US3384772A

US3384772A - Method and apparatus for controlling breadown voltage in vacuum

Info

Publication number: US3384772A
Application number: US501782A
Authority: US
Inventors: Rabinowitz Mario
Original assignee: Individual
Current assignee: Individual
Priority date: 1965-10-22
Filing date: 1965-10-22
Publication date: 1968-05-21
Anticipated expiration: 1985-05-21

Description

May 21, 1968 M. RABINOWITZ 3,38 7

METHOD AND APPARATUS FOR CONTROLLING BREAKDOWN VOLTAGE IN VACUUM Filed Oct. 22, 1965 2 Sheets-Sheet 1 FIG.5.

INVENTOR FIG.|. FIG.2. FIG.3.

May 21, 1968 Filed Oct. 22, 1965 M RABI NO WITZ METHOD AND APPARATUS FOR CONTROLLING BREAKDOWN VOLTAGE IN YACUUM 2 Sheets-Sheet :v

| NVENTOR United States Patent 3,384,772 METHGD AND APIARATUS FOR CONTROLLING BREAKDQWN VOLTAGE 1N VACUUM Mario Rabinowitz, 403 Kingston Drive, Wilkins Township, Pittsburgh, Pa. 15235 Filed Oct. 22, 1965, Ser. No. 501,782 21 Claims. (Cl. 313147) ABSTRACT OF THE DISCLOSURE Apparatus and a method for controlling the breakdown voltage between a pair of spaced, electrically charged electrodes disposed within an evacuated housing are provided. The electrode faces are separated by a fixed distance and at least one of the electrodes is shiftably mounted on the housing so as to move relative to the other electrode without changing their separatiomMagnetic means is provided to establish a magnetic field of a given magnitude and direction in the space between the electrodes. The breakdown voltage is affected by changing the relative positions of the electrode faces since this changes the local heating of the electrodes as well as gives rise to inertial forces on particles as they leave the moving electrode faces. Also the breakdown voltage is affected by the magnitude and direction of the magnetic field of the magnetic means.

This invention relates to methods of increasing or decreasing the breakdown voltage of electrodes in vacuum without changing their separation, and thus in particular by increasing the breakdown voltage without the usual decrease in electric field strength that has been established between them.

The phenomena of electrical breakdown in vacuum is still not well understood even though it was observed as early as 1897 by R. W. Wood, and has been the subject of many investigations since. The reason for this may in part be due to the complexity of breakdown mechanisms which may all be at work in a given situation with one or more mechanisms dominating. In order to understand the way in which the breakdown voltage may be controlled, i.e., increased or decreased, it is desirable to make a brief review of the various proposed vacuum breakdown mechanisms, and to define the term electrical breakdown in vacuum.

By increasing the potential between electrodes separated by vacuum, the current between the electrodes steadily rises until a voltage is reached at which the current suddenly increases by orders of magnitude. A correspondingly rapid decrease in the voltage between the electrodes also occurs. The maximum voltage just prior to the voltage drop is called the breakdown voltage. The steep rise in current is accompanied by a catastrophic spark. As used here, the terms electrical breakdown in vacuum, vacuum breakdown, or more briefly breakdown, refer to these phenomena. As used here, the term vacuum denotes pressure less than 10- torr, and preferably less than l0 torr; or more generally that the mean path between charged particle and neutral gas particle collisions is greater than the electrode separation.

In 1952, L. Cranberg suggested that breakdown is initiated when a charged clump of material is removed from one electrode surface under the influence of the electric field, strikes the opposite electrode, and thus causes suffciently high temperature to produce local evaporation. This led him to conclude that for plane parallel electrodes the breakdown voltage, V, and the electrode gap length, d are related thus: V=Kd where K is a constant characteristic of the electrodes. In 1957, I. N. Slivkov modified this hypothesis slightly, and derived the relation: V=.Kd for plane-parallel electrodes.

The next hypothesis postulates that evaporation of the anode is produced by bombardment of an electron beam issuing from the cathode, which leads to breakdown in the anode vapor. A number of people are associated with this hypothesis, some of whom are L. B. Snoddy (1931), I. W. Beams (1933), and J. A. Chiles (1937). More recently A. Maitland in 1961 derived an equation from this hypothesis of the form V Kd. He gives an equation for a in terms of the electric field and the gap, d.

