TRIUMF: Canada's national laboratory for particle and nuclear physics

Progress report on the feasibility of using the ISR magnets in a TRIUMF Kaon factory Blackmore, Ewart William; Botman, J.; Bugg, D. V. (David Vernon); Craddock, M. K. (Michael Kevin); Hereward, H. G.; Joho, W.; Laxdal R.; McKenzie, G. H.; Reeve, P. A.; Reiniger, K.; Richardson, J. Reginald; Teng, L. C. Jun 30, 1983

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


51833-TRI_KAON_1983_01.pdf [ 36.66MB ]
JSON: 51833-1.0355475.json
JSON-LD: 51833-1.0355475-ld.json
RDF/XML (Pretty): 51833-1.0355475-rdf.xml
RDF/JSON: 51833-1.0355475-rdf.json
Turtle: 51833-1.0355475-turtle.txt
N-Triples: 51833-1.0355475-rdf-ntriples.txt
Original Record: 51833-1.0355475-source.json
Full Text

Full Text

TRIUMF TRIUMF \ PROGRESS REPORT ON THE FEASIBILITY OF USING THE ISR MAGNETS IN A TRIUMF KAON FACTORY The following have contributed to var~ous aspects of this study: E.W. Blackmore J. Botman D.V. Bugg M.K. Craddock (Chairman) H.G. Hereward W. Joho R. Laxdal G.H. Mackenzie P.A. Reeve K. Reiniger J.R. Richardson L.C. Teng 4004 Wesbrook Mall Vancouver, B.C. V6T 2A3 , . June 1983 ii. SUMMARY TRIUMF is considering the construction of a "kaon factory" post-accelerator to take the present 100 ~ proton beam (6 x 10 14 pis) f rom 500 MeV to energies in the tens of GeV. This would provide beams of kaons, antiprotons, neutrinos, etc., 100-1000 times more intense - or alternatively cl eaner - than those available at present. Such beams would open up new fields in both nuclear and particle physics in the same way that the pion factories LAMPF, SIN and TRIUMF have done at lower energies; in particular the enhanced ability to study rare processes could throw light on mass regions far beyond those accessible with any presently conceivable super-high energy accelerator. Since it is understood that there are no definite plans to use the dipole magnets from the Intersecting Storage Rings at CERN af t er closure of the ISR in 1984, a study has been undertaken to assess the feasibility of incorporating them in a TRIUMF kaon factory. If these magnets were available at nominal cost, then for the same total funds it should be possible to build a kaon factory of considerably higher energy and physics capability. Out of various options the one selected as most suitable in terms of its physics capability, site requirements and cost would use one ring of magnets for a 30 GeV 50 ~ synchrotron and the other for a 30 GeV dc stretcher ring. . The magnets would be reconfigured in a regular 48 cell FFODDO lattice, apart from 4 straight sections. The two rings would be mounted one above the other in the same tunnel, along with a dc accumulator ring at the injection energy. This arrangement would have the advantage of providing both low and high duty factor beams from the outset. The synchrotron magnets would operate with a triangular waveform at 0.5 Hz, using flywheel/motor generator energy storage. In the longer term a 60 GeV superferric or 120 GeV superconducting ring could be added in the tunnel. iii. To provide 50 ~ average current a charge of 100 ~C or 6 x 10 14 protons must be accelerated in each pulse. Charges of this magnitude, corresponding to a circulating current of 32 A, are routinely injected into the ISR from the PS and stored for many hours. Injecting them at low energy, however, introduces much stronger space charge defocusing effects, while the bunching of the beam adds a new set of possible instability modes. To avoid longitudinal space charge ~~~ aefocusl~~ms at transition, quadrupoles are added to the lattice with a superperiodicity of 12, just above the horizontal tune value Qx' in order to drive the transition energy above 30 GeV. To limit the tune shift due to transverse space charge defocusing to 6Qy '0.25 for 6 x 1014 protons an injection energy) 3 GeV is required. The 3 GeV booster accelerator would have a 30 m diameter (twice that of TRIUMF, one-ninth that of the 30 GeV ring), a separated-function lattice with 4 straight sections, and a repetition rate of 4.5 Hz (or possibly 22.5 Hz with 5-turn stacking in the large ring). Since 100 ~ average currents can be accelerated in the TRIUMF cyclotron while only 50 ~ are needed by the main ring some protons remain available for use at either )430 MeV or 3 GeV. In the latter case the booster would be run at a higher repetition rate, say 6 Hz (30 Hz), and 3 pulses would be sent to experiments for each 9 going to the main ring. Schemes for matching the cw beam from TRIUMF to pulsed accelerators have been discussed in published papers. Packets of 160 turns with acceptable momentum spread can be formed by lowering the energy gain per turn locally using auxiliary rf cavities. The packets are separated by time gaps programmed from the ion source. Pulsed electrodes deflect the packets of H- ions vertically for extraction by electrostatic septum and magnetic channel. The preferred extraction energy is 430 MeV, to avoid H- loss by electromagnetic stripping. The iv. packets are injected into a 430 MeV dc accumulator ring mounted above the booster ring by H- stripping in a carbon foil. Several thousand packets are collected during each 111 ms cycle for pulsed ejection into the booster, multiple scattering being limited by kicking the proton orbits away from the foil. The purpose of this progress report is to keep interested parties informed of the progress of the study and to provoke some feedback in the form of criticisms, corrections, additions and improvements. While the most critical problems involved in adapting the ISR magnets to a kaon factory have now been solved, a number of features remain to be examined in more detail before the proposal can be regarded as complete, e.g.: magnet lattice designs, especially the inclusion of straight sections inj~ction and extraction systems rf accelerating systems beam instabilities and their control correction magnets cost estimates, including civil engineering. Work will proceed on these items, among others, during the summer, and if, as expected, the studies continue to progress favourably, it is anticipated that a more complete study and a Letter of Intent will be submitted to CERN for the September or October meetings of the Committee of Council. The Letter of Intent would express TRIUMF's formal intent to CERN to negot i ate acquisition of the magnets and to undertake the requisite steps toward the funding of the accelerator system from the Government of Canada. v. CONTENTS Summary ii. Introduction 1. Beam characteristics required 3. Schemes involving the ISR magnets 5. Pulsed operation of the magnets 6. Lattice design for high transition energy 8. Booster synchrotron 10. Matching the TRIUMF cyclotron to the booster synchrotron 12. Conclusion 16. Acknowledgements 17. References 18. Table 20. Figures Appendices 1. INTRODUCTION Two frontiers may be recognized in the uncovering of new phenomena in subatomic physics - those of high energy and high intensity. The race to higher energies has always left ground behind it only partially explored - a factor which the pion factories LAMPF, SIN and TRIUMF have successfully exploited over the last few years with their 200-800 MeV high intensity accelerators. These have made possible a large variety of experiments, in particular observations of very rare processes and high precision comparisons with theory. Some recent examples at TRIUMF include a search for neutrinoless muon conversion,1 ~ + A + e + A (measured with a time projection chamber to a new world lower limit below 10-11), which would violate lepton number conservation, and an accurate measurement of the endpoint decay spectrum for highly polarized muons 2, which has put a lower limit of at least 380 GeV on the mass of any right-handed charged W boson. A similar opportunity now arises for accelerators in the tens of GeV - the so-called "KAON factories" - capable of generating beams 100-1000 times more intense than those available at present. (The limitations of this term are somewhat relaxed if KAON is understood as an acronym