. .. . •' • 1 OF 1 ORNL P 1374 ht . . :. . MICROCOPY RESOLUTION TEST CHART NATIONAL BUREAU OF STANDARDS -1963 here ORN-P-1374 CONF-630935-5 JUL 20 130 in Blect.con-Cyclotron Plasma* Energetic Neutral Injection Into An Electron-Cyclotron Plasma* W. B. Ard A. C. England R. A. Dand. G. M. Haas N. 1. Lazar Oak Ridge National laboratory Oak Ridge, Tennessee, U. S. A. A beam of twenty-keV hydrogen atoms has been directed through an electron- cyclotron plasma (ECP) contained in a magnetic mirror. The observed density is more than a factor of ten greater than that obtainable from neutral gas trap- ping under comparable conditions. The measured currents of charge-exchanged 20-keV protons and the decay of these currents are in experimental agreement with the calculated neutral density inside the ECP. The experiments were per- formed with a large volume (50 liters) ECP in order to ascertain the effective- : ness of plasma shielding against thermal neutral gas. The extension of this work to higher densities of hot ions by the employ- ment of much derser ECP and the effect of such critical parameters as thermal neutral reflux and energetic beam current are discussed. The density, power, and stability bounds on ma wavelength ECP's currently under investigation are also described. * Research sponsored by the U. S. Atomic Energy Commission under contract with the Union Carbide Corporation. .- -. - PATENT CLEARANCE OBTAINED. RELEASA TO THE PUBLIC IS APPROVED. PROCEDURES ARE ON FILE IN THE RECEIVING SECTION. LEGAL NOTICE The report w epand na account of Comanum sponsored work. Kaldhor the wed mamo, nor the counterton, por me porno acting a heall of the Commission: A. Makes my m oty or opracodation, pred or lapted, mul impect to the ACCT- rusy, completed, of watalions of the lefarado con lod ta do roport, or that the we of way baloration, entus, methods or prono deslood be the moport wyne intot pinily wed me of D. A ur the won ma meet to the won or for deres moltes from the man baloration, mento, machod, or acons declared to the foport Alwd in the mon, "perna they a half of the Corint" trobudno my plogna or controller at de consumaton, or phone al wel contractor, to the extent that med employee of contracte de contacten, op wployee at mod contractor propera, dermatoa, or pronto scooter og bormation and to wo wyloge or contract woh the comington, or Melegant mu mal contractor. DRAFT ENERGETIC NEUTRAL INJECTION INTO AN ELECTRON-CYCLOTRON PLASMA 1. Introduction During the last year an experimental program has been in progress at ORNL designed to trap 20-keV ions in an electron-cyclotron plasma (ECP). The program has two aims. The first is to measure the charge-exchange decay time of the trapped particles as a means of ceasuring the neutral density inside a macroscopically stable ECP. The second aim, contingent upon favorably low neutral density, is to employ high injection currents to build up a fairly dense (2 200 ions/cm3) hot-ion plasma in order to ascertain the stabilizing effect of the high electron temperature. If instabilities are seen similar to. those observed in other injection experiments, the influence of electron temper- ature on the threshold of instability can be studied. As an extension of the second part of this program, studies directed toward much higher hot-electron and hot-ion densities are in progress. For hot-electron densities greater than 10+ ions/cm", millimeter wavelength, microwave power sources will undoubtedly be necessary, and appropriate scale experiments at mom. wavelengths are under way. At the present time experiments are under way in the ELMO facility using 8-mm microwave power to generate an ECP in a variable mirror geometry. Although the temperature and density have not yet been measured, power absorption by the plasma is efficient from the two l-kw cw traveling-wave- tube oscillators. A macroscopically stable hot-electron plasma has been achieved. Experiments to produce an ECP using up to 1 kW cw of microwave power from a 5.5- mm traveling-Wave-tube amplifier will begin soon. Assuming the obtainability of hot-electron densities >1013, a calculation 18 given in Section 5 which deter- mines the energetic neutral beam current as a function of the parameters of an electron-cyclotron plasma. 