IEEE TAan4action6 on Nuctea2 Science, Vot.NS-22, No.3. June 1975

DESIGN CONSIDERATIONS FOR THE ORNL 25 MV TANDEM ACCELERATOR

J. K. Bair, J. A. Biggerstaff, C. M. Jones, J. D. Larson,J. W. McConnell, W. T. Milner and N. F. Ziegler

Oak Ridge NationaZ Laboratory *Oak Ridge, Tennessee 37830

Summary

Construction of a 25 MV tandem electrostaticaccelerator is now planned as part of a new heavy-ionfacility at the Oak Ridge National Laboratory. Thedesign of this accelerator incorporates several unusualfeatures. The most important of these are a foldeddesign, in which the low-energy and high-energy accel-eration tubes are contained within a single columnstructure, and a digital control system. Motivationsfor these design features are discussed in conjunctionwith a brief description of the accelerator.

Introduction

A new heavy ion accelerator facility' is now beingdesigned for construction at Oak Ridge National Labora-tory. The facility is expected to be built in twophases. Phase I, which is now under way, will consistof a new 25 MV tandem electrostatic accelerator, im-provements to and modifications of the existing iso-chronous cyclotron, ORIC, and a building addition tohouse the tandem accelerator. In Phase I it will bepossible to operate the two accelerators independentlyand also in a coupled mode in which beams from thetandem accelerator are injected into the ORIC forfurther acceleration. In Phase II, another more power-ful booster will be added for coupled operation withthe tandem accelerator. In this paper we will discussgeneral properties of the new electrostatic acceleratorwith emphasis on its unique features. Companion paperspresented at this conference describe beam transportthrough the tandem accelerator2 and conversion of ORICto accommodate injected beams3.

The 25 MV tandem accelerator will be purchasedfrom a commercial manufacturer. General design philos-ophy for the accelerator has been developed in consul-tation with prospective manufacturers during prepara-tion of specifications for the accelerator. Thesespecifications form the basis for the present dis-cussion.

Several criteria strongly influence the design ofthe tandem accelerator. The most important are a ter-minal potential variable in the range 7.5 to 25.0 MVand acceleration of ions in the mass range 12 to 250amu at intensities up to 1 particle microampere. Inaddition, the intended utilization of the acceleratoras an injector requires high reliability, the abilityto coordinate operation of the accelerator with opera-tion of other accelerators, and production of pulsedbeams. ORNL has retained responsibility for beampulsing and bunching so that the specifications, inthis respect, are only addressed to isochronous beamtransport.

In Fig. 1, we show a simplified, preliminary lay-out of the tandem accelerator and in Table 1 we present

*Operated by Union Carbide Corporation for the ERDA.

some parameters& selected from specifications. As canbe seen in Fig. 1, the accelerator has a folded con-figuration in which both "low-energy" and "high-energy" acceleration tubes are contained within asingle column. In this configuration negative ionsare injected into the low-energy acceleration tubeand accelerated to the high voltage terminal whichis maintained at positive potential. In the terminal,the ion beam first passes through a stripper, becomingppsitively charged. After stripping, one charge statecomponent is bent by a magnet through an angle of 1800and injected into the high energy acceleration tube

Fig. 1. A simplified layout of the tandem acceleratorsystem.

1655

for further acceleration back to ground potential.The essential point is that a folded tandem acceler-ator requires only one column structure in contrastto a conventional or "linear" tandem accelerator whichemploys two column structures, one on each side of thehigh voltage terminal.

Table 1

Selected Specified Parameters

Pressure VesselInside diameterInsulating GasMaximum operating pressureGas cycle time

Insulating ColumnLength (excluding terminal)Diameter

Charging System

Major Dead Sections

Lenses Within PressureVessel

Strippers

> 33 ft.SF6125 psia10 hrs. in or out

> 62 ft.> 11 ft.

dual, independent chain-belt with total capacityof 600 uA plus provisionfor later addition of3rd unit.

2 at X, 1/3 and 2/3height.

2 in terminal, 1 inlower low energy deadsection.

gas and foil in termi-nal, foil in high energyupper dead section.

The basic size of the accelerator is determinedby the voltage gradients which occur at the maximumoperating potential, 25 MV. Consistent with its roleas an injector, the accelerator has been sized so thatthese gradients are conservative. In particular, ifthe length of the major dead sections totals 8 feetthe remaining column length will be at least 54 feetand the corresponding maximum average longitudinalgradient will be 25 MV/54 feet = 0.46 MV/ft. (1.51MV/m). The macroscopic radial gradient on the surfaceof the column near the terminal will be no longer than4.14 MV/ft (13.6 MV/m) assuming the values of columnand tank diameter given in Table 1. To place thesegradients in perspective we note that comparablegradients have been employed in the ORNL 5 MV electro-static accelerator since its installation in 1951.

In several respects the new machine will bedifferent from existing tandem accelerators. Thesedifferences include the folded configuration, the useof a large number of ion-optic and beam diagnosticelements at high potential, and the use of digitalrather than analog systems for transmission of controland monitoring information. Since the ion-opticsystem is discussed in another contribution2 we willconfine our discussion to the folded configurationand the control system.