for the major particles produced - Kaons, Antiprotons, Other hadrons, Neutrinos). Indeed there is additional justification for such projects in that grand unified theories are predicting particles with rest energies far above the practical aspirations of any laboratory - and therefore observable only indirectly through higher order effects (like those responsible for rare decay modes) in experiments involving the largest possible number of particles. The kaon itself remains one of the most fascinating of the elementary particles. Over the past thirty years its behaviour has led to a number of crucial discoveries in particle physics: strangeness, parity violation in the weak interaction, the violation of CP invariance and the existence of a fourth "charmed" quark, first suggested by the suppression of decays such as K~ + ~+~-. Today the kaon continues to promise fundamental insights not only into particle physics but also into nuclear physics. Nevertheless the beams of kaons available at present are frustratingly weak (~105 K-/s) and heavily contaminated with pions (~10 n/K). Many desirable experiments are just not 2. feasible. A similar situation holds for the neutrinos, antiprotons, hyperons and other secondary particles produced by GeV accelerators. The same was true 'for pion and muon physics before the advent of the pion factories. If anything the situation is worse for K and p beams, which are of poorer quality than the n and ~ beams were 15 years ago. Consequently there is a strong interest in building kaon factories to produce beams much more intense than those available at present or, at the sacrifice of some intensity, much less contaminated. Experimental topics of particular interest at a kaon factory include -Rare kaon and hyperon decays; CP violation; Neutrino scattering and oscillations; Hyperon production, scattering and reactions; Meson and baryon spectroscopy; Hadron-nucleon interactions (nN, KN, NN, YN); Antinuc1eon interactions; K+-nucleus scattering; Hypernuclear physics; (K-,K+) double strangeness exchange reactions; Resonance propagation in nuclei; Exotic atoms; Muon physics (muon fluxes will be an order of magnitude higher than at the pion factories). The case for kaon factories has been made at a number of recent workshops: Brookhaven 3 (1976), TRIUMF 4,5 (1979, 1981), LAMPF 6- 8 (1981, 1982) and Santa Cruz 9 (1983). Proposals for kaon factories have come in the main from the existing pion factories, the reason for this being that these machines alone have adequate energy and current to act as injectors. Support for these proposals has come not only from pion factory users and neutrino physicists, but also from an enthusiastic community of K and p users from high energy laboratories where the lower energy channels are gradually being closed down 3. through financial exigency. The strong continuing interest of this group is evidenced by the series of international conferences held at Zvenigorod 10 (1977), Jablonna 11 (1979), Rome 12 (1980), and Heidelberg 13 (1982). BEAM CHARACTERISTICS REQUIRED The energy chosen for a kaon factory accelerator will depend on the secondary particle species, momenta and intensities desired, perhaps tempered by consideration of what funds might be available. What beam momenta are required for the major secondary particles (K,p,v)? Nuclear studies with kaons require primarily slow or stopping beams. Hadronic interactions and spectroscopy need beams of a few GeV/c, while some rare decay studies would be best served at -5 GeV/c. The antiproton situation is complicated by the LEAR project, the impact of whose cooled and pure beams cannot yet be fully assessed. In the short term, interest in p beams at a kaon factory would probably be restricted to momenta >1.3 GeV/c, the maximum attainable by LEAR; in the long term, special p cooling and storage systems would presumably be built. To clear up some uncertainty in the production cross sections for kaons and antiprotons, an experiment was mounted at the CERN PS in 1981 involving scientists from TRIUMF, LAMPF, CERN, Rome and Saclay.14 Measurements were made for proton energies of 10, 18 and 24 GeV, 1 em thick targets of carbon, copper and tungsten, and ~-, K-, and p momenta of 0.4, 1.0 and 1.4 GeV/c. For all particles and targets the data are consistent with a linear increase in cross section with proton energy over this range (Fig. 1). It is of interest that this increase is much more rapid than that predicted by the Sanford-Wang formu1a 1S with kinematic reflexion applied (although the shape of the momentum spectrum for a given primary energy does agree with the formula). While proton energies ~15 GeV are probably sufficient for good fluxes of <2 GeV/c kaons, it appears that at least 25 GeV is needed for antiprotons and higher momentum kaons. For neutrinos there is no clear energy threshold or limit, but production and reaction rates increase linearly with primary beam energy, favouring the highest energy possible. 4. The primary beam currents under consideration are 50-100 uA (3 - 6 x 10 14/s). This would give about two orders of magnitude improvement over existing beams in this energy range. Such currents would seem to be technically feasible and would be sufficient to make possible the significant new experiments on which the project is predicated - either by straightforward increases in rate, or by improving beam purity through more selective separation. 16, 17 Experimental requirements on time structure span the whole spectrum from sharply pulsed to dc. For neutrino experiments very sharp pulses on a macroscopic time scale are required (~10-5), whereas for many-particle coincidence experiments dc beams are preferable. The microscopic time structure of the beam could be very valuable for particle identification, a pulse repetition period in the range of 20-50 ns being most suitable. 5. SCHEMES INVOLVING THE ISR MAGNETS With the above specifications in mind, the news that the Intersecting Storage Rings at CERN were to be closed down in 1984, and that there were no definite plans for further use of the dipole magnets, immediately suggested that they might be employed in the kaon factory which TRIUMF had been considering. Both the energy capability and circulating charge in the ISR are in the range of interest for a kaon factory. Each ring of magnets is capable of bending and focusing a proton beam of 31.4 GeV. The maximum circulating current in a ring has exceeded 50 A, correspnding to a charge of 160 ~C (10 15 protons). If this charge could be held through all phases of acceleration and the magnets could be pulsed at 0.5 Hz, like the CERN PS, then an average beam current of 80 ~ would be realizable. If (again) the magnets could be obtained at low cost a very important saving could be achieved in the overall expense of the project - or a more energetic accelerator with greater physics potential could be built for the same available funds. In view of the likely costs, previous plans for a kaon factory at TRIUMF had been based on a 15 GeV 100 ~ machine - either a superconducting cw ring cyclotron 18 , 19 or a 30 Hz rapid cycling synchrotron 20 , 21 (see Appendix II). The ISR magnets, being designed for 30 GeV, suggested consideration of a higher energy. Three different ways of using the magnets were considered (with intensities normalized to the same beam power - assumed roughly proportional to particle production - as the 15 GeV project): 1. A 60 GeV 25 M synchrotron, using all the magnets in one ring 2. A 30 GeV 50 M synchrotron, plus a 30 GeV dc stretcher ring 3. A 25 GeV 60 M synchrotron, a 25 GeV dc stretcher ring, plus all the proton switchyard dipole magnets (144 0 total bend). Of these t~ 30 GeV option was chosen as the most suitable on the basis of physics capability, site requirements and cost. It provides sufficient energy for good production of antiprotons, neutrinos and 5 GeV/c K-, and, most importantly, would make a high duty factor beam for counter experiments available from the outset. Its compatibility with the site can be seen from