US3519942A

US3519942A - Apparatus for providing short bunches of charged molecular,atomic,or nuclear particles

Info

Publication number: US3519942A
Application number: US542361A
Authority: US
Inventors: Ralph C Mobley
Original assignee: RALPH C MOBLEY
Current assignee: RALPH C MOBLEY
Priority date: 1966-04-13
Filing date: 1966-04-13
Publication date: 1970-07-07
Anticipated expiration: 1987-07-07

Description

3,519,942 GED July 7, 1970 R. c. MOBLEY APPARATUS FOR PROVIDING SHORT BUNCHES OF CHAR MOLECULAR, ATOMIC, OR NUCLEAR PARTICLES Filed April 13. 1966 3 Sheets-Sheet l OW) mmJDl NN WN mumnom zo @WWW/.Mm

July 7, 1970 R. c. MOBLEY 3,519,942

APPARATUS FOR PROVIDING SHORT BUNCHES OF CHARGED MOLECULAR, ATOMIC, OR NUCLEAR PARTICLES Filed April 15. 196e sheets-sheet 2 l K35 Q \\7//////,

I I l i 1| l SWW/

5H

52,53 iB-54 35 C- s -y---ETQ i u' ON l l l I I PATH l l :d n l a i 5o l 5o ///M/ INVENTOR. 4 RALPH C. MOBLEY BY GRAY, MASE 8.DUNSON ATTORN EYS BWQLWI. www

July 7, i970 R c. MOBLEY 3,519,942

APPARATUS FOR PROVIDING SHORT BUNCHES OF CHARGED MOLECULAR, ATOMIC, OR NUCLEAR PARTICLES Filed April l5, 1966 3 Sheets-Sheet 5 1 5 DRIFT TUBE [L \\i ////////////////4 r/ Gl G2 7 |ON PATH LMTS I ENS l- ENs \v LENS l POTENTIAL ON GRID 82) R F SIGNAL SOURCE INVENTOR.

RALPH c. MQBLEY (ALL GROUNOS ARE LABORATORY GROUND) BY GRAY, MASE 8. DUNSON ATTORN EYS United States Patent O 3,519,942 APPARATUS FOR PROVIDING SHORT BUNCHES F CHARGED MOLECULAR, ATOMIC, OR NU- CLEAR PARTICLES Ralph C. Mobley, 2585 Stoodleigh Drive, Rochester, Mich. 48063 Filed Apr. 13, 1966, Ser. No. 542,361 Int. Cl. H013' 23/00, 23/34 U.S. Cl. 328-233 26 Claims ABSTRACT OF THE DISCLOSURE lons from a source 8 (FIG. 1) are focused `by a lens 9, fed at low velocity in a column 10 into the region 1, and al1 are accelerated momentarily to high velocity therein. Decelerating the ions as they enter the region 2 greatly foreshortens the column. Similar momentary acceleration of the entire column while in the region 3 and deceleration of the ions upon entering the region 4 foreshorten the column 10 still more. Further deceleration in the

regions

5 and 6 provides even more foreshortening. The column is momentarily accelerated in the region 6 just before it enters the region 7, where the same rate of acceleration is maintained to add an effective time bunching of the column on the target 32.

This invention relates to charged particle bunchers. It has to do particularly with apparatus for providing short bunches of charged molecular, atomic, or nuclear particles such as ions, electrons, positrons, and the like.

Apparatus according to the present invention is especially useful for providing a very high bunching ratio with ions, while utilizing a substantially higher proportion of the ions supplied to the apparatus than has been possible with prior devices. The apparatus can also -be made much smaller and lighter than the presently known types of ion bunchers.

Typical apparatus according to this invention for providing short bunches of charged particles may comprise:

(a) Means for providingsaid particles at low velocity to a rst region so as to accumulate and form a column therein of substantial length;

(b) Means for momentarily accelerating all said partices in said column substantially equally to higher velocity in said rst region; and

(c) Means for decelerating each said particle to lower velocity as it enters a second region, and thus to foreshorten said column therein to a small fraction of its initial length.

Such apparatus may comprise also:

(d) Means for momentarily accelerating all said particles in the foreshortened column substantially equally to higher velocity in a given region; and

(e) Means for decelerating each said particle to lower velocity as it enters a subsequent region, and thus to further foreshorten said column to a small fraction of the length it had just before entering said subsequent region.

Apparatus of this invention may comprise also:

(f Means for accelerating all said particles in the further foreshortened column substantially equally to higher velocity in a given region, immediately before they enter a subsequent region; and

(g) Means for continuing to accelerate all said particles in said subsequent region at the same rate of acceleration as that provided by the means (f), and thus to provide an effective time bunching of said column on a target beyond said subsequent region substantially proportional to the increase in velocity of said column between said given region and said target.

ice

The apparatus may also include other preferred and alternative features as disclosed and claimed herein.

In the drawings:

FIG. 1 is a largely schematic and partially sectional view of typical apparatus and waveforms therein according to the present invention.

FIG. 2 is a largely schematic sectional view of typical electrodes and associated structure according to this invention.

FIG. 3 is a largely schematic sectional view of other typical electrodes and associated structure and components according to the invention.

FIG. 4 is a largely schematic and partially sectional view of other typical apparatus according to the invention.