The Folded Configuration

The folded configuration was first proposed byAlverez4 in a note suggesting the tandem accelerator

concept. However, to our knowledge only one suchmachine has been built. This is a 4 MV tandem accel-erator built at the University of Auckland, New Zealandand described by Naylor5. The Auckland machine, how-ever, is a special case since it was built with anexisting column which happened to have an unusuallylarge diameter to length ratio. We have recentlylearned6 that the University of Oxford has proposedconversion of their single stage injector into a foldedtandem accelerator. In general, it seems clear that afolded configuration is only attractive in tandemaccelerators of rather large size, a limitation re-lated to the space required in the high voltage ter-minal to bend ions of interest through a net angle of1800 and, for a simple 1800 terminal magnet configura-tion, the space required for separation of the low-energy and high-energy acceleration tubes. In sub-sequent paragraphs we discuss this problem in greaterdetail using the latter assumption.

Let B be the maximum magnetic field in kGmaxwhich may be achieved in the 1800 terminal magnet. LetQ be the charge state of the component of the beamwhich is bent in the magnet. Let S be the tube sepa-ration in cm. Let M be the ion mass in amu and let Ebe the ion energy after terminal stripping in MeV. Inmost systems E will, to a good approximation, be equalto the terminal potential V in MV.

whenIn this approximation, ions may be transmitted

SQ > 288BMax

(1)

We now distinguish between two cases: fully ionized(generally light) and partially ionized ions. Forfully ionized ions Q/M is approximately equal to 1/2(except for tritium) and equation 1 reduces to

S 565/76 (cm)Max

or

S > 41.1 (cm)

if we assume B = 14 kG. The worst case occurs whenMaxM = 2 and

S > 29.1 /v (cm).

For partially ionized ions, we consider theseparation required to accelerate ions in the mostprobable charge state, Q*, emerging from a terminalgas stripper. To estimate Q, we make the approxi-mation that Q* is equal to the average charge Q anduse an expression consistent with data presented byBetz.7

Q = 0.481 v = 0.481 vZ Z0.55 rv Z0

where v = = 2.188 x 108 cm/sec, v is the ion

velocity in cm/sec, and vr is defined as a reduced

velocity. This expression is valid for vr < 1.0 which

at 25 MV corresponds to Z > 19. Substitution of thisexpression in equation 1 along with parameterizationof M in terms of Z shows that in this approximation,S is essentially independent of terminal potentialand an approximately linear function of atomic numberZ. Fig. 2 shows the magnet radius of curvature p

1656

in reduced insulating gas inventory, reduced size ofthe gas handling and gas storage systems, and reducedbuilding costs.

2) The electrostatic stored energy in a linear tandemaccelerator is app.roximately 40% greater than in afolded tandem accelerator of comparable dimensions.

3) The 1800 magnet required to reverse the beam di-rection in the high voltage terminal serves as anexcellent charge state separator for selection of a

or zsingle charge state after stripping.

4) A tandem accelerator of the size contemplated hereis more easily built in a vertical rather than hori-zontal orientation. With a linear configuration, theroom housing the injection system must be located ontop of the tower which houses the accelerator pressurevessel. This creates problems related to a) transportof personnel, equipment, and control information toand from the injector room, b) inflexibility due tolimited room size, and c) possible differential motionof the injection system with respect to the accelera-tor pressure vessel which in general is supported atits base. In a folded configuration, the injectionsystem is located near the base of the accelerator andall these problems are naturally solved or ameliorated.

5) Bremsstrahlung produced in or near the high volt-age terminal is directed away from both acceleration

0 20 40 60 80 100 tubes.z

Fig. 2. The radius of curvature, p(cm), for theterminal magnet in a folded tandem acceleratoris shown as a function of atomic number, Z,for the case in which ions of the most prob-able charge state emerging from a gas stripperare accelerated. The assumed magnetic fieldis 14 kG.

S/2, for an assumed maximum magnetic field of 14 kG.As can be seen, the required separation for uraniumis approximately 200 cm. This basic result, inde-pendence of V and dependence on Z, would be obtainedwith other strippers and charge state selectioncriteria.

For light, fully ionized ions the required sepa-ration is proportional to the square root of terminalpotential while for heavy, partially ionized ions therequired separation is independent of terminal poten-tial. Conversely, column diameter, as determined byelectrostatic considerations, scales as the firstpower of terminal potential. Thus as the maximumterminal potential is increased there is a cross-over

and the folded configuration may be used without an

artificial increase in column diameter. For the con-

ditions cited, uranium ions in the most probablecharge state emerging from a gas stripper, the cross-

over occurs at a terminal potential of about 22 MV.This is the worse case. With other constraints suchas lower ion mass or higher charge state, the cross-

over points will come at a lower terminal potential.

In our view, the principal advantages and dis-advantages of the folded configuration are as follows:

Advantages

1) Use of one column structure rather than two reducesthe length of the pressure vessel by 20% to 30%. Thislength reduction has important economic consequences,not only in the cost of the pressure vessel but also

Disadvantages

1) The 1800 terminal magnet is heavy, consumes a fairamount of power, and must be operative for theaccelerator to be functional.