In 1936, A. J. Ahearn proposed the following model of local heating of the cathode to explain initiation of breakdown. Before breakdown, the field emission currents come primarily from a few of the microscopic cathode surface projections, where the local field is intensified by their presence. In addition the force on the cathode surface due to the electric field is greatest at these projections. There is local resistive heating, which depends on the size and geometry of these protrusions, as well as their thermal contact with the body of the cathode. As the electric field is increased sufiiciently, a rupture occurs at the protrusion where conditions of electric force, resistive heating, and tensile strength are most favorable. This rupture then can lead to breakdown. Modified versions of this model have also been proposed.

The basic assumption of the next hypothesis is that at a critical voltage a free, charged particle upon striking an electrode produces an avalanche of charged particles by secondary emission; with photo emission also playing a role. For example electrons striking the anode release positive ions which in turn release electrons upon striking the cathode, etc. In 1947 J. G. Trump and R. J. Van de Graatf proposed the positive ion avalanche hypothesis to account for the inititiation of vacuum breakdown. In 1948, J. L. McKibben and R. K. Beauchamp proposed the positive ion-negative ion avalanche hypothesis.

In 1963, W. D. Owen and F. Llewellyn-J ones proposed the hypothesis that vacuum breakdown results from an ordinary Townsend discharge in the gas liberated between the electrodes from the gas adsorbed on the electrode surface.

As you can see there are a multiplicity of hypotheses to account for vacuum breakdown, each of which is based on a difierent model of the initiating mechanism. Though each hypothesis may be partly correct, my experimental investigation has shown that no hypothesis can account for all the experimental facts. This has led me to introduce a new hypothesis which is capable of predicting at least qualitatively many of the experimental results not predicted by previous hypotheses.

My hypothesis provides a simple concept that can predict at least qualitatively the known experimental results. No particular model in terms of processes that are assumed to occur is associated with this hypothesis. Rather, it is concerned with the energetics of the problem. The initiation of breakdown and gap conduction can be extremely rapid. The voltage drop can occur in the order of nanoseconds which is so fast compared to the time constants of most breakdown circuits that only the capacitively stored energy of the electrodes and supports discharges within this time. After breakdown has occurred, most of the energy of the power supply is dissipated elsewhere in the circuit because the arc voltage is so low compared to the voltage drops across the other circuit components.

Hence it is reasonable to assume that the energy available to initiate electrical breakdown in vacuum is equal to some traction of the capacitively stored energy of the electrodes. If the available energy is less than the needed energy, breakdown cannot occur. It is the object of this invention to control breakdown by controlling the efficiency of the breakdown process, or the fraction of the energy available to initiate breakdown. As described in my paper in Vacuum, 15 (1965), 59, for the plane parallel electrodes this hypothesis leads to the equation:

where V is the breakdown voltage; W is the energy needed to initiate breakdown; s is the permittivity of vacuum; A is the area of an electrode; d is the distance between the electrodes or gap length; and f is the fraction of the capacitively stored energy avail-abie to initiate breakdown. From this equation we see that by changing f we can change the breakdown voltage, either increasing or decreasing it.

For a better understanding of my invention, reference may be had to the following drawings taken in conjunction with the accompanying descriptions wherein:

FIG. 1 is a fragmentary perspective view of two electrodes rotating in the same direction.

FIG. 2 is a fragmentary view in axial cross-section of two rotating electrodes showing the configuration of the magnetic field which is produced by their rotation.

FIG. 3 is a fragmentary sectional view of two rotating electrodes wherein the two axes of rotation do not coincide.

FIG. 4 is a fragmentary perspective View of two electrodes rotating in opposite directions about the same axis showing particles leaving the electrodes due to inertial forces.

FIG. 5 is a view in axial cross-section of two electrodes in a vacuum chamber showing a liquid metal seal of the electrode supports at the chamber walls.

FIG. 6 is a fragmentary cross sectional enlarged View of an electrode and support, showing in detail a liquid metal seal.

FIG. 7 is a fragmentary view in axial cross-section showing another method for rotating the electrode by means of a magnetic coupling of the shaft inside the chamber with a magnetic drive outside the chamber.

FIG. 8 is a sectional view of two electrodes in a vacuum chamber illustrating a reciprocating bellows action for oscillational motion of the electrodes.

Whereas particular embodiments of my invention are shown and described, it is to be understood that various changes, modifications, and alternative construct-ions may be employed without departing from the true spirit and scope of my invention.

An important point to bear in mind is that if the breakdown voltage of given conditioned electrodes inside a vacuum chamber is increased by increasing their gap separation, the electric field between them is decreased since the breakdown voltage varies as the gap distance by a power less than 1, V=Kd 0 a 1; and the electric field is, E=V/d. My invention allows one to increase or decrease the breakdown voltage without changing the electrode gap. Hence for a given gap :a larger electric field could be produced than could otherwise be produced and maintained.

