of the interaction region. Drift Conversion probability .U6 .UB -50 .5k chambers are used for position determination. 7 The dominant source of direction uncertainty Energy IOBS in converter 13 H 15 19 for low energy y's is the multiple scattering of Overall resolution with 5.8% 3.35! 1-3JS 1.2* the converted electrons in the active converter. Pb-glass The rms radial position uncertainty at the detec tor for a single electron of =:iergy E passing 18 Angular resolution (mr) 35 8-9 3.5 through a thickness d-, of so.terer of radiation due to MCS only length XQ followed by a drift t^ace dg to the detector is 2Xo E 100 £00 U00 1000 Thus, this contribution to the angular resolution y goes .Inversely as E. This has been used to esti Conversion probability -TO .72 • 7k .79 mate the average angular resolution as a function 7 of y energy of the configuration shown in fig. 7. 40 Energy loss in converter' 32 50 70 The resolution has been averaged over conversion Overall resolution with 9.OS 5.1? 3.0* 1.6S points, energy partition "between the conversion Pb-glass pair, and photon production angle and ranges up to several times the average value shown. We assume Angular resolution (mr) 77 38 19 7.7 that the event contains charged particles so that due to MCS only the vertex is known. For high energy photons, the jffip; ""-Prut Chamber jg =3- was a radial segmentation of each prism into two -SJmplt Ball parts, separated, at about 3XQ» to allow some n-e separation. The first interacts with the active converter just described. Since it would be in- ' Avtragi Angular Ruolutlon of practical to shield its phototubes from a several M field* these two options could not be mixed. I Aetivt Convtrttr Since we felt that the arguments for photor angular resolution are stronger than those for a field, the matter vaa not pursued to a practical design. Two methods of segmenting a prism were sug Etfimati Including Shuwir Divtlopmtnt gested, one based upon polarization of the light from the inner segment, and the other based upon a sharp cutoff filter, tuned to split the spectrum. Both require doubling the number of phototubes In the system hut thi? is possible with 1" tubes. C.Multipta Scattering and Orift Ctiambtr The polaroid idea would work nicely if only total internal reflection were used to propagate the light in the prism. This, however, leads to a -li:l ratio of light collection efficiency from one end to the other, and so Is intolerable. Thus, diffuse reflectance is required, and this is ex pected to lead to some depolarization. Without Fig. 9 XHI, 759 836] experimental work, it is impossible to speculate on the achievable separation with this method. The second method Is to separate the two segments with a sharp cutoff filter, and put a resolution is determined by the size of the showers similar filter in front of one phototube. Suitable after 2 r. 1. This has not been evaluated, but filters for splitting the spectrum are available. experience with a similar arrangement at DESY sug gests 10 mr as -the limiting resolution. Figure 9 shows the expected average angular resolution, CONCLUSIONS for photons. There are several difficulties with this It appears practical and desirable to augment system which we can point out, hut have not fully the SPEAR version of the Crystal ball with M\ evaluated. First, the segmentation into only 2k external calorimeter for use at PEP. It is ex parts implies serious overlap between uncorreleted pected that the resulting system would give a photons, -iven if they are isotropically distributed. calorlmetric measurement of total energy, and so Figure 5 shows the probability that no two photons would yield a total hadronic cross section from one are in the same active converter segment as a photon exchange with different biases than the mag Amotion of the number of isotroptcally distributed netic spectrometer methods. We find that an active photons in an event (curve with f = ^.3$). In the converter is geometrically feasible, but its de frequent case that two or more photons fall in the sirability is not so clear. A definitive evalu same active converter element, we will ~je unable ation of these additions must await Monte Carlo to disentangle the part of their energies lost in studies, but the "back of envelope" deoigns seem the active converter. Thus, we anticipate that promising enough to warrant more detailed work. the overall energy resolution with this system Finally, the calorimeter clearly does not deteri will be seriously deteriorated. A second problem orate the response of the system to low energy is that the size of the shower at photons for which most other instruments &re rela the drift chamber will fluctuate strongly de tively blind; a active converter is much more pending upon the conversion depth in the active problematic in this regard. converter. Thus, the expected average angular resolution of 10 mr is not a good measure of the ability to resolve two nearby photons. For ex REFERENCES ample, the puir from a late converting photon can ! be easily lost in the shower of a nearby, early 1. Stanford Linear Accelerator Proposal Ll-. -. converting one. In conclusion, the desireability 2. "The Study of Heutral Particles at PET" in of an active converter depends upon a more precise this proceedings. definition of physics goals than given here, but we 3. J. Engler, et al., Nuclear Instruments and think we have identified the mnst importp.ct ex Methods log, 189 (1973). perimental trade-ofrs with such an addition. I. F- J. Sciulli, California Institute of Technology Beport Ho. CALT 68-50^ (19'iTO 5. E. Bloom and M. Richardson, private Other Modifications communication. 6. R. L. Walker, California Institute of We considered two other possible modifications Technology Report No. CALT-68-3?0 and F. Bulos, to the simple ball. One was the addition of a "Resolving Overlapping Gammas in a Hodular small magnetic field to the inner cavity to pro Neutrals Detector", in this proceedir;-s. vide the sign of charged particles, and the other T- K. Hngel, Ann. Physik 186, 313 (196.0. A LIQUID ARUON NEUTRALS DETECTOR (LAND) FOR PEP A. Eisner, G. Hanson, D. Hitlin, U. Koetz, M, Marshak, T. Mast, J. Matthews, C. Peck and D. Yount Summary the relatively coarse angular resolution, and the We discuss the physical effects limiting the lack of flexibility coming from the expense and gamma energy resolution of a liquid argon calori fixed geometry of the crystals. An alternative meter without passive converter plates. An example which we have studied is a new detector using of such a detector based on the General User's liquid argon ionization chambers. Magnet designed at this Sumner Study is given. We were motivated by the successful construc tion and operation of liquid argon calorimeters during the past year.1 Oui initial goal was to de 1. I-DTTVATION AND GOALS sign a detector with energy resolution for low en ergy gammas (less than 500 MeV) that was compar The general motivation for an apparatus sensi able to that of Nal. In addition we wanted modest tive to low energy photons at PEP has been dis momentum resolution for charged particles, flexi cussed in "Neutrals Detection at PEP," a paper from bility in terms of being able to match given geo this Summer Study. That paper describes our general metries of other detectors and magnets, and hope ideas about neutrals detection, the physics avail fully low cost. Our initial vision was a small sole- able to a neunrals detector, and the multiplicity, noidal magnet surrounded by a large pure liquid energy, and angular distributions of gammas ex argon gamma detector (i.e., without plates of pas pected at PEP. Our emphasis on low energy phutons sive converters). In addition to the above featu.cs is motivated by the realization that general detec such a detector would have some nice capabilities tors will provide adequate resolution and sensi coming from the longitudinal information available tivity in the GeV range. in a liquid argon counter. This information and the good energy and momentum resolutions would yield One approach to observing the expected large excellent electron-hadron discrimination. In addi number of low energy photons would be to use the tion the longitudinal information in the argon Crystal Ball detector presently being built for provides measurement of the ionization energy los; SPEAR. This detector has some nice properties jnd for charged particle identification. some obvious disadvantages, both of which are de scribed in the paper "The Crystal Ball at PEP" from The major effort of the group was not the de this Summer Study. The major advantage is the ex sign of a specific detector, but instead the study cellent energy resolution and nearly 4TT solid . of the considerations that would go into such a angle coverage. In addition the small size makes it design and the feasibility of achieving the goals very feasible to surround it with a hadron calori described above. In Section 2 we describe the re meter. The major disadvantages are the lack of mea sults of that study. In Section 3 we do describe surement of the charged particle sign and momenta, very briefly one possible configuration for such Table 1. Properties of shower counter materials Sef, 3} Interactions on Moliere radius dE/dx Radiation protons/X 21.2 X /E (MeV) n n c Density Material (MeV/cm) length Critical (Op in mb) (an) g/cc Eneruy,Ec Argon 2.11 14.1 29.8 15.6 10"4 o 10.0 1.40 Lead Glass 5.38 3.22 17.3 9.31 10"4 o 3.9 3.61 (F2) P Aluminum 4.37 9.00 39.3 23.5 10"4 0 4.9 2.70 Lead 12.8 0.56 7.17 2.55 10"" o 1.7 11.35 Nal 4.84 2.5S 12.5 5.5 10"" o 4.4 3.67 Copper 12.9 1.45 18.7 9.0 10"4 0 1.0 8.96 Iron U.6 1.77 20.5 10.2 10-4o 1.8 7.87 a detector, I'iiialty wo suniuariie in -Section 4 our ym Hesphition conclusion;; on the U-asihiltty ol" using liquid argon for a high resolution neutrals detector, and lilectromagnctic slrawers create ioiuraiion in we outline tht* work that we think needs to lit* done the liquid argon; the collection of electrons from before -i detailed design can he made. this ionization then induces a signal in the de tector, and finally this signal is processed by the electronics, he use this scries of events to 2. uLUiA(;ij',Kis~iics oi iiniiiii AUUIN mrn.crifiKS classify the various contributions to the energy resolution as Follows: The use oi' liquid argon hi high energy physics was originally studied with the view of achieving very high spatial resolution for charged par A. Fluctuations in ionization coming irisa ticles.2 Mure recently considerable success has fluctuations in: been achieved in using it in had route and elec U) energy ius'. out the back nd, tromagnetic s!::;wor detectors.' lliis work of the (21 enetgy loss in a magnet i' I and past few years is well described in the literature ciewar. and we will not review it here. Ive have summarized 13.) energy loss in jussive com iter in Tali I e I some of the characteristics of liquid plates, argon and other material-, used in shower counters. (41 energy loss transverse I *.• out ot a shower fiducial vol tone, (5) tracklength induced h> eleciitmag- The detectors used to date have contained plates of passive convei ters (lead or steel) in netic luidran production. order to reduce the dipths ol the counters, Com It. I-luctuations in the pulse induced in the pared to a pure liquid argon counter these plates electronics coming from fluctuations m sinmei seriously degrade the energy resolution for low track positions. energy gammas. Thus our main emphasis was on a de tector without such plates, and we simply accept C, lileetronic Noise. the deep counters which thi< implies. However, we were concerned with the relative]*- large transverse We discuss each of these below. In c.ener.il, -.pread ol" a shower when these plates are absent. we adopt the attitude that fluctuations at tin. 1'i Consequently some work was de.: Ftuctuotions of Ionization Lois for Gammas in Liquid Argon • •ex. tigma»(width containing 0 67 of tvinti)/2 100 1C00 GAMMA ENERGY (MtV) I/2X.AI at SX. I/2X.AI at OX. tor convtrtlna aamniat only 100 1000 PHOTON ENERGY (MeV) XHI. :r)ll)-H45S liquid ar^on counters of various thickness. The results are shoiui in Fig. 1 as a function of gamma Monte Corlo Gammas in Liquid Argon energy. For the low energy showers that we arc RMS of Energy Lost in Each Loyer emphasizing, ahout It) radiation lengths are suf ficient to maintain an energy resolution of <_ 2%, • -50 MeV Energy loss in a magnet coil and dewars. + '70 MeV During the past year considerable progress has been * * 100 MeV made in designing and testing thin superconducting 200 MeV magnets.'' A solenoid coil presently heing construc " * ted will have a thickness of 0.33 radiation lengths. , o . 500 MeV Considering the oblique angle of some gammas and + • » - 700 MaV the liquid argon dewar thickness we consider the effects of 0.5 radiation lengths of passive mater ial (aluminum.). With the magnet coil in front of the shower detector about 27': of the gammas convert in the coil,and the rms energy loss for these converting gammas is plotted in Fig. 2. Shown for comparison are the resolutions, for lead glass, sodium iodide, and that expected for the Mark II SPEAR Magnetic Detector. Since we do not want to degrade the resolu tions of low energy gammas, we have considered the case where the oil follows 5 radiation lengths of gamma detector. (The motivation was to make the length large enough to get most of the shower en ergy and small enough to minimize the wasted mag netic field.) The results are also shown in Fig.2. These latter results are based on the fluctuations in energy loss at various depths in the shower shown in Fig. 3. From this calculation, we con 2 3 tinergy lo^s in passive converter plates. the original electron-positron pair will be formed Al though the" ailill tloif or very th in i ead plates to and then come t" rest through ionization loss. If a pure liquid argon detector will add energy loss the tracks are perpendicular to the collection fluctuations, it may also decrease the electronic latcs, there will be about 25 track gaps. Fluctuat noise and nMive I la- tmnsversc spread of tlie fions in the relative lengths of the tracks give showers (See Sections J. 15 unci 2.3 below). Thus rise to track-gap variations of 1/2 track gap. it is important to calculate the size of these fluc The rms over the mean is then(l/2/T7)/ 25 - .006. tuations as a function of the plate thickness. The If the tracks are at IS" to the plates,this is in usual rules of thumb for these fluctuations (see creased to .008. We conclude that this effect will for example Kef. 11 are not valid for very thin be small,and we relegate it to the general unre plates. Unfortunately we have not found an analytic solved background of 1% effects. expression for those fluctuations which agrees with both intuition and existing data. EitheT an analytic Cylindrical-wire geometry. Wire collection or Monte Carlo solution of this problem is needed electrodes are attractive for mechanical reasons for an optimized design. However, on the itasis of in a practical counter. We have therefore con some rough estimates we feel that enough material sidered some of the effects associated with this to significantly reduce the shower spreading would option. probably significantly degrade the energy resolu tion. In a cylindrical volume of radius b with a wire of radius a,the voltage pulse induced by the collection of an electron originating at radius Fluctuations in energy loss transversely out r equals (e/2n£)ln (r/a) = k In (r/a).6 For a uni of a sTiower T i JiuJ a J \rt ij lane" In practice one will form distribution of ion pairs the mean voltage add' together a" lar.ne enough number of readout cells equals k(ln (b/a)-1/2) and the (root mean square to contain the s.hower. As this number is increased deviation about the mean)/mean = l(Jiiiih/a)-l) the fraction of energy escaping and the fluctuations The latter varies from 0.18 at b/a = 25 to .07 at in this energy art' reduced. However, the electronic b/a = 2000. (A 10 nta radius volume with a 5 micron noise increases as tlie square root of the number radius wire has b/a = 2000.) Thus if a shower con of cells added (to be discussed later). Thus there sisted of only one ion pair the resolution would is some optimum number of Lells to be added together, be poor. This resolution is improved by the square and this number depends on the particular config root of the number of randomly distributed ion uration of the detector and the energy of the pairs, and since this will be quite large for most shower. In Section 7..1 of this paper we discuss .-bowers, this contribution to the resolution will the electronic noise for various configurations, be negligible. For low energy showers (50 MeV or and in Section 3 we discuss the transverse size less) this effect will contribute significantly to of the shower. However, we have made no detailed the resolution in a geometry with the. wires longi optimization to determine the size of shower fidu tudinal along the gamma direction. .Such a geometry cial volumes and this would probably have to be de wus considered by A. Odian at last year's Summer termined by the experimental calibration of each Study7 and rejected for this reason. We conclude particular detector. that wire collection electrodes transverse to the shower axis, replacing the conventional parallel pl*.te geometry, will not contribute significantly Fluctuations in track length induced by elec to the energy resolution for the reasons considered tromagnetic hadron production. At some low level here. However, it must be pointed out that we have there will be fluctuations in the ionization due not considered the effect of inductance in the to the production of Iiadrons by electrons and wire geometry and this bears on electronic noise. photons in the shower. This sets a floor under Indeed it may be that the characteristic impedance the achievable resolution of any electromagnetic of a wire collection electrode (treated as a trans shower detector. For the low energy showers that mission line) is large enough that the lumped com we are emphasising here.we expect this effect will ponent analysis used here is not applicable and be below the 1° level. considerations such as given in Ref. 8 should be used. We have not done this. {B) Fluctuations in the Pulse Induced in the Electronics (C) Electronic Noise The puis- induced by the collection of an electron in an ion chamber depends on the original We have considered three possible collector position of the electron, i.e., the position of geometries. the original ionization.° Since the position of tracks will fluctuate.the above effect will give (1) Squares. These were suggested at last rise tc fluctuati ns in the output. We have esti year's Summer Study by A. Odian.'The hope was to mated the size of these effects for parallel plate connect them in tower arrangements (Shish-kelwh and cylindrical-wire collection geometries. geometry) so each detector defines a unique solid angle and each has a low capacitance. We have not Parallel plate geometry. For most showers devised a convenient way to connect the high volt the niuiiher'bT tViick-tyips wifl" be very large and age and readout squares, tn additf.m then' would the I'luctuatii'iis in the position of the ionization be an enormously large number of elemcr.es re within the gaps will In- reduced by the square root quiring electronics and so we considered this of tliis large murilvr. So we consider an extreme arrangement no farther. case, a 50 Me\ I'.amma in a pure liquid argon detec tor with 1 cm i'.ups. Since the critical energy is (2) Strips. These uill be easier tn construct 30 MeV. we haw .ibont two tracks. In other words and have a relatively low capacitance. Inn they will have the problem that more than one garana will hit The energy resolution for each of the geom the same strip, etries is shown in Fig. 4, and the resolution of Mill is shown for comparison. In the low energy (3) Wires. These will be similar to strips region liquid argon is comparable to Nal for con and perhaps easier to construct. figuration f. The resolution increases as the square root of the length of the strips, so for 3 The capacitance for the wires and strips are merer strips these resolutions are increased by 1.7. not very different,so we use strips as an example The resolution also varies as d"3/2 so consider for the following discussion of electronic noise. able improvement can be achieved if J, the spacing between electrodes, is increased. One pays linearly We follow the electronic noise analysis of in collection time and the required high voltage Willis and Radeka* based on a system where the ;is d is increased. signal from the collectors passes through a trans former into an FET preamplifier. The resuJts of the Attempts have been made to reduce the collec analysis are shown in Table 2. The first two col tion time by increasing the drift velocity through umns define the geometries considered, d is the the introduction of polyatomic organic compounds liquid argon gap length and w the strip width. to cool the electrons. A group (Ref. y) at LBL Column 3 shows the capacitance (C.) for strips 1 has increased the drift velocity in liquid argon meter long and with enough strips connected in (at 10 kV/cm) by 305 using 500 ppm methane. An parallel to make 1 radiation length in depth experimental program to study this problem has (= 1 cell). Column 4 gives the collection time been started in Berlin by Gerhard Knies.10 assuming a drift velocity of 200 nsec/tmi. The opti mum transformer ratLo (column 5) is given by One way to reduce the capacitance of the de n = (toA\) where C, is the FET, transformer, tector, which was used by Knies and Neuffer (Ref.11), and strav capacitance.^Ye follow Ref. 1 and take is tu couple the plates within a cell in series in C = 30 pF. A stead of parallel. We have not made an analysis of the noise nor the constructional difficulties for Column 6 gives the rms energy fluctuations this configuration. from electronics noise for 1 cell. = (52.8 MeV lost/signal electron detected) Electronic Noise with Passive Converter 4 e , CC C )1/2A1/2, where e is the FET ft ni/ d A n Plates. As plates of passive converter are added noise, 1 nV/H , and \ is the resolving time, which to pure liquid argon, the number of gaps per radi we take as 1.5 •*• collection time. In order to con ation length decreases,and thus the capacitance tain the entire shower we add together the cells for and electronic noise per radiation length is re a depth of 12 radiation lengths and for a transverse duced. However, since the amount of energy being distance of 60 cm. The noise from these cells adds deposited in the argon is also reduced, the size of the signal is reduced,and eventually the frac in quadrature and is given in column 7, Finally, tional electronic noise increases. I'nr the elec column 8 gives the ratio of the noise without a tronic noise with lead plates we find that the ex transformer to that with a transformer pression used above for the electronic noise in = 0.5(n + 1/n). For the last geometry it may be pure argon must he multiplied by the factor 1 2 possible to eliminate the transformer. F = (1 + 6.07 r)/(l + 25.2 r) ' where r is the Tabic 2. Electronics noise for various configurations of a pure liquid argon detector (1) (2) (3) (4) (51 (6) (7) (») Type Collector Capacitance Collection Optimum rms energy rms energy Katio of dimensions Tor 1 meter tine transformer fluctuations fluctuations noise without w d by 1 X (pF) (us) ratio per cell per shower transformer to (mi) = 1 coil (MeV) (MeV) that with it a 10 2 4956 0.4 12.9 0.93 26 6.5 b 10 5 793 1.0 5 0.24 6.6 2.6 c 10 10 198 2.0 2.6 0.083 2.3 1.5 d 20 S 1586 1.0 7.3 0.33 6.6 3.7 c 20 10 396 2.0 3.6 0.12 2.3 1.9 f 20 20 99 4.0 1.8 0.042 0.82 1.2 r Eltcironic Noitt in a Pur* Liquid Argon Dtitctor 100cm x 60cn) ENERGY(MeV) XBI. 7510-1457 ratio of the length of lead [cm) to the length of l+6.07f argon (an) in each gap. The numerator in F comes 1.5 from the reduced signal >r>cn in argon and the de nominator comes from the reduced capacitance per y/\*25.2p radiation length. This factor is plotted in Fig.5 and shows a shallow minimum at r = 0.08S. For a gap of 2 cm in liquid argon, this implies a thick ness of 1,7 mm of lead,which would introduce fairly large energy loss fluctuations in the lead. Our conclusion is that the noise reduction due to adding lead plates is small,and we expect that it would be more than offset by the energy loss fluctu ations in the lead. 1.0 I WILLIS If no transformer is used,then the functional 1 MARK D dependence of the noise on the capacitance changes 1 2 from (CJCA) / to (Cj + CA), and the effect of adding pJates is different. The factor relating the resolution in this case to that of pure liquid argon with transformer is given by [(eye,)"'/( ! • 25.2 r) 0.5 This is plotted for two values of Cj/CA in Fig.b, and again the electronic noise is not strongly re duced by the addition of lead plates. In the above considerations we have not in cluded the reduction in electronic noise that cones from the fact that the transverse shower spread may decrease as lead is added and thus fewer cells need be added to conta in the shower. _1_ 0.5 1.0 2.2 MAGNETIC l-IIJ.H lil'IIXrrS The Electronic Effects As noted earlier,signals from liquid argon are transmitted via fcrrite core transformers ex- HOTEriESIS CURVE 25 •c C./C..2 < / / 0*C f P - / J' >.' '4f C./C,»26 /'»i _ f t 2000 r i ^ j j I T1 ' til I, 0.5 ri P *t XBL 7510-8448 r. ._ n IE cept in the unusual circumstance of very low ion H (OERSTEDS) chamber capacitance. The signals are small, and thus the excursions on the characteristic Iiystr- esis curves are also small. The response of the transformer then depends only on the slope of the hysteresis curve at the ambient excitation. The Sampling Detector. To be speuific, we con- characteristic curve for the 3D3 ferrite, shown sider a sampling detector consisting of lead plates in Fig. 7, (Ref.1'2) appears to be linear for and liquid argon gaps. The gain is determined by fields up to about 1600 gauss. Typically, a high- the fraction of the ionization that occurs in the permeability material in air will gather flax argon. This fraction is normally small so that, from an area of order four times the cross section for example, doubling the gap doubles the signal. perpendicular to the field. Thus a 3I>3 ferrite in We then ask; To what extent does the presence of a an ambient field of 400 gauss would shunt a field magnetic field alter the fraction of the ioniza of order 1600 gauss. We conclude that a 3D3 ferrite tion that occurs in the liquid argon sensitive would have an output independent of ambient field voLimc? up to about 400 gauss-'providing that the core operates always on the same hysteresis curve~ This The pathological example shown in Fig. 8 implies that such cores could be used routine!>• illustrates that the fraction of the ionization in fringing fields, possibly with some mayntttic sampled can be influenced by a magnetic field. shielding, but they would not operate in the high Case 1 corresponds to a load-liquid argon detector field region of typical magnetic detectors. A with no field, and Case 2 corresponds to the same further point is that calibration pulses should device in a field so large that no charged shower be injected into the electronics in situ upstream particle reaches the next lead plate. Evidently, of the transformers, rather than downstream, to incorporate any magnetic field effects into the calibration. Shower-Track Sampling Effects Pure Liquid ArRQn Detector. In a total .«h- sorption detector, such as liquid argon or si'dimi iodide, virtually all of the shower energy ap pears ultimately in the form of ions, photons, ;t ui fTce electrons. It makes no difference whether shower tracks arc straight or curved due to a mai:- netic field. In particular, the number of ion^. CASE I CASE Z free electrons, or photons is independent of tIn field. -101- thc amount of detectable ionization is proportional controlled by locating the cores in a low field to the width of the liquid gap for Case 1 and in region for which the price may be increased ca dependent of it for Case 2. This shows that the pacity and inductance in the coupling cable. Dis presence of the field docs change the response and placement of the shower tracks would occur through that the magnitude of the effect can be large. out the large volume of the detector itself, and For example, it can be larger than a factor of 10. the results are harder to anticipate or correct. Clearly it is important to operate liquid argon A rule of thumb for track sampling is that the and iead-liquid argon detectors in a magnetic field magnetic deflection of shower tracks at the criti before large devices of either type are constructed. cal energy of the absorber plates should be small compared to the spacing between plates. For ex 2.3 Spatial Resolution ample, if the 3bsorber plates are of lead (crit ical energy of 8 MeV) and if the plate spacing is The spatial and angular resolution for in £ 'on, then fields of a few Thousand gauss would dividual gapmas will be determined by taking the cause significant changes in gain. centroid of the shower on the strips, and we expect the spatial resolution will be equal to or less If a large lead-li<\uid argon detector is op than a strip spacing of 1 or 2 cm. For the LAND erated in the fringing field of a magnet and if detector described below this will give an angular the field is high enough to cause sampling error? resolution of about 10 milliradians or better. of the type just described, then variations in the field would lead to non-uniformity of the detector However, a matter of some concern in a pure- response as a function of position. Calibration liquid argon shower detector is the transverse and correction of such an effect would be tedious si-e of the showers since this leads to non-neglig and difficult in a large device. ible overlap probabilities for the expected gam ma multiplicities of 10 to 15. Analytic sltower I'lectron Collection theory*3 gives the result that about 90S of the energy in a shrwer is contained in a cylinder, of To what extent does the collection of elec diameter 3.6 fc. (36 cm), where R.. is the Moliere trons depend upon the magnetic field? radius In the case of integrating detectors, such R. = r^fey X0 = 10 cm for liquid argon, as the ionization quantameter, the integration time is very lout; and fully covers the time interval during h'hicn induced current flows from both the where Kc is the critical energy and XQ the radia primary ions and electrons. Any recombination of tion length. This shower model assumes that elec ions and electrons would be sensitive to the tromagnetic processes lead to negligible angular ambient magnetic field, hut this is minimized by displacements and only the effect of the multiple using an utilization medium, gas or liquid, of suf Coulomb scattering of the shower part icles is con- ficiently high purity. Thus this tyjie of device stdered. can be node to June no magnetic field dependence due to this source. lixperimental work (Ref. 1J 15), however, in dicates that in low Z materi;... the transverse In a detector sensitive to individual tracks sizes are larger than those which analytic theory (or events) such as we have been considering, the predicts. This is attributed to the annihilation electronic integrating time is long enough to in radiation, but the experimental evidence is in clude the induced current from the motion of the dependent of this interpretation. In Figs. 9 and 10 primary electrons, but so short that it includes we show the observations of transverse shower sise almost none of the current from the much slower in units of fyj. Extrapolating smoothly in Z to motion of the heavy positive ions. This is just argon leads us to expect that 90°, of the shower is the effect that gives rise to the dependence of contained in a cylinder of radius 5 KM = 3U an at output signal on position of ionization considered 185 McV and J.25*1^ = 23 an at 900 MeV. in Section 2.1.2. As is illustrated for an ex treme case in Fig, 8, the equivalent location of In a shower detector of the kind we are con ionization is at the center of the gap in Case 1, sidering, the radial distribution is not measured. but displaced from the center in Case 1. Thus in Instead the transverse projection of this onto addition to the effect on output signal due to strips is detected. If we take the radial distri the change in energy deposition in the liquid dis butions implied by the data in Figs. 9 and 10 and cussed above, the magnetic field will also give project them transversely onto the strips, we see rise to a change in the output signal induced by 90% of the energy in a transverse distance of this ionization. •% (2.8 %) for'the 18S Mel' (900 MeV) shower, 'llius for the low energy showers, the transverse Summary size is about 40 cm. It appears that liquid argon detectors will This relatively large transverse size lias two be sensitive to magnetic fields via: {1} saturation consequences. First, there will he random over of fcrritc cores in the readout, (2) change of laps of tht showers from unconelated gammas and track length of low energy shower tracks in the secondly, the two gammas from a high energy v° liquid (applicable when lead plates are used), will be strongly mixed. Walkcr'h has given a (3) change af the equivalent location of the ioniza method based upon second moments which allows one tion between the plates (applicable whether or nut to reconstruct the invariant mass of an overlapping lead plates tire used). The first of these can he gamma pair, and thus search for u°*s. The analysis 185 MeV 900 MeV • -Kantz and Hofttadttr • i Crannell J* 3 _J_ 30 40 50 60 70 80 90 z -I l_ _1_ _J l_ _l Fig. 9 •"" 0 10 20 30 40 50 60 70 80 90 z Fis. 10 XBI. 75ION45J he gives assumes that the photons strike the de tector perpendicular to the readout planes and this will probably not be the case in a PEP detector. The probability of no overlapping gairanns in However, the longitudinal development of a shower an event with N photons with isotropic distribution is measured so that a correction can be applied for non-normal incidence. The general problem of overlapping showers is considered in detail by Bulos in "Resolving Overlapping Gammas in a Modular h a - no, Neutral Detector," a paper from this Sumner Study n=2 and it seems feasible to use an analysis similar where f is the fraction of the total space used by to Walker's. each photon. It is interesting to calculate the fraction of events in which there is no overlap in either of two orthogonal readouts for the 1 liquid argon detector described in Section 3. The detector is broken into octants, split longitudinally along the beam into two 2-meter sections, and the shower maxima occur at a radius of about 2.2m. The value of f for the strips parallel to the beam is (0.40)/(2X2TTX2.2) = 0.0145 and for the transverse strips (0.4O)/(8x4.0) = 0.0125. Figure 11 shows the probability per event of no overlap on a single longitudinal readout and on -neither of the two readouts. It is seen that with the assumptions made here, essentially every event will have at least one pair of gammas overlapping for the ex pected photon multiplicities, even ignoring the correlations between them implied by the jet model. The-energy resolution for these overlapping gam mas will be degraded,and the importance of this needs to be studied further. 5 10 15 20 3. A NEUTRALS nifTECTOR FOR 110: GENERAL IJS1-R N (NUMBER OF GAMMAS) MAGNET The General User Magnet (CUM) is described in detail in a paper from this Summer Study. It has a Fig. 11 segmented coil to produce a solenoidal magnetic 103- OUW FLUX RETURN field region ol" 1-meter radius, and this field region is instrumented with drifl chambers, h'c describe here a possible design for instrumenting Che region outside t':o coil with a pure liquid argon detector (sec fig. 12). The GUM has a lumped coil which leaves 781 of the solid angle down to 0 = 30° unobstructed in 6 and 821, of the $ coverage unobstructed. Thus 561 of the total solid ai.glc, or 71* of the coverage down to $ ~ 50° has a minimum of material. Ol course in practice this region has some ma they each weigh 24 metric tons when filled. This terial; 0,05 radiation lengths for scintillation is quite heavy, but does not constitute an assembly trigger counters, 0.05 radiation lengths for cryo problem, since the unfilled tanks will weigh only genic insulation of the superconducting coils, a few tons. We have not estimated the cost ol" the 5.10 radiation lengths for the structure and in detector, but we note that the cost of the liquid sulation of the liquid argon tank. In the region argon at S0£/liter is 58501) per module and so this covered by the coils and their support structure part of the cNpense is negligible. With this large there i s very poor or no gamma energy measurement. In volume, a refrigerator is clearly preferred over a contrast, for a uniform solenoidal coil D.33 Xfl heat exchanger ?hat boils UN. thick, the energy resolution is degraded through out the entire region for those gammas which con The readout consists of 100 planes of cells, vert [about of the gammas). This degradation 11% tligh voltage wires or planes spaced 2 cm in depth is described in Section 2.1.1. A disadvantage of are stretched along the 4-m dimension in each the GUM design is that the coil support and end module. Each longitudinal half of the module is cap flux return make it very difficult to achieve read out separately. Readout wires are placed mid good neutrals coverage for 0 < 30°. It would be way in each layer, 2 cm apart. These wires are possible to put a more compact photon detector, alternately at 0°, +45° , and -45°. About 10 such as lead glass/liquid argon sandwich,in front wires of a given orientation are grouped in depth of the vanes, but we have not attempted this in and connect^ to a single amplifier. Thus there our design. are six. channels in depth (two for each of three orientations) and an average of 130 transversely. = h'c have divided the liquid argon detector Thus, each module requires 2*6x130 *^W< channels into eight identical modules fitting between an-1 of readout; the eight modules then require 12,500 suspended from the flux returns of the G1JM. 'the channels of electronics, which probably makes it modules arc trapezoidal in cross section [see economically necessary to multiple tin- ADC's. Fig. 13), I m (14 radiation lengths) in depth, I.IS m in their trunsver..*- dimension inside sm« The energy resolution for this lU-tivicr (coming 3 n: outside. Their volume is 1" m-Vinodule and swlel/ from electronics noise) is shown in fig. H. -104- Energy Resolution for GUM-LAND 1,3) lixpcrimental measurements of the effect of magnetic fields on liquid argon chambers will establish constraints or freedoms in the design of a practical PEP detector. (4) A more detailed study should be made of effects we found here which con tributed to the resolution at about the K level. References 1. W. J. Willis and V. Radeka, Nuclear Instruments and Methods 120, 221-236(1974). 1. S. 1". Derenzo, et al. Nuclear Instruments and Methods 122, 319 (1974). 3. Particle Data Group. Plivsics Letters, SOB PHOTON ENERGY (MeV) (April 1974). XBI. 75HI.H450 I. M. A. lireen. "The Large Superconducting Sole Fig. H noid for the Minimag Experiment." Lawrence Berkeley laboratory Report LBL-3677. 4. CONCLUSIONS .*». T. 5. Mast and J. E. Nelson. 1974 PEP Summer Study, PEP-153. We conclude that it is feasible to build a pure liquid argnn detector for PEP that will have li. H. H. Staub. Experimental Nuclear Physics. energy resolution-; for electrons and gammas above Volume 1, E. SegrO, editor"John Wiley, New York about 100 MeV* that are comparable to those achieved (1953). by Nal. In addition, such a detector can measure the longitudinal development of a shower which is 7. A, Odian. 1974 PEP Sumner Study, PRP-lSf.. useful for electron-hadron discrimination. The de tector is flexible in the sense it can be designed 8. V. Radeka. IEEE Trans. Nucl. Sci. NS-2I, 51 to fit into a variety of geometries and is inex (Feb. 1974). pensive enough to he built around a magnetic charged particle detector with good momentum reso 9. Private communication, S. Derenzo, A, Kirschbaun, lution. T. Tometani, H, Zaklad. h'e have discussed in this report various as 10. G. Knies, private communication, pects of the design of a pure liquid argon counter but have not presented a detailed design of a com II. G. Knies and D. Neuffer. Nuclear Instruments plete detector system for PEP. The following work and Methods 120, 1 (1974). needs to be done before a detailed design can be made. 12. "Ferroxcube Linear Ferrite Materials ant.' Com ponents." Ferroxcube Corporation, Saugerties, (1) A better understanding of the effect Sew York, 2nd edition, pp. 1-1? (1974). of lead plates on the energy loss fluc tuations and the transverse spreading 13. Grieson, Progress in Cosmic Ray Physics, vol.5 of showers is needed to know whether (19S6). such plates can be used to reduce the gamma overlap problem without destroying 14. C. J. Crannell, Phys. Rev. 161, 310 (1967). the good energy resolution. 15. Kantz and Hofstadter. Phys. Rev. 8£, 607 (1953) (2) More experimental work needs to be done and Nucleonics 12, 36 (1954). on increasing the electron drift vel ocity through the use of organic 16. R. L. Walker. California Institute of additives. Technology Report CALT-68-330. RESOLVING OVERLAPTNG GAMWS IN A MODULAR NEUTRALS DETECTOR Abstract It is shovm here that in a segmented neutral p(zj = P(-z);i.e., p(iO is an even func- detector the method of moments can be used to de termine whether a connected region of energy de posit involving several modules contains one, two, We will assume that the areas of the individu or more y's. Up to 2 y's can be handled analyti al segments of the detector are equal or known, cally to solve for their energy and position. Fur and so the areas will not appear explicitly. ther it is shown that if the positions of the y's are determined externally,then the method of mo Two cases will be dealt with here: Case I ments can be used to solve the problem of more than where no external determination of the y position 2 Y'S analytically. is made and Case II where Y positions are known. It was shown by Walker et al (Calt-68-330) CASE I that the method of moments can be used to deter mine the ratio M/H for a particle depositing en Let (Xj.yJ he the centers of energy deposit ergy (via its v decayj in a connected region of E; in module i with respect to co-ordinate system a segmented neutral detector. Essentially,if all X7V. parallel to the plane of detector (i.e., the the overlapping y'a come from the same particle Z axis is along the longitudinal direction of the (e.g.) 2 Y'S from u^the ratio M/C for the par modules). ticle can be determined and hence one can identify the particle. Ej is known from the pulse height of the module and x±,y^ can be taken to be the center of The method however cannot separate the Y'S the module. and hence is not suitable for recognizing individ ual (isolated] Y rays. The analysis described here Then the following averages are known (developed with the help of J. Friedman, SLAC computation Group) is an attempt at separating overlapping Y'S analytically. The principle involved has been verified ex perimentally and with Monte Carlo calculations namely: Consider s Y (C) of energy E. incident on a slab of radiator of a depth along the y direction sufficient for total absorption. The fractional en ergy Er deposited in a cylinder with its axis along the y direction of radius between r and r + dr and length equal to the depth of the total The equations of the major and nrnor axis of radiator (i.e., energy is integrated longitudenally) the energy deposit and its major and minor vari is independent of the y energy En and is a func ances are given by the eigen vectors and eigen tion of r only: values obtained by diagonalizing the following matrix: r Er ?