Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Calhoun: The NPS Institutional Archive Theses and Dissertations Thesis and Dissertation Collection 1955 Experiments on improving the efficiency of the Bevatron Ion Source. Stone, Troy Edward Monterey, California: U.S. Naval Postgraduate School EXPERIMENTS ON IMPROVING THE EFFICIENCY OF THE BEVATRON ION SOURCE TROY EDWARD STONE j/L'^r.fi.'.'-ir','-'-'' •5i«;Jji^^?5;-3rp^F^-^;'^-V'r^^ si':; I. ^^ If. S. School Naval Postgraduate Monterey, California DUDLEY KNOX LIBRARY NAVAL POSTGRADUATE SCHOOL MONTEREY. CALIFORNIA 93943-5002 '">'. EXPERIMENTS ON IMPROVING THE EFFICIENCY OF THE BEVAXRON ION SOURCE by Troy Edward Stone Lieutenant, United States Navy Submitted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE IN PHYSICS United States Naval Postgraduate School Monterey, California 19 5 5 xJ tn irrffi niif"^ 'b'T'. Ji>-_ >lJli- hef+h^ This work the thesis is accepted as fulfilling requirements for the degree of MASTER OF SCIENCE IN PHYSICS from the United States Naval Postgraduate School ip9i si^^dl 9AI 10 H ;4 CUj-C bailttU f Jhrsrj ' ' A ' A BS T RA C No^''"' ^' "''(^'•aH.jflte study has been conducted to determine the focal properties of an ion source that is typical of those used with the Bevatron, Beam-intensity patterns have been obtained for various focusing conditions. of School The effects space -charge repulsion and lens aberration have been investigated. Losses through Coulomb scattering on residual gas molecules have been observed, and a curve showing variation of these losses with variation of energy through the focus electrode has been obtained. Pressure throughout the accelerating column was found to be an important factor, and losses occasioned through slight increases in the normal operating pressure have been recorded. All previous ion sources for the Bevatron have been designed for presentation of a real image to the subsequent injection system. A virtual-image source has been designed, constructed and tested on the Bevatron. 11 as taqoiq i ytxansJnx-mfisa aiDs^ls 9xiT /i- i .noilsvafl snrrmaJsb oxJB-iiadB n99d 9v6xi 89JuD9Jorn sb^ i6t/bxa^i no 9:ijL;aa9T'5 ybuJe A SDiuoa noi naad 9v«rf aatsiiBq ai roiifiiriBv rfJxw asrl ^aorft lo iBoiqyi ex JsxfJ arii Hixv/ .baiBgxJasvnx xiaad ,ToJDfil bsiDw^noo nsad oj noxalx/qai sgiBrfo-aDBqa io ai oD §r A'guoidi aaaaol aesd* lo noxlBxiBv gxixworia avii/D s JnBjToqmx ,b9V"i9ado fanB ot arfJ ffgiroixfJ .t xtb 9d oJ bnirol aBw gnrjBigqo Ismioa odi nx a9a£ nrnx/lOD gnxlBisIgDDB sxfJ ila HgjjOTxll agaaoJ Ygi9n9 lo iuod^uoidi bdnoisBooo aaaeo/ bn£ .bsbioosi n99d gvBrf 9ixjaa9iq lol b9ngxa9b nsgd 9VBx1 nc ' srfl 11;^^ e9Diitoa noi ax/oxvaiq ?aya xioitD9(;nx indupf^adxia ^dS oi dgsnii Ibs sdl no bota-^i bns baibx/iJanoo ,b9ngxa9b naad i ^ io noxJelnaaaiq BBif SDii/oa 9gBaix-Isi/jiiv .notJBvafl xt PREFACE A magnitude machine beam of the output of the itself during the past approaching 10 Bevatron. the output and (or) one way to increase the Refinements in the year have resulted in a current intensity protons per pulse, a most significant improvement; but considerably better performance than this is made continuing and intensive effort is being improving the efficiency of attaining more is envisioned. Increasing Bevatron" ion source of the work has been con- beann, and to this end ducted during the past year by the writer. This experimentation was conducted under the guidance B.B. Kinsey, formerly of the of Dr. Chalk River Atomic Energy Project and during the past year a visiting scientist at the University of California Radiation Laboratory. Frequent recourse has been nnade edge and experience of Bruce Cork is of the Bevatron staff. to the knowl- The writer particularly indebted to them for their enlightenment, assistance, and encouragement during the course The writer wishes to of this work. thank Professor N. L. Oleson of the U. Naval Postgraduate School for his assistance in the S. preparation of this paper. This work was conducted under the auspices Energy Commission while the writer California Radiation Laboratory. «i was attached of the U.S. Atomic to the University of .m tua ax airiT SDnaxraqx.'J b bas 9gbs JnarnsgBixroDna bnB >xlT grrij s airi lol JoorfD^ sJjSirbBTigJao*! Isv&Vl laqfiq alxrfw flora axmrnoD locffiJ noii«ibijH ygt^^nS fixmolxIsO TABLE OF CONTENTS Item Page Title Abstract ii Preface iii List of Illustrations v Chapter I Introduction Chapter II Focal Properties Chapter III Losses through Space -Charge Repulsion, Lens Aberration, Coulomb Scattering, and Residual Gas Chapter IV Design and Test Ion Source 1 of a Typical Ion Source of a Virtual . . 5 10 -Image 16 Bibliography 20 IV JHAT e r-:5>.1I oli/oO o J af rfciBi LIST OF ILLUSTRACTIONS Page Figure \ 1. Block Diagram 2. Cross Section 3. Focusing Action 4. Block Diagram 5. Original Faraday Cup, Shield with 3/8 -inch Hole 25 6. Faraday Cup, Long Shield 25 7. Faraday Cup, Special Shield 25 8. Mechanical Advantage 9. Beam-Intensity Cross Sections, Variable-Focus Electrode Voltage 27 Spreading of Originally Parallel Proton Beami Due to Space Charge 28 11. Ion Trap 29 12. Beam-Intensity Cross Sections, SpaceCharge Effects 30 13. Spherical Aberration 31 14. Chromatic Aberration 31 10. 15. of Ion of Source 21 Arc Channber of of 2Z Three Electrodes Bevatron Injection System of 2 3 .... Cup Movement 26 Beam-Intensity Cross Sections, Aberration 32 Effects 16. 24 Total Current vs. Energy, Coulonnb Losses .... 