US2724057A - Ionic centrifuge - Google Patents

Description

1955 J. SLEPIAN 2,724,057

IONIC CENTRIFUGE Filed Jan. 21, 1944 2 Sheets-Sheet 1 [Zecifarz z'c Fewer Sup 02y 7'0 Vacuum Par/7p WITNESSES: 39

INVENTOR 4/ ATTORN Y United States Patent IONIC CENTRIFUGE Joseph Slepian, Pittsburgh,.Pa.. .assignor to Westinghouse .Electric Corporation, East Pittsburgh, Pa., a corporation of Pennsylvania Application January 21, 1944, Serial No."519,185

18 Claims. (Cl. 250-413) My invention relates to electrical discharge apparatus, and has particular relation to apparatus for separating molecules or ions having diiferent masses, such, for example, as the molecules or ions comprising isotopes of the same element.

Heretofore, ions or molecules having different masses have been separated from composites containing, them by such methods as diffusion, mechanical centrifuging and employment in effect of high vacuum mass spectrometers and magnetrons in which the massesto be separated were, after ionization, projected electrically through magnetic fields. In general, the degree in which the resulting mixture can be enriched in respect to a component of one mass is roughly proportional to A? is where s is the natural logarithmic base AW is the difference in the kinetic energy of the average translatory motion which is imparted by the separating means to the masses to be separated, T is the absolute temperature attained by .the molecules in the course of the separating process, and k is Boltzmanns constant. Where the attempt is made to separate molecules differing from each other in mass by an amountwhich is only a small fraction of their actual mass, which situation is characteristic of many isotopes which his desirable to separate, it can be shown that in thecase of the diffusion and centrifuging methods, AW is a small fraction of 'kT; hence the value of M kT which indicates the degree of enrichment per unitoperation, is extremely small. Hence, the diffusion and centrifuging processes are extremely slow and uneconomical for practical purposes. While in the case of both the mass spectrograph and the magnetron, AWcan be raised to many times kT and the above-mentionedenrichment factor is correspondingly high, the ion current traversing the magnetic field must be held to an extremely low density in order to prevent the effects of space charge and collision between ions from distorting the trajectories of the ions through the magnetic field to such an extent as to deflect the ions away from the proper collecting electrodes.

In my application Serial No. 447,679, filed June 19, 1942, entitled Ionic Centrifuge, and assigned to the as- .signee of the present application, I have described an ionic centrifuge in which large ion currents are caused to flow through a vacuum in a magnetic field, and to deposit ions of different mass on sepaarte electrodes in which, instead of following paths in which substantially no collisions occur as in the mass spectrograph and the magnetron, the ions may be so numerous thatthecollisions characteristic of conditions in gases at higher pressures occur, and are. even an important factor in separaice tion of the particles of different mass. Further study of the behavior of ions, under the conditions subsisting in the arrangement just referred to, has evolved the possibility of substantially increasing the effectiveness of separation of molecules even of the slight difference of mass characteristic of many isotopes; and this application contains a description of the modifications of structure shown of the earlier apparatus which are effective in obtaining the improved results just mentioned.

One object of my invention is, accordingly, to provide apparatus for efiiciently separating molecules and ions of different mass.

Another object of my invention is to provide apparatus for eificiently separating isotopes of the same element.

Still another object of my invention is to provide apparatus for separating'molecules of different mass which shall be efiicient in action and capable of rapidly yielding large quantities of the separated materials per unit size of the apparatus.

Still another object of my invention is to provide a novel source of gaseous ions.

Still another object of my invention is to provide a novel source of ions of isotopes of uranium.

Other objects of my invention will become apparent upon reading the following description, taken in connection with the drawings, in which:

Figure 1 is a view in vertical diametrical section of a preferred form of apparatus for practicing my invention;

Fig. 2 is a sectional .view along the line II-Ii in Fig. 1; and

Fig. 3 is a graphical plot of curves useful in explaining my invention.

In devising the arrangement disclosed in my abovementioned application, the interactions between ions and electrons traversing the magnetic field were treated as of negligible importance. In my latest studies, I have developed theoriescoveringthe interactions and, as a result, I have discovered the principle governing certain alterations of the apparatus structure, and certain methods of operating that structure, which greatly increase the yield of an ionic centrifuge of given spacial dimensions. I will first describe the structure which I prefer as illustrated in the drawings; and will then discuss the bearing of the principles I have discovered on the mode or" operation of the structure.

Referring then in detail to Figures 1. and 2, an ionic centrifuge embodying the principles of my invention may comprise a cylindrical casing 1 which may be of insulating material, but which may also, as shown here, be of metal. The respective upper and lower ends of the cylinder are closed by end plates .3, 4 which are preferably connected to the cylinder 1 by flanges provided with suitable gaskets for rendering their junction vacuum-tight. The interior of the cylinder 1 is intended to be operated at pressures widely different from that of the atmosphere, and so all joints thereto are preferably of types which are tight againstgas leakage. In the center of the top plate 3 is fastened one end of a Sylphon 5 the upper end of which supports a so-called Wilson seal, through which an electrode rod 6 projects downwardly into substantially the center of the cylinder 1. The Wilson seal is, 'as' shown, a conical rubber diaphragm 8 having a hole through which the rod 6 may be moved up and down. Such a seal is substantially tight against gas leakage provided the rubbing surfceas against the rod 6 arecovered with stop-cock grease or other suitable viscous material. When the pressure below the diaphragm 8 is less than that of the atmosphere, the latter presses the edges of the diaphragm 8 tightly into contact with the rod 6 and forms, with the aid of the stop-cock grease, a substantially tight seal. Nevertheless, the friction between the edges of the diaphragm 8 and the rod] 6 is sufficiently sl ght so that the rod 6 may be slid up and down through the diaphragm 8. V

