• I

- ­.--.

,•

I·iI·

I

I

1

. ....

LEGI81L·ITYNOTICE. .

:-

A major purpose of the Techni­cal Information Center is to providethe broadest dissemination possi­ble of information contained inDOE's Research and DevelopmentReports to business, industry, theacademic community, and federal,state an·d local governments.

Although a small portion of thisreport is not reproducible, it isbeing made available to expeditethe availability of information on theresearch discussed herein.

.' r-

.'.L- " /\ .

12070 Jefferson Avenue

Newport News, Virginia 23606

(804) 8i5 .000i14'l-'l/()iJ

CEBAF Design Report

DOE/ER/40150--T1

DE90 010523

Continuous Electron Beam Accelerator Facility

May 1986

DISCLAIMER

This report was prepan:d as an account of work sponsored by an agency of the United StatesGovemmenL Neither the United States Government nor any ageucy th.:rcof. n.:lr any of theiremployccs. makes any warranty. express or implied. or assumes any Icgalliability or respousi­bility for the accuraC"j. completeness. or usefulness of any information, apparatus, produa. orprocess disclosed. or represents that its use would not infringe privately owned rights. Refer­ence herein to any specific commcrcial produa. process, or service by trade name, trademark.,manufacturer. or otherwise docs not necessarily coDStitute or imply its endorsement, recom­mendation. or favoring by the United States Government or any agency thereof. The viCIIISand opinions of authors expressed herein do not necessarily state or reflect those of theUnited States Government or any ageucy thereof.

M~SlERIIE-' \S UNUM\lH

O\S1R\BU1ION Of TM'S OOCU

¢

I I

Notice

This Design Repon contains confidential commercial infonnation that shall be used or duplicatedonly for official government purposes. and this notice shall be affixed to any reproduction or abstractthereof. Disclosure of the confidential commercial infonnation contained in this report outside thegovernment shall not be made without the advice of counsel. The restrictions contained in this noticedo not apply to any data or infonnation in this report which is not commercial infonnation or toinfonnation generally available to the public on an unrestricted basis.

This book was prepared as an account of work sponsored by an agency of the United Statesgovernment. Neither the United States government nor any agency thereof. nor any of their employees.makes any warranty. express or implied. or assumes any legal liability or responsibi!ity for the accuracy.completeness. or usefulness of any infonnation. apparatus. product. or process disclosed. or representsthat its use would not infringe privately owned rights. Reference herein to any specific commercialproduct. process. or service by trade name. trademark. manufacturer. or otherwise. does not necessarilyconstitute or imply its endorsement. recommendation. or favoring by the United States government orany agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflectthose of the United States government or any agency thereof.

This work was supported by the U.S. Department of Energy under contract DE-ACO-84ER4(}I50.

11

Executive Summary

This book describ~s the wn(.~ptual design of. and th~ planning for. th~ Continuous Electron B~am

Accderator Facility (CEBAF). which will be a high-intensity. continuous-~avo.:electron lino.:ar accelerator(linac) for nuclear physi~_ Its principal scientific goal is to understand th~ quark structure. behavior.and clustering of individual nucleons in the nuclear medium. and simultaneously to understand the forco.:sgoverning this behavior. In accordance with the recommendations and endorsement of the l'uclearScience Advisory Committee. CEBAF has been identified as the highest priority new construction projectfor nuclear physics.

The linac ""ill consist of 1 GeV of accelerating structure. split into two antiparallel O.5-GeV segments.The segments will be connected by a beam transport system to circulate the electron beams from onesegment to the other for up to four complete passes of acceleration. The maximum beam energy will be~ GeV at a design current of 200 microamperes.

The accelerator complex will also include systems to extract threo.: continuous beams from the linacand to deliver them to three experimental halls equipped with detectors and instrumentation for nuclearphysics research. The accelerating structure will be kept superconducting within insulated cryostats filledwith liquid helium produced at a central helium refrigerator and distributed to the cryostats via insulatedtr:lOs{er lines. An injector. instrumentation and controls for the accelerator. radio-frequency powersystems. and several support facilities will also be provided.

Prototype units of CEBAFs accelerating structure have been developed and tested both by CornellUniversity's Newman Laboratory and by industry. These five-cell niobium cavities operate at 1500 MHzand have a capacity to carry about 50 times the current required by CEBAF. All prototypes have exceededCEBAFs design specifications for accelerating gradient (5 MV/m) and quality factor (3 x 10'\

A cost estimate based on the Work Breakdown Structure has been completed. Assuming a five­year construction schedule starting early in FY 1987. the total estimated cost is $236 million (actual yeardollars). including contingency.

I I

List of Figures

Contents

VII

1. Project Overview _ _. 11.1 Project Scope _ _ _ _ _.......... I1.2 Project History __ _ _.......................................................... 21.3 Superconducting CW Linac _....................................... 51.4 Summary of Alternatives 8

2. Scientific Justification .. 112.1 Introduction.............................................................................................. 112.2 Physics Requirements for the CEBAF Design 142.3 Advantages of the Design 192.4 Highlights of the Proposed Initial Experimental Program 28

3. Superconducting CW Linac 433.1 Design Overvit:w . 433.2 Work Breakdown Structure 473.3 Beam Stability 493.4 Radiation and Safety 533.5 Technology Base 55

4. Machine Vacuum Components (WB~ 1.0)* 614.1 Introduction.............................................................................................. 614.2 Accelerating Ca,,;ties 614.3 Cryostats.................................................................................................. 654.4 Cryo-Units.. 654.5 Cryomodules............................................................................................. 704.6 Cryomodule Interconnection 744.7 Vacuum System 744.8 Cavity Procedures 764.9 Assembly and Testing 774.10 Operation................................................................................................. 80

5. Beam Transport (WBS 2.0) 835.1 Overview 835.2 Linac Lattice ,. 835.3 Recirculation Arcs 875.4 Beam SpreaderslRecombiners and Beam Extraction 1035.5 Beam Switchyard .,. III5.6 Beam Transport Magnets 115

6. RF Power System (WBS 3.0) 1216.1 Injector.................................................................................................... 1216.2 Linac RF System 127

7. DC Power System (WBS 4.0) 1317.1 Overview 1317.2 Power Supply Specifications 132

8. Instrumentation and Control (WBS 5.0) 1358.1 Control Requirements 1358.2 Beam Instrumentation 1368.3 Control System 138

• Work Breakdown Structure number.

v

9. Cryogenics (\VBS 7.0) 1419.1 Background _................................................... 1419.: System Rt:quirements 1414.3 Operating Temperature Selection 1424,4 Cycle Design 14495 Central Helium Refrigerator 1489.6 The Cryogenic Distrihution System 1499.7 End Station Cryogenic System 1519.~ System Operation........................... 1539.9 System Stability and Helium Supply Regulation 155

10. Experimental Equipment (WBS 6.0) 15710.1 Introduction.............................................................................................. 15710.2 Magnetic Spectrometers.............................................. 15710.3 Large Acceptance Detector 166lOA- Variable Acceptance Spectrometer 17110.5 Targets..................................................................................................... 17310.6 Layout of the Experimental Halls........................................................ 174

11. Conventional Construction (WBS 8.0) 17711. 1 Introduction ,. . 17711.2 Site 17711.3 Accelerator............................................................................... 18011.4 Experimental Areas 18111.5 Support Facilities 18611.6 Utilities.................................................................................................... 19511.7 Codes. Standards. and Guides 198

12. Project Management (WBS 9.0) 20112.1 Project Management Objectives 20112.2 Organization and Responsibilities 20212.3 Method of Performance 20512.4 Project Documents and Reponing 20712.5 Project Management Plan 20812.6 Manpower 21012.7 Environment Assessment 21012.8 Safety Assessment 21112.9 Quality Assurance 212

13. Cost13.113.213.3

Introduction .Cost Estimate .Contingency Analysis .

213213220224

14. Schedule '" 22514.1 Critical Path Analysis................ 225

AppendicesAppendix A National Advisory Board for CEBAF 231Appendix B CEBAF Workshop Participants 233Appendix C SURA Membership 237

vi

1.1

1.21.31.41.52.12.22.32.42.52.62.72.82.92.102.11

2.122.13

2.142.152.162.172.182.19

2.202.212.222.232.242.252.262.272.282.293.1(a)3.1(b)3.23.34.14.24.34.44.54.64.74.84.94.104.115.1(a)

List of Figures

Electron accelerators from 0.1 to 5.0 GeV maximum energy currently in operationor under construction 2Schematic of CEBAF conceptual design as of May. 1985 3Recirculating Iinac conce?t 5Schematic of present CEBAF conceptual design 6CEBAF-Cornell accelerating cavity 7Carbon nucleus. circa 1960 12Carbon nucleus. modern view 13Zirconia. as seen by an electron microscope 14Nuclear matter viewed through an imaginary microscope 16Schematic representation of a coincidence experiment 17Current profile of a pulsed accelerator 17Demonstration of why continuous beams are required for coincidence experiments 18The variables used to describe electron scattering 20Recent Saclav data for WLand WT from 40ea 21Kinematics diagram showing how Q2 and v vary for fixed E and e 22Missing-mass plot showing how high resolution can be used to rick out truecoincidences 24High-resolution coincidence data on 2O!lpb taken at NIKHEF 241.5-GHz RF wave with electron bunches designated for each of the three endstations 25The inclusive cross section for the scattering of electrons from protons 26Kinematic regions accessible to 4- and 6-GeV electrons 27Dynamics of the d(e,e'p)n reaction 28Saclay data on deuteron momentum density 30Missing-energy spectrum in Saclay )Ie(e,e'p)X experiment 31Momentum density distribution for the 2- and 3-body sectors of the )Ie wavefunction ... . . . .. .. . . . .. . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. .. . .. . . . .. . . . . .. .. . . .. . . . . . . . . . .. . . . . . . . . . . . . . . .. . . .. .. . . . 31Variables for the (e,e'p) experiment 32Diagrams illustrating how some (e,e'p) structure functions can be separated 33Illustrative spectrometer settings for (e,e'p) measurements 33Dynamics of the (e,e'2N) reaction 34Illustrative spectrometer settings for (e,e'2N) 35Quark structure of the N and ~ 36Chart of known and predicted excited nucleons 36Diagram of the photun-proton-~interaction 37Illustrative spectrometer settings for (e,e'-rrN) 37Charge distribution of the 35 shell in 206pb 40Single linear accelerator 43Multiply recirculating linac 43Schematic of the CEBAF recirculating superconducting CW linac 44Simulation of beam breakup driven by localized structure 52A pair of CEBAF-Cornell accelerating cavities 61Residual high-field Q and accelerating field 63Comparison of instability probabilities 64CEBAF cryo-unit 66End view of a cryo-unit 67Frequency tuner for the CEBAF-CorneII cavity 68Schematic plan view of a cryomodule 71Side view of the end of a cryo.i:oduIe 71Subatmospheric cryogenic transfer bayonet with guard vacuum 73Schematic of major beam-pipe vacuum components 75Cavity pair as assembled in a clean room 78Beam envelope. linac first pass 84

vii

5.I(b)5.I(c)5.I~d)

5.25.35..+(a)5.4-(b)5.5(a)

5.5(b)

5.6(a)5.6(b)5.7(a)5.7(b)5.8(a)5.8(b)5.95.105.115.125.135.145.155.165.175.185.195.205.215.225.235.245.255.26(a)5.26(b)6.16.26.36.46.56.66.7

6.8

6.96.106.117.17.27.38.18.29.19.2

viii

Beam envelope. linac second pass R-+Beam envelope. linac third pass S5Beam envelope. linac fourth pass 85Recirculating Iinac layout showing recirculation arc beam lines................................. 88Structure of. and lattice fl;nctions for. recirculation arc beam lines 91Tune per arc as a function of momentum offset ~p;p" 92Matched betatron functions as a function of momentum offset ~p/p" 93Horizontal phase space during beam transport through high-energy recirculationbeam line 94-Vertical phase space duri!1g beam transport through high-energy recirculationbeam line 94Cumulative induced ~E at start and end of each beam line 96Cumulative induced uEIE at start and end of each beam line 97Transverse phase space at end of high-energy beam line 98Longitudinal phase space at end of high-energy beam line 99Emittance increase from final beam line as a function of energy 99Induced energy spread from final beam line as a function of energy 100Dogleg path-length adjustment........... 102Schematic of CEBAF recirculating linac 103Beam extraction region 103Layout of a spreader system 105Detailed view of a spreader layout 105Layout of a recombiner system 106Deflection of microbunches by RF separator 108The Karlsruhe-CERN superconducting S-band RF separator 109Transport of microbunches after RF separator 109Side view of extraction system 110First segment of experimental area transport 112Layout of experimental area transport system 112The + I achromat concept 113Bending dipole magnet cross section 115Quadrupole magnet cross section 116Sextupole magnet cross section " 117Septum magnet cross section 118POISSON calculations of the field produced by the septum magnet 119Magnetic fields on either side of the septum 119Schematic view of the l00-keV section of the CEBAF injector 121The CEBAF injector 122Details of the l00-keV electron gun 122Water-cooled apertures for the loo-keV injector line 123Details of the rectangular subhannonic chopper cavity 124The beam pattern produced by the chopping cavity at the aperture A-3 124Calculated transverse beam profile anc ~ongitudinal phase of the beam as it passesthrough the injector 125Calculations of the phase and energy spectrum of the electron beam at the end of thepreaccelerator 126RF drive system 127Typical klystron DC distribution 128RF control module 129DC power supply locations 131Typical DC power supply circuit for dipole magnets 132Typical DC power supply circuit for quadrupole magnets 133Supervisory level. I&C "................ 138Local control level. I&C 139Total heat load as a function of temperature 143Normalized refrigeration costs ........................................................................•... 143

• •

9.39.~

9.59.69.79.S10.110.210.3lOA10.510.610.710.810.910.1010.1110.1210.1310.1410.1510.1611.111.211.311.411.511.611.711.811.911.1011.1 i12.112.212.312.412.513.113.213.314.114.214.314.414.5

Block diagram of CEBAF refrigerator 144Schematic diagram of CEBAF refrigerator _ _.. _ 146Pressure-temper;J.ture plot _ _.,. _..... 1-+8Cryogenic distribution system 150Cryomodule flow schematic _ _.. 151Transfer line cross sections 152Typical dimensions of modular elements for the spectrometers ._ 161The .+-GeV/c and 2.5-GeV/c thin target spectrometers....... 162The '+-GeV/c reimaging spectrometer _ 163The '+-GeV/c and 2.5-GeV/c moderate resolution spectrometers 163Side view of the 1.2-GeV/c spectrometer 1MSection of the 1.2-GeV/c spectrometer 1MLarge Acceptance Scintillator Hodoscope (LASH) 165Scatter plot of p", vs. 9", for F3~ decay 167The Large Acceptance Detector (LAD) 168Radial dependence of the <b component of the magnetic field in the LAD 169Time-of-flight differences for pions. kaons. and protons in the LAD 170Computer display of a single event seen by the LAD 170The Variable Acceptance Spectrometer (VAS) .. 172Time-of-f1ight differences for pions. kaons. protons. and deuterons in the VAS 173Layout of end stations A and C 175End stations A and C as they might appear after expansion 176CEBAF site plan 178End stations A and C. floor plan 182End stations A and C. elevations and sections 183End Station B 184SREL Building floor plans 187SREL Building section and elevation 188Office and Computer Center floor plan and elevations 190Central Helium Refrigeration Plant section and elevation 192CHRP floor plan 193CEBAF site AC power distribution to 12.47-kV substation 196Power distribution to facilities from substation 197CEBAF project organization chart 202Proposed management relationships and interface with DOE 203CEBAF organization chart 204Flow chart. CEBAF baseline control process 206CEBAF manpower projection (FTE) 210Cost estimate input data sheet ,. 214Obligation budget profile 223Cumulative obligation budget distribution 224Schedule logic diagram for critical path 226Timelines for critical path 227Summary network diagram 228Summary timelines 229Major milestones 230

ix

•

1. Project Overview

1.1 Project Scope

The Continuous Electron Beam Accelerator Facility (CEBAF) wil; be a high-intensity. continuouswave (CW) electron accelerator for nuclear physics. Its principal scientific goal is to understand the quarkstructure. behavior. and clustering of individual nucleons in the nuclear medium. and simultaneously tounderstand the forces governing this behavior.

To achieve this goal. the mission of CEBAF is to design. build. and operate a world-class electronaccelerator facility to serve the nuclear physics community. Implicit in the mission is the support offacility users. the conduct of an in-house research program using CEBAF and complementary facilities.and continuous improvement of the accelerator and experimental areas to keep pace with the needs ofthe users.

Scientific requirements. summarized above and described in detail in Chapter 2. define the scopeof CEBAF. It is to be an electron accelerator providing a continuous beam of electrons at any energybetween 0.5 and 4.0 GeV to one or more of three equipped experimental stations. To serve more thanone experimental station simultaneously. the current or intensity of the continuous electron beam is tobe 200 .,A.• The beam will have a duty factor of 100%. (See Table 1.1.) Finally. the accelerator is tobe completed and available for physics research in 1992 for a total estimated cost of $236 million (3ctualyear dollars).

Table 1.1CEBAF Project Scope

Ac:ceIerator SpecificationsParticleEnergyCurrentDuty factor

Additional FeaturesThree equipped experimental areasOther buildings and facilities necessary to

support the accelerator. users. and staff

Cost and ScheduleTotal estimated cost (actual year dollars)Construction period

·GeV=billion electron volts; IloA = millionth of an ampere.

electron0.5 to 4.0 GeV200 .,A100%

$236 MFY 1987-1992

CEBAF Design Report

This conceptual Design Repon describes the accelerator and its conventional facilities. Chapter 2presents the scientific justification. Chapters 3 through 9 describe in detail the conceptual design for asuperconducting CW electron linac meeting CEBAFs specifications. The report also includes a descrip­tion of a possible initial complement of equipment for the experimental areas (Chapter 10).

The conventional facilities included within the project scope are described in Chapter 11. Chapter12 summarizes plans for project management. Chapter 13 discusses cost. a.'1d Chapter 1~ discussesschedule.

The remainder of Chapter 1 relates the history and context of the CEBAF project. introduces theCW Iinac design. and considers alternative designs.

1.2 Project History

The scientific need for a high-duty-factor. high-energy electron accelerator to explore the quarkstructure of the nucleus has been recognized for more than a decade. Such an accelerator was foreseenby the first Long Range Plan for Nuclear Science (1979) developed by the DOEJNSF Nuclear ScienceAdvisory Committee (NSAC) and endorsed by subsequent panels. It would provide a capability notavailable elsewhere in the world (Figure 1.1).

t5

> 4SLAC Low-Energy o CEBAF (USA)CD Injector (USA)

">-en...CDl:CD 3E~

ExCll~

2

Darmstadt(W. Germany)

Illinois

o Mainz upgradeNBS (USA)O

(Japan). Lund(Sweden)

Bates upgrade (USA)~Bates (USA)Saclay (France)NIKHEF (Holland)

1

0% 50% 100%

•••

Duty factor •

lay> 100 pA

10 < I < 100 p Aay

1 < I < 10 l1Aay

Hollow symbols:proposed orunder construction

FIgUre 1.1. Electron acce;.;rators from 0.1 to 5.0 GeV maximum energy and a current in excess of I ~. HoUow symbols indicatemachines currently UDder construction or proposed.

2

I

Project Ove",,-jC\lo"

Recognizing this need in the mid-1970s. Prof. James S. McCarthy of the University of Virginiabecame interested in designing and proposing a new electron accelerator for nuclear physics. He convenedthe First University of Virginia Conference on Electron Accelerators in 1979. and set in motion thecollaborations that led to the incorporation of the Southeastern Universities Research Association (SURA)in August. 1980.

With a limited budget and staff. McCarthy designed an accelerator to produce continuous beamsof ~GeV electrons. In December 1980. SURA submitted a proposal for construction of this acceleratorto the DOE. The design of the SURA accelerator relied primarily or. well-proven technology. It was apulsed linear accelerator (linac). of the same design as the Stanford Linear Accelerator Center's (SLAC).with a pulse stretcher ring (PSR) to stretch the pulses into a continuous beam (Figure 1.2). The circum­ference of the PSR would be exactly equal to the length of the electron pulse from the linac. Thus onelinac pulse would fill the PSR completely with beam. As this loop of electrons would circulate aroundthe PSR it would be extracted slowly and continuously. during the interval between linac pulses. Theexperimental stations would receive an essentially continuous beam. SURA called this accelerator pro­posal the National Electron Accelerator Laboratory (NEAL).

Linac

Recirculation

Linac

Phase II

PSRJ u

End stations

B

c

L

A

•

FJgIU'e 1.2. CE&AF conceptual design as of May. 1985-a pulsed linear accelerator with pulse suere"'e! ::g. This design hassince been superseded by the superconducting CW design.

The submission of the SURA proposal triggered significant activity with in the electromagneticnuclear physics community. both to prepare a national scientific justification and to develop alternativedesigns. Much discussion centered on selecting the appropriate energy for the continuous electron bea.::ns.In the fall of 1982 an NSAC subcommittee. chaired by Prof. Peter D. Barnes of Carnegie-MellonUniversiry. recommended 4 GeV. Shortly thereafter, SURA submitted the second NEAL proposal.Argonne National Laboratory, the University of Illinois, the Massachusetts Institute of Technology(MIT), and the- National Bureau of Standards (NBS) also submitted proposals. Both Argonne's andSURA's proposals were for 4-GeV accelerators; the other institutions' designs were for lower e!!er;ies.

During the winter of 1982183 the NSAC Panel on Electron Accelerator Facilities (Prof. D. AllanBromley, chairman) reviewed and evaluated the proposals at the request of the NSF ana DOE. As partof the deliberations, the panel reconsidered the choice of energy and affirmed the selection of 4 GeV.NSAC endorsed the panel's recommendations to build a facility based on the SURA proposal. In May1983, SURA submitted a proposal to DOE for preconstruction R&D.

3

I

!!!!!

CEBAF Design Repon

Subsequently the agency and the nuclear physics community have reviewed and consistently reaf­firmed the various decisions. including the choice of energy. the priority of an electron machine. andthe selection of the SURA proposal. In this process. the name of the new facility was changed toContinuous Electron Beam Accelerator Facility (CEBAF). Fiscal Year 1987 was set as the year forconstruction authorization.

In May 1985. Dr. Hermann A. Grunder became director of CEBAF. In preparation for constructionDr. Grunder initiated a study to review the project's basic technology in light of innovations and tech­nological progress made since the design was originally proposed in 1980 and refined in 1982.

The objectives of CEBAFs Technology Review were to identify the most appropriate and cost­effective technology choices for CEBAF commensurate with an FY 1987 construction start. At thatpoint. superconducting RF technology emerged as a promising alternative to the pulsed linac with PSR.Its advantages are numerous: inherently continuous beam with 100% duty factor. improved beam quality.considerable power savings during operation. ability to deliver simultaneous continuous beams at threeenergies. conceptually simpler design (no PSR). and potential for significant upgrades in energy andoperational flexibility.

The key component of a superconducting linac is the superconducting accelerating cavity. Since theearly 19605 pioneering groups have been developing such cavities and the technology to support them.An early ieader was Stanford University's High Energy Physics Laboratory (HEPL). Now there aremany centers of superconducting RF research. Initial results were encouraging. but progress has beenslow until recently. when several breakthroughs occurred. Within the past two years. improvements inniohium quality. cavity fabrication. and processing have allowed the performance of the cavities to morethan double. Prototype cavities now consistently exceed generally accepted specifications for applicationin electron accelerators.

Within the past year. designers of two major accelerators under construction in Europe and one inJapan have elected to use this technology. These accelerators (Table 1.2) are expected to becomeoperational when CEBAF does. Industrial capability (especially in Europe) has grown rapidly. and thecompanies working with the European accelerators have built and delivered components that exceedspecifications.

Table 1.2Worldwide Plans for Cavities Produced by Industry

Number Electrical, of length Date.~ Lab Machine cavities (m) running

U. Darmstadt Linac 8 8 1986

DESY HERA (30 GeV) 16 20 1987

DESY HERA (upgrade)

CERN LEP 16 27 1988

CERN LEP II 200 340 1991/92

KEK Tristan 130 200

CEBAF Linac 400 200 1991

Reviews and workshops involving technical experts and nuclear physicists have confirmed thatsuperconducting accelerator technology is ready and appropriate for CEBAFs construction start in FY1987. In October and November 1985, CEBAF prepared a Scientific and Technological AssessmentReport (STAR) on the superconducting CW linac. From November 20 to 22, 1985, DOE reviewed theSTAR. On Decembe~ 4. 1985, the Department directed CEBAF to complete the preconceptual andconceptual designs for this machine. This Conceptual Design Report was reviewed on February 11 and12. 1986, by DOE.

4

- -Project Overview

1.3 Superconducting CW Linac

CEBAFs continuous wave (CW) linear accelerator (linac) is a straightforward way to producecontinuous high-energy electron beams. One simply injects a steady beam from an electron gun, preac­celerates it in the injector to nearly the speed of light, accelerates it to the desir~d energy in the mainaccelerator, and sends it to an experimental area.

The acceleration is accomplished electromagnetically using a metal structure to establish the electricfield gradient that accelerates the electrons. When this metal structure is copper, as is the case at existingelectron linacs, the electric field produces enormous resistive heating in the copper. The resistive heatingis so large that for practical reasons such a linac is oper.ite-d only in a pulsed mode, with brief high­current pulses of electrons separated by long intervals between pulses. However, if the acceleratingstructure were supercondueting-that is, with vanishing electrical resistance-the resistive losses wouldbe very small, and acceleration of a continuous beam would become feasible.

Some materials, for example niobium, become supercondueting at temperatures close to absolutezero. For niobium this transition occurs at 9.2 K. Since no material is a perfect RF superconductor,there are small resistive losses in a superconducting niobium accelerating structure. The quality factorQ is a measure of these losses. A superconducting structure has a very high Q; its losses are smaller bya factor of around a million than the losses in a room-temperature copper structure.

The prerequisites for a CW linac, then, are supercondueting accelerating structures, known ascavities, and a cryogenic system to keep them below their transition temperature.

In two ways, the CEBAF design differs from the simple concept sketched earlier. First, the electronbeam is passed through the accelerator four times in a process called recirculation. Second, the linac issplit into two segments so that half of the acceleration during each pass occurs in one segment while theother half occurs along the return path (figures 1.3 and 1.4). Each linac segment can accelerate the beam

A

B

c

Eight C.S-GaVlinac sections

D

I

Io"igare 1.3. Recirculating linac concept. A 4-GeV linac (A) could be divided into eight O.5-GeV sections (8) and wrapped intoa compact spiral (C). With recirculation (D), the four sections on each side are merged into one that the beam passes throughfour times.

5

CEBAF Design Repon

Li,-:ac tunnelcutaway

50-MeVinjector

(13= 0.99995)

Extracted beamto end stations

~""Linac '\.(0.5 GeV) ""

"RF separators

Arc tunnelcross section

(dia. = 3.43 m)

r2.93 m

Linac tunnelcross section

FJgUre 1.4. Schematic of CEBAF 5uperconducting CW recirculating linac.

by 0.5 GeV. The paragraphs below describe how the linac works by following electrons through themachine from injection to the experimental areas.

In the injector. a CW electron gun produces a continuous stream of electrons which is bunched tothe linac·s RF frequency (1500 MHz) and preaccelerated to 50 MeV for injection into the first linacsegment. As the beam passes through this linac segment it sees a repeating pattern of superconductingaccelerating cavities and magnets. The cavities are the basic accelerating structures. and the magnetsfocus and steer the beam.

The CEBAF linac will use a superconducting cavity designed. developed. and tested at CornellUniversity (Figure 1.5). Four prototypes of this 15OQ-MHz cavity have been built by Cornell. and allhave exceeded performance criteria now being specified for the CEBAF linac: they must support anaccelerating gradient of at least 5 MVlm at a Q of 3 x 109

• The active length of each cavity is 0.5 m.so each cavity provides 2.5 MeV of energy gain.

During FY 1986 CEBAF is working with industry to fabricate and test additional prototypes. therebyqualifying industry for cavity fabrication. The first two cavities were completed and tested in February.and both exceeded these specifications (Table 1.3).

The cavities are paired and enclosed in a cryostat. Within the cryostat. the cavities are completelyimmersed in liquid helium to keep their temperature at 2 K. RF power from CW klystrons feeds throughthe cryostat into the cavities. Four cryostats are joined together in series within a cryogenic modulecontaining eight cavities. The cry/)genic modules are linked together in a cryogenic network suppliedwith liquid helium chilled at a central refrigeration plant. The m2gnets. beam instrumentation. andvacuum attachments are placed in the warm sections between cryogenic modules. In each of the twolinac segments the beam passes through 25 cryogenic modules containing. in total. 200 superconduetingcavities to accelerate the beam by 0.5 GeV.

At the end of the first linac segment. the beam enters a recirculation arc to transport it to theentrance of the linac segment on the return path. The second linac segment is identical to the first. andcan accelerate the beam an additional 0.5 GeV to a second arc that carries it back to the beginning ofthe first linac segment for further acceleration.

6

Project Overview

Figure 1.5. A CEBAF·ComeU accelerating cavity.

Table 1.3Perfonnance of Prototype Cavities

Achieved gradient Achieved QGradient as percentage of Quality as percentage of(MV/m) design gradient factor (Q) design Q

CEBAF design109specification 5.0 3 x

Cavity #1109(vendor 1) 6.1 122% 5 x 167%

Cavity #2109(vendor 2) 7.9 158% 6 x 200%

Cavity #3109(vendor 1) 6.8 136% 7 x 233%

Electrons can make as many as four full circuits of the accelerator. Since the beam is continuous,there are simultaneously electrons at four discrete energies in each linac segment. For example, whenthe machine is set to produce a 4-GeV beam after four passes. halfway down the first segment there areelectrons at 0.30 GeV. 1.30 GeV, 2.30 GeV. and 3.30 GeV. These four beamlets must be spread apartfrom each other and put into separate beam lines to negotiate the recirculation arcs connecting the twolinac segments. In each beam line the field of the bend magnets controlling the beam's trajectory mustbe matched closely to the energy of the beam. Therefore a magnetic device, called a spreader, is locatedat the end of each linac segment to spread the beams vertically so that they can follow separate paths

7

CEBAF Design Repon

through the arcs. Each path has a lattice of magnets to steer and focus the beam it contains. At thebeginning of each linac segment. another magnetic device-a recombiner-recombines the beams sothat they can travel as one through the linac segment (Figure 1.4).

Beams at three energies can be extracted from the linac and directed to the experimental areas.One or more of these beams may be the highest-energy beam. which has made four circuits of the linac.This beam is separated from the others by the spreader at the end of the second linac segment and senttoward the experimental areas. To extract a beam at any of the other three energies present at the endof the second linac segment. a device called an RF separator is installed in each beam line just downstreamof the spreader. This device can deflect specific bunches of electrons so that one can be diverted to theexperimental areas while the next two continue into the arc for further acceleration by the linac. Thebeams directed toward the experimental areas are sent through the beam switchyard for distribution tothe appropriate experimental areas.

Table 1.4 summarizes the baseline specifications for the CEBAF linac design.

Table 1.4Baseline Specifications for the CEBAF Recirculating

Supercooduc:tiDg CW LiDac

Beam charac:temticsElectron energy E [GeV]Average current [J,LA]Transverse emittance [m-radians]"Energy spreadtDuty factorSimultaneous beamsSimultaneous correlated energies

Linac parametersConcept

Structure type

Number of linac segmentsNumber of beam passesEnergy gain per passFrequenq [GHz]Design gradient [MY1m]Design residual QNominal injection energy [MeV]Arc average radius [m]

0.5::s: E::s: 4.0200::s: 10-8::s: 2 x 10-4100%33

RecirculatingCW linac

SuperconductingRF cavity

241.0 GeV1.55.03 x 109

5080

·Emittance = phase space area containing 95% of beam. divided by 'ff

tFWHM (full \l,idth at half maximum)

1.4 Summary of Alternatives

The CEBAF Preconstruction Technology Review evaluated alternative technological approachesfor meeting the CEBAF Scope. The results of this review are described in the workshop report fromthe Technology Review. Three approaches were considered viable: the superconducting CW linac. thepulsed standing wave linac with pulse stretcher ring (PSR). and the pulsed traveling wave linac withPSR. Other accelerator designs are unable to deliver the current or the energy specified in the projectscope.

The original design for CEBAF was a recirculated, pulsed traveling wave linac with PSR. TheTechnology Review judged this design to be sound and able to meet the specifications. However. it doesnot represent the latest design concept for high-average-current linacs. Conservative in most aspects,

8

Project Overview

this design has challenging RF requirements: klystrons with 4O-MW peak power and 13O-kW averagepower.

Technology Review participants (Appendix B) considered designs using alternative room-temper­ature structures. The most cost-effective solution uses a modem standing wave structure and 42 klystronsof 27 MW peak power and 60 kW average power.

The pulsed linac with PSR achieves a duty factor around 80%. Its upgrade potential is limited to 6GeV by the size of the PSR; at higher energies quantum excitation destroys the beam quality by greatlyincreasing the emittance.

Table 1.5 shows the comparison between the pulsed linac with PSR and the CW linac. Both designsmeet the original project requirements. of course; but as can be seen. the CW approach is superior inseveral ways. It provides a macroscopic duty factor of 100%; that is. it provides a truly continuous beamwhen the accelerator is on. It has better beam quality. as measured by emittance and energy spread. Itoffers experimenters simultaneous CW beams at different (but correlated) energies for each of the threeexperimental stations.

Table 1.5Comparison of Pulsed LinacIPSR and CW Linac

Duty factor

Emittance (-rrm)

Energy spread (~EI£)

Simultaneous CW beams at different energies

Design energy (GeV)

Possible upgrade energy (GeV)

Current (~A)

Amount of accelerating structure (GeV)

Number of recirculations

Passes through linac

Pulsed/PSR

>80%

2 x 10-7

2 X 10-3

1

0.54

-6

200

2.0

1

2

CW Baseline

100%

<10.8

<2 x 10-4

3

0.54(6)

-16

200

1.0

3

4

The superconducting CW linac is also superior to the pulsed linac with PSR in terms of maximumenergy-both design energy and possible upgrade energy. With the conservative component specificationsselected to ensure meeting the scope with superconducting technology, the CW linac may exceed the4-GeV energy specification by as much as 50% shortly after commissioning. Moreover, it has substantiallymore potential for upgrade in energy. Whereas the pulsed linac with PSR was expected to be able toreach 6 GeV after upgrade (its energy is limited by tlle circumference of the pulse stretcher ring), theCW linac is expected to be able to reach significantly higher energies when accelerating cavities achievinghigher gradients become available.

This upgrade potential is important. An accelerato~ facility is a major, long-term investment for ascientific field. Thus the scientific community and the accelentor designers must anticipate a desire tomodify the facility to keep pace with future scientific interests of the community. In the case of CEBAF,nuclear physicists have already expressed an interest in using polarized beams, and one can foresee thatthey eventually may want to push to higher energy, increase the number and variety of end stations,and use positrons. The extent to which the CW linac permits cost-effective upgrades is a clear advantagein the selection of technological approach.

2. Scientific Justification

2.1 Introduction

The study of nuclear physics began in 1911 when Rutherford observed the scattering of alpha particlesfrom a thin gold foil and thereby discovered that the majority of the matter and energy contained in anatom was concentrated in a tiny center 100,000 times smaller than the atom itself. The discovery of theneutron in 1932 confirmed that nuclear matter (made of protons and neutrons, collectively known asnucleons) is fundamentally different from atomic matter. and that a new force was needed to explainthe properties of nuclei. Yukawa's conjecture in 1935 that this force was transmitted by a new, unobservedparticle was confirmed by the discovery of the ..-meson (or pion) in 1947.

During this period. nuclear and particle physics followed the same track, and it appeared that itmight be possible to understand the nuclear (or strong) force from first principles in the early 1950s.Ho....ever, pion-nucleon interactions turned out to be more complex than people thought, and the dis­covery of the first excited state of the nucleon, the delta (6.) resonance with the mass of 1232 MeV,"and the subsequent discovery of large families of "elementary" particles led to a separation betweenparticle and nuclear physics. In the interval between the mid-1950s and the late 1970s, nuclear physicistsconcentrated on studying the systematics of nuclei, developing the theory of many-body systems, andinventing sophisticated models capable of treating the complicated dynamics of a strongly interactingsystem. (See Figure 2.1.)

Among the major achievements of this period were the development of the meson theory of nuclearforces, calculation of the properties of the simplest two- and three-body nuclei from this force, directexperimental measurement of these properties to very high precision, and the theoreti~ developmentand experimental confirmation of both the shell structure and collective effects in nuclei.

The discovery of quarks in the mid-1960s, and experimental confirmation of their existence in themid-1970s, opened a way to understand the bewildering zoo of "elementary" particles. It is now believedthat nucleons and mesons are themselves made of quarks, and that the force between quarks is generatedby the exchange of a particle called the gIuon. This fundamental force is described by the theory of colorinteractions, known as quantum chromodynamics, or QCD. These discoveries are having a dramaticimpact on nuclear physics.

While there is now a unique opportunity to develop a fundamental theory of nuclei based on quarksand QCD, the path which should be followed to achieve this goal is far from clear. The forces describedby QCD are very simple at extremely high energies, where they are weak and where perturbativetechniques ~em to work. but they are not simple when applied to the study of particles and nuclei. Herethe confinement forces. which are responsible for confining quarks into colorless triplets (such as neutronsand protons) or quark-antiquark pairs (such as the ..-meson) are very strong and playa very important

• Uni:s of energy used in this chapter are multiples of an electron volt (eV). which is the energy a panicle with the charge of aneleeuon acquires when accelerated through a potential of one volt. Commonly used multiples of this unit are keV (1000 eV),MeV (1000 keV), and GeV (1000 MeV). Mass and energy are related by Einstein's famous relation E = me:. so that masseswill also be given in energy units.

11

I

I

CEBAF Design Report

FJgIU'e 2.1. A picture representing the carbon nucleus (six protons and six neutrons) as it was viewed during the 19505 and the19605. The protons and neutrons were known to have size and structure. but details of this structure were unknown and it wasnot incorporated into nuclear models. The exchange of pions was known to be an important pan of the nuclear force; the overlapof two nucleons W'1S not believed to involve new pbenomena essential to the understanding of nuclear structure.

role. Confinement seems to be a consequence of the theoretically beautiful but mathematically compli­cated gauge structure of QCD, and a detailed understanding of the role of confinement in nuclear matteris a challenge of the first magnitude. It appears that when nucleons are far apan, their constituent quarksremain with the parent nucleon, and do not mingle (see Figure 2.2). In this case the nuclear force istransmitted by the exchange of Yukawa's pion, a mechanism not currently understood in the languageof QCD. However, when the nucleons get close together. the confining forces may weaken. allowingthe quarks to modify their configuration and dynamics. These modifications may include expansion ofthe nucleon volume, percolation of quarks from one nucleon to another, or, when the nucleons overlap,the formation of entirely new composite six-quark systems, also shown in Figure 2.2. The meson theoryof nuclear forces becomes very complicated at such short distances, and probably no longer correctlydescribes the interaction; this is the region where it is hoped that a description based directly on quarksand QCD will be simpler, and will permit the development of a new fundamental understandir.g of thenuclear fo:-ce. Because nucleons sometimes overlap in nuclei, nuclear matter is a good laboratory inwhich to study these short-range phenomena.

The fascinating opportunities offered by these developments have been recognized in the LongRange Plan prepared by the Nuclear Science Advisory Committee (NSAC) in 1983. The understandingof the nucleus in terms of quarks. and the use of the nucleus as a laboratory for the understanding ofQCD, were stated as major long-range goals. For the investigation of this quark-nucleus interface, NSAChas assigned the highest priority to the construction of a continuous electron beam accelerator with anenergy of 4 GeV. Complementary tools. in particular a relativistic heavy-ion collider identified as thenext major construction project after the electron accelerator, and a kaon factory. were discussed aswell.

12

Scientific Justification

FIgW'e 2.2. A picture representing a modem view of the carbon nucleus. Each proton has two "up" quarks and one "down"quark. while the reverse is true for the neutron. Gluons (represented by corkscrew lines in the figure) are exchanged between thequarks inside the nucleon. When the nucleons are far apart. th::y sometimes exchange To-mesons. which contain a quark and anantiquark (not shown). It is believed that nucleons sometimes get excited in the nuclear medium: one such excitation. the ~. isshown (a deformed state "'ith three ··up·· quarks). Finally. when two nucleons overlap they may form new states of mattercontaining six quarks. labeled with the question marks. One of the missions of CEBAF is to study these new states.

The new electron accelerator must be a research device with remarkable power. It must be able tostudy nuclei down to distances of a fraction of a nucleon diameter. and must give experimental datawhich are precise and easy to interpret so that the complexities of the nuclear system can be unraveled.CEBAF has been designed to meet these demanding requirements.

CEBAF will be an electron accelerator capable of producing a high-intensity. continuous beam ofelectrons of any energy from 500 MeV up to 4 GeV. and beyond. No accelerator exists anywhere in theworld capable of producing high-intensity. continuous beams in this energy range (see Figure 1.1). Thesecontinuous beams will open up a major new program of research. and will make CEBAF a unique.world-class facility for basic nuclear physics research.

In broad general terms. the principal scientific mission of CEBAF is to study the extent to whichindividual nucleons change their size. shape. and quark structure in the nuclear medium. to study hownucleons cluster in the nuclear medium. and to study the force which binds quarks into nucleons andnuclei at distances where this force is strong and the confinement mechanism is important.

The scientific reasons for building CEBAF have been summarized in a number of reports andproposals prepared over the last five years. and are also discussed in Research Program at CEBAF, thereport of the CEBAF 1985 Summer Study Group. The sections below are a self-contained summary ofthese considerations. Section 2.2 explains in general terms why CEBAF is an ideal tool for the study ofthe quark structure of nuclei. Section 2.3 describes the significant scientific advantages of the CW design.and Section 2.4 describes highlights of the extensive scientific program proposed for CEBAF.

13

CEBAF Design Repon

2.2 Physics Requirements for the CEBAF DesignElectrons as Probes

Using an electron microscope. it is possible to look deep inside of ordinary matter and see theunderlying molecular and atomic structure. The smallest objects which can be seen with an electronmicroscope are individual atoms. about 0.2 nm- in diameter. considerably smaller than the wavelengthsof visible light (which vary from 'ibout 400 to 800 nm). Figure 2.3 shows a picture of zirconia (ZrO:!) in

FJgUn 2.3. Picture of zirconia taken with the Atomic Resolution Microscope. an electron microscope at Lawrence BerkeleyLaboratory (counesy of Lawrence Berkeley Laboratory).

a regular solid array taken with an electron microscope using I-MeV electrons.(!) To see the nuclei atthe center of these atoms requires electrons with much higher energies.

The electron is an ideal probe for nuclear matter in three related ways. First. it is. as far as is known.a point particle with no internal structure of its own. Any structure observed in electron scattering istherefore related to the structure of the nuclear target and cannot be confused with the structure of theprojectile itself.

A second useful characteristic of the electron is that it interacts only with quarks. and this interactionis through the electroweak force. which is the best understood force in nature. Since nuclear matter isthought to be an extremely complicated soup composed of about equal parts of quarks and gluons. it isinvaluable to know that the electron "sees" only the quark content. Hadronic probes. such as n-mesonsand protons. interact with both quarks and gluons and give complementary information. Experimentswith hadrons are more difficult to interpret. however. given the internal structure of these probes.

A third characteristic of the electron is that it interacts with matter very weakly. so that an electronis likely to scatter only once as it passes through the nuclear medium. This scattering is through theexchange of a single quantum. a virtual photon. Theoretical ideas based on single-scattering events workvery well. so that the results of electron scattering experiments ·can be interpreted with clarity and

• Units of distance used in this chapter are nm (lO-~ meter). and fm (10-" meter).

14

•

-... Scientific Justification

precision. By selecting the electron scattering angle and energy loss. the experimenter can carefully adjustthe energy. momentum. and polarization transferred to the nuclear target by the virtual photon. so thatthe resolving power (wavelength) and the extent to which the electron excites the target can be inde­pendently controlled. Finally. since a single scattering is just as likely to take place deep inside thenucleus as on its surface. the electron probes the entire nuclear volume.

Need for High EnergyIn common with all matter. an electron has an intrinsic wavelength which decreases as its energy

increases. Since details much smaller than the wavelength cannot be resolved by any optical instrument.a high energy is needed to study very fine details. The effective wavelength of 100-MeV electrons isabout 3 fm. just sufficient to "see" the sizes and shapes of nuclei. as illustrated in Figure 2.4. Investigationsof nuclear size were initiated with accelerators in this energy range at Stanford Ul~iversity and theUniversity of Illinois in the 19505 (actually. by 1956. the Stanford accelerator had reached an energy ofabout 600 MeV). and this research opened up a new field of nuclear structure studies. But to study theconfiguration of the nucleons which make up the nucleus. it is necessary to resolve distance scales stilllower. to somewhat less than the nucleon diameter of 1.6 fm. For these studies. electrons with energiesof about 500 MeV. which have an effective wavelength of about 0.6 fm. are required~ such programsare currently being carried out at the Bates Laboratory at MIT. Saclay in France. and NIKHEF in theNetherlands. Finally. study of the underlying quark structure of the nucleus requires that even smallerdistance scales be resolved-scales much smaller than the size of neutrons and protons. This requiresenergies in the multi-GeV range. the range planned for CEBAF.

What energy must CEBAF reach in order to successfully carry out a study of the quark structureof nuclei? This important question has been discussed extensively by the scientific community since thelate 19705. when plans for a high-energy continuous electron beam accelerator began to take shape. Thequestion was addressed by the Barnes subcommittee of NSAC in 1982. and by the Bromley panel in1983. Both groups recommended an energy of about 4 GeV. The ad hoc NSAC subcommittee chairedby Eric Vogt in 1984 also endorsed the choice of 4 GeV. This choice is a compromise between thescientific need for electrons with short wavelength (high energy). and the cost of the accelerator facility.which grows with energy.

The CEBAF design meets the scientific requirements by providing a continuously variable beamfrom 500 MeV up to 4 GeV. The ability to produce beams below 1 GeV is important in providingcontinuity and overlap with previous and proposed new lower-energy programs. and the capability togo to 4 GeV is necessary to carry out investigations into the quark structure of nuclei. A more detaileddiscussion of the physics to be done in this energy range is given in sections 2.3 and 2.4 below.

Need for a Continuous BeamWhen a high-energy electron scatters from a nuclear target. the nucleus reacts in one of two general

ways: sometimes the nucleus recoils as a whole. remaining intact (either in its ground state or an excitedstate); and sometimes it breaks into two (or many) fragments. While it is sometimes possible to identifyspecific individual events in the first class of reactions by detecting the scattered electron only. distin­guishing individual events in the second class of reactions usually requires that one or two of the nuclearfragments be observed in coincidence with the scattered electron. Coincidence experiments are a powerfuland general way to study the details of specific reactions.

A coincidence experiment is most easily done with an accelerator that has a continuous beam. Sincenone of the existing accelerators with energies in the range of a few GeV have continuous. high-intensitybeams. only a few coincidence experiments have been carried out. Thus. CEBAF will open an entirelynew program of experimental study.

To see why continuous beams are necessary for coincidence experiments. consider the simplest suchexperiment. illustrated in Figure 2.5. in which one spectrometer observes the scattered electron while asecond spectrometer observes the scattered nuclear fragment (which will be taken to be a proton in thefollowing discussion). If these two observations occur within some minimum time difference -r (theresolving time). they are simultaneous within the experimental uncertainty. However. it is possible forelectrons from one event and protons from a different event. separated by times less than T. to beconfused with true coincidences. in which the proton and electron come from the same event. Suchoccurrences are referred to as accidental coincidences. or accidentals.

15

CEBAF Design Repon

I- 10 fm • I 2fm -I

...1-----0.2 fm I- 0.05 fm • I

F"JIUn 2.4. An anist"s conception of nuclear matter viewed through an imaginary microscope with larger and larger magnification.In the first picture (field of view of 10 fm) the size and shape of the nucleus can be seen. a:ld the motion of its neutrons andprotons in a ··mean field·· created by the other nucleons can be studied. At larger magnification (field of view of 2 fm) the exchangeof pions between neutrons and protons can be studied. and some hint of the underlying quark structure may begin to appear. Ifthe magnification is ten times larger (third figure). six-quark states resulting from the overlap of two nucleons may be studied.Here g1uon exchange between the quarks (represented by the corkscrew lines) will be imponant. and complicated effects such asthree·g1uon interactions (shown at the top) and the production of quark-antiquark pairs (shown at the bottom) may be imponant.This is the region that CEBAF will study. Finally. the founh figure shows that unexpected new IJhenomena may be discoveredby a further fourfold increase: in magnification. such as would be possibl' with a future upgrade in the energy of CEBAF.

16

•Scientific Justification

•

Incidentelectron-------

~/~~d I, electron

., 7 (Resolving time)

FJgIIn 2.5. Schematic representation of a simple coincidence experiment. The scattered electron is detected in spectrometer 5,.and one of the hadronic fragments. a proton. is detected in spectrometer 5:. In this panicular example. two other badronicfragments are not detected in any spectrometer. Detection of the electron and proton is simultaneous if these observations occurwithin the resol'ing time ':'.

To avoid confusing accidental coincidences with true coincidences. the experimenter will usuallyneed to tum the beam current down so that the ratio of accidentals to true coincidences is not muchgreater than one. This is the maximum current which can be used. If the beam is continuous, data canbe collected continuously at the rate detennined by this maximum current. If the beam is pulsed, asshown in Figure 2.6. data can be collected only during the (short) time interval (t in the figure) duringwhich electrons are actually striking the target; during the remainder of the cycle the equipment muststand idle waiting for the next pulse. It will therefore take significantly longer to perfonn a givenexperiment with a pulsed accelerator than it would with a continuous accelerator.

T ----·-1....---- T ~I

FJgIIn 2.6. Time dependence of the current in a pulsed accelerator. The pulse of electrons arrives during the (shon) time intervalI, which is repeated every T seconds. (The duty factor in this example is about 15%. The duty factor describes the extent to whichthe beam is pulsed.)

If the experimenter tries to get around this limitation by increasing the current, the number ofaccidentals will increase faster than the number of trues, making it difficult to pick out the true coinci­dences. In popular tenns. if the current is too high. the detectors will be "blinded" and thus not be able

17

CEBAF Design Report

to "see" the coincidences. Figure 2.7 gives a visual demonstration ofthis fact. In Figure 2.7(a) the beamis continuous. and the current has been adjusted so that the true coincidences can be di~tinguishedfromrandom events which take place in each detector. In Figure 2.7(b). the experimenter has a pulsed beam.and has tried to circumvent the limitation by turning up the current so that the same number of eventsare collected during the shon time interval (I in Figure 2.6) during which electrons strike the target. Ineach case there is one true coincidence. but in the second case the true signal is swamped by the accidentals.

I

(a)

Electrons

Protons

(b)

Electronsj1J'-----_~~Protons mr IT]

..Time

-

Resolving time h)

FJgUI'e 2.7. In both figures time iocrcascs uniformly from left to right. and tho: vertical lines mark the times at which electronsor protons arrive at tile spectrometers. The resolving time (T) of the system is shown. In the upper figure (a) the beam is continuous(100% duty factor). and eight electrons are detected in the electron spectrometer and nine protons in the proton spectrometerover the time interval shown. Most of the counts arrive at random times. but there are one true coincidence and one accidentalcoincidence. In the lower case (b) the duty factor is 10%. and the experimenter has tried to circumvent the limitation discussedin the text by increasing the current so that the detectors will receive the same number of counts in the shorter time intervals.Note that the closer spacing between the counts makes it harder to distingcish the true coincidences from the accidentals. andthat withio the resolving time assumed in the figure several accidental coincidences occur.

The duty factor measures the extent to which the beam is pulsed. It is the ratio of the duration ofthe pulse to the time between pulses (lIT in Figure 2.6). The discussion in the previous paragraph leadsto the quantitative conclusion that. if all other variables are fixed. the time required for a coincidenceexperiment is inversely proponional to the duty facto!'. Therefore. only coincidence cross sections whichare very large can be measured with an accelerator with a duty factor as low as 1-2%. A successfulprogram of such coincidence measurements is being carried out at Saclay and at NIKHEF. which haveduty factors in this range. but lower energies. The 100% duty factor at the energy of CEBAF is neededto extend these programs into the region where nucleons overlap and coincidence cross sections are low.

Need for Higb CurrentIn common with all microscopic interactions. electron-nuclear interactions are subject to statistical

fluctuations. The error in any measurement is proponional to llYN. where N is the number of eventsobserved. so that 10.000 events must be collected in order to achieve a statistical accuracy of 1%. Becauseof the weakness of the interaction between electrons and nuclear matter. it generally takes a long time(several months) to collect this number of events. The actual time required for a specific experimentdepends on many details. but if all other variables are held constant. it is proponional to the numbetof electrons which strike the target per second. This number in tum is directly proponional to the current(sometimes referred to as intensity) produced by the accelerator. The CEBAF design of 200 microampscorresponds to 1.2 x 1015 electrons per second. This current is several times greater than currentstypically produced by other electron accelerators. and is sufficient to supply a few end stations with thehigh-intensity beams needed for the study of rare events.

18

I

,; .

{- .-

Scientific JustificatiOD

Need for Three End StatiODSThe productivity and flexibility of the CEBAF physics program will be directly related to the number

of end stations (the experimental halls where the scattering takes place) which can be used simultaneously.Even with the high currents CEBAF can provide, typical experiments will take several months to completeonce they are set lip, and they sometimes will require a similar amount of time to set up. It is th~refore

extremely important to be able to distribute the beam simultaneously to more than one end station, sothat a number of experiments can be done at the same time, or so that one experiment can be set upwhile another one is running. The choice of three end stations is a compromise between the need forhigh productivity and the need to contain costs.

2.3 Advantages of the Design

This section uescribes several scientific advantages of the CW linac design described in this report.Briefly. these advantages are in the areas of

• simuitaneous continuous beams wit.ll correlated energies,

• high beam quality,

• low power cost.

• stable low-current beams, and

• possibility for significant upgrade in energy.

The following subsections discuss each of these advantages in general terms, and then with moretechnical detail (which can be skipped on a first reading). These di~ionsalso serve as an introductionto some of the scientific material in Section 2.4.

Simultaneous Continuous Beams with Correlated EnergiesThis CW lin2.c can deliver continuous beams to each of the three end stations simultaneously, and

will allow the experimental group in each end station some latitude in choosing their beam's energy.Specifically. if the maximum energy being delivered to any of the end stations is E rrwx , one beam willuse ErroJJ<, another beam can use Ern:u., 3/4 ErroJJ<, 112 ErroJJ<, C'r 1/4 ErroJJ<, and the third beam can use eitherE~ or whichever of the three fractions of Emu the second beam is not using. Thus, experimenters canchoose from among various combinations of correlated beam energies, and furthennore can have anyof the three selec::ed energies delivered to any of the three end stations.

When the linac is delivering a 4-GeV beam to one experimenter, a second experimenter has thefreedom to select the same energy, or choose one of the lower (but correlated) energies of 1, 2, or 3GeV. As it turns out, this freedom is almost as useful to the experimenter as would be the freedom tochoose the second energy completely independently of the first. This is because the exact energy of theincoming electron is not as important as the energy (and momentum) of the virtual photon transferredby the electron, and since the latter may be varied continuously (over a limited range) even when theincident energy of the electron is fixed, the correlated energies a,,·ailable offer a choice broad enough sothat almost any experiment can be done with one of them. Hence, the freedom of the second experimentalgroup to choose anyone of the correlated energies allows it to conduct an experimental program thatis essentially independent of that being undertaken by the first experimental group, almost as if eachexperimental hall were fed by its own accelerator. Furthennore, a tlird experimental group may als..>choose 4 GeV, or any of the lower energies not selected by t!le second experimental group. This freedom,while somewhat less than that enjoyed by the first and second experimenters, is still quite significant. Insum, then, the flexibility in choice of energy which can be supplied to each of the three experimentalhalls significantly increases the productivity of the overall program at CEBAF.

To explain why the freedom to vary the energy over a wide range is important, and why the freedomto choose any particular energy is less important, it is necessary to introduce variables used to describeelectron scattering; these variables are shown in Figure 2.8. The incident eleetton has energy E andscatters, in the laboratory, through angle 9 to final energy E'. Two combinations of these three inde­pendent variables are of special importance. The first is the energy v lost by the electron in the laboratorysystem,

v = E - E'

• I

(2.1)

19

•

CEBAF Design Repon

The second is the square of the four-momentum transferred by the electron through the virtual photonto the hadronic target. which is

(2.2)

where q is the momentum of the virtual photon. Note that Q~ is )lQsitive. It is sometimes convenient torep!dce v by W. the square of the relativistic mass of the final state. which depends on Q~ and v

(2.3)

where M T is the mass of the target.

E--~

Figure 2.8. A schematic diagram showing the imponant variables in inclusive electron scattering. The electron (represented bythe dashed line) emits a vinuaJ photon (represented by the wiggly line) which is absorbed by the hadronic target (with four­momentum Pl. When the high-energy photon strikes the target. it breaks it into fragments. which have a total mass w.

A fundamental quantity measured in inclusive electron scattering is the differential cross section.which is pro;x>rtional to the probability that a given nuclear reaction will take place. The differentialcross section for inclusive electron scattering. in which only the scattered electron is detected. is

~_!!H..[2 ~ ]dE'df! - MT

VL W L (Q .v) + V T WdQ • v)

where

(2.4)

2VL = P

VT = !p + tan2!9

(2.5)

The scale for the scattering is set by the Mott cross sectio~

(2.6)

and W L and W T are the longitudinal and transverse hadronic structure functions. The quantity Cl inEquation 2.6 is the electromagnetic fine-structure constant. which is 10 to 100 times smaller than thecorresponding constant in the strong interactions. and is the ;>rincipal reason for the small size ofelectromagnetic cross sections. The formula for exclusive (coincidence) measurements will be given inSection 2.4 below; it is similar to Equation 2.4 but involves more structure functions.

All the information about nuclear structure which can be obtained from inclusive electron scattering

20

•

Scientific Justification

(2.7)

is contained in the hadronic structure functions. For example. the longitudinal structure function. WL.

is a measure of the distribution of charge in the nucleus. while the transverse structure function. WT • ismore sensitive to currents and magnetization densities (spin). Note that these (unknown) functions depend(in the one-photon-exchange approximation) only on the twe independent variables Q2 and 11. and thatthe dependence of the differential cross section on the third variable, 9. is known (the e dependence isisolated in the known functions UM. VLand V T). If the cross section is measured only at forward angles,where 8 is small, only one combination of these two functions is obtained, which is called W2

W2(if,lI) = ~T [p2WL(if.v ) + !p WrtQ2,1I) ]

From this function alone it is not possible to trace the origin of any discrepancy between theory andexperiment to either W L or W T. By measuring the differential cross section at the same (Q2,V) point butat cwo different values of e, both WLand WT can be separately detennined. Recently this has been donefor several nuclei; data for 40ea taken at Saclay(2) are shown in Figure 2.9. When these data werecompared with theory, unexpected discrepancies in the longitudinal structure function were observed,leading to a flurry of recent speculation that the charge distribution of bound nucleons inside nuclei issubstantially different from when they are free.

Iql=L.10MeV/c

a

0.02

0.08

0.06

0.0L.

o~:'-_I.-_-""--_......L-__....L.__"""""

so 100 150w{MeVJ

200

40Ca(e.e·) b

0.06

0.04

0.02

so 100 150

Iql=410MeV/ c

\ 1+1t.200 w(MeV)

FJgUre 2.9. Recent data from Saclay (as plotted there; see Reference 2) for WT (R T in the figure) and WL (RL in the figure).The solid and doned lines are theoretical calculations of these quantities. Note that the theory does an adequate job of explainingW T but does not do weD with the longitudinal structure function. W,.. (The quantity ... is the same as the quantity 1/ defined inEquation 2.1.)

21

•

CEBAF Design Repon

These recent experiments show that it is very important to independently measure all structurefunctions similar to W L and W r at a sufficient variety of points in the (Q:!.v) plane so that the cluesnecessary to unravel the underlying structure of the target will be available.

To separate W L and Wrat a particular (Q:!.v) point. the precise values of eat which the cross sectionis measured are not important; it is sufficient to choose one value of ewhich is small (forward scanering)and one value which is large (backward scanering). Given a choice of any two of the four correlatedenergies £ = 1. 2. 3. or 4 GeV. it is alwa.,ys possible to choose two such angles. as shown in Figure 2.10.This figure shows those points in the (Q-.v) plane which can be measured using the illustrative anglesof 30° and 150°. (To reach points in between those shown with the same choice of correlated energies,angles somewhat different from 30° and 150° must be chosen.)

6.0 ,......---.~"'T""T--"'T""""'T'T-.,...-....---r----r----,

6.05.04.0

III 9=300I

~ - -- - 8=1500

IIE=4I

3.02.01.0

3.0

0.0 L-_-=o&....-_~::.L-__;:>L...__~__.....1..__~

0.0

5.0

2.0

1.0

4.0­N-U->GJo--

N

a

v (GeV)

rJgUl"e 2.10. The solid diagonal lines show how if and v vary when the incident electron energy E is fixed and the electronscattering angle 9 is 30". The dotted lines show the same variation for fixed incident electron energies and a fixed electron scatteringangle of 150". The figure shows. for example. that separation of the structure functions in the vicinity of the point (if.v) = (2.5.1.6) requires measurements with a 4-GeV incident beam at a 30" scattering angle and a 2-GeV incident beam at a 1500 scatteringangle.

High Beam QualityTwo measures of the quality of the beam produced by an accelerator are the energy spread, 13..£1£,

and the transverse eminance, which is the product of the spatial width of the beam and its angulardivergence. The eminance remains constant as the beam is focused or defocused through the beamswitchyard, so that if two beams have the same diameter, the one with less angular divergence will havethe smaller emittance. The beam produced by the CW linac will have excellent quality; its energy spreadwill be about 10- 4

, and its emittance will be about 10- 9 m-rad.Briefly, the advantage of the small energy spread is that it expands the overall capability of CEBAF

in two ways. Measurements which require the resolution of closely spaced energy states become morefeasible and. in cases where the unobserved final state is narrow, coincidence experiments can be extendedinto regions where true coincidences cannot be separated from accidentals by timing methods alone. Theadvantage of low eminance is that it is easier to deliver a beam to the end stations without an accompanyinghalo of electrons, so that backgrounds in the end stations are reduced, less shielding is required, and

22

Scientific Justification

greater flexibility in the design and placc::ment of detectors is permitted. The improved emittance andlower energy spread of the CW linac will also reduce the cost of the experimental equipment to be placedin the end stations.

The (elation between resolution and energy spread deserves funher discussion. The energy spreadis directly related to the precision with which the masses of unobserved final states can be determined.For inclusive scattering.

EE'~w = M - ~E' - - sinO ~O

W(2.8)

where ~E. j.E'. and ~O are the experimental uncertainties in E. E'. and O. This approximate relationshipholds for low-lying states excited from massive targets. Table 2.1 shows how small j.W mast be in orderto allow study of various physical systems. To resolve excited states of heavy nuclei. it would be desirableto be able to measure W to an accuracy of tens of keY. which requires that the energy spread of a 4­GeV electron beam be about 10-5

• and that the final electron energy and scattering angle also be veryprecisely measurec.

