ON THE POSSIBILITY OF CREATING THERMONUCLEAR REACTIONS IN A GAS DISCHARGE*)

I. V. Kurchatov

Among the most important problems of modern technology, a special place in terms of its significance is occupied by the problem of the energetic utilization of thermonuclear reactions. The unusually interesting and at the same time very difficult task of controlling thermonuclear processes is now attracting the attention of physicists in all countries of the world.

Research in this field is being conducted under the direction of Academician L. A. Artsimovich at our institute. The leading role in the development of the theoretical questions belongs to Academician M. A. Leontovich.

As is known, thermonuclear reactions can arise in the event that the temperature of matter is so high that there appears a noticeable probability of overcoming the Coulomb potential barrier in thermal collisions of atomic nuclei. Of especially great interest is the excitation of thermonuclear reactions in deuterium and in a mixture of deuterium and tritium, since in this case a relatively lower temperature is required to obtain a noticeable effect.

Physics owes its first information about the processes of interaction of deuterons to the great founder of the modern doctrine of the atomic nucleus—Ernest Rutherford. In one of his last works he investigated nuclear reactions arising as a result of the collision of two deuterons. At that time it was impossible even to suspect that the new facts he had discovered would bring closer the prospect of mastering the sources of energy hidden in the hot interiors of the sun shining above us and of the distant stars.

The intensity of thermonuclear reactions in deuterium must increase very rapidly with increasing temperature—up to temperatures of the order of several billion degrees.

*) Lecture delivered on April 25, 1956, at the British Atomic Research Center at Harwell. Journal Atomic Energy, No. 3 (1956).

An idea of the conditions necessary for the experimental detection of thermonuclear reactions can be obtained by considering concrete examples. At a density of matter corresponding, under normal conditions, to a solid body, in order to obtain one neutron per second in 1 g of deuterium, a temperature of about \(2\cdot 10^5\) degrees is required. In a highly rarefied gas, at a concentration on the order of \(10^{13}\) atoms per \(1\ \mathrm{cm}^3\), to obtain the same effect from 1 g of deuterium it is necessary to create a temperature of about \(5\cdot 10^5\) degrees in a volume equal to \(30\,000\ \mathrm{m}^3\).

Thus, in order to approach at least the threshold for the occurrence of thermonuclear reactions, we must raise the temperature of the substance to a very high level. At this temperature level, deuterium under stationary conditions must be a plasma with ionization close to \(100\%\).

The reserve of energy that must be concentrated in the plasma in order for its temperature to rise to values at which thermonuclear reactions become sufficiently intense is relatively small. At a temperature of \(10^6\) degrees, the thermal energy accumulated in 1 g of deuterium amounts to only a few kilowatt-hours. Approximately the same amount of energy is required to boil water in a large family samovar.

Therefore, if one finds such a method of heating the plasma in which heat losses are practically reduced to zero, then even with the aid of a comparatively low-power energy source it is possible to cause the onset of intense thermonuclear reactions. The chief difficulty, however, consists precisely in eliminating heat losses, which increase very rapidly with increasing temperature, since the thermal conductivity of plasma is proportional to \(T^{5/2}\). When matter is heated only to several tens of thousands of degrees, these losses, in the absence of thermal insulation, become so large that a further increase in temperature proves impossible.

When matter of high density is heated, yet another serious obstacle appears: it is necessary somehow to overcome the enormous mechanical forces that arise because of the increase of pressure with temperature. In trying to heat solid or liquid deuterium, we find that already at \(T = 10^5\) degrees the pressure exceeds a million atmospheres. Therefore, in matter of high density, a thermonuclear reaction can be excited only for a very short interval of time, and such a process will always have the character of an explosion (perhaps, however, a harmless one) or of a brief pulsation.

Considering possible ways of carrying out controlled thermonuclear reactions of high intensity, we find before us a very broad horizon of different directions that may be followed in trying to solve this problem.

On one edge of this horizon lie directions connected with the development of methods for obtaining stationary thermonuclear reactions; on the other—a path based on the idea of an instantaneous rise in temperature in pulsed processes of very short duration. However, whatever direction of research we choose, we always encounter one and the same question: how to isolate plasma heated to a very high temperature from the walls of the vessel in which it is enclosed. In other words, how to keep fast particles in the plasma for a period of time sufficient for a noticeable fraction of them to have time to react with one another.

