© 1975 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE. IEEE TmactioMd on Nuc.kvc Science, VoLNS-22, No.3, June 1975
SELF COLLIDING BEAMS ("MIGMA") AND CONTROLLED FUSION
B. Maglich, M. Mazarakis, R. A. Miller, J. Nering, S. Channon, and C. Powell Fusion Energy Corporation, Princeton, N.J. 08540
and
J. Treglio* Stevens Institute of Technology, Hoboken, N.J. 07030
Khile much of the early work on colliding beams the lead in the development of THERW MIGMA was done in the U.S., NUCLEAR this technique is now held by Europe. The most spec- -a tular being the only colliding beams of nuclei in the /e.J !%;,,, BREE[XR Intersecting Storage Rings at CERN. It may be for this reason that the idea of using colliding beam technology as a means of achieving fusion has not reappeared until just recently. The idea of using colliding beams for fusion is nearly as old as is the interest in fusion as a source of power, but the problems of low reaction rate and high coulomb scattering initially seemed in- surmountable. This work describes recent work done at Fusion Energy Corporation which attempts to overcome these problems. Before turning to our own work, let us examine the well know facts about CERN ISR in a manner suggesting comparision with traditional plasma thermonuclear ma- chines. The confinement time of the ISR is 235 days, that is, a beam stored in the morning has decreased by less than a percent in intensity by evening. The pa- rameter, n7, density times confinement time can easily be found to be about 10" cmm3 set compared with pre- sent plasma values of no better than lO'","Lawson Cri- terion" (see below) values of 10'4, and values of 10" required when the Lawson Criterion is corrected by a IOkeV nlnnber of realistic considerations (see below). An- / IhJECTlCN ACHIEVED f, ~ ENB/IGYT,- other parameter of interest to plasma researchers is IONTEMP’s - t !i, = Sn nkT/??, which is the ratio of kinetic energy PLASMA MIltMA DEVICES MOOELI density to magnetic energy density. In magnetically -,--IQ-~~ I T710-3s confined plasmas, this ratio must be less than 1. At ! i ISR the ratio is 36 in the curved sections and as much DESlGPr PLASMA MIGMA i[ as 108 in the straight sections. The difference may ION TEMP’s + ~19801 (19761 he a:tributed to the high degree of order of the moticn in a beam compared with a plasma. Figure 1 shows another advantage of beam technol- oep simply in the higher energies attainable. In this Fig. 1. Reaction parameter
Fig. 2. Comparison of migma with a -; Y hypothetical pair of intersecting 11 storage rings. r w
,013i , I “1 dq ?r 1, /
0.41 4 h?lP 12 14
Fig. 3. Monte Carlo caiculation of relative fusion rate as a function of radius in a migma.
For reasonable vertical confinement the scattering time is found to be longer than the fusion time. Migma has some superficial features in common with mirror-confined plasmas, including the DCX series which was an attempt to heat a mirror-confined plasma by a beam of up to 600 keV. The differences are actu- ally many and important. To begin with, migma is high- ly organized, resulting in the high central density and predominance of near head-on collisions already men- tioned. The gradient of the magnetic field is large across a migma orbit, while it is essentially zero across the lower energy plasma orbits. Among other things, this means that migma is focussed rather than confined by magnetic pressure. Because migma is fo- cussed, a ratio of kinetic to magnetic energy density, t?, greater than 1 can be achieved. Because ions are injected directly, MeV energies can be attained, com- pared with roughly 1 keV in present plasmas. In ad- dition to the advantages discussed in connection with -Izpp 11. Fig. 1, higher energies also allow the use of non- polluting advanced fuels. 