Table 2.1Typical Energy Differences in Nuclear Physics

System ~w

Proton-j. mass difference

Neutron separation energies

Typical level spacings near closed shells

Typical level spacings in rotational bands

Closely spaced levels in heavy nuclei

300 MeV

1-10 MeV

0.3-2 MeV

100-300 keY

-10 keY

Such precisions are very difficult to achieve with electrons in the GeV range. and for many exper­iments such a good resolution is not needed. However. the better the resolution which can be achievedthe broader the physics program which can be undertaken. and it is therefore desirable to achieve thehighest resolution possible for CEBAF and its spectrometers. The (-y.K) program discussed in Section2.4 is an example of a proposed CEBAF experimental program which could benefit from the highestpossible resolution.

A number of techniques (such as the use of energy compression systems. dispersion matching, orintermediate focusing methods) exist for achieving high resolution in W starting with medium resolutionof the incident beam. These techniques can be applied to any accelerator. but are expensive and oftendifficult to implement. The advantage of the CW accelerator is that its higher beam quality means thatfe'Yer of these costly techniques are necessary to achieve a given resolution. or alternatively. that em­ployment of such techniques will produce a system with still higher resolution.

Another advantage of high resolution is that it also increases the extent to which the true coincidencescan be seen above the background of accidentals for cases in which the unobserved hadronic state isvery narrow. The experiment d(e,e'p)n illustrates this point.· The unobserved neutron is a narrow state.If this process is studied in a region of phase space where it is very improbable. the true coincidencesmay be very hard to distinguish from the accidentals, even with a 100% duty factor. In this case theseparation of trues from the accidentals can be improved by plotting the data as a function of the missingmass. which must equal the neutron mass for the true coincidences but can be arbitrary for the accidentals.In such a plot, shown in Figure 2.11. the accidentals form a smooth continuous background. while thetrue coincidences will appear as a peak centered around the neutron mass. M". The width of the neutronpeak will depend on the resolution of the system. The area under the true neutron peak is the same inboth cases. so that if the resolution is low. the peak is fairly broad. but if the resolution is high the peak

"The notation A(e.e'x.lIjZ denotes the reaction e ... A - e' .,. x .- .v .,. Z. and indic..es that the scattered electron e' and thefinal-state hadronic particles x and.v are all observed in coincidence.

23

I.~

•

CESAF Design Repon

(a)

II

~(b)

1Mn

F"1gIII"r 2.11. In case (a). the missing-mass resolution ~W is large. so the neutron peak is broad and not easily distinguishablefrom the background. In case (b). the missing-mass resolution is small. and the neutron peak is clearly discernible.

will stand up clearly above the background. High-resolution techniques have been used successfu!.!}: atNIKHEF to enhance their program of coincidence experiments. Figure 2.12 shows some recent """"pbdata,(3) where the resolution in missing energy is about 100 keV. This is sufficient to separate the narrow,closely spaced peaks which arise from individual shells, and to measure the structure of each shellindependently. For a discussion of how such data are interpreted, see Section 2.4.

208 Pb (e,e'p) 207 Tl

NIKHEF 1985

~ t·c~

::.....1III...-:a...1III

ECl.

ELI.I

en

zQ....uz~

"-

...J

-<a:....uW11.en

3.'/2

J~..~ Pm. UJMev/e.~

~

e...-~,;; (0.35) (1 65)

o 2

Ex (MeV)

Pm • 11)0 MeVIe

2dSn

3

3. '12

P",. 220 MeY Ie

(1.35 )

1h~2

F"1gIII"r 2.U- Recent NIKHEF data as ploned in Reference 3. High resolution enables the experimenters to distinguish dil1erentstates of the final =11 nuclear system. corresponding to knockout of protons from different shells in the lead nucleus. The fourpeaks which can be seen lie along the four dashed diagonal lines in the figure. The dependence of the peak height on the momentum.P_. enables the experimenter to determine the spatial shape of the shell (see the discussion in Section 2.4).

24

I

Sci('ntific Justification

Low Power CostSince AC power costs are a major operating limitation of existing nuclear and high-energy physics

accelerators. one of the most satisfying features of the superconducting CW linac is that its powerconsumption is lower than that of a room-temperature accelerator with comparable capabilities. Anoptimized energy-efficient. pulsed room-temperature linac with PSR would con~ume about 16 MW perhour of operation. whereas the CW linac consumes only 8 MW per hour. Therefore. for a given annualbudget for AC power. the CW linac would operate twice as long. and provide major additional oppor­tunities to the experimental program.

Stable Low-Current BeamsOne of the three CEBAF end stations (designated B) will be dedicated to physics which requires

the use of a Large Acceptance Detector (LAD). For final states with more than three panicles. an LADis the only practical way to observe the event fully. Such a detector sees a large fraction of all the paniclesproduced by all reactions that take place. and. unless the beam current is very low. is easily saturatedwith too much data. This experimental hall. which will carry out about one-third of the initial CEBAFscientific program. therefore requires stable beams with intensities of the order of tens of nanoamps.

The CW linac would produce stable low-intensity beams very simply. The linac operates at a fre­quency of 1.5 GHz. generating an electromagnetic sine wave with a period of 0.67 nanoseconds. as shownin Figure 2.13. Panicles are accelerated in small bunches which ride on the crests of the sine wave. The

A B c A B c

f.- 0.67 ns ..jr...4t------- 2 ns ------~

• High-current bunches

x Low-::urrent bunches

Fagure 2.13. A representation of a 1.5·GHz electromagnetic sine wave in which high-eurrent and low-eurrent bunches ride oncrests. The time between successive bunches delivered to a single end station is 2 ns.

normal mode of operation calls for the injector to load the bunches in sequence. such that every thirdcrest carries electrons destined for a specific end station (designated A. B. or C). Since the injector candeliver currents varying from bunch to bunch by as much as a factor of 100. high-current and low-currentbunches can be accelerated independently. and the intensity of low-current beams will be independentof possible variations in the intensity of high-current beams.

Since the time difference between successive pulses delivered to a single end station (2 ns) iscomparable to the resolving time of the detectors (typically not much less than 1 ns). the delivery ofevery third pulse to an end station is at a high enough repetition rate (0.5 GHz) to look continuous tothe experimental equipment.

Possibility for Significant Upgrade in EnergyThe CW linac is expected produce energies greater than 4 GeV after commissioning. given the

conservative design parameters specified to guarantee 4 GeV with a new technology. A significantadvantage of this design is that. if the nuclear physics community should consider it desirable at a latertime. it can be upgraded to energies considerably in excess of 4 GeV as progress expected in thedevelopment of superconducting cavities of high gradients is made. The prospect of extending the CEBAFresearch program to higher energies is a great asset. and is a major scientific advantage of the CW linac.

2S

•

CEBAF Design Report

It guarantees that significant new physics opportunities will be available to CEBAF in the future, andensures that the facility will have a Ion£. and productive lifetime.

The discussion of the need for high energy in Section 2.2 explained in a general way why it is ofscientific importance that CEBAF be able to be extended to higher energies in the future. These argumentswill be developed in somewhat greater detail now.

A plot of the structure function vW~ for the proton (so that M r = M p in Equation 2.7) is shown inFigure 2.14. This figure, which is similar to the one published in the reports of the Barnes subcommitteeand Bromley panel. is constructed from data obtained at SLAC in the late 19605 and early 19705.(") Notethat for Q~ < 3 (GeV/c)~ the three proton resonances, ~ = 1232 MeV, N; = 1520 MeV, and N;= 1675 MeV, stand up clearly above the background. but that in the region where W> 2 GeV and Q~> 1 (GeV/c)~. known as the deep inelastic region. the structure function assumes a smooth featurelessshape which depends only on the scaling variable w' and is independent of Q~. This scaling behavior isnow understood to result from incoherent scattering of the electrons from pointlike quarks which makeup the proton. Study of electron scattering in the deep inelastic region is a major source of currentknowledge about quarks and their distribution in nucleons, and the EMC measurements(~)have dem­onstrated that t~e effects of the nuclear medium on these distributions do not completely disappear evenat very high Q- (or at very short distances). Perhaps the most striking feature of Figure 2.14 is thetransition from the first region. where the scattering excites specific proton resonances whose structurefunctions decrease rapidly with increasing Q~, to the second regi~n (deep inelastic) where individualresonances disappear and the scattering becomes independent of Q-. This is referred to as the transitionregion. One of the primary missions of CEBAF. with its continuous beam. is to explore this region andto study the specific final states which contribute to the structure functions.

Limit of a 6-GeV beam0.4

v W2

0.2

O~__H-

1.0

1.0

Limit of a 4-GeV beam

...'ll~};r..~..:.t.•,·.~.~,;.·.·::.·.~.~..,j;,·.:.,.;,.,:....:.:,.,_.:..".::.•:..,~-.:~:.:.,•.~.::....,..[::..p"'::~""·••'·'..',.,t,~..:~.".~.~.~·;.• ...

·:;'.~.r.·~,~,'.~:.;.·:.f.~.;.:e,~.i:·,.".·,'. '. '. r:~ \'~.. ".' ." ~ ~ ~ '. ·e'~:~.,'."':~····: " .

5.0

(4' = 1 + w2Q2

10.0

I

Figun 2.14. The proton inelastic structure function "W; shown a~ a function of Q; and the scaling variable w', At large Q', "Wbecomes a function of the scaling variable ",' only. a behavior referred to as scaling. The regions which can be studied by a+.GeV accelerator and a 6-GeV accelerator lie to the left of the limilS shown.

26

Scientific Justification

The main reason the Barnes. Bromley. and Vogt committees considered 4 GeV to be the appropriateenergy for the next electron accelerator for nuclear physics is that 4-GeV electrons have sufficient energyto cover a major part of the transition region. Figure 2.14 also shows that as the energy of the electronbeam is increased. a bigger portion of the region can be studied. Electron beams at 6 GeV. for instance.would permit the study of the proton structure functions into the deep inelasdtic region. where quarkeffects dominate.

The same i~formation is presented in a somewhat different way in Figure 2.15. which shows theregion in the (Q-.v) plane accessible with 4-GeV and 6-GeV electrons. In this plane inelastic scaneringto final states with a fixed mass W is represented by straight lines parallel to each other. and the deep

Maximum kinematically allowed region4-GeV beam 6-GeV beam

. . . . . . .. .. . . . . . . .. .................... ...................... ............. . . . . . . . . ....................... .............

..))}«(() :

II

IIIIII

CVI/I I ......~ I II

I -TIIIII

IIIl

6~-~-r-...,......-...,.,....,...""""-""""~""",

5

,

C\I(J

-4>a>

C)::::3

C\I

°2

o , 234v (GeV)

5 6 0 , 2 3 4v (GeV)

5 6

Region where aM > ,62 ~/st4-GeV beam 6-GeV beam

65234v (GeV)

,

II

IIII

cv l/I I~I

II

IIIIIII

II

I.

6 05234v (GeV)

,

IIIIII

CVI/I I~I

IIIIIIIII

II

o

5

,

6

C'i.~4>a>

~3

rlgW'e 2.15. The same region of the (Q=.,,) plane is shown in each figure. The solid diagonal lines which slope from the lowerleft to the upper right represent the loci of points for elastic scattering (labeled by P) and excitation of nucleon resonances(.1. N;. N;) from the proton. The top two figures show the maximally allowed kinematic region. which is to the left of the heavyline labeled e = lSOO. Note that a ~GeV beam can probe the entire transition region. but only a small fraction of the scalingregion (the shaded area). while a 6-GeV beam (upper right-hand figure) can probe a much larger fraction of the scaling region.(In the scaling region. W > 2 GeV and Q= ~ 1 (GeVlc)=.) The bottom two figures show the region which can be probed if it isrequired that the Mott cross section be larger than the value indicated on the figure. Again. the 6-GeV beam can explore a largerfraction of the interesting area.

27

•

CEBAF O<:sign Report

inelastic region (referred to as the scaling region in the figure) is also shown. The Mott cross section.defined in Equation 2.6 above. decreases as the electron scattering angle increases. so that counting ratesnear the boundary defined by 18D-degree electron scattering decrease rapidly. While a 4-GeV electronbeam is sufficient to cover most of the transition region. counting rates at the right edge of the regionwill be very low. so it will be difficult to study much of the deep inelastic scattering region. A 6-GeVelectron beam can be adjusted so that the Mott cross section will be larger throughout the transitionregion: such a beam also permits a larger portion of the deep inelastic scattering region to be studied.

Experience has shown that after 10 to 15 years cf research with any accelerator. the physics com­munity often desires to increase its energy. Given continu:~g progress in superconducting technology.CEBAF can eventually be upgraded to energies considerably in excess of 4 GeV. ensuring that significantnew physics opportunities will be available to CEBAF in the future. Thus the facility should remain atthe forefront of scientific development for a substantial time.

2.4 Highlights of the Proposed Initial Experimental Program

Over twenty experiments planned for CEBAF are described in the 1982 SURA proposal. and inthe report of the Workshop on Future Directions in Electromagnetic :'~uclear Physics(6) published in1981. This section will describe five key experimental programs which illustrate both the science and theexperimental requirements for the end stations. The first one or two paragraphs describing each programare intended for the general reader: the subsequent discussion is more technical. The key programs are:

• coincidence measurements of single-particle densities.

• two-body correlations in nuclei.

• production of excited nucleons.

o production of hypernuclei. and

• charge structure of the neutron and deuteron.

Coincidence Measurements of Single-Particle DensitiesMeasurements in which a proton ejected from the nucleus is detected in coincidence with the scattered

electron give direct information about several different aspects of nuclear structure. First. these meas­urements show that the nucleus contains protons. Next. they tell how the charge distribution (and hencethe quark structure) of a proton bound inside the nucleus differs from that of a free proton. Finally.they give detailed information about the motion of bound prot(".ns. In particular. the proton momentumdistribution (which is the probability that the bound i"fGton will have any given momentum) can beextracted from these measurements.

To illustrate how these quantities can be inferred from (e.e'p) experiments. the simplest possiblereaction leading to proton knockout from the simplest nucleus (the deuteron) is illustrated in Figure2.16.

,I

EII

I ;(,(p-q)

-(p-ql

Figun: 2.16. Schematic diagram of the d(e.e·pjn experiment. The ,,-irtual photon emitted by the scattered electron strikes theproton when it is moving away from the neutron. Because the outgoing proton has momentum p and the virtual photon carriedmomentum q. the experimenter deduces that the initial momentum of the proton Wa'i p-q.

28

Scientific Justification

If magnetic scattering is neglected (for simplicity). the differential cross section obtained from thisdiagram is given by

(2.9)

where. as usual. the scale of the overall cross section is set by the Mon cross section. F(Q) is the formfactor of the bound proton (the Fourier transform of its spatial charge distribution), and 1!I(p - q) is thewave function of the deuteron. The wave function depends on the relative momentum of the twonucleons.and its square is the proton momentum distribution. To determine F(Q) and 11!I12

se~arately.measurements of different values of the outgoing proton momenta p can be made at a fixed Q . and ofdifferent Q2 at a fixed value of p - q. (In practice. interactions between the proton and neutron in thefinal state. or interactions in which the electron scaners directly from a meson being exchanged betweenthe two nucleons. introduce corrections to this simple picture which :nay be significant.)

For comparison. the same simple theory gives the following expression for the single-arm inclusivecross section

tfadf!d£'

(2.10)

This expression differs from Equation 2.9 in two important respects. First, all directions of themomentum of the final proton contribute. which means that the expression will alway<; be dominated bythat part of the phase space where the wave function has a maximum, making it difficult to extract thewave function in regions where it is small. A second difference is the sum over ~, where ~ labels thedifferent Fock space components of t:le wave function which described the behavior of different com­binations of constituents which the deuteron can contain.

This means that the larger components will dominate. Coincidence measurements allow the exper­imenter to fix the momentum of the wave function at a particular value by choosing a particular directionof the ;nomentum of the outgoing proton. and to select a small component of the wave function forstudy, such as the M component of the:: deuteron. Therefore, coincidence measurements give importantnew information not obtainable from single-arm inclusive experiments.

Programs of (e.e'p) coincidence measurements have already been started at Saclay, NIKHEF, andBates. but they are limited both by low energies (Saclay = 720 MeV, NIKHEF = 500 MeV, and Bates= 750 MeV) and low duty factors (about 1-2%). Figure 2.12 (in Section 2.3) showed results of recentNIKHEF measurements on lead. The peaking of the coincidence cross section (proportional to thespectral function S shown in the figure) at definite energies corresponding to specific removal energiesfor the proton, combined with the characteristic momentum dependence of each energy peak, providesbeautiful, direct evidence for the shell structure of nuclei.

Results also exist for few-body nuclei. Figure 2.17. taken from Reference 7, shows the results of(e.e·p) measurements from the deuteron. The dashed curve shows roughly how th,," momentum distri­bution would fall if the deuteron had only an S-state; clearly the data are very sensitive to the D-statecomponent which, in tum. is a measure of the strength of the tensor force. Similar results have beenobtained for the three-body system from the Saclay 3He(e.e'p)d and 3He(e,e'p)np data. Figure 2.18,taken from Reference 8. shows how the two- and three-body contributions can be separated from oneanother. and Figure 2.19 shows the momentum-space dependence of each of these separate components.

29

I

CEBAF Design Repon

CEBAFs high energy and 100% duty factor will permit this exciting program of coincidence meas­urements to be extended to much higher values of momentum transfer. and will permit the measurementof small components of the wave function which have coincidence cross sections too low to be measuredat a low-duty-factor facility. For example. the deuteron S-state has a second maximum in the regionfrom 500 to 700 MeV/c. and small relativistic components are also expected to become important in thisregion. High energy and high momentum transfer are needed to fully explore the three-nucleon system.Neutron knockout experiments (e.e'n) with cross sections 10 times smaller than (e.e'p) are harder tomeasure because of the inefficiency of neutron detectors. but promise to be as rewarding as (e.e'p)measurements. They will require the full capability of CEBAF if they are to become a viable program.

A detailed kinematic analysis(9) of the (e.e'p) experiment shows that the cross section can be wrinenin the form

~ .7~ 10QI~

~

Q.c _8.~ 10-:::J~...-'"'1:l>- .9

- 10";;;CQI

'1:l

E~c _~10o~

d (e,e'pJ ne:500MeV

ttulr",- Y.m.'iJucn,

h'~"-'

100 200 300p (MeV/c)

FJglII"e 2.17. Data from the Saclay measurement described in Reference 7 compared with four theoretical models for the deuteronmomentum density distribution. A dashed line. showing the S-state contribution only. has been added to the figure.

30

I

Scientific Justification

-Nf

L. 6~NI

>Q)

~.0.5

4....0..c:

"'0c:"'0....0..

"'0 2....w"'0-t:l

<Q

"'0

00

He (e.e' p)

9 ·52.2

9 p ·64.41

E ·528 MeV

E' ·428 MeV

Two-body breakup

break up

Em (MeV)

FIgUR 2.18. Distribution of missing energies. showing how the two-body p .,. d final stale can be separated from the three-bodyp .,. p + n final state.

3He (e,e 'pJd

>-';;; 10c...",

e:::0

c:...e0

%:

-,10

­'-">..g

.:§- z..:'10

3He (e,e'pJpnEm < 20 MeV

l..11-

• 11-, ~

1 ~:-..

h\, "I ".""It"

\I \

I \

1\• + \, ~,

­'-">..gfa.~10zc.!:!";

:E'""6>-;;; 10c..

.",

e:::0

c:..eo

%:

o ?OO 300I~: (P'IeV/c]

o 100 200 300Ipi [l'1eV/c1

FIgDn 2.19. Momentum density distribution for the two-body and the three-body sectors in inelastie electron scanering from)ie.

31

I

CEBAF Design Repo"

{VLw'~j+ VTw'~ + VTTw'~cos24>1

+ VTLw'~L cos4>. + h V TL , w'~L' Sin4>l} (2.11)

where 4>1 is the out-of-plane scanering angle of the proton. as shown in Figure 2.20. The kinematic tennsare:

( p ==\

VT = ip + tan2ta

VTT=-ip ,

1VTL = '\I2p V p + tan2ta

1VTL, = - '\12 p tanta

(2.12)

The Wi are structure functions that depend on four variables. which can be taken to be if, 11, PI,and 91, where 91 is the angle of the detected proton with respect to the direction of the three-momentumtransfer q in the plane defined by Phq as shown in Figure 2.20. In Equation 2.11, h is the polarizationof the electron in the longitudinal direction; this fifth tenn, which contains additional information on thescanering process, is zero unless the incident electron beam is polarized and measurements are madeout-of-plane (4)1 =1= 0 or 1r).

Facare 2.20. Exclusive-l electron scanering. Here the z-axis is chosen to be along Q and the electron scanering occurs in theu-plane so that the y-axis is defined by u,.-Xx X'/IXx X'I. The panicle detected in coincidence with the electron bas momentump, whose direction is given by angles (8.. 4».) in this coordinate system.

To separate W~ and w'~L' requires measurements where the ejected particle is not in the (e,e')reaction plane (referred to as out-of-plane), but the other three structure functions can be separated bymeasuring the differential cross section both at forward and backward electron scanering angles 9, andby observing the scanered proton on both the "left" and "right" sides of the three-momentum transferq, as illustrated diagrammatically in Figure 2.21.

32

•

I

~1:lO

E

-,-­II

~8-:~ /! ~\-: -----: )_.....--I ....-- 8.

E

lbl

I

Scientific Jus,ilication

FIglIn 2.21. Diagrams illl!Strating how the structure function WK can be separated from the others by measuring the outgoingproton at the same angle 9, on either side of the direction oi the three-momentum transferred by the electron. In case (a'. 9, =o (c:os4l, = 1) and in case (b). c!>, = "11' (so that cosc!>, = -I). Taking the ~ifference in the 'ounting rates in these two cases willisolate the struet".lre function WK.

As discussed above, it would be desirable to extend the measurements shown in Figure 2.17 to twicethe :elative momentum. about 700 MeV/c. and to make these measurements at different values of Q2in order to study whether the proton structure is influenced by the momentum distribution in the deuteron.Illustrative configurations which accomplish this at both forward and backward electron scattering anglesare shown in Figure 2.22. Note that 4-GeV/c and 1.5-GeVlc spectrometers are sufficient to carry out thisillustrative measurement.

.., lei 10'

FIglIn 2.22. Figures (a) and (b) show the magnitude and direction of the momentum of the outgoing proton in the case whenthe re:ative momentum of the initial Slate being probed is 700 MeV/c. and if is 1 (GeVlc)~ for electron scattering angles of 30"and 150". In both figures. the momentum corresponding 10 c!>, = 0 is dashed. and it is assumed that the electron is detected by a4-GeV/c spectrometer at bot" "- 'Ward and backward scattering angles. and that the proton is detected in a l.S-GeV/c spectrometer.Figures (c) and (d) sioow th" .,arne situation for a if of 2 (GeV/c)~. These kinematic diagrams illustrate how the spectrometersmust be placed in order to carry out this program.

33

I

I

CEBAF Design Report

Two-Body Correlations in NucleiMany nuclear phenomena can be understood by assuming that the nucleons in nuclei behave as

independent particles moving in some mean force field. Ironically. this overly simple picture is to someextent useful because of the strong short-range repulsive force between nucleons. which prevents nucleifrom collapsing and guarantees that many-body forces will be small. clearing the way for a mean-fieldapproximation. This strong short-range repulsion should lead to strong correlations between nucleonsat short range. but very little is known about such correlations. What is known is based primarily onwhat has been inferred from the behavior of free nucleons. It is therefore important to observe thesecorrelations experimentally and to study how they are modified by the nuclear medium. If unusual six­quark states are important in nuclei. they would be expected to influence the short-range correlations.

The observation of two-nucleon knockout through the process (e.e'2N) appears to be the mostpromising and direct way to study these nucleon-nucleon correlations in heavy nuclei. In most cases.observation of two outgoing nucleons in coincidence with the scattered electron will require the use ofthree spectrometers. However. if the two nucleons come out parallel to each other. they can be detectedin a single spectrometer. If the two outgoing nucleons have a large momentum with respect to the restof the nucleons in the nucleus. final-state interactions between the observed pair and the rest of thenuclear fragments will be reduced. and Figure 2.23 should describe the reaction approximately. The

2N

FJgm'e 2.23. Schematic diagram of the (~.~·2N) reaction. The electron (represented by the dashed line) emits a virtual photon(represented by the wavy line) which strikes a nucleo!l with momentum p, and changes its momentum to P,. so that it may emergetogether with the nucleon with momentum P,. It is assumed that the two outgoing nucleons ioteract with each other. but that thepair does not interact significantly with the rc,iduaJ A·2 nuclear system. The reaction is proportional to the probability that twonucleons in the initial system wiU have momenta p, and P,.

differential cross section obtained from this simple di~gra:n is proportional to the probability that thetwo nucleons detected in the final state had initial momenta PI and P2. The cross section is directlyproportional to the ~uareof the momentum-space cor:-dation function. which is a function of the relativemomentum between the two nucleons. P = ! (PI - p~;.

If the electron kinematics are adjusted so that

2

x=-.Q:..=22Mv

and all binding effects are ignored. it can be shown that the cross section is largest when

34

•

(2.13)

(2.14)

•Scientific Justification

so that the two nucleons in the detected nucleon pair have equal momenta in the direction of the three­momentum transferred by the electron. In this case the final-state interactions between the pair are largebut can be easily calculated.

With these kinematics. the two-nucleon correlation function can be measured to momenta in excessof 1 GeV using the spectrometer configurations illustrated in Figure 2.24. Note that 1.5-GeV/c and 4­GeV/c spectrometers are sufficient for the program.

II

lal

E =0.99 - 3.20

III1-160 I I

I -I "I II I'

1-250

I "I ~,

1/,'I I'

I "I '

I "1.13I

IIII

E =0.48 - 1.80

Ibl

FJgW? 2.24. Diagrams showing the energy and scattering angle of the electron. and magnitude of the momenta of the outgoingnucleon pair. as discussed in the text. The range of variation of the incoming and outgoing electron energies. and proton momenta.correspond to varying if from 1 to 4 (GeVlc):. It is assumed that the 4-GeVlc spectrometer is used to measure the scatteredelectron at both the forward (9 = 45°) and back""afd (8 = 120") angles.

Production of Excited NucleonsNuclei are, to a significant extent, collections of three-quark clusters similar to free nucleons. As

these clusters move and collide with each other in the nuclear medium, they may change their size,shape, or internal motion. Study of this behavior is an important part of the study of the role of quarksin nuclear structure. Insight into what changes in size, shape, or internal structure may occur can beobtained by studying the behavior of free nucleons when they are struck by an external probe, such asan electron or a pion. The study of such elementary collisions is therefore an important part of modemnuclear physics. and CEBAF is well suited to carry out those measurements which involve electroncollisions.

When struck by an electron or pion, a nucleon may be excited to certain metastable states, whichhave a characteristic mass, lifetime, and internal structure. The most prominent of its excited states, andthe first to be discovered, is the ~ with mass 1232 MeV. The appearance of the ~ as a strong, broadresonance in low-energy pion-nucleon scattering suggested initially that it should be viewed as a compositeof a nucleon and a pion, but it is now knOW'll to be primarily an excited state of three quarks, similar tothe nucleon but with the spin of one of its quarks ~ipped (Figure 2.25). The existence of the many excitedstates of the nucleon, shown in Figure 2.26, is direct evidence of the fact that the nucleon is a compositesystem of three quarks.

35

CEBAF Design Report

N

FlJ:Ure 2.25. Artises conception of a nucleon (:>OJ and a delta (~). showing their internal quark structure. In the nucleon. two ofthe quark spins are aligned. while the third points in the opposite direction. In the delta. all three spins are aligned: this hassuggested to some theorists that it will have a deformed shape. as shown. In the modern view. the nucleon and the delta aretherefore very closely related. differing only by the extent to which the spins of the quarks are aligned.

The ~ plays an important role in the intermediate range nuclear force (the two-pion-exchange part).and it is also expected to have a strong influence on the structure of nuclei. It is of interest to know whatfraction of the nuclear ground state consists of ~·s. how they propagate through the nuclear medium.and how they are produced when electrons scatter from nuclei. One aspect of the investigation of these

100 - -,~~ I ee 1'~~ e e(e) I •• -- 0I .... I

900 I •••• I • -• •••

800 [U UjI I700

I I ••• • ••••••

600

500~

400 ~

2

'"'"oE

20

,+N'2

Fagure 2.26. A chart of excited nucleons (taken from Reference 10). The labels N and ~ refer to two different alignments ofquark spins. as explained in Figure 2.25. while the fractions refer to the total spin of the system. The open boJreS with asterisksinside correspond to resonances which have been seen in ToN scanering (the greater the number of asterisks the more certain arethe observations). The solid bars are predictions using the quark model of Reference 10. The short solid bars label states whichare not expected to couple strongly to the -rrN channel. but could be seen in photoproduction experiments. as described in Reference11. Observation of these ··missing·· states at CEBAF would be an important confirmation of the quark model.

36

Scientific Justification

questions is the study of the electroproduction of ~'s from free protons; the study of the transition formfactor or the structure of the photon-proton-~vertex function shown in Figure 2.27 is a question offundamental interest in its own right, strongly sensitive to the details of quark dynamics. In particularit is believed that the excitation is mostly magnetic (Ml), and that the amount of electric (£2) excitationis a sensitive indicator of the shape of the ~.

p!--.......---- ,.... 9y............ 1....

/ --FJglII'e 2.Z7. Schematic diagram of the structure of the nucleon-photon-delta vena function. or form factor. The vinual photonis represented by the wavy line. and the delta by the double solid line. The delta cannot be seen directly; evidence of its presencecomes from observing its decay products. the nucleon and pion.

The electroproduction experiment (e,e'-rrN) is a triple-coincidence experiment in which a pion anda nucleon are detected in coincidence with the scattered electron. The kinematics for the electroproductionof ~'s from the proton are determined in part by the fact that the mass of the final state, W, lies somewherein the region of 1232 :!: 50 MeV, where 50 MeV is the half-width of the ~ resonance. A striking featureof the decay is that for if bigger than 1 (GeV/c)2, the proton angle, 6p , is restricted to very smallvariations around the direction of q, and almost all pions scatter outside of the small forward coneoccupied by the protons. If a 2.5-G-::V/c spectrometer with an acceptance of about 20 msr were placed6° to the "left" of q, it would accept a large fraction of the protons. A 1.5-GeV/c spectrometer placedon the opposite side of q at an angle of about 200 (acceptance from 15° to 25°) with respect to thedirection of q. would pick up a fraction of the pions with the same eM angle. If a 4-GeVic spectrometerwere then used to detect the forward-scattered electrons (Figure 2.28), the system would have a reasonably

IIII Protons

~_:~o

£.152

FJgUre 2.28. Schematic illustration of spectrometer placement for study of the eleeuoproduction of deltas at if = 2 (GeV/c)2.

37

-

CEBAF D<.-sign Rcpon

large acceptance for electroproduction of .:l·s in the forward direction. Under these conditions systematicstudies of the t:xcitation of .:l·s and higher resonances in nuclei become possible. The influence of thenuclear medium on these resonances. a central question for both QCD and nuclear physics. can bestudied.

The.:l is only the first and most conspicuous excited state of the proton: a study of the basic propertiesof the other excited states (shown in Figure 2.26) is a!so offundamental importance. The method describedin this section can be extended to any state which decays substantially into the -rrN final state. However.the higher-lying excited states tend to have widths larger than the sepa:-ations between them. so that theisolation of the contribution of a single resonance requires the use of polarized particles and the meas­urement of the angular distributions of the hadronic decay products of the resonance. (The problem issimilar to electron scattering in the giant reSOi1anc~ region.) To obtain experimental information sufficientfor a multipole decomposition. a number of independent experiments are required. including differentialcross sections. beam and target asymmetry. recoil nucleon polarization and double polarization exper­iments. A complete determination of all the independent amplitudes would require measurement of atleast 7 independent observables for real photons. and 11 for virtual photons.

Some excited states of the nucleon do not couple strongly to the -rrN channel. and therefore cannotbe detected in elastic -rrN scattering (see Figure 2.26 a~ain). The search for these resonances is an importanttest for QCD-inspired nonrelativistic quark models. 11) Several states are predicted to decay to the -rr-rrNchannels through states like -rr.:l or pN and still show a reasonable coupling to the -yN channel. In thiscase. electromagnetic excitation becomes the only practical channel available in a formation experiment.

To carry out a complete experimental study of these resonances requires a maximum energy transferof about 3 GeV and momentum transfers in excess of 1 (GeV/c)z. CEBAFs maximum energy of 4 GeVis sufficient to produce the required monochromatic and polarized real and virtual photons. The exper­imental equipment required includes polarized proton and deuterium targets. and a Large AcceptanceDetector (LAD). needed to achieve high data rates and low systematic errors.

Production of HypernucleiOne of the novel programs proposed for CEBAF is the study of the excitation and the decay of

hypemuclei (nuclei in which one nucleon is replaced by a strange baryon. usually a A). Such nuclei areinteresting because they contain one strange quark. a type of quark not found in ordinary nuclear matter.This quark (and the strange baryon to which it is attached) can be thought of as an impurity in thenuclear system: its interaction with other quarks (or nucleons) is different from the usual interaction.and it can be distinguished from its neighbors. Study of the structure of hypernuclei, which is of interestfor its own sake. is also interesting because of the insight it should give into the structure of ordinarynuclear matter. and into the nature of the forces between nucleons and strange baryons.

Our present knowledge of the level scheme of hypernuclei comes entirely from (K- .-rr -) exchangereactions.(IZ) in which a neutron is converted into a A by the reaction K- + n -+ -rr - + A. In thisreaction the A-nucleon mass difference can be perfectly compensated by the K - - -rr - mass difference.leading to recoilless A-production and therefore to large cross sections for the production of hypemuclei.The best overall energy resolution that has been achieved is 2 to 3 MeV. mainly governed by the targetthickness. (13) The big disadvantage of the (K- .-rr -) reaction is. however, the large distortion induced byboth the incoming K- and the outgoing -rr - • making the study of heavy hypernuclei very difficult andquantitative interpretation of the results uncertain. Only a small class of levels, ones in which a A issubstituted for a valence nucleon. are accessible. .

In contrast, the electromagnetic excitation of hypernuclei offers the advantage that the nucleus ispractically transparent to both the incoming probe and the outgoing K~ . The clean production mechanismand the low distortion in the final state make the electromagnetic production of hypernuclei an excellentprogram for CEBAF.

The sizes of the cross sections for the excitation of hypernuclear levels are determined by the crosssection for the elementary reaction -y + p -+ K+ + A and by the momentum transfer to the final nucleus.which determines the size of the transition form factor. Small momentum transfer (q = 250 MeV/c) isonly reached for photon energies E-y ~ 1.5 GeV and small K~ laboratory angles (OK -:s;; 6°). At thesemomentum transfers the population of substitutional states is reduced and transitions to states involvinga large change in the angular momentum are emphasized. This permits states with low principal quantumnumber to be excited in heavier nuclei. giving access to the full spectrum of hypernuclear states.

The behavior of A's in ~eavy nuclei is of special interest. If the A is regarded as a struetureless pointparticle. it may occupy the lowest possible energy level. Viewed as a composite system of quarks. it is

38

I

Scientific Justification

possible that the Pauli principle would suppress this process slightly. Due to the expected closeness ofhypemuclear levels. the experimental resolution for the excitation energy of the hypemucleus should beof the order of 100 keY in order to derive maximum benefits from the program. Detailed cross section~cula~~ons h~ve been performed by several authors.(loS) !p>ica} cro.ss ~ct~ons (e.g., for the. transition-C - -.\B(1 g.s.) at E., = 2 GeV) are of the order 10 cm-/sr. mdlcatmg that the expenments are

feasible.At present. three different schemes are considered for the study of the electromagnetic excitation

of hypemuclei:

a) (e.e'K+) with 0.. and OK >100

This setup requires the full beam current and the full energy of CEBAF.

b) (e,e' K+) with 0.. and OK near 00

Due to the much higher photon flux for the 00 virtual photons. this experiment requires onlylow current (I == 100 nA). Since the target can be made very thin. the K+ energy loss will notcontribute to the final resolution. Also. the energy of the primary electron is lower (E = 2.3GeV). Therefore. the relative resolution can be a factor of two worse than in (a). The low beamintensity makes it possible to achieve very high resolution by letting the target intercept only asmall fraction of a dispersed beam.

c) (-Y.K+) with tagged photonsEnergy requirements are similar to (b) and the maximum current is determined by the timeresolution of the tagging system. Since photon production and photon reaction are now separatedprocesses. the target thickness can no longer be arbitrarily reduced without an intolerable lossin counting rate. Therefore. this method may not be suitable for very high resolution experiments.However. the low background in tagged photon experiments will make it possible to study thedecay modes and the lifetimes of hypemuclei.

The requirements which must be met to achieve an overall energy resolution of 100 keY are extremeand have not been met before at these high energies. In this respect the excellent beam quality and thelow halo of the CW design presented in this report offer hope that such resolution will be achieved.

Charge Structure of the Neutron and DeuteronAs discussed in the previous sections. the distribution of charge inside nuclei can be measured

directly from electron scattering. Early measurements of these distributions gave data on nuclear sizes;later. with increasing electron energy and resolution. more detailed information about shell structureand sha~sof deformed nuclei were obtained. Figure 2.29 shows the charge distributions of the outermostshell in -06pb inferred from recent measurements.(IS) and is a good illustration of the precision and clarityof interpretation possible with the electromagnetic probe.

Two charge distributions of fundamental importance to nuclear physics have not yet been satisfac­tori!y measured. The distribution of charge inside the neutron. which is electrically neutral. is of fun­damental importance for two reasons. Not only is it sensitive to the distribution of quarks in its interior.but precise knowledge of this quantity is needed in order to extract information about nuclear structurecontained in all high-momentum-transfer electron-scattering data. Similarly, the charge distribution ofthe deuteron is of fundamental importance because it is the simplest nucleus. and one of the few nucleiwhose structure can be calculated from first principles.

Charge distributions can be obtained in principle from a separation of the elastic differential crosssection

da E'[ 2 2 2]dO' = aM£: A(Q) + B(Q ) tan!O

where, for spin one-half targets such as the neutron, the structure functions A aI'!d B are

(2.15)

1 + T

(2.16)

39

CEBAF Design Report

206Pb

20STI

35

FJgIU'e 2.29. The charge distributions of~b and =n together with the charge density of the 35 proton shell. The units arearbitrary.

where G£ and GM are the electric and magnetic form factors related to the Fourier transform of thecharge and magnetization distributions respectively. However. at high Q2 the structure function A(Q2)is dominated by the magnetic form factor G...,. making it very difficult to extract accurate values for thesmaller G£ from Equation 2.15. This situation is particularly severe for the neutron. Since the neutronhas zero charge. its electric form factor. G;, is swamped by the neutron magnetic form factor for allvalues of ~. G; is only very poorly known up to Q2 = 1 (GeV/c)2, where measurements suffer fromlarge statistical errors and unknown (but large) systematic errors, and is completely unknown above Q2= 1 (GeV/c)2. .

40

•

I

Scientific Justification

A somewhat different problem occurs for elastic scattering from the deuteron, which is a spin-onesystem and therefore requires three form factors for a full description. These form factors are the electricmonopole and quadrupole form factors. Gc and GQ • and the magnetic form factor, G..." which enter thestructure functions of Equation 2.15 according to the relations

., ., .,A(Q:!) = G~- + ihl G:~ + §or) c.:;

( -~),,- 4Md

(2.17)

Since only these two structure functions can be measured in single-arm elastic, unpolarized scattering,it is impossible to separate the monopole and quadrupole form factors of the deuteron in this way.However. it is extremely important to know all three form factors. For example. it is widely believedthat the smoothness of the structure function A for the deuteron is a result of the contributions of twofunctions (Gc and GQ ) which oscillate in a diffractive fashion in such a way that the maximum of oneoverlays the minimum of the other. An experimental measurement o( Gc or GQ alone would be muchmore sensitive to the detailed structure of the deuteron than the measurement of A.

With a longitudinally polarized electron beam, and either a polarized deuteron target, or polarimeterscapable of measuring the polarization of recoil neutrons and deuterons, it is possible to measure thecharged form factor of the neutron, and to separate the monopole and quadrupole electric form factorsof the deuteron. If a polarized target is used. the experiment must be done with a low beam current,and would be best car..:ed out using the Large Acceptance Detector (LAD) or a large acceptancespectrometer. such as the VAS described in Chapter 10. If it is decided to measure the recoil polarization,the experiment would be done with full beam current using spectrometers in the heavily shielded endstations. Both of these methods have their advantages and disadvantages, but both appear feasible andboth should probably be undertaken.

Measurement of the recoil polarization requires that the scattered hadron be measured in a polar­imeter in coincidence with detection of the (forward) scattered electron. To increase the counting ratesat high Q:! it is best to work at small 9. where the Mott cross section is large. In the neutron case, thedoubly differential cross section is

(2.18)

where cDl is the azimuthal angle of the secondary scattering, defined with the same coordinate systememployed in Figure 2.20. h is the longitudinal polarization of the electron, A y is the analyzing power ofthe polarimeter material, and pz is the recoil polarization which depends on G£ through the relation

_ - 2V..O + ..) G.., G E tan!9pz - (1 + ..) (A + B tan2i9) (2.19)

Note that longitudinally polarized electrons are necessary for this measurement, and that the recoilpolarization depends linearly on G£ (a great advantage) so that both the sign and magnitude can bedetermined by this experiment. The formula for the deuteron case, which is somewhat more complicated,can be found in Reference 16. In this case three recoil tensor polarization quantities can be measuredwith unpolarized electrons, <111d if the electron beam is longitudinally polarized, an additional vectorpolarization transfer quantity can be measured. With a polarized electron beam it is possible to obtainan absolute measurement of the ratio of Gc to GQ by looking at asymmetries (the cl>l dependence of thesecondary scattering) only.

41

.....

CEBAF Design Report

References

1. Picture taken with the Atomic Resolution Microscope at Lawrence Berkeley Laboratory.2. Z. Meziani. et al.. Phys. Rev. Lett. 522130 (1984). and 54. 1233 (1985).3. P. DeWitt Huberts. invited talk presented at the XI Europhysics Divisional Conference. "Nuclear

Physics with Electromagnetic Probes:' Paris. July 1985.4. See. for example. E.D. Bloom and F.J. Gilman. Phys. Rev. 04.2901 (1971).5. European Muon Collaboration. J.J. Aubert. et al.. Phys. Lett. 123B. 275 (1983). A. Bodek. et al.•

Phys. R~v. Lett. 50. 1431 (1983) and 51.534 (1983).6. Report of the Workshop on Future Directions in Electromagnetic Nuclear Physics. organized by the

Bates Users Group. P. Stoler. Chairman.7. M. Bernheim. et al.. Nucl. Phys. A365. 349 (1981).8. E. Jans. et al.. Phys. Rev. Lett. 49.974 (1982).9. T.W. Donnelly. in Research Program at CEBAF: Report of the i985 Summer Study Group.

10. R. Koniuk and N. Isgur. Phys. Rev. D21. 1868 (1980).I!. N. Isgur. published in the Proceedings of the CEBAF 1984 Summer Workshop.12. See talks by B. Bassalleck. L.S. Kisslinger. and R.A. Eisenstein. published in the Proceedings of

the Su:amboat Springs Meeting. 1984. AlP Conference Proceedings # 123.13. T. Walcher. Nucl. Phys. A434. 343C (1985).14. See. for example. A.M. Bernstein. Proceedings ofthe Conference on Hypernuclearand Kaon Physics.

Heidelberg. West Germany. 1982. and T.W. Donnelly. Proceedings of the Workshop on Electronand Photon Interactions at Medium Energies. Bad Honnef. West Germany. 1984.

15. B. Frois. published in the Proceedings of the international Conference on Nuclear Physics. Florence.Italy. 1983. and Nucl. Phys. A396. 409 (1983).

16. R. Arnold. C. Carlson. and F. Gross. Phys. Rev. C23, 363 (1981).

42

3. Superconducting CW Linac

3.1 Design Overview

The mission of CEBAF is to design. build. and operate a world-class electron accelerator facilityto serve the nuclear physics community. The accelerator is to provide a continuous beam of electronsat any energy between 0.5 and 4.0 GeV. To serve more than one experimental station simultaneously,the current or intensity of the continuous electron beam is to be 200 ~. This chapter describes arecirculating superconducting CW electron linac which meets these needs.

The key component of a superconducting linac is the superconducting accelerating cavity, whichallows for continuous acceleration of beam without excessive power loss in the cavity wall. The CEBAFlinac will use a superconducting cavity designed. developed, and tested at Cornell University. Fourprototypes of this 1500-MHz cavity have been built by Cornell, and all have exceeded performancecriteria specified for the CEBAF linac. as have three prototypes that have been built by industry forCEBAF. Superconducting RF (SRF) technology is now sufficiently mature that construction of a high­efficiency CW superconducting linac is practical.

The remainder of this section of Chapter 3 introduces the conceptual design for a multiply recir­culating superconducting CW electron linac for CEBAF. Later sections present an outline of the workbreakdown structure for the construction of the accelerator, a discussion of beam stability, a discussionof radiation safety. and an analysis of the technology base for the linac design. The conceptual designfor the systems and subsystems required for the linac i~ described in chapters 4 through II.

Multiply Recirculating LinacThe most straightforward CW accelerator is a single linear accelerator which the beam traverses

one time. With an accelerating gradient (EQ ) of EQ = 5 MeV/m, 800 m of active accelerating structurewould be required to reach an energy of 4 GeV (Figure 3.1(a». Even a cursory estimat~ shows this tobe an economically inefficient approach. A more cost-effective solution is obtained by passing the beamrepeatedly through a shorter linac structure by means of recirculation paths (Figure 3.1(b».

4-GeV linac

o .. '-- r---.. 4 GeV

FJgUre 3.1(8). Single linear accelerator.

Three-energy transport

4 GaV ---41..----1

Four-energy transport

Fagure 3.l(b). Multiply recirculating linac.

43

CEBAF Design Repon

This approach is made possible by the fact that electrons are fully relativistic. i.e .. move at verynearly the speed of light at quite modest energies. At only 50 MeV. for instance. the electron velocityis 0.99995c. Once fully relativistic. an electron's velocity is essentially independent of energy. Thus.beams at different energies pass together through the linac structure. all maintaining the proper phaserelative to the accelerating RF field. The recirculating linac approach has been used successfully at boththe HEPL superconducting linear accelerator and the MIT-Bates conventional pulsed linear accelerator.and will be employed at the Wuppertal-Darmstadt superconducting linac. currently under construction.

Since the electrons are fully relativistic. the cavities need not have the same accelerating gradient.One cavity can have a higher gradient than its neighbors without the bunch RF phase being degraded.Strongly performing cavities can make up for weaker ones since the RF design allows for individualcontrol to excite each cavity to its maximum field. The total beam energy. therefore. reflects the averagegradient of all the cavities and is not severely limited by the weakest unit.

The number of r>asses through the linac must be determined through cost optimization. More passesreduce the capital cost. Fewer passes reduce the operational complexity and enhance beam stability.Based on the cost of our cavity prototype (the CEBAF-Cornell five-cell cavity) and the design choicesfor the recirculation arcs. the optimum number of passes is four or five.

A four-pass system has been adopted for CEBAF. This accelerator will COnSi5! of two O.5-GeV linacsegments joined by recirculation arcs that bend the beam. transporting it from the exit of one linac tothe entrance of the next (Figure 3.2). Beam will be injected at an energy of 50 MeV.

In this four-pass system. four electron beams. each at a different energy. are simultaneously presenton the same trajectory in the linac segments. As each beam is at a different energy. each will require aseparate recirculation path tuned to accommodate the particular energy of that beam. There are. there-

Linac tunnelcutaway

50-MeVinjector

(13= 0.99995)

Extracted beamto end stations

';#!;'" "

"Linac '\.(0.5 GeV) '\."

"RF separators

Arc tunnelcross section

(dia. = 3.43 m)

Linac tunnelcross section

-

figun 3.2. Schematic of the CEBAF recirculating supcrconduCling CW linac.

44

-

Supercondueting CW Linac

fore. several recirculation beam lines within each recirculation arc. These beam lines are separatedvertically. Upon exiting a linac. the beams at different energies are spread apart vertically and transportedto the appropriate recirculation beam lines through use of a device we have tenned a spreader (or. if atthe required energy. extracted and transported to the experimental areas). At the end of the recirculationarc. the beams are transported from each recirculator beam line. recombined. and injected into the nextlinac through use of a device we have tenned a recombiner.

The path of a single electron bunch from injector to experimental area therefore comprises fouracceleration cycles. each of which is as follows:

1. injection into first linac segment.2. acceleration by first Iinac segment (for 0.5 GeV additional energy).3. "spreading" to the proper beam line in first recirculation arc.4. transport through recirculator to recombiner for injection into second linac segment.5. acceleration by second linac segment (for 0.5 GeV additional energy).6. spreading to the proper beam line in second recirculation arc. and7. transport through recirculator to recombiner for reinjection into first linac segment.

On the fourth (and final) cycle. this sequence is interrui>ted at step 6 where the beam. after spreading.enters an extraction beam line. Extraction elements in each of the recirculation arc beam lines aiiowextraction at step 6 on any of the preceding cycles as well.

Energies between 500 MeV and 4 GeV can be achieved by lowering the energy gains per linac andextract:ng bunches of the desired energy at step 6 after an appropriate number of cycles. For example.a beam of 1.6 GeV may be obtained by using an energy gain of 0.4 GeV per linac segment and byextracting the beam at step 6 of the second pass.

The extraction elements for beam extraction at cycles 1. 2. and 3 are RF separalors, time-dependentdeflecting devices that allow one out of every three passing bunches to be steered into the extractingbeam line rather than into the recirculation arc beam line. On the fourth cycle the beam is directed toa fourth RF separator. As will be shown in subsequent chapters. this configuration makes it possible toserve all three experimental stations simultaneously with beams of up to three different energies andindividually adjusted current levels.

Key Beam Dynamics IssuesThe key beam dynamics issues for a recirculatinglinac are beam stability and beam quality. Both

issues have been resolved in the conceptual design of CEBAF.The first issue is the problem of insuring against beam breakup. which limits the amount of beam

current a Iinac can accelerate. In a recirculating linac. the most likely cause of beam breakup is thegrowth of transverse deflecting fields in the RF cavities generated by repeated passes of beam bunches.This process is called multipass regenerative beam breakup. Analysis of multipass regenerative beambreakup in CEBAF predicts that the beam current required is below threshold by well more thm anorder of magnitude.

The second issue is the problem of c"nserving the emittance and momentum spread. In a recirculatinglinac two factors can degrade the beam: synchrotron radiation during bending in the recirculation arcs,and possible mismatch of the electron beams upon reinjection into the linac segments. Beam quality canbe maintained and reinjection mismatches avoided through proper design of the recirculation arc beamlines. The features of CEBAFs beam lines are characteristic of approaches employed in the design oflow-emittance storage rings. They include careful control of beam path length, isochronicity. achroma­ticity. and careful correction of chromatic effects to facilitate reinjection after each arc. Strong focusingminimizes emittance growth caused by quantum excitation due to synchrotron radiation.

Accelerator ParametersThe main parameters of the CW linac are summarized in Table 3.1. To describe the machine

components. we follow the Work Breakdown Structure (WBS) (Table 3.2) that is described in greaterdetail in Section 3.2.

4S

•CEBAF Design Repon

Table 3.1CEBAF SRF CW LiDac: Design Parameter LN

I

46

Beam cbanIcleristicsElectron energy E [GeV)Average CWTent [J,IA]Transverse emittance (95%. 1 GeV) [mJEnergy spread [95%!Duty factorSimultaneous beamsSimultaneous energies

LiDac paraIDder'CConcept

Number of passesNumber of linac segmentsSegment length em]Maximuill energy gain per pass [GeV)Recirculation time per pass [IJ.S]FocusingPhase advance per cell (pass 1)Half-ceU length em]Number of cavities per !la1f-ceUNumber of half-ceUs per segmentVacuum (before cooldown) [torr]

Cavity paramda'sTypeFrequency [MHz]Eleenic iength [m]Shunt impedance (rIQ) [ohmlm]Design gradient [MV/m]Design residual QTypical HOM QencnoaIOear aperture [mm]Transverse HOM rlQ [ohmslm1

RFsystemNumber of klystronsKlystron R.I:" power coupled

to beam [kW]

lDJector .....-.enGun energy (MeV)Injection energy [MeV)Average current [J,IA]Transverse emittance (at 0.1 MeV) [mm-mr)Longitudinal emittance [keV-degrees]Bunch length [degrees)Pulse capability [IJ.S]

Recirculatioa an:sNumberMagnetic radii em]Phase advance per periodPeriods per arc

CryoaaUc sysIaDTotal RF load at 2.0 K [WITotal beat load at 2.0 K [W)Total heat load at 40-60 K [W)

•

0.5'" E". 4.02002 X 10.9

1 x 10'"100%3"'3

SuperconductingCW recirculatinglinac4

22351.04.2FODO120"9.482S10'9

Superconducting15000.5960.05.03 X 109

Hi'to 10'70". 16.4 x :0·

418

2.0

0.10SO2001< 15-Jr< 1.50.05 to 10

711.5 to 28.6211'(5/4)4

240032006000

SupercoDducting CW Linac

Table 3.2Work Breakdown Structure by Level

Levell

Level 2

Level 3

CEBAF

1. Machine Vacuum Components2. Beam Transport3. RF System4. DC Power Systems5. Instrumentation & Control6. Experimental Stations7. Cryogenics8. Conventional Facilities9. Project Services

.1 Linac

.2 Recirculation Arcs

.3 Switchyard

.4 Injector

.5 Experimental Station A

.6 Experimental Station B

.7 Experimental Station C

.8 Distributed Systems

.9 Conventional Facilities

3.2 Work Breakdown Structure

CEBAFs Work Breakdown Structure (WBS) divides the accelerator facility into areas of functionalresponsibility (level 2) and machine sector (level 3) (Table 3.2). D~perWBS levels break these areasdown into assembly units and subunits and cost components (Table 3.3).

Table 3.3Work Breakdown Structure Levels

Levell.Level 2.Level 3.Level 4.LevelS.Level 6.Level 7.

ProjectFunctional ResponsibilityMachine SectorMajor AssemblyAssemblySubassemblyCost Components

As reflected in the WBS, the main accelerator parts are machine vacuum components (WBS 1.0),beam transport elements (WBS 2.0), RF system (WBS 3.0), DC power (WBS 4.0), instrumentation andcontrol (WBS 5.0), and cryogenics (WBS 7.0). WBS element 6.0 is experimental equipment, WBS 8.0is conventional facilities, and WBS 9.0 is project services. The following subsections provide overviewsof the nine WBS elements.

Machine Vacuum Components (WBS 1.0)Machine vacuum components include the beam vacuum as well as insulating vacuum in cryostats.

In particular, the SRF cavity is included here. The CEBAF CW linac will employ a 1500-MHz, 5-eellcavity with waveguide, beam pipe, fundamental power and higher order mode couplers that follow strictlythe proven design of the Cornell cavity. Specifications call for a minimum accelerating gradient of 5MV/m and a residual quality factor (Q) of at least 3 x 10

9• Recent experience shows these specifications

to be conservative. Two cavities will be installed in a cryostat to form a cryo-unit, the smallest self-

47

~·0 '

CEBAF Design Repon

contained modular unit of the linac. Four cryo-un:ts containing a total of eight cavities form a cryogenicmodule. placed between warm linac sections containing linac quadrupoles. steering magnets. beaminstrumentation. and vacuum equipment. Chapter 4 describes the machine vacuum components.

Beam Transport (WBS 2.0)Beam transport elements (Chapter 5) include the dipole. quadrupole. and sextupole magnets required

for beam steering and focusing in the linac and in the recirculation arcs. as well as the RF separatorsand septum magnets for extraction. The magnets are relatively small bore (typically with a 38-mm gap).and have low to medium field « 1 T). They are of laminated steel construction with copper coils. RFseparators at 2.5 GHz follow closely established designs (e.g.. CERN-Karlsruhe S-band separator).

RF System (WBS 3.0)The RF system' (Chapter 6) powers the cavity. For maximum flexibility and control without me­

chanical devices. one klystron per cavity is envisaged. This will maximize machine availability and willallow each cavity to operate near its maximum capabilities in the most straighi.!orward manner. Includedin the RF system is the injector. whose key parts are the gun. subharmonic chopper. and a buncher.Beam quality depends on these components and their proper functioning. Differential bunch loading.i.e. the creation of beams of different current levels. is accomplished by the injector.

DC Power Systems (WBS 4.0)DC power systems (Chapter 7) are the power supplies and their controls required to energize the

magnets in the linac. recirculation arcs. and the beam switchyard.

Instrumentation and Control (WBS 5.0)Instrumentation and control (Chapter 8) includes all diagnostic and control (computers. instrumen­

tatim!. interfaces. and software) functions required to operate the accelerator. A key feature of CEBAFscontrol architecture will be heavy reliance on hardware vs. software. and on distributed intelligence. Mostcontrolled components will be equipped with ILCs (intelligent local controllers). devices that will maintainproper operation with minimum flow of data to and from higher levels. Chapter 6. in which RF controlsare described in greater detail. illustrates how this particular system could remain operational even ifcommunications to the next higher level were temporarily interrupted.

Experimental Stations (WBS 6.0)CEBAF will have three end stations (Chapter 10) for experiments. The major pieces of experimental

equipment are the magnetic spectrometers. the Large Acceptance Detector (LAD). the Variable Ac­ceptance Spectrometer (VAS). and targets.

Cryogenic System (WBS 7.0)The cryogenic system. which is essential to operate 1~e ~~ f_ .... 'li!;~s. is described in Chapter 9. The

SRF cavities must operate at cryogenic temperature. An optimization has led to the choice of an operatingtemperature of 2 K. The cryogenic system includes a central heli:lrn refrigerator. storage vessels. and adistribution system to keep the cavities at 2 K and the heat shields in the cryostats J,elow 60 K. The totalcryogenic load in the linac is 3200 W at 2 K. and 600 W at 60 K. The AC power requirement of thecryogenic system is 5 MW. The experimental end stations may use superconducting magnets operatingat 4.2 K. If supplied by the central refrigerator. their requirements will add another 0.5 MW of power.for a total of 5.5 MW cryosystem input power.

Conventional Facilities (WBS 8.0)CEBAFs conventional facilities wiII include all of the buildings and structures for the accelerator.

the three experimental areas. and the various support systems and utilities. as well as for the users andSLaff. Chapter 11 covers this element of the WBS.

Project Senices (WBS 9.0)CEBAFs management objectives are to design. build. and operate a facility meeting the project

scope. Chapter 12 summarizes the project responsibilities and management tools and procedures plannedto accomplish these objectives.

48

•

Superconducting CW Linac

3.3 Beam StabilityIntroduction

The beam current of RF linacs has been limited by beam breakup phenomena. In normal conductinglinacs single-pass cumulative and regenerative beam breakup have threshold currents of tens to hundredsof milliamperes. Since higher-order-mode Q's of the superconducting RF cavities of tbe CEBAF linacwill be reduced by couplers to levels (10") characteristic of room-temperature cavities. these single-passeffeC"..s should not be of significance at the few-hundred-microampere-per-beamlet values of CEBAF.Computer modeling supports this conclusion. Multipass regenerative beam breakup, on the other hand,has limited the beam current of superconducting CW linacs such as the Stanford recyclotron to tens ofmicroamperes. The damping achieved in the 55-eell cavities used at Stanford. however. reduced HOMQ's to values still in excess of 10

7• Considerably higher threshold currents, in excess of 10 mAo can be

expected with the five-eell cavity considered here.The following discussion presents a detailed investigation of multipass regenera:ive beam breakup

for the CEBAF CW linac. Computer modeling of beam breakup has been developed, and includesaccurate accounting for time delays. mode frequency spread. and lattice function variation. This simu­lation was found to agree with a more specific theoretical analysis which had successfully described beambreakup at the Stanford recyclotron. Application of this simulation to the long array of five-eell cavitiesof the CEBAF design indicates that beam breakup occurs only when currents exceed ten milliamps perbeamlet-more than an order of magnitude above CEBAF"s design current of 200 J.LA.

Phenomenology of Beam BreakupConsider the excitation of a single parasitic mode of a cavity by a charge q passing at transverse

position r off axis. A charge e passing. behind will experience a transverse momentum kick of

ric. - 20'''~ f..p == W (.) = 2Q q r sm (wo) e (3.1)

where Z' is the transverse coupling impedance (ohmslm3) and L is the active length of cavity structure.

Consider now a series of bunches spaced by !b with average current 10 and charge per bunch ofq = loI!b. Let Woeos(wt + cb) be some initial wakefield excitation of the cavity. A bunch will experiencea momentum kick

f..p = %Wo cos(wt + cb)

After recirculation, this bunch returns to the cavity with a transverse displacement

(3.2)

(3.3)

where B" is the transfer coefficient (momentum to transverse position) between the i th 2Ild jth passes.On traversing the cavity the second time. an additional wakefield

Z' -~T

f..W = 2Q Le q (B 12 f..p ) sin (wo) e 2Q (3.4)

is excited. Steady state is maintained if the bunch excitation compensates for losses in the cavity wallsand HOM couplers. At higher current levels there will be instability. With the assumption of a most­pessimistic conspiracy of time delay. the steady-state condition yields the threshold condition

(3.5)

I

49

..

I

CEBAF Design Repon

In an n-pass recirculation configuration. there are n beamlets passing through the cavity. One beamlethas been kicked (n-l) times and has a perturbed transverse position described by the transfer coefficientsB I " through B(n.I)". Another beamlet has been kicked (n-2) times with corresponding coefficients BI(,,_I)

through B(".2)(".I). and so forth. Thus. for an n-pass recirculating linac. the threshold current is (again.assuming a worst-case phase conspiracy)

(3.6)

where the maximum values of the Bil are taken to obtain a most-pessimistic result.Determination of the threshold current in CEBAF requires estimates of both the transfer functions

B,j of the linaclrecirculator and the expected transverse impedance of the RF cavities. The HOM imped­ance of the Cornell five-eell structure is well documented in the literature. The Z' and Q for the fourworst modes are listed in Table 3.4.

Table 3.4Strongest Transverse Modes of 5-CeU Cavity

Frequency (MHz)

1888 1969 2086 2110

Q (10") . 3.2 0.4 1.0 1.3

Z'/Q (104 ohm/m3) 6.95 16.4 5.0 10.0

Z'l (109 0hm/m2) 445.0 131.0 100.0 260.0

(Z'l)cff (109 oh :n/m2) 89.0 94.3 47.0 104.0

I = 200 m for CEBAF CW linac

The linac will consist of some 400 of these five-eell (0.5-meter active length) cavities and have atotal active length of 200 meters. Manufacturing tolerances will yield mode frequency variations of about1 MHz full width. Thus. there will not be perfect coherence of 400 modes of a given type since modewidths are typically less than 1 MHz. In addition. cryo-units are separated by small-bore (38 mm) stainlesssteel beam pipes which limit cavity-to-cavity mode coupling. At threshold there exists a steady-statemodulation of the linac beam at some frequency. and it will be the value of the impedance at thisfrequency (modulus the bunching frequency) that will drive the beam above threshold. In light of this.the total impedance offered by the RF cavities is

Z'(Z'l)cff =Q . Q . 200 m . F(Q) (3.7)

where F(Q) estimates the fraction of modes of a particular type acting coherently. Computation of theimpedance sum of randomly distributed resonators gives F(32000) = 0.2 and F(4000) = 0.72 for auniform distribution of modes centered at 2 GHz with full width of 1 MHz. The last two entries in Table3.4 give the total Z'l (assuming no manufacturing variations) and the effective (Z'l)cff (assuming I-MHzvariations) for the four modes listed.

