! !. . . 4+. URL: Vi LLC AT M UNCLASSIFIED ORNL 1 IT 2 . 1 el P 278 . ORN 2-P-218 CP9NF- 6807 –) DTIE-3 : $9P2 1 1962 ION SOURCE AND COLUMN PERFORMANCE AT ORNL* G. G. Kelley and O. B. Morgan 0 Oak Ridge National Laboratory Oak Ridge, Tennessee Most of the experimental work in controlled fusion in Oak Ridge is based on the trapping of high energy molecular ions in a magnetic "bottle" by disso- ciation of the ions in the bottle. For the past six years our group has been concerned with the development of injectors for these experiments. Large DC currents are needed and large current densities are desirable. Since we are not interested in short pulse performance and do not have to match emittance shape to an accelerator, our approach has been difl'erent from that of the people making preinjectors. We have not been measuring emittance but have concerned ourselves with passing a beam through the smallest possible channel in specific geometries. Our beams are pulsed on and off, but generally we are not concerned with the first 100 usec after turn-on or turn-off. On the other hand, we have the problem of dissipation of very large amounts of power at high power den- sities. Two power supplies and two test stands have been used for our studies. One test location is in a small laboratory. It has available up to 100 kv at. 3 amp. Here we have made source studies and have done magnetic deflection beam analysis. The other test facility is in a large open area and is coupled to a 600 kv, 1 amp supply. At this site we have tested accelerator tubes and made - - - studies of beam profile and of targets. "Research sponsored by the U S. Atomic Energy Commission under contract with the Union Carbide Corporation. - - - We have been using the duoplasmatron of von Ardenne. We started with what was practically a Chinese copy of a source described in his books, but we could not get sufficient heat transfer from the tungsten anode insert to the . anode to permit operation at high DC arc currents. A current of about 5 amp was the most that could be maintained reliably for long periods. After we found that we could get good performance with nonmagnetic anodes, we changed to solid copper. A solid molybdenum should work somewhat better, but it is less convenient to use. One other change had to be made to permit operation at steady arc currents greater than 10 amp. The tip of the intermediate elec- trode needs to be cooled very well to prevent heating beyond the curie tempera- . rárne. 3- ..ioo..... mm. same ture. A water cooled copper block is brazed to this electrode just beyond the the than tip. Fig. 1 is a cross sectional view. of one of our present sources. We are . tiv using filaments of a type designed by C. D. Moak of the Physics Division of ' c ORNL. They are made from 40 mil tantalun wire covered with a platinum gauze and wrapped in a bi-filar spiral. Care is taken to see that no part of the f'inished filament is directly on the axis of the assembly. The filaments are then dipped in a barium strontium carbonate solution, the standard cathode dip used in tube manufacture. They require no activation procedure but should be outgassed in a separate system to prevent dirtying the source. Operating cur- T. . , . D E rent is about 20 amp. These filaments have a life in the hundreds of hours when they are kept away from the high energy electrons which stream back through the anode aperture from the accelerating gap, The methods we have used for determining the mass ratios in the plasma TE from the source will be described later. Our main interest has been in the production of molecular ions. When a source is run with a small spacing between intermediate electrode and anode--about 1/16 of an inch--and if the . source 18 run gas starved, i.e., at ourficient arc voltage and at low cnough source pressure that the desired output current is insensitive to variations in arc voltage, then the Ha* ion component 18 at least 60% up to 100 ma and at least 95 ma of Ha* has been obtained at correspondingly higher total cur- rents. The proton yield can be increased by increasing the intermediate electrode to anode spacing to at least 1/4 inch, and by the use of high elec- tron densities--by reduced anode aperture size and inc: 'eased arc current for a given output. A relative proton yield of 90% has been obtained at moderate current, and proton yield seems to be even more favored at higher currents. .... .. ... . -.ir . . . - : ...