Figure 2, where the 6. tunnel is designed to accommodate both rings (Figure 3) . This layout was chosen from among several investigated 22 for general convenience and because it allows beam to be fed back into the existing experimental areas as well as into new ones. The 60 GeV option, though offering some physics temptat i ons, would have had to extend beyond the boundaries of the University of British Columbia campus and appeared to be outside reasonable funding expectations. In any case Option 2 still permits a high intensity 60 GeV beam to be obtained at a later stage by installation of a superferric synchrotron ring in the same tunnel; with superconducting magnets even 120 GeV could be considered, though the cycling rate and intensity would be lower. The chief problems associated with using the ISR magnets in a kaon factory stem from their potentially low repetition rate compared to that of a rapid cycling synchrotron at 30 Hz. Assuming that 0.5 Hz can be achieved (which we justify below) then even allowing for the required current being only half that in the RCS the charge per pulse must be 30 times greater. This introduces serious problems with space charge, both longitudinal and transverse, and with beam instabilities. In addition the mismatch between the synchrotron and the 23 MHz cw TRIUMF cyclotron is enhanced by a factor of 60. These problems are discussed in the next four sections. Fortunately, and perhaps surprisingly, all of them, so far as they have been investigated, appear to be capable of solution. PULSED OPERATION OF THE MAGNETS Although built for dc operation the magnets are in fact of laminated construction and so can in principle be run in a pulsed mode. Indeed CERN itself is considering their use in a 40 GeV "Replacement PS" which could be installed in the ISR tunnel. The question of crucial importance, however, is not so much whether, but how fast they could be pulsed. Limitations on the pulse rate could arise either in the magnets themselves or from their power requirements. 7. The magnets, illustrated in Figure 4, are constructed from 1.5 mm laminations of a very low carbon mild steel; these are held together by welded stainless steel bars lengthwise and by 5 em thick plates of low carbon mild steel at the ends. The pole shape is profiled to give dipole, quadrupole and sextupole components. There are also poleface windings connected in series around the ring which can be used for zero harmonic field profile corrections. The main coils are directly water cooled and insulated with fibre glass epoxy. The insulation should be good for 10 9 rad although no coils have yet received as much as 10 7 rad. The eddy currents and hysteresis effects associated with pulsed operation could give rise to both heating and field perturbation problems. 23 The extra heating turns out to be negligible for realistic pulse rates - a 10% effect for 1 Hz. Field perturbations are more significant, particularly phase lag due to the end plates and the copper heat shields, and saturation effects above 1 T. Multipole correction magnets should be able to deal with these problems, but a magnet measurement programme would be required to determine the solution in detail. Energy storage and power requirements would seem to set more stringent limits on the repetition rate f. Thus for a triangular waveform with no flat top or bottom (as required by a kaon factory to achieve maximum f) the maximum power required for one ring is 85 f MW. Five different methods of energy storage have been considered. 24 1. Mechanical energy storage in a flywheel 2. Direct connexion to the utility grid 3. Resonant circuit 4. Capacitive storage 5. Inductive storage The last two methods are too slow, while at frequencies where the power is reasonable, a resonant circuit would require unreasonably large banks of capacitors and chokes. Direct storage on the grid is attractive in principle but would not be so in practice, because of TRIUMF's location >20 km on the wrong side 8. of Vancouver from the nearest 500 kV substations. Discussions with the British Columbia Hydro and Power Authority have indicated that installation of a special 500 kV line would be prohibitively expensive, whether buried or overhead through the city, or by submarine cable around it. This leaves the flywheel/motor generator option. For the CERN PS, with a peak input power requirement of 41 MW, this provides a maximum repetition rate f = 0.6 Hz. For the ISR magnets a value of 0.5 Hz would seem reasonably attainable; the technological advances of 25 years might even make 1 Hz achievable. LATTICE DESIGN FOR HIGH TRANSITION ENERGY As noted above, to get an average current of 50 ~ with a repetition rate of 0.5 Hz requires a charge per pulse of 100 ~C (6 x 10 14 protons). This is 40 times greater than has been achieved in the CERN PS, so that space charge defocusing forces and beam instabilities must be carefully considered. Longitudinal defocusing effects are most critical near transition, where Yt = (~)-1/2, ~ being the average horizontal dispersion. In the PS, passage through transition imposed a serious limitation on beam intensity until the Yt-jump technique 25 was introduced using strategically-placed pulsed quadrupo1es to speed the process. Because the speed required depends on the charge it has been considered more appropriate in the kaon factory to raise Yt above the top energy Ymax ( L e. send n-+-O). It has been pointed out (see e.g. Ohnuma 26 ) that nand Yt can be conveniently shifted without affecting the Q-va1ues or B-function significantly by adding quadrupo1es to the lattice with a superperiodic i ty n close to the value of the horizontal tune Qx. The change 6n is given by R Q2 - n 2 x where R is the machine radius and an is proportional to the quadrupole strength. 9. For our purposes a convenient basic lattice is one with 48 FFODDO cells (straight sections and higher order effects have not as yet been considered). With radius R = 135 m (an even multiple (18) of the 7.5 m radius for 430 MeV orbits in TRIUMF) the horizontal tune Qx = 11. Quadrupoles can then be placed with superperiodicity n = 12 and always fall between two short-F magnets, where 8 reaches its maximum. To raise Yt to 34.6 the quadrupole strength needed is 9.2 Tim x 0.4 m. The lattice functions for this case are illustrated in Figure 5. Although the &-function is hardly affected by addition of the quadrupo1es the dispersion n(s) develops strong negative and positive peaks in the cause of bringing n + o. From Courant & Snyder 27 the change in the peak value ~ a n Comparing the linear dependence on quadrupole strength here with the quadratic dependence for ~~ it can be seen that the peak dispersion can be restrained by not bringing Qx too close to n and by using the strongest quadrupoles possible. " For the case quoted n = 7 m, while for injection at 430 MeV there would be an 8 MeV energy spread (see Figure 9) corresponding to ~p/p = 1%. The maximum dispersive displacement would therefore be 7 cm, well below the 14 em available. For higher energy injection (see below) the situation would be even more favourable. The solution above is simply meant to demonstrate the feasibility of raising Yt and has not been optimized in other respects. Straight sections will be required for injection and extraction and possibly for acceleration and collimation. Resonances will have to be carefully considered in making a precise choice of working point Q-values, and higher order effects will also have to be taken into account. 10. BOOSTER SYNCHROTRON Transverse space charge/ de focusing effects are most serious at low energies, in particular at injection. The incoherent tune shift 28 t1J. =-y NRr p 'lTV + [.2.