One of the ideas arising in connection with this question is to use a magnetic field for the thermal insulation of the plasma. This was first pointed out in 1950 by Academician Sakharov and Academician Tamm. In a sufficiently strong magnetic field, electrons and ions can move freely only along the lines of force. In the plane perpendicular to the field lines, the motion of particles will take place along circles of small radius. The centers of these circles can shift only as a result of collisions, and at each collision only by a distance of the order of the radius of curvature of the particle trajectory. If the radius of curvature of the trajectory is small in comparison with the mean free path, then particle diffusion and the thermal conductivity of the plasma in the plane perpendicular to the magnetic field will be sharply reduced. The theory of processes occurring in a fully ionized plasma shows that, for large values of the field strength \(H\) and high temperatures, the coefficient of transverse thermal conductivity is inversely proportional to \(H^2\) and is many orders of magnitude smaller than in the absence of a magnetic field. Under these conditions, however, energy losses due to radiation must be taken into account.

The magnetic field necessary for thermal insulation can be created by passing a sufficiently strong electric current through the plasma. As the current passes, the plasma will also be heated owing to Joule losses and the work of electrodynamic forces. These considerations served as the basis for the development of theoretical and experimental studies of the physical processes occurring in a plasma when a strong electric current passes through it.

Let us first dwell on the initial theoretical ideas that preceded the accumulation of experimental facts. During the passage of current, compression of the plasma must occur under the action of electrodynamic forces (the attraction of parallel currents). In this case the temperature of the plasma should rise. If, under the action of electrodynamic compression, a plasma cord is formed, detached from the wall of the chamber, then the temperature in the cord can be estimated from the condition of equilibrium of electrodynamic—

chemical forces and pressure. A simple calculation shows that, in such a quasi-stationary compression process, the temperature of the plasma must increase in proportion to the square of the current strength. If the electrons and ions are in thermal equilibrium with one another, then, as is known, the plasma temperature is expressed by the following formula:

\[ T=\frac{I^2}{4Nk}, \]

where \(I\) is the current strength in the electromagnetic system of units, \(N\) is the number of ions of one sign per \(1\ \mathrm{cm}\) of the length of the discharge chamber, and \(k\) is Boltzmann’s constant. Investigation of the conditions of thermal equilibrium showed that at \(N \sim 10^{17}\) the temperature of the electrons and ions should be practically the same. At substantially smaller values of \(N\), only the electrons will be heated.

A plasma cord detached from the walls can exist only during that interval of time when the strength of the discharge current is increasing. At constant current strength the cord must fall apart and touch the walls.

It is obvious that by passing current through a plasma one cannot carry out a thermonuclear reaction with a constant yield over a long time. One can count only on the periodic occurrence of cycles of plasma heating with the excitation of intense thermonuclear reactions in that phase of each cycle which corresponds to the maximum value of the current. Calculations of the expected thermonuclear effect led to a result that at first glance was paradoxical. It turned out that the total number of elementary acts of nuclear interaction in one heating cycle, for a given maximum current strength, should not depend on the duration of this cycle. It was therefore possible to hope for the excitation of very intense thermonuclear reactions in short pulsed discharges in deuterium, if the current strength was sufficiently large. On the basis of theoretical calculations it was to be expected that already at a current strength of about \(300\ \mathrm{kA}\) appreciable neutron radiation of thermonuclear origin should appear. At a current strength of several million amperes it should reach an extremely high intensity. Such were the theoretical predictions that existed before the beginning of the experimental work.

The further development of ideas about the nature of the processes that take place in plasma during the passage of large currents was determined entirely by new facts discovered as a result of the experimental investigation of powerful pulsed discharges.

The results of these experiments completely changed the character of the picture that had been created by the first attempts at a theoretical study of the problem.

Experimental investigations of powerful pulsed discharges were carried out over a wide range of variation of the basic parameters.

characterizing the initial conditions of the discharge*). The processes of current passage through hydrogen, deuterium, helium, argon, xenon, and gas mixtures (deuterium—helium, deuterium—argon, deuterium—xenon) with different component contents were studied. The initial gas pressures were varied from 0.005 mm Hg to 1 atm. The principal experiments were carried out with straight discharge tubes. The length of the discharge gap in different experiments varied from several centimeters to two meters, and the diameter from 5 to 60 cm. The discharge was supplied by a voltage of several tens of kilovolts. The maximum current in the discharge ranged from 100 ka to 2 million a, and the rate of rise of the current varied within the limits from \(10^{10}\) a/sec to \(10^{12}\) a/sec. The maximum instantaneous power released in the plasma reached 40 million kW. The sources of electrical power for the discharge were banks of high-voltage capacitors. The busbars collecting the current from the capacitors and delivering it to the discharge gap were designed so as to reduce to a minimum the parasitic inductance of the electrical circuit, which limits the magnitude of the current and the rate of its rise. At a voltage of 50 kV and a total capacitance of the capacitor banks reaching several hundred microfarads, the parasitic inductance of the circuit together with the switching device was brought (in those cases when processes with the maximum rate of current rise were studied) down to 0.02–0.03 microhenry.

