US2591350A - Traveling-wave electron reaction device - Google Patents

Description

April 1, 1952 E. J. GORN TRAVELING WAVE ELECTRON REACTION DEVICE Filed April 26, 1947 5 Sheets-Sheet 1 F/QS April 1, 1952 E. J. GORN 2,591,350

TRAVELING WAVE ELECTRON REACTION DEVICE Filed April 26, 1947 5 Sheets-Sheet 2 F/Qd April 1, 1952 E. J. GORN 2,591,350

TRAVELING WAVE ELECTRON REACTION DEVICE Filed April 26, 1947 5 Sheets-Sheet 3 April 1, 1952 J GORN 2,591,350

TRAVELING WAVE ELECTRON REACTION DEVICE Filed April 26, 1947 5 Sheets-Sheet 4 A ril 1, 1952 E. J. GORN 2,591,350

TRAVELING WAVE ELECTRON REACTION DEVICE Filed April 26, 1947 5 Sheets-Sheet 5 6 7 83 6/ 8/ 0 o o a o a o a o as Q\Q wuwwuuw Patented Apr. 1, 1952 TRAVELING-WAVE ELECTRGN REACTION DEVICE Elmer J. Gorn, Newton, Mass., assignor to Raytheon Manufacturing Company, Newton, Mass.,

a corporation of Delaware Application April 26, 1947, Serial No. 744,143

Claims. (01. 315-39) This invention relates to an electron-discharge device in which electrons are caused to react with an electromagnetic wave in a waveguide so as to give up energy to said Wave or to abstract energy therefrom. In the first case the device could be made to operate as an amplifier, oscillator, or the like, while in the second case the device could act as an electron accelerator.

Heretofore, the velocity, at which the phase of a wave radiated through a waveguide progresses along a direction parallel with the longitudinal axis of the guide, has been termed the phase velocity of the guide for the particular wave under consideration. Such phase velocity, which will hereafter be characterized as the linear phase velocity, is, in the case of, a simple waveguide, in excess of the speed of light, which it approaches as a limit. In order for an electron to react in the manner described above, it is necessary to preserve the electron in a substantially fixed position relative to the phase of the wave throughout substantially the entire travel of the electron through the device, since otherwise the electron at one time would be imparting energy to the wave and at another time would be abstracting energy therefrom, giving a substantially zero net effect. It is impossible to cause the electron to travel at the linear phase velocity of a simple waveguide, since, as pointed out above, such velocity is in excess of the velocity of light.

An object of this invention is to cause an electron to travel along such a path in a waveguide as to" preserve a substantially constant phase relation with respect to a wave therein whose linear phase velocity exceeds the velocity of light.

Another object of this invention is to devise an arrangement in which such constant phase relation exists and which operates as an amplifier of microwaves.

A further object of this invention is to produce a device of this kind which operates as an electron accelerator.

A still further object is to devise an arrangement for causing an electron to travel in a predetermined non-linear path without the necessity of using large transverse magnetic fields.

The foregoing and other objects of this invention will be best understood from the following description of exemplifications thereof, reference being had to the accompanying drawings, wherem:

Figs. 1, 2 and 3 are diagrams of the cross-sections of various shapes of waveguides excited in various modes to illustrate certain aspects of the invention;

Fig. 4 is a cross-sectional diagram taken along line 4--4 of Fig. 3;

Fig. 5 is a cross-sectional diagram similar to Fig. 4 but illustrating another aspect of the invention;

Fig. 6 is a central vertical cross-section of a microwave amplifier representing one embodiment of this invention;

Fig. 7 is a perspective View, partly broken away, of the waveguide component of Fig. 6;

Fig. 8 is a diagram illustrating certain aspects of the operation of the device shown in Fig. 6;

Fig. 9 is a cross-sectional diagram of an electron accelerator embodying this invention;

Fig. 10 is a cross-sectional diagram of a modified electron accelerator;

Fig. 11 is a cross-sectional diagram of a further modified type of electron accelerator embodying this invention; and

Fig. 12 is a perspective diagram of the electron accelerator of Fig. 11;

If, instead of confining our consideration to the linear phase velocity of a Wave radiated through a waveguide, we consider certain angular pathsand consider the angular phase velocities involved, we will discover that such angular velocities are within the range at which an elec-- tron can travel.

It is desired to be explained at this point What is meant herein by the expression, the angular phase velocity of a wave in a waveguide. The angular phase velocity is the velocity at which the phase of an electromagnetic wave in a waveguide propagates along a non-linear path with respect to said waveguide, said non-linear path being so related to the longitudinal axis of said waveguide as to provide a substantially periodic angular component of motion with respect to said axis for a point moving alon said nonlinear path at a uniform rate.

Let us consider, first of all, a cylindrical waveguide excited in the TE1,1 mode as shown in Fig. 1. It will be seen that if we consider the direction and magnitude of the electric field around the inner circumference of the waveguide we obtain a sinusoidal variation in the field relative to the axis a-a. In other words, as we progress from the axis aa around the circumference of the waveguide the field increases in magnitude from zero to a maximum, back to zero, increases in the reverse direction to a; maxi-' mum and then falls back to zero. If we select an arbitrary point e in the wave pattern and if we move the point so as to preserve its relation to the wave pattern as referred to a reference point such as the center 0, the'velocity of the point would be the phase velocity of the wave. Thus, if the point e were moved along a direction at right angles to the plane of Fig. 1, the velocity would be the linear phase velocity which, as has been pointed out above, is equal to or greater than the velocity of light.

Let us now rotate the wave pattern of Fig. l in the direction of the dotted arrow at an angular velocity of w=21rf (1) where f is the frequency of the wave. This is the condition of circular polarization. If we now move the point e in the plane of Fig. 1 in the direction of the dotted arrow at an angular velocity equal to or we find that the relation between point e and the wave pattern is preserved when related to the reference point 0. In other words, thefield at point e always points away from the point e, for example. Thus, the angular phase velocity of the wave in the plane of Fig. 1 is equal to no.

If we now, in addition to its angular motion, impart to the point e a velocity 'Ue along the waveguide, we can derive a general expression for the angular phase velocity along any resultant path with reference to an axis passing through 0. At a time i=0, the relationship between the phase of'the angular motion of point e and the phase of the electric field of the wave will have some constant value. There is no loss of generality in assuming this initial phase difference to be zero. At a later time t and at a distance x along the waveguide from the initial point under consideration the phase w of the wave is and )\g= the wave length of the wave in the guide. At the time t, in order for the point e to be at the same point x,

The phase e of its angular motion is qbe wet (4) where we is the angular velocity of the point e.

In order to preserve a constant phase relation between .w and e these values are equated as follows:

The well-known expression for the linear phase velocity Up is w w U 01 n B up I Substituting this expression in Equation 8 and simplifying, we have Equation is a general expression which gives the angular velocity of a point traveling along a waveguide with any given linear velocity in terms of the angular velocity of the wave and the linear phase velocity of the wave, where phase equality is preserved.

Circular polarization of the wave was referred to above as an aid to a graphic visualization of the relationship involved. However, the generality of Equation 10 is not affected as the polarization becomes elliptical and reaches linear polarization. The generalization is also clearly applicable to other types of waveguides such as those of rectangular cross-section. The only special consideration that must be met is that the excitation mode which is set up in the waveguide must be such that a circular path within the wave pattern set up at any instant in the waveguide must include at least one complete alternation of the electric field as referred to any given axis or point of reference. Equation 10 represents the general condition where one such alternation of field is encountered.

