First-Principles Determination of the Space-Charge-Limited Current in High-Current Synchrocyclotrons (Rainwater's Matching-Radius Problem)

The large synchrocyclotrons of the 1950s-60s delivered time-averaged proton currents of only about 0.4-5 microamperes. Rainwater (Rev. Sci. Instrum. 37, 262, 1966), building on MacKenzie and Lawson, showed that this ceiling is set by space-charge beam blowup at small radius and compressed it into a semi-empirical limit formula, $I\le(5{-}10)\nu_z^2 f$ microamperes, whose constant rests on empirically fixed beam-geometry parameters (half-height $a\approx5$ cm, azimuthal spread $\Phi\approx2$ rad) and an empirical matching radius $r_m\approx10$ cm. The formula predicts 60-280 microamperes for machines with spiral-ridge vertical focusing ($\nu_z^2\approx0.04{-}0.09$) and a 300 Hz frequency-modulation rate, a gain of up to about 100. The Nevis synchrocyclotron was actually rebuilt with spiral sectors (1970-72) and reached about 12 microamperes by 1975; the full predicted regime was never built, and the semi-empirical constants were never derived from beam dynamics. The problem is to determine, from self-consistent space-charge beam dynamics of the frequency-modulated synchrocyclotron - including the phase-oscillation-modulated charge density at small radius, the resonance conditions $2\nu_z=\nu_r$, $3\nu_z=\nu_r$, $3\nu_z=1$ crossed once $\nu_z^2>0.1$, and beam-induced excitation of parasitic modes of the dee structure - the limiting current with the matching radius as an emergent output, and to establish whether stable operation at 60-280 microamperes is possible at all. The 1966-1975 follow-up theory and machine literature, and the modern validated intensity-limit theory and self-consistent multi-particle simulation program for isochronous machines, leave this core open: the adjacent isochronous-machine theory does not treat the frequency-modulated case.

Background

A synchrocyclotron is a cyclotron in which the radio-frequency (RF) accelerating voltage is frequency-modulated (FM) to follow the decreasing revolution frequency of the ions as their relativistic mass grows with energy; this lets a single magnet accelerate protons to several hundred MeV, but the beam is captured in short bunches once per FM cycle, so time-averaged currents are small. The accelerating structure consists of hollow D-shaped electrodes ('dees'); the beam gains energy at the gap between them while circulating in the magnet gap. In a classical weak-focusing synchrocyclotron the average magnetic field decreases with radius, which provides vertical focusing measured by the vertical tune $\nu_z$ - the number of vertical betatron oscillations per revolution - and the radial tune $\nu_r$ plays the analogous role for radial oscillations. In the large machines of the 1950s-60s, $\nu_z^2$ was only about 0.005 near the center, and time-averaged currents sat at 0.4-5 microamperes (for comparison, the isochronous sector cyclotrons later built at PSI, the Paul Scherrer Institute, today's leading high-intensity cyclotron laboratory, now operate near 2 milliamperes).

MacKenzie (1964) showed that this current ceiling was set not by ion sources or RF power but by space charge - the mutual Coulomb repulsion of the beam ions - which blows the beam up vertically and radially where the density is highest, at small radius during the FM capture period. Lawson extended the theory to ordinary (fixed-frequency, isochronous) cyclotrons. Rainwater (1966) then reconsidered Lawson's treatment in detail for synchrocyclotrons - where the added complication is that phase oscillations - the periodic swing of each ion's radio-frequency phase about the accelerating phase, which alternately compresses and dilutes the circulating bunch - make the charge density at small radius strongly time-dependent - and obtained a simpler semi-empirical limit. In reduced form it reads $I\le(5{-}10)\,\nu_z^2 f$ microamperes, where $f$ is the FM repetition rate in cycles per second. The numerical constant embeds three quantities fixed by hand rather than by theory: the beam half-height $a\approx5$ cm, the azimuthal spread $\Phi\approx2$ rad, and a 'matching radius' $r_m\approx10$ cm, the radius at which the space-charge-dominated early motion hands over to the focusing-dominated regime (defined quantitatively as the radius at which the space-charge-induced vertical tune shift becomes comparable to the external vertical focusing, $|\Delta\nu_z^2|\sim\nu_z^2$). With the measured $\nu_z^2\approx0.005$ the formula reproduces the best observed currents (about 2 microamperes at Nevis, about 5 microamperes at Rochester), and substituting $\nu_z^2\approx0.04{-}0.09$ with $f\approx300$/s predicts 60-280 microamperes, a gain of up to about 100.

'Spiral ridges' (an azimuthally varying field, AVF) are iron sectors with spiral-shaped edges added to the pole faces; the azimuthal field variation ('flutter') and the spiral angle add vertical focusing without lowering the average field, raising $\nu_z^2$ toward 0.1. Raising $\nu_z^2$, however, brings the operating point toward known orbit-resonance conditions - $2\nu_z=\nu_r$, $3\nu_z=\nu_r$, and the walkout resonance $3\nu_z=1$ - at which driven betatron oscillations can eject beam, so the focusing gain must be traded against resonance loss. At the predicted high currents, space-charge forces on radial and azimuthal motion and the beam-induced excitation of parasitic RF modes of the dee structure were also flagged as untreated.