+ fm£2 ] ~ B8 2y 3 h2 g2 1 where N is the number of protons in the ring, 2'ITR the circumference, rp = 1.54 x 10- 16 cm, B the bunching factor, 8 and y the relativistic velocity and energy factors (at injection), a and b the semi-major and -minor radii of the beam cross section, hand g the half heights of the vacuum chamber and magnet gap, £1 and £2 geometrical factors and fm the circumferential fraction of magnet steel. At low energy «<1 GeV) the self-force term dominates and N 82y 3, but at higher energies the image force terms take over and N ~ y. Figure 6 illustrates the energy dependence of N, showing the two terms separately as well as their net effect - which varies almost linearly with kinetic energy. The parameters used are those for the ISR with heating pads and poleface windings removed from the magnet aperture (see Figure 4), a taller vacuum chamber (16 cm x 9 cm I.D.), beam 14 em x 7 cm, t1J.y = 0.25 and B = 2/3, roughly the bunching expected after injection from TRIUMF (see Figure 9). At 430 MeV, the preferred energy for extraction from TRIUMF, the space charge permitted is only 11 ~C. To achieve the 100 ~C desired the injection energy must be raised to 3 GeV with a booster synchrotron. [Note that if the pole face windings were left in, h would be reduced from 4.5 cm to 3.5 cm and a 5 GeV booster would be needed; if the heating pads were also left in, bringing h down to its present 2.6 cm, the booster energy would have to be 6.7 GeV. In this range only the image force term is important.] The field correction function of the poleface windings would be fulfilled by multipole correction magnets, as is already necessary on the ISR at maximum energy, and as is conventional on recently constructed synchrotrons. 11. A suitable radius for the 3 GeV booster would be 15 m. twice that of TRIUMF and one-ninth that of the main ring. To fill all the RF buckets in the main ring the booster must cycle 9 times as fast. at 4.5 Hz. Figure 7 shows the time structure of the various machines. Note that both booster and main ring synchrotrons are fed by dc accumulator rings so that beam can be collected continuously. As for the main ring. the booster transition energy should be above the top energy (Yt > 4) to avoid longitudinal defocusing problems with the high space charge. For a regular lattice. where Yt ~ Qx. this implies a relatively high Q and 8. features which are also desirable for high acceptance and shift. (Overall. these specifications are somewhat similar to those of SATURNE II although the repetition rate is rather higher.) To keep the phase shift per cell ~ n/2 a periodicity of 16 has been chosen. The latt i ce functions for a simple 16-cell separated function FODO lattice for this machine are illustrated in Figure 8. The transition energy is 3.32 GeV with Q-va1ues ~5. This is a very preliminary solution and the next steps will be to include straight sections for injection and extraction. adjust the Q-va1ues with respect to resonances and investigate higher order effects. In view of the tightness of the latt i ce it may also be necessary to consider a design using combined function magnets. An option which has been suggested to help spread out the cash flow. without affecting performance significantly. would be to schedule construction and installation of the booster one or two years after the main ring. Although the latter would at first be space charge limited to 5 ~ (3 x 10 13 protons/s) this is already a high intensity by present standards and is possibly all that would be allowed or desired for the first years anyhow. while operational experience was being gained. On the other hand there would be addi t ional expenses and difficulties. For instance. for the main ring to operate down to 430 MeV rather than 3 GeV its rf system would have to be capable of a frequency swing of 28% rather than 3% and the magnet system would have to pr ovide a good field down to 3% of its peak value. rather than 12%. These technical problems and their associated costs will have to be weighed carefully against the cash flow benefits of delaying the booster. 12. MATCHING THE TRIUMF CYCLOTRON TO THE BOOSTER SYNCHROTRON The TRIUMF cyclotron is a six-sector isochronous machine accelerating H- ions to a maximum energy of 520 MeV. It routinely produces cw beams of 100 ~A within an emittance of 2n rom mrad. Five H- bunches per turn (h = 5) are accelerated, the bunches being separated by the rf period of 43 ns (23 MHz) and having a width of 5 ns (±200). This width can be reduced to less than 2 ns (±8°) with some reduction in beam intensity. The mismatch between the TRIUMF bunch frequency of 23 MHz and a synchrotron repetition rate of 0.5 Hz is a large one - a factor of 46,000,000 - but not unscalable. In the first place only 50 ~ is needed at 30 GeV so the cyclotron need only supply the main ring with a duty factor of 0.50. Thus 23,000,000 TRIUMF bunches (4,600,000 turns) must be collected into one synchrotron pulse. Further factors of relief come from the larger circumference and higher rf frequency of the synchrotrons. The former ensures that 18 "boxcars" - unwound turns from TRIUMF - can be fitted around the ring (in practice 9 boxcars of unwound booster turns, each consisting of 2 TRIUMF turns). The latter means that by operating the booster at 46 MHz at injection the bucket spacing (21.7 ns) is only half that of the bunches from TRIUMF. Consequently the bunches can be interleaved around the rings and double the number accommodated in each boxcar. Together these factors of 18 and 2 bring down the number of TRIUMF turns to be fitted into each boxcar to 128,000. This can be achieved by a combination of two techniques - extraction of multiturn packets from TRIUMF and multiturn injection into the booster accumulator by H- stripping. Pulsed Multiturn Extraction of H- Ions from TRIUMF Beam extraction from TRIUMF is normally carried out by stripping the H- ions in a thin carbon foil, thereby producing a cw stream of proton bunches at the radiofrequency. To extract the H- ions whole, and in pulses, is a very different matter. Nevertheless schemes to do this have been conceived and published 20 ,30 (see also Appendix I) and appear to be feasible using presently 13. available technology. These schemes can be considered in three stages: turn compaction, pulsed deflexion and finally steering out of the cyclotron. To compact of the order of 100 turns closely together i n TRIUMF all that is needed is to decrease the energy gain per turn. This may be achieved either by lowering the dee voltage locally (by modifying the dees or installing X/4 coaxial line decelerators) or by slipping to a non-accelerating phase (by means of a magnetic field bump). Both methods have been investigated theoretically and appear to be capable of providing over 100 turns within an acceptable emittance (12n mm-rad) and momentum acceptance (±0.5%). The rf method is favoured since it produces a longitudinal emittance shape better matched to the synchrotron bucket (Figure 9). Conservation of longitudinal phase space requires that the turn compaction be accompanied by a phase expansion - a feature which of course limits the compaction factor, depending on the initial phase width needed to hold the required beam intensity. The mechanism for this phase expansion has been shown 3 to be the time-varying magnetic field associated with a spatially decreasing accelerating field which gives a time (phase) dependent radial kick. The resulting change in orbit length expands the phase interval occupied by the beam bunch. Computer simulation shows that a reduction of the dee voltage from 82 kV to 11.5 kV will broaden an initial phase width of ±6° to ±50°. One hundred and sixty turns may be compacted over 37 mm, equivalent to 8 MeV or ~p/p = 1%. In practice it is proposed to decelerate the beam by rf cavities outside the dees at the extraction radius (Figure 10). Two X/4 loaded coaxial transmission lines would be laid back to back in push-push mode. They would be shaped to match the extracted orbit and slotted horizontally to admit the beam. At 23 MHz each would be 4.8 m long. At present calculations are beginning to estimate the fields, including leakage field, and determine their effect on the beam. The existing tenth scale model of the vacuum tank and rf system will be used to study the interaction of the proposed cavities with the rest of the cyclotron. 