As used herein, the term reference point, as used in connection with alternations of the electric field in a circular path within the wave pattern, is intended .to mean any point with respect to which an electron, considered as a negative charge, experiences a force due to the electric field present; the term a completealternation of the electric field is intended to mean the presence of a single pair of such forces, one of the pair pointing toward and the other of the pair pointing away from the particular reference point under consideration. i

Let us now consider a more complex wave pattern, as for example that represented by the TE1,1 wave in a square waveguide, as shown in Fig. 2. It is clear that the circular path represented by the dotted circle in Fig. 2 includes two complete alternations of the electric field in the field pattern set up by the TE1,1 wave when ref-erred to the point 0, since there are two pairs of forces present with respect to point 0, one force of each pair pointing toward point c and the other force of each pair pointing away from point 0. In other words, an electron in proceeding around the dotted circle in Fig. 2 experiences four forces with respect to point 0, two of which tend to force it toward point e and the other two of which tend to force it away from point 0. Since there are two complete alternations of the electric field in the field pattern of Fig. 2, referred to point 0, included in the dotted circle of said figure, the angular velocity of a point e in such case need be only one-half of the velocity in the case of a single field pattern alternation in order to maintain the constant phase relation with the wave. Thus, the angular phase velocity is inversely proportional to the number of complete field alternations in a given circularpath through an instantaneous wave pattern. Therefore, Equation 10 can be expressed in a still more generally applicable form i 1 .11) V w 1);, v

where p is the number of complete field alternations as described above. 7

If instead of selecting point e as the reference point in Fig. 2, some other point such as d were selected, only a single complete alternation of the electric field pattern would be included in the circular path referred to above, since with respect to point d there is only a single pair of oppositelydirected forces (represented by the substantially vertical field lines in Fig. 2) which would'tend to move an electron toward and away from point d. In this case, the substantially horizontal field lines are ineffective, since they would tend to move an electron only in a direction parallel to a horizontal axis passing through point d, and not toward or away from said point. Similarly in Fig. l, the circular path would not encompass a complete alternation of the field pattern with respect to the point at, since all the forces are directed away from said point and there is therfore no pair of forces one of which pair is directed toward point (1 and the other of which pair is directed away from point 03. With respect to point 0, there is a single pair of forces (represented by the substantially vertical field line in Fig. 1) one of which is directed toward point 0 and the other of which is directed away from point 0, so that with respect to this latter point the circular path includes a single complete alternation of the electric field through an instantaneous wave pattern.

Thus, in determining the value of p in Equation 11 the reference point or axis to which the wave pattern is referred must be kept in mind.

Although the examples cited above have involved transverse electric modes of excitation, the generality of Equation 11 also applies to longitudinal modes of excitation. Let us consider, for example, the E1,2 mode in a rectangular waveguide, as shown in Figs. 3 and 4. In this case the point e, in moving through the dotted circle, includes one complete alternation of the electric field with respect to the point 0, since with respect to this point there is a single pair of forces (represented by the substantially horizontal field lines in the upper and lower portions of Fig. 4) present, one of this pair being directed toward point e and the other of this pair being directed away from said point; the angular phase velocity for synchronism in this case is given directly by Equation 10, since is equal to 1. The helical motion of point e, resulting from the circular orbital motion thereof combined with the trans latory motion thereof along the waveguide at a velocity U8, is indicated by the dotted curve in Fig. 4.

If an electron were caused to follow the path of the point e in each of the above instances it would persist in the particular phase relation with the wave at which it is injected into the waveguide. If the phase is such that energy is abstracted from the motion of the electron by the electric field, amplification of the wave traveling along the waveguide may be expected, whereas if the phase relation is such that energy is imparted to the electron by the electric field, the electron will be accelerated. If such acceleration is along a path which permits the electron to travel a substantial distance without being captured by the waveguide boundaries, relatively high electron velocities may be attained before the electron is intercepted by the boundaries of the guide. In some cases the acceleration may vary in direction so as to impart non-linear motion to the electron involved.

Angular velocities of a fixed and predetermined value may be imparted to electrons by magnetic fields in accordance with the known expression tion of the electron in synchronism with the angular phase velocity of the wave.

If we assume 10: and 118:0 we obtain, in the case of 3.5 cm. wave, a value of B equal to approximately 3000 gauss. Equation 13 indicates that if v8 were increased, B could be decreased. However, for moderate values of We the decrease in B is rather small. Electrons injected into the waveguide with a voltage of about 3000 volts will permit about a ten percent reduction in the required field. The injection voltage needed to produce greater reductions in B rises very rapidly, especially since relativity considerations are involved.

From Equation 13 we see that if p is greater than 1, relatively large reductions in B might be obtained. Thus if p could be made 2, or 4, the field required would be reduced to one-half or one-quarter respectively, giving in the case of the 3.5 cm. waves fields of 1500 and 750 gauss respectively. However, if we are dealing with zero injection voltage for the electron, in considering any wave pattern to determine the number of field alternations present in the circular orbit of an electron we must not lose sight of the fact that at microwave frequencies, the diameter of the electron orbit is necessarily small, otherwise the linear velocities tangential to the direction of electron travel tend to approach the velocityof light. The increased mass resulting from such high speeds would in turn require higher magnetic fields to preserve phase synchronism. Thus in the case of the TE1,1 mode of excitation with a 3.5 cm. wave, if we limit the tangential velocity of the electrons to one-tenth of the speed of light, the diameter of its circular orbit is about one millimeter. If p is increased to 2, the diameter of the electron circular orbit increases to about two millimeters at the same tangential speed. As we go to longer wavelengths (or lower frequencies) the required magnetic field decreases, thus simplifying the problem. However, shorter wavelengths (higher frequencies) require still higher fields.

In the case of an electron accelerator, we are free to choose any convenient wavelength of propagated wave. Thus with a ten centimeter wave we should obtain the desired phase coincidence, at least in the front end of the waveguide, with a field of about 1,070 gauss. Here again, an increase in 22 might be useful.

In the above discussion of the magnitude of the magnetic field, we have been assuming that the velocity of the electron along the waveguide and the mass of the electron relative to its orbital motion remained constant. Such a condition might be attained in an amplifier excited with a TE mode and with the electron injected with a tangential velocity of about 3000 volts or less. In such an arrangement energy would be abstracted fromthe orbital motion of the electron causing it to travel in a helix of decreasing radius but of constant pitch. Thus none of the conditions determining the value of B would change either across or along the waveguide. If, however, we assume an amplifier excited in an E mode and with the electron injected a velocity along the waveguide of about 3000 volts or less, then energy will be abstracted from the linear component of the electron motion and the electron velocity along the waveguide will progressively decrease. This will require an increase of the magnetic field along the waveguide, as can be seen from Equation 13. In the example assumed, however, the maximum in crease from the beginning of the waveguide to the end will be 11.1 per cent, this maximum being required if we assume that the electron speed reaches zero immediately ahead of the end of the waveguide.

In an electron accelerator, we are interested in feeding large amounts of energy to the electrons and this therefore will produce increases in both 10 tial velocity of the electron exceeds about 3000 volts, the electron mass will increase. From Equation 13 we see that such increase in mass will require a corresponding increase in magnetic field. To the extent that the increase of the orbital electron mass increases the mass of the electron relative to its motion along the waveguide, the velocity of the electron along the waveguide will decrease since presumably no accelerating force along the waveguide is present to compensate for such increased mass. This may likewise tend to require an increase in the magnetic field along the waveguide. However, with small or substantially zero drift along the waveguide the variation in magnetic field along the waveguide may be neglected. Under such conditions, the

increase in magnetic field due to the increase in the orbital electron mass would have to be secured by a radial increase in the magnitude of the magnetic field. An advantage of such an arrangement would be the possibility of using a short air gap for the magnetic field structure, thus facilitating the production of the required magnitude of magnetic field.

If we consider an accelerator excited with an E mode such as shown in Figs. 3 and 4, the value of De in Equation 13 increases along the waveguide, which tends to reduce the value of B. However, as the speed of the electron exceeds that corresponding to about 3000 volts, its longitudinal mass increases, due to relativity effects.