FIG. 5 is a largely schematic and partially sectional view of other typical apparatus and waveforms therein according to the invention.

FIG. 6 is a view similar to FIG. 5 of still other such typical apparatus and waveforms therein.

FIG. 7 is a schematic diagram of a type of modification that can be made in the circuitry in FIG. l.

FIG. 8 is a schematic diagram of another type of modication that can be made in the circuitry of FIG. 1.

FIG. l shows a typical embodiment of the invention as designed for bunching positive deuterium ions. Some typical dimensions and electrical quantities are included for convenience to the reader, but the gure is largely schematic and is not drawn to scale. Positive deuterium ions are provided at high velocity with an energy of about 5 kilo electron volts (kev.) from an ion source 8, such as a Moak type radio frequency ion source (described in Nucleonics 9 no. 9, 18 (1951) by Moak, Reese and Good., or see Fast Neutron Physics, Part I., by Marion and Fowler-Interscience Publishers 1960pages 535-538). The ions are focused by a lens 9 into a large diameter parallel beam 10, and are decelerated to a lower velocity with an energy of about electron volts (ev.) before y entering an equipotential region 1 through a grid 11. This would normally be accomplished by operating the ion source 8 so that the electrode (probe canal in an R.F. ion source) out of which the positive ions emerge from the source is at a suflciently negative potential that when the ions reach the ground potential of the grid 11 they have only 100 ev. left. For example, -1900 volts for positive emerging ions with a kinetic energy of 2000 ev. (For negative ions the polarity of the retarding potential would be reversed.) The ions thus enter the region 1 through grid 11 at a relatively low energy and hence low velocity.

For about 7/s microsecond ions are allowed to accumulate, partially filling the region 1 and forming a column 10 of monoenergetic ions about 8.6 centimeters long. A 1A; microsecond 13 kilovolt positive pulse is then applied from a pulse generator 21 through a capacitor 22 to the grid 11, creating a substantially uniform electric field between the

grids

11 and 12 and causing all ions in the region 1 to accelerate substantially equally. Just before the front of the column 10 reaches the grid 12 the pulse on the grid 11 terminates. The region 1 returns to its normal equipotential condition, with a resistor 23 and a diode 23' maintaining the

grids

11 and 12 at substantially the same potential in the intervals between pulses. All ions now move at the same constant higher velocity with an energy of about 2500 electron volts, and the length of the column or bunch 10 remains constant at about 8.6 cm.

The ions exit from the region 1 through the grid 12 and enter a drift and deceleration region 2. In the region 2 the i-ons are decelerated individually to about 100 ev. by the 2400 volt potential difference between the

grids

12 and 13 which is provided by a direct voltage supply 24 the negative terminal of which is connected to the terminal ground 25, to which the grid 12 also is connected. The

positive terminal of the power supply 24 is connected to a grid 14 and through a resistor 26 and a diode 26 to the grid 13. The reduction in kinetic energy by 25 reduces the ion velocity by and causes a foreshortening of the bunch as it accumulates in the region 3 by a factor of 5. The length of the 100 ev. ion column or bunch 10 in the region 3 is then about 1.7 cm. The 11 cm. spacing between the

grids

12 and 13 is also chosen to delay the entrance of the bunch 10 into the region 3 until just before the arrival of the next pulse provided from the pulse generator 21 through a capacitor 27 to the grid 13. The region 2 between the

grids

12 and 13 then serves both to decelerate the ions and to delay the bunch 10 for proper phasing upon entering the region 3.

The uniform acceleration -of all the ions in the region 3 to about 6300 ev. is followed by separate or individual deceleration in the region 4 to an energy of about 100 ev. causing an additional bunching of about 7.9. The proper retarding potential is furnished by a direct voltage supply 28 through a resistance 29 to the grid 15. A resistancer30 is connected at one end to the resistance 29 and the voltage supply 28 and at the opposite end to the grid 16. A diode 30 is connected in parallel with the resistor 30. Thus the grids and 16 are maintained at substantially the same potential, in the intervals between pulses, to provide an equipotential drift region 5 to achieve proper phasing of the bunch 10 as it enters the region 6.

A third stage of acceleration is provided in the region 6 by another positive voltage pulse provided by the generator 21 through a capacitor 31 to the grid 16. The twice foreshortened column 10 enters the accelerator tube 7 at the peak of the pulse on the grid 16, at which time the gradient in the region 6 matches that of the accelerator tube 7, making possible an additional effective time bunching 4of about 55 for an accelerator potential of 300,000 volts (70 for 500,000 volts). This effective further biinching comes about because the gradient in the region 6 is chosen to match that of the accelerator tube 7 as the bunch 10 leaves the region 6. The bunch length of 0.2 2 cm., therefore, does not change as the bunch proceeds down the accelerator tube 7. Arrival time on the target 32, however, decreases in proportion to the ratio of final to initial velocity. For an initial ion energy of 100 ev'.r in the region 6 and a final ion energy of 300,000 ev. .upon exiting from the accelerator tube 7, the velocity increases by a factor of 300,000/ 100=54.7 with an effective time bunching to this same extent. The overall bunching then is 5 7.9 54.7=2l60. For an initial bunch of 7A microsecond, the nal burst length on the target 32 is about 7A: microsecond/2l60=0.4l nanosecond. With a pulse :repeltition rate of l mc./sec., ideally 87 percent of ion source output would be bunched. y

This embodiment of the invention, therefore, comprises a light compact buncher of very high bun'ching ratio and of very high ion source utilization eiciency, suitable for installation in a wide variety of accelerators and tandem injectors. Variation of the initial ion energy and pulse height make a given buncher also adaptable to a wide range of ion masses.