2. Description of the EPA Facility and Beam The EPA facility is shown schematically in Fig. 1. It consists of a large vacuum chamber, vacuum pumps, and two mirror coils. Up to May 1965, the coil pair in use gave a 3:1 mirror ratio and all the data presented here was taken with this mirror ratio. Presently, there is a new coil pair giving a 2:1 mirror ratio and, as a result, a less steep radial gradient on the midplane. A micro- wave cavity is located between the two mirror coils. This high mode cavity con- • sists of a 36-in.-diameter perforated copper cylinder and associated end plates. Continuous-wave microwave power 18 fed into the cavity from three klystron ampli- fiers operating at 10.6 GC. Power levels up to 50 kW cw have been obtained. The details relating to the construction and operation of this microwave system have been described elsewhere. - The 20-keV neutral beam is also shown in Fig. 1. The beam is formed from a duoplasmatron ion source. An accelerate-decelerate system is used instead of a simple accelerator. Typical operation 18 with +40 kV applied to the source and -25 kV applied to the accelerator electrode. Deceleration to tank potential takes plase in a second gap. Conversion of the 40-keV , * ious to 20-keV rº atams takes place along the entire length of the 20-ft.-long beam tube. Since the gas influx into the EPA tank necessary to operate the ECP 18 of the order of 1 atm. cm/sec, the addi- tional gas influx from the beam tube does not appreciably affect the cavity pressure. The gas evolution from the target, located on the opposite side of the cavity, also does not affect the cavity pressure. Up to 40 mA of neutral --- - .--- . - - - -- - particles have been injected into the cavity through a 2-in.-diameter aperture. The angular divergence of the neutral beam 18 ~ 1º. Equilibrium between the * and rº beam 18 reached at beam tube gage pres- sure of ~ 5 x 10-5 torr. Based on published cross sections, 9 this value 18 in reasonable agreement with expectations. Approximately cmerº 18 injected for each * accelerated. The angle and direction of the beam injection can be adjusted to same - - -- . --. . ev ... degree. In the initial experiments, the beam was aimed with its center line -. . passing 2 in. above the magnetic axis on the midplane. Presently, it 18 aimed - so that the beam center line passes 4 in. above the magnetic axis. At a field of 200 gauss on the axis at the midplaney particles can now be trapped with their guiding centers on the magnetic exis: * 3. Description of the Detectors The detectors in use at present are commercially available silicon barrier diodes. These are used to measure total current rather than to detect individual particles. A schematic diagram of the detector circuit is shown in Fig. 2. The silicon diodes are sensitive to light and x rays as well as the protons. Hence, the diodes are shielded from the x rays by lead collimators and from the light - - by thin-aluminum windows. The Al windows also act as a barrier against slow plasma which might drift into the detector. The Al windows are made by evapo- rating 600 Å layers of aluminum on Zapon lacquer films and dissolving the Zapon. The aluminum films have been supported on copper grids. They have also been supported on copper apertures which supply additional collimation for the pro- tons. The lead collimator has a 3/8-in.-diameter, 2-in.