The machine side of this program was actually carried out, but only part way: the Nevis synchrocyclotron (Columbia University) was rebuilt in 1970-72 with three-fold spiral-sector iron - implementing the spiral-ridging proposal - raising the energy to 560-570 MeV with a design current of about 20 microamperes external at 300 Hz and roughly 80 percent extraction efficiency; it reached a projected 12 microamperes by 1975, roughly an order of magnitude over its pre-conversion 0.4-1.6 microamperes but far short of the 60-280 microampere regime, which no synchrocyclotron ever entered. The high-current frontier passed instead to continuous-wave isochronous sector cyclotrons, for which a quantitative intensity-limit theory (Joho's sector-model limit, scaling as the cube of the effective accelerating voltage, set by turn separation at extraction and by uncompensable nonlinear space-charge energy spread) has been validated against detailed three-dimensional particle-in-cell simulations - the standard multi-particle method, in which the beam is represented by many charged macroparticles deposited on a spatial grid and the self-consistent space-charge field is solved on that grid at each step - and against machine measurements at PSI. Modern synchrocyclotrons are compact superconducting medical machines running at nanoampere currents, for which space charge is not a design driver. What was never done, for the frequency-modulated synchrocyclotron specifically, is to replace the calibrated constants of the semi-empirical limit by a first-principles derivation from self-consistent beam dynamics, and to establish whether the predicted 60-280 microampere regime is stably reachable at all.

Problem Statement

Determine, from first principles of self-consistent space-charge beam dynamics in a frequency-modulated synchrocyclotron, the space-charge-limited time-averaged beam current as a function of the machine parameters (vertical tune $\nu_z^2$, frequency-modulation repetition rate $f$, magnetic field, RF program), such that the quantities Rainwater (1966) fixed by hand - the matching radius $r_m$, the beam half-height $a$, and the azimuthal spread $\Phi$ - emerge as outputs of the dynamics rather than fitted inputs; and determine quantitatively, independently of the limit function, whether stable beam confinement and acceleration are possible in the spiral-ridged high-focusing regime ($\nu_z^2\approx0.04{-}0.09$, $f\approx300$/s, predicted currents 60-280 microamperes), in which the beam must cross or skirt the orbit resonances $2\nu_z=\nu_r$, $3\nu_z=\nu_r$ and $3\nu_z=1$ - with the local $\nu_z^2$ varying with radius, so the resonances are skirted within this window and crossed once $\nu_z^2$ rises above about 0.1 - and in which the analysis must quantify the phase-oscillation modulation of the charge density at small radius, must quantify the space-charge-driven radial, vertical and phase-space distortion of the bunch, and must quantify the beam-induced excitation of parasitic modes of the dee structure.

These results are required jointly and neither implies the other: the first-principles limit with $r_m$, $a$ and $\Phi$ emergent does not by itself establish stability in the spiral-ridged regime, so the quantitative stability verdict there must rest on the resonance, beam-distortion and parasitic-mode analysis. The derivation and the verdict must make contact with the historical calibration data - existing machines at $\nu_z^2\approx0.005$ delivering about 2 microamperes (Nevis) and about 5 (Rochester), and the rebuilt Nevis machine with three-fold spiral sectors reaching about 12-20 microamperes at 300 Hz - and must state explicitly the machine class, parameter range, and current regime for which their conclusions hold.