In the above explanation the effects of a number of factors have been omitted for simplicity. They are: energy spread in the ions from the ion source, space charge effects, energy spread introduced by the third bunching stage, and ion losses to the various grids.

The energy spread of the ion source 8 tends to be magnied in each stage by the bunching ratio of that stage. Because of the small extent (2.2 mm.) of the foreshortened bunch 10 by the time it reaches the region 5, ions of slightly different energies in the region 1 will have become separated physically in the region 5 with the higher energy ions being closer to the grid 16 and the lower energy ions closer to the grid 15. Application of a small square wave positive pulse to the grid 15 from a pulse generator 33 through a capacitor 34 after all the ions are in the region 5, and maintenance of the pulse until all the ions have exited through the grid 16 into the region 6, gives more energy to the low energy ions than to the high energy ions. It is thus possible to compensate, at least to first order, for the ion source energy spread and at the same time, to some extent, for the effects of space charge spreading. The use of a large diameter parallel ion beam throughout the buncher also minimizes space charge effects. Because the ions are subjected to a large electrical gradient after they have entered the region 6 and because, unlike the preceding stages, this gradient is maintained until the ions exit from the accelerator tube 7, the ions at the rear of the bunch 10 fall through a larger potential difference than the ions in the front of the bunch by an amount equal to the gradient in the region 6 times the length of the bunch 1G therein. In the above example this potential difference is 0.22 cm. 4.65 kv./crn.=1.02 kv.

Because the energy spread is linearly distributed along the bunch it can be removed if necessary as the ions exit from the accelerator tube by a small amount of inverse velocity modulation at that point. This can lbe accomplished in at least two different ways, either lby the inverse of klystron velocity modulation or by the inverse of the bunching action of region 6 of FIG. 1.

The first technique is illustrated in FIG. 5 for the parameters of the previous example, i.e., 300 kev. deuterons which would have a velocity of 5.4 108 cm./sec. The ion bunch 10 emerging from the accelerator tube 7 is focused by a lens 71 (such as a stron-g focusing electrostic or magnetic quadrupole pair) to a focus at the center of the gap G1. A lens 72 refocuses the bunch at the center of the gap G2 and finally a lens 73 focuses the bunch 10 on the target 32, which is not shown in FIG. 5. The bunch 10 also is not shown in FIG. 5, only the envelope of its path as dashed lines. The drift tube 74 is driven by a radio frequency signal source which for example could have as one possible combination an amplitude of 0.51 kv. at a frequency of 405 mc./sec. This voltage appearing across the gaps G1 and G2, if properly phased with the passing lbunch, will remove the 1.02 kv. energy spread introduced as the bunch 10 emerges from the accelerator tube 7.

This comes about in the following manner. If the gaps G1 and G2 are small (say 1 mm., which is a practicable separation for smoothly polished and rounded surfaces in a good vacuum at the voltage and frequency mentioned), an ion traversing a gap will acquire or lose an amount of energy equal to the average potential across the gap during its passage across the gap. If the gap transit time is limited to the 27 of the R.F. cycle in this example or less, then if the ion bunch passage is timed to traverse the gap during the essentially linear portion of the R.F. cycle such as from a to Iz for G1 and from c to for G2 (FIG. 5), then the average potential for any given ion as it traverses the gap is essentially that which exists across the gap as the ion crosses the center plane of the gap. For 300 kev. deuteron ions which travel at approximately 5.4 108 cm./ sec., the time t required to traverse the center plane of the gap would be t=extent of bunch/v=0.22 cm./5.4 :m./sec.=4.1 10-10 sec. If the energy spread of the ion bunch is 1.02 kev., as in the example, then if each gap removes half this spread, the voltage across each gap will have to change by 510 Volts as the ions in the bunch cross the center plane of the gap- Considering the rst gap G1, and remembering that thtx ions at the front of the bunch 10 are lower in energy than those at the Iback by 1.02 kev., then the leading ions ought to see -255 volts on the drift tube 74 (point a on the R.F. cycle in FIG. 5) thus gaining approximately this energy, while those at the rear of the bunch 10 ought to see +255 volts on the drift tube 74 (point b), thus losing this amount of energy. The difference between points a and b then being 510 volts. If this swing occurs between 0=30 and +30 of the sine Wave on the drift tube 74 as the bunch 10 passes through the gap G1, then this corresponds to 1A; of the R.F. cycle and since this must also be the bunch transit time through the center plane of the gap G1 or 4.1 1010 sec., then the R.F. period T =6 4.1 10r1 sec.=2.46 109 sec. and the frequency =T1=405 megacycles. Also, since sin 0:1/2, the amplitude of the R.F. voltage would be 2X255=5l0 volts. Other combinatons are also possible, for instance, the transit time 4.1 10-10 sec. might be made equal to 1/12 cycle, whence the frequency would be 202.5 mc. and the amplitude approximately twice or 1.02 kv. or 101.25 mc. and 2.04 kv., etc.