-{X)* XY-X 7 ^ = f(r). If f(r) is projected on an axis z,one obtains ^XT-JCT Y^-CH2] f(z) which has the same property, i.e., Let tnc eigen axes and eigen variances he u,v and A^, \. respectively. Figure 1 helps illustrate"the method. Note that the eigen vec • fU). tors pass through the center of gravity and for the case of two Y'S their centers must fall on The two functions differ only in normaliza the major axis. tion. Instead of the fractional energy* the frac tional energy density (Ur/A E^l = p(r) will be To differentiate one, two,or more Y'S,one used here mid its counter part p(z) with: proceeds as follows a) \ = \ indicates one y or a con spiracy. p(r) 2TIT dr = 1; p(z) dz * 1 b) For two y's, WA„ and further!• vcthe distribution is skewed along the major axis tin- Only one of the equations (£q. (3)) is not immediately obvious. Its proof becomes clear in Case II below. We thus have four equations with four un knowns E-i,E- ,vj,V2» i.e., the two energies and positions of1 tne Y'S. (Recall that u\ = U2 = 0 ENERGY DEPOSIT an additional equation which has been already AREA used in Fig. 2). Hence the case of two Y'Sca » be solved completely. For more than two Y'S we do not have enough equations. This is discussr.-d in K,T"(C.G) Case H below. Fig. I CASb' II. Effect of position measurement on the sep aration of individual y'sT less the two y's are equal. The third moment M5 Let x ,yn,IL be thecoordinance and the en with respect to the e.g. co-ordinate system where ergy deposit measured in the n-th module (these u is the major axis, can be used to measure quantities are known). Then the first moments give skewness (M3OJ) = IE^UJ/E™..) . Hence T.E x EE x n n . EE„ E .\s I" Jj, HjM t 0, M3(v) * 0 T(m*A L indicate two Y'S °f unequal energies 7 = ^n ^OTAL The second moments give Xs * \ • M3(U) = °' M3(V) = ° ™ J- indicate two y's of equal energies or conspiracy. c) *s t At, M3(u) 1 0, M3(v) t 0 indicate more than two Y'S. xvti Complete Solution of Two Y'S With the help of Fig. 2 for two Y'S one can Thus we have the following known quantities: write the following set of equations (X,T) (center of gravity), E^, ox, o" * B, E (1) "1 * "2 TCT EJVJ * E,v7 = 0 U) Consider i y's whose positions are known (from an external converter - position identifier). Let (EJVJ • EjvlVEjrjj. -\- A (5) s the co-ordinates of Y'S wirh_respect to a co- J - c\ * EjJ/E^ =• o2 (4) oidinate system translated to X,7 as center be s Uj,^. These are known. where o is the (width)2 of the single y linear We now use the properties of the linear frac fractional density function p(z) described above. Re tional density function of a single Y which was member that a is a constant for all Y'S,known from described before. p(z) either experimentally or by Monte Carlo. XBl. 751l).s:lH5 tet E(u) be the continuous linear energy de 2!:E IE posit in the clust.rr projected on the u axis-, °u- l" i* i"fl/%r E(v) is projected 01. the v axis. It is easy to see (Fig- 3) that "v" l<,2sEi* EEivi'/EiOT E(u) - IEJ pdyu) E(v) - Efit p(vx-v) where o is the known constant width of single distribution function. jr— f EWu2 du We thus have the following general equations °l- (compare with Case I) «J" |— [ E(v)v2 Jv IE1,EIOT Thus we can write ((I E. p(u.-u)u2)du Eror la' IE. * lE.lif ] u ETOT E 2 °v ' E^,f | (I i Kvrv)v )dv <- k^ i°2 EEi* JEi"i' For each term in the integrand we write u * (u-uO+Uj and similarly for v. Substituting Tlie only unknowns are the E-. In principle in the integrands and recalling that p(z) is an then the case of up to five Y's whose positions even function of z, wc obtain are known can be handled analytically- REPORT 07 THE GENERAL PURPOSE DETECTOR GROUP A. Barbaro-Galtierl, W. Bartel, F. Bulos, R. Cool, G. Hanson, U. Koets, R. Kotthaun S. Loken, D. Luke, A. Rothenberg Abstract In this report ve describe a general purpose system* In particular, detectors using phototubes detector for PEP. The main components of this would be severely handicapped and central detectors detector are a 1 meter radius, 15 kilogauss super requiring a uniform field, such as the time-pro conducting adenoidal magnet with drift chambers jection device,1 would be eliminated* Moreover, to detect and measure the momentum of charged the strongly nonuniform field would place on particles, a liquid argon neutral detector and unnecessarily heavy burden on a pattern recognition hadron calorimeter, and a system of Cerenkov and system for the expected high-multiplicity events time-of-flight counters for identification of at PEP energies. charged hadrons. A major consideration in the design of this detector was that it be flexible: For the reasons outlined nbo--e, wc- have the magnet coil and drift chambers form a core chosen to study the use of a uoleno-ial field around which various apparatus for specialized configuration with its axis parallel tc the tear, detection can be placed. axes for use with a general detection facility. We recognize that the configurations which we l.uvc I. INTRODUCTION rejected for our purpose may in fact be rej'.lred in other specific experimental situations.^ A general purpose detector for PEP should have the following properties: Finally, the 30lenoidal field o.ny be generated either by a continuous, uniformly wounu coll > vt-r 1. Large solid-angle coverage for detecting the required length or by a set of "lurpcd" coila and measuring both charged and neutral particles; spaced along the axis, with nc iwitvrinl between coils other than that required to restrnin iho 2. Good momentum ind Angular resolution for magnetic forces. We consider here lh< continuous charged and neutral particles over a large srlid solenoid, while the "lumped" coil coml^urnilon is angle; the subject of another report.' it Good charged-partlcle identification over as large a solid angle as possible; Two possible designs for general detection !t. Good muon and electron identification facilities using continuous solmoid coils wen- ever a large solid angle. presented in the PEP Sumner Study Ri-port cf 197-," This report should therefore be considered as »r. The problems involved in designing such a updating of the lp/fU study based on the SPEAR detector were considered separate^ by three sub physics results which have become available with groups of the General Purpose Detector Group: the in the year, the decision by SLAC to build >i new Streamer Chamber Subgroup, the Time-Projection Mark II detector fcr SPEAR, and, of cours?, new Chan.ber Subgroup, and the Wi*-e-Chamber Subgroup. techniques and new ideas which have been advanced. The first two subgroups have submitted separate reports to this Surimer Study. We present here the Our starting point is the assumption that the conclusions of the Wire-Chamber Subgroup. Mark II SPEAR general detector system will be available as a facility at PEP. A general detector II. MAGNETIC FIELD CONFIGURATION system specifically designed for PEP must there fore be complementary to Hark II and be designed Three possible field configurations were taking into account the much higher energies which considered in relation to their suitability for a will become available. general detection facility for PEP. A magnetic- field transverse to the beam axes was rejected During the past year, studies of new techni because a large field integral Is required ques for building superconducting coils requiring {"* 10 - 20 T-m) and the synchrotron radiation a small fraction of a radiation length of material would produce a serious source of background. The have made considerable progress at CERN, LBL, and possibility of using a toroidal field was also elsewhere. Descriptions of this work given during given some consideration. It was concluded that, the Summer Study^ lead us to conclude that, on the the toroidal field is not optimal for a general PEP time scale, we can foresee with considerable detection facility for the following reasons: .-onfidence its use for high-field (-* 1!» kO) sole (1) the forces on the conductors to produce the noids with radii of about one meter, based upon required field integral are very large, (2) as 'i the data available to us, we will therefore assume result of (l), the material required LO restrrlii that a continuous solenoid coil of 1 n radius with the fcrcea will obscure too large n. fraction of Vj ktl can be constructed, and that the I.otal material Uie ap'-Tturs, and (j) the fringing magnetic field el the conductors and cryostat can be held to about Irr -i relatively open conficuratlon would place i-me-half radiation length in thickness. There i3 sevT" limitations on the choice of detection n distinct advantage if this assumpti' r. is enrrcet . sT* Ltff'T.1 " / wagpjj „; ,',i I FiC. \ Si*io vli-v cf ti.r n-nlrdi rcgicr, of * contijr.iC-'J=ly vr'Uttit, a4p?rcea:i'4«lifi£ End vi*-* of the proposed ctgncl ccn figurJ.ll««. »eru»N YCK£ •A^..]-v. i SIDE VJEW Fit;. *• :.iie vicrf • I » t'Oii'.itiuousIjr rfound, Mfi. <* ttngnctlc Mux sap for thv pvposri 3\tpvrc*TMi'.icti:.(; rolenc-id engnet phoning cngiiet configuration si :•.-••• inc lines i u.» proposed f.- cfifis<-tie flux return. const ant trngnetic field OAII. Since PEP will prod'ie, Liii.tr r. ocvtttur. EecondnrieB that the angnet design be compatible witl. upstream than SptAH/ » larger f i- Id integral is dt-iiirnblc and dovnsf rcntr. external detectors* '.;«• l.nve adopted to tinl:!.iy.r tt»- i-rrrr IT t.lj. wr-mtuc. particles. theite suggestions as design criteria. This 1-irger livid int- (ml crm u- achieved mi a smaller roiling iuni, ritice tl:.- e©'-t of detectora is A possible design for a tnagnet •••l.iel. appears typlcuiiy prepot'liauhl *-«.- U.. s-ninrc of the radius, to noet the above criteria is ahovr. in figs. 1, 2 potential ccat a-tvlnr.a me eonsidi'mltle. and 3» The cryootat and coil are assuwii to repre sent n totnl of about one-half radinLir-n length. To supplement f-sirk II, it uould be desirable The physical thicfciienK of * 15 an is taken from for a now PEP device Lo crivi- n largf.;- solid angle the CERll ISR solenoid design* In Vi,;. U, the vLtliin which irt*»mire*M"itfl <>f ehnrged-particlc mapnotio flux fr.ap is displayed. The wigm-tic field aoewnta and /-ray vniTflv.- emi tie mad-: aith adequate is uniform to within t# in the central region of occurncy. Considerations of r-j phydicr, which the detector. are discussed elsewhere in this report, * (strongly suggest that u com- of - l'/1 opening angle bo left It should &e noted that the inn i-rtigiit-tle open In tl»* forward nji.i lnekwjw\l directioitu mid flux return shm/n is, in n tense, -luite ariibrary. TABLE I CEtnUAL BETECKB PAHAKETEKS • 8 = 30 Sz = I cm •9 = 30° 82 = I mm Magnat (coDtlmiDUB soltnoiJ) • 9 = 90* Radius of coil Coil thickness 0 ' 1 1111.1..J Length of coil Radius or TVturn (a) _ Magnetic field 5 t B. Drift Chambers 4 1 Radius of Chamber (cm? 3 -w / t / f / A' 2 \ \ 1 Spitlal n-jPlutii £ n 1 1 aiiculhal u loiii'.l' iJlrml •.ilm-tP :. • 1 1 (b) _ 5 6 -1 b 1 v 5 " " "'' y •:• -In ;• •= 1 Or X - L.W» >rt 1..'- ' 1.1 •' A h -v, X * i>.01 > 1. 1* !.,.« 1.' J 3 •. \ / — tips, the lrcti return path is aln.ost completely 4 _ 2 *"••. M^S* '"" arbitrary and sliruld have ncjilifiblp effect on the internal field so lon^ as the iron cross section is adequate to carry the flux. Both the end 1 pieces and ycV.f e;in be rearranged to suit almost any external detector configuration. Cost would n 1 11 1 appear to be the only decisive consideration. The specific: arraivuinenl shown in compatible with the 0.1 0.2 0.5 I 2 10 20 external detection systems we have considered. p (GeV/c) 111. CENTRAL DETECT® Fig. 5 Komentun resolution a(p)/p for charged particles in a 1 meter radius aolenoid For the purpose of measuring charged particle with l;.- kG field. The position resolu trajectories, we have chosen to consider cylindri tion in the drift chambers is assumed cal drift wire chambers (WC). During the past year to be XyO mr. Curves are shown for considerable progress lias been achieved in this i-eaolutiot. along the bean, direction of field and a number of new techniques are under test. values Sz = 1 mm and 62 -- 1 cm for two Progress reports from CERN, LBL, and SLAC were polar angles. The effect of multiple presented durinR the studv. The basic fact which scattering in 8 double drift chambers permits designing a device around a small, high- is indued. Calculations are shown field magnet is that an accuracy of about 150 um for two dir.eiiOt thicknesses of in a direction perpendicular to the field lines material p**r chamber: (a) X = O.OOJj appears to be achievable in large systems. We radiation length and (b) X= 0.01 assume it in our calculations. It is loss certain radiation lepgth. whether or not the optimal design of EWC will use a separately-shaped drift field. New designs at The p >le tip'- clearly must be removable for access LBL and SLAC now under test eliminate this feature to the inLcrinil c*Tilr"!l detector. They are there and minimize the total material of the detector. fore flexible and li.t.erchanceable. Those shown Our view S.s that it is highly probably that the leave open u 15° eor.c. for 2-T physics and as a total material can be kept in the range between result produce H:e lield nonuniformity of Fig. U. 0.10|j and 0.01 radiation length per cjian.ber. Our calculations are made for Doth values tc indicate A closed pole tip wo.ild produce an almost uniform 1 field, while ?• rr.'ill, vestigial pole tip would tli' sensitivity of o(pl/p to this quantity. severely distort tl.i- internal field wtile leaving ^ i0° open, in .ill L-.^'S, the fringUu field I'reir. The method and occuracy of the Ion; itudinal jO < •) < VA ° is ridii-.Jil. Apart from the polo (z-coordinattj) Mcasureiicnt is Lbs n.oat c^ntro- recognition for realistic PEP r.ultiplirlt l-s. In particular, careful Kent': Orlo studii .. si ould Le carried out for the competing lu-'thod:;. IV. CHARGED HADRCN IDEHTIFXCATiai We propose to identify charg •-. J.'tJron." (ifr, K", p) with momenta up to 1;. h\"t u/im tiiw-or-fliiili' (TOKl measuresnt and h-r-r.i. v tayginy. Complete hadron separation is p m-it li- in the solid angle range not covered by ihi- j] & return iron, i.e, - A. TOF GysU-m 't?0 scintillation counter;- fclO \ . •.' * : !•:;-') are arranged cylindrically at a dlst'u.o- •! -A1 cm from tho interaction pcint. Each COJ.V. • r i. vi-'wea bi1 2 photomultipliers (2" phetocathodi-a, e.r., XCA-dr/pj). With a similar system of TOF counters the DASP I'ollaboration at DESY achieved e TCF resolut ion of JT = 0.26 nsec." The TOF measurement was started by the signal of a small scintillation counter close to the beam pipe. At PEF with the bunch crossing signal (AT~- 0.13 nsec) a more pre cise start signal for TOF measurement, will he avail able! We therefore expect a TOF resolution of oT = 0.22 nsec. Figure 6 shows the ji-omentum pmax at which hadrons of different masses fire separated by >-T as a function of the TOF path length. Fig ure 7 shows pmax vs. polar angle 0 fov yiia proposed TOF counter arrangement. At -3 = Q0L' n and K ore separated up to 1,'j Gev/c, K and p up tc ;Vf GeV/c, and JI and p up to 2.8 GeV/c. In the range above these momenta particle masses will be determined Fig. 7 Mass separation by time-of-flicht as by Cerenkov tagging. « function of polar anf-li" f1 for 't m puth length. B. Cerenkov Counter System verGial. Three methudn were discusscd> namely, ft system of three threshold Ccrenkov counters delay lines, small angle stereo with wires, and CI, C2 and C3 will serve to identify hadrons in large enyle stereo with cuthode strips* Eicli of the momentum range not covered by the TOF measure these methods is now under development and test. ment. Table II lists the detector parameters and Since no firm conclusion seems now possible, we Fig. 8 shows how the counters sre used in identi have made our j(p)/p calculations for two extreme fication. The combined Cerenkov counter and TCP cases to show the sensitivity; (l) v;, •= -- 1 en which Information produce unambiguous particle sifinatuivs seems now achievable and (:.') *.2 = t 0,1 cm which except for the momentum range of 2.U to it,9 GeV/c, appears possible* where K and p cannot be separated. If desired, this gap can be narrowed by increasing the TOF Calculations of the momentum resolution are path length or using a higher Index of refraction given in Fit;, 'j for the chamber configuration shown Cerenkov counter instead of for in addition to) in Fi(5- 1. The characteristics of the central Cl. Silica aerogel radiators^ could be used to detector are summarised in Table I. separate K and p in the momentum range not covered by the proposed gas Cerenkov counters. Gar study concludes thai. IWU'a are likely to te a leading candidate for the central detector. For the development now underway, wo would empha Each Cerenkov counter is subdivided into size the importance of two factor:': (ll the optical cells small enough to be viewed by one '}" minimization of material both lor low mrit.eiituir, photomultiplier tube. The light collection system charged particle and for low eneiyy y-ray measure has not been designed in detail, but the optical ments, and {?.) the import;inc<' cf the z-coordinate cell siv.e and arrangement have been cho^i'i! in :\ K.oasuremeiiL technique in iv la'.inn to pattern way that the optics outlined in Ref. 6 c;u, Le -4 \i-*« W.)U -Atf rot »'t'W-'" '• •?-•*. . _.Qi-C-CS £•• T - OOJIOW.IO pH--t • s -! ci-e-cj :->>' 11;. = o p identification '.a(-t;lin: with threshold ft*rer:r ' •. irc'intn-s CI, Cr, mid C-'. ltesV.ii Cfnaldcrim-iita tl- prcssiiri:-. .i < CI in < i-dfj- •-.. cinir.i.-.' tl •• press w ' i.- ••;•. 'Hit fl" t: iito :• i •> 11 ij.-, ceSJitiry f'-r ;t •"ii :•. -. I t-t ln.rlidi. •ivtl.s 1. in T'.i !•• II i.avt utfoQxxk •rdiiir T :£yC>xX> n,. wi I red t-> U :>' i t lour (> " ._!._. - -1 1 - ein i''. . For CI ;md c: u-ed IJ0 = .... wl. id uot.si-rvitti :: Ut.l- J'.-r (.11 Jj" PI oton.ul .ipllL-i- u S< n ('llPSI'll. Fift. 9 Side view of the Cerenkov and tirre-of- Up = l'.'O a: ij neo'.ss-ii'y l'or C_; in order flifi.t counter arrangement. not to ir>«kc '.)." counter prohibitively lonf,. This cnn be ueliii-v i by r.nV.int: use ol a hifh-nensitl- vii.y pliotoc.ultlplif;r ir wtivi-leiu;Ui sltiit-'iT. TOF Cl WOJ: pl;*ccd closest Lo the internction r",-.iiii in nrder to iT.inirnir.e the area to 1e covered 4* 'a- wit.li pressurized colli'. A crude cost estimate, including scintil - lator material, photorcultipliers, voltage dividers, mechanics, discriminators, ADC's^ and TPC's (only for tl:e TOF counters) rives: VV S/ C3 riOOO per TOF channel (T channels/counter) -ridOO per Cl cell N mOO per C", C; cell. - C2 A c Tot.Jil cost is: Coil TOF systfin: "J^O x 11000 = -•' B^tO k i 1 1 Cl : L4c x ltiou = lit I* k re : ?:- 't x 1100 * 3itfa Jt C; : y.'O x 1100 = ¥j>7! k Fi£> 10 End vie/; of the Cerenkcv and tii:.>'-r'f- 'I'otiil cost ni- p'irLicle rilfiit counter arrangement. ide»it.^fication ayalor. = •fl^U? k. 1 ' At Elwctron-Photon detector (ii A Tlif Tirol blo<:h of the culcric*:tcr (;ce Fig* 11) conaiBia of a lead-liquid argon detectrr. between each 1.** c*. aheet of load Is c - w. ell " " of liquid ar^on with a pltr- cf vires for ion _. collection in the niddlc (awe Fit. 12}. The wires 'in.- bussed together to (Vrc channel-* .,j " •". "*'„" ' ' " ••or.stam. iiimU-. The U6c «>J wire pinner rui);<.T ' ii. t ,,'.')".,. than strips sicplifii-s '•ci)8tiu<'lir>n ft *.!.<• vtiii- obli'-wi.iHi channels used to keep t-onat-'int nn*']-- nOCeptfcliCe. Ti' achieve j;ocd rtsnl-jLl- j. in hot I pr.lat- uu ' nzlcuUml anglec, »t- ;i3t' otf.-rloppint cl nnn-ls in •-• and ; • Readout in •> is alternated v.-iU. ., -i. indicated in Fi<> 1^- Tin- enercy read-itim. : J-'ic. U fM •..!•••• •••].• •'••I-«CII. t,lis sycl.or. is expected to ve (E)/E ^ 7? ^ g (E in r-V). Ci crursi-r the photons wl.i'-i. .<:.-.• c in the angina coil (ace Kit. I;i0 will ..'.v.- *• • ;> v. .'./j.'iu:;. •v:am:.i;t- -ner&v rosriution (dee vu-.. i-b). ru- ie^i..ti ;. sl.o-m iii Fir. l.'ij should t- cadej in ^.mdi-j^..r- /• nil-ill" •• L i i. • t- r ii; idea '<• deleft '.o the nl"!V*> •.•••silution frr those photn.s wj ic:. elect (vi,a, ju .(.., •:. : <\- !T< • 1 Wii w- itrttl liUvi- (.-jnve-rt. r-'iis< The fi- ;•' : " 1- .; • : •],•• Vl ;e" <-i' '.re si nvii ill TM If 1IJ. i. •• :•• ; ;i- is dflf-e<-nVtii,- i ana In rrd> r M. f.<-lp si.-pRi-.-ite n's ani t-'i ..e >}(V'-,:: 1/1> : t- .;. Ii . .-mil- a' U.r r-m ral scgn.enl tin. detector ii. d-pu.. TU- first ti.rw ^i-ti'i-:' r {i;.tJ • •> •••A :i\). r&diutit.n l**nrths are otw.hitk J to oitai:. f-/*t- r-ii;-- Ld plute.4 total, vac), ; radiation length «= 1,U i - ti identical to 3 - 9 identical to 2 - d channels identical to 1 - •$ displaced by A $> /? «-....»... , - d channels displaced by d8/2 - $> channels of constan1- u? - 9 wires joined to form channels of 1 ' ' constant A 9 5 cm at 9 = 90° 10 cm at S = 30° To collect ionization, we sura all channels of the sane angle In two parts: one up to 3 radiation lengths, and one from k to 12 radiation lengths. For example, one sum is 1 + 5 + 9 another is 13 + 17 + ?1 + 25 + 29 ... + 1»5 Total readout: 20 x p 0 cells (l;0 mrad) 22 x ? 0 + At»/2 16 x 2 ? lti x 2 (p + AC|t/2 Total = 1^2 channels Ceiii at $ UK// channel = $l^.?k Fig. 12. ShowiT Detector 0.4 X0 2.0 X0 1 1 1 0.8 - la) ^_ "- = 0.6 b i»v» . .•..-/ /i '' -1 ( t- \d\ m 0.4 — \ m< o r / X £ 0.2 X 0 0.3 - (b) 0.2 0.1 1 1 " T^^^^s^ 0.05 0.1 0.5 1.0 3.0 PHOTON ENERGY (GeV) rig. 13(a) a,b,c. Probability for photons to convert in an aluminum coil of Indi cated thickness in radiation lengths (XQ) with one or more electrons of energy ^ £ MeV coming out of the coll. d,e,f. probability for photons to planes. The readout is again divided intc i-ells convert ii. the coil with no electrons of constant A0 and .i? (in this case 100 mrad). of energy ^. 2 KeV coming out of the Improved resolution in angle is achieved by u&inL* coil. a second set of cells offset by ±9 /£ and -lv/,'. By combining these channels weighted by the pulse Fig. lj(b) RQS energy loss for photona traversing height we can find the centroid of the shower ir. an aluminum coil of indicated thick- both the lead and iron detectors. Theae results are based c Monte Carlo The second and third sections consist of ^0 en. calculations. of iron each: SO plates of £ cm thickness in the second section and 10 plates of ^ cm thickness in height, and the next nine are combined to obtain the third. Wires are divided into 100 mrad cells u second pulse height. in 9 and The first section consists of 20 plates each He have presented a general purpose detectc; 1 err. thick separated by liquid argon and wire design lor PEP. The core detector consists of n X'j kG, l-mt>ter-radi«3 routinuoua-coll super should he ascertained within the coming year conducting coli-rmidal r.fltfnct inside of which "re through developmental work now being carried out eight drti't wire i'l.rir-.-.-ru to emBBuru the positions ut various laboratories. and tr.otn.-ni.ii oi ii;;,i ('LJ particles. 3V.a con- detect- or covera 0,'i x ;*n :;r solid angle* The momwituir. resolution for chare*'d particles 1B calculated to he o(p)/p * o.y? Htp-1 GeV/c, H * 9C°, and 0«OCV> rudiutirr. length of material per choitbcr. Idt--ntlfiuutl.il. of cl.arced hadroiis over 0,kj x ^n ur D. R. Nygrm, PEF-l'*'*, Proceedings of the ia accomplished usim; titr.e-of-flight and Cerenkov 1W PEP :Jurj*r Study, p. ;#; 19TJ PEP counters* n, tf., mtd p ure separated river the Surrju'r Study. comploto r.cmentnt. r'Mt'v ap to Vj OeV/c except (or See for --xumple the experiment^ Of the the range ;v< tn •.," i^-V/c where K and p are not Weak iHtetwtiono reports in thir. v&luroe separated. A leaa-lron, liquid-argon calorimeter (197'- FKi: Summer Study). covering o,U'H -..T ,;r (eight modules) is used to Report or tlie General Ueer Kaf-iet Group, detect ptirtoiis will, energy resolution OtO'f/JZf mid tin(:ular resolution j{ti) 5 JU) a A. Barbarn-Caitieri, et^B^., PEP-lWJ, Pro 'j.(J rj-ai. FJ->!r :. -il- cinetry lr, provM- •'•• '-ilh ceeding vi the l^k PEP Summer Study, p. lie. ^(Ehad'/iiMKj 0.'i1/^ EJihd* n/e rejection of Set for example H.A, Green, LBL- ^677 (1975),', 10-^ to lv'-J tor ircr-.nta from O.J) to X,'j GeV/c and A O-ner-il Users MUgnet Design (PEP-I89) is obtained ty s^c-ut lng the front part of the \n vhir vtlioni: (1975 PEP Summer Study). calorimeter rad.' ally • In addition, the large See for i-xBD.plc, C. Buschhorn, et al., path of irri, pnvid.'j; 'in excellent rr.uon filter I-EP-l1"-, Prcceedings of the lS'T^PEP Summer with netjlif.it 1«- pun^-h-through,. The only ambiguity Study, p. at. in muon identification will come from decays in flight. Ctnt'S ff 1° [-peninc angle forward and Report oj the Two-Photon Physics Group, 1H7^> backward l.nvt- l-ci. l.-n open for two-photon PEP Sui:jT.er Si tidy. DES'i Report Ts/V*, Kay l«75f and submitted to physics. Photnn-ta(yini; apparatus as described in Hef. 7, or possiilj,' a simpler version, can he Phys. L-'tt-.rs. placed in this r>-,-i ti. M. Cnntin, F. Bulofl ABSTRACT Starting from a set of desirable features* a 3. n-e separation general shape of the detector is chosen. To decide 4. Real flexibility over 2* azimuth. on the size 4nd magnetic field 4 cases are compared. 5. Provisions for fast trigger A good compromise is chosen and illustrated here B. Limited Solid Annie Features - Additive General Features 1. Finer measurements for jr'a The general features chosen were divided into 2. Particle Identification extensions two groups: Basics that should be present over 3. Provisions for a forward tagger to help as near 4ff solid angle aa possible, and others identify 27 processes as recommended by that would be costly and complicated such that the 21 group study they would be used in a fraction of the total solid angle. However the magnet flexibility should allow with these features in mind a solenoldal adding more coverage at will magnetic field is chosen. The flux return la a set of 8 ribs capable of returning the flux with A. " 4"" Solid Angle Features - Basic out further additions. A rib must block as little 1. Good momentum measurement space as possible azimuthally. In any case, 2, Energy and position measurement for r's enough space is provided between the coil and the Magnetic Field (K.G) -— 5 12 15 Coil IP Outers) -— —..——— 1.5 1.0 .85 No. of removable flux return sections -- Cross-sectional area of Mag Field (SQ. Meter3>- Cap area required (5Q. Meters): 1 Field in iron 18 k.g. 1.96 2.1 1.89 1.48 2 Field in iron 20 k.g. 1.77 1.9 1.70 1.33 Cap area available (SQ.Meters) Width of flu* return rib (Meters) (Thickness fixed at 20 cms. 1 Field in iron 18 k.g. 1.25 1.32 1.2 .95 2 Field in iron 20 k.g. 1.10 1.2 1.05 .85 X-sectional area for radial coverage out from coil surface of 40 cms. and full azimuth: A.5 3.3 2.6 Mom. Meas. capability (Mark II Units) 1.0 1.0 .57 AH M0 k.g. coils are superconducting. Mark II as presented by B. Richter during PEP Study is based onM115 cr.id radial track measurement, 200 microns dr'.ft chs, resolution and 5 k.g. field. Quoted momentum resolution .75 tlcev, ) , (VP)c( H„, c FiK. 1— Cross section of h.isii- dele. n>r 1/4 shown only. A Space available (at pipe cntr and prop chs. 3 Drift chs - momentum analysis. C Triggc- cntrs. D Coil st E1 I part of shur cntt lead-scint - pmp t lis sandwich. E_ 2nd si,wr cntr lead scint E,+ i:, j;£v.' rnergv meaS. Division helps jr-r separation. F Flu* return rib, G, and G- Struts for structure. (;. is separated from end cap by sp.ire tor Ii)):( guides - opening 0. H, a^ H, End cap and end flux rib. I Shower:* Either liquid arnon «r IJ. in J Prop chs - Coarse digitization - n-. I .-. forward trig. cntr. K Front tagfier opening do' ). it ions - 1/2 shown If licht guides arc used tti, shown exploded- through K or I) whf.h i .in be purpose. ribs to allow continuous coverage with most of the Since the magnet shape is the same for all additiv? detectors invisagod. The thickness of cases and case III seems like a good compromise, the rib azimuthally was chosen to be 20 cms. this case was chosen for illustrations. Figs. 1, 2 show the structure of the magnet including all To decide on the radius of the cot 1 and the the basic features listed in A. above. value of the magnetic field 4 (Nises were considered all bavins the same features above. The results arc- shown in Table I. Flexibility played a major part. It LS to see that any experimentalist can build on From the table one concludes that case IV, more fluK return sections (Iron is relativi-I the smallest magnet, is not useful because of the cheap), equip it at will with additive foamr loss in momentum measurement capability and the test it without interfering at all with an a marginal cap area available. experiment. When ready the new sect ion/sect can replace uhe old ones in a very short ti"i Cases I-II1 arc acceptable. However case I The replaced sections van be used similarly the largest magnet is costly from the point of physicists next in line. Although no* shown view of additive features {l,2). Case II is also Figures, it is trivial to give the physicist still costly for the same reasons and also requires option of removing the shower counters togei a rather wide flux return rib. Case III nei»ds with the flux section or leaving them in pl;n careful consideration by magnet experts to deter (for example if he is interested in adding ^ mine the shape of end cap ribs. tion only). THE STREAMER CHAMBER AS A GENERAL DETECTOR G. Barbiellini, R. Kotthaus, S. Poucher A. Seidl. F. Villa, and D. Yount Abstract A general purpose streamer chamber facility lode- (3) High Information density of about 2 streamers per scribed. . The multiplicities and momenta of charged cm (useful In pattern recognition, momentum, and particles are determined in the central region of the d'J/dx measurements). chamber over a solid angle of about 0.85 • Ait, The re mainder of the visible volume is devoted to irVe* sepa (4) Large solid angle. ration and to JT° and gamma ray detection. The former (5) Good gamma ray angle and energy measurement, is achieved by observing electromagnetic shower devel allowing precise determination of the mass of par opment and the latter by measuring both the conversion ticles that decay Into two photons. point of each primary interaction gamma ray and the mo mentum of each shower track produced In an array of (6) Electron Identification via electromagnetic shower coaxial lead-oxide plates. There are 9 such plates, development. spaced 8 cm apart, and the plate thickness is 0.6 radi (7) Good separation of hadrons at low energies and ation lengths. Separation of ir*/e* Is effective above also partially in the region of the relativistic rise 250 MeV/c, and gamma ray detection Is practical above In Ionization, via dE/dx . 100 MeV by this method. The solid angle of the lead- oxide shower detector is about 0.5 • 4ir. Muons above (8) Muon Identification. about 1 GeV are distinguished from charged hadrous by a single layer (which gives three track coordinates) of multlwlre proportional chambers that determine whether a given particle has interacted strongly in the colls or The advantages listed above suggest that the flux return plates of the magnet. This "external muon streamer chamber would be particularly well suited to identifier" alBO subtends a solid angle of about C. 85 • 4ir. the following types of experiments: Good separation of ir/K, K/p, and ir/p is obtained by ionization below 0.7, 1.0, and 1.4 GeV/c, respectively, (1) Any investigation requiring measurement of high and useful separation (about 1.6 standard deviations for momentum. ir/K, 0.7 standard deviations for K/p, and 2.2 standard (2) Strange particle production (V's can be recon deviations for ir/p) is achieved in the region of the rela- structed). tivlstie rise in ionization. The ionization measurements are supplemented by time-of-flight, which is eifective (3) High multiplicity studies (pattern recognition). for TT/K, K/p, and ir/p at momenta up to 1.0, 1.7, and (4) Search for new particles. 2.0 GeV 'c, respectively. (5) Genera] class of experiments in which one particle or more Is identified (e.g., it inclusive). INTRODUCTION (6) Inclusive spectra of particles (x°, i), ...) that de The streamer chamber has been discussed as a cay into two gamma rays. possible central track detector in a previous PEP Sum mer Study Weport.i Advances in technology during the (7) Inclusive spectra of charged particles. 2 past year, particularly in spatial resolution, in film- (8) Total cross section. less readout,2 and in particle separation in the region of the relativistic rise in Ionization"* have added signif (9) Multiplicity of charged hadrons. 1 icant!" to the already impressive arguments favoring (10) Two-gamma-ray experiments (gamma ray tagging this ..^hnlque. We have therefore re-examined the must be added). streamer chamber approach with two basic questions In mind: (1) What auxiliary equipment can be added to the GEOMETRY central detector already discussed to make it a general- purpos" facility? (2) How can the recent advances In The streamer chamber facility is shown in top view technology be implemented to optimize system perform In Fig. 1 and in beam view In Fig. 2. Two streamer ance? chambers are placed back-to-back, and two coils in a Helmholtz configuration (coil separation equals coll ADVANTAGES radius) produce a rather uniform magnetic field of about 15 kG parallel to the beam pipe. The electrodes of the Among the advantages that we see for the system chambers are perpendicular to the beam pipe and yield desciihcd below are: electric fields parallel to the beam pipe and thus to the magnetic field. The chambers are driven by Blumleins (1) Superb spatial resolution, possibly better than a = and Marx generators on one side of the magnet, left 40 fi in the plane perpendicular to the magnetic open for this purpose. field and ir = 120 a along the magnetic field.2 The inner core of the detector, extending from 25 (2) Excellent multiple-track efficiency. cm to 90 cm in radius, is devoted to dE/dx and momen- Fig. 1 turn analysis of charged particles. The region from 90 OPTICS cm to 180 cm in radius is provided with lead-oxide plates that permit 3i0/y and ir°/e* separation, as well It has been demonstrated experimentally in Kef. 2 as measurement of electromagnetic shower energy. A that the residues of the streamer position measurements single layer of multiwire proportional chambers over a one-meter-long track can be as low as 40 pm in (MWPC) with delay-line readout surrounds the Iron flux space. These Tesults were obtained with a digitizing return plates and is used to identify muons via their sensor (Nocttcon tube, Thompson-CSF TH 9655) at a failure to Internet strongly in the iron. Streamer track demngnificatlon of about 100. In the present application, data (including brightness) are digitized octant-by-octant charged-coupled devices (CCD-211 by Falrchild Solid using arrays of self-scanning "charged coupled devices" State Division) could be used. These devices have a dy (e.g., CCD-211 by Fairchtld Solid State Division). Drift namic range of about 1000 fo I and a light sensitivity of chambers just inside the magnet end plates provide some about a factor of ten higher than that of the most sensi sensitivity to charged particles in the smaIl-$ region tive film used to record streamer tracks. Equally im not accessible to the streamer chamber. The trigger portant, they can be operated in high magnetic fields. consists, for example, of a beam crossing signal, o Thus the charge-coupled devices can be installed with two-particle signal from beam pipe scintillators, and a the remainder of the optical system within the high-field two-particle signal from an array of scintillators lo region at a distance of about 40 cm from the nearest cated outside the streamer chamber and covering as electrode. large a solid angle as is practical. The latter counters The number of CCD sensors required can be cal are used in identifying particles by time-of-flight, sup culated from the sensor dimensions, the field of view, plementing the ionization measurements. and the demagniff cation. A possible layout compatible -120- MWPC FOR EXTERNAL MUON IDENTIFICATION Fig. 2 with the required spatial resolution is shown in Fig. 3. rates. In the outer trigger counters, we require two or The total number of sensors needed for the facility is more signals at polar angles 0 that are consistent with 192, and the number of lenses is 112. those recorded by the beam-pipe scintillators. If the outer scintillators are replaced in the trigger by multi- + _The rich Information to be extracted from a typical e e~ annihilation at PEP can be stored in about 120,000 wire proportional chambers, the correlation in 9 can bytes on a disc or fast tape. The reading time of a self- be made rather restrictive. scanning CCD is about 30 msec when amplitude informa With the trigger outlined above, the Mark I detector tion is recorded. Thus the 192 sensors must be read has trigger rates that are typically ten times the out In parallel. eventually-analyzed event rate. For a total cross sec TRIGGER tion of fftot = 20 nb and a luminosity or S£= 0.25 x 1032/cnrvsec, our event rate is 0.5 per sec. If the The streamer chamber trigger consists, tentatively, total cross section falls as 1/s, then the value at 15 GeV of a three-fold coincidence involving: (1) beam crossing, would be about 1 nb, yielding 1.5 events per minute and (2) a double layer of scintillators (or mulUwire propor probably less than 20 triggers per minute. We believe tional chambers) surrounding the beam pipe, and (3) an that trigger rates as high as 5 per second are manage outer layer of scintillators just inside the magnet coll able, and we therefore conclude that die expected rates and end caps. The outer scintillators are used in coin are quite satisfactory. cidence with the beam-crossing signal to provide time- Should the background rates prove to be substan of-flight information. The outer scintillators could be tially higher than expected, a layer of scintillators replaced or supplemented in the trigger by multiwlre could be installed just inside the lead-oxide plates. proportional chambers, and another layer of scintilla Timing between various scintillators will provide ade tors could be installed inside the streamer chamber quate cosmic ray rejection. just in front of the lead-oxide plates. After the chamber is pulsed, some type of thresh From the inner trigger counters, we require sig old requirement on the total light output should allow us nals In both layers in at least two positions. Experi to reduce substantially the number of non-events logged, ence from the Mark I magnetic detector at SPEAR indi Just as a cut requiring several drift chamber hits would cates that this substantially reduces the background reduce the background In the Mark n detector. CHARGED MASS KKSU1.UTION The momentum resolution of a streamer chamber is Riven by4*r> U'l I P L, I p h ISy? .. 4, (17 • lll'^n 4 ,8p+fn (145p/M)l + *"Tfls IPI./Jeos2* . 1.7C- lo'GHan2> + —ns is due to multiple scattering and dominates below about 700 AleV/e, and where lAp \ _ 1.44p c 1.2-10~g e sta A m H L* cos A L c06 A is the result of measurement setting error and domi nates above about 700 MeV/e. In these equations, p Is the momentum in MeV/c, M is the particle mass in Fig. 3 MeV/c-2, ft = p/E Is the velocity of the particle relative to the speed of light, H Is the magnetic field in kG, L is the measured track length In cm, A is the dip angle An estimate of the cost of scintillators, photo multi- (90° - A = 0, where 0 Is the polar angle), and E is the pit era, and electronics for the trigger counters and spatial measurement error in microns. time-of-flight detectors is tabulated below. The total The results lor a streamer chamber in the present cost appears to be in the neighborhood of $260,000 to geometry with c = 40 fim and H = 15 kG are plotted as a $330,000. If the inner scintillators are replaced by function of momentum and dip angle A in Fig. 4. In multiwire proportional chambers (e.g., four layers this calculation, we have not included the interaction with 5-mm spacing and 250 wires per layer), the cost point. The momentum resolution le excellent ox accept of this component can be reduced by a factor of two able over the full momentum range of PEP and over our (from $50,000 to about $25,000). full angular acceptance, COST OF TRIGGER COUNTERS AND ELECTRONICS No. of Array Dimensions Dollars No. of tubes Counters Outer Scintillators 64 2x19.6x360 cm3 128 90K Inner Scintillators 32 2x 7.8x150 cm3 64 45K Central Scintillators 32 2x17.6x380 cm 25K (optional) 45K Two End-Cap 64 2* £\ x 150cm3 25K 45K Scintillators 2 Total with Central 192 2 cm x 90 m 105K 320 225K Scintillators 2 Total without Central 160 2 cm x 70 m 80K 256 180K Scintillators The precisions in dip angle (AXJ and azimuthal where the other symbols are defined above. The re angle (A^) are given uyr sults obtained with the usual parameters arc plotted in Fig. 5 and 6. The two-body mass resolution is given in terms of the momentum and angular uncertainties by1 2_ 1.2 • 10"" e ' cos > 1. 76 • 10"" L (AM pV (AM) ^oosd) 4i + G 2 3 ^N"*' L12 3.8- 10" e J 1.3 • 10 " L + PjPo sin a (/if-)" i i r r| I i l l 1 1 n | - X'60"i . " X*0" - - X--45" _ X ' 60° - - 55° " - / " V \ - 50°" \ \ 451 - ^0" ii "1 i TrTTTTi TTTT^ - ~ ~ T ~ ~r--T-n--r T-f^r--^= 0 (GeW:) p (Gev/c) Fig. 4 Fig. 5 chamber serves as a pair spectrometer consisting of 9 where a Is the angle between particles 1 and 2 and Act plates spaced 8 cm apart and having a total thickness of Is the corresponding error. 9 x 0.61 * 6 radiation lengths. In Fig. 7, the mass resolutions calculated, respec The "visible" energy spectrum obtained by Monte tively, for a phase-space model and for a jet model at Carlo calculation for 3-GeV incident gamma rays is Vs" = 30 GeV/c are plotted. Figure 8 (jives the mass re shown in Fig. 9. A low-energy cut-off of 10 MeV has solution for the ?(3100) as a function or momentum for a been assumed. The horizontal scale gives the fraction dip angle of A^ = 30°. The streamer chamber yields ex of the incident energy actually materialized, and the cellent mass resolution, which is necessary in finding vertical scale indicates the number of gamma rays (out narrow states in the presence of a possibly large back of 2000) in each bin. The half-wldth^at-half-maximum ground. is i 7% of the most probable visible energy. In Fig. 10, the energy uncertainty (HWHM) in per NEUTRAL MASS RESOLUTION cent Is plotud versus photon energy in GeV. The Monte Gamma ray energies are determined over a solid Carlo results at 1, 3, and 6 GeV can be summarized by angle of about 0.5 • 4ir by measuring die moment?, of the expression positrons and electrons pair-produced in an array of AE, /E. = a+ bE (5a) lead-oxide plates. In effect, this region of the streamer (5b) 12 - - x-- o° 1 1 1 1 10 X = 45" - A = 60° - 1.6 - - !° - 1.2 " " ieV~"^ a 6 - 0.8 " - 0.4 \. - 1 I 1 1 1 1 *-T|-r~|- -|-i-lTrlr — ^^^=-^ T 2 p (GtWc) M„ (GiV/t ) Fig. 6 Fig. 7 [ iii- 1 1 1 150 ~1—T-T" T—I—i—l—i—I—i—I—"—|—i—I—i—I—r 12 S E = JGeV 10 - ^^ ~ r 5 B - ^^ ^^ _ M = 3.IGev/cs I _____—-^-"^ B£s "a* * RES'3°° - l« - - yWtffUCMEhificHj 2 - i ± L 1 1 1 1 1 Fig. 9 PARTICLE IDENTIFICATION The usual ir*/e* rejection fattor for particles of fig. 8 known momentum incident on a total-absorption shower detector is of the order of lo~3. This limit is set In Fig. 11, the mass resolution AM/M in percent mainly by v~ charge exchange via the reaction JT~ + p — is plotted versus energy for a neutral particle decaying TT° + n. The ir^/e rejection for the streamer chamber into two gamma rays. The maas resolution was ob pair spectrometer should be about one order of magni tained from the expression** tude better since the high resolution permits one to dis ,,2 tinguish a single gamma ray from two gamma rays by _JM_ the small opening angle of the TT° decay. In addition, M " (6) some ff*/e* discrimination is provided by the dE/dx measurements made in the central region of the cham fi?