-long hole which 18 designed to shield the detector from the x rays originating from the mirror throats. The silicon diodes are mounted in metal cans and view the plasma through beyond-cutoff waveguides to prevent microwave leakage into the diode and its associated circuitry. These cans are outside the cavity and see the plasma through high transparency (~ 70%) perforated metal windows to further reduce the microwave leakage. Two detectors are currently in use and are shown in Fig. 1. Both detectors are movable and can scan the plasma both perpendicular to and parallel to the magnetic axis. Other detectors have been used from time to time. Generally, they have been fixed in position and mounted at larger radii so as to be better protected from x-ray and microwave irradiation. The amplifiers are transistorized dc linear amplifiers (ORNL Q-2773-2), which have an adjustable gain. Currently, a gain of 1 volt out for 10° amp in 16 being used. The response time 18 5 5 usec. A 4-volt detector bias battery insures & very shallow depletion depth in the S1 diode to decrease the signal produced by the x rays from the ECP. The 16-volt bias battery shown is used to null the current into the amplifier. The detector is A-C coupled to the amplifier to eliminate the slow drift due to x-ray and proton irradiation. Measurements made by C. F. Barnetton the 600 Å Al films, which we have made and supplied to him, have shown that the film absorbe 13 keV of the inci- dent proton energy. The 7-keV particles which reach the diode produce ioniza- tion less efficiently than higher-energy particles. At this energy an average of 10 eV per electron-hole pair 18 required, as compared to ~ 3.5 ev per electron- hole pair measured for higher-energy particles. Hence, this process produces a current gain of about 700. The energy loss measurements by Barnett are not in agreement with those by Ormrad.* However, Barnett calibrated foils wade by this group with a tech- nique identical to the technique used to measure the proton current from the EPA. Hence, his technique essentially calibrated the system used in EPA and provides internal consistency. 4. Preliminary Experimental Results Fig. 3 shows the effect on the proton charge-exchange time, T, of increas- ing the microwave power. These data were take:a early in the experiment at a time when only 21 kW of microwave power were available. The detector collimator was pointed at the magnetic axis. Also shown is the detector signal amplitude, I. The amplitude appears to decrease with small amounts of power then increase with more power. The product It of signal amplitude and decay time also shows decrease with small amounts of microwave power followed by a monotonic increase. Fig. 4 shows a scan riade with the movable detector at higher power. The product of the detector signal, I, and the charge-exchange decey time, t, is plotted against the tilt angle of the collimator. This product should be pro- portional to the trapped proton density. The scale is arbitrary. The magnetic field was charged between each scan and the curves shown on the figure reflect the large charges in trapping due to the magnetic field. All the curves shown were taken with the same amount of microwave power and the gas pressure adjusted at each current setting to peak the microwave temperature. The data for the IT product at various magnetic fields without microwave power are not shown. However, in these experiments the IT product without microwave power did not exceed 15 on the abscissa. Hence, the data show that - ------ - - --- under optimistic conditions, the IT product increases by a factor of about 7. Under specific conditions of microwave power, magnetic field, and gas pressure, the IT product has been obse. Pred to increase by a factor of 2 10. In these experiments, both I and r increase; however, the a generally increases more. 5. Medium Energy Atoric Neutral Injection into a Mirror-Contained Electron-Cyclotron Plasma The fożlowing calculation attempts to determine the average trapped density of 20-keV protons resulting from the injection of a