33 STtig/ 3«OtD 0^5 .Si'lJs it S£ •{ . qaiT 9S it IK '' -^ no: :"f9riq8 ' oxJBxnoirfr ^ -i. CHAPTER I INTRODUCTION An ion source provides the flow of ions which are the particles of an accelerator. The kinds of ions that bombarding are utilized by pres2 1 ent accelerators are protons (,H ), deuterons (,H ), and alpha parti4 cles (^He ). Lithium ions have been used infrequently. In some cases, ions of tritium (,H 3 or light helium ) (^He 3 ) are desired, but these are generally obtained by nuclear processes rather than from an ion source. Sioce the Bevatron designed for the acceleration of protons, this is paper will be confined to ion sources that deliver protons. The original ion source used as that developed by in the Gow and Foster Bevatron was basically the sanae Q] for the Radiation Laboratory, University of California 32-Mev proton linear accelerator. With only minor modifications use. this design, shown in Fig. 1, is still in arc chamber used in this source, shown in detail in Fig. 2, type generally known as a Penning or Phillips ion gauge C2J arc chamber was chosen because it The is of the . Such an has one of the most simple geometri- Aside from simplicity, cal forms its advantages are low-pressure operation, high efficiency, no filament, that will produce a stable discharge. and small physical size. cause the simplest it is Axial extraction of the ions means of withdrawal. is employed be- The three focusing elec- trodes are placed in typical electron-gun fashion, and are referred to throughout this paper as the probe, the focus, and the accelerating electrodes. Atonnic hydrogen gas that source is is to be ionized in the arc chamber of the obtained by passing hydrogen gas through a palladium leak. This atomic gas is imnnediately introduced into the arc chamber, where a strong electric field is utilized to strike an arc between anode and cathode. An electron released from coated aluminum, chamber is is a cathode, which is accelerated toward the anode. made of un- As the entire arc centered in a coil which provides an axial magnetic field of high intensity, radial motion of the electron is constrained by this 10 MC ;. Divo^< -fb-xficfmod 9ri 5 ladi a; • BrfqiB hns OB nl i dri sfo mi/iriiii.-; id .anoloiqio noiisi9J9D-»: aiill ns anoJoiq .{"^Hf) >b ef idi 93^18 r iw i9<p jjoa noi ol baniln* .vtoJBic >ibBfi 3di ro'i [jj oO yd b^qolsvab • arnio^xlBO ylno riixW .gil arii rui iji nwoiie .ngiaab ^i'i ai liSi^b ni nwoi lo rfDwa r ilqrriia taom ,1[i9mBiil on aj& -m lonxn^ bsau t9diaxBdo diM «93 sqy* 10 gninxid^i TiBdD dib adorf ootq Jliw .saiBrioaib elc ififtf Y*'«»i»vxnU Iko axrii ar. di at ar* aoil sbraA lo anoi il Jsrii am bJB aJi ,y: -ad bgyoic i ariJ • oJ fa9Ti9i9"i 9TB bfiB -D9l9 gnijBisieDDB 9ill riJxw lo liBixia ;ta9lqmxa ddl ax t hnB Ji »«i/fi3 n0§-noiiD9l9 iBoxqyJ nl fo93Blq 9t& a9boi;t ,noxx{B£l bns aydq lo noii'j&tlxs iBxx.^ ,8XJ3ol 9xfJ ,ecfoiq »di bb isqBq guoirfi air .BsbotJ 9x1} lo i9df .Assl 16 edi j:^i b9 oidi ti dmBxi »xf:t d oJ 8X *sx(l aBg n930-rbyff afig xiggoTbytf gnraafiq oinx b9o;. yd yf9jBxb9mrnx oicfiolA, bsxiiBJdo ar soTx/oa ax afig oimoiB aidT (9X1 DXTll09i9 bx. I a A j biBv .dboifiBD ^lalsDDB ax .xrn/nxmxjJB bgJBoj fioxifw ) A gnOTlB B nx b9nxB"iianoD ax aotioel^ sAi lo xioxjomt IbI bgisJnsD ai "i9dcn£ct txan9Jnx dgxrf lo The constraint field. may be such that the electron is accelerated into the anode, coasts through the field-free region within, loses envirgy through collisions with the gas, and emerges from the other end with somewhat less energy than gained in it reflected by the electric field, its oscillation of 35 electron several hundred volts, is within the anode during by the magnetic come it it Since the initial energy of the produces approximately ten ions These ions are also constrained its oscillations. field so that their principal to one order of the below ionization energy and falls taken up by the anode. is It is During its axial oscillation. produces several ion pairs, losing energy it volts per pair, and finally subsequently acceleration. constrained again by the magnetic re-enters the anode, and continues field, they is its initial motion Eventually, is axial. end of the anode, where they are accelerated into the Some corresponding cathode. accelerated into the forward of those cathode pass through the exit hole and become useful proton beam. A proton of several hundred volts energy incident on the cathode surface has a finite probability of releasing a secondary electron. trons so released will have life cycles as described above. Elec- No good data are available on the magnitude of the proton- to-secondary-electron ratio for aluminum between five and first electron, Gow and Foster oxide, but ten. If more (llj estimate that it lies ions than this figure are produced by the the discharge will increase in intensity until limited by other mechanisms. A fundamental problem within the arc chamber Let us call free path for ionization. it by the cathode the probability mean and d the nnean length \., an electron path from cathode to anode. is that of the of For a single electron emitted of ionization is approximately d/. , i where d is considerably less than >. . ber were equal to that in the accelerator, hundred meters. If this exceedingly low. If we were take If the X. pressure in the would be equal arc cham- to several the case, the rate of ionization would be d equal to ten centimeters and \. as 200 meters, we will get approximately one ion for every 2000 electrons. And since the exit hole in the cathode through which ions are drawn inch in diameter, the rate of removal of icns only 0.050 low. Considering probabilities of is is generally production and of extraction, we will no a t btt e {1»q n vd get relatively few useful ions. To the state it simply, the pressure in the injection system and through Bevatron must be sufficiently low to allow the accelerating ions to travel fronn the source to the target without colliding with residual neu- On tral atoms. have a high probability the arc by the cathode naust the other hand, electrons emitted chamber in of colliding order quirements are met by to give a (a) with atomic hydrogen atoms within These re- high rate of ionization. pumping system utilizing a high-speed which pernnits relatively high pressure within the arc chamber, and (b) forcing the electrons to follow complicated paths before reaching the anode, as described above. Pulses kilovolt of current through the arc chamber are provided by a three power supply that is capable of three amperes of current during millisecond pulses, such pulses coming twice per second. Thus pulses mixed with ionized molecular hydrogen, are emitted from of protons, the exit hole in the cathode with the The presence of ionized same frequency. molecular hydrogen is, of course, undesirable. Such ions are not acceptable for the Bevatron, which for the charge and tons in the ion