For the study of high-power pulsed discharges, methods were developed and used for oscillographic measurement of various parameters characterizing the state of the plasma during the passage of the current. In addition to oscillographic techniques, ultrahigh-speed motion-picture photography (up to two million frames per second) and photography by means of Kerr cells equipped with special electroexplosive shutters were also employed.

In addition to the magnitude of the discharge current and voltage, the intensity of individual spectral lines of the plasma glow, the intensity of neutron and x-ray radiation, pressure pulses recorded by means of piezoelectric elements, and also instantaneous values of the strengths of the magnetic and electric fields at various points of the plasma were oscillographed. For measuring magnetic and electric fields, probes in the form of miniature coils, loops, and needle electrodes were used, which could be placed at different locations inside the discharge chamber.

The numerous results obtained in this cycle of experimental investigations do not fit within the scope of the present communication. Here one can briefly touch upon only a small part of this experimental material.

Of greatest interest is the study of the first phase of the pulsed discharge, during which the current in the plasma rises from zero

* In England in recent years pulsed discharges have been studied by Craggs with his collaborators Cousins and Ure, among others.

to a maximum value. In the experiments discussed here, the duration of this phase was from 3 to 30 μsec. At the very initial stage of the discharge, after gas breakdown, a smooth increase of the current and voltage in the discharge gap occurs. Then, at some moment in time, a decrease in the voltage is observed, taking place as an abrupt jump. On the current oscillogram, at the same moment in time, there appears a more or less strongly pronounced kink (see Fig. 1, which schematically shows the general character of the change in current and voltage during the process, as well as the oscillograms in Figs. 2 and 3). After the first disruption, the voltage at first rises very rapidly, and then again drops sharply. At the moment corresponding to the second voltage disruption, the next kink appears on the current oscillogram. Sometimes three to four successive disturbances of the smooth course of the current and voltage are observed in the first phase of the discharge.

These characteristic features of pulsed discharges with a large current strength are expressed especially clearly in those cases when the discharge occurs in a gas of small atomic weight (hydrogen, deuterium, helium) and the initial pressure is low.

Fig. 1. General character of the change of voltage and current in a pulsed discharge and the neutron pulse accompanying the discharge.

At a current rise rate of the order of \(10^{11}\) A/sec, the time interval from gas breakdown to the moment of appearance of the first voltage disruption amounts to several microseconds.

The duration of this time interval is a regular function of the parameters characterizing the initial conditions of the discharge. For a given diameter of the discharge tube it varies approximately as the fourth root of the mass of gas per 1 cm of length of the discharge gap.

In a pulsed discharge with a large current rise rate, the inductive voltage drop in the plasma considerably exceeds the active one. Therefore, using the current and voltage oscillograms, one can find the dependence of the inductance of the plasma cord on time and, from these data, determine how the radius of this cord changes at different stages of the process. Such an analysis shows that in all cases the very initial stage of the process is characterized

with an increase in inductance caused by the compression of the plasma in the direction of the axis of the discharge tube. The plasma is compressed the faster, the greater the initial rate of rise of the current (i.e., the value of the derivative \(\frac{dI}{dt}\)) and the lower the gas density. At the moment when a kink appears on the current oscillogram and the inductive voltage drop begins to decrease sharply,

Fig. 2. Oscillogram of current and voltage for a discharge in deuterium at
\(V_0=40\) kV and \(P_0=5\cdot10^{-2}\) mm Hg.

Fig. 3. Oscillogram of current and voltage for a discharge in deuterium at
\(V_0=40\) kV and \(P_0=0.2\) mm Hg.

the inductance begins to decrease. This means that this instant of time corresponds to the maximum degree of compression of the plasma cord. Immediately after this, a rapid expansion of the plasma occurs. If on

if several breaks are observed on the current oscillogram, this means that successive compressions and expansions of the cord are taking place.