Choosing the initial combination, 405 mc., and 510 volts for further calculation, it is now necessary to remove the other 510 ev. energy spread from the bunch as it emerges through the gap G2. For this to occur it is necessary for the drift tube voltage to swing from c to d or from +255 volts to -255 volts as the bunch 10 crosses the center plane of the gap G2. This shift from the front side to the back side ofthe R.F. cycle in going from the gap G1 to the gap G2 is necessary because the direction of the electric field is opposite in the two gaps for a given voltage polarity on the drift tube 74. To assure that this be the case, the ion bunch 10 must spend (n+1/2) cycles in traversing the distance L between the centers of the two gaps, where n=0, 1, 2, etc., i.e., a whole number. Now in (n+1/2) cycles, the ions will travel a distance L=(n-I1/2) TX v=(n'l1/2) 2.4 109 sec. 5.4 l08 cm./sec.=(n+1/2) 1.32 crn. If we let n=20, then L=27 cm., a quite reasonable value.

The reason for the three lenses in FIG. 5 is to permit the gaps G1 and G2 and their aperture to be as small as possible. Also, since these gaps will generate quite strong time varying electrostatic lenses, a beam crossover in the center of such lenses will nullify this effect. This can be readily seen from an optical analogy.

An alternative possibility is to form the gaps by two very closely spaced (l mm.) grids. The lenses would not then be necessary. Whether large grids with such small spacing (1 mm.) between grids and the much closer spacing of wires in a given grid (perhaps 0.1 mm. or 0.1 of the gap) needed to make the grids sufficiently smooth equipotential planes for this gap spacing are practical either physically or electrically is not clear. The arrangement in FIG. 5 appears to be more practical.

With this particular arrangement the length L is related to the velocity v of the ions and hence to their energy and to the number of R.F. cycles n the ions spend in the drift tube. For a fixed buncher frequency (the most convenient operating condition technically at present) and a fixed drift tube length L, the velocity of the ions and hence their energy could be changed only in discrete steps corresponding to changes in the value of n by l. In the previous example, this Would be a 4 percent change in velocity or 8 percent change in energy. Alternatively keeping n constant and changing the frequency from 405 mc. (405th harmonic of the buncher frequency) to 406 mc. (406th harmonic) would correspond to a 0.25 percent change in velocity and hence a 0.5 percent change in possible ion energy. Varying both n and f gives many more possibilities, but all corresponding to discrete ion energies. Modifcation of the buncher oscillator to allow a small change in frequency (1 percent would be more than enough) or construction of the drift tube so that L could be varied by l0 percent would make it possible to handle ions of any energy throughout the design range of this energy demodulator.

The sine wave technique for prevention of velocity modulation of the ion bunch upon emerging from the accelerator is essentially that of region 6 of the buncher (FIG. 1) in reverse with a sine wave substituting for the pulse. FIG. 6 shows a way in which this technique can be used to prevenlt the 1.02 kev. energy modulation in the ion bunch. A first gridded electrode 81 is grounded to the laboratory ground while a second gridded electrode 82 is driven with a sine wave from the R.F. signal generator 83. A diode 84 serves to bias the grid 82 so that its potential only swings down from zero as in the adjacent graph in FIG. 6. The signal generator 83 is so phased with respect to the arrival time of the ion bunches 10, that the bunches 10 enter region y80 (between grids 81 and 82) when the voltage on the grid 82 is at e, the bottom of its swing, and exit from the region through the grid 82 when the voltage on it is zero (point f). Now if the gradient in the region 80 matches that of the accelerator tube 7 when the bunch 10 enters and drops to zero just as it leaves, then all ions in the bunch 10 are subject to the same electrostatic forces at all times and will therefore all have the same velocity upon emerging into the region 90.

This technique does not remove velocity modulation, it doesnt allow it to start in the first place. As a specific example, let the frequency of the R.F. signal source be 10() megacycles/sec. Clearly it is necessary that this frequency be some integral multiple of the buncher frequency of l megacycle/ sec. if proper phasing is to be maintained between the signal source and bunch entrance into the region 80. For f= mc./ sec. the bunch will move 5.4 10S cm./sec.

#103 cycles/see. :2'7 cm' itx during one-half cycle of the voltage on the grid 82. The separation s between the grids 81 and 82 then needs to be 2.7 cm. In order to match the accelerator tube gradient of 4.65 kv./ cm., the peak to peak R.F. voltage would have to be 2.7 cm. 4.65 kv./cm.=l2.5 kv. A peak R.F. voltage of 6.25 kv. would therefore be needed. In arriving at this result it is assumed that the electric field between the grids 81 and 82 is uniform.

The entire buncher and associated system from the ion source y8 through the accelerator tube 7 in FIG. 1 are all operated under high vacuum (10-5 mm. of mercury or better) maintained by high vacuum pumps (diffusion, getter, etc., not shown). The insulating spacers 35 between the electrodes 11-417 normally are used both for mechanical support and as part of the wall `36 for maintaining the vacuum in the system. The accelerator tube 7 is similarly constructed with insulating sections 37 between the electrodes. Resistances 38 are connected from each electrode 39 to the next, to divide the potential evenly from the high voltage terminal at 25 to the laboratory ground 40 of the accelerator tube 7.