The linac lattice is modeled by a 9.4-meter half-cell-length latt!ce with a phase advance per cell of120" for the first-pass energy. The recir:::ulator is assumed to offer an identity transfonnation. Without

50

• . I

• •

Sup.:rconducling CW Linac

any lattice optimization. a maximum vaiue of ~IBI = 56.0 cm'(MeV/c) was found. Insertion of this valueand the effective (Z'l)eff of Table 3.4 into Equation (3.6) yields lower bounds on threshold currents.between 1.5 and 3.3 mNbeamlet for the four Cornell modes.

These values represent a worst-case analysis. with effectively all passes assumed to have time delaysyielding maximal coherence and with the cavities localized on the largest value of B,,. However, eventhese worst-case thresholds exceed the design current of 200~A by nearly an order of magnitude.

Breakup at the Stanford RecyclotronA standing wave theory of multipass regenerative beam breakup of Vetter was confirmed by a series

cf experiments performed at the Stanford superconducting recyclotron. Although the theory containsboth singlepass and multipass effects in the form of integrals over the cavity fields. it was found that animpulse approximation of the multipass effects alone was sufficient to describe the instability thresholJs.This limit is consistent with the analysis of the previous section and the work of Herminghaus. TheStanford experiments were well described by the threshold condition

KQ (R1'Y.n)

R = ~ ~ R'fr 1.'n 1) sin (0/-1 - o,·d/ ,<-::/ 'Yin + I - 'Yo

(3.8)

where 0.. is a phase shift for the U'h recirculation. 'Yin is the injection energy (in units of rest mass), 'Yu isthe energy gain per pass, R'fr is the transfer matrix element (in TRA]';SPORT notation), and K summarizesthe cavity field integrals. For the stror:gest mode at 1863 MHz

K = 4.0 . 10- 5 ~: (A_m)-I (3.9)

(3.10)

where N is the number of cells active for the mode and L, is the effective length of the mode. From thedefinition of K in terms of field integrals as given by Vetter. we have

~l = 2 ( m;2 )K ;

and the equivalence of equations (3.6) and (3.8). Numerically.

Z'/Q = 11.7 . 104flIm

3 (3.11)

is obtained for the 1863 MHz mode of the Stanford structure. This value is comparable to the valuesfound for the Cornell structure itemized in Table 3.4. The damped HOM Q's at Stanford. however,were at the 10

7level as compared to the 10

4level for the Cornell cavities. and are primarily responsible

for observed threshold currents of about 10 ~A.

In summary, the experience with multipass regenerative beam breakup at the Stanford recyclotronindicates that the phenomenon is described by conditions of the form of equations (3.6) or (3.8) withtransverse impedances Z' consistent with the values listed in Table 3.4. On the other hand. because ofthe high Q's associated with the Stanford structures, there was not mode frequency overlap as would bethe case for the proposed CEBAF cavities. Some uncertainty therefore rem<tins concerning the effectsof cavity-to-cavity mode frequency variations and the distribution of interacting cavities along the lengthof the linac in improving the .hreshold currents. These issues among others have been addressed with acomputer simulation which is discussed in the next section.

Computer Simulation of Beam BreakupFrom the Stanford experience it appears that the dominant mechanism for multipass regenerative

beam breakup can be modeled by an impulse approximation. This is also the regime appropriate tosingle-pass cumulative beam breakup as observed in SLAC where extensive computer modeling has been

51

CEBAF Design Rcpon

successful. The SLAC program (authored by R. Helm) has been modified at CEBAF to include multipassrecirculation with both random and systematic bunch displacement as initial conditions. For the single­pass cumulative breakup the chief diagnostic is the final bunch position at the output of the linac. Athreshold is determined by a scraping condition at the beam pipe wall. For multipass regenerative breakup.there will be exponential growth of cavity fields. For a panicular amplitude of initial condition. beamdisplacement may remain small for the duration of the run even though cavity fields are growing ex­ponentially. However. for CW operation these fields will eventually grow to the point of producing beamloss. Therefore. the chief diagnostic in determining threshold currents is the observation of exponentialgrowth of the cavity wakefields.

The analytic work as presented above represents a worst-case analysis in that time delays wereassumed to conspire for maximum effect and the largest values of the machine transverse transfer functionwere taken even though the cavities are uniformly distributed along the linac. The simulation offers thepossibility to study more realistic distributions of time delays and lattice functions. and. in addition.allows for investigation of the effects of modal frequency spread.

The analysis summarized in Equation (3.6) is appropriate to a localized. single mode with nofrequency spread. Since the threshold current for this case can be simply calculated including exact pass­to-pass time delays. it offers a clear test of the simulation. Both 2-pass and 4-pass threshold currentvalues were calculated for a FODO linac lattice structure with a mode frequency of 1890.0 MHz. Theseconfigurations were also modeled with the simulation code. Simulation and analytic threshold currentvalues were found to agree well within a factor of two. Figure 3.3 shows results for a four-pass configurationwith a localized transverse impedance equivalent to 1 m of CEBAF structure with no mode frequencyvariation.

100000

10000

1000

100>-Cl"-CD 10c::CD

>-

>1

C'Cl<.>

0.1

0.01

0.001

0.0001

o

",-"

,----""",

"-"

20

Time (revolution periods)

----- 100 mA 40 mA

FIgUre 3.3. Simulation of beam breakup drivcn by localized structurc. Analytic threshold current is 70 rnA.

52

Supercondueting CW Linac

A second study compared the single 1890.0 MHz (Q = 32(00) threshold with that of 100 modesu!2iformly distributed between 1890.0 and 1891.0 MHz. Each mode was given 1% of the single-mo~e

transverse impedance arid all were localized at the same point on the lattice. The threshold current hasbeen found to be a factor of 6 higher for the distributed modes and is in reasonable agreement with thefactor-of-5 estimate included in the worst-case analysis presented earlier. This result firmly establishesthe benefits of mode frequency spread in easing beam hreakup.

Runs were performed to model the full CEBAF cavity array distributed along the linac with a FODOlattice. Run lengths corresponded to over 200 damping times of the cavity modes. Threshold currentsfor the four Cornell cavity modes are summarized in Table 3.5. All threshold currents are greater thanlOmA/beamlet.

Table 3.5Computer Simulation Estimates or Beam Breakup Thresholds

1888 MHz12 mA

1969 MHz13 mA

2086 MHz22 mA

2110 MHz11 rnA

A final set of runs was performed to observe any interaction of modes of significantly differentfrequency. Each of fifty ··supercavities·· was allowed to have four modes at frequencies of 1890. 1969.2108. and 2120 MHz with no cavity-to-cavity frequency spread. Each mode was given an impedanceequal to the 1890 mode. The observed threshold current and growth rate were within 25% of single­mode values and do not indicate any significant mode interaction.

SummaryA computer model of multipass regenerative beam break up that includes mode frequency spread

and lattice variation has been developed. For a single. localized mode. this simulation is found to be inagreement with analytic estimates which have been successful in describing beam breakup in existingrecirculating linacs. For the CEBAF design. a threshold current for beam breakup in excess of 10 mAlbeamlet is found. This estimate exceeds the design goal of 200 j.LA by more than an order of magnitude.In addition. further improvements in threshold behavior can be expected with optimization of recirculatordesign.

3.4 Radiation and SafetyOverview

Any electrons lost from the electron beams. normally confined within the accelerator beam pipesand cavities, will constitute a radiation source. A major source is the beam dump where the electronbeams are finally deposited after passing through the experimental target. Electrons, when they strikesomething such as the sides of the beam pipe. magnets, the experimental target. the beam dump and soforth. produce penetrating radiation which must be shielded. Based on assumed electron losses, thefacility is shielded so that levels of radiation external to the shield and hence the exposure of workersand members of the public is below accepted limits. In general. specifications for thick shielding againstthis radiation are dominated by the very penetrating high-energy neutron component. Although the doseequivalence for the high-energy neutron component is a relatively small fraction of the total yield, itrequires greater thicknesses of shielding to reduce it to the low levels required in occupied areas.

In order to specify adequate shield thicknesses it is necessary to define certain basic parameters,which -=an be summarized as follows:

1. loss term.2. source term for each contributing type of radiation (with high-energy neutrons dominating the

specifications for thick shielding. as mentioned llbove).3. external dose to be achieved. and4. removal mean free path for the appropriate source term.

53

•CEBAF Design Repon

The loss tenn (1) is the most difficult to specify with any certainty in any location other than beamdumps. which are intended to cope with the full beam power in any given channel. The loss tenn willbe discussed below under each of the accelerator component parts. The source tenn (2) is a complicatedmeans of converting loss into radiation dose in such a way that it can be attenuated by a given quantityapproximately defined in the case of high-energy neutrons by a constant inelastic cross section for theshield material. In actual practice the source tenn and relevant attenuation lengths are largely based onexperimental measurements. Parameter (3) is the external dose to be achieved by the shielding. It isnecessary to design shielding to achieve radiation levels in occupied areas lower than the limits forradiation doses to people laid down by appropriate authorities. For CEBAF we have designed to radiationdose levels of 1 rem/year just out-side shielding (compared with the limit for occupational personnel of5 rem/year). We have not included any allowance for occupational factors. thus giving a further measureof safety. Radiation limits to individual members of the public are lower than for people who work withradiation. and dose limits for prolonged exposures of the population are lower still at 100 mrem/year.Therefore. CEBAF will be designed and operated so that dose rates at the property boundary will beless than 100 mrem/year.

We consider each part of the facility separately:

Beam DumpsThese. although requiring the most shielding. provide the simplest beam loss tenn-the total average

beam power of the channel which the beam dump serves. The main beam dumps A and C will beadequately sized to cope with the full machine power. Beam dump B is sized to cope with I% of the fullbeam power since end station B will only be used for low-intensity experiments. The beam dumps willincorporate specific safety features to overcome any problems of dump coolant failure or other mal­function. The assumed full accelerator capability is 1.2 MW corresponding to 6 GeV. but for shieldingpurposes we take an annual average of 600 kW or 2 x 1022 electrons accelerated to 6 GeV per year.

End ScacionsThe major loss in the end stations arises from electrons scattered from the experimental targets.

Furthennore. the end stations have to be large to accommodate the angular movement of the spectro­meters. The two-meter-thick concrete walls and one-meter-thick roof are designed to achieve the requiredexternal radiation levels and yet pennit experimenters reasonable freedom to perfonn a broad mix ofexperiments. Typical experiments require the insertion of a thin piece of material into the electron beamso that the various electron nucleus events may be studied. However. these thin targets will give rise tosome scattering of the electrons so that not all the electron beam finally ends up on the beam dump.Instead. a small fraction of the beam is unavoidably lost in the end station itself. The amount of beamscattered depends on the thickness of the experimental target (the thicker the target. the greater thefr3.ction scattered) and the energy of the incident electron beam (the lower the energy. the greater thefraction scattered). It is of interest that for a target of given thickness. the overall source tenns get largeras the energy reduces since the fraction of electrons scattered increases faster than the reduction in totalbeam power. The average power loss in the end stations per mrem/year on the nearest property boundaryis 4 watts for A and C. For end station B the power loss must be restricted to 0.05 watts to achieve the1 rem/year just outside the shield wall with a resulting boundary dose less than 10 mrem/year. Thesecalculated power losses pennit a reasonable average mix of experimental conditions. while keepingboundary dose rates less than the limit of 100 mrem/year.

The LinacIt is not expected that the lina.: will suffer significant continuous loss of electrons. It has a large

aperture and the beam has a very small emittance. The pressure in the superconducting cavities will helow. Thus there will be insignificant losses from gas scattering. Other possible mechanisms for beam lossare being studied. but it is thought that any continuous beam loss will be very small. On the basis ofexperience with other accelerators such as SLAC. and given the good beam and aperture of CEBAF. aloss tenn of 0.05% has been used for linac shielding. This is a very conservative assumption. in that itexceeds any conceivable estimate of concinuous beam loss by several orders of magnitude. This aspectis considered in Section 5.3.

Arcs and Beam SwicchyardAgain on the basis of the low beam emittance and good aperture. a continuous loss tenn of 0.5%.

based on experience from SLAC, is considered to be highly conservative. A similar loss tenn is used for

54

Supcrconducting. CW Linac

the beam lines and switchyard. Again. the highest fractional beam loss would occur during tune-up (i.e ..low-current operation) and would therefore lead to a low total loss.

In steady state lower relative loss terms are to be expected. Most of the loss would be forced tooccur by judiciously placed halo scrapers placed in spreaders. halfway through recirculation arcs. and inrecombiners. These would eliminate halos corresponding to less than 1O·~ of the total beam. force theselosses to occur locally (where they are most easily shielded). and prevent any further loss downstream.

ShieldingA description of the shielding structure is given in Chapter 11. A summary ()f the numbers used for

shielding is given in Table 3.6.

Table 3.6Beam Loss and Shield Thickness

Component

Beam dumpsA&C

Beam dumpB

Arcs & BSY

Linac

End stationsA&C

End stationB

Loss(%)

100

0.5

0.05

Energy(GeV)

6

6

6

6

Average annual loss

(power) (electrons/year)

600.0kW 2 x 1O~

6.0 kW 2 X 1O~1I

3.0 kW 1 X 10~11

0.3 kW 1 x 1019

4.0 W·

0.05W

Shieldthickness (ft)

31 (earth)

22 (earth)

21 (earth)

17 (earth)

6.6(concrete)

2(concrete)

·per mrem (boundary)

Note: ThL"'SC beam loss terms are averages assumed for shielding calculations only.

MonitoringIt must be stressed again that during commissioning. beam currents will be kept low and only

increased slowly toward maximum after stable conditions have been fully established. Any major faultconditions or excessive beam loss will be detected by installed instrumentation. These fast-responseinstruments will rapidly cause the injector to be disabled and shut the beam off. Beam diagnostic andradiation loss instruments are described in Section 8.2.

In addition to the beam-loss monitors. other radiation monitors will be installed in occupied areasaround the facility. These instruments are designed to measure radiation at the levels designated for thegiven area. They will not only operate as stand-alone instruments. but will also provide data to a centralmonitoring/recording location; they will also have level alarms to warn of any high external radiationconditions which might occur under exceptional conditions. Monitors at site boundary will also beprovided. The combination of shielding and monitoring provides one layer of the total safety system.Other layers will involve the definition of categories of radiation areas and the necessary administrativeprocedures to control the entry of personnel. Where high radiation areas inside shielding are concerned.all access ways will be fitted with a high-reliability system of locked and interlocked closures.

3.5 Technology BaseIntroduction

Radio-frequency superconductivity is an attractive technology for several types of accelerators. fromheavy-ion machines to electron machines. The technology for low-beta acceleration (heavy ions) issufficiently different from that for machines with 13 = 1 to allow this discussion to concentrate on the

55

CEBAF Design Report

latter. However. experience operating superconducting low-beta systems will be mentioned where it isapplicable.

SuperconduCling accelerating structures with ~ = 1 have intrinsic advantages for several types ofelectron accelerators: storage rings. CW linacs. TeV-scale linear colliders. and free-electron lasers. Thebasic element of the accelerating structure for each of these applications is a superconducting cavityconsisting of one to tens of cells. each half as long as the RF wavelength. For the past 20 years. R&Don such cavities has focused on bringing their performance to a usefu: level with regard to acceleratinggradient and current-carrying capacity. Simultaneous efforts have aimed at developing a cavity fabricationprocess capable of producing cavities that consistently meet performance specifications.

ElectronIPositron Storage RingsThe main recent motivation for the development of superconducting accelerating cavities has come

from interest in high-energy ~lectronlpositronstorage rings. For these machines. superconducting cavitiesoffer a practical means of compensating for synchrotron radiation losses. which increase with the fourthpower of the energy for a given machine. Several laboratories (CERN. DESY. Cornell. KEK) operatingor planning high-energy electron storage rings have thus invested significant effort and funds in developingRF superconducting technology. Within the past two years. this effort has paid off. with all four labo­ratories fabricating cavities that have achieved accelerating gradients exceeding 5 MV/m. Table 3.7summarizes the recent progress.

Table 3.7Laboratory Results on Multi-Cell Cavities Designed for Storage Ring Use

E ..... E" CouplingLab Year MHz Cells (MeV/m) Qo (MeV/m) holes Couplers Limitation

CERN 1983 500 5 5.0 0.74 x lOy 5.0 Yes No Defect.-main coupling hole

CERN 1985 352 4 6.0 3.3 x lOy 5.0 Yes No

Cornell 1984 1500 5 8.9 7.0 x lOy 8.9 Yes Yes Defect. locationundetermined

Cornell 1984 1500 5 8.0 3.0 x lOy 8.0 Yes Yes Defect. cell 3.2.5 cm from eq.

Cornell 1984 1500 5 15.3 2.0 x lOy 15.3 Yes Yes Defect or fieldemission. cell 1

DESY 1983 1000 9 6.6 0.9 x 109 6.6 No No Defect orelectron loading

DESY 1983 1000 9 > 6.7 0.9 x 1O~ 6.7 No No Available power

KEK 1983 508 3 > 5.2 0.5 x lOY 4.0 Yes No Available power

'Spots of high resistivity. such as weld im~rfeetions. inclusions. impurities. or surface contamination.

Since August 1983. four superconducting cavities have been tested in storage ring beams at DESY.KEK. and Cornell. They have yielded accelerating gradients uj) to 6.5 MV/m. exceeding the goal of 5MV/m. and have handled currents up to 22 rnA. This progress is noteworthy. since achieved in-beamgradients were only around 2 MV/m just two years ago.

As a consequence. several laboratories now have definite plans to use superconducting cavities inelectron storage rings on a large scale (Table 3.8). CERN and DESY expect to issue bid packages soon.European industry-Interatom (Siemens) and Dornier (Daimler-Benz)-is prepared to respond.

56

I

Supcrcondueting CW Linac

Table 3.8Plans for the Application of SC Cavities to Storage Rings

Storage Energy Freq. Cells per Ea

Lab ring (GeV) (MHz) cavity Cavities (MeV/m) Q

CERN LEP 1 55-62 352 4 8-16 5--7 3 x 10'1

CERN LEP II 95-104 352 4 384 5--7 3 x 10'1

DESY HERA 30 500 4 16

KEK Tristan 33--35 508 5 40

KEK Tristan .w 508 5 120-144

CW Electron LinaesExperience with superconducting CW linacs began about 21 years ago with a pioneering effort at

Stanf6rd's High Energy Physics Laboratory (HEPL). Because this machine never achieved its designspecificC1tions. the technology has been largely ignored in the United States. However, much has beenlearned from the development and operation of that facility, and its achievements are noteworthy.Specifically. HEPL demonstrated the successful operation of a large 1.8 K refrigeration system. it achievedrecirculation, and it produced a beam of outstanding quality (with low emittance and momentum spread).

With its 34 meters of accelerating structure, HEPL is the largest superconducting electron linacoperating to date. Since 1976 the University of Illinois has been operating a superconducting microtronusing one 6-meter HEPL linac section. Currently, the Universities of Wuppertal and Darmstadt in WestGermany are building a small recirculating CW linac. Cavities produced by industry for this machinewithin the past year consistently achieve gradit:nls in excess of 5 MV/m.

HEPL's problems. namely low gradient and low current capacity, are both understood and solvedtoday. The gradient was limited to just over 2 MV/m by multipactoring, a regenerative process causedby secondary emission. The current capacity was limited to 92 IJ.A for two passes and 20 IJ.A for threepasses (confirmed well by theory) by transverse particle motion induced by inadequately damped higherorder modes (beam breakup). Solutions to these problems are described below.

Advances in Superconducting RF TechnologyThe problems that plagued HEPL all relate to the design and processing of the superconducting

cavities. Over the past few years there have been significant improvements in these areas:

Optimization of Shape and Coupling

• Spherical or elliptical cavity shape prevents serious muitipactoring.

• RF coupling and higher-order-mode damping at the beam pipe prevent cell wall penetrations,which cause field enhancement and can also induce local muitipactoring.

Better Quality Superconductor

• Improved ni_obium purity may reduce the density of defects and does improve thermal conduc­tivity, which stabilizes remaining defects against dAiving the cavity normal.

Better Manujacturing Procedures

• "Rhombic raster" or defocused electron beam weld minimizes weld defects.

• Improved cleanliness in assembly reduces dust contamination.

• Better surface insp.:.:ction and cavity testing with thermometry allow dirt and defects to be identifiedand repaired.

• Increased rinsing speed following acid cleaning reduces the formatior. of insoluble compounds.

57

I

CEBAF Design Rcport

Understanding and Cure of Current Limitations• Deliberate heavy damping of higher order modes significantly decreases excitation of transverse

deflecting modes of the beam. thereby allowing significantly increased currents to be accelerated.

Cavity PerformanceTo build a CW superconducting linac. the following major issues related to the cavity and its

performance must be addressed:

1. accelerating field strength and Q of the cavity.2. long-term performance.3. tuning of the cavity and RF coupling to the cavity.4. RF frequency. and5. damping of higher order modes and beam breakup.

Accelerating Field Strength and Q of the CavitySince the HEPL experience. it has been recognized that SC linac technology would become practical

when accelerating gradients of 5 MV/m or higher could be achi.::ved routinely. In a paper presented atthe 1985 Accelerator Conference. H. Piel (Wuppertal) observed that groups working on RF supercon­ductivity at various frequencies and at several laboratories all have achieved this goal (Table 3.9).

Table 3.9Performance of Superconducting RF Cavities

Laboratory CERN KEK DESY Cornell Wuppertal-Dannstadt

Accelerator LEP TRISTAN PETRA/HERA CESR l30-MeV Recyclotron

Matcrial Nb Nb Nb on Cu Nb Nb Nb Nb Nb-5n

Frequency inMHz 350 500 500 500 1000 1500 3000 3000

Operatingtemperature 4.2 K 4.2 K 4.2 K 4.2 K 4.2 K 1.8 K 1.8 K 4.2 K

Single-ce/l cavitiesE. (MV/m)·· 10.8 13.0· 10.8 7.6" 5.5 22.8" 23.1" 7.2Q at E. 1985 1.8 x IOv 0.7 x IO

v0.4 x IOv 0.6 x IOv 5 x iO" 2.5 ll: IOv 1.2 x 10v 1.1 x IO

v

E. (MVlm)" 4.7 7.6 6.5 5.5 8.5 10.0Q at E. 1983 5 x IOv 3.6 x IO

v4.1 x IO

v 5 x 10" 7 x IOv 4 x 10"

Multice/l resulIs 4 cells 5 cells 3 cells 9 cells 5 cells 5120 cells 5 cellsE. (MV/m) 7.5" 5.0 5.8 5.5 15.3· 12.3nA 4Q at E. 1985 2.2 x 10" 0.7 x IO

v0.6 x IOv

0.5 x IOv 2.2 x 10" 3.5/1.2 x 10" 4.5 x 10"

E. (MV/m)"" 2.8 2.5 5A 3.7Q at E. 1983 x IOv 5 x 10" 4.5 x IOv I x 10"

·Cavities fabricated from high-thennal-conduetivity niobium ··Under continuous wave operationSource: H. Pie!. Wuppertal.

He concluded. therefore. that it would be conservative to design a superconducting accelerator basedon niobium cavities with a gradient of 5 MV/m and a Q of a few times 109

• This advice is confirmed byrecent results from Cornell. where the latest four 5-cell cavities (1500 MHz) all achieved gradientsexceeding 6 MV/m (average: 8.3 MV/m).

Long-Term PerformanceExperience shows that supercondueting cavities perform for long periods without degradation of

gradient or Q (Table 3.10). Cavities have been kept in-beam for modest to long periods at DESY.Cornell, CERN, and HEPL. Illinois, and Argonne (low beta). without adverse effect other than due to

58

Supc:rconduCling CW Linac

a few specific accidents. The most impressive test. however. was an accelerated lifetime test done atCornell in 1974nS. The vacuum in the Cornell synchrotron was degraded deliberately to expose thecavity to the equivalent of 20 years of operation at 10 - III torr (2 x 10''J torr-years of gas exposure). Theperformance of the cavity decreased by only about 10%. On one occasion this same cavity was accidentallyexposed to titanium dust from a sputter ion pump. The 3 performance dropped drastically in gradientand Q. On examination. the inner surface of the cavity was found to be fully coated with titanium oxidesand nitrides. After washing with detergents and rinsing with solvents. but without acid cleaning. thecavity regained its original gradient and Q.

Table 3.10In-Beam Experience with Multi-Cell SC Cavities

Number Electricalof length

Lab Machine cavities (m) Date

Routine OperationStanford HEPL-SCA 7 34 1972-present

illinois Microtron 6 1976--present

CERN SPS (separator) 2 5.5 1977-79

Argonne ATLAS (low beta) 7 to 13 1978-present

~~onybrook Linac (low beta) 7 1984-pl'"esent

Lifetime and Performance TestsCornell Synchrotron 1 0.6 1974-75

Cornell CESR 4 (separate) O.S (ea) 1982. 1984

Karlsruhe DESYIPETRA 1 0.3 1982

CERN DESYIPETRA 1 I.S 1983

DESY PETRA 1 1.5 1985

Cavity Tuning and RF CouplingSuperconducting accelerating cavities. properly coupled and loaded. are remarkably broad bana.

They can be tuned mechan;cally. simply by squeezing or stretching the structure. For RF coupling, it isessential not to disturb the symmetry of the cells; thus all coupling is done at the ends of the cavity. asclose as possible to the beam axis. Waveguides are used on the 15OQ-MHz cavities, and broad-bandcoaxial couplers are being developed for lower frequencies at CERN and DESY.

RF FrequencyCavities operating at RF frequencies between 350 MHz and 3000 MHz have been fabricated suc­

cessfully (Table 3.9). For CEBAF. we have selected a frequency of 1500 MHz. This choice representsa satisfactory compromise between the combined criteria of fabrication ease. material costs. and quality.all of which improve with high-frequency (small) cavities. and operational issues. which favor low­frequency (large) cavities. Dominant operational issues are deflecting impedance. which causes beambreakup and increases as the cube of the frequency (for similar shapes of cavities). and the number ofcavities and controls required per unit length of the accelerator. An additional factor favoring ISoo MHzis that Cornell has developed prototype cavities at this frequency that meet or exceed all specificationsrequired for CEBAF. The Cornell group is willing to collaborate with CEBAF and assist in transferringtheir technology to industry.

Damping of Higher Order Modes and Beam BreakupEngineering to suppress the Q of all important HOMs has succeeded in obtaining typical loaded

HOM Q's of 104

in cavities developed at Cornell. These Q's are approximately 1000 times lower thanthose achieved in the HEPL structure. This i'i accomplished by using special couplers at one end of the

59

I

-

CEBAF Design Repon

cavities to extract HOMs and by using the fundamental power coupler to extract HOMs at the othercnd. According to measured and computed HOM impedances and external Q·s. existing Cornell cavitieswith HOM couplers could support recirculating CW linac output currents around 10 rnA or higher (afterfour passes). with no beam breakup (see Section 3.3). The CEBAF baseline output current is 200 JLA.

Future DevelopmentsAs a result of ongoing and future R&D. improvements in achievable gradients are inevitable.

Currently. gradients are limited by defects in the superconductor or by field emission: thus the achievedperformance of 5 to 15 MV/m in multicell cavities is far below the theoretical limit (-50 MV/m forniobium: -SO MV/m for niobium-tin). As superconductor quality improves. achieved gradients will rise.

Within the next decade. Nb (or Nb:"Sn. with its high critical temperature) promises to achievegradients between 15 and 25 MV/m. Corresponding increases in residual Q would be useful to minimizeRF losses in the cavity. thereby avoiding increased need for refrigerator power. By replacing CEBAFscavities when the technology is ready. the electron energy of the linac ca.n be increased significantly.

60

I

I

I

• I

4. Machine Vacuum Components

4.1 IntroductionThis chapter includes an overall description of machine vacuum components: the accelerating cavity

and cryostat. and the vacuum system for the injector. linac segments. recirculation arcs. and beamtransport lines.

4.2 Accelerating Cavities

OverviewCEBAF uses superconducting radio-frequency (RF) cavities as the basic element for accelerating

the beam to its final energy of 0.5 to 4.0 GeV. The principal characteristics of these cavities are that theyoperate continuous wave (CW) at ~ 5 MV/m with a residual Q ~ 3 X 109

• The high Q is required toIJ'Ijnimize the heat dissipation that must be removed by the cooling system. For purposes of the workbreakdown structure (WBS). the cavities inciude the welded cavity units and the reference probes thatare il1stalled on these cavities.

Pbysical DescriptionThe niobium superconducting RF cavities are identical in design to cavities developed and tested at

:--;ewman Laboratory of Nuclear Studies at Cornell University. A pair of CEBAF-Cornell acceleratingcavities is shown in Figure 4.1. These cavities have five cylindrically symmetrical accelerating cells. witha waveguide at one end of the cells which acts as the fundamental power input coupler and as a couplerfor extracting some of the higher order modes (HOMs) generated by the beam current. At the otherend of the cavity. two waveguides serve as coup!ers to extract higher order modes; these waveguides are

F~ 4.1. A pair of CEBAF-Cornell accelerating cavities In the configuration used at CorneU.

61

I

•CEBAF O....sign Rc:pon

perpendicular to each other and to the beam axis. The cavities are referred to as "elliptical" because.in dorsal section. the inner surfaces of the cells comprise elliptical segments. The operating frequencyis 1497 MHz. and each cell has a length of 0.1 m. CEBAF will use 200 of these cavities in each of twoparallel linac segments. plus 18 cavities in the injector.

The functional characteristics of these cavities are described in the following section. Their fabri­cation. initial inelastic tuning. and processing are described in Section 4.8. Assembly and testing are des­cribed in Section 4.9. The heat dissipated in these cavities is described in Section 4.10.

Functional Characteristics and Basis for SelectionAll research and development on the CEBAF-Cornell cavities necessary to meet all of CEBAFs

requirements has been completed. Four prototypes of this cavity have been built at Cornell University.and all of these cavities have met or exceeded all requirements for CEBAF (Table 4.1). The finalverification of the suitability of these cavities for use in CEBAF was provided by a beam test of a pairof these cavities in the CESR e ~ e - storage ring at Cornell in November 1984.

Table 4.1Parameter Choices for CEBAF

P3l"3C1eter CEBAF RequirementDemonstrated CapabilityCEBAF-Cornell Cavity

Frequency [MHz)

Gradient [MV/ml

Residual Q

Cavity current transport [(J.A)

Output current.BBU limit [(J.A)

Power into beam [kW/m)

Input coupling. QCXI

HOM power extractionrequirement [W/m)

900-3000

5.0

3 x 1O~

800

200

4

2.2 X lOt>

< 0.5

1500

8.2 (avg.)

3.9 x 1O~ (avg.)

22.000

10.000

26

7.000 to 1 x 10K

> 280

FrequencySince the beam extraction system can remove every third microbunch for delivery to a particular

end station. and since it is desired that the bunches arriving at a particular end station be not more than3.3 ns apart. the CEBAF RF frequency must be at least 900 MHz. An upper limit of 3000 MHz wasestablished because no developmental work has been done on practical superconducting acceleratingstructure above this frequency. Keeping the number of modules from becoming too large favors lowfrequencies. as does the fact that low frequencies permit the use of somewhat higher cryogenic temper­atures. High frequencies are favored by the ease of handling individual modules and avoiding theircontamination. and by the fact that the lower cavity surface area associ~ted with higher frequenciesreduces the probability that there will be a significant defect present on the surface. The Cornell prototypesoperate at 1.5 GHz. and this has been selected as the frequency of the CEBAF-Cornell cavities.

GradientThe accelerat:lg gradient specified for the CEBAF-Cornell cavities is 5.0 MV/m. In a total of ten

tests conducted at Cornell on the four pro:otypes. each with a separate surface treatment. the averagegradient reached was 8.2 MV/m: these ten tests represent all (including the worst) tests conducted onthese cavities following the removal of initial design defects. Among these ten tests. the first cooidownof each cavity met or exceeded the CEBAF gradient specification. The accelerating field achieved ineach of the four prototypes was between 6.0 and 15.3 MV/m (Figure 4.2). all exceeding the CEBAFrequirement of 5 MV/m. The beam test in CESR demonstrated the feasibility of reaching gradients inexcess of 5 MV/m in the presence of substantial b~am current. Since the gradient in each cavity at CEBAFwill be controlled separately (see Chapter 6). the maximum capability of each cavity can be fully exploited.

62

•

E16""-

>~ 14

"012

(I) 10

Cl8

c:: 6as 4...(I)

2(I)0 00« LE5-3 LE5-4 LE5-5 lE5-6

Machine: Vacuum Components

CEBAF-Cornell 5-cell cavities

m-e 12

010

"08

(I)-I 6~

Cl 4~

as 2~

"0 0U)(I)

c:LE5-3 LE5-6

CEBAF-Cornell 5-cell cavities

Figure 4.2. Residual high-field Q and accele:r,lling fie:ld. The: dashe:d lines represent design goals. The: residaal Q shown for LES­S (marked by an aste:risk) was me:asure:d at 12.0 MV/m: all others were me:asured at the grddie:nt shown.

Residual QThe RF losses in a superconducting cavity consist of two components: residual losses and BCS

(Bardeen. Cooper. and Schrieffer) losses. as discussed in Section 4.10. The CEBAF requirement for theresidual Q is 3.0 X 109

• The average residual Q obtained in the ten independent tests at Cornell was3.9 x 109 at high field. The residual Q obtained in each of the four prototype cavities. together withthe corresponding gradient. is shown in Figure 4.2. These values range from 3 x 109 to 10 X 109

•

equaling or exceeding the CEBAF requirement of 3 x 109• The beam test in CESR 2.1so demonstrated

the ability of these cavities to exceed the required Q in the presence of the beam.

Current TransportThe current-handling capability (e.g.. without cavity breakdown) required for CEBAF is 200 JLA

per beamlet times 4 beamlets. or 800 JLA. The current sustained by the cavity in the CESR beam testwas 22 rnA.

The output current required without beam breakup in CEBAF is 200 JLA. Based on the knownimpedances of the CEBAF-Cornell cavity. the beam breakup threshold exceeds 10 rnA. as discussed inSection 3.3. The heavy damping achieved for the important higher order modes in this cavity wasdeveloped as follows.

63

(:EBAF Design Report

Higher Order Mode SuppressionThe HOM coupling on this cavity was designed to provide heavy damping of HaMs. with panicular

emphasis on those with high intrinsic impedances (r/Q). The lengths of the shoned stubs on both theHOM coupler and the fundamental power coupler were determined on this basis. ?.s were the dimensionsand locations of the couplers. Intrinsic impedances and Qc", values wl::re measured using bead pulls andprobes on copper models. and computed using the programs SUPERFISH and URMEL. The accuracy ofthe r/Q and Qcx. values was confirmed by measuring instability thresholds in the beam test in CESR;resonance widths observed in the beam test provided funher verification of the Qcx. values of imponantHaMs.

All imponant HaMs are damped successfully to 5 x 1O~ ~ Q ~ 1.7 x lOs. The cavity thresholdinstability current was measured in the CESR beam test by mechanically sweeping the resonant fre­quencies of the cavity during the beam test in order to provide an indication of the instability probabilityone would encounter if the HOM frequencies were selected at random within a panicular frequencyrange. This measured probability distribution was compared to one computed by a Monte Carlo program(written by Roben Siemann). The input to this program consisted of HOM impedances. frequencies.and external Q's obtained by measurements and by computations using SUPERFISH and URMEL. Theagreement between the beam measurement and calculation is shown in Figure 4.3. Note that the measuredinstability threshold current is within a factor of 2 of the computed current at all probability levels. Thisis evidence that no serious errors or omissions occurred in determining the cavity impedances.

•4>-

-.0 .3as.00...a. .2>-

-.0as .1II)

c:

C0.1 1.0 10 100 rnA

Figure 4.3. Comparison of instability probabilities measured in Cornell beam tcst with predictions based on bench measurements.

The HOM power extracted through the fundamental power coupler is absorbed by an HOM filter.The two arms of the HOM couplers are at right angles to each other to couple effectively to bothpolarizations of the deflecting modes. Quadrupole and sextupole modes are also adequately damped bythese couplers. The waveguides leading out of the HOM couplers are cut off at 1900 MHz. so no separatefilters are required on these couplers to reject the fundamental power (1500 MHz).

The HOM couplers on the CEBAF-Cornell cavity provide heavy coupling for both accelerating anddeflecting HaMs. Only transverse instabilities are of concern in CEBAF. since the orbits are isochronous.However. the amount of HOM power generated in CEBAF is dependent on the QC:XI values of theaccelerating modes. Since the HOM power generated by the CEBAF beam is less than 0.5 W/m­sufficiently low to be absorbed at liquid helium temperature-the HOM waveguides are designed to beterminated in the liquid helium vessel within the cryostat. as described in Section 4.4.

Fundamental Power CouplingAt full beam current and an accelera[ing gradient of 5 MV/m. 4 kW/m (or 2 kW per cavity) of

fundamental power must be coupled into the beam in CEBAF. The power couple-d into [he beam in theCESR beam test was 26 kW/m.

64

I

I

Machine Vacuum Components

The input coupling (fundamental mode Q) required for CEBAF is 2.2 x 10". This value representsa factor of 3 overcoupling at 5 MV/m and 200 ~A per beamlet. and was chosen because. at the designcurrent. it requires only 20% added incident power. permits the cavity to be operated stably at as littleas 0.83 MV/m. and triples the control bandwidth of the cavity. The input coupler of the CEBAF-Cornellcavity is designed in such a way that. by varying the location of the step between wide and narrowwaveguides. the QC1. value can be varied from 7.000 to I x 10".

Other ParametersAs previously mentioned. the amount of HOM power induced by the CEBAF beam is less than 0.5

W/m. In the CESR beam test the cavity extracted more than 280 W!m of HOM power.The minimum acceptable beam hole diameter in the cavity is 3.9 cm. to ensure that the cavity

aperture is not the smallest aperture in the linac. The CEBAF-Cornell design has a beam hole diameterof 7.0 em.

Other properties of the CEBAF-Comell cavity are an intercell coupling of 3.3% and an rlQ valueof 960 ohms/meter.

In summary. all properties exhibited by the CEBAF-Comell cavity meet or exceed the requirementsofCEBAF.

4.3 Cryostats

Cryostats serve the function of providing a suitable environment for the cavities. The cryostatsremove heat to maintain the cavities at a temperature of 2.0 K. and they provide the cavities with anambient magnetic field of less than 5 milligauss. Other devices. such as waveguides and beam pipes.connecting the cavity to objects outside the cryostat are also considered to be part of the cryostat. Forpurposes of this report. a cryostat is considered to be the cryogenic housing for one pair of five-cellcavities. plus the auxiliary equipment just described. A cryostat with cavities installed. as discussed inSection 4.4. is defined as a cryo-unit. CEBAF requires 209 cryo-units: 200 for the two linac segmentsplus 9 for the injector.

The components and design of the cryostat are described in Section 4.4.

4.4 Cryo-UnitsOverview

As stated in Section 4.3. a cryo-unit is a cryostat with two cavities installed (Figure 4.4). A cryo­unit. with the simple addition of a pair of end caps. is the smallest individually testable acceleratingcomponent in CEBAF; relatively small units have been chosen intentionally to maximize the probabilitythat each unit will meet specifications when first assembled and to minimize the number of neighboringcavities subjected to the risk of contamination should an adjacent cavity require servicing. As discussedin Section 4.5. four cryo-units plus some other components comprise a cryomodule.

Physical DescriptionAs previously stated. a cryo-unit is a cryostat with two installed cavities (Figure 4.4). All internal

auxiliary components associated with both the cavity and the cryostat. with the exception of the referenceprobes on the cavities and their mounting hardware, are defined to be part of the cryostat. The principalcomponents making up the cryostat are as follows.

The cryostat consists of an outer vacuum vessel, a 40 K to 60 K helium gas cooled radiation shield,a two-layer magnetic shield, inner <Jnd outer superinsulation blankets, a liquid helium vessel, radial andaxial support rods for the helium vessel. a feedthrough tube for 2.2 K helium, two dynamic tuners forthe cavities, three niobium beam-pipe extensions, two stainless steel bellows beam-pipe extensions. twobeam-line gate valves, four niobium waveguide elbows for the HOM couplers. four waveguide loads forthe HOM couplers. two stainless steel supports for the outboard ends of the cavities, two Kapton windowsand short niobium waveguide extensions for the fundamental power couplers, two copper-plated stainlesssteel fundamental power waveguides with integral Kapton windows. two external copper-plated stainlesssteel fundamental power elbows with thin ceramic windows. infrared detectors, and arc detectors, in­strumentation cabling and feedthroughs. two sets of bellows-sealed rotary feedthroughs and shafts fordriving the tuners, one heater, two thermometers. and a complete set of assembly hardware.

65

-

CEBAF Dc._ign Repon

*H

FiRUre 4.4. Top view of a CEBAF cryo-unil. Asterisked items shown only once. CA. Vacuum ~hell flange and captured seal ring;B. HOM load; C. Inner magnetic shield; D. Outer magnetic ~hield: E." 2 K return helium connection and helium vessel: F. 401060 K radiation shield: G." Beam·pipe flange surface on end valve: H." 2.2 K helium supply line; I. Outboard cavity suppon:1." Shield ~uppon bumper; K. Ca"itl': L." Axial suppon: M. Fundamental waveguide; N. Rotary feedthrough; 0." Tuningmechanism: P. Superinsulation: 0." Helium v~1 ~uppon rod.)

Functional CharacteristicsWith reference to Figure 4.4. the outer vacuum vessel. made of stainless steel. houses the insulation

vacuum. The flanges at the ends of this vessel are machined after welding to permit alignment of successivecryo-units without the use of adjustable devices. A study of the manufacturing tolerances shows that acryomodule can be assembled and installed within the tolerance specified in Section 5.2. The method ofjoining one cryo-unit to the next or to an end cap is described in Section 4.5. The vacuum vessel alsohas a "top hat" on the side to permit installation of waveguides and other instrumentation devices. Thecylindrical cover of the top hat is O-ring sealed to the remainder of the top hat to facilitate access.

Tnermal radiation from room temperature is intercepted by a radiation shield, cool::d to 40 K to60 K by helium gas supplied directly by the refrigerator. The shield is made of two layers of stainlesssteel sheet. spot-welded together in a quilted pattern. and hydraulically expanded to provide passagesbetween the two sheets. The ends of the shields have the two sheets welded together to form a vacuum­tight seal. Tubulations are welded to the shield to permit cold helium gas to enter at one end of theshield and to exit from the other end. The opening in the shield at the top hat is closed by a cop~r

cover plate.Proper magnetic shielding is important during the cooldown of the cavity as it passes the critical

temperat:Jre to guard against magnetic flux trapping within the niobium cavity that could cause a severedegradation in Q. Magnetic shielding can be accomplished either actively by a few wires. arranged tocompensate for all exterior magnetic fields. or passively by adding magnetic shielding around the cryostat.The passive method has been chosen. because the active method requires that the counteracting fieldbe present at all times. With the active method. turning off the bucking field. followed by an accidentalwarmup to a temperature above the superconducting transition temperature, followed by re-eooling,would cause flux to enter the cavity and be trapped. Two layers of an iron-nickel alloy such as Coneticare planned. The inner layer is wrapped on the helium vessel, and the outer layer lines the interior ofthe thermal radiation shield. The magnetic shielding is also continued through the top hat region by theuse of separate pieces which have adequate overlap of the main shield.

66

Machine Vacuum Components

A blanket of superinsulation consisting of 60 layers of aluminized mylar separated by nylon spunbonded cloth covers the outside of the shield. This blanket greatly reduces the thermal radiation loadreaching the shield. and also reduces the thermal conductivity of any gases which may be present butnot cryopumped. A blanket of 20 layers of superinsulation is installed on the inside of the thermalradiation shield and serves the same functions as the outer blanket. but does so with respect to thethermal load on the liquid helium vessel.

The liquid helium vessel contains a pair of cavities. and during operation is filled with liquid heliumat 2.0 K to somewhat above the bottom edge of the end pipes at the top of the vessel. Helium vapor at0.031 atmosphere is present above the liquid. A machined and accurately positioned plate is welded intothe cryostat at the center of the side. A frame is welded around the perimeter of this plate. and a skirtis welded to the frame before the plate is machined to ensure that effects of weld warpage are negligible.During welding. a fixture holds this assembly in place in the main pipe forming the helium vessel. Afterthe cavities and their associated equipment have been installed in the main pipe. the end domes andend-dome centers are welded in place.

The liquid helium vessel is supported at each end by four stainless steel rods mounted between thehelium vessel end dome and the vacuum vessel flange. As shown in Figure 4.5. the rods are positioned

A B c D

K

J

F"agun 4.5. End view of a cryo-unit showing suppons for tile liquid helium vessel. CA. Vacuum shell flange: B. Inner magneticshield; C. Outer magnetic shield: D. 2 K return helium connection and helium vessel: E. Superinsulation: F. 40 to 60 K radiationshield: G. Shield suppon bumper; H. Beam-pipe /lange surface: on end valve; I. Helium vessel suppon rod: J. Fundamentalwaveguide: K. Rotary feedthrough shaft; L. 2.2 K helium supply line.)

in a double-X (XX) panern such that pitch. roll. yaw. and radial translations are constrained. The mountsat the end dome are beyond center such that the contraction of the vessel towards center is slightly lessthan the graded contraction of the entire rod upon cooldown. Thus the rods go into slight tension uponcooldown and the liquid helium vessel remains centered in the vacuum vessel. Two axial supports (item

67

CEBAF Design Rc:pon

L. Figure 4.4) are used to prevent axial (=) motion of the helium vessel; each of these supports lies inthe same plane as the beam axis. All of these supports are tied to the 40 to 60 K heat shield by meansof coppc:r braids to reduce heat conduction into the 2.0 K bath.

The cryostat houses part of the transfer line for the 2.2 K supercritical helium supply circuit. Thepipe for this line passes through the liquid helium vessel rather than around it. This is accomplished withnegligible heat transfer during the cooldown cycle by isolating the tube within another internally super­insulated tube that tunnels through the helium vessel. Thus the vacuum vessel and 40 to 60 K heat shieldneed not be made larger in diameter. The tunnel-tube. welded to the helium vessel. has a hydroforrnedbellows at one end to accommodate dimensional change due to temperature differentials.

Figures 4.4 and 4.6 illustrate the tuning mc:cl1anism used on the cavity. Dynamic tuning of the cavityis supplied by a mechanical tuner which a!"plies axial compression to the cavity between the centers ofthe first and fifth cells. A collar is fitted to each of these cells. and each collar is pinned to a yoke. Onone side of each cavity. a differential ball screw is used to adjust the dimension between yokes. On theother side of each cavity. a piezoelectric stack provides fine adjustment of the same dimension betweenthe yokes. The design is such that the piezoelectric stack is in compression when the linkage betweenthe yokes is in tension. A stainless steel worm gear on a bronze spur gear provides torque multiplicationto drive the differential ball screw. The worm gear is driven by shafts which are linked through bellows­sealed rotary feedthroughs to a stepping motor outside of the vacuum vessel. The reaction torque of thegear reducer is balanced by the off-center axial force provided by the drive linkage. The operationalaspects of the tuner are described in Section 4.10.

Worm gear reducerand right angle drive

Shaft fromrotary vacuumteedthrough

End

Note: Couplers not shown

ball screw

Piezoelectricstack in link

Pin connectionto collar

Fi~re 4.6. Frequency tuner for the CEBAF-Comc:ll callity.

68

Machine Vacuum Components

As shown in Figure 4.4. niobium beam-pipe extensions are used between the two cavities. and atthe outboard end of each cavity. RF fringing fields from the cavities are cut off exponentially in thebeam pipes. and the surface must be superconducting for 14 cm from the last cell iris at each end ofeach cavity in order to avoid excessive heat dissipation due to these fringing fields and their associatedsurface currents. The spacing between cavities within the cryo-unit is maintained at 1.25 wavelengths(2.5 cell lengths) to reduce cross talk between the cavities to a negligible level. This spacing also makesit possible to power a pair of cavities from a single power source. and to divert the reflected power intoa dummy load. should this capability ever be needed.

Outboard of the outer niobium beam-pipe extensions. which are 7 em in diameter. a pair of 3.8-cm(1.5H

) 10 stainless beam-pipe extensions is added. These beam pipes contain integral bellows to com­pensate for differential thennal contraction between the cavity and helium vessel. and to pennit thecavities to change in length due to tuner operation. The relatively small diameter of these pipes reducesthe amount of blackbody radiation which can enter the cavities at the ends of a cryomodule. and alsoprotects the cavities in adjoining cryo-units from cross talk via those HaMs which are not cut off by thediameter of the niobium beam pipes. This cross talk is also severely inhibited since HaMs are not tunedto a controlled frequency. and the probability that two consecutive cavities will have HOM frequenciesclose enough to fonn an effective collective mode is very small. The stainless steel beam-pipe extensionsare pre-welded to integral dished disks which match and are welded into the domes on the ends of thehelium vessels. These disks have a larger diameter than the gate valves (described next) and are necessaryto pennit clean assembly (without disturbing the cavity vacuum).

A gate valve is installed at each end of each cryo-unit. and is used to keep the cavities under vacuumafter initial testing of the cavity pair (described in Section 4.9). The gate valves also pennit any cryo­unit to be removed from a cryomodule for servicing without destroying the vacuum in adjacent cryo­units. The gate valves have a 3.8-em (1.5 H

) clear diameter. and are all-metal construction except forViton a-ring gate seals. Viton a-rings are used because wann sealing is required but sealing along thebeam line at cryogenic temperatures is not. These valves have manual actuation mechanisms. The valvesare mounted such that the seals face the exterior of the cryo-unit.

Niobium elbows are used on each of the four HOM coupler waveguides in a cryo-unit. These elbowsmust be superconducting because the fundamental field amplitude is cut off slowly in these waveguides.and is too large at the location of the elbows to pennit use of nonnal conducting waveguides. The elbowsredirect the HOMs being extracted so that the required diameter of the liquid helium vessel is no largerthan necessary. Copper braid straps connected to the one elbow which protrudes above the liquid heliumlevel enhance cooling of this elbow.

Waveguide loads connect to the waveguide elbows just mentioned. These loads have a VSWR(voltage standing wave ratio) less than 1.5 between 1.9 GHz and 5.0 GHz, and are used to absorb partof the HOM power geneated by the CEBAF beam. This power is less than 0.5 VI per active meter ofcavity. These loads will be sufficiently clean that they can be connected directly to the cavity vacuum.

The outboard ends of the cavities are supported in the liquid helium vessel by stainless steel sheetswhich hold the ends of the cavities precisely in the center of the liquid helium vessel. These sheetsprovide considerable rigidity in the transverse direction (both vertically and horizontally). but are flexiblein the longitudinal direction to permit motion due to differential thermal contraction and due to tuneroperation.

Kapton windows are used on each of the fundamental power waveguides to pennit the cavities tobe sealed at an early stage of assembly. Kapton is chosen because of its small RF reflection. good vacuumproperties. and excellent radiation resistance.

The Kapton windows in the fundamental power waveguides are followed immediately by shortlengths of niobium waveguide so that seals at the ends of these extensions can be made and brokenwithout disturbing the seals on the Kapton windows.

Additional lengths of niobium waveguide are used to connect the waveguide extensions to the plateat the side center of the liq~id helium vessel. The waveguides are made of niobium because they aresurrounded by helium at 2.0 K, and the use of superconducting material in this location minimizes heatlosses due to RF power dissipated in the waveguide. This waveguide is annealed at::s: 10.7 torr and 1800°Cto render the niobium dead soft before welding to its flanges (which must remain hard). The flanges onthe cavities and on the waveguides are all machined after welding to obtain flat. smooth sealing surfacesin the correct locations. Care must be taken not to introduce a bend in the cavity during inelastic tuning,described in Section 4.8. With all of the flange locations precisely determined, and with the location of

69

I

CEBAF O..."ign Rqxm

I I

the plate in the side of the helium vessel precisely determined. the two cavities are attached to the heliumvessel without the requirement for any flexibility other than the slight flexibility provided by the annealingof the waveguides. This procedure was used successfully in the beam test of a pair of cavities in CESR.The end of each cavity near the center of the helium vessel is thus supported by its fundamental waveguide.and the outboard end of each cavity is supported by the stainless steel sheets previously described. Theniobium waveguide is made slightly short at room temperature to compensate for the fact that thestainless steel helium vessel contracts more than does the niobium waveguide: the cavity has sufficientflexibility to accommodate this differential contractlol'\ without inelastic deformation.

Stainless steel waveguides are used between the liquid helium vessel and the vacuum vessel. Thesewaveguides are plated on the inside with several skin depths of copper (to minimize RF losses withoutcausing too large a thermal conduction along the length of the waveguide). Each has an integral Kaptonwindow with a I-em-diameter hole in it. This window is located at the waveguide flanges. which aremaintained at 60 K. and serves to intercept most of the infrared radiation transmitted down the waveguideso that it does not get absorbed at 2.0 K. The hole in the window permits vacuum pumping on bothsides of the window without use of an external pumping manifold to connect the two sides. Thesewaveguides pass through the plate on the vacuum vessel's top hat. and are vacuum sealed to this plateemploying hydroformed bellows which accommodate relative changes in length and transverse displace­ments.

The sum of the heat conducted through the waveguide from room temperature and of the heatgenerated by dissipation within the waveguides is minimized by placing the 60 K heat intercept 8 emfrom the 2.0 K end of the waveguide.

Two copper-plated stainless steel fundamental power elbows are used outside the insulation vacuumvessel. Each of these elbows is equipped with a thin ceramic window which separates evacuated waveguidefrom air-filled waveguide. Ceramic is used here because Kapton has some permeability to most gasesand a high permeability to helium gas. Each elbow is also equipped with an infrared detector and avisible arc detector to detect anomalous conditions at any of the windows. and also has a pumping portto permit the waveguide to be pumped as described in sections 4.5 and 4.7.

The stainless sh:el plate at the side of the liquid helium vessel and the flat plate at the side of thevacuum vessel are equipped with instrumentation feedthroughs. These feedthroughs include two coaxialSMA bulkhead feedthroughs (on each plate) for the semi-rigid cables leading from each of the referenceprobes on the HOM couplers. There are also feedthroughs for two sets of four leads for thermometers.two sets of two leads for the piezoelectric tuning adjusters. and one pair of leads for a heater. Also inthis region are the two bellows-sealed rotary feedthroughs for operation of the tuner.

The heater in the cryo-unit is to calibrate the heat load measuring instrumentation (described inSection 4.10). and to accelerate warmup of the cryo-unit when desired.

A thermometer is attached to the cavity. and is used to determine the temperature. This is particularlyuseful during cooldown and warmup. Another thermometer. located at the bottom of the cryostat.provides temperature information during cooldown and warmup.

Assembly hardware includes nuts. bolts. and spring washers where required. Also included are thevacuum sealing materials. At room temperature. a-ring seals are used on the vacuum vessel. Indium isused to seal between niobium and niobium. or between niobium and stainless steel. It has been foundthat indium seals are reliable if the flange faces are flat and smooth. and if the flanges are at least 1.0em (0.375") thick so that they do not distort excessively. It is also required that spring loading of theseal be maintained. Cryogenic joints between stainless steel and stainless steel are made with knife-edgeseals using copper gaskets. Any joints which do not require disassembly for servicing purposes are weldedwherever possible.

4.5 CryomodulesOverview

A cryomodule normally consists of four cryo-units connected together. with an end cap on each end(figures 4.7 and 4.8). One cryomodule is used between each set of magnetic focusing elements. CEBAFuses 50 cryomodules in the linac segments. and two standard cryomodules. together with a cryomodulecontaining only one cryo-unit. in the injector. It is planned that the first seven cryomodules will bepretested and installed in the linac. To service any particular cavity. its entire cryomodule would beremoved from the linac and a spare unit installed in its place.

70

...

I

Machine Vacuum Components

cap

2.2 K supplvand 60 K shieldU-tubes

..:: .:.-

Transfer line

. - - ... -

Beam handlin g devicesat room temperature

.. Fast-closing valveevery five modules

.to; y o-unitCavity pitch

0.5 m 1.91 m(1.64 ttl (6.27 ttl

Ciyomodule pitch1----------- 9.40 m --------_.-l(30.e4 tt)

. - [ .. ' ~!----'-~~I­3.35 ml(11 ttl

Figure 4.7. Schematic plan view of a cryomodule installed in the Iinac. (RF equipment not shown.)

Spool

Captured seal ring

Floor

Liquidhelium surface

Bayonet in end cap

Helium vapor

0.686 m(27 in.>

J stands-~~-\~~-.. _ - :.. ",'"

o -,' 0" ..

Figun 4.8. Side view of the end of a cryomodule.

71

II II •• • •

•CEBAF Design Repon

Physical DescriptionIn addition to four cryo-units. which are described in Section 4.4. each cryomodule includes three

spool pieces which span a 15-cm space between cryo-units. Within each of these is installed a beam pipewith bellows to accommodate thermal contraction. a pinch-off pump-out tube. a pipe with bellows toconnect the liquid helium vessels. a pipe with bellows to connect the 2.2 K supercritical feed lines.connections for joining the two 40-60 K thermal radiation shields. a set of thermal radiation shield bridgepieces. two sets of magnetic shielding bridge pieces. bridge wraps for the 6O-Iayer superinsulation blanket.and a bridge wrap for the 20-Iayer superinsulation blanket. A 7.6-cm (3") diameter pumping manifoldruns along the outside of each cryomodule for pumping on the fundamental power f~ed waveguides; thisapparatus is described in Section 4.7.

A supply end cap is bolted to the vacuum vessel on one end of the cryomodule. and contains thebayonet sockets for two V-tubes (Figure 4.9). one supplying 2.2 K supercritical gas and one supplying4V-6O K shield gas. The first is internally connected through a IT valve to the input of the helium vesselin the first cryo-unit. and also continues in parallel with the IT valve into the transfer line passing throughthe first cryo-unit. The 40-60 K gas input is connected through a thermal shield in the end cap into theshield in the first cryo-unit. This end cap also contains all the relief valves and rupture disks on both thehelium circuits and the vacuum tank. A beam-pipe extension with thermal expansion bellows and atransition from 2.0 K to room temperature is also included in this piece. as are closures for both layersof magnetic shielding and closures for the superinsulation blankets. Vacuum vessel closure is providedby a dished cover which passes over a motorized gate valve located on the beam pipe.

At the other end of the cryomodule. a return end cap is similarly attached to the cryomodule. andcontains bayonet sockets for three V-tubes. One of these sockets is for continuing the 2.2 K supercriticalgas supply to the next cryomodule. one is for continuing the shield gas to the next cryomodule. and oneis for connecting the 0.031-atmosphere helium gas exhaust line to the return transfer line. The componentsin this cap are similar to those in the supply end cap. except that the additional V-tube connection isconnected into the pipe at the top of the helium vessel and contains the module's liquid-level gauge.Another exception is that the connection between the 2.2 K supply line and the liquid helium vessel isabsent. Pumping on the insulation vacuum is provided at only one end. as discussed in Section 4.7.

Each cryomodule also has two detachable support stands cradling the vacuum vessel at the innerend of the first and fourth cryo-unit. Each stand's feet contain the adjustment mechanisms necessary toposition the module on the beam-line axis.

Functional CharacteristicsThe principal characteristics of the cryomodules are the same as those described in Section 4.4; only

the features not characteristic of individual cryo-units are described here.As shown in Figure 4.7. the cryo-units are assembled with the top hats in a left-right-right-left

pattern. The purpose of this pattern is to reduce to a negligible value the effect of small transverse beamkicks caused by the asym."11etric fie ids in the fundamental power couplers.

The thermal radiation shields and magnetic shields between cryo-units are attached to one cryo-unitand are free to slide with respect to the other one to accommodate thermal contraction. With the inclusionof the end caps. a nearl;' closed volume is provided with respect to thermal radiation shielding andmagnetic shielding.

The spool pieces. which slide over one of the adjacent cryo-units during assembly. have their endsmachined to provide accurate alignment of the cryo-units within the cryomodule without adjustablealignment devices. The spool pieces are mounted in compression between seal rings bolted to the backface of the vacuum vessel flanges. Connection is provided by a series of threaded rods between sealrings. These seal rings each contain two O-rings. with one sealing to the cryo-unit vacuum vessel flangeand the other to the spool piece.

The JT valves are adjusted to maintain the level of liquid helium above the bottom of the pipesconnecting the helium vessels of the cryo-units. and below the center of these pipes. Vpon passingthrough the IT valves. the supercritical 2.2 K helium is expanded to become 2.0 K liquid. Additionaldetails are included in Section 4.10.

72

I I

Machine Vacuum Components

Removable U-tube

To guardvacuum valve

To valve andrelief valve tiedto guard vacuum

Taperedsupennsulation

1He flow

I

U-tube vacuum

O-ring static seals

Chevron seals

To 3 atmHe valve

shell

Labyrinth seal

III................. ",.- 77 K intercept

Cryostat vacuum

FIgIUe 4.9. Subatmospheric cryogenic transfer bayonet with guard vacuum.

73

•CESAF Design Rc:pon

4.6 Cryomodule Interconnection

I

OverviewCEBAF uses two antiparallel linac segments. each containing 25 cryomodules and therefore 100

cryo-units. or 200 five-eell cavities. Each of the two linac segments is designed to produce an energy gainof at least 0.5 GeV per pass. The linac segments are connected via warm recirculation beam lines. Alongeach linac segment the beam pipe is warm between cryomodules.

Physical DescriptionIn addition to elements previously described. the linac segments include warm 3.8-cm-diameter beam

pipes passing through the magnetic elements and connected between cryomodules. These room-tem­perature beam pipes have pump-out provisions and bakeout provisions. and are continuously pumpedby 60 liter/sec ion pumps. They have 2.5-em-diameter collimators to provide the smallest apenuresformed in the vicinity of the cryomodules; these apenures are many times the diameter of the beam.This ponion of the vacuum system is funher described in Section 4.7.

The linac segments also contain fast-elosing valves every five cryomodules.At each end of each linac segment. there is an additional 1 m of beam pipe which is maintained at

2.0 K.

Functional CharacteristicsCondensable gases from warm sections must be prevented from entering and contaminating the

cavities. This is accomplished by baking out and pumping out the warm sections prior to cooling downthe cavities. In addition. the sections are pumped continuously by 60 liter/sec ion pumps. For funherprotection. there are fast-elosing gate valves every five cryomodules. The vacuum system and bakeoutprocess are described funher in Section 4.7.

The shield gas temperature increases from 40 K to 60 K as it passes through either 12 or 13cryomodules in series. (The 25 cryomodules of a linac segment are split 12 and 13 for this panicularpurpose.)