-. .on. Triatomic ions are produced by operating at high gas pressures and low arc .. voltage. A maximum of 37 ma of Hz* has been produced. . The plasma streaming through the anode aperture in the source consists ..... of rather energetic electron--up to 100 ev--and considerably lower energy ions. The ions have a directed energy of the order of about typically 8 ev. the internetu ministeren hoe meer reinigen we're When a strong electric field is created in the region beyond the aperture, the electrons are repelled and the ions accelerated. There results a plasma- m anentinin in e beam boundary which forms at such a place that the space charge of the ion beam shields the plasma surface from the extracting field. Since the current density of ions in the anode aperture may be as high as 100 amp/sq.cm (It 18 kept high to make the gas efficiency of the source high--typically greater than 90%. ) and since the maximum current density that can be supported with physically realizable extracting fields 1s under about 2 amp/sq.cm, the plasma will expand into the region below the aperture. For a long time we thought that the arrangement using the highest possible field and correspondingly the smallest amount of expansion was most desirable. Recently we have been p ar I . experimenting with large cup arrangemento found to give better beam quniity by the Leningrad group and others. * This arrangement has the further advantage .. . of more reliable voltage break-down characteristics. .. . .. . . The maximum current that can be obtained with a given muekimwn wilowable : ... : .. ... - - . . divergence of the beam after extracting depends only on the extraction voltage. Since the current density for space charge limited current is given, for an - . .. .. -- . - - .. infinite plane beam, and for a parallel cylindrical beam using Pierce geometry, - .-** - ... .-. - by the expression .. - ; . ... j - 5.44 x 10-8 73/2 . . . . M2/2 22 . . - - . * I where M is the mass number, j 16 in amp/sq.cm, Ø is in volts, and 2 is the . .s. ** electrode spacing in centimeters, the total current depends only on ø for a . * '* a given ratio of spacing to beam diameter. We have found that this expression - u * predicts the maximum current density even when the electrode shape is far from that which woulii be expected to produce a parallel beam according to the deri- vation of Pierce. (To get good agreement, however, it is necessary to make an empirical correction which consists of increasing the value of x by the radius of the aperture in the accelerating electrode.) For a spacing of twice the radius of the beam which seems to be a reasonable choice, the maximum current is given by I = 19v3/2 where I 18 in ma and V is tens of kv for protons. The maximum current then which can be obtained in a beam of moderate divergence at 150 kv is slightly over 1 amp. A. I. Solayshkov et al., "Current Injector for a Strong Focused Linac," Proceedings of the Dubna Conference (1963). See M. D. Gabovich, Review Article, "Extraction of Ions From Plasma Ion Sources and Primary Formation of Ion Beams.," struments and Experimental Techniques, No. 2, pp. 195-206 (March-April, 1963). See also N. B. Brooks et al., "Produccion of Low Diver- gence Positive Ion Beams of High Intensity," RSI 35, 894 (July 1964). Yin - . . E SVT A current of 500 mo noo heen extracted from a jource on the timeli toot otand at 100 kv. A current of 400 ma was extracted from this jource contin- uously for a period of four hours. Our 600 kv oupply has a 170 ma bleeder with taps at 150 kv, 300 kv, and 450 kv. We extract from the source plauma at 150 kv and accelerate the beam in three more 150 kv high-gradient, close-spaced steps. An ion from the source is accelerated to the full voltage in approximately 12 inches. Elx. 2. is a cross sectional view of the accelerator tube. There are four alunina insulators six inches high by 14 inches ID. These are fastened to stainless steel ringe by the vinyl seal technique. The ringo provide electrical connec- tion and alignment. Viton O-rings are used between the metal pieces. Skirts of unfilled epoxy molded to the insulator sections provide a large external break-down path. Some of these skirts have been in use for over two years with no trouble due to break-down through the interface between epoxy and ceramic. The electrodes are designed to keep the metal surface area having a strong field at a minimum. We made tests which showed that voltage cleanup problems become much greater when the linear dimension of the surface.at high field 18 large