In vertical al'gnment with the rod 6 is a second rod 9 which projects through a tube entering the center of the bottom plate 4 and connected at 11 to a vacuum pump. A Wilson seal 12 is arranged to permit the rod 9 to slide through a substantially gas-tight joint, and permit vertical movement of its upper end within the cylinder 1. The lower end of the rod 6 carries an electrode 15 which is preferably composed of a mixture of uranium and uranium oxide, pressed into a cylindrical form. A shield 16 of quartz or of glass capable of withstanding the proximity of the are between the electrodes 15 and 16 is arranged to prevent the projection of material from the arc toward the upper end of the cylinder 1. The upper end of the rod 9 supports an electrode 13 of metallic uranium or a mixture of uranium and uranium oxide. Below the electrode 13 is what may be called a catcher 14 which is roughly, saucer-shaped, and for which glass may be used.

Supported in staggered and overlapping relation below the top plate 3 by suitable insulating bushings 17 is a series of rings 18, 19. The staggered and overlapping relation accomplishes several purposes. First, it causes the end plates 3 and 4 to be completely shielded against the arrival to them of ions and electrons from the discharge in the main space of the ionic centrifuge. It allows relatively free passage of the gases to be drawn from the vessel by the vacuum pump. It permits the spaces between adjacent rings to be large enough so that bridging of adjacent rings by small conducting particles from the arc is unlikely. I have shown, for purposes of illustration, seven such rings, but it will be recognized by those skilled in the art that the number may be varied in accordance with the overall dimensions of the cylinder 1, a greater number being used for cylinders of large diameter than for those of small diameter. The width of the respective rings 18 and 19 is preferably made such that their total exposed areas are the same, the rings of smaller diameters thus being of greater width as shown. By the term exposed area of the rings 18, I mean that part of the actual area which would be exposed to the incidenceof ions projected parallel to the magnetic field and passing through the annular gaps between electrodes 19. Projecting downwardly from the lower face of each ring and electrically connected to it are annular bands 21, the axial dimension of these bands being somewhat greater than the distance between them. While I have shown buta single band attached to each ring, it will be recognized that under certain circumstances it may be desirable to attach two or more of such bands to each ring.

Spaced away radially inside the cylindrical wall 1, by a distance suiiicient to withstand the operating voltages to be described in more detail below, is a pair of metal cylinders 22 and 23 which are insulated from each other by a gap at their midpoints which is sufiiciently wide to withstand the operating voltages described below. The annular space between the cylinders 22, 23 and the wall I, is broken up by insulating walls 23A into nearly completely separated smaller spaces or cells. These insulating walls may be glass or porcelain tubes, pushed down between the cylinders 22, 23 and the wall I. Breaking up the annular space into cells greatly increases the voltage which can be impressed between the cylinders 22, 23

and the wall I, without a discharge taking place, in the presence of the magnetic field. The gap between the cylinders 22 and 23 is preferably shielded on its interior side by a projecting ledge 24 on one of these cylinders.

axial length of the cylinders 22 and 23 should be sub stantially the same. The upper cylinder 22 may be supported by one or more insulating bushings 25 projecting through the top plate 3 and provided with an inleading conductor 26 by which its potential may be fixed as desired. The lower cylinder 23 may be supported in one or more insulating bushings 27 projecting through the bottom plate 4, and be provided with an in-lead 27'.

Normal to the interior wall of the cylinders 22 and 23 and electrically connected thereto are positioned a series of metallic bands 28 of annular form. The radial breath of these bands is not critical so long as they do not occupy a major portion of the space within the cylinder 1, and their axial spacing is made somewhat less than their radial breath.

Supported from the bottom plate 4 by insulating bushings containing inleading wires, substantially similar to the bushings 17 already described as located in cover plate 3, are staggered and overlapping rings 18, 19 substantially identical with the rings 18, 19. The rings supported from bottom plate 4 are provided with upwardly projecting bands 21' substantially identical with the bands 21 already described in connection with the rings 18, 19.

An arc struck between the electrodes 13, 15 is arranged to be supplied with current from a direct-current source 31 through circuit connections embodying a rheostat 32, the positive terminal of the source 31 being preferably connected to the upper rod 6 and its negative terminal to the lower rod 9. The upper end of the rod 6 is pivoted to a lever 34 having a fulcrum 35 and biased upward by a spring 36. A dashpot 37 is preferably arranged to damp the movements of the rod 6. The lower end of the rod 9 is connected to a nut 38 supported on a screw 39 which is supported on a portion of the foundation which supports the cylinder 1. The lower end of the screw 39 is provided with a pair of bevel gears 41 which are arranged to be turned by a motor 42 under control of a manually operated switch 43. When the switch 43 is closed, the electrode 13 is slowly raised to a higher position within the cylinder 1. The purpose of this arrangement will be described in more detail below. The strength of the bias spring 36 is arranged to be such that when the cylinder 1 has been exhausted to a high vacuum, air pressure on the rod 6 plus its weight is sufficient to depress the electrode 15 into contact with the electrode 13. Current then flows from the source 31 through the rheostat 32 which acts through relay 33 to energize a solenoid 44, thereby moving the lever 34 and raise the rod 6, and draw an are between the electrodes 13 and 15. The arc so drawn is not perfectly stable, and may frequently go out. When this happens, the atmospheric pressure will. again depress the electrode 3.5 into contact with the electrode 13. and the. arc will be restruck. This repeated periodic striking and extinction of the are between electrodes 13. and 15 will go on periodically at a rate limited by the adjustment of the dashpot 37.