-H*ft)-e» ber. where A0yy is the uncertainty in gamma ray opening Particle identification via streamer brightness or angle in raalans, MQ is the mass of the decaying neu streamer density (dE/dx) was discussed in detail in tral particle, and AEJ/EJ and AE2/E2 are the uncer last year's PEP Summer Study. * Good separation of tainties in the respective photon energies given by Eq. v/K, K/p, and n/p can be obtained by ionization below (5a, b, c). In this calculation, the uncertainty i n open 0,7, 1.0, and 1.4 GeV/c, respectively. In the region ing angle (less than 1 mllllradian} is negligible, and the of the relativistlc rise, the separations of TT/K, K/p, result is therefore independent of neutral maas MQ, and 71/p are about 1.6, 0.7, and 2.2 standard deviations. The mass resolution of the streamer chamber pair Recent unpublished streamer chamber data3 com spectrometer is comparable with that of a crystal ball paring dE/dx for electrons and pions in the region of (sodium iodide sphere with parameters similar to those the rehvHvistic rise are shown in Fig, 12. The data given in Hef. 6), although the solid angle of the latter is were obtained fay measuring track Lrlghtnees, rather about a factor of two larger. than by counting streamers or gaps. The total amount GAMMA RAY ENERGY (GeV) Fig. 10 Fig. 11 With the combined information from Ionization and time-of-flight, the following hadron separations can be A Electrons From y—*-e*e" achieved: 2.0 • Pions From Kj—-7T*TT' ir/K/p below 1 GeV/c, 1 Meter Track Length Minimum Expected Rise Fc n-'s ;r/ REFERENCES 1. G. Buschhorn, H. Meyer, A. Odian, and D. Yount, "A Streamer Chamber Detector for PEP," PEP Summer Study Report: PEP-151 (1974). 2. J. Badier, E. Delacour, R. Marbot, J. Y. Parey, AM/M FOR A. Romana, andR. Vanderhagen, "On-Line Digi ELECTROMAGNETIC tization of Low Luminosity Streamer Tracks," PARTICLES LPNHE-Ecole Polytechnique preprint: LPNHE/ X/75-05 (1975). _ CRYSTAL Y8ALL 3. F. Villa (private omrounication). 4. F. VUla, "The Streamer Chamber," SLAC pre STREAMER print: SLAC-PUB-1250 (presented at the Interna CHAMBER tional Conference on Instrumentation for High Energy Physics, Fraecatt, Italy, May 8-12, 1973). 5. C. M. Fiflher, "Optimization of Bubble Chamber Design Parameters: Measuring Accuracy for AM/M FOR Charged Particles," CERN Report: CERN 67-26, CHARGED Vol. 1, page 25 (1967). PARTICLES 6. T. Mast and J. Nelson, "Some Design Considera tions for a Large Solid Angle Charged Plus Neutrals Detector for e+e_ Storage Rings," PEP ' PARTICLE Summer Study Report: PEP-153 (1974). IDENTIFICATION 7. F. A. Harris, S. I. Parker, V. Z. Peterson, y,e,/i,7T,K,p D. E. Yount, and M. L. Stevenson, "Muon Identi fication Using Multiwire Proportional Chambers," Fig. U Nucl. Instr. and Meth. 103, 345 (1972). THE TIME-PROJECTION CHAMBER-1975 Dave Nygren Abstract and a. is the mean free path, o can be rewritten in more fundamental terms. New experimental results relevant to the per formance of the time-projection chamber (TPC) are In the absence of a magnetic field, the trans presented. These include certain properties of the "ball-wire" readout scheme, the suppression of verse distribution is given by electron diffusion by magnetic fields, and the ef fects of E*B components. The challenging problem oc particle identification is discussed in the SB (4) framework of a TPC. : rm, which can be compared with the strong magnetic * * * field case (WT »1) Last year in the 1974 PEP Summer Study Re port! i discussed a new concept for the detection of charged particles, the time-projection chamber. The interested reader should refer to this report tco2 for a presentation of the basic ideas. Lacking at that time, however, were data on electron trans In the strong magnetic titld regime, the best reso port in various gases relevant to the PTC. There lution is obtained with a gas giving the longest fore, an experimental program has been carried out mean free path, whereas just the opposite obtains to determine those properties. without magnetic field, tn addition, the resolution I. DIFFUSION SUPPRESSION obtained with strong magnetic fields depends much more sensitively on the electron temperature. Central to the feasibility of the TPC is the Equations (4) and (5) indicate that a minimum a existence of substantial suppression of the dif will exist for some value of the drift field E fusion of electrons as they drift through a gas. since low fields correspond to small W values, and This apparent conflict with basic motions of ther high fields lead to increased V due to non-thermal modynamics can in fact occur by the simple act of agitation energies, fhese qualitative features are imposing a strong magnetic field parallel to the born out by the data presented in Figs. 3 and 4, drift electric field. If the radius of curvature which were obtained by l^e technique described be of the electrons in the magnetic field can be made low. small relative to the electron mean free path, then the diffusion suppression can be large. The ratio Figure I shows schematically a small drift of the diffusion coefficients D for diffusion chamber designed to measure directly the distribu transverse to the fields is given quantitatively tion of single electrons after a 15 cm drift inter by the expression^ val. This is done by counting with a proportional counter those electrons which "fall through" a 70 D(B) _ micron wide slit in the bottom plane. The slit and (1) DW- ,.2,2 counter are mounted on a micrometer-driven stage so that the relative counting range versus x can Here to is the cyclotron frequency of the electron, be mapped out. At the bottom plane, the distribu eB/mc, T is the mean collision time of the elec tion of electrons is expected to be of the form trons in the gas, and D(B) is the diffusion co efficient as a function of magnetic field. d"N . The spatial distribution of an ideal point dxdy (6) swarm of electrons after a time T is described by a Gaussian form of width The slit effectively integrates over y, so that the a - /2DT (2) resulting dN/dx distribution directly determines o. The point source of the electrons is a 25 micron where a is the rms normal distance to an arbi diameter pin-hole in the top plane, in reality a trary plane containing the origin of the swarm. gold foil of 50 micron thickness. In the presence of an electric field (2) can be rewritten approximately as Electrons are generated by ionization in the space above the foil by a 140 pcurie (244QJ) a aa (3) particle source. In the source region, a separate, M- variable electric field can be imposed to regulate Here L is the drift length, and W is the drift the maximum counting rate at the proportional velocity in the direction of the electric field. counter below. Background rates are easily deter Since D = VH/3, where V is the speed of the elec mined by reversing the source field, driving the trons (not to be confused with the drift velocity), electrons up and away from the pin-hole. In To-HtMStMitervvoluge •*- To HV (or drill field Fig. 1. Cross-section view of chamber for measure ments of transverse diffusion of electrons in various proportional gas mixtures. Free electrons are continuously generated in the space above the gold foil by alpha particles born in the curium source* A weak electric field of either polarity can be created in this space by a battery which rides with the potential of the gold foil. Thus electrons can be driven either towards the pin hole or away from it to study signal/background ratios, typically 100 to 1000 to 1. Not shown in this figure is a micrometer which moves the proportional counter assembly remotely. practice, signal-to-noise ratios of 100 to 1U00 A 20% 002-80% Argon mixture was evaluated for were typically obtained. The entire apparatus fits comparison with the methane mixtures. Predictably, into the 8 inch-high aperture of a Bevatron it shows the smallest effect as the magnetic field 18*36 LPH magnet. Figure 2 shows a typical dis is applied. tribution, along with a Gaussian fit to the data. The systematic errors are typically 1 to 2 per It is worth emphasizing that the diffusion cent. A fitting program has been used to find the along the direction of the fields is unaffected first and second moments of the data. as the magnetic field is applied. However, it turns out not be safe to assume tiiat the trans An intriguing possibility noted in the orig verse diffusion obtained without magnetic field inal paper-* on the TPC is the beneficial impact should be the same as the longitudinal diffusion.4 of the Ramsauer-Townsend effect. This is the For pure argon (without magnetic field) the ratio quantum mechanical phenomenon of a very deep min of Di/Dy can be as large as 1/7. For all other imum in the electron-atom collision probability molecules or atomr. che ratio is close to unity, at electron energies of about 1/3 eV. The effect but no calculatJ.uns or experimental comparisons is significant for noble gases and simple molecules appear in th' literature for methane. Because like methane, which has an electron shell structure methane iF a good moderator, the ratio can be ex similar to neon. As Eq. (5) suggests, the very pected 'o be close to one in the quasi-linear or large mean free path due to the transparency of low F.'p regime. The measurements of Cochran and the noble gases may result in the best resolution Fo'aster5 predict a transverse diffusion of being obtained from a mixture of methane plus a 9S0±100 microns after IS cm drift for pure methane noble gas. The data in Figs. 3 and 4 bear this at an E/p of 1.1 and atmospheric pressure. Our out, although the mixtures show a rapid deteri own Tesults, referred to atmospheric pressure oration at high E/p due to increased agitation yield a transverse diffusion of 860S30 microns. energy. A Saclay group*" (E/p =1.1, atmospheric pressure, pure methane] obtain a longitudinal resolution of 295±60 microns for is cm drift. Their result, At another extreme is carbon dioxide, which however, may not represent the single-electron has a very small mean free path for electrons. 20% CH4 s\ 80% Ar EJP •> 0.20 B - 20.4 kG Field OkG Qrilt length. ISci u - 0.0073 in. Pr«* 600 Tor P = eOOTorr I / ^ 20 24 28 32 3G 40 Micrometer reading (0.001 in.) (zero arbitrary) Fig. 2. Typical data and fit for a particular set of conditions. This data represent the probability Fig. 5. Transverse diffusion results without mag- distribution for single electrons to enter the nctic field us a function of E/P. Without magnetic slit a$ a function of slit position. Systematic field, COT mixtures give smallest diffusion. Lines errors are typically 1 to 2%. This set of conditions are drawn to guide the eye. Addition of argon to does not correspond to the best resolution obtained. methane increases the diffusion, as expected. transport value because their electronics may have a substantial suppression of transverse diffusion been integrating over the arrival of several elec has been experimentally demonstrated for methane trons. In any case, the distinction between longi and argon-methane mixtures under conditions tudinal and transverse should be maintained even suitable for high-energy physics. The accuracy in the absence of magnetic fields. of the data is estimated conservatively as ±5o, and can be safely extrapolated to larger or smaller The diffusion suppression has been measured drift intervals. as a function of magnetic field for various argon- methane mixtures. The results are given" in Fig. 7. II, § x § EFFECTS The data of Figs. 3 and 4 have been combined In practice, the attractiveness of a TPC according to lit). (1) to give values of WT and T will depend in part on the extent to which non-per shown in Fig. 5. For convenience, the experimen fect alignment of the E and B fields distorts or tally observed? drift velocities in argon-methane degrades the track information. Since the fields mixtures are presented in Fig. 6. In general, the are strong, one naively expects the effect to be maximum wt is reached at E/p values less than potentially serious. This effect has been recently- those corresponding to the maximum IV values. explored by mounting the diffusion chamber in a cradle which can rotate the chamber body [and To give some specific numbers, a 1.6 meter hence E) relative to B. These measurements in (90 cm each direction) long TPC should give a volve mapping the movement of the mean impact resolution for single electrons of 500 microns point in the direction perpendicular to the plane at 20 kgauss using pure methane at an E/p of 0.5. defined by E and B as the angle between the fields If the Saclay groups' results can be a guide, the is varied. Some preliminary results are shown for resolution for particle tracks may be substanti selected conditions in Fig. 8. These data do not ally better than this figure. seem to be easily understandable in terms of simple models. Generally, the higher the IOT product, the To summarize this section, the existence of smaller the Ii * B effect. -129- 1 1 - II 1 1 1 1 L—45 S 7V / >A " - rjJ / ^^ " ^^--25_ " o CH - ill 1 10% 4 A 20% CHj, CH 0 40% 4 0 60% CH4 J CH X? 80% 4 20* A co2 m 100% CH4 1 1 0.4 as as W i i ,11! E/P (tfo1» cm -Toirl E/P (Vdti/em-Torr) XB1. 75129779 XBL 7512-976] Fig. 4. Companion set of trans- ,je diffusion re Fig. 6. Values of W, the drift velocity, for vari ous argon-i.iethane mixtures, as taken from English sults, with magnetic fie...i. Wore C02 is the poorest choice, and mixtures of argon-methane give best and Hanna, Can. J. Phys. 31, 768 (1953).The ac curacy of these curves is probably not better than results. ± 15%; other measurements indicate that the maxi mum drift velocity is closer to 10 cm/usec, shown here as a convenience in estimating TPC parameters. - 10* CHA The most significant result implied by the 20* CH A A 4 data of Fig. 8 is that the E x g effects are O •to* CH,, V - small, much smaller than some simple models would 0 60% CH, •\ predict. One way to view the problem is to com ICO* CH 279 _ 4 pare the shift introduced by ExB to the resolu- A 20% SL :ion introduced by diffusion. - 4 In the example given above, the diffusion is - 530 microns in 90 cm, or 1 part in -1800. To ^b^ •> 03 ^ - \ ^ 0 1 1 i 1 Fig. 3, The effect of nonvanishing £ *B components for three mixtures at full and half field. These XBL 7512-9782 results are very surprising for two main reasons; 1J the effect on the electron trajectory is very Fig. 7. The dependence of the transverse diffusion small- note difference in scale between electron on magnetic field. The results Follow Eq. 1 to an trajectory angle and field angle; 1) the magnitude accuracy of about ± 53. of tlie effect increases as the field is reduced. At quite low fields, = 1000 G, the sense of the effect changes to what one ordinarily expects. To summarize this section, the impact of the £ x5 effect has been experimentally shown to be remarkably small, and should not place severe de be studied are the timt response in a magnetic mands en the engineering specifications of the field as the electron impact parameter is varied, magnetic field design. what diameter of ball gives optimim performance, and what geometry is mechanically convenient. ill. THE BALL-WIRE READOUT SCHEME The magic mode occurs in argon-isobutane In last year's Summer Study Report 1 dis mixtures for isobutane concentrations less than cussed one of several possibilities for readout about 83, as indicated by flowmeters. It manifests at the end-caps,viz, the ball-wire detector. This itself at high gain and is most simply explained geometry is attractive for several reasons: as a serai-Geiger type of response occurring over, 1) very high gain can be obtained, with good but limited to, the surface of the ball. The pulse stability; 2) ball-wire data is intrinsically height spectrum in the magic mode is shown in 3-dimensional, eliminating the N- ambiguity prob Fig. 9. At higher ates, some broadening appears lem of pattern recognition; 3) the readout plane due to space charge effects. The purpose of men has (in principle] a simple mechanical design. tioning this phenomenon here is that the magic mode provides a very convenient tool for exploring Several, but not all, of the interesting the properties of the detectors, and to point out aspects of the ball-wire readout have now heen that detectors using bail-wires cm have simpler explored. A so-called magic mode has been dis and cheaper electronics because hii;h thresholds covered, in which extremely uniform pulse heights can be safely employed. on the order of 10 millivolts can be obtained even across loads as small as 100 ohms. The recovery The recovery time has been studied in two time has been studied.. The dependence of the pulse ways, one, by observing the dependence of the height on the length of the wire above the plane of pulse height of the second of two single-electron the dielectric has been determined, Remaining to pulses on the time interval between them. This Drt'jnn of tall from . XBL71W-9.8S J'ig. 10, The relationship between pulse-height and the distance of the hall-wire above the dielectric plane. This relationship is shown for a constant magic mode output, 10 MV into 1ZS ohms; the ordi nate is the IfV needed to maintain this output as the distance is changed. As these curves are for 3 Ctwg* O l*rUtr«r uoi single ball-wire, the corresponding curves for a cluster may be somewhat different XBI. 7512-9764 Fig. 9. The pulse-height spectrum observed from a providing superb pattern recognition. However ball-wire detector under "magic mode" conditions their ability to provide momentum resolution com (see text). The counting rate here was approxi petitive with drift chambers is not yet clear. mately SOD ll . z Further studies are planned. recovery time depends on the mobility of the posi tive ions moving away from the ball, and is typ ically 200 micro seconds, using the high-gain magic mode. However, the recovery time in practice IV. THU SUPER TIME-PROJECTOR (PARTICLE IDENTIFICA depends on the sensitivity of the electronics and TION REVISITED) to only a small extent, the gain at the ball. Using fast circuity with 250 micro volt threshold Last year, I discussed a geometry of approx and 2K ohm input resistance, recovery times have imately one meter diameter, and noted that while been always less than 250 nano-seconds, for a dE/dx information isavaiiablcthe - .5 meter path variety of conditions. The explanation is that length is not long enough to provide clean particle while high gain operation requires a very long identification in the region of the relativistic time to fully recover, the low threshold doesn't rise. Time-of-flight techniques are not useful require much gain in the first place. Low gain above a CeV/c or two, and several layers of operation leads to correspondingly smaller space- Cerenkovs may not turn out to be a very attractive charge effects, so that low-gain operation also solution either, for a variety of reasons. In recovers very quickly. The conclusion to be drawn principle.- one can identify e_ectrons, picas, from these remarks is that a ball-wire can re kaons and protons using dE/dx alone except where solve separate tracks which cross within its the dE/dx curves cross (Fig. 11). In these regions, sensitive volume if they are separated by ~ 2 cm a time-of-flight measurement or a Cerenkov must be along the drift axis. employed to resolve the ambiguity. The voltage required to maintain a 10 mW To utilize dE/dx with time-of-flight to pulse height/electron has been studied as the identify these four types of particles, one is length of the wire beyond the dielectric plane has driven to substantially larger diameter detectors been varied. The results are shown in Fig. 10. It to obtain the requisite resolution. A four-meter does not appear necessary to have the wire pro diameter detector will give 51. rms dE/dx resolution trude more than ~ 4 mm to achieve good response. using argon-methane.9 TT-K time-of-flight dif ference at the ambiguous point is about 0.7 nsec; Much more is known about ball-wire detectors K-p time-of-flight difference is only .3 nsec, a than has been presented here, interested parties very difficult separation to achieve. The availa should feel free to contact the author. bility of micro-channel plate photomultipliers would alleviate this problem, but their evolution To summarize this section, the properties to the stage of off-the-shelf competitively-priced of the ball-wire detector show no particularly commercial stock is by no means guaranteed even undesirable aspects and should be capable of within the PEP time-scale. Fig. 11. The most probable number of ion pairs/cm lor argon axid xenon at STP, as a function of P(,i;eV/c) for pions, kaons, and protons. Hot shown arc the curves for electrons, which would be essentially flat at the asymptotic values goverened hy the density effect. XBL75l2-96rtO The resolution obtainable in a 2 meter flight Fig. 12. A possible configuration of a super TPC path through argon-methane still allows several end cap, 36 hexagonally shaped modules of wires percent misidentification due to overlap in TT-K, fill an end cap 1.9 m in radius. Each module might and K-p separation. It is therefore worthwhile to contain « 50 wires ( ^1 cm spacing) for a total consider xenon for two reasons: I) it gives steeper of < 2000 wires/end cap. Each wire will be cap slopes in the relativistic rise region.*0 and, able" of a pulseheight read-out for several tracks 2) dE/dx is approximately 3 times larger than for within the STPC drift-time. Every fifth wire can argon, thereby gi' ing much better resolution in a be equipped to provide read-out along the wire, so given path length. I assume that 2 nt of xenon will that tracking and momentum information can be ob give at least the resolution obtainable in 6 m tained. In the central hexagonal region, through of Argon: 2.5* rms using the optimum sampling which the beam pipe passes, arrays of ball-wires thickness (3 cm Ar or 1 cm Xe). If this resolution can be placed to ensure 1001 pattern recognition can be achieved in practice, TT-K separation should capability for any multiplicity. n be > 99 0 efficient, K-p approximately 90$ (in the relativistic rise region). In the low p region, either time-of-flight or dE/dx can provide close to 100$ identification. Electron identification at all energies should be essentially 1002, due to the STPC to background from ~ 10-15 extra beam their extreme relativistic nature. crossings. How severe a problem this would be in practice is not safely predictable without exper ience at PEP itself. However, background studies A xenon + methane filled STPC of 4 m diam done at last year's PEP Summer Study have led to eter and 4 m length (2 m drift length) can be rather sanguine conclusions, especially that by readout again in several ways. The diffusion in suitable scraping and punping, particle losses can xenon-methane is unknown, but is probably larger be very suppressed near the intersection region. than in argon-methane. A rough estimate indicates that multiple scattering is still the dominant contribution at 8 Kgauss field for a 4 meter di As part of the program to evaluate the TPC ameter STPC. concept, plans are being laid to study ionization loss in xenon mixtures, in order to measure the slope of the relativistic rise, the intrinsic reso Readout at the end-caps in the STPC will lution, and the optimum sampling thickness. probably be made of modules of wires, in the plane of the end-caps arranged so that the drift time to V. CONCUJSIONS the wires, the pulse height on the wires, and the position along the wires (i.e., current division, The TPC concept remains a potentially induced signals, or delay lines) can be read out. superior approach for the detection of cnarged Approximately 4000 wires will suffice to read out particles at PEP. Much work remains to be done be the entire end-caps. The mechanical advantages of fore a critical comparison with drift chambers or this geometry compared with a drift chamber system streamer chambers, etc. can be made with clarity. of comparable volume are worth emphasizing. Fig. 12 We conclude by noting that in recent times, the shows a possible hexagonal module pattern for the physics now accessible to study, made possible by end-caps; squares are also possible. great strides in accelerator technology, has been seriously hampered by inadequate and/or very costly Cosmic ray fluxes will not contribute detector technology. The pursuit and development noticeably to the background in an STPC. However, of new ideas in this arena of physics deserves the maximum drift times of 20-30 microseconds open more attention than it regularly receives. I wish to acknowledge the numerous con 5. L. W. Cochran and D. W. Forester, Phys. Rev. tributions of Peter Robrish to this work. He has 126, 17S5 (1962). The accuracy of these re contributed not only his time, but many ideas as sults are regarded with considerable reser well. vation by Kuxley and Crompton in "The Diffusion and Drift of Electrons in Gases", John Wiley REFERENCES and Sons, 1974, p. 405. 6. See Ref. 3 in Reference 1 of present work. 1. D. R. Nygren, "The time-projection chamber- 7. W. N. English and G. C. Hanna, Can. J. Phys. a new 4n detector for charged particles." 31, 768 (1953). The reliability of these PEP Note 144, in 1974 PEP Summer Study, measurements is probably not better than 10%. (PEP 137). 8. Private communication, Dr. Klaus Halbach, 2. See Ref. 7 in above reference. LBL. 3. D. R. Nygren "Proposal to investigate the 9. M. Aderholz et al., N. I. M. 118_, 419 (1974). feasibility of a novel concept in particle 10. Somewhat contradictory evidence exists: see detection." LBL internal report, Feb. 1974. A Rousset et al. Nuovo Cimento 14, 365 (1959). 4. James H. Parker, Jr., and John J. Lowke, See also R. M. Sternheimer and S7 F. Peierls Phys. Rev. 181, 290,302 (1969). Phys. Rev. B3, 3681 (1971) for theory. COMPARISON OF TIME-PROJECTICfl AND DRIFT CHAMBER DETECTORS J.A.J. Matthews and A. Rothenberg Abstract DETECTOR DETECTOR- Time-projection and drift chamber detectors COIL ENDC A are suggested for the central region of a small- TRAcy \ , 7 radius, high-field solenoid spectrometer, A general comparison of these detectors is made, along with a quantitative study of their resolution. He find that at present time-projection chamber and drift chamber detectors give essentially equivalent performance. 1. INTRODUCTION Time-projection chambers (TPC) have been pro posed1 as a substitute Tor drift chambers (DC) for TPC 1 TPC 2 use as the central detector in solenoid spectro meters at PEP. Hith experimental data now avail able on the suppression of diffusion transverse Lvws^vv^w^^^^^^^ to the electric (and parallel magnetic) fleld^ a quantitative comparison of TPC's and DC's is now possible. A comparison of the TPC and DC resolution will be discussed first. In particular, we will con Fig, 1 Magnet with TPC's, sider the recognition of, and resolution'for, tracks originating within the detector, such as from neutral V's- This will be followed by a may substantially improve the resolution in the general discussion of pattern recognition, dE/dx future. measurements, solid angle coverage, etc. For comparison to the TFC, several DC systems will be considered: 2. DESCRIPTION OF TPC AND DC DETECTORS a) 10 chambers distributed uniformly in r, or The operation of a TPC requires a uniform b) 5 groups of 2 chambers distributed uni magnetic field of ~* SO kG. Such a magnetic field formly in r, as shown, for example, i,i Fie- '<• can be made using an aluminum stabilize!?, super Two levels of longitudinal resolution are also conductor with a continuous coil design, comple considered: delay lines parallel to sense wires, mented by appropriate steel end plates and associ ated flux returns. The necessity of a high field, plus the desire to limit the total extent of the TABLE 1: Chamber Parameters TPC along the beam, recommends the TPC for those experiments requiring a compact detector. There fore, for our comparison we take: a) B = 20 kC, and b) a detector sensitive region of fO cm radius (from the beam), and of length 2 m (see Fig, l). The resulting detector has a solid angle of -0.8 x kit with full field. We will assume, arbi trarily, that the first 20 cm (radically) are not available for chambers, being occupied by the bean; pipe, proportional wire chambers, trigger scintillators, voltage distribution for the TPC, etc. The TPC i3 divided into two 1 m modules to reduce the maximum drift distance. The spatial resolution of the TPC has been estimated using the experimentally measured transverse and longi ••(-•"•) tudinal diffusion for a 15 cm drift path.2 These values are scaled^ to the present geometry and -Or)' added in quadrature with a 200 u overall error (see Table l). He assume a TPC measurement for every cm of radial distance such as could be realized by a ball-wire detector1 at each end of the apparatus. Other envisioned schemes,** e.g. arrays of short tangential wirea at various radii, ~i 1 1 rrr\ r rfc I, —r vlg. 2 DC 5 x 2 showing decaying particle. with 6z ~ 1 cm, or cathode strips orthogonal to the sense wires, with fiz ~ 0.3 cm. The assumed final resolutions, wire spacings, and numbers of readouts are given in Table 1, 3- RESOLUTION Momentum and angular resolutions for the TPC and DC systems are calculated following Glucks- tern^ (see Appendix 1). The predicted resolutions in azimuth, momentum, and polar angle are shown in Figs. 3s> 3b, and 3c, respectively, as a function of the radial starting point of the track. Tracks coming from the primary vertex start at a radius of zero, those from neutral decays start at a radius greater than zero. With the measurement resolutions given in Table 1, we observe that the TPC and DC systems should measure momentum and azimuth equally well; the TPC provides a distinctly better determination of the polar angle. For general solenoid geomet ries, s direct quantitative comparison of the two systems is obtained by simply taking the ratio of thr, predicted resolutions in Appendix 1. For example, to obtain equal momentum resolution (see Appendix l); &P, 'TPC \pc TPC . = 1 0 10 20 30 40 50 60 70 TRACK RADIAL STARTING POINT (cm] -. €DCJ or the spatial resolution; 3 Resolutions for TPC and DC systems as a function of track radial starting point for (a) azimuth., (b) momentum, and (c) polar angle. * O.032 cm DC. Thus, the principal advantage of the TPC, in for 5° TPC measurements compared to the DC 5 X £ car example, is that for radii r J W cm the Du detector. 5 X 2 system has only two point-pairs left and a One of the advantages of the TPC is the ease track^ cannot be reconstructed. of detection of neutral V's due to the large The contributions of track momentum and polar number of points on each track. However, the angle uncertainties to two-body mass resolution resolution degrades rapidly with decreasing radial {at a mass of 2 GeV) are shown in Pig. k and dis- path length L, cnBBed in Appendix 1. Except for the DC's with delay lines; (bz ~ l cm), the momentum uncertainty is predicted to dominate the angular uncertainty at M12 ~ £ GeV, but only dominates the TPC angular uncertainty at H^g * 1 GeV (not shown). Thus the superior polar angular resolution of the TPC may (see Appendix l), with the result that the TPC yield somewhat better two-body mass resolution than the present DC detectors. has mainly a pattern recognition advantage over the —I 1 1— TABLE 2: Maximum Solid Angle Acceptance 50 cm Path Length 20 kG Field '•' •",••3 !••*• 2 4 6 8 10 "PI; Pf * I MOMENTUM (M,2) «SeV/c) Contributions to two-body mass resolution might be 10 ns minimum ccuiit and 13 us maximum due to track-momentum and polar-angle count, ouch a readmit requires a few more tits uncertainties at a mass of 2 GeV. than a DC readout, but a much lever clock rate. 5. conclusions '•*. QUALITATIVE DISCUSSION OF TPC AND DC Taking reasonable values for the TPC and DC a) Pattern recognition: Space points will resolution, we find that both TPC and DC detectors probably be essential for DC detectors and were give essentially equivalent performance. Drift assumed in estimating the number of eadouts, etc., chambers are therefore recommended for the near in Table 1. The large number of points/track fcr term on the basis of reducpd amounts of electronics the IPC's should be an advantage In high multipli and volume of data to be handled. Continuing city events, but may be degraded by the presence of development of the TFC {in particular, tests to old tracks due to the leng chamber memory time (see determine the spatial resolution for real tracks), Table l). of streamer chamber readouts, and t f large cylindri b) d£/dx: Possible for the TPC in the low cal systems, leave unresolved thf finfi choice cf momentum region, p 11 GeV/c (where TOF may provide a PEP central detector. ade^j late separation), but not in t\.n region of relacivistlc rise.6 REFERENCES c) Solid angle cov-, c-e: DCs and TPC's appear e^ual when resolution is also censidered (see Table 1. D. R. Nygren, PEP-14!*, Proceedings of the 197** 2 and examples cf V resol'i ion in Fig j). We note, PE; Uummer Study, p. ^'i. however, that increased spatial rescluticn for both 2. f. H. Nygren, Communications tc the 197^ PEP TPC and TC si'stetr.s only increases the solid angle Summer Study. coverage for the TPC's, since the Ws were geometri 3. Assuming tract's uniformly fill the detector cally limiting, Note that for II large, the abore parameters are »11 proportional to 1 ^l 1 -vrc . Thus, (or exaaplc, —•• = ^.vg In the aoleneidal geooetrj. IT'* Pj Ei,3'2 for the inclusion of nultiple scattrrlne see Gluekatern. The resulting tsio bMy nasa resolution for synnetric decays 'pl = ^S* ,nd "^ ~ ^ ~ °I i« appro* last ely: (i) P^ " P aln & {see FJB- I tor coordinate ayaten) (It) h 13 thu "projected track Ici^th", MC, the radial nparatlon of the ftral and. lest points or: the track. ^to*W'-(*)' (Ill) a, Bi are the measurement uncertainties In a: lmith and along ubere e,p la the laboratory angle betveen p. and p, HIE MARK II MAGNETIC DETECTOR FOR SPEAR Rudolf R. Larsen The Mark II Detector, presently in the de sign stage, is a SLAC/LBL detector project to re place the Mark ll»2 now in operation at SPEAR. While similar in concept to the Mark I it will have improved momentum resolution, shower detection, solid angle coverage for both triggering and tracking and a magnet design providing easier access to those particles transmitted through the aluminum coil. MOMENTUM RESOLUTION By using an array of 16 cylinders of drift chambers in the 5 kG field volume we expect to achieve a momentum resolution of op/p 0)=p (GeV/c) at 90°. This resolution will deteriorate- by about * 10 at small angles (~ 15 ). SHOWER DETECTION Design and prototyping of a system of liquid argon and Pb ionization chambers is proceeding. The goal is a special resolution of order 1 an and en ergy resolution of order 10? at 0.5 GeV/c. Two Fig. 1. Mark II Detector without shower counters. arrays are contemplated: one array of modules sur rounding the solenoid magnet coil (~ 1 r.l. of THE MAGNET aluminum) and another assembly with planes normal to the beam direction located at each end of the The magnet is a "conventional" solenoid with coil. Angular coverage extends over 2ir in azi field uniformity of a few percent at large polar muth and polar angles 2 15°. angles and, unlike the Mark I, the flux return iron will be located only on the top and bottom of the coil. This provides greater external access (after TRIGGER AND TRACKING SOLID ANGLE removing some shower detectors) to those particles transmitted thru the coil, and it should be a As in the Mark I, a cylindrical array of valuable asset for "outrigger" experiments. scintillators at a radius of - 1.5 m will provide time-of-flight information and, in conjunction with It is expected that the Mark II will be in the shower detectors and additional contemplated operation at SPEAR by Fall 1977. proportional chambers, will provide a triggering solid angle of -0.95MTT. The potential of trig gering on tracks reconstructed on-line from the drift chambers is being studied. The tracking solid 1. SIAC Proposal SP-1. angle will be essentially the same as the trig 2. J.-E. Augustin, et al., Phys. Rev. Letters 34, gering solid angle. 233 (1975). — DOTATION 0]: HIGH MCTMBTIUM PARTICLES WITH [DHNTII-ICATION OP THE FINAL STATE U. Becker, R. Cashmore, E. Groves, I, Keller, S. C. Morehouse, S. Poucher, and M. Strovink Abstract many of the charged final states that include high- momentum particles. Capabilities such as detection Two detectors tor PEP are described which will of neutrals and lepton identification would be measure and identify charged particles with good developed to the extent practical. momentum resolution in the range 0.2-15 GeV/c. Com plete identification is attempted for almost all RATES AND PHYSICS charged hadrons as well as for a significant frac tion of. elections, muons, and photons- In this way, Reconsideration of event rates and physics events involving high momentum particles can he goals in the light of developments since the 1974 studied nearly completely. Whereas both detectors PEP Summer Study has guided our approach to high- use solenovdal magnets with Cerenkov and time of momentum particle and final state detection. New flight counters and drift chambers, the more favor data from SPEAR and DORIS are in general agreement able design assumes light sensitive detectors op with the semi-empirical formula for inclusive rates erating within ;i magnetic field. A comparison to given last year by Richter,^ as is the limited p^ the detector designed by the 1974 group is made, model developed recently.3 Figure 1 demonstrates and ideas suggesting substitutions for part of this agreement of inclusive rates vs. x = 2p/Js as apparatus are discussed. predicted by the two models.4 These rates are low. l:or cxamplr, particles of any type with momenta ex INTRODUCTION ceeding 10 CJeV/c at PEP are as rare as = lnO/day. Particular types of particles such as antiprotons e+e* annihilation has offered many surprises could comprise <. 101 of these events, while the in 1974-75. In particular, many states with heavy range of extrapolation from present observations masses and thus highly energetic decay products have been observed, as well as indications of jet structure. Therefore, energetic secondaries de i i - - i - • i \ • T serve particular interest, as do the other particles which complete these final states. Apparatus for ' detection of high momentum particles and identifi cation of the final state at PEP may aim to provide various levels of information: ^^ 3D GcV - 20.0 f 1. Velocity and momentum analysis of charged V _ particles within a relatively small solid angle 1 adequate for measurement of inclusive single-parti 10.0 z Z cle yields. Part of the remaining solid angle may Nv be devoted to tracking without magnetic field. B Hin p p 4 Apparatus representative of these goals is in op v ^ - e '? eration at the 1SR, Sl'liAR, and DORIS. 5.0 - >^ ~ " \. - \v 2. Analysis of momenta and of highest veloc " ~ ities over a largo fraction of the solid angle, with measurement of lower velocities (ft < 0.998) " ^\ - over a smaller aperture adequate for inclusive yield measurements. Such an apparatus would det»>: - 1.0 ; mine inclusive particle spectra with additional E \, charged particle correlation information, but would not identify all the charged particles in the final 0.5 state. A detector of this type was studied in the 1974 analog1 to this report, Appendix A of this : paper describes the possibility of achieving some i i 1 1 1 1 similar capability by adding Cerenkov counters to the Mark [I SPEAR detector now under construction. li XH1.75I2-98M 1. do/dx vs x = 2p/i/s at ^s = 30 GeV according 3. Pull velocity and momentum an;'-.-sis of to tlie I iinitcd-transvcrse-momentum jet model charged particles ovi>r most of the solid angle. The (Refs. 3,4). Hie dashed line is the semi- prescnt report is fneussed upon this type of de empirical formula used bv the 1974 group tector, which wouhTcompletely analyze and idem i i'y (Kef. 2}. Tig. 2. Typical mulTihadron final states at *T= 30 GeV generated according to the limited-transverse-momentum jet model (Refs. 3,4). properly introduces overall rate uncertainty at the detector cells will be required to detect the nearly order-of-magnitude level. These event yields de- collinear parallel and antiparallel members of the •nand large solid angle for identification of charged jet. Examples of simulated jet structures4 at PEP particles over the full momentum range. As much as energies are found in Fig. 2; Fig. 3 is the simu possible "of the final state accompanying these rare lated distribution of opening angles between parti high-momentum particles should be identified and cles. We shall return to ihe study of cell granu momentum-analysed. The rates do not permit piecing larity in greater detail. together the needed information from fragmentary measurements of a large number of events, nor from measurements of correlation functions. PEP SIMULATION, JET 30CEV COS. OF ANCLE BETW. CH. PIONS Logarithmic extrapolation of existing e e" data «.X> = -3.6 7-10 "; 1 05287 ENTRIES suggests charged multiplicities of 7-8 at Ss = 30 GeV and neutral multiplicities of 4-7. The y multi plicity is assumed to be nearly twice the latter figure. Concurrent detection of neutrals clearly is necessary tor apparatus attempting full reconstruc- tion of a significant fraction of final states. ['articular emphasis should be placed on pre cision in high-momentum resolution for these studies. With event rates decreasing approximately as exp(-8x), accurate measurement of large x is essential. Further, uncertainty in high-momentum measurement will dominate the error in momentum balance of the final state. With a relatively good resolution of £p/p2» 0,005/(CeV/c), Ax would be •0.07 at x = 0.9, ^s = 30 GeV; the uncertainty in recoiling momentum would be S910 MeV/c. New evidence for jet structure in e e an nihilation at high s intensifies the interest in COSINE OF ANGLE OF PAIR high-momentum secondaries, which in the "sphericity" >'BL 7511-9312 analysis^ largely determine the jet a;is. If the jet structure and associated polarization phenomena Fig. 3. Distribution of opening angle cosines be- exhibit further s-dependence at PEP energies, the between final-state hadrans at *^s = 30 GeV comiwsition of the jet particles will be studied generated according to the 1imited-transverse- intensively. Clearly, fine granularity of individual momentum jet model (Refs. 3,4). As further motivation for study of a detector with the capabilities underlined above, one might imagine the role its analog would have played in the exploration of properties of the new particles5 at the existing e+e" storage rings, The best examples cannot be cited, since information from such a device to a large extent is not available now. Obviously, a much larger sample of U J(3095) would exist than has been detected6 in the HASP aperture. This example stresses the importance of precise charged particle momentum and Y conversion XBL 7S6-3200B point measurement over a large solid angle. The alternative decays Fig. 5. Side view of PEP'74 high-moment urn hadron detector. The right half is a section through one of the flux return legs. U.'