2) The 1800 terminal magnet is a source of timedispersion for pulsed beams.

3) With a simple 1800 terminal magnet configuration,the acceleration tubes are considerably displacedfrom the column axis and thus may possibly be more

subject to damage under sparking conditions.

4) For equal sizes and similar construction a single-ended column will be less rigid than a columnsupported at both ends.

We believe that the advantages cited outweighthe disadvantages for tandem accelerators of the 25MV class.

Digital Control System

In the early stages of the development ofcriteria for the design of the control system for theaccelerator it was realized that we had a propitiousopportunity to start with a computer-based controlsystem rather than to try later to computerize a

"conventional" control system. Relative to a con-

ventional system, a computer-based system offers theadvantages of 1) ease of installation and maintenance,derived in large part from the use of multiplexedsignal paths which actually reduce the complexity ofthe system outside the computer itself, 2) inherentexpansion capability, 3) ease of implementing themultiple control consoles required for operation as

an injector, and 4) a more tractable ground loopsituation.

Actual computer control of the accelerator(putting the computer inside feed-back loops) is notseen as an early requirement. Only manual, computer-

1657

120

100

80

C-

0-60

40

20

0

assisted control will be implemented initially. Aftercompletion of the installation we expect to proceed ata deliberate pace to develop computer logging andretrieval of operating parameters to assist in setup,computer surveillance of operating conditions todetect and correct abnormalities, and, ultimately,actual computer control.

Significant features of the system required bythe specifications are as follows:

1) There will be two co-equal control consoles. Onewill be used when the accelerator is being operatedindependently; the other when the accelerator isbeing used as an injector.

2) All control signals and monitoring informationwill be digitized, stored in a common computer memory,and transmitted from place to place over multiplexedserial data links. A large reserve capacity will beavailable.

3) A second computer will be available for off-lineprogram development. It will also have direct accessto the data stored in the control computer. It isplanned to use this computer for the logging, sur-veillance, and control tasks mentioned above.

4) Extensive use of "CAMAC" hardware throughout thesystem will provide for easy system maintanance andmodification.

5) Careful attention will be given to avoidance ofground loop problems.

20

10

E

z

z

5

2

0 40 80 20ION MASS (omu

160 200 240

Performance

In Fig. 3 we show ion energy vs ion mass func-tions based on the assumption that the most probablecharge state component is accelerated after eachstripping process. The functions labeled "singlestripping" are for operation of the tandem acceleratorindependently with only a terminal stripper. Asmentioned above, the accelerated beam intensity is ex-

pected to be as much as one particle microampere witha gas stripper. The function labeled "double strip-ping, tandem alone" is calculated on the assumptionthat the tandem accelerator is operated independentlywith a terminal gas stripper and a foil stripperlocated in the upper dead section. The functionlabeled "double stripping, tandem + ORIC" is calcu-lated on the assumption that the tandem accelerator isoperated as an injector with a terminal gas stripperand that the beam is then accelerated by the ORICusing a foil stripper in the ORIC to perform thecapture function. Maximum intensity in both of thesemodes is expected to be about 0.1 particle microampere.

Acknowledgements

We wish to express our appreciation to staffmembers of the High Voltage Engineering Corporationand the National Electrostatics Corporation for theirhelp and cooperation. We also wish to express our

appreciation to our colleagues at various acceleratorlaboratories for their generous help during the pastyear. We are especially indebted to the staff of theNuclear Physics Laboratory at Daresbury who havepioneered many of the concepts incorporated into our

accelerator.

References

1. J. B. Ball, Bull. Am. Phys. Soc. 19, 1112 (1974).

2. W. T. Milner, G. D. Alton, D. C. Hensley, C. M.Jones, R. F. King, J. D. Larson, C. D. Moak andR. 0. Sayer, "Transport of DC and Bunched Beamsthrough a 25 MV folded Tandem Accelerator", IEEETrans. Nucl. Sci. NS-22, No. 3, 1975 (this volume).

3. R. S. Lord, J. B. Ball, E. D. Hudson, M. L.Mallory, J. A. Martin, S. W. Mosko, G. S. McNeilly,K. M. Fischer, R. M. Beckers and J. 0. Rylander,"Energy Boosting of a Tandem Beam with the OakRidge Isochronous Cyclotron", IEEE Trans. Nucl.Sci. NS-22, No. 3, 1975 (this volume).

4. L. W. Alvarez, Rev. Sci. Inst. 22, 705 (1951).

5. H. Naylor, Nucl. Inst. and Meth. 63, 61 (1968).

Fig. 3. Ion energy (MeV/amu) vs ion mass (amu) forseveral operating modes discussed in thetext.

6. K. W. Allen, private communication.

7. H. D. Betz, Rev. Mod. Phys. 44, 465 (1972).

1658

DOUBLE STRIPPING (GAS-FOIL)

TANDEM + ORIC-,-'TANDEM ALONE

SING

GAS