Current flow in the are between 13 and 15 is originally initiated by closing manual switch 43 to move electrode 13 upward until it strikes electrode 25. Manual switch 4-3 is then opened.

By means of an electromagnet M-M, a substantially uniform magnetic field parallel to the axis of the electrodes 6, 9 is produced throughout the interior of the cylinder 1. By means of the voltage source 46 having its positive end connected to the rod 8 and provided with a plurality of tapped potentiometers, various voltages may be impressed, through inleading conductors passing through'the bushing 17, on the rings 18, 19, and the similar rings 18, 19 and also on the two cylinders 22, 23. However, where the device is to be operated to separate the isotopes of the element uranium as is shown in the drawing, it will, for reasons to be pointed out below, usually be preferable that the cylinder 23 and the bands 18', 19' shall be electrically disconnected and freely i when it is negligible.

flo'ating in electricalpotentiali Howevennnder certain. circumstances of operation of the': ionicficent-rifu'gaw for.

example, when it is desired to separateisotopes of the lighterelements, it'will be found; in accordance with the principles pointed out below, tobe desirable to impress certain potentials on both of'the cylinders 22 and 23 and onboth sets of rings 18, 19and 18 19. Provisions for the impression of such potential have, accordingly, been described above.

Assuming that some or all of the rings connected to the end plates 3 and 4 and the cylinders 22and 23 are given potentials relative to the cathode 13 of the arc operating between the electrodes 13, 15, arr-electric field will be produced having a radial component in'the'space' surrounding these electrodes. I will call' the radial component of this electric field B.

When the arc between the electrodes 13 and is burning and if the vacuum is good enough, it serves as a source" of positive ions and electrons; both of which pass radially outwards from the arc, irrespective of the polarity of the radial electric field E. This seems paradoxical,

but has been conclusively demonstrated experimentally by me. I believe this effect is due to a" strong interaction between electrons and ions in the space, and I have worked out a mathematical theory based on this interaction which accounts for allthe effects which I have observed. While I shall not present this mathematical theory here in all its details, I shall try todescribe this theory qualitatively in a physical way.

By the interaction between the electrons and ions I mean the following. If in a portion of space occupied by both electrons and ions, .the electrons are caused to move in any direction relatively to the-ions, then these electrons in so moving exert a dragging force upon the ions and make them tend to follow the' motion of the electrons. Similarly if the ions are moved by any. means relatively to the electrons, they exert a dragging force upon the electrons making them tend to follow the motion of the ions. Of course this dragging effect of the electrons upon the ions in any region ofrspace-is equal in magnitude and opposite in direction to the dragging effect of the ions upon the electrons in that same region of space.

The degree of this interaction between ions and electrons is very much greater than would be anticipated by those skilledin the art. In cases whichI have examined, I find the interaction of the electrons with the ions tobe more than a million times as great as the interaction which takes place between electrons and an equal density of electrically neutral molecules. I shall not attempt to expla'inthis high degree of interaction betweenelectrons and ions here, but shall accept it as an experimentally determined fact.

This interaction between electrons and ions causes the average paths of the ions and electrons to'be outwardly winding spirals, where otherwise theym'i gh't be concentric circles about the central axis, as I shall presently explain. The paths of the ions thendeviate f'romthose contemplated in my above mentioned earlier application, where the density of ions and electrons must be assumed to be so low that the eifects of the electron-ion interaction are negligible. It then turns out that the conditions for optimum separating eifect are diflerent, when the ion-electron interaction is considerable, than The establishing and maintaining of these conditions is the purposeof thisinvention.

To see why both ion and electron paths spiral radially outward regardless of the sign of the electric field E consider first the case where we have an outwardly directed positive electric field, tending to draw ions outwardly, of magnitude E, and less than the quantity which is known as the Larmor gradient e being-thecharge For the electronwe may set m equal Zero in the Formula 1, and we find that if-thereis' no interaction between electrons and ions, the electrons may move in equilibrium circular orbits with peripheral velocity We in the same rotationalsense as v and given by Comparing (1) and (2) we see that the electrons go around their orbits more slowly than the positive ions.

In going around their orbits with the peripheral velocity v8 given by (1), there is a balance between these forces on=the ions; the centrifugal force radially outwards, the pull of the electric field, eE, radially outwards, and' the reaction of the magnetic field, evaH, radially inwards.

Now suppose there is interaction between the electrons moving with the slower velocity We, and the ions moving with the greater velocity v6. This reaction causes the ions to move more slowly than v8 given by (l) and upsets the balance between the forces acting on the ions moving in the circular orbit. Itmaybe shown that this upset is such as to give the ions a radially outward acceleration, or in other Words,,the ion will leave its circular path; andmove outwards on a spiral.

Similarly-with the electron. With the peripheral velocity V's given by (2 there isequilibrium between the pull of the electric field, -eE, directed inwards, and the reaction of the magnetic field -eHv'@ directed outwards. Now, the interaction of the ions with the electrons drags the electrons forwards, so that they move with a greater peripheral velocity than v'a. of Equation-2. Hence the outwardly directed magnetic reaction -eI-Iv is increased while the inwardly directed force -eE is unchanged. The result is that the electrons move outwards on a spiral.