(3(i84) — XC3410J (IT or K ?) + (71 or K ?) observed at SPHAR would be identified properly and an "open" cylindrical magnet with flux return vanes the parent mass determined independently. The ob and a nonuniform field reaching 8 kfiauss at the served6 scale-breaking excess flux of charged prongs interaction point. Two layers of Cerenkov counters produced hadron separation over a fraction of the at /s = 4.1 GeV and x = 0.4 would lie identified momentum range and 65$ of 4TT stcradians. No imion as to particle type and composition of the accom or high-energy electron identification was attempted. panying final state. Again, this is the more im Within itslimits, the design represents a well-con portant at high s, where the rate will be so low ceived approach to the experimental problem. that the fraction of recognizable events must be enhanced. Our attempts to modify that magnet design were aimed toward increasing the polar angular aper MAGNET STRUCTURE ture, improving the field uniformity in order to ease computation problems in reconstruction, and Initially we attempted to use as a starting raising the magnetic field in order to improve mo point the detection systan studied by tho high-mo mentum resolution. These efforts were frustrated hy mentum hadrons group* at the 1974 I'lil* Summer Study. the fact that both the pole tips and flux return Their detector, shewn in Figs. 4 and 5, employed vanes were nearly saturated in the existing design. Other, quite different magnet systems were con sidered. High field superconducting toroids without iron were rejected hecause of excessive forces on the coils." Two magnet designs were selected for further study: Larfte volume low magnetic field. Because Ap/p- « Be^, a solenoid of radius T~= 3m can pro vide good momentum resolution with a modest central field of B - 3kGauss. A large (600 ton) magnet is implied, with very small forces on the coil. The photosensitive devices implied by a particle identi fication system would need to be located within the magnetic field. These could be the micTOchannel photomultiplicrs described in the following report." "Conventional" solenoid, flic lumped51 super- conducting coil would provide 15 kGauss central field within a radius of 1 m. All phototubes would XBL 75h-3200A be located in low field regions outside the coil. Fig. 4. Iind view of PEP'74 high-moment inn fmdron de In addition, an aluminum-coil solenoid like tector. The Ccrenkov counters labelled C[, tjie "existing" Mark 11 magnet amid he equipped with 0?, C, have refractive indices similar to Cerenkov counters over a small fraction of the solid those described in Fig. 8. angle. This possibility is explored in Appendix A. Figure 6 contains sketches of the two magnet the events will suffer at least one ambiguity. FJC- configurations with a comparison to the scale of amples of such systems are Cerenkov and tune-of- the magnet proposed by the 1974 group. If the re fllght (TOF) counters, and wire chambers, which turn yoke is included, their radii are similar. The tend to havi.' fewest elements close to the beam. The 1975 magnets are longer along the beam direction, probability that the entire detector will unambig reflecting a desire to study for example beam polar uously detect an event is the product of probabili ization effects over a Larger range of polar angle. ties like the above for each independent subsystem Economics could force a retrenchment in this atti of elements. Fcr this reason it may be advantageous tude. to arrange layers oT for example Cerenkov cells to subtend the same elements of solid angle (ignoring curvature In the trajectories). The problem is exacerbated by upward fluctua tions in event multiplicity; additional track." from •y conversions* 6 rays and strung interactions in detector material; and correlations in particle di rections introduced by energy-momentum conservation and by dynamics such as observed at SPIiAR and de 3 scribed hy the limited- pA jet model mentioned earlier. For example, for 12 tracks the 100-cleiiient detector yields at least one ambiguity 501 of the time. Figure 7 displays these probabilities for variousmultiplicittcs and cell granularities. ED Earlier in Pig. 3 a jet-model simulation of the distribution of particle opening angles at PEP 75 >-s = 15 CeV was shown. It implies a concentration Conventional o of particle density in the jet direction which is s 4 times that expected without correlations be tween particles. Effectively, the number of detec tor cells is reduced by a similar factor! Tending detailed Monte Carlo st'u!\, we have construiiied layers of proposed detection system to possess >. ' 100 cells. Coarser system.; surely will not work; liner systems probable will be necessary. E^..^ * 1V 1.0 • N ' 5 Xiil. 75 I2-9B04 - 9 0.8 Fig. 6. Comparison of the siies and shapes of the soIenoid.il magnets (a) and (b) studied in • 12 tfii:i report, and (c) the magnet envisaged by the 1974 group. 0.6 = IB 0.4 IiLTliCi ION SYSTOfi: GENERAL CONSIDERATIONS BcFore considering details of the detectors associated with each magnet configuration, certain 0.2 common features are more briefly described below. Cell Granularity on \ i i r • i i i i i i ->- 30 50 100 The probability that no two charged particles of multiplicity N = 7 are counted in K = 100 un No. of devices iformly illuminated cells is XBl. 7512-9865 N-l • Fig. 7. The probability of no double hits, given N uncorrelated tracks and 30-200 devices (cells). That is, even in systems with 100 elements, 19% of Velocity Measurement ment of the latter figure requires use of sptoal phuiwC.tthodo and phototube window nuileri;:!, or For complete ir\p separatum in tin* range he- wavelength shifter,1'1 or noth. I'hotutube and col low 15 (.IcV/c, five types of velocity iiiivisuromeilt lector sir.L'S were chosen considering I>otli the size are needed (.Fig. 8): of the cWonkov ring and the track curvatuivs at lowest relevant n* niomunta,'s Most phototubes a) Time-of-flight oi'er 3-in flight juth witli illustrated in the figures have a photoc;tt!u*h> o = 0.22 :isec.i° Maintaining this timing accuracy area roughly the same as that of a standard 5i- inch over huge hodoscopes ;it this radius hutild provide tube. In these figures and in Fig. R, countors lb), * 3a -iK separation below 1 3 (JfV/i: and Kp separ (c), (d) , and (e) have been dubbed Q, C~, i'.*, and ation below 2.1 GeV/c. _, Cj respectively. hj Aerogel threshold Ceicukov counter with re fractive index n <. 1.04. This material is in tlic development staye.il Aerogel with u = l.flti has We would be happy to replace this array of been produced and iiiciices down to 1.025 are thought apparatus with simpler means of velocity measure to be possible. The latter index would he ideal, ment. Appendix li is an initial attempt to explore the use of microchannel - plate phototubes to re providing nK separation between D.b and 2,.2 GcV/c cord Ccrenkov ring radii. Unforiunately, in the and Kp separation between 1.2 and 4.2 GcV/c. In 2 the large volume low 15 detector, - 0(10 cells roughly example chosen, 25 m of 1 cm- photosensitive ele 15 mi on a side would be necded--an enormous ments are required. We did not seriously consider quantity by present standards, (n the conventional relativistic-rise dE/dx velocity measurement, as solenoid at larger radius* at least five times the necessary sampling lengths seem excessive. more would be required! liven in the more favorable Puls5-hei(>ht analysis on Cerenkor counters with case, *he use of aerogel on this scale is regarded large light levels conceivably could provide only as a possibility. , additional information sufficient to eliminate one layer of Cerenkov counters. c) Isobutane threshold Cerenkov counter, n = 1.006 at 4.1 atmospheres, providing nK(Kp) separation between 1.3 and 4.5 (4.5 and 8.5) CeV/c. Triggering Six atmospheres of propane could substitute for the isobutauc, at the expense of thicker vessel walls, We approach the intractable problem of de Frcon is not_ an acceptable substitute in this case signing a trigger with sufficiently low background of wide angle optics, because it scintillates ex rate by assuming that the FJiP vacuum is as good cessively ( - 1-7 plmtons/cm).12 as that at SPEAR and that beam-gas scattt ing con d) Isobutane or pentanc threshold Cerenkov tinues to contribute 5-10e of the multi-lm. J11 ic counter near atmospheric pressure with n = 1.0015 e-vent rate. In this case an adequate and flexible or 1.0017, providing nK (Kp) separation between trigger scheme would make use of 2-3 coaxial cy 2.4 and 8.5 (8.5 and K..1J CcV/c ifor n - 1.0017). lindrical proportional counters (or other fast e) CO? threshold Lerenkov counter at one at high-resolution devices) close to the beam pipe*. mosphere (ft = i.00045), providing TTK separation Azimuthal coincidences of hits at different radii above 4.7 GcV/c and en separation below 4,7 in these counters, in time with the beam crossing GeV/c. and TOF* counters, would form a fast bias-free - 1 kHz pretrigger. Flexibility in setting ma jority requirements on the number of aiimuthal co The radiator lengths wer? chosen to give com incidences would be retained. Seme cosmic ray con fortable light levels except in :he case of the figurations would fail to satisfy the TOF coin- COT counters where a minimum of Np ; 5 photoelec- c idence. trons would be produced 'f N. = ICQ 1.6*. Achieve . 1 1 1—W-l ) 1 f- Rtdutw TOF * lenatti lor Nvne 13 m) J P *it - 100 e i - 5 !.1(tn c* n - 1.036 4.2 cm C3 n-1.006 , I.OGdii 17 cm CO? c, _ III 1 1—*-J > " .« .5 .« .7 1.0 3 4 5 6 7 XBL 7*11-98*6 Fig. 8. Sysnti of 5 counters for particle identification (described in the text). The time-of-flight (TOF) solid lines labelled n.K, and p indicate the momentum ranges over which err, TTK, and Kp respec tively are separated by less than 3a. The threshold Cerenkov (^1>Q,03,1:4) solid lines indicate the momenta above threshold. Fig. 9. Side view of one quadrant of the large volume low magnetic field detector. After the prctrigger, a scheme similar to remaining 50 cm are used for position and energy those in use at PLUTO anil envisioned for the SPEAR measurement of showers (discussed brlow). The 50 Mark II detector would perform fast track recon cm thick endcaps leave a 10° cone for electron struction within -100 ijsec. Only drift chamber tagging of the YY process. A very homogeneous wire occupancies (rather than digitized coordinates) s^.cnoidal field is expected. Total wight of the would be used. Control of momentum cutoffs, multi iron is I'OO metric tons. plicity, and topology would remain flexible. The two-level trigger is intended to provide a logging rate of - I ilz with < 10°0 dead time. Flux return LARA: VOLUME I-OK MAGNETIC FIELD DETF.CirjR Idles (1RU and ilrifl dumber gas at STP The 3 kGauss solcnoidal magnet, with a coil of radius 3 m and length 8 m, already has been mentioned and its outline shown in Fig. 6(a). Figures 9 and 10 depict portions of its side and nC0 end sections in greater detail, along with the de 2 tector array. The allocated coil space is sufficient either for a superconducting coil or for a simple Drift (3 lav"*) aluminun winding. Th? superconducting approach is C2 1 atm pentane made attractive by the low field, minimal coil forces, and Low stored energy (~ 7 MJ), but not Drill (5 layers) by the large cryostct. A scaled-up "MINIMAL" de- Aerogel C4 0 signl- seems too expensive. The coil area is isobutane x 4.3 that of the Mark II magnet and the field Oiili 16 or TPCt 58% as large. .iith half the Mark II current density a conventional Al coil would draw 3 MW and still Trigger 13 Overs) present less than one radiation length. The magnet flux is returned on eight sides of an octagon by laminations of 1 radiation length steel plate, with total steel thickness 30 cm and Fig. 10. End view of one octant of the large total lamination thickness 80 cm (Fig. 10). The volume low magnetic field detector. {'article Ident if iv;it ton The a1iii">plk-i pressure counters Cy and C? are interleaved wit h drift chamber layers outside The prime advantage of this detector is that the aerogel {figs, '.' and 10). The integral tracking the Cerenkov com.lets C- and C. (see Fig. K), wlmh and momentum analv is is intended to discriminate have highest index and large cost per unit volume, between original •-« •eondaries and daughters arising may be kept manageable in size by placement near from knock-en and •iron• g interactions in C3 and C4. the beam crossing ixiint. This choice combined with i!, subtends H-l\, ol' •\v t while Cj is extended to the low magnetic ileJd minimizes the effect of i'R'i of 4TI because ol" the interest in clean identi- track curvature on their optics. The counters lie I'icatjon of singji-electrons with x £ 0.3. Afore in the region 50 cm • radius < 95 cm and subtend the energetic electron?, will be identified by the range 32" < 0 < 148" (H4* of 4ir). Their location is slmwcr counters use I for detection of neutrals shown in Fig. !' (sideview of one quadrant) and (described he low). Fig. 10 (endview of one octant), with optical de tails in Fig. 11. Using overpressure safety factors Muons are identified with 855 confidence over in f —ess of 3, their material(mainly AB amounts 73; of 4;T hy comparing the coordinates of charged to 0.24 radiation lengths. Despite the relative trucks outside 30 em of steel with extrapolations compactness of the >ierogel counter, it still re of their trajectories within the solenoid.*7 This quires 608 elements roughly IS an on a side. This confidence level is not badly matched to the levels volume of aerogel may exceed projected production of decay background; for example:, the 5-GeV K^ capabilities, leading to a large reduction of the decay probability is 0.06 at 6 =• 90° . "u2 solid angle of the counter. In this case a hole in Kp separation between 2.1 and 4.5 GeV/c may be Sensing Cercnkov Light within the Magnetic F-* J created in the otherwise full momentui) range for hadron identification. liach of the Cerenkov counters mentioned above employs single reflections from aluninized surfaces The time-of-flight scintillation counters onto photomultipliers located within an ambient subtend 82° of 4TI steradians just inside the coil. 3 kGauss magnetic field. The necessary photocathodc The length of this system (8.S m) probably makes it area of each indicated light-sensitive device is necessary to split the coil about the plane z = 0 roughly that of a 5-inch phototube. Obviously, (not shown in Fig. 9) for phototube readout at magnetic shielding of conventional ph^jmuitipliers z = 0 in addition to z = ± 4.25 m. The number of would be too unwieldy. TOF elements would exceed 10U and the number of phototubes (conventional or microchannel) would The assumption sustaining our interest in exceed 200. These counters are inaccessible to this detector is that, by 1980, microchannel-plate particles with pj_ < 135 MeV/c. Possibly, addi photomultipl icrs {"IJCIW'S") with sufficient in- tional Kp separation at lower momenta could be ob sensitivity to magnetic fields will exist in ade tained with TOF information from drift chambers at quate quantities ( - 400 excluding aerogel smaller radius. counters). A brief introduction to pCPM's is .Mil 7iI2 ~ \^ Outer 2/3-"—-—*• Full detector " Certain early photomultiplier design*; used - magnetic fields as a necessary part of ty focusing structure. Although requiring specific i -tat ion 1 with respect to the field, such devices • :'. be -*- worth reincarnating for our purposes. An alternative to light detection within the field is to pipe the light out. Systems XBL75I2-9B70 of light pipes opening into the Cerenkov gas volumes can be engineered to require few bounces Fig. 12. Momentum resolution of large volume low and obstruct a minor fraction of the drift chamber magnetic field detector (a) vs. momentum p, tracking area. Unfortunately, even quite an flab- (b) vs. polar angle 0, assuming sagitta orate framework of light pipes wouiu service only resolution cf 200 urn. Below 10 GeV/c only two layers of Cerenkov counter, which is not an the outer 2/5 of the detector is used in advance over the 1974 detector.1 order to minimize Coulomb scattering effects. Tracking Charged particle trajectories are detected Although we have described Dj as a set of by seven sets of drift chambers, detailed in conventional drift chambers, the time-projection Table 1(a) and Fig. 9. These are divided into two chamber described by Nygren17 is a candidate for main groups. In the cone 32°< 6 < 148°, cylindrical use in this region. The reason is that information chamber sets Di-D* w'th small-angle stereo in se of sufficient density for local track reconstruction quence +5°,-5,0° are planned. Cell sizes of 4 cm is sought in an annular region only 30 cm thick. are used except in the innermost set of chambers, The location, centered on a point of high symmetry where 2 cm cells may give fine enough granularity. in the solenoid, would provide exceptional magnetic A particle passing through D1-D4 intersects a total field uniformity1** since the time-projection chamber of 18 layers of drift chamber. For momenta less than would occupy only a tiny fraction of the (already than 10 GeV/c, momentum measurement will use D£, uniform) field volume. Dj, and D4 only, in order to avoid the effects of Coulomb scattering in C3 and C*. The chamber set The regions 10° < 8 < 32" and 148° < 9 < 170° D] provides track direction information at radii are serviced by Di and by drift chamber sets Ds,D6 inside C3 and C4 for interpretation of these v and D7. Each of the latter sets is a -60", 0°, 60 counters, and for improved momentum resolution configuration for highest information content of anchors the trajectories of particles with momenta each individual plane. The inner 10° cones are ten exceeding 10 GeV/c. Assuning (conservatively) tatively left open for access and to prevent beam- 200 yra resolution in sagitta of tracks projected gas background at low 9 from creating too many upon the z = 0 plane, the momentum resolution, as extra wire hits. shown in Fig. 12, is better than 6$ at p = 15 CeV/c, 6 = 90°, and improves further in the range 20° < 9 < 90°. The total solit' angle for charged particle track reconstruction is 983 of 4TT. -147- T.abl e 1. Dr ift Chamber Wires a (a) Urge Volunc Low-S Cell r Length Layers size(aa) Angles Wires Total (cm) •5° 2 D 32.5112.5 111) e S3" 102 612 l 0° 1? D 120110 4 188 2 505 6 0° 1128 I? D 180110 650 4 283 3 3 0° 849 15° D 840 4 442 4 282±10 3 0° 1326 130* D5 180 6 4 45 270 130" D 350 4 87 348 6 4 90" 130° D 550 4 136 408 7 ^ 90° 4.9 K (b) Conventional 2018 100 4 1 15* 63 252 °1 3012 150 1 2 Del ay line 95x3 288 15° D, 55S10 275 S 2 0" 141 707 410 4 3 15° 200 h 9515 0° 15° D 3 3 260 780 4 12514 ilO 0° 0° D 30018 600 0 10 60° 24D 1920 5 »210 120° 4.8 K aDoes not include shower Jetector. -148- Detection of Neutrals for typical conversions in the coil, and ±0.5(1 + (O.S/E)2}'* mrad The one-radiation-length ("lX^") steel Flux return plates, subtending 73t of 4* steradians, for typical conversions in the iron. ideally would be sampled by a liquid argon system (of enormous area). Since the choice of sampling Electromagnetic shower energy resolution suf device is governed by cost analyses which at this fers contributions both from the shower statistics stage are not complete, we instead shall describe an and from delta rays in the gas.20 For the case of example of what may represent a manageable system. IXQ sampling in iron with 2.5 cm of argon at STP, It has the advantage of emphasizing precision in these contributions to o(E)/l; are estimated to be direction of the neutral particle, which to a cer 0.22M: andO.&Minns, respectively. Thus the an tain extent has been neglected even by proposed ticipated y energy resolution is o(E)/E* 0.45/M;, detection systems tliat concentrate on neutrals. For or z 6 times worse than may be achievable with this purpose the location of these counters at 3 liquid argon and thin Pb plates. meter radius is a distinct advantage, not only be cause of the large lever arm ff* direction measure Above 3 GeV, this y energy resolution com ment but also because those Y'S which are tightly bined with the high angular precision will achieve collimatcd have an opportunity to spread out and >3o separation of ir° from n°. Once identified, the produce separate showers. Y pair's energy resolution can improve dramatic ally if opening angle information is used. Let M The chosen sampling medium is drift chamber and E be the u" or n° mass and energy, and k, gas [mainly argon) at STP. Thicknesses of 5 cm fol and k? the Y energies. Then low each of the first four 1X0 radiators (the coil counts as the first radiator). Thicknesses of 2.5 cm follow the next ten IXQ slabs, and alternate E = (M/6) (1 + kj/k,)'* (I + kj/kj)1*; IXQ slabs in the final four radiation lengths. Each 2.5 cm of gas is contained within one drift chamber, with cell spacing designed to minimize cost, con k 2 sistent with granularity requirements. Jfost drift £ss> o2m+( l ^W^)_ a tk2A wires run in the "z direction, except for the outer E 2 2 2 most chamber and half of the eight chambers within e WE A k "Ify the first four radiation lengths, which have wires in a rectilinear direction approximating $. A 3-GeV TTU or n with k. = k, has better than 1% The innermost pair of chambers is azimuthally seg energy resolution, [f k. * 2k."the resolution rises mented to follow the coil shape. The innermost eight to 185. " chambers have wires staggered and at slight stereo angles in order to resolve ambiguities, and are Events with up to four Y'S that cannot be read out individually with electronics giving 1.5 identified as ?ru or n(1 may still be fully recon mm spatial resolution, while preserving pulse height structed, provided that the y's fall within the information. Corresponding wires of the next five 755 of 4TT devoted to neutrals detection and the and five chambers, respectively, are connected to charged particles lie within the 98 o of 4TT which common electronics channels. The outermost two cham is instrumented for charged particle reconstruction. bers are read out individually in order to identify Since the Y energies arc not used in the fit, the muons. The intention is that the first 4 Xp portion number of constraints drops from 4 to 0 as the of the counter detects conversion electrons with number of Y'S rises from 0 to 4. Even the OC events 1.5-mm resolution both in Z and r^ ; the first 14XQ are subject to the condition that the fitted portion measures electromagnetic shower energies energies be consistent with their independently with lXg sampling; and the entire 18X.. (2L k ) a s measured values. If IT" or ri" are presumed to be counter observes hadron showers with^2XQ sampling the Y"ray parents,-1 one niiiy demand with good am! identifies muons. accuracy that two independent t pairs each match one of these parent masses. Each of eight octants by this prescription would contain 20 drift chambers of a size exceeding Separation of e" and «" is performed by the 2x8 m, with electronics for the equivalent of 12 shower counter only above 4. ftCcV/c , because of the chambers. We imagine that the iron laminations existence of the CO2 Cerenkov counter (4. For IT* themselves would supply most of the frame stiffness that feed <, 1/3 of their energy into electromagnetic and gas enclosure for these chambers. Nevertheless, showers, the separation of e- from -n- will exceed without considerable design innovation their con 4o. struction cost would be staggering. Total drift chamber electronics channels would not exceed the At energies below 15 GeV, the 30 an of Fe will = 51)00 allocated to the inner detectors. Their Low contain > 805 of the hadron shower energy.22 level of required performance needs to be exploited Crudely taking into ncccint cont;uninmcnt effects fully in order to keep the cost within reason. and 6-ray confusion in the drift chamber gas, we compute an abysmal hadron energy resolution The conversion point resolution is limited a(E)/E : 2/y^(GeV). Neutron and K'P directions are by Coulomb scattering of the produced pair within measured by the muon identifier chambers. This the radiator at low energies. As a function of en class of particles can be identified with low ergy E{GeV), at 9 - 90° the projected y direction (- 505) efficiency and high confidence if they first is known to interact in the outer half of the shower counter. We note that 30 cm of steel is not a bad K£ -*- Kp. ±0.5(1 + (3/E)2)^mrad regenerator. -I4y- DETCCrOR USING "OONVEW1IONAL" SOJJWOIU to the relatively small coil obstruction of 101. M*o the cryostat is manageable. The field is re The side view of one hemisphere of this de turned through 6 legs of 3.4m inner radius which tector Ls shown in Fig. 13 and its end view in Fig occupy 1/3 of the azimuth. By these demands, the 14. The right-hand quadrant in Fig. 13 shows a sec magnet turned out to be similar to the General tion between the laminated flux return legs, while User Magnet. For details of a typical design, the the left-hand quadrant shows a section within the reader is referred to their report." legs. Magnet Particle Identification The magnet was chosen to have IS kGauss over The detector uses a time-of-flight system lm radius for sufficient momentum resolution. A that is similar to that previously described, ex superconducting coil was selected for power econ cept that the counters at 3 m are inaccessible to omy, and preference given to a lumped coil to keep particles with momenta s 225 MeV/c. Because of Y conversion and hadron interactions localized the very large volume of radiator that would be End pin* ~\*^r~--\ 1 1 1 H fml 0 1 2 3 4 4S5 XBL 7512-9871 fig. 13. Side view of "convcutional" superconducting solenoid detector. l:ig. 14. lind view of "conventional" superconducting solenoid detector. -tso- required, an aerogel counter covering an appreci possesses an extra layer of delay-line drift cham able fraction of the solid angle is not envisaged. ber cells for improved pattern recognition. The The gas Cerenkov counters Cj» C2, and C3 (Fig. 8} chambers in DV are almost square in shape and perform the same particle separation functions with equipped with wires at +60°, 0°, -6H".These the same refractive indices described earlier. The provide additional tracking precision as well as high-pressure isobutane counter (C3) lies inner identifying Y conversions in the Pb. The total most, beginning at a radius of 1.4 m. It and the number of drift chamber wires is 4900. surrounding atmospheric-pressure C0,» counter (Cj) subtend 821 of In steradians. These counters per form TTK separation above 1.3 GeV/e, and Kp COMPARISON OF DETECTORS separation between 4.S and 8.5 GeV/c. Kp separation above 8.5 GeV/c is provided over 2/5 of the azi Size and Equipment muth by the atmuspheric-pressure counter C2- An immediate comparison of the size of the two magnets discussed above with their 1974 counter Technically, the counter C3, which produces plenty of light, employs double-reflecting optics, part^- lias been displayed in Fig. 6. Greater detail consisting of outer and inner spherical mirrors is provided in Table 2. Although the magnets have with respective r;idii of curvature 1.8 and 0.6 m. grown by - 2S% radially and up to 501 longitudinally The momentum dispersion created by the magnet since 1974, they still are accommodated by the 4-m field integral uf 450 MeV/c makes it necessary to pit floor clearances anticipated for fi^e experi use two 5-in. tubes per cell in order to make the mental areas. More than half of the interaction light collection efficiency momentum-independent. area length (20m) will be available for compensating Inexpensive tubes like KCA4525 may be used,24 coils, electron tagging, etc. Magnet tonnages also since tuning and rate are no problem. The photo have grown by 50-1501. However, most of the tubes and mirrors will be enclosed by a cylindri ton weight of tiie largest magnet is composed of cal pressure vessel of 0.7 cm Al wall thickness.25 plates that also serve as media for shower develop ment. Photons will not trigger the TOT scintilla tion counters and thereby can be distinguished The 1975 detectors each substitute z 3000 from charged particles. The drift chamber outside drift chamber channels for = 3000 MNPC wires plus the IXQ Pb absorber will determine coordinates of 42000 spark-chamber capacitive-readout channels a charged particle or the vertex of a y conversion used in the 1974 design. The large volume low B with 50* efficiency. The Pb absorber will produce detector's shower counters would require up to S00O new 5 rays while absorbing those produced at additional low-resolution drift chamber channels smaller radii. By constraining charged particle with pulse height capability, while the "conven trajectories, this chamber will recognize interac tional" detector would add 5000 proportional tubes. tions in the walls of C3 and especially in the Total scintillator area has dropped by 20°. The lumped coils, which is a problem specific to this number of Cerenkov layers has risen from 2 to 3 design. (or 4 with the low £ aerogel option), while the number of Lcrenkov cells has trebled to - 400 (excluding aerogel). Most of the latter increase Muons are identified with 971 confidence over is needed to solve ambiguity problems, rather than 271 of 4TT by drift ch.ijiibers outside the flux re to service the extra layer. One measure of dif turn legs {not shown). HLectrons below 4,7 GcV/c ficulty in constructing the explosive 4.1-atmo are identified by C-. over 821 of 4TI. Higher-en sphere Ccrenkov counters is the total area of the ergy electrons arc identified over 311 of 4r by innermost pressure windows. This is 24 in- for the 5000 crossed proportion:!I tubes interleaved with 1974 detector, 48 m2 for the 1975 "conventional" 2 iXn iron laminations in the flax return legs and detector, and 6 m for the 1975 low & apparatus. shower counters close to the pole faces, (Fig. 13). These laminated detectors identify neutrals in a manner similar to that discussed previously for Performance Summary the large volume low 15 detector. Table 3 summaries the relative performance Tracking of the 1974 and two 1975 detectors. Listed are solid angles for tracking, various ranges of TTK Five sets of drift chambers D.-DV record the and Kp separation, and electron, muon, and neutrals trajectories or charged particles. D^-D- subtend detection. The 1975 detectors have only 10-301 of the 1974 tracking inefficiency. The high-momentum 901 of 4rr. Using the partial information in Di-D2 gives less precise momentum reconstruction over TTK separation (C^) extend over 981 or 821 of 4TT, as 961 of 4TT. The momentum resolution using the sets opposed to 48% in 1974. Intermediate momentum irK D D and Kp separation (C ) subtends 821 for the low I 1~ 4 (subtending 821 of 4n) has the high-energy 2 asymptote detector and = 601 for the others. Lower-momentum separation (C,) has the largest gain: > 8I[1 of 4tr, as opposed to 161 in 1974. The low B detec Ap/p" = 0.0055/GeV tor retains the possibility of an aerogel Cerenkov layer (C4, 821 of 4TT) which overlaps the TOP separ at 9 = 90° , again with the assumption of 200 urn ation region. The TOT systems in all three detec sagitta resolution. tors cover similar solid angles. The details of these drift chambers are found in Table l(bl. H1.-D4 arc sets of drift chamber Identification of muons and high-energy elec layers in sequence +5", -5", 0° , except that D* trons, capabilities absent from the 1974 detector, are performed over ~ 75 or - 30"*, of Hv . Hie con Simpler Ccrenfcov optics improve counter fidence levels for muon identification range from performance and reduce cost, 'I'lie set of Cerenkov 80S to 90°«. Photons are detected with excellent counters is more complete. accuracy in direction over 75" or 30% of iv , and with modest energy resolution a(V.)/i^ = 0.-15 Location of high pressure counter at smaller GcV'i. Some K£ and neutron identification is pos radius decisively improves the design ol the large sible. Capability for neutrals detection again was volume low IS detector. Its high pressure isohutane absent from the 1973 detector. counter has a o-m2 inner-pressure window subtending 84% of 4TT. The corresponding 1974 counter had a Summary of New Design Features 24-m- inner window subtending lt°„ of 4TI. Essential new features of the 1975 detectors [TILL IDE.WIFICATION 01= llffi FINAL STATE ? Homogeneous magnetic fields tremennously With the detectors specified, the report con simpli fy the tracking software. cludes with a brief estimate of their ability to achieve the original goal--full identification of e,YiM, and neutral hadron detection capabil a non-negligible fraction of the final states ities are added with lamination of the iron return accompanying the high-momentum particles. We yoke i consider an event as Js = 30 GeV with 6 velopn charged particles and 4 TTO'S. LOSS or ambiguity in particle identification may occur because of: Full use of drift chambers reduces chamber electronics requirements while make possible Finite solid angle for charged particle "slow" trigger decisions. tracking and ferenkov aperture. For each momentum, Table Z. I'etector Comparisons: Scale PEP'75 (Large volume PEP'75 low « field ] [Conventional] Outside dia * length (m) 6.5x6.74 7.8x9 Magnet Peak field (klauss) 8 3 IS Yoke weight (tons) 230 600 330 Coil weight (tons) 5 23 Cerenkov detectors-uo. of cells a C02 (1 atm) 72 176 Isohutanc or pentane (1 atm) 32 144 112" High pressure (n = 1.006) 96b Total 5 in photomultipliers Total pressui 24 6 Drift wires 1960 4900 + shower MWPC wires 3500 -500 Spark chamber channels 42,000 - Scintillator area (m2) 210 170 160 Proportional tubes 5000 .cells, OPTION: Aerogel*C (n = 1.025) "POT ' Number of phototubes may be larger. 192 phototubes. CA11 phototubes approximately equivalent in area to 5 in dia. TSxcludes time-of-flight counter phototubes (approx. 200 in. each case), ^licrochannel-plate photomultipliers Inside window only. Table 3. Detector Comparisons:Performance PEP175 I Large volume PEP'75 low B field] [Conventional] I. &S1/4TT Full tracking 0.90 (partial tracking) (0.96) II.Afl/4ir for particle identification (a) C,: ffK separation (p>4.6 GeV/c) 0.48 0.98 0.82 TlK fbl C - separation (2.4 ir\ r .riTK separation (1.3 f0.84)a (d) C4: Kp separation (2.2 rTiK separation (p III. Electron identification - An/4Ti (a) Cj JK4.6 GeV/c 0.98 0.82 (b) Shower counter p>4.6 GeV/c 0.75 0.31 IV. Muon identification (a) an/4Tr 0.73 0.28 (b) Punch-thru + IT prob. 8 5 GeV/c - ;0.15 += 0.06 :0.03 •= 0.06 V. Photon measurement (a) &!/4» 0.75 0.31 (b) iE/E^CGeV15) 0.S 0.5 (c) (ae or Lf)y 1 mrad 1 mrad VI. £ and neutron identification M1/4TT = 0.75x=0.25 efficiency t£/E* 2 tkf/1 aAerogel costs may be unrealistic, typical; see text. -153- two part. ir Cerenkov layers must he crossed. could be reconstructed unambiguously. In both de The tracking probiihil il>' for the (aj low B or (b) tectors, (good 01: + pscudoconstraint) fits can be conventional deleftor is 0.98 or (partial tracking) made to ( <. 4> + charged) final states. The strong 0.90 respective J/. I lit,' Ccrenkov solid angles are dependence of recoastruction efficiency on y near 0.83 of 4;r sicrudjans except for Kp separation multiplicity means that most reconstructed events in the conventional I detector at u > 8.5 UeV/c. Jet will be drawn from downward fluctuations in neu structure which correlates the particles may lead tral multiplicity rather than fluctuations within to -2X reduction in the effective number of inde a fixed multiplicity as in our example. pendent isotropic struck directions, except for fine- structure in the Cevenkov acceptance due to 10* Results. Hie numbers quoted above are meant coil lumps in the conventional detector. These as guidlines rather than as estimates appropriate effects give efficiencies to all possible mtiltiparticle configurations. The products of these crudely estimated efficiencies for reconstructing and identifying this o-particlc c. = (0.«8J3* (0.83) = 0-54 charged final state are 0.15 and 0.10, respectively, for the large volume low B and conventional detec 3 3 6 tors. Detection (not reconstruction) of 8 y's eb = (0.96) * (O.S5/0.90) x (0.90) = 0.39, plunges these efficiencies below 104 and O.IS, re spectively. It is probably safe to conclude that Multiple hits in detector cells. We assume full reconstruction and identification of "typical" that drift cnamher redundancy and staggered cell charged plus neutral final states at /s = 30 GeV azimuth confine this problem to two Cerenkov will not be made with more than * 10% efficiency. counters whose cells occupy the same solid angle Of course, the events containing one or two am element. Six particles plus two converted gamma biguously identified particles are still rich in rays share only SO cells due to jet effects. The correlation information; many complicated multi- efficiencies are particle states of known baryon number, strangeness, and invariant mass will become available for study. c»- %•«••*• AOfNQKLEDGBflOTS Disabling of Cerenkov cells by interactions. Delta-ray background is worst in the high pressure It is a pleasure to acknowledge the contri counter C3 at the level of z Z\. Its walls will butions of D. Coyne, R. C. Field, M. I. Green, create 5 11 6-ray background in C2. Inelastic G. Hanson, K. Halbach, P. LeComte, P. Oddone, strong interactions occur at the zZ% level in these D. Nygren, V. Perez-Mendez, and M. L. Stevenson to walls. We have already discounted the effects of this report. coil lumps. The efficiencies are e = £ = (0.95)6 = 0.74 a b Appendix A Decays. K decay is dominant. Perhaps two of INSTRUMENTING THE MARK II DETECTOR WITH the particles' are K's with /p) = 2 GeV/c and decay HIGH-MOMENTUM PARTICLE IDENTIFICATION probability (out to 2m) = 15%. Hence GENERAL £ = e. = 0.72. a b The Mark II, proposed for general purpose de Soft particles. Tracks with momenta smaller tection of multi-hadron events, has no provision than 135 MeV/c or 225 MeV/c, respectively, fail to for charged hadron identification above 0.6 - 1.0 reach the TOF counters. A (model-dependent) guess GeV/c momentum. This note sketches a detection is that 5% or lOfc of events, respectively, have system over the available solid angle. a particle in these ranges, yielding The Mark II detector as now envisioned con e = 0.95; e = 0.90. sists of a 3 m. diameter, 3 m long solenoid magnet a b filled with tracking chambers. This cylindrical Detection of neutrals. The detection effi- magnet is enclosed in an iron flux return box ciencies are 0.75 and (0.31 + 0.56 x 0.5) = 0.59 open on two sides. Each open side allows TT/2 in respectively [0.5 is the detection efficiency of azimithal angle and TT/2 in polar angle free at the iron flux return, or 0.18 of 47r solid angle, if the the lXn PbJ. Assume that the effects of the jet model and nu -* 2-j decay reduce the number of open sides are oriented in the horizontal, both effectively isotropic tracks from 8 to 3. Then could be instrumented with particle identification; if the open sides are at the top and bottom, then only the top could be instrumented. As presently 3 ea = (0.75) = 0.42 designed, the Mark II coil is surrounded by .liquid argon sliower detector modules. The modules occluding % = (0.87)3 *(0$]8 = 0-007. the solid angle subtended by the particle identifi cation package would have to be removed. The major disadvantage to instrumenting the Mark II with Reconstruction of neutrals. For the conven particle identification modules is the fact that tional detector, reconstruction of 8 y's is the modules can be brought no closer than 1,5 m virtually impossible since crude energy informa to the source, requiring enormous volumes of de tion is available only for - 3 of them. In the low tectors. B detector we speculate that 5-10% of 8 y systems IDENTIFICATION SCHEME Particles will be identified in a triple tandem array of Cerenkov counters. The counters are primarily threshold devices, but pulse height detection will increase the range of separation. The three counters, labeled II, 12, 13, have re fractive indices 1.0045, 1.008, 1.05 respectively. A chart of Fig. A-l shows where particle separation is made. With this scheme n-K-p separation is accom plished between 1.6 GeV/c and 10 GeV/c, Pulse height comparisons are required in the aerogel, 15, to separate K-p from 3.2 GeV/c to 4.2 GeV/c and in the propane, 12, to separate K-p from 8.5 GeV/c to 10 GeV/c. Separation of ir from K and p is complete from 0.6 GeV/c to 15 GeV/c. K-p separa tion ends at about 10 GeV/c, and perhaps lower depending upon the ability to measure pulse heights in 12. Assuming a conversion efficiency of Cerenkov photons produced to photoelectrons of 0.1 in each counter type, a B <* 1 particle yields 8 photoelec trons in the 2 m II, 40 photoelectrons in the 50 cm 12, and 100 photoelectrons in the 10 cm 13. Fig. A-2. Plan view of particle-identification system which could be added to the SLAC-LBL Mark II detector. a ^"roSS" { 1 >3 ' , "i1" ( 1345STBB10 Monwmurn |G*V/C] Fig. A-l CONFIGURATION The particle identification module is shown in Fig. A-2 (plane view) and Fig. A-3 (elevation view). It consists of four different elements. We describe a module for one of the open quadrants; numbers of elements refer to one module subtending 18* of 4ir. There are two large tracking chambers with position resolution of - lain to flag hadrons interacting in the magnet coil and S-rays produced in the coil. The chambers each have an area of 1,7 in *3.5 m = 5.8 m2. Next cixnes a 10 cm thick - layer of aerogel at a distance of 2 m from the interaction region. The 12.6 m^ area is composed of a 16x 20 array of aerogel modules, each 20 cm x 20 cm on a face with a 30 cm light-collection Fig. A-3. Elevation view cf particle-identification assembly behind leading to a phototube. 320 of system which could be added to the SLAC-LBL these 60 cm deep elements will be required. Fol Mark II detector. lowing the aerogel layer comes a layer of 8 pres sure vessels of diameter 1.3 m and length 3.5 m. Each vessel contains 8-atm projiane and three optic ally separate Cerenkov systems with 50 cm active will' contain a total of 24 separate Cerenkov length. The mirror area in each cell is 4.5 mz, counters. The final elements of the system are 24 divided into three 1.5 m2 mirrors, one for each atmospheric CO, Cerenkov boxes each 2 m long with separate optical system. The 8 high pressure cells a 2.4 m * 2m mirror on the back. The distance of the last Cerenkov mirror from interaction region is 6 m and the area of .he rear of the array is lows for time of flight (TOF) measurements. The 12m x9.4m = 113m2. There are a total of ~ 1D00 TOF and uCFM must be followed by an atmospheric drift wires and 368 phototubes. Realistically the gas Cerenkov counter to separate high momentum module probably covers only 15% of 4TT talcing into n and K. account support frames, etc. The 2 cm thick .aerogel with aim drift DECAY AND MIsrUENriFICATION region and 450 mrad Cerenkov angle gives a 9 mrad wide circle of Cerenkov light. The ti-K Cerenkov Decay. In detectors of the size vie are con cones are separated by 9 mrad at 5 GeV/c and the sidering, a crude table helps to evaluate decay K-p Cerenkov cones are separated by 9 rarada t probabilities- 9 GeV/c. Thus the identification is: ^L Prob. Decay 1. TOF IT K to D.6 GeV/c K-p to 1.2 GeV/c 2. Aerogel n-(Kp) 0,4-1.2 GeV/c 2 20% Tt-K-p 1.2-5 GeV/c; 5 10% K-p up to 9 GeV/c 10 51 3. 1 Atm CO, n-(Kp) above 5.5 GeV/c. 15 3% The tracking chambers give momentum and direction The kaons which decay will yield muons which will of the particle to be identified and a search is radiate Cerenkov photons. These rates can be re made of the three possible Cerenkov riiigs. Assuming liably calculated once the spectrum of surviving photocathode efficiency of 0.2 and transmission ef kaens is known. ficiency in aerogel of 0.7, approximately 15 photo- electrons are produced with which to define the 5-Rays. In the counters we are proposing Cerenkov rings. 