of the

(1)

HENRY H U R W I T Z , Jr.

General Electric Research Laboratory, Schenectady, New York

I n t he public mind it is taken for granted that fusion power wil l inevitably replace fission power j u st as fission power is n ow replacing power from fossil fuels. Although scientists engaged in fusion research continue to be optimistic about t he long-range prospects of fusion power, they recognize that it is still too early to predict whether, or on what time scale, fusion power wil l reach t he stage of practical importance.

Progress during t he past year has made it appear m u ch more likel y that a controlled thermonuclear reactor which produces more electrical energy than it consumes can ultimately be constructed.

But it is by no means certain that even after this milestone has been achieved, t he reactor can be developed to a point where it would be economically attractive. Indeed it is quite possible that t he reactor which first produces a net power o u t p ut wil l not be of a type which lends itself to t he development of an economic machine.

I t has been argued that t he world's need for power is growing so rapidly that withi n a relatively few generations t he accessible supplies of fissionable materials as well as fossil fuels wil l have been exhausted.

M o r e careful studies have, however, shown that this wil l not neces- sarily occur. By making use of t he breeding process which converts the a b u n d a nt isotopes of u r a n i um and t h o r i um to thermally fissionable isotopes of u r a n i um and plutonium, t he efficiency of utilization of fertile materials is greatly improved. T h is is t u rn may make it economically justified to process extremely low grade ores so that the supply of fissionable material may in effect become unexhaustible.

Hence, even from t he very long-range point of view, it appears that fussion power wil l become important only if it can compete on an economic basis wit h fission power. As D r. Weinberg of Oak Ridge has expressed it, t he long-range choice is between " b u r n i ng t he rocks and burning t he sea." But regardless of h ow this decision turns out, fusion research continues to be one of t he most interesting areas of modern science.

* Summary of talk by H. Hurwitz, Jr., on fusion research.

289

(2)

290 H. HURWITZ, JR.

I t is often forgotten that the fusion process was discovered in 1932, and so predates fission by 6 years. Rutherford and his associates, who made the original discovery, were quick to recognize the great difficult y of developing the fusion process into a source of useful energy. I n Rutherford's experiments a beam of accelerated deuterons was focused on a deuterium bearing target. Although fusion reactions could be detected by observing the products of these reactions, the fusion energy release was less than a millionth of 1 % of the beam energy. T h is is because the deuterons were slowed down rapidly by ionization processes before more than a minute fraction of t h em could cause nuclear reactions.

I t was almost a decade after Rutherford's discovery that scientists began thinking seriously of avoiding the loss of energy by ionization and Coulomb scattering by replacing the beam and target combina- tion by a single volume of deuterium heated to several h u n d r e ds of millio n degrees.

T he concept of using magnetic fields to confine the hot reacting gas was a natural, but bold one. I t is based on the fact that at the temperature required for the thermonuclear reaction to be self- sustaining—several tens of kilovolts—the reacting gas is completely ionized. T he stripped nuclei and electrons making up the plasma interact strongly wit h the magnetic field and, wit h appropriately chosen field configuration, can be contained in a finite volume of space for the times required for significant fusion b u r n up to occur.

T he motion of the charged particle in the magnetic field results i n a microscopic electric current which tends to cancel the magnetic field. T h us the plasma pressure which can be contained by a magnetic field is limited in accordance wit h the requirements of the Maxwell stress relationships. T he equivalent magnetic pressure, B

2

_/STT,_can, however, be large and in practice is limited only by the physical strength of the coils which provide the magnetic field.

I t was recognized at an early stage that the existence of equilibrium configurations involving b o u n d ed plasmas in magnetic fields does not i n itself ensure the validity of the concept of magnetic confinement.