Suppose, now, the electric field is directed inwards. Then the ions, in the absence-of-interaetion, may move in circular equilibrius paths with VB still given by (1) but with E now having a minus sign. The electrons likewise may move in circular orbits with v being still given by Equation 2. But now, because of the change of sign of E, the electrons move with a greater peripheral velocity than the ions. Hence, if there is electron-ion interaction, the electrons are slowed down, and the ions are speeded up. The slowing down of the electrons lessens the now inwardly directed magnetic reaction KV'QH, while leaving r the outwardly directed force eE unchanged. Hence, under the influence of interaction, the-electrons will spiral outwards. Similarly, the speeding up of the ions, due to the electron interaction, increases the outward centrifugal force negative.

As explained in the aforementioned earlier application, the ionic centrifuge separates the iostopic ions in virtue of the difference in their centrifugal force arising from their diiference in mass. For this difference to be effec tive, the'ions must make many revolutions before they reach the various electrodes upon which they deposit. That is, the ions must travel outwards from the arc in spirals of very small pitch. If the pitch gets as large as 45, that is if the spiral path-makes an angle of 45 with the radii which it crosses, the mathematical theory indicates that the separating effect becomes zero. An object of this invention is to make the ions travel outwards from their source in spirals of small pitch.

When the interaction between ions and electrons is negligible, as in my aforementioned earlier application, the motion of the ions in close spirals is accomplished by causing to be set up a radial electric field E, which is everywhere nearly equal to the Larmor gradient If the electric gradient is much less than this, the ions do not spiral out at all, but stay on small circles close to the arc, and thus do not come out any appreciable radial distance before they are caught by the electrodes. If the radial electric gradient is much greater than the Larmor gradient, then the ions move out on spirals which are of large pitch, so that the ions arrive at electrodes at larger radii, but there is practically no separating effect. Hence, when there is little interaction between electrons and ions, the radial electric field must be held close to the Larmor gradient .3. 2 em T if there is to be any considerable separating effect.

In my present invention, I contemplate operation of the centrifuge with ion and electron densities large enough so that their interaction is no longer negligible. In this case, according to the foregoing explanation, it is no longer necessary to operate at the Larmor gradient since the ions and electrons will spiral outwardly even with much smaller electric fields. Actually, in this case, the electric gradient must not become as large as the Larmor gradient, as then the spirals will become of such large pitch that little separating eifect will be obtained, It is an object of my invention to provide means for operating at such electric gradients that the ions will come out in spirals of sufiiciently small pitch that good separation will be obtained.

The mathematical theory which I have developed expresses results in terms of a parameter P which is defined as I f F- where Ir is the total radial current flowing across the space from the arc, and Ip is that portion of the current which is carried by positive ions. The total radial electric current is of course the algebraic sum of the electric currents carried by the positive ions and the electrons in their radial motion away from the arc. Ii. and I may have any ratio depending on the manner of operation, as I will now explain. If the radial electric gradient is very small, the radial speeds of the electrons and ions in their spiral paths are very .nearly equal, and the radial electron current is nearly equal to the positive ion current. The total radial current Ir is then nearly zero, and F is likewise nearly Zero. As the radial electric field is increased in the positive direction, the radial speed of the positive ions increases faster than that of the electrons, and F increases. At a sufficiently large value of the electric field, the electron radial velocity becomes zero and F=1.

For negative electric fields the electron radial velocity exceeds the positive ion radial velocity and F is negative.

The mathematical analysis previously mentioned shows that the pitch of the spirals on which the ions move is determined by the value of F. For small values of F, the pitch is small. As F increases, the pitch increases, and when F=1, the pitch angle becomes equal to 45 at which pitch angle the separating eifect becomes zero. It is then an object of my invention to operate. with a value of F large enough to make the spiral ion paths of sufficient pitch to bring the ions out to the desired radius, but not so large that the ion spirals pitch angle becomes so large that little separating effect results.

Besides being important as a convenient parameter in the mathematical theory, the quantity F is important practically, because by very simple devices it is possible as I will now explain to control the excitation of the electrodes 18, 19, 22, 23, etc. so as to make F have any desired value, even though the ion output of the arc and the degree of interaction between ions and electrons varies considerably. It is also possible by simple devices to control the excitation of the electrodes 18, 19, 22, 23, etc. so as to make F always take such values that the ions come out to a rather definite radius.

This comes about because the distribution of the positive ion currents flowing to the rings 18, 19, 13, 19' and the cylinders 22, 23 is determined by the pitch of the ion paths; in other words it is determined by the magnitude of the parameter F, and the latter is susceptible of control in a very simple way. In the space in which the ions and electrons move, relatively large radical electric fields, E, which have been the subject of the preceding discussion, can exist without giving the electrons enormous radial velocities because the magnetic field continuously converts the radial motion which the field tries to impart to the electrons into circumferential motion. In the axial direction, however, the electrons may move exceedingly freely, because there is no reaction of the magnetic field upon electron motion parallel to itself. Hence, the axial component of the electric field in the space is very small, or practically zero, for if it were not, large axial motions of the electrons would take place in such a sense as to reduce to zero the electric field which is tending to produce them.

Because the axial component of the electric field is nearly zero, the axial velocities of the ions remain quite small, and are those which they possess initially on leaving the arc, or which they acquire from their interactions with the electrons. I find that in the size of ionic centrifuge with which I have experimented, the axial motions of the ions corresponds to a kinetic energy of only a few volts, whereas the energy of the circumferential motion of the ions is several thousand volts. The axial motions of the ions are random, some up, and some down, and approximately equal numbers of the ions deposit on the electrodes in the lower and upper halves of the ionic centrifuge.