0.2-1% "of slew particles will count in the Cerenkov medium itself due to 6-ray production. The low index materials (atmospheric pressure gases) have the Q.l% confusion level; the high index materials (high pressure gas or aerogel) fl have the ~ l 0 confusion. In addition, pressure vessels contribute 2-3% confusion in high pressure gas counters. Tracking chambers with - 1 mm resolu tion positioned 10 an behind 6-ray producing ma Inrmctuo ^ terials (such as magnet coils) will reduce the ogion -** J 5-10% confusion so generated to - 0.5a. Appendix B LARGli SOLID-ANGLI: DIFFERENTIAL CEHENKOV COUNTFR USING MICRtxntANNa WIDTlMn.TIPLIERS AND AEHOGlih This note expJores the possibility of a large-sol id-angle differential Cerenkov counter. Such a device is impossible with presently available photosensitive detectors which, because of their expense, require optical systems to collect photons produced. No optical system has been devised which focuses unages directed over 4n solid angle. How ever, if a photosensitive device with a large and segmented urea cnuld be obtained, no optics would be necessary and a differential counter could be imagined. MicroChannel photomulcipliers (uCIW] could be the answer to the large-area device re quired. However, scaling up from existing 5-10 cm- Fig. IM. Scheimit ie end view of possible future dif areas to 2f> m- and J-an jwsition resolution re ferential l.erenkov counter system using quired make this hyjiothetical for the near future. aerogel radiator and microchannel photo- niulttplier detectors with 1 cm- area covering JS m-. Assume that the \A'.M is availahlc in lavy.' areas witli a 1-cin square cell readout, Then a dif ferential counter could he constructed by placing a Cerenkov radiator of «cfracting inde.\ n - 1.1 at small radius surrounded by a drift region to allow the Cerenkov cones to separate in space, followed C. Hiiseliliom et aU, PUM46,8t> (1974). by the IJCI'M to record the positions of the phot mis. A detector (Fig. 17) could consist of an inner pro IS. Hichtor, 1Y.JM 40,24 (1974). portional counter, a J-cm layer of u - 1.1 aerogel, and aim drift region containing tracking chunbers li. Hanson, reported at J975 SMC Summer inside the |i(."l\M. Assuming the uClW records posi Institute in ('article I'hysics, :ind private tions only, i layer of scintillation counters i«>l- conemmicat ion. 4. li. Hanson and P. Oddone, private communication. 17. Similar to the Fennilab iJUcniJil Muon Identifier. F. A. Harris et al., Nucl. Inst, 5. J. J. AuU'rt et al_., Phys. Rev. Letters 35, and Meth. 103, 345 (1972)." 1404 (1974); IFM. August In et al., Phys. Rev. Letters 33, 1400 11974). 18. D. Nygren, PEP-197, in This vol mm-. I). Sl'fcih and DORIS contributions to 1975 Lepton/ 19. The 3 kGauss field would be too weak to provide Photon Symposium at SLAC. significant transverse diffusion suppression of the type Nygren has measured. This may or 7. K. Halbach, private communication. may not be a problem depending on the informa tion required from :Jie chamber (tracking vs. 3. P.LeComte and V. Perez-Mendez, PEP-201, in momentum fitting) and the readout method used. this volume. 9. Similar to the "General User Magnet," PEP-188, 20. Katsura et al., Nucl. Inst, and Meth. 105, in this volume. 245 (19727. ~~ 10. The DASP TOF system achieved o = 0.26 nsec 21. We are cautioned by discoveries of the radia resolution which, with a beam crossing time tive decays of *'(3684) and J(3095) against determined to 0.13 nsec rms as at PEP, would wholehearted embrace of this prejudice. It improve to 0.22 nsec. must be noted, however, that measurements of Y direction rather than energy were central 11. M. Cantin et al., Nucl. Inst, and Meth. 118, to these discoveries. 177 (1974)": 22. F. Sciulli, rapporteur talk at Fermi Lab 12. CERN-Heidelberg AB group, private communica Calorimetry Workshop (1975). tion. 23. l'liP-188, in this volume. 13. The scintillation properties of pentane were not measured by the nthors of Ref. 12. 24. These have all the properties of more expensive 5-in tubes with conventional -photocathodes, 14. We are indebted to R. C. Field for addressing except for a long risetime which is of no con our group on this subject. sequence here. These considerations imply larger light col 25. The inside window will be ribbed in order to lectors than were used, for example, in the minimize its unsupported length and thickness. high pressure counters of Ref. 1. Where possible, the ribs will be in the shadow of the coil lumps. The outside window 16. M. A. Green, (LBL-3677 [197511 » and private is necessarily corrugated (Fig. 13), giving communication. added strength. USi: 0I: MICROCHANNEL ELECTRON MULTIPLIERS IN HIGH ENERGY PHYSICS P. Lecomte and V. Perez-Mendez Lawrence Berkeley laboratory PROPERTIES OH THE MICROCHANNEL ELECTRON MULTIPLIERS MicroChannel plates were developed for mili tary purposes, mainly for night vision applica tions, Imt soBtr of their characteristics are of ; great interest in high energy physics. He will summarize briefly the properties of the available plates, bearing in mind that they are designed for ?/ image intensifiers and that they could be further optimized for our particular applications. These plates can he presently obtained with useful diameter between 25 and 75 mm, channel > - diameters from 'J to 51) microns, and electron gain per plate around 1(P to lir' (Pig, 1); the transit time of the electron is about on. nanosecond, and the time spread is well heluw 1QU ps. The plates arc extremely compact and are not affected by mag netic fields, but arc sensitive to heavy ions, lorf - energy electrons, soft -x-rays and UV light. f/ 1 enwiv Gains in excess of 10', with noise of 3 counts/cm2 sec are obtained when cascading two plates. A "cheiron" arrangement is needed to re I 1: duce the positive ion feedback, which could damage tlie photocathodc; spatial resolution is then di vided by two. As an alternative, plates could be - manufactured with twisted channels. I'hotomult ipl iers incorporating these micro- channel plates could be manufactured in various / shapes; they should he extremely compact and, in tin so-called Wafer design, where the photocathode / is in close proximity to the MicroChannel plate, 1 (extensively used in image intensifiers and in : / sorm? microchannel photomultiplicrs, Fig. 2 and 3) they should exhibit good inseiisitivily to magnetic fields and extremely low time jitter for single electron response. It should be noted that this type of tube is not well suited for energy measurements, the pulse height distribution of one plate being cither Multiplier Voltage exponential or saturated. Th? iiiaging capabilities of these plates nre also of flre.itinteres t and several met Hods; of Typical Characteristics electronic readout arc presently developed, in L/dc = 42 spired by the technique:* employed with proportional d = 14.5 p chambers. c OAR » 60 % R = 3 ^ 1Q8 U SOME POSSIBLE APPLICATIONS XM.7! fa) To our knowledge, no other device can presently ':omp«.-t CONCLUSIONS , / Phosphor itrt From the preceding, it is apparent that micro- / CEMA channel photomultipliers and electronic imaging devices are of great interest in high energy Pholocolhode jnJ input window physics. Dlpl miiytt •dCEMAmlensrfier Several manufacturers have the ability to XBL 759-83M produce such devices, but, in order to obtain products well adapted to our applications at a reasonable cost, we must define the exact specifi- cacionc we require and convince '. "ie manufacturers that a satisfactory market exists. Preliminary conversations show that, or.ee this is done, prototypes could be obtained within some months. CHANNEL PLATE REFERENCES PH0TOCATH0QE A good summary of the properties of micro- channel plates can be found in "MicroChannel Plate Imaging Detectors", by p.J. Ruggieri. IEEE trans actions on Nuclear Science •jS-.lg, i\o, 3 (1972). > COAXIAL OUTLET Applications to image intensifiers are des cribed in: "The Proximity Focused Image Intensi- fer", Bendix Technical Memorandum EOTfl 72ul (197Z). Some other interesting papers are "High Gain Imaging Electron Multiplier", W.B. Colson, et at., Rev. Sci. Instr. 44., No. 12 (1973). XBL 759-8:134 "Photomultipliers and the Transfer Techni que", J. Nussli, Philips Technical Review 30, be achieved only with energy losses of 10 to 30 MeV 23D. in the scintillator. "The Channeltron Electron Multiplier Array (b) Conventional photomultipliers could be and Its Applications to Low Level Detection replaced by MicroChannel plates as long as a good Devices" First European Electro-Optics Markets energy resolution is not required, where space is and Technology Conference. critical or where high magnetic fields are present. "Solving Intensifier Problems with Channel (c) The imaging capabilities of the micro- Plates", .1. Adams, Electro-Optical Syst. Design, channel plates, even at ultra low light level and November 1969. high speed, allow them to compete favorably with such devices as return beam Vidicon and image "Fast Timing Measurements us-ng a Chevron dissectors. They could he used for pattern recog MicroChannel Plate Electron Multiplier", Mike nition in Cerenkov counters, that is to record t. Green et al. N.I.M. 126, 175 (1975). PHOTON-PHOTON PHYSICS G. Barbiellini, A. Benvenuti, K. Berkelman, A. Courau, F. Foster, K.-W. Lai, F. Lobkowicz, J. Matthews, N. Mistry, ^.d T. Rhoades "... Further, when one considers the ex treme speed with which light spreads on every side, and how, when it comes from different regions, even from those directly opposite, the rays trans verse one another without hindrance, one may well understand that when we see a luminous object, it cannot be any transport of matter coming to us from this object ..." Christian Huygens I. INTRODUCTION Experimental research with e e~ colliding beams has been centered on the study of the one- photon mechanism and the production of C = -] states. With the advent Tagging kinematics: the relations among the of high-energy high-luminosity facilities such as yy center-of-mass energy, the center-of-mass PEP it will also be possible to study the two- motion, and the laboratory photon energies. photon process and thereby Definitions and formulas are given in appendix I. probe not only the realm of C = + 1 states , but also the inner structure of the photon itself.3-5 This i: the report of a study which extends the work of the 1974 PEP study groups on two- photon processes, mainly by considering the experimental cuts in tagging measurements re quired to obtain high-quality data, the resulting effects on the. tagging efficiency, the effects of transverse e+ and e" beam polarization, a possible design for a parasite tagging apparatus, and a suggestion for a prime photon-photon experiment. A systematic explanation of the kinematics and quantum electrodynamics of tagging, including polarization effects* is given in the appendix to this report. In the two-photon prpcess we a^e concerned ith the reaction e+e~+ e e" + hadrons, which we view in terms of a head-on collision of two almost-real photons: OB 1.0 e •+ e y ] YY •* hadrons. e" - e" y J XBt."»tl3& The process is identified by detecting in coinci + dence the outgoing e , e" ,'and presumably at 2. Deep inelastic kinematics: four momentum trans least one hadron. The e+ , e- detection fixes fer as a function of tagging angle and photon the energies E' and E'2, polar angles 0, and 62, energy. Definitions and formulas are given in and azimuthal angles the lost events with o < 10 mrad do not amount to 1 I 1 ' 1 ' 1 ' 1 ' 1 a very large fraction of the total. This 1s about the minimum tagging angle that one can achieve L Virtual Photon Spectrum viithout magnetic deflection, using detectors just outside the beam pipe (about 10 cm radius) 10 l^s. /** max 0.04 meters from the 1i teraction point. RVS. / N(E x,6) d8 The counting rate for a tagging detector sensitive over the range 9 e is 'min » < m-v P^- portional to In (8 /e ) (appendix II). Since W\\ Wv\ for E=l5GeV m1n this decreases rather gradually with 6 , it is 0.03- max difficult to decide on an -upper limit. In prac tice, the decision will depend on the desire for deep inelastic data, economic considerations, and compatibility with other parts of the experiment. ^-300 mrad. 0.02- We will show (in section IV) that reasonable deep •S^iobX. inelastic rates can be obtained out to about 9 = 300 mrad. l50v>O\ =IO^NS\ B. Energies 0.01 8 -2 _ ""^OJ There will be some minimum ?nyrgy E'j below which it will be impractical to tag the outgoing + w e or e- . It is difficult to distinguish elec max ^s. \1 trons by shower detection below E' ^ 100 MeV; multiple scattering in the beam pipe can degrade 0.2 0.1 0.6 0.8 1.0 i 1 . 1 . 1 . 1 . N the 8 and E' resolution; and background electrons w/E and photons from beam-gas collisions7 will cause random coincidences if E1 min Since it 3. The virtual photon energy spectrum integrated 1s rather difficult to predict what E' , will over tagging angle from zero to 6 - for have to be, it is fortunate that the YY luminosity various values of o . since the dependence is not very sensitive to tumax. d£ /dW depends on on beam energy E is only logarithmic, these "m-ix on*y for tne M9ter portion of the W range curves are approximately valid for a wide range of E. See appendix II for more details. (see A10, A12). For rate estimates we will there fore use E' .„ = 0, ww = E. min max e e". These variables together with the incident e+ or e- energy E determine the energies u, and One will also have to set a maximum energy uy* the four-momentum-squared qfand q„2, and tiie cut E' , either 1n the trigger electronics or polarization of the tagged virtual photons. The in the data analysis, There are several reasons laboratory photon energies u. and u« determine the for this. YY center-nf-mass energy W and the Tab velocity B of the center-of-mass. The relevant formulas 1) Elastic and radiative e e" scattering are given in append'" I and plotted in Figs. I gives a high rate at E'^E.' Although this re and 2. action is distinguishable by having w, and u,, both \iery small and col 11 near e e" [or at least coplanar, if there is radiation before scattering), II. TAGGING CUTS one may be able to exclude some accidental coin cidences by rejecting hitjh V tags. 2) The effective photon spectrum N(w/E, e min, The backgrounds expected to be important in max goes to zero as ui/E . 0 (Fig. 4). the tagging system are discussed in detail in the 3) For Me case Ej » E; the re sol 1974 PEP study report'. Any detector sensitive to W is given electrons which have lost energy but are emitted by at 9 = 0° will have 1002 occupancy from beam-beam bremsstrahlung. Even at very low beam currents 11 or at 100™ beam duty cycle, the coincidence rate W 2. E-"i/ '.) from the 0° double bremsstrahlung8 would com (t pletely dominate over the desired two-virtual- It diverges as Ei -> E. photon tagging rate, fortunately, however, these and other backgrounds decrease much more rapidly 4) For the case V, » EA the YB of the photon- with angle (typically o-3) than does the e-' photon center of mass relative to the lab is distribution for virtual photon tagging (see appendix II). Tagging ato= 0° is unnecessary. It is clear yfi from Fig. 3 that a? long as u/S is not too small, £/ECIFE7) and diverges as £j * E. Note that YS= cot >, where u is the laboratory angle of a relativistlc | l | . ~T- particle emitted at 90° in the center of mass. 1 yy Lymtnosily If rfl » 1 the hadron angular distribution will be considerably distorted by the cm. -» lab trans formation. 6 9 • , 1 max , 2 m* . In- |n -i 1 1 1 1 1 1— °l mfr» B2m(n Virtual Photon Spectrum t*ffla* E,> • lO-ZOOmrod. U.005/P . O.tO - 0.10 - ~ - 013 • • . / 0.2 " "min >w . 0.06 - ' ,1 1 1 1 TT- The photon-photon luminosity (defined in appen dix II) with the dependence on angle cuts divided out. The various curves correspond to different tagging energy cuts. The counting rate for ee -+ eeX in a bin of YY energy from H to H + Aw* is obtained by multiplying the value from the curve by the ee luminosity (in r^sec"'), the angle cut factors i angli i!n{0lma>!/6,lmin, A,^e^^•«'/e9»'",• tne normalized YY energy bin width, . tion a,„, for yy •* X. 4. The virtual photon energy spectrum integrated YY ever tagging angle from 6m* to om: See appendix II for details. Taking all effects into account, it appears = that E'max/E -80 to .95 is probably reasonable. For most of the rate estimates we will then use wE = •'• C. Single Tagging It is clear from Fig. 1 that if one could integrate over all angles 9, for one of the e+ or e~ by not requiring its detection, one would ga*n in tagging efficiency, especially for low values of the corresponding photon energy ay This would certainly increase d III. PHOTON-PHOTON ENERGY RESOLUTION £(^£-'). A. Dependence on Tagging Resolution varies from .055 to .076 for q2= .1 to 10 GeVZ. This would imply (for t=0) a AE'/E' between .01SE Since the YY center of mass energy is given and .14%, a completely negligible contribution to by the experimental energy resolution. The statement was made in the 1974 PEP Study" that "The Inherent resolution of the tag is limited by . . . radiative effects (1-3S)". the energy resolution 6W is This we believe 1s misleading. The question 1s how one characterizes the resolution of such a pathological function (see sketch). Although the average AE'is several percent (because of the W VU e^F?/ U e-ea7 . long tail), most of the energy losses are less Since any measurement at fixed W covers a range of than a small fraction of a percent. We believe ..j and w values (Fig. 1), the data will include that when looking for resonances and other struc 2 ture in energy, the latter criterion Is more events with a range of W resolution. It will be relevant. The long tall contributes to the back best for symmetric tags, u, = w„ = W/2, and worst ground but does not affect the "resolution". If for the most asymmetric tags. In the latter case this were not true, the masses of the $> and *' the contribution to 6W from the highest E' will would not have been measured so precisely. The dominate. important inherent limit is probably the spread in initial e+e" energy B. Intrinsic Limits on SE'/E' The e or e~ energy detected in any electro magnetic reaction will always be degraded some what because of radiation of an extra photon or photons in the reaction itself or later in the material through which the e+ or e" passes on its way to the detector. If P{k)dk is the probability of an energy loss between k and k+dk we have for the integral'" where q is the four-momentum-transfer-squared in the reaction and t is the number of radiation lengths (assumed « 1) of intervening material. E' is the e+ or e" final energy in the absence of extra radiation. P(k) itself has the typical k~ behavior as shown in the sketch. f Approximate dependence of the virtual photon polarization (see appendix IV) on photon energy. Otherwise, the radiation will contribute to the W= 15 GeV. Measurements of inclusive distribu electron shower and no energy will actually be tions rather than exclusive final states will be lost. The energy scale of a shower detector will of increasing importance at the higher W. Such be self-calibrating because of the Bhabha scatter distributions may be characterized by the usual ing events. variables, P,, and XF * Pfi /P^ (* denotes center of mass}. Assuming that the two photons IV. SOME POSSIBLE EXPERIMENTS are almost real, the physics devides roughly into A. 3 to 20 GeV Survey two regions: lj Central region (XF| < A Suppose we assume the following conservative 2) Photon fragmentation region |XF| > .1 cuts in tagging angles and energies, If one of the photons has q » 1 GeV2, there may 10 mrad < 8,, 6, < 300 mrad, be a third region, the so-called parton fragmen tation region. Under the assumption that the X- .IE < uj, »J2 < E, distribution is flat 1n rapidity, we would expect and taU: E = 15 GeV and o = .3 ub (actually, it similar total rates 1n the two regions. Figs. 9 probably will decrease with energy like o (w)= and 10 show the relation between laboratory angles and XF, Px for two choices of W, yB: a "worst" .24 (ib + (.27 pb-GeV)/W.). We can then derive case 1n Fig. 9 and a more typical situation in from (Al) and Fig. 5 the expected number of counts Fig. 10. The central region of X corresponds to for an experiment in which |j" dt s l(r°aif* (see F J ee i 32-2 -l o > 30° except at the lowest Fj. . Furthermore, Fig. 8). Assuming an average Z^^-10 cm sec , the fragmentation region for the "slow" photon such an experiment would take 46 days. The integrated number of counts over the whole W range 1s 2 x 10*. This should be compared with 5 x TO* w-ioMv. ?#*i.ss the expected number of ee+hadrons events, assuming R = 6. Fig. 8 also shows the maximum value of Y6 for the motion of the two-photon center of mass (see also F1g. 1). The angular distribution of hadrons in the center of mas'i will presumably be similar to that observed in nadron-hadron colli sions. We can probably expect the charged hadron multiplicity in the final state to Increase loga rithmically from 5U2.8 at W= 4 GeV to n«5.4 at Laboratory-frame Peyrou plot for yy collisions with W = 10 GeV, yB s 1.33. Curves of fixed center-of-mass fractional longitudinal momen : tum xc This repre sents a worst case in terms of distortion of the angular distribution by center-of-mass motion, for the experiment described in section IV-A. Estimated number of events per GeV interval in W for an experiment with E* 15 GeV, *> tc*v> V "z' Laboratory-frame Peyrou plot for yy collisions The dashed curve = with W = 7 GeV, yB -5. Curves of fixed xp shows the maximum *r6 of the center-of-mass are indicated. This 1s a rather typical motion for the same experiment. See section situation in the experiment described in sec [V-A for details. tion IV-A. -164- (i.e., XF < 0) also corresponds to large lab A simpler argument leading to the same conclusion angles. By symmetry the inclusive distributions is made by noting that a "typical" hadron 1s cannot depend on the sign of Xp, so that one emitted at a center-of-mass angle given by would not lose any inclusive Information by Ig <8> = sln'Vo^/ ) noring the more difficult to detect fast photon 1 - s1n" (<:Pi> n/w) fragments. Note that the interesting high Px events, where fcinematically allowed, always appear = sin"7 (.3 GeV - 10/10 GeV) at large angles. = 17°. The resolution SW will be no worse than .4 GeV multiplied by the percent tagging resolution 6E7E' (see Fig- 11). This is certainly appro priate for a survey experiment in which the data might be grouped in flW = 1'GeV bins. B. Resonance Hunting The above facts Imply that such an experiment is quite appropriate for running as a parasite Since the 1974 PEP study we have come to with a single-photon experiment using a "conven suspect the existence of many resonances 1n the tional" detector with an added tagging system energy range 2 GeV < W = 5 GeV. Some should have which is interrogated every time a hadron trigger C= +1, accessible only 1n YY process, (or through occurs. Whatever the host experiment 1s doing decay of hiyher 1" resonances). He list here a (o. t, inclusive spectra, jets, correlations, number of known and suggested C= +1 (and J>*1) particles and their properties. high Pj_ , neutrals, etc.) the YY experiment will also do -- with comparable statistics, but with a p NAME M(GeV) J rtot(HeV) decay distribution of energies W. The parasite experi iykev) ment also measures the YY contamination in the o 6 7.8-10" YY host experiment, of course. IT .135 0" .0078 It is also clear from the Y6 curve in Fig. 6 n .550 0" .30 .001 YY» 7r1n^ + and from Figs. 9 and 10 that for an experiment in 2 e .7 0 1.3?' 600 nit which one wishes to see several or all of the final state hadrons one would prefer to have an n' .958 0' 4.5? (SU3) .023 YTTT.PY angular acceptance for hadrops which extends to + 12 much smaller angles (say ^ 5 J than is customary f° 1.27 2 6? 170 mi for a detector designed for single-Y experiments. 1.31 2+ ? 100 PIT, nir "2 A° 1.64 2" 7 300 fit *3 I'll n 37 0" 13?13 c i 5" YY. • . rr CM Energy Resolution The ener gy-i ntegrated pr sduction cross section 0.5 assuming —^ = 1 •/• - O.IE<» 0.2 - as) i where SW Is the r.m.s. resolution in W. Could we find the n resonance in a YY experi best*^^ ^V ment at PEP? With the conservative tagging cuts 0.1 - - and ring energy E assumed in the previous section, the YY luminosity (Figs. 5 and 8) goes to zero at W= 3. We consider three alternatives: la) reduc ing the dt 10 1"* the total number of tag J*ee • ' coincidences from n. production will be The resolution in the YY center-of-mass energy for the experiment described in sec tion IV-A implied by 1% resolution in the "•-flflVI- 4&l) — fa- C. Deep Inelastic Electron-Photon Scattering One can divide the yy physics program into three areas according to the characteristics of the virtual photons: both photons almost real (already discussed), c almost real photon and one far off the inass sho!, and both photons far off the mass shell. This division is in increas ing order of experimental difficulty, since the •• n & 34(1 counts. yy luminosity contains a factor q~2 for each (thfs should be reduced a factor of 4, since photon (equation A14). 3L is proportional to E% in order to com The second area of research may be the most plete the experiment in the same time.) exciting, since 1t allows us to probe the inner 3 n = 85 counts structure of the photon itself . The idea is to use the flux of almost real photons emitted by one of the beams as a target for large-angle These rates are rather low but not impossibly low. scattering by the other. The reaction studied W.iat about resolution? Suppose we assume is then + + chat the npn-resonant background cross section e + Y -> e" + hadrons; and the cross section is given in (A18) in terms ° = -3 vb • In order to get o. = a for the nr of photon structure functions. n , we can compute the required resolution from (5): Wreq = 3D MeV (r.m.s.J. 3000 The expected average tagging resolution can be estimated by averaging equations (1) and (2) and assuming optimistically that iE'/E1 = U: a) 6W = 240 MeV b) fiW = 390 MeV c) M * 170 MeV In all of these examples the expected energy resolution is insufficient to see the n with the assumed properties at a signal/background of better than one. There are several ways out. 1) Run the experiment at a much lower energy E or at a lower energy machine. But although the resolution will improve, the expected number of ? counts will be lower, taking into account jf«E expected for PEP and the lower X of other mach ines. 2) Reconstruct the resonance mass from the detected decay products instead of from the tagged e+ and e" energies. This will not be easy. Hany events will have high y6 values (see Fig. 1), im plying small lab angles; and 0" resonances usually have large branching ratios into neutral modes. The requirement that the final state be recon- structable will certainly reduce the number of useful counts. 3) Use single tagging. This improves the rate but since the energy of one of the initial photons is unknown, one still has to reconstruct toe resonance mass from the decay products. The backgrounds will be worse, since a single tag is not enough to guarantee that it is a two-gamma t\j2 <«V2) event. 4) Give up on low-energy resonances and rake the search at higher masses where the resolution (Fig. 11) is much better. Here again, the rate The number of deep inelastic events per GeV will be lower unless P increases sufficiently interval in q§ for the experiment described to offset the M" in the cross-section (equation in section IV-C, assuming various values for 3) and the W"1 in the luminosity (Fig. 5). the minimum v. In order to estimate counting rates for a (or vj, are given (A19) in terms of the structure r T deep inelastic scattering experiment, we assume functions W,, W2 » W/. The transverse polarization E= 15 GeV, we assume that the "target" photon is asymmetry o. is separated by noting the dependence tagged with 1 < cj < 14 GeV and 10 < 0, < 50 mrad, of the counting rate on the angle $ between the we assume that the "deep" spacelike photon is two photon polarization planes (appendix IV}. tagged with I < E^ < 14 GeV and with 8, fixed by To separate the transverse oT and the longi <& (although always d„ > 10 mrad), and we take a tudinal a. we must measure a twice at the same naive guess at the values of the structure func 2 L YY tions (appendix V). The predictions are then q« and W but with different values of e*. This plotted in Fig. 12 as a function of q2 and v . . is actually nowhere near as difficult as it is in The event rate per GeV, shows several thousand electron-nucleon scattering. The longitudinal polarization E2 is mainly a function of just events at reasonable q9, v values for a typical f -p 38-2 1) u, = 14, t^ = 4 GeV, 02 = 200 nirad events are predicted to occur for 70< B?< 400. mrad. There are three structure functions that can e, = .916 be measured in an experiment in which the photons 2) u,- 5.6, wg • 70 GeV, can be transversely and longitudinally polarized. In order to discuss the possibility of separately .394 measuring-the three functions, we use the equiva The counting rates are essentially in inverse lent form of the cross section in terms of u^ YY ratio of the angles 6, and are therefore not very (A17): different. In this example, using the <(>-averaged = a + E: a +£, C0S 2 mp^Buremenis o(l). a(2) we have °Yy T 2 L £2°A % The o , a,, a > which are functions of qi i T A c>(t) - oC2) fl2 (mra '•, < ^>" s Every PEP experiment should have a two-photon ' Pb GLASS S.C. OR No I Sf*CEFOB COMPOiSffllNG NHWSNET 3 4 5 6 T 7 METERS FROM INTERACTION POINT 14. One quadrant view of the proposed parasite tagging system described in section V. It is symmetrically arranged on both sides of the interaction point and around the beam line. energy accuracy (^.08 GeV}, at least for C. large-Angle Tagger (40-300 mradJ tagging at small angles), 3) angular resolution of the order of 2% Here again we use lead-glass for energy r.ra.s, or better, resolution, proportional chambers for angular 4) shields and scrapers to minimize back resolution, and scintillators for timing. To grounds from beam-gas collisions and minimize the area, the distance to the Interaction synchrotron radiation, point is made as small as the detector of the 5) low cost, if possible. host experiment will allow. This distance and An apparatus meeting these requirements 1s the size of the open cone in the host experiment shown in Fig. 14. The detector on each side of determine the actual upper limit of the acceptable the interaction region consists of two parts, angular range. The lead glass covers a total dividing the angular range at 40 mrad. area of about 7 mS 14 radiation lengths deep, for both sides of the interaction point. Since that is a considerable amount of lead glass (and proportional chamber area), one may not wish to equip every interaction region with such a large detector. A smaller e may be sufficient for B. Small-Angle Tagger (10-40 mradJ many experiments. At the larger angles one can also group chamber wires and still have the re The resolution requirements can 6e met by quired resolution. using lead-glass (or sodium Iodide if one needs the ultimate resolution)17. A total frontal area An atmospheric of 1 m and a thickness of 14 radiation lengths is included to give pion rejection. will be sufficient to cover the angular range on both sides of the interaction region. The angu D. Constraints on the Host Experiment lar resolution can be accomplished with multiwire proportional chambers with 2 mm wire spacing. It 1) The antral detector must have an open may be advisable to include a layer of scintilla cone of at least 10° half angle in both beam tion counters to improve timing resolution to directions, corresponding to the tagging aper suppress backgrounds. Ideally the detector should ture. A solenoidal field in this region does not be placed close to the fl-st quadrupole magnet significantly perturb the tagged e+ and e" tra Q3, 9.5 m from the interaction point, however jectories; the effect can be taken into account compensating magnets may have to be Installed, so 1n the analysis. that a m may be more likely. The actual e^ will 2} Compensating magnets, if required, must be determined by this distance as well as by the be placed beyond the tagging counters at the ex minimum clearance between beam and counters that treme ends of the 20 m straight section. The can be tolerated by the beam and by the counters. maximum field B in a solenoid compensator, and The double tagging rate varies as In (e 7e -j )- max m n therefore the minimum length & is limited by the Since these detectors will be located very following rule17, which insures thai. fi\> remains close to the beam line, they will be especially less than .02 per compensator: vulnerable to backgrounds from beam-gas collisions and synchrotron radiation getting past the scrapers st the far end of the straight section and scat (6,i'.)Bj. < |2 tered from the scrapers at the near end. Some lead shielding will certainly have to be placed between the detectors and the jeam pipe. As soon as possible, tests of these backgrounds should be where B& for the compensator is i BJt for the made to optimize the placement of detectors> central solenoid, say ^ x 40 kG-m for the General shields, scrapers, etc. 113 The small-angle tagging system is essentially User Magnet; 6= 500 ir, and Pits-m'Pbew/- " the same as that proposed In the 1974 PEP study10. 500 kG-m at 15 GeV. This implies B <_ 30kG, which General User Magnet (!5kG Solenoid) Dipole Magnet Pb-A Shower Detector Muon Chambers Muon Chambers* Compensating Solenoid Inner Detector (Cylindrical Drift Chambers) Superconducting FluK-excMing Pipe Inleraclion Point One quadrant view of the prime photon-photon experiment described in section VI. It is symmetrically pliiceJ on both sides of the interaction point and around the beam tine. means that the field of the compensator can be in the wrong direction. The energy resolution of 20 kG and its length one meter. The limitation the Nal detector is assumed to be SE/E-0.02/E74 becomes more serious at lower beam energies. If (FWHM).1' The resulting resolution in wean be the compensator is only one meter long, one will obtained from Suction III. have to lower all the solenoid fields for beam The large-angle tagging system spans the energies below 12.2 GeV. range about 75mr to 350mr. This angular range is covered by the "forward" detectors consisting of VI. PRIME PHOTON EXPERIMENT large dipole magnets straddling the beam-pipe, with "flux-excluding" pipes to shield the circu We describe here a detector system designed lating iPams. The wall thickness of such a specifically for two-photon processes, to be run superconducting pipe, including coil, coolant, as an independent experiment. The system is insulation, and vacuum vesse", is expected to be identical in the electron and positron beam direc about 4 cm.19 A 30W He-refrigerator would be tions, and symmetric around the beam axis. Only required for the two sections. Trt large dipole one quadrant is shown in Fig. 15. magnets, of aperture 3m x 3m, are suitable for use at 90° with the General User Magnet as high A. Tagging Systems momentum analyzers. Two tagging systems are provided. A small- The forward detectors also serve to detect angle tagging system, in the angular range about and identify hadrons produced at small angles to 12mr to 50mr, utilizes an array of Nal crystals the beam direction. Sets of drift chambers (trans arranged around the beam pipe. The front face of verse to the beam axis) provide trajectory infor the Nal array is 7.5m from the interaction point. mation with high resolution (0.2mm). The momentum It is assumed that 1.5m - long compensating sole resolution in this angular range is chosen to noids will be installed at the ends of the inter match the momentum resolution in the central sole action region as close to the quadruples (Q3) as noid detector for tracks at about 30° to the beam possible. The beam pipe in this region has an axis. With a dipole field integral of 10 k Gauss- outer radius of 8cm. with a sheet of lead around meters, and trajectory information from a lever- it to shield the Nal from soft X-rays. Two sets arm equal to the length of the magnet (1=1.5m) as of drift chambers spaced one meter apart provide well as within the magnet, the resolution i- trajectory information for the scattered electrons. expected to be Sp/p* -\. 0.3*(GeV/c) . The mater Scintillation counters wittngood timing resolution ial in the inner detector before the forward distinguish electrons from photons and suppress trajectory chambers is restricted to about 0.02Xo. background due to scattered electrons travelling Note that a special beam-pipe with "saw-tooth" corrugations should be Installed for a length of for use as the central detector for a 2Y-experi- about 1.4m to prov.de minimum material 1n the path ment. Its 1 rge aspect ratio (L/R = 4m/7m) allows of electrons of 75mr < 6 < 350mr. good acceptance and resolution for forward - going hadrm states which are produced in asym B. Central Detector metric 2Y collisions. The iron end - plate 1s specifically designed with an open cone up to The proposed General User Magnet {PEP - 1975, e =iQ° (350mr) to allow tagging and detection of elsewhere in this Report) is particularly suited very forward - going hadrons. The solenoid 1s 16. Section view of a portion of the prime photon-photon experiment in a plane normal to tho beam line. -170- expected to produce a reasonably uniform 15kGauss in depth,with 3 full cells making up approxi field, with momentum resolution 6p/p2 ^ 0.56% mately one radiation length (1.07X ). These (GvV/c)-1 at 90°; this assumes a radial distance cells are grouped for readout. The first three of 0.8m for trajectory measurement within the ceils are grouped together (about 1X0). The solenoid. The resolution improves as sin 6 for next two cells are read out in * and 9 strips forward tracks up to 8 i 30° and then deteriorates respectively. The C. Small Angle Detectors At 0 < 20°, particles will pass through the and y, while coarse segments divide the counter- holes in the solenoid end-caps and will be an in * and ft segments: <\$ •- 2V16, it- ••- lOOmr. The alysed in the dipole magnets. Trajectories are fine resolution is sufficient to identify the two determined by a set of transverse drift chambers photons of a 10 GeV/c r° decaying with minimum as shown in Fig. 15. Each chamber provides an opening angle. unambiguous space point, with 0.2mm resolution in The hole in the solenoid end-cap can be the bendinq plane and a few mm in the other di filled with a gas-cerenkov counter, 1.2m long, mension. With a chamber at 5m from the inter preferably at atmospheric pressure with thin action paint, the angular resolution in the non- windows, the total material being restricted to bend plane should be < 2mr. The momentum reso about .02XQ. COo at atmospheric pressure will lution, with a centraT field of 4.5k Gauss, separate electrons from pions up to about 4.6 (8L2 ^ 10k Gauss - meter2), is ^p/pz -* 0.3i GeV/c, and pions from kaons up to about 16 GeV/c. (GeV/c) , with these errors in momentum and The yield at »= 1 is about 10 photoelectrons for angle, the mass resolution obtained from re 100!- light collection. constructing two-body decays of a mass Mx moving The detection of forward muons suffers from along the beam with high momentum px = 10 GeV/c the background from decays of pions and kaons, is shown in Fig. 17. Thus a pD-meson at because of the 6m flight path. This rejection p % 10 GeV/c would be reconstructed with an limit is now 6xl0'z a(Hadronic), for muons above error 6M^ ^16 MeV. Note that particles at 3 GeV/c, so that an optimum of 3Aa(jS should 50mr < t) < 75mr are absorbed in the material of suffice to keep the hadron rejection at the level the flux-excluding beam-pipe. Those at 0 •. 50mr of 6%. Thus 50cms of Fe in two slabs, with two travel in a field-free region once they are beyond seti of spark chambers is sufficient. the solenoid, and are detected in the Nal tagging counters, where they deposit about 250 HeV in t>. logger 20X0. This represents a lower limit on the acceptance of tagged energy. We emphasize that The trigger for 2y events is a double-tag. YY collisions at q2= 0 may resemble pp collisions in any combination, of one small-angle Nal counter at the ISR, with typical small transverse momenta or one large angle shower-counter segment. In p = 300 MeV/c, requiring good resolution and each range of O, a thin scintillator is included identification at smali angles. in the trigger to demand a charged particle. As Following the drift-chambers in the dipole shown in Section II, the cuts in a require that is an array of shower detectors. These consist the scattered electron energies lie between of segmented Pb-Argon cells as in the central E' = 0.1E ard 0.9E. The trigger hardware should detector, with a total depth of -X0. The fine demand limi s slightly outside these, so as to resolution is provided by 6cm c.Je strips in x allow clean :uts in the analysis. VII. RECOMMENDATIONS cU f 'Icr (ft- — e*X) 1. Every PEP experiment should have a two- photon tagging system, such as the parasite apparatus we have described. This requires that = ^if a(y^^x). (M) every experiment be open at each end (at least 10°) to accomodate the tagging aperture, and that any compensating magnets be placed at the ends of d_£yj the 20m experimental straight section. dy contains the QED factors for the fluxes 2. A two-photon experiment optimized for of the two virtual photons. hadron detection at angles down to about 4° should be done. We suggest a possible setup using the proposed General User Magnet supplemented by transverse field magnets at the ends. 3. As early as possible, tests of the back ground rates with various tagging arrangements, shields, scrapers, etc., should be tried to APPENDIX II. « DISTRIBUTION OF YY LUMINOSITY optimize the details of the tagging system before the final design. In the Weiszacker-wniiams approximation we have21 _L J£" N(E.*,A)N(e.t,A) APPENDIX I. KINEMATICS AND NOTATION Measured quantities E = initial e+ or e" energy where Ej, Ei = final e+ or e" energy 9,, e~ = final e e~ polar angle (we will t0 oe usually take 62 the larger one) Ue have already integrated over $. and Note that since 8(l-x)ss flu, one should avoid W= l/^uj,!!)- - q^E'i/E (assuming using approximate formulas where B is large and q* «^2, m«E) x is small; the implied e is unphysically large. : total YY cm. energy In the limit of small x the 6 distribution is very strongly peaked at 8 = 0, although not 2 ^o)jtu2 if qj, qz « 4LJ1U)2 actually divergent, as the approximate expression implies. As x •* 1 the forward peaking becomes gentler. velocity of YY c.m.s. relative If only the distribution of £yy in U and/or r°>l to lab 6 is desired and not the distribution in Q? or Oimensionless normalized quantities Oo, we need the integral of N(E, x, 0) over 6: Xj * oi./E (similarly x„) 2= U/2E =7x,x2 Q^= Q?/E2 •= (1-x) a\ (similarly oj) The relations between z,£ and X,, x,, are plotted in Fig. 1, and the relation between Q, and xr 8, in Fig. 2. Fig. 3 shows plots of this integral for E=15 GeV (the corresponding plot in the 1974 PEP study6 is Photon-photon luminosity in error by about 102). Let y = any variable or variables (such as To integrate the photon spectrum N(E, x, o) given in (A3) over 9 between e,- and e„ , the x,, z, Q^, . . .). tillm n maax most accurate procedure 1s to take the difference between two evaluations of the integral (A5) taken The counting rate per unit y can be expressed If one also imposes a cut on a (^ee sketch), N(.«.&....eM.)--%[(h(, O^nl^ limit becomes '2-0% «•«• o-rie. H -1"«'*(.-«>6, ;?•]. (A6) This is plotted as a function of x for several Equation (A9) is plotted in Fig. 5, with the choices of fl.