mass stream. of the proton, is synchronized and they take the place Several investigators C3J have sought of pro- to in- crease the output of similar sources by introducing impurities that ionize at low voltage, such as mercury vapor. They found impurities generally give a nnanifold increase of the magnetic analysis shows that almost Useful current is all the ions beam arise that these current, but from the impurities. actually decreased. The proton-to-molecular-hydrogen ratio can be varied within however. In general, higher arc currents, linnits, stronger magnetic fields, and lower arc -chamber pressures give the more favorable ratios. der good operating conditions with the type of chambers Un- utilized in these experiments, 75 percent protons are obtained, but under normal operating conditions a The size promise. crease in common figure is of the exit hole in the arc 60 percent protons. chamber is necessarily a conri- Although the number of protons emitted increases with inaperture size, the increase realized is somewhat less than i^ViiJi { aB rii iiov u- >g X9 9 rii i( rri vori directly proportional to the area of the aperture tA2 And • a larger hole results in higher pressure throughout the subsequent system, which is obviously undesirable, and offers a larger object to the focusing sys- tem, which results in a larger image spot. Larger holes also compli- plasma emission surface, such control being accom- cate control of the plished through adjustment of the probe electrode voltage. voltage, the ion. plasma surface protrudes from With no probe the exit hole in tongue fash- Increasing the probe voltage causes the plasma surface to retreat back into the hole under the most favorable emission conditions, until, the surface is perpendicular to the axis of the source. If the holt is nnade too large, however, the plasnna can no longer be forced back prop- erly by probe voltage, the plasma surface The exit-hole diameter current decreases. the best compromise kilovolts from that has been found to be of the three electrodes is shown in Fig. Pro- 3. the exit hole are accelerated to approximately by the probe electrode, decelerated the focus electrode, and accelerated to celerating electrode. beam perpendicular, and 0.050 inch. is The focusing action tons emerging is not 45 20 to about eight kilovolts in 50 kilovolts in the ac- or With these voltage settings, a real image is ob- tained approximately one foot beyond the accelerating electrode. The output beam from Cockcroft-Walton the ion source goes into a colunnn, (refer to Fig. 4), which accelerates the protons to volts. A turning magnet then deflects the buncher, a resonator and decreases its width. to be raised to ten Mev the beam and sends 20 which increases the height The beam then enters and, finally, is turned 500 kilo- of the it into proton pulse the linear accelerator 35 by an electrostatic inflector and injected into the Bevatron proper. Each of the stages subsequent to the ion source constitutes a focusing system, and the connbination great complexity. It is of these systems is a thus to be expected that the compound lens maximum of current reading immediately beyond the ion source or at the end of the Cockcroft- Walton column will not necessarily result current output Indeed, the requirements of the Bevatron are unique, for the Bevatron. and the efficacy of a settings while is it in the highest source can only be determined by trial and error pulsing into the Bevatron. 'X d oliTl CHAPTER II FOCAL PROPERTIES OF A TYPICAL ION SOURCE The purpose of the prove the design some experimentation herein described was not to im- of the arc chamber, whose operation is detail in the previous chapter, but rather to study trolling the beam obtained from such a chamber. for this study is identical to the one employed described in means of con- The source selected Bevatron with the in the exception of the diameter of the focusing electrodes, which are two inches in the former and 1-1/4 inches in the latter. Variation in total output with variation of the focusing -electrode voltages was of interest and was investigated, but the main effort was directed toward obtaining beam-intensity patterns that would show the focal properties of the three electrode system. 1. Operating conditions. The pressure in the arc chamber during these experiments was 125 ± 25 microns, and the pressure at the end of the accelerating elec- trode was about 3 x 10 _5 millimeters net coils surrounding the arc of mercury. chamber varied from Current 0.5 1.2 to depending upon the setting needed for stable operation, and in a nnag- amperes, this resulted magnetic field strength within the channber varying from 1000 gauss. in the 500 to Voltage across the arc during a pulse was of the order of 300 volts, with the setting of the arc pulser varying from one to two kilovolts as required for stable operation. to three Arc current varied fronn one amperes, depending upon the requirements of the experiment. High-current arcs over long periods necessitate frequent replenish- ment of the oxide layer on the aluminum cathodes. running the arc on oxygen for a half hour or so. the aluminum cathodes are found bides of aluminunn, on their polished. Complications what variable, but within ten percent. in of This is done by And, from time to time, form compounds, believed to be carennission surfaces and must be removed and this sort nnake settings for a stable arc someto most cases results obtained were reproducible iTjec; li fit&w bfi£ "^ ^ 5SI iuod^ 89 npjBm ao mo -jfriJ • r ?> s ** "» •* n hn A '=-f")rfi,. - '^•f r f <^ -or if oJ Measurement techniques. 