compared to the electrode spacing. A solenoid focusing magnet 18 provided just below the accelerator tube to converge the beam. It 18 cap- able of operation at 2.4 x 105 Amp turns and has six-inch diameter throat. The site of the development of this tube 18 shown in Fig. 3 and Fig. 4. The beam 18 passed into a long cylindrical tank where probe studies and visual observation can be made. By means of an extension on the bottom of this tank the beam can be allowed to travel about 17 feet before striking a target. A profile measurement at a number of points along the beam gave an extrapolated value of 2.9 inches for the dirmcter of the born in the lonc at i total beam .. ... current or 180 me. Thio measurement was made quite some time ago. We bclieve with large pourco cupo now that we can make the beam considerably omaller large bcum currents can . . . < as . .. . be maos analyzed by thio teot facility by making usc of the different focal . . .::. lengtho of the magnetic lens for the dirrerent maos components. These can be focused successively through a small aperture and the power on a target beyond measured calorimetrically. There is an essentially complete sels-neutrai120. tion of the space charge of the beam by electron trupping. The beam profile shows no space charge spreading down to the lowest operating pressure which can be obtained in the system--2 x 10-6 mm Hg. We found, however, that the beam cannot be passed through a crossover in the region below the lens without being seriously disrupted. When a beam component is focused in the observation tank, it appears to get brighter as it gets smaller down to a diameter of . about 1 cm and then becomes less bright but continues to converge to a sharp *5 point. Nothing 18 teen of the beam below this point. When the beam 16 allowed to fall on a target and the lens strength 18 increased, the spot size becomes smaller and smaller and more and more intense down to no more than a pin point. With a further increase in lens strength, the spot disappears. We intend to study this phenomenon in greater detail. . VES We have operated this accelerator at a total power supply drain of 330 mo. At the same time we were able to account calorimetrically for about 300 ma. We do not know what became of the other 30 ms. It did not flow to any of the electrodes in the tube. These currents were measured to be no larger vinerii BNI . than 1 ma. At the same time the power dissipated by the source anode war being monitored. An electron current of 0.5 me at the full energy should have **** been detectable. No power difference was measured when the accelerating voltage was turned on or off. * The beam can be switched on and off by an electronic switcricoi, device in the source-arc supply. This switch is operated through a crater imponoto- multiplier light beum link between ground and the 600 kv level. Toc Dcum cao be turned on and off in less than 2 msec. The instantaneous beam current is transmitted from the 600 kv level to ground through a FM radio link. The system uses two commercial. I'M receivers having AFC, with slight modification. The arrangement has excellent linearity and low drift and has a time resolution of about 13 msec. We have found, as have others, that an accel-decel arrangement can be used to permit neutralization of an ion beam. To be successful the arrangement must provide an electron-repelling field only in a small region around the beam. The exit electrode should be at ground potential and the beam beyor.d should not be able to "see" other potentials. We have neutralized a 50 a, 70 kv beam after the beam had passed through an Einzel lens. A potential of -3 kv, with aperture sizes of 1 to 1.5 inches, was enough to prevent electron 1068 from the beam. .. . . Fig. 5 shows a cutaway view of DCX-2, the largest of the experimental .... ::::.. devices at Oak Ridge. In this device 600 kev Ha* ions are injected through magnietically shielded channel into a uniform 12 kilogauss field. This field increases to 39 kilogauss at points 81 inches either side of the midplane. Ions enter the field 9 inches from the longitudinal axis at such an angle that they have a helical trajectory passing from a point near one magnetic mirror to slightly beyond a corresponding point near the other mirror. The orbit diameter 16 10.3 inches. These ions reflect and return to the injector . after having traveled a distance of the order of 100 meters. During their flight some of them are dissociated citier oy backgrour.d zas ano passam