„, 0„„ in Fig. 4. Note that dependence on 0 and e divided out. in i n ma x , ax N(x. o . , ii „) is independent of z ; N(x=0)= 0 A simpler but cruder approximation tc min max d<£ /dz, accurate within about 20%, is obtained and N(K>....f..„)*^[MO-''Ji.,^ by taking {cf. A7) 5* f 6. (An) P~ » -•> eJ»„ . We now have Thi1s give£i£s » —r in ; i..\ -== (AS) r* i2> j_ _<«:„_ _ 2NC',r,.-f-,.^lN(lr',g-,...f,-.\ An accurate integratiun over x. can only be done APPENDIX III. numerically, but a rather good approximation, exact in the limit xm,-„ » '•'-,. can be obtained The y$ factor of the motion r>f the photon- mill ITIrf X photon center of mass is a conv ,iient measure of by usino (A7): the distortion of the hadron a 'gular distribution. For any choice of x , , x, erd z one can read i« "rfi" ' 'f1 ",•-..,. "A..... £ off the maximum i£ from Fig. 1. We rewrite (A8) in terms of z and yK. using the approximation •^'-••'A'-s;-^)'1'' (All): Ac CA'» That is, at all z the Hminosity is distributed The limit of integr.-'.ion x cdn be sesn in the in vU almost uniformly {see Fig. 6} up to the sketch. We integrate along a curve of constant maximum value. + z within the upper left triangle bounded by If one of the tagged e or e" is defected at dashed lines (there is a factor of 2 to account a moderately large angle (say 0,), ttie corres for the other triangle). ponding photon will be off the mass shell, with a spacelike four-momentum-squared Q|. The y, lumin osity for such collisions is obtained from (A?) and A{3). We reexpress N(E, x„, e?) in terms of E, x„,anc * % an<^ """tegrate N(E, x,, t^) over '•, as in (A6). The resulting function of x,, x?, Q? is integrated over x,, x~ at fixed z (with some simplifying approximations) to get Jf„ e* and e- beams has apparently not yet been done. analogous to the W, and W? defined for the pro- Some generalizations can be made however. 1) The rate of tagging counts will be uniform Srx«. E Ef/u), 1n azimuth *, and $- even when the e+ and e* beams ds dE-*U«£ 22 a ?,«, ^..gfc are transversely polarized. 2) If the e* and e~ beams are unpolarized, thereof ?1 be a polarization of the virtual photon *%w?*(>-%)»W with E in the plane pf-tanging. The degree of •0 polarization will beb^3,Z4 <-£,j-(l- 3) In the limit x1 •* 1 the photon polariza The parton model predicts^ 2 tioion j""[e"l is 700% in the plane of the e" or e" spipiin 23,24. For intermediate x. values the situa- tion is presumably some mixture of the two limit fx^'-W.'' 2*.£<.\*-HxG-«). ing cases. On genera! grounds the YY cross section can be decomposed as follows:5 where x= qpZv (0< x< 1) <\\, *} - oT(W, q^ ) + ^(W, q* ', ,W. 4 : ) _. 4 fie2 °A*W' qPco s2* * (A17) £e- = sum of fourth powers of parton charges. where qf « ql In order to estimate rates we assume (without angle between photon polarization planes justification) that \i!= g- \ftil (i.e., o,= 0), and eA = transverse polarization of the vWj * (o /o ) uWJJ * (1/300) (.3). *" photons Yp pp •= E,, e„, if e and e" are initially REFERENCES unpolarized (A16) 1. C. Huygens, Treatise on Light (The Hague, K °L> 1690). 2. S. J. Brodsky, Journal de Physique 35, C2-69 ^A = \ oL = a for the collision of a longitu 6. G. Barbiellini, G. Rlngland, 6. Shen, W. dinal (scalar) photon and an almost Toner, W. Vernon, 1974 PEP Summer Study, PEP- B7, p. 525. real unpolarized transverse photon: Ref. 6, p. 534. 0 at <\% •• 0. P. DiVecchia, M. Greco, Nuovo Cimento 5GA, 319 (1967). Ref. 6, p. 530. APPENDIX V. DEEP INELASTIC L. H. Ho, Y. S. Tsai, Rev. Hod. Phys. Ref. 6, p. 529. If the final e+ or e" is detected at a large T, F. Walsh, Lecture notes, Maria Laach, 13 angle 0„, the process can be described as the Sep. 73. -175- 13. E. Elchten, K. Gottfried, T. Kinoshtta, 18. J. M. Paterson (prlv. (omn.) J. Kogut, K. D. Lane, and 1. M. Van, Phys. 19. S. St.-Laurant (prlv. conn.} Rev. Letters 34, 369 (1975). 20. Cantin et al., Nucl. Instr. 112. 1// (l'l/4). 14. T. Appelquist, A. deRujula, H. D. PoHtzer. 21. 5. J. Brodsky, T. KinosMta, H. Tci.i.-.i™. and S. L. Glashow, Phys. Ri-v. letters 34, Phys. Rev. D4, l!,3? (1971). 365 (1975). ?2. S. J. Brodsky, G. jranmar (priv. HHIII..) 15. H. Terazawa, Rev. Hod. Phys. 45, 615 (1973). 23. N. Dombey, Rev. Hod. Phys. 41, 236 (l'l,'i). 16. Ref. 6, P. 531. 24. J. 0. Bjorken, Phys. Rev. Dl, 1176 ll'i'n). 17. E, Bloom et al., 1974 PEP Sumner Study PEP-1: p. 276. REPCRT OF THE NEW PARTICLE GROUP A. Carroll, B. Cox, A. Eisner, K, W. Lai, F. Lobkowicz, M. Marshak, J. Marx, J. Matthews, N. Mistry, C. Morehouse, J. Poucher, A. Rothenberg, A. Seidl, Eid D. Yount Abstract these signatures and iwthods lor decrfn3lnc the We review the properties of possible new backgrounds relative Ic the signature signals. A particles, and conclude that there are some general New Particle Detector Facility is proposed which features that a New Particle Detector con exploit. compromises the new particle signatures as little We investigate backgrounds in the n.ost lik«ly signa as possible relative to a mere general culti- tures and propose a New Particle Detector Facility. particle detector. Finally, a few 3pi-c: inline J Several more specialized detectors are also de detectors are described. scribed. II. SUNMAHY OF HEW PARTICLE INFORMATION Table of Contents (1) Heavy leptons • excited, sequential, gauge I. Introduction theory, and qu; :i-sfjble lieuvy leptons II. Summaries of New Particle Properties {from (2) Charged Intermediate Vector Bosons I974 PEV Summer Study) (3) Charmed Herons III. Signatures and Experimental Parameters (h) Resonances 1. Quarks (5) Quarks 2, Magnetic Monopcles 3« Tachyons (6) G1UOI13 1*. Heavy Leptons (7) Honopoles 5. Stable Charged Particles 6. Stable Neutral Particles (6) Higgs Scalers 7. Charmed Particles (q) Neutral Intermediate Bosons 8. Scalars 9- Vectors (10) Lee-tfick Particles 10. 2~y Production (11) Tachyons 11. Compton Production The following abbreviations pre used in the tables: IV. Hadronic Backtround to Lepton Signals ir = either e or n 1. Introduction h = hadron 2. Misidentification of Hadrons as Electrons MM = missing mass 3. Misidentification of Hadrons as Muons M = new particle mass k. Conclusions m* = invariant mass V. New Particle Detector Facility COP = coplanar ACP = acoplanar 1. Detector - General COL = colinear 2. Magnet 9 = velocity of new particle in center-of- 3. Inner Detacl-or mass system h. Outer Detector R = a(e+e—* hadrons )/c(e+e- -ji+u.-) 5• Summary FF = form factor VI. Specialized New Particle Detectors III. SIGNATURES AND EXPERIMENTAL PARAMETERS 1. General User Magnet (GUM) 2. Split Chamber 1. Quarks 3- Liquid Argon Neutrals Detector (LAND) k. Off-Axis Solenoid Wo consider the detection of stable quarks of 5. Streamer Chamber fractional charge and unstable quarks. Stable Appendix: PEP Cross-Sections Estimates from integer charge quarks are considered elsewhere. Hadron Machinec Quarks are assumed to be produced in pairs so that they carry the beam energy, EQ. I. INTRODUCTION Fractionally Charged Stable Quarks It is desirable in searching for new particles to have some general signatures which would reveal Detection requires charge and/or mass measure the existence of more than one type of new particle. ment. Since dE/dx — Z^/02 anomalous ionization can It is apparent from the excellent tabulation of be measured if the quark velocity is high enough. "conventional new particles" in the 19?4 PEP Summer Landau fluctuations in a given ionization measure Study (included here in Section II) that the pres ment introduce fluctuations in dE/dx of - 20J, so ence of leptons is the most general signature for that one should Independently sample ionization at a search. We analyze background contributions to least 3 times with each sampling required to be Production (10.-,38" . Decay Modes/ Experimental Mechanism Branching Ratios signatures Present Limits PI.P Limit (1) HttVY LEPTONS [1] 1) acp pair n * < 2 Gev 'It) (A) Excited ui* x (J8gC) / r I* i~ leptons + 1 or 2 hard r's X2 ^ ln"3 2) bump in m C,t*~... *r vtyt... ;*; 250 x (se^.) If vt massive, 1) acp it pairs m < 1.2-2.2 GeV < Ik GeV t3) »,-•/(/ v,i 2) / * h's -. / * h's 3) > 2 charged rth-sj/ru) ~ R /•s U) missing p. and E (B) Sequen IO" x (Jtef) r ,V 1 1) acp 1/ pair m, < lU GeV tial - J(«'; ' 2) 1 * h's Leptons 191 [10] r(h's)/r(i) ~ R (:,). (:;i- [1] General searches: M. L. Perl, SLAC-PUB-lo(j2 (1972); Decays and Production: V. s. Tsai, Phys. Rev. D>t, 2821 (1971)i X, J. Kim and y. s. Tsai, Phys. Lett. U09, t c a " 1 V* «:! si s .A, Ib'fliA^vfl — 1) J -I "'•" • -'• ..; ; rcCic-.icr. " "•Cr., K W &•?*•-IT-"-r.".t»i Prescft*. Si." ;'trM.Clx '•Kd.Ta.i&T . ' * r- : :-•:.•-; :r-.. >r.:;.- -"—-.'--">-• Ueli; Liri-.: • .• (£) CHAB0E3 <») eV * *V 1W. Ccr ." •'-• v ' O *lftv. .".t>j Frr*ent: • 1- >v VECTCB • t '"'"•.! _ ••- v,v •. •• . K, i !-• -» •. . BOSCKS '""* " """ "' '-* * * -' " ('. * ch'-ive fr? ^ *'.'•".:':.^ >y. m-Mi la! _ ruc-.cr . :, ^-, -u, *?,: *.*. fij; r.: rt.,-u-:'.e ?fti*-? if. ra'-if • ...... rr . I.-X 2SB •.. liiil nr ;siijr4- : • '-•••? pol- M=«r.V. J : (l! •%" .« «f* ^ o >.••: :•;, • 1;.- :«".' '*' vl'*" •rl-.J ^ .-.sji.V j. v.: £ 'c) «V .V*V" Prcfafttlv -.c.iviiy U* r r.r'.«f. • i: >v !E. 4 intc t.ntir.fis '.nSrcRic crost fvec'-or prtr'.cr:s; V <, sc'T-.i j. --i-.i. .- Sfe frr extir-plc K. Cabifccf * + (3) CHARM 1) e e"-*CC - 1 - 100j C - S * h's 1) Apparent viola 2 GeV «- M ' )0 GeV tlf2#3J tion of strange = A A r -- cos o„ ,_ Ul_ ness if c -* h*s (non- C -. h's (non- _ strange) strange J c -. h's + -35 for W = : - sin *>c strange hadron lOGeV C 2) e+e" >CC»X ~ 1000 [ I* j T ~ sin & cos 9„ above thresh electromag old 7£ inclu c . s - {*»,) <;) Anomalous change netic sive baryon in production of :Viv)/!''h's) - production strange particles 10''- 'JO1- as EQ passes through Charm threshold y 3) e"V -+CX ^ 20 for r - cos &Q -12 weak T - 10 sec M 10 GeV 3) Signature 2) with zo ° an associated charged lepton [1] G. A, Snow, Nucl. Physics B^, 1+V,-U;.U {IJJli. [2] M. K. Gaillard, NAL Conf. 7UA3-THY (April 1T?U). [ 3j S. L. Glashow, IV Intl. Conf. on Meson Spectroscopy. [h] J. Marx, private communication. Production Rate . 10? cm" Decay Modes Experimental Particle Mechanism per , 10 \i+ix- Branching Ratios Signatures PEF Limits 1 (U) RESONANCES (a) Formation In principle, v For vector mesons, Peaks in c vs En 30 GeV + formation could v* •* hadrons, -1005; e e" - X for particular final account for a v' -* t^i \ 5 few f>. staves. large fraction of a'e+e" •*• hadrons) m S.g.f high-mass /ector meson laughters (v1) of D,u, and • [l] 5 (b) Tvo-photon 5 x 10 Depends on specific r'rr fixed E0, 30 OeV exchanKe with r»sonan?e peaks in X for 2 + [2.3] ory inclusively, for v A any X. (c) Production 'Depends .-n Depends on specific (v') Invariant 30 GeV .specific resonance mass peaks in resonance multi-h, Y-hf i-h, y-9.t 2.-E, etc., N. J« combinations. E.g., see e# in + e M(ey) for ee~-»e**i i (e*-*ev) (ii) Hissing mass peaks. [.'] F. M. Renard, Hontpelier (France) Univ. Report No. PM/7V3 (Feb. 197M [2] David «. 0. S. Leith, SLAC-PUB-lMO (June I97M [3] J- Hosner, SUC-PUB-1391 (197M Production ' 10J Experimental Mechaniam Decay Mode Signatures Present Limits (5)QUARKS' V Low mass; a /incnalous 10 cm for ray be quas iE,dx for •' l^_GeV for stable fractionally « qqX at charged Pine = 300 Gev/c, Htqh-J quarks) may decay : , e/3 or ^e/3) ' into hadron b, Massive jets COL particle 0 -r 3 '10 * cm , pair at low Z = e/'J, Mq < ^2 GeV; velocity ; *•• I y. 10" ''h cm2, c COL hadron 2 - 2e/3, Mq < 13 GeV lets : for pp -* qqX at s =» ^500 GeV^j '3] < 10-39 cm2, 2 = e/3, *\ < **.T GeV; 7 < 2 x lo"37 cn^ Z = 2e,'*, Mg < U.T GeV ; for p • A? -* qqX at p. = 70 GeV/cJ [U] .1; For quark search review see L. W, Jones, Phys. ioday . -1 L. B. Leipuner et al,, Phys. Rev. Letters j_l> 1 .. • 1- ';• M. Bott-Bodenhausen et al., Phys. Letters .JOB, -'"•; 1- ; [h] Yu. H. Antipov et al., Phys. Letters ^^OB, :7- . . V>*>9;. [Rate Production I (10.-3" cm" Decay Modes/ Experimental Mechanism Branching Ratios Signatures Present Limits PEP Limits (6) GLUQNS [11 (A) Neutral Possibly Not known s dependent ' i. K [1.1, structure in J (like ot i, = i- 15 e'e" • h's i Color singlet & h'sj T SUo singlet *" e*e" --*-"/ if G is not J o^ e e" -KKi, (B) Vector [2] SU^ singlet, etc. Color SU, or if synmetry Multiple! breaking [ 3] [1] S. Weinberg, Phys. Rev. D&, 605, 1(482 1 iyf'jLL]Fritsch, Gell Mann, Leutwyler, Phys. Lett. ,u [3] Chanowitz, Drell, Phys, Rev. Lett. JO, 807 (19"";), Phys. Rev. E^t .,-0'. 1 •"';,. [k] Dalitz, Talk given at PEP summer Study, 19?U. Rate Production (loS8 cm2 Decay Modes/ Experimental Particle Mechanism Branching Ratios Signatures Present Limits PEP Limits (7) MONOPOLES Tl.2"! e+e~ -» D+D~ No method to [|»] (A) Magnetic }(pN -. p'N'DTS) calculate rate (A) Free monopoles— charge— GeV + -- l.U - lo'^W /DYOHS because of the stable (rather accelerated for M «- > o*. 0" strong magnetic unlikely) in magnetic D A \ charge (B) Monopoies anni field. [7! V hilate and y's V 2 (g /r.c * 137) are emitted. /BJ Heavy ion [21, I'll Or, for Dyo.is, ization hadrons may be emitted. (c) Cerenkov radiation with elec tric vector field direc e" N tion rotation V of 90° from charged particles [81 (D) NO reliable estimates on M , but large masses ,- 6 GeV to 137 X PL.) have been suggested [2,3 \ 'E) Large number of y's from D*D" annihilation '1.' i [1] Dirac, Phys. Rev. jkt 817 U9H0). [2] Schwinger, Science 16^, 757 U969). [3] t»Heoft, CERN PREPRINT TH-I876 '197*0. [h] Kewmeyer and Trefil, Phys. Rev. Lett. 2i, 1^09 (1971.; Kuovo ciwento j^t \o- 1?.'.. ; and Trefil, private communication. [5] Alvarez, Eberhard, Ross and Watt, Science 167, 701 (1970). [6] Fleischer, Price and Woods, fhys- Rev. JL8U, ljgfi (1969). [7] Adair, Proceedings of the 16th international Conference on High Energy Physics, Batavia, kp ;07 (197--. [8] Shi-:herbakov, Proceedings of the lfith International Conference on High Energy Physics, Bativia, 2, *h0 ( IV Rote A Prcducticn 7sr-.it.iv Mechanise 1 • (5) HlaoS SCAUP. PARTICLE -1C2 • t 'c K - - K-. : (• ~102 + long. X. ? ^" pa. y <\ ;5^< Petti-, tr. -irsiiy (&) 1.1DTF.M. INTEKKEDIATE a >** -.h*3 B0S08 [1] ~w> D lRT.'r!--!-«'n.v + - H !>•.•:,•• i i. (10) LEE-HICK e e" -»l> -»e e" + - PARTICLE + - i-pr I t. •:.. in * v -HI M • hf»drcri3( 5) i,t.f-il-.r ;.; - {Heavy Ul M-. -..(.f --11 Photon) (+) ror Lee- !rit -t i< i. :n .•f" - .+* • e e (u J wick (-) for Heavy irr j;: • ••,! Photon 4 1 [1 Htir.t'.-rt, Pt.ys. Rev. I *, i:6t- (W67), Saloir, Elerr..>ntt»ry Pm-'.icl.- physics, -d. 3v«,'f. In i:-- t-v.t !r J'--"' , p. lc . Hlgge, Phys. Rev. Letters 1^, ';o6 (1^6'*). 1 :) J. Kirkby, talk given at PEP Sumner Study f lc- M- '" . Frfn K[ d^cay, see [jj. '; Lf-e> nrid Wick, flucl. Physics t», ?0O fWfrl), A. Littw-, Pt.lU T>.«-FU, r*t*rv*>r ; "i.iv-r.-. 1 , '. [•-. fcs.rrt.ay, S., Phyu. Rev, fllO,T;i; (lWM. Production Decay Modes/ Experimental Mechanism [ranching Ratios Signatures Present Limits til) MCHVOH [1] (a) T~ emit Photoproduction Cerenkov light (E = 1.2 HEv£ at ~ go° in electric field tor (b) v > c 0.5 e < tachyon charge < 1.9 e (c) M^ < o [2] (d) COL parcicle pair [1] 0. M. P. BilomiaV et al., Art. J. Phys. JO, 718 (1962); G. Feinberg, Phys. Rev. 159. 1069 0967). [2] T. Alvager and M. N. Kreisler, Phys. Rev. 171, 1357 (1968); M. B. Davis, M. N. Kreislsr, T. Alvager, Phys. Rev. JlSj, 1132 (1969). several (probably 3) standard deviations from proton, we find that at 15 GeV the method works mjniirum ionizing. Thus we require for M > ^.B GeV, at 10 GeV and at 5 GeV for M > £ GeV. Thus to Bearch for relatively low mass AdE/dx -: N|dE/dx(Z=l, 0«l) - dE/dx(quark)| objects by TOF one reduces the machine energy. Cerenkov identification of low velocity co- linear pairs can te accomplished by setting the Cerenkov threshold well below that for protons of AdS/dx Hifr ,, energy Eo« This method is limited by the ineffi s dB/dxl quark "'jS " *! ciency of the C* counters and by the necessity to eliminate almoBt colinear events with enough miss where 6 is the quark veli 'ty, Z its charge, and ing energy so that a pair of conventional particles li the number of standard deviations required in is slow. Considering these difficulties and the the effect. Using AdE/dx = 0.20 dE/dx and K o 3 cost of large solid angle coverage with Cerenkov we obtain counters, the TO? method ij preferable. In the case of inclusive quark production 0 ^ <1.07)Z* both searcher for anomalous ionization and for apparent momenta greater than Eo/c would be effect ive. Time-of-flight would also oe effective in 2 searching for high mass particles if combined with 3 - 1 - 4 calorimetry to determine their energy. Unstable Quarks H « (1 - l-OTZ2)1^. Unstable quarke would, be difficult, if not impossible, to identify. One would look for co- linear hadron jets each of which carries the full If Z = 2/3 we have M e 0.'(2F,Q. Of course, above this value of M/E (P < 0.67) the velocity beam energy. The distribution of P^ within each 0 jet would have to be quite limited ( i 7 j £ 350 can be determined to an error of .02 with 100 ps ± of timing resolution over 5C cm flight path. HeV/c or so). To identify such a complete jet Thus by combining TOF and dE/dx the full kinematic requires kii detection of charged and neutral range in quark masses can be covered* hadrons. Such a detector horders on being pro hibitively complicated and expensive, and even if A second method for identifying fractional such a complete jet structure were seen it would charge is to use magnetic analysis to search for be difficult to prove that they were due to un particles with an apparent momentum greater than stable quarks rather than some dynamic:. Eo/c (due to their fractional charge reducing the bend angle in the magnetic field). If the quarks In all quark searches it is important to produced in pairs have the energy of the beam, realize that one should observe a threshold in the this method is sensitive for masses auch that signal .dth S. Such a threshold would confirm the quark mass and give an indication that the signal is not the result of a class of background which *o& itself doesn't have a threshold. Con 'lusion The requirement that both members of the pair have "supermomenturn" would eliminate most back' ground due to errors in momentum measurement (due Fractionally charged stable quarks can be to inefficiency + accidental in a chamber). detected by anomalous dE/dx (possibly combined The identification of colinear pairs with low velocity is useful as a signature for any 1 1 massive stable particle (including quarks of any 1 1 1 charge). The relevant velocity measurementa can be performed by TOF or using Cerenkov counters. The uncertainty in the mass using TOP is - a - AM£2M - '•H-7) E»l •GeV where d is the flight path, E the energy and At 'he time resolution. The mass is determined for colinear pairs by H = E jl-&* . Values of A t - 100 pa can, in principle, be obtained by 10 using segmented solid Cerenkov radiators coupled to photosensitive multichannel electron multiplier - 5 plates. Such plates are positioned — 1 cm behind a photo cathode in order to achieve this excellent time resolution. In Fig. III-l we depict AM 1 ^~ 1 1 l^~~~i versus M for At = 100 ps and d = 50 ca for E = 5 CeVf 10 GeV and 15 GeV. Since we eonsf-der colinear pairs, this energy is the beam energy. If we require AM £ 2 M so that we car tell a Fig. ixl-l Error in Calculated Mass vs. Mass for massive object of mass greater than 2 GeV from a Collineer pairs (at = 100 ps, d = 50 cm) -188- with velocity measurement), anomalous momentum monopoles discussed above (if the monopole mass is (greater than the beam energy) or by the obser below 5 GeV). Thus, one should search for mono- vation of slow massive colinear pairs using TQF pclcs indirectly by detection events where the or Cerenkov counters. Unstable quarks would be full center-of-mass energy appearB in a large quite difficult to identify in a model-independent number of uncorrelated photons. manner* Such a detector requires 4it detection of y's with adequate energy resolution and special reso 2. Magnetic Monopoles lution to reconstruct n0,1;. •*• so that each hadronic event can be eliminated from the sample The theoretical basis for the predicted of monopole candidates* The question of n° and r\ existence of magnetic monopoles is diecussed in reconstruction is discussed in PEP-153 where form PEP-174, In the unlikely situation that mono- ulas are given for the n° mass resolution in terms poles can be produced as free particles, they can of the position and energy resolution of ay de be detected from the Emf whita reoults when they tector. The resulting identification efficiency pass through a superconducting coil (a la. Alvarez for Hal and Pb-glass is given in PEP-154, Figs. and Eberhard). A second handle on their detection 9 and 10. If a monopole multigamma signal were is to search for extremely heavy ionization seen It would be important to study its S-depend- ( •• Uo GeV/gm). Unfortunately, a free monopole will ance in order to map the threshold behavior. The be so heavily ionizing that it will never penetrate energy-dependence of a multigamma signal would very far into a detector. The following table help to enhance its credibility and would allow gives the range in matter for monopoles produced at an estimate of the monopole mass. the full PEP energy. We conclude that the best monopole detector in the unlikely possibility that free monopoles Table III-l are produced is a superconducting coil around the beam pipes at a distance from an intersection Konopole Mass Monopolc Range In region containing a solenoid magnet. A useful (GeV) Aluminum Scintillation Steel lower limit on free monopole production at PEP 2 1.2 mm 2.3 mm 0.4 mm from proton experiments is 2 GeV. Indirect evi 5 0.9 mm 2.5 mm 0.3 mm dence of monopole production could come from events 10 0.5 mm 1.3 mm 0.15 mm observed in a 4K photon detector where all of the 15 0 0 0 center-o:'-mass energy appears BB uncorrelated photons. In addition to this massive energy loss, mono- 3. Tttchyona poles suffer acceleration to energies of 4 GeV/ Kg-m of magnetic field. Thus monopoles emitted at We describe briefly the properties of parti low kinetic energies will be accelerated along the cles hith velocity >c, called "Tachycns"!, from beams if solenoid magnets are used. The appro the Greek work Taxts, meaning swift. If we assume priate detector for free monopoleB is then an that the normal relativistic expressions for E and Alvarez coil wrapped around the beam pipe several p held for tachyons, then we must require that the meters downstream from a solenoid magnet. The mass m be an imaginary quaui.il/ in order that E and magnet acceleration gives the coil ft large ef p, which are measurable quantities, be real. This fective solid angle. is logically consistent, since the particles can The best limit on the monopole production not be brought to rest, thus m is not a measurable cross section in accelerator experiments is 1.4 x quantity. We only require that if P > 1, it always lO"*1 in 70 GeV proton collisions. At 70 GeV the remain so. Thus cross section for the production of virtual photons can be calculated as a function of the photon mass m = vu (n real) (see Appendix). Photons of mass greater than 4 GeV ar» needed to produce 2 GeV monopoles. This cross 2 2 , „ UC i , UV section is ~ 4 x 10"35 cm . For 5 GeV moncpoles r r and E = »JP ; P - \fr the corresponding cross section is —2 x 10-38 cm?. Q g When this is compared to the limit of 10-Ul cm2 we can say that the probability of a massive virtual photon making monopole pairs is less than 1/4 x where _ _v lp| as usual. 10-6 for 2 GeV mcnopoles and less than 2 x 10"3 for 5 GeV monopoles. If a typical PEP experiment represents an integrated luminosity of 1038 cm-2 _2 2 P ? 4 sec-1 and if am = 20 nb at PEP, then an experi Wc see Lhat B" - p c = -u c , ment consists of the production of 2 x 10° massive so that the four-momenta are always space-like. virtual phctons. Thus we conclude from the proton Further, when P = •», E = 0, p =• pc, so that as experiment that at most, —1 event will be seen if tachyons lose energy, they speed up to infinite the monopole mass is 2 GeV while the upper limit velocity. Beth energy and momentum vary smoothly for 5 GeV monopoles is 4000 (.'vents. with &x and at & = 1, E =o>, |p| =uc. Since the In any case., it is rather unlikely that free four-momentum is space-like, the sign of the energy monopoles will be produced because of the strong component can be changed by an ordinary Lorent2 attractive force between monopole pairB. This transformation. States with positive energy turn results in the emission of many photons from the into states of negative energy. However., the same radiative damping of the n^nopole pair, litis transformation changes the sign of a time-interval damping can be thought of phenomenolcgieally as the so that a tachyon traveling in one direction may reason for the low coupling of virtual photons to r.ppeer in another system X, moving such that -189- i\tif> J, ti. be an tritiLauhyon traveling in the It may be that only neutral tachyons exist.£ opposite direction.? In that case the only evidence would be through Tachyons are assumed to be spinles;?, however, tachyon exchange processes, for example in Bhabha Fcinbergl linds thut scalar tachyons fields are scattering, where a narrow sharp resonance would quantized using anticoaminieation relations, and be observed at fixed q** = ^ < 0 , independent of are therefore fermions and not bosons. beam energy. Finally, since r* > 1, tachyons can emit Cerenkov radiation in vacuum* The energy loss BEFEHEHCES per unit length is 1. G. Felnberg, Phys. Hev. 3^£, IO89 (19 V 2 2 ^. Heavy Leptons = p- u -jy , where n = 1 in vacuum and 2ft p~ Ze = tachyon charge. Heavy Lepton Signatures and Detection RequirementB Tt •- rate of energy loss in vacuum is enormous, and within a distance less than 10"2 cm, even high The characteristic signature of relatively energy tachyons (i.e., 1^ GeV) are "accelerated" short-lived heavy leptons is the presence of one to < 1 eVi Thus o.z '...ehyons emerge from the pro or more e* or u* in the final state. (Loiii-livt-d duction point, the energy is clo3e to zsrr . but heavy leptons are treeted elsewhere.) While V- — cc. Cerenkov lifiht emitted at this sti ge is hadronic decays may dominate for some heavy leptons, produced at -90° to the direction of motion. final states in which all of those produced decay Tachyons cannot be produced singly by ,'ormal to hadrons plus neutrinos are detectable only via particles! but can be produced in pairs. _"n thresholds. Other particles with substantial particular, they can be produced through one- purely-leptonic decays (such as W* if by soite photon annihilation at PEF, as lonfi as the to al chance they can be produced at PEP) may also be four -moment uir.-squa red of this pair is positive. considered here. There is no threshold, since the pair can be pro The following subsections detail the require duced at any energy, even Epflir = 0. K< wever, ments on acceptance, resolution, anu particle since |p,n| > E™, if tachyons ore produced with identification, using specific assumptions for the beam energy EQ each, their momentum will b^ production and decay of various particles. Ironi |Pm| -* Eo- In faot,j since t|:eir energy is reduced cally, the most serious background to the decay of to zero rapidly, the momentum of an emer^iiu; any one heavy lepton (aside from excited heavy tachyon will be the minimum value |p| = nc lentons, for which mass plots may be formed) could It" tacbytns interact normally with an electro- well be the decay of another new p&rticle. Scans niaenetic field (at they must, if they are produced versus machine energy - to search for thresholds - by photons), then they can be n:ade to reach a may be the only feasible way to handle such a D an problem if it exists. 3tcady state unergy %\ y electric field such that the energy lost through Cerenkov radiation is balanced by that gained in this field t. a) Excited Heavy Leptons. These may be pro ducedin~^aTrsTT*ET7r^sultIn"g in a final state e~? e 7 or n~7 M 7» F°r a particle coupling to 1 Ze the Jhoton, there will be ^10^ of these for an integrated luminosity < of lo3o cm-P sec-l^ Mith Hie Cerenkov radiation emitted in a field of 3kV/cm a rapid turnon (slightly faster turn ;-•) above is sufficient to be observable in an ordinary photo- threshold. If the decay width is sufficient, pro rnultiplier (spectral response 2.5 eV - >.^ eV). duction of pairs like L*t+ (resulting in one fewer This effect has been used to search for tachyons final state y) may also occur. produced by 7-rays from radioactive decays.3 If these occur at the rate indicated, they The same method may be used at PEP. However, thJ should be obvious as threshold effects in their intense flux of very soft X-rays near the bear, final states, and from -t? mass peaks. (If one 7 pipes precludes placing a small detector immediately goes undetected in an -t -t ?? final state, there surrounding the interaction point. A larjje detector is still enough information to reconstruct the s-me meters away from the bean; pipe would be event.) He require good n, e, and 7 identification, impracticable becnuse of Door light collection with y's over as close to hn as possible. The presence or absence of hadrons and extra ?'s should (ec = yo°). It may be possible to time the flight 'fa also be known over %, but determinations c hadron tachyon through limited-aperture Cerenkov detect momentum, type and charge are not necessary. Energy resolutions of ~10 - 15^ are probably adequate; ors. However, Feinberf: states that a t.nchyon 0 wave-function cannot be localized, and doubts that angular resolutions of 3 - b will then give com the concept of a measurable "velocity" can be parable or smaller contributions to mass resoiutians applied. above 5 GeV. (If the.<»* L*"'s exist at much lower It in not obvious whether tachyons can ionize masses, they should aim 1st certainly he seen at atoms when ('-=•>, since the time spent in the SPEAR and DORIS.) vicinity of any atom approaches zero. Thus zero Requirements on hadron misidentification as energy tachyons (P~™) may not show trucks in u's or e's are less stringent than for other types scintillators or track chambers. of heavy leptons. Even if all 2-prong tisdronic events ( < lOjt of total, according to Monte Carlo predictions*) were candidatea for misidentifi- cation, one could have mlsidentification probabil ities of .1 v/c/'cny ( **•! for pointlike production) and still have < 1$ backgrounds without even using the mass-peak or 7•counting information* Ordinary QED backgrounds can be made negligible ty requiring acoplanar leptons and staying away from 0°. b) Chased Heavy Leptons (and W~) in Purely Leptonic Decay Modes. We consider pair-produced charged heavy leptons which both decay by L-*\ signal of .05 ou„ for pointlike production. Thus >. < .15 should allow clean identification of such processes. However, in order to study the nature Fig. III-2 Lepton Distribution from Heavy Lepton of the process (and allowing an extra factor of 10 Decay (M Rest Frame) in case of suppressed production), one should aim for 7\ & .01. These limits should really be momentum- where P and p are the L and •£• At-momenta, and E ia dependent; this will be discussed in the section on the -t'a energy. For 15 GeV beams tnd three differ backgrounds. However, the numbers given are prob ent values of H, the lab momentum spectra J ussuroing ably very conservative (Perl^ picked out a signal pair production) are shown in Fig. III-3/5. The with ?\ a* »2 and far from ^JI coverage on 7's). S-body decay of a pair-produced object givea rise Kith 5°° eventa, we can begin to measure some of the properties of the new particles, such 0.14 1 as type, masB, whether there is more than one, etc. „ 1 1 1 1 1 1 Three handles on such properties are s-dependenoe, ~N lepton momentum spectra, and lepton-lepton angular N M = 2 GeV 0.12 correlations. These methods are complementary, and E0=l5GeV all should be used. \ V-4 Consider the lepton momentum spectrum. Figure 0.10 III-2 shows a rest frame de^ay distribution for V+A L-.-oi('£-v ), assuming massless neutrinos. The two \ J curves represent the standard (unlike) V-A L*^ , 0.08 coupling and a V+A coupling suggested as a possi bility in certain gauge theory models.3 The V+A \s. case corresponds to a Michel parameter P = 0, as 1 0.06 opposed to the usual p = j/k. A 2-body -C-v decay \ \ of a W- or some other weakly-decaying boson of \ \ mass H would shew as a 6 function at M/2 on this 0.01 - N X - plot* \ X \ X. In an arbitrary frame, normalized decay 0.02 distributions for the 3-body decays are \ \^ - 1 ! 1 2 2 'ZF E • P . pjM + 2(P - p) ) V-A 0 2 4 6 8 10 12 14 eor/it;:?") w.„ Fig. III-3 Lepton Distribution from Heavy Lepton 2 p • Ptf" pf Decay (Lab Frame) -M™ <* P GeV/c , -£* Eoeam - 15 CeV r • I i i i i i 0.20 ; v E0 -15 GeV t \ M =13 GeV 0.18 ~ V-A V+A o.ie II \ i I \\ 0.14 ^ fI \ \\ \ 0.12 i \ Id :, \ \ p \\ \\ 50.10 1 \ \ l \ \ 0.08 \ \ 1 \\ \\ 0.06 \ \ \\ \\ 0.04 \ \ \ \ 0.02 fl 0 i.i, NA 2 4 6 8 10 12 ?. 4 6 8 10 12 e Of /i EIGBV) tor fi E(GeV) Lepton Distribution frcw. Heavy Lepton Fly. III-^ Lepton Distributl?n froc. Heavy Lepton Decay (Lab Frwi*)* l'.^ = 11 GeV/C, Decay (Lob Promt) - K,, = 1; GeV/'c^, to a flat spectrum between K-dupendent kinemttlc iit) ? and hadron detection over virtually litr.its. The distributions plotted were obtained by '•a. assuming '•n angular acceptance, and are thus Independent of the L*L~ production distribution. (iii) e and p .noi&entun: resolution (ip/p?) rifts (The latter depends on U.e specific nature i,f the ^ .71$ (and preferably ip/p < .0>j tit new particles - it is p-likc- for "ordinary" L's - 10 GeV, at least for electrons); charges. an? also on transverse bea:n polarizations.) These (iv) y position and enerej resolution not spectra sUouid be taJ.^n only as indications of the important. types of things a detector should be expected to Hurt out, and not as detailed predictions. For fv) Hadron type and kinematic neasurencnts cxaitple, with M " 11 fieV, the V-fl and V+A coses necessary only crudely (or frois other lead to -- UO - 100 andl >0 - :J0 events above 6 GeV. cxptt.?)ln order to help understand With these statistics, mosnentum resolution ip/p^n backgrounds. .00"/> can probably distinguish then; adequately. !(• wever, it is clearly c.ocnta. For e's, a shower high p^Centa. detector can provide an excellent # res at 10 GeV (vlll l.nt'r.iae energy scans < if a signal (but a pjicnct ™J8t still give the sign). easts at high S). Angular correlation* between the decay leptons cun probably provide even pore information than Mjfccntuis spectru, if biases are avoided. (Kepsnd- e) O'arKtl Heavy Lfcptcns wlt.it Hadron Decays* ence on bean polarlzntlcns ran also be used here, For pair production of pointllhc heavy leptons,, «<-• but not In distributions of sinfilc-L decays, oir.ee can detect 22000 events (per 10* MH) if KC require the averse palarir.atirn of ouch L Is tfero. only one purely leptor.ic decay (plus one defray t< v • lisdronf)• If leptonic branching ratios are K-4ch esnller than assu&cd this aifdit be the only fe^;:«ie detection tethod. (Certain typeo of gnufe- To auer.nri£< tlie re'|u*rotwnla: lh*«r'}' J-.thvy Xoptons cay decay in states containing (I) o rutd )i detfeiit-n ever ** ion:*,- a K #v G- (rather titiui flv) -* lindi-ons. Tttcse would r- •"iselbl--. for mtr wtd *•& n* be soeiMrttBt trnsler to detect inaacucli as final btafirs. *t*tt*-e weuld cftcr. contain iw t's.) -192- The requirements are »inilw to thofe in d) Heuti-al Heavy Leptons (and Heavy Meutrinoa) Section (b), with the following addition!: Using the 197^* Summer Study's weaK production estimates for single L°v) and pair production modes (i) Information on the identities and pro and decay branching ratios similar to those ssauned perties of hadront nay be useful; if we for charged leptons, one can expect £50 and 520 see the aingle-lepton signal, we will purely leptonic events in these two modes, if these want to know what else Is in the final L°'s are of a type which couple to e's or ji*s. The state. signatures are then 2 (acoplanar) or k charged All hadronic events are now potential ordinary -t'a, respectively, with no other visible .<"> backgrounds. If we restrict our particles. The requirements including those on attention to momenta >20J& of the beam hadron misidentification needed to study charged energy, one expects^ ~1 charged hadroc/ heavy leptons (P'-& subsection b) should be adequate event, so that we need a hadron mis- to detect these decays. Because of the low rates, identification probability backgrounds from decays of other new particles are .So potentially a major difficulty. .002 For IrL° production, one can increase the **$!=&- sample considerably by measuring events in which for 10:1 signal noise* These back either or both L°'s decay to ji or e + hadrons (if grounds must really be considered vs. we can adequately handle the likely low-energy momentum. For convenience in back particles), and events in which one L° decays ground comparisons. Fig. III-6 shows leptonic&lly and the other to v + hadrons. All of the conventional (V-A) decay spectra of these lead to final states with 2 -t's. Fig3. III-3/5 for all energies SE vs. Certain other gauge-theory neutral leptons^ E, normalized to .20^ times an appro will either decay to charged heavy leptons, or, if priate threshold factor. (For heavy they're too light, may decay undetectably or not leptons decaying to e or (i T hadrons, at all. a ^ a* .DO1* would allow us to study final state properties of a final stste produced at the -CSo^ level.) (iii) The signature of missing energy and G« Hanson, private communication. momentum has been suggested, but the M. L. Perl et al., to be published; M. L. Perl, level of calorimetry needed to utilize this may not be feasible. SLAC-PUB 1592 T1975). A. de Rujula, H. Georgi, and S. L. Glashow, to be published. 1 1 1 1 J. D. Jackson, in Elementary particle Physics : and Field Theory, Brandeis Summer Institute, = 15 GeV : : E0 1962, W. A. Benjamin (1963). M 2 GeV • M = II GeV 13 GeV p. Stable Charged Particles 10"' r We defined heavy stable changed particles as having the following characteristics: . 1) T > UQ-° sec , decay outBide \ -. \ apparatus 2) M > 5 GeV. If they are lighte:-, \ \ \ they would have been disccered else b* 10"' •: where 3) Charge of ±e, fractional charge will be * *• \ uidcussed under the quarks section. - \ ". \ Without the additional signature of the fractional * V \ charge, the cross-section limits at proton accel * *. 1 erator and storage rings in o£ IO-" cm2.1 Hence I 1 •. '. \ t though RIAL and the ISR have energies comparable V • -._ to PEP, the proper combination of low production - i : crosB sections and hadronic backgrounds might mean i : that it is still possible that new heavy stable \ 1 particles would be discovered at PEP. The most sensitive region for searching for heavy stable particles is where their velocity is • less than 0.5. In this region one can use a 1 1 It combination of TQF, range and simple lucite derenkov 10-4 counters for identification. Particles with TQF 3 6 9 12 15 differences greater than 1 nsec should be almost EtGeV) completely free of random background due to ~5 cm bunch length at PEP. Fig* III-6 Cumulative Muons or Electrons at The mass resolution that one can obtain with p 2: E from Ifj -»•£• u 5 separate momentum and TQF measurements assuming -193- ncgltgible error in the momentum is pretence could be discovered by looking at the conversion depth distribution in a hadrop calori meter. An excess converting at larger than 5 d s neutron interaction lengths in tine with the bean >/T-a crossing would be an Indication of such a new neutral particle. More refined calorlmetry com where d is the flight path and c the velocity of bined with TOF would la principle yield a mass light* If with aiinro-channel detectors one can determination. achieve time resolution of 0.1 naec, then over a flight path of 0.5 a for a 10 GeV particle with 7. Charmed Particles p = 5 GeV/e, c,M » .3U GeV/c?. More discrimination can be made by range IncJ jive pair productions is the most likely measurements. For particles with & = 0.5, the 2 source of charmed particles at PET. Present esti range ia —(MH/MpJZ^gm/cm , where Mh la the mass mates of charmed particle masses are 2 GeV £ Mc& of the new heavy particle relative to the proton 10 GeV with the lower limit in some trouble due to mass, Mp. If \.he new particle has no strong inter the nonohservance of charmed particles at SPEAR. actions or has an interaction length more than Assuming the conventional decays into strange three times as loug as the proton then there should particles, the clearest signature involves strange be distinct range which can be seen in a segmented particles associated with a single charged Jjepton hadron calorimeter. Such a range measurement would (one charmed particle decaying seml-leptonically, also be a good cross check on the mass determin the other decaying into hadrons). However, it is ation made by TOF. possible to construct models in which the strange A simple lucite Cerenkov counter with a particle decay is supressed. If this is indeed the threshold 0 - 0.7 would serve to reduce backgrounds case one ia forced to look for peaks in Invariant from ordinary hadrons. mass spectra associated with the other peaks or Jf the new particles have a lifetime in the single leptonB. range of 10-& to 10-6 seconds, the observation of Taking cross-sections and multiplicities from of a delayed coincidence due to a decay would be PEP-1^0 we expect ~7 x 105 hadrcnig events (B « 6) another extremely powerful means of identification for an integrated luminosity of lo38 ,^-S ^^j, a of a new state. mean charged multiplicity of ~7«5 charged paxtt- It is most probable that these new heavy cles/event. This gives ~7.U x 10° neutral 2- particles have quantum numbers such that they must body mass combinations. Monte Carlo calculations be produced in pairs. Then in order to find these give a combinatorial background of .756 (phase particles if their production cross-section yields space model) and .5% (Jet model) of the total a few events In ay experiment with an Integrated combinations in a 50 HeV bin at M = 3 GeV. Thus luminosity of 10^ cnP, one must have a rejection we expect «5x 101* background events/50 MeV at a probability of ~3 X 103 on each particle which mass of 3 GeV. ThlB large background makes it seems reasonable with the above techniques. To unlikely that bumps in raw mass plots will be achieve lo7 rejection on single partieleB in a observed. nearly it* geometry is not nearly so certain because the possibility of an unlikely combination of mis- If we look at invariant masses associated measure&ent of the raomentum and Ttff. However, the with 3ingle charged leptone it la possible to presence of 5 to 10 events in e single 1 GeV mass supress the 'background by ~lo3 over a large solid bin would probably be identifiable above back angle (I0_l*it/e rejectionc seem to be possible). ground. Estimates of T(c —tu + hadrons)/r(c - hadrona) range from ~1036 to ~5°£» For a seml-leptonic branching ratio of 2^f> we expect ~£0j6 of the events REFERENCES containing charmed particles to have single e± signatures. If we assume inclusive charm production 1. L. B. Leysuner et al., Phys. Rev. Lett. 31, » inclusive baryon production {~100O events/ 1226 (1973). IOJV cm"2) -we can get a signal-to-noise ratio of M. 0. Albrow et al., submitted to Hucl. Phys. ~1/1 st- & mass of 3 GeV with a mass resolution of B, and A* Seasons, private conmunlcations. AH * ± 50 MeV. G. fliacomelli et al., PhyB. Repts. IgC, 223 Thus we need good mass resolution AH — «01M (1975). and good hadron/lepton rejection —lO-1* If charmed particles •*• to be found at PET' with l/l signal 6. Stable Heutral Particles noise ratios in 2-body mass distributions. Similar conclusions hold for decays to higher Long-lived neutral particles are very diffi multiplicity final states. cult to identify. If they only interact weakly (heavy leptons), their presence can only be dedu 8. Scalers ced by the presence of missing transverse momentum. If calorimeters could be placed within ~10 or of The Biggs scalar meson1 is suggested In gauge the beam as for the 2y tngging detector, then theories. It may be found In e+e~ collisions, In missing transverse momentum of & 1 to £ GeV could pairs via one-photon electromagnetic interactions, be seen. The absence of electrons or muons or singly via Sy or weak interactions. In either accompanying the event would indicate that the case, the rates are expected to be lew, of the misBing transverse momentum was unlikely to be order of »010uu . The decay will be electro the result of ordinary neutrinos. In all prob magnetic, so tne width ought to be small* Bxperi- ability such particles are more likely to be seen atntally the signatures would be: (1) a sharp In neutrino experiments. peak In the total cross-section ai> a function of If the long-lived neutral particle had an center-of-maaa energy, but on top of a substantial anomalously low hadronlc cross-section then its background (expect for the longitudinally-polarised beams case, where the ordinary electromagnetic Many of the above states may also be produced processes are suprcsaed); and (2) a peak In the with other particles, for example missing-mass spectrum for the "2y events e+e"-« 1 e+e- + X. a) gluons \*^J l "f^" I gluons and REFERENCE +/ V_-' -...7 \ hadrons 5 1. P. Higgs, Phys. Rev. Lett. 13, 5°8 U96U), and b) heavy photons, vector mesons (in 2y geometry): 9. Neutral Vector Statea We classify neutral jPc • 1" states with the following notation;1 a) high mass vector mesons (p,ll,V111,) which have small coupling to e+e"(u"V~); e e b) heavy photons (B°), having a large where e* ip an excited electron. In these cases coupling to e+e~j the resolution of the detector may well be the c) neutral currents, Z° 2; and limiting factor in observing these states. d) vector gluons.3 The states (exclusive of SU(3) singlet gluons) couple directly to e+e~/u+u"Aiq", and can potenti tot ally appear as an enhancement in oee or oee *pn at the resonance masa, S = M2. Local variation^ For a general description see D. Berley et al. in the charged/neutral ratio, baryonic or strange 197^ PEP Summer Study, p. U$0 {197*0. particle yields, average momentum , and jet S. Weinberg, Phys. Rev. Lett. 19, 1261). (1967)5 structure in the hadronic decays provide sif?natures A. Salam, in Elementary Particle Theory, for those states with significant hadron widths. edited by N. Svarthohm(1960); 0". D. fijorken, The heavy photons, coupling strongly to e+e~, C. H. Llewellyn Smith, Phys. Rev. D7, 887 n+u"» produce deviations from QEE in e+e* -»e+e-and (1973). fc''»~ -» u+u" reactions* Neutral current effects are C. H. Uewellyn Smith, Phys. Rev. pjt, 2392 probably best studied looking for j - Z° inter (1971); H. Fritzsch, M. Cell-Mann, H. Leut- ference with or without longitudinally polarized weyler, Phys. Lett. Jt7B, 365 (1973); M. S. beams.