We know that there is also some increase in the transverse mass of the electron, although to a lesser degree than in the longitudinal mass. The value of m in Equation 13 is the transverse mass in this case. mass tends to increase the value of B. No'attempt is made herein to compute the net result of these eifects. However, such computations could be made in order to determine what variations if any should be made in the magnetic field along the waveguide to preserve the desired phase synchronism. Since there is probably no feeding of energy into the angular momentum of the electron, it is not believed that its orbital radius will increase substantially with the increase in G transverse mass.

Although in the examples'above, the orbital or circular path of point e has been considered to lie in a plane perpendicular to the longitudinal axis or center line of the waveguide, the generality of 5 Equation 11 also applies to circular'paths which lie in other planes. Let us consider next the E1,1 mode in a rectangular waveguide, with the circular path of point e lying in a plane parallel to the longitudinal center line or length of the waveguide, as shown in Fig. 5. In this case the point e in moving through the dotted circle includes one complete alternation of the electric field with respect to the point 0, since with respect to this point there is a single pair of forces (represented Such increase in the transverse 50 by the substantially horizontal field lines in Fig. 5) present, one of this pair being directed toward point e and the other of this pair being directed away from said point; the angular phase velocity for synchronism in this case is given directly by Equation 10, since 10 is 1'.

To'recapitulate, the above discussion shows that it is possible to produce a motion of electrons in a waveguide in synchronism with the angular phase velocity of the wave being radiated therethrough, and also shows that when such synchronous relation exists, a net energy interchange can take place between the electrons and the electromagnetic wave in the waveguide.

Figs. 6 and 7 illustrate a practical microwave amplifier by means of which the above principles may be put into effect. Numeral I generally designates an electron-discharge device according to this invention. A pair of hollow waveguide,

sections 2 and 3 are joined and hermetically sealed together with the longitudinal axes of the two guides at right angles to each other (see Fig. 7). The interiors of these two guides are evacuated and are in communication with each other. and in order to provide a hermetically-sealed envelope portion for the electron-discharge device, a sealing member 4 is placed in the interior of the vertically-extending waveguide 2 a substantial distance above the junction of the two guides, or on the wave-input side of the device, while a sealing member 5 is placed in the interior of the horizontally-extending Waveguide 3 a substantial distance to one side of the junction of the two guides, or on the wave-output side of the device.

The lower end of waveguide 2 is at the junction between the two guides, while waveguide 3 extends on both sides of said junction. The interior of waveguide 3 is sealed ofi from the atmosphere at the side of said waveguide opposite from the wave-output side of the device, by means of a metallic end wall 3a, integral with said waveguide, which closes off the interior of the waveguide 3 and also terminates said waveguide at this side of the junction between the guides 2 and 3.

As shown in Fig. '7, waveguide section 3 has an H-shaped cross-section throughout its length, while waveguide section 2 has an upper portion 2a of H-shaped cross-section and a lower portion 2b, of rectangular cross-section in its midsection, but having transition sections above and below its mid-section. These transition sections connect the rectangular part of portion 2b to the upper H-shaped portion 2a, on the one hand, and to the H -shaped guide 3, on the other hand. Fortion 2a and the lower extremity of the lower transition section of waveguide 2 have exactly the same cross-sectional dimensions as does waveguide 3, while the rectangular mid-section of portion 2b has an increased cross-sectional area relative to portion 2a and waveguide 3. The lower extremity of the lower transition section of waveguide 2 matches the cross-section of waveguide 3, is aligned with it front-to-rear in Fig. '7, and is joined and sealed to it at right angles as aforesaid, the interiors of waveguides 2. and 3 being placed in communication with each other by an H-shaped aperture 6, having the same cross-section as waveguide 3 and being out through the top wall of waveguide 3. With the position of the electron-discharge device or wave amplifier l illustrated in Figs. 6 and 7, the longitudinal axis of guide section 2 extends substantially vertically, while the longitudinal axis of guide section 3 extends substantially horizontally.

An external magnetic structure, designated generally by numeral 1, is positioned outside of the sealed and evacuated portion of waveguide section 2, and extends downwardly somewhat below waveguide section 3, as shown in Fig. 6. Most of the magnetic structure has been removed from the device in Fig. '7, in order to show the waveguide components of said device more clearly. Structure 1 includes a pair of oppositely-disposed U-shaped permanent magnets 8 and 9, together with three vertically-aligned pole pieces I0. II, and I2, each of which has a central opening extending therethrough. The upper legs of magnets 8 and 9 are north poles, as indicated, and the lower legs of said magnets are south poles. All of the poles pieces have the same maximum external dimensions; upper pole piece I is in intimate contact with the upper north poles of the magnets, and lower pole piece I2 is in intimate contact with the lower south poles of the magnets.

The upper pole piece l0 entirely surrounds upper portion 2a of Waveguide 2, the hole through said pole piece being shaped accordingly, and the lower end of said pole piece is spaced slightly above the upper end of central pole piece II to provide an air gapvf'l3 in the region immediately adjacent the lower end of waveguide portion 2a; pole piece I0 is tapered down to have a reduced cross-section at gap I3, to give high magnetic flux density thereat. The inner surface of pole piece II] is in firm contact with the outer wall of portion 2a of waveguide section 2.

Central pole piece II surrounds portion 21) of waveguide 2, including the upper and lower transition sections which connect the rectangular part of waveguide portion 2b with the upper H-shaped waveguide portion 2a and Waveguide 3, but is spaced slightly from the outer wall of portion 2?); the hole through said pole piece is shaped accordingly, so that said hole has a central region of maximum size, tapering at its upper and lower ends to upper and lower areas of decreased size. The upper face of pole piece II is tapered to a relatively small area which is substantially equal to that of the lower face of pole piece Ill, which faces it across gap I3, while the lower face of pole piece II is tapered to a relatively small area substantially equal to that of its upper face. The lower face of pole piece I I rests on, and is in firm contact with, the upper wall of waveguide section 3, and has a corresponding configuration.

A coil I4, of substantially the same length as pole piece II, is wound around waveguide portion 212 and within pole piece I i, in the space between the outer wall of said waveguide portion and the inner wall of said pole piece, that is, in the central opening of said pole piece. Opposite ends of coil 14 are connected, by conductors I5 and I6, to opposite terminals of a source I1 'of direct current, for example, a battery. The current flowing through coil I4 establishes a magnetic field within waveguide portion 2b, in a direction parallel to the longitudinal axis of said portion.

The lower face of pole piece II is spaced from the upper surface of pole piece I2 to provide an air gap I8 therebetween, and in this air gap is positioned the waveguide section 3. The upper face of pole piece I2 is hermetically sealed to the lower wall of waveguide section 3 and has a corresponding configuration, and, since the lower face of pole piece Ii is in firm contact with the 10 upper wall of waveguide section 3, said waveguide section defines the lower air gap I 8.

It should be brought out at this juncture why it is preferable to use an H-shaped waveguide for waveguide section 3. It has been found that, for the same cutoff Wavelength, an H-shaped waveguide can have smaller overall dimensions than a rectangular waveguide; since the dimensions of waveguide section 3 determine the length of air gap I8, this means a shorter air gap is possible with an H-shaped guide than with a rectangular guide having the same cut-off frequency. It is desirable to decrease the length of the air gaps in the magnetic circuit as much as possible, in order to decrease the total reluctance of the magnetic circuit, since rather high magnetic field strengths are necessary 'in devices of this character, as explained above.

The central openings through pole pieces II), II and I2 all have a common longitudinal axis or center line, these pole pieces therefore being vertically aligned. Lower pole piece I2 is tapered upwardly to an area at its upper face which is smaller than the area of its lower face, but the area of its upper face is somewhat larger than the area of the lower face of pole piece I I, which faces it across gap I8, for a purpose to be explained hereinafter.