5 A graphic demonstration of the difference between migma and plasma is shown in Fig. 4. The two sets of lines plotted are flux lines and migma energy envelopes Are obtained from the requirement of conservation of canonical angular momentum and are the boundaries of migma ions of given energy. Plasma, on the other hand will be bound to the flux lines. One consequence of this is that migma does not obey the Alfven mirror relation that plasma does. As mentioned earlier, Fig. 1 seems to indicate that around 1 MeV is the favorable energy region for dd fusion. However, dd plasma reactors were targeted for about LOO krV. This difference is puzzling and is rooted in the "Lawson Criterion"." The Lawson Cri- Fig. 4. Magnetic flux lines compared with the terion ~2s developed by J. D. Lawson almost 20 years more sharply curved migma energy envelopes for ago. It is derived from a simple energy balance equn- the SGme magnetic field. (See text). tion and includes a single conversion efficiency for
1791 all forms of energy emitted by the reacting medium. Lawson clearly stated that this criterion was "ideal- ized" and "by no means sufficient for the successful operation of a thermonuclear reactor.'16 Reference 1 includes 18 effects not in the original Lawson Cri- terion used to derive a more realistic criterion. Fig. 5 shows the resulting values of nr for energy break- even for a set of parameters representative of plasma ; /I t, SET-1 ,,I’ SETII('MIGMA") 11;’ mew - SEPTUM PIMW ’ 1rPLASMA”) i ;\ , , , I Ill 7c2= j ST4TlON j, ItI ~ I al
IO” KCELERATOR -* Sub ‘1 ~ q 111 [\ ,/J/g f
r c2 III ,.;~sG--qqpP~p3o’ 0.9999 i ;!i\ ,’ c 7 5 / i 10 A.! II ‘\ /!P”,,q.,:;;r ” E r 0.999. i/i!\ \ / ~T+;;;i> W1hm \; / 4 I 3 2 T!’ I 1/ ,&y-l”.* / g”p, tg ‘38 I I ;0’5 ,YIIIzJ Set I sca,e 1 -‘:ys~,;;o;;24:.i;p= / 1 b, \ 1 ,.,.. I I 1 set 1 scale i i ” s:,;g @~;v~;y*~b Tc2: CONFINEMENT EFFlCIENCY ii_-.;’ : ;:B-, SF%----TP*140 W.’ 7; SP. woe pm, 10 I Ill1111 I ) ~“‘,#I”‘” Ts - sub= rlfcnl”” Subilmafar C.01 0. I 0 0) ION ENERGY, T, (MeV) Fig. 6. Experimental set-up at Rutgers. Fig. 5. Plot of nr as a function of energy from reference 1.
and a set representative of migma for the dd reaction. Reference 1 should be consulted for a discussion of the choice of parameters. Note that the minimum value of n? for plasma is more than 10" cmm3sec compared with approximately 10" ~103~~set given by the Lawson Criterion for the dd reaction. This minimum, however, requires a confinement efficiency of 99.99%; that is, OE SCAN for every 10,000 deuterons scattered through 90", only Kicker at 14.3cm t 1 can escape. The migma curves, however, show a min- 10 imum of about lo1 5 cmd3 set at about 600 kcV for a con- finement efficiency of only 60%. a It should be noted that there is no serious plan ;! to build a dd plasma reactor; all present plans call for the dt reaction. It is an advantage to be able to z 1.0 use the dd reaction, because tritium is rare, radio- s I active, and expensive. More importantly the dt reac- tion has as a product a 14 MeV neutron which carries : 1 most of the energy. These neutrons pose serious tech- 0.1 nological problems of activation, radiation damage and thermal pollution because their energy can only be con- verted thermally. 1 If only positive ions are injected into a migma- oat, I 1 1 1 cell, their density is limited by the space charge lim- 60” 50” 40” 30” 2v 100 0” -10” -20” -300-40” -50” I it to about 1O'a cm-a. At that density the fusion power is a snail fraction of a watt. In order to a- chieve useful power densities it will be necessary to Fig. 7. Typical scan with one of the movable neutralize the positive charge with electrons. Two probes. schemes have been proposed to accomplish this. The first scheme is simply to impregnate the nigma with a non-linear electrostatic kicker was used to inject the :hermal electron gas. It is shown in reference 7 that beam into the chamber. Two curved scans with tungsten the electrons will be limited to about 100 keV by radi- probes shown in Fig. 6 intercepted the beam at a number ation and that at that temperature they will be only of points and the path of the beam could then be recon- weakly coupled to the ions by multiple coulomb scatter- structed by computer. Fig. 7 reproduces a typical scan ing. The second scheme is to have ordered electron by one of the movable probes. Fig. 8 shows a typical motions as well as ion motions by inducing vertical computer fit tc a scan, but not the scan of Fig. 7. oscillations of electrons along the magnetic field The kicker injects the beam into the self-colliding lines.* configuration we call "figure of 8". In this config- In order to develop these ideas the experimental uration the Dz ' molecules dissociate and form atoms sc-t-up shown in Fig. 6 KS constructed.' A 108 keV which are then well confined by the magnetic field. D, + beam is accelerated, analyzed, and focussed in the The lj0 peaks in Fig. i are at exactly the location and migma chamber, which is shown in more detail below. By have the symmetry expected of atomic migma orbits. means of differential pumping and cryosublimators a Solid state counters were placed near the interaction vacuum of lo-' torr was obtained in the chamber. A region and the proton and triton peaks from the dd re-
1792 Fig. 1 and Fig. 5. Provisions are made for differen- tial pumping to reach ultra-vacuum. A special super- conducting magnet has been designed and built which produces 50 kG at the center and 60 kG on the windings Data taking will be semi-automated through the use of digital drive scanning probes. Electrostatic quadru- poles will be used exclusively for beam transport. Instabilities that may develop in a neutralized migma will be more akin to those in accelerators and storage rings than those in plasmas. Studies of in- stabilities are under way but are beyond the scope of this brief report.
References
* Reserach supported by Fusion Energy Corporation.
1. B. Maglich and R. Miller, "Generalized Criterion for Feasibility of Controlled Fusion and Its Ap- plication to Non-Ideal dd Systems," to appear in J. App. Phys.
2. B. Maglich, J. Blewett, A. Colleraine, and W. C. Harrison, Phys. Rev. Letters 11, 909 (1971).
3. B. Maglich, Nucl. Instr. and Meth. 111, 213 (1973).
Fig. 8. Computer generated orbits fit to 4. R. A. Miller, Phys. Rev. Letters 2, 1590 (1972). observed interceptions of beam by probe. Upper and lower plots are for non-linear 5. J. Treglio, Fusion Energy Corp. preprint FEC-02-75. kicker off and on, respectively. 6. J. D. Lawson, Proc. Phys. Sot. E, 6 (1957),
7. R. A. Miller, Nucl. Instr. and Meth. 119, 275 (1974). action were observed. However, proof that these come a. B. Maglich, "The Principle of Time Average Neutral- from beam-beam interaction by showing quadratic de- ization of Fully Ionized Matter " pendence, will be obtained with the new experimental set-up. Fig. 9 shows the new experimental installation at 9. B. Maglich, M. Mazarakis, 3. Galayda, B. Robinson, Fusion Energy Corporation, which is now nearing comple- M. Lieberman, B. Webber, A. Colleraine, R. Gore, tion. A 3 MeV van de Graaff accelerator will allow D. Santeller, and C. C. Chieng, Nucl. Instr. and experimentation in the MeV region as is suggested by Meth. 120, 309 (1974).
.- .- - ,--t--T- ,w---- r---T~- ,d.!.r i.JJJ, ,i: .*/IleI L--.. * ; ! .~ 1 cam--_=------== - - -
_--__---_ KCELE? ATOR L_-_------.-. -L__, .-- --. UPERmJ BUCTING
%,." , / .,-a
--.--__-.r-l .~_ - _ .-c-_.I----~--_ .- ,--. . .._._. -~ _.-..- -_----- \ NIGMA CHAMBER
Fig. 9. Diagram of the experimental installation now nearing completion at Fusion Energy Corporation, Princeton, N..J.
1793