^ Possible particle signatures and cross- Chanovitz, S. D. Drell, Phys. Rev. Lett. 30, sections are summarized in Table III-?. 807 (1973). TABLE III-S POSSIBLE SIGNATURES OF NEUTRAL VECTOR STATES CHARACTERISTIC VECTOR PHCPERTY GLUONS LEFTCHIC DECAYS small if SU(3) •inglet 1 HADRCNIC DECAYS T0T 1 /LOCAL CHANGES IK a* , \ . - , small k.x^ [ R, CHARGED/NEUTRAL RATIO,] i HADRCKS KR \ , «te. / JET STRUCTURE OF Prob. different j Prob. different HADHCWIC DECAYS ! from off i from off j resonance j resonance EFFECTS QED TI9TS LONGITUDINAL POLARIZATION EFFECTS - 1 (deptmdn on \°ee --• UM/ total width) h, C. K* Llewellyn Smith, D. V. Nanopoulus, Kucl. < , k^E /. Sao,*!3 1, ("j) Phys. B78, 205 (197*0; Report of Longitudinal signal - P , " =• I« srrr ' •=-to 1 =-=-/• Polarization Group, PEP 1975 Summer Study. ? imin 2 E n v/(E-E1)(E-Es)\ > "V 10. 2y Production • 1038 on2 . t a in In the search for pseudoscalar and scalar c = + 1 objects which can be generated by the 2- has been used to calculate the signal photon production process one important signature is a state consisting only of a tagged e+e~and a OI (w = 2 /(E-E )(E-E ) ) . neutral boson decaying into y's, JI S, or T,*a: V I 2 U. Search for New Particle States at B Sinple Beam Energy (coapton Production) *0 The radiation of a hard photon by an electron or positron before annhilation into hadron* provide >* 7Ji°n a method for searching for JPC «= 1 resonances over a large mass range et a sinsle beam energy. The mass range accessible at PEP for observing a The process considered is shown kinematically significant effect in 2y final states with the below: w proposed 2y tagging systems (see y-f group report) is Bo .f-S GeV< N < 5.0 GeV at E * 15 GeV e -#- In setting this limit a significant effect was defined to be at least 25 events signal in a •£e+e~ = 10+39 cm"2 run and a signal-to-noise ratio The corresponding Feynman diagram is: greater-than or equal-to 1. These resonances are .& likely to be of the order of 10 KeV in width. In the calculation of the observable cross sections it is estimated that the measuring apparatus dis cussed at the Sumrcev Study would be capable of + "-'*-f Hx better than 1 GeV resolution in invariant mass ^f e tf X >- the photon pair over this mass range. This process is clearly reduced by a factor & The backgrounds for such resonance production relative to the normal single-photon annihilation.^- would arise from the 300 mb 2-photon cross- However, the annihilation cross-section to be con section (as opposed to the much smaller single sidered. Is at the reduced value of S* = M| and photon cross-section) from two sources: hence considerably larger than at S >= 900 GeV9. The spectrum of hard photons recoiling against 1. Events in which the 2-photon procea.T the mess Mx should show corresponding structure. gave rise to a number of 7t°'s, two of The kinematics are shown below: which were identified. 2. Events in which one photon from each of two n°'s is lost, giving rise to a false ; £' 3XE 2-photon signature. O It has been assumed that the mass spectra li E arising from such false signatures would be uni <• c form and that the dominant process would be 2 because of the expected momentum spectrum of the n°'s find the angular resolution for resolving = c separated showers in the liquid-argon type detector. With these assumptions the ratio of signal-to- *S„ background behaves as shown in the following table; = Table III -3 Thus X = K^ /S^ . Mass (GeV/c?) Signal Background Ratio The photon energy spectrum is .7 775 33 2J.5 1.0 905 78 11.6 2.0 267 92 2.9 E? - » - E„ • li.O 52 72 .72 5.0 25 53 M 6.0 13 lio .33 - U-X)E„ •H> where (or pseudoscalar or scalar resonances The limits ere: » dw = ^- (2J*l)r (r ~ 10 keV, J"0) - 0 when M - 2E x o and the approximation • E when H • 0 , i.e., a photon. Unfortunately, the range 0 < Mj-c 9.5 GeV is and improves rapidly with M* a° that prominent high all compressed into the region 15 GeV < L < 13.5 mass resonances around 10 GeV/ca will stand out. GeV, ."or E = 15 GeV. However, even though the The only background giving i.igh energy tagging resolution is mass for any realistic photon energy- photons is e+e- -»JI°XJ since Bhabha scattering is resolution is likely to be terrible, the advantages eliminated by requiring a neutral "tag" anil a ol a crude mass-search at a single beam energy are hadron signature in an apparatus like the 2y aya- considerable. tem.3 if we consider the n° inclusive spectrum „Tlie cross-section is expected to be of the scaled from SPEAR, we expect 60 events/lO-» cm~£ forrr.c at p o > 13.5 GeV/c. This corresponds to an aver do age croaa-sectlpn in the range 13.5 GeV - 15*0 GeV of ~ 0.1* x 10"-*" cm /GeV, and is presumably iso tropic. where G(x) is the probability of finding an The problem remains of tagging at 0°. If the e.\e-L.'on or positron of fractional energy x, and smallest angle possible is 10 mr, as the Sy case, o(S') is the normal hadronic cross-section at tht gain due to-fcn (So/hf) is lost. However, S = S' =XS0 . enough sensitivity is probably retained in looking for priminent resonances such as the Y, 1', etc. Integrals over photon angular ecceptancea need to G(X) • be evaluated before the senaltivity can be estab •fMfc)-*J lished. 1. The reduction is not cf as in PEP-17'', p- 17. 2. Stan Brodsky, private communication. 3. 2? Process Report, PEP-175. h. See PEP-l*i6 in 1971* PEP Jummer Study. i 2 The tn(S0/Me ) term comes from the integration over vcy small angles of emission of the photon. Assuming for the moment that these photons can be IV. HADROHIC BACKGROUND TO LEPT0N SIGNALS detected down to Qy —> 0, we can evaluate the loga ? rithm at 80 = 900 GeV . 1. Introduction 1 The rejection needed for lepton experiments 0.051 at PEP depends on exactly which type of new parti H^- } cle one is hunting. For pair production in which both leptons are observed, signals >.01 0 might and expressing be expected. For single lepton experiments the level might be £.1 o^. In both cases one is dealing with lepton momentum spectra that fall hadronie (S*) = R(S')o = R(S') • '(GeV2) rapidly with momentum. Furthermore to determine the mass of such leptons one wants to run near get threshold where the signal is suprecsed by an addicional factor. Also signals from the leptonic decays of new hadrons could be at a lower product ion cross-section with nomentum spectra that peak at lower momenta due to the multi-hadron final state in a semileptonic decay. Reduction of back 2 ground" to ~10~ oUil seems to "r.e a reasonable goal for the single lepton experiments. This implies that a background rejection against pions and kaons ~lo3 la required. Figure IV-3 shows this background level at s • 2-00 and the integrated lepton yield from the decay of either of a pair of If we take the example of the ¥(3100), where m • 11 GeV heavy leptons. For the two-lepton + E 2X00, unobserved v'n signature a considerably lower re jection ~1£T/K should be sufficient. do . . - , —- , g. Hieidentification of Kadrona as Electrons =-" 3 nb/GeV at the peak. Rejection against charged pions can be achieved using the following techniques: This, of course, assumes a good mass resolution as in the original V experiment. With a resolution of 0) p • E (Hcirintiun of psrtlcie - Energy of about 500 MeV/c the effective value of R is particle deposited into a total absorption reduced to about 2 and thus the ¥ would be lost in electromagnetic shower counter.) the background. The mass resolution is given by b) Shower growth - pion showers {hadronic cascades) in shower counters have differ ^x ^o" Aw ent radial and longitudinal distributions than electron showers* Hx " c) Cerenkov - piona below threshold (—5 GeV) IF do not give Cerenkov light in atmospheric 197- preasure Up counters (n - 1) *• j X lO"4. d) Transition radiation - pious (jive much less transition JC-ruy yield than elec trons. The last technique hur been developed to the point where a rejection oi io can he achieved in a ^0 cm long module.1 These modules means that if pn > \>,b CeV they will be B background. Belcw this momentum, it/e rejections of ">7 < 10- have been achieved in experiments.-1 b) Showers and 6 rays (rr.ade in a solenoid coil) allow below-threnhold it's and K's to be counted os t-'s. A highly-segmented cnunter can sif.nit icanlly reduce the chower problem. These are sufficiently annuyirii' ior u;; to not con sider then, any lurt.her. A combination of p » E and 3l.ower growth criteria appears tn tie the best way to achieve the necessary bficK<'X<>uiiii rejection. Neither technique is sufficient in it:;ell. A sln.ple p = E require ment can only t'.ive u rejection ol 10' against charged pions. Shower tjrowth criteria can have B rejection cf 10- 'J(>uin2t n's having p = E. This Segmental Liquid rejection power din,inisi.es for pirn.1? <•! p > £. Argon Co(ofiiti«tii There is an additional background cnti.iijg from u charged |.[rticle entering the shower counter in t Ft*'. IV-1 Three Types of Electron Identiii. at lm the snp.e region us an entrgeiir f.:inp:;i ray. Requir ing p « L" mid dividing ti;e counter iutt: small solid tingle set^r^ntr cun eliminate this background. Several iliflerent combination!: el' counters can be Therefore for the function r-f electron identi used tc give this re.ieei ior,. fcr example, see fication the best systen. appears to be > if propon Fig. 1V-1, ent:- are right) t lend-li'juid argon calrriir* ter. To completely contain the shower it length if 1^ - a) A syslen rcr^isting .H * radintion lengths IB radiation lengthr would be required I -,",) en). of lend g.".ars in one block Jellied by Vj radiation It should be segmented into a least two but pre lengths In 'i second i.lock achieved si r«,;'?ctior. of ferably more lonEitudinai segments. The eriteria £lil." on p = E plus .ihower growth -ijteria.J This for seftientation in the transverse direct ieii are waa done without a pas-'w -•'"..*eiter (solenoid normally dicteted by the spatial resolution required e^il) in front, i The divUicn c) A system cons is tiny of segmented lead- several GeV/c. PL would pessimistically be - D X liquid urgen ahnwer counter is potentially the mixinum transverse modentum of a trapped particle cost promising:. It seems reasonable that appro ( — .3 GeV/c). (Trapped particles are those whose priate shower growth criteria would ^Ive rejections radius of curvature is less than 1/2 the solenoid octter or equal to those in the systems studied radius.) For pe2 > pr, we observe two oppositely above. However, there doesn't seem tc be any charged electrons with 0° opening angle. If they available data yet. rMrtheraore, the coot of leod- come from a conversion they will verticize away Uquld argon shower counter* has been estimated to from the interaction region. be 1/3 - V? and the overall thickness 1/? - 3/h To estimate this background we can do a cal of lead-glass ayrietm, culation at 90° to the axis of the solenoid. The probability of a n° of energy pQ giving a back ground pair after assuming a conversion or Dalitz decay ia: J dx - 2.1 X 10 (1 - x) (x<.5) Jf o_o = ° i ** 0«'a charged {a pessimistic Prob, n assumption), and we compere the ratio oi tl-.' ~/ Vr '*"*,) Oalitz background cross-section to o at s = »in(P jP ) 2 w e 0 900 CeV , min(P+P ,P ) °P,"3,P then — - = .012 The amount of background froir conversions PL *»»!.( P P ,P„) 1+ e depends on the thickness ^f the vacuum pipe. P 200u stainless pipe is .011 radiation lengths* In this case the conversion background from the pipe This function is shown in the ?lg. IV-?. The is roughly equal to the Dalitz background. fact that p. has a lew value results in a strong The above conversion background can be almoBt suppression of this background. The value of p. = entirely eliminated if the following configuration 2 .5 is equivalent to pL = .3 at an angle of 37° • is used; As one ",oes to decays that are further forward the 1. Place a MPWC or drift chamber immediately background becomes worse for a fixed minimum de tectable transverse momentum* This suggests the after the bean pipe. use of additional electron detectors in the sole 2. Follow it with a highly segmented scintil noid to cover small angles* It also says that the lation counter which is used to measure background is liable to be lowest near 90°« dE/dx. If we assume that n° production is given by Then we require the presence of a signal in the chamber and that the pulse height in the scintillator be < 1*5 minimum ionising. Thus a S |S . 1.5 * 10" -7.4X"n, b - GeV (x < .5) 10°F 1 1 1 1 - io dCTctiatgt dp r - "f dp 0>j^ - e"s from M= II Gev Lepion " ~—~*tf> - a Dalitz Decay) for Ee 2 E -i*w- conversion in the pipe °r chamber is rejected for that both are oUuined by 5C«tl.u-g the inclusive pulse height and the conversion in the scintillator charged hadron distrit dtion3 Iroc. ilPEAR energies. is rejected because there ie no track. Real tingle A parametrization good to K10* fn X =•• E/E ^p to electrons pass both cuts except for the CU that :s.U (£ e bear, energyyis then Btart to shower or give a i ray In tht pipe or chamber. Even tl' the finely divided counter is too much of a ltx.iry i'or a given experiment the chamber should be retained lor i •j'-etinn of conversions after t.ie bsair pipe. 1'tr n',1 and K's separately (but :; m-fin,- u u Thus the Dalit/ and r> nvei::lrn lackgreund can charges)** Mr St r.odels dc, not pivile es i i,i. n be k»pt below the level of the pion background fraction cf K's, so the thern.odyi.iirie I.E. -ij.-pti- i. assuming a rejection of l.1-1. probably represent;; a wrest cine. Tills Is shown in Flf. IV-,- •.;!,-re we assure If momentjm ir.eacar'-n.ento an- VM'H nnly leltre that the conversion background la i-qu-il to the decays, K'S contrib-ite '*.{ tin.es i>S i:>n;; J.'E US UC Kill*;, decay background and that the kaon hack- a'c at given apparent E (ineludiji,. Isn-tcis tor ground rejection (chatted K's sin.uluting o's) is decay rates at the sane E, = E, mid tt.e Y. • (*- no grouter than the pion backcriand rejection* brnnching ratio). Actually, t-J* i. cc unit . also veil below this level If sc-iling holds. fil GeV or so will range out. At Ec = 1^. GeV, cue obtains (per rater of decay path J j 3. Klaldentiflcntion ol Hudrcwn ai; Kuons A hadron can be misiicntifit-d a-3 a muon by two mechanisms: llBdron decay (primarily the two- body decay modes of charged n's and K's), and 1 meter o hadron simulation of a mucin signature. He will consider only energies -_• GeV, and assume that The fraction of mlsidentifjed hadrrns (n'r •* K's) is tlms .00^/x at E = 1^ GeV. This is i« rather muon-hiidron discrimination is U he done in. a 0 layered absorber + detector arrangement. For large 1,7^ at j GeV (but more lifc. l.'.r if wr example, a muen signature might be a ninititum- exclude ranged-out u's), .J' at 'W,- n«V, «nd .& at lonizing signal in each of a number of scintillator IP GeV. planes. Kuons of Z2 GeV should reaaiii mlnlmus.- On the othEr hand, if the ^ it^-r«nts are r..ea- ionlzing through a typical ab^t rl>er of 1 refer of sured, the K contribution at a fiven EM > d.<- GeV iron. Crudely, the signature r.i a identification is 2.1 times the n contribution (since K's spread (punch-through) probability is e~n, where n is the their u's over a wider energy range tl,;_-j dr si's); number of inelastic interaction lengths for a at lower E^, the ratio beconies sotiLevhat l«rr_er, typical hadron (i.e., a. pion). He should choose n because the highest-energj' n'z can no longer \ N E :- *-.\ *s : : \\ \ : \\ \ * \° \ 10-2 r \ \ \ \ i '•., \ ^ \ 10-3 \ v * i i i i i 0 0.2 0.4 0.6 0.8 1.0 0 3 6 9 12 15 x=E„/E0 E (GeV) Fig. IV-1* Muon tomenturaSpectr a from Meson Decays Fig. IV-5 Kucn Backgrounds and Typical Signal (1 Keter Distance - See text for for Eu 2 E assumptions) single-muon (+ hadrons) signature. One way co do will decay to the other; and we also substantially better is to roughly remeasure the ji momentum using increase the number of events this vay. magnetized iron. Another way is to shorten the We cannot conclude that for processes leading decay path by beginning u-hadron discrimination as to two charged leptons, it should flot be difficult close to the interaction region as possible. (See to obtain the necessary rejection against misidenti the section on detectors for suggested methods. fied hadrons for either e's or u's (above = 2 - This has the disadvantage that the resolution 2-1/2 GeV). For the detection of single leptons needed for stulyinc less difficult signatures may (especially those resulting frcm deca,v: of new not be achievable.) hadrons), the higher rejection obtainable from In order that the punch-through background be electrons may prove importart, and can be crucial no greater than the decay background up to about if the region below 3 GeV is critical. It should 11-12 GeV, this requires e~" £, .002, or n =s 6. be emphasized that the muon background curves are If this is the number of pion inelastic interaction based on pessimistic K to n ratios, and that the lengths, — 1.35 meters of iron is needed. Punch- electron curves may be technologically pessimistic. through of 10"3 has been indicated in Fig. IV-5« (The latter provides a cushion against hadronic (If one believes that for any of the reasons given yields much higher than anticipated.} above, the decay background Hill in fact be lower, then one should probably design n = 7.) REFERENCES Finally it should be noted that for beam energies E0 < 15 GeV the decay backgrounds relative 1. J. Fisher et al., BKL 20063 (submitted to HIM). 1 to Oup will be larger by a factor 15/E0 at any 2. F. Busser et al., PI. b3B (197 *), PL 56B, 212 fixed X. (However, the lower momenta will make an (1975). approach to the analog of the solid curve in Fig. 3. B. Blumfeld et al., NIK 9_7, U£7 (1971). IV-5 somewhat easier.) This may be significant if F. Busser et al., PL ^8B, 371, 377 (197M- it is necessary to search for the threshold of a *+. L. DiLella, private communication. low-mass state. V. THE NEW PARTICLE DETECTCR FACILITY h. Conclusions 1. New Particles Philosophy One clearly wants to design a new particles experiment to detect both u,'s and e's. We cannot The prediction of properties of undiscovered be sure that every particle which decays to one particles is a risky business. However, there are features of moat of the commonly discussed new return yokes and as passive hadron absorbers. High particles on. which & search for new particles can resolution time-of-flight detectors based on prox concentrate. The most common In the decay of new imity-focused multichannel plate phototubes and particles Is the presence of leptons. Neutrinos dE/dx counters surround the beam pipe and magnetic con be detected only by kinematical effects caused field volume. A superconducting coil of the type by their absence in the measurement of the final used by Alvarez for inonspole searches is located state - ucoplanarity of lepton pairs, large mrmentum- 'iround the beam pipe at a position ~Q m away from momeutum loss transverse to the beam in multi- the intersection point. In addition, the Kew hadron events. Charged lepton detection is of Particle facility should be equipped with a sophis course straight-forward but complicated at PEP ticated Py tagging system to cover polar angles up because of feed-down from the normal tiadronic to lf° to i;ive tagging in search for scalar parti final states. A less universal signature of new cles produced by ?? collisions and to increase the particles is an increase in strange particle pro solid angle coverage for lepton identification to duction. This signature is most easily exploited include email polar angles. Of tiecessity, certain in neutral strange particles where only momentum design compromises hive had to be made in order to and angle measurements are required to reconstruct achieve the versatility in detector characteristics the unstable parent. Observation of threshold and emphasis on lepton identification which is so bchBVior in the total hadronic annihilation cross- desirable in seurching for unpredictable phenomena. section can signal the presence of new particles He have for example, foresaken the possibility of possessing no distinctive decay produrts. Of n-K-P separation above the momenta where TOF is course stable new particles must be detected by useful. Such aeparation would be quite defliresble effects of their mass and charge (ionization, dE/ if detailed studies of hadronic final states cf dx, TGF, deflection in magnetic field). new particles are to be made. Perhaps the further evolution of this detector would include compact It is possible to construct a detector which Cerenkov detectors (e.g., aerogels) for ^uch parti sacrifices r.ost features to emphasize one parti cle identification. In any case, we feel that the cular new particle siKiature. This type of detect Solenoid New Particle Facility offers a conceptual or might contribute to the discovery of a new solution to the problem of new particle searches at particle but would probably not reveal much about PEP. the properties of the new particle. Since any search experiment will be time-consuming and even In reading the following, reference should be specialized detectors will be complicated and made to Figs. V-l/jf which give detailed schematic? expensive, it is felc that the best detector would of the proposed facility. Figure V-? depicts be reasonably general-purpose bat with emphasis on further details of the modular blocks for electron good lepton detection. It is felt that other and muon identification shown in Figs. V-l/3- general-purpose detectors should include as good J— lepton detection as is possible. The detection triggers should be as loose as possible allowing an ex-post-facto searcli for anomalous effects rather than presupposing those effects. £ti Detector - General The general features of new particle signa tures and decays described above indicate the v?~ feasibility of e new particle facility which ^t::. emphasizes lepton identification and momentum :rkV^rU--'1i-gg measurement, photon energy and direction measurement l^.,:,^.; I (for -t + 7 mass reconstruction), vee reconstruction and identification, dfi/dx (for anomalous ioni H~ L. zation determination), time-of-flight (to identify slow massive particles), a superconducting coil aurrounding the beam pipe (for monopole detection) Fig. V-l New Particle Solenoid Detector and 2/ tagging (for 0" production). In this sect Octant General View ion we describe in detail the design and character istics of such a facility. The design has not been fully optimized in that detailed Monte Carlo studies are required to optimally configure the various calorimeters, drift chambers and magnetic field volumn. The detector consists of a high- field solenoid magnet whose radius is kept as small as possible consistent with adequate lever arms far momentum resolution and especially enought sensi tive volume to be able to reconstruct most lambdas (vee's). The small radius is desirable if one is to obtain good muoti identification over most of lin by placing a hadron absor _-rs^ close as is feasible to the intersection point (consistent with other requirements). The marnetic field voluni.' is essentially surrounded by liquid argon electron identifiers (which also serve as positlon-sensiLivf photon colorimeters), and by active hadron absor lit't-s except for p;irt of the • iidc;ip;? which SITV :is Hew Particle Solenoiu Detector - End Cap ;md Pattern Recognizer Detail -202- MMMWHMW UODUtt / •cm Ft »I>TE 6cm F* PL»TE 9 mm Thick 6cm F* PLATE SWItriUMr 1 K„ • * UNI ClK'ronicIt l|i 4cilM Togtlhr - ifiam.1 Ullll-H T99it!m -iO. 20CT>— t*MS»l6 2tmSi'*i-i5Ch,|NB ADC I Fig. V-3 New Particle Solenoid Detector • {•1117-60 TSM">t> -JCx, Cap End View ElECrCO D«4cf« Wt>*J* • IS Xc 1' E.ptC toatenRectum »I0"S 3. Magnet Ptioion Peti'un M*oh«*mw« ii, The detector consists of an inner core which D is a superconducting solenoid magnet with minimum useful length of uniform field of 2 m and a dia meter of 1.3 m. The beam pipe is centered in the magnet and is £200 microns of stainless st«ael or its equivalent. This thickness was chOBen to re duce the level of gamma conversion in the pipe to that of• direct Dalitz decays of JT0,S. The beam Eiptct b» 1980 Sueft TuM> pipe diameter is to be a minimal value consistent IDM Ct>w, Co"»ee' with the beam stay-clear region so as to maximize AI~IO"'°Mc the number of vee's decaying in the visible region of the magnetic core. The beam pipe is surrounded Fig. V-5 Hadron/Muon, Electron, and TOF Modules by a multiwire FWC and scintillator for y conver sion suppression, dE/dx and time-of-flight measure hope that with increased experience the coil thick ments. These elements are described in detail ness for such a magnet can be reduced to 0.5 below. radiation length (minimag has a O.33 Xo coil). The magnet is essentially a copy of that The system -itores ^3 H joules of magnetic energy designed by W. Morpurgo for the CCRO experiment and would consume — 20 liters/hour of liquid at the CESH ISR. The coil i9 constructed of helium. —1000 turns of aluminum-stabilized niobium- The magnetic flux return (see Figs. V-2 and titanium conductor carrying 3300 A and generating V-3) is through eight voe-shaped iron end plugs at e field of up to 20 k.G and uniform to 1$ in the each end of the solenoid which butt up against the fiducial volume. The coil designed for the CCRO hadron/muon calorimeters to complete the return experiment is ~1 radiation length thick but we path. The calorimeters, which are described be low, contain sufficient iron to serve as return yokes. This magnetic field does not hinder their function as calorimeters. Further study may indi cate that fewer than eight vee end plates are required so that a larger fraction of the end caps can be devoted to active (hadronjiiuon modules) rather than passive hadron absorber. The vee- shaped end plugs are 12 ed to facilitate placement of photomultipliera a' i to maximize the amount of end cap devoted to active absorber. A crude map ping of the magnetic field configuration is given in Fig. V-h. The cost estimate for the CCRP magnet is appropriate to our design.. In 197^ this estimate, including iron, superconducting coil, cryogenics, power supply, mechanical supports, liquid helium dewers and transfer lines came to 1.62 M Swiss francs or — $0,5 M. We estimate a cost for the Hew Particle Facility solenoid of $0.8 M to account for inflation, further R and D on thin coils, contin gency, and ether unforeseen cost increases. Table Fig. V-k Hew Particle Solenoid Detector • V-l summarizes the properties of the magnet and Magnetic Field Configuration the CCRO cost estimate. TABLE V-l 9 [degrees) 50 40 30 Design Paraaetera of Superconducting SoIanc-iJ Kagnet 1 1 1 1 1 1 (Hi Useful Field Diameter l30om 4ir " Useful Field Length 200cm >™*--\ 3 . "fr Volume 3.0ra 7^- — Magnetic 20KG on -^ -- •^ Magnetic Energy ~ 3MJ ^\^f " Coil Material Al- L stabilizeJ Ni-Ti /'' 1 1 1 0 I 2 3 4 & 6 number of Turns 1000 a Current 3000A Fig. V-6 Solenoid Geoinetry Characteristics Total Height ~80 Tona v t. Mtgneil: ...1. Liquid He Consumption litres/hour CCRO Cosz Estimate (197^ in 1Q-* Swiss Francs) Iron (72 Tons) kSF 1»00 •=€Z) Al-stablilized Ni-Ti conductor 350 Cryogenics 200 Power Supply 60 Supports 200 Transfer Line (15m long) 50 Two 1000 litre dewars 120 Helium Expansion Ballon 25 Recovery Line for Helium Gas (estimated) Miscellaneous TOTAL kSF 1615 h. Inner Detector The problem of the optimum geometry for Ac identification was studied by Hitlin, Marx and Yamin (PEP-161*) who concluded that a solenoid geo Fig, V-7 Vee Detection Efficiency in VariouB metry maximizes the yield (assuming an angular Geometries distribution of A*s given by 1 + cos26 and momentum spectrum as given by Bjorken and Kogut for p*s efficiency which determines the shape and volume of at PEP). Figure V-6 indicates the solid angle the magnetic field required. accepted by a solenoid of various aspects ratios The volume inside the solenoid magnet is and Fig. V-7 depicts the vee detection efficiency filled with position-sensitive detectors Joo form calculated by Hitlin et^ al. Their conclusions, a "pattern recognizer". Two configurations were which were applied to solenoids of 3 m length and considered - - an array of cylindrical drift 1.5 m diameter with a 5 KG field are almost dir chambers and the tine projector detector. In prin ectly applicable to this detector because of the ciple a streamer chamber would serve as such an high field and shorter visible traclt length required inner detector. For track recognition and momentum for reconstruction. It is the vee reconstruction measurement we will propose the canonical drift -204- TABI£ V-2 MWPC and DC Parameters Length Chamber Radius Planes •ctors Wire § # of Spacing of Cathode (ran) Wire* Readout* MHPC 8,J+2 S 2 250 1 14,80 2 54-90° delay 1: 15 120 108 Construction 2 26,180 2 120-90° strips 15 220 240 - *300 K 38,180 21.0 3 2 120-90° atrlpa 15 320 Electronic! It 50,180 2 12r-J0° strips 15 420 240 3 $100/readout - $260 K 5 62,180 2 120-90° strips 15 520 240 TOTAL 160C 1068 *560 K chamber system.1 There are several different l.j mr and o(ie) ~ 17 *in?8 nr for 0{4z) • 1 cc. schemes for large solid angle solenoid systems Somewhat better resolution* can be obtained by that are currently entering development or final fitting observed tracks to the reconstructed construction. In a few years a more realistic event vertex* choice will be possible. At this time there is The basic drift chamber design consists of more speculation than anything else. ~£0O cm long chambers (I80 cat active) with drift For a new particle detector the momentum wires every 1*5 cm along the circunferer.ee. The resolution does not have to be pushed to its drift wires would, of course, be staggered in each limits, flp/p - .01 should be sufficient. Further pair of gaps in the usual manner (see Fig* V-8). more, we do not have to worry very much about The individual gaps could be 1 cm although the multiple scattering errors. A contribution of practicality of such a geometry in 20 kO fields o ~ 1 - 2? should be tolerable. It is also true still needs some testing. There would be no azi that the polar angle 9 does not have to be measured muthal field shaping and the time measured should extremely well or the mass resolution would be correspond to the distance of closest approach of dominated by momentum measurement errors (see the track tc a wire. The maximum drift time woold section on ey or A). Therefore 50 err. SECTOR $TftuC!uM£ OF CM*Mfl£«S m r>*14 *«• • St*** Wc H«*»l 3 * SH'fl . ^ DfliFl CMfiMBEf MOCULC SlS&KlUflt Mo V-tf ttrift Ci.ncti-i J^ruc'-yre* flu? ccdulcjf cuuld U- t-cnstructed na shewn in Pi«. V-0. Each todule la divided Itvic sectors e-1 arc liriuttli -*•»*? «. The fdbr leal left ol 4t-vrr.il identical sellers for each radius and l):c c»re swin^-ablc size or the individual scctora sheulrt simplify the construction tmd lover the cool of Fig. V-V Urtdo Sesolmlon id livv particle thi?sc eharbern. "he sector* for each 1/? cylinder Deiectcr can be held toother at tin- end* by Seal-circular support rinea. TI;c ti?n hnlf-cfcarbcrs cw> i>c hunt on vertical t;uppori.r rtt carl: cud of the solenoid. unliable particle* which arc produced at PEP. The Ne etploy a uiui.l-.iicl. construe*ien usinr. cither Kinl£'JC decay dlstar.ee which can be observed la foao or Kexccl as a lo* density tutrix which la limited by the finite bete «i£c of e„ * US ec, given rigidity by akiui f:lj*'d nloti^ the tap faces. Vith * VICUUR pipe cf * .01 radiation length at Thir osscfcbly nan bc cade into n cylindrical shell 10 cc: radius, and 0.2 CK resolution in the tracking in an appropriately designed preaiu The aV.lna device, the displacement at the Vea» is scalier (mylar, Kaplon, or rii.<-r,.lJ>«,--? wii; !•< ;i.- -I'-l.c-i- than the bean sice. atrip printed circuit In t.r canci (stilus 7 , t> in Jn order to obtain a feeling for ths cass and fig. V-ti). Tliia circuit i^ identical fcr the outer coaentiic resolution possible incidv* a 70 cm radiuc four modules. Tlio other four ah inn »111 *»• covered pattern reco^jiiEer we consider the case of:', decays. with n plain copper or aluminum luyer. The Cuter Ve assure that the.'. *3 decay in * plane perpendi- two (1, b) can then jierve 'is part of a Faraday culnr LO the beais axis, at their tfean decay dis cage for tlte entire chamber. tance. Ve use curves for the itoirentue and angular The signals from the cathode strips of each resolution given in pig* V-o. In the region of /. ecoentutc frue 0.8 to 5. J CeV/c the is deter- sector can be token out on minaturc cc-nx ( ^-1 mm f mag* OP) nt tbe edge M eicli wrtor. per all lv.il the elncd to — i 2 KcV, but above 'j GeV/c the resolut smallest radius the outputs of the cathodi- strips ion rapidly beccr.es worse due to the long decay at a given zee can be OB'cd for each hnlf-'-y Under distance and reduced coiecntun: resolution. Kost of before being amplified and digitised, tor the the tic-c»ntufi. resolution of the A cores from the ••. first module we want to look at both .>uda of each ceoEurement of the proton ROEantiw., so the A delay line. Those lines have tr» be terminated at w>»entuB, resolution follows the curve of ip = their enda and the signals car be taken out after 0.0075P2, up to the point wl.ere the decay distance an emitter follower. All nienala con In.- taken out to co tie a important at 'y GoV/c. We have not included of the solenoid on co-ax cables to external ampli the fact that for A 'e with long decay paths, we fiers. There ie no need to htu? any power dissi have the additional constraints fron. the A direct pation In the eclenoid. Tliio in important because ion. temperature gradient*: can effect tne alignment of for massive particles produced nearly at rest, the chnsibers and, therefore the wire petitions. the moss resolution is approximately equal to the Particles which have lifetimes in the range momentum resolution of the two outgoing particles. 10~H to 10"" see can be observed aa visible Hence 0 'J GeV particle produced at rest will have a "kinks" or "vees" within the detector. Tho-ac could mean resolution of ± 1*0 MeV, and a 10 CeV particle either be the known kaona and hyperous, or new will have a resolution of * 160 HcV. -206- The complicated experimental question of various momentum and mass. The figure shows the und«rit«ndlng the origin of a new particle signal range of mass and momentum where stable new parti is discussed elsewhere in this report,. This dis cles can be recognized by such TOF resolution cussion leads to a possible criteria for momentum (M > 2 GeV, iM * 2 M so that the new objects can resolution in the pattern recognizer i.. that one be separated from the proton). K-n separation to Ctn use the momentum distribution of leptons from 2 is achieved up to 1.7 GeV/c and K-P separation heavy lepto.; decay in order to learn about the to So is achieved up to 3 GeV/c. For time resol structure of the currents responsible for the decay. utions other than AT= 10"*-° sec or distance Since some models now prefer heavy leptons coupled differing from d = 1m the \ ertical scale in Fig. to V + A currents require sufficient momentujn me V-10 can be scaled by&T/d. resolution to resolve V - A and V + A heavy lepton Scintillators and conventional phototubes decays from the measured momentum distribution of for dE/dx measurements (to search for anomalously the final state charged lepton. As indicated in ionizing particles) are located as depicted in Section HI.*» a value of AP/P? - 0.75$ is suffici Figs. V-l and V-2. If we assume that each TOF ent to separate the V + A from the V - A hypothesis counter and dE/dx cost $1K /channel of PMT and adequately. This criteria may not be the most digitizer and scintillator 1B $1 K/m2 the cost of sensitive physics choice since angular distribu TGF, dE/dx and trigger coulters is $io0 K. tions may contain sore information, but it has been useful as a solvable quantitative problem for The importance of lepton identification and evaluating whether the resolution of the pattern the background levelB associated with single lepton recognizer ia of any value other than for mass identification have been discussed earlier in the reconstruction (A, £7, ...). Further studies of report. The New Particle Facility uses lead plates the new particles facility should include Monte and liquid argon for electron Identification and Carlo simulations of particles decays and attempts 7-position measurement and calorlmetry. Figures to analyse these simulated events with resolution V-l and V-2 depict the placement of these electron effects included. Unfortunately this is too time- modules and Fig. V-5 depicts the configuration of consuming a chore to be attempted during the Summer one such module* Study. Perhaps if such calculations are done by We have chosen liquid argon ion chambers with the designers of Mark II, they can be adapted to lead plates as an electron-photon detector because the new particle Facility in order to answer quest of its compact size and its good n-e separation ions of the extraction of detailed physics, rather capabilities. Each cell in the detector will have than Just a signal, in light of the resolutions 1.5 mm of lead and 2.5 mm of liquid argon, so 23f> we hope to achieve. of the shower energy will be deposited in the argon. This relatively small percentage of energy into the active converted implies less than optimal 5. Outer Detector energy resolution, but we have traded resolution for compactness, since the detector is only 3k cm Possible configurations for TOF counters are deep, not including the cryostat. The rf-esepa depicted in Fig* V-9 and their placement in the ration is enhanced by our ability, with a liquid detector is shown in Figs. V-l and V-2. The TOF argon detector, to segment the detector longitudi is based on the -uailability of proximity-focused nally. multichannel p..->totubes. These tubes contain Physically, the detector consists of 60 cells, flat multichannel plates ~1 mm from a flat photo with each cell comprised of a 2.5 mm liquid argon cathode so as to give a Jitter in response time gap and collection plates embedded with 1*5 mm of of less than 100 ps {as experimentally determined lead. The electron collection time for pure argon from phototypes). The multichannel plates have is VjO nsec.-3 The total detector depth ia 17.1 gains of —10" and thus produce signal levels simi lar to those produced by proportional wire chambers. The tubes would be read out by an anode (possibly i I I [ I T~~ segmented) at the back of the plate and PWC-type K-TT Sepaio'ion 2a- for p< 1.7 GeWc readout electronics. If such tubes become extreme K-p Sepcralion 2c for p £ 3 GpV/c ly cheap by —1980 {an unliKely possibility) one can envision flat "tubes" with large areas (—0.5 cm thick) which can be placed in arrays behind plastic scintillator to give high-resolution (aT~ 20-10 sec) segmented TQF hodoseopes. If these tubes are too expensive to cover 10 - 15 m2 of area, the scintillator can be viewed edge-on so that slabs of scintillator placed parallel to the cylinder axis as shown in Figs. V-l and V-2 can be read out from loth ends by such tubes. By correl ating the time of signal arrival at each end the impact point along the axis can be determined so that the TOF from the interaction point can be calculated. If single photoelectron threshold levels are used one would be sensitive to Cerenkov light from the scintillator from particles with [i >.7. For scintillator several mm thick this should allow time resolutions matched to those achievable with multichannel phototubes. Figure V-10 depicts the mass resolution achievable if AT = 10 "^ sec over a distance of 1 m for particles of Fig. V-10 Time-of-Flight Resolution for d = 100 cm, 6t = 0.1 nsec -207- radiatlon lengths villi yivf. of the radlntlon lengths in the lead. In thf tr«navcr*e croBfl-aeetion, each module is 60 cm by I'O en, so that the cnp,-icitance of each module is M at* The longitudinal devel- opemt-ft of the shoner will be monitored by electri 0.3 - cally aividing the module into 'j regions with different depths. Each of the first lour radiation lengths will constitute an individual region and the remaining 13 radiation lengths will bi- the final region. Since the momentum of an electron 0.2 - going into the detector will be known, this .longi tudinal segmentation should reject hadrons better U J "s. than 1 in 10, The transverse position rl the bJ shower will be determined by segmenting the electro About 9^ of the photons will convert before Er er liquid »r(?-!i> eovrr the lull loild tingle or the 1.0 Jet ••elf r except in tin- end cups ./here for part of the mva tie nujwfi end platea replace the ilrot tfcveo Interact iwi lengths if L»dron/cuon modules* 0.8 Thiso active hurfrcn »t>tiorbcro have fine-enough parrpiint: to essentially give corplete nu«n identi fication to the level of litt*~" p.mcJ-ilirojt*>t. A 0.6 tmveraing cucn (-lveB unlforn. pulse in-i^M (except fir Lrnidim fluctuations) while uny htiiron inter I action produces lnrjjo energy d^po-ilion and very 0.