14. The compacted beam is kicked vertically out of the beam plane by pulsed electric kickers. For its size, TRIUMF's focusing is weak and a vertical field of 30 kV/cm over 1 m would displace the beam 18 mm 90° downstream. The vertical width of the TRIUMF beam is about 10 mm so a kicker plate separation of 20 mm should be adequate. The fringe field of the plate would perturb orbits close to, but not in, the stacked region. We therefore impose a macro-duty cycle on the beam injected into TRIUMF so that 80 out of 240 turns are not populated. As shown in Figure 9 this provides a 30 mm space between the inner edge of the kicked packets and the leading edge of the advancing beam front 80 turns away. Since the duty factor obtainable with this technique (160/240 = 67%) is greater than that needed to provide 50 ~ at 30 GeV (50%) one quarter of the packets would be available for experimental use at other energies, either (430 MeV or 3 GeV. In the latter case the booster would be run at a higher repetition rate, say 6 Hz, and 3 pulses would be sent to experiments for each 9 going to the main ring. It is also necessary to restrict the initial phase width to ±6°. It has already been experimentally shown that the use of emittance restricting apertures can provide a beam ±6° wide and that the magnet and rf system are close to having the stability required to maintain a radial distribution such as that shown in Figure 9. A program of ISIS and cyclotron development is underway to determine what beam intensity can be obtained in this phase interval. Following the vertical kick two radial kicks, such as those shown for a static case in Figure 11 (60 kV/cm over 1 m) would generate ample radial displacement for the beam to enter an air core magnetic channel and be conducted out of the machine. To avoid unnecessary electromagnetic stripping of the H-ions, and to make use of the Qr = 3/2 resonance, the preferred energy for extraction from TRIUMF is 430 MeV. Lorentz stripping in the magnetic field rises sharply above this energy to a total of about 10% at 500 MeV. 15. Multiturn Injection into the Booster by H- Stripping Stripping H- ions to protons in a thin foil has recently been employed with success at several synchrotron laboratories to improve the efficiency of multiturn injection. The charge changing process circumvents Liouville's theorem so that beam may be injected into the same phase space over an unlimited number of turns. The only fly in this perfect ointment is the multiple scattering that will also take place if the injected proton beam continues to circulate through the foil. To reduce the number of passes through the foil Teng 21 (see Appendix II) has suggested that after each boxcar is filled it be kicked into a storage orbit not intersecting the foil. The first packet entering a boxcar would then make 128,000 turns before being kicked out. Averaging over later packets (x 0.5) and allowing for returning protons not hitting the foil, which need be little bigger than the incoming beam (conservatively x 0.5) we assume 32,000 passages through the foil. Using the standard formula 33 and assuming the net effect to be the same as that of a plate 32,000 times thicker than the 200 ~g/cm2 (carbon) that is needed we find the rms scattering angle 60 = 6.6 mrad. If protons are lost for displacements Ym ) 45 mm at the maximum 8m = 12 m (Figure 8), while the stripper is placed where 8 = 1.5 m, then the beam loss would be 2.2% (assuming a Gaussian distribution, valid at these angles). This loss is tolerable but could stand improvement. Some alternative approaches are being investigated. One would involve stacking 5 booster turns side by side in the transverse betatron phase space of the main ring using a quarter-integral Q-value. This would enable the booster to be pulsed 5 times faster at 22.5 Hz so that correspondingly fewer packets would have to be stripped into each booster boxcar. 16. CONCLUSION The purpose of this progress report is to keep i nterested parties informed of the progress of the study and to provoke some feedback in the form of criticisms, corrections, additions and improvements. While the most critical problems involved in adapting the ISR magnets to a kaon factory have now been solved, a number of features remain to be eamined in more detail before t he proposal can be regarded as complete, e.g.: magnet lattice designs, expecially the inclusion of straight sections injection and extraction systems rf accelerating systems beam instabilities and their control correction magnets cost estimates, including civil engineering. Work will proceed on these items, among others, during the summer, and if, as expected, the studies continue to progress favourably, it is anticipated that a more complete study and a Letter of Intent will be submitted to CERN for the September or October meetings of the Committee of Counci l . The Letter of Intent would express TRIUMF's formal intent to CERN to negotiate acquisition of the magnets and to undertake the requisite steps toward the funding of the accelerator system from the Government of Canada. 17. ACKNOWLEDGEMENTS This study has benefitted considerably from discussions with various staff members at CERN, particularly Y. Baconnier, o. Bayard, P.R. Bryant, L.Coull, F. Ferger, W.E.K. Hardt, Y. Marti, D. Neet, J.D. Pahud, K.H. Reich, and L. Resegotti. We are also grateful to CERN Management for facilitating our enquiries. From TRIUMF E.G. Auld, D. Axen and M. Zach have given invaluable assistance . Expert computing aid has been provided by F. Jones, C.J. Kost and R.Lee. R. Servanckx has also been invaluable i n helping us to get his second order lattice code DlMAT operational. Finally, Janet West has carried out the typ i ng and preparation of this report with great efficiency and good humour in the face of a very tight schedule. 18. REFERENCES 1. T. Numao, H.L. Anderson et al., 21st Int. Conf. on High Energy Physics, Paris, 1982. 2. J. Carr, G. Gidal et aI, LBL-16183, 1983 (to be published). 3. Proc. Summer Study Meeting on Kaon Physics & Facilities, Brookhaven, June 1976, ed. H. Palevsky, BNL-50579 (1976). 4. Proc. Kaon Factory Workshop, Vancouver, 1979, ed. M.K. Craddock, TRI-79-1. 5. Proc. 2nd Kaon Factory Workshop, Vancouver, 1981, ed. R. Woloshyn and A. Strathdee, TRI-81-4. 6. Proc. Workshop on Nuclear and Particle Physics at Energies up to 31 GeV: Los Alamos, 1981, ed. J.D. Bowman et al., LA-8775-C. 7. LAMPF II Workshop, 1982, ed. H.A. Thiessen, LA-9416-C. 8. Proc. Second LAMPF II Workshop, ed. H.A. Thiessen et al., LA-9572-C. 9. Proceedings of the Theoretical Symposium on Intense Medium Energy Sources of Strangeness, Santa Cruz, March 1983 (to be published). 10. Proceedings of the Seminar on the Kaon-Nuclear Interaction and Hypernuclei, Zvenigorod, September 1977, ed. P.A. Cerenkov (Science Publishing House, Moscow, 1979). 11. Proceedings of the Meeting on Hypernuclear and Low-Energy Kaon Physics, Jablonna, September 1979, ed. J. Pniewski (Warsaw University, Warsaw, 1980). 12. Proceedings of the Workshop on Low- and Intermediate-Energy Kaon-Nuclear Physics, Rome, 1980, ed. E. Ferrari & G. Violini (D. Reidel, Dordrecht, 1981). 13. Proc. Int. Conf. On Hypernuc1ear and Kaon Physics, Heidelberg, 1982, ed. B. Povh, MPIH-1982-V20. 14. J.F. Amann et a1., LA-9486-MS (1982). 15. J.R. Sanford and C.L. Wang, BNL-11479 (1967). 16. P. Birien, Ref. 13, pp 371-8. 17. D.E. Lobb, TRI-PP-83-48 (to be published). 19. 18. M.K. Craddock, C.J. Kost and J.R. Richardson, IEEE Trans. NS-26, 2065 (1979). 19. J. Botman et a1., TRI-PP-83-21 and IEEE Trans. NS-30 (to be published). 20. J.R. Richardson, IEEE Trans. NS-26, 2436 (1979). 21. L.C. Teng, IEEE Trans. NS-30 (to be published - see Appendix II) 22. D.V. Bugg, "Considerations for Beam Layouts for a Kaon Factory", TRI-DN-83-l4. 23. P.A. Reeve, "Properties of ISR Magnets in Pulsed Mode", TRI-DN-83-9 24. K. Reiniger, "Power supply requirements for pulsed operation of CERN ISR magnets", TRI-DN-83-15. 25. W.E.K. Hardt, Proc. 9th Int. Conf. on High Energy Accelerators, Stanford, 1974, p.434. 