US2925522A - Microwave linear accelerator circuit - Google Patents

Description

Feb. 16, 1960 M. G. KELLIHER 2,925,522

MICROWAVE LINEAR ACCELERATOR CIRCUIT Filed Sept. 30, 1955 2 Sheets-Sheet l Z DISTANCE ALONG WAVEGUIDE FIG. 2

ELECTRON VELOCITY DISTANCE ALONG WAVEGUIDE 3 Feb. 16, 1960 M. G. KELLIHER 2,925,522?

MICROWAVE LINEAR ACCELERATOR CIRCUIT Filed Sept. 30, 1955 2 Sheets-Sheet 2 ELECTRON VELOCITY DISTANCE ALONG WAVEGUIDE FIG. 4

FIG. 5

I FIG. 6

. 2,925,522 MICROWAVE LINEAR ACCELERATOR cmcurr Maurice Kelliher, Arlington, Mass, assignor to High Voltage Engineering Corporation, Cambridge, Mass., 'a corporation of Massachusetts Application September 30, 1955, Serial No. 537,825

3 Claims. (Cl. SIS-5.42)

This invention relates to microwave waveguide structures in which energy is transferred to an electron beam from a traveling electromagnetic wave, and in particu-' lar to a method of an apparatus for reducing phase error and increasing energy stability in such waveguide structures. The invention is particularly advantageous in microwave linear accelerators having, in addition to the main accelerating waveguide, a buncher waveguide; and, accordingly, the invention will be described with particular reference to such microwave linear accelerators. However, the invention is not limited to this particular application, but may be used to advantage in other microwave devices wherein electrons are injected at velocities substantially less than that of light into a waveguide structure in which a traveling electromagnetic wave issustained.

In a waveguide structure wherein there exists a tray/el ing electromagnetic wave and into which electrons are injected at velocities less than that of light, a transfer of energy to the electrons from the wave increases the velocity of the electrons, but does not alter the phase velocity of the wave. The waveguide structure is therefore designed so that the phase velocity of the wave changes, as the wave travels along the waveguide structure, in accordance with the change in velocity of the electrons as they absorb energy from the wave. In this way, the electrons are kept in the proper phase relationship with the wave throughout the length of the waveguide structure.

Now, the rate at which the velocity of the electrons increases at any particular point along the waveguide structure depends upon the electric field at that point and not directly upon the electron current. However, the electric field at all subsequent points along the waveguide structure does depend upon the rate at which the electrons absorb energy at the first point, and this rate of energy absorption is proportional to the square of the electron current. Therefore, the rate at which the velocity of the electrons increases at all subsequent points along the waveguide structure depends upon the electron current. 'velocity of the electrons will be less than that which they would have had if the electron current had not been increased, and this velocity deviation increases as the electrons travel along the guide. Of course, changing the electron current has no effect on the phase velocity of the wave. Consequently, the waveguide structure can be designed only for a particular electron current, and

any deviation in electron current results in phase error and hence unsatisfactory operation. Thus, forexample, the buncher of a microwave linear accelerator is designed for a particular electron current and a particular radio- Thus, if the electron current is increased, the

in electron current, by means of an inverse-feedback circuit to be described in detail hereinafter. Briefly stated, the invention comprehends feeding into the wavefrequency power flow, audit is not possible to vary the electron current output of the linear accelerator over wide limits without causing the buncher to operate unsatisfactorily and thereby to deteriorate performance of the linear accelerator.

The invention reduces phase error, despite variations frequency power fed back results in a decrease in the radio-frequency power entering the waveguide structure. It will be recalled that increasing the electron current in the waveguide structure not only decreases the electron: velocity so as to cause phase error, but also increases: the amount of energy absorbed from the wave. The radio-frequency power emerging from the output end of the waveguide structure is therefore decreased, so that. the effect of the inverse-feedback circuit is to increasethe radio-frequency power fed into the waveguide structure. This additional radio-frequency power tends to increase the velocity of the electrons in the waveguide structure, and thereby tends to compensate the velocity decrease initially caused by the increase in electron cur-- rent. In this manner the invention reduces phase error;. In particular, the invention enables a microwave linear accelerator having a buncher waveguide to be operated over a wide range of current output.