2. The first device day cup with making used for current measurements was a shielded Fara- 3/8 -inch entrance hole (see Fig. a It was found bias would decrease current readings 15 percent. still Such a fluctuation that electron currents of the electron below the in the readings. Be- obtained, the magnitude currents must be known and their effect minimized. The purpose formed on major role a beam current readings can be fore meaningful 30 percent current readings showed clearly of were playing Additional positive farther, at a decreasing rate, 90 volts bias resulting in current readings no-bias figure. that would de- the cup six volts positive with respect to its shield crease current readings by as much as with 5). of positive bias is to draw back secondary electrons the cup and prevent their passing out the entrance hole, being accelerated back through the source, and causing erroneous beam- A current readings on the high side. tive should be sufficient to decrease in The further decrease dicates that secondaries drawn all secondaries to the cup, and the observed current for such a bias to this action. ing reclaim bias of the order of six volts posi- to the formed on in is almost assuredly due current with larger biases in- the lip of the entrance hole cup and causing erroneous readings In order to provide a long field-free region and of secondary electrons formed Fig. 6 was constructed. at the or^ the Tests on this device showed that multitudinous Ekraday cup were obtained for a all to be So many shown holes makes it it 7 in current readings, was developed and has been used for the bulk The field-free region between the two entrance unlikely that secondaries formed on the lip of the outer hole will find their cup makes Clearly this was than before. in Fig. of the data collected. electrons recorded that negative current readings To minimize both positive and negative errors a device of the shielding of these but an axial position of the cup. much poorer design the effect entrance hole, a device shown in cup when the cup was inclined to the beam. into the low side. minimize secondary electrons were formed on the inside surfaces came were be- way to the cup. A positive bias of six volts on the unlikely that secondaries formed on the cup will escape and gc back through the source. Readings were found to vary less than tHo'^:} irf^m-?- •h bf- nw '^w Bfl'rd :}fl BB ft \ .r^ fe'f il >a .0 :Tr ^TOTTT S" iO 15 percent for the range of biases up 90 volts. to Small permanent magnets placed around the entrance hole device might further reduce errors. magnetic uTsing a field of several of his target cup. Thonemann CbI] of this reports success hundred gauss parallel to the surface secondary electrons, which In this strong field the are relatively slow, are immediately sent back to the cup. 3. Beam-intensity patterns The measuring device described above was mounted on the end rod (see Fig. means which passes through a vacuum seal and provides a 8), of positioning the be pushed into the from of a cup within the vacuuna chamber. vacuum chamber The rod can until the cup lies only two inches or pulled out until the cup the exit of the accelerating electrode, vacuum -chamber wall and is 15 inches from the accelerating electrode. When in the latter position, the exposed end of the rod moves 5.6 inches for each inch of movement of the entrance hole lies next to the on the cup shield. Thus by plotting positions of the free end of the rod, one obtains quite accurate knowledge of the position of the entrance hole within the chamber. ings for beam To take advantage of this magnification, all read- intensity patterns withdrawn position. of the individual The patterns Total were made with beam currents were the cup at its fully obtained by integration currents throughout the observation plane. that are shown as Fig. 9 were obtained by leaving all settings constant while varying the voltage on the focus electrode. is seen to be a sharp focus at five kilovolts with the pattern flattening and spreading out for voltages above and below this setting. mum ampere through an aperture 0.042 inch A The mini- size of the pattern is of the order of one -half inch in diameter, and for this pattern the central peak carries a current is There peculiar property of these patterns in of O.IZ nnilli- diameter. is the definite noticeable for focus -electrode settings of 6.2 ring effect that and 7.1 kilovolts. Similar rings are observed with a thick lens in conventional optics, but there they are most prominent when the observing plane lies outside the focal plane of the lens. The patterns herein obtained show pro- nounced rings only when the observing plane lies inside the focal plane £99. .€ OffR h /Jno h. ^ <»rft moil lyt/os laoq larnsvorn sH:f "io eU n^tifif om Ic; ga'tvB9l 's yd Ja quo >o !6Tir'>o(K et'r.'f lUt\ &tr Jb q^ / ban; yif^TTf/.o rrrB^d 1/ Tworfff • is *>>iB l^rfl «»rrc»r»> nwBtbd*rw en Ji/o ;70T:0| «!nfbi39Tqa bn« aid* Tol bne • I?*! : • 10 'jvr»rf? no qifida s ad toi s^f' t<>a arfi enx£Jdo ©no ^'cr -i .allov. boi aflioiiJBq y-trert^^nr -rtBdd foTt 8§ni bsnrjsJo 3e gnii>5T9 I*? '>r{i Hnoiiieoq p >fft .? aaii it •!«.'• system. of the lens Similar rings were noted other times when data They are even more noteworthy were recorded. of this type at virtual-image source, which described is in Chapter IV. Probe electrode length. 4. It was believed the distance from the probe hole might be greater than necessary and that a shorter might result the probe it for the was to sharper focus. drift distance within To carry out this experiment, determine whether displacement of the magnet rear would adversely affect the operation of the arc chamber. to the was found the necessary first in a electrode to the focus that It moving the magnet back one and one-«ight inches decreased beam current by Runs were also made 25 percent. effect displacement of the what to see magnet has on proton content of the beam. This was done with a magnetic deflecting device which permits separate readings of proton and molecular hydrogen currents. magnet displacement did It was found that not significantly affect the ratio of protons to molecular hydrogen. A probe 5/8 inch shorter than the usual one was constructed and The arc chamber was moved toward installed. ing distance. The magnet, which the probe a lies flush against the correspond- back of the moved forward, but the effect of this variation of strength through the arc chamber had been discounted as noted source, could not be field above. Intensity patterns for this arrangement disclosed a decidedly poorer focusing ability for the system, with focus -electrode voltage of little The with less "pancake" fashion A system notably