26. S. Ohnuma, Fermilab p Note No. 105 (1980) 27. E.D. Courant and H.S. Snyder, Ann. 'Phys. l, 1 (1958). 28. L.J. Laslett, Proc. Brookhaven Summer Study on Stor age Rings, BNL-7534, (1963), p.324. 29. Deleted. 30. R. Laxdal et al., TRI-PP-83-24 and IEEE Trans. NS-30 (to be published -see Appendix I). 31. W. Joho, "Proposal for a TRIUMF Storage Ring-TRISTOR", TRI-DN-82-13. 32. W. Joho, Particle Accelerators, ~, 41 (1974). 33. Particle Properties Data Booklet, April 1982, p.95. Final Energy Initial Energy Radius Frequency - Initial Final Repetition Rate Ha rmonic Number Lattice Type Lattice Structure Number of Cells Number of Superperiods Dipole Length Dipole Field Quadrupole Length Quadrupole Gradient Transition Yt Tune Qx Tune Qy Chromaticity ~x Chromaticity ~y Maximum Amplitude ex Minimum Amplitude ex Maximum Amplitude ey Minimum Amplitud ey Maximum Dispersion nx Minimum Dispersion ry 20. TABLE 1 Synchrotron Parameters Main Ring 30 GeV 3 GeV 135 m 61. 7 MHz 63.6 MHz 0.5 Hz 180 Combined Function FFODDO 48 12 2.44 m 1.35 T 0.40 m 9.2 Tim 34.6 11.13 10.1 -1.6 -1.0 45.2 m 4.2 m 32.6 m 6.3 m 6.7 m -4.0 m Booster 3 GeV 430 MeV 15 m 46.1 MHz 61.7 MHz 4.5 Hz 20 Separated Function FODO 16 1.57 1.60 0.75 m T m 10.9 Tim 4.5 5.19 5.54 -0.45 -0.53 11.7 m 1.0m 12.3 m 1.0m 1.1 m 0.5 m N.B. These preliminary lattices contain no straight sections and have not been optimized with regard to resonances or higher order efects. ,..... ~ (!) 0 -b c. "0 N c:: "0 "0 "'-,..... W _. Q. Nb "0 "Cc:: ~ I em CARBON TARGET 1.40· GeV/e SECONDARIES --- SANFORD-WANG + KR ! PRESENT DATA 10 P 8 6 K-4 -1T 2 10 18 24 E PRIMARY (GeV) Fig. 1. Energy dependence of production cross sections on 1-cm carbon target normalized to lO-GeV cross section. h @ --' ---=--~ ~. -r ~ " \ '. _~,t.:J ",.\ \. .. -' \). I .~ ---.. ~ ~ , .\:: ~ ~ . \b -' -' .1 .... \ '" '" .:.QC;\"''' V".' ~ (oI~~~ .,j t. \ " " " , " :: .I; , I' .. ' ' -j-I .:P ~J ,,0, .s ,] . , .., r II .' . ') ., ,0 . .. " ' ... \. " • C':'Jf' PU' fO·:u""C'·P'Y ~ ... . -': .~ .-- .. 1 d -'-". -- -" .--.\--'--f"" - .-.-'_" J r' ... r 4 .. .... _. 11 ==-c ~ , \ - '!_ . . _\, t __ ~_-' ~"-t.:'-""_-"". _ ,: b • t " " , .. \ ~\ ;.. .~~ -.~ -;:::-:: : -~-->(- ::.-~.-'I ' .' \' . -l . t • - \ , \ .. .. 'It: : --_\ -:;; \]" ·\' ·\ ·~ . t .,~ .-V: " , "' . .). ~ . -- l-:-'-~ I . : : l' ... · 1 · " .. ~ \' . ..... . - 1.1 -- ..... : ~ .. . . ~ . \ I W _~ v ~' __ . _ " , \\ .• .•. _ . .. _IA, ._ ------ r~Ll ' : -· · I 1\ '\ I , '. . ~.. '. I :, , . .... ',\ ~---=-- ~J--::'I , ' \ ! 'I ' .. . / ~.. " '- -- ' -,. =; """ . 1,":1!1·:';..t.. ,:..'i ..... .  -~~ . r::--:: 'I 'I -. I, ' . ~Sc'?' _ _____ ' " :; I . />. .., ---;;:;.:~ ..... , I ' -- .. 1 \ . I / . . . "'... - , - '/--7" , '. • \ / ~ 1 ! I ~~ I ---._--_ , r ;" [I \ \ '---::':~_ . ' .... _ l ) -rC0....~,) ~ jJ.J I Ii C=-=J, I i :L r- / I __ ._ . t .;' : -: - . . ~ --~ ~~ i-.... 'l ' ilt:.:::J :i \ .' .' :,,?' -- ...t. I · . I . n . ~ ! " ~ I~ , .. ~ , ;1 e;-:.; . ~. ' I.: I ! BJ" =-i~~~9;=-':-=::"' ___ j :! y _ . .. ~I . , j , I - I I L ~~{-- .. ~ • __ • , .. , I ' I 'W' . '-_ 1 . , , , I.. , . \ '-- r -. -r; .' J . ' " - .  l -, \ -d.. , \ \ _ I , ' , . • • . ' , ' C; .. w " . • . I . I ,.' '" .~ . '" :j ..... L.:: .- ~ " '. , . , • IX' fi L . _., - -, -- . -- . i I . " 7 . L ' -';i - -;- . . " . i l. . ! /T ...... 7~ '\ /y " . . . . , -. ---; \, /, . . . 1/ \ / ' Fig. 2. . . / .' /",~~~ /:/ Possible site layout and 30 GeV main ring ' ... 'v / '''''':t,I,' / ' . .. ~ : ' . I~.: Y,;", /' .... . ""'/ j/' ........ <- / /. / ... ..... ~'. , /. ' "'... . ~ "''' , ,," . / "}.. ,/ . / "I' . / .' / SCALf : , -1000 - 2" INTEJNIIL5 for 3 GeV booster synchrotrons. ~ ..... '(. ----.., .. ,1'II"ItJci . ~,,-cIo6DO . f .. :....-:._-'----' ""'-- l . ', ' . , " , . . >, ... ~. " .. ' C1CZA.u,JIo.I'i!!.. -~'". , ~~ -,-Fig. 3. Cross section through the main ring tunnel showing the synchrotron and stretcher ring magnets. .. Fig. 4. ISR Magnet -Perspective and Cross section. Lng ~._ . __ Tullo d. r,froidi ... rMnI I I I i I ";'.ti';" . ___ ---+.389,& _ ___ _ -1 50 40 (l(rn) 30 ?1(rn) 20 1 0 o -10 30 GeV TRISR. 12 Superperlods. 15-~UN-IS83 13.03 FS o t x-- 1• 60 t y ·-I . OI lox = I I. I 3 I ~y -10. 19 E -31.6GeV tr C-848.23111 L ~~~ I' ~68~ - L. Bill ax , __ .. _ DOl ...... 1.-1 0 20 30 40 50 Fi g . S. L.:l tti ce fun c tions f or the 30 GeV rin g . 60 70 S (m) .. 80 200 • 1 80 o 160 () 1 40 1 20 100 80 60 40 20 o I ........ O. 0 N (j.J.C) 2 3 (J 'Y term 1 term 0.5 1.0 Fig . 6. 1.5 2. 0 2.5 3. 0 3. 5 4. 0 Energy( GeV) Transverse space charge limit for the main ring as ·a function of injection energy. DC STRETCHER } 30 GeV j ............ ~ '-...... 0.5 Hz '\ .~ "-~ ~ y y 3 GeV . ~-)\ )\ ~, )\ )\ )\ !\ )\ )\ )\ )\ J 4.5 Hz ( \ \ I l r Illlllllllllllllllllllllllllllllll.~ ,~ I I ~~,~ ~~ ~ ~ ~ ~ I~I~ I 1111111111111111111111111 ! lllil ~! ~ ~ ~! Illl~ ~ ~~I~I ~'II t 430 MeV 20kHz 23 MHz Fig. 7. Time structures in the proposed accelerator chain. 3 GeV. 16 PERIODS '5-~UN-'9B3 12.19 1 4 12 ~ /''\. E -3. 32GeV tr p(rn) 1 yx / V:y / ~ 1 a Oy=5.54 71 (In) 8-:1 \ / \ / ~X--. 45 t y --· 53 -t \ / \ / 6 4J X X C-94.2SI11 L.BIII8X-8.14T 2 a a 2 3 4 5 6 7 8 S (m) Fig. 8. Lattice functions for the 3 GeV ring. PHAS 1 80 (deg) 150 ~ TRIUMF 1 20 90 60 30 Synchrotron 46 MHz bucket /x\/ I· , I .. · .. ..... \, \ n = 1620 n = 1460~ ... · . -. -. - .~~-. Y-0 - · /\ \ '. \ \ / ~ ~\ ..... '. \ n = 1380 /" -(\ \ '. \ ~ \ ". \ \ ~ .. .. " . ,\\ \\, • I , \ .. • • • • • ' ... ,i ll \ \ \ \ \ .. . . ....... . •• -. • , •• 1 1\\ , .. \ ', \ r- • • 1. • • eoc:w • I.' . . . . . . . . 0 "- ....... L ! ' I \ i I \ \ . , · 'v .-.-.----.----.-.~"------30 , 425 8MeV ~ . ~ .. 30mm 37mm 430 435 440 445 450 455 E l'J E F< (~ Y ( 1:f e \7 ') Fig. 9. RF turn compaction in long itudinal phase space (n = turn number in TRIUMF) 460 / hJ ~~ fILE 8 HE II PROBES /1 1 BEAM ® J A -A . ?Z- --OP=-:A:-:-N~E~l- / A C !'20oKCRY r_'1 / -~ \ / - - -..:=:::' S::::l--' / ,.. l:::-----_=r=-_ I 1:::::2 _ I p., / L..........; . ~ VACUUM COAX liNE DECELERATOR TANK WAll Fig. 10. Auxiliary decelerating cavity for RF turn compaction. P r perturbed -..., , i , , i i 20 o -20 -10 Fig. 11. P r unperturbed SEPTUM 1 magnetic. channel 300 0 -5 o 5 10 R perturbed R unperturbed (c m) Extraction without stripping: deviation of the deflected orbit from the unperturbed orbit. The beam ellipse is plotted every 30° of cyclotron azimuth. Sufficient clearance is obtained to enter a magnetic channel at about 290°. Appendix I Appendix II APPENDICES Matching an isochronous cyclo t ron to a synchrotron to provide a high intensity injector. R. Laxdal, M.K. Craddock, W. Joho, G.H. Mackenzie, J.R. Richardson and L.C. Teng . Kaon factory with TRIUMF as an injector. L.C. Teng. Note that the following reports were written with reference to an earlier kaon factory concept based on a fast cycling synchrotron. Nevertheless many of the techniques described remain applicable to the present concept. Paper presented at 1983 Particle Accelerator Conference Santa Fe, Harch 21-23 APPENDIX I TRI-PP-83-2" Har 1983 HATCHING AN ISOCHRONOUS CYCLOTRON TO A SYNCHROTRON TO PROVIDE A HIGH INTENSITY INJECTOR R. Laxdal, M.K. Craddock,* W. Joho,t G.H. Mackenzie, J.i.Richardaon* and L.C. TengS TRIUHF, 400" Weabrook Hall, Vancouver, I.C., Canada V6T 2A3 SUlIIIIIBry Schemes are deacribed for injecting the 23 MEz ew S-beam from the TRIUHF 500 HeV cyclotron efficiently into a 30 Hz rapid-cycling aynchrotron. We have conaidered extraction of 100-200 turn from TRIL~F followed by multiturn 1njection into a dc accumulator ring mounted 1n the aynchrotron tunnel. An alternative il to build a amall ilochronous atorage ring to compact leveral thousand turns. Computer .imulations of the various extraction will be prelented. Introduction The TRILO/~ cyclotron l is a six-.ector i.ochronous machine accelerating H- ions to a maxim~ energy of 520 ~:e\'. It routinely produces beams of 130 uA cw within an emittance of 2~ ~-mrad. Five H- bunches per turn (h-5) are accelerated; each bunch separated by "3 ns (23 MHz) and with a bunch width of 5 ns (~O"). This