The invention also provides increased energy stability: that is, the invention stabilizes the energy with which electrons issue from the waveguide structure, despite variations in electron current. Without the inverse-feedback circuit of the invention, an increase in electron current causes a decrease in electron velocity, as hereinbefore described, so that the energy of the electrons at the output end of the waveguide structure is decreased. Since the inverse-feedback circuit of the invention tends to compensate any deviation in electron velocity caused by changes in electron current, it also tends to stabilize the energy of the electrons. In a microwave linear accelerator having a buncher waveguide to which the invention is applied, it is only the energy with which electrons issue from the buncher waveguide which is stabilized by the invention, and not the energy with which electrons issue from the main accelerating waveguide.

When the invention is applied to the buncher waveguide of a microwave linear accelerator, the invention provides increased efficiency in still another way. In both the buncher waveguide and the main accelerating waveguide, optimum operation is achieved when a certain amount of radio-frequency power, usually about ten percent of .the input power, is left over at the output vention, the power left over from the buncher waveguide is delivered to the main accelerating waveguide; thus, if the buncher waveguide uses ,P, the total loss to the single resistive load is 5% of P. i

The invention may best be understood from the following detailed description thereof, having reference to the accompanying drawings, in which:

Fig. 1 is a diagram showing a microwave linear accelerator, having a buncher waveguide in addition to the main accelerating waveguide, and having an inverse-feedback circuit constructed in accordance with the invention; Fig. 2 is a graph showing the axial electric fieldinthe buncher waveguide of Fig. 1 as a function of distance along the waveguide and illustrating the bunching process;

Fig. 3 is a graph showing, as a function of distance alongthe buncher Waveguide of Fig. l, the phasevelocity of. thetraveling electromagnetic wave; (unbroken line)? and. the velocity of tl1e-- electrons at very highsbeam our; rents. with (dashed linel-aud without. (dotted-lin the."

inverse-feedback circuit of the. invention; a r

Fig. 4is agraph, similar tothatofFig. 3.,,shotwing the phase velocity of the wave andsthe electron velocitiesxas. functionsof distance along the Waveguide, where the. electron current, while greater-than that .for which the buncher waveguide is designed, is not excessive;

Fig. 5 is a graphillustrating. the rangeof-energyand current output ofthe microwavelinear acceler-ator-ofi-Eig. 1 wit h (horizontai shading) and without (verticalshading) the, inverse-feedback circuit of the invention; and.

Fig. 6 isa graph, similar. to thatof Fig. 5, showingthea the buncher waveguide land enter themainaccelerating waveguide 3, they are travelingat .very nearly the velocity of,light., The bunched electron beam. continuesto gain energy .asit travelsthrough the main accelerating wave: guide 3, but without anyappreciable change in;velocit'y, the increased energytaking the forrnof an increase in. the electrons mass.

In the main, accelerating waveguide 3 (Fig. 1), both.

the phase velocity of the wave and the velocity of the electrons are very nearly the velocity of light, so that the electrons remain in the same phase relationship with the Wave throughout the length of the waveguide 3. The phase position of the electrons in the main accelerating waveguide '3 may be adjusted atwill, by means to be de- V scribediinl detail hereinafter.

The. bunching process inthe buncher waveguide is. I

illustrated in the graph of, Fig. 2, wherein the sinusoidal line shows the axial electric field in the bu'nchen waveguide 2' as a function or" distance along the buncher wave-- guide}. The horizontal axis (distance alongwaveguide),

is partly broken away, so that the graph of Fig. 2fshows the axial electric fielddistribution at four spaced segments s of the buncher waveguide 2. Positive values of E' indicate an electron-accelerating field, and hence a negative electric fieldin the conventional sense. The dots on the sinusoidal line indicate the electron distribution alongthe axis of the buncher waveguide 2, each dot representing a.