improved is indicates that the spreading out of the mainly due to length was installed. beam A probe one it was decided to explore inch longer than the custo- Intensity patterns taken with conditions duplicating previous tests with the customary probe indicate that current decreases slightly and focusing ability somewhat. Neither effect in space-charge effects. shorter probe was decidedly detrimental; the operation of a longer one. mary for additional patterns with the shorter fact that the focusing ability of the beam current patterns the rule and use in sharpening them up. Lower arc currents were used probe. flat is of the pronounced enough beam system improves to establish whether the longer probe is preferable; the likelihood is that either will serve equally well. Probe -electrode aperture 5. The size of the size. probe hole that lies over the arc -chamber exit hole has customarily been about 0.1270 inch in diameter. determine the gross effects to this hole by 50 percent. was tested by a factor first, A probe from seemed desirable and decreasing the size beam current was down normal that realized with the hole. Tests with 0.0760 -inch hole were equally unrewarding, the total beam current in this case being down by a creases in beam current are great enough factor of three. size probe hole as in the neighborhood of 0.1270 inch. to establish the These de- optimum Probe -tip and exit-plate shapes. 6. to form is customarily a 90 The probe has customarily been machined its entrance hole at the apex. The probe hollow cone. tip of the Foster fits probe spaced about [^IJ beam Tests were made ventional 90 The exit plate a l/l6 inch from the exit hole. Gow and probe. to strike the probe. with the exit plate essentially flat and with the con- No noteworthy a hemispherical surface on its end hemispherically shaped exit plate seemed cone with concentrically into the exit plate with the alteration in beam-current out- put or in focal properties of the system was observed. sign 90 report that less acute cone angles cause an appreciable fraction of the it of 0.1890 inch diameter with a hole of and under optimum conditions of eight a probe having a of enlarging It to be as good in all A probe was developed and used with of slightly larger radius. with a This de- respects as the others, and in addition gave less trouble through sparking between the probe and exit plate. 9/U f iJ lo oaie sriT d diixnmslsb ol yd esq 05 U '. .T£9d tq 95X8 !3na ... aJrl ... ,.»j r,i-i qrJ-adoT*^ adotq 9dT , ...,j ,,., .^t'' enoo wollori JToqsi aol {_-., .m 919W '^ifn'xq <^ 0^ ....-.- V-.c . Jf I'" _'dc F' ' ' rr-i i.'lOV s !•{ ..•rr«oit«ii f? ., -^1 1i CHAPTER III LOSSES THROUGH SPACE-CHARGE REPULSION, LENS ABERRATION, COULOMB SCATTERING, AND RESIDUAL GAS The desired product distortions. It is beam occur while only the of this focusing it is A being focused. is an innage free of minimum further desirable that a the extent of the object if system attenuation of the perfect image of the path is so be obtained and image as well as the inclination image -forming proton rays are exceedingly small, of the protons is uniform, and may if if of the velocity the proton concentration at all points small that the mutual repulsion of the protons has negligible effect. If these three conditions are not fulfilled geometrical aberrations, chromatic aberrations, and space-charge defects arise, the result of each being to decrease useful obvious source of image imperfections is beam A further current. mechanical misalignment or defective construction of the electrodes an (or) the exit aperture. 1. Space-charge repulsion. The ability of a focusing in large measure upon system to obtain a sharp focus is dependent the effect of space -charge repulsion on the beam. Space -charge effects as observed for charged particles have no parallel A in light rays. strong concentration of protons at any point has the effect of reducing the potential or decreasing the speed of the individual protons. In a beam of given cross section, this concentration is in- versely proportional to the velocity. Thus space charge plays a role primarily in low-velocity, high-current beams. spreading Fig. of the beam, resembling Here it produces a the effect of a negative lens (see 10). Let us consider two identical charged particles traveling together at the (1) same speed. They exert on each other two different forces; an electrostatic repulsion 2 '<''){') 10 and a magnetic attraction, because they (2) correspond to parallel currents It may be seen that ^A V 2 e u '^'^ = o (fl energy 20 kilovolts, In the case of protons of meters per second, so F that / F./F„ that x 10 2x10 and we see , same energy. of the For either electrons or protons far 3 -5 This would not be the case for electrons can be neglected. . about is v is only from energy 20 kilovolts, the speed is the charge density is x 10" 1.2 8.5 x 10 their charge density would be 7 Coulomb per hand, protons with the same energy have a we see mutual repulsion For a one-milliampere beam negligible. of the effect of 4.5 x 10 of electrons with an meters per second and naeter. much lower beam of the On Thus nneter. beam same energy, because the other speed, and Coulomb per that the space -charge repulsion of a proton than for an electron is is much higher the charge den- sity is higher and the attractive force is negligible. A complete treatment Gibson in "The Production Their results parallel beam may a with problem has been made by Fowler and of Intense Beams of Positive Ions." be summarized as follows: centimeters in radius; repulsion becomes 0.00280A, of this 2a, 3a, 4a its Consider an D^ . initially radius due to space -charge distances given by 0.00188A, at 0.00358A, where U expressed in volts and elementary positive charges and Fowler and Gibson I in anaperes; M Z the molecular is the mass state that, for instance, for a proton 11 number of of the ions. beam of iaII/i-Tiiq 5ns oJ brjoqasTioo yariJ >1 •t o •» o / > r • t ; rfT ^ rr *) r, 4-^ .i r^-T i:; o n : n sot;-,. .^tAdo ri^dl - rT^ r •'I =) ^jij ^„ fij ff Tr» ; T/1 f, K*^/^ t r i rt /x 'v > ' I -.,fT' rT-, . •, -3 I T I^ rt » K 'vr:: M am ibIuoqIo"-' "rT* it IL tMi St r?'0 '•r-i; T'-ir ^(-•'5 20 kilovolts initially distances are 0.375 centimeter in radius, the corresponding 24.9, and 31 9 centimeters. 