These conclusions, obtained from analysis of current and voltage oscillograms, are fully confirmed by data obtained with the aid of high-speed cinematography of pulsed discharges in tubes with transparent walls. In the accompanying photograph (Fig. 4) four successive frames are shown, obtained in filming a pulsed discharge in deuterium at a pressure of \(0.1\) mm Hg and a maximum current of about \(200\) ka. These frames, following one another at intervals of half a microsecond, cover a small interval in the development of the process near the instant of time to which the break in the current and the drop in voltage correspond. The minimum diameter of the plasma cord corresponds precisely to this instant of time (the cine frames are phased with the current and voltage oscillograms).

Fig. 4. Frames from cinematography of a pulsed discharge.

The next photograph (Fig. 5) was obtained by using the cine apparatus in the mode of continuous photographic recording. A narrow slit, arranged perpendicular to the axis of the discharge tube, cuts out a short section of the discharge gap, and its image is swept at high speed along the motion-picture film. In this way a continuous picture of the change in diameter of a small section of the plasma cord is obtained on the film. The photograph shown was taken for a discharge in deuterium with a maximum current of about one million amperes. The initial gas pressure was \(10\) mm Hg. The instant of maximum compression and the subsequent development of the process are clearly visible.

Figure 6 shows a photograph of a contracting plasma cord obtained with the aid of a Kerr cell.

Valuable information about the principal physical processes occurring during an intense pulsed discharge is provided by measurements of the magnetic and electric field strengths in the plasma. On the basis of measurements of the magnetic field one can draw the following

Fig. 5. Streak photograph of the discharge in deuterium at a pressure \(P_0 = 10\) mm Hg. The electrodes are hemispherical. The distance between the electrodes is 45 mm. Sweep scale: \(1\ \mu\text{s} = 18\) mm. Chamber diameter: 180 mm.

Fig. 6. Moment of compression of the discharge. Exposure \(0.2\ \mu\text{s}\). Photograph of the discharge with a Kerr cell. Discharge in deuterium at pressure \(P_0 = 1\ \text{mm}\). Distance between the electrodes \(45\ \text{mm}\), chamber diameter \(180\ \text{mm}\).

Fig. 7. Photograph of the discharge obtained with a Kerr cell \(2.2\ \mu\text{s}\) after the start. Exposure \(0.2\ \mu\text{s}\). Initial deuterium pressure \(P_0 = 1\ \text{mm Hg}\). Distance between the electrodes \(40\ \text{mm}\).

...the pattern of current distribution in the plasma. Immediately after breakdown, the region occupied by the current is a thin cylindrical layer adjacent to the walls of the discharge tube. The inner boundary of this layer at first slowly, and then more rapidly, contracts toward the axis. Owing to the motion of the inner boundary of the current, after a certain interval of time it fills the entire tube. The instant at which the current reaches the axis practically coincides with the first kink in its oscillogram. At this moment the current density near the discharge axis exceeds the average current density over the cross-section of the tube by several tens of times. During subsequent expansions and contractions, the current density in the central region, with a radius of several centimeters, remains very high at all times, although it undergoes noticeable oscillations.

The distribution of current density over the cross-section of the discharge tube at different instants of time is shown schematically in Fig. 8.

Fig. 8. Distribution of current density over the cross-section of the discharge tube at different instants of time.

On the left is shown the distribution of current density at the very initial stage of the discharge. The drawing in the center corresponds to the instant at which the current moves toward the axis. On the right is given the distribution of current density after the first compression of the plasma cord. An interesting feature of this stage of the process is that in a certain zone of the discharge the current changes direction and flows in the opposite direction. The quantity that directly characterizes the dynamics of the processes of a pulsed discharge is the velocity of motion of the ionized gas. In a plasma with sufficiently high conductivity this velocity is determined by the ratio of the longitudinal electric-field strength \(E\) to the magnitude of the magnetic field \(H\):

\[ v=\frac{cE}{H}. \]

Measurements of \(E\) and \(H\) show that in a pulsed discharge with a rapid rise of current the velocity of radial motion of the plasma can be very large. For gases of low density, the maximum velocity during compression and expansion of the plasma cord...

reached hundreds of kilometers per second. This means that the kinetic energy of the directed motion of the ions in the plasma reaches several hundred electron-volts.