If the spiral path of the ions is of large pitch, then the radial energy of the ions is a thousand volts or more, whereas their axial kinetic energy is only a few volts. In this case then, practically all the ions will reach the cylindrical electrodes 22, 23, and very few will reach the ring electrodes 18, 19, 18', 19. If, however, the spiral paths of the ions is of so small pitch that their radial kinetic energy is only a volt or less, while their axial kinetic energies are several volts, then with proper tions of the ionic centrifuge as indicated in Fig. 1, practically all the ions will be collected by the innermost ring of 18, 19, 18', 19', and almost none will reach the outer rings or the cylinders 22, 23. Thus it is evident that by properly adjusting the potentials impressed on the electrodes just mentioned it is possible to predetermine the portions of the electrode system on which the ions will be deposited. I v

If, in theoperation of the ionic centrifuge, a ring electrode, is disconnected from the power supply sothat it is left electrically floating, then the total current't'o it is neces sarily equal to zero. In this case, as many electrons must reach the ring as positive ions. For such a ring, the electron current received equals in magnitude the positive ion current received. Such a :ring will float at very nearly the potential of the space above it.

We may, however, instead apply such large negative voltages to a ring electrode, that all the electrons approaching it are rejectedorturned back, and only positive ions arereceived. For such an electrodethe electron currentreceived is zero, and the totalcurrent deposited there equals the positive ion current. Such an electrode will have a potential negative relative to the potential of the space opposite it. This difference between the potential of the ring and thepotential of thespace will exist across a relatively thin sheath in which :the electrons are turned back, and the ions are accelerated. In practice it will not be necessary to make theelectrode more than a few volts negative with respect to the space opposite it to turn all the electrons back. However, the adjustment is not critical, since a larger negative voltage on the ring continues to turn all the electrons back, and merely receives the ions with the greater kinetic energy acquired by 'fallingthrough the axial field in the sheath next to the electrode.

It is now evident, that by allowing some of the electrodes to float electrically, so that theirtotal currentis zero, and by energizing the others sufiiciently negatively so that the electron current to them is zero, it is .possible to cause the value of Wemay energize the upper " ring electrodes 18, 19, each sufliciently negativelyso that they collect no-electron current; we allowthe lowerring electrodes, 18', 1 9!, to float electrically, so that theyco'llect electron currents and ion currents equal in magnitllde'; we-energize 'the upper half cylinder electrode 22, negatively, so that it receives only positive ion current, and we allow thelowerhalf cylinder electrode, 23 to float electrically so thatit receives electron current equal in magnitude to the positive ion current. When this is done, since the ions go 'symmetri eallytto the *upper and lowereleetrodes, the total electron current will equal oneihalfrthe total ion current and When operated in this way, the radial electric fields in the space take such values as to make the radial electron current equal to one half the positive ion current, and the excess negative potentials applied to the excited rings are consumed in the thin sheaths which form adjacent to them.

It is clear that instead ofenergizing all of the upper rings negatively, and letting all the .lower rings float, we could energize alternate upper rings, and let the other upper rings that, and energize alternate lower rings negatively, and let the otherlower rings .floa-t. The energized electrodes then still will take /2 the total positive ioncurrent, and we will have In .general, we may arrive at .any -desired value of F .by impressing potentialfrom source 46in Fig. lonthe traction ;F of the totalelectrodetarea, and we may allow for nonauniform distributioniof theqions overithe various i0 electrodes, by dispersing the energizedf raction F over the whole available electrode area.

In the foregoing I have described the results, which I have been led to by experiment and theory, relating to the average or mean motions of ions and electrons in the ionic centrifuge when the density of ions and electrons is so large that interactions between them cannot be neglected. When the ions are made up of more than one isotope species, each species Will follow paths consisting of spirals, but the pitches of the spirals will be different for the different isotopic species. Thus the proportion of the ions which depositon the innerrings, as compared to those which deposit on the outer rings or cylinders 'will be diiferent for the difierent isotopes, and thus the possibility of isotopic separation arises. The lighter isotopes will go out on spiral paths of small pitch and deposit on the inner rings, and heavier isotopes will go out on spiral paths of larger pitch, and deposit on the outer rings.

The spiral paths mentioned in the foregoing, are mean or average paths of the various ions. The actual paths of the individual ions will deviate from these average paths in random ways, due to the fortuitous variations in the detailed interactions between electrons and ions, and also the variations between ions and ions. The total motion of the ions and electrons may be described as a following of the average spiral paths described above, but with a random helter-skelter motion superposed on this average motion. The motion may be likened to the motion of the molecules of a moving gas, where all the molecules partake of the common motion of the gas as a whole, but in addition possess a random or helterskelter motion otherwise called their heat ortemperature motions. For this reason, I assign a temperature to the random motions of the ions, letting be the mean energy of the random motion, or deviation from the average motion, of the @ions, where k is the Boltzmann constant, and T is the temperature assigned to thexions.

Experimentally, I find that as the energy of the ions in their average spiral paths is increased, their temperature also increases, but not in proportion. When the temperature of the ions becomes so high that kT equals several electron-volts, then the temperature increases very slowly as the energy in the average spiral path further increases. With uranium ions, when the energy of the ions in their spiral paths is 1000 volts, the energy of the random motion, kT, is still less than 10 volts.