T h i s is because of the additional consideration that only stable configurations can be expected to confine a plasma for times which are long compared to the dimension of the system divided by the sonic velocity in the plasma. T he theory of hydromagnetic stability has been the subject of intensive investigations. M a ny useful results have been obtained in the magnetohydrodynamic approximation in

(3)

which the plasma is treated as a conducting fluid. M o re accurate methods have also been developed in which it has been possible to take into account the fact that the plasma is made up of particles having finit e radii of gyration.

I t was predicted theoretically that specific configurations such as the stabilized pinch and the stellarator geometry should be stable against perturbations involving gross motions of the plasma. Early experiments indicated a definite correlation between the behavior of laboratory plasmas and the theoretical predictions in that configurations conforming to the theoretical stability requirements exhibited less turbulence than configurations not conforming to these requirements.

On the other hand, even when the recognized stability conditions were satisfied, quiescent contained plasmas did not appear to be established. T h is has been interpreted as being a consequence of more subtle instabilities which arise from the fact that the velocity distribution of particles in the plasma is not Maxwellian. I n parti- cular, because of large macroscopic currents in the plasma the velocity distributions are neither isotropic in direction nor monotonie functions of particle energy. T h e o ry predicts that such deviations from the Maxwellian distribution can lead to instabilities in which space charge and magnetohydrodynamic waves may grow spontane- ously in a m a n n er familiar from the operation of traveling wave amplifier tubes. T he final amplitude of these waves is determined by nonlinear effects which have not yet been full y analyzed. I t is, however, reasonable to assume that the velocity space instabilities can lead to the development of turbulence and a consequent drift of the plasma across the magnetic field lines.

Recently R. F. Post and W. A. Perkins have reported experiments i n a magnetic mirror geometry in which it was possible to vary the symmetry of the electron velocity distribution and to demonstrate that enhanced plasma diffusion across the magnetic field did in fact occur when the anisotropy of the velocity distribution was large enough to fall in the region of predicted instability.

I n another magnetic mirror experiment F. Coensgen and his associates succeeded in producing a plasma wit h ion temperature in the 3-kev range by adiabatic compression of plasma produced by an occluded gas source. T h is plasma was confined for the order of 100 ^sec during which time fusion reactions occurred at a rate consistent wit h the ion temperature and plasma density. Since the

(4)

confinement time is far larger than that in which instabilities would manifest themselves, it is concluded that disruptive instabilities were, i n fact, absent.

T he ratio of plasma pressure to magnetic field pressure in Coens- gen's experiment was of order 1/200. A key question which remains to be resolved is that of how m u ch this ratio can be increased without encountering instabilities.

Another investigation of magnetic mirror confinement is being carried out in the D CX machine at Oak Ridge. I n this experiment 600-kev hydrogen molecular ions are injected into the mirror region where they are dissociated either by collisions wit h background gas molecules, wit h previously trapped particles, or wit h the atoms in an auxiliary carbon vacuum arc. T he 300-kev proton which remains after the dissociation has a smaller radius of gyration than the original molecular ion and t h us is trapped in the region between two magnetic mirrors.

By careful attention to reducing background gas pressure and by virtue of developments in ion source technology, it has been possible to build up a plasma density of about 10

8

ions per cubic centimeter and confine it for as m u ch as 7 sec. Although the plasma pressure is only of order 1 0

-5

of the magnetic pressure, it is reassuring that certain types of electrostatic instabilities which could develop i n this regime have not been serious. Progress is well along in the construction of a larger D CX machine in which the attainable plasma . density wil l not be limited by source strength so that a direct deter-

mination of the limitin g stable plasma pressure should be possible.

Still another class of magnetic mirror investigations are the Scylla, or #-pinch experiments being conducted at Los Alamos, the Naval Research Laboratory, the General Electric Research Laboratory, and other places. I n these experiments m u ch higher ratios of plasma to magnetic pressure are attained. T he time scale of these experiments is only a few microseconds so that the existence of instabilities cannot be definitely excluded.