high current ( ~ one ampere) 20-keV atomic neutral beam into a dense (3 x 10+5 cm 3) electron-cyclotron plasma (ECP) in magnetic mirror contal.nment. The ECP 18 assumed to have a size sulfi- cient to provide a thermal neutral shielding distance of at least five mean free paths for ionization of 10 and * The calculation takes into account the. reflux back into the plasma of those neutrais which go out to chambe: walls. The ECP required for a device of the sort visualized would have a volume of several liters in a magnetic mirror field of approximately twenty kilogauss. The scaled ECP, because of the scaled-up density, will require magnetic fields of 20 kg. Correspondingly, the cyclotron heating source will necessarily have a wavelength of approximately live and one-hall mm. A toroidally closed mirror- coutained ECP is diagrammed in Fig. 5. The microwave power required for this ECP will be several hundred kW cw; the exact figure for engineering scaling is in the process of determination. It 18 anticipated that this high power will be required for such a small volume for two reasons. First, the power will be required to support the increased density, and second, it will have to support the increased mirror 1088 rate resulting from this increased density. Obviously, the purpose of the following calculation is to demonstrate, in what 18 felt to be a relatively conservative way, the build-up of a significant hot-ion density in the system described. Since the EP can be operated in regimes of macrostability, it 18 felt that this system should not be dominated by nacroinstabilities for hot-ia densities significantly less than the ambient bot-electron density. The ratio of plasma radius to hot-ion larmor radius will be interestingly large ( 10). Bence, the experiment should permit the investigation of a steady-state, very-bot-ion, very-bot-electron plasma at significant densities.' The womal rate equation for the build-up of bot-100 denoity, ry 18: J1 • Ja • Is (1) Moer: Ja • Average current of energetic ions trapped 10 a wit volume. 32 - Average current of energetic cous charge transferring out of a wit volur. 1, • Average current of energetic lons Coulomb scatter-lost out of a wit volume. the calculation of , 18 post conveniently accomplished by averaging the total current of bot ions over the volume filled by the hot lons. However, J, Lo calculated by first calculating is, where Jr 18 the trapped hot-ima curent per wit volume at the radius r. The differential volume av = 2x r dr z 18 filled with hot ions from a trapping length dL 2 dr. We AB Bume a cylindrical plasma with 2 = + R/2. The factor 2 in av results from the precession of small orbits around the annulus formed by the assumed dif- ferential volume (see Fig. 6). i . he way explore 2 dir 2x r dr R ir RA ... Note that since de aer, the peak hot plasma density would be expected to : be at the center of the plasma. The average energetic injection current , • : 28 R +R/2 I r do dr az r R N 2. I. . P a RJ- 0 0 2 A21 kW. The trapping of energetic neutrals in the case considered Involves two daoinant processes: . trapping by jonisation on hot electrons, the bean free path of which 16 me, and trapping by charge transfer to cold lans, the mean free path of which 18 desso del De ose We assume that at all times De • Dhyo . Examination of the cross sections for 20-kev hydrogen atoms shows to be much the smaller and therefore the more dominant term, 1.e., 2 Io de Oax : and 8; " EL ana Substituting (2), (3), and (4) låto (1): het en dat hy Pet. A mó Dor a 2.0 sin steady state tate The neutral, density in the center of the plasmas changes as da + 13 + 1 + 15 16 12 18 the unit-volume-current-gain of slow neutrals made by charge transfer from energetic neutrale to cola ions. 18 18 the unit-volume-current-loss of slow neutrals by charge transfer to hot protons. . 