decreases. Since the buncher waveguide 2 is designed. so:

that the phase velocity of the wave is initially the same as the injection velocity of the wave, the electrons cluster in the accelerating half of the wave, ahead of the peak, as at "Z -Zg. Suppose now that the phase velocity-of the wave is causcd to increase-along the waveguide 2, by proper design thereof, in such a way as to keep pace with electrons at the point P on the wave. Then, electrons lying. between the peak of the Wave and the point :P'

will be accelerated morethan the, wave, and will .move forward towards the pointP. Electrons lying ahead-of thepointPywill be accelerated less than therwave, and will drop back towards the point P. Hence, the electrons wilLbunch at the point P, as at Z Z As theelectron velocity and phase velocity increase, the bunchingprocess I increasenzsoth-at the electrons issue from the :waveguide. 2 in 0 ml bunches traveling at very nearly-the .velocity of l sh tZnZs.

is that; absorbedby; the electrons.

Referring again to Fig. 2, it will be noted that the amplitude of the:wavedecreases as-the wave proceeds along the buncher Waveguide? This is because thewave is continuously losing energy, andm'ost of the energy'loss It has been statedthat the buncher waveguide 2 is designed so that the phase velocity of the wave increases along the'g'uide in step with the increase in velocity of the electrons at the point P. Now, the electron acceleration at the point P is proportional to theielectric field-at the point P, and the-electric field at the point P decreases along'the guide at aIrate; which is proportional to the rate of absorption-of energy- Since any increase in electron current;

by-the electrons. increases the rate of absorption of energy from the wave,

.such an increase in electron cu-rrentalso increases; the

rate'at whichthe electricfield at thepoint P decreases along the guide v2. As a result, such-an increase in-electron; current. causes electrons at the point P to be ac-- celerated less than the wave, thus impeding the formation ofanelectron bunch at the;point P.

The .foregoing phenomenon is illustrated by the graphs;

of. Figs. 3 and 4, wherein the phase velocity of thewave asa function of-distancealong the waveguide is shown by: the unbroken line. 7 At the electron currentfor which buncher waveguide 2 is designed, the velocities of all theelectrons lie very-close to this line, since theylcluster' about .the point P (Fig. 2) on the wave. Consequently,

the. unbroken linealso shows the velocity of the electrons asafunction of distance along the waveguide, where. the. electron current is that for which the buncher wanfeguidev 2 is designed.

If the electroncurrent is very great, not only ar bunches improperly formed, but the poorly formedbunches eventually. move into a decelerating field andare lost to the walls of-the guide. This is true, at very'high beam currents,- whether or not the inversesfeedbackfcir cuit of the invention is used, as shown. by'the graph of; However, when the inverse-feedback circuit of? the invention is used, the beam disintegrates at a greater Fig. 3.

distance along. the Waveguide than otherwise wouldj'be the case, as clearly appears frorn the graph of Fig. 3.

Hence, for a given waveguide length, the inverse feedback circuit of the invention permits the use of higherbeam' currents Without beam disintegration.

If the electron current is greater than that for which the. buncher Waveguide 2 is designed, but not excessive, and if the inverse-feedback circuit of the invention isfnot used, thevelocities of the electrons will become increas v ingly less than the phase velocity of the wave (until. the phase velocity approaches the velocity oflight), as shown by the dotted line in the graph of'Fig. 4. As previously noted, this impedes'proper bunch formation. Moreover, it is clear from the graphs of Figs. 3-andf4 that increasing the electron current decreases the energyof the electrons, since the dotted line lies bel0w tli'eunbroken line all along the waveguide, except atthefex treme input end. K

similar device.