A circular aperture I9 is cut through the central portion of the lower wall of waveguide section 3, said aperture having its center aligned with the longitudinal axes or center lines of the openings in pole pieces I0, II and I2. Th upper end of the central opening in pole piece I2 has the same diameter as aperture I9, and from this end said central opening is enlarged as we progress downwardly, by an outward taper, to a diameter approximately equal to that of the central portion of the central hole in pole piece I I. When this diameter of the hole in pole piece I2 has been reached, it is maintained constant from the point where it is reached, downwardly to the lower end of said pole piece.

A hollow metallic cylinder is hermetically sealed into the lower end of the central opening in pole piece I2, and projects downwardly therefrom below the lower end of said pole piece. A sealing member or disc 2| closes and hermetically seals the lower openend of cylinder 20. From the above-described structure, it may be seen that a hermetically-sealed enclosure is provided, constituted by the following elements: disk 2I, cylinder 20, pole piece I2, waveguide 3 with sealing member 5 and end wall 3a, and waveguide 2 with sealing member 4. This enclosure is evacuated and hermetically sealed to provide an evacuated metallic envelope.

A coil 22 is wound within pole piece I2, in the central opening thereof, this coil having a length such that it extends from the lower external wall of waveguide section 3 to a point adjacent the upper end of cylinder 20. Opposite ends of coil 22 are connected, by conductors 23 and 24 sealed through disk 2|, to opposite terminals of source Il'; current flowing through coil 22 establishes a magnetic field within pole piece I2, in a direction parallel to the longitudinal axis of the central opening thereof. Conductors 23 and 24 are sealed through disk 2|.

An electron gun and injection structure, designated generally by the numeral 25, is positioned and mounted within the central opening in pole piece I2, in a location an appreciable distance below aperture I9. Electron gun structure 25 comprises a central electron-emissive cathode T1 element 26, a concentric hollow open-ended cylindrical primary anode element 21 radially spaced from element 26, and a coaxial deflecting or secondary ring anode element 28 which is spaced beyond (above) the upper ends of the elements 26 and 21. Cathode 26 is provided with an electron-emissive coating of any suitable type, this coating being heated to the temperature of thermionic emission by a suitable heater (not shown), the opposite ends of which are connected, by means of leads 29 and 30, to opposite ends of the secondary winding 3| of a heater transformer 32, the primary winding of which is connected to a suitable source of heating current. The coaxial line 33, which supports the cathode element 26, passes out of the evacuatedinterior of the device through disk 2! and is sealed therethrough, the inner conductor of this line being hermetically sealed into the outer conductor thereof by means of a vitreous seal 3 Lead 30, which is connected to the outer conductor of coaxial line 33, is electrically connected to the negative end of a direct current source 35 of high voltage, so as to maintain cathode 26 at a suitable negative potential. Anode element 2! is connected, by a conductor 35 which is sealed through disk 2 i, to the positive terminal of source 35, so as to maintain said anode at a high positive potential with respect to cathode 23. Conductor 36 also supports element 21 in position.

A lead-in and support wire 3?, which is sealed through disk 2|, connects ring anode 28 to an intermediate point, on source 35, which is positive with respect to cathode 26.

Finally, to complete the electrical connections to the device I, a conductor 33 electrically connects cylinder to an intermediate point, on source 35, which is positive with respect to cathode 26 and is nearer the positive end of source 35 than is the point to which ring anode 28 is connected. Since metallic cylinder 20 is fastened to pole piece 12, since said pole piece is fastened to waveguide section 3, and since waveguide section 3 is fastened to waveguide section 2, all of these fastenings being hermetic seals, a potential positive with respect to cathode 26 is applied to waveguide section 2.

The output end of waveguide section 3 may be brought out in another plane than that of Fig. 6,

in order to clear magnet 8; the showing of said output end and said magnet in the same plane, in Fig. 6, is merely for convenience of illustration.

The magnetic circuit including the permanent magnets may be traced as follows: from the north pole of either magnet 8 or 9, through hole piece l0, gap [3, pole piece ii, gap i8, and pole piece 12, to the south pole of the same permanent magnet. It will be seen that magnetic flux substantially parallel to the longitudinal axis of waveguide 2 is provided across gaps l3 and it mainly by this permanent magnet circuit, while magnetic flux substantially parallel to the said longitudinal axis, between gaps is and i8, is provided mainly by coil id, and magnetic flux substantially parallel to said longitudinal axis, below gap I8, is provided mainly by coil 22.

Having described the structure of the electrondischarge device I, we will now proceed to a consideration of the mode of operation thereof as a microwave amplifier.

A copious supply of electrons is thermionically emitted from cathode 26, these electrons originally proceeding toward primary anode 21 because of the positive bias thereon. As described above, coil 22 produces a longitudinal magnetic field along the axis of the central opening in pole piece 12. Due to the transverse motion (from cathode 28 toward anode 2i) of each electron across this magnetic field, each electron is caused to move in an orbital substantially circular path'around the longitudinal axis of the opening through lower pole piece l2.

The above discussion has neglected the effect of ring anode 28. However, due to the presence of this element which is biased positively with respect to cathode 2%, each electron is caused to move or drift upwardly axially along the opening in pole piece I2 with a velocity 'Ue determined by the potential on anode 28. Each electron-as a result of its circular orbital motion combined with its longitudinal motion, follows a helical path; all of the electrons taken together can be pictured as a hollow substantially cylindrical cloud which rotates and moves upwardly along the longitudinal axis or center line of pole piece 12. This cloud can move beyond or above element 28 because said element is toroidal, having a hole in its center.

Now referring to Fig. 8, this figiu'e represents a TE01 mode wave radiating or propagating through the TE'o,1 wave amplifier of this invention, this wave propagating downwardly along waveguide 2 in the directionof arrow A, turning the right angle at the junction of guides 2 and 3, then propagating along waveguide 3 in the direc-- tion of arrow B. Termination or end wall 3a of waveguide 3 is constructed and arranged to terminate said waveguide at this closed end in its characteristic impedance, thereby preventing any appreciable reflection of waves from this end of said guide; this termination is represented as a resistance in Fig. 8.

The arrows extending between the walls of the waveguides 2 and 3 in Fig. 8 represent the electric field lines for an instantaneous wave pattern. It will be noted that, at this instant, right at the junction between guides 2 and 3, as indicated by arrows C and D, respectively, the electric force has a maximum value downward at an angle at the left side of aperture i9, and a maximum value upward at an angle at the right side of said aperture.

Now considering the changes in the wave pattern with respect to time which take place along any particular line, such as line D, there is a sinusoidal variation of the electric field along this line at a periodicity corresponding to the linear phase velocity of the electromagnetic wave. In other words, the angular position or location of line D does not change with respect to time, but the force represented by said line varies only from a maximum to zero and then to a maximum in a direction directly opposite to that indicated by the arrow on said line in Fig. 8; all of these changes occur along the line D.

As the rotating cloud of electrons, which is drifting upwardly axially of pole piece i2, reaches the vicinity of aperture i9, it reaches a region of increased longitudinal magnetic field strength. The aforesaid cloud passes through aperture l9 and proceeds toward aperture 5 at the upper side of waveguide 3. In the vicinity of aperture 6, at the upper end of gap l8, the value of the magnetic field strength B is sufficient, due to the presence of pole pieces II and 12 on opposite sides of this gap (the lower surface of pole piece ll having a relatively small area), to satisfy Equation 13, so that in this region the electrons are moving in synchronism with the angular phase velocity of the wave. This region, in the vicinity of aperture 6, may therefore be termed a synchronous field region. Due to the somewhat larger area of the upper face of pole piece l2, in the lower part of waveguide 3 (that is, in the vicinity of aperture l9), the value of the magnetic field strength B is somewhat less than that necessary to produce electron motion in synchronism with the angular phase velocity of the wave; in other words, the electron angular velocity is less in the vicinity of aperture l9 than in the vicinity of aperture 6 (see Equation 12).