This random or helter-skelter motion of the ions causes adifrusion of the ions which tends to undo any separating efiect which the motion in the average spiral paths would produce. This effect of diffusion may be excepted to limit the enrichment of the light isotope on the inner rings to a factor ,n 'kiT where AW is the difierence between the: kinetic energies per ion of the two isotopes in their average or spiral motion, and

is the average energy of random motionper-ion.

We may then estimate the separating effect of the average motion in the spiral paths, but expect the effect of the diffusion arising from the random ion motion to set a limit, so that the enrichment factor cannotexceed ,a r lcT The resultsof mycalculation ofthe separating effect of the motion in the average spiral, with *theeffects of scribed before.

diffusion arising from the random motions of the ions neglected, are shown in Fig. 3, as carried out for the two isotopes of uranium of masses 235 and 238. The abscissa is F, the ratio of the total radial current to the positive ion radial current. The ordinate is J1 Am where y is the difierence in radial velocities of the two isotopes divided by the average of their radial velocities. Evidently, when y is greater than 2, perfect separation of the particles will result, as then the lighter isotope cannot have a positive radialvelocity, and therefore must all deposit on the innermost electrode. In this case, the enrichment of the lighter isotope on the inner rings will be limited by the diffusion factor to 91K kT For uranium, when y is equal to 2,

it Am is equal to 159, and a horizontal line has been drawn for this valve of i Am in the figure. The curves shown in Fig. 3 were calculated for various values of kI-I, where k is the mobility of the electrons relative to the ions and is a reciprocal measure of the degree of electron-ion interaction, and H is the intensity of the magnetic field. We see that if the degree of ion-electron interaction is such that kH is greater than 500, perfect separation is obtained with F equal to /2. With larger values of kH, Fimay become more nearly equal to l, to obtain perfect separation, but in no case must F be allowed to equal 1.

Where perfect separation is indicated by Fig. 3, the actual enrichment in the lighter isotope on the inner rings as compared with the outer rings will be limited by diffusion to an IcT If, as has been indicated by experiment, With uranium, AW may be 10 volts, and kT, also 10 volts, the actual enrichment factor may be as large as 2.72.

Where the ion yield from the arc is constant, operation may be carried on with a definite value of F, obtained by energizing a definite fraction of the collecting electrode area, and leaving the rest floating as has been de The value F should be chosen so as to make the separation perfect according to curve of Fig. 3, and at the same time bring a considerable ion current out to the outer rings by impressing on a large enough area of the outer rings a sufiiciently negative potential.

'12 I would represent ions of other mass, the only difierence being in the scale for ordinates l'L Am and in the values of kH with which the respective curves are labelled. The general conclusions as to the values of F to be preferred in operation are thus valid for ions generally.

Fig. 3 indicates the possibility of obtaining complete separation of isotopes for a wide range of values of F provided the proper value of kH is employed. However large values of kH correspond to operation with small current through the central are 13, 15 and hence with a slow yield of ionized particles to be separated. On the other hand operation with smaller values of F result in reduction of the enrichment factor IcT since AW is proportional to F. Therefore in practice a compromise should be struck between these two conflicting considerations and a value of F should be chosen which will yield the most desirable result for any particular isotopes. Conditions may arise in which the decreased enrichment factor corresponding to lower values of F will be less important than complete separation with low values of kH, and operation at low or even negative values of F may then be justified.

Where the ion yield from the arc is variable, the following method of energizing the electrodes is preferably used as it gives a varying value to F, larger when the ion This is desirable because diffusion limits the enrichment factor to Z AT shows that, with larger values of kI-I, larger values of F may be used and still give perfect separation according to Fig. 3.

While Fig. 3 represents conditions for the uranium isotopes of mass 235 and 238 curves of the same form yield and interaction is small, and smaller when the ion yield and interaction is large. According to this method, all the rings 18, 19, 18', 19 are energized with negative voltage, so that they receive only positive ions. Thus cylinders 22, 23, however, are left floating so that they always collect an electron current equal to the positive ion current which reaches them. With this electrical connection, the electric fields in the ionic centrifuge take on such values that a considerable fraction of the ions reach the rings, 18, 19, 18', 19, and also a considerable fraction of the ions reach the cylinder electrodes 22, 23. We arrive at this conclusion by showing that the hypothesis that only a very small proportion of the ions reaches the rings 18, 19, 18, 19' leads to a contradiction, and that the hypothesis that only a very small proportion of the ions reaches the cylinders 22, 23 also leads to a contradiction.

Suppose only every small proportion of the ions reaches the rings 18, 19, 18', 19'. Then practically all the ions will reach the cylinders 22, 23. But since cylinders 22, 23 are electrically insulated, or floating, the electron current to them must be equal to the positive ion current to them. The total current out from the arc must then be nearly zero, or

must be very small. But if F is very small, the pitch of the spiral paths of the ions must be very small and the ions will mostly reach the rings 18, 19, 18, 19. Thus we are led to a contradiction.

Suppose only a very small proportion of the ions reaches the cylinders 22, 23. Then practically all the ions will reach the rings 18, 19, 18', 19'. Since these rings are negatively energized, the ion current to them will be unaccompanied by electron current. Hence P will be nearly equal to 1. But if F is nearly equal to 1, the pitch angle of the spiral paths of the ions will be nearly 45, and most of the ions will reach the cylinders 22, 23. Thus again we are led to a contradiction.