width can be reduced to 2 ns (~.) with .ome loss of bea~. A 5 mg~/c~2 carbon foil is used to Itrip H-to ~ for extraction. \.ie are presently considering using TRIL"MF as an injec-tor to a hisher-energy machine capable of producing proton bea~s of the order of 100 uA at energies auit-aLle for the production of kaon, antiproton and neutri-no beams of greater intensity than presently available. In this paper we will address the proble~ of matching the 23 ~!r.z beaa: from TRIL'MF to a 30 Hz fut-cycl1ne synchrotron. To utilize the TRIL'MF beam current effi-ciently, bealL must be collected over a large fraction of the synchrotron cycle. At 30 Hz up to 770,000 TRIL'l':F bunches would be available for each accelerating cycle. On~ prorosal is to collect bealt froc; TRl1.:1:F over -100 turns (500 bunches) and extract either B-, h O or H+ particles in 215 ns long packets. Almost 1540 of these would be stored in an accumulator before injection into the synchrotron. Another approach is to extract H-particles fron. TRlnlF cw into a leparate de itorage ring. from this ring --60,000 TRIL'l-lF bunches would be extracted in a sin£le packet 215 nl lonb and injected into an accumulator. Turn Compaction in TRlt:~:r To get the equivalent of 100 turns/pulse from TRIL1:F 1n a usable emittance requires increasing the turn density over the extraction region from 16 turns/inch to )100 turns / inch. This can either be done by driving the ion out of phase with a magnetic perturbation or by reduc-ing the effective dee voltage locally. A vertical kick would drive the dense be~ out of the midplane for subsequent extraction. 1. ~:Ilsnetic Turn COlLpaction 2 A Imall perturbation, !Bz ' in the ilochronous field, Bi50ch' over the extraction region caules a deviation from isochronism per turn given by d~/dn --21fh &lz / 'Y2siaoch ' As + approaches 90· the energy gdn/ turn (4e\"dcos~), where Vd 16 the dee voltage, decreues until at 90· the ion begins to be decelerated back towards the cyclotron centre. For a single particle maxir::ur. turn density 1& achieved by a &low approact, to 4-90 0 • However, for a finite phase spread this results *also at Physics Dept., Univ. of British Columbia. ton leave from SlK, Villigen, Switzerland. ton leave frolL Physics Dept., univ. of California at Los Angeles . SFer~ilab, Batavia, IL. in a larae radial apread in the turnaround radii and a Ireater aensitivity to field, frequency or voltage fluctuations. Therefore a more rapid approach to ain~l 1. desired. The optimum compaction occurs vben the phue alip arad1ent dain+/dE • 26ain./~e where Aiin + 11 the 1ni tial phase apread and l£e the extrac-tion interval. 3 One problem with the magnetic compac-tion technique 1s illustrated in Fig. 1. The tran-aition from acceleration to deceleration produces a boomerang-shaped emittance with a turn-front which is elongated 1n radius and difficult to match efficiently to the 46 KHz synchrotron bucket. A significant trade-off must be made between turnl/pulae and total extrac-ted current. For bunches of t6" initial phase width a phase slip of dsin~/ dE • 0.05 He\-l allows 90 turns / pulse with no losses, or 120 turns/pulae with 104 loss for a 3.8 em radial extraction interval. Figure 1 also ahows how the bunch width has expanded to t50· (12 ns). Computer stud i es show that the existing circular tri~ coils cannot produce. for the scalloped orbits, a suf-ficiently rap i d approach to sin~l, but new coils fol-lowing the orbit ahape would do ao. The coils would be best placed to augment one of the existing phase oscil-lations such u that at "30 HeV whid, would a1&o pro-vide a low electromagnetic stripping loss rate. In one case a acalloped triplet t25 em from the midplane with 190 At and 58-4 G produced d'in~/dE-O.05 Me\'-l while t.vz<0.07. 2. if Turn Compaction The aecond method of stacking be~ is to reduce the accelerating voltage thus decreasing the racius gai n per turn of all phases. A de~ gap voltage decreasing with radius produces a time varying magnetic fielc iz-~V /dR)sin ~/ "tf wt:1ch expands the bunch longitucin-ally such that (dE/ dn)t.sin¢ - constant. ~ The AS TeR code frolf, SIt; which simulates thili compaction procedure indicates our belt case is with 2.51 rf flat ~ topping and with the fundamental dee voltage dropping from 82 kV to a constant value of 11.5 kV. Under these con-dition 180 turns with an 1nitial spread of to· could be compacted within 7.5 He\' and t50' (Fig. 2). This is about a factor of two better than magnetic turn CO~p8C­tion, basically due to the better matchin~ of the emit-tance shape to the aynchrotron bucket. For maxilLu~ extraction efficiency an 110chronous phase history is needed over the region of the voltage gradient so that extra trim coils may have to be added to the cyclotron to flatten out the ex16ting fluctuations of 66in ~-:c..2. Two schemes have been propoaed to reduce the effective dee voltage at the extraction radius. In one Icheme S a grounded ahielding plate is inserted into the existing dee atructure just outside the desired extraction radi-us. One possibility would be a plate extending fro~ one resonator root to the other (Fig. 3). SlPERFISE t calculations ahow that this geometry will still reso-nate at the TRIUMF frequency of 23 KHz. In addition experimental tests on a 1:10 rf model of TRIL"MF have ahown that the plates do not perturb the frequency. reduce the Q-value or aeriously increase the rf leakage from the accelerating gap. J Alternatively decelerating rf cavities could be installed outside the dees at the extraction radius. 8 l/" loaded coaxial transmission lines would be laid back to back 1n push-push mode (Fig. "). They would be Ihaped to match the extuctior, orbits and would be slotted horizontally to admit the beam. At 23 KHz each would be ".8 m long. rib. 1. A be~ of t6 0 initial phlie spread undergoin~ magnetic turn co:paction. 3. Pul6ee Extraction Pulsed electroitat1c plltes would deflect the compacted bear vertically 10 that it would either intercept a Itrippinr foil for HO or ~ extrlction or enter a racial electroitatic deflector for H- extraction. The plate6 would be puhed on for one turn to u pty the ex-traction region and then be off for -100 turn6 while the extraction rebion filIi up again. A field of 10 kV / c: extending over 10' of azi~uth (130 cm) wo uld be adequate. To reduce 10iies due to the rldial frinsing fiele of the plates the beu injected into TRIt"Mf woule have a Jr.acre duty cycle of -7 5~, Illy 10C turns on and 30 turns off. The be~ quality fr~ both methods of co~paction i~ siltllar: longitudinal bunch ~"idt h ~O · (1" ns ) , ~·7 . 5 !'1e\' , eltittances of b " mm-mrld horizon-tal ane 2 .. u:r.-:rac vertical. Accult~lation of COltpacted Bea~ Pulsed extraction yields 100 turn packets consisting of a Iilring of 5 bunches -14 ns ~'ide at 43 ns intervals. Allroo5t 15"0 of these packets, Irrivin[ at 22 IE inter-vals, would then ha ve to be Itacked before acceleration by the 30 Hz synchrotron. ArI accUlllu1ator ring could be built in the synchrotron tunnel for this purpose. 9 Synchrotrons of various sizes have been considered but for the purposes of this paper we will assume a radius of -76 : . This would allo~ 10 TRIl~ packets to be placed end to end in -boxcar- fashion around the cir-cultference with another 10 packets interleaved in the s r aces betwee n bunches . This interleaving of bunches is made p05sitle by running the accumulator rf at 46 MHz, $t11[LD HOI II!" lIE SONAT ORS DH /:,AP SIDE VI£W fig. 3. Rf shielding plates for rf turn completion. ? 120 90 60 30 o -30 425 430 10.. J1.1t1 435 440 445 450 ENERGY (WeV) 455 460 rig. 2. A bear of !60 initial phlse .pread undergoin~ rf turn compaction. twice the TRll~ frequency. 