A preferred form of the pulse generator 21 is disclosed and claimed in the copending United States patent application of the present inventor, Ser. No. 663,820, filed Aug. 28, 1967. Part of the circuit is shown schematically in FIG. l. The circuit is basically a push-pull amplifier with a high Q delay line load. The amplifier tubes are driven alternately to cut off by short negative pulses, the drive to the two tubes being periodic at a repetition rate of l mc./ sec., chosen to match the transit time down the delay line and back. A pulse traveling along the line is in a sense resonantly reinforced by first one tube and then the other as it reaches each end of the line and is reflected. Since the output voltage pulse is substantially sinusoidal in waveform and is delivered from the generator 21 during only one-eighth of each cycle, the B+ voltage from the power supply 41 theoretically could be as low as one-sixteenth of the voltage at the peak of the output pulse. Because of losses, the supply voltage must be higher than that, but still only a small fraction of the pulse voltage.

The waveform and phase of the pulses at the grid 13 are shown in the upper curve 42 at the bottom of FIG. l, while the waveform and phase of the pulses at the

grids

11 and 16 are shown on the same time axis in the lower curve 43. Of course the output over a complete cycle is not sinusoidal. This condition for this particular pulse generator is caused by the cutoff characteristics of the delay line. The shape of the pulse is not important as long as it is short enough and has the proper average value. An 8 7 kv. square pulse would do equally well. Since the condenser 22 blocks D-C, and since the resistor 23 is tied from the grid 11 to the terminal ground 25, the average Voltage on the grid 11 is zero. Since the pulse generator 21 forces the grid 11 to rise 13 kv. during 1/s microsec. in the example (average value of the pulse is 8 kv. and is thus equivalent to a square pulse of this height), then for the potential to average zero it would actually swing up to +l2,000 volts, then down to -1,000 volts forthe remaining %p. sec. for a total swing of 13 kv. For best operation the grid y11 should swing down to, but not below the grid 12 or zero. Similarly for the

grids

13 and 14 and the

grids

16 and 17, respectively. This can be accomplished in either of two ways.

A high voltage diode, preferably a vacuum diode, can be placed across each

resistor

23, 26, and as shown in FIG. 1. This .is probably the simplest solution. It can be arranged to hold the grid 11 within a few volts of terminal ground. By connecting the anode of the diode 23 to a slightly positive adjustable voltage with respect to terminal ground, as at 23" in FIG. 7, the grid 11 can be made to assume an average potential very close to zero between pulses. Similarly for the

grids

13, 14, and 15, 16. With capacitance 22 of 1000 paf. or more, the voltage variation between pulses is 1.5 volts or less, which is an acceptable value.

An alternative is to insert a power supply of +1000 volts in series with each

resistor

23, 26, and 30, as shown at 23 in FIG. 8 for the resistor 23. Here the grid 11 has an average potential of +1000 volts swinging up to +13 kv. and down to zero. Again with a capacitance 22 of 1000 gaf. or more, the time constant of the circuit is 1 millisecond or more and the voltage change on the grid 11 between pulses varies only from 0.7 to +0.7 volt as a maximum, again a very acceptable value.

The invention is applicable generally to charged particles of any type, such as electrons, positions, etc., as well as ions. The configuration shown in FIG. 1 illustrates the bunching principle involved using deuteron ions as the charged particles. This is a form of ion buncher of great practical value to nuclear physicists. Changing the pulse height and initial particle energy on the entering grid 11 makes it possible to bunch (at separate times) particles of widely different masses. Going from deuteron ions to electrons or positions requires such a large change in the parameters, however, that a change in dimensions of the buncher is advisable.

The grids 11-17 preferably comprise parallel wires carefully aligned optically from grid to grid along the ion paths in order to make all the wires in each grid following the grid 11 fall in the shadow of the corresponding wires in the preceding grid. Grids are advantageous to assure the uniform fields between electrodes needed for the bunching process.

A truly uniform field can be had between plane electrodes only if they are infinite in extent. Such electrodes preferably should extend sufficiently far from the ion path radially to assure an essentially uniform field. This requires electrode diameters of several times the separation between electrodes, which would be rather unwieldly.

It is preferable to thicken the electrode as the radial distance from the beam path increases, as shown schematically in FIG. 2. This places the opposed

inner surfaces

44 and 45 of the electrodes 11' and 12 somewhat closer together near their edges to compensate for the tendency of the field to fall off with distance from the center of fiat electrodes of finite extent. The insulator also serves as part of the vacuum wall needed for the buncher, which operates under the relatively high vacuum characteristic of accelerators. Reducing the electrode size in this way has, beyond the obvious one of compactness, which is very important, the additional advantage of reducing nterelectrode capacity and the loading it might throw on the pulse generator that drives it.