« noii-unilern. pulse heights. Except for punch- 1 tlm'ugh the hadrcn reject lor. should if —10*J or better. For six interaction lengths b,i' *^I|t in this facility' il ** pun •!-: i rough probability iss 7*^ ». 10*-'. As indicated in the section of this report describing backgrounds to muon identification, a haJron absorber need only be so thick us to give punch-through probabilities 0 2 4 6 8 10 ceiching the buci-.grounas due to meson decays in M (GeV) the central free space. In the detector, meson decays dominate for n.uon ^omenta le.«a than t< OeV/e Fig. V-1T Resolution in toss Hcconstructicn tor and the two backgrounds are approximately ejual V •.£ + . above 6 CieV/o muon momentum. Figure V-lj depicts the backgrounds to single muons as was done for Muons are identified by the hadron/muon modules electrons. The electron signul is seen to te low whose placement is shown in Figs. V-l, -2, -3 and er in background than the IT.UI n signal by a factor •rhose configuration is depicted in Fig. V-'J* A of 10 at 3 OeV/o and 3 at ';> c;eV/c muon momentum. single module is 1.13 interaction lengths deep, We note that such of the iron in the l.adron/ made of three iron plates, each 6 cm thick with muon modules is magnetized in its role as return scintillator betw.=(-n the plates* The scintillator yoke for the solenoid magnet. The presence of is read out by phototubes at each end (into ADC's drift chambers as the most rut3ide elenent of the and TLC's) SO that pulse height information and detector will thus provide a crude ( — .'Ot) meas timing to determine position along the counter is urement of muon momenta outside of the pattern extracted. The centers in the azimuthal detector recognized so that the melon decay background are segmented into 16 azimuthal segments (wo ptr appears at the muon momentum rather thin the octant) and segmented along the solenoid axis into larger parent meson momentum. This £ postlori measurement of the muon momentum thus has the effect of decreasing the effective single muon background at higher muon momenta. The ability of the hadron/muon modules to give crude hadron calorimetry helps to suppress beam-gas background in total cross-section bump hunting. The outer drift chambers have been chosen to contain 5 cm cells in order to match the multiple scattering of muons in the absorber and to limit the sensitive time T - 1 ^S. At ilOO/drift wire these chambers would cost — $100 K. The hadron/muon modules contain —300 tons of iron plates, 400 m2 of scintillator, and 160 channels of 5W PMT, TDC, and ADC. If we assume the iron plates cost $1 K/ton, the scintillator #1 K/m2 {liquid scintillator would be substanti ally cheaper), and $750/channel for PMT, base and electronics, the hadron/muon modules cost —$820 K. The Solenoid New Particle Facility requires two types of auxilliary detectors. First, a super conducting coil of the sort used by Eberhard and Alvarez in their monopole searches. Such a coil would circle the beam pipe at a distance of -8 m from the interaction region and would detect stable monopoles which are accelerated in the beam direct ions by the solenoid magnetic field. Such a Fig. V-13 Background/Signal for Elepton > E in technique is detailed in the literature and else Leptonic Decay of One of 11 GeV Heavy where in this report. The second auxiliary detect Lepton Pair (Ebeain = 15 GeV) or is a Sy tagging facility which would serve -209- TABLB V-•3 UelUt ha^M Cor Laptev identification 95* x **n Madron/Electron Dejection Sio"-1 Hadron/Kuon Rejection S2.<* x 10"3 A Reconstruction Efficiency >80jE /. >>// .75* 1. fc/E (?) TJ/JT ; position (f) 1 cm A Kaao Resolution See Pie* v-9 d> Kasa Resolution See Pig. V-12 j) Mass Resolution See Fig. V-15 TOf Henolutlon 10_l sec. See Pig. V-IO Background to Single Leptons See Pig. v-13 Weight ~ 430 Tone Size Qrllndert redius > 2m, Length - 5.2i No, PWC Wires S50 No. Drift Chamber Wires 1600 Cost: Magnet $ 800 K TOF + dE/'dX + Trigger ?eo Electron Modules 300 Hadron/Muon Modules 820 Outer Drift Chambers 300 Pattern Recognizer 560 SUBTOTAL $ 3060 K ENGINEERING, SUPPORT STRUCTURE CONTINGENCY 3530 K TOTAL ^590 K several purposes* It is necessary for total cross- resolution is rather discouraging, independent of section measurements (and bump hunting for J?c = decay configuration. 1 states like I^'S) and for tagging Sy production of pseudoscalar states. In addition it provides 6. Summary lepton identification at polar angles below 12°. The use of 2? processes to search for new parti Although the Solenoid New Particle Facility cles is described elsewhere in the report. described here is optimized for the observation of If we use the formula of the yy study group new phenomena at PEP, it can be used to study for the missing mass resolution derived from other, expected, physical processes whjse char measurements of the tagged electron and positron acteristics are matched to the detector's capabili and Bssume 1$ measurement of the electron and ties. Much of this physics is described in detail positron energies for a particular event in which in other PEP reports. These include A and Kg the scalar or pseudoscalar is made in the absence inclusive studies, total hadronie cross-section of other particles, we can calculate the resolution measurements and especially weak interaction experi in the masa of a yy Bystem of mass W. Figure V-l*f ments using lambda longitudinal polarization as depicts this resolution as a function of W. The a polarimeter to search for parity violating effects shaded area spans the space of possible 77 decay Such effects and their measurement were described configurations. ,Tor values of W below h GeV the by Hitlin, Marx and Vamin (PEP-161*) and are of i -r i r i ?. U»ing fDrssalne from GliK-'kstvrn, HIH £>•, j6l (l'X>3) u V. Willi* ana V, fmdeka, HIM 1£U. .'tl (l'V/J.) h c. J. Ci-nnnell, Phy*. Rev. 161, JiO (1907) fijtn (w t 0.32 . t t . - VI. SPECIALISED NEW PARTICLE DETECTORS 1. general (Jier Magnet 0.28 - The ide.il charmed particle detecter sho*n in —Moil symmtH* pholon Fito V-l could in principle be adapted to the 0.24 configuration proposed General User Hagnnt. The tr.uin differ (Leas' probable) ences beweon the two magnets are the larger sizes of flUV. J 0.20 \ (*")<» oun tfew Particle Detector L^-Most asymmetric photon R - l.S tr againet 1 n 1 r conftgurolion L « <* m against 2 m : o.i6 \ (Most probable) B * 15 kg against ru kg * §0.12 2 IV.* n.ettpm.uri reoolution cf the CUM is slightly better than that of our ideal detector (Bq* • UM" ' ,0 kgo? against 7*20 kgm2). Its larger size o.oe \ implies that the external detection equipment tarpon counters and y-li idtntiflera) would be 20-»i tware expensive* Also the larfiev distance 0.04 \^ transversed by any n before it can be identified as such implies that the n/p. rejection will be slightly {20} ) worse. A possible arrangement 1 i i i i using GUM as the central detector is shown in 0.08 0.16 0.24 0.32 0.40 Z Fig. VI-1* If a continuous coil rteslgn is chosen 2.4 4.fi 7.2 9.€ W IGeV) for GUK, the radiation length presented by the z = W/30 coil would be the seme. The end cap cf GUK can be can be ad&pted to act as MA identifier. Should Resolution in T7 Mass from Tagging light detectors using micro channels become com Apparatus mercially available, it may well be advantageous to put the TOF inaide the coil, thus making the two outer detectors even more comparable in size. obvious interest in their own right. They are also (These devices presumably would work equally well relevent for new particle search in that, quasi- in a magnetic field.) stable hadrons ind heavy leptons can be expected to decay into strange particles through parity There is a point to be made about the use of violating transitions. For this reason, this a general purpose detector as a new particle dis detector was scaled to observe A°'s - not only to coverer. That is, most attempts at identifying observe anomalous strangeness production from new hadi'onic quantum numbers at PEP involve charmed particle decay, but also to use the A° studies of statistical distributions; e.g., K/rt polarization to spot thresholds due to weak decays ratios, thresholds in scaling violations at amall of new particles. The obvious utility of total x, mass bumpB, etc* It is nlso possible to dis hadronic cross-aections and lepton pair crods- cover a new quantum number, like charm, with one sections to new particle searches has been demon fortuitous event. Ckie need only find a violation strated at SPEAK. of strangeness conservation* Consider an event In Table V-3 we summarize the particle where a charm-anticharm pair are produced (with characteristics of the Solenoid Hew Particle or without other particles). If both members of Facility. It is a large solid angle detector with the pair decay hadronically, one to a strangeness a multitude of characteristics of special relevence zero final state and the other to a state with non to new particle searches. As such we feel it is zero strangeness, then a detector which can detect the kind of system best suited for ntudies of the all secondaries in the event and identify all "expected" new phenomena at PEP, and, more Import strange particles would tell us that strangeness antly, it is a system which may optimize the violation (and perhaps the existence of charm) chances of being at the right place and time to has been found. Che would see an event which is observe the entirely unexpected phenomena which a i(-C fit with non-zero strangeness. The General accompany all new accelerators. Purpose Detector proposed at the 1975 PEP Summer Study [hn charged detection and hadron identifi cation by Cerenkov techniques) could be such a detector. The ideas for the system described here are heavily influenced by: 2. "Split Chamber" a. CCRO Solenoid Detector - high p leptons and hadrons at the ISH It is very well possible that excited states b. LBL MINIMAG Detector of the muon exist. Such states, apart from having c. CERN- Scandinavian -Harvard-LBL-Ruther ford- the decay p.* ~* p + y, would also show the electro high p hadrons at the ISR magnetic decays n* -»n + 2K, H + 3n, etc. In order LOTWXItW, Rltum Hadton Muon MOOuHs New Particle Facility in General User Magnet (GUM) Fig. VI-2 OFF-AXIS Solenoid Detector to study such decays, the rt/n rejection in the present detectors may not be adequate. If we capability* The trade-off for the best energy assume that the n" are produced without a form resolution is a large percentage of liquid argen factor, and the branching ratio into hadrons ia compared to any passive heavy radiator which makes B - 1*, then the LAND e'etector quite large. External muon identification is, therefore, not possible. The group investigating LAND-* did nc;t reach oU+e- -»n'u -~w*nn)^ B tf IQ-3 an absolute conclusion about the inclusion of a(e-*e- -h) ~R passive radiators within the liquid argon. These radiators degrade the energy resolution by intro Threshold effects would further depress this ratio ducing a fluctuation resulting from the variable and since one of the two muons from the reaction amount of energy deposited in the liquid argon. would typically have low momentum ( =15 Gel// The pure argon detector, on the other hand, permits No* IT'S + 1) the n/ji rejection of the detector may the shower to spread radially over large distances. not be sufficient. However, it would in such a From the Koliere radius of 8.5 cm, one can estimate case be possible to remove part of the inside track that a region 0.5 m in diameter will be required to ing device and replace it over an azimuthal range capture QO? cf the shower energy.£ If a new parti 6 ft— JI/2 - n/3 by a JI/JI identifier similar to those cle is protJut. sd in a complicated state with many used outside. The material of the now rather small photons, the pure liquid argon detector may have internal n/u identifier would have to be non hopeless ambiguities. Pure argon does have the magnetic. Its purpose would be to yield for single advantage of short absorption length compared tc fast muons a n/V rejection of 1-2 x 10"3. This radiation length. A 15-radiation length detector, rejection factor, when multiplied by the rejection for example, represents 2.6 absorption lengths. factor JT/JI — 2? of the slower muon, would enable Even without external calorimetry, this detector one to study such processes with only a small back would give about £00 to 1 rejection of JT pairs as ground. u pairs. The energy resolution of a pure liquid argon 3. Liquid Argon Neutrals Detector detector is somewhat speculative. In principle, it is limited only by the capacitance of the detector, LAND (Liquid Argon Neutrals Detector) empha the amount of radial spreading of the shower and sizes the high energy-and-positinn-i'esolution the energy which leaks out the back. Since there detection of neutrals without dome of the dis is no passive radiator, in a pure liquid argon advantages of other kn neutral detectors such as detector there is no errsi- due to sampling. Liquid the Nal (T&) Crystal Ball. The major advantage argon also does not have the low photon statistics over the Crystal Ball is the presence of a magnetic problem of lead glass nor ths light collection field and high resolution drift chambers iiunedi- and crystal homogeneity problem of WaI(T*t). Of ately surrounding the interaction regions, which course, "gremlins" may appear but an energy resolu permits tracking and momentum resolution on charged tion of 3£/E should be possible. The resolution particles. This momentum measurement and the depends on E rather than JH because it is domi longitudinal segmentation possible with liquid nated by electronic noise. (This resolution argon also improves the JI - - discrimination requires a detector to insure that the gamma has p IGev/c) „ Fig. VI-3 Momentum Resolution In Streamer Chamber Fig. VI-4 Multipsrticle Matt Resolution In Streamer Chamber not converted in the solenoid coil.) Adding pas sive heavy radiators worsens this resolution and Cue way to reduce the decay path is to parti changes the energy dependence of the resolution to ally fill the interior of the solenoid with a fE except at the lowest energies. For example, if dense non-magnetic material. This essentially only 20% of the shower energy is deposited in the wastes the magnetic field in that region. An liquid argon, as is the case for the electron- equivalent effect can be achieved by moving the photon module in the general new-particle detector, beam pipe OFF the axis of the solenoid without then the.energy resolution above 500 MeV is about the loss of the magnetic field region. This allows an improved momentum resolution for particles in Because the general new-particle detector the opposite hemisphere. It also should be possi involves the use of lead-liquid argon modules and ble to use the same detectors as the normal sole a magnetic field surrounding the interaction noid configuration if they are designed in advance region, it is instructive to compare a pure argon for this option. The problem of compensating for LAHD with the new particles detector. The two are the effect of the magnetic field is probably equi quite similar, but the LAND has better energy valent to that for the normal solenoid. The radi resolution for the gammas traded for poorer muon- al field t ,at occurs in the beam pipe near the end hadron identification. The transverse position caps is small if the iron is not saturated and can resolution for the gammas is also somewhat better be further reduced by surrounding the beam pipe for the LAND as is the n-e separation, both be with a n-metal jacket in that region. cause of the greater segmentation of the detector• A detailed study would be required, however, to p. Streamer Chamber as a Core Detector for New quantitatively estimate these differences. Particle Searches In summary, for the new particle quest, LAND is superior to the general detector only for the The principal report done on the streamer study of heavy electrons which decay e* -» e?. The chamber di'.ring the study is included in the General muon-hadron separation and the inability of LAND Detector Group report. Here we review the proper to measure total energy by calorimetry are severe ties of such a system and note its characteristics disadvantages in new particle searches. Indeed, as a new particle detector. unless the e* has a pathological dscay scheme, it Due to the high information densities possi is likely that the y will have sufficient energy ble with current electronic readout techniques, the for the general detector to do an acceptable energy streamer chamber can be used for very precise measurement. He conclude that LAND has only limited charged particle moment1::., measurement. With a use for a new particle search. chamber of length l.;j m, outer radius .9 m and inner radius .25 m in a magnetic field of 15 kG, a measurement with fip/p < 1% up to p = 10 GeV/c appears feasible over .75 X **n solid angle (see See the report on LAND elsewhere in this Fig. Vl-3). Angular measurements of comparable Summary Study Volume. precision can be made, with resulting excellent multibody mass resolution {see Fig. VZ-k), Carol Jo Crannel, F..vs. Sev. l6l, 310 (19&7). Particle identification by ionization is Off-Axis Solenoid possible over approximately the same range as usually done by time-of-flignt. In addition, e/«, Clie of the problems of single u. search at PEP e/K and e/p separation can be done for .? < is t'ie background from meson decay. The effect of p < 10, p > .8 and p > 1.5 GeV, respectively, due the -lm decay path typical of most solenoid de to the relativistic rise in particle ionization. tectors is to make a single \x search for p^, < Some JI/K and jr/p separation is also possible for 3 GeV/c very difficult. The easiest way to reduce p • 2 and p > 3 GeV/c, respectively, by the same this problem is to reduce the decay path. technique. Excellent e/hadron separation and good y detection and energy measurement over a somewhat mensionleas form factor; S is total center-of- reduced solid angle can tie obtained by Insertion tnasa energy in the F-P collision. If scaling holds of PbQg plates in the outer region of the streamer then F ii a function of the variable q /S. chamber, which, however, requires either more Figure 1 depicts F(x) s a function of x « radial space or reduced momentum measurement accur q^g for all existing data for P-nucleon experi acy, as veil as substantially increased data ments (Leaerm&n and Ting *t BHL end W. L^e at handling capacity. (It 18 already bad*) RIAL). The Eolid curve is the prediction derived Two access problems require solution. First, from the parton distribution functions extracted the chamber charger line must be brought through from deeply •inelastic le*'ton scattering. the coil, probably requiring a hold of 20 cm dia meter through the coil. Second, XY-. BOiid state readouts require perhaps 1 m (less with some lenses) of transparent spatial separation from both ends of the streamer chamber, although appar The production cross-section for pairs com ently they need not be outside the magnet pile pared to u+u" from tiee-lixe virtual photons la caps. A further requirement ia that the trigger must be relatively low rate, with 5/sec or less probably being acceptable. The Marx charging tlms (2) is of the order of 100 msec and readout time should be less than that. A two-particle trlgtrfir with *--(9iw[-^r« inner MItfPC, outer trigger counters and beam cross where K is the mass of a new particle and q is the ing should be satisfactory. virtual photon mass An apparatus built around a streamer chamber clearly has great advantages for searching for new 2M2 particles whose decay modes must be studied con for fertnien.8 sidering: (a) high momentum particles; (b) Miss ing neutral p and E; (c) high, cultiplicity, particularly in jets; (d) mas3 reconstruction of multi-boay states; (t-) dE/dx; (f) electron identi •(ft fication. Many of the classes cf particles sought hove signatures detectable and studjable with such information. Thus, the total paiiairr cross section is a function of the mass of thiee new particles APPENDIX 1 I'E? Cross-Sections Estimates from Madron Machines pair H'.'w particle searches at prcton machines have ••"•©*wi-*r#'"' established limits on their production cross- sections. These li.rj.ts i.aV'j relevance to such searches at PE? if a n.oael c^n be derived to re late h&dron collisions to e+e" annihilation. (4) If we ufj-siice that all new particle production at hadron machines is due tc pair production from virtual time-liXe photons we can use measurements cf the cross section for p + P -1+ : • anything This integral can be evaluated numerically as a measure of the virtual photon flux in hadronic for a given value of li using Fig. A-l ana the value collisions. If this flux is then intf^atcd over of S appropriate for a gived hadronic experiment. the appropriate kinematic range for new particles In this way one can translate an experimental of a givf-n charge, mass, etc., the crosr section of upper limit for new particle production in hadronic verticil photons capable cf producing a given collisions into a lower liir.it on the new particle particle in an experiment, can be estimated find mass which pair is produced. This mass limit can compared to the experiments! limit. In this way then be used with Eq. 2 tc predict an upper lin.it on one can estimate 'in upper limit to the number of now particle production at PEP (q2 = 900 GeV2). new particles which can be produced at PEP for a A quick method cf estimating whether it is given integrated luninnsity by extracting u:ass reasonable to search at PEP frr particles of given limits from the hadronic experiments. mass in light of hadrcn beam cross-section limits From scaling arguments (dimensional analysis? can be useful. In terms of the variable x = qf/S the cress section for p + p *l* 1"+ anything a ir is given by Eq. j. can be written The quantity F(x)/ ^ is plotted in Fig. S so that t;.e pair cross section in a hadronic experi ment with center-of-mass energy S can be simply - F'u.% S) (1) calculated from folding the kinetic factors with Fig. a for x = 1 to x = UH^/S. The ratic of the experimental limit to this predicted cross section where q is the invariant mass" of^the leptcrj pair gives the effective suppression factor for verti (i.e., the virtual photon) and Ftq-,3) is a di- cal photons to produce pairs of particle mass = M. -214- 11 lO"' 25 IO" I p + p—— I r + onylhing El p +p—Q*Q~ + onylhing - \ 2 2 dq 3* S z I0"26 _\ dq' 3x S x Ledermon BNL - \ x Ledermon BNL 10"' x • Ting BNL • Ting BNL \ • W. Lee FNAL : \\ N« io-" r \x 1 W. Lee FNAL \ — Parton Distribution X \ Parton Distribution Li_ " \\ I0"28 10- \ " \ 29 * \ \ I0" \ \ \ \ sx in-30 1 1 1 N 1 1 . 0 0.1 0.2 0.3 0.4 0.5 0.6 0 O.t 0.2 0.3 0.4 0.5 0.6 2 x=q /s * =q2/S F(x) from Hadron Beams (see text for F{x)/x from Hadron Beams (see t»xt for use) use) Thia suppression factor, whether viewed as a form better to include thetn and to calculate numerically factor, or as due to some dynamic suppression of the integral. the co'iplin^, can in- used trivially t"» estimate an It should also be noted thai; the higher the upper limit on the number of events possible at mass the lower a pair so that the suppression PEP. factor for a given experimental limit decreases For exan.plc, if the predicted n pair = with increasing mass. Another aspect of this pro 10" & cm1" for a ir;iss of ? GeV and an S appropriate cedure is that the hadron-hadron estimation of the to an experiment which gives a cross-section limit virtual photon flux actually measures the product ot 10-i*l crtf? for the production of a particular of flux and the partial width of the virtual particle', then the suppression factor in cross- photons to lepton pairs. It may be as in the case 1 section is 1C-(J. Eincf D typical PEP experiment of the y and Y that the branching fraction to with an integrated luminosity of 10-^ cm~^ produces •""pton pairs of a resonance state is small. This 2 x 10b virtual phrtons (if o> =. 20 mb), the results in a relative insenaitivity to the state hadronic limit predicts only few events at PEP. in a lepton pair experiment in hadron beams while •trie should, however, be very careful in neglecting storage ring experiments could still observe the kinematic factors in the integral. It is large effects. PEP page I REPORT OF EXPERIMENTAL AREAS GROUP A. Carrol, B. Case, D. Coyne, F. lister, F. I,obkowicz, F. Martin, C. Morehouse, P. Oddone, C. Prescott, L. Keller, and G. Manning 1. INTRODUCTION The arguments for having the same interaction height for phase I and phase II is that experi* The 1974 PEP Summer Study made rccommenda- mental equipment will in practice grow to occupy tions concerning the experimental areas. The ma the full 4.8 m available below the beam and it will jority of these have been incorporated into the up then be very difficult to lower the equipment by dated PEP design. The description of the proposed the required 80 cm when the proton ring is added. experimental areas presented to the Summer Study Many detectors will be reused in phase II either is appended to this report. In the appended drawings to continue with e*e~ physics or because they are the elevations are shown relative to the tunnel suitable for ep physics. midplanc. At the time of the Summer Study the pro posal was to construct the e+e" ring 80 an above The arguments for not having the same inter the tunned midplane. The proton ring, located in action height are the extra cost and inconvenience the tunnel midplane, would be added later and the of longer supports for the electron magnets and the interaction point would then be lowered bv 80 cm. fact that the whole of the electron ring has to be moved when the proton ring is added rather than It will be seen that Area 6 is an open con only half if the electron ring is placed 80 cm crete pad of 40 x 40 m, 4.8 m below the proposed above the proton ring height in phase I. e*e" interaction point. The equipment is housed in a light flexible building that can be modified to From the experimental viewpoint the same satisfy experimental demands. Area 4 is deep below height for e+e" and ep interaction points is still the surface and access is only through a tunnel of strongly recommended.1 3 x 3 m cross section that ends at the machine tunnel height, which is ~ 2 m above the pit floor. 3. SIM1ARY QF DETECTORS PROPOSED DURING 1975 SOWER STUDY Areas 2, 8 and 12 have access from the side Table 1 gives a swmary of relevant param 4 m above the pit floor level. Area 10 is very eter- of the detectors considered during the 1975 limited and is not intended for experimental PEP Sumner Study. physics. The following explanatory points need to be All interaction pemts have 20 m between the high & quadrupoles giving t 9.7 m free for experi mental equipment. Areas 2, 4, 6 and 12 have the a) Item 1 (the parastic Zy tagging system) e+e" interaction point 4.8 m above the pit floor is not a full detector, but a possible and Area 8 has 6.8 m. As proposed the ep inter 2y tagging system that can be added to action point for ep would be 80 cm below the any other detector that leaves adequate e+e~ interaction point.i room. Those experiments listed as using it need the equipment they specify plus 2. RECOMMENDATIONS FROM 1974 PEP SttMER STUDY NOT that given in row 1 of the tablfi. INCORPORATED INTO UPDATED DESIGN b) Item 2 (general user magnet) is not a full experiment but is a magnet system There are only two significant recommenda plus a central drift chamber assembly tions made by the 1974 Summer Study that are not for measuring momentum of charged parti included in the updated design. The first of these cles. Space is left for the addition of was the recommendation that three of the six in other detectors required for specific teraction regions should have 10 m between the high experiments. Experiments using it need B quadrupoles rather than 20 m -- the object being the equipment they specify plus that to get a higher luminosity in these regions. This listed in row 2 of the table. recommendation was not included because of the cost. It is intended to consider this modification as a The following comments can be made on the possible improvement program. information gathered. The second recommendation not adopted was 3.1 The 1974 study included twenty experi that the e*e" and ep interaction heights should mental systems, the 1975 study includes ten. A be the same. To achieve this it would be necessary possible extrapolation to 1976 could be five de to either install the electron ring magnets for tectors which matches the five interaction regions— phase I at the planned height of the proton ring this may not be so unreasonable because the ten for phase II or to deflect the e+ and e" beam down systems from this study include considerable dupli by 80 cm in the long straight sections. The foimer cation; the number could easily be reduced by a method is simpler. factor of two in removing this duplication. Table 1. Requirements of the detectors designed by the 197S PEP Sumner Study. The symbol ti means that the requirements of the device following the • symbol need to be added to the basic detector. I'horoluhcs t>irr* iluim-li IOW ItnoO sTTt^-.T ISO) 1000 IDdO 3.5 l.S 5.5 -217- 3.2 A "typical" experiment has 300 photo 3.6 Eight o£ the ten experiments use super tubes, 3000 cell; of drift chamber and weights conducting magnets. Hence, in general, the power 400 tons. Adding the parastic 2y tagging system requirements are low, even though large volumes increases the requirements by about 1000 photo of magnetic field are used. tubes and 14,000 wires of MtfPC. A simplified and cheaper parasitic Zy system is surely possible 3.7 There will be considerable requirements and should be designed. for refrigeration. For superconducting magnets there are up to 24 tons of coil at A'K. For liquid 3.3 Figure 1 is a scatter plot of the total argon calorimeters there are up to 200 tons at length and total width of the twenty detectors 78 K. Not all groups have estimated the refrig from the 1974 study shown as dots. The de eration requirements and further work is needed in tectors from the I97S study are shown as crosses. this area, The lengths shown are ZL given in Table 1, where the figures are those outside the brackets. This 3.8 Most groups require areas for fast trig is the length required by the detector alone. The ger electronics close to the detector (within 10 figures in brackets are the length required in m) and an area of 10 ra* seems adequate, Groups will cluding addition of 2y tagging. also require about 100 m2 of counting room which can be more remote—within 50 m. General facilities Including these requirements it is clear that will require rather more room because more than most experiments have expanded to fill the avail one electronics system may be present at the same able interaction length and the height available. time. 3.9 All groups require gass*s used for MUTC ^"T or drift chambers. Several groups require addition al gasses for Cerenfcov counters. These require ' ments are shown in the table. 20 - - 4. RECOMJENPATI0NS • 4.1 The e e" and ep interaction points '• should be at the same height above the experi mental area floor.1 10 4.2 Counting rooms should be provided al • lowing 10 m2 within 10 m of the detector plus 9 100 mZ within SO m. The former will have to be in the experimental areas, but the larger areas a should be in separate counting houses as the ex «• • • 9» perimental areas are expensive real estate. Gen £>••• • eral purpose detectors used by more than one ex a periment will require more counting room area. All groups will require access to light workshop space. 0 i 1 O 10 20 4.3 The lower energy range of PEP should be LENGTH IN METERS decreased to - 4 GeV to provide an overlap with SPEAR.1 4.4 Survey should be made of the cryogenic Fig. 1. Plot of length vs. width of detectors from requirements of detectors and how they can be most 1974 (filled circles) and 1975 (crossed economically met. circles) PEP Summer Studies. 4.5 A survey should be made of the computing requirements for PEP for both on-line and off-line All experiments fit within a 4 m pit. Were experi use. There are obvious advantages in standardiza ments able to obtain higher luminosity by de tion of on-line computers. creasing the interaction length, it is probable that some of than would accept either no Z-y tagging 4.6 Provision should be made for access to or a lower 2-y tagging efficiency. the experimental areas not only for large heavy loads but also for easy access for light loads- 3.4 All detectors have been persuaded to say of less than 2 tons and area less than 2 * 2 m. manage with a crane capacity of 30 tons, but sev eral would benefit if a higher capacity were avail These requirements are adequately met in able. region 6. Region 4 is special and it is recommended that the tunnel access should finish at the level 3.5 Groups were asked t:- estimate the size of the pit floor. of the largest single piece to be taken into the experimental area. These values are given in the Regions 2, 8 and 12 are similar in that table and show that access is required for large access is from the side and is 4 m above the pit objects. Several experiments will not fit into floor. Four methods were considered for access for region 4 for this reason. light loads: -218- (a) A cage picked up by the overhead and 8. The alcoves are 20 n long, 7 m wide and 5 m bridge crane. high. Some e*-e" experiments will almost certainly (b) A cage picked up by a light capacity need some alcoves, and the implementation of Phase crane supported from the side wall. II with 200-GeV protons would require than. The (c) An elevator. cost of adding then later would be many times (d) A ramp. larger than that of providing them initially. Areas 4 and 6 are extended naturally in the CW Item (d) is not recommended. Item (a) is proton direction, as described below. easily provided but will be inconvenient because a crane driver will be required for each minor Some of the special features of each area load and often the bridge crane will be in use for follow: other work. Area 12 (Figs. A-5a and A-3b) Item (c) is preferred to item (b) but is probably more expensive. This area is located in moderate terrain with the interaction point 5 m below natural Three methods were considered for access for ground elevation. Access to the experimental hall large and heavy loads: is from a yard located on the outside of the ring. The experimental hall is offset laterally frxi 1. A hole in the roof with a lightweight the beam line. This allow.* set-up of very large cover. apparatuses in the 17 x 20 m area adjacent to the 2. Proposed doors and a removable platform beam. A typical device, and how it may be arranged under the bridge cranes. in the area, is shown in Fig. A-4. The most utili 3. A removable side wall to the building with tarian mode of bringing materials into the experi the bridge crane extending beyond the pit. mental hall is through portable loading platforms placed under crance coverage, as also shown in Item 1 is not recommended as it is expensive, Fig. A-4. The area has an alcove in the CW di is not compatible with the 4 m of earth shielding rection. required for the ep system, and is not essential. Item 3 is more expensive than item 2 and will make Area 2 (Figs. A-5a and A-5b) it very difficult to provide adequate shielding for the ep option. Item 2, which is the proposed This area is also located in mMerate terrain system, is recomnended providing sufficiently large with the interaction point 6 ra below the tv airal doors are provided--doors about 6 m widi and 5 m land elevation. The experimental hall and yard are tall seem desirable. similar in dimensions to Area 12, except that the hall will be placed symmetrically with 12 m on 4.7 Most experiments will require a 2y tag each side of the beam line. This would allow op ging system and it is recommended that suitable eration of two experiments alternating in a vacuum chambers and synchrotron shields be de push-pull configuration. The area would also be signed so that all interaction regions can have very useful for two experiments not requiring full this option. solid angle coverage or a single experiment needing large transverse dimensions on both sides of the beam line. The area is extended by an alcove in the FOOTNOTES CW direction. 1. Following the 1975 Summer Study recommendations a Area 4 (Figs. A-6a and A-6b) tentative decision has been reached to build the electron-position ring in the midplane of the tun Area 4 is on the side of a steep hill v.ith nel. Similarly the lower energy capability of PEP the interaction point 12 m below natural ground will extend to 4 GeV. See Appendix 2 of this report. elevation. The hall will be positioned asymmetri cally along the beam line, favoring experiments needing length in the CW direction. The experi Appendix_l mental hall is narrow (12 m) and long (40 m). Access is via a tunnel from an adjacent clear area DESCRIPTION OF THE EXPERIMENTAL AREAS AS OF near grade level. Because the CW direction is JULY 1975. favored, no alcove is included. By convention, we nunbur the areas according Area 6 (Figs. A-7a and A-7b) to the hours, with 12 o'clock to the north. Figure A-l shows the contour lines of the site and Fig. This area is in the lowest terrain, with the A-2 an elevation cut along the ring. The topograph interaction point nearly at natural ground level. ical variation is cr. advantage for the varied ex Taking advantage of the low terrain and a good perimental program that we are considering. Each sandstone base, the proposed area consists of a area has strengths and limitations, and together very large pad (40*40m) providing great flexibility. they fit both the land and the varied needs of the Portable cranes, light buildings, portable program. The basic parameters of the experimental shielding and ad-hoc arrangements will be used to areas are shown in Table A-l. rig and service experiments. The area is offset in the CW direction, and access is possible from both The land pattern favors the clockwise (CK) sides of the beam line. The flexibility of this direction for the future proton beam. Alcoves in area is especially advantageous for future de the CIV direction have been included in Areas 12, 2 velopment. Area 8 (Figs. A-Ba and A-8b) Area 10 The interaction point in Area 8 is 5 m below This area is deep underground, with the in natural ground level. The area is similar to Area teraction point 20 m below ground level. The de 12, offset laterally relative to the beam and velopment of this area will be very modest. The having an alcove in the CW direction. The special tunnel will be widened to About 7 n for a length feature a£ this area is the extra depth below the of 20 m. The area will be devoted primarily to beam (6 m) and the crane height above the beam machine physics experiments. Anyone who has ob line (8 m). Some large cxperinental devices will served the prograss of operating effectiveness in have particular need for this feature. modern accelerators, particularly the ISR and SPEAR, will appreciate the excellent benefit to all experiments that accrues from the investment of machine time in storage ring improvement studies. Several experiments designed by the 1974 Sinner Study would fit in this area, but the apparatus would have to be carried in small pieces. 'JU'|*"iJ4J tiv\». Y;-^ •&$ ". V-V :-s^-' - a •V I REGION 4 = REGION 8., \_?TO VK ^'< ?.•:';'' Jgjjii^j/ •;• y>'tf AX :}y)m ---•?>•= ,^\ y?~ . \- SL' Fig. A-l, Topography o£ the PEP Site at SLAC . 100 f^e^»_Hpe**ontaf and vertical scales the same -\\SN\\\\\\\W PEP PROFILE 2 •7 in TUNNELS AND EXPERIMENTAL AREAS 0 200 400 600 S00 1000 120J 0 140I 0 160I 0 180I 0 200L 0 Meters ,_ XBL 756-3OT3 Fig. A-2. Topographical profile of the PEP Site. -221- Table A-i. Specifications for Experimental Areas. Are. Specification 12 2 4 6 a 10 Length (•) Clockwise (C») 10 10 30 25 10 10 Counter Clockwise (CCK) 10 10 10 IS 10 10 Width (•) Inside 7 12 5 20 10 3.5 Outside 17 12 7 20 20 3.5 Height (•) Up (Crane Clearance) 5 5 4 8 3 Down3 4.8 4.8 4.8 4.8 6.8 2.8 Crane (tons) 30 30 30 No 30 No Nearby Space for Yes Yes Yes Yes Yes NO Counting House Alcove Yes Yes No No Yes No Length (Ctt) (a) 20 20 20 Width i (a) 3.5 3.S 3.S Height {* W 3 3 i - (a) 2 2 2 Power (MW) Installed Capacity 3 3 3 3 3 0 Distributed 1 1 1 1 1 0 Loading access Outside Outside Outside Both Outside Beam (yard) (yard) (tunnel) sides (yard) tunnel aDrops by .8 m when the proton, ring is added. 'Available at pad. The total for all areas is 10 MW. "1 t T I n J M0»»-MCTIOU -1 REGION (Z LOHBITUOINai. •CCTIDM Fig. A-3a. Isometric view of Area 12. Fig, A-3b. Architectural view of Area 12. PORTABLE lOADING PLATFORMS BEAM A. AREA 12: POSSIBLE EXPERIMENTAL SET-UP Fig. A-4. Possible experimental set-up for Area 12. 1 •-- - fl1>.w 1 KLCOIC 1 «,.«.;' *+cfl CROSS • 3CCTIOM Fig. A-Sa. Isometric view of Area 2. Fig. A-5b. Architectural view of Arcn 1. pw ^ L «. -1 LO..J LONGITUDINAL SECTION {fevili.;—i:^9 REGION A METRES Fig. A-6a. Isoraetric view of Area 4. Fig. A-6b. Architectural view of Area 4. REGION 6 Fig. A-Ta. Isometric view of Area 6. <$*.,. "V-. Fig. A-7b. Architectural view of Area 6. Tn. H? JO j jg__ J LONGITUDINAL 5ECriON Fig. A-3a. Isometric view cf Area 8. Fig. A-8b. Architectural view of Area 8. -228- Appendix 2 OHICI MIMOIANOUM • STANFORD UHIVEIIITV • OFFICE MEMORANDUM • STANFORD UNIVERSITY • OffICI MEMORANDUM DATE: October 8, 197S PEP Distribution F*OM J. R. Rees : r SWECT: Working Decisions on Operating Energy Range and Beam Elevation One of the recommendations of the 1975 PEP Summer Study was that the storage ring should operate down to a beam energy of 4 GeV 1n order to overlap the SPEAR operating range. This appears to be a reasonable and Inexpensive criterion which we should adopt. It is therefore adopted as a working deci sion pending any outcries from the PEP group. Another recommendation was that the beam height of the e+-e" ring should be the same as that planned-for the proton ring. In our proposal, we were inconsistant in that the magnets in the tunnel were depicted as hanging just below the roof, 80 cm above the elevation planned for the eventual proton ring, while the beam line in the Interaction regions was shown at the elevation of the proton beam. Whether to accept this recommendation has been the subject of many spirited discussions regarding cost, crane-hook clearance, shielding problems, the puerile irrepressttllity of the average user and the passionately- held Intestinal feelings of the storage ring builders. The working decision reached was that the beam height shall be that planned for the eventual proton beam. These working decisions should be reflected In the Conceptual Design Report. JRR:c cc: G. Manning, Rutherford Laboratory K. Strauch, Harvard University ^ For the host experiment the tagging provides not ?tT * only the obvious physics bonus -- a coiroarable '/ 00 number of events of a new and fascinat q process f s' l~"——i 5 ! . . t -- but also serves to monitor the YY banground. A general-purpose double tagging system should have the following characteristics: 1) detection of electrons with energies between about 1 and 14 GeV and at angles between about 10 and 300 mrad, 13. The deep inelastic cross section for e+ve+ 2) energy resolution of the order of .3 GeV anything as a function of the cutgoing elec r.m.s. or better (if one wishes to hunt tron energy and angle. Lines of constant q* for C= +1 resonances, the resolution are also indicated. should be as good as the ultimate machine S.C. PbGLASS