Due to the fact that the cloud of electrons originally is a complete circle, there would be just as many electrons abstracting energy from the wave at any instant on one side of the circle as there would be imparting energy thereto on the other, because on one side of the center of the waveguide there would be forces tending to decrease the orbital radius of the electrons, while on the other side of the center there would be equal and opposite forces tending to increase the orbital radius of the electrons. Therefore, in order to change the net energy level of the wave, it is necessary to select the electrons having only a particular phase with respect to the wave, and to cause only these electrons to react with the wave. The aforesaid selection is made by an action which may be termed angular launching of the electrons.

Consider now a circular cloud of electrons :5

which arrives in the vicinity of aperture 6 (that is, in the synchronous field region) at an instant when the electromagnetic wave in the waveguide has the wave pattern represented in Fig. 8. At

this intsant, the forces C and D have their maximum values in the directions indicated by the respective arrows thereon. The right-hand half of the aforesaid circular cloud of electrons is pushed upwardly by force D, while the left-hand half of said cloud is pushed downwardly by force C.

When the upwardly-pushed half of the circular cloud has rotated so that it is on the left-hand side, forces C and D have reversed in direction,

since in this region the electrons are moving in synchronism with the angular phase velocity of the wave. At this time, therefore, force C is directed upwardly, and force D is directed downwardly. The upwardly-pushed half of the circle is therefore now given another upward push,

while the downwardly-pushed or opposite half of the circle is given another downward push. This process repeats itself as the electrons continue to rotate, so that the half of the circle of electrons which has a certain predetermined phase with respect to the wave is given repeated upward pushes, while the opposite half is given repeated downward pushes. As a result, substantially half "of the circle of electrons proceeds upwardly into waveguide portion 2b, the other half of the circle proceeding downwardly toward the somewhat weaker magnetic field region in the vicinity of aperture I 9.

The end result of the above action may be likened to splitting a hollow cylinder lengthwise along its longitudinal axis and displacing one above described, into the region of somewhat weaker longitudinal magnetic field in the vicinity of aperture l9. In this region, the angular velocity of the electrons is decreased (see Equation 12), so that these electrons lag behind the angular phase velocity of the wave. Eventually the lag becomes equal to one complete cycle of the Wave, and at such a time these electrons have the certain desired predetermined phase with respect to the wave, so that these electrons are pushed upwardly into waveguide portion 2b to be effective for purposes of amplifying the traveling wave, as described hereinafter.

If desired, the potential on element 23 may be made such that the downward push of the electrons of improper phase, by the wave, will be insufficient to force such electrons clear down to the lower wall of waveguide 3, there to become trapped because of the positive potential of said wall.

The above-described events which take place in gap l8 may be termed the process of angular bunching, since only a semi-cylindrical group or bunch of electrons having a certain predetermined phase with respect to the wave travels upwardly into waveguide portion '21).

The angularly-bunched electrons move upwardly from the vicinity of aperture 6 into waveguide portion 21). Because these electrons are moving out of the high-fiux-density gap I8 into a region of lesser magnetic flux density, the angular velocity of the electrons will decrease, in accordance with Equation 12, and the motion of the electrons will no longer be in phase synchronism with the angular phase velocity of the wave. However, these bunched electrons will maintain the same positions relative to each other, because the angular velocities of all of such electrons will be simultaneously decreased by the same amount.

Because the electrons in this region are not moving in synchronism with the angular phase velocity of the wave, there is a substantially zero net energy interchange or reaction between the electrons and the electromagnetic wave in said reg1on.

The magnetic field B in the upper air gap l3, due to the presence of pole pieces II) and H with their reduced cross-sections facing each other across this gap, is suflicient to satisfy Equation 13, so that the region of this air gap may also be termed a synchronous field region.

The tangential velocity of the electrons is held fixed. The interior of waveguide portion 21) may be termed a drift space, and in this space the angular velocity of the electrons decreases, as explained above. Since the tangential electron velocity is held fixed, and since the angular velocity of the electrons decreases, the diameter of the circular electron orbit must necessarily increase. To allow for this increase in orbital diameter, the waveguide portion 2b has an increased cross-section, as previously described.

In order for the rotating upwardly-drifting half-cylindrical angularly-bunched cloud of electrons to give up energy to the wave in the Waveguide, it is necessary that such. electrons have a predetermined appropriate phase with respect to the wave in the synchronous field regicn l3. Waveguide portion 21) may have any convenient length, and in order to achieve the necessary and appropriate phase relation of the electrons and the wave, waveguide 2 maybe given a twist around its longitudinal axis. The aforesaid phase relation may be varied by varying either the length of waveguide portion 21) or the twist of waveguide 2 between aperture 6 and gap l3. Since the wave pattern is in effect rotated by twists in the waveguide (that is, it follows twists in the waveguide), while the electron path is unaffected by twists in the guide, the phase relation of the wave and the electrons can be varied by twistingthe guide. It should be apparent that the said phase relation can be varied also by varying the length of waveguide portion 21), since the length of said guide determines the number of circular orbits traveled by the electrons, for a given velocity along the guide and for a given electron angular velocity.

The angularly-bunched or half-cylindrical cloud of electrons drifts upwardly along waveguide portion 2b, the angular velocity of the electrons first decreasing as they proceed upwardly from aperture 9, due to the decreasing longitudinal magnetic field, and then gradually increasing as the electrons approach gap l3, due to the increasing longitudinal magnetic field.

As the electrons enter the synchronous field region at gap 13, their angular velocity increases sufficiently so that the motion of the electrons is in synchronism with the angular phase velocity of the wave. Under these conditions, and since at this position the angularly-bunched electrons have the proper phase relation to the wave to impart energy thereto, energy is abstracted from the orbital motion of the electrons by the electric field of the wave, resulting in an increase in the energy level of the wave and a decrease in the orbital radius of travel of the electrons. This increase in energy level of the microwaves traveling along the waveguide is, of course, amplification of the microwaves.

Since the angular-bunching action selects and causes to react with the wave only those electrons which have the proper phase relation with the wave to impart energy thereto, there is a net increase in the energy level, or an amplification, of the wave.

The synchronous field region extends substantially throughout gap I3, so that the motion of the electrons is in synchronism with the angular phase velocity of the wave substantially throughout this gap; therefore, during substantially the entire travel of the electrons through this gap,

they impart or give up energy to the traveling wave. Electrons at the upper end of gap E3, or therebeyond, have given up substantially all of their available orbital energy, and such electrons are traveling substantially axially of pole piece I; these electrons eventually wander to the walls of waveguide 2, which are supplied with a positive potential by lead 32, and return to the cathode.

To summarize, the microwave amplifier of this invention functions as follows: the electromagnetic wave produces angular bunching of the electrons in the region l8, the bunched electrons then traveling upwardly and giving up energy to the electromagnetic wave in the region l3.

Now referring to Fig. 9, this figure shows an electron accelerator according to this invention. A waveguide 39 is hermetically sealed and evacuated to provide an envelope for the electrons to be accelerated. One end of said guide is closed by an end wall 39a, onthe interior surface of which is mounted a suitable target 49. Two concentric lines 4| and 42 have inner conductors 43 and 44, respectively, which extend into waveguide 39through suitable apertures provided in the end wall of said guide opposite from wall 39a. Conductors 43 and 44 are so arranged and positioned with respect to guide 39, and are so supplied with microwave energy, that they act as exciting rods to establish an E1,2 mode wave in said guide. The field pattern of this mode of wave is shown in Figs. 3 and 4. The guide 39 is evacuated and a vitreous sealing member 45 seals off from the atmosphere the opening in guide 39 through which conductor 43 extends, while a similar sealing member 43 seals off the opening through which conductor 44 extends.