The additional one-meter beam pipes maintained at 2.0 K at both ends of each linac segment adsorbcondensable gases which could otherwise enter the linac segments and contaminate the cavities. Thesepipes must be cooled down before the cryomodules are. and warmed up after the cryomodules are.

4.7 Vacuum SystemOverview

A high vacuum is required in an accelerator to minimize the amount the beam is scattered bycollisions with gas panicles. In a superconducting accelerator the vacuum also protects the cavities. byminimizing the raL~ at which material accumulates on the superconducting surface, thereby degradingthe Q. An additional function of the ....acuum system in a cyrogenic machine is to maintain the insulatingvacuum in the cryomodules. Figure 4.10 is a schematic diagram of CEBAFs beam-pipe vacuum system.

Vacuum System: Linac Segment

Physical DescriptionThe beam-line pumping of the linac segment is described in Section 4.6. Pumping on the fundamental

power waveguides is provided by a 7.6-cm (3") diameter manifold that runs the length of each cryo-nodule.This manifold is equipped with a valve for initial pumpdown. There is also a 60 liter/sec ion pumpmounted near the center of the mCl'lifold. Jumper lines connect the manifold to each of the eightfundamental power elbows. '

Each cryomodule has a valved pon fo~ initial pumpdown of the insulation vacuum. Also includedis a thermocouple gauge to monitor vacuum quality. In case of a small helium leak to the vacuum space,this pon acts as a connection point for a 50 liter/sec turbomolecular pump with backing roughing pumpwhich would be installed. Provision has been made to purchase five pump systems for this use.

Functional CharacteristicsIt has been empirically determined that the pressure in a superconducting cavity should be ~ 10

07

torr at the time the cavity is cooled down in order not to suffer degradation of performance due to theadsorbed gas layer. Once the cavity is cold, the surfaces which were contributing to this pressure byoutgassing will no longer outgas at a significant rate. However. any surface which remains warm and

74

-

•

Typical cold beam pipe.6 places

\...Typical beam-line pump.30 places

•

,

.1. I •

Typical roughing manifoldin recirculation arcs,

32 places \

Injei::l:or system

Experiment

I •

Machinc Vacuum Componcnts

.t.. Fast-closing valves

•

FIgIU'e 4.10. Schcmatic of major beam·pipe vacuum componcnts.

which remains directly connected to a supercondueting cavity is at a pressure of =!iO: 10-9 torr before thecooldown is started. Experience shows that at this pressure outgassing and gas adsorption are sufficientlyslow that the cavity Q would be degraded only 10% in a period of 2 years, if the outgassing rate remainedconstant during the period. The 10% degradation is the maximum that would be expected, since theoutgassing rate decreases with time.

The cavity pairs are evacuated to a pressure of =!iO: 10-7 torr immediately following assembly, andremain evacuated thereafter. When the beam pipes between crye-umts are evacuated, they will beevacua~d to a pressure of =!iO: 10-7 torr before their pinch-off tubes are sealed. Since the temperature ofthese pipes approaches 2.0 K, their outgassing is eff~etively stopped during operation. The beam pipesbetwe..':n the end aye-units in a cryomodule and the cryomodule's end gate valves experience a tem­perature gradient between 2.0 K and room temperature. Accordingly, with the gate valve on the crye­unit closed and the one on the cryomodule open into a gump, this section of beam pipe is baked atl00"C to remove desorbable gas until a pressure of =!iO: 10· torr at room temperature can be achieved.The end valve on the cryomodule is then closed and the one on the crye-unit opened. The end cap coverof the cryomodule is then installed.

The warm beam pipes connecting cryomodules are baked a. l00"C with the end valves on thecryomodules closed, until a pressure of =!iO: 10.9 torr at room temperature can be achie""ed in this pipe.At this time, the end valves on the cryomodules are opened.

Fast gate val·...es are placed after every fifth cryomodule to limit damage to the machine in the eventof a sudden vacuum accident. Such an accident could increase the amount of gas and particulates in thebeam pipe, and they could coat the supercondueting surfaces, causing sharp degradation in performance.These valves do not seal tightly, but present a near-seal to give the tightly sealing motor~ven valvesat the ends of each cryomodule time to close without severe degradation to the downstream cavities.The fast-dosing valves are placed every fifth cryomodule to limit damage to 10% of the cavities.

The pressm"e at the pump which maintains the vacuum in the waveguides will be =!iO: 10-7 torr priorto the beginning of cooldown. This pressure is specified to be relatively high, even though part of thesystem remains warm, because the fields in the affected components are much lower than those in thecavities themselves.

75

I

CEBAF Design Repon

I I I

A pressure of :s:; 10'" torr in the insulation vacuum at the beginning of cooldown of the cryomoduleprovides adequate thermal insulation.

Vacuum System: Recirculation ArcsThe major elements of the vacuum system for the recirculation arcs are the vacuum pipe with

associated pumping ports. the roughing manifolds and roughing pumps. and the high-vacuum ion pumps.The stainless steel vacuum pipe consists of 7.tH:m (3") 00 pipe with interspersed regions of altered

cross section because of the magnets and the beam collimators. To reduce gas desorption. all materialbombarded by synchrotron radiation is hydrogen degassed in a vacuum furnace. Connections betweenpipes and magnets are generally made at flanged joints.

The roughing system consists of 32 pumping manifolds and 16 portable 160 liter/sec turbo pumps.The manifolds have valved connections to the vacuum pipes of each of the vertically stacked recirculationarc beam lines. They also have a vacuum test port and a valved connection to a turbo pump port. Theturbo pumps are backed by 210 liter/min mechanical pumps. Both the turbo pumps and the mechanicalpumps are shared between linac segments. arcs. and the beam switchyard for respective roughing pro­cedures.

To allow modular isolation of selected regions of the system. three gate valves are placed in eachrecirculation arc beam line. two at each end and one in the middle. Four additional gate valves areplaced between the spreaders/recombiners and the cold beam pipes which isolate the linacs.

High vacuum is obtained using seventy triode ion pumps operating at 30 liter/sec. They are directlyattached to the vacuum pipe by means of a flanged port. Individual ion pump power supplies are usedand system gauging is provided by ion pump readouts and thermocouple gauges distributed throughout.In addition to the other equipment. a total of 88 beam collimators are used to prevent beam loss insidethe superconducting linac segments.

Peak gas loads in the recirculatior. arcs will occur inside and downstream from the dipole magnetsas a result of synchrotron radiation striking the beam pipe wall. The ion pumps are placed to minimizethe scattering of beam electrons by background gas. After a week of running at full current. the beampower loss is less than 0.1 W in each of the arcs (Section 5.3).

Vacuum System: Beam LineThe beam-line vacuum system is similar to the system in the recirculation arcs. Thirty ion pumps

operating at 30 liter/sec are used to maintain the high vacuum. with roughing ports attached to sixteenof the pump mounts. Fast-closing valves are installed at four points in the beam switchyard and theexperimental areas. Seven additional beam-line gate valves provide modular isolation of selected regionsof the system. Such choices give vacuum performance similar to that of the recirculation arcs.

4.8 Cavity ProceduresFabrication

The CEBAF-Cornell cavities are fabricated from pure niobium sheet and plate. The niobium isinitially inspected for surface inclusions or scratches and for delamination. Any such defects. if found.are ground away when possible: otherwise. the material is rejected.

Niobium used for the cups used in the five cells is specified to be of higher purity than r':a:::tor grade(which is 99.85% pure). and to have an RRR (residual resistivity ratio) value of at least 120. lhis materialis deep-drawn into cups. The cups are then "yttrified" by exposing them to yttrium vapor at 12500C forfour hours: after cooling. the yttrium which has sublimated onto the niobium is removed with nitric acid.The purpose of the yttrium treatment is to getter dissolved gas from the niobium and hence raise thethermal conductivity by approximately a factor of 3. which in turn increases the size of defect the cavitywill tolerate while operating at 5 MV/m. The edges of the cups are machined. the material is cleanedwith solvents and acids. and the cups are electron beam welded. iris to iris (i.e.. small end to small end).to form ""dumbbells." The electron beam welds are to exhibit no residual cracks. either inside or outside.This set of welds is then ground with an aluminum oxide composition wheel to remove any roughnesscaused by the welding. The dumbbells are then subjected to an additional chemical clea-,ing..

The parts for the couplers are made from reactor-grade niobium. These parts are either deep-drawnor bent in a brake. and their edges are machined. Aanges. stiffening braces. and a few other parts aremachined from plate. These pans are then chemically cleaned and electron beam welded together. Thereare three iterations of machining. chemical cleaning. and electron h~am welding required. The final

76

.... II r

I I

Machine Vacuum Compo:lenl~

..

iteration of machining is for the purpose of ensuring that the waveguide and beam pipe flanges are flat.smooth. and in the correct locations. For the couplers. it is also important that there be no inside cracksin the welds. The couplers are completely welded together before begInning to weld them to cups otherthan the end cups. The final five welds join the cells at their equators. and are done with either a "rhombicraster" scanned electron beam. or a defocused electron beam. These welds are done from the outside.and must be full penetration and smooth on the inside. When any part of the cavity is at a temperatureof;:?; 200°C. the cavity must be in a vacuum of :s: 5 x 10.5 torr. or the thermal conductivity of the niobiumwill be degraded.

Inelastic TuningInelastic tuning is performed on the cells of the cavity to ensure that the field profile is flat and the

frequency is correct. The field profile is measured with a bead pull. and the frequency is determined byphase-locking an oscillator to the cavity resonance. with the cavity filled with 20"C nitrogen gas. and bymeasuring the frequency with a frequency meter.

The cavities are tuned to a room-temperature frequency of 1494.2 =0.1 MHz. This frequencycompensates for subsequent frequency shift due to cooldown contraction. for material removal duringfinal chemical polishing. for dielectric constant differential between nitrogen and vacuum. and for changein differential pressure between outside and inside the cavities. The operating frequency of the cavitiesat 2.0 K is 1497.000 MHz.

The tuning corrections required for each cell are determined by a perturbation theory calculationbased on the bead pull. The field flatness must be within =5% of the median.

Tuning corrections are made by inelastically deforming the cells; the cells are either stretched orcompressed axially. Plates which fit between cells are used to apply the required force. For the end cells.plates which grip the equatorial region of the cells are used instead. Care must be used in the tuningprocess not to introduce an angle between consecutive cells.

Fine adjustments to the coupling strength of the input coupler are made by inelastically deformingthe body of the coupler. The input coupling is to be adjusted to within =20% of Qc:JCt = 2.2 X 106

•

Elastic tuning processes are described in Section 4.10.

ProcessingFollowing inelastic tuning. the cavities are degreased. and then chemically cleaned by immersion in

a mixture of nitric. hydrofluoric. and phosphoric acids. Approximately 100 microns of niobium areremoved from the surface. This is done in two stages to prevent overheating; the cavity is mechanicallyagitated during the chemical processing to avoid the collection of stagnant ;>recipitates or bubbles on thesurface. Following re:noval from :he acid. the cavity is rapidly rinsed using 20 gallons per minute of purewater. sprayed at 10 psi onto the cavity surface. The water is purified by reverse osmosis. ion exchangeresins. an ultraviolet sterilizer. and 0.2-micron filters. It is stored under a nitrogen atmosphere and iscontinuously recirculated through the distribution piping. The cavity is ultrasonically rinsed in pure waterfor at least an hour. It is then rinsed with electronic-grade methanol in a class 10 clean roo~. and allowedto dry in a horizontal position.

4.9 Assembly and Testing<herview

The test protocol will evolve as the project progresses and CEBAF identifies and eliminates thecauses of ~ommon problems. It is planned that the first 50 cavities will be tested thoroughly as singlecavities. cavity pairs. and cryo-units. They will then be tested as cryomodules in the SRF facility on theCEBAF site. and again as installed cryomodules to ensure that their performance was not degraded bymoving them into the tunnel. Eighteen of these cavities will be installed in the injector at an early stageto permit system operation to be tested with actual RF systems and controls. plus beam. With experiencegained from this testing. it may be cost-effective to modify the test procedure by eliminating some ofthe tests.

It is also planned that two or three cryomodules will be tested in a string as they are installed in themain linac tunnel in order to assure proper system operations and to uncover possible system problemsearly. These tests will be accomplished using the actual RF system. the injector. and refrigeration systemfrom the test facility.

77

• ' '.~ ,~' •• " '~ • > >~, ...' , •• ' • "..

CEBAF Design Repon

CavitiesIndividual cavities are assembled in a class 10 clean room by placing niobium blank-off flanges over

the HOM waveguides and the beam tube at the fundamental power coupler end of the cavity. A referc:nceprobe is installed on the cavity. and its coupling is preadjusted on the bench to couple out approximately100 mW when the cavity is operated at 5 MV/m. A shorted niobium waveguide of appropriate length(13.018 em) is attached to the fundamental power coupler. The cavity is attached (in a class 100 cleanroom) to a test stand with variable coupling and pumping provided through the beam pipe at the HOMend of the cavity. The cavity assembly is evacuated. leak checked. and further evacuated until a pressure:$ 10.

7torr is reached. The cavity is then installed in a vertical cryostat. and cooled to the operating

temperature range of 4.2 K to 1.8 K. An oscillator is phase-locked to the cavity resonance. and the inputcoupling is adjusted to provide critical coupling. The Q of the cavity is determined from the decay timeconstant of the RF field when the input power is turned off. The accelerating field is calculated from thenumber of watts of power dissipated in the cavity (determined from the power incident on the cavity.and corrected for coupling line attenuation and reflections. if any). combined with the known geometryof the cavity and the Q of the cavity. The power coming out of the reference probe permits the calibrationof the reference probe with respect to cavity field to be established. As the incident power is raised. theQ of the cavity may acquire a field dependence. at which point the incident power. reflected power. andpower transmitted through the reference probe are used to determine both the field in the cavity andthe Q at that field. The maximum achievable field is measured. Measurement of the Q as a function oftt:mperature permits the residual and BCS contributions to be separated. The frequency of the cavity at2.0 K is measured. A calibrated probe on the shorted niobium waveguide attached to the fundamentalpower coupler permits the input coupling of the cavity to be measured.

Cavity PairsAn assembled cavity pair is shown in Figure 4.11.

HOM coupler

Niobiumbeampipe

Central portionof vessel head!

(shown in section>!

Reference probe

Isolation valve

F'agure 4.11. Cavity pair as assembled in a clean room.

A pair of cleaned cavities is assembled in a class 10 clean room as follows. All parts that are to beconnected to the cavity vacuum are to be equally clean. Kapton windows and short extension waveguidesare attached to the fundamental power couplers. With the short extension waveguide faces held againsta flat. metal-backed Teflon plate. the two cavities are connected together using the niobium beam pipewhich goes between them. Niobium elbows and HOM loads are attached to the HOM coupler flanges.

78

I I I I

•

IiIllI

Machine Vacuum Components

Niobium beam-pipe extensions. stainless steel bellows. and gate valves are anached to the cavities atboth ends. Preset reference probes are installed on the two cavities. Pump-out chambers are added toboth gate valves. Shorted niobium waveguide extensions of appropriate length are added to both of thefundamental power waveguides: these extensions are equipped with variable coupling probes. pumpinglines. and a precalibrated reference probe. The various pumping ports are connected to a test stationtop flange in a class 100 clean room via precleaned flexible hydroformed stainless steel bellows. Thecavity pair is then installed in either a horizontal "bathtub" cryostat or a vertical cryostat. The remainderof the test proceeds as described below. except that there are two cavities to be tested (one at a time)and the coupling is adjusted to be critical using the coupler which couples into the waveguide hat. Thegate valves are closed after the test is completed. When the indium seals to the pumping chambers arebroken. particularly on the beam line. the joint is faced downward to avoid having metallic flakes fallinto the gate valves or other parts which will later be opened to the cavity vacuum.

Cryo-UnitsA cavity pair is installed in a cryostat to form a cryo-unit as follows. The cavity pai:-. under vacuum

and supported by a carriage. has fundamental power waveguide extensions added. Indium is pre-formedonto the ends of these extensions. Tuners and instrumentation cables are added. The cavity pair is thenmoved into the central portion of the liquid helium vessel. with the cavities as far from the waveguidemounting plate as possible in the vessel. At this time. the helium vessel is externally supported by afixture which maintains its circular cross section. When the cavities are centrally located. they are movedtoward the waveguide mounting plate. and the indium-sealed waveguide joints are bolted. The stainlesssteel sheets which support the outboard ends of the cavities are added next. The cavity support carriageis then removed. and the various instrumentation. mechanical. and vacuum feedthroughs are connectedto the c~ntral plate. Leak checking and instrumentation checks are performed at this stage. Next. theend caps of the liquid helium vessel are welded both to the main part of the liquid helium vessel and tothe dished disk which is part of the cavity pair assembly. The welds are designed to be ground off andre-made as many times as necessary. Leak checking of the liquid helium vess~1 is :10W performed. Thefirst magnetic shield and superinsulation blankets are then wrapped around the vessel. The secondmagnetic shield is installed inside the thermal radiation shield and the helium vessel is then installed intothis shield. The second superinsulation blanket is now wrapped around the shield. This assembly is nextmoved into the vacuum vessel. and the support rods at each end of the liquid helium vessel are attachedand adjusted. The axial supports are then added. and the remaining waveguides. heat exchanger con­nections. mechanical shafts. and instrumentation cables are built up and completed in the top hat. Thethermal radiation. magnetic shields. and the superinsulation in the top t: . are closed and completed.and the vacuum vessel cylindrical cover is added. The waveguide elbows outside the vacuum vessel areadded. together with provision for pumping on the waveguides. Further leak checking is done at thisstage. End caps are added to the vacuum vessel. with special pumping chambers added to the gate valvesso that the gate valves can be opened during testing.

After adequate pressures have been reached in the various vacuum chambers (see Section 4.7). theshield is cooled down either by passing liquid nitrogen through it or by supplying 40 K helium gas to it.and the helium vessel is cooled to 2.0 K.

The cavities can be powered either by klystrons. or by lOO-wan transistorized amplifiers by usingan iris in the waveguide to resonate the waveguide between the cavity and the iris. By knowing thecalibration of the reference probes. the attenuation of the cables leading to these probes (which can bedetermined by reflectometry). and the power incident on the cavities. the input coupling is checked toconfirm that it has not been disturbed by the assembly process. Since the external Q of the cavity is nowvery low compared to its intrinsic Q. the intrinsic Q of the cavity can no longer be determined by thewidth of the resonance. In this configuration. the Q is determined by the increased rate of helium boiloffwith the RF on. Such a measurement requires that the pressure of the helium be accurately regulated.or the time constant with which the pressure changes ......ill be too long (of the order of an hour) to obtainan accurate reading: with suitable regulation. the time constant can be of the order of a second. Theoperation of the tuners and the ability to reach the correct frequency (1497.000 MHz) is checked at thistime. In addition. the maximum achievable field. the static heat leak. the component of heating due toRF dissipation in the input coupler (by its longer time constant for the heat to reach the 2.0 K bath).and rnicrophonics are all measured.

79

CEBAF Design Repon

CryomodulesCryomodules are assembled from four cryo-units and other components. as detailed in Section 4.5.

All beam-pipe components are precleaned and installed under clean room conditions. using either a locallaminar flow hood or a large clean room. Evacuation of the beam pipes is done according to therequirements stated in Section 4.7.

Cavity performance is again measured as it has been for the eryo-units. The static heat leak is re­measured. since there are different cryostat end and joint conditions than in the cryo-unit test.

Testing After installationThe cryomodules are installed in the tunnel using removable sun:.:=y monuments at each end of each

cryomodule. The precleaned warm beam pipes between cryomodules are attached to the cryomodulesusing portable laminar flow hoods to provide a clean environment.

If a cryomodule is connected to the common refrigerator. Q measurements must be made as discussedin Section 4.10. Other tests can be made as in the cryomodule case discussed above.

4.10 OperationRF and Beam

Operational aspects involving RF and beam are discussed in more detail in Chapters 6 and 8.Initial operation at a particular energy is established by bringing consecutive cavities on-line one at

a time. after deciding in advance the gradient at which each cavity is to be run. A low-current beamwith bunch trains of the order of 1 ~s long is used for tuning purposes. As each cavity is turned on. itsphase is adjusted by applying audio-frequency phase modulation until the energy variation caused bythe modulation is minimized and the energy is maximized (as determined by downstream beam positionmonitors). The gradienT in each cavity is then adjusted by an appropriate correction factor to make theoutput energy agree with the desired value as determined by a high-dispersion. high-precision locationaJong the beam line. Doglegs leading to successive turns are adjusted to maximize the energy gained onthose turns.

One algorithm for adjusting the gradient in each cavity is to raise all gradients uniformly. but tostop raising gradients in particular cavities when their maximum gradient capability is approached.

The gradient in each cavity can also be adjusted in a way that minimizes the RF heat dissipation(cryogenic load) in the linac by making use of the Q vs. gradient curve for each cavity. All gradientscan be raised in such a way that the dissipation is minimized for a particular sum of gradients. Thegradients would continue to be raised in this way until some cavities approached their maximum gradients.at which time the gradients in those particular cavities would no longer be raised.

If a cavity breaks down (i.e .. if a rapidly growing portion of the surface of the cavity becomes heatedabove the superconducting transition temperature of 9.2 K). the oreakdown can be detected in threeways: the ratio of power coupled out of the reference probe to power incident on the cavity will dropbelow its previous value; the power coupled out of the reference probe will itself suddenly drop: andthe rate of boiloff of liquid helium will suddenly rise. Under this circumstance. it is necessary to shutthe RF off on the offending cavity immediately. and to shut the beam off immediately. Unless the cavitywhich broke down now continues to do so at the same gradient.•he RF can be turned back on withina second. and the beam current can be ramped back up within another second. If the cavity continuesto break down. a different set of cavity gradients will h?ve to be selected. If a particular cavity is to betaken out of service. its frequency can be run several bandwidths off resonance using the mechanicaltuner.

TuningDuring operation. the cavities are tuned to within =200 of phase. which corresponds to 1.497.000

MHz =124 Hz. This is equivalent to a dimensional change range of =0.25 microns. Tuning is accom­plished by both coarse mechanical and fine electronic methods. Including the effects of the differentialball screw and the worm gear. the above range of the mechanical adjustr.1ent tolerance corresponds tohalf a tum of the external shaft. On the opposite sidt: of the yokes. a link whose length is changed bya piezoelectric stack provides fine adjustment. This capability is provided in the design. but may beunnecessary if adequate control can be obtained solely with the mechanical tuner.

The cavities are permanently tuned when warm to the high side of the frequency range. such that

80

, . . ~ .-- ; ... - . ~'

I

Machine Va...-uum Components

the tuning range is spanned exclusively by ca"ity compression. This avoids an insensitive dead bandaround zero. where the mechanisms may have backlash.

Q MeasurementA very important capability of the system is the ability to measure the Q vs. gradient for each cavity

from time to time. This capability is important to determine whether any general or localized degradationis taking place. to permit the gradients at\vhich various cavities are operated to be re-optimized. andto identify any cavities that have severely deteriorated in Q and therefore require servicing. This meas­urement will be accomplished by using the low-pressure helium shutoff valve in a throttling mode toregulate the pressure in each cryostat to a value slightly high~r lhan the standard 0.031 atmosphere; afterthe temperature in each cryostat has stabilized at the slightly higher value. changes in the position ofthe low-pressure shutoff valve will provide a measurement of the changes in helium boiloff rate withand without RF in a particular cavity. Heaters in each cryo-unit will facilitate calibration of the valvesfor the measurement of flow. During these measurements. the JT valves are closed.

In normal service. if the valves are left in a position where they produce a negligible pressure drop«< 1 torr). they will still provide good sensitivity to very high boiloff .ates. such as those associatedwith breakdown. and can thereby contribute to an effective system of interlocks.

Heat LossesThere are several sources of heat entering the cryostat; this heat must be removed by the refrigerator.Static heat leak is due to conduction or infrared transfer of heat to low temperature from room

temperature. This source of heat can be minimized by use of an intermediate temperature shield andby minimizing the thermal conduction through components connecting the helium vessel to the outsideworld. These losses are effectively temperature-independent.

Residual losses are due to an imperfect surface on the superconductor. These losses are essentiallytemperature-independent. but are proportional to the square of the accelerating gradient. The residuallosses may exhibit some additional dependence on gradient.

BCS losses are due to the presence of normal conducting electrons in the superconductor. Theselosses are a strong function of temperature. doubling typically every couple tenths ,:,f a degree. BCSlosses are also proportional to the square of the gradient. The choice of optimum operating temperaturedepends on a compromise be:ween increased BCS losses at higher temperatures and increased refrigeratorcapital and operating costs at lower temperatures. 2.0 K has been chosen as the operating temperaturefor CEBAF. Although this temperature is very slightly below the cost optimum. it will be closer to theoptimum if residual losses are reduced and operating gradients are raised above the design values. Suchimprovements in cavity performance are fully expected to accompany the further evolution of SRFtechnology.

Coupler losses are associated with the RF dissipated in the copper-plated stainless waveguide; partof this heat is conducted into the 2.0 K helium. These losses are proportional to the coupler couplingstrength and are proportional to the square of the gradient. Heat transfer into the 2.0 K helium isminimized by the 60 K heat intercept on the fundamental wavegUIde.

HOM losses. which are intentionally dissipated in the liquid helium because they are so small. areproportional to the square of the beam current.

The heat loads are discussed quantitatively in Chapter 9.

I I II

5. Beam Transport

5.1 Overview

The beam transport system focuses and steers .he electron beam as it travels from the injectorthrough the accelerator to the experimental areas. It is composed of five optically matched subsystems.The first is the linac lattice. a periodic structure that focuses the beam as it traverses each linac segment.The second subsystem. the recirculation arcs. consists of seven beam lines (four in one arc. three in theother). which transport the beamlets from one linac segment to the next for further acceleration. Thethird subsystem (discussed as part of Section 5.3) is a set of path-length adjustment modules ("doglegs")which allow precise control of the recirculator path lengths and hence provide for individual RF phasematching of all beamlets from linac to linac. The fourth subsystem. spreaders and recombiners. performsthe tasks of separating and recombining beamlets of various energies and providing optical matchingbetween the linacs and other beam optical subsystems. This subsystem also extracts the beam from theaccelerator. The fifth subsystem. the external beam switchyard. transports the extracted beam from theaccelerator to the experimental end stations. This chapter presents the details of each subsystem andindicates the manner in which they are integrated to make up the beam transport system of the accelerator.(The magnets for all these subsystems are described in an additional section. Section 5.6.)

5.2 Linac Lattice

The linac lattice provides focusing for the beam as it is accelerated. to ensure beam stability in thepresence of magneto-optical perturbations and collective effects. Each linac segment lattice is a periodicstructure of 12V2 FODO cells. in which a total of 200 cavities in 25 cryomodules are embedded (one 8­cavity cryomodule per half-eell). The length of each cell is 18.8 meters; the quadrupoles are in a FODOstructure of alternating gradient. Adjacent to each quadrupole are a beam position monitor for diagnosticpurposes and an orbit-bump dipole to correct for beam steering due to misalignments.

The quadrupoles in each cell are adjusted so that the beam will execute 1/3 betatron oscillation whiletraversing the cell on the first pass through the linac. Beam dynamics studies demonstrate that this first­pass phase advance of 120 degrees per cell will provide excellent beam stability; in particular, the BBU(beam breakup) thresholds are quite high (see Section 3.3). This choice of phase advance is readilyachieved through the use of quite modest quadrupole fields. as indicated in Table 5.1.

Table 5.1T)-pical Linac Lattice Quadrupole Fields

I

First-pass energy

0.050 GeV

1.050 GeV

Length

30 em

30 cm

Bola

1.03 kG/m

21.6 kG/m

83

..

CEBAF Design Repon

On the second and subsequent passes of a beam through the linac. the phase advance per cell isreduced. due to the higher beam energy. This situation potentially can create betatron mismatches ofthe beam to the linac lattice. resulting in large-amplitude betatron oscillations or "beating of the beamenvelope." However. the recombiners include quadrupoles that match the betatron functions to the linaclattice. After the first pass. these quadrupoles adjust the beam envelope for reinjection into the linacsto insure that proper matching of the beam to linac focusing occurs. A possible matching scenario. withonly minimal "beta-beat" on the second. third. and founh passes. is illustrated in Figure 5.1. The positionof the J in each plot denotes the end of the first linac segment and the beginning of the second linacsegment.

0.10

0.08

E Horizontal - - -0

CD 0.06Vertical

Q.0CD>

Ic: I

(CD 0.04 ~ I

E " '. ~I.as .,CD

•CD0.02 ~ I,

" I, II I I,

0.000 100 200 300 400 500

Longitudinal (m)

FJgUre 5.I(a). Beam envelope. linac firsl pass (0.05 10 1.05 GeV).

•

0.08Eo

~0.06oCD>c:CD 0.04EasCD

CD0.02

100

Horizontal - - ­

Vertical

I

400 500

F~ 5.1(b;. B.=am envelope. linac second pass (1.05 10 2.05 GeV).

84

~, • .- - • .. r ."- • • • ~ '. • • • I '. • ..

I

Beam Transpon

500400

Horizontal - - ­

Vertical

I

200 300Longitudinal (m)

100o.0 0 ~ -'-""'-- --'- """""'-'

o

(I)

g- 0.06(I)

>C

(I) 0.04ECO(I)

CO

0.02

_ 0.08Eo

Fil:Un 5.l(c). Beam envelope. Iinac third pass (2.05 to 3.05 GeV).

o.10 r"""T""""'""T"""T-r""""'""T"""""""""""""T""'''''--''''''''''""'T''""''T''""'''''''''''''''''''''-r-''''''''-'''-"",-,

500400

I

Horizontal - - ­

Vertical

200 300Longitudinal (m)

100o.0 0 .........",....................L...-.'I.........&.--'-..J-.&-.l.....l.-ZIo.............L...-.'..........................&-.l.....l...........J-........

a

~0.C6o(I)

>c:(I) 0. G11-­E:':I ,-(I)

en0.02

_ 0.08Eo

Fil:Urc 5.l(d). Beam envc:lope. linac founh pass (3.05 to 4.05 GeV).

As described above. the optimum focusing scheme is provided by adjusting the linac lattice quad­rupole fields to values suitable for the lowest-energy (first-pass) beam. Succeeding. t-..igher-energy (second-.third-. and fOl:rth-pass) beams thus have less. although adequate. focus. Linac misalig;tments and mis­alignment correction have been evaluated based on the lattice described above. A correction dipole nearand upstream of each quadrupole is used to correct for misalignments. These orbit-bump dipoles areoriented to move the beam horizontally for horizontally focusing quadrupoles and to move the beamvertically fo!" vertically focusing quadrupoles. A beam position monitor (BPM) near and downstream ofeach quadrupole is used to determine both the vertical and horizontal position of the beam.

The computer code TRA:'IOSPORT(I) was used to determine the possible beam env~lopesresulting from

85

r I I • r

•CEBAF Design Repon

I I •misalignments. Given the admittance of the linac. a maximum uncorrected transverse excursion of thebeam envelope of 1 em was chosen as the criterion for determining the necessary alignment tolerancesfor the quadrupoles and accelerating sections. Each element type (quadrupole or accelerating section)was simulated as being misaligned in each of the six possible degrees of freedom: horizontaIly (x),venically (y). longitudinally (z), rotationaIly about x (x'), rotationaIly about y (y'). rotationally about z(z'). The x and y' misalignments of the quadrupoles and the y' misalignment of the accelerating sectionseach contribute to the horizontal displacement of the beam centroid. The y and x' misalignments of thequadrupoles and the x' misalignment of the accelerating sections each contribute to the venical displace­ment of the beam centroid. Table 5.2 delineates the resulting alignment tolerances for the quadrupolesand accelerating sections assuming that each of the three contributions in each plane contributes equally.AIl tolerances are readily achievable.

Table 5,2Alignment Tolerances (TRA.~SPORT)

x y Z x' y' z·Element type (mm) (DUD) (mm) (mr) (mr) (mr)

Quadrupoles 0.17 0.17 20 2.2 2.2 35Accelerating

sections 0.42 0.42

Table 5.3Misalignment Correction Simulation

Deviation from Central Orbit (mm)Linac pass Xn- y....... x...... y- Xmia YaaiD

1Uncorrected 4.4 2.1 9.5 5.6 -12.0 -5.1Corrected 0.17 0.16 0.35 0.65 -0.72 -0.30

2Uncorrected 1.2 0.45 1.8 0.96 -2.7 -1.0Corrected 0.70 0.70 1.2 1.5 -1.5 -1.5

3iJncorrected 1.6 0.60 2.3 0.90 -3.0 -1.2Corrected 0.84 0.66 1.2 1.4 -1.4 -1.8

4Uncorrected 1.4 0.69 1.7 1.2 -1.5 -2.8Corrected 0.99 0.48 1.4 0.87 -0.95 -1.9

The computer code D1MAD(2) was used to simulate the correction of a misaligned linac. (Due tolimitations of the code, misalignment of the accelerating sections was not simulated. However. mis­alignment criteria for the accelerating sections from Table 5.2 show that this will not be a severe problem.)The quadrupoles and BPMs were randomly misaiigned assuming a truncated (six-sigma) gaussian dis­tribution with sigmas of 0.2 mm and 2.0 mr for the three positions and the three angles respectively.The corrector dipole values were found from a linear least-squares fit which minimized the BPM valuesand thus the beam deviation from the central orbit for the first pass. Second and higher passes weresimulated using the corrector values used to minimize the first-pass beam excursion due to the misalign­ments. Table 5.3 provides the results of this study for the corrected and the uncorrected misaligned linaclanice. The uncorrected lattice would allow the passage of all beams. The corrected lattice is withinspecified tolerances.

From Table 5.3. it may be seen that the beam deviations from the central orbit are greater for thesecond pass than for the first pass. and remain approximately the same for all higher passes. This isbecause the matrix for the accelerating sections between the quadrupoles is a function of the ratio oftne input and output energy. This function is nearly constant for the second and higher passes. but differs

86

I I

I

Beam Transpon

markedly from the first to the second pass. Therefore. the corrector values may be chosen to minimizethe beam centroid deviations for the first pass or the second through founh passes. but not both. Althoughthe corrector values which minimized the second-pass beam excursion from the central orbit were alsostudied. the sum of the rms excursions of all four passes was found to be approximately two times largerin this case. Therefore. the minimization of the first-pass beam excursion is the preferred solution.

5.3 Recirculation ArcsIntrOC:uction

The recirculation arcs transpon the beam from one linac segment to the next for funher acceleration.They were designed to meet the following criteria: minimal degradation of beam quality. reliance onmechanically simple magnetic components. and compatibility with a tunnel length that could accom­modate the potential upgrade of linac energy within the existing recirculator housing.

Arc Parameters and ConfigurationThe requirement of minimal beam degradation implies two types of recirculation arc latt;ce con­

straints. The first is that the lattice be achromatic. isochronous. and imaging. and in length equal to amultiple of the RF wavelength. These features will facilitate reinjection of the beam into the linac. andinsure that no dilution of the beam phase space due to betatron or longitudinal mismatch will occur.Achromaticity and isochronicity may be achieved by proper transpon and matching of the dispersionfunction through the arcs~ imaging may be achieved by a choice of optics in which the betatron phaseadvance leads to a whole- or half-integer cumber of betatron oscillations through each arc.

The second constraint :s that the arcs introduce minimal excitation due to radiation. The radiation­induced energy spread 0'E and emittance increase .:le occurring in a ISO-degree turn through a recirculationarc lattice are given by the following expressions (MKSA units are assumed):())

:! 1 IS? X 10-"" GeV:!-m:! L_'O'E= • - 'p

Here

with

and

5

.:le= 1.3270 x 1O-:!7 m:!-rad ~ (df)p

(df) (lIL) Jtis {(l/~) [T]:! + (~TJ' - ~WTJ):!])bend"

L orbit length in bending magnets

p = orbit radius in bending magnets.

(5.1)

..

As is clear from the above expressions. the induced energy spread is a function of the bending radiusonly. while the transverse emittance increase may be controlled. to some extent. by a proper choice oflattice function (dI).

The use of a split-linac acceleration scheme (in which 0.5 GeV is added to the beam energy by eachpass through a linac segment) will require a total of seven recirculation beam lines in two arcs. The"E(ast)" arc comprises four beam lines. which will initially operate at energies of 0.5. 1.5. 2.5. and 3.5GeV. while the "W(est)" arc comprises three beam lines operating at energies of 1.0. 2.0. and 3.0 GeV.We denote these beam lines as E1. E2. E3. E4. and WI. W2. W3. respectively. (See Figure 5.2.) Toset the parameters of these recirculation arcs. we observe that the final relative energy spread O'd£ andfinal emittance increase .1.e produced by seven traversals of beam lines bending through ISO degrees aregiven by the following expressions:

S7

I •

CEBAF Design Rc:pon

(5.2) L

FIgUre 5.2. Schematic machine layout showing recirculation arc beam lines. (Nominal design energies shown.)

•••

NOi"theastNorthwest spreaderrecombiner

-E(astl- arc-N( orthl" Iinac-W(estl- arc - E1: 0.5 GeV

W1: 1.0 GeV E2: 1.5 GeVW2: 2.0 GeV E3: 2.5 GeVW3: 3.0 GeV "S( outhl- linac E4: 3.5 GeV

4.0 GeV -To beam Southwest Southeastswitch yard spreader recombiner

In these expressions. "Yk is the value of "Y at the nominal operating energy of the kth beam line. Pkis the bending radius in the kth beam line. (.d+)k is the average of.d+ in the bends of the kth beam line.and E rinal and "Yrinal are the electron final energy and "y. respectively. The strong dependence of bothaE/E and.:lE upon energy suggests that the bending radii of the lower-energy beam lines may be reducedwith only minimal degradation of the final beam quality.

Parametric studies based on Equation (5.2) also indicate that the final fractional energy spread israther weakly dependent on the choice of bending radius for the first six beam lines. For example. wemay use Equation (5.2) to compute the relative energy spread at the final energy of 4 GeV for two typesof recirculation arc. As a first case. we consider an arc design in which the bend magnets operate atdifferent fields in each beam line. so that the bending radius is a consta:lt from beam line to beam line.In the second case. we consider an arc design in which the bending field is constant from beam line tobeam line. so that the bending radius varies linearly (with energy) from beam line to beam line. If P7 isthe bending radius of the final beam line in either case. the constant-bending-radius beam line inducesa fractional energy spread of 2.8 x 1O- 4/P7 at 4 GeV. while the constant-bending-field beam line inducesa fractional energy spread of 3.0 x 10 - 4/P7 at 4 GeV. It is therefore apparent that the last recirculationbeam line (at the highest energy) sets the magnitude of the momentum spread. while the lower-energybeam lines will change the final momentum spread by only a small factor.

The choice of bending radius for the final beam line is driven by these radiation excitation consid­erations and by the desirability of accommodating linac energy upgrades within the tunnel defined bythe recirculation arcs. Since an upgrade to 6 GeV shortly after machine commissioning will be practicalif SRF accelerating cavity technology continues to develop as expected. we define the bend radius byrequiring that the induced relative energy spread. 4 x (aE/E). be on the order of 1 x 10-

4at 6 GeV.

Use of Equation (5.2) indicates that this is readily possible if a final bending radius on the order of30 m is chosen. The requirements of minimal radiation excitation are thus easily met. Moreover. thebend packing fraction in an achroma"tic. isochronous lanice of the type re~uired here is only 50% orJess. Hence. a tunnel of mean radius 60 to 80 m will be required. Howevei. since lattices capable ofrecirculation at even higher energies in a tunnel of 80 m mean radius have been generated. the choiceof 30 m final bending radius is also consistent with accommoGating such an upgrade within the acceleratorhousing.

Choices for bending radii for lower-energy beam lines are constrained by a desire for operationalsimplicity. which suggests that all recirculator beam lines be designed in a similar. if not identical. fashion.

I I

Beam Transport

This is readily achieved if the beam lines are designed in a modular fashion. so that low-energy beam­line designs may be generated from the high-energy design by simply replacing bending magnets withdrift spaces. resetting beam-line focusing magnets to retain the desired optical properties. and adjustingdrift lengths to yield a total beam-line length that is a multiple of the RF wavelength. The bending radiifor the lower-energy beam lines are thus constrained to be fractions of the radius for the final beam line.with the fraction being the ratio of the number of bending magnets in the two beam lines. A convenientnumerology for all bending radii is given by noting that a bending radius of approximately 30 m isprovid~d by the use of 90 m of bending in the final beam line. This may be achieved through the use offorty dipoles of 2.25 m magnetic length in the final arc. It is then convenient to use 32 similar bends insome lower-energy arcs. and 16 similar bends in the lowest-energy arcs.

Parametric studies using Equation (5.2) indicate that the distribution of dipoles among beam linesthat is given in Table 5.4 produces acceptable energy spreads up to energies of at least 6 GeV. lhischoice also provides a degree of uniformity in dipole field from arc to arc (with a variation of only afactor of four in field between dipoles of lC'west field and of highest field) while reducing the total numberof dipoles required for the recirculator.

Table 5.4Dipoles per Beam Line Yielding

4 x (aEtE) == 1 x 10-4

at 6 GeV

Beam line

E1WIE2W2E3W3E4

Number of dipoles

16161616324040

Bending radius (m)

11.459211.459211.45C;211.459222.918328.647928.6479

All bends will require fields of less than 8.75 kG. even for final energies of 6 GeV: a typical valuefor a dipole field is 6 kG.

The configuration of recirculation arc beam lines is illustrated in Figure 5.2. In addition to thenominal design scenario with a 4.0-GeV final energy. we have also considered an upgrade scenario withinthe context of the present recirculation arc lattice. This scenario is an upgrade to 6 GeV final energy.in whi\~h the E arc bez.rn lines would run at 0.75.2.25.3.75. and 5.25 GeV while the W arc beam lineswould run at 1.5.3.0. and 4.5 GeV. Results for emittance and energy spread increase. together with atabulation of the required magnetic elements. are given in this section for both cases.

Recirculation Arc Beam-Line OpticsThe transverse emittance increase may. as noted above. be controlled by a proper choice of lattice

functions. Specifically. we seek to minimize the function (d-I) of Equation (5.1) insofar as is possibleconsistent with the other constraints. This is achieved. together with achromaticity. isochronicity. andimaging. through the use of a four-period struc.[ure in each beam line. Each period consists of a modifiedfour-cell FODO structure with "missing magneb.·· In suct"l a structure. the regular pat!ern of alternatingdipoles and quadrupoles is interrupted: dipoles .':"~ replaced by drift lengths to allow the off-momentumorbit to cross the design orbit and create an i!>oJchronous structure. The use of four periods per arcreduces the maximum and the mean values of the dispersion and betatron functions. while allowing areasonably high bend packing fraction. This leads to a decrease of (d-I). while keeping the arc length ata moderate value. Use of more than four periods per arc. although it would decrease (.;;...If.) still further.would reduce the packing fraction. while use of fewer than four periods would allow (d-I) to increase tounacceptable values.

Each of the seven arcs has been designed with the number of dipoles specified in Table 5.4. Asnoted above. this allows uniformity in bending field while reducing the [otal number of bends requiredin the seven beam lines. As all beam lines consist of four periods of "missing-magnet"' FODO structure.the focusing and bending structures are quite similar in all beam lines. This insures that all beam-line

89

..c .,~. . • I. ;;... -: -: ' ' : ~ '" , .'

•

CEBAF Destgn Repon

magnets lie in approximately the same location at the same path length. simplifying installation in thetunnel. To insure that ~ams of differing energy arrive at the linac reinjection points ... i the same phase.all beam lines have path lengths that are multiples of the 2o-cm RF wavelength.

Imaging is achieved by tuning each period to a phase advance that produces a unit transfer matrixfor the four-period beam line. We have chosen a tune of 5/4 ~tatron oscillations in each plane for eachperiod. This choice. in addition LO producing the desired imaging behavior. yields well-behaved lanicefunctions. Minimal variations in maximum and minimum values of betatron functions occur throughoutthe lanice. despite the strong modulation of the off-momentum orbit. This choice of phase advance alsoallows simple elimination of all linear chromatic aberrations through the use of Brown's sextupolecorrection schemes. (") The arc beam lines thus function as second-order achromats. (The choice of fourperiods per arc was also driven by this consideration.) Single periods for all beam lines have been matchedto the same betatron function values. 35 m in the horizontal plane and 3.5 m in the vertical plane. Thisallows each beam line to be treated as identical to all others insofar as matching to the linacs is concerned.Lattice specifications for each type of recirculation arc beam line are presented in Table 5.5. The structureof one period of each type of beam line. and the Ianice functions for each type. are shown in Figure5.3.

Table S.sLattice Specifications for Each Type of Recirculation Arc Beam Line

Low-energy Mid-energy High-energyProperty beam lines beam line beam lines

EI, WI, El, W2 E3 W3, E4

Beam-line length (m) 253.2 252.6 252.0

Orbit radius in dipole (m) 11.4592 22.9183 28.6479

Magnetic length of dipole (m) 2.25 2.25 2.25

Number of dipoles 16 32 40

Minimum focal length ofquadrupole (= Bp/B'I formax. B') (m) 4.56 4.30 4.36

Length of quadrupoles (m) 0.3 0.3 0.3

Number of quadrupoles 32 ~? 32"'-Horizontal phase advance 2'1r(5/4) '!'Ir(5/4) 2'1r(5/4)

Vertical phase advance 2'1r(5/4) 2'1r(5/4) 2'1r(5/4)

Horizontal beta (m):max/min/matched value 37.7/2.8/35 43.512.0/35 61.0/1.9135

Vertical beta (m):max/min/matched value 33.4/2.213.5 31.212.7/3.5 35.312.213.5(dI) (m) 0.114 0.167 0.269

Horizontal chromaticity(uncorrected) -7.62 -8.62 -9.32

Vertical chromaticity(uncorrected) -8.11 -7.76 -7.93

Peak sextupole strength(B"/Bp) (m - 3) 6.07 4.54 5.00

Sextupole length (m) 0.25 0.25 0.25

Number of sextupoles 8 8 8

90

•

75.00 t---o!J_60.00.§>-45.00

co.

.<30.00C!l.

15.00

·0

8.00

6.00 :§x

4.00 ~

2.00

Beam Tran"port

0.00 ., 0.000.00 7.03 14.06 2110 2S.13-35.1642204923 562663.30

Longitudinal distance (m)

F"1glU'e 5.3(a,. One period of IO""-energy beam line.

75.00 10.00

E 60.00 8.00

>-~co. 45.00 6.00x

>< 30.00 4.00 ~en.

15.00 2.00

0.00 0.000.00 7.01 14.03 2105 28.06-35.08 42.10 49.n 56.1363.15

Longitudinal distance (m)

F"1glU"e 5.3(b). One period of mid-energy line.

75.00 10.00

60.00 8.00E>-45.00 6.00 :§

co.. x

>< 30.00 4.00 ~co..

15.00 2.00

0.00 0.000.00 7.00 14.00 2100~.00~5:0042.0049.0056.oo 63.00

Longitudina1.Qistance 1m)

F"1glU"e 5.3(c). One period of high-energy line.

F"JgDn 5.3. Structure of. and lattice functions for. recirculation arc beam lines.

91

An unfonunat~ featun: of such strongly focusing. lattices (in which th~ off-momentum orbit ismodulat~d in ord~r to achi~v~ txlth isochr,mi~;ty and achromaticity) is a rath~r high level of chromaticaberration. Although not as g.r~at a problem in a sing.I~-pass d~,,·ic~ (such ;.:s an arc beam line) as in amulti-turn storage- ring. such chromatic abt:rrations can produc~ mismatch~-sof lattice functions for off­mom~nt~m panicles. which in turn can kad to a dilution ofth~ beam phase space. W~ have. consequently.introduct:d a chromatic corre-ction sche-me- for the- rccircuhtion arc beam-Iin~ lattices. It is based on the··second-ord~rachromat" conc~pt of Brown.''>' and provides a complet~ correction of all linear chromaticaberrations of th~ lattic~ and a suppression of geometric aberratil)ns. Spccifically considered was a two­family interleaved scxtupolc correction scheme in which a pair of scxtupoles were placed in each of thefour pcriods of each beam line. Since th~ tune pcr pcriod was rational (5/4 betatron oscillations) andsince the total tunc over each beam line was an integral number of wa'l.'e1engths (4 x 5/4 = 5 wavelengths).all linear chromatic aberrations canceled over the length of the beam line. while no geometric aberrationswere introduced (to first order in the scxtupole strength).

Use of the program DIMAD to study the t:ffect of this scheme on· the chromatic and geometricaberrations of the recirculation arc beam lines leads to the following conclusions: First. over the physicalmomentum acceptance of the beam line (i~p/plJi ~ 0.003 for the l-em-radius beam pipe. given a peakdispersion of approximately 3.5 m). the betatron functions in either plane vary by less than 0.1 %. Hence.the beam :ines will remain well matched for all momenta. Second. the tunes per beam line vary imper­ceptibly over the physical acceptance of the machine. This is not critical for a single-pass device. but isan assurance that the chromatic correction is very effective. Finally. and perhaps most importantly. thefrequent occurrence of correction WIth this interleaved scheme minimizes a geometric (i.e .• amplit:Jde­dependent) orbit distortion. that of increasing path length with amplitude. This is a critical point. in thatif the orbit of a particle lengthens with either amplitude or momentum. the particle will fall out of phasewith the acceleration field in the linacs. and a dilution of the longitudinal phase space will occur. (Forexample. this creates the need for an isochronous linear lanice.) We find that with the interleavedchromatic correction scheme. large-amplitude particles-ones at the physical aperture. with a largemomentum (~p/PIJ = =.0.002) offset-accumulate a path length increment equal to only a few percentof the bunch length. Hence. no dilution of the longitudinal phase space due to RF phase mismatch shouldoccur.

Figure 5.4 illustrates the dependence of linear lattice functions on moment'.lm. As described above.the chromatic correction removes all linear dependence of the lattice functions on momentum for each

0 Low-energy beam line. S.,0 Low-energy beam line, 8v

6 Mid-energy beam line. OV+ Mid-energy beam line, ov

X High-energy beam line,8v

0 High-energy beam line.8v

.0020 0 v

.0015

.0010

.0005

-.0005

-.0010

-.0015

-.0020

F'~ 5.4(a). Tune per arc as a function of momentum offset ::.pIp" [ov • v (::.pIp,,) - v(O».

92

Bc:.m Transpon

ty~ of beam lin~. Th~ principal r~maining chromatic aberrations ar~ clearly quadratic in nature. andcould be suppressed. if necessary. by use of a two-family OC'tupole correction scheme. Since the variationof lanice functions over th~ physical acceptance is not large. it is not anticipated that such a correctionwill be required.

0 Low-energy beam ~ne. 6.l3 x / ax0 Low-energy beam ~ne. 6 13 y 113 Y

.0100 6 p/J3Ii Mid-energy beam line. 6 Dx Ie x+ Mid-energy beam line. 6 Dy 113 Y

.0075 X High-energy beam line. 6 I3 x Ie x0 High-energy beam line. 6. l3 y 113 Y

.0050

.0025

6 pIp o.004

-.0025

-.0050

-.0075

-.0100

FJgDre S.4(b). Matched betatron functions as a function of momentum offsct ~/po [~~p • (~(~/po) - ~(O»/~(O».

Use of geometric analysis features of DlMAD indicates that the interleaved chromatic correction alsosuppresses sextupole-induced geometric aberrations. as expected. Specifically. tracing of a bundle of raysdistributed on ellipses from start to end of each beam line produces an undistoned image. under thebeam-line transfer map. of the initial ellipse to well beyond the physical apenure. This is illustrated byFigure 5.5. At six times the physical apenure. an octupole-induced aberration does begin to appear.This is driven by the same pairing of sextupoles that drives the quadratic dependence of tune on mo­mentum discussed above. Although it is almost cenainly unnecessary to do so (because of the largeamplitude at which the aberration becomes apparent). this aberration could be suppressed through theuse of octupole correctors.

The path-length aberration mentioned above is related to an additional issue in the recirculatordesign. Although each period is globally achromatic and isochronous. if an electron experiences a mo­mentum error within the period (by emission of a photon) it will accumulate both path-length andtransverse-positional errors. The magnitude of these errors will be governed by the maximum and averagevalues of the dispersion functions in the bending magnets. by the magnitude of the betatron functions.and by the spectral composition of the radiated photons (i.e .• the momentum errors). If the dispersionfunction is kept small (by keeping the number of periods large) and if the bend radius is maximized. wemay control the magnitude of this effect. Here four periods are used in an effon to minimize (d-I).Results from Monte Carlo simulations of this radiation-induced path-length error indicate that it will. inthese lanices. produce entirely negligible effects. These simulations are discussed in more detail below.

93

I

•

X 1m).08

+ Image of initial conditionswith momentum ollset

L1 p/po· ...002

Image of initial conditionswith momentum offset

~ p/po· -.002

• Initial conditions

o Image of initial conditions(on momentum)

.04 .p•

p

.p

.". ". ".p

.02

X' (mrad)• +J • .p.

+!

.. 4t ~.p ...... .

.0012

.0006

~ .".".-.0018

-.0012

.0018•.,.~ .p.,

~

••••••

~ -.04~

.~

-.08

F"J2W"e S.5Ca). Horizontal phase space during beam transport through high-energy recirculation beam line. Small ellipses corre·spond to a beam at approximately 150% of the physical aperture (indicated t-y solid ellipse): large ellipses correspond to a beamat approximately 6OO'n of the physical aperture. Octupolar distortion at ia:ge amplitude is evident.

.006 yl (mrad)

~ •.p• .p .p

.p• .p

.+JP.p.

.p •

-.024

...P • .p• .p

/~..p..p.-PoP.~• .p -.012• .jj

.jj

.of'

4'.p -.004

,p.,p ~ .p

.006 .012 ~ •~ .~ .

Q;l

.e:to

Y (m)

.024• Initial conditions

o Image of initial conditionsion momentum)

+ Image of initial conditionswith momentum offset~ pIp E ...002

o

Image of initial conditionswith momentum ollset

A pIp = -.002o

F"1gIU'e S.5(b). Vertical phase space during beam transport through high-energy recirculation beam line (same case as Figure5.5(a».

94

•

Bcam Tran~pon

Recirculation Arc Beam-Line ComponentsThe components for each of the three types of beam line are specified in Table 5.6. The orbit-bump

dipoles are specified by requiring that they cancel the integrated dipole fidd produced by a O.6-mmmisalignment of a tully excited quadrupole;: in the particular beam line under consideration. This cor­responds to a comple;:te correction of the error introduced by a 30' misalignment of any quadrupole(assuming that the survey and alignment of focusing elements is carried out to an rms accuracy of 0.2mm). Within any particular family of beam lines (low-. mid-. or high-energy). the only variation frombeam line to beam line (and from final energy to final ener~ry) is a change in peak field. as demandedby the varying beam rigidities in each case::.

Table 5.6Single Beam-Line Component List for Recirculation Arc Beam Lines

32328

32

Elementtype ~umber

Low-energy beam lines EI. WI. El. W2Rectangular dipoles 16Quadrupoles 32Sextupoles 8Orbit-correction dipoles 32

Beam position monitors 32

Mid-energy beam line E3Rectangular dipolesQuadrupolesSextupolesOrbit-eorrection dipoles

Beam position monitors 32

High-energy beam lines W3. E4Rectangular dipoles 40Quadrupoles 32Sextupoles 8Orbit-eorrection dipoles 32

Beam position monitors 32

Total Components for All Beam Lines

Element type

Rectangular dipoles

Quadrupoles

Sextupoles

Orbit-eorrection dipoles

Beam position monitors

Magneticlength (m)

2.250.30.25

2.250.30.25

2.25C.30.25

Max. pole tipfield (kG)

at 6 GeV final

8.740.91 (at 1.25 em)0.047 (at 1.25 em)0.0126 (8 x length.

kG-m)

5.461.20 (at 1.25 em)0.044 (at 1.25 em)0.0175 (8 x length.kG-m)

6.111.67 (at 1.25 em)0.068 (at 1.25 em)0.0241 (8 x length.kG-m)

Number

176

224

56

224

224

No. of independentpower supply

circuits

152

32

152

32

152

32

Number of independentpower supply circuits

7

35

14

224

I I

Table 5.7 provides specifications for peak fields in each beam line for each of two final energies.4.0 GeV and 6.0 GeV.

95

I

IS:] Start

~ End

Table 5.7~1agnetie Element Specifications for Recirculation Arc Beam Lines

PeakPeak Peak integrated

Bending Dipole quadrupole field sextupole field correctionBeam radius Energy field (at 1.25 em) (at 1.25 em) dipole fieldline (m) (GeV) (kG) (kG) (kG-m) (kG-m)

El 11.459 0.50 1.4555 0.1516 0.0079 0.002190.75 2.1832 0.2274 0.0119 0.00328

WI 11.459 1.00 2.9109 0.2939 0.0158 0.004381.50 4.3663 0.4549 0.0237 0.00657

E2 11.459 1.50 4.3663 0.4549 0.0237 0.006572.25 6.5495 0.68"..3 0.0356 0.00985

W" 11.459 2.00 5.8217 0.6065 0.0316 0.00843.00 8.7327 0.9097 0.0475 0.016

E3 22.918 2.50 3.6386 0.8031 0.0296 0.01163.75 5.4579 1.2047 0.0444 0.0175

W3 28.648 3.00 3.4931 0.9525 0.0391 0.01384.50 5.2396 1.4288 0.0586 0.0207

E4 28.648 3.50 4.0752 1.1113 0.0456 0.01615.25 6.1129 1.6669 0.0684 0.0241

Recirculation Are PerformanceTwo issues are critical to an assessment of the perfonnance of recirculation arc beam-line lattices.

First. the effects of radiative excitation must be established. Second. the sensitivity of each beam lineto magneto-optical perturbations must be detennined.

Effects of Radiative ExcitationThe perfonnance of the proposed recirculation arc lattices in the presence of radiative excitation

has been studied using both analytic tools and Monte Carlo numencal simulations. The results of analyticestimates based on equations (5.1) and (5.2) are displayed in Table 5.8 and Figure 5.6. The analytic

3.4..,.------------------------,3.2

32.8

"0 2.6III 2.4...E2.2- 2

:: 1.8o 1.6x 1.4

CD 1.2<J 1

0.80.60.40.2

O.......""'T'"--~-----'~'I..-.L...Jo.._+'_...:I- ......'"""t'~-""""'_+L..o<l'--'~~

1 2 345Beam line il:.Jmber

6 7

F"Jg1II"e 5.6(a). Cumulative induced ~ at start and end of each beam line (4-GeV final energy).

96

-

1.2 -r------------------------,1.1

0.9 rs:Jit)

~ 0.8 IZn)( 0.7w

0.6w

b 0.5

0.4

0.3

0.2

0.1

O .......-F:ZI.-.......o:::f"';.",a........L..:I""'"f';.",a........L..:I""'"f';.",a........L..:I""'"f';.",a........L..:I""'"f';....a.......,I,.,;'-f'....:I

1 2 3 4 5Beam line number

6 7

•

F"1glIn S.6(b). Cumulative induced 0'..1£ at start and end of each beam line (~GeV final energy).

estimates indicate that beam quality in the longitudinal and transverse dimensions is maintained to finalenergies of up to at least 6 GeV.

Table 5.8Analytic Estimates of Performance of Each Recirculation Arc Beam Line

Emittance Induced energyBending blowup spread

Beam (H) radius Energy ~E x 10.0 CJ'i: X 108

line (m) (m) (GeV) (m-rad) (GeV2

)

El 0.114 11.459 0.50 0.0000323 7.73 x 107

0.75 0.000243 0.0000132WI 0.114 11.459 1.00 0.00103 0.0000989

1.50 0.0078 0.00169E2 0.114 11.459 1.50 0.00785 0.00169

2.25 0.0591 0.0:89W2 0.114 11.459 2.00 0.0331 0.0127

3.00 0.249 0.2163E3 0.167 22.918 2.50 0.0369 0.0151

3.75 0.384 0.258W3 0.269 28.648 3.00 0.0948 0.0346

4.50 0.720 0.592E4 0.269 28.648 3.50 0.205 0.102

5.25 1.56 1.74

Final Total Total Final~E x 1010 ' 11 4 x (CJ'EtE)energy CJ'E x 10

(GeV) (m-rad) (GeV:!) x 10"

4 0.293 0.166 0.4126 2.23 2.83 1.14

97

I

I

CEBAF [)e.;ign R~"P"rt

Monte Carlo simulation of the single-panicle dynamics. including radiation. verifies the analyticestimates. Bundles of panicles ~'ere launched on the ~esign orbit of each beam line. and traced throughthe line. IUdiation effects were simulated at eacn bend using statistics for the number and energies ofquanta appropriate to the energy of the beam in the panicular beam line of interest. The results for thesimulation (the beam phase space at the end 'of the highest-energy beam line) are displayed in Figure5.7(a) and Figure 5.7(b). As is clear from the two-pan figure. the beam quality is maintained even inthis highest-energy line. Summarized results for both longitudinal and transverse excitations are presentedin Figure 5.8: these confirm the analytic estimates and demonstrate that beam quality is main!ained tofinal energies well beyond 6 GeV.

2

C/) beamc:as"0 1as 95% of beam:i. •

a>C) •c: 0as

•as-c:0.......0 -1J:

25(microns)

50 75

FJglU'e 5.7(a). Trans~ersc phase space at the end of the high-energy beam line. The 60% ellipse correspoDds to an iDducedeminance ~ of 1.9 x 10- 11 m·rad.

The Monte Carlo simulations also demonstrate that one possible source of error is actually negligiblein these lattices. As noted previously. the beam lines are isochronous from period to period. They are.however. not isochronous between arbitrary points. Thus. if an electron radiates a photon in the beamline. resulting in a momentum error. the electron can accumulate a path length. and hence RF phaseerror. through the remainder of the period in which the radiation occurs.

The high periodicity of the beam lines ensures that this effect is minimized. This is demonstratedby the Monte Carlo simulation. in which path-length errors of only microns accumulate. despite theradiation effects. (See Figure 5.7(b).)

Effects of Magneto-Optical PenurbationsThe effects of two types of lattice penurbations upon the beam-line opti~ have been studied. The

first type are errors that cause deviations of the design orbit from its nominal path through the centersof all magnetic elements. Principal contributors to such effects include quadrupole misalignments anddipole powering errors. Analytic estimates indicate that rms quadrupole m;"3.J.ignments of 0.2 mm willproduce design-orbit errors of less than 10 mm in either plane; hence the beam should remain withinthe physical apenure of even the uncorrected machine. This result is confirmed by numericai simulation.At the end of the beam line. we find that the nos penurOed design-orbit displacement (in an ensembleof 1000 machines with misaligned quadrupoles) is 4.2 mm horizontally and 1.3 mm vertically. Simulation

98

I

Beam Transport

2

o

of beam

(80% of beam

of 9 0% of beam/ 95% of beam

oo

oo

-1 0 1Longitudinal position (microns)

30

CO0

20xc::0- 10ce><I)

"0 0

>- 0c:n...<I)

c::<I)

<I) -10>-ce<I)

a:-20 0

-30-2

FIpI'e 5.7(b). Longitudinal phase space at the end of the ltigh-cnergy beam line.