Another way to attain a uniform field is to use flat electrodes with a poor conductor in place of the insulator. An equivalent alternative way is to sub-divide the region 3 with additional ring electrodes 46 and with resistors 47 Vbetween them as shown in FIG. 3. Both of these latter arrangements suffer from the large amounts of power that would be dissipated in the resistors 47 or the poor conductor. One might use condensers rather than resistors, but there would then be large amounts of reactive current. Either would heavily load the pulse generator. The arrangement in FIG. 2 is therefore generally preferred, but other configuration may be useful for some purposes.

Besides helping to form a uniform field in the region 1 (FIG. l) during the first phase of bunching, the grid of the electrode 11 also serves the very important function of preventing ions from the source 8 from entering the region 1 during the pulse (by repelling them back toward the source). This chopping action, if not performed by the grid 11, would have to be (and could be) performed by some other means to the left of the region 1. If additional ions were to enter the region 1 while the pulse is on the grid 11, the additional ions would acquire variable amounts of energy and therefore would not bunch properly. Ideally Ms of the ions are bunched and 1/8 thrown away in the example given. Bunching Vs of the ions from the source is a superior achievement not approached by any other known bunching apparatus.

The presence of the grids in the electrodes 11-17 in FIG. 1 causes some loss of ions traversing the buncher. Except possibly for the grid in the electrode 11 Iwith its chopping action, all the other grids in the buncher could probably be eliminated by increasing the spacing between the electrodes considerably, and applying pulses only when the bunch of ions 10 was far enough from the electrodes both at the beginning and the end of the pulse to be in a region of essentially uniform field during the' pulse. Both the acceleration regions such as 1 and the deceleration regions such as 2 could probably be made minimal in length through the arrangement of FIG. 3. In any case a buncher with such electrodes would tend to be longer than that of FIG. 1 and the transition from acceleration to deceleration regions would be accompanied by strong electrostatic lens action on the ion bunch 10 because of the curvature in the equipotential surfaces produced by the change in potential between regions not having grids between them. The lens effect might be adequately offset by an axial (solenoidal) magnetic field directed along the beam axis of the buncher or by the stronger focusing action provided by ring magnets arranged coaxially with the beam poth with their magnetic field directions alternating axially from magnet to magnet (as in the permanent magnet focusing utilized in traveling wave tubes).

Another, but probably less satisfactory, solution would be to use quadrupole focusing magnets distributed along the beam path as with the ring magnets mentioned above. Quadrupole focusing magnets are widely used with accelerators in nuclear physics to focus beams.

When a completely gridless buncher can be successfully built is not clear. If all grids cannot be. eliminated, however, some very likely can be. It then becomes a question of whether the added complexity of the compensation techniques outweighs the added efiiciency obtained without grids.

True physical bunching in the apparatus of FIG. 1 occurs only in the regions 1 through 4. In the region 6 and the accelerator tube 7 the process is quite different. Here the bunch 10 does not get shorter physically as it moves from the region 6 into and down the accelerator tube 7, but remains the same physical length and merely moves faster. The effect in terms of arrival time of the ion bunch 10 on the target 32 at the output of the accelerator 7, however, is the same. The faster it goes, the shorter the arrival time. This is not Klystron bunching. Klystron bunching relies on velocity modulation, which does not exist here at all until the bunch 10 emerges from the accelerator tube 7 and then what does come about isl not appreciable.

Where a bunch of particles is sufficiently short, as in the region 5, further bunching can be accomplished using ordinary sine waves instead of pulses. A means for doing that is shown in FIG. 4.

Electrodes

51 and 53 are lgrounded. The electrode 52 is driven by a sine wave oscillator 60 through a coupling capacitor 61, and is biased by a diode 62, so that it swings only positive from ground. The particle bunch 50, being short, is admitted during the time the electrode 52 is at ground potential. With the frequency and voltage of the sine wave and the spacing of the

electrodes

52 and 53 properly chosen, the bunch 50 is accelerated from the electrode 52` toward the electrode 53 and arrives at the electrode 53 just as the electrode 52 swings back to ground potential. The bunch 50 then exits from the region A when the field in this region is zero.

Deceleration in the fixed potential retarding region B then causes bunching as in the region 2 of FIG. 1. With the spacing of the

electrodes

53 and 54 in the region B properly chosen the bunch exits from the region B through the

electrodes

54 and 55 into the region C when the electrode 55 is at its minimum potential, which is equal to the potential of the electrode 54. At the moment the particle bunch 50 enters, the region C has a potential equal to that of the voltage of the bias supply 63 as needed to decelerate the ions in the region B to the velocity they had on entering the region A. The process is then repeated with the region C playing the part of the region A and the following region that of the region B, etc. A diode 64 biases the electrode 55 so that it swings only positive with respect to the bias supply 63 and hence also with respect to the electrodes S4 and 56. Coupling capacitors 65 and 66 connect the alternating voltage source 60 to the electrode 55 and subsequent electrodes and a capacitor 67 bypasses alternating current around the bias supply 63. The bunching in the present apparatus is clearly different from that which occurs in a cyclotron or a linear accelerator. Inthe latter devices all regions or sections of the Iaccelerator are at the same average D-C electrical potential (zero) with acceleration and so-called phase `bunching occurring under the action of R-F fields, only, which at the same time produce progressively increasing velocities. In the present sine-wave buncher the average D-C potential is increased from stage to stage by providing a higher bias voltage on each successive stage (as in FIG. l), while the average velocity throughout the buncher remains constant. This form of the buncher may be especially suitable for electron bunching. Also, a couple of stages of this type Imight profitably be used following the region 5 of FIG. 1, where the ion bunch is very short.