An electron-emissive cathode element 41 is positioned centrally of guide 39 near the waveinput end thereof, this element being supported in position by means of a pair of lead-in conductors 48 and 49 which are sealed through the side wall of said guide. A hollow open-ended cylindrical primary anode element 59 is positioned in guide 39 concentrically of cathode 41, element 59 being spaced from cathode 41 and being supported in position by a lead-in conductor 5| sealed through the side wall of said guide. A coaxial deflecting or secondary ring anode element 52 is positioned in guide 39 on the side of cathode 41 toward target 49, element 52 being supported in position in said guide by a lead-in conductor 53 sealed through the side wall of said guide.

A coil 54 is wound around the waveguide 39, adjacent turns of said coil being spaced further and further apart as it progresses from cathode 4'! toward target 49. This coil has a length sub stantially equal to that portion of the length of the guide, from the end of cathode 41 closest to the wave-input end of the guide, to wall 39a.

The connections from the various elements of the Fig. 9 accelerator to the external voltage sources are similar to those of Fig. 6; that is, cathode 41 is connected in the same manner as cathode 26, anode 59 is connected in the same manner as anode 21, anode 52 is connected in the same manner as anode 28, coil 54 is connected in the same manner as coils [4 or 22, and waveguide 39 is connected by lead 55 in the same manner as waveguides 2 and 3.

By the above-described connections, a longitudinal magnetic field is established by coil 54 along the axis or center line of guide 39, and, when the electron gun and injection structure 47I50-52 and the coil 54 are energized, a hollow substantially cylindrical rotating cloud of electrons, moving to the right along waveguide 39 with an initial velocity We, is produced, the action in this respect being substantially similar to the action of the electron gun and injection structure of Fig. 6.

It is desired to be made clear at this point that there is a difference between the action of an amplifier of this invention .and the action of an electron accelerator according to this invention. In the wave amplifier, in order to increase the energy level of the wave, it is necessary to bunch the electrons, in order to select and to cause to react with the wave only those electrons-which have a particular phase with respect to the wave, since otherwise at any instant there would be just as many electrons abstracting energy from the wave as would be imparting energy thereto at the same instant. In the electron accelerator, on the other hand, it isnot considered necessary to bunch or select only particular electrons. Although some of the electrons will impart energy to the wave, thereby being decelerated, others will abstract energy therefrom and will be accelerated as desired; it is immaterial whether the net energy level of the wave is changed or not, and

electrons are ordinarily present in sufflcientnumbers so that it isentirelysatisfactoryto accelerate only some but not all of them, the number of those accelerated being ample to accomplish the desired end result, whatever: that may be angular phase velocity of the wave, in the vicinity;

of anode 52, there will be aninterchange of energy between the electrons and the electromag: netic wave radiating through waveguide -39,substantially half of the electrons abstracting energy from the electric field or wave and substantially half of the electrons givingup or-imparting energy to the electric-field or wave. Be-

cause of the Emode excitation of the waveguide, those electrons, of the hollow cylindrical rotating cloud, which abstract-energy from the wave,-are

given a push along the longitudinal axis of the waveguide, in the direction of target 40. In other words, they are accelerated-along the guide, thereby increasing their velocity us.

As the velocity veincreases, the quantity inside the parentheses in Equation 13 decreases, so that the magnetic field B necessary to maintainthe electrons in motion in synchronism with the angular phase velocity 'of the wave decreases. Therefore, as we progress along the guide from anode 52 to target lfl the spacing between adjacent turns of coil 54 is increased to decrease B, corresponding to the increase of 'Ue, in such a manner as to maintain -synchronism between the motion of the-electrons and the-angular phase velocity of the --wave, throughout substantially the entire travel of the electrons from anode 52 to target 40. Those electrons which are injected into the waveguide with such a phaserelationtothe wave as to abstract or -absorbenergy-therefrom, are therefore acceleratedthroughout substantially their entire travel, from the vicinity of anode 52 to the target end of waveguide 39, so

that they strike target with a very high velocity, thus producing high voltage X-rays, for

example, if said target is made of an appropri- Such electrons can return to the ate material. cathode by means of lead 55.

The electrons which are injectedinto the waveguide with such a phase relative to the. wave as, to give up energy thereto, are thereby-deceler-. atecl and subsequently either drop into phase with subsequently-emitted electrons .beingaccelerated.

or wander to the side .wall ofthe waveguide. and return to the cathode. In either case, they do not interfere appreciablywith the accelerated,

electrons, and the device of Fig. 9 thereforefunctions very efiectively and simply as an electron accelerator, with the waveguide thereofexcited with an E mode wave.

The interior of waveguide 39 is. hermetically. sealed off from the atmosphere and evacuated, in

accordance with conventional electron-discharge device practice.

It hasbeen disclosed above,. in the general dise cussion of the theory on which this inventionis o based, that an electron accelerator excited I in a a TE 5 modewill add energy in the' orbital com ponent of the electron path, and -as-the-"-tangential velocity of the electron exceeds about 3,000 volts, theelectronmass willincrease.- With small or substantially zero drift'along thewaveguide;

the increase in magnetic field required by Equation 13 due to the increase in the orbital electron mass,- will have to be secured by a radial in-- crease in the magnitudeof the magnetic fieldi Fig. 10 illustrates an electron accelerator which operatesin accordance with the principles just stated. A hollow waveguide 56 is closed at both ends and has, near thelower endwall' 56a thereof, an exciting rod 51 extendinginto the in ter-ior of-said guide through-a suitable opening in the side wall thereof. Rod 51 is sealed through theside wall of said guide, and is adapted-tube connected to a source of microwave -energy-toestablish or set up a TEo,1 mode waveinsaid" The end of waveguide 56 oppositefromendwall 56a is closed, and said waveguide is terminated at this. end by awave impedancemeans 58 which functions to prevent any apuide.

preoiable reflection of microwave energy from this end of the waveguide.

Upper and lower pole pieces 59 and 60. respectively, surround waveguide 56 in the vicinity of impedance means 58, said pole pieces being axially aligned with each other a and being vertically spaced from each other to provide a short air gap 6|. Pole pieces 59 and 66 are coupled to a source of magnetomotive forc (not shown),

which may be a permanent magnet for'example,

in such a manner that there is longitudinal mag -netic flux ingapGI. Pole-pieces 59 and60 are each tapered toward gap 6i sothat thesurfaces whichface each other across said gap have rather small cross-Sections, thus producing a very high flux density at said gap.

in waveguide 56. Upper'pole piece' 59, which surrounds waveguide 56 adjacent the upper end thereof, is solid above the end of said guideybut has a central aperture 63 extending upwardly? from the lower end thereof a short distance thereinto to accommodate therein the upperend-portion of guide 56; above this a pyramidalopening 62, the lower end of whichmerges into-the walls defining aperture 63, extends upwardly into pole As a resultof the piece 59 a short distances above-described pole piece configuration, a substantially longitudinal magnetic field is provided inside guide 56, this fieldbeing weakest at the center of saidguide and increasingin strength as one "progresses radially outwardly from the center of. said guide; this field is strongest at the side wall. of saidguide.

It is 'to be understood that waveguide '56 isentirelyclosed and hermetically sealed from the atmosphere, so that the interior thereof can be evacuated.

An electron-emissive cathode element 64 is mounted centrally inside waveguide 56 in a plane-- immediately below the lower endof gap 6!; said. element is supported in position by the-lead-in" conductors 65 and 66.2which are sealed through the side wall of guide 56. Ahollow open-ended cylindrical primary anode element 61 is mounted in guide 56 concentrically of cathode 64, element 6! being radially spaced from said cathode and being supported in position by alead-in conductor 68 sealed through .the side wall ofsaid guide- A grid structure 69, ofsuch size as-to cover substantially. .all' of. the interior hollow Lower pole piece 66 has a central hOle extending entirely therethrough toaccommodate therespace. inside guide 56, is mounted horizontally inside said guide, in a plane immediately above the upper end of gap 6|; this grid is supported in position by a lead-in conductor 10, which is sealed through the side wall of said guide.