16,8, While the above rigorous discussion Spangenberg, and Helm of Q7_] ends of " well be accurate, Field, have shown that positive -ion neutralization space charge can be attained even "ion trap. may This trap (see Fig. 11) at low pressures by means of an consists of suppressor rings at the electrode tubes with a sufficient positive bias to contain positive ions within the electrode. Under these conditions, the investigators re- port, positive ions are unable to leave the tube, and accumulate in such number that the electron space charge, which would ordinarily depress the potential in the center of the tube, is almost exactly neutralized. Their results show, for instance, that an electron current of amperes, traveling a pressure of l/2 the ion trap the This ion trap 36 centimeters with an energy of micron, suffered no spread beam spread is in its 130 milli- 5.7 kilovolts at diameter. Without diameter. to twice its original equally applicable to proton currents, obviously, and the principle further raises the question as described by Fowler and Gibson ions within the tubes (76]] were not taken whether the limiting currents, , have any meaning, for negative Perhaps there are into account. adequate numbers of negative ions within the focusing electrodes of the ion source to neutralize space -charge effects of the beam and lead to very high limiting currents. To study space-charge repulsion patterns were obtained for varying voltages constant. effects in the test source, intensity beam currents with all electrode Experimental results show that smaller currents definitely result in a sharper pattern (see Fig. 12). beam Increasing current from 6.1 to 8.7 milliamperes, for instance, cut down the peak current at the center of the pattern by a factor of almost two while the diameter of the apparent that, pattern increased by a factor of one -half. even though the spreading appears be predicted by the rigorous treatment, there through space -charge effects in this source. employ the "ion trap" to lessen these effects. 12 is It is thus to be less than would considerable loss No attempt was made to w IS IT aj5 finox aq . ov nrf9 2. Lens aberration. The primary contributions to lens aberration for an electrostatic system are from spherical aberration and chronaatic aberration. former, sometimes called aperture defect, The only geometric aber- is the ration which causes unsharpness of the image on the optic axis (refer to Fig. This aberration results in a circle 13). image ture a is the which point, same over confusion about every proportional in diameter to the cube of the aper- is imaging pencil. of the of In magnitude the spherical aberration the entire image. It is caused by the fact that rays passing through the outer parts of the lens are refracted somewhat more strongly in proportion to the separation from the axis, returning to the axis at a smaller distance fronm the lens than rays having a smaller in- clination to the axis. Chromatic aberration results from of a lens acts for a shorter the fact that the deflecting field period of time on a proton of greater velocity. Fast protons are thus less affected and focus the lens than do slow protons. produced by as this action, The effect of spherical A spread at a of the shown by Fig. greater distance from image along the axis is 14. aberration was investigated by obtaining in- beam cur- tensity patterns with identical operating conditions and total rents for varying beam-input cross section. the source operated in the usual fashion. 15), In the first case (see Fig. In the second case, a 3/32 -inch-diameter coUimating hole was placed in the probe from the its tip, beam as and this collimator served to reduce the cross section of Peak current intensity at it entered the focusing system. the center of the pattern coUinnator. was increased by a factor The diameter collimator in as with it of the pattern out. was A large portion of four current available with it in use is down by These experimental results indicate at the beam is maximunn beam of the available a factor of three. that spherical aberration con- tributes considerably to the spreading of the beam. more pronounced by using the half as great with the screened out by such a collimator, however, and the effects are 3/4 inch edges Since aberration of a lens field, it might be expected that increasing the diameter of the electrodes would improve 13 l'X9dB 3/9 ns :[o) 8n9 I .S noil iBfloi q 9;B£rnl a 9tU ai .essq .ri ayf anal ari;} s^eib isilfitna « l£ aixfi xnc •iidds DJcJsmo'iilD iq B "to fj . /ola ob nBdi ansl )DB airii Tiddfi i£3iidilqa enolQ-rq iajsl dx{i yd bsoffboiq k itl\ 9r ol alrrsi -> 90tUOB 9di ^h all ,(5 I moil T«9d 94^ jraidlaya dJ >0 rgtsrr; svB arii 00 i ;)913B lBin9tnit9qxf> sa^riT iaaeo ^aludiii ciB&d adi 9101Tr r,ib sift t I 91B aJ39ll9 snie^sTDni l&di baJogqxa the performance of the source by reducing the spreading of the beam. But increasing the diameter of lenses also reduces their focusing strength, and thus the electrode diameter must be a compromise between these two factors. Chromatic aberration means was found is doubtless present in the over-all problem, but no for independently investigating its effect. Such aber- ration is very likely present in the above space-charge and spherical- aberration experiments, but we assume that it does not play a prominent role in the effects attributed to them. Coulomb 3. scattering. The residual gas nnolecules within the ion source constitute a target material, albeit quite thin, which the protons must pass through. proton in the beam should residual molecules, suffer Coulomb scattering it may Thus it is not a a one of these loss of energy and change of direction would its almost assuredly be great enough for beam. off If it to be eliminated from the useful question of whether this effect causes losses, as be in the case of space-charge effects, but rather of how large such losses are. Throughout the runs where beam-intensity patterns were recorded with focus -electrode voltage the only