Of course, voltages having polarities opposite to those shown in the drawings can be applied to the electrodes at the other ends of the various regions to provide the same actions in those regions. For example, in FIG. 1, negative pulses can be applied to the

electrodes

12, 14 and 17, instead of positive pulses to the

electrodes

11, 13, and 16, to accelerate the column 10 in the

regions

1, 3, and 6, respectively.

While the forms of the invention herein disclosed constitute presently preferred embodiments, many others are possible. It is not intended herein to mention all of the possible equivalent forms or ramifications of the invention. It is to be understood that the terms used herein are merely descriptive rather than limiting, and that various changes may be made without departing from the spirit or scope of the invention.

What is claimed is:

1. Apparatus for providing short bunches of charged particles comprising:

(a) means for providing said particles at low velocity to a first region so as to accumulate and form a column therein of substantial length;

(b) means for momentarily accelerating all said particles in said column substantially equally to higher velocity in said first region;

(c) means for decelerating each said particle to lower velocity as it enters a second region, and thus to foreshorten said column therein to a small fraction of its initial length;

(d) means for momentarily accelerating all said particles in the foreshortened column substantially equally to higher velocity in a given region;

(e) means for decelerating each said particle to lower velocity as it enters a subsequent region, and thus to further foreshorten said column to a small fraction of the length it had just before entering said subsequent region;

(f) means for accelerating all said particles in the further foreshortened column substantially equally to higher velocity in a given region immediately before they enter a subsequent region; and

2. Apparatus for providing short bunches of charged molecular, atomic, or nuclear particles comprising:

(a) means for providing said particles at LOW VE- LOCITY in a parallel beam of large diameter through an opening in a first electrode in said apparatus into a FIRST REGION, lbetween said first electrode and a second electrode spaced therefrom in the direction of movement of said beam;

(b) means for maintaining said first and second electrodes normally at substantially EQUAL POTEN- TIAL to permit said particles to accumulate and form within said first region a column of said particles at least about one-half as long as said first region;

(c) means for applying to said first electrode a large electrical AIDING PULSE to create a substantially uniform electric field between said first and second electrodes when said column has been formed and thus to accelerate all said particles substantially equally, said pulse ending just before the forward end of said column reaches said second electrode, so that substantially all said particles move through an opening in said. second electrode at substantially equal and constant higher velocity, and with the length of said column remaining substantially constant, into a SECOND REGION, between said second electrode and a third electrode spaced therefrom in the direction of movement of said particles; and

(d) means for providing a RETARDING POTEN- TIAL DIFFERENCE between said third and second electrodes to decelerate said particles to lower velocity, and thus to foreshorten said column to a small fraction of its initial length, as said column moves through said second region and passes through an opening in said third electrode into a THIRD REGION, between said third electrode and a fourth electrode spaced therefrom in the direction of movement of said particles.

3. Apparatus as in claim 2, comprising also:

(e) means for maintaining said third and fourth electrodes at substantially EQUAL POTENTIAL as said column enters said third region;

(f) means for applying to said third electrode a large electrical AIDING PULSE to create a substantially uniform electric field between said third and fourth electrodes when said column has passed through said third electrode and thus to accelerate all said particles substantially equally, said pulse ending just before the forward end of said column of particles reaches said fourth electrode, so that substantially all said particles move through an opening in said fourth electrode at substantially equal and constant higher velocity, and with the length of said column remaining substantially constant, into a FOURTH REGION, between said fourth electrode and a fifth electrode spaced therefrom in the direction of movement of said particles;

(g) means for providing a RETARDING POTEN- TIAL DIFFERENCE between said fifth and fourth electrodes to decelerate,said particles to lower velocity, and thus to further foreshorten said column as it moves through an opening in said fifth electrode into a FIFTH REGION, between said fifth electrode and a sixth electrode spaced therefrom in the direction of movement of said particles; and

(h) means for maintaining said fifth and sixth electrodes at substantially EQUAL POTENTIAL to continue to foreshorten said column as it enters said iifth region and to maintain it at a constant length as it moves further through said fifth region and through an opening in said sixth electrode into a SIXTH REGION between said sixth electrode and a seventh electrode spaced therefrom in the direction of movement of said particles.

4. Apparatus as in claim 3, comprising also:

(i) means for maintaining said sixth and seventh electrodes at substantially EQUAL POTENTIAL as said column enters said sixth region; and

(j) means for applying to said sixth electrode a large electrical AIDING PULSE to create a substantially uniform electric `ield between said sixth and seventh electrodes when said column has passed through said sixth electrode and thus to accelerate all said particles substantially equally, so that when said pulse reaches its peak said column moves through an opening in said seventh electrode into an ACCELERA- TING DEVICE having a potential gradient in the direction of movement of said particles substantially EQUAL to the POTENTIAL GRADIENT present between said sixth and seventh electrodes when said pulse is at its peak, the matching of said potential gradients maintaining the length of said column substantially constant and providing an effective time bunching of said column on a target substantially proportional to the increase in velocity of said column of particles between said fifth region and said target. S. Apparatus as in claim 2, wherein a GRID of parallel wires is provided across the opening in said first electrode.

y6. Apparatus as in claim 4, wherein a similar grid of parallel wires is provided across the opening in each said electrode, the corresponding Wires of all said grids being parallel and in register as viewed in the direction of movement of said particles.