Hence, with the electrical connection described, ions will arriveat both the rings 18, 19, 18, 19, and the cylinders '22, 23 in considerable numbers. Thus even with very wide variations in the ion-yield from the arc, with the accompanying wide variations in the degree of interaction, of electrons and ions, the ion paths will always be of sufliciently small pitch that a good separation of the isotopes is obtained between the deposits on the inner and outer electrodes, while at the same time, the pitch of the spiral ion paths is never permitted to become so small that practically all the ions reach only the inner rings.

1 shall now describe the functions of the cylindrical projections 21, attached to the rings 18, 19, and the ring projections .28, attached to the cylinders 22 and 23. When ions of a metal arrive at an electrode surface and give up their charges, the neutral metal atoms so formed do not all remain attached to the electrode surface, but a considerable fraction of those arriving are ejected from the surface as neutral molecules. This phenomenon is known as ion sputtering. The neutral molecules so ejected travel to other electrodes and deposit there.

If the projecting electrodes 21, and 28, were not used, practically all the molecules sputtered from any one ring 18, 19, 18, 19', 22, 23 would travel to remote rings, and so this phenomenon of ion sputtering would cause a considerable remixing of the material brought to the electrodes by the ions. The projecting electrodes 21 and 28 largely prevent this remixing. If the projecting electrodes 21, and 28, are made of very thin material, very few of the ions will arrive at their outer edges. Those that arrive at the face of a projecting electrode 21, .28, will, of course, cause neutral molecules to be sputtered away from these faces. However, the greater part of these sputtered molecules will be caught by the opposing faced the adjacent projecting electrode, 21, 28, and only few sputtered molecules will pass out between the projecting electrodes to travel greatdistances before finally depositing. Likewise, the molecules sputtered from the ring electrodes 18, 19, 18, 19?, 22, 23 willmostly deposit on the adjacent projecting electrodes 21, 28. In this manner, by reducing the distance travelled bythe sputtered molecules, the projecting electrodes 21, 28 reduce the remixing effect of these molecules.

In the foregoing I have described the general principles governing the operation of this invention, and have given preferred mannersof utilizing these principles. However, I do not wish to be limited to these preferred means, and I contemplate all manners of operation as are evident from application of the principles which I have discovered and here disclose. For example, Fig. 3 shows that good separating effectmay also be obtained when is negative, that is, when the electrodes 18, 19, 18, 19 are given positive potentials so that they draw larger electron currents than positive ion currents. Operation of the ionic centrifuge with positive potentials on the electrodes is then included within the scope ofmy invention.

In practicing my invention, the ionic centrifuge is operated, until a considerable amount of material has been transferred from the arc electrodes to the ion-receiving electrodes. Then the ionic centrifuge is taken apart, and the material collected by the receiving electrodes is removed by scraping or by chemical agents. The material from the inner receiving electrodes will be found to be richer in the lighter mass constituent or isotope, and the material from the outer receiving electrodes will be richer in the heavier mass constituent or isotope.

The material from the inner or outer electrodes respectively may then be used to make new are electrodes for use in the ionic centrifuge again for further enhancement of the separation of the light and heavy constituents or isotopes.

While I have described a particular embodiment of my invention, it will be recognized by those skilled in the aft that its princi les are of "broader a plication "in ways that will be apparent such persons.

I claim as my invention:

1. An electrical discharge device comprising a closed container'having at least a portion of its. walls of conductive material, a pair of electrodes centrally located therein, supply circuit connected to produce an arc discharge between said electrodes, means for insulating a conductive portion of the container Wall and permitting it to float electrically, and means for impressing a negative voltage relative to one of said pair of electrodes on the remaining conductive wall portions of said container.

2. An electrical discharge device comprising a closed container having conducting wall portions which line substantially its entire interior surface, two closely spaced arqproducing means located substantially in the "central region of said container, said conducting wall portions being divided into a'plura'lity of e'lectrically'separate constituent areas, and means for fixingat will the potentials of all of said separate areas relative to said arc.

3. An electrical discharge device comprising a closed container having its entire interior surface subtended by conducting sheaths, means for producing an electric arc in the central region of said container, said sheaths being subdivided into electrically insulated areas, means for maintaining certain of said areas electrically insulated from said are, and means for impressing on the remainder of said areas electrical potentials relative to said arc.

4. An ionic centrifuge comprising a cylindrical closed container enclosing two electrically insulated conducting cylinders shielding its internal side walls and having its end faces completely shielded by sets of electrically insulatedrings of conducting material, means for producing an electric are within said cylinders between a pair of electrodes at least one of which contains: an element having isotopes to be separated, means for impressing an axial magnetic field throughout the interior of said container, and means for connecting said electrodes'through a voltage source to certain of said rings, and means for maintaining certain others of said rings and cylinders insulated from said are.

5. An ionic centrifuge comprising a cylindrical closed container enclosing two electrically insulated conducting cylinders shielding its internal side walls and having its end faces completely shielded by sets of electrically insula'ted rings of conducting material, means for producing an electrical are within said cylinders between a pair of electrodes at least one of which contains an element having isotopes to be separated, and means for impressing an axial magnetic field throughout the interior of said container, said rings being provided with cylindrical extensions extending axially of said container.

6. An ionic centrifuge comprising a cylindrical closed container enclosing two electrically insulated conducting cylinders shielding its internal side walls and having its end faces completely shielded by sets of electrically insulated rings of conducting material, means for producing an electric are within said cylinders between a pair of electrodes at least one of which contains an element having isotopes to be separated, and means for impressing an axial magnetic field throughout the interior of said container, said cylinders being provided with portions extending radially inward.