19 18 the unit-volume-current-gain of slow neutrals from outside the plasma due to vall reflux of blow atanic neutrals coming out of the plasma. 14 18 the wit-volume-current-gain of slow neutrals from outside the i. plasma due to wall reflux of energetic protons lost by charge exchange. . 19 is the unit-volume-current-gain of slow neutrals from outside the plasma due to the neutral gas input IBg. Of course, this flux : should be made as small as possible. *.16 18 the unit-voline-current-1088 of slow neutrals due to their drift; to tibe cavity walls. :. . 2 Io Ro de ocx ... .. . . . 11 . . . . . . . .... Va ni no Ucx V1 2 K " .. • .. Notice that in 12 through 18 the pessimistic asoumption is made that no atanic neutrals (other than that fraction of energetic beam trapped) are ionized by the electron-cyclotron plasma. However, the complete . ionization of thermal Hº to He* 18 assumed to take place in the first ..., few mom of the outer surface of the ECP, and 12, 14, and 15 result from molecular hydrogen entering the surface of the ECP. Tae factor 2 in the denominator of 13, in, and 15 results in the consideration that he 18 first lonized to k*, so only one-half the i. neutral atoms enter the plasma. The es from the fraction of H2* dissociatively lonized after excitation (Franck-Condon process). It also includes the favorable effect of plasma pumping of molecular locis before dissociative ionization. Substituting (7), (8), (9), (10), (11), (12), into;(6), and solving Por no in the steady state :; A * l ois - tikinau...1..ie wirwa - ...et • : : 210 RO De Pox+ (2 to Ro ne cx + IBg) V my OCX Va t (va of Oct. Vg) - .. - . ------ -- . . ---- . --. re..therandsadora 12 In order to simplify the americal consideration of the previous expressions, the following assumptions and parameters are incorporated. VOV - R . A - la R R R = 20 cm ocx - 6 x 20-26. No 3 x 1025 . K-3 : I. 16 expressed 10 amps and Igg 18 expressed in terms of cavity pressure P. (torr) outside the electron-cyclotron plasma, without energetic injection. Since 18gbo and since no - P. 3.2 x 2016, then IBS 1.9 x 202* Po . Incorporating these factors into (13) gives: 1.6 x 1030 (10 + Po 1.3 x 108) Ps (a1 + 3 x 1012] (24) ::. Then substituting this relationship for a int (5), (Tg calculated : .. to be 1.5 x 10-2 seconds for 20-kev protons in 3 x 1023 density plasma In a 3:1 nissor 9.... . . . - 2018 /[2.05 Io + 9.35 x 10"Po + 1.5j2 + 30.9 Io ... [2.05 I. + 9.35 x 10° Po + 1.5) Hot-ion density vs energetic (20 kev) neutral injection current 18 plotted in Figure 7. The parameter for these curves 18 neutral pressure. Po ranging from 2005 to 10-7 torr. Pressures in the region of 20-0 torr. should be realizable outside electron-cyclotron plasmas and beam currents in the : . 13 . neighborhood of one-ball amp should be realizable with the few constraints imposed on them by this system. The 20-6 torr curve in Fig. 3 then indicates an average 20-kev proton density of 1.7 x 1012.cms. Since no advantage has been taken of the reduction of neutral flux by electron-cyclotron plasma lonization of atomic neutrals, and since & somewhat pessimistic scatteringe · 108b-relationship was employed, the above donsities seem plausible. It should be pointed out that at low pressures la 10° torr) and low beam currents (~0.1 amp) the plasma density 18 mirror-loss limited. At high beam currents (~ 10 emp) and all pressures, the plasma density 18 limited by charge exchange. Under these conditions the hot-ion lifetime 18 considerably longer than . the scattering time so that any velocity or energy spread in the resulting ...: trapped hot-ion plasma must be built in by the nature of the injected ion beam. However, since ultra high vacuum 18 not required for the experiment, consider- able flexibility'18 allowed in the design of the ion beam. In fact, if the 1.sotropy in the resulting plasma is obtained by the use of several beams injected at various angles, present lon beam technology. 