radio-frequency power beingemployed. The, cl0ckwise distance along the rat-race 9'from arm 5 to arm 6, from arm 6 to arm 7, and from arm 7 to arm 8, is AL. Radiofrequency power of wavelength L is fed into arm .5 from a suitable power source 10, such as a magnetron. Half of this radio-frequency power travels around the rat-race 9 in a clockwise direction and half in a counterclockwise direction. At arm 6, the clockwise-traveling power has traveled AL from arm 5 and the counter clockwise-traveling power has traveled 1%1. from arm 5; the clockwise-traveling power thus arrives at arm 6 in phase with the couuterclockwise-traveling power, and so radio-frequency power is transmitted into arm 6. At arm 8, the clockwise-traveling power is also in phase with the counterclockwise-traveling power, since each has traveled AL from arm 5, so that radio-frequency power is transmitted into arm 8. However, the clockwise-traveling power arrives at arm 7 out of phase with the counter-clockwise-traveling power, since the former has traveled /2L and the latter L from arm 5, and so no radio-frequency power is transmitted into arm 7.

The radio-frequency power which creates the traveling wave in the buncher waveguide 2 is fed into the input end of the waveguide 2 from the waveguide bridge 4 through arm 8. The radio-frequency power remaining at the output end of the waveguide 2 is fed back into the waveguide bridge 4 through arm 7. By suitable design of the waveguide bridge 4, or by providing a phaseshifting device 11 in arm 7, or by any other suitable means, it is possible to pre-set, prior to operation of the microwave linear accelerator, the phase difierence at arm 8 between the radio-requency power arriving from arm 5 and that arriving from arm 7. Regardless of the adjustment of the phase-shifting device 11, or similar preset element, none of the power entering the rat-race 9 through arm 7- can leave the rat-race 9 through arm 5, the reason being the same as that which prohibits the transmission of radio-frequency power from arm 5 to arm 7.

It will be recalled that radio-frequency power entering the rat-race 9 from the power source is transmitted through arms 6 and 8. In the absence of any other power entering the rat-race 9, some fraction of the ratiofrequency power from the power source 10 will be transmitted through arm 6 and the remainder through arm 8. The radio-frequency power fed back into the ratrace 9 through arm 7 divides into a clockwise-traveling component and a counterclockwise-traveling component which are in phase with each other at arms 6 and 8, in accordance with the analysis hereinbefore given with respect to the power entering the rat-race 9 through arm 5. However, by pre-setting the phase-shifting device 11, the radio-frequency power from arm 7 may be caused to arrive at arm 8 with a phase difference, with respect to the radio-frequency power arriving at arm 8 from arm 5, of any desired value. Thus, if the phase- .shifting device 11 is pre-set so that the power from arm 7 arrives at arm 8 180 out of phase with the power from arm 5, the power entering arm 8 will be a minimum, and, in that event, the power from arm 7 will arrive at arm 6 in phase with the power from arm 5, so that the power entering arm 6 will be a maximum. In general, it is preferable to design the waveguide bridge 4 so that the ratio of the waveguide bridge 4 is such that the desired radio-frequency power input is delivered to the buncher waveguide 2. In the case of the rat-race assembly, the bridge ratio is adjusted by varying the cross-section of the rat-race 9 in a suitable manner, as is well-known in the microwave art. Other techniques, well-known in the microwave art, may be employed in the case of magic-Ts and other types of waveguide bridge. Having thus fixed the bridge ratio, the phaseshifting device 11 can be used to maintain this ratio in the event of small frequency changes of the radiofrequency power source 10.

Suppose that the waveguide bridge 4 has been designed 6 so that the power P which remains at the output end of the guide 2 is 10% of the power P' which is delivered to the input end of the guide 2, that the guide 2 is constructed to operate best with a power input of P and an electron current I, and that the guide 2 is operating at an electron current I. Now, suppose that the electron current is increased to 1+1. As previously explained, this tends to reduce the radio-frequency power remaining at the output end of the guide to some value such as P -p However, it will be recalled that P =P-P where P is the power arriving at arm 8 from arm 5 and P is the out-of-phase component of the power arriving at arm 8 from arm 7, P being proportional to P Hence, reducing P reduces P and increases P Now, increasing P increases the electric field .throughout the guide 2, and therefore increases the electron velocity throughout the guide 2. Referring again to the graph of Fig. 4, the dotted line shows the electron velocity without the inverse-feedback circuit of the invention. The elfect of the inverse-feedback circuit of the invention is to increase the electron velocity throughout the guide, as shown by the dashed line. It is not possible, by means of the inverse-feedback circuit of the invention, to cause the electron velocity to equal the phase velocity of the wave throughout the guide. However, this ideal condition is very nearly approximated when the electron-velocity curve first rises above and then falls below the phase-velocity curve in such a manner that the total area under the one curve is equal to that under the other--i.e., the line integral of the electron velocity along the length of the guide equals the line integral of the phase velocity along the length of the guide.