7. An isotope separator comprising a closed chamber having a source of ionized vapor containing the isotopes to be separated positioned in its central region, means for producing a magnetic field through the space within said container, said container having a lining of conducting material divided into electrically insulated areas which subtend all portions of the interior Wall of said container, and means for connecting certain of said areas to said source of ionized vapor through an electrical current source, and for leaving the others of said portions electrically insulated from said source of ionized vapor.

8. The arrangement described in claim 7 in which said insulated portions are about equal in area to the portions connected to said current source.

9. The method of separating molecules having different masses which comprises providing a source of ionized vapor of said molecules in an enclosure having conducting wall portions, impressing on said vapor a magnetic field and an electric field at an angle thereto, insulating a fraction of said conducting wall portions from said source, and impressing on the remainder of said wall portions a negative potential relative to said source.

. 10. The method of separating molecules having diiferent masses which comprises providing a source of vapor of said molecules in an enclosure having conducting Wall portions, impressing on said vapor a magnetic field and an electric .field at an angle thereto, insulating substantially one-half of said conducting wall portions from said source, and impressing on the remainder of said wall portions a negative potential relative to said source.

11. The method of separating isotopes of an element from each other which comprises providing a source of vapor containing said isotopes in the central region of a container which completely encloses said source with conducting walls, insulating a fraction of the area of said walls from said source and impressing on the remainder of said walls a negative potential relative to said source, whereby an electric field having lines of force with a radial component about said central region is produced, and producing a magnetic field having lines of force nor mal to said radial components.

12. The method of separating isotopes of an element 7 from each other which comprises providing a source of vapor containing said isotopes in the central region of a container which completely encloses said source with conducting Walls, insulating substantially one-half of the area of said walls from said source'and impressing on the remainder of said Walls a negative potential relative to said source, whereby an electric field having lines of force with a radial component about said central region is produced, and producing a magnetic field having lines of force normal to said radial components.

13. An isotope separator comprising a closed cylindrical chamber having a source of ionized vapor containing the isotopes to be separated positioned in its central region, means for producing a magnetic field parallel to the axis of said cylinder, cylindrical walls of conducting material within said chamber at least partially surroundingsaid source and provided with means for regulating their electrical potential relative to said source, a plurality of annular electrodes at each end of said cylindrical chamber, said annular electrodes being insulated from each other and being arranged at each end in two stages axially displaced from each other, the radial width of the electrodes in each stage being greater than the radial gap between the electrodes of the other stage, and the annular electrodes of the one stage being in staggered relationship so as to subtend the gaps between the annular electrodes of the other stage.

14. An "isotope separator comprising a closed cylindrical chamber having a source of ionized vapor containing the isotopes to be separated positioned in its central region,,means for producing a magnetic field parallel to the axis of said cylinder, cylindrical Walls of conducting material within said chamber at least partially allel to the axis of said cylinder, cylindrical walls of conducting material within said chamber at least partially surrounding said source and provided with means for regulating their electrical potential relative to said source, a plurality of annular electrodes at each end of said cylindrical chamber, the annular electrodes positioned at at least one end of said cylindrical chamber being arranged in two stages axially displaced from each other, the electrodes in one stage being staggered so as to subtend the gaps between the electrodes of the associated stage, and the radial width of the respective annular electrodes being such that their exposed areas facing toward the middle region of said chamber are equal to each other.

16. An isotope separator comprising a closed cylindrical chamber having a source of ionized vapor positioned in its central region, cylindrical electrodes surrounding said central region within said chamber and provided with means for fixing their electrical potential relative to said source, means for producing a magnetic field parallel to the central axis of said cylindrical chamber, barriers of insulating material subdividing the space outside said cylindrical electrodes into separate cells.

17. An electrical discharge device comprising a closed container having its entire interior surface subtended by conducting sheaths, means for producing an electric arc in the central region of said container, said sheaths being subdivided into electrically insulated areas radially displaced from each other, and means for impressing an electrical potential on those areas at less than a predetermined radial distance of such magnitude and sign that only positive ions deposit thereon, the electrode areas outside said predetermined radial distance being electrically insulated from said-arc.

18. An electrical discharge device comprising a closed container having its entire interior surface subtended by conducting sheaths, means for producing an electric arc in the central region of said container, said sheaths being subdivided into electrically insulated areas radially displaced from each'other, at least a portion of said sheaths being impressed with a potential sutficiently positive relative to said arc so that a net current of negative sign flows to them from said arc.

References Cited in the file of this patent UNITED STATES PATENTS Great Britain Oct. 9, 1939

Claims (1)

1. AN ELECTRICAL DISCHARGE DEVICE COMPRISING A CLOSED CONTAINER HAVING AT LEAST A PORTION OF ITS WALLS OF CONDUCTIVE MATERIAL, A PAIR OF ELECTRODES CENTRALLY LOCATED THEREIN, SUPPLY CIRCUIT CONNECTED TO PRODUCE AN ARC DISCHARGE BETWEEN SAID ELECTRODES, MEANS FOR INSULATING A CONDUCTIVE PORTION ON THE CONTAINER WALL AND PERMITTING IT TO FLOAT ELECTRICALLY, AND MEANS FOR IMPRESSING A NEGATIVE VOLTAGE RELATIVE TO ONE OF SAID PAIR OF ELECTRODES ON THE REMAINING CONDUCTIVE WALL PORTIONS OF SAID CONTAINER.

Images