2.5 5 7.5 10Electron energy (GeV)

12.5 15

FIpI'e 5.8(8). Eminance increase from final beam line as a function of energy.

99

I I

><

~oIt)

enoenocoo~

-0w-ow<J

2.5 5 7.5 10 12.5Electron energy (GeV)

15

FIgW'e S.8(b). Induced energy spread from final beam line as a function of energy_

with the program DIMAD indicates that the effect of quadrupole misalignments may be corrected throughthe use of a single orbit-bump dipole and an x-y position monitor at each quadrupole; the residual rmsorbit error after correction is on the order of 0.1 mm.

Dipole powering errors have less effect. As the lanice is achromatic. a systematic dipole poweringerror will not lead t-:> any orbit offset at the end of the beam line. Numerical simulation confirms this;a systematic dipole offset of 0.01 % produces an orbi~ offset on the order of 1 ~"Tl or less at the end ofthe beam line. Random dipole powering errors produce more significant effects; if the dipole fieldfluctuates from magnet to magnet with rms deviations of 0.01 %. analytic estimates indicate that thedesign orbit will be offset by an rms horizontal value of 1.5 nun in the lowest-energy beam line (the mostsensitive beam line. since it has only 16 dipoles). This result has been confirmed by numerical simulation.This offset is not readily corrected. as it is due to random variations in field from dipole to dipole. andthus may change in time (if. for example. the dipole fields ripple independently). The offset is. however.avoided by powering the dipoles in series; any powering errors are then systematic and will. as indicatedabove. have no significant effects.

Analytic calculations and numerical simulations with DIMAD indicate that neither systematic norrandom dipole powering errors generate path-length errors of more than a few tens of microns. Thiswill not produce any significant RF phase mismatch. Quadrupole misa!ignments can produce a path­length offset; simulations with DIMAD indicate. however. that this offset ::; on the order of at most a fewmillimeters for ;on uncorrected orbit. As this is a time-independent effect. it may be compensated throughthe use of a ··dogleg" path-length adjustment (described later in this section). Moreover. DIMAD sim­ulations indicate that the steady-state path-length offset for a corrected orbit is on the order of 0.1 mm.Hence. only minor path-length corrections will be necessary when orbit correction is employed; nosignificant path-length error or RF phase mismatch will occur.

The second type of errors investigated are those producing effects upon motion about the designorbit. These include quadrupole power supply errors and sextupole errors (both systematic and random)in dipole magnets.

We find that rms quadrupole power supply errors of 0.1 % will produce rms betatron functionmismatches of 0.4% of ~ and on the order of 0.004 in a (both horizontal and vertical). as well as rmstune variations on the order of 0.002. If these betatron mismatches are propagated through the machine.the: generate an increase in the emittance of the beam at final energy of less than 6%. The errors aretherefore negligible.

100

I

• •

Bc:am Transport

The sextupole order errors generate only higher-order (nonlinear) aberrations. If the systematicsextupole component of the bend produces a ~B/B of 10 - .. at 1 em horizontal displacement. the effectivecontributed chromaticity 'Will be on the order of 10% of the natural chromaticity of the beam line. Thisis readily comp.::nsated with the standard chromaticity-correetion sextupoles. Simulation with D1MAD

indicates that only 10% additional sextupole strength is required to compensate such a systematic sex­tupole in the bend. and that no significa'1t impact on the nonlinear optics is observed. Only slight variationsin the dependence of tune and lattice functions on particle momentum are generated when the additionalchromaticity is compensated.

As the nns random sextupole component should be an order of magnitude smaller. we anticipateno untoward effects whatsoever from the random sextupole. This has iJeen confinned through numericalsimulation with DIMAD.

Vacuum Considerations with Regard to Beam Loss and Beam QualityGiven a moderately good vacuum. beam quality will not be affected by residual gas scattering. Beam

loss due to residual gas scattering will be iess than 1 W for 800 kW beam output power. The basis ofthese statements is (following Helm. Lowe. and Panofsky). [he expression

tL = ~~:r[ 1 - ( 1 + :L )-I~ ( 1 + ~n ) - I~ ]

where -yeo = ZI/3/137. /L is the fraction of electrons scattered with

(X' )2 ("")2-- + -- ;;a: 1~.. a.

a""", = max (a". ay ). anun = min (ax. ay). dN/ds is the number of scatters per unit volume. and the othersymbols have their conventional meanings.

In evaluating this expression. we neglect the cryogenic section of the linac where the pressure is10- 10 torr.

By setting 8" = 0 + and ay = 0 + • we compute the total fraction of all electrons undergoing anyinteraction at all. The scattering probability is 1.9 x 10-

4for the pressures expected in the arcs. This

indicates immediately that beam quality will not be affected. even at pressures a few times more thanassumed. and that multiple scattering processes need not be taken into account.

To compute the fraction of the particles lost by the scattering. note that to prevent beam loss in thesuperconducting linac segments the collimators should remove particles with

(i = x, Y )2~""", ~i(S)J

2r,

where rf is the vacuum chamber radius at the focusing quads in the linac and~ is the beta functionthere. The results displayed in Table 5.9 are obtained assuming 200 !LA of beam after one week ofoperation. The table indicates that the total continuous beam loss in the arcs is less than 1 W.

Table 5.9Expected Beam Loss in Arc Collimators

Beam IiDe Fraction lost Power lost (W)

El 9.0 x 10-7 < 0.09WI 4.7 x 10-7 < 0.09E2 2.3 x 10-7 < 0.07W2 1.9 x 10-7 < 0.08E3 1.2 x 10- 7 < 0.06W3 8.3 x 10-8 < 0.05E4 6.5 x 10- 11 < 0.05

Total 2.0 x 10-6 <1

101

I

CEBAF Design Report

I I • I

Path-Length AdjustmentIn order to compensatt.: for path-length errors induced by sources such as quadrupole misalignments.

each recirculation arc beam line has associated with it a ··dogleg.·· or orbit path-length bump. This systemis a simple achromatic orbit bump used to adjust the recirculation arc path length of each beamletindependently and to insure that all beamlets are properly matched in RF phase to the linac segmentsat reinjection.

A dogleg resides at the end of each west-arc beam line and at the beginning of each east-arc beam­line. Each dogleg consists of four rectangular dipoles 2.25 m in length separated by 0.5 m drift spaces.A single dogleg is depicted in Figure 5.9. This configuration allows achromatic path-length adjustmentsin the hotizontal plane of over 20 mm total range: to achieve the maximum path-length differential.dipole fields of 1.67 kG per GeV of beam energy are required.

P =20 m

P =20 m

P =20 m

P =20 m

Fagan 5.9. Dogleg path-length adjustment. The solid curve represents the beam path at 6 GeV ir. a field of 10 kG. This path isover 20 mm longer than the trajectory in zero field (represented by the dashed line).

The complete orbit-path-Iength adjustment system for the machine thus comprises seven doglegs of10.5 m total length. containing 28 dipoles 2.25 m in length driven by 7 independent circuits. A total of30 m of length has been allocated. in order to allow for future expansion of the path-length adjustmentsystem in the event of an energy upgrade.

102

I

I

Beam Tl"3JlSl'On

5.4 Beam. SpreaderslRecombiners and Beam. Extraction

Optical systems are required to spread and recombine beams of different energies and to extractbeams from the linac system. Beams are spread for transport through the various beam lines of the tworecirculation arcs and ror extraction. They are recombined for reentry into the linac segments from therecirculation arcs and for delivery to the experimental area beam switchyard. (See figures 5.10 and 5.11.)

Linac tunnelcutaway

50-MeVinjector

1/3= 0.99995)

Extracted beamto end stations

0'"Linac '\.

10.5 GeV) ",,

separators

/~

// l .. )--- ~. /~

--'k~

Arc tunnelcross sectionldia. = 3.43 m)