variable, total beam current was observed effect was not obvious, however, tested. manner, it was noted that less as the focus voltage decreased. until the virtual The -image ion source was This source operates in an acceleration-acceleration-deceleration the final or accelerating electrode (using previous terminology) being used as the variable or "focusing" electrode A plot of total beam current versus accelerating -electrode voltage for the virtual -image source showed a much more noteworthy decrease of current with decreasing voltage. to Coulomb scattering of charged particles by target materials is known vary in such a way that a plot of 1/E versus log total bombarding current is a particles. straight line, To correlate creased with losses due the arguments 1/E where E Coulomb and log total energy beam current the loss in to is the of the as voltage scattering, a plot beam 14 bonnbarding was de- was made with current, where E is the energy riv ioe of the accelerating electrode. This plot shown as Fig. is and 16, its close approximation to a straight line confirms that the major contri- butor to the observed losses The ratio between total is Coulomb scattering. beam current for accelerating-electrode voltages of ten kilovolts and of five kilovolts is a factor of ten. This clearly demonstrates that the losses incurred through long drift dis- tances at low energy are significant. In the conventional source the losses from Coulomb scattering are not so pronounced as those discussed above because the drift distance at low energy, provement pressure in i. e. the length of the focus electrode, is not great. performance of this in the focus electrode source probable, however, is Im- if can be reduced, or voltage increased, or both. 4. Pressure losses. The losses through increased pressure in the focusing system are similar in magnitude to and inseparable from Coulomb scattering losses. But it was of interest to observe the gross effects of pressure increase in the systenn, and data were taken for the conventional and for the virtual -image sources. An increase of pressure from three system caused a decrease of to six beam current of tional source operating at its usual settings. pressure with the virtual -image source of 70 percent of the beam. that low pressure is of x 10 _5 mm in the^focusing 20 percent for the conven A similar increase in in operation resulted in the loss These results confirm a fact well known, prime importance. 15 iJiovoii/T nit. ' « ov ion TO . b/ife jalu 'JO a 8^ • r.'Ofc ei SJjSn sijc/aaanq n • : afidiDsb £ baa s ani: rrfiad *.ion '«norJ sdJ lo JiiSDtaq C lixraadiq wo/ l«rfi CHAPTER IV DESIGN AND TEST OF A VIRTUAL-IMAGE ION SOURCE Experiments conducted during the autumn of 1954 showed that short- ening the length of the Cockcroft-Walton column so as to provide a drift distance between the sovtrce and column of about two feet resulted in a significant increase in new arrangement the exit hole of the arc of the object in the feet from beam current throughout the system. chamber, which With this is the position composite lens system, was situated more than three the entrance to the Cockcroft-Walton column. This suggested the possibility of employing a virtual-image ion source, which would eliminate the drift distance introduced between source and column and optically provide a virtual image of the exit hole at the same position, relative to the column, as the actual exit hole had been with the con- Such a source would permit the physical position ventional source. the exit hole to be at a minimum Cockcroft-Walton entrance. And distance, about if it 18 inches, were possible from of the to duplicate the optical characteristics of the conventional source as seen by the Cock- croft-Walton, considerable gain might be realized as the result of eliminating two feet of drift distance, for drift distance is, of itself, neces- These reasons, and sarily injurious. the fact that a virtual -image source had never been tried on the Bevatron, prompted the attempt to achieve significant improvement in the useful output through use 1, of a of the ion source virtual-image focusing system. Design. The arc chamber was incorporated that had been used for the previous experiments into the virtual-image source (see Fig. 1 ), and the physical dimensions of the housing of the focusing electrodes were retained. Thus the new source could be used interchangeably with the The changes made to old. achieve virtual -image performance were in the length and diameter of the electrodes. A three -inch diameter for the tubes was adopted with the hope that gains from lower pressure and less spherical aberration would be greater than any losses occasioned through u ix son .M weakening One inch was added of the lenses. to the length of the focus electrode, and the accelerating electrode was retained at length by permitting source than before. probe the tip it to usual protrude an inch farther from the end of the The probe length was maintained, but the shape and exit plate were hemispherical instead To achieve its of 90 of cones. a virtual image, the electrodes are energized to provide an acceleration-acceleration-deceleration action on the protons. approximation calculations on the focusing systenn, it Through was predicted that, with 20 kilovolts on the probe electrode and 40 kilovolts on the focus electrode, parallel rays would be obtained for an accelerating electrode voltage of about At some higher voltage, something less 12 kilovolts. 