18° more than gdequate .. to provide the necessary current. Acknowledgments : The authors wish to express their appreciation to A. H. Snell for his interest and encouragement. The conscientious technical assistance of M. C. · Becker, A. 0. Eason, M. W. McGuffin, R. L. Livesey, H. C. Hoy, and O. D. Matlock is gratefully acknowledged. The accel-decel arrangement and the large cup design on the source anode was based on suggestions of 0. B. Morgan, R. C. Davis, and G. G. Kelley and their assistance 18 also gratefully acknowledged. Definitions of Symbols in · 2 - Hot-ion plasma length ..Ro - Radius of BCP R = Radius of hot-ion plasma Do • Thermal neutral density outside plasma . mo - Neutral density inside plasma :D - Hot-ion density .. no - Hot-electron density • My * Dicola : 04 • Ionization cross section for hot electrons on i energetic neutral atams. (6 x 10-19 cm) : 2 x 10-16 cm? ocx - Chargə transfer cross section for 20-kev protons on neutral atomic hydrogen (6 x 10-16 cm) ..VA • Velocity of 20-kev rº or + 2 x 108 cm/sec Vo - Velocity of thermal bydrogen - 2 x 205 cm/sec 1.Ve = Franck-Condon or slow coa velocity - 3 x 106 cm/sec V . Volume of ECP VI - Volume of hot-ion plasma .10 Current in part/sec of 20-kev rº To - Coulomb scattering time for 20-kev prota Pa = Mean free path for process a i: otong. . : . . :.. . . . . .. :... : .. .: .. . - . .- 25 - ...... REFERENCES 1. DANDL, R. A., et al, Nuclear Fusion 4 (1964) 344/353. BARNETT, C. F., et al, Atomic & Molecular Collision Cross Sections of Interest in Controlled Thermonuclear Research, ORNL-3113 (revised) p 17, (Aug. 1964) Unpublished. 3. BARNETT, C. F., private communication. 4. ORMRAD, J. A., et al, Can. J. Physics 43 (1965) 275/284. 5. FOWLER, T. K., Phys. Fluids 8 (1965) 459/470. 6. POWIER, T. K., Plasma Physics (Journal of Nuclear Energy Part C) 6 (1964) 513/514. 16 FIGURE CAPTIONS 1. Diagram of the EPA Facility and Neutral Beam. 2. Schematic Diagram of 81licon Barrier Detector circuit. 3. Detector signal, I, Decay Time, T, and the IT Product as a Function of Microwave Power.' The IT Product as a Prinction of Detector Tit Angle. When the detector 18 tilted 18°, it 18 looking at the equilibrium orbit for 20-keV protons in a 2-kG field. 5. Sketch of One Possible Cold Plasma Feed and Hot Plasma Contain- ment Scheme. ORNL-DWG-65-1038 ORNL-DWG-65-1039 22 6. Diagrammatic Cross Section of the Plasma. 7. Hot-Ion Density as a function of 20-keV Neutral Injection Current with Pressure as a Parameter. ORNL-DWG-65-1040 . . .. .. DUOPLASMATRON ION SOURCE -X-RAY SHIELD UUUUU - MICROWAVE ATTENUATOR FOCUS MAGNET DETECTOR MOUNTS - ACCELERATING TUBE : MICROWAVE GENERATOR BEAM PATH — CPUM 10.6 GHZ -TARGET - PLASMA VACUUM TANK- -MICROWAVE CAVITY TO PUMPS 1 : : ! - - - .- * * * BLANK PAGE T data center oscilloscope -- TII+ - - aperture 600 Å film 220K silicon barrier , detector 2.2 Meg 7.5 mfd. Į 10amp/volt amplifier charge-exchange protons +, - LEH 4.05 V. 70 mfd. battery IK lead x-ray shield Fig. 2 JUNE 2, 1964 FIELD CURRENT 1800 AMPS GAS PRESSURE 2.0 x 10-5 TORR DECAY TIME, alusec) 1, DETECTOR SIGNAL AMPLITUDE (ARB. SCALE) IT (ARB. SCALE) 5 10 15 MICROWAVE POWER, kw 20 25 Fig. 3 ST ? k BLANK PAGE . FEB. 8, 1965 30 kW MICROWAVE POWER CRITICAL GAS FEED AT EACH FIELD CURRENT 15 ma BEAM CURRENT 120 ....... ... .. - . - -- - ... 2200 AMPS - 1800 AMPS x - - - - - IT (ARB. SCALE) - - - . .- - - ... - - - - ... - - - 3000 AMPS . C 2000 AMPS 1 - - -- - ----- - -- og 1 10 11 12 13 14 15 TILT ANGLE, DEGREES 16 17 18 19 Fizet 20 KEV_H BRAN TOTAL PONER 3750KN 43 MM. UN SOURCE - 2:7 ANSNETIC MIRROR REGION CONTAINING HOT PLASMA DİAT COW PAWSANED NOBLO MAGNETIC MIRROR REGION SUPPLYING COLD PLASWA REED ZO. HOT. AASWA MIRPOR ... Fig. 5. ORNL-DWG-65-1038 --- -- . .. ..... … . - - --- -- - - - - - - - - - - - - - - - E.C.P. BOUNDRY -PROTON ORBITS ALONG OLAM BEAM OF HO ထ " " "" I . . - - • • • 1. 6. ONLDMC-65-1039 • • • • END . DATE FILMED 11/ 9 /65 +