fI

2.93m

~~~L1Linac tunnelcross section

FJgIII"e S.IO. Schematic of CEBAF .~rculatinglinac.

RF Magneticseparators septa E!

........~~~~§§~~§~~§§~==§i;_. To BSYSpreader Beam extraction

r====~~====;:E Recirculator

FJgIII"e S.II. Beam extraction region.

- Magneticsepta

Spreader

RFseparators Beam extraction=~BSY

Rec;,c~~.,

"103

•

•

CEBAF Design Repon

The beam leaves the injector region at a nominal energy of 50 MeV and is combined with threerecirculated. pre,,;ously accelerated. and therefore higher-energy beams. The four be.lms. including theinjected beam. are brought into a single line in space by a recombiner. Following this optical system.the four combined beams are each accelerated 0.5 GeV by passage through the first linac segment. Atthe end of this linac segment. the four beams are vertically separated by another optical system: aspreader. Once vertically separated. the beams are treated independently and are injected into an opticallyindependent recirculation lattice a.!> described in Section 5.3. At the end of the recirculation arc. the fourvertically separated beams are combined into a common line in space and injected into the second linacsegment. where they each receive an additional 0.5 GeV of acceleration. At the end of the second linacsegment. the four beams are again spread vertically in space. The highest-energy beam is not recirculatedbut is transported to the experimental area. The three lower-energy beams are transported through abeam extraction region where some of the available current from any two of the three lower-energybeams m~y be directed to the experimental area. with the remaining fraction transported to the recir­culation arc ....earn lines and accelerated to higher energies.

This section. then. wiU describe three optical systems:

1. the spreader. a system that takes four collinear beams of specific energy ratios and verticallyseparates them.

2. the recombiner. a system that takes four vertically separated beams of s~cific energy ratios andrecombines them into a collinear beam. and

3. the beam extraction system. which extracts current from anyone of the four available beams ofdifferent energy and directs that current to anyone of the three experimental areas.

Beam SpreadingfRecombiningAs described in Section 5.3. the recirculation arc beam lines are stacked vertically. The recirculation

lattices transport beams of different energies in separated beam lines with essentially identical projectionsonto the horizontal plane. Therefore. the operations of b~am spreading and recombining are executedin a vertical plane containing the linac axis. At the end of each linac segment. a vertical bending magnetdeflects the four different energy beams by different angles; the angular deflection of a specific beam isdependent upon the energy of that beam. This leads to a vertical separation of the beams a few metersdownstream. At this point. separate magnet systems act individually on each different beam. reestab­lishing the original beam direction (parallel to the linac axis) at the appropriate vertical height. Thebeams are then parallel. but move on orbits displaced vertically one from the other (figures 5.10 and5.11).

The strengths of the bending magnets. their precise locations. and the addition of focusing magnets(quadrupoles) create the desired vertical separation of the beams. provide optical matching from linacexit to recirculation arc entry. and ensure that the system is achromatic. The process need not beisochronous. because the nonisochronicity introduced is small and because it may be compensated forby the recirculation arc optics. Figure 5.12 provides a layout of a spreader system. Near the beginningof the system. the optical elements are closely spaced. Figure 5.13 is an enlargement of the area specifiedin Figure 5.12 showing that there is adequate space available.

The method of recombining the beams is simply a mirror image of the beam-splitting process. Thebeam optical considerations are. therefore. identical; and the recombiner beam lines are simply reflectionsof the beam spreaders. Figure 5.14 provides a layout of a recombiner system.

104

Beam Transpon

I

1.5

E 1.0

­...Q)

>0.5

15 20 25Longitudinal (m)

spreader

Figure 5.12. Layout of a spreader system. with venical axis exaggerated by a factor of 25. (Unlabeled magnets are quadrupoles.)

0.4

EO.3

ctIo

CD 0.2>

0.1

o 2.0 4.0 6.0 8.0Longitudinal (m)

10.0 12.0

Figure 5.13. Detailed view of a spreader layout (the southwest spreader). with vertical axis exaggerated by a factor of 25.

105

CEBAF Design Repon

1.5

­...\%)

>0.5 .....-+--ff-"'"""*--ilHII--it-+l.

o 5 10 15 20 25Longitudinal (m)

30 35

II

FJgUre 5.14. Layout of a recombiner system. with venicaJ axis exaggerated by a factor of25. (Unlabeled magnets are quadrupoles.)

Two spreader systems are required---one at the end of each linac segment. The first separates thebeams at the end of the first linac and injects them into the four recirculator lattices. The second fourfoldspliner separates the beams at the end of the second linac, matching three of them into the secondrecirculator system while transferring the fourth beam to the experimental area recombiner. Threerecombiner systems are necessary. Two are used to recombine the beams before reinjection into thelinac segments after recirculation; the third recombiner system is used before injection of the extractedbeams into the experimental area beam switchyard (BSY). Table 5.10 specifies the number. type, andparameters of the magnetic elements necessary for the five recombiners and spreaders.

Table S.IORecombinerlSpreader Parts List(For one system; five required)

Max pole tip Number ofLength Width Gap field (kG) @ independent

Element type Number (m) (em) (em) 6Gev circuits

Dipole 5 0.9 7.5 2.5 5.0 5

Dipole 1 1.8 7.5 2.5 2.4 1

Magnetic septum 1 1.8 2.5 2.5 2.4 1

Quadrupole 3 0.3 2.5 4.7 1

Quadrupole 3 0.3 2.5 4.0 1

Quadrupole 3 0.6 2.5 4.0 1

Quadrupole 24 0.3 2.5 4.0 24

106

I

I

Beam Transport

Beam ExtractionThe CEBAF beam extraction system can provide CW beams simultaneously to each of the three

experimental end stations. The extraction system provides three simultaneous beams as specified in Table5.11. The current in each beam can be adjusted continuously and independently over a dynamic rangeof 100 by adjusting the amount of charge in those microbunches contributing to each beam through aprocess described in the injector section of Chapter 6 (Section 6.1).

rfable 5.11Production of Multiple Simultaneous CW Beams

Design Capabilities

Fraction of highest beam energy Number of

0.25 0.50 0.75 1.0 1.0 1.0 simultaneousbeams*

x x x 3

x x x 3

x x x 3

x x x 3

x x x 3

x x x 3

x x x 3

"Dynamic variation of current in individual beams: factor of 100

The current is extracted at the microbunch level. It is important to note that the beam is in fact notexactly continuous. but rather consists of bunches of electrons (microbunches) which are approximatelya picosecond long and are separated by two thirds of a nanosecond. With an RF separator it is possibleto deflect alternate microbunches in alternate directions and thereby separate them. and hence. extractsome of the available current. Figure 5.15 displays the basic pr.nciple. The system depicted at the topof the figure is phased so that one of the microbunches receives a deflection opposite to the other two,and hence is capable of splitting the available current into two separate beams. one of which may beextracted while the other is reinjected into the recirculation arc beam lines to be accelerated to higherenergies. The system depicted at the bottom of the figure involves precisely the same concept exceptthat the phase of the microbunches relative to the RF has been adjusted to provide for the generationof three distinct beams. The extraction scheme employed is based upon these two systems.

The RF separators will operate at 2.5 GHz. This frequency has been selected because it is a multipleof 0.5 GHz. Lower-frequency devices have the disadvantage of larger physical diameters. There is notechnical difficulty inherent in this choice. Figure 5.16 shows a cross section of a 2.856-GHz RF separatorwhich has operated for several years with fields significantly greater than that specified for CEBAF.

The RF separator is used with a system of magnets as shown in Figure 5.17. The extracted current,depicted by microbunch one. is deflected around the main-line magnets and transported to the appropriatearea. That current not extracted, der-~cted by microbunches two and three, is brought back on linethrough the use of correcting elements and injected into the recirculation arc, the first element of whichis shown. All deflections in the figure are horizontal. There are three such systems, one for each of thethree lower-energy lines. one above the other.

Figure 5.18 provides a side view of the extraction system where the system displayed in Figure 5.17is represented by X-ed circles. That current which is extracted from the three lowest-energy beam linesis horizontally deflected to the left or right or remains undd'iected. This deflection is quite small andtherefore does not affect the aperture requirements of the recombiner system following. However, bythe end of the recombiner system the three beams to the three experimental areas have a beam-to-beamseparation cf approximately one em. and therefore may be handled with individual magnetic elements.The highest-energy beam may receive the same horizontal deflection. However, this deflection will bedue to an RF separator phased to split the beam into up to three different beams.

107

I

CEBAF Design Report

Frequency ~ 2.5 GHz

1 2 3 1 2 3

Frequency = 2.5 GHz

123 1 2 3

RFseparator

RFseparator

FJgUn 5.15. Deflection of microbunches by RF separator. generating two distinct beams in one case (top) or three distinct beamsin the other (bottom).

108

Beam Transport

14------- 134-----~..

1+---- 2b-122,5-----I welding coolingchannel

rJgW'e 5.16. The Karlsruhe-CERN superconducting S-band RF separator. a model for the CEBAF 2.5-GHz separator. (FromReference 5.)

RF

...o~

as--c:oN

Dipoles

?--~ 1sePtumDi~.... , ..~ J.....0.45 m

separator 1 ......e::=t'O '.----_:-~- -- ~ --=-=%-- ~ ~ ~-- ~

....,.....-------~- ~-... Correcting F~ To recomt?inerCorrecting d' I . It' and experimentalIPO e reclrcu a or " d -

septum dipole en stations

Klystron

..&....-------- -48_8 m ----------t~~

rlgUn 5.17. Transport of microbunches after RF separator (top view).

109

I

CEBAF Design Repon

• End station B

To o End station A

recirculation <:> End station Carcs

1 GeV

il ..:--~;_:;:_-~ :~_n~~~~'6' <:> ....... _,.....4 GeV -([)-__------ --~.::--_-_...~"T_O_B_S_Y~

• -------.---~ 0e , I....

Recombiner

DC dipole

High-energy splitterRF separator

Two-beam RF separator system

F"Jgan S.18. Side view of extraction system. (Circles with X's refer to system shown in Figure 5.17.)

Table 5.12 provides the type, number, and specifications of the magnetic and RF elements necessaryfor the extraction system.

Table 5.12Extraction System Parts List

(One required)

Magnetic Elements

Max pole tip Number ofLength Width Gap field @ independent

Element type Number (m) (em) (em) 6GeV circuits

Magnetic septum 1 1.5 2.5 2.5 0.9 1

Magnetic septum 1 1.5 2.5 2.5 1.8 1

Magnetic septum 1 3.0 2.5 2.5 1.4 1

Magnetic septum 1 0.3 2.5 2.5 0.25 1

Magnetic septum 1 0.3 2.5 2.5 0.50 1

Magnetic septum 1 0.3 2.5 2.5 0.75 1

Dipole 2 0.6 7.5 2.5 0.13 2

Dipole 2 0.6 7.5 2.5 0.25 2

Dipole 2 0.6 7.5 2.5 0.38 2

Dipole 1 2.2 7.5 2.5 1.0 1

110

•

ElemenJ: type

Dipole

Dipole

Dipole

Dipole

Dipole

Quadrupole

Quadrupole

Element type

RF separator

RF separator

RF separator

RF separator

Beam Transpon

Table 5.12 (continued)Extraction System Parts List

(One required)

Magnetic Elements

Max pole tip Number ofLength Width Gap field @ independent

Number (m) (em) (em) 6GeV circuits

1 2.2 7.5 2.5 2.0 1

1 2.2 7.5 2.5 3.0 1

3 3.0 7.5 2.5 1.2 1

3 3.0 7.5 2.5 2.4 1

3 3.0 7.5 2.5 3.6 1

6 0.3 2.5 1.0 3

24 0.3 2.5 4.0 24

RF Elements

Number ofLength Frequency independent

Number (m) (GHz) RF circuits

1 0.75 2.5 1

1 1.5 2.5 1

1 2.3 2.5 1

1 3.0 2.5 1

RF separators are all 5uperconducting.

5.5 Beam Switchyard

The experimental beam transport lines comprise the final beam optical system required. These beamlines are necessary to transport the extracted beam from the linac area to the three experimental endstations: end stations A, B, and C. End station A is centrally located; end station C lies to the north ofstation A; end station B lies to the south of station A. Each end station requires a separate beam line,and hence three separate transport systems.

As described in the previous section, the beam extraction system provides three beams that arehorizontally separated by approximately one centimeter between beams. The beam to end station A isnot deflected by the extraction system. The beams to end stations B and C are deflected in equal andopposite directions by the extraction system. Therefore, the beam path to end station C will lie approx­imately 1 em to the north of the beam path to end station A, and the beam path to end station B willlie approximately 1 em south of that to A.

The horizontal separations resulting from the deflections supplied by the extraction system aresufficient to allow placement of a pair of magnetic septa which increase the angular separation of thebeams to end stations B and C from the undeflected beam path to A. Following the septa, the beampaths to all three end stations will have achieved sufficient horizontal separation to allow the introductionof additional magnetic elements into each of the three beam paths to transport the beams to the individualend stations. These three sets of magnetic elements comprise the three experimental beam lines. Figure5.19 shows the first segment of the experimental area transport system.

111

CEBAF Design Repon

112 cm Magnetic septa

-S· achromatic system

To -A· achromatic system

To ·C· achromatic system

4 Recombiner

Fagan 5.19. Fir.;t segment of experimental area transpon (lOp view).

Figure 5.20 provides a layout of the total experimental area transport system. Because of the highbeam quality. the experimental beam-line designs require only that care be taken not to degrade thebeam phase space. No provision for momentum definition through the use of high power slits in adispersive region is necessary. Consequently. the only required major beam-line components are magneticelements.

End station S

End station A

End station C

Fagan 5.20. Layout of experimental area transpon system.

Once the beams have been separaled. the beam transport system is based optically upon a + Iachromat system. The + I achrornat concept is shown in Figure 5.21. The + I images the beam fromentrance to exit and allows chromatic correction exact to second order with two families of sextupoles.The number of + I units. the phase advance per cell in the + I units. and the total horizontal deflectionof each beam line are given in Table 5.13. A brief description of each.b--am line follows; a total componentcount is provided in Table 5.14.

112

Beam Transpon

r 1celli

06DO~DO~DO~D~-------------... IoFocusing quadrupoleoDefocusing quadrupole

~ Dipole

rlglU'e 5..21. The ... 1 achromat concept.

Table 5.13BSY Optics

End station

A

B

C

# of cellsper I

6

4

6

60

90

60

Horizontalbend per

cell(j

-3.33

+3.75

-4.73

#ofI

units

2

1

2

Totalhorizontaldeflection·

e)-40.0

+18.2

-60.0

"Relative to Iinac centerline

Table 5.14Beam Switchyard Magnetic Elements Parts List

LengthElement type Number (m)

To End Station ADipole 12 3.0Quadrupole 24 0.3Sextupole 24 0.25Quadrupole 8 0.3

To End Station BMagnetic septum 1 3.0Magnetic septum 1 4.6Dipole 1 2.2Dipole 4 3.0Quadrupole 2 0.3Quadrupole 1 0.6

Width(em)

7.5

2.52.57.57.5

Max pole tip Number ofGap field (kG) @ independent(em) 6GeV circuits

2.5 3.8 12.5 2.0 22.5 0.24 22.5 4.0 8

2.5 0.3 12.5 1.1 12.5 2.5 12.5 4.3 12.5 3.0 22.5 3.0 1

(Continued on next page)

I

113

•

CEBAF Design Repon

Table 5.14 (continued)Beam Switchyard Magnetic Elements Parts List

Max pole tip Number ofLength Width Gap field (kG) @ indepeadent

Element type Number (m) (em) (em) 6GeV circuits

Quadrupole 8 0.3 - 2.5 3.2 2Quadrupole 4 0.3 2.5 4.0 4Sextupole 8 0.25 2.5 0.12 2

To End Station CMagnetic septum 1 3.0 2.5 2.5 0.3 1Magnetic septum 1 4.6 2.5 2.5 1.1 1Dipole 1 2.2 7.5 2.5 2.5 1Dipole 12 3.0 7.5 2.5 5.4 1Quadrupole 2 0.3 2.5 3.0 1Quadrupole 1 0.6 2.5 3.0 1Quadrupole 24 0.3 2.5 2.5 2Quadrupole 4 0.3 2.5 4.0 4Sextupole 24 0.25 2.5 0.12 2

Transport Line to End Station A:As the beam path to end station A does not receive a horizontal deflection from the extraction

system. the transport line to end station A is the simplest of the three experimental beam lines. Down­stream of the extraction system. beams to end station A travel directly into a system of eight quadrupoles.These are used to adjust the beam envelope at the target and maintain the beam envelope through adrift region. Radiation safety requires that the beam to end station A be deflected from the line extendingalong the axis of the second linac. Therefore. following the four-quadrupole beam-matching system. thebeam is transported through a 4O-degree bend by a second order achromatic transport line. This lineconsists of 12 chromaticity-corrected FODO cells tuned to 60 degrees betatron phase advance per cell.The 4O-degree angle of horizontal bend is sufficient to insure that no beam can reach the end stationunless the transport line magnets are energized. Following the bend. the beam is transported into endstation A.

Transport Line to End Station B:Beams directed to end station B receive a transverse deflection during extraction. Downstream of

the extraction system. this deflection results in a horizontal displacement of the beam by approximately1 em to the south of the undeflected trajectory (which goes to end station A). At this point, a pair ofmagnetic septa are introduced, creating an additional I.fKtegree separation of the beam away from theLfideflected beam path. This bend is rendered achromatic through the use of a quadrupole triplet and asingle dipole magnet that produces another 1.6 degrees of angular separation. (See Figure 5.19.) Thebeam is then transported through a four-quadrupole system, which is used to adjust the beam envelopeat the target. Final transport to end station B is accomplished by bending the beam through 15 degreesand delivering it to the experimental hall. The 15-degree bend and transport to the experimental stationare provided through the use of a second-order achromatic beam line made up of a single + I achromat.

Transport Line to End Station CThe transport line to end station C is similar in design to the line employed for end station B. Beam

directed to end station C is horizontally deflee;ted during extraction; downstream of the extraction systemthis deflection results in a I-cm northerly displacement of the beam from the undeflected beam pathleading to end station A. At this point. -the angular separation between the beam to station C and the(undeflected) beam path to station A is increased by use of the same achromatic septum/quadrupoletripletldipole system as was used in the beam transport line for end station B. (See Figure 5.19.) (Notethat in this case. the bending is toward the north; in the previous case. the bending was to the south.)Following this final separation of the beams. a four-quadrupole matching system is used to adjust thebeam envelope at the target position. The beam is then delivered to the target position in end station

114

I

Beam Transpon

C through the use of a second-order achromatic transport line. This line. which consists of two + Iachromats. bends the beam through a total northerly angle of 60 degrees while transporting it to thetarget location within the end station.

5.6 Beam Transport Magnets

Several types of magnets are used for the CEBAF beam transport system. The larger main dipoleand quadrupole magnets used throughout the transport system are made of laminated iron cores withwater-cooled epoxy-impregnated coils. and with integral vacuum beam tubes. Smaller magnets such asIinac quadrupoles. linac steering dipoles and beam-line sextupo:es are constructed with split-ring casesand bolted pole tips with wrap~d and impregnated air-cooled coils. Water-cooled DC septa are usedin the spreaders and recombiners and also in conjunction with the RF separators to e1ttract the threedifferent electron beams and direct them to the appropriate end stations.

Bending Dipole MagnetThe cross section of the bending dipole magnet is illustrated in Figure 5.22. The same la..nination

shape will be used for each of the four different lengths of dipole magnets (0.61,0.91.2.25, and 3.05m). The dipole has a 25.4-mm gap and a 96.5-mm pole face. The magnetic code POISSON was used tocalculate :he profile shown in Figure 5.22. This profile produces a good field region, !:J.BIB ... 10-4, overa region of 50 mm about the central region of the pole. This permits any of the different lengths ofdipole magnets to be built without sagitta and still maintain a good field region of 8-mm radius alongthe electron beam path through the dipole magnet. The cores are made from 1.57-mm, 1004-1008 steellaminations, supported by steel angle plates. End plates (38 mm) are machined to match the pole faceand give clearance for the coil turnaround. The coil packages are made from hollow, 8.7-mm squarecopper conductor coils. packaged in four two-layered pancakes connected in series. Turn-to-turn insu­lation and ground wrap of fiberglass impregnated with epoxy is cured in the magnet for maximum strengthand coil clamping.

284 mm ..,

228 mm

V "\r96.5 mm-,

I ~r -+ 1- l J

25.4 ~mj

~ ~ ~

rJg1ln 5..22. Bending dipole magnet aoss section.

115

I

CEBAF Design Repon

The bending dipoles are operated at a number of field strengths. as outlined in Tables 5.7. 5.10.5.12. and 5.14. The highest field required is 0.87 T (for 6-GeV operation) which uses an input currentof 440 amps at 29 volts for a 2.25-m dipole. This corresponds to a current deru;ity of 7.7 AJmm2

• Mostof the dipoles operate at current densities of less than 5 AJmm2 for 6-GeV operation. Weights of thedipoles are as follows:

~(m)0.610.912.253.05

weight

(kg)265397970

1323

Quadrupole MagnetsTwo types of quadrupole magnets. water-cooled and air-cooled. are required. The water-cooled

quadrupole has a laminated core and a hollow copper coil with a 25.4-mm bore. The cross section isshown in Figure 5.23. The core is made with 1.57-mm. 1004-1008 steel laminations. stamped and stackedas quarter cores and then welded. The ma6J1et and coils are assembled and the tie plate welded togetherwith the vacuum tube in place. Once assembled. the unit is epoxy-impregnated. The quadrupole is madein three lengths: 30.5. 40.6. and 61 mm.

r~-----228.9 mm------~·I

228.6 mm

FJgUn 5.23. Quadrupole magnet cross section.

116

I

Beam Transpon

The maximum field strength for the quadrupole magnets (for 6-GeV operation) is 0.4 T. This requires260 amps at 35 volts with a current density of about 3 Almm2

• Most of the quadrupoles operate at lowerfields than this with correspondingly lower curr~nt densities. The weights are as follows:

~(mm)305406610

weight

(kg)

118157236

The air-cooled quadrupole has a split ring and bolted pole tip with a 38.1-mm bore and 305-mmlength. The split case permits the magnet to be installed on the beam tube without breaking vacuum.This magnet is used as the linac quadrupole. with one placed before each cryogenic module. It is designedto have a maximum pole tip field of 0.1 T. and to weigh 90 kg.

Sextupole MagnetThe sextupole magnet is used in the recirculation arcs and the beam switchyard. It has a split case

so that it can be installed on the beam pipe without breaking vacuum. The coils are air-cooled. woundon the pole tip. epoxy-impregn~ted.and the pole tip is bolted into the case. The maximu~ pole tip field;" 0.1 T and requires 0.9 Almm- for this operation. Figure 5.24 shows the cross section of the sextupole.it is 250 mm long with a 25-mm bore.

88.9 mm 00 tUbing

.....--76.2 mm 10---.1

Fi~ 5.24. Sextupole magnet cross section.

Steering Dipole MagnetsAn air-cooled. solid iron. lSO-mm-long magnet with a bolted core is used throughout the beam

transport system as the steering dipole magnet. It is made in two gap sizes. The magnet used in therecirculator system and beam switchyard has a 25-mm gap. and the magnet used in the linac has a 38­mm gap. To keep costs down. both will have the same coil. pole tip. and back leg plate. with 12.5-mmlarger side plates for the 38-mm gap. The fields have been calculated with POISSON. and the larger gaphas a good field region of lO-mm radius with t:.B/B ~ 2.0 x 1O-~ at the maximum field 0.04 T.

117

CEBAF Design Report

Septum MagnetMagnetic septa are magnetic elements that produce a strong dipole field on one side of a thin

magnetic shield. and a nearly field-free region on the other side. They are used in the beam extractionsystem. the recombiner/spreader systems. and the beam switchyard to increase the separation betweenthe electron beams. Tables 5.10. 5.12. and 5.14 show the number of septa required in these regions.There are several different types of septa used for different purposes in these regions with differentlengths and strengths. Figure 5.25 shows t.l:te cross section of one of these magnets.

Coil

IField region 101 mm

24 mmI

Field-free region

1.25 mm septum

f------102 mm ----~...

FIgW'e 5.25. Cross section of a water-cooled septum magnet.

Continuous-duty-current septa were chosen for this application because the bend in the horizontalplane is compatible with the rest of the transport system. Current septum design represents a compromisein field strength, septum thickness, and power density in the current-carrying septum. The septum shownin Figure 5.25 represents a reasonable choice of these parameters. The copper septum is made fromthree layers of 5.1-mm-thick by 7.1-mm-high water-cooled conductor. A field of 0.15 T is produced inthe central region with 960 A in each conductor, at a current density of 42 A/mrn2

• The length of theseptum, 1 m, is determined by the limited flow rates in the small water-cooling channel. The three coilsare cooled in parallel, with a flow rate of 0.014 liter/sec at a velocity of 4.9 m/sec. The temperature risein each coil is about 31°C. The total power in a I-m septum is about 5500 watts.

The long septum magnet required in several regions will be made by pairing two septum magnets:one such as that illustrated in Figure 5.25, followed by another similar to this but with a copper septumthat is twice as thick, and with twice the field. The first septum produces sufficient separation of thebeams to permit the use of a wider and stronger septum in the second stage. These will be mounted ina common vacuum vessel.

Figure 5.26 shows the results of POISSON calculations of the fields produced by the septum shownin Figure 5.27. The upper half shows the field lines produced by POISSON; the lower half shows the fieldstrength. The left-hand ordinate scale refers to fields on that side of the thin (1.25 mm) iron septum.and the right-hand ordinate scale refers to that side of the iron septum. Note the large suppressed zeroon the right-hand scale.

118

Beam Transpon

F"agure 5.26(a). POISSON calculations of the field produced by the septum magnel.

1550

§ 1.0OlQ)

"0Q)

·7~00.5

....J

-2 o 2Magnet length (em)

4

o1500 "0

Q)

E:J-a.Q)

en1450

6

F"agure 5.26(b). Magnetic fields on either side of the septum.

119

CEBAF Design Rcpon

Two septa are required that will permit three beams to be further divided. These win be similar tothe magnet shown in Figure 5.25, with a mirror image reflected to the left of the centerline. A series410 stainless steel beam tube rather than the iron septum will be used in the center to produce a field­free region. Both the left and right coils will be independently controllable to handle different beamenergies.

References

1. K. L. Brown, et al.. "TRANSPORT: A Computer Program for Designing Charged Particle BeamTransport Systems," CERN-8~ (March 18, 1980).

2. R. v. Servranckx, et al.• "User's Guide to the Program DIMAD," SLAC Report 285 uc-28(a), (May1985).

3. Sands, M., SLAC-121 (Nov. 1970).4. Brown, K. L., I.E.E. Trans. Nuc. Sci. NS-26, 3 (1979).5. Nuclear Instruments and Methods 164 (1979).

120

6. RF Power System

6.1 InjectorBackground of the CEBAF Design

Any of several existing high-quality injectors for CW electron accelerators could serve as a modelfor the CEBAF accelerator. The injectors used at the superconducting linear accelerator at the HighEnergy Physics Laboratory (HEPL) at Stanford and at the room-temperature microtrons at the NationalBureau of Standards (NBS) and the Nuclear Physics Laboratory (NPL) of Illinois have many commonfeatures that will be used in the CEBAF design. The lOO-kV section of the NPL injector (which. in tum.was modeled on the NBS injector) is the closest in design to the CEBAF injector and serves as a goodmodel.

Overview of DesignThere are four different sections of the injector that have to be considered in the design: a lOO-keV

line to produce. collimate. and bunch the electrons; a region to capture the electrons and acceleratethem to nearly the velocity of light: a preaccelerator; and an acceleration section to increase the energyto a nominal 50 MeV before injection into the rest of the linac. The main features of the injector areoutlined schematically in figures 6.1 and 6.2. The lao-keV section consists of an ele=tron gun. a com­pensated RF subharrnonic chopper. and a buncher. The other elements of this line are solenoidal lenses.steering coils. and apertures to control the transport of the low-energy electron beam into the firstaccelerating section. The RF elements of the injector up to and including the capture section are room­temperature components operating at 1500 MHz. or subharmonics for the RF chopper. The lOO-keVline produces an electron beam with a velocity of 0.55 c. bunched into la-degree bunches for injectioninto the capture section.

A-1 A-2

Beam profilemonitor

cavity

Figure 6.1. Schematic view of the lOO-keV section of the CEBAF injector. This section produces a chopped and bunched beamfor further acceleration in the rest of the injector.

121

Superconducting RF

CEBAF Design Report

Room temperature I

I,100-Kev line~ .~

Capture' \.. 1 \. ) \.. ). I -....".-- """-----~-----.... ....------.....r--------"

section I 2 Cavities 8 Cavities 8 CavitiesI

Fagure 6.2. The CEBAF injector.

The capture section increases the velocity of the electrons to nearly the velocity of light and furtherreduces the width of the bunch. Experience at the Darmstadt superconducting linac has shown thatinjection from a l00-keV line into a ~= 1 superconducting capture section did not work well. Therefore.a room-temperature capture section will be used to increase the energy of the electrons before injectioninto a cryogenic preaccelerator. Two sections of room-temperature accelerator will be used. graded inbeta. and each powered by 5 kW of RF. This would produce an electron beam energy of about 950 KeV(~=O.937) that will be injected into two standard CEBAF-Comell cavities used as a preaccelerator ina dedicated cryost:lt. A small analyzing magnet will be used after the first injector cryostat to deflect theelectron beam into a diagnostic region for beam tests and to set up the injector before sending the beaminto the rest of the linac. A lens is used after the injector cryostat to refocus the electron beam into thenext section of the injector line. Two cryogenic modules are used as an injector accelerator. as shownin Figure 6.2. These will increase the energy of the electron beam to a nominal 50 MeV before it isinjected into the main part of the linac.

Electron GunThe electron gun produces a high-quality DC electron beam at an energy of 100 keY. The basic

details of the gun are shown in Figure 6.3. The gun has a Pierce geometry with a 1-mm-diameter dispenser

Control electrode

Electron beam

Anode

Focusing

Cathode

FIgW'e 6.3. Details of the tOO-keY electron gun for the CEBAF injector.

122

I

RF Power System

cathode that will produce 2 rnA of electron beam current at a reasonable current density. A modulatingnonintercepting electrode is used to add fast pulsing capabilities to the gun; this is a useful feature forsetup. other beam diagnostic experiments, and some experimental programs. The eminance of a Piercegeometry electron gun at this low perveance and small cathode diameter should be about 1 mm-mradat 100 kYo The cathode and anode radii are designed for a 2: 1 ratio. The beam is focused by a lens topass through aperture A-I (Figure 6.1), a I-mm-diameter aperture built as shown in Figure 6.4. This is

Water-cooling

Electron beam-

dipole

FJglIl'e 6.4. Water-cooled apertures for the lOO-keV injector line. The shallow impact angle that the electron beam makes withthe aperture surface bas been found to reduce cathode poisoning.

insulated from the beam line using the Kapton insulated conflats developed at Chalk. River NuclearLaboratories, and water-cooled with an external cooling coil as shown. These apertures are simple toconstruct and were found to produce very little poisoning of the dispenser cathode when used on thehigher-current ETA injector at Chalk River. The second aperture, A-2, is 4 rom in diameter, and located0.5 m downstream. This combination of apertures limits the eminance to about 1 mm-mrad. Althoughthe electron gun is capable of about the same eminance, the apertures actually define the transverseeminance of the injector. A simple air-eore dipole is used around A-I to provide an in-line beam dumpfor conditioning the gun cathode.

RF ChopperThe beam is then focused through the RF chopper to :;:.1 image at the chopper slit, A-3. The CEBAF

injector will use the NPL injector's technique to chop the beam at subhannonic frequencies in order toload different bunches at different levels. The RF chopper is a rectangular cavity driven at 1rJ2 in thevertical direction and 10/3 in the horizontal direction. Figure 6.5 shows the details of the chopper cavity.The RF inputs shown on the figure will produce the scanning panern shown in Figure 6.6 at the chopperaperture. A-3. Variable slits at the chopping aperture can be used to reduce the bunch length from 60degrees to 0 degrees for each of the three succeeding bunches. These differentially loaded bunches areseparated at the end of the accelerator to produce up to three different beams of different currents thatcan be delivered to the end stations. A compensating lens similar to that used at the NPL injector willbe used to avoid rotation of the scanning panern. The second chopper is used to compensate for theradial momentum introduced by the first chopper, and completes the chopping system. Tests at the NPLinjector have shown that the beam current transmined by this system can be controlled over a range ofgreater than 100:1.

123

• I

CEBAF Design Repon

(r:--RF in

(fO

= 750 MHz)

TE201

mode

~TE102mode

RF in (fO

= 500 MHz)

......----T~/-i---------r./

II

III

0.3997 m

0.5996

pickup

Figure 6.5. Details of the rectangular subharmonic chopper cavity.

y (em)

0.5

o

-0.5

""'---.......------'----'----::~ X (em)-0.5 o 0.5

Figure 6.6. The beam pattern pro<!uced by the chopping cavity at the aperture A-3.

BuncherThe electron beam is then transmitted into a bunching cavity that uses the same principle as a

klystron buncher to further bunch the beam from the 60 degrees to 10 degrees before it enters the capturesection. The buncher is a TMuIll room-temperature cavity at the same frequency (1500 MHz) as thefundamental of the accelerator. The electrons are accelerated or decelerated depending on their phasewith respect to the buncher RF field. After a drift distance. the higher-energy electrons reach the lower­energy electrons. producing a compression of the longitudinal phase of the beam.

124

I

RF Power System

Capture SectionThe bunched beam is then injected into an RF cavity that acts as a capture section. Ideally. this

section should have a ~ < 1 and be graded in beta. Experience at the Dannstadt superconducting electronlinac has shown that injection from the l00-keV line into a ~= 1 superconducting capture section doesnot work well. Therefore a room-temperature capture section will be utilized. In order to use the samelow-power klystrons and RF control systems that are used for the rest of the accelerator. short. five-cellstanding wave accelerator cavities similar in design to the NBS capture section will be used. The injectorhas been modeled with .he c(·ce PARMELA. and it was found that two 5-cell room-temperature cavities.graded in beta. produced a good m2~ch into the first superconducting cavity. The beam energy is 0.95MeV (~ = 0.94) after the two 5-cell cavities. with an energy and phase spread of 16 keV and 1.7 degrees.The beam is then focused with a solenoid and injected into a cryogenic preaccelerator.

PreacceleratorThe preaccelerator uses two 5-cell CEBAF-Comell cavities in a dedicated cryostat. The beam energy

is 5.8 MeV (~ = 0.997) with an energy and phase spread ("If 16 keV and 1.2 degrees after the preaccelerator.A bending magnet will be used after the preaccelerator to bend the beam into a diagnostic region forbeam tests and tune-up before injection into the rest of the linac. A lens will be used after the preac­celerator to refocus the beam before injection into the injector accelerator. Figure 6.7 shows the results

0.50

Buncller

0.25 IE0

0x

-0.25t t t t t tt t t t t

L1 L2 L3 L4 LS L6 capture L7 PreacceleratorA1 A2 A3 section

-0.500 100 200 300 400 500 600 800

Distance from gun anode (em)

40

20

ClCI)

"0 0C%l

-20

I I I I I I I

-

hllil III " ,_III "' I.... -

-

I I I I I I

100 200 300 400 500 600 700 800

Distance from gun anode (em)

Figure 6.7. Calculated transverse beam profile and longitudinal phase of the beam as it passes through the injector.

125

CEBAF Design Repon

o-2 2-1 0 1

6 4> (deg)

_..'{.

~\

r ::'..IIII

-10

-20-2

w 0<l

10>CD~

20

2-1 0 16 e::p (deg)

! ~.,,lr hr

o

'It: 24

­...~ 48

96

C/) 72~o

6015 30 45:# of particles

20 ....-----------.16

12 r---~=t==> 8.: 4

w 0<J -4

-8-12

L-.z-------'ioil-16-20 ... ...

o

0.125 i--~--_+--+_-_.

0.250 ----------_

-0.125 ....-~--_+--+_-__I

-0.250 ...._-"__..1-0_..... ...

-0.250 -0.125 0 0.125 0.250X (em)

Eo

>-

5.0

2.5

...E>- 0

a>

-2.5

/i./

5.0

2.5

...E

oxa>

-2.5

V./

-5.0-0.250 -0.125 0 0.125 0.250

Y (em)

-5.0-0.250 -0.125 0 0.125

X (em)

0.250

Figure 6.8. Calculations of the phase and energy spectrum of the electron beam at the end of the preaccelerator.

of PARMELA calculations performed at Los Alamos for the I()()-keV line and the capture and preaccel­erator sections. The beam current was 200 IJ.A and calculations were performed both with and withoutspace charge. which produced only a small effect. Figure 6.7 shows the transverse dimensions of thebeam. and also shows the phase. as a function of distance from the gun anode. The position ·of the

126

RF Power System

elements is shown on Figure 6.7(a). A 6O-degree beam was started from the gun to simulate the chopperto avoid starting a large number of p;.Lrticles that would be lost at the chopper. Figure 6.8 shows thecalculated phase and energy spectra at the exit of the preaccelerator. A bunch of less than 1.5 degreesand 26 keV has been prepared for injection into the rest of the accelerator.

Injector AcceleratorThe rest of the injector will consist of two cryogenic modules to increase the energy of the electron

beam up to about 50 MeV before injection into the main accelerator. These will be identical to thecryogenic modules used in the rest of the linac. The injector line will be built in-line with the rest of thelim.:. and a chicane of dipole magnets will be used to bend the beam around the recombining dipole atthe end of the recirculator. Small correction dipoles will be used on each of the recirculation paths tocorrect for the small deflections produced by the last injection dipole.

6.2 Linac RF System

The fundamental objective of the linac RF system is to provide the energy necessary to acceleratean electron beam. In order to accomplish this. precise power and phase control of the cavity is required.Additionally the RF system must protect itself. and the accelerator. from the effects of various failures.Finally. the RF system must provide information for determining the proper tune and operation of thesuperconducting cavities and must report cavity parameters tC" the higher-level control hierarchy foroperator evaluation. Generically. the system consists of individual RF systems driving each of the cavitiesinstalled in the linac. Each system is autonomous in function and is coordinated through a hierarchy ofcontrol computers. RF frequency references. and real-time monitors and diagnostics. Each system consistsof the waveguide and coaxial cable connections to the cavity. a 5-kW output klystron. a higher ordermode (HOM) absorber. a dual-directional coupler. and an RF control module (Figure 6.9).

X418I

I

I

I

II

---------1I

III

____J

HOM absorber

X21DC power

"==:=:.1 s uppIyII1-- - -I

IIII

...J

- - - - - - - - - --- - -,1-------1,) :-------------1

I

I

RFcontrol

Systemcontrol

:X418

I "'=====~I

Ir- - --IIII

,+-c:::::r-----i So-

To computers

To computers

RFreference

rIIIII

IL_

FIgW'e 6.9. RF drive system.

127

I

CEBAF Design Repon

An imponant requirement of the linac RF system is the ability to control the accelerating fieldgradient and phase of each cavity individually. This allows utilization of each cavity's maximum gradientperformance to achieve the desired machine energy. Individual cavity controls allow precise set-pointcontrol of gradient and phase; they also simplify the structure of the control system while maintainingmaximum flexibility in the operation of the RF system. Installation. maintenance. and initial turn-on arefacilitated by having this flexibility.

The RF power required to drive each cavity is given by the formula:

The f1I'St term of the equation represents the power coupled to the electron beam with four passes.The second term represents the RF losses in the cavity due to the cavity intrinsic Q. The last termrepresents the power losses in waveguide components. The 13 is the coupling coefficient of the coupler.Thus. for a cavity accelerating potential of 2.5 MV. a delivered current of 200 ~A. a 13 of .75. andexternal losses of 20 W. the input RF power required is:

RFinpuI =(2.5 x 10~'4'(200 x 1O-~ + 2 (2.5 x lOy/ (960 x 3 x 109

)

.75+ 20

= 2693 W

The klystron will be capable of delivering 5-kW RF output at 1500 MHz. This power level meetsthe minimum requirements for the accelerating gradients plus some design margin for losses and growth.It is anticipated this unit will be a permanent magnet-focused device. eliminating the need for an elec­tromagnet and associated power supply. The high-voltage DC will be supplied from central supplies.Each DC power supply will service 20 klystrons. isolated by power conditioning and high-voltage switchingcircuitry to minimize ground loops and cross talk and allow the removal of defective or unwanted klystronsfrom the system (Figure 6.10).

Single klystron circuit

Power conditioning

buss480

3

Filamentisolationtransformer

20 total taps

rlglU'e 6.10. Typical kJystron DC distribution (with one klystron circuit shown).

128

•

I

RF Po....er S~tem

The key to the overall system is the RF control module. which has all of the circuitry required tocontrol the entire RF chain. It provides set-point control. gradient leveling. phase control and stability.quench detection. reflected power foldback. and drive for the klystron. It also provides outputs pro­portional to field gradient. forward klystron power. reflected klystron power. and ph3Se. It will be builtin CAMAC (lEEE-583) format using a 3-wide module. Utilizing this format. the control hierarchy needonly deal with this one module to control any of the parameters of the cavity. This structure simplifiesthe control software and allows replication of software routines throughout the system (Figure 6.11).

Gradient monitor

OACI----t

Reflectedpower

Forwardpower

Ref

Levelingamp

Phasecontrol

Gradientadjustment

F"JgIIn 6.11. RF con:rol module.

Intermediatepower amp (2 Wl

The advantage of individual control lies in the ability to operate each cavity at any power level upto its maximum operable field gradient. In the case of high perfcrming cavities this means operating atenergies above the specified 5 MV/m gradient. For cavities operating below 5 MV/m. this means operatingat the maximum possible gradient without breakdown. Since each cavity can be set individually, strongerperforming cavities can make up for weaker ones with individual cavities able to operate at greater thanthe lowest-common-denominator gradient.

System availability is always an operational concern. For a system composed of 400 RF drive chains.each with an estimated MTBF (mean time between failures) of 4000 hours, the probability of failureduring an operating cycle of 160 hours (6V3 days) running and 8 hours (1 shift) of maintenance with nointervening maintenance is .039. Assuming an operating energy of 4 GeV. and assuming that the totalsystem can compensate for a 6.5% loss in RF drive, then 26 drive systems can be lost before the acceleratorwill be down. Using a POISSON distribution. the probability of 26 or more systems going down in theoperating cycle described above is .013. Therefore. there is a probability of .9f57 of successfully completingan operational cycle as described above without having to stop for maintenance on the RF system.

I

7. DC Power System

7.1 Overview

The DC power system will drive the beam transport magnets of the linac segments, the beamspreaders and recombiners, the doglegs, the recirculation arc beam lines, a...-.d the external beam lines.The effe-'"tS of magneto-optical perturbations as discussed in Section 5.3 set the regulation requirementsfor the various DC power subsystems. Also in light of these effects, the dipole magnets of each of theseven recirculation beam lines and of the three external beam lines will be powered in series to reduceto negligible levels the impact of random dipole powering errors on orbit offset.

DC power supplies for the linac segments, the recirculation arcs, and the external beam lines aredistributed among the service buildings (SB) and the end station structures. (See Figure 7.1.) Quadrupole,sextupole, septum, and correction magnets will be individually powered from power supplies located inservice buildings distributed around the inner radii of the recirculation arcs and in the external beamswitchyard. The recirculation-beam-line dipole power supplies will be located in service buildings W5and ES, which also serve as acc~s buildings for equipment access to the tunnels. Some DC power suppliesmay also be located in a service building for the injector.

~o

S6W5

North linac segment S6

South Iinac segment S6

F"JgIII"e 7.1. DC power supplies for the linac segments, the recirculation arcs. and the external beam lines are disaibuted amongthe service buildings (58) and the end station structures.

131

CEBAF Design Repon

7.2 Power Supply Specifications

Typical power supplies for the dipole magnets will be solid-state. water-cooled. free-standing units.The smaller units deliver 3 kW at 100 volts with 0.01 % regulation. The larger supplies typically' deliver400 kW at 800 volts with 0.01 % regulation. Figure 7.2 is a circuit diagram of a typical system. For thequadrupole magnets the typical supply is 300 watts at 10 volts with 0.1 % regulation. Fifteen of theseindividually controlled units will be driven from a common mainframe supply of approximately 4.5 kW.The circuit diagram for a quadrupole system is shown in Figure 7.3. Other small power supplies for thelinac quadrupole. linac steering dipole. beam-line sextupole. and correction magnets are rack-mounted.air-cooled units with 0.1 % regulation. All power supplies will have both local and remote control toenable testing in the service building on local mode and in the machine control center on remote mode.

400 kW power supply

pp

ILocal Remotel Control

'1 0 MachineReference control

0center

I0

I0

DC

~outPi..t- ...

II

.c-lnterlock s

DiDole~ ~ ~"""\,

Lj II I 1I I I I

l1.So dia. water-cooled co er buss

480·VAC3 0

-

Fagure 7.2. ·fypical DC power supply cireuit for dipole magnets.

132

I I I •

I

DC Power System

I

480 VAC

3 e

Multi-outputbipolar power supply

Remote

Local •

1

6

Mainframepower supply

Control

ReferenceMachinecontrolcenter

F"JIIIR 7.3. Typical DC power supply circuit for quadrupole magnets.

DC power to the beam transport dipole magnet strings of 8 to 40 magnets in series will be carriedby 500 MCM cable. In those locations where dipoles must be ;x>wered independently, 250 MCM or 500MCM cable will be used. Quadrupole magnet power is taken to !he magnets on cable sized for therequired power. With distributed service buildings, all cable runs are as short as possible: approximately150 feet for quadrupoles, sextupoles, and correction magnets, and approximately 850 feet for dipoles.

13f!14

I

8. Instrumentation and Control

8.1 Control Requirements

The control system must maintain normal accelerator operations. monitor conditions which mayaffect the operations. provide information to the operators to enable them to take timely steps formaintaining normal conditions or averting emergencies. interpret operator commands to the acceleratorequipment and insure that they are carried out. and provide for personnel and equipment safety in apotentially hazardous environment.

Monitoring Status .'Several subsystems must be carefully controlled in order to deliver a useful beam to the target

stations. They include vacuum. cryostat. RF. beam monitor. and magnet systems. Under normal con­ditions. the major requirement of the operating system is to monitor the machine parameters to be surethat they remain within preset tolerance limits. When a parameter drifts outside preset tolerances, thecontrol system will take corrective action. In many cases the response must be quite fast; the feedbackloop will be short, and operator intervention will not be required.

The vacuum system will include a set of vacuum gauges and valves which will be distributed alongthe entire length of the machine. Since the pumps will run continuously when the machine is on, theirfeedback network is intended only to shut them off when necessary for their protection. Fast-closingvalves will seal off the cryogenic sections in the event of vacuum failure in other parts of the machine.

The cryogenics system will monitor the helium pressure (temperature) and level. It will also insurethat adequate flow of liquid helium is maintained in the cavities.

The RF system will provide individual phase and amplitude control for each klystron and separator,and will control the beam timing of the injector. Once satisfactory conditions are established, the correctset of phases and amplitudes will be maintained by a rapid-response feedback network.

The beam control system will have several parts: beam position monitors, beam current monitors,synchrotron light monitors, and spectrometers. Because of the high beam power, no intercepting deviceswill be employed in the accelerator or recirculation arcs. Even though viewers, for example, could beused at reduced currents. the interlocks which would be necessary to protect them at high currents wouldoffset any advantages they provide over nonintercepting methods. A moving-wire profile monitor willbe used in the beam switchyard, in a region where the beam size is large.

The magnet systeYJI (dipoles in the recirculation arcs and beam switchyard, quadrupoles and cor­rectors in the entire beam transport) will require close regulation. Calculations indicate that the bendersmust be regulated to about one part in 104

• The quadrupole regulation is not quite so stringent: aboutone part in 103

• The correctors are weak and do not require close regulation; they have only about 0.01times the effect of the benders, so 1% regulation is adequate for them. The magnet cooling system willalso be included. The output water temperature will be kept below a preset upper limit by means ofthermal switches, and the water flow from the mains will be checked by a set of flow meters and switches.

Data DisplayAny monitored function of the machine can be made known to the operators on demand. Since

operator response is inherently slow, messages sent during normal conditions will be primarily for iL-

135

CEBAF Design Repon

fonnation and display purposes only. Status displays can be either binary (on-off) or numerical. asappropriate. Some of the numerical data-for example. the beam position and current as a function ofposition along the machine-can be made graphical. if desired.

The display system wiil provide several levels of message priority. Simple status messages requiringno operator response will be written only to a file. to be called up for display when and if desired. Alog file will be retained to get a complete picture of machine status at hourly intervals. or more often ifdesired. At another level are messages which require response: depending on the severity of the conditionthey report. these may merely require acknowledgment. or they may have absolute priority and requireimmediate action.

The display system will use color TV monitors. with color-coding part of the program design (thatis. blue will indicate off-line status. green satisfactory. yellow marginal. red unsatisfactory. and flashingred emergency).

Operator ControlIn routine operation the operators will have as little to do with controlling the machine subsystems

as possible. The control architecture will rely on distributed intelligence that will maintain proper com­ponent operation with minimum flow of data to and from higher levels. The actual control will beexercised by CAMAC modules. which are themselves controlled by microcomputers with the correctCAMAC commands in their memories. When it is desired to change the machine status. the operatorcan take control by issuing the correct commands from his console. In order to minimize the chancesfor erro:". this will be done as much as possible by responses to a menu.

The supervisory computer itself will then interpret the operator's instruction. put it into correctCAMAC fonn. route it to the microcomputer(s) which control(s) the intended function. and then waitfor the response from the affected monitors to show that the instruction has been carried out. Failureto close the loop in this fashion will result in an abnonnal-status message. which mayor may not requirefurther action.

Personnel SafetyA primary concern of the operating system must be the safety of people. either the general public

or the operators.The major potential hazard to the general public is radiation. which is monitored at several positions

at the site boundary to insure that it is below specified levels. If levels get too high. the machine willimmediately be shut down through an injector fast-cutoff link until corrective action is taken.

Operator protection places more stringent requirements on safety systems. Operators are exposedto larger radiation doses than the general public. and may have to contend with the common industrialdangers of electrical shock. fire. or oxygen deficiency hazards (ODH). The operating system is designedso that the beam is shut off whenever personnel are at risk. Hard-wired interlocks prevent entry intoradiation areas when electrons are being accelerated. no matter how small the beam current. CEBAFwill develop fonnal procedures based on industrial experience and safety regulations to minimize operatorrisk due to the other hazards.

Because of its importance. radiation safety is covered in a separate section of this report. Section3.4. Chapter 12 also discusses safety.

8.2 Beam Instrumentation

The control and measurement equipment is. with few exceptions. standard. We will as far as possibleuse equipment that can be supplied by industry. in order to minimize the costs of design and development.The few items which cannot be purchased are the various beam monitors. Fortunately. they are similarto devices which have been used in other laboratories. so that at least the principle of operation isestablished.

Injector Fast-Cutoff LinkFor reasons of both personnel and equipment safety. if the radiation level becomes too high. it is

necessary to tum off the beam. Because one mechanism for increasing the radiation is catastrophic beamloss caused by steering the beam out of the pipe. the response must be quite fast. The beam will burnor melt a hole in any solid object it hits in about 25 microseconds. We cut the beam off by breaking the

136

.- - I

• I

Instrumentation and Control

injector permissive signal transmitted on a fiber-optic cable running above ground (to avoid radiationdamage) the length of the machine. Fiber optics are chosen tCl minimize attenuation in the kilometer orso of transmission path. and to eliminate interference and spurious signals. (The Newport News area isvisited on average by 37 thunderstorms each year.)

The input signals will come from a set of aluminum-cathode electron multipliers. acting as radiationdetectors. or from the beam current monitors to be described below. acting as beam loss monitors. Inorder to deliver the signal as fast as possible. hard-wired analog circuitry will be used for the analysis inorder to avoid digitizing delzy. Even so. propagation delays of up to 5 microseconds are unavoidable.and about 15 microseconds of beam is already in the beam pipe.

As a backup protection system. to be used only if these two systems fail. we will use a set of ionizationchambers. These are inhereatIy slow devices. however. with response times of almost a millisecond. Itis likely that the machine would sustain some damage before they could respond.

Beam Current Monitor (RCM)Current monitors serve to determine beam current level. measure transmission through machine

sections. detect and localize beam loss. and provide a normalizing signal for the evaluation of beamposition monitor signals. We chose a resonant device because the increased sensitivity easily offsets thedisadvantages of reduced bandwidth and slower response. Beam current is monitored by a cylindricalcavity resonator oscillating in its TMolO mode. So long as the beam does not move very far from the axis,the output signal (voltage) is proportional to the beam current. In order that the cavity respond fairlyrapidly to changes in the beam current. it is loaded so th.. ~ its quality value Q is only about 100. Loadingis done by the antennas. The output signal is divided. with one path sent to the CAMAC crates foranalysis. the other to an analog comparator which compares its signal with that from the BCM at theinjector. A low signal is evidence of beam loss. It is also possible for the beam intensity to be too greatin some areas. such as a low-counting-rate target station. The monitors will therefore define an allowedcurrent window; the injector will be inhibited if the current is either too large or too small.

Beam Position Monitor (RPM)Each BCM is associated with another cylindrical cavity resonator oscillating in its TMuo mode. The

field amplitude is proportional to the dipole moment. the product of the beam current and its displacementfrom the axis. Four straight wires project into the cavity from one of its fiat faces and act as antennasto measure the voltage of the resonant mode. They break circular symmetry about the axis, so resonantmodes corresponding to displacements parallel to the x and y axes are decoupled.

Both the BPMs and the BCMs are designed to oscillate at 3.0 GHz rather than the machine frequency,1.5 GHz. This is done only to reduce their sizes. The beam has enough harmonic content that no problemsare introduced.

Synchrotron Light Monitor (SLM)Although we cannot use any intercepting devices for measuring beam quality (spot size, emittance,

pulse length). we can use the synchrotron light emitted in the benders. We can also use it as input to abeam position monitor. but it would not be as economical as the cavity resonators discussed above. Inorder to be useful for the CEBAF CW beam, resolution of the object must be about 0.1 mm or better.Because of the angular distribution of synchrotron radiation, the telescope we use as the optical systemmust work in the far ultraviolet, at about 50 nm or less.

The synchrotron light monitor will be a reflection telescope with overall magnification of about 10.Aluminum coatings will not work in the EUV; experiments made by NASA astronomers indicate thatthe mirrors must be coated with silver or indium. The beam cross section will be imaged on a charge­coupled device (CCO) for immediate digitization. Readout will use standard techniques, also developedfor NASA. Data collection rates will be a few per second.

The beam pulse length can also be measured by means of the synchrotron light. The duration ofthe light pulse is very nearly equal to that of the electron bunch, and can be measured by means of astreak camera. At present. streak cameras with a resolution of 1.5 to 2 ps are available, which is sufficientto resolve the bunch length of 3 ps.

Beam Analysis SpectrometerA simple spectrometer will be set up near the injector end of the accelerator. It will consist of only

a bending magnet, a quadrupole, and a scintillation screen. Its purpose is to observe the momentumspread in the beam at an early part of the cycle.

137

CEBAF Design Repon

MisceUaneousA few broad-band. high-frequency current and position monitors (gap monitors and strip-line elec­

trodes) wiII be instalIed for the study of fast transient beam behavior. Together with network and spectrumanalyzers. they will enhance beam diagnostic capabilities and complement the main set of beam instru­ments used for steady-state operation.

8.3 Control System

The control system will consist of a hierarchy of computers and CAMAC controls. in a structurequite common in laboratories like CEBAF. The control system will provide minicomputers. local areanetwork (LAN) lines for computer interconnection. and CAMAC crates for the various operating sub­systems. The minicomputers will be of a high-speed. 32-bit architecture. These computers are relativelylow cost. off-the-shelf items. The multicomputer design will prevent anyone computer failure fromshutting down all machine controls. and quick replacement capability minimizes downtime. The LANlines (IEEE standard 802.3) will interconnect computers throughout the system with 10 MB data rates.

Multicomputer and multi-LAN line structures provide for subsystem autonomy and high-speed dataflow. CAMAC crates (IEEE 583) allow for modularity. and thus for quick repair and ease of expansion.Many CAMAC modules are commercially available. and they provide for high data transfer rates (>600 kB).

The I&C equipment is to be structured at two levels. the higher of which is the supervisory leveloutlined in Figure 8.1. It consists of several operator consoles located at the machine control center andinterconnected on an LAN. Each console contains a minicomputer. color graphics monitors. and keyboardand touchscreen interfaces. Their function is to provide monitoring and display of equipment data. andto form the interface for operator control of their assigned machine subsystems. The lower five operatorconsoles in Figure 8.1 indicate how they are to be defined. The I&C Supervisor console shown hasmonitoring capability over all systems.

To local computerI

I&C supervisor

isory

Injector Linac Cryo/vac Switchyard Safety ops.operation ~ operation ~ operation ~ operation ~ supervisorconsole console console console console

I I I I ILocalcomputer LANs

Super vLAN

FIgUn 8.1. Supervisory level. I&C.

Each of the supervisor computer consoles is then connected via LAN lines to its computers on thelocal control level, shown in Figure 8.2. This set of computers will be located at the service buildings.and assigned to operate and monitor a sector of a machine subsystem. They will update data to andreceive commands from their supervisor computers and provide local control for testing and maintenance.While interconnection on an LAN does allow for these computers to communicate among themselves,the controls are designed to operate in a master/slave mode. with the supervisor computer as the master

138

-

I

Instrumentation and Control

I

Supervisory ILAN II I

Subsystemoperation I

computer( s) Iconsole I LocalLocal

Icomputer GPIBLAN

I CAMAC crates I

Figure 8.2. Local control level. I&C.

to which all data will be sent and from which all commands are issued. This is designed to minimizeLAN data traffic and maintain single-point control for overall system operation.

The local computers are then tied to the CAMAC crates. the last items provided for by the I&Csystem. Most of these crates will be located at the service buildings and tied directly to the local computersvia a general-purpose interface bus (GPIB. IEEE 488). Certain distributed systems will be tied to thelocal computers by a fiber-optic serial highway. The number of CAMAC crates connected to anyonecomputer varies by system from 6 to 15.

-

I

9. Cryogenics

9.1 Background

A superconducting linac requires a cryogenic system to cool the machine from ambient temperatureto the operating temperature. This operating temperature must be below the superconducting transitionpoint for niobium. 9.2 K. The maintenance of the temperature requires a balance between the total heatload of the cryomodules and the capacity of the cryogenic system to remove the generated heat. Additionalspare capacity is required for reliability as well as cooldown.

There are two types of resistive losses in a superconducting RF cavity: residual resistance. and BCSresistance (Bardeen. Cooper. and Schrieffer). The residual resistance is caused by localized resistiveareas where defects. impurities. or surface dirt disturbs the superconductive properties. The BCS re­sistance increases with increasing frequency. and decreases as the operating temperature decreases. Othersources of 2 K heat include static heat leak. conduction of heat dissipated in the input waveguide. andabsorption of higher order mode power generated by the beam current. For CEBAF. an operatingtemperature of 2.0 K is an economic optimum.

9.2 System Requirements

The helium refrigeration system at CEBAF must provide an adequate flow of 2 K helium to com­pensate for resistive heating in the niobium and for heat leaks in the cryostat and distribution system.In addition. it must provide helium at 40 K to keep heat shields in the cryostat below ® K. Table 9.1summarizes the calculated heat losses for CEBAF. assuming an accelerating gradient of 5 MeV/m at aresidual Q of 3 x 109

•

The CEBAF cryogenics system is designed to handle 150% of the calculated load at 2.0 K and 200%at 40 to 60 K. In addition. superconducting magnets are likely to be used in the experimental equipment.These magnets may require helium at 4.4 K to handle a cooling load of 130 literslhour (260 literslhourwith 100% design overcapacity). The option exists to meet this requirement either by purchasing acommercially available 4.4 K helium refrigerator for the experimental areas. or by designing the centralhelium refrigerator to handle this additional load: we have chosen the latter due to its lower requirementsfor operating manpower. Table 9.2 presents the total cooling requirements (including experimental equip­ment) for CEBAFs refrigeration plant.

The refrigeration system should be designed to provide easily for fast cooldowns. power failures.and various repairs and decontaminations. The distribution system must be able to handle instabilitiesand cryostat replacement without seriously disrupting linac operations. Furthermore, the design shouldbe amenable to cost-effective upgrades by additions of compressors. expanders. and cold compressorcapacity.

141

I

CEBAF Design Repon

Table 9.1Heat Load Summary

2.0 K 40-60 K

RF Heat LoadsRF residual lossesHOM absorberRF coupler jointBCS losses

Total RF

Static Heat LeaksRadial supportsLongitudinalRF coupler conductionBlackbody radiationMLI radiationV-tubes (2.0 K)Inlet IT valveVacuum shutoff valveShield V-tubesTransfer line

Total heat leaks

Total

1800 Wl00W75W

457W2432 W

50W50W

142 W64W44W

270W13W

135W

67W

785 W

3217 W

250W250W

1535 W655 W860W804W25W

402W318 W870W

5969 W

5969 W

Table 9.2Cooling Requirements

He temp. Calculated Refrigeration Ratio Pressure(K) load capacity (%) (atm)

Linac cavities 2.0 3200W 4,800 W 150 0.031

Linac heat shields 40.0-60.0 6000W 12,000 W 200 3.0

End stationliquefaction 4.4 130 l/hr 260 l/hr 200 1.2

9.3 Operating Temperature Selection

The choice of operating temperature affects the BCS component of the cavity Q, and thereby theRF heat load, as well as the refrigeration costs (both capital and operating). The BCS losses vary inverselywith the cavity Q, approximately doubling every 0.2 K. Figure 9.1 shows the total heat load as a functionof temperature. The refrigeration costs vary inversely with the temperature; in addition capital costsincrease with the 0.7 power of heat load, while operating costs increase to the 0.85 power. The net effectis shown in Figure 9.2.

CEBAF workshop participants have chosen 2.0 K as the operating temperature. The BCS losses,while an exponential function of temperature, are still a small fraction of the total heat load at 2.0 K.Figure 9.2 shows that the refrigeration capital costs are flat to 0.5% between 2.0 and 2.2 K. An operating

142

Cryogenics

6000

5000

~ 4000

...Q)

~

~ 3000 '"""1.._--BCS Losses

2000 Temperature-independentRF losses

1000

Static heat leaks

O+----~------,r__---""""!""----..,...---......,

1.8 2.0 2.2 2.4Temperature (K)

2.6 2.8

FIgUre 9.1. Total heat load as a function of tcmperature.

1.15

1.10

1.05

1.00

Normalizedcapitalcosts

2.82.62.2 2.4Temperature (K)

2.00.95 +----+----.....,r__---....,.----"T'"---......,

1.8

FIgUre 9.2. Nonnalized refrigeration costs.

143

CEBAF Design Repon

temperature below 2.0 K would not be cost-effective: moreover. it ",ould be technically difficult due tothe very low vapor pressures (less than 0.031 atm). Above 2.5 K (0.1 atm) we could delete one stage ofvacuum pumping. but the BCS losses are so large that it would not be economical.

This leaves us with an operating range of 2.0 to 2.5 K. We have chosen to size the distributionsystem to be optimized for 2 K operation with a flow safety factor of two times the calculated heat load.Since possible future higher cavity gradients will tend to shift the optimum toward lower temperatures.this will permit future beam energy increases without requiring replacement of the relatively expensivedistribution system.

The refrigerator will be specified to be a 2 K unit. This provides us with two advantages: First. ifthere are problems during initial commissioning of the system we have a 0.2 to 0.5 K safety factor inthe pressure drops. and second. the exchangers will permit future upgrades.

Two-phase helium becomes a superfluid at 2.177 K. While we do not expect superfluid problems(vacuum leaks. increased heat leak. or oscillations). we plan to commission the accelerator at a tem­perature of 2.2 to 2."s K with a few percent higher operating cost. It is our intention to operate at 2.0K after the initial commissioning period.

9.4 Cycle Design

The CEBAF refrigeration system is shown in block diagram form in Figure 9.3 and in schematicform in Figure 9.4. The primary systems are the screw compressor system. a standard cold box. the4.4 K dewar system. the distribution system. and the cold compressor system. Table 9.3, which is keyedto Figure 9.4, provides a conservative set of process points. Helium flowing to the linac is superimposedon a helium pressure-temperature plot. Figure 9.5.

We have chosen this configuration because it almost completely decouples the standard refrigeratorfrom the subatmospheric system. This decoupling of the cycles has several advantages. From a procure­ment standpoint it breaks the cryogenics into a standard off-the-shelf refrigerator and a "high tech"subatmospheric module. which in turn also simplifies the operation and controls. The requirement fordouble seals with a guard vacuum to eliminate air leakage therefore only applies to the subatmosphericmodule.

COldcompressors

Screwcompressors

Gas returncompressor

/;../.l------....L.,

Standard helium system /---------_-/Subatmospheflc system

Cold box

Figure 9.3. BI.xk diagram of CEBAF refrigerator.

144

Cryogenics

Process PointsSeveral key process points of the process cycle (as seen in Figure 9.4) are as follows:

Point 0Helium cOlT!pre-ssed to 20 atmospheres and recooled to 300 K is first purified through three stages

of oil removal: bulk. mist. and vapor. The bulk ~paration is accomplished by velocity reduction andmacroscopic barriers. while the mist is removed by a coalescing-type filter. The vapor phase is removedby a charcoal absorber. The compressed helium is introduced into the refrigerator.

Point IHelium is cooled to 80 K by heat-exchanging the returning helium gas with about 5% of the flow

passing through a separate nitrogen precooler. which is used to prevent nitrogen freezing and crackingthe exchanger.

Point 2Helium is cooled to 60 K by returning cold helium gas and by helium that has passed through the

first expansion engine. 0 •. The main flow splits into two parts. with about 15% passing L~rough the firstexpander and the remainder (the main stream) bypassing the expansion and proceeding directly to Point4. The flow portion sent through the expander is reduced in pressure by doing work on a gas turbine.and is cooled to 40 K at 3.5 atmospheres.

Point 4Another flow split passes 50% of the total flow of the helium through Oz. which operates between

approximately 20 K and 12 K.

POlnt6Helium cooled by heat exchange is reduced in temperature to approximately 9 K. The entire flow

~s introduced into the final expansion engine W. where it is expanded to 2.8 atmospheres and cooled to5.6 K.

Point 8The helium passes through an additional exchanger and then through a 1.2-atmosphere subcooler.

Its flow splits to the linac. the end stations. and the dewar subcooler, which connects to a 3O.000-gallondewar.

Point 9Helium is cooled to 2.2 K by counterflow of the 0.031-atmosphere helium returning from the

cryomodules. The supply is supercritical gas at 2.8-atmosphere pressure. and is sent through the transferline for distribution to the cryomodules.

Point 9//2Each cryomodule contains 8 cavities and has a IT valve. which is operated by a liquid-level control.

The helium surrounding the cavities is maintained at 0.031 atmosphere and 2.0 K.

Point 391/zHelium gas at 0.03i atmosphere is returned from the cryomodules, intercepting some additional

heat leak, primarily the return bayonets and U-tube.

Point 38After heat exchange with the supply stream the helium gas is compressed by cold compressors to

1.2 atmospheres in three stages of compression.

Point 3424 K helium is returned to the cold box to be used for the counterflow heat exchange.

Point 20Helium at 1.05 atmospheres is compressed in two successive stages in the screw compressors to 3

atmospheres and then to 20 atmospheres.

Additional FeaturesSome additional features worth noting are that the refrigerator may operate as a conventional 1.2­

atmosphere, 4.4 K helium refrigerator by simply turning off the cold compressors and passing the flowaround them. The refrigerator may operate at reduced capacity if any of the expanders are off for repair.or it can operate at close to full capacity for up to three days by consuming liquid.

145

I

CEBAF Design Rc:pon

146

•

Purifier

,. - - - -r--04""';;;;"'I

,",I \

I Dc \L,J

L. - - - -L=r--DirG-'1''--t_'''''''::=l

30.000-galHe cewar

8"2L....__---t Primary load

FIglIn 9.4. Schematic diagram of CEBAF relrigcrator.

34

36

I I I

Cryogenics

Table 9.3CEBAF Refrigerator Process Calculations

Pr~""Ure Temp. Enthalpy FlowPoint (atm) (K) (JIg) (g1sec) Notes0 20.0 JOO.OO 1579.0 1713OA 20.0 300.00 1579.0 162508 20.0 300.00 1579.0 88I 21J.() XO.OO 434,4 1713IA 20.0 RO.OO 434,4 162518 20.0 80.00 434,4 882 20.0 60.00 329.0 17132A 20.0 60.00 329.0 146028 20.0 60.00 329.0 253 Eff = 75% o = 30.335 W:- 20.0 38.50 213.11 14604 20.0 20.00 108.9 14604A 20.0 20.00 108.9 612

..- 48 20.0 20.00 108.9 &48 Eff = 70% Q = 30.757 W5 20.0 12.50 61.54 612b 20.0 9.00 38.25 612 Eff = 70% Q = 7.124 W

7 2.80 5.61 26.61 6127! 2.Hl.l 5.50 22.53 6127~A 2.110 5.50 22.53 6127!B 2.80 4.73 12.73 08 2.80 4.50 I!.36 612SA 2.l\O 4.50 11.36 24088 2.Hl.l 4.50 11.36 368HC 2.80 4.50 11.36 4 120 literslhour9 2.Hl.l 2.20 5.10 240

~ 0,()31 2.00 5.10 240

lOA 3.0 287.50 1509.0 1101il 3.0 78.00 420.4 llOI12 3.0 57.69 314.6 llOI12A 3.0 57.72 314.8 986128 3.0 57.40 313.1 il5 ~H = 104 0= 11.960 W13 3.0 37,49 209.1 98613A 3.0 37.49 209.1 848138 3.0 37.49 209.1 138

13C 3.5 37.49 209.1 ll5130 3.5 37.49 209.1 253

14 3.0 16.39 97.31 848IS 3.0 11.94 72.63 84S

20 1.05 287.50 1508.0 60821 1.08 78.00 420.0 60822 1.11 57.73 314.6 60823 1.14 37.50 209.3 608.,. 1.16 16.11 97.31 368_..25 1.18 12.00 75.41 36826 1.20 5.23 36.68 36826A 1.20 5 .,~ 36.68 368•.t..>

268 1.20 4.50 30.76 0 ~H = 19.40 Q = OW27! 1.20 4.424 29.94 368 Gas28 1.20 4.424 11.36 368 2 phase

34 1.20 23.82 137.9 240 Eff = 68% Q = 15.408 W

35 0.30 11,45 73.70 240 Eff = 70% Q = 6.230 W

36 0.10 6.42 47.74 240 Eff = 60% Q = 3.845 W

38 0.031 3.32 31.72 240;19 0.031 2.15 25.46 240 ~H = 0.81 Q = 194 W39! 0.031 2.00 24.65 240 ~H = 19.55 Q=4692W

40 1.00 275.00 285.1 267 ~H = 376.9 Q = 100.632 W41 1.20 78.95 -91.8 267 4.0 atrn liq.

Percent of Camot = 15.3% with compressor 55% isothermal

147

..,..

.. I I

CEBAF Design Repon

30

o

Warmcompression

Third-stagecold compression

r---------J 20

Second-stagecold compression

point

Wet expansion

6

First-stagecold compression

8

Lambdaline

93

10

20

0.10

0.01 -+--"'--I""'-'-I""'-'..,......,....__I""""I"'.,..,.---'t""""-'t""""-r-~...,....T""T...,....,---"T""" ..........

0.03

E 1.0-as VaporCD....

-I::lenenCD....0- 0.3

He JI

1 2 5 10 20Temperature (Kl

50 100 300

FJgUn 9.5. Pressure-temperature plot.

9.5 Central Helium Refrigerator

Both efficiency ar.d economics of scale support a decision to use a central helium refrigeration plantrather than a distributed refrigeration system. This plant must be capable of providing helium at three

148

I

Cryogenics

temperatures for the entire CEBAF facility: 40 K. 4.4 K. and 2.0 K. (The cooling capacity required ateach temperature is given in Table 9.2.)

The central heli;Jm refrigerator consists of a large section that is just a common commercial heliumrefrigerator and a subatmospheric system to lower the temperature to 2.0 K. There is a further separationof the ha;-dware into the heat exchanger box. which will sit outdoors. and a pair of cold boxes thatcontain all of the valves and rotating machinery. This arrangement will result ir. a cost-effective use ofthe refrigerator building. The separation of the refrigerator into two sections will insure that if futuremodifications to the 2 K cycle need to be made. they will not affect the majority of the equipment.

The central heliu;n refrigerator will include the following main components: compressors. heatexchange;-s arranged in a vacuum vessel called a cold box. expanders. cold compressors. and a controlcomputer.

The compressors will be two stages of oil-flooded screw machines. which have become the stc.ndardfor large helium refrigerators and are extremely reliab!e. The design will be such that the addition ofcompressor capacity at a later date will provide increased refrigeration capacity. The first stage uses threeunits operating at 92% of full load and a standby spare. The second stage uses three units operating at87% of full load: with one off. the interstage pressure increases to 4.0 atmospheres and one can continueto operate at close to full capacity.

The expansion engines are oil-bearing turbines. They have enjoyed an excellent record of reliability.and large units have efficiencies in the range of 70 to 80%. The CEBAF refrigerator will use two expandersoperating between 20 atmospheres and 3 atmospheres to optimize the overall cycle efficiency. A thirdexpander in series with the main refrigeration flow approximately doubles the final output. comparedwith refrigerators with a Joule-Thompson (JT) valve in this location. This final expander will outputhelium in a supercritical state to prevent two-phase flow "instabilities.

The warmest expander. 0 1• will be redundant: each will have its own independent inlet filter. Bothturbines are used for cooldown. with only one in use during normal operation. The independent inletfilters are required to trap low-level contaminants such as water and dust. The remaining two turbinesare not redundant. but all fOllr are installed in a configuration permitting the replacement of the wheelor cleaning of the filter with the refrigerator operating. During repair the system can be kept at fullcapacity by consuming liquid from the storage dewar and converting it into refrigeration.

The subatmospheric subcooler must be large enough to handle the low-pressure helium (0.031atmosphere) returning from the linac cryomodules. but due to the low temperatures it is relatively small.The conceptual design of CEBAFs refrigerator is based upon the use of three cold compression stagesfrom 0.031 atmosphere to 1.2 atmospheres. This design avoids the difficult and expensive problem ofextremely low pressure heat exchangers. which would become huge due to the low pressure drop re­quirement

The cold compressors are at the forefront of helium refrigeration technology. The requirements ofhigh throughput and exceptional reliability demand the use of rotary compressors. These machines arenow in use as a result of extensive development by several manufacturers. namely Rota-Flow and Crearein the U.S. and L'Air Liquide and Sulzer in Europe. These essential components provide the low­temperature compression to maintain both the operating conditions (0.031 atmosphere. 2.0 K) in thecryostat and positive pressure in the ambient suction piping and compressors (1.05 atmospheres).

The cold compressors are fully redundant but not removable from the refrigerator while operating.due to the subatmospheric operation. The three pairs are all used during the transition from 4.5 to 2.0K. Whether the redundant units are turned off or kept idling will be determined by operating experience.

9.6 The Cryogenic Distribution System

The distribution system must be sufficiently flexible to allow a wide range of operating conditions.It must be able to handle contingencies. such as the replacement of a cryomodule while maintaining thesystem in a standby condition. We have selected a solution that provides the required flexibility and alsominimizes costs: in addition. it permits the accelerator to operate while a cryomodule is either beingwanned up or cooled down.

The system depicted in figures 9.6 and 9.7 is based upon using the string of cryomodules as part ofthe supply transfer line. and a transfer line for the return flow. The cryomodules are series-eonnectedin an "H" pattern utilizing V-tubes and internal flow to distribute 2.2 K helium at 2.8 atmospheres and

149

I

I I

CEBAF Design Repon

40 K helium at 3.5 atmospheres. Each cryomodule (Figure 9.6) is connected to the return cold vacuumline to maintain its 0.031-atmosphere internal pressure. and the shield flow is returned to the transferline at four places. one at the end of each arm of the "H". This series-parallel system minimizes the costof the distribution system. Table 9.4 is the list of distribution process points and is keyed to Figure 9.6.

Table 9.4Distribution System Process Points per Module

Point Temp. Press. Enthalpy Mass flow(K) (alm) (g1sec)

9 2.20 2.800 5.10 4.8

9! 2.00 0.031 5.10 4.8

39! 2.00 0.031 24.65 4.8

39 2.15 0.031 25.46 4.8

Be 37.5 3.5 209.1 28.75

12B 57.4 3.0 313.1 28.75

Supplytransfer

line

Returntransfer

line

60 K return==::;~:::¢~=::::;;jj:=!::;:;::::;::::::====:::==========:::::;--_.*"];(l

39 ~~ Vacuum 0.031 atm 2 K return~~ shutoff

valve

-I 40 K shields 40 KL _13C

I~

---, r--------- -----------,I I 39 1/2' I

: : 1 0.g3~ ~tm I 112 or 13 cryomodules I ,...----L-.:I I ~ 9 1/2 ~ ""--r-----J II ~ I "'L9ttJT val~e ~-r2.2 K 2.8 atm supply I

.....'"----:-~:-'--IIE-~_L.-..-_~-3I:-.&--•••-----~_3IIl....----E_---(:x:i--r'_Jo_'T'....,I ~I: ~ :: I

3.5 atm supply I_________ ..-J

________1 _1.5 atm 300 K return

FJgUn 9.6. Cryogenic distribution system.

If a cryo-unit must be removed. the cryomodule containing it would be isolated from the supply byremoving the V-tubes at each end of the module. These V-tubes are replaced by a V-tube which spansthe gap created by the cryomodule and allows the helium supply to the remaining modules to be resumedin a short time. The CI);omodules in the other three arms of the "H-- are completely unaffected by thisoperation. Those in the affected arm must rely upon the large helium inventory in each module tomaintain the temperature during the very short transition time. The modules upstream of the isolated

150

I

I • ---Cryogenics

Vacuumshutoffvalve

0.031 atmguardedU-tube

From end

0.031 atm 2 K return

Transfer line

60 K shield return