I n the General Electric experiments, an extremely low inductance 400-kjoule capacitor bank is used to apply 50 kv to a single t u rn coil surrounding a previously ionized deuterium plasma (see Figs. 1 and 2). T he coil current rises to almost 4 Χ 10

6

a mp in 5 /xsec, thus producing a peak magnetic field of over 100 kgauss. If an initial bias field of a few kilogauss is applied in the reverse direction, yields of up to 2 Χ 10

9

fusion reactions per discharge are attained.

(5)

T STATUS OF THE FUSION PROBLEM 293

FIG. 1. Magnetic compression experiment at the General Electric Research Laboratory. This picture shows the low inductance 400-kjoule capacitor bank, switching equipment, and the single turn coil surrounding a ceramic tube in which the plasma is produced.

(6)

T he magnitude of the fusion yield and its variation with capacitor voltage suggests a plasma temperature between 1 and 3 kv. T he fact that the fusion reaction rate reaches a peak shortly before the time of peak magnetic pressure, and the role of the initial reverse magnetic field in the initial stages of plasma heating are still incompletely

FIG. 2. Neutron yield in General Electric Research Laboratory rapid compression experiment. The oscilloscpe traces show neutron yield (n) and magnetic field (B) applied to the plasma in a rapid magnetic compression experiment using the equipment shown in Fig. 1. The plasma is initially preionizçd by an auxiliary discharge and a reverse bias magnetic field is applied before the main compression. The neutron yield persists for aprroxi- mately 5 /xsec centered about the time of peak magnetic pressure. T h e cor- responding number of fusions is between 10

s

and 10

9

. The spectrum of deuterons in the plasma is not well known so that a theoretical extrapolation of the performance to still higher powers cannot be made with certainty.

1, Start of bias bank; 2, firing of preionizing discharge; 3, firing of main bank.

Scale: 5 μ-sec/division.

understood. T h e se questions are being pursued by exploring the effects of various modifications of the geometry and operating conditions of the equipment and also by utilizing diagnostic techni- ques such as measurements of the magnetic field as function of position and time in the plasma, and various spectroscopic observa- tions.

Although magnetic mirror experiments are receiving particularly great attention at the present time, many other magnetic confinement

(7)

geometries are also being actively explored. Several groups are continuing their efforts in finding means of overcoming the difficul - ties encountered wit h the stellarator and toroidal stabilized pinch configurations. Some ingenious ideas have been developed for this purpose. I n addition, interest is growing in various types of magnetic cusp geometries which offer the promise of stably confining higher pressure plasmas than is possible in other geometries.

T he prospects of developing a fusion reactor wit h net power o u t p ut have been given an unexpected boost by the recent advances i n the field of superconductivity. T he fact that it now appears technically feasible to construct superconducting coils which p r o- duce magnetic fields in the 100-kgauss range introduces the possi- bilit y of obtaining net power o u t p ut u n d er conditions in which the ratio of plasma pressure to magnetic pressure is low—possibly not m u ch higher t h an that which has already been explored in the Coensgen experiment. On the other hand, it is by no means assured that such low pressure devices would be economically attractive.

Hence the motivation for exploring configurations in which the plasma pressure is not small compared to the magnetic pressure is as strong as ever. Indeed, even t h o u gh the high pressure devices may not provide the quickest path to developing a machine wit h net energy output, they may well prove to be more practical in the long run.

I n s u m m a r y, the concept of magnetic confinement of high t e m- perature plasmas has progressed from a theoretical idea to a laboratory reality. Major advances have been made in both the experimen- tal and theoretical aspects of plasma physics, and the area of correlation between theory and experiment is growing rapidly. Never- theless, the properties of magnetically confined plasmas are extremely complex, and despite the great progress made in the last few years m u ch remains to be learned about the subject. I t is reasonable to expect that withi n the next 5 years our knowledge of high t e m p e r a- ture plasma physics wil l have progressed to the point where a definitive conclusion concerning the ultimate potential of controlled fusion power wil l be possible.