One of the most interesting effects observed in powerful pulsed discharges in light gases is the occurrence of hard radiations. In 1952, soon after the beginning of experimental investigations of pulsed discharges, it was discovered that, at a sufficiently large current intensity, a discharge in deuterium becomes a source of neutrons.

The first experiments set up to study this phenomenon showed that neutrons appear under such conditions when the maximum current intensity in the discharge reaches 400–500 ka and the initial pressure of deuterium is about 0.1 mm Hg. Neutron radiation was observed in a rather narrow pressure interval, and its intensity increased rapidly with an increase in the voltage on the discharge tube, i.e., with an increase in the maximum value of the current. The neutron indicator in these first experiments was the radioactivity of a silver target placed in a paraffin block near the discharge tube. Since it could be assumed that the observed neutron radiation is connected only with the bombardment by accelerated deuterons of deuterium adsorbed by the electrodes or the walls of the tube, control experiments were set up which did not confirm this simplest supposition.

At an early stage of the research it was quite natural to admit that the neutrons arise as a result of thermonuclear reactions in the plasma heated to a high temperature. This effect had been expected in advance, and in favor of such a point of view spoke, above all, the circumstance that it had been discovered under conditions fully corresponding to a priori theoretical predictions. The dependence of neutron radiation on pressure and current magnitude observed in the first experiments was in qualitative agreement with the assumption that this phenomenon is caused by a thermonuclear mechanism. However, after a very short time serious doubts arose as to the correctness of such an attractive hypothesis. They appeared after it was established that neutrons also arise at comparatively small values of the current intensity in discharges, with a maximum current of about 150 ka. According to the initial calculations, the intensity of thermonuclear reactions at a current intensity of about 150 ka should have been practically equal to zero.

In subsequent experiments, scintillation counters with output to an oscilloscope were used for the registration of neutrons. With the aid of this method it was established that neutrons always arise before the second kink on the current oscillogram, i.e., at the moment when the plasma undergoes secondary compression (Fig. 9). At the moment of the first compression of the plasma, neutrons do not arise. The emission of neutrons always has the character of a short pulse with a steep

front. The growth of this pulse occurs over several tenths of a microsecond. These principal results of the oscillographic investigation contradict the initial assumption that the emission of neutrons is the result of quasistationary heating of the plasma, in which the temperature increases in proportion to the square of the current strength.

In the course of further study, many interesting facts concerning the neutron radiation of the plasma were discovered.

Fig. 9. Oscillogram of the current and neutron pulse for a discharge in deuterium at \(V_0=40\ \text{kV}\) and \(p_0=5\cdot10^{-2}\ \text{mm Hg}\).

In particular, it was established that in discharge tubes of special design neutrons can arise at fairly large values of deuterium density, up to initial pressures on the order of several tens of millimeters of mercury. This fact indicates that neutron radiation is a very nontrivial effect.

It was found that a pulsed discharge is a source not only of neutrons, but also of hard X-rays. Hard X-ray radiation arises when large currents pass through hydrogen, deuterium, and helium. The radiation in discharges in deuterium always consists of short pulses. The pulses caused by neutrons and by X-ray quanta can be precisely phased on oscillograms. It then turns out that they arise simultaneously. The energy of the X-ray quanta appearing in pulsed electrical processes in hydrogen and deuterium reaches \(300\)–\(400\ \text{keV}\). It should be noted that at the moment when quanta with such high energy arise, the voltage applied to the discharge tube is only about \(10\ \text{kV}\).

The theoretical analysis of the complex phenomena that occur in the plasma of a pulsed discharge, pulsating under the action of electrodynamic forces, is still at a stage where a number of facts have not yet received a satisfactory explanation. However, the general picture of the process is gradually becoming clearer, and some characteristic features of the phenomena are becoming comprehensible.

At the present time it is quite clear that the processes of compression and expansion of the plasma are not quasi-stationary processes, for which equilibrium between the forces of external and internal pressure is characteristic.

In the equations describing the dynamics of the pulsed process, the principal role is played by a term that takes into account the change in the amount of motion of the ionized gas under the action of magnetic-pressure forces. Therefore the kinetic energy of directed motion may, at certain stages of the process, considerably exceed the thermal energy concentrated in the plasma.