Such an electron-velocity curve is shown by the dashed line in the graph of Fig. 4.

Thus the inverse-feedback circuit of the invention reduces phase error and increases energy stability in microwave devices such as the buncher waveguide 2 of Fig. 1. However, it will be noted that the portion of the radio frequency power delivered by the power source 10 which is not transmitted to the buncher waveguide 2 is transmitted to arm 6 of the waveguide bridge 4. Unless some use can be made of the power in arm 6, reduced phase error and increased energy stability will have been achieved only by sacrificing large amounts of radiofrequency power. Fortunately, in the microwave linear accelerator of Fig. 1, excellent use can be made of the power in arm 6, for it is fed, via a second phase-shifting device 12, into the main accelerating waveguide 3. As previously noted, traveling-wave electron-accelerating devices generally operate best when some fraction, such as 10%, of the power input remains at the output end of the guide. This left-over radio-frequency power is usually absorbed by a resistive load, and the load 13 is provided at the output end of the waveguide 3 for this purpose. Such a load is not needed in the buncher waveguide 2, since the left-over power in that waveguide is fed back, as previously described.

The purpose of the second phaseshifting device 12 is to control the energy of the electron beam delivered by the microwave linear accelerator. Since both phase velocity and electron velocity are very nearly equal to the velocity of light in the main guide 3, all electrons in the accelerating-field part of the wave will remain throughout the guide 3 in the phase at which they entered the guide 3. If the electrons enter at the peak of the wave, they will absorb a maximum amount of energy from the wave; if they enter at zero field on the wave, they will absorb no energy from the wave. Thus, by adjusting the phase-shifting device 12, the amount of energy absorbed from the wave may be adjusted so that the electrons have an energy of any desired value between the energy with which they are injected into the main accelerating waveguide 3 by the buncher waveguide 2 (no energy gain in the main accelerating waveguide 3) and a maximum which is set by the fact that, unless at -of Fig. 5, the minimum energy output at any current output is equal to the energy output of the buncher waveguide 2. Without inverse feedback, this energy decreases with increasing current. With inverse feedback,

this energy output remains substantially constant. Without inverse feedback, increasing the current decreases the radio-frequency power remaining at the output end of the buncher waveguide 2, and when the radio-frequency power left over falls below a certain small value, the buncher waveguide 2 ceases to function. This constitutes the maximum current attainable. With inverse feedback, increasing the current has less elfect on the radio-frequency power left over at the output end of the buncher waveguide 2, since the radio-frequency power input increases so as to handle the increased current. Hence the maximum current attainable is increased to a'large extent. The main accelerating waveguide 3 can handle any current injected into it, no matter how large (in principle), because the second phase-shifting device 12 may always be adjusted to lower the energy'gain in the main accelerating waveguide 3 to compensate for any inc'r'ease in current. The maximum'energy attainable decreaseswithincreasa greater number of electrons. Without inverse back, in addition to the loss of about of the power to main guide 3 which is dissipated in the load l3, thereis an additional loss of about 10% of. the powerinput to the buncher guide 2 which is dissipated in a similar load (not shown). With inverse feedback, the power remaining at the output end of the buncher guide 2 is not dissipated, but is delivered to the main guide 3. Hence the maximum energy attainable with inverse feedback at any given current is always greater than that attainable without inverse feedback at the same current.