25 kilovolts, the calculations indicated that a virtual image would than be achieved about two feet behind the exit hole. that were to be ennployed, available to insure 2. With the power supplies however, adequate latitude optimum placement of the virtual of settings was image. Test-stand data. A series of runs was made with probe voltage set at 14 kilovolts and Although a crossover focus was expected focus voltage at 40 kilovolts. with about eight kilovblts on the accelerating electrode, total beann current was so diminished by Coulomb scattering the setting for best crossover in this low-energy region that beam have been made was effectively masked and currents were deceptively low. If measurennents could total within the accelerating electrode rather than at a distance two feet beyond, more indicative patterns and higher currents would undoubtedly have been achieved. Settings of ten kilovolts and slightly above on the accelerating electrode gave much higher tering, total currents, as would be expected with less scat- and the patterns were spreading out and flattening in a manner indicative of parallel rays, which would be expected with these settings. No ready means was of a virtual image, and further information on the source could only be achieved by operating 3. available on the test stand for determining positions it with the Bevatron. Bevatron runs. Before new power supplies for use with the virtual -image source were 17 ti&di jjs us jodik lo »^«jit>v a asw 'ni siorn .bnoy >um 9VBg >rfi fan :9j srjxfani breakdown available, a source of the installation of the virtual source. It in the Bevatron necessitated the was operated time with at this acceleration-deceleration-acceleration settings, and the best results were obtained when voltages were such as The image. to give a real best performance available with this arrangement showed the current at the end of the Cockcroft-Walton with the usual source, but total down by a factor or be comparable with that obtained beam two or three. replaced in the Bevatron when to it out of the linear accelerator Therefore, the usual source was was again operative. Later tests with the virtual-image source operating showed a definite maximum maximum from test information a minimum 48 kilovolts-- point at with accelerating voltage at crossover was expected this indicates that the first designed current through the Cockcroft-Walton with accelerating voltage at eight kilovolts, a and a higher in its 18.5 kilovolts and focus voltage fashion- -probe voltage was maximum was 13 kilovolts, 19 kilovolts. at about Since 8 kilovolts, the crossover or real-image minimum a case where the rays were essentially parallel, larger maximum the desired virtual -image operation. Maxi- condition, the and the mum current through the Cockcroft-Walton with the accelerating volt- age in the neighborhood of is somewhat less than the 19 kilovolts was nine milliamperes, which 12 milliamperes obtained The maximum available ditions with the usual source at this position. current at the input to the linear accelerator was which compares even less favorably with the can be obtained with the usual source. exit of the linear accelerator the conventional source the virtual a factor of may 6.0 2.2 milliamperes, milliamperes that Current measurements were ridiculously low, register under good con- hoN^ever. at the Whereas 300 microamperes at this point, image source could produce only eight or ten microamperes, 30 down. The new source was clearly not an improvement. Reasons why the virtual-image source did not get more current through the system are obscure, just as the focusing characteristics of the injection system are, but indications are that the beam was diffuse, certainly in size and probably also in energy, for accelerated effectively on through the injection system. 18 it to too be Smaller tubes • is •0JS d3 ni 9di x^m nj&dli e»»l -iHis iati. lib 8.-) Hi di IS yidf >6 81 or other changes might improve performance somewhat, but it is quite likely the systenn prefers a real -image source to any virtual image source that might be designed. 19 BIBLIOGRAPHY 1. Gow, J. Foster, D. and J. S. , Jr. A HIGH -INTENSITY PULSED ION SOURCE, The Review Instruments, pp 606-610, of Scientific Vol. 24, No. 8, August 1953 2. 3 Penning, F. M. Physica 4, 71 (1937) Hoyaux, M. and COMPARATIVE SURVEY OF ION Dujardin, GUNS, I. Nucleonics Research Laboratories, Ateliers de Constructions Electriques de Charleroi, Belgium 4 Pierce, J. THEORY AND DESIGN OF ELECTRON BEAMS, R. D. 5. Thonemann, P. 6. Fowler, R.D. and Gibson, G. E. 7. Field, C. L.M. Spangenberg, K. and Helnn, R. Van Nostrand Company, 1949 Nature 158, 61 (1946) Phys. Rev. 46, 1075 (1934) CONTROL OF ELECTRON BEAM VACUUM BY IONS, DISPERSION AT HIGH Elec. 121, 8. Comm. 24, No. 1, pp 108- March 1947 Zworykin, V. K. Morton, G. A. ELECTRON OPTICS AND THE ELECTRON MICROSCOPE, Ramberg, John Wiley and Sons, Inc. E. G. J. and Vance, A. W. Hillier, 20 , 1948 :>Jeo •£ T n K noiiT ffT'^/'n: bii6 f^n TWTT .X ,g .? " 1 ^^ " "^ ARC T\ CHAMBER ACCELERATING -- FOCUS PROBE i ELECTRODES'^ - 1 1 1 1 1 30 1 15 1 k V. k.v. 1 1 ^_1_ — . _J_-'- 60 ! MU-9546 Fig. 1 Block Diagram of Ion Source. Diameter for Electrode Tubes, Lengths of Electrodes: probe, focus, 2 in. 2. 5 in. ; ; accelerating, 4 21 in. 2 in. ; J CATHODE ANODE V^ MU3409 Fig. 2 Cross Section 22 of Arc Chamber n ( : 3C U V,<Vj V. j-^ »" ^ v,>v •*" MU-9547 Fig. 3 Focusing Action 23 of Three Electrodes Fig. 4 Block Diagram System of Bevatron Injection 24 ?*^ SHIELD 'i- FARADAY CUP 3/8' I khWIRE MESH I I 0.042 MU-9548 Fig. 5 Original Faraday Cup, Shield with 3/8-inch Hole Fig. 6 Faraday Cup, Long Shield Fig. 7 Faraday Cup, Special Shield 15 VACUUM CHAMBER SOURCE AXIS to SCREEN MU-9549 Fig. 8 Mechanical Advantage 26 of Cup Movement ^ .H Si'i "= 0.172" MU-9550 Fig. 9 Beam-Intensity Cross Sections, Variable-Focus Electrode Voltage 27 MU-9551 Fig. 10 Spreading Beam Due of to Originally Parallel Proton Space Charge 28 9BHHD TRAP ELECTRODES p ^ ^ I -.I.I+ <W^-J MU-9552 Fig. 11 Ion Trap 29 ^ t ^l|H IM ki 6.1 m.a. 8.7 m.a. 1"= 0,172" MU-9553 Fig. IZ Beam-Intensity Cross Sections, Space-Charge Effects 30 .i ^^^ LENS MAGE PLANE OBJECT FAST PROTONS I ginw PRhTON f; MU-£554 Fig. 13 Spherical Aberration Fig. 14 Chromatic Aberration 31 .-^. 103180 ! eWOTOfiq T2A1 -. wnp with l"= 3/32" col I i motor 0.172" MU-9555 Fig. 15 Beam-Intensity Cross Sections, Aberration Effects 32 I \ij} /'i^ — 1000- liJ 100 3 O UJ CO < O CD O 10 20 10 .2 i/e X 30 _3 10^ MU-9556 Fig. 16 Total Current Versus Energy, Coulomb Losses (E is energy of Accelerating Electrode) 33 — JL 19 ST INTERLIB 2847 Stone on ^P^^'^^the efficiency +.<= JH A^i Stone Experiments on imnro-^nn^ efficiency of the bevatron ion source. +''» thesS765 Experiments on improving the efficiency 3 2768 002 02042 2 DUDLEY KNOX LIBRARY •*.^^':-.^