Leads 65, 66, and 68 are connected to suitable direct current sources (not shown) in such a manner that anode 61 has a positive potential with respect to the electron-emitting cathode 64. Conductor 10 is connected to a source in such a way (shat grid 69 is positive with respect to cathode 6 Electrons emitted from cathode 64 proceed toward anode 6'! because of the positive bias thereon. These electrons are made to travel in circular orbits as a result of their transverse motion across the longitudinal magnetic field established by the substantially vertical leakage flux between pole pieces 59 and 60. The electrons thus traveling in circular orbits are caused to have a small upward drift or velocity along guide 56 by a small positive bias applied to grid 69.

Thisgrid also has another function which will be described hereinafter. The magnetic field strength B in the central portion of gap 6| is sufficient to satisfy Equation 13,.-so that the motion of the rotating hollow cylindrical slowlydrifting cloud of electrons is originally in synchronism with the angular phase velocity of the H wave radiating upwardly along waveguide 56. Since this condition of angular phase synchronism exists, there will be an interchange of energy between the electrons and the electromagnetic wave radiating through waveguide 56, substantially half of the cloud of electrons abstracting energy from the electric field or wave and substantially half of the cloud giving up or imparting energy to the electric field or wave. Because of the H mode excitation of the guide, those electrons of the hollow cylindrical rotating cloud which abstract energy from the wave, are given a push toward the side wall of the guide, so that in effect their orbital radius is increased and they begin to travel in larger orbits. Energy is added, therefore, in the orbital component of the electron path, and the tangential velocity of such electrons increases, since their angular velocity we remains constant.

Each line of magnetic flux from pole piece 59 to pole piece 60 is directed downwardly and also radially outwardly, due to the aperture 62 in pole piece 59 and to the annular shape of pole piece 60. As some of the electrons begin to travel in larger orbits, the path of the same is substantially a spiral, since the electron upward drift is small. The magnetic flux lines tend to exert a guiding action on such electrons, so that, in the absence of any positive'bias on grid 69, such electrons would tend to slide downwardly along the flux lines as their orbital radius increases, since the flux lines are directed downwardly as they progress radially outwardly. However, grid 69is given a sufficient positive bias to counteract this tendency toward downward motion of the accelerated electrons; this bias is, in fact, slightly greater than that necessary to counteract the downward-motion tendency, so that the electrons have a net small upward drift or velocity along guide 56.

As the orbital radius of those electrons. which.

have the proper phase is increased due to the addition of. energy in the orbital component of the electron path, the tangential velocity of such electrons increases, since their angular velocity we remains constant. Below values of electron velocity we is constant, so, that theelectron motion stays in synchronism with the angular phase velocity of the wave propagating along guide 56. However, as the tangential velocity of the accelerated electrons exceeds about 3,000 volts, the electron mass will increase. From Equation 13, it may be seen that this increase in orbital electron massm requires an increase in-magnetic'field Bin order to maintain the electron motion in synchronism with the angular phase velocity of the wave. Since'in Fig. 10 the tangential velocity of the accelerated electrons increases as their orbital radius increases, the orbital electron mass increases as-the orbital radius increases, so that the required increase in magnetic field has to be secured by a radial increase in the magnitude of the magnetic field; as described above, the configuration of pole pieces 59 and 60 is such that there is a radial increase in the strength of the magnetic field, as required to maintain the motion of the accelerated electrons in synchronism with the angular phase velocity.

From the above description, it may be seen that a very effective electron accelerator has been deyised according to the invention, the energy necessary to accelerate the electrons being provided by a TE mode wave in a waveguide.

Let us now consider a waveguide which, instead of being straight, has imparted thereto some, curvature. This enables us to devise arrange-- ments for imparting curved paths to. the electron. under the influence of the field pattern. In, Fig. 2, let us adopt the reference point (1 and es-- tablish the conditions of angular phase synchro nism. Thus, 10 reduces to l, and from Equation. 11 we see that with We equal to zero the angularfrequency of the electron equals the angular frequency of the wave. Thus, when the electron is at thetop of the waveguide it will be pushed up. When it reaches the left side of the waveguide, the voltage of the wave will be zero and there will be no sidewise push. However, in traveling through this quadrant there will also be. a component of force pushing the electron to the right. When the electron reaches the bottom of the guide, it will again be pushed up, while in traveling through the proceeding quadrant it will be pushed to the left. As we progress around the waveguide we find that the push on the electron is always up and alternately to the right and left under conditions of synchronism, resulting in a net push upward. If the electron is traveling at a given velocity at right angles to the wave pattern of Fig. 2, the resultant forces on the electron will cause it to move substantiallyin a circle, which it will do provided we bend the. waveguide in a circle of the proper radius. Such. an arrangement might be useful in devices where. it is desired to cause an electron traveling at high speed to travel in a circle, the field pattern in such cases supplying the centripetal force.

Now let us'consider a special case of the analysis of the angular phase velocity of a wave in a waveguide, the principles of which are applicable to a microwave amplifier, as well as to an electron accelerator.

Consider a waveguide with an excitation in which the field pattern has at least one complete reversal along a closed path transverse to the waveguide, such as that shown in Figs. 3 and 4, for example. If an electron is injected into this which the waveguide is given a complete half turn relative to the path of the electron. When the electronhas traveled throughthe distance g the wave should have propagated beyond the position of the electron by an amountequal to.

M2 so that, the reversal due to the, twist of the waveguide will orient the field to maintain the same phase relation withrespect to the. electron as at the zero pointof distanced. If Up is the phase velocity of the wave. 11 is the velocity of the electrqn along the guide and )\g isthe wavelength in the waveguide, then, we can, calculate the time t during which the wave has; progressed ig/2 with respect to. the electron as h 2 Fur. But

g vet (15) Substituting the value of i from Equation 14,

0,). new? (16) If We wish to produce an electron accelerator we can excite a rectangular Waveguide with a longitudinal electric mode of the necessary configuration to give at least one reversal of the field pattern, as pointed out above, and to produce acceleration of the electron along the guide. With a given velocity of electron injection, we can calculate the pitch of the waveguide twist at the point of injection, from Equation 16. However, the electrons of proper phase will start to accelerate, thus reducing the pitch progressively along the guide. Thus, even with moderate injection speeds, due to the relatively large accelerating voltages which can be produced in waveguides, the pitch of the twist, even within the initial length of waveguide, can be made reasonably small. The magnitude and variation in pitch can be calculated from the characteristics of the waveguide as related to the frequency of the wave, as well as the voltages within the waveguide. We can designate the pitch of the waveguide in terms of turns per A; as

If De increases by an increment U1 per unit length then the pitch at a point a distance. L. along the waveguide from the point of injection is given as the. electron, relative thereto, the pitch-of the twist. in the, waveguide can be reduced if: the electron is traveling in. a sense oppositeto that or. the Waveguide twist. If we give tothe. electron an orbital period. we, in opposition to the. twist in the waveguide, then we can. multiply the distance G by a. factor It where Now referring to Figs. 11 and 12, these figures illustrate. an electron. accelerator according to this invention, this, accelerator utilizing. some of the twisted waveguideprinciples.above set forth. Fig. 11 is a longitudinal crosssection through. the electron accelerator, while Fig. 12 is a perspective diagram of the waveguide itself. Waveguide 1| which may. have a rectangular crosssection. as. shown, is hermetically sealed and evacuated to serve as an envelope for the electrons to be. accelerated. One end of the. guide is closed by an end wall Ha, on the interior surface of which is mounted a suitable target 12 on. which the accelerated electrons may impinge with a high velocity. Two concentric lines 13 and 14 have inner conductors 15 and 16, respectively, which extend into. waveguide ll through suitable apertures. provided in the end wall of said guide opposite from wall He. Conductors 15 and 16 are so arranged and positioned with respect to guide H, and are so. supplied with microwave energy, that theyact as exciting rods or probes to establish an E1,2 mode wave in said guide. The field pattern of this mode of wave is shown. in. Figs. 3 and 4. Conductors l5. and 1,6. aresealed through the end. wall of guide H into the evacuated. interior thereof.