The fact that the buncher waveguide 2 bunches satisfactorily only at the current I for which it is designed, results in a greater energy spread in the electron beam without inverse feedback than with inverse feedback at currents different from I This effect is shown by the graph of Fig. 6, wherein the shaded areas show the energy spread in the electron beam as a function of beam current, assuming operation at maximum energy, horizontal shading indicating operation with inverse feedback and vertical shadingindicating operation without inverse feedback. With inverse feedback, the energy spread remains more or less constant at its minimum value. Without inverse feedback, the energy spread deviates from the optimum value by an amount which increases as the optimum current I is departed from.

Having thus described the method of my invention together with a preferred embodiment of apparatus for carrying out the method, it is to be understood that althoughspe cific terms are employed, they are used in a generic and descriptive sense and not for purposes of limitation, the scope of the invention being set forth by the" following claims.

i claim:

I. In combination with a microwave power source and a microwave waveguide structure in which electrons, injected into one end thereof at a veiocity substantially less than that of light, are accelerated by the absorption of energy'from a traveling electromagnetic wave which in g current, since the available power is produced in said microwave waveguide structure by said microwave power source; a magic-T type of waveguide bridge having a first input arm, a second input arm, a first output arm and a second output arm, a first transmission line connected between the power output of said microwave power source and said first input arm, whereby substantially all of the radio-frequency power from said microwave power source issues from said waveguide bridge through said first output arm and said second output arm and not through said second input arm, a second transmission line connected between said first output arm and the input end of said microwave waveguide structure, a third transmission line connected be tween the output end of said microwave waveguide structure and said second input arm, whereby substantially all of the radio-frequency power from said microwave waveguide structure issued from said waveguide bridge through said first output arm and said second output arm and not through said second input arm, the electrical length of said third transmission line being such that at least some of the radio-frequency power from said microwave waveguide structure issues from said first output arm out of phase with the radio-frequency power from said microwave power source issuing from said first output arm, a load, and a fourth transmission line connected between said second output arm and said load.

2. A microwave linear accelerator comprising in combination: a main accelerating waveguide, into which electrons are injected at approximately the velocity of light,

a "buncherwaveguide, into which electrons are injected less the velocity of light and fl fqu which the electrons are injected. into said main acceleratin g wafveguide in the form of a bunched beam, arnicrowave power source, a magic-T type of waveguide bridge having a first input arm, a second input arm, a first output arm and a second output. arm, a first transmission line connected between the power output of said microwave power source and said first input arm, whereby substantially all of the radio-frequency power from said microwave power source issues from said waveguide bridge through said first output arm and said second output arm and not through said second input arm, a second transmission line connected between said first output arm and the input end of said buncher waveguide, a third transmission line connected between the output end of said buncher waveguide and said second input arm, whereby substantially all of the radio-frequency power from said buncher waveguide issues from said waveguide bridge through said first output arm and said second output arm and not through said second input arm, the electrical length of said third transmission line being such that at least some of the radio-frequency power from said buncher waveguide issues from said first output arm out of phase with the radio-frequency power from said microwave power source issuing from said first output arm, and a fourth transmission line connected between said second output arm and the input end of said main accelerating waveguide, the electrical length of said fourth transmission line being such that the radio-frequency power transmitted thereby arrives at said main accelerating waveguide in proper phase relationship with the bunched beam of electrons.

3. Electromagnetic waveguide structure comprising a non-dissipative waveguide bridge assembly having a pair of conjugate" input arms and a pair of output arms, a utilisation waveguide connected to one of said output arms, a feedback waveguide connected between the output of said utilisation waveguide and one of said input arms, ili source of wave energy connected to the other of said input arms and power dissipative means connected to the other of said output arms, the output arm connected to said utilisation waveguide being positioned at a point on said bridge at which at least a portion of the inputs from said input arms cancel one another, the output arm connected to said dissipative means being positioned at a point on said bridge at which the inputs from said input arms cancel one another in the steady state condition.

References Cited in the file of this patent UNITED STATES PATENTS Tiley Mar. 3, 1953 Robertson-Shersby-Harvie- Dec. 28, 1954 Arams Nov. 13, 1956 Miller et a1. Oct. 22, 1957 Chodorow Nov. 19, 1957

Images