In the initial stage of the discharge the internal pressure in the plasma is very small; therefore the electrodynamic forces produce in the plasma an acceleration directed along the radius toward the axis of the discharge tube. The work of the electrodynamic forces is thus expended not on raising the temperature, but on imparting kinetic energy to the converging cylindrical layer of plasma. At this stage the discharge tube acts as a peculiar accelerator, in which the particles are accelerated by the magnetic field. Since charged particles of different signs move with the same velocity, the ions thereby acquire a large kinetic energy, whereas the energy of the electrons, because of their small mass, remains almost unchanged. From the point of view of gas dynamics, the compression process should be regarded as a phenomenon in which a cylindrical shock wave is formed in the plasma, converging toward the axis. In front of the inner front of this wave there is at first neutral gas. As the wave moves, the gas is carried along together with the charged particles of the plasma, and at the same time ionization of its atoms takes place. Therefore the mass of matter set in motion gradually increases, and the total number of electrons and ions in the plasma rapidly increases.

If one calculates the velocity acquired by the compressed gas as a result of the work of the magnetic forces, then one can determine the duration of the compression process. It turns out that it must be approximately proportional to the quantity

\[ \sqrt[4]{\frac{M}{V_0^2}}, \]

where \(M\) is the mass of gas per unit length of the discharge tube, and \(V_0\) is the initial voltage. This is in full agreement with the empirically found dependence characterizing the duration of the time interval from breakdown to the first kink on the current oscillogram.

The last stage of cumulative compression begins when the plasma accelerated by the magnetic field reaches the axis. At this moment, a considerable part of the energy of directed motion is converted into heat, which leads to a sharp increase in the pressure and temperature of the plasma. In the phase of maximum compression, the plasma temperature reaches a value on the order of a million degrees. The nature of the processes occurring at the moment of maximum compression is not yet very clear; however, it is obvious that after the moment of maximum cumulation there must arise a diverging shock wave that carries the plasma toward the walls. When the diverging wave arises, a rarefaction zone must form inside it. The diverging wave must be rapidly slowed under the action of compressive currents of electrodynamic forces, owing to which a compression phase again sets in. It differs from the first phase in that, during the repeated compression, the density of matter in the internal zone of the discharge is small, and the gas in this zone is apparently practically completely ionized. Owing to this, in the phase of repeated compression conditions are created that favor acceleration, in the longitudinal electric field, of some group of ions and electrons located near the axis of the discharge, i.e. in the region where the magnetic field is small. Here one may see a certain analogy with the acceleration mechanism proposed by Fermi in the theory of the origin of cosmic rays. The plasma, possessing high electrical conductivity, moves together with its magnetic field and, for the particles in the internal zone, plays the role of a converging magnetic wall from which the electrons and ions trapped inside it are reflected many times, increasing their energy each time. Acceleration of ions and electrons in the longitudinal electric field near the discharge axis is also possible and is the cause that gives rise to the appearance of neutrons and hard X-ray radiation. The intensity of the longitudinal electric field in the phase of the second compression is very large. It may exceed many times over the value determined solely by the external voltage applied at that moment to the discharge tube.

It should be noted, however, that by no means all aspects of this acceleration mechanism have been clarified. It should be noted that, under certain conditions, owing to the influence of fields created by space charges, acceleration of ions in the longitudinal electric field proves possible even outside the central zone of the discharge. Certain types of instability characteristic of a plasma cord may play an important role in the process of accelerating particles in the plasma. In particular, one of the types of this instability observed experimentally may be of great importance for electron acceleration. It consists in the spontaneous generation of a longitudinal magnetic field in the plasma as a result of vortical twisting of the plasma cord. If, after the second compression, several more radial oscillations of the plasma cord occur, then the process of particle acceleration may

may be repeated several times. In the experiment it has so far been possible to observe no more than three successive oscillations. This, however, may be explained by the fact that at a certain moment the interaction of the plasma with the walls of the discharge chamber begins to make itself felt, leading to evaporation of the wall material and to the appearance of an appreciable amount of foreign gases in the volume.

We have considered some features of the phenomena discovered in the study of powerful pulsed discharges in gases of low density. The prospects for further work in this direction depend to a considerable extent on whether it proves possible to create conditions under which the plasma column, during the rise of the current, will undergo multiple oscillations without touching the walls. There are serious doubts that this can be achieved.

In assessing the prospects of various directions that may lead to the solution of the problem of obtaining thermonuclear reactions of high intensity, we cannot at present completely exclude further attempts to reach this goal by using pulsed discharges. At the same time, we must carefully study other possible approaches to the problem. Of considerable interest are those in which stationary processes can be used.