The previously described hypotheses suggest that if there were a relative motion between anode and cathode, the breakdown voltage would be increased. This relative motion and the ensuing effects are the essence of this invention. For example, local heating of the anode by an electron beam coming from the cathode would thus occur at continually changing fresh surface :areas of the anode rather than be confined to a stationary spot. This would result in less heating at any given spot. The relative motion, without changing the electrode separation could be achieved in a number of ways. For example the electrodes could be rotated in opposite directions around a common axis. Or they could be rotated in the same direction, with their axes displaced as in FIG. 3. Even rotating only one electrode or the other would be etfective in this ease. Relative motion could also he achieved by a re ciprocal sideways oscillation of one or both of the electrodes; or by other means. In this example, the electron beam does not focus on the same electrode area, and similarly neutral and charged particles released from the anode do not move back toward the electron emission site on the cathode responsible for their release thus decreasing mutual interaction between active anode and cathode sites leading to an increase in breakdown voltage.

Relative motion as described in the preceding paragraph can be effective even when mechanisms other than electron beam heating of the anode are important. The relative motion of the electrodes gives rise to inertial forces on particles as they leave the moving electrode surface. This tends to either throw them out of the electrode gap and miss the opposite electrode or to strike the opposite electrode at more of a glancing angle and hence be less effective than otherwise. That particle inertia effects can occur and are present to a significant degree is shown in the photographs of FIGS. 10, 11 and 12 of my paper in Vacuum, 15 (1965), 59.

Since electrical discharges in vacuum are essentially supported by vaporized metal vapor atoms and other particles removed from the electrode surfaces, any radial motion of the electrodes perpendicular to the electrode axes will be imparted to these particles as they leave the electrode surfaces. For example, if a particle leaves a moving electrode with a radial velocity 1 its trajectory in a uniform axial electric field E will be parabolic. if the particle leaves an electrode rotating with angular velocity, 0:, at a radial distance r, v=wr. The particle will be radially displaced a distance,

21nd 1/2 i: a

where m is the particle mass, q is the charge on the particle, d is the gap separation, and E is the electric field. This may cause it to either miss the opposite electrode, or to hit it at a glancing angle. When the particle hits the opposite electrode its momentum vector makes an angle with respect to the electrode axis, making it less effective in knocking out secondary particles, in sending them in the direction of the electrode axis, and in returning them to the particles original site. V is the voltage between the electrodes. If the particle is not charged, it will move radially out of the electrode gap, neglecting polarization effects.

In addition to the obvious effects of changing areas which face each other, an inertial forces on particles as they leave the electrode surfaces, there is also the more subtle effect that a magnetic field is generated by the charge distribution on the moving electrode, which further complicates the situation. This method of producing a magnetic field between the electrodes circumvents the necessity of either using electromagnet coils near the electrodes with the associated current carrying wires that must be brought in, or of using permanent magnets near the electrodes. If the electrodes are to be baked out at high temperatures both the electromagnet coils and the permanent magnets placed just behind the electrode surfaces would present limitations which the production of the magnetic field by motion of the charged electrode would not be vulnerable to. There may well be additional reasons where a magnetic field is desired between the charged electrodes and where one can not use permanent magnets or electromagnet coils. In those situations where conditions permit, permanent magnets and/or electromagnets may be used next to the electrode faces to increase or change the magnetic field between the electrodes.

Since the electrodes are oppositely charged, rotating them in opposite directions causes the fields to add together and produce an almost uniform magnetic field, While rotating them in the same direction will produce an almost radial magnetic field. For example, consider a plane circular disk electrode of uniform surface charge density, rotating with angular velocity, to. In cylindrical coordinates, p, z, and 15, the components of the magnetic flux density, B, are given by:

w is the angular velocity of the disk in radians/sec;

R is the radius of the disk in meters;

0 is the surface charge density in coulombs/meter ,u is approximately the permeability of the electrode material at regions near the electrode, otherwise it is the permeability of free space in henrys/meter; and p and z are measured in meters. The origin is at the center of the disk and the disk is perpendicular to the z-axis. A and G are complete elliptic integrals of the first and second kind:

From these equations for B we see that when two such electrodes of opposite charge, separated by a distance d, rotate in the same direction, the axial fields l3 almost cancel, while the radial fields B, add together. When electrodes of opposite charge rotate in opposite directions the axial field B add, while the radial fields B, almost cancel.