One cryo:"..odule (one of 50)~Helium relief

r----------- --------1I 2.0 K cryostat ~

I I Vacuum

2.2 K I --- ""II!" ~ ~ ""i'" a::::::;::;c- ~ I relief

supply I IU-tube I ! I

I JT-v;ive --------;.2K interna(~upp~line-- II II I

D.....;'~-l........-_--------'7'----------_---:....'-t><JE-­

I 40 K shield J I

40 K ~UPPIY U-tu~ - - - - - - - -- -- - - - - .....J

F"JgUn 9.7. Cryomodule flow schematic.

module may still be supplied with 2 K helium. while the downstream modules must rely upon theirhelium inventory of 1500 liters each to keep cool. The removal and replacement of the V-tubes will nottake more than 10 or 15 minutes, which is much less than the several-hour stand-alone capacity of eachcryomodule.

In this system. the transfer lines are a simple coaxial design. which can be mass-produced easily andeconomically (Figure 9.8). The system will be easy to control. because it has few control valves. eachwith a well-defined function. A control valve at the end of each branch of the "H" will maintain theshield at a temperature between 40 K at the inlet and no more than 60 K at the outlet. A control valveat each cryomodule will maintain the liquid level in each module in the full state. while the parallelconnection to the cold vacuu...n line will keep the pressure in each module at 0.031 atmosphere for 2 Koperation.

9.7 End Station Cryogenic System

The present ~esign for the CEBAF end stations includes several large superconducting dipoles.quadrupole doublets, and toroidal subcoils. As a basis for the conceptual design. they are assumed tobe pool-boiling magnets. These magnets are very simple to control cryogenically. as they need onlyliquid-level control. The magnets will be cryo-stable. with the possible exception of the quadrupoles, sothat quench detection and protection is reduced to a manageably simple system.

151

CEBAF Design Repon

60 K

2.0 K return0.031 atm

Return transfer line

in. SCH 10 pipe

in. tUbe, 0.120 in. wall

a in. SCH 10 pipe

6 in. SCH 5 pipe

2.2 K supply

40 K supply ~t::===::::::::::--.....

Supply transfer line

10 pipe

2 in. SCH 5 pipe

1 1/2 in. SCH5 pip~

3/8 in. SCH10 pipe

Superinsulation60 layers

Superinsulation20 layers

End station transfer fine

Fagure 9.8 Transfer line cross sections.

152

..

Cryogenics

The helium system should appear as a utility to the end stations. rather than as an overhead operationwith which they must be intimately involved. This concern and the desire to reduce overall system costssuggest that it would be desirable for the central helium refrigerator to provide for this magnet cooling.Thus. supercritical cold gas will be delivered to the end stations and distributed locally via transfer lines.

Each magnet would have a liquid-level control that would run an inlet JT valve. Thus the problemsassociated with distribution of liquid helium and two-phase flow would be eliminated. The end stationshould have a small compressor and a suction buffer tank to return the warm helium gas from the magnetsthrough a small high-pressure line. The gas will pass through the nitrogen-cooled utility purifier beforeentering the central refrigerator. Table 9.5 summarizes the end station magnets.

Table 9.5End Station Magnets

Lead FiowQuantity Type Liquid Required Heat Load

End Station A2 dipoles 4 llhr 10 lIhr

4 GeV 1 Ig. quad 2lfhr 5 llhr1 sm. quad 2 lfhr 4 llhr

1.2 GeV2 dipoles 4lfhr 10 llhr2 quads 4lfhr 8 llhr

End Station B8 toroidal coils 24 lfhr 20 lfhr

End Station C2 dipoles 4lfhr 10 lfhr

4 GeV 1 Ig. quad 2 lfhr 5 llhr2 sm. quad 4lfhr 8 llhr

50 lfhr 80 lfhrDesign load 130 lfhrRefrigerator capacity 260 lfhr

9.8 System Operation

The system initial cleanup, start-up, cooldown, and warmup are integral factors in the overallcryogenic system design and can affect the cavity-cryomodule design. It is also essential to establish earlyin the design process what the cooldown limits are. In the absence of a thermal-mechanical stress limit,the limit is established by the pressure drops, refrigerator capacity, and return compressor capacity.Pressure changes can be dealt with in most cryogenic system designs in a straightforward way, and infact, since most pressure drops affect efficiency. they are reduced as a matter of course. Capacity, however,is an economic question and is related to the operation modes that are desired. We plan to keep CEBAFcold for periods of up to a year with only partial warmups of small subsystems to allow replacement. Acooldown cycle for the whole of CEBAF of two days' duration is a reasonable expectation, and thus isthe basis for this design. Tests at Cornell at very high cooldown rates have shown no degradation ofperformance: therefore there are no constraints on the system imposed by thermal stresses at the cooldownrates that we will achieve. For single cryomodules we plan a 30-minute cooldown for the shields and asix-hour cooldown for cavities.

The initial cleanup process consists of removing debris and water from the system. The parts of thesystem that can take high pressure will be pressurized and blown down to remove dust and debris. Thisis to be followed by dehydrating the system, with a once-through flow of warm nitrogen-a process thatmay require several weeks. With the turbines and cold compressors removed and special debris-catchingfilters installed, the system will be run at maximum flow rate for a few days. The turbines and coldcompressors are then to be installed, and the system pumped and backfilled with helium. This one-timeprocess has become an industry standard procedure for commissioning large refrigerators.

153

•CEBAF Design Repon

The start-up process consists of the steps outlined in Table 9.6 and explained in the text that foHows.

Table 9.6Start-Up Process

Operation

Leak testing

Pump and backfilI

Purif.cation

60 K cooldown

5 K cooldown

2 K cryomodule filI

Time

- 4 hours

8 hours

24 hours

7 hours

5 hours

- 8 hours

Leak TestingThe system is pressurized to 1.5 atmospheres. and both the cryostat vacuum and guard vacuum

systems are monitored for ieaks. Warm connections may be tested by bubble testing. and the overaHintegrity gauged by long-term pressure measurement. Bubble testing' of alI warm joints is a continuingprocess that must be pursued even after the system is in operation. since most day-to-day loss of inventoryand diffusion of contamination is through this mechanism. The only way to reduce inventory loss to atolerable level is by careful continual attention to this detail.

Pump and BackfillThis stage of purification involves the bulk removal of atmospheric gases by vacuum pumping and

the subsequent dilution by backfilling with pure helium. Cryogenic systems always have long deadendtubes that usualIy have very low pumping speeds. and cryogenic systems always have smaH leaks. Thesetwo constraints place a lower limit on the purity level that can be attained by this technique. UsualIyseveral hundreds to 1000 ppm is the best that can be achieved in a large distributed system with threepump-backfilI cycles.

PurificationOnce initially dehydrated. the system is always kept at positive pressure. even during long shutdowns.

Some water leaks into the system during the pump and backfilI cycle. especially if it is done slowly. Thiswater plus any that diffused in during operation wilI be removed by the dual dehydration beds on theinlet to the utility purifier. AlI contaminants other than neon and hydrogen wilI be removed by the dualnitrogen-cooled charcoal absorber beds.

While the system is at pressure, alI of the deadend tubes in the system are flow purged. This processis repeated three times. When the system reaches < 0.2 ppm water and < 0.5 ppm nitrogen, it is definedas clean. These extremely low levels are required for long-term operational reliability. The cold com­pressors are now started to reduce the load pressure to 0.8 atmosphere, and the nitrogen contaminationlevels are monitored for air leaks. The cold compressors are taken to minimum speed after the systemis confirmed to be leak tight.

Initial CooldoWD to 60 KHelium gas compressed to 20 atmospheres is precooled by LN2 and used to cool the plant to the

80 K temperature level. The turbines are operated at half speed to purge out all trapped contamination.This produces 60 K helium which can then be introduced to the cryomodules. The bulk of the heatcontent of the system is removed in this step. At a cooldown flow rate of 400 gisec for the entire linac,we wilI accomplish this cooldown to 60 K in 7 hours, assuming a 25% thermal efficiency in utilizing theenthalpy of the helium.

CooldoWD to 5 KAs the vent temperature reaches 65 K. helium is returned to the cold box at point 22. The expansion

engines are taken to operating speed and the output of the refrigerator lowered to 5 K. Liquid is addedfrom the dewar to match the density changes in the linac. The flow rate possible is 160 g/sec at thistemperature level; consequently we can accomplish this phase of the cooldown in just over 5 hours.

154

..Cryogenics

2 K Cryomodule FillLiquid helium pumped at up to 500 g/sec from the storage dewar will be transferred to the cavity

cryomodules. The flow rate will be governed by the refrigerator's ability to reliquify the boil-off gas.since the warm gas storage is insignificant compared to the liquid volumes.

When the pump is at full speed the cold compressors are speeded up and the cryomodule pressurereduced. the speed again being limited by the reliquefaction rate. The maximum rate is limited by boththe cold and first-stage warm compressors to a rate of 9 g/sec per cryomodule; the duration of the processwill be 8 hours.

The total duration of the start-up and cooldown process is about two days and is not unreasonablefor periodic warmups with our projected one-year interval. The details of the warmup procedure havenot yet been determined. but one-day warmup can be accomplished by a combination of breaking vacuumand circulating warm gas. This turnaround time would be acceptable during an initial shakedown periodif warmup and cooldown were required once a month. Individual cryomodules can be removed from thesystem and from the linac tunnel while maintaining the system in a state of readiness. The fill andpumpdown of a precooled cryomodule will take much less time than the mechanical reassembly thatmust accompany a cryomodule replacement.

9.9 System Stability and Helium Supply Regulation

The CEBAF cryogenic distribution system has 50 IT valves that demand a portion of the heliumsupply based upon the maintenance of liquid level in the cavity cryomodules. The valves modulateindependently and respond to an independently changing demand with a supervisor program that preventsoverloading the cold compressors. The cold compressors regulate the 0.031-atmosphere pressure in thecryomodules. with a one-minute time constant. The first-stage screw compressors regulate the 1.05­atmosphere suction. The second-stage screw compressors regulate the discharge pressure. which is ad­justed to match the load requirements. Interstage pressure is regulated by the inventory control.

The supply liquid system will be kept stable at 2.8 atmospheres by pumping 10 to 20 g/sec from thedewar and dumping the unused liquid back into the dewar subcooler. point 28. This subcycle producesa heat load less than 2 Joules/gram; in addition a 3 kW heater located in this line will be used duringthe commissioning phase to deal with the varying RF heat load.

An additional subcooler will probably be added and tied to either the inlet or outlet of the C2 coldcompressor. This will improve the cycle efficiency, but more importantly it will help the cryomodulepressure stability with respect to imbalanced flows. Table 9.7 is the current list of possible system upsetsand their remedies. This list will expand as the system design progresses.

Table 9.7Upset Conditions

Condition

Single-cavity RF shutdown

Single-cavity higher heat load

Single-cavity extremely highheat load

Cryomodule high heat load

Inability to excite RF

Leak to insulating vacuum

Loss of insulating vacuum

Cause

Varied (contamination.electronic, etc.)

Local impurity; local low Q

Above

Above (not restricted to singlecavity)

Poor cavities; electronic

Small leak

Major leak

(Continued on next page)

Cryogenic consequence

None

Larger flow required incryomodule

Tum off RF

Higher flow of He reqaired

RF off; He flow reduction

Apply extra pumping; increasehelium flow

Bypass cryomodulecryogenically

155

I

CEBAF Design Repon

Condition

Loss of beam ....acuum

Loss of vacuum

JT ....alve stuck open

IT valve stuck closed

156

Table 9.7 (continued)Upset Conditions

Cause

Internal leak

Unpumpable leak

Mechanical or control failure

Above

Cryogenic consequence

Isolate cryogenically and pumpout He spaces

Replace cryomodule

Valve design limits flow totwice normal

Cryomodule will warm up;valve may be serviced inplace

• ••

10. Experimental Equipment

10.1 Introduction

To give as much time as possible for the latest scientific developments to be taken into considerationin planning the CEBAF experimental program, the design of the CEBAF end stations and conceptualdesigns for the major pieces of experimental equipment will not be frozen '.:ntil the summer of 1987.Planning for the experimental equipment is, therefore. phased a year and a half behind the planning forthe accelerator itself. Therefore the plans and designs presented in this chapter are preliminary.

The major pieces of experimental equipment discussed in this chapter are the magnetic spectrometers(Section 10.2). the Large Acceptance Detector (LAD) (Section 10.3), the Variable Acceptance Spec­trometer (VAS) (Section 10.4), and targets (Section 10.5). The section on spectrometers is largely thework of three contributors to the CEBAF 1985 Summer Study Group (I. Blomqvist. J. Mougey. and R.Neuhausen). The chapter concludes with a short discussion of the proposed layout of the experimentalhalls, and of possible plans for future upgrades of these halls (Section 10.6).

While many of the ideas presented in this chapter were discussed during the 1985 Summer Study.this is the first time a comprehensive plan for experimental equipment for CEBAF has been assembled.Thus, this plan has not yet received the extensive review necessary before conceptual designs are frozen.These ideas are being discussed extensively with the scientific community. which will playa major partin developing the final conceptual designs and the plans for the research program. The CEBAF 1986Summer Study will provide one opportunity for significant community participation in this effort.

The experimental equipment complement and re~a:.chplans will be reviewed fonnally by the PhysicsAdvisory Committee. The first such review will occur in 1986, and a second review will take place beforethe conceptual designs are frozen. CEBAF welcomes any comments on the designs and approachesdescribed here. and invites interested physicists to collaborate in the design and eventually the constructionof the experimental equipment.

10.2 Magnetic Spectrometers

Physics RequirementsFor nuclear studies, the electromagnetic prot>.:: is characterized by its ability to provide precise

quantitative information, owing to the well-known and relatively weak electromagnetic interaction. Forthis reason. high-precision instruments like magnetic spectrometers make up the major part of experi­mental equipment at electron accelerators; thus the CEBAF physics program will rely heavily upon theexcellence of its magnetic spectrometers.

An important part cf the experimental program defined in Chapter 2 consists of exclusive experimentsof the type

e + A - e' + A-

e + A - e' + x + B*e + A - e' +x+y+ C-

157

• I I • • I I ' r .-

I

CEBAF Design Rcpon

in which particles x and yare detected in coincidence with the scattered electron e'. When the finalundetected nuclear system A *. B* and C* is left in a bound state. the complete identification of theprocess requires a high-precision determination of the energy and angles of alI other particles. TypicalIy.an overall resolution of approximately 100 keY is needed for heavy nuclei. and about 1 MeV is necessaryto separate the bound deuteron from unbound n-p pairs in reactions involving two- and three-nucleonsystems.

For experiments in which the final undetected nuclear product is either an elementary particle or acomplex nuclear system in a continuum state. such very high resolution is not needed. However. con­sidering the need to keep the value of the true-to-accidental ratio reasonably large. and the way in whichelectromagnetic cross sections vary with the kinematics. an energy definition not worse than about 5MeV is most often desirable. Examples of such experiments are:

• (e,e'p) reactions on deeply bound nucleons

• (e.e''7l') experiments in the .:l region

• d(e.e'p)n reaction

• 3I-1e(e.e'2p)n reaction

Keeping in mind t:le physics requirements outlined in Chapter 2. a family of spectrometers wasdefined and preli:;;i;:ary designs were developed during the CEBAF 1985 Summer Study meeting. Themost critical parameters determining the quality of a magnetic spectrometer are listed bdow:

• maximum momentum

• momentum acceptance

• momentum resolution

• solid angle

• angular resolution

• angular range

Values of these parameters for the preliminary spectrometer designs are listed in Table 10.1. Thedesigns were done according to the folIowing specifications:

• The family of CEBAF spectrometers should include high-resolution spectrometers of maximummomentum equal to 4. 2.5. and 1.2 GeV/c. as welI as moderate res91ution ones reaching 4 and2.5 GeV/c.

• AlI high-resolution spectrometers should have a momentum resolution of 10-4 or better. Moderateresolution should correspond to 10.3 or better.

• The 4-GeVIe spectrometers should be able to reach forward angles of the order of 10°.

• The high-resolution 4-GeV/c spectrometer should have a solid angle of 5-10 msr. The high­resolution 2.5-GeV/e as welI as the 4-GeV/c and 2.5-GeV/e moderate-resolution ones should havesolid angles of 10-20 msr. The 1.2-GeV/c spectrometer should have a solid angle of 30-40 msr.

• The momentum acceptance should be about 10% for. the high-momentum spectrometers and 15­20% for the 1.2-GeV/e one.

• Some spectrometers should be able to handle liquid or gas targets that extend 10 to 20 cm alongthe beam direction.

• Some detectors should be able to measure particles emitted at large angles (greater than 45°)from the horizontal plane.

General design considerations as welI as a brief description of the present designs are given in thefollowing pages. In addition to magnetic devices. a large scintilIation hodoscope is also described. Thiscould be used in coincidence experiments when large solid angles are necessary. and when a iesolutionof about 20 MeV is sufficient.

158

Experimental Equipment

Table 10.1Preliminary Design Parameters for CEBAF Spectrometers

Maximum ADguIar Length ofSpectrometer ceotra1 Momentum Momentum resolution ceotral

IIIUIIe IDOmentum acceptaDc:e resolution Solid aDgIe (e.~)· (e.~) trajectory Weight(GeV/c) (~/p iD %) (Op/p) [msr. (mrad. mrad)) (mrad) (meters) (metric

lons)

~GeVlc

reimaging ~ 10 ... 10-' 10.8 (60. ISO) 0.5.0.5 35 880

~GeVlc

thin target ~ 10 ... 10-' 10.8 (60. ISO) 0.5.0.5 23 550

~GeV/c

moderateresolution ~ 10 ... 1O-~ 10.8 (60. ISO) 0.5. Z.O 15 480

2.5-GeV/cthin target 2.5 10 ... 10-' 20 (SO. 250) 0.5.0.5 23 550

2.5-GeVlcmoderateresolution 2.5 15 ... 1O-~ 20 (SO. 250) 0.5. 1.5 12 :!SO

1.2-GeVlc 1.22 20 ... 2 x 10-' 35 (160. 250) 1.0. 1.0 15 700

"9 refers to the deflection plane a.ld ~ to the transverse plane.

General Design ConsiderationsThe design of a spectrometer is an iterative process, involving complex trade-offs to maximize

performance and capabilities while trying to minimize cost, weight. size, and complexity. The specifi­cations listed in the preceding paragraph are quite extreme when compared to high-energy spectrometersbeing used at places like SLAe. DESY. and Bonn. On the other hand. extrapolation to 4 GeV of low­energy. high-resolution designs (MIT. Saclay, NIKHEF-K) would be too expensive, so original solutionshave to be developed.

There are major advantages in designing spectrometers for a CW supercondueting linac. The mainadvantage is that the much smaller energy spread means that 10-4 resolution can be reached withoutresorting to complicated dispersion-matching techniques. Also. the improved beam emittance leads tosubstantial savings on multipole elements needed to correct for higher-order aberrations. It also permitsthe use of spectrometers which bend in the horizontal plane and which have a considerably lower totalweight because they require a smaller dipole aperture.

Some of the considerations which led to the present iteration of spectrometer design are discussedbelow.

"Software" vs. "Hardware"Modem magnetic spectrometers require complex detector systems to completely determine the

particle trajectory. In principle, use of these systems in combination with high-speed computers makesit possible to correct any medium-quality spectrometer for aberrations if each particle trajectory isdetermined ("software" option). However. the "hardware" option of designing the !:lest possible magneticsystems was chosen for these designs so that the software corrections could be used to achieve higherresolution or to deal with long gas targets.

Horizontal vs. Venical BendingSome of the spectrometers deflect the particles horizontally. others vertically. Horizontal deflection

is best suited when it is necessary to have the lar~est acceptance in the vertical plane and keep the totalweight as low as possible. A practical rule of thurr.b is that the ideal verticallhorizontal angular acceptanceratio is 3/1 for a horizontally bending spectrometer. Moreover, for large systems, the shielding andsupport structure is easier to build in a horizontal configuration. However. this configuration has twomajor disadvantages: it mixes the measurement of the scanering angle with the determination of the

159

I

I

CEBAF Design Repon

particle momentum. and it occupies a larger floor space. reducing the angular range and increasing theminimum possible angle between two spectrometers around the same pivot.

Reimaging SpectrometersThe complete determination of the position of the particle trajectory in the field-free region behind

the spectrometer (4 parameters) is in general not sufficient to determine where it originated in the target.as the particle momentum enters as a fifth parameter. However. when working with thin targets andsmall beam sizes. one can assume a pointlike source of particles. which requires only 3 parameters pertrajectory. If the source is extended in only one direction. a complete reconstruction is still possible ifthe bend is in the other one. In the most general case. which occurs when using a horizontally bendingspectrometer with a long target. the position of the trajectory must be measured in another place todetermine the fifth parameter. To minimize dispersion effects due to multiple scattering in this inter­mediate detector. it has to be located at an intermediate focus. This leads to the so-called reimagingspectrometer design. In principle. when performing a coincidence experiment. only one instrument ofthis type is needed. The extracted information giving the location of the event vertex can then be appliedin software corrections to mca5ured trajectories in the other spectrometers. The disadvantages of thereimaging spectrometers are their length. cost. and complexity.

ModulariryThe general approach taken here has been to build spectrometers made up of modules. in order to

reduce costs and allow for possible further reconfigurations. All dipoles. quadrupoles. and multipoleswould be slight variations of a few standard designs. This would also make future upgrading easier­like turning the moderate-resolution spectrometer into a high-resolution one--or expanding a spectro­meter to accommodate a higher maximum momentum.

Magnet TechnologyTo ensure the field homogeneity needed to meet the high resolution requirements. iron-dominated

dipole elements have been chosen. The use of superconducting coils would permit dimensions and weightsto be minimized. but conventional copper coils have other advar.cages. and both solutions are underconsiderations. For the large aperture. highfield gradient quadrupoles. use of superconducting technologyis likely to be the only possible solution.

Figure 10.1 shows examples of the modular elements for the family of 4-GeV/c and 2.5-GeV/cspectrometers.

The dipole elements have rectangular pole pieces. no curved boundaries. and the field is uniform(parallei pole faces). Maximum field at the pole tips is assumed to be 1.7 T. Coils are saddle-type tominimize the size of the yoke. Typical deftectiol! angles will be 15° at 4 GeVlc and 24° at 2.5 GeVlc tokeep the length of the dipole to about 2 m.

Quadrupoles will have rectangular aperture shapes following the originai design of Hand and Pan­ofsky.(1) Maximum field close to the coils will be on the order of 2 T. The return yoke will be enclosedin the cryostat. Although such large superconducting quadrupoles have never been built. reasonableextra polations can be done from existing ones. like the Satume II superconducting quadrupole. (2) Thisone has a 40 x 16 em useful aperture with 2.8 T at the coils. Front quadrupoles will have to be shapedto allow the spectrometers to be used at small scattering angles.

Multipoles of higher order could be of the type developed by Ikegami et at. (3) Maximum fields ofapproximately 1 T at the coils have been assumed.

Current Spectrometer DesignsThe layouts for each of the spectrometers mentioned in Table 10.1 are shown in figures 10.2 to 10.6.

Their designs are based on first- and second-order beam optics calculations performed during the CEBAF1985 Summer Study program.

• The high-resolution. thin target. 4-GeV/c and 2.5-GeVlc spectrometers (Figure 10.2) bend hor­izontally. They consist of an essentially symmetric QQDDQQ configuration. with the particletrajectories parallei to the spectrometer axis in the central region. Point-to-point focusing withunit magnification is realized in both directions. Second-order angular aberrations are completelycorrected. The focal plane is tilted to approximately 50° from the central trajectory. allowingprecise momentum reconstruction. Note that if the energy resolution of the initial beam were10'3. the design would have been more complex (leading to 850 tons. 33 m ir.stead of 550 tons.

160

I I

.. I I L

Experimental Equipment

180

~.-.----160----~

Quadrupole

Dipole

250

I.......t--------- 400 ----------1

FJgIII"e 10.1. Typical dimensions (in en) of modular elements for the 2.5-GeV/c and 4-GeV/c spectrometers. Fr<:':It elements willhave to be shaped to allow the spectrometer to reach small forward angles.

23 m for the 4-GeV/c design). The reason is that in order to achieve about 10-4 resolution, thebeam would have to be vertically dispersed ~r. target ("energy-loss mode"). Moreover, moresecond-order aberration corrections would have to be made, due to larger spot-size on target.Variations of these basic designs can be easily developed. For example, momentum dispersionmay be traded for angular resolution by reversing the polarities of the last two quadrupoles. Notethat these two spectrometers are built from the same basic modules.

161

r

I

CEBAF Design Repon

4 GeV/c

2.5 GeV/c

01

012345m

I

FJgIU'e 10.2. Design of the 4-GeVlc and 2.5-GeVlc thin target spectrometers.

• The reimaging 4-GeV/c spectrometer (Figure 1O.3) is certainly the best solution for dealing withextended targets and/or extended beam spots. as it allows a complete reconstruction of the pointfrom which the particle left the target (the event vertex). It has an intennediate focus in thebending (horizontal) plane. where additional counters can be 10cated.This configurdtion minimizesthe effects of multiple scattering in these counters on the momentum resolution. Complete tra­jectory reconstruction. which implies some software iterations using the information from the twosets of counters. allows definition of ··software collimators" (for example. to avoid contributionsfrom gas or liquid target walls) and correction for the "depth of focus" problem that will be quitesevere at very forward angles when using long targets.

162

M1

012345m

Drift Chambe~ 02

~~.f8=~'::J:. .• r.--r-.o

FIgUn 10.3. 4-GeV/c reimaging spectrometer.

Experimental Equipment

• The 4-GeV/c and 2.5-GeV/c moderate-resolution spectrometers (Figure 10.4) are QQDD andQQD devices which bend vertically. They have parallel-to-point focusing in the horizontal planeand point-to-point focusing with a magnification of 5 to 7 in the vertical plane. The momentumdispersion is slightly more than 1 cmI%. They are relatively simple spectrometers with no second-

4 GeV/c moderate resolution

01

2.5 GeV/ c moderate resolution

o 1 2 3 4 5m

~~I \I \I 0'\ \I \

Figure 10.4. The 2.S-GeV/c and 4-GeV/c venicaJ bending moderate-resolution spectrometers. They are built from the samemodules used for their high-resolution cousins.

163

I

CEBAF Design Repon

order hardware corrections. The resolution of about 10-3 comes from software corrections andthe good qualities of the CW beam. The solid a..rlgle is eventually determined by the dipole gaps.With some loss in resolution. these spectrometers can be upgraded to higher momenta by simplyreducing the bend angle.

• The 1.2-GeVlc spectrometer (figures 10.5 and 10.6) is a vertically bending system similar toexisting high-resolutio;) spectrometers used at lower energies. It has a unique combination of verygood momentum resolution a.,d very large solid angle. The first-order resolution of 3 x 1O-s

(- 40 keY) is calculated assuming a beam spot size of 0.05 em. The resolution may approach10-0; due to the small beam emittance of the CW design. The solid angle is 35 msr. This spectrometerdoes not use the "standard" modular approadl of the other designs.

rA

QQQ.!2.Lenglns In em

---------------- -

ShIelding house

60 em coner.leS em Doron layer~ em 1••C1

..,B

.-- 02 : I. ,l-----f--,--~- _l_

e :II,,,

rJgW'e 10.5. Side view of the I.2-GeVlc high-resolution spectrometer.

01-f---

C>

01 02 •I ".. -.h_.--------__,.....,...--1-L-l- 100~ 60 ~ so -: 60~30 G N

: _1_]_~;~~:::-- -- -,

ITa,get ... :::,'-::----......... -.... - -­

...................---............... --

................... - ..... - ----,5-..... ---

................... - ...............__ ..... 20·-,

' ......... 25·

rJgUn 10.6. Section C (with reference to Figure 10.5) of the 1.2-GeVlc high-resolution speetromet..:r.

164

I I

I I I

Experimental Equipment

• A Large Acceptance Scintillator Hodoscope (LASH), which could be used in conjunction withthe spectrometers, is shown in Figure 10.7. The basic building block is a 0.2 x 0.2 x 2 meterplastic scintillator bar with 5-inch photomultipliers at both ends. Although it has only a moderate

Counter C

Helium bag

Target tE~-----_+H---------------------+_+_+_-

Ii s.75m

I

1111......------------15 m --------------­I

I

E30 elements0.20 x 0.20 x 2 m

scintillation counterst 5" h tit' r

~

30 elements0.20 x 0.20 x 2 m

scintillation counters+ two 5" photomultipliers

H~

+ wo p oomu Ip lers

30 elements0.20 x 0.20 x 2 m

scintillation counters+ two 5" photomultipliers

1.-1------ 6 m----~~

Counter C

~1.S m--i

TE

Counter B31) elements

0.05 x 0.01 x 1.5 mscintillation counters

+ two 2" photomultipliers

Counter A30 elements

0.05 x 0.01 x 1.9 mscintillation counters

+ two 2" photomultipliers

FJIlIR 10.7. The Large Acceptance Scintillator Hodoscope (LASH).

165

CEBAF Design Report

resolution of about 5 x 10-3 (for protons of 800 MeVlc and a 15-m flight path) and will requirereduced beam currents of 1 J,LA or less to operate effectively. its large solid angle and versatilitywill be extremely useful. Moreover. it can be used in experiments like (e,e'n) or (e,e'pn) reactionsin which neutral particles need to be detected. A particular arrangement suitable for chargedparticle detection is shown in Figure 10.7. In this setup. thinner scintiIIators have been added infront to provide trajectory localization and independent time-of-flight measurement.

10.3 Large Acceptance DetectorPhysics Requirements

A significant fraction of the CEBAF experimental program will require a Large Acceptance Detector(LAD). Some typical experimental problems that are best handled using an LAD are:

1. The detection of multiple-particle final states (high detection efficiency and a model-free analysisfor these events can only be achieved by detectors with a complete coverage of the angular andenergy range for all outgoing particles). Such reactions include (-y = real or virtual):

• Excitation of the higher nucleon resonances as discussed in Section 2.4 including-y N -+ N" -+ .. .1. -+ .... N

-+Np-+ .... N-+NW-+1T'!r1TN

• Reactions on the deuteron"{d-+NN1T

-+ .1. .1. -+ N N 1T 7l'

• Reactions on nuclei-yA-+NNX

-+N"X-+ K A" X

To illustrate typical requirements for the LAD, Figure 10.8 shows a scatter plot for p.,.. vs. 8.,..for pions from the decay of a photoproduced F3s(l975) nucleon resonance into two pions and anucleon. All of the available phase space is filled. Due to the Lorentz-boost the forward-goingpions have (on the average) higher momenta than backward-going pions.

2. Measurements at limited luminosity (target density x beam intensity). This limitation can bedue to the target (rare or dangerous elements, active targets, polarized targets with low radiationresistance) or due to the beam:

• Limitation due to the beam intensity. All programs using a tagged bremsstrahlung beam (in­tensity = 107

-y/sec) fall into this category, independent of the number of panicles in the finalstate.

• Limitation due to the use of a polarized target. Every experimental program involving ahydrogen or deuterium target will require at some stage the use of a polarized target. Especiallyimportant are the measurements of the charge distributions of the neutron and deuteron,described in Section 2.4.

The general properties of a Large Acceptance Detector (necessary for a broad range of experiments)are listed belnw.

• Homogeneous coverage of a large angular and energy range for charged particles (magneticanalysis). photons (total absorption counters), and neutrons.

• Good energy and angular resolution (for all particles).

• Good particle identification properties.

166

r

Experimental Equipment

1.2 ~----T"""----'T""----~----..,..-----r----....

o->CDo

+I:

Q.

0.9

0.6

0.3

· - ----.- - . . --. --· --....-. .-._._ ....----_ .. -.-- ..•.••...... -._._ ..... ­..•.•••••......-.- -.-.--- -•••......-_ .•.••........••••••••••••••••.......••.._._-.­......••.•........._-.......•.......-- ......•.............•.•..•...... __ ..•••••••••••••••••••••...........-_...-..........•.••!E•••.........•••••••• • •••••••••••••········1 - ........ -- .•.•..........._._.- .•.................. . --_•••...... - ..- __ .a•• _ _ _ •••••••••••.........•.•. ._---••••......... ::::::::= • • =-===::::::::::::::.-.......••••• -- ...•.............................•---_ _--- -_ .. __ .•••••••••••••••••••••••••••••••••••••••••••••••· •••••••••••a ••••••••••••••••••••••••••••••••••

••••••••••••• •••••••••••••••••••••••••••••••· . . . . .. _.. --. -. . . . . . . . . . . .0.0 30.0 60.0 120.0 150.0 180.0

Fagure 10.8. Monte Carlo calculation for pion momentum (vertical axis) versus pion Ial:'oratory angle (horizontal axis) for '11'+from the reaction"'t + P -.. F),(l97S) - '11'- -+- ~ + + - '11'- + '11'+ + p. The size of the boxes is proportional to the Dumberof particles per bin.

• No transverse magnetic field at the beam axis (to avoid sweeping (e+ ,e-) pairs into the detector).

• Large field-free s~ace around the target to allow for the installation of complicated targets (cry­ogenic, polarized, track sensitive, etc.) or auxilliary equipment (polarimeter, 'Y-converter).

• Symmetry around the beam axis to facilitate event reconstruction.

• Larger f Bdl for particles going forward than for particles going sideways, to account for Lorentz­boost.

• High count rate capability: the detector will be required to work in a tagged photon beam(N~ • 10

7per sec) or a weak electron beam (e.g., 10 nA on a 0.1 g/cal target - luminosity ==

103 /cm2-sec).

Description of LAD

The requirements listed above can be satisfied by using a toroidal detector consisting of eight coilsarranged around the beam line to produce a magnetic field essentially in the 4Klirection. Charged particlesare tracked by drift chambers; scintillation counters are used for the trigger and for time-of-ftight; photonsare detected by shower counters. A schematic figure of the detector is shown in Figure 10.9. A d~-ription

of its main features will be given below.

167

CEBAF Design Report

Shower counters

Drift chambers

Beam - Target n ~ -.--- n ~ .--

I ( 4m >/

figure 10.9. Design of the LAD described in the text. An end view (cross section perpendicular to the beam direction) is shownon the left. while a side view (cross section parallel to the beam direction) is shown on the right.

Magnetic FieldThe magnetic field is generated by a set of 8 superconducting coils. Each coil carries a current of

500.000 amps. This results in a maximum field of 0.75 tesla; the f Bdl for particles emitted at 90" is about0.7 tesla-meter. The radial dependence of the et> component of the field is sketched in Figure 10.10. Thestrong forces that pull the coils towards the axis (about 50 tons/coil) can be handled by using a coldsupport ring around the outside of the coils. The outside dimensions of the vacuum chamber are 4 m(length) by 4 m (diameter). About 20% of the c!>-ra.:lge is obstructed by the vacuum chamber.

Tracking ChambersIn each of the eight segments, charged particles are tracked by six planar wire chambers consisting

of several layers of staggered sense wires. The segments are instrumented independently to achieve highcount-rate capability. The estimated momentum resolution is better than, or about, 1% in the momentumrange of interest; the initial et> and e of the track can be determined to better than 5 mrad.

Scintillation CountersThe outer planar drift chambers are cC'/ered by scintillation counters. The barrel counters consist

of 8 x 8 counters, each 400 em long, 20 em wide, and 5 em thick. The counters are viewed by 2-inchphototubes at both ends for improved timing and position resolution. Both endcaps are covered by8 x 4 counters, each viewed by one photomultiplier. The scintillation counters serve the double purposeof providing the trigger and the time-of-ftight information. Also, a fraction of the high-energy neutrons(= 5%) win interact in the outer scintillation counters and will thus be detected.

Shower CounterThe detector is surrounded by shower counters for the detection of high-energy photons from

Compton scattering, and 11'0 and TJ decays. Due to the size of the counter ( = 80 m2, = 100 tons)

inexpensive material and construction techniques have to be used. The typical energy resolution of thesedevices is erIE < 0.15/'VE (GeV); the position resolution is typically a few centimeters. The showercounter also helps to identify charged particles; high-energy electrons (that are difficult to separate frompions using time-of-ftight) will show up especially clearly in the shower counter. To avoid dead spaces,the inner side of the vacuum chamber can be covered by a high-density photon detector (e.g.• 10 radiationlengths of BGO with photodiode readout).

168

Experimental Equipment

Toroidal field0.80 r--------,:-------,-----""T""----""T"""----r-------,

3.02.52.01.5Radius (m)

1.00.50.00 L __::::........L.. ......L .L...- ....&- ....::::::L:::::::::=:::._-.J

0.0

0.20

coCf.Ien

0.60

~0.40CD

Fagure 10.10. Radial dependence of the cb component of the magnetic field in the LAD.

Particle ldentificalionThe combination of momentum and time-of-Bight is used for the identification of charged particles.

As demonstrated in Figure 10.11. pions and kaons can be separated up to 1.5 GeV/c, kaons and protonsup to 2.5 GeVlc. By using the pulse height in the shower counter in addition, 1r/~ and ~e separationcan be achieved.

The perfonnance of the LAD can be predicted using Monte Carlo methods to generate multiple­particle final states and to test particle tracking and reconstruction. As an example, the probability ofbeing able to reconstruct the complete 1t' + 1t' - P final state from the F3s (1975) decay has been found tobe 30% (most of the events are lost because one of the three particles hits a coil). Figure 10.12 showsa Monte Carlo generated event from the reaction

as it would be reconstructed by the detector.

The LAD will occupy the lightly shielded end station B. It can be used in combination with a low­intensity electron beam for electron-scanering experiments or with a real tagged photon beam for pho­toproduction experiments. The bremsstrahlung tagging system will typically cover the photon energyrange Ey = (0.5 to 0.95)·£ with an energy resolution of t:1E-y = 5 MeV.

169

I I

CEBAF Design Repon

Time difference2.5 ..----r----'r--................,....---~-__..__._...._-_r_-----_,_-----___,

----

.....-....

p

I,,",I",I", ,~ ~ "..I II I

• II II I, I,I ,, I, I, ', ..

I I, "I I, .

I •

I '.I '.I '.

...... ..........._-- ......................... _- ...._-- -......

-----------------.-::::~,=~~-------------~.-::-~--~-~-~---~:J--.--...... -------=:.:.:.... ......... - - -............ - ......

.............- __ _----------_ .. _-_ _-_ _----------

2.0

0.0

1.5

0CDCI)

c: 1.0.....<J

0.5

2.52.01.51.00.5

- 0.5 t....- ---L ........... ..L.-. ----JI-.... --I

0.0p (GeV/c)

Fig1In 10.11. Time-of-flight differences between ~= I panicles and pions. kaons. and protons (venical axis) for a 3-meter flightpath as a function of momentum (horizontal axis). The error bands correspond to a 500 psec (FWHM) resolution.

#1

ITarge. :J I~ '------

Event #12

P Ep 8 (/J

TT- 0.44 0.32 74.5 190.07T+ 0.70 0.57 37.9 324.0proton 0.99 0.43 19.5 78.0

I • J

_I 1,-------,

FIgure 10.12. Singie-event display of a Monte Carlo generated event for the photocxcitation of the F,,(1975) baryon resonance.as described in Figure 10.8. The energy of the photon is 1.6 GeV. The left-hand side shows a view in the direction of the incidentbeam; the right-hand side shows the eight individually instrumented segments of the detector. presented in counterclockwise orderstarting with number one.

170

..

I I I • - .-Experimental Equipment

10.4 Variahle Acceptance SpectrometerMotivation

An appreciable fraction of the experiments which have been proposed for CEBAF do not requirethe use of very high resolution spectrometers. but can be performed with spectrometers having moderateresolution only. These experiments include. among others. the following measurements:

• the one- and two-nucleon knockout processes.

• form factor measurements of deuterons and nucleons in a coincidence arrangement.

• the production of single pseudoscalar mesons (pions. etas) from free nucleons.

• the electro-disintegration of the deuteron.

• the coherent pion and eta production from A =2,3 nuclei.

In most of these experiments it is generally desirable to detect and measure the outgoing hadronsin a large solid angle and a large momentum range, in order to cover an appreciable frae-..ion of thehadronic phase space. To separate the various contributions to the coincidence cross section, it is necessaryto perform "out-of-plane" measureIr.ents where one of the outgoing hadrons is detected at azimuthalangles which differ considerably, from 0 or 180 degrees.

The general layout of the detector is shown in Figure 10.13. For various reasons it appears highlydesirable to keep the spectrometer as flexible as possible. Some of these reasons are listed here:

• Operating the detector with polarized targets in low-intensity beams (1 to 50 nA) as well as withlong liquid or gas targets at moderately high intensities (several ~). In the first case, where alarge acceptance is desirable, it will be possible to operate the detectors with direct view of thetarget. In the latter case one could eventually live with a smaller solid angle, but the variousdetector elements (in particular. the scintillator hodoscopes) would have to be properly shieldedto screen out the background of photons emerging from the target.

• The performance of out-of-plane measurements at reasonably high luminosities is of great im­portance for many experiments. in particular for studies of nucleon resonances, deuteron disin­tegration and two-nucleon correlation experiments.

• Many experiments require the use of polarized solid state targets. Associated with these targetsis a high magnetic field which may have to point in any arbitrary direction to create the appropriatepolarization axis. Since the integrated field length may be as large as 0.75 T-m, serious deflectionsof the particle trajectories may occur which may prevent low-momentum particles from beingdetected in the spectrometer.

It is planned to allow the magnet to be movable in all three dimensions. The first drift chambershould be attached to the magnet yoke, whereas the remaining chambers and the scintillation hodoscopesshouid-be kept moveable. It is also important to have the possibility of changing the relative position ofthe DO-HI H2 package with respect to the H3 H4 one.

AcceptanceThe solid angle coverage and momentum acceptance depends on the actual experimental setup

chosen. For the standard arrangement (distance of 1 meter from target to magnet entrance. B = 1.5tesla) the solid angle is about 50 msr for momenta between 0.75 and 2 GeV/c. At 3 GeV/c the acceptedsolid angle is still approximately 25 msr.

Particle IdentificationThe detector provides charged particle identification up to very large momenta by combining mo­

mentum measurements and time-of-flight measurements (Figure 10.14). For the purpose of electrondetection. a long gas Cherenkov counter could be installed behind the last scintillator hodoscope.

Photon and neutron detection is provided over the full solid angle using the scintillatorllead-glasssetup. Both particle types can be separated by "investigating the energy deposition in the front and rear

171

I I I

•CEBAF Design Rcpon

eTarget

1 meter

Pb-glassarray

Side plate

Superconductin gcoils & cryostat

H3 H4

Scintillatorarray

Superconducting coil(mirror plate removed)

Frontmirrorplates

Beam

\

F"JglIR 10.13. Thc Variablc Acceptance Spectromctcr. Top: sidc vicw. Bottom: front view (Icft) and top vicw (right).

pans of the detector. The longitudinal subdivision enables a good separation between electromagneticshowers (induced by photons or electrons) and signals induced by interacting hadrons. At sufficientlylarge energies. the probability for detecting photons will be close to 100%. and for neutrons. greaterthan 60%.

For the scintillator array and for momenta above 1 GeV/c the effective solid angle for measuringthe nO via two photon decay increases with the momentum of the pion.

Trigger CapabilitiesThe employment of detectors with a large momentum acceptance at high current electron machines

may bring about serious dead-time problems for measurements of rare processes. These problems may

172 ., I

I I -Experimental Equipment

7

6 '\0 '\G)

\ '\(/) \c: \ \ '\ '\. Deuterons5 \ \ \ '\ "G)

\ \ \ \ '\

"0 '\c: \ \ \ \ "G) \. "..4 \ \ \ '\ " "G)- \ \ \ \ Protons " "-i5 \ \ '\ '\ " " ......

\ \ \ " " .....u. 3 \ " ......0 \ \ ......

"......

f- \ '\ '\ ...... "..........

\ Kaons...... ......

\. ..... .....2 '\ " " ..... ....

'\ " "- - "-Pions '\...... ...... --- ....

"" " ...... _- ' ..... -----1 , .... " ---- ~--- ...... .......,:::..-----.... _-,-- - --- .... --....... ---- -0 1 2 3 4

Momentum ( GeV/c>

FJgUre 10.14. Time-o!-ftight difference compared to ~= 1 particles. A flight distance of approximately 6 meters fro:n the targetto the last scintillation hodoscope is assumed.

be overcome. to some extent. by employing evolved fast trigger systems at the hardware and at thesoftware level. This. ofcourse. requires a detector system with reasonably high granularity. The scintillatorhodoscopes. for example. may be used to select events within a given momentum bite. Such a ""momentumtrigger" may be realized by requiring hits in certain combinations of hodoscope strips. This could bedone either at the fast electronic level. or in a later stage at the microprocessor level. Another possibletrigger specification could be the selection of certain charged particle species, like pions, protons, etc.,by investigating the time-of-flight information of the hodoscopes and the hits in the drift chambers.

10.5 Targets

The technology required for the different types of targets--sOlid, cryogenic liquid, high-pressure,and polarized-to be used at CEBAF has been largely developed in the past. In panicular. techniques .for high rates of heat extraction by forced cooling, movement of the target, and rastering of the beamhave been established. an~ can be used at CEBAF. This section discusses polarized targets, which areof particular interest to CEBAF. in more detail.

Some of the most fundamental experiments to be conducted at CEBAF require polarized electronsand polarized hydrogen and deuterium targets. The other option of internal targets in the storage ringwill not be available with a CW linac. at least in the initial phase of the project.

The existing, state-of-the-art designs for polarized hydrogen or deuterium targets are based ondynamic RF spin coupling to induce nuclear polarization. While there are a variety of materials thathave been used. NH3 (ND3 ) is now preferred. The major problem with these targets is beam-inducedradiation damage and background counts from the nitrogen atoms.

The polarization of the sample can be increased by lowering the temperature. Unfortunately. theavailable cooling power also decreases with temperature. and :his limits the usable beam current. Theuniformity of the magnetic field over the sample volume should be kept within a tolerance of l1BIB lessthan lO-a· There are superconducting coil arrangements that produce fields of 5 T with l1BIB less than2 x 10-5 over volumes of greater than 6 em3

•

173

I

CEBAF Design Repon

II •

II

Table 10.2 lists some average characteristics that have been achieved in existing systems.

Table 10.2Characteristics of Existing Polarized Targets

NHJ ND3

Temperature (K) 0.25 to 1 0.25 to 1

Magnetic fields 2.5 to 5 2.5 to 3.5(tesla)

Target density 1023 1023

(H or 0 nUclei)(cmo2)

Maximum current 2 x 109 to 3 X 1011 2 X 109

to 3 x lOll

(electrons/sec)

Luminosity 2 x 1032 to 3 x 1034 2 X 1032 to 3 x 1034

(cmo2

secol

)

Polarization 0.4 to 0.8 0.07 to 0.5

Tensor polarization 0.2

The luminosities of existing targets are already adequate for the physics program using polarizedexternal targets. It is possible to improve upon the performance by lowering the temperature whilemaintaining a reasonable cooling capacity by using a ~elHe dilution refrigerator with a very highpumping speed.

The use of external polarized targets of hydrogen and deuterium for the CEBAF research programis clearly viable. In fact. for several experiments it would be the preferred experimental setup. even ifinternal targets were available.

10.6 Layout of the Experimental Halls

The present preliminary design of the two high-current experimental halls A and C is shown inFigure 10.15 and in Chapter 11. These halls share a common wall. which is sufficiently thick (3 m) toallow either hall to be occupied safely while the other is receiving beam. This configuration maximizesthe useful experimental area available. Both halls have one wall built of removable blocks to allowexpansion in the future.

The location of the target and the angle of the beam through end station A were chosen so thatthis larger hall could house the 4-GeVic high-resolution thin target spectrometer to the right of the targetwithout placing restrictions on its range of angular motion. The hall contains a pit to hold the 1.2­GeVlc spectrometer. which can move through a full circle· about the target. A tOP view of these twospectrometers. as they would be placed in hall A. is shown in Figure 10.15. and the placement of the1.2-GeV/c spectrometer in the pit is shown in Figure 11.3. The plan to place t~e 1.2-GeVlc spectrometerin a pit is driven by the fact that this arrangement permits the height of hall A to be lowered considerably.making it possible t::> increase the floor area of this hall significantly without increasing its cost. In addition.the pit provides ~ convenient place to locate the support structure for the spectrometer where it wouldnot interfere with the 4-GeV/c spectrometer. If this plan is adopted. great care must be taken to makethe pit large enough (in particular. to accommodate sufficient shielding); once constructed. it cannot beenlarged.

The smaller hall (hall C) could contain the 4-GeVlc moderate resolution spectrometer together withthe Variable Acceptance Spectrometer (VAS). The pivot is located at the center of the room. allowingmaximum flexibility in the angular ranges.

Hall 9. which is separate from halls A and C. is sized to hold a photon tagging magnet and detector.the LAD. an analyzing magnet for forward-going particles at small scattering angles. and appropriatebeam dumps. See Chapter 11.

174 r

I I I I I

Experimental Equipment

"North

F'JgUre 10.15. Outline (top view) of the high-intensity end stations A and C showing direction of beam lines through the buildings.The silhouettes in end station A show the 4-GeVlc high-resolution spectrometer and the 1.2-Gevlc spectrometer, which bendsdownward into a pit (:lot shown) as discussed in the text. The ~'est wall of end station A and the north wall of C are made ofmovabl.: blocks as shown in the figure.

UpgradesThe design of the experimental halls should be flexible to accommodate future upgrades in the

experimental equipment or to ailow for special experiments which do not make use of standard CEBAFequipment. While the layout of experimental halls A and C shown in Figure 10.15 appears to be areasonable starting point, it does not allow performance of high-resolution triple-eoincidence experiments.To allow for addition of this capability in the future, more floor space is needed; therefore. the westwall iu hall A will be made of blocks (as shown in the figure), so that it could be taken away withoutblasting and the hall expanded. Accordingly. the beam dump will be located sufficiently far away. Thisexpansion will also provide envugh space to the left of the target for the 4-GeVie reimaging spectrometer(shown in Figure 10.3). or alternatively. a higher-energy spectrometer. Figure 10.16 gives an outline forthe new expanded halls A and C, showing how three high-resolution spectrometers could be laid out inhall A. The angle which the beam makes with the walls would permit the new larger spectrometer togo from 100 to about 5Go.

An upgrade of haII C should allow an improvement of the 4-GeVIe spectrometer to higher resolutionor I-aigher momentum. !t should also accommodate an additional spectrometer !or triple-eoincidenceexperiments. The back wall of this building is also bl,;jlt of blocks to provide the extension needed for

I..

175

•

CEBAF Design Rc:pon

I I i

~North

I

F'ipre 10.16. Outline of end stations A and C as they might appear after they have been clLpdJlded in the future. End stationA has been expanded sufficiently far in the westerly direction to make room for the 4-GeVlc high·resolution reimaging spectrometershown in Figure 10.3. Hall C has been expanded in the northerly direction.

this purpose. The angle which the beam makes with the walls was chosen so that the longer spectrometerswhich Inight be installed after an upgrade would have the maximum angular range possible for a buildingof that shape. Hall C. which houses the smaller spectrometers and does not have a pit. would be thebest place to locate experiments requiring special equipment.

References

1. L. N. Hand and W. K. Panofskv. Rev. $c. lost. 30 (1959).2. R. Auzolle el al.• IEEE Trans. Nucl. Sci. NS 28, 3228 (1981).3. H. Ikegami el al.• Nucl. lost. Meth. 175, 335 (l98!"J).

176

II" - I ..

I I

11. Conventional Construction

11.1 Introduction

CEBAF will be built on a site in Newport News. Virginia. It will comprise berm-eovered beam­line-enclosure for the accelerator. three experimental stations. and various support facilities includingseveral pre-existing buildings. This chapter describes the plans for overall use of the site. includingconstruction. utilities. and building renovation. An extensive list of codes. standards. and guides to befoilowed is included at the end of the chapter. Conventional constructi~n planning has '"'een done atCEBAF in conjunction with the A&E contractor. Daniel. Mann. J.:>hnson. and Mender. ·~2, (DMJM).

11.2 Site

Figure 11.1 is the site plan for CEBAF.

Site Description and Overview

SettingThe CEBAF site is located in the city of Newport News. Virginia. on Jefferson Avenue, Route 143,

adjoining the northern boundary of the Oyster Point Industrial Park. a community of small businessesdealing in research. technology. and manufacturing. The site consists of approximately 200 acres andadjoins 67 acres of state-owned land dedicated for the project. The site is zoned for research anddevelopment. Nearby transportation resources include Interstate Route 64. the Chesapeake & OhioRailroad. and Patrick Henry International Airport.

Current ConfigurationThe larger parcel of land includes the decommissioned NASA Space Radiation Effects Laboratory

(SREL), a 4O,SOO-SF high-bay structure with an attached 16,SOO-SF. two-story office building. The 67­acre parcel includes the Virginia Associated Research Campus (VARC) building, a one-story. 31.535­SF structure. as well as a one-story 2892-SF service buildiilg. These structu:-es bve utilities. services.and established roadways. walks. and parking areas in place. The structures themselves will be renovatedand remodeled.

Tne project site is heavily wooded except for the areas around the existing VARC and SRELbuildings. Ground surface elevation.; vary by approximately 10 feet. with an average elevation of ap­proximately + 34.5 feet MSL. Drainage on the site is interrupted by a network of old roadways.

A rea GeologyThe Newport News area is located within the Atlantic Coastal Plain geologic province and underlain

by approximately 2500 feet of marine and estuarine deposits. The subsurface stratigraphy in this area isgenerally divided into major subdivisions based essentially on geologic age. The Columbia Group ofPleistocene Age is composed of normaily consolidated to lightly overconsolidated clays and !>acds. andextends to depths of 20 to 40 feet below the ground surface. Locally. these soils are referred to as the

1 ""'-j . I "

177

."

I • II

CEBAF Design Repon

t~--_·_------·----_·_-_·"

i

iiI

I i: $OUTH t; ACCESJ!BUILDIN~i i! I

IjII1I

END STATIONS

NORTHL1NAC

MACHINECONTROLCENTER

~ '~T(.ROJECT NORTH ' t~

SREL : I I~B1,.===;;::::7/ 1t.Yi,N~~;Hp .gr/ ~ti)lt_~~

OFFICE AND1rFnoCOMPUTER .: I ~CENTER Ls

-----_ .._----...;.~\

') { , f

~ PATRICK HENRY AIRPORT JEFFERSON AVENUE DOWNTOWN NEWPORT NEWS ~

Fagun 11.1. The CEBAF site plan. VARC and SREL are existing buildings with access and parking lots in place: all else in thefigure is to be constructed.

Norfolk Formation. The majority of these soils were deposited in an estuarine or shallow marine en'.;­ronment. An important fea!Ure of an estuary is that tidal currents ar~ very effective in distributingsediments; thus frequent changes in soil stratification and composition (which may significantly affectengineering properties) occur across short horizontal as well as vertical distances.

Geotechnical StudyLaw Engineering has completed the initial geotechnical study for the project. Their initial report

includes a rev:ew and evaluation of the proposed CEBAF constructivn. previously obtained information.and data recently obtained at the site, The conceptual design for conventional construction used thisreport as a basis. During spring and summer 1986 the detailed geotechnical study is being conducted.

178

•

I I I

Conventional Construction

GroundwaterWater levels measured in ~8 borings during the preliminary geotechnical study in the summer of

1985 averaged 7 feet below existing grade---corresponding to a mean groundwater elevation of + 28 feetMSL. Auctuations in the groundwater level should be expected due to tidal variations. seasonal changes.and occasional heavy precipitation. Well points have been instal1ed to measure these fluctuations. Meas­urements made in the autumn of 1985 indicated that the water table had risen I If:z to 2 feet. Based onobservation of the head differential. a preliminary groundwater movement is estimated to be 2.5 metersper year in a southeasterly direction.

Site Development Plans

ArchitectureTo create an architectural identity appropriate to a principal international center for scientific re­

search. a consistent architectural theme will integrate and unify the facility. AI1 new buildings except theend stations. which wil1 be cast-in-place concrete. wil1 consist of a light steel-frame structural system withdeep-fluted corrugated steel siding mounted horizontally. A series of smooth steel panels wiil be usedhorizontal1y to organize the locations of louvers. doors, windows. and graphics. The steel siding is aninsulated sandwich panel and is prefinished on both outside and inside surfaces. Interior partitions willbe either concrete masonry or gypsum wallboard on steel studs.

AccessSite access ""il1 be provided from Jefferson Avenue (Route 143), a four-lane highway that is schedliled

to be widened to six lanes divided between 1985 and 1989. Oyster Point Road, north of the site, isplanned for widening to four lanes with an interchange proposed at Interstate Route 64. On the eastside of the ~ite an emergency access will be provided to Bunker Road.

The internal road network wil1 provide access to all facilities by means of a primary road leadinginto the center of the accelerator complex. secondary access road.:; and service drives, and vehicle parkingat each building. The present access road from Jefferson Ave;;.;e to the SREL Building wil1 serve as theentrance to the accelerator. the Machine Control Center. and al1 the support buildings. This wil1 be arestricted technical entrance for use by scientists, engineers. technicians, and service vehicles. Similarly,the present access road to the VARC Building (which will continue to serve as the CEBAF administrationbuilding) wil1 be the main public entrance to the site and to the administration and visitors' area. Theroad network wil1 not only separate the types of access to the site, but wil1 also permit convenient on­site circulation.

Site Preparation and GradingWooded area wil1 be preserved where possible. Erosion and sedimentation control devices will be

placed in conformance with Virginia Soil and Water Conservation Commission criteria to minimizeenvironmental damage.

Grading operations will comply with the recommendations and requirements set forth in the geo­technical reports. Soils will be identified according to the Unified Soil Oassification System. Materialtesting and compaction requirements wil1 meet the requirements of applicable American Society forTesting and Materials (ASTM) publications.

DrainageThe site presently drains to Brick Kiln Creek. which discharges to Big Bethel Reservoir. No perennial

streams are present on the site; however. an ephemeral stream handles discharge from old borrow sites.Overall. the site is poorly drail"ed.

Runoff wil1 be calculated by the rational method. Channels, culverts. and short closed piping systemswill be designed to handle a 25-year frequc:ncy storm. The lOO-year storm will be evaluated to determineany adverse effects to the facilities.

Site drainage will be intercepted by existing channels that ultimately flow to the Brick Kiln CreekWatershed in accordance with a U.S. Army Corps of Engineers deed restriction. This restriction requiresapproval by the Corps of Engineers of any development to ensure that there is no contamination orreduction in flow to Big Bethel Reservoir.

Roads. Walks. and Paved AreasA primary road leading into the center of the ring will provide access to all buildings by means of

service drives and parking areas. This road wil1 connect to the existing road system in front of the SREL

179

• •

•CEBAF Design Repon

Building. Street lighting will be provided for both safety and security. Walks. bicycle paths. and joggingtrails will be provided to minimize the on-site usage of automobiles.

FencingTne secure areas of the site will be enclosed by a chain link security fence 7 feet high. Access-control

gates will be provided at the gate house and at the emergency egress road that leads to Bunker Drive.

Utilities CorridorA dedicated utilities corridor to serve the accelerator and related support facilities will be provided.

with lateral corridors serving the individual facilities. All utilities will be underground. and access man­holes will be provided at intersections of the main corridor and the laterals as well as at the bend pointsin the corridor.

LandscapingDisturbed areas that are not paved. graveled. or otherwise covered will be landscaped with shrubs.

trees. grass. or other types of ground cover. Lannscape vegetation will be compatible with local conditionsto minimize the use of either fertilizers or water. Where essential. sprinkler devices will be used toestablish and maintain vegetation. Ground cover will be planted on the earth berms to provide a main­tenance-free slope stabilization.

11.3 AcceleratorGeneral

The CEBAF linac will be split into two antiparallel segments connected by recirculation arcs. Thebeam-line-enclosure for the two linac segments and for the injeCior provides an envelope 11 feet wideby 8 feet high. The enclosure for the recirculation arcs provides an envelope 8 feet wide by 8 feet high.(The injector will be located in an extension of the north linac segment.) Access to the linac for personneland equipment is by means of two drop access points located in access/service buildings at the west endof the north linac segment and the east end of the south linac segment. Additional personnel egress isprovided at the midpoints of the linac segments and recirculation arcs. Utility service to the tunnels isthrough the service buildings. which are fed from a utilities corridor.

StructuraIThe beam-line enclosure tunnels may be constructed of poured-in-place reinforced concrete or

corrugated metal arch sections. With the varying thickness of the compressible soil layers (clays) belowthe tunnels. some differential settlements are anticipated due TO the surcharge effect of ttoe shieldingberms. It is anticipated that ~ith preloading most of the settlement wiH occur before final alignment ofthe machine.

MecillmicalThe linac and switchyard tunnels will require humidity control. The relative humidity is specified to

be under 55%. During both operation and shutdown the air will be circulating in the direction of thebeam. Exhaust fans will be located in strategic locations for purging the entire accelerator in 10 minutes.For air tempering, heat recovery from power supplies may be utilized when appropriate. Air temperaturein the beam enclosures will be maintained at a minimum of 55°F.

Duplex pumps will be installed in tunnel sump pits at each exit point to service the enclosure andit" exterior foundation drain tile.

ElectricalAn illumination level of 30 foot-candles will be provided in the tunnel by fluorescent strip lighting.

Utility power will be supplied by l2O-volt single-phase outlets and 480-volt three-phase outlets. Cabletrays will be provided for accelerator power and control cables. Sump pit high-water alarms will beconnected to the site-wide supervisory system. Battery-operated emergency lighting units will be installedin each egress alcove with remote heads mounted in the enclosure. Fire alarm manual pull sta~ions willbe installed at each egress alcove.

180

I I

Conventional Construction

11.4 Experimental AreasGeneral

The three end stations house the experimental detector equipment. Contiguous counting houses willcontain space for data collection. electronics. power supplies. mechanical equipment. kitchens. toilets.and sleeping areas. Provision IA.;II also be made for parking for experimenters· trailers and for the requiredumbilicals to the high-bay experimental halls and counting houses.

Each end station has the following requirements:

• Floor of high-bay area 10 feet below centerline of beam.

• Fifty-ton-capacity bridge for a pair of 25-ton hoists with hook height of 40 feet above the beamline. Having two cranes allows independent use at 25-ton capacity as well as combined use at 50­ton capacity.

• Exterior shielded access doors 16 feet wide by 16 feet high. These doors are roller-mounted withelectric drive me~hanism to facilitate operation.

• Hinged platfonn at grade level to pennit drive-in service by vehicles for unloading by the interiorcranes while preserving the floor of the end station exclusively for experimental equipment.

Personnel access will be provided to each end station by means of a shielded opening at grade levelat the entry to the counting house stairway.

End Station BThis will be the smallest experimental area. Its purpose is low-eurrent experimentation.

End Stations A and CBecause high-current experiments will be conducted in these two end stations. it is necessary to

construct them with concrete walls 2 meters thick and with concrete roofs 1 meter thick. All penetrations,shafts. and doorways must be shielded with double bends. See figure~ 11.2 and 11.3 for definition of endstations A and C.

Structural

End Sta/ions A and CEnd stations A and C are reinforced concrete box structures. The roof and wall thicknesses are

dictated by radiation shielding requirements. The roof is a one-meter-thick poured-in-place concrete slabwhich acts compositely with and is supported by built-up steel-plate girders. Vertical roof loads arecarried by the 2-meter-thick walls. A continuous corbel will be cast into the wall to support the overheadcrane rail. Seismic and other lateral loads are resisted by the concrete roof diaphragm and the box sidewalls. Each .,)f these two end stations has one wall built of removable blocks to facilitate future expansion.

The base slab for the building is of reinforceJ concrete supported on piles placed in a unifonn grid.Additional piles will be placed under the walls to support the weight of the walls and roof. Each pilewill be precast. prestressed concrete with a 4O-ton capacity.

End Station BSince the roof is only one foot thick. it will be supported on non-eomposite rolled steel beams.

Because low-current experiments will be conducted in this end station, the shielding requirements forthe building construction dictate 2-foot-thick concrete wal!s and I-foot-thick concrete roof. See Figure11.4 for definition of End Station B.

This end station is essentially a smaller version ofend stations A and C and has the same requirements,except that the crane will have a 35-ton capacity with a 35-foot hook height. This crane will be supported.however. by steel beams bearing on concrete pilasters cast integrally with the wall.

Counting HousesEach end station has a contiguous counting house. The substructure is reinforced concrete and the

bac;e slab is a continuation of the end stztion foundation. Building columns will be supported on spreadfootings. The counting house superstruc::.ue is a steel-framed one-story building. Lateral building ~taolli~y

is provided by X-bracing and the thick end station walls. The roof is a sted diaphragm supportea onsteel beams.

181

I

CEBAFDesig:JRe;x>~

iJ

~lI

! I

I II J, II j

f\I "

Fipre 11.2- Aoor plan• end statiODS A and C.

182

I

Con\lentional Construction

----=---~ ,

cC

:

I-

l-

II...

- ~. :..,.,.,I .,.,.,

\~ H"

I/

II

I

f­l­I­l­I­f­l­f­I-------

,-IrI,, ," I

I. !

I

.J

, ':I'

..i

:...:..:1

: I

, i

1',, ,: ':I '

Ii

i

==

I

!II

I~I

I [J

II ,

! i\ ~

~ IIIIII\~ I

I:....=

FIpn 11.3. Elevations and sections. end stations A and C.

I I

183 I

CEBAF Design Repon

I i 'I

Ii·.,. h• ••• .. I.

~:£Ja::..>Iea::...

•184

~G:I ,-' .,e " _Zz

:5i5::» lL

i §-'-' ...

FIpre 11.4. End StatiOD B.

• I

Conventional Construction

MechanicalThe end station experimental halls will not be air conditioned; they will be heated and ventilated

only. Either unit heaters or infrared radiation units (fuel or energy source to be determined) will maintaina minimum working level of 55°F. Air-handling units. with heating elements and appropriate duct work.will distribute fresh air uniformly throughout the building with a ventilation rate of six air changes perhour for a volume of 10 feet above the floor. Radiation shielding will be provided for air intake louvers.

The counting houses will be provided with HVAC for all occupied areas as well as heating andventilation for electrical and mecha:-ical equipment rooms. The counting room. with a raised floor. willbe equipped with a computer-room-type self-contained unit with humidity controls and economizer.

Fire protection for both the experimental halls and counting houses will be in accordance with NFPAOrdinary Hazard. Group 1 occupancy. Separate. hydraulically calculated sprinkler systems will be in­stalled in the experimental halls and counting houses. Dry pipe systems will be installed in the experimentalhalls. The access floors of the counting houses will be protected by Halon systems (see below).

ElectricalEach counting room will have a 48O-volt switchboard, 6O-ampere 3-phase 48O-volt receptacles. as

well as 120-volt receptacles. There will be an uninterruptible power supply for data processing equipmentas well as required wired telephone and computer terminal outlets.

Data processing equipment and general space in each counting room will be supplied by isolated­ground receptacles whose grounding terminal is connected to the isolated-ground buss in the access floorspace. The isolated-ground buss will be wired directly to service entrance neutral ground.

Counting rooms will be illuminated to 100 FC (maintained) by 2-foot by 4-foot recessed three-lampparabolic reflector luminaires in suspended ceilings. Illumination will be fully dimmable from zero to100 Fe.

Individual Halon systems will provide fire protection for each counting room's access floor spacewith its data processing cables. In case of fire. the Halon system will provide up to two minutes of audio­visual alarm signal before releasing measured Halon gas. The Halon system will use ionization-typesmoke detectors connected to the central alarm system.

Beam Dumps

Beam Dump BEnd Station B requires two beam dumps. one witnin the building and one downstream of but

contiguous to the building. The beam dump inside the building consists of a shelved trench with provisionsfor placement of concrete shielding blocks. The beam dump downstream consists of a concrete tubecontaining beam-stop material which can be replaced internally through the beam pipe. The entire beamdump is then protected by earth shielding.

Beam Dumps A and CA major consideration is housing the specially fabricated high-power beam dumps, capable of

continuously absorbing and dissipating the full beam power produced by the linac over a wide range ofenergies. The dump vessel may be mounted on a mobile frame to allow removal at a later date.

It is proposed to use a 100foot-wide by 7-foot-high concrete box structure to house the beam dumpequipment. This will allow the necessary space around the dump vessel to accommodate future access.The jacket housing the electron beam. which connects the end station with the beam dump vessel chamber.will be housed in a 7-foot-wide by 7-foot-high concrete structure. This will allow removal of the beamdump vessel through the end station area via this chamber. A utility shaft \l<ill be required in the vicinityof the beam dump vessel chambers to provide access to the miscellaneous piping and other utilityconnections.

To provide the desired degree of control against radiation. the beam dump housing structures willbe shielded by high earth berms. Given the requirement of attenuating the secondary radiation. and theloading requirement imposed by the berms, it is proposed to use 4000-psi reinforced concrete structures(28-day strength). Special additives will be specified to prevent concrete dusting. The bottom half of thebeam dump vessel structures wili be lined with stainless steel to prevent radioactive water spillage fromentering the pores of concrete or escaping into the outside ground water through the joints.

185

I

I I

CEBAF Design Repon

I III •The structural foundation system for the beam dump tunnels will be supported on select backfill

and crushed stone. The preliminary geotechnical data suggest a possibility of one-inch to two-inchdifferential settlement between the beam dump chamber and the end station wall.

If the system supporting the electron beam jacket and the beam dump vessel can be desigr.ed toaccommodate this type of long-term differential settlement. the tunnels beyond the end station walls canbe designed to be supported on a stabilized elastic foundation. If this is not possible. the tunnels can besupported on concrete piles. Based on the method of construction chosen for the end stations. theconcrete box structures may be constructed using anyone of the following options:

• Open cut method with sloped excavation

• Trench method with strutted supports

• Cast-in-place concrete box structure or precast concrete box sections

The sump pit method will be used for excavation dewatering. In all cases. special attention will bepaid to detailing construction joints. waterproofing membrane placement. use of special sealant materials.and concrete mix design.

11.5 Support FacilitiesSREL Building

GeneralThis existing building' contains a high bay (40.500 SF) and a 2-story office wing (16,500 SF). The

high bay contains the support system and girders for a 50-ton-eapacity crane. The exterior walls of thehigh-bay part of the building are precast concrete with sprayed asbestos insulation in the upper portion.The exterior walls of the office wing are aluminum and glass curtainwall with brick panels in betweenand at the end walls.

The approach to the renovation design of this building is to make maximum use of the existingpartitions and structural support systems for both the high bay and the office wing in order to expediteconstruction and minimize construction costs. The exterior envelope will be modified only to accom­modate the current energy conservation criteria. See figures 11.5 and 11.6 for definition of the SRELBuilding.

The high-bay area will be apportioned as follows:

• Receiving. staging, and shipping areas

• Storage for machine-related components

• Magnet assembly area

• RF test system area

• Linac cavity work area

• Cryostat and transfer line assembly area

• Oear. room

• Accelerator welding shop

• General machine shop

The two-story wing will contain the following:

• Offices

• Conference room

• Electronics shop

• Light machine shop

• Drafting room

• Central utility plant

186

I • I

I

Conventional Construction

j .,.. _ r

•

" I~I: :

~ ~ I". ..: 8 ::: ~ ;I I I i

,'"II

Ci'

--------,I II II I

III____J

zc...A-

...Z..~..lD

:i

Ill'I I

I ~I .I I

II I

) ;

;

~.=-

z tz

c~ ...A- !II: i;II:

!0g0...

j... ......

...0!!z..

·····

II

c-

Fipre U.5. SREL Building floor plans.

• I

187

I ..,.

I

CEBAF Design Repo~

...

j. --'

I· - Ic

~~UIII•

_,.-1I ' I:

~i'P

;1'

= I ':ide:1 ': i I'e

• u r-' ~ -; iI a!l, i; I :s ~:.I. ~ I~III' ;1 .

~ c !el•• ..I" 1:. 5:;.~ 1:11

u" I i: i ;1 'Iii iI

zo

aIII..III

IIIUZCIIC..ZIII

188

F"JIIIR 11.6. SREL Building section and elevation.

-

Conventional Construction

Two 25-ton cranes will be adapted to existing crane rails which extend over the full length of thehigh-bay area. permitting the cranes to operate either independently or in combination to provide a SO­ton capacity when required.

Test functions which may generate radiation are located to take advan:age of existing shieldingwalls. In addition. maximum use will be made of existing precast concrete shielding blocks previouslyused in this area and presently stored on site just east of the SREL Building.

StructuralProposed-use areas in the SREL Building are such that no structural modifications are required.

Heavy-load uses such as magnet assembly and storage. large machine shops. etc.• are assigned to theold experimental area, which has a thick. heavily reinforced pile-supponed slab. The suppon wing ofthe building houses the light-load uses such as offices, electronic shop, and light machine shop.

MechanicalThe high-bay area will be heated in winter to 55"F and the rate of ventilation will be based on six

air changes per hour at 10 feet above the finished floor.The offices will be comfon-conditioned to 7SOF db 50% R.H. in summer and 7O"F db 50% R.H. in

winter. The HVAC requirements will be developed with due consideration for architectural design,building materials. energy conservation criteria, local building codes, and life-cycle costing.

Since this building is 20 years old, our approach will be to rejuvenate the old fixtures if possible orreplace as required. The existing chillers will be examined and either refurbished or replaced. The existingboilers will be refurbished and used to generate hot water. The building renovation will not requireremoval of existing duets, diffusers, and registers. Asbestos will be removed or encapsulated prior tooccupancy.

For fire protection. the SREL Building will be divided into convenient zones and treated as OrdinaryHazard. Group 1. Sprinker systems will adhere to NFPA requirements and local codes.

ElectricalThe existing electrical fixtures, panels, outlets, and wiring will be checked and retained where

possible.

0fIice and Computer Center

GeneralThis multi-function building will contain space for the following general functions. See Figure 11.7.

• Computer center

• Instrumentation & control

• Offices• VISiting scientists' office

The net area for these functions is 18,700 SF. By applying an efficiency of 65.5% to the net area,a gross area of 28,550 square feet is therefore assigned to this building.

The building will conform to the architectural theme of light steel-frame construction with deep-ribmetal-insulated sandwich siding panels. I!:terior partitions will be steel stud with gypsum wall board.

StructuralThe core of the building consists of 20-foot by l00-foot structural steel modules consisting of roof

bar joists, girders, and steel columns. These modules are repeated until the overall building size isgenerated. The 20-foot module was chosen to optimize the span of the deep-rib horizontal siding. Lateralbuilding stability is provided by a steel-diaphragm roof deck and reinforced concrete masonry shearwalls. The shear walls are located around the outside of the building core. thus becoming walls for themechanical room, computer room, and so forth. The on-grade floor is a wire-mesh-reinforced floatingconcrete slab, thickened at the edges, under the shear walls and at column footings.

MechanicalBecause this is a mixed-use administrative and operational building, the various spaces will be

conditioned differently. The computer room will have computer-room units complete with glycol dry

189

., I

CESAF Design Repon

190

~<C

~ 5i § I""'---- ~

II••:1.1 t

II I

~<C

5~

-.:-

FIgIIn 11.7. Office and Computer Center Boor plan and elevations.

•

Conventional Coru.truction

coolers equip~d with humidity control for \l.inter cooling. l'ecessar~; plumbing will be provided forcomputer room air conditioning units. its humidifie•. H&V units. and AHVs. The conference rooms willbe provided with 10 CFM ~r person of outside air for ventilation: all the general occupallC)' spaces \l.illrequire a minimum of 5 CFM per person. The office areas will be evaluated for the air conditioningload. It is proposed to provide a water-S<)urce heat pump system with electric or hot water backup.

This building will be fully equip~d with sprinklers. and the Halon suppressant system \\ill be usedfor tire protection beneath areas which have access floors.

Compwer Center ElectricalThe building will have a 480-volt switchboard: voltage for power and general lighting will be 2TII

480 volts.Data processing equipment will be supplied by isolated-ground receptacles whose grounding ter­

minals are connected to the isolated-ground buss in the floor access space. The isolated-ground buss willbe wired directly to service entrance neutral ground. Uninterruptible power supply will be provided fordata processing equipment.

The computer room will be illuminated to 100 FC (maintained) using 2-foot by ~foot parabolicreflector luminaires in a suspended ceiling. Illumination will be fully dimmable from zero to 100 FC.Mechanical equipment rooms and stores will be illuminated to 30-50 FC maintained using low-bay high­pressure-sodit:m luminaires.

Separate Halon systems will provide fire protection in the computer center access floor space. AHalon system will provide up to two minutes of audio-visual alarm signal in the room before releasingmeasured Halon gas in case of fire. The systems will use ionization-type smoke detectors connected toa central alarm system.

Central Helium Refrigeration Plant (CHRP)