7. Apparatus as in claim 3, wherein means are provided for applying a SMALL electrical AIDING PULSE to said fifth electrode when said column has passed through said fifth electrode and for maintaining said pulse until said column has passed through said sixth electrode into said sixth region; thus to substantially COMPENSATE for any spreading theretofore, between particles having different energies, by imparting more energy to the particles having lower energy than to those having higher energy.

8. Apparatus as in claim 7, wherein said pulse has substantially constant amplitude.

9. Apparatus as in claim 4, comprising also means for providing inverse velocity modulation in said column after it leaves said accelerating device, to substantially eliminate any differences in velocity between the particles in said column.

10. Apparatus as in claim 9, wherein said means for providing inverse velocity modulation comprises means for focusing the particles in said column to a point; and means for providing, in a region having said focus point substantially at its center, an electric yfield that varies substantially linearly from a predetermined accelerating potential gradient, as said column enters said region, to a predetermined decelerating potential gradient, as said column leaves said region.

11. Apparatus as in claim 10, comprising also at least one further combination of said focusing means and said electric field providing means.

12. Apparatus as in clairn 4, comprising also means for substantially eliminating any differences in velocity between the particles in said column as saidY column leaves said accelerating device.

13. Apparatus as in claim 12, wherein said difference eliminating means comprises means for providing, in a region adjacent the exit end of said accelerating device, an electric field that varies from a potential gradient that is equal to the potenial gradient in said accelerating device, as said column enters said region, to zero, as said column leaves said region.

14. Apparatus as in claim 4, wherein said electrodes are conductive and said means (b), (e), (h), and (i) comprise RESISTANCES and diodes connected between said respective electrode for maintaining them normally at substantially equal potential.

15. Apparatus as in claim 4, wherein said PULSES in (c), (f), and (j) have the same polarity as the charge on said particles.

16. Apparatus as in claim 4, wherein said pulses of (C), (f), and (j) are REPEATED PERIODICALLY and the repetition rate, amplitude, and duration thereof and the velocities of said particles and lengths of all said regions in said apparatus are such that successive columns of particles are formed and moved periodically through said apparatus with the proper phasing during their movement as specified in claim 6.

17. Apparatus as in claim 16, wherein said pulses of (c), (f), and (j) have a DURATION of about 0.1 to 0.15 of the period between successive pulses.

18. Apparatus as in claim 4, wherein the amplitude of said pulses of (c), (f), and (j) and the potential differr' encesof (d) and (g) are such as to respectively increase and decrease the velocity of the particles sufficiently to FORESHORTEN said column by a factor of at least about 5 in said second region and by a factor of at least about 8 in said fourth region, and to provide an effective TIME BUNCHING on said target of at least about 50; and thus to provide a total bunching of at least about 2000.

19. Apparatus as in claim 2, wherein said COLUMN in said FIRST REGION is about 0.6 to 0.8 as long as said first region.

20. Apparatus as in claim 19, wherein the diameter of said column equals about 0.2 to 0.5 of its length.

21. Apparatus as in claim 2, wherein said rst and second ELECTRODES are positioned transverse to, and coaxial with, said column of particles, and are so shaped that the SPACING between their opposed surfaces decreases substantially with increasing distance from said column.

22. Apparatus as in claim 4, wherein said electrodes are positioned transverse to, and coaxial with, said column of particles, and are so shaped that the spacing between their respective pairs of opposed surfaces in said first, second, third, fifth, and sixth regions decreases substantially with increasing distance from said column.

23. Apparatus as in claim 2, including at least one combination of further foreshortening means, each said combination comprising:

(a) means for accelerating all said particles in the foreshortened column substantially equally to higher velocity in a given region; and

(b) means for decelerating said particles to lower velocity as they enter a subsequent region.

24. Apparatus as in claim 23, wherein each said region includes an upstream electrode and a downstream electrode and:

(c) wherein each said accelerating means (a) comprises means for applying to the upstream electrode of said given region a large electrical aiding voltage to create a substantially uniform electric Ifield between the upstream and downstream electrodes of said given region when said column has passed through said upstream electrode, said voltage ending just before the forward end of said column reaches said downstream electrode; and

(d) wherein each decelerating means (b) comprises means for providing a retarding potential difference between the downstream and upstream electrodes of said subsequent region.

25. Apparatus as in claim 24, wherein said means for applying an aiding voltage comprises a source of sinusoidal alternating voltage and half-wave rectication means connected between said source and said electrodes.

26. Apparatus as in claim 25, wherein said source of alternating voltage is connected to a capacitor in series with said electrodes and said rectification means comprises a unidirectional conducting device connected in parallel with said electrodes.

References Cited UNITED STATES PATENTS 3,333,142 7/1967 Takeda et al. B15-5.41 X

ROBERT SEGAL, Primary Examiner U.S. Cl. X.R. 313-63