77 packets would be stacked in each boxcar. ~ote thlt the boxcars in the Iccumu-lator mike 10 revolutions, while wliting for the next packet, lome particles making I totll of 15,400 revolu-tions per synchrotron pulse. Protons, neutral h)dro&E n ItO:S and H- ionl have all been considerec for transfer of the be&lt from TRll~ to the accum ulator ring. 1. Proton Transfer Proton transfer retains the advantages of strippin~ for the extraction from TRllo/.F and offers the convenience of chargee particle bea~-handling. Injection inte thE accumulator, however, 15 more difficult since uC'h packet must enter. lepar.te volume of phase spI.e . A sc heme has bee n inve5tiiate~ in wt-.1c h 5 pac;"ea wel- l e be stacke~ in transverse (betatror.) pha5f spa:f anc 11 in IDOlroentur. space u6inf rf .tacking. 1 0 K:ic ker ugnea pu15ed at 46 kHz would steer the packet~ 1nte t hE apFropriate phale Ipace rei1on. One o~t of fiVE ef t hE TRIl,.:r bunches would have to be suppreuec! at tr,e ior, source to all0. for the k.1cker rhe time. The ovuall efficiency of this .cheme is low (661) Ine the kicker magnets Inc power .upply would require a significant development effort. 2. Charee Exchanee Injection H- or HO extraction would allow char~e exchangE in-jection into the accumulator thu6 .voiding the restric-tion of Liouville's theorem. EO extraction u v be feasible by pa"age throub~ very thin toil5. 1l" A sir--Cle pus through 30 ugIt / cm 2 carbon would yield 55 :. r. 0 Ind 2l~ H+; multiple pissaie through thinner toils would increase the fraction of HO Ind decrease that rig. 4. Auxili.ry decelerating c.vities for rf turr. coltpaction. o 50 tOO t50 200 250 CARBON TARG[T THICKNESS (loO&m ! Cm') Fig. 5. Charge statel emerging fr~ a carbon foil as a function of itl thicknels. of h~ to a lo~er limit of -4: aet by double charge e~­change (fig. 5). The HO bea~ lile 60 m dovn&trea~ ~oule be 6>13 cc: 2. The H- option requires an efficient met hod of extraction to be developed for TRIl~f; no difficulties are foreseen. If the be~ circulating in the acc~ulator passes con-stantly tt:rougt: the foil multiple scattering ~ould be excessive; the leading packets of each cycle would make 15,400 passes. Instead an accumulator could be con-structed ~itt: two closee orbits: a filling orbit which traverses a stripFing foil and a Itorage orbit which does nvt. )2 A boxcar 1n the filling orbit ~ould be fillec \o"ith 154 packets, the firat 77 normally and the next 77 interleaved by actuating a 22 ns dogleg 1n thE injection line. A fast rise time kicker would then de-flect it tv the Horage orbit and the TRII.11f _extraction F:ates woul e delay 215 ns to allow filling of the next boxcar. This IcheQe reduces the kicker freGuency to 30e Hz and the nu~ber of foil traversals to a maximu~ of 15~ (. The average number of traverlals would be <25~ of tt:is value increasing the transverse e~ittance in ~uacrat u re by 15 ~~rad to Ii mm-mrad, still suit-able fer the 32 ~-mrad acceptance of the accumulator. External Phase Expa nsivn Ring 1. n:s,;u, Ar. alternative scheme 13 ~hich could compact many more turns into one packet reducing the number of packets and their rate of arrival at the accumulator, would be to build an auxiliary .torage ane compaction rint. Such a rin~ (TRIS70R) has been proposed with TRIl~f raeius anc an operating energy of 430-440 MeV. The energy gain/turn falls from 3.5 k\' at 430 t;e\- to 0.5 H at extraction to provide phale expansion and turn com-paction. The ring would receive H- cw from TRIl~~ and co~;:act 12,000 turns vithin 7 Me\' (33 _) and tioS· vith a 3000 turn Kap before the next packet to allow clean extraction. The repetition rate of the accumulator injection kicker would thul be reduced to 300 Hz. Ten proton packets would be injected into the accumulator on each cycle, filling every other 46 MHz bucket and allowing 32 ns between bunches for the kicker rise time. ~ stripping il relerved for injection to the storage ring where it would be elsential because of the high turn density. The chief problems with the design are the high degree of i.ochronii~ required in the ring (an order of magnitude better than in TRIl~F) and clean extraction from TRIl~f of 125 \IA H- ions within ~ •• 3 -10 -s o S 10 R peri ...... - R .... r1wr ... (em) Fig. 6. Extraction without stripping; deviation of the deflected orbit from the unperturbed orbit. The bea: ellipse i. plotted every 30· of cyclotron azi~uth. 2. Continous H- Extraction frolt TRln:f In a acheme pre.ently under .tudy the last turn (430 Me\') 15 perturbed rac!1ally with two electrostatic deflectors to deflect the bee: into a magnetic chann~l. Figure 6 showl the deviation from the normal equilibri-um orbit of a beam elliple palsing through the t~ de-flectors. The fint deflector of 60 H/clI: and I II: length pUlhei the beam out radially. At the second deflector 87.5' dovnHreall: the orbit 11 displacee by 2 cm. Th15 deflector, of -60 ltV/a and leneth 1 11:, pushes the bea: inward and after revolving -210· te the magnetic channel (vr -I.5) the beam 11 at its maximuc raCial dilplacement of over 8 Cit. The larBe disrlace-ment, should pe~it us to reduce the positive voltaf~ on the first deflector which may be more difficult to sustain. The field change from a PAssive magnetic channel (iron pipe) is too larBe to be co:pensatee for by the ex16ting nUn:f trim coill. \.:e woule use ar active channel ~ith windings arrangee to proQuce only Imall field changes at inner radii.l~ Since 'Wr-1.5 at 430 Me\" a coherent olCilluion e~ual to the normal turn .pacing (1.5 m:) woulc! modula t ~ tre turn spacing to give a maxim~ jurp of 3 mr. The 6 mr wide 100 ~ IRll,.IF bUill would have to be lignificantly reduced in order to clear the first .ept~. A pre-septum stripping foil could direct unaccertable n- 10ns to a uleful dump or beall: line. References 1. J.R. Richardlon et al., IEEE Trans. ~S-:2 ( 3), 14C2 (1975) • 2. R.E. Laxdal, F.. Lee and C.H. Mackenzie, n.r::n internal report TRI-D~-S2-10. 3. M.K. Craddock, TRIL'MF internal report TRl-t'~: -E2-11. 4. 1.:. Joho, Particle Accelerators 6, 41 (19 74) . 5. K.L. Erdman, private communication. 6. K. Halbach et al., Proc. 1976 Pr6ton Linear Acc. Conf. (CRh"L, Chalk River, 1976), p. 122. 7. D. Dohan and T. Enegren, private com:unication. 6. J.R. Richardson, private comll!unication. 9. L.C. Teng, TRIL1IF internal report TRl-t-K-Sl-16. 10. T. Katayama, private com:unication (19S2). 11. R. "artman, R.E. Laxdal, C.H. Hackenzie and ¥..K. Craddock, laon ractory - Neutral leam Injection, TRIl'MF internal report. 12. L.C. Teng, TRIlJllf {nternal report TRl-t'~-S2-25. 13. w. Joho, TRIl~ internal report TRl-th-E2-14. 14. R.E. lerg and H.C. Blosler, IEEE Trans. ~S-12, 39~ (1965). .~ __________ . _____ -'APPE..t:mJX II ____ _ . __ _ ON' -FACTORY- WITH ' TRIUMF- AS 'NJECTOR . . ! . : . . l. C. Teng Fermi National Accelerator Laborat ory· P.O. Box 500 Batavia. IL 60510 Introduction : I I . Beam Packet from TRIUMF ! I ; I ! A kaon factory is defined as a proton accelerator lin the energy range of 15-30 GeV and the average cur-rent range of 10-100 ~A such that it can produce kaons copiously. A linear accelerator is certainly capable bf this performance but tends to be rather costly. At : ~he current unit cost of -10 eV/S a 15 GeV proton linac ' Would cost well over one billion dollars i L For a circular accelerator one can get higher I . am current with fixed magnetic field. Fixed field ~lternating gradient (FFAG) accelerators have been : studied exhaustively by the Midwes~ern Universities Re- , ~earch Association l. The microtron also uses a fixed : ~gnetic field to recirculate the beam many times : ~hrough a linac. These studies indicate that for ener- ' Qies above 15 GeV all types of circular accelerators . kith fixed field are very difficult technically and ~u1d also be very costly. We are, thus, left only with the pulsed field circular accelerator, namely the ~ynchrotron. This is not surprising if one remembers ~hat hi stori ca l1y the synchrotron was invented to extend ~he energy into the GeV range and the alternating gra- ; ~ient synchrotron made energies much higher than 10 GeV. realistically attainable. The average beam current of : noo ~ is just poss i ble with a fast cycling synchrotron; ~ay, pulsing at 30 Hz and with 2xl013 protons per pulse. Jhe design of such a synchrotron with a linac injector i(high current pulsed accelerator) is straightforward3. po obtain .a long beam sp~ll fo~ ex~