GeneralThe arrangement and design of these structures is determined by their function. The buildings will

conform to the architectural theme of light steel-frame construction with deep-rib insulated-metal sand­wich siding panels. See figures 11.8 and 11.9.

The refrigeration building must be specially designed to accommodate the cold box. which is 14 feetin diameter and 35 feet high. This box must be set in a 15-foot pit and requires 12 feet clear space allaround. The building will contain a 5-ton bridge crane with a 32-foot hook height and will have a removableroof hatch permitting the entire box to be removed. Motor control centers. a control room. and officesare also located in the refrigeration building.

The compressor building. another separate structure in the CHRP. houses the co;npressors. whosehandling will require a 5-ton bridge crane. Because of the high noise generated by these compressors.special acoustical treatment must be provided.

The liquid helium tanks will be mounted on a frame set on an exterior concrete 'pad. The locationof these tanks (next to the refrigeration building) must permit access by truck to facilitate reloading.Similarly. the air-cooled radiators will be pad-mounted and located adjacent to the compressor buildingwith provisions for easy service access.

StructuralThe foundation of the compressor building consists of a 3-foot-thick reinforced concrete mat. whose

primary function is to provide sufficient mass to attenuate the vibration associated with heavy reciprocatingequipment. Building columns will be supported directly on this mat.

The refrigeration building houses the cold box and the cryogenics plant control facilities. The coldbox area is a high-bay. steel-framed structure. The roof deck is not a diaphragm because of the largehatch provided for cold box removal. Lateral roof stability is accomplished with an X-bracing. Buildingcolumns are supported on spread footings. The below-grade portion of the cold box will be housed in areinforced concrete pit. This pit will be waterproofed in the same manner as other grade concretestructures. The pit walls carry the weight of the cold box to the bottom slab. which acts as a mat foundationto support the cold box.

The control facilities area is a light steel-framed structure with conventional spread footings and aconcrete slab on grade.

191

CEBAF Design Report

I •

!CIi 'I ...... I

~

I .

I~I..~II

: ....... 1 - ~. I IIi il11 ~I~ I-,

0

0 CI

0 ~ i ~c

I~5

i~.-...~

FIpre 11.8. Central Helium RefrigeratioD P!aot (CHRP). sectiOD aod elevatioD.

192

Conventional ConstruCtion

--<j . __ 11

~ - I

~ !.. ' ; .. , i.~-- - - -- ----~

:1: : : II I~ I ! ,

; li.. ~. .. ,

! g;: I !, II ' . --, ,

, .. .. .. U! Ii---- - -----_., , ,-

: ,!:: ! I ...

i : I/ ~::----- -------\ ! i~ , , '" 1_: \a \ : .. ;

\ ~---- ::-----".:-1 •I

!C-----mu

-

u

---' )

~

I II :J :: :]:t1 ~---------------~

, ..... '

Jl II 11. _ P"

II r-.. -- i

I roni

:t-. , ,II W ~ §liI!!------:

I~I !'-4 I"!' ~

5 ii

I ICJ:------i~-------,i""iiiC-~_ ~

CJ0

Ie~CJ

§i

I §CJ Ii ~

Itl

I 10 01~ 10 01s

gIe 01 ~

J !r ,

icd

FIgure 11.9. CHRP floor plan.

MechJJni.calThe compressor building will be provided with industrial-type fans for ventilation based on a rate

of six air changes per hour. This building may not require winter heating beyond that generated by thecompressors. However, during shutdown or when all compressors are not working, heat will be providedby unit heaters.

193

I I

CESAF Dolg.n Repon

The refrigeration building is a high-bay structure requiring exhaust fans for the upper space. Air­handling units will maintain 85°F in summer with provision for dehumidification. The control room andoffices will be air conditioned.

ElectricalCompressors and central liquefiers will be supplied at 4.16 kV with separate motor controllers. A

~volt switchboard will pro"ide 277/480 volts for power and general lighting. with 120-volt and~volt receptacles in all equipment areas. Wiring for computer te&T.1inals will be provided as required.Mechanical equipment and electrical rooms will be illuminated to 30-50 FC using low-bay high-pressureluminaires.

Service BuildiDgs

GeneralThe service buildings are distributed around the inner perimeter of the accelerator. with the exception

of the building serving the injector and the buildings that are needed to house the power supply for thebeam switchyard area. The service buildings are equipped with minimal building appurtenances; theirmain function is to protect equipment from the elements. Two Iwge service buildings 12 feet wide byapproximately 1000 feet long will house the klystrons. power supplies. control cabinets. and instrumen­tation for the linacs. The service building for the injector will be a 16-foot by l00-foot building locatednorth of the injector. Eleven small service buildings. four around each arc and three at the beamswitchyard. each 12 feet wide by 30 feet long. will house the DC power supply and switch gear for themagnets. Additional electric substations wiU be located on pads adjoining the access/service buildings.

Equipment access to the beam-line enclosure floor level will be by means of two access/servicebuildings. one located at the west end of the north linac and the other at the east end of the south linac.A 7Vz-ton bridge crane will be positioned so that it can unload a truck and lower equipment into thetunnel labyrinth through the access opening. A removable steel grating and· removable railing will beprovided for personnel safety. There will be an access stair from the ground level to the tunnel level toserve as an emergency exit from the tunnel. The control center which will adjoin the northwest accessIservice building contains a room for machine control equipment as weU as offices. toilets. and a smallkitchen.

StructuralThe service buildings have identical superstructures of conventional steel framing. The overhead

crane in the access building is supported from the building columns. OveraU building stability is providedby X-bracing. The deep-rib siding spans from column to column; hence girders are required only toframe out doorways and other openings.

MechanicalThe control center and the service building for the injector will be air conditioned. The other service

buildings and the access/service buildings will be heated (to prevent freezing) and ventilated. exhaustingthe power supply heat loads. Fire protection for the service buildings. control center. and access/servicebuildings will include ionization-type smoke detection. and both dry C1:Jd wet charge sprinkler systems.Manual fire alarm pull stations will be provided as weU as exterior fire hydrants. All systems will reportto the site-wide supervisory system.

ElectricalExterior. pad-mounted substations will supply 480-volt switchgear in each building. Utility power

will be provided by 12O-volt. 2OB-volt. and 44O-volt outlets. IUumination will be provided in confonnancewith building function. Conduits for the telephone and supervisory systems will be provided betweenbuildings and exterior communication duct banks.

Gate House

GeneralThis 100foot by 100foot building. located at the entrance to the perimeter road for the machine

facility. will control access to the fenced area surrounding the machine. The building will conform to thearchitectural theme of light steel-frame construction with deep-rib metal-insulated sandwich panels.Interior partitions will be steel stud with gypsum wall board.

194

I

•

Conlientional Construction

StTUClUTalThis ",ill be a light steel-frame structure with steel siding and steel roof deCK.

MechanicalThe gate house will have a separate air conditioning and heating unit ",ith automatic controls to

maintain comfortable en...ironment throughout the year.

ElectricalThe gate house will be illuminated to 30 Fe (maintained) using suitable fluorescent luminaires.

Power will be supplied for heating. ventilation. and room receptacles. A telephone outlet will be provided.as well as a computer terminal for printout facilities. An alarm facsimile board with representation ofeach building or facility will furnish an audio-visual alarm signal for the entire accelerator facility. Securitylighting at the gate house will be provided as part of the site lighting.

11.6 UtilitiesSite AC Power Distribution

Figures 11.10 and 11.11 are the site AC power distribution schematics.The existing Virginia Power Company 12.47-kV service to the SREL Building will not be affected

by the new construction and may remain in operation to supply construction power and early start-uppower for the machine. However, it is inadequate to serve the estimated 15-MW to 25-MW total siteload and will be replaced by a new 34.5-kV overhead line. Virginia Power will install a new 34.5/12.47kV substation to serve an adjacent CEBAF master substation, consisting of outdoor metal-clad switch­gear, to supply the underground 12.47-kV distribution system throughout the main site. This system willinclude concrete-encased PVC duets, manholes, and 15-kV cable to supply 12.47-kVl~Vpad-mountedtransformers at the various buildings. The transformers will supply ~volt main switchboards in eachbuilding.

The existing 12.5-kV direct-buried cables supplying the SREL Building have been in service formany years and should be replaced by pennanent feeders. routed through the new duct bank distributionsystem, from the master substation switchgear. Virginia Power should install a new 34.5/12.47 kV sub­station, with a minimum of 5 MVA capacity, on Jefferson Avenue to supply standby power through theexisting SREL Building substation switchgear, as a backfeed to the master substation and to providelimited power to the entire main site in the event of a failure of the normal 34.5-kV supply. In addition,an emergency generator will be installed to supply minimum power to selected areas in the unlikelyevent of failure of both the normal and standby sources, which are supplied by separately circuited andlooped 115-kV Virginia Power sources.

The selection of the source of standby power may be affected by the final Virginia Power Companydesign decision on voltage level and source for the normal power, as well as by a cogeneration systemnow under study.

Emergeoc:y Power SystemsCentralized and local eme:-gency power systeJ'l1S will be inst'.alled throughout the site as required for

life safety and to support critical machine operations. Automatic banery-operated lighting units will beinstalled in all buildings and enclosures. The master substation, fire alarm, and security systems willinclude automatic emergency power backup. The control, computer, and communication centers, as wellas the site-wide supervisory and control system, will all be provided with automatic emergency powersystems. A detailed operational and safety analysis will establish the emergency power requirements forsump pumps, vacuum pumps, and emergency ventilation.

Acx:elerator CooIiDg SystemsAccelerator cooling systems will provide cooling for magnets and beam dumps with low-conductivity

water (LCW) of 98"F to 138"F range and for electronic equipment with LCW of 1050f" to 125°F. Includedwill be a cooling tower, air-cooled heat exchangers, pumps, piping, tanks, de-ionizers, and control systems.The LCW systems are closed-loop and cooled by air through air-cooled heat exchangers or by cooling­tower water through water-to-water heat exchangers. Air-cooled heat exchangers will be used wheneverthe temperature range is above 105° to reduce evaporative loss of water.

195

tjmtil;l>...,

~'I<l:>

i::1

_ UIl POLII

!l

1.DI'1"M) ..... , IV ~~,_=...= _- ,­IIIII

__.!!!.'!!!!'!......Ja.., O'II-...D'"

t .. ,

&'J,"

--------.... nPeo "I U on_aD UIII--------1

r -I" -"-~I I vuco

I ,.

I J '10 ,,,'UUI ,IUllla"o_

I 'L __ !

TO 12.47 KV MASTER SUBSTATION

VEPCO METERING COMPARTMENT

f..:-~

m~laoil

~

1.,

....~

r - - - -t-~- 1, 11. I

I ""r II II ,

: ~ ~ II IL~~ P!.1I!ll!! J

[gog.g

SN~~~

~a.o!'

TO 12.47 KV 800 AMP SWITCH ELECTRICALLY OPERATED

~--

11r-r---'I II I ••••pco

~I ,It~':-Ul¥, _a_I ,L J

"'iiliWOO-au~\"""""""'"

,.~Ii !:-.1 II.1 .

z0i=<~>CD~

ICD ....~.CD • -C\l0-

~o- f~ .. ,w'; i>..,1~< • Iw> iz::E I ' !

00 · -~ .. • •

~=~ II

•II .

dl I

, 11 .

L._.

FIpre 11.11. Power distribution to facilities from 12.47-kV substation.

Conventional ConstrUCtion

III

197

CEBAF Design Rc:po"

Communications. Controls. 3:1d InstrumentationA system of duct penetrations for accelerator control circuits will be installed between the control

center and the injector. between all service buildings and the linac tunnel. and between the end stationcounting rooms and the experimental halls.

A communication system duct bank will be installed in the utility corridor. coordinated with thepower duct bank. with separate hand-hole access compartments built into the power manholes forcommunication circuits. This duct bank win contain telephone circuits and Supervisory Control and DataAcquisition (SCADA) circuiting for utilities and process systems. fire alarms. security. combustible gasdetection. and radiation monitoring.

The SCADA system will monitor electrical power demands at the master substation; LCW con­ductivity. temperatures. pressures. and flow; sump pit water levels; critical operating parameters fromthe CHRP compressor building and the end stations; HVAC system operati:lg parameters; fire protectionsuppression and detection systems; combustible gas detection systems in the end stations; radiation andradiation alarms; and security alarms. The SCACA system will operate independently from the accel­erator control system and will have masking capability .0 separate and direct alarm messages to thecontrol center. the SREL Building for maintenance suppon. to the communications and security centerand to the municipal fire depanment.

11.7 Codes, Standards, and GuidesGeneral

Codes

• 1985 Virginia Uniform Statewide Building Code

• 1985 Virginia Uniform Fire Prevention Code

• National Fire Protection Association (NFPA)

• Occupational Safety and Health Act (1971 OSHA)

• Federal Propeny Management Regulations (FPMR) (handicapped)

Standards

• American National Standards Institute (ANSI)

• American Society for Testing and Materials (ASTM)

• Underwriters' Laboratories. Inc. (UL)

• Factory Mutual (FM)

Guides

• DOE General Design Criteria Manual (1983)

• Federal Construction Council (FCC) Federal Construction Guide Specifications

Civil

Codes, Standards, and Guides

• American Water Works Association (AWWA)

• American Concrete Pipe Association (ACPA)

• American Association of State Highway and Transportation Officials (AASlITO)

• Institute of Transportation Engineers (ITE)

• American Society of Civil Engineers (ASCE)

• Virginia Department of Highways and Transportation (VDlIT)

• Soil Conservation Services (SCS); USDA

• V:rginia Soil and Water Conservation Commission Erosion and Sediment Control Handbook

• U.S. Depanment of Commerce (Weather Bureau) Technical Paper No. 25

198

Conveutional Coosttuction

Architectural

Codes, Standards, and GuidesNRCA Roofing & Waterproofing Manual (1983)FM Factory Mutual 1-28-Insulated Metal

Deck.19TISMACCNA Architectural Sheet Metal Manual (1968)NFPA 101-81 Life Safety Code

Fare Protection

Codes, Standards, and Guides

• American Society Cif Mechanical Engineers (ASME)

• American Water Works Association (AWWA)

• Hydronics Institute (In)

• National Electric Manufacturers Association (NEMA)

• American Society of Plumbing Engineers (ASPE)

• National Fue Protection Association (NFPA) rue Protection Handbook and National rue Codes

EIec:tricaI

Codes, Standards, and Guides

• National Electrical Code, NFPA-70

• Nationa! Electrical Safety Code, ANSI C2

• National Electrical Manufacturers Association (NEMA) Standards

• Institute of Electrical and Electronic Engineers (IEEE) Standards

• illuminating Engineering Society (IES) Lighting Handbooks

• Insulated Cable Engineers Association (ICEA) Standards

Heating, Veotilatiag, aod Air ConditioDiDg (~rt F8Cl1ities)

Codes, Standards, and Guides

• American Society of Mechanical Engineers (ASMEr-Boiler and Pressure Vessel Code Require-ments

• The National Standard Plumbing Code

• Associated Air Balance Council (AABC)

• Air Moving and Conditioning Association (AMCA)

• Air Conditioning and Refrigeration Institute (ARI)

• ASHRAE Standard 9OA-I980, Energy Conservation in New Building Design, and Standard 62-1981, Ventilation for Accepl/lble Indoor Air Quality

• American Water Works Association (AWWA)

• National Environmental Balancing Bureau (NEBB)

• National Electric Manufacturers Association (NEMA)

• National Fue Protection Association (NFPA) Standard 9OA, Air Conditioning and VentiltuionSystems

• ASHRAE Handbook Fundamentals (1981)

~ Air Force Manual (AFM) 88-8, Chapter 6, ··Engineering Weather Data"

• American Conference of Government Industrial Hygienists Industrial Ventilation Manual

• GSAlPBS Energy Conservation Guidelines for Existing and New Office Buildings

199

I

3000-400O psi~SOOOpsi

S()()()-«)OO psiASfM A-615 Grade 60ASfMA-I85ASfMA-36ASTM A-242 or A-S88ASfMA-32SASfMA-SOl

CEBAF DcsigD Repon

P1umbiog

Codes, Standm'ds, and Guides

• PHCC National Standard Plumbing Code (NSPC)

• American Society of Mechanical Engineers (ASME)

• American Water Works Association (AWWA)

• Cast-Iron Soil Pipe Institute (CISPI)

• Construction Specification Institute (CSI)

• Hydronics Institute (HI)

• National Electric Manufacturers' Association (NEMA)

• Air Force Manual (AFM) 88-8, Chapter 6, ··Engineering Weather Data"

• American Society of Plumbing Engineers (ASPE)

• Thermal Insulation Manufacturers' Association (TIMA), Economic Thidcness Manual

StnIdDraI Materiab

Codes, Standards, and Guides

.. Aluminum Construction Manual, The Aluminum Association

• Building Code Requirements for Reinforced Concrete, ACI 318, American Concrete Institute

• Practice for Concrete Floor and Slab ConstTuction. ANSIIACI 302

• Code Requiremenzs for Nuclear Safety Re1aled Concrete Structures, ANSIIACI 349

• Prestressed Concrete Institute Standards

• Building Code Requirements for Concrete Masonry Structures, ANSIIACI 531

• Building Code Requirements for Masonry, ANSI/NBS 211

• Building Code Requirements for Reinforced Masonry, ANSIINBSD Handbook H74

• Recommended Building Code Requirementsfor Engineered Brick Masonry, Structural Cay Prod­ucts Institute

• Specific.aJion for the Design and Cons:lTUCtion ofLoad Bearing Concrete Masonry, National Con­crete Masonry Association

• Specifiauion for Design, Fabriazlion. and Erection of Structural Steel for Buildings, AmericanInstitute of Steel Construction

• Suuadard Specifiauions, Load Tables and Weight Tables for Steel Joists aru:!. Joist Girders, SteelJoist Institute

• Steel Deck Design Manual for ComposiJe Decks, Form Decks and RoofDecks, Steel Deck Institute

• Steel Deck Institute Diaphragm Design Manutll

QuaIiryofMaterioJsConcrete for structuresPrecast concretePrecast prestressed concreteReinforcementWelded wire meshStructural steelStructural bigb-streDgth steelHigb-strength boltsStructural tubing

200

-

12. Project Management

12.1 Project Management Objectives

The U.S. government, through the Deparnnent of Energy (DOE), has established as its first priorityin nuclear physics a high-intensity, high-energy, continuous-beam electron accelerator to perform fund­3r.)ental research. SURAICEBAF entered into a partnership with DOE to develop and construct thatfacility for the U.S. government and the nuclear physics community. CEBAFs matlagement objectivesin this partnership are to design, build, and operate a facility meeting the project scope (Table 12.1) atthe budgeted cost, within the agreed upon schedule, and in a manner consistent with all applicable laws,contracts, regulations, codes, orders, and policies. To meet these objectives, CEBAF is developingproject management policies and procedures that will be compiled into a Project Management Plan tobe submitted to the Department of Energy.

This chapter summarizes the project responsibilities and management tools and procedures plannedby CEBAF to accomplish the objectives.

Table 12.1CEBAF Project Scope

Accelerator SpecificationsParticleEnergyCurrentDuty factor'

Additional FeaturesThree equipped experimental areasOther buildings and facilities necessary to

support the accelerator, users, and staff

Cost and ScheduleTotal estimated cost (actual year dollars)Project startP:"('lject completion

"Macroscopic duty factor when the beam is OD.

electron0.5 to 4.0 GeV200 .,.A100%

S236M2nd quarter FY 19872nd quarter FY 1992

201

CEBAF Design Repon

12.2 Organization and Responsibilities

Figure 12.1 shows the CEBAF project organization. The Director provides the interface with SURAand with DOE. He is responsible for the overall management of the construction project and the facilityand for assuring the quality and safety of its activities and products. The proposed interface with DOEis shown in greater detail in Figure 12.2. The Director also interacts with the nuclear physics community.and solicits advice from that community through two standing advisory committees: the CEBAF NationalAdvisory Board (NAB) and the Program Advisory Committee (PAC). The NAB reviews CEBAFperfonnance and progress and makes recommendations on scientific policy to the Director. The ProgramAdvisory Committee evaluates proposals for experimental programs and equipment. and. when thefacility is operating. will make recommendations to the Director and the Scientific Director on the efficacyof planned experimental equipment and on the allocation of beam time.

DOE HQ DOE OROSURA

MA, ER. HENP DOE CSO

I I! I

I PAC• CEBAF Director • •I NAB

Scientific Director

ExecutiveIQA/Safet y : ManagementBoard

.-- ------,------I I

Accelerator Physics &CEBAF

Procurement ExperimentalTechnology Design

Project OfficeAdministrative Equipment

Assoc. Directors for ~ ~ Support DesignAccelerator Physics Projoct Manager Assoc. Director for Assoc. Director for

& Technology ~ Administration ~ Research/Users

I I

A & E In-Ho1Jse

Contractor Engineering, Fabrication Vendors& Installation

FIgure 12.1. CEBAF project organization chan.

202

I I I

Project Management

Department of EnergySecretarv

Deputy SecretaryUnder Secretary

Office of Energy Research

Division of Nuclear Physics

Office of High Energyand Nuclear PhysicsDivision of

Construction.Environment & __~ .

Safety

I1 ,

1I1IIII

CEBAF Director

Oak RidgeOperatior.s

Legend

Project Direction

Project Reviews andCommunication

Technical Direction

ManagementDelegation

Figure l2.2.. Proposed management relationships and interrdCC with DOE.

To assist him in managing the construction project and to advise him on its progress. the Directorhas established an Executive Management Board. Members of this board are the Director (chairman).the Project Manager (vice-chairman). the associate directors of all CEBAF divisions (see Figure 12.~).

and the Civil Construction Manager. The head of Quality Assurance and Safety is an observer. Thisboard will meet regularly during the construction project to review designs for all major subsystems.review all design changes that will affect the technical performance of the accelerator. review otherchange requests that exceed the authority of the project manager. monitor construction progress. andrecommend corrective actions if progress is inadequate. In addition, the board reviews documents andreports prepared by the project for submission to DOE. The board will utilize technical experts asconsultants. as needed.

The CEBAF Project Office (CPO) is headed by the Project Manager, who reports to the CEBAFDirector. His responsibilities are to direct the construction activity and to implement the project's technicalscope v.ithin cost and on schedule. To meet these responsibilities. the Project Manager

• selects and supervises project staff;

• coordinates engineering design;

• directs construction activities consistent with guidance from the Director. the Executive Man­agement Board, and DOE;

203

I

CEBAF Design Repon

SURA Council of Presidents

SURA Board of Trustees-----------Executive Committee

SURA PresidentHarry D. Holmgren

SURA Visiting Committee I- _

for CEBAF

CEBAF Director

ProgramAdvisory

Committee

Hermann A. Grunder

--- ----------- ----Scientific Director

J. Dirk Walecka

...........-t: Director"s Staff

NationalAdvisory

Board

IResearchDivision

F. L Gross(Acting)

IAccelerator Physics

Division

C. W. Leemann

IEngineering/Project

ManagementDivision

A. K. Chargin

IAdministration

Division

J. E. Coleman

FJgIII'e l2.3. CEBAF organization chan.

• establishes and implements project infonnation, reporting, planning, and control systems andprocedures coordinated with the Work Breakdown Structure (WBS) and consistent with DOEguidelines;

• informs the Executive Management Board of project progress;

• defines estimating procedures and approves cost estimates;

• prepares project status reports for DOE;

• implements an effective quality assurance and safety program; and

• controls project cost, manpower, and schedule.

204

Project Managemeot

Project direction includes overseeing all in-house engineering. fabrication. test. and installationactivities. and the performance of vendors subcontracted to fabricate and assemble accelerator compo­nents and experimental equipment. Through the Civil Construction Manager. the Project Manager willoversee the activities of the A&E contractor and the general construction contractor(s) performing thecivil construction.

The Project Manager will ensure that detailed system and component designs are consistent withthe requirements stated in the CEBAF Design Handbook. a controlled document. This handbook is theworking document. prepared and updated by the Accelerator Physics and Technology divisions (for theaccelerator) and by the Research Division (for the experimental equipment and experimental areas).that describes the facility design. Changes to the Design Handbook are made only after approval at theappropriate level (see Section 12.3).

The CEBAF organization is shown in Figure 12.3. The Director oversees the activities of thedivisions: Research. Accelerator Physics. Engineering and Project Management. and Administration.

The Scientific Director reports to the Director and assists him in providing scientific leadership forCEBAF. Specifically. he guides the research program. maintains CEBAFs interface with the nuclearphysics community. and oversees the activities of the Research Division. On behalf of the user community.and with its close collaboration. the Research Division sets the hardware specifications for experimentalareas and equipment. describes them in the CEBAF Design Handbook. and assures their implementationthroughout the construction project. When the project is complete. personnel in this division will conductin-house research. assist users. and maintain and improve the experimental equipment to serve evolvingscientific needs. Responsibility for these activities lies with the Associate Director for ResearchlUsers.who heads the Research Division and reports to the Scientific Director.

The Associate Director for Accelerator Physics manages the Accelerator Physics Division. Hisresponsibilities include translating the project scope into system requireme!1ts. describing them alongwith their conceptual designs in the CEBAF Design Handbook. specifying and coordinating R&D re­quired for the construction project. reviewing the detailed designs developed by the Project Manager toinsure that they meet the requirements and incorporate the results of the R&D. and participating in the.testing and commissioning of the accelerator as it is assembled and installed. The Associate Director forAccelerator Physics -Nill be responsible for operating the accelerator when the facility is completed.

:be Project Manager also manages the Engineering and Project Management Division. as its As­sociate Director. to insure that he controls construction manpower. He is responsible for cost and schedulecontrol. project reporting. conformance to applicable codes. component engineering. fabrication (in­house or by industry). testing. acceptance. assembly. and installation. and for conventional construction.After the project is compieted. this division will provide engineering and shop support for acceleratoroperations. facility improvements. and physics research.

The Associate Director for Administration is responsible for providing administrative support tothe construction project and the facility. This support includes budgeting. accounting. personnel admin­istration. procurement. travel. plant maintenance. security. and library services. He also insures thatrelevant laws, contracts. regulations, codes. and policies are complied with.

12.3 Method of PerformancePhilosophy

CEBAF intends to manage the conventional and technical aspects of the construction project. Thepreferred approach will be to contract with industry as much as possible, using competitive bidding andnegotiated fixed-price contracts. where feasible. Subcontracts will be phased. as necessary. to insure thatthe construction project proceeds on schedule without incurring obligations in excess of budget authoritygranted by DOE.

CEBAF will employ a prime A&E contractor to produce Title I and Title II designs for conventionalfacilities. This contractor will also assist CEBAF in overseeing the general contractor(s) selected on thebasis of competitive bidding to perform the conventional construction. The general contractor(s) willemploy additional subcontractors as appropriate and as specified in their contracts.

Accelerator components will be fabricated by industry where practical. Component functional per­formance and interfaces will be specified in detail in bid packages. along with necess2I)' quality assurance

205

CEBAF Design Repon

procedures and test protocols. This approach allows industry to use its engineering expertise to developdetailed designs for CEBAF components, and gives the vendor a share of the responsibility for achievingperformance. CEBAF intends to contract with at least two qualified vendors for critical componentsrequired in quantity, and to adjust the orders according to price and performance. Vendor performancein achieving cost, schedule, technical, and quality objectives will be measured according to the samestandards applied to the CEBAF project as a whole.

~eCoDtrol

The project baseline includes a technical scope, a cost, and a schedule, as specified in Table 12.l.All three aspects will be controlled via procedures described in detail in the Project Management Plan.

Figure 12.4 is a flow chart depicting the baseline control process. There are three progressive levelsof change approval within CEBAF: System Engineers, Project Manager, and Director. Each SystemEngineer is in charge of one level 2 element of the WBS. Approval authority at each level within CEBAFwill be agreed upon with DOE and specified in the Project Management Plan.

-Within authority--Exceeds authorityTESCA-Techni:::al Effect. Cost, Schedule Assessment

Yes

No

No

No

No

No

NoProjectServices

All SystemsEngineerslTESCAl

FJglU'e 12.4. CEBAF baseline control process.

The Project Services Manager tracks and documents all change requests and their disposition andfacilitates their implementation. Three documents are central to the baseline control process:

• Change Request

• Technical Effect, Cost, and Schedule Assessment (TESCA)

• Change Order

The change req!.lcst describes and justifies the proposed change.The TESCA inclr!des an analysis of how the proposed change would affect the cost, schedule, or

technical performance of all systems. The person making the change request must insure that the TESCAis completed.

206

Project Management

The change request accompanied by its TESCA is reviewed and approved or denied by the SystemEngineer. Project Manager. and Director. as appropriate to the magnitude of the cost. schedule. andtechnical effects of the change.

To advise him on proposed changes. the Project Manager has a Change Control Board. whichincludes the system engineers managing each level 2 WBS element, as well as the Engineering Managerand the Civil Construction Manager.

Changes exceeding the Project Manager's authority are considered by the Executive ManagementBoard. which advises the Director on their merits. He approves or denies requests or submits them toDOE for approval. if they exceed his authority. Changes affecting the project scope. performanceobjectives. project completion dates. major milestones. or total cost must be approved by DOE, alongwith contingency allocations exceeding the authority delegated to CEBAF.

Approved change requests become change orders. Change orders are incorporated into projectdocumentation, including cost. schedule, and PMS data bases and the Design Handbook, as appropriate,and they are implemented by the system engineers.

Perfolrmance MeasurementCEBAF will use computerized data bases and software for measuring project progress. The system

will be based on earned value and will operate on microcomputers. It will track the budgeted cost ofwork scheduled (BCWS), the actual cost of work performed (ACWP). and the budgeted cost of workedperformed (BCWP) or earned value. Operational Performance Measurement System (PMS) reports willbe issued monthly to allow CEBAF project management to monitor and manage the progress closely.

12.4 Project Documents and Reporting

The major project documents to be prepared by CEBAF for DOE are listed and described in Table12.2. In addition, CEBAF will submit routine reports documenting project progress (see Table 12.3).

Table U.2CEBAF Project Documents for DOE

Document

Construction Project Data Sheet

Conceptual Design Report

Preliminary Design Report

Project Management Plac

Procurement Manual

Procurement Plan

Quality Assurance Plan

Safety Analysis Reports

Safety Plan

Failure Mode Effects Analysis

I

Purpose

Summarize project, its purpose, cost, and obligation profile

Describe complete conceptual design for projeci, including projectjustification, and summary of bottoms-up cost estimate and schedule

Describe the preliminary design forCEBAF: Title I for conven­tional facilities, engineering designs for technical components

Describe project objectives and management responsibilities, in­terfaces, QA and safety policies, change order system, cost andschedule control system, and performance measurement system

Describe procurement policies and procedures

Describe plan and schedule for procurements

Describe QA policies, procedures, record keeping and auditingplans

Analyze safety issues related to construction and operation ofCEBAF

Describe procedures for insuring project and facility safety

Identify failure modes and analyze their effects

207

CEBAF Design Report

Table 12.3Routine Project Reports Prepared for DOE

Report

Cost Estimate

Contingency Allocation Report

Progress Report Narrative

Fmancial Report

Performance Measurement System(PMS) Report

Frequeucy

Semiannually

Semiannually

Monthly

Monthly

Monthly

CEBAF also prepares operational reports and documents. The C£BAF Design. Handbook will beamong the most important of these. It will be a controlled document. specifying the requirements andfeatures of CEBAF systems. subsystems. and components. It will be the basis for detailed engineeringdesigns. and will be updated as required throughout the project.

Table 12.4 lists the operational reports and documents prepared by CEBAF. maintained in theproject records. and utilized by the Executive Management Board.

Table 12.4Operational Reports and Documeuts

Documeut

Design Handbook

Procurement Schedule

QA Reports

Weekly Critical Path Schedule Outputs

Monthly PMS Reports

Minutes of Executive Management Board Meeting

Weekly Accounting Reports

Monthly Expense Statements

12.5 Project Management Plan

The Project Management Plan describes project management objectives and responsibilities. It alsotreats organizational interfaces, quality assurance, safety, and the performance measurement system.Tabie 12.5 presents the draft table of contents of the Project Management Plan.

208

Project Management

Table u.sCEBAF Project Management Plan

Draft Table of Contents

I. Introduction1.1 Project Management Plan1.2 Other Project Dueumentation for DOE

II. Project Objectives2.1 Project Scope and Scientific Program2.2 Project Schedule Objectives2.3 Project Cost Objectives

III. Management Organization and Responsibilities3.1 Responsibilities of Participants3.2 Communicating Official Project Business

IV. Work Plan4.1 Work Description4.2 Work Execution4.3 Quality Assurance4.4 Environmental Assessment4.5 Safety Analysis and Review

v. Work Breakdown Structure5.1 Overview5.2 Work Breakdown Structure

VI. Schedule and Logic6.1 Schedule6.2 Critical Path6.3 Logic Diagram for Management Activities6.4 Major Milestones

VII. System Requirements

VIII. Cost and Manpower Estimates8.1 Schedule of Obligations and Costs by Fiscal Year8.2 Profile of Obligations and Costs by Fiscal Year8.3 Manpower Estimate

IX. Project Functional Support Requirements

X. Project Management, Measurement, Planning and Control System10.1 Control Systems Overview10.2 Decision Making Apparatus10.3 Work Authorization and Control Documents10.4 Cost and Schedule Perfonnance Measurement Systems10.5 Technical Control Policies and Procedures

XI. Reporting and Information11.1 Documents and Reports Prepared by ORO/CSO11.2 Documents and Reports Prepared by CEBAF Project Office11.3 Design Handbook11.4 Reviews11.5 Meetings

XII. Project Procurement Plan12.1 Introduction12.2 Procurement Planning12.3 Subcontracting for Architects and Engineers12.4 Subcontracting for Construction12.5 Subcontracting for Component Fabrication

209

CEBAF Design Report

12.6 Manpower

Figure 12.5 shows CEBAFs projected manpower. including scientific visitors. for the duration ofthe construction project and into operations. On-site contractor personnel are not included in the figure.When complete. the facility will employ around 250 people. and about 50 visiting scientists will be onsite at any time. The manpower plan during the project allows the staff to build in an orderly way toreach full staff just prior to project completion.

350

93

Operations

92

,,,,,,,,\

\\\

\

\

91

Pre-Ops

89 90Year

Construction

~---..,.._ ......

8887o 86

50

150

100

200

300

250

FIpre 12.5. CEBAF manpower projcctioD (FTE).",-

12.7 Environment Assessment

Environmental compliance. safety. and quality assurance are the responsibility of the CEBAF Di­rector. Each issue will be treated in the Project Management Pian and covered in detail in topicalassessment reports, plans. or manuals.

-'.

Environmental AssessmentThe final draft of the Environmental Assessment Report (DOElEA-0257) shows no significant

environmental impacts. Following is a summary of the minor impacts expected from construction andoperation of CEBAF.

Air quality in the vicinity of the CEBAF site will be slightly decreased by local, short-term emissionsof fugitive dust, vehicle and equipment exhaust. and solvent fumes. An insignificant amount of coolingtower drift may deposit downwind of the site.

Groundwater levels at the site might be lowered b)t dewatering during construction. Proper designand careful operation of the facility (i.e., shielding, containment of the beam. and storage and monitoringof adjacent groundwater prior to release) will minimize any potential for induced radioactivity in ground­water and soils; no significant impacts are expected. A groundwater monitoring and mitigation plan willbe implemented. Storm water runoff and discharge will be controlled by utilizing good construction andengineering practices and a well-planned and well-implemented erosion and sedimentation control pro­gram.

Some erosion and alteration of topography is likely. About 60 acres of hardwood and hardwood­pine forest will be cleared. resulting in a proportional loss of habitat and displacement of resident wildlife.However. the species present are not endangered or threatened.

210

I I

•

Project Maoagement

Ambient noise is expected to increase during operation of construction equipment and. duringCEBAF operation. from compressors and cooling tower and ventilation fans. Increases will not affectany sensitive receptors.

No significant atmospheric releases of radicnuclides are expected from operation of CEBAF.

12.8 Safety Assessment

A preliminary safety assessment indicates the CW linac can be constructed and operated withinacceptable safety and health parameters. Seven major areas were considered in the safety assessment.

1. Radiation Safety: Personnel must be protected from radiation produced by the primary electronbeams and other radiation sources. Protection of personnel from potential radiation sources (e.g.•beam dumps. end stations. linac. etc.) is based on the establishment of radiation controlled areas.defined by appropriate shield walls and barriers. together with a system for occupancy control.This system. coupled with area and personnel radiation monitoring. permits personnel doses tobe controlled to the required levels.

2. Crvogens: Cryogenic systems will be used for the superconducting accelerator. Potential problemsfrom liquified gases include cold burns. explosion. and asphyxiatio:l. CEBAF plans to incorporateappropriate safeguards and practices for storage and transport of liquid gases. Ventilation systemswill be design~ and incorporated into the appropriate machine areas to mitigate explosionpotential as well as asphyxiation potential. Oxygen deficiency monitors will be installed alongwith emergency breathing sets.

3. Chemicals: AIl chemicals and solvents at CEBAF will be used according to currently acceptedgood practice. Vessels containing hydrogen (including deuterium) will be appropriately ventedand secured from combustible materials. Only explosion-proof. non-sparking motors will bealiowed to operate in the vicinity of such vessels. Any tritium brought onto the site will becontained securely and safely.

4. Electrical Safety: Installation ofelectrical equipment and accessories will conform to the applicablenational. state. and local codes. Safety procedures will likewise support and reinforce the re­spective codes to assist in maintaining safe working conditions. Approved safeguards. warningsigns. and operating procedures will be established for laser use and in magnetic field areas.CEBAF will utilize the experience. standards. and practices of other operating particle physicslaboratories to ensure electrical safety.

5. Ere Safety: Ere safety concerns range from ordinary combustible materials to combustible gasesand liquid solvents. Combustible materials will be stored and secured safely using acceptedpractices. The conventional facilities will be designed to Building Officials and Code Adminis­tration (BOCA) standards. The Newport News municipal fire department will be summoned toCEBAF in case of fire emergency.

6. Natural Phenomena: CEBAF facilities will be de."igned to BOCA specifications and standardsthat are applicable to the Newport News area. This will provide the required protection againstnatural phenomena.

7. Environmental Safety: Environmental aspects of CEBAF have been discussed in DOElEA-0257,which showed there to be no significant environmental impacts.

The continuous wave (CW) linac should pose no significant safety hazards to the operating personnelor the general public. In compliance with DOE directives on health and safety. the following actionswill be performed:

A Preliminary Safety Analysis Report (PSAR) on the CW design. as required by DOE Ol'ders5481.1A and 5481.1B, will be initiated. This document will systematically identify safety design criteria.provide analyses of potential hazards of the operation of the facility. indicate proposed measures foreliminating. controlling. or mitigating any such hazards. and evaluate the potential risks of operation.

Preliminary and final engineering design documents will be reviewed by qualified safety experts toensure requirements are met.

Project personnel and subcontractors will be advised of and trained in safety procedures pertinentto their activities.

Operational safety procedures will be developed and implemented based on the analysis of thefacility's operational hazards. (This will be incorporated into the rmal Safety Analysis Report (FSAR)

211

.11 I

CEBAF Design Report

if this report is required.) These procedures will include provisions for a systematic and comprehensiveenvironmental health and safety review of the accelerator and each experimental area (end stations)before they become operational.

Safety issues are discussed further in sectiom 3.4 and 8.1.

12.9 Quality Assurance

All personnel performing activities on the CEBAF project are responsible for the quality of theirwork. The CEBAF Director, however, is responsible for quality assurance (QA) and the formulationand implementation of a QA program based on a written plan. This program will incorporate QA as asystematic and integrated part of the design, procurement, fabrication, assembly, testing, and operationof all components and the entire facility. On QA issues, the Director will be assisted by a QA specialist,who will

• report to the CEBAF Director;

• establish a Guality assurance program based on a written QA master plan;

• assist the Project Manager and system engineers in assessing the risk or consequences associatedwith the failure of each component to conform to design requirements;

• assist the Project Manager, associate directors, and system engineers in developing an imple-menting subordinate QA plans tailored to their specific activities;

• plan and execute regular QA inspections and audits;

• establish and maintain QA records, including audit results;

• assist the Project Manager, associate directors, and system engineers in establishing and main­taining QA records on their activities and components; and

• serve as ex officio member of the Executive Management Board.

212

-

13. Cost

13.1 Introduction

All costs for the CEBAF project were developed from the Work Breakdown Structure (WBS). TheWBS encompasses all elements of the construction phase of the project. The accelerator design wasdeveloped by the Accelerator Physics Division and the EngineeringlProject Mamtgement Division. Theexperimental equipment design was developed by the Research Division and the EngineeringlProjectManagement Division. Both designs were derived from workshops and study groups held at CEBAFand numerous contacts with experts from around the world. The conventional construction designs weredeveloped by the CEBAF staff in conjunction with the architectural and engineering firm of Daniel,Mann. Johnson, and Mendenhall (DMJM).

Level 1 in the WBS is the CEBAF project. Level 2 in the WBS is the responsibility level (Table13.1), which groups technologically similar items of the project under the responsibility of one individual.the System Engineer. who "'ill oversee the realization of those items within the project. Level 3 in theWBS is the machine sector identification, which allows tracking of costs of major machine components.such as linac. recirculation arcs, etc. Level 4 in the WBS represents major assemblies such as powersupplies. helium refrigerator, etc. Level 5 represents assembly-level components, such as vacuum man­ifold. dipole magnet. quadrupole magnet. analog-t<Kligital modules. magnet power supplies, etc. Level6 represents subassembly components. Each item is estimated at the appropriate WBS level; the costestimate was developed using an input data sheet (Figure 13.1) that includes all anticipated cost com­ponents for the item being estimated. These cost components represent the lowest level in the WBS forthat item. If the estimate was made for a sub3SSembly, the cost components are level 7.

Table 13.1Work Breakdown Structure (WBS)

Levell.Level 2.Level 3.Level 4.Level 5.Level 6.Level 7.

ProjectFunctional ResponsibilityMachine SectorMajor AssemblyAssemblySubassemblyCost Components

213

CEBAF DesigD Report

_. __ rlalt&ale'

~-'_._._.-_. .....,-/-/-

""""'1:......... 2:--..,.....,. 3:==::

r• -· _.oo• DeftIE _ •

==-~-..-=='lR--=- ­~ .....-

. -'---o _. I -i4n __ I. I~

~A

•• Deft I bil1-.:rial _AO _.I~

• • I 'ft"-..IED

.-..a: .a... f'ncw_c. (<AO.OOO) " __ .-r_ "_-.....- "-~r I

taIaI. : -

~. 11 ••••••••••••••••••••••••••••••••••••••••••••••••• __

__aacr~ ' __CW' I ' __

FIpre 13.1. Cost estimate input data sheet.

The WBS names for levels 1. 2. and 3 are shown in Table 13.2. The complete WBS is shown inTable 13.3. The details in this table show the levels at which the estimates were made.

Table 13.2WBS Levels 1, 2 aDd, 3

I

214

Levell

Level 2

Level 3

CEBAF

1. Machine Vacuum. Components2. Beam Transport3. RF System4. DC Power System5. Instrumentation & Control6. Experimental Stations7. Cryogenics8. Conventional Facilities9. Project Servi~

.1 Linac

.2 Recirculation Arcs

.3 Switchyard

.4 Injector..5 Experimental Station A.6 Experimental Station B.7 Experimental Station C.8 Distributed Systems.9 Conventional Facilities

Cost

Table 13.3Complete Work Breakdown Structure

WorkBreakdown

Number

o11.1

1.1.11.1.1.11.1.1.1.11.1.1.1.21.1.1.1.31.1.1.1.41.1.1.1.51.1.1.1.61.1.1.1.71.1.1.1.81.1.1.21.1.1.2.11.1.1.2.21.1.1.2.31.1.1.2.41.1.1.2.51.1.1.2.61.1.1.2.71.1.1.2.81.1.1.2.91.1.1.3L1.1.3.11.1.1.3.21.1.1.41.1.1.4.11.1.1.4.21.1.1.51.1.1.5.11.1.1.5.21.1.1.61.1.1.6.11.1.1.6.21.1.1.71.1.1.81.1.1.8.11.1.1.8.21.1.21.1.2.21.1.2.31.1.31.1.3.11.1.3.21.1.3.31.1.41.1.4.11.1.4.2

Element Description

CEBAFMachine vacuum componentsMachine vacw.Jum components,linacCryomoduleCavity PairCavity pair integrationCavityNb parts for cavity pairEnd valvesNon-Nb parts for cavity pa=rCavity pair assemblyIndividual cavity testsCavity pair testsCryo-unitCryo-unit :ntegrationStainless steel vesselsBall screwPiezoelectric translatorRotary feedthroughsPurchased partsFabricated partsAssembly. leak and pressure testsAcceptance testsCryogenic bridge materialsPurchased partsFabricated partsEnd capsModule end capsTest enJ capsCryomodule auxiliary partsFurchased partsFabricated partsCryomodule assemblyToolingAssembly tasksCryomodule laboratory testsCryomodule tunnel commissioningTransfer & insulationAcceptance testsValvesFast-acting valvesVacuum angle valvesSupportsIon pump. 5'tandLinac cryostat standsKlystron supportPumpsIon pump & power supplyRoughing pump

WorkBreakdown

Number

1.1.71.1.7.11.1.7.21.1.7.31.21.2.21.2.2.11.2.2.21.2.2.31.2.31.2.3.21.2.3.31.2.3.41.2.41.2.4.11.2.4.21.2.71.31.3.21.3.2.1

1.3.2.21.3.41.3.4.11.34.21.3.71.3.7.11.3.7.21..e

1.4.11.4.21.81.8.11.8.222.12.1.12.1.22.1.32.1.42.22.2.12.2.1.12.2.1.1.12.2.1.1.22.2.1.1.32.2.1.1.42.2.1.1.52.2.1.3

Element Description

Vacuum accessoriesVacuum piping & fittingVacuum gauge & readout. TCMiscellaneous vacuum partsMachine vacuum com;l., recir. arcsValvesVacuum valves. 1.5 inchesVacuum valve. slide gateCold beam pipeVacuum manifoldsVacuum manifold (four valve)Vacuum manifold (three valve)Vacuum pipe & fittingsPumpsIon pump & power supplyRoughing pumpVacuum gauge & readout. TCMachine vacuum components, BSYValvesVaCl1um valve. switchyard beamlineFast-acting valvesPumpsIon pump & power supplyRoughing pumpVacuum accessoriesGeneral vacuum fittingsVacuum gauge. TCMachine vacuum components,injectorInjector. integrationInjector. cryogenic moduleRefrigerationEnd caps for testingRefrigeration for testingBeam transportBeam transport, linacSteering dipole. 1.506AQuadrupole. 1.5Q12APositionII monitor & electronicsLCW systemBeam transport, recirculation arcsRecirculation ring 1Dipole magnets, ring 1Dipole magnet. 1024Dipole magnet. 1036. ring IDipole magnet. 1072Dipole magnet. 1088. ring 1Dipole magnet. 10120. ring 1Quadrupole magnet. lQ12A. ring 1

215

I I I

CEBAF Design Repon

Table 13.3Complete Work Breakdown Structure (CODtinUed)

-

WorkBreakdown

Number

2.2.1.42.2.1.52.2.1.62.2.1.72.2.1.82.2.1.8.12.2.1.8.22.2.22.2.2.12.2.2.1.12.2.2.1.22.2.2.1.32.2.2.1.42.2.2.1.52.2.2.32.2.2.42.2.2.52.2.2.62.2.2.72.2.2.82.2.2.8.12.2.2.8.22.2.32.2.3.12.2.3.1.12.2.3.1.22.2.3.1.32.2.3.1.42.2.3.1.52.2.3.32.2.3.42.2.3.52.2.3.62.2.3.72.2.3.82.2.3.92.2.3.9.12.2.3.9.22.2.42.2.4.12.2.4.22.2.4.32.2.4.42.2.4.52.2.4.62.2.4.72.2.52.2.62.2.6.12.2.6.2

216

ElemeDt DescriptioD

Quadrupole magnet. lQ12Sextupole magnet, ISlO, ring 1Steering magnet, ring 1RF beam switch, ring 1Magnetic septum, ring 1Magnetic septum, lSp30Magnetic septum, lSplORecircu.latioD ring 2Dipole magnets, ring 2Dipole magnet, 1024Dipole magnet, 1D36Dipole magnet, 1D72Dipole magnet, 1088Dipole magnet, 10120Quadrupole magnet, 1Ql2AQuadrupole magnet. 1Q12Sextupole magnet, ISlOSteering magnet, 106ARF beam switch, 60 inchMagnetic septum, ring 2Magnetic septum, ISp30Magnetic septum, lSp10Recircu.latioD ring 3Dipole magnets, ring 3Dipole magnet, 1024Dipole magnet, 1036Dipole magnet, 1D72Dipole magnet, 1088Dipole magnet, ID120Quadrupole magnet, 1Q12Quadrupole magnet, 1Ql2AQuadrupole magnet, 1Q24Sextupole magnet, lS10Steering magnet, I06ARF beam switch, 90 inchMagnetic septumMagnetic septumMagnetic septum, lSplORecircu.latioD ring 4Dipole magnet, 1036Dipole magnet, 1072Dipole magnet, 1D88Quadrupole magnet, lQ12Quadrupole magnet, lQ12ASextupole magnet, ISlOSteering magnet, 1D6AMagnetic septum, lSp30Beam safety equipmeDtPositionII monitor & electronicsProtection collimator

WorkBreakdown

Number

2.2.72.2.82.2.92.2.9.12.2.9.2

2.2.9.3

2.32.3.12.3.1.12.3.1.22.3.1.32.3.1.42.3.22.3.2.12.3.2.22.3.32.3.52.3.5.12.3.5.22.3.62.3.72.3.82.3.8.12.3.8.22.3.8.32.3.8.42.3.933.13.1.23.1.2.13.1.2.23.1.2.33.1.2.43.1.33.1.3.13.1.3.23.1.3.33.1.43.1.4.13.1.4.23.1.53.1.5.13.1.5.23.1.5:33.1.5.43.43.4.1

ElemeDt DescriptioD

LCW systemBeam switch RF type 10 ftSupports & traysCable trayMagnet stand, dipole &quadrupoleMagnet stand, dipole & quad,4 highBeam transport, switehyardDipole magnets, switehyardDipole magnet, 1D6ADipole magnet, 1D72Dipole magnet, 1088Dipole magnet, 10120Supports & traysSupport alignment eqaipmentCable trayLCW systemQuadrupole magnets, BSYQuadrupole magnet, lQl2AQuadrupole magnet, 1Q12Quadrupole magnet, 1Q24Sextupole magnet, 1S10Beam safety equipmeDtBeam dump, low powerBeam stopProtection collimatorPositionII monitor & electronicsMagnetic septum, lSp30RFsystemsRF system, IiDacKlystroD, DC power, etc.DC power suppliesKlystronsWaveguide componentsDC conditioning modulesLow level CODtrolsContr~1 moduleCoaxial componentsRF diagnosticsMaster refereDce systemMaster oscillatorsReference DistributionSystem coDtrolContact monitorsOutput registersAnalog-to-digital modulesControl power supplieslDjedorlDjedor system

COSt

Table 13.3Complete Work Breakdown Structure (continued)

WorkBreakdown

Number

3.4.1.13.4.1.23.4.1.33.4.23.4.2.13.4.2.23.4.2.33.4.33.4.3.13.4.3.23.4.43.4.4.13.4.4.23.4.53.4.5.13.4.5.23.4.5.33.4.63.4.6.13.4.6.23.4.6.33.4.6.43.4.6.544.14.1.14.1.24.1.2.64.1.34.24.2.14.2.1.14.2.1.24.2.1.34.2.1.44.2.1.54.2.1.74.2.1.84.2.24.2.2.14.2.2.24.2.2.34.2.2.44.2.2.54.2.2.64.2.2.74.2.2.84.2.2.94.2.34.2.3.1

Element Description

Elecrron gunGun power supplyModulatorPumps & vacuum systemChopper cavitiesRF driveApertureBuncher systemBuncher cavityRF driveCopper acceleratorCopper cavitiesRF driveInjector diagnostic30 deg. magnetMagnet power supplyDetectorBeam transportBeam monitorsLens magnetsLens power suppliesVacuum pipesVacuum pumpsDC power systemDC power, linacSteering mag. power supplyQuad power supplyMagnet power supply, 2.4 kWCable trayDC pwr, recirculation arcsPower supplies, ring 1Dipole mag. power supply. 20 kWDipole mag. power supply, 2.5 kWMagnet power supply, 60 WMagnet power supply, 100 WMagnet power supply, 1.5 kWMagnet power supply, 75 kWMagnet power supply BP, 60 WPower supplies, ring 2Magnet power supply, 1.5 kWDipole mag. power supply, 2.5 kWMagnet power supply, 60 WMagnet power supply, 100 WDipole power supply, 75 kWMagnet power supply, 250 kWMagnet power supply, 2.4 kWMagnet power supply, 400 kWMagnet power supply BP, 60 WPower supplies, ring 3Dipole mag. power supply, 20 kW

WorkBreakdown

Number

4.2.3.24.2.3.34.2.3.44.2.3.54.2.3.64.2.3.74.2.3.84.2.44.2.4.14.2.4.24.2.4.34.2.4.44.2.4.54.2.4.64.2.4.74.2.54.2.64.2.6.14.2.6.24.2.6.34.34.3.14.3.1.14.3.1.24.3.24.3.2.14.3.2.24.3.2.34.3.2.44.3.2.54.3.44.3.555.15.1.15.1.1.15.1.1.25.1.1.35.1.25.1.2.15.25.2.15.2.1.15.2.25.2.2.15.2.2.25.35.3.15.3.1.15.3.1.2

Element Description

Magnet power supply, 2.5 kWMagnet power supply, 60 WMagnet power supply. 600 WMagnet power supply. 1.5 kWMagnet power supply, 2.4 kWMagnet power supply, 400 kWMagnet power supply BP. 60 WPower supplies, ring 4Magnet power supply, 20 kWMagnet power supply, 2.5 kWMagnet power supply, 60 WMagnet power supply, 860 WMagnet power supply, 400 kWMagnet power supply, 1.5 kWMagnet power supply, BP, 60 WBeam switch supply & controlPower components, recircPower bussPower cablePower cablesPower, switchyardPower components, BSYPower cableDC bussPower, switchyardMagnet power supply, 2.5 kWMagnet power supply, 75 kWMagnet power supply, 20 kWMagnet power supply BP, 60 WMagnet power supply, 1.5 kWMagnet power supply. 60 WMagnet septum power supplyInstrumentation & controlInstrumentation & control, IinacAccelerator controlsLinac supervisorRF control systemsMagnet control systemsLinac beam diagnosticsBeam pos. & current monitorsInstr. & CDtrl., recirculation arcsRecirculator controlsMagnet local controlsRecirculator beam diagnosticsBeam pos. & current monitorsSynchrotron light monitorsInstrumentation & control, BSYSwitchyard controlsBSY supervisor consoleBSY local control system

217

CEBAF Design Report

Table 13.3Complete Work Breakdown Structure (continued)

WorkBreakdown

Number

5.3.25.3.2.15.45.4.15.4.25.85.8.15.8.1.15.8.1.25.8.1.35.8.25.8.2.15.8.2.25.8.2.35.8.2.45.8.2.55.8.2.65.8.2.75.8.35.8.3.15.8.3.266.55.5.1

6.5.26.5.36.5.46.5.56.66.6.16.6.26.6.36.6.46.6.56.76.7.1

6.7.26.7.36.7.46.7.56.8

6.8.16.8.377.17.1.17.1.1.1

218

Element Description

BSY beam diagnosticsBeam position monitorsI&C, injector controlsInjector supervisor consoleInjector local controlsI&C, distr. sys.I&C. support equipmentI&C supervisor consoleDevelopment stationTest equipmentSafety systemSafety supervisor consoleLocal safety monitor systemsSafety interlock systemBeam loss monitor systemWDTIHSDI moduleRadiation monitorsOxygen deficiency alarm systemCryogenidvacuum systemCryo/vac supervisor consoleCryo/vac local systemsEnd stationsExperimental ball A4-GeV/c magnetic spectrometer,HR1.2-GeV/c spectrometerBeam linesSupport equipmentCounting house A equipmentExperimental ball BLarge acceptance detectorPhoton taggerBeam linesSupport equipmentCounting hause A equipmentExperimental ball C4-GeVIe. spectrometer, moderateres.Variable acceptance spectrometerBeam linesSupport equipmentCounting house C equipmentExperimental stations, distr.systemTargetsComputer equipmentCryogenicsCryogenics, linacCentral helium refrig.Cold box, refrigerator

I

WorkBreakdown

Number

7.1.1.27.1.1.37.1.27.1.2.17.1.2.1.17.1.2.1.27.1.2.1.37.1.2.1.47.1.2.1.67.1.2.1.77.1.2.1.7.17.1.2.1.7.27.1.2.1.7.37.1.2.1.87.1.2.27.1.2.2.17.1.2.2.27.1.2.2.2.17.1.2.2.2.27.1.2.2.2.37.1.2.2.2.37.1.2.2.3.17.1.2.2.3.27.1.2.2.47.1.2.2.57.1.2.2.67.1.2.2.6.17.1.2.2.6.27.1.2.2.6.37.1.2.2.77.1.2.2.87.1.2.2.8.17.1.2.2.8.27.1.2.2.97.1.37.1.3.27.1.3.2.17.1.3.2.27.1.3.37.1.3.47.1.3.57.1.3.5.17.1.3.5.27.1.3.5.37.1.3.5.47.1.3.67.1.47.57.5.17.5.1.1

Element Description

HeliumCryogenic system engineeringTransfer linesTransfer linesSupply transfer lineExpansion boxShutoff valve, ballV-tubeVacuum eqUipmentRelief valvesRelief valveRelief valve. shieldRelief valve, heliumBayonet. pressureT~'3I1Sfer linesTransfer lineTransfer boxesTransfer line junction boxEnd boxTransfer box, dualVacuum valvesVacuum control valveVacuum shutoff valveInstrumentationVacuum equipmentRelief valvesRelief valve, vacuumRelief valve, shieldRelief valve, heliumV-tubeBayonetsBayonet, pressureBayonet, vacuum guardedShutoff valve, ballValves & bayonetsBayonetsBayonet, pressureBayonet, vacuumShutoff valve, ballControl valveRelief valvesRelief valve, vacuumRelief valve, shieldRelief valve, helium lineRelief valve, heliumExpansion jointsTunnel string test refrigeratorCryogenics, expo station ATransfer linesTransfer lines

Cost

Table 13.3Complete Work Breakdown Structure (continued)

WorkBreakdown

Number

7.5.1.1.17.5.1.1.27.5.1.27.5.1.37.5.1.3.17.5.1.3.27.5.1.47.5.1.4.17.5.1.4.27.5.1.57.5.1.67.5.1.77.5.1.7.17.5.1.7.27.5.1.7.37.5.1.7.47.5.1.7.57.5.288.98.9.18.9.1.18.9.1.28.9.1.38.9.1.48.9.1.58.9.28.9.2.18.9.2.28.9.2.3

Element Description

Transfer lineHelium gas lineV-tubeStorageDewar. nitrogenGas storageHelium equipmentPurifierCompressor. heliumVacuum equipmentTransfer boxTransfer linesRelief valve. vacuumRelief valve, lineShutoff valve, ballControl valveBayonetsEnd station cryogenicsConventional facilitiesConventional facilitiesAccelerator facilitiesNorth linac tunnelSouth linac tunnelRecirculation arcs tunnelBeam switchyard tunnelTunnels to experimental stationsExperimental facilitiesExperimental station AExperimental station BExperimental station C

WorkBreakdown

Number

8.9.38.9.3.18.9.3.28.9.3.38.9.3.48.9.3.58.9.48.9.4.18.9.4.28.9.4.38.9.4.48.9.58.9.68.9.799.89.8.19.8.1.19.8.1.29.8.1.39.8.1.49.8.1.59.8.29.8.2.19.8.2.29.8.2.39.8.39.8.3.19.8.3.2

Element Description

Support facilitiesOffice and computer buildingRefrigerator plantSREL building renovationService buildingsMiscellaneous structuresUtilities and site prep.Accelerator coolingElectrical distributionSite utilitiesSite preparationConstruction ManagementPurchase SREL buildingStandard equipment purchaseProject servicesDistributed systemsProject managementProject managementCost monitoringSchedule monitoringQuality assuranceSafety assuranceProject services, snpportOerical servicesBuyersMaterial handlingProject services, tech.System integrationAccelerator system design

The cost estimates include all costs incurred following construction approval to meet the projectspecifications. Specifically, the construction scope includes the accelerator facility, the experimentalequipment, and the conventional facilities, along with associated Engineering, Design, Inspection, andAdministration (EDIA). The construction was as presented in Table 13.3 does not include the pre­construction R&D, or preoperational activities required for commissioning, or physics research. Pro­jections for these costs are presented in Table 13.4.

219

CEBAF Design Report

Table 13.4CEBAF Budget Projection

(Actual year M$)

PriorYears 87 88 89 90 91

Totalto

1991 92

Construction-Related Budget (Obligations)Construction

line item. TEC 0.3 25.0 65.0 65.0 55.0 26.0Preconstruction

Ft~I> 9.5 6.2 6.0 4.5Total project cost

236.3

26.2262.5

Operation-Related Budget29.1 (20.5")Pre-opslops 1.1 7.1 11.9 18.6 38.7 ..

Physics research ... 2.0 3.3 6.0 7.3 18.6 9.2 (6.5), ..Capital equipment 1.2 2.0 3.2 5.7 (4.0)GSa ~ mgmt. fee ... ... 1.4 2.8 2.0 3.5 9.7, ,

Total 31.2 75.5 82.7 76.1 57.4 44.0.FY 1986 dollarstInciuded in preconstruction R&D

13.2 Cost Estimate

All cost estimates include EI>!A and hardware deliverables. FY 1985 dollars are used throughout.except where indicated. Escalation is added after the budget profile is applied. Contingency is addressedat the estimating level and will be discussed in a later section.

The level 2 WBS elements consist of hardware deliverables. except for element 9, which is a servicefunction. They are defined below:

• 1.0 Machine vacuum components consists of all accelerator components that hold vacuum,including RF cavities. cryostats. valves, and vacuum accessori~.

• 2.0 Beam transport system consists of magnets, RF beam switches, magnetic septa. and beamposition monitors.

• 3.0 RF system consists of components required to accelerate the beam. such as the electrongun. klystrons, waveguides, and power supplies., It also includes the local control module.the master oscillator, analog-to-digital modules, and control power systems.

• 4.0 I>C power systems consists of components required to deliver I>C power to the magnets inthe linac, the recirculation arcs, and the external beam lines.

• 5.0 Instrumentation and control consists of components required to provide master controlhardware and software to all the facility systems. The local-eontrol hardware is includedwithineach individual system; however. all of the integrating software is in WBS element # 5.

• 6.0 Experimental stations consists of experimental equipment used for physics research in theend stations, such as spectrometers, targets, large acceptance detector. scintillation counterarray, and beam dumps.

• 7.0 Cryogenics consists of th~ central helium refrigerator and helium distribution system for thelinac and the end stations.

• 8.0 Conventional facilities consists of tunnels for the accelerator, end station buildings for theexperiments, support facilities for the whole project, utilities. and site preparation. CEBAFconstruction management and A~Emanagement services are also included here. This WBSelement includes the purchase and renovation of the SREL building.

220

I I I

Cost

• 9.0 Project services consists of costs associated with the functions of project management. costmonitoring, schedule monitoring. systems integration (including overall physics support).clerical services. reporting. and procurement. The quality assurance and safety issues directlyrelating to project construction activities are costed in this WBS element.

Input data for the cost estimates come from several sources. such as vendor catalogs. discussionswith appropriate vendors. CEBAF staff experience. and accelerator construction experience at otherlaboratories. For non-standard items-the superconducting RF cavity is the major example-CEBAFdouble-checked vendor estimates by considering material cost and fabrication labor.

Tables 13.5(a) and (b) show the details of the cost estimate.

Table 13.5(a)Project Cost Detail (KS)

A. Engineerir.g. Design. Inspection & Admin.1. Conventional construction (at 21.6%)2. Accelerator components (at 28.5%)3. Research equipment (at 27.6%)

B. Construction costs1. L3nd & land rights (SREL)2. Conventional construction

a. Accelerator facilitiesb. Experimental facilitiesc. Support facilitiesd. Utilities and site preparation

3. Technical componentsa. Accelerator componentsb. Research equipment

C. Standard equipment

D. Contingency (at 25%)Total

8.60015.30010.20015,700

67,00030,400

ItemCost

10.80019.1008.400

1,70049.800

97.400

TotalCost

38.300

148,900

1,900

46,900

236,000

Table 13.5(b)Project Cost Profile

by Major Cost Components (MS)

Accelerator componentsAccelerator EDIA

SubtotalResearch equipmentResearch EDIA

SubtotalConventional facilitiesConv. fac. EDIA

SubtotalSREL purchaseStandard equipmentContingencyTotal

67.019.1

30.48.4

49.810.8

86.1

38.8

60.61.71.9

46.9236.0

221

III

CEBAF Design Repon

Table 13.6 shows the cost breakdown from level 1 through level 3 in FY 1985 dollars. beforecontingency and escalation have been applied. The highest-valued item in the table is the conventionalfacilities (WBS 9.9). The estimate is $53.1 M. which was derived after extensive design and cost estimatingstudies by the A&E firm. DMJM. Considerable documentation exists to back up the estimate. Chapter11 describes all of the conventional facilities.

Table 13.6Cost Breakdown Summary to Level 3

(Rounded FY 1985 M$)

Mach. Beam Exp. Conv Proj.vac. trans. RF DC I&C sta. Cryo. fac. servo1. 2. 3. 4. S. 6. 7. 8. 9. Totals

Linac .1 23.2 0.6 12.1 0.1 1.4 9.4 46.7

Recirc. arcs .2 2.2 7.0 2.8 1.0 13.0

BSY .3 0.6 1.2 0.5 0.5 2.9

Injector .4 1.0 1.5 0.3 2.8

Exp. sta. A .s 11.5 1.9 13.3

Exp. sta. B .6 8.8 8.8

Exp. sta. C. .7 7.0 7.0

Distr. sys. .8 1.7 0.8 5.3 7.9

Conv. fac. .9 53.1 53.1

Totals 27.0 8.8 13.6 3.4 4.9 28.1 11.3 53.1 5.3 155.5

The highest-valued linac components are machine vacuum components (WBS 1.1). RF power system(WBS 3.1). and the cryogenic system (WBS 7.1). These items are discussed below.

• WBS 1.1 The machine vacuum components (linac) estimate is $23.2 M. We have completedthe detailed design layouts of this element. which includes the RF cavity and thecryostats. The superconducting RF cavity has been fabricated at Cornell UniversityNewman Laboratory. Manpower time estimates for cavity fabrication are extrapolatedfrom Cornell experience to production. These estimates have been confirmed withfour interested vendors. who have studied the cavity drawings. and are involved inCEBAFs FY 1986 cavity prototyping program. The estimate of the RF cavities' costis based on CEBAF analysis of part-by-part fabrica~;on cost. (::omell experience. andthe cumulative input from all four prototype vendor;:,. The detailed fabrication drawingsfor the cryostats. which will contain the RF cavities. completely identify the cryostatcomponents and their design shapes for estimating purposes.

• WBS 3.1 The RF sysrem estimate is $12.1 M. All of the components within this system arewithin the capability of conventional RF technology. Many vendors and other largelaboratories were contacted to contribute data for the cost estimate.

• WBS 7.1 The cryogenic system (linac) estimate is $9.4 M. All of the components designed intothis system have been fabricated to date; however. the operating experience with coldpumps is limited. Cold compressor testing is underway at Fermilab. Brookhaven Na­tional Lab. and vendors.

222

Cost

Table 13.7 shows the construction budget profile by fiscal year. This table indicates the derivationof the escalation amoucts. using escalation factors provided to the project by DOElDCES staff. TheSREL purchase price of $1.66 M. determined by the GSA. does not have contingency or escalationapplied to it. The construction budget profile is graphed in Figure 13.2. The plot of the cumulative budgetdistribution is shown in Figure 13.3.

Table 13.7CEBAF Construction Budget Profile

(M$)

45.06 42.45 33.79 13.83

6.94 9.55 10.21 5.33(15.4%) (22.5%) (30.2%) (13.85%)

13.0 13.0 11.0 6.84

Construction budget in FY 1985dollars (w/o SREL purchase)

Escalation each year with respectto FY 1985

Contingency

SREL purchase

Total budget, in actual-year dollarsTotal

FY1987

18.59

1.69(9.1%)

3.06

1.66

25.0

FY1988

65.0

FY1989

65.0

FY1990

55.0

FY1991

26.0

Total

153.72

33.72(21.9%)

46.90(25.0%)

1.66

236.0236.0

M$

80

70

60

50

40

30

20

10

086 87 88 89

Fiscal year90 91

o Contingency

~ Escalation

~ Construction

rJglU'e 13.2. Obligation budget profile.

I

223

CEBAF Design Rcpon

MS255240225210195180165150135120105907560453015Ol.......l....e::~__-'-- ........ --L L- ->-- -86 87 88 89

FisC2' ;tear90 91

rJgUre 13.3. Cumulativc obligation budget distribution.

13.3 Contingency Analysis

Following the format of Table 13.6. Table 13.8 shows the contingency allowance at WBS level 3.The contingency allowance was estimated after evaluating four factors: 1) design completeness. 2) un­foreseen and unpredictable conditions. 3) uncertainties and risks. and 4) market conditions-prices andcompetition. Contingency factors are considered on each input data sheet (Figure 13.1). and assignedas a cost component at that level.

Table 13.8Contingency Breakdown Summary to Level 3

(In each column. percentages are on left. rounded FY 1985 MS on right.)

Mach. Beam Exp. Conv. Proj.vac. trans. RF DC I&C sta. Cryo. fae. servo

1. 2- 3. 4. 5. 6. 7. 8. 9. Totals

Linae .130 20 30 15 25 30 30

7.0 .1 3.6 0.0 0.3 2.8 13.8

Recifc. arcs .220 20 15 25 19

0.4 1.4 0.4 0.3 2.5

BSY .320 20 15 25 20

0.1 0.2 0.1 0.1 0.5

Injector .4 30 30 25 300.3 0.4 0.1 0.8

Exp. sta. A .s 19 30 212.2 0.5 2.7

Exp. sta. B .619 19

1.6 1.6

Exp. sta. C .719 19

1.3 1.3

Distr. sys. .825 20 20 21

0.4 0.2 1.0 1.6

Cony. fae. .925 25

12.8 12.8

Total (%) 29 20 30 15 25 19 30 25 20 25

Total (MS) 7.8 1.7 4.0 0.5 1.2 5.3 3.3 12.8 1.0 37.6

224

I

14. Schedule

14.1 Critical Path Analysis

The critical path is established by examining the checkout needs of the nearly completed projectand then the delivery schedule required for the hardware. The critical path logic diagram is shown inFigure 14.1, and the associated timelines are shown in Figure 14.2. CEBAF plans to check out theexperimental equipment in the end stations and the south (second) linac segment in parallel for sixmonths (months 52 to 57). The plan also allows for stand-alone checkout of the south linac for anadditional six months (months 46 to 51). In the preceding year, the cryo-units are installed in the southlinac, and the north linac is in the checkout stage (months 34 to 45). Within the year prior to this phase,the south linac cryo-units are in fabrication and the north linac cryo-units are in installation (mo!lths 2Sto 33). In the preceding year, the cryo-units for the north linac are in fabrication (months 12 to 24). Sixmonths prior to that the vendors are preparing the fabrication facilities (months 6 to 11). Bid time forfabrication of cryo-units is planned at 5 months (months 1 to 5). To be able to begin installing the northlinac hardware, site preparation bidding and construction will be accomplished during the first sevenmonths. Concurrent with the site preparation, the north linac tunnel bidding process (months 5 to 8)will be completed, with the beam enclosure construction (months 9 to 30) scheduled to start upon sitepreparation completion.

According to the described schedule, the project construction duration will be 57 months. Projectcloseout, including payment of all obligations, final project documentation, and acceptance by DOE maytake an additional half year.

A summary network diagram is shown in Figure 14.3, which corresponds to the timelines exhibitedin Figure 14.4. The narrow lines following the dark lines in Figure 14.4 show the float available for theactivity. The klystron delivery schedule shows some slack, but it is not far from the critical path. In theconventional construction area, the initial tunnel is on the critical path to meet our goal for early checkoutof the north linac. Other conventional construction activities are distributed throughout the projectconstruction period and are scheduled to meet the installation needs of the accelerator and experimentalequipment components. Figure 14.5 presents the project's major milestones by WBS, consolidated ona single timeline per WBS. This figure identifies the major tasks within a WBS and shows the correspondingtime frame during which the event is to be accomplished.

22S

•

- ~0\

f;:~

- - - Crillcal - - - Crillcal

8~

If::1

15Itylonftl H Finishng

3/30192110 LF a 03/30/92

r------ ----- ,..----- -----7 8 11 12

Bid Fabricate Fabricate. Install- RF cavilles RF cavilles RF cavilles cryostatshnac 1 IInac 2 Iinac 2

5 LF a 06/01/87 19 LF a01l02/89 12 LF a01l01/90 9 LF a10101190

r------ -----.----- ----- ...... -1 4 6 6 9 10 14 14A

Conllnue Bid Tunnel Install Test and Test and FacStart Tille II tunnels construcllon cr~ostats commlsslon commission opera

design nac I Knac 1 Inac 2/BSV teso LF a12/29/86 5 LF a 09/07187 4 LF a 01/04/88 12 LF a 01/02/89 10 LF a 10/30/89 11 LF a10/01l90 12 LF a09/30/91 6 LFaO

3 3ASite Beam encl.

preparallon preloadlng

3 LFa 10/05/8 7 9 LF a 01/04/88

0-.,

I~6'~,

~

~

9.['2!7

ig.l:l

fiii'Ei'8

!

-~

StartBEGIN PROJECT

Bd Rf CIBID Rf CAVITIES

f Rf CI flINITIAL fAB Rf CAV LINAC

frst CI DELIVERY Of FIRST CAVITY

~RF CI fl

FABR REHAIN RF CAV LINAC 1LiUC 1

BEGIN INSTALLATION OF LINAC 1r RF CI JlINST CRYOSTATS/CAVITIES LINAC

t RF CI F2FABRICATE RF CAVITIES LINAC 2

Telt LilBEGIN TESTING LINAC 1

6' Co,." Li 1.. TESTING AND COHHISIONING LINACB. Linlc 2BEGIN INSTALLATION LINAC 2

~ RF Ca 12INST CRYOSTATS/CAVITIES LINAC 2

J Tilt Li2BEGIN TESTING LINAC 2

CO"" Li2 TESTING AND COHHISIONING LIHAC 2Inl CO"..

INSTALLATION COHPLETEFlc Tune

FACILITY OPERATIONAL TESTEnd

PROJECT COHPLETE

Jan Jul Jln Jul Jan Jul Jan Jul Jln Jul Jln, iii iii I Iii I,

••

••

••

•

I

I

CEBAF DesigD Report

N

.....c~

l' l'N N...............

•

r T of f , f

~--Uc

lI""'" IN ..- ~ =... - ....c ...

'= ~...

~ Il' ! ! ! ! !:A •.. .. ... ... ... t ... =... .. .. • :; ... .. •~ c

... ..~ II! ...

r " ... ...f f

..i

~t¢ ~ I..- : • ........ '" ... .. :; II:

U C ..~.- " .. ..- - ... •

c ... ~ ;::~ ! • •.. ..~

f..c ....I II

l'~ .. ..... ...

.=: ..... ~ .:: ...- ... ...

±

1...-....i

..:.:::.

....

wt ± t~ - ~w ~ ~_ E

r-~L" ± ±. t N _ N.... .. .. - -~ It ~ •

~ = W ~ I ac ~.. c

r i L--.. ~u ...... ..

1I ~,... ~ -... .. ..• u::;; a::.

FIpre 14.3. Summary DCtWOrk diagram.

228

la RF C.

SREL C,)oft

la ST Pr

I .. d Crw

Conc~ ES

ICI _.'.-rI.a IIlh"Toa"l. II... T.. ..,

I" AF PSBel Eel 5.

~.". iii",.

C"C", E5

~ E~..

:'t' P.'.r... T.. FI

RFPSFl

RF Ca FI

Crw FaDi

1(1 ... F'.abl

I.d T"ft

Pr.l:)~

1.... 5tlF'

1"ft St."Tun COft

Tuft l ..dZ

r:""." c.RF Ca F1

Tuft Con2

I." E_C.. Ea 5"

r:.t [ .I .. R.f....

L .. ...c 1

RF Ca FZ

~ Ca n1'1- FaDZ

Cr... F'aDZ

~ PS FZ

In.'' L.l... T.. FZ

1..,." LatCOf'W" Lal

1.. D.... 2

LanK 2

~ Ca IZ

I""" L.. 2... T.. nE"" 5..P.. T.. 13

I". E••

' ••" L .. 2C~... L.IZ

I""C~

J'.c T"...

IEI;IN PROJECT

liD RF C""'ITIE5

RENOV..TION or SREL

liD SIT PllEP WOIlIl

.ID CRYOST..T5

DEVELOP ColNCEpT FOR EXP EOIII",,"T

lID REFRICERaTDR i DISTRIIUTIDN

.ID KLYSTRONS

CDIoTINUE TI TLE II DESICN

liD 8£_ TRMSPORT PARTS

.ID RF POWEIt SuPPLIES

liD END ST..TIONS

CLE..RINc;. (;llUIIIIIl:. c;R"DIIIl:

CDNPLETE CONCEPT FDA ElCP EOIIIp

DESIGN E"PEltIP'ENT"L E')vI~NT

F"I ItEFplCEltaTDA S DI~TltlluTIDN

F..I ..." TIt"NSPDRT PARTS LI C 1

.. """ SPf'I.' F.....IC..T1OM LI C I

INITIAL ""1 RF CAV LIN"C I

,.....,ICaTE CRTOST..,S LINAC I

F.... IC..TE KLYSTRONS LI ....C I

lID INITIAL TUtl'!ELS

8£_ ENCLOSUl'E PRELlIADIIIl:

INSTALL SJIl' Lal

STAll' TUNNEL CIJNSTRUCTIDN

TUNNEL CONSTRuCTIDN

lID TUNNEL LINAC ZDELIVERT lII' Fl"'5T CAVITT

F.... RE.... IN RF CM' LIMC 1

TUNNEL CONSTItUCTIDN LINAC 2.ID EXPERINENT"~ EOIIIPNENT

CONSTRUCT END ST..TIYN5

F..... IC..TE E"PEFINENTAL EOIIIPNENT

INSTALL REFRIc; ~ I:ISTRI. LIIlAC 1

IEc;IN IrlSTALL'" 10M or LINAC 1

F"BRIC"'E RF C"VITIES LINAC ZINST CIt'OST..TS:CAVITIES LINAC

F"BRICATE KLYS~RONS LINAC ZF"!PIC..TE CR'OST..TS LINAC ZCDNP\.. RF PWR SuPPLT FaIRIC..TIQM

INST LINAC 1 '" i ·.~ISC C_'TS

F"••E.... TRANSPOPT PARTS LIMC Z

IEc;IN TESTlIIl: LI"'C 1TEST IN(; AND CDNNISIDNIIIl: LINAC 1

INST"LL LH[ DISTRIIUTION LIMC ZIEc;IN INSTALL..TIDN LI"'C 2INST CRTOST..TS/C..VITIES LINAC Z

INST LINAC Z RF i "1st C_'TS

F... 8£,," TRANSPORT PARTS LINAC Z

iE;:N INSTALLaTION E~P EOIIIP

INST....L 1ST CDNPONENTS

INST<tLL i TEST E""ED' L EDUIPNENT

IEc;IN TEST INC; LINAC Z

TEST III(; AND CONNISIONINl: LINAC Z

INST"LL..TION CONPLETE

F"CILIT. C/PER"TlONAL TEST

PPOJECT CO""LE'[

.. ' ..

•

Jan,

•

Jul

•

Jo"i

•

.... 1

•

•

Jul

Schedule

Jan

•­ •

229

•

~

f~~~.

[

I

Cy JAN JAN JAN JAN JAN JAN87 88 89 90 91 92

1. Machine Vacuum~ellgn Bid Febr Clviliel & crro Unlll

A:uemble i Tell,

:

Install I N Unlc , Till N L1nlCI

Till S L1nlc Op TIIIICommi S Llnlc

, ,

2, Beam TransportDIIlgn Bid Flbrlclle

Install ... N L1nlc = S L1nlC - BSYI E/W Arc

, I

3, RFpIIIlgn Bid Flbrlclll

Inllill N L1nlc Inllall S L1nlc

4. DC Power,..Ign Bid Flbrlcillon

Inllill & Tesl

5. Controls Bid Flbrlclle

conc1g,lue~Inllill N L1nlcllnj. Inllill S L1nlc/E Arel Inllill w/Arc/BSY Enll 511

6. Exp. Equip.Desl n DIIlgn Bid Flbrlclle In51all & Tell

7. Cryogenics~lIlgn Bid Fib Relrla Inllill TI'III

Design N Unlc S un~e E ArcTransfer Lines Accel T.L. Bid Fib Inllill Inlill Inlilil

8. A) Acce!. Fac, (Bm. Encl,) DlIlgnConllrucl B~:Bid N/S Unlc Conllr E/W Arc

B) Exp. Fac. (End Sta,)'B' AlC

DlIlgn Bf·Vi ' Bid I Conllrucllon ,

C) Site Prep.Bid Conllrucllon

D) Roads/Utilities IDlllipn , Bid , Conllrucl ,

E) SREL Bid ConlUucllon Bt?Inllill

SRF

lrltll

~oQ.~

f::l

Appendix ANational Advisory Board for CEBAF

Ernest Henley (University of Washington), Co-ChairmanRichard Kropschot (LBL), Co-Chairman

Lowell BoIlinger (ANL)Richard Briggs (LLNL)

Lawrence Cardman (Illinois)Harold Jackson (ANL)

Steven Koonin (Caltech)Stanley Kowalski (MITlBates)

James LeissHerbert Lengeler (CERN)

John Lightbody (NBS)Boyce McDaniel (Cornell)

Roger Miller (SLAC)Ralph Minehan (University of Virginia)

Ernest Moniz (MITlBates)Helmut Piel (WuppertaI)

Ingo Sick (Basel)Ronald Sundelin (Cornell)Peter Vander Arend (CO)

I

Appendix BCEBAF Workshop Participants

Summer Workshop, June 3-7, 1985

G. Adams (u. of South Carolina)H. Arenhovel (Mainz)R. Arnold (Stanford)J. Becher (ODU)B. Berman (GWU)W. Bertozzi (MIT)C. Bingham (U. of Tennessee)I. Blomqvist (MITlBates)W. Buck (Hampton U.)C. Carlson (Wm. & Mary)D. Cassel (Cornell)D. Cline (U. of Wisconsin)J. Comfort (Arizona State)H. Conzett (LBL)H. Crannell (Catholic U.)D. Day (U. Va.)M. Deady (U. Va.)J. Distelbrink (MIT)W. Donnelly (MIT)R. Eisenstein (U. of IUinois)M. Epstein (California State)M. Farkhondeh (MITlBates)M. Finn (Wm. & Mary)H. Funsten (Wm. & Mary)D. Geesaman (Argonne)C. Guaraldo (INFN)M. Harvey (Chalk River)J. Hiller (U. of Minnesota)H. Holmgren (SURA)R. Holt (Argonne)P. Hwang (Indiana U.)C. Hyde-Wright (MIT)D. Jenkins (VPI)S. Koonin (Caltech)J. Lieb (George Mason)J. Lightbody (NBS)R. Lindgren (U. of Massachusetts)R. Madey (Kent State)J. McCarthy (U. Va.)A. McInturff (FNAL)

Summer Study Group, June I~August30, 1985

I. Blomqvist (MIT)W. Buck (Hampton U.)V. Burkert (Bonn)C. Carlson (Wm. & Mary)C. Chang (Maryland)C. Ciofi degli Atti (Rome)J. Comfort (Arizona)

•

R. McKeown (Caltech)M. McNaughton (LAMPF)H. Meyer (Indiana U.)W. Meyer (Bonn)R. Milner (Caltech)R. Minehart (U. Va.)J. Mougey (Saclay)N. Nayestanaki (MIT)R. Neuhausen (Mainz)J. Noble (U. Va.)J. Nolen (Michigan State)J. Norbury (ODU)B. Norum (U. Va.)J. O'Brien (Montgomery College)C. Papanicolas (U. of lllinois)S. Park (Hampton U.)J . .Parmentola (West Virginia U.)C. Perdrisat (Wm. & Mary)V. Punjabi (Wm. & Mary)C. Rangacharyulu (U. of Saskatch.)E. Redish (U. of Maryland)P. Riley (NSF)D. Robson (Florida State)A. Rosenthal (Western Michigan U.)A. Saba (U. Va.)F. Smoot (Hampton U.)D. Sober (Catholic U.)J. Spencer (SLAC)P. Stoler (RPI)G. Tamas (Saclay)S. Thornton (U. Va.)D. ToiL. Townsend (NASA)H. Weber (U. Va.)M. Weyrauch (U. Va.)1: Williams (Washington & Lee)C. Williamson (MIT)1: Wl5e (U. of Wisconsin)L. Zumwalt (N.C. State)

R. Milner (Caltech)R. Minehart (U. Va.)T. Mizutani (VPI)J. Morgenstern (Sa-clay)J. Mougey (Saclay)R. Neuhausen (Mainz)J. O'Connell (NBS)

233

CEBAF Design Repon

S. Cotanch (N.C. State)H. Crannell (Catholic U.)W. Donnelly (MIT)M. Finn (Wm. & Mary)B. Frois (SacIay)H. Funsten (Wm. & Mary)S. Hanna (Stanford)R. Holt (Argonne)P. Hwang (Indiana)D. Jenkins (VPI)B. Keister (Carnegie-Mellon)L. Kisslinger (Carnegie-Mellon)J. Lightbody (NBS)E. Lomon (MIT)R. McKeown (Caltech)

Technology Workshop, August 1~IS, 1985

R. Alvarez (LLNL)G. Biallas (CEBAF)J. Bisognano (CEBAF)I. Blomqvist (MIT)T. Chargin (CEBAF)C. Chang (U. of Maryland)J. Coleman (CEBAF)D. Day (CEBAF)W. Diamond (CEBAF)G. Doody (DMJM)D. Douglas (CEBAF)'W.. Eukel (Brobeck)F. Gross (CEBAF)H. Grunder (CEBAF)J. Harris (SLAC)B. ~artline (CEBAF)D. Hendrie (DOE HQ)H. Holmgren (SURA)D. Jenkins (VPI)G. Krafft (CEBAF)v. Lakdawala (ODU)C. Leemann (CEBAF)J. Lightbody (NBS)E..Lomon (MIT)A. Maschke (TRW)

Cryostat Worksbop, October 1-3, 1985

G. Biallas (CEBAF)R. Burnette (Brown-Boveri)A. Chargin (CEBAF)G. Chenault (NN Industrial Corp.)B. Diamond (CEBAF)K. Dreitlein (Babcock & WLlcox)J. Fugitt (LBNL)H. Grunder (CEBAF)H. Heinrich (Wuppertal)

234

•••

C. Papanicolas (U. of Illinois)J. Parmentola (W. Va. U.)C. Perdrisat (Wm. & Mary)G. Peterson (Massachusetts)R. Pollock (Indiana)D. Robson (Florida State U.)B. Sapp (MIT)W. Sapp (MIT)I. Sick (Basel)J. Spencer (SLAC)W. Van Orden (U. of Maryland)H. Weber (U. Va.)M. Weyrauch (U. Va.)B. Zeidman (Argonne)

J. McCarthy (U. Va.)B. Mecking (CEBAF)R. Miller (SLAC)M. Molen (ODU)T. Moore (CEBAF)J. Mougey (SacIay)G. Neil (TRW)B. Norum (CEBAF)J. O'Connell (NBS)C. Papanicolas (U. of Illinois)G. Peterson (U. of Massachusetts)H. Piel (Wuppertal)A. Rider (DMJM)E. Ritter (DOE HQ)B. Sapp (MIT)W. Sapp (MIT)S. Schriber (LMTL)T. Smith (HEPL)B. Stein (Ind. Rad.)P. Steinke (DMJM)R. Sundelin (Cornell)J. Weaver (SLAC)R. York (CEBAF)B. Yunn (CEBAF)

M. McAshan (Stanford)B. Meadors (Modem Machine & Tool)T. Moore (CEBAF)G. Neil (TRW)J. Perry (Sulzer Bros. Inc.)L. Phillips (Cornell)H. Piel (Wuppertal)D. Proch (DESY)A. Roberts (Dynamic Engineering)

1: Hodge (Modern Machine & Tool)K. King (TRW)J. Kirchgessner (Cornell)U. Klein (Interatom)S. Kleve (TRW)R. Kramer (DEI-East)S. Loer (TRW)

C. Rode (FNAL)R. Sundelin (Cornell)H. Thompson (NN Industrial Corp.)E. Tilles (CEBAF)T. Walton (Babcock & Wilcox)R. York (CEBAF)

Appendix B

Appendix CSURA Membership

Alabama

Auburn UniversityUniversity of Alabama at BirminghamUniversity of Alabama

Delaware

University of Delaware

District of Columbia

American UniversityCatholic UniversityGeorgetown UniversityGeorge Washington University

Florida

Florida State UniversityFlorida Institute of TechnologyUniversity of MiamiUniversity of Florida

Georgia

Georgia Institute of TechnologyUniversity of Georgia

Kentucky

University of Kentucky

Louisiana

Louisiana State University

Maryland

University of Maryland

North Carolina

Duke UniversityNorth Carolina State UniversityUniversity of North Carolina

South Carolina

Oemson UniversityUniversity of South Carolina

Tennessee

University of Tennessee

Virginia

College of William and MaryGeorge Mason UniversityHampton UniversityJames Madison UniversityNorfolk State UniversityOld Dominion UniversityUniversity of RichmondUniversity of VrrginiaVirginia Commonwealth UniversityVirginia Polytechnic Institute & State UniversityVirginia State University

West Virginia

West Virginia University

237