So for electrodes with an electric field, E, between them and rotating in the same direction about the same axis, B is approximately zero and B, is proportional to Ewe p. where E is measured in volts/ meter, and s is the permittivity of free space in coulomb /newton-meter In this case the electric and magnetic fields are essentially perpendicular to each other which can lead to an increase in breakdown voltage.

For electrodes with an electric field, E, between them and rotating in opposite directions about the same axis, B is approximately zero and B is proportional to Ewe In this case the electric and magnetic fields are essentially parallel or antiparallel to each other which can lead to a decrease in breakdown voltage as in a Penning discharge. However, by using permanent magnets, and/or electromagnets, just behind the electrode faces, with like poles opposite, the magnetic field between the electrodes can be made essentially radial even when the electrodes rotate in opposite directions leading to an increase in breakdown voltage. Conversely if the permanent magnets or electromagnets are placed just behind the electrode faces, with unlike poles opposite, the resulting magnetic field will be increased axially.

The equation of motion of a charged particle in crossed electric and magnetic fields is given in vector notation by d A A m= =qE+qvAB The motion is complicated and depends on the initial velocity. Qualitatively one can say that when the fields are perpendicular the charged particle moves in a cycloidal like .path with net displacement perpendicular to E and B, and when they are parallel the path is helical with net displacement parallel to 1 5 and The effect of the magnetic field can usually be neglected when the magnetic flux density is small. This needs to be determined in each situation by the angular or linear velocity of the electrodes and the other variables which give the magnetic flux density in relation to the electric field. The magnetic field in the gap region can be increased by using electrodes of high permeability material such as iron; or by placing a slug of such material immediately behind each electrode face or by placing permanent magnets or electromagnets immediately behind each electrode face.

The relative motion of high voltage electrodes in vacuum can be beneficial to a great variety of devices. The following list is illustrative, but does not represent all such devices: high energy particle separators used in connection with accelerators; low loss high frequency vacuum capacitors; high intensity pulsed over-voltage X-ray tubes; vacuum electronic tubes; high energy particle accelerators; photo-multiplier tubes; microwave tubes such as klystrons; controlled nuclear fusion devices; ion propulsion engines; electrostatic voltmeters; outer space electrostatic shielding against high energy charged particles; and vacuum switches.

For example in the case of vacuum switches, the relative motion of the electrodes can be beneficial during arcing as well as to withstand a higher voltage following arcing. It is important in vacuum switches to avoid undue electrode erosion. In accordance with this invention this objective is accomplished by the relative motion of the electrodes. This motion continually presents fresh electrode surfaces to the arc and helps to prevent pocking erosion at any one point on the electrodes. In addition the tangential velocity imparted to electrode particles with ensuing inertial effects as they enter the arc helps to improve the arcing characteristics and the extinguish the arc.

The novel features which I believe to be characteristic of my present invention are set forth with particularity in the appended claims. My invention itself, however, both as to its organization and method of operation, together with further objects and advantages thereof, will be better understood by reference to the following descriptions taken in connection with the accompanying drawings which represent particular embodiments of my invention.

Referring now to FIG. 1, there are shown two axially spaced coaxial electrodes 1a and 1c embedded in a vacuum chamber or the vacuum of outer space. The electrodes 1a and 1c rotate with angular velocities ca and w respectively about their common axis 2 shown as a dotdash line. There is a slight depression 3c at the center of electrode 1c, and a similar depression 3a at the center of electrode 1a to avoid axial effects, i.e., to insure that a potential breakdown site will occur otf-axis. The electrodes 1a and 10 as shown are of cylindrical conformation about axis 2; however they may also have other shapes. As shown, the electrodes are rotating in the same direction. This will bring into being inertial effects, as well as a radial magnetic field as shown in FIG. 2. However this invention may be practiced by the rotation of only one of the two electrodes, or by the rotation of both electrodes in opposite directions as will be described in connection with FIGS. 3 and 4.-The electrodes are supported, respectively, by the conducting support arms 4a and 4c.

As shown in the axial cross-sectional view of FIG. 2, two electrodes 1a and 1c, in vacuum, are rotating in the same direction around their common axis 2. There is a slight depression 3a and Scat the center of each electrode to insure that breakdown will not occur on axis. The electrodes are supported by the conducting supporting arms 4a and 4c to which the power source is connected. Shown embedded immediately behind the face of each electrode 1a and 1c are

permanent magnets

5a and 5c to enhance the radial magnetic field. Of course electromagnets could be used in place of the permanent magnets. As shown, these magnets 5a, So have like poles facing opposite each other to produce a radial magnetic field. If an axial magnetic field is desired then unlike poles would be put facing each other. In situations where neither permanent magnets nor electromagnets could be used, a radial magnetic field would still be produced by the rotation in the same direction of the two charged electrodes 1a, 10. This magnetic field could be enhanced by placing inserts of high permeability material such as soft iron or silicon .eel immediately behind the face of each electrode in, 1c in the position shown occupied by the permanent magnets Sa, 50. Alternatively, the electrodes la and la themselves can be made of a high permeability material. The resultant magnetic field distribution is indicated schematically by the three representative magnetic flux lines shown as dashed lines. There are of course many such flux lines making up the overall field.

As shown in FIG. 2 the magnetic field fringes radially outward. In a preferred form of this invention, the faces of the electrodes and 1c are slightly convex with

depressions

3a and 3c at the center so that they are in accordance with the shape of the magnetic flux lines and substantially parallel to adjacent fiux lines. Since the electric field lines are perpendicular to the electrode faces, this helps to make the electric field lines and magnetic field lines mutually perpendicular and assures maximum interaction of a charged particle with the fields. However the electrode faces may also be plane or any other shape. In addition to the magnetic field action, both neutral and charged particles acquire a radial velocity as they leave the electrode face due to the rotation of the electrodes. The benefits from these effects leading to an increased breakdown voltage have previously been described.

FIGURE 3 shows a sectional view of two electrodes 1a and 1c, in vacuum, rotating in opposite directions about their respective axes 2a and 2c which are displaced radially from each other. This displacement helps to minimize the effects of the magnetic field that is produced. The electrodes are supported, respectively, by the conducting support arms 4a and 46. Since the angular velocities w, and w are in opposite directions, both the inertial effect and the effect of continuously changing electrode areas are brought into play. It should be noted that even if the electrodes rotate in the same direction about their respective axes, given locations on one electrode, face opposite continuously changing locations on the opposite electrode because the axes of rotation do not coincide.

In FIG. 4, the electrodes 1a and 1c, in vacuum, are rotating in opposite directions about their common axis 2. There are

depressions

3a and 30, respectively at the center of each electrode 1a and 1c to insure that potential breakdown sites will occur ofit axis. The electrodes in and 1c are supported respectively by the conducting support arms 4a and 4c. Two particles 6a and 6e are shown respectively leaving the electrodes 1a and 10 due to inertial forces arising from the motion of the electrodes imparted to the particles as they leave the electrode surfaces. The rotation of the electrodes in opposie directions causes given locations on one electrode to face opposite continuously changing locations on the opposite electrode, leading to an increase in breakdown voltage. Even though the rotation of the electrodes in opposite directions tends to produce a magnetic field in the axial direction parallel to the electric field which would decrease the breakdown voltage, by proper choice of geometry (electrodes radius, depression size, spacing, etc.), charge density, angular velocity, and other factors as previously discussed this magnetic field can be made very small so that its effects are negligible and the breakdown voltage increases due to changing area and inertial effects. And as previously described, permanent magnets and/ or electromagnets can be placed behind the electrode faces to make the magnetic field either essentially radial or essentially axial independent of the relative motionof the electrodes. A radial magnetic field would help to further increase the breakdown voltage.

Referring now to FIG. 5, there are shown two axially spaced coaxial electrodes 1a and 1c embedded in at hermetically sealed vacuum chamber. As shown, electrodes 1a and 1c rotate in opposite directions with angular velocities w, and w about their common axis 2. There are depressions at 3a and to insure that breakdown cannot occur near the axis 2. The electrodes in and 1c are supported, respectively, by the conducting support arms, 4:: and 4c. Wound respectively around the hollow supporting arms 4a and 4c, and just behind the electrodes 1a and 1c are two electromagnet coils 7a and 7c which serve the same purpose as the permanent magnets of FIG 2, but which add versatility to the system. When the fields of the two electromagnet's produce like poles opposite each other, an essentially radial magnetic field is produced between the electrodes as shown in FIG. 2, which helps to increase the breakdown voltage. When the fields of the two electromagnets produce unlike poles opposite each other, an essentially axial magnetic field is produced between the electrodes which tends to decrease the breakdown voltage. The effects of the electromagnets 7a, 7c are independent of the motion of the electrodes 1a, 10, i.e., these effects are present whether or not the electrodes 1a and 1c are in motion. The effects of the fields produced by the electromagnets 7a, 7e superimpose on the motional effects of the electrodes la, lie when both are present. The current to the

electromagnets

5a and 5c is brought in by the leads 8a and through the hollow support arms 4a and 40 so as not to interfere with the rotation of the electrodes 1a, 10. This may also be accomplished by the use of a commutator or slip ring arrangement.

As shown in FIG. 5, the electromagnet 7a and 7c are respectively powered by the power sources 9a and 90. As shown these power sources 9a and 9c are direct current, but for some purposes they may be alternating current. The power source 9a is turned on and off by the simple switch 10a. The power source is turned on and off by the double pole double throw switch Title, which also permits changing the polarity of the power source 9c and hence the polarity of the magnetic field produced by the electromagnet 70. This makes it easily possible to change from a radial magnetic field to an axial magnetic field or vice versa permitting one to readily decrease or increase the breakdown voltage. Of course the combination of one electromagnet and one permanent magnet behind each electrode in and 1c can also serve this purpose. However this does not allow for as much flexibility as when two electromagnets 7a, 70 are used, wherein both the magnitude and the direction of the applied magnetic field between the electrodes 1a and 10 may be easily changed by varying the power supplies 9a and 9c accordingly. If the switches 10a and are left open, there will be no applied magnetic fields, and the effects are due to the motion of the electrodes 1a and 10. Alternatively, the electrodes may be stationary with the effects entirely due to the applied magnetic fields. Two electric lines L and L make electrical connection by slip rings or similar method to the supporting arms 4a and 4c to provide high voltage to the electrodes 1a and 10.

Also shown in FIG. 5, are the other components of the hermetically sealed vacuum chamber, including the dielectric insulating wall 11, with sealed on end flanges 12a and 120. A liquid metal seal 13a and 13c in made with each supporting arm 4a and 4c at the flanges 12a and 120. This seal, which is described in greater detail in connection with FIG. 6, permits both rotation and axial motion of the support arms 4a and 40 to which the electrodes 1a and 1c are attached. Such seals would not be necessary if the motion producing devices were inside the vacuum chamber. A vacuum chamber with seals for rotation and axial motion would not be needed if the electrodes la, 10 were in the vacuum of outer space.

FIGURE 6 shows an enlarged view in detail of the liquid metal seal mentioned in connection with FIG. 5. There is shown the electrode 1, rotating about the axis 2, supported by the supporting arm 4. A liquid metal 20 provides the seal between the support arm 4, and the flange wall 12. Preferably the liquid metal is one of low vapor pressure, low melting point, wets the metal surfaces at the seal, and is non-corrosive. A liquid metal alloy made of 62.5% gallium, 21.5% indium, and 16% tin with melting point of about 10 C. and vapor pressure less than 9 10- torr up to 500 C. can be used for this purpose even though it is corrosive at elevated temperatures. To prevent corrosion, the supporting arm 4 may have a shaft clad 21, and the wall may have an insert 22 that are resistant to corrosion by the liquid metal 20. In the case of the gallium-indium-tin alloy just described, the shaft clad 21 and the insert 22 may be made of tungsten or tantalum for high temperature use, or stainless steel where the main use is at room temperature. There is approximately a 0.005 inch gap all around for the liquid metal between the shaft clad 21 and the hole in the insert 22, On the atmospheric side of the flange 12, a forepump vacuum of about 10 torr is maintained to reduce the pressure differential on the liquid metal 20. Inside the enclosure 23 the pressure is 10 torr. This enclosure 23 may be sealed with rubber O-rings 24 as shown. Inside the enclosure 23 are ball bearing races 25 to hold the supporting arm 4 in alignment to permit it to rotate freely.

FIGURE 7 shows an alternate means for providing rotation of the electrodes. There is shown the electrode 1 rotating about the axis 2, and supported by the supporting arm 4. A permanent magnet 3% is attached to the end of the support arm, perpendicular to axis 2 and parallel to the flange. The support arm 4 and the magnet are held by the sleeve 31. Ball bearing races 32 hold the support arm 4 in alignment and permit it to rotate freely. An electric line L is connected to the outside of the sleeve 31 to provide voltage to the electrode 1. An external magnet 33 couples its rotational motion to the internal magnet 30. The same effect can be achieved by several fixed electromagnets with oscillating fields as is done by the field coils in some motors.

FIGURE 8 shows a way in which linear oscillational motion may be imparted to two electrodes 41a and 410 in a hermetically sealed vacuum chamber 40. The electrodes 41a and 41c oscillate back and forth in the directions respectively indicated by the arrow tipped lines 42a and 420 in the plane of the paper, which motion is permitted by the reciprocating members 43a and 430 shown as bellows. The electrodes 41a and 410 are supported respectively by the conducting support arms 44a and 440 which are attached to the bellows 43a and 43c. The bellows 43a, 43c are attached to the end flanges 45a and 450. The bellows 43:1, 430 allow for a bending motion in addition to the linear motion, if desired. The

end flanges

45a, 45c are separated by an insulating wall 46. Electric lines L and L are attached to the support arms 44a and 440 to provide voltage to the electrodes 41a and 410.

In FIG. 8 the linear motion of the electrodes 41a, 410 in opposite directions relative to each other causes given locations on one electrode to face opposite continuously changing locations on the opposite electrode. The motion of the electrodes 41a, 41c, also imparts a velocity perpendicular to the electric field as particles leave the electrodes which results in the previously discussed inertial effects. The linear motion of the charged electrodes 41a, 410 also generates a magnetic field between the electrodes which can be made negligible by proper choice of parameters similar to the rotational case. As before, a magnetic field may be applied, if desired, by placing electromagnets or permanent magnets immediately behind the electrode faces 41a and 41c. As before, when the like poles are placed opposite each other an essentially radial magnetic field perpendicular to the electric field is produced between the elcctrodes 41a, 410 which acts to increase the breakdown voltage. When opposite poles are placed opposite cach other, a magnetic field is produced between the electrodes 41a, 41c essentially parallel to the electric field which acts to decrease the breakdown voltage. Another mode of oscillation is in which the two electrodes 41a, 41c oscillate in phase so they move together back and forth in the same direction at the same time. In this case, the changing area effect is eliminated as the same locations would continue to face opposite each other on the two electrodes 41a, 410. However the inertial and magnetic effects may still be present.

While I have shown and described specific forms of the present invention it will of course be understood that many modifications'and alternative constructions can be made without departing from its spirit and scope. I, therefore, intend by the appended claims to cover all such modifications and alternative constructions as fall within their true spirit and scope.

What I claim as new and desire to protect by Letters Patent of the United States is:

l1. An electronic device comprising: an evacuated housing; a pair of electrodes; means mounting said electrodes at fixed, spaced apart locations within said housing and including structure permitting at least one of the electrodes to shift along a preselected path relative to the housing as the spacing between said electrodes is held constant, whereby the electrodes are movable relative to each other; means coupled with the electrodes for establishing an electrical field therebetween, having a predetermined value whereby the electrodes will be oppositely charged; and means coupled with said one electrode for shifting the same relative to the other electrode along said preselected path at a rate sutficient to control initiation of electrical discharge between said electrodes when said electric field is maintained at said value.

2. An electronic device as set forth in claim 1, wherein said one electrode is rotatably mounted on said housing.

3. An electronic device as set forth in claim 18, wherein said one electrode is mounted for rectilinear movement on said housing.

4. An electronic device as set forth in claim 1, wherein said housing has a pair of opposed walls, each electrode having a support arm rotatably mounted on a respective wall and held against axial movement, said electrodes being secured to the inner ends of respective arms, said shifting means being coupl d to said arms for rotating the same.

5. An electronic device as set forth in claim 4, wherein said shifting means includes structure for rotating said arms in the same direction.

6. An electronic device as set forth in claim 4, wherein said shifting means includes structure for rotating the arms in opposite directions.

7. An electronic device as set forth in claim 4, wherein said arms are in axial alignment, said electrodes having opposed faces, each face having a central depression, the depressions being in alignment with each other.

8. An electronic device as set forth in claim 4, wherein said electrodes have opposed faces, said arms being parallel and axially offset to thereby ofiset said opposed faces from each other.

9. An electronic device as set forth in claim 4, wherein is provided means coupled With said electrodes for establishing a magnetic field in the space therebetween.

10. An electronic device as set forth in claim 9, wherein said establishing means includes a permanent magnet for each electrode respectively, the magnets being within said housing and adjacent to respective electrodes.

11. An electronic device as set forth in claim 9, wherein said establishing means includes an electromagnet for each electrode respectively, the electromagnets being within said housing and adjacent to respective electrodes, and electrical power means coupled with each electromagnet for actuating the same.

12. An electronic device as set forth in claim 11, wherein said power means for one of said electromagnets includes a circuit having a current reversing switch, and means for varying the electrical power supplied to said electromagnet.

13. An electronic device comprising an evacuated housing; a pair of electrodes, each electrode having a face; means mounting said electrodes within said housing with said faces being separated by a predetermined distance and held against movement toward and away from each other; means coupled with said electrodes for establishing an electric field therebetween having a predetermined value, whereby the electrodes are oppositely charged; means coupled with said electrodes for establishing a magnetic field therebetween having a magnitude and direction to control the initiation of electrical discharge between the electrodes as said electric field is maintained at said predetermined value.

14. An electronic device as set forth in claim 13, wherein said establishing means includes an electromagnet for each electrode respectively, the electromagnets being within said housing and adjacent to respective electrodes, and an electrical power source for each electromagnet respectively, one of said power sources including a circuit having a reversing switch and means for varying the electrical power to the respective electromagnet.

15. A method of controlling the breakdown voltage between a pair of electrodes spaced a fixed distance apart in a vacuum comprising the steps of: coupling said electrodes to a source of electrical power to oppositely charge the electrodes and thereby to establish an electrical field therebetween having a predetermined value; and moving a path and at a rate sutficient to control the initiation of a path and at a rate sufficient to control the iniiaion of electrical discharge between the electrodes .as said electrodes are maintained separated by said fixed distance and as said electric field is held at said predetermined value.

1 6. A method as set forth in claim 15, wherein said moving step includes rotating the electrodes in the same direction.

17. A method as set forth in claim 15, wherein said moving step includes rotating the electrodes in opposite directions.

18. A method as set forth in claim 15, wherein said moving step includes reciprocating said electrodes along respective, rectilinear paths.

19. A method as set forth in claim 15, wherein is included the step of establishing a magnetic field between the electrodes.

20. A method of controlling the breakdown voltage between a pair of electrodes spaced a fixed distance apart in a vacuum comprising the steps of: coupling said electrodes to a source of electrical power to oppositely charge the electrodes and thereby to establish an electrical field therebetween having a predetermined value; generating and maintaining -a magnetic field in the space between the electrodes with the field having a magnitude and direction sufficient to control the initiation of electrical discharge between the electrodes as said electrodes are maintained separated by said fixed distance and as said electric field is held at said predetermined value.

21. A method as set forth in claim 20, wherein said magnetic field has a component for each electrode respectively, with each component having a preselected polarity, and wherein is included the steps of selectively reversing the polarity of one of said components, and changing the magnetic field intensity of a first of said components.

References Cited UNITED STATES PATENTS 660,613 10/1900 Baker 313-149 2,116,218 5/1938 Seil 313-149 2,562,031 7/1951 Gerber 313-149 2,712,097 6/1955 Auwarter 313- X 2,793,325 5/1957 Wenzel 313-149 X 2,831,994 4/1958 Schering 313-149 X 2,831,995 4/1958 Agule 313-149 X 2,985,788 5/1961 Opsahl 313-153 X 3,020,408 2/ 1962 Martin 313-61 X 3,082,307 3/ 1963 Greenwood 313-153 X 3,097,292 7/ 1963 Kugler et al. 219-121 3,320,462 5/1967 K-awiecki 313-154 DAVID J. GALVIN, Primary Examiner.

STANLEY D. SCHLOSSER, JAMES W. LAWRENCE,

Examiners. R. L. J UDD, Assistant Examiner.

UNITED STATES PATENT OFFICE CERTIFICATE OF CORRECTION Patent No. 3,384,772 May 21, 1968 Mario Rabinowitz It is hereby certified that error appears in the above numbered patent requiring correction and that the said Letters Patent should read as corrected below.

In the heading to the printed specification, lines 4 and 5, for "403 Kingston Drive, Wilkins Township, Pittsburgh, Pa. 15235" read Station A, P. O. Box 695, Menlo Park, Calif. 94025 column 3, line 4, the equation should appear as shown below instead of as in the patent:

column 4, line 49, for "an" read and column 5, line for "field" read fields line 60, the equation should appear as shown below instead of as in the patent:

column 6, line 27, for "the", second occurrence, read to column 7, line 50, for "opposie" read opposite line 57, for "electrodes" read electrode column 8, line 56, for "in" read is column 9, line. l8,cafter "alignment" insert and ;column 10, line 29, for the claim reference, numeral "18" read 1 column 11, line 22,

after "moving" insert one of the electrodes relative to the other electrode along Signed and sealed this 31st day of December 1968.

[SEAL] \ttest EDWARD M.FLETCHER,JR. EDWARD J. BRENNER \ttesting Officer Commissioner of Patents