References

1. J. Rainwater, 'Increasing Synchrocyclotron Currents and the Space Charge Limit', Review of Scientific Instruments 37(3), 262-267 (1966). DOI: 10.1063/1.1720154, https://doi.org/10.1063/1.1720154
2. K. R. MacKenzie, 'Space charge limits and cyclotron beam enhancement', Nuclear Instruments and Methods 31, 139-146 (1964). DOI: 10.1016/0029-554x(64)90335-0, https://doi.org/10.1016/0029-554x(64)90335-0
3. K. R. MacKenzie, 'Invited Paper - Synchrocyclotron Conversion', IEEE Transactions on Nuclear Science 13(4), 220-225 (1966). DOI: 10.1109/TNS.1966.4324209, https://doi.org/10.1109/TNS.1966.4324209
4. H. G. Blosser, 'Synchrocyclotron Improvement Programs', IEEE Transactions on Nuclear Science 16(3), 405-414 (1969). DOI: 10.1109/TNS.1969.4325262, https://doi.org/10.1109/TNS.1969.4325262
5. D. J. Clark, H. Kim, K. R. MacKenzie, J. T. Vale, 'Modification Studies for the Berkeley 184-In. Cyclotron', IEEE Transactions on Nuclear Science 13(4), 235-239 (1966). DOI: 10.1109/TNS.1966.4324214, https://doi.org/10.1109/TNS.1966.4324214
6. R. Cohen, 'The Nevis Synchrocyclotron Conversion Project', IEEE Transactions on Nuclear Science 13(4), 240-242 (1966). DOI: 10.1109/TNS.1966.4324216, https://doi.org/10.1109/TNS.1966.4324216
7. J. Rainwater, 'Status of the Nevis Synchrocyclotron Modification', IEEE Transactions on Nuclear Science 18(3), 262-267 (1971). DOI: 10.1109/TNS.1971.4326027, https://doi.org/10.1109/TNS.1971.4326027
8. 'Columbia Builds Meson Factory From Nevis Synchrocyclotron', Physics Today 21(10), 57-59 (1968). DOI: 10.1063/1.3034540, https://doi.org/10.1063/1.3034540
9. J. Rainwater, 'Status of the Nevis Synchrocyclotron Modification', AIP Conference Proceedings 9, 165-178 (Sixth International Cyclotron Conference, Vancouver, 1972). DOI: 10.1063/1.2946357, https://doi.org/10.1063/1.2946357
10. D. W. Storm, R. C. Cohen, G. Dugan, A. M. Sachs, J. Rainwater, Y. Au, R. A. Hawrylak, J. Shill, F. G. Tinta, 'Nevis Synchrocyclotron Beam Status Report', IEEE Transactions on Nuclear Science 22(3), 1408-1410 (1975). DOI: 10.1109/TNS.1975.4327896, https://doi.org/10.1109/TNS.1975.4327896
11. S. Kullander, 'Perturbed phase motion in a synchrocyclotron due to space charge', Nuclear Instruments and Methods 62, 169-172 (1968). DOI: 10.1016/0029-554X(68)90430-8, https://doi.org/10.1016/0029-554X(68)90430-8
12. J. D. Lawson, 'Space charge and vertical motion near the centre of the cyclotron', Nuclear Instruments and Methods 64, 214-216 (1968). DOI: 10.1016/0029-554X(68)90202-4, https://doi.org/10.1016/0029-554X(68)90202-4
13. S. Kullander, 'Studies on capture efficiencies of synchro-cyclotrons', Nuclear Instruments and Methods 81, 61-71 (1970). DOI: 10.1016/0029-554X(70)90611-7, https://doi.org/10.1016/0029-554X(70)90611-7
14. D. Thouroude, 'Filling of the Phase Space in the Central Region of a Synchrocyclotron', AIP Conference Proceedings 9, 226-231 (Sixth International Cyclotron Conference, Vancouver, 1972). DOI: 10.1063/1.2946365, https://doi.org/10.1063/1.2946365
15. D. Thouroude, 'Space Charge Effects in the Vertical Plane during Injection in a Synchrocyclotron', IEEE Transactions on Nuclear Science 20, 880-881 (1973). DOI: 10.1109/TNS.1973.4327277, https://doi.org/10.1109/TNS.1973.4327277
16. W. van Genderen, W. van Hattem, 'The influence of carbontetrachloride on the beam current in a synchrocyclotron', Nuclear Instruments and Methods 80, 293-295 (1970). DOI: 10.1016/0029-554X(70)90775-5, https://doi.org/10.1016/0029-554X(70)90775-5
17. W. Diamond, R. C. Cohen, D. W. Storm, 'A Pulsed Synchrocyclotron Ion Source Using a Single Cold Cathode', IEEE Transactions on Nuclear Science 26(2), 2171-2174 (1979). DOI: 10.1109/TNS.1979.4329830, https://doi.org/10.1109/TNS.1979.4329830
18. V. I. Danilov, I. V. Puzynin, A. N. Safonov, B. V. Feoktistov, 'Restriction of the beam current by space charge in the JINR synchrocyclotron', English translation of a JINR (Dubna) report, ANL-TRANS-646. https://inspirehep.net/literature/1098825
19. Y. J. Bi, A. Adelmann, R. Dölling, M. Humbel, W. Joho, M. Seidel, T. J. Zhang, 'Towards Quantitative Simulations of High Power Proton Cyclotrons', Physical Review Special Topics - Accelerators and Beams 14, 054402 (2011). DOI: 10.1103/PhysRevSTAB.14.054402, https://doi.org/10.1103/PhysRevSTAB.14.054402 (preprint arXiv:1012.0718)
20. A. Kolano, A. Adelmann, R. Barlow, C. Baumgarten, 'Intensity limits of the PSI Injector II cyclotron', Nuclear Instruments and Methods in Physics Research A 885, 54-59 (2018). DOI: 10.1016/j.nima.2017.12.045, https://doi.org/10.1016/j.nima.2017.12.045 (preprint arXiv:1707.07970)
21. D. Winklehner, J. M. Conrad, D. Schoen, M. Yampolskaya, A. Adelmann, S. Mayani, S. Muralikrishnan, 'Order-of-magnitude beam current improvement in compact cyclotrons', New Journal of Physics 24, 023038 (2022). DOI: 10.1088/1367-2630/ac5001, https://doi.org/10.1088/1367-2630/ac5001 (preprint arXiv:2103.09352)
22. W. Joho, 'High Intensity Problems in Cyclotrons', Proceedings of the 9th International Conference on Cyclotrons and their Applications, Caen, France, 337-347 (1981). https://inspirehep.net/literature/1301374