An. electron-emissive cathode element 1.1 is

positioned eccentrically in guide H near the wave-input end thereof, this element being supported in. position by means of, a. pair oi lead-in conductors l8 whichare sealed through the side Wall of said guide. A ring anode element [9 is 6 positioned in. guide H adjacentto and aligned withcathode 11, on the side of cathode H toward target 12, element 19 being supported in position in said guide by a lead-inconductor sealed through the side wall of said guide.

A coil 8| is'wound, around the waveguide II, this coil having a length substantially equal to that. portion. of said guide from cathode 11 to wall Ha.

Cathode. leads 18 are connected to a source (not shown) in. such a manner as to heat said cathode. to the temperature of thermionic emission and as to apply av negative potential to said cathode. Anode lead 80 is connected tothe same source as is cathode 11, in such. a manner as to make the anode 19 positive with respect to cathode ll. Leads 82 and 83, connected to opposite ends of coil8l, are connected to a suitable source, to energize said coil so that the same supplies a moderate substantially uniform longitudinal magnetic field inside guide ll.

Electrons emitted from cathode 11 are attracted toward anode 19 because of the positive potential thereon, and travel along substantially a straight line fromrcathode 151 to target 12. This line ecc ntri ith... respe t o t e longitiidinal axis of the waveguide, due to the eccentric position of cathode 11. The moderate longitudinal magnetic field provided by coil 8| is utilized to align the electron fiow along a line parallel to the longitudinal axis of guide H, thereby to prevent wandering of the electrons transversely.

As represented in Fig. 12, waveguide II is twisted along its length at a certain pitch; the length along which the waveguide is given a complete half turn relative to the path of the electron, which path is indicated by the dotted line F in Fig. 12' and is substantially a straight line, is denoted by g The line or path F,

although substantially linear with respect to space, is non-linear with respect to waveguide 1 I since said guide is twisted along its length. 'The length g, at the point of injection of the electrons, has the proper value determined by Equation 16 for the electron injection voltage used, so that the electrons travel along the waveguide H in phase with the phase velocity of the wave.

Since the electrons are moving in synchronism with the angular phase velocity of the E mode wave, there will be an interchange of energy between the electrons and the electromagnetic wave radiating through waveguide I I, those electrons having the proper phase abstracting energy from the electric field or wave and thereby being v accelerated. Because of the E mode excitation of the waveguide, those electrons which abstract or absorb energy from the wave are given a push along the longitudinal axis of the waveguide, in the direction of target 12. In other words, they are accelerated along the guide, thereby increasing their velocity vs. The increase of velocity De will increase the length 9, as seen from Equation 16, so that the twist of the waveguide may be reduced progressively along the guide. The variation in waveguide pitch along the guide may be calculated from Equation 18, and guide II is designed in accordance with this equation.

Therefore, the travel or motion of the electrons, along the guide 1 I, is maintained in synchronism or in phase with the angular phase velocity of the wave, throughout substantially the entire travel of the electrons from anode 19 to target 12.

Those electrons which are injected into the waveguide with such a phase relation to the wave' as to abstract or absorb energy therefrom (here' termed with the proper phase), are therefore accelerated or given a longitudinal push throughout substantially their entire travel, from the vicinity of anode 19 to the target end of guide I I so that they strike target 12 with a very high velocity. Because of the eccentricity of path F with respect to the longitudinal axis of the waveguide, these electrons are given a longitudinal push in the same direction throughout substantially their entire travel. Such electrons can return to the cathode from target 12 by means of lead 84, which is connected to waveguide H and to the source to which cathode 11 is also connected.

As should be apparent from the above description, the device of Figs. 11 and 12 functions very effectively as an electron accelerator, with the waveguide thereof excited with an E or longitudinal electric mode wave.

Of course, it is to be understood that this invention is not limited to the particular details as described above, as many equivalents will suggest themselves to those skilled in the art. It is accordingly desired that the appended claims be given a broad interpretation commensurate with the scope of this invention within the art.

"1. In combination: a wave" guide structureadapted to be excited by electromagnetic energy having a predetermined angular phase velocity;'

a source of ele'ctrons; means, adjacent saidsource gular velocity of said orbital component of motion of said electrons to substantially equal the. angular phase velocity of said electromagnetic energy. I

2. In combination: a wave guide structure adapted to be excited by electromagnetic energy having a predetermined angular phase velocity;

a source of electrons; means, adjacent said source of electrons, for injecting electrons from said source into said Wave guide structure with linear and orbital components of motion; and means} adjacent said wave guide structure. for establishing a longitudinal magnetic field therein; said magnetic field having a magnitude determined by the expression 10wm fa 6 v,

where w is the angular phase velocity of said electromagnetic energy, m is the electron mass, e is the electron charge, Be is the linear velocity of said electrons along said wave guide structure, and u is the linear phase velocity of said,

electromagnetic energy.

3. In combination: a wave guide structure adapted to be excited by electromagnetic energy having a predetermined angular phase velocity;

a source of electrons; means, adjacent said source of electrons, for injecting electrons from said source into said wave guide structure with an orbital component of motion; means, coupled to said wave guide structure, for separating said electrons into substantially in-phase and out-ofphase groups with respect to said electromagnetic energy; and means, adjacent said wave guide structure, for adjusting the angular velocity of said in-phase group of said electrons to substantially equal the angular phase velocity of said electromagnetic energy.

4. In combination: a wave guide structure,

adapted to be excited by electromagnetic energy having a predetermined angular phase velocity; a. source of electrons; means, adjacent said source' where w is the angular phase velocity of said electromagnetic energy, m is the electron mass, e is the electron charge, Ue is the linear velocity .of said in-phase group of said electrons along said wave guide structure, and Up is the linear phase velocity of said electromagnetic energy.

5. In combination: a wave guide structure including a pair of communicating wave guide sections which are perpendicularly disposed with respect. to. each otherand which are adapted to;

ill

be excited by electromagnetic energy having a predetermined angular phase velocity; means providing a magnetic circuit having a first gap in the region adjacent the junction between said wave guide sections; means, spaced from said lastnamed means, providing a magnetic circuit having a second gap in a region along the length of one of said wave guide sections; a source of electrons; means, adjacent said source of electrons, for projecting electrons from said source across said first gap in the direction of said second gap with linear and orbital components of motion; the strength of the magnetic fields across said first and second gaps, relative to the linear component of motion of said electrons, being such that the angular velocity of said electrons is substantially equal to the angular phase velocity of said electromagnetic energy; and the effective length of the path of said electromagnetic energy betweensaid first and second gaps being such as to provide a predetermined phase rela- REFERENCES CITED The following references are 01' record in the file of this patent:

UNITED STATES PATENTS Number Name Date 2,122,538 Potter July 5, 1938 2,233,126 Haeif Feb. 25, 1941 2,241,976 Blewett et a1. May '13, 1941 2,242,888 Hollmann May 20, 1941 2,289,756 Clavier et a1 July 14, 1942 2,300,052 Lindenblad Oct. 27, 1942 2,408,903 Biggs et'al Oct. 8, 1946 2,409,991 Strobel Oct. 22, 1946 2,454,094 Rosenthal Nov. '16, 1948

Images