OM by a vacuum arc run between electrodes at opposite erds of the machitoe. We protons resulting circulate between reflection points and precess tie ECE. reflection, but some of them are deposited in the field in such & way this i they do not return to the injector in spite of this precession. The mechanical and magnetic design of the injection channel 1s quite difficuit. ii consists of a hyperco cylinder with overlayed windings which compensate externaiiy for the effect of the cylinder and also cancel the longitudinal component of mag- netic field along the cylinder. The problems of design of the injection duct make a small channel very desirable. The value chosen was i 5/8 inches. Fig. 6 18 a schematic view of the beam path. Stcering magnetic fieids are provided in the pumping chamber just below the lens to compensate for small misaiirn- ments and for the effect of the stray, magnetic field. This fieid is reduced along the beam path by the use of ferromagnetic materials in the eiectrodes and in other hardware wherever possible. It probably is the effect of the smali residual field which has prevented injection of more than 50 ma oi Ha* into DCX-2 in spite of the fact that, as has been said above, 95 ma have been passed through an identical structure in the test stand. The losses probably will be reduced by repair of the 600 kv supply which has been producing an abnormally large voltage ripple. - - 0-40 Among FILAMENT SUPPLY MAGMET suPMY - - - 1 ARC SUPPLY 0-450V COIL DOWN INTERMEDIATE ELECTRODE (ARMCOY www.mw 序LE . . ---- LOADIOS DLL 9000 COOLINS PYREX 110 V ACCELERATON SUPPLY 0-100 KV NI EXTRACTOR PIESE MODIFIED DUOPLASMATRON ION SOURCE 9. XXL ' 7. Su .. P. EU . K 11 "Mr. E . . !, 1 ha MIN WWW EX. +4 . W. 1 VNI Mb px .. TO STAINLESS STEEL memes NICKEL NICKEL- ce tjecalo je porodica .ALII badoo MILD STEEL winning 10,17 KM ION SOURCE PLEAD 25- - Margarannapurnarnumannnnnnnnnnnnnnn oo ooo II 17"" kv Accelerator Tube 300 kV- 450 KV- 150 KVM 600 KV ALUMINUM CORONA RING DORCELAIN INSULATORS 6* LONG -144" 1.0. ia EPOXY SKIGTS 460.2- Fig. 2. 0 -E DOXY . . * YA a Quim w . w 27 . . . 660 LACCHICATOR MIRCM ASSIMA G IN " * 2 . 113 R. SOLO * . . 1 1 M narin -meterinario onde entier...", porn.orm ...... ....... ......vai. 1 WAT. .: i n tensitate the reason is dit niet van de stemmer x vihmone RA . . . RE 1 . mere . ..** . AU .. THY 7 r A . . AC ! 33 ws SV + A T identit e tin e Til pics S Vin . he --- } ar . 2. ! 19 21. ZA * PN 2 1 . . . . . . N " ** SNS iti Fig. 3. Vacuum Chamber for Accelerator Test Stand " LY 4 . . NY . . my . . T WN " . Y- ... 1. , YLE i Det NI '' MINI . w 10 2 SLA MV 2 NON HUN . ** n . " When 1. V s no e . yok SEE - - . ' SA . . --. . . . .. . SU 22 , " 4 . . . arrivat % AULA a . . . i bio . i . 1 + to . met . 1 .. . 1/ .. ws I *** C : in nje 2 1 . . Fig. 4. Top View of Accelerator Test Stund V . XIN U . " N i . IT . LINOL TV Y W . LUT TY W AII, WTF, # ! PINTER ..44 TY NL - IN ' , 7 2. ORT ' 11 W ANY INC SAMO 1. . COLT WS '" P KAFETY SM: YM UT1. ' P CH ., . U u 51 . 2 A P1 - . WNU. UL TA ZAM WWW MA NAME - - - . I Lt .. . :: :: . . . . :: .: . . A . . . . : . . 0 - 1 . 2 . . . . CN I . 1 . Age Fig. 5. Cutriwo View of DCX-7 ORNL-LR-DWG 69712 -200w 0.2 RADIANS, Z = 118 IN. - Z= 100 IN., D = 3£ IN. - - - - Hạt -H* -- Z = 70 IN. -GUIDING COILS El -Hat COOLED MAGNETICALLY SHIELDED DUCT - - . . - . . that -2 = 0, D = 4 IN. DCX-2 INJECTION SYSTEM Fir " T " KL *** 12 In my wr YAPTI S . A . w V WMV i M LAR. DATE FILMED 11/ 24 /164 14 the W LMT in LA . :- A . . T . . OM 2.22 . LEGAL NOTICE The report was prepared as an accrue of Governance opored worth betohet dhe Ontted Skates, dor the counterton, not my person wcthang on behalf of the contenuto A. Maiin my writy or representation, opened or notled, wat reipact to the new racy, completree, or wetteline of who tutoration contend to the rupt, or the one of sy fuormation, mento, madrach, or Doce declared the these regert au sol battestato privately owned e , or B. A s muy Ibabuindles who rupect to the one of, or for denne ruting from the wes of any adoration, porta, nadhod, or procederon tanto reporth Ao wed the ne bove, porno it ou be of her commentarer time to my ptoga or contactor of the Commutaton, of employee of met contractor, to the extent dans kich toplogut Of conductor of the Content , or employee of met attractor proper, dienennetes, Os provide reconto, may tatlondon who comptent or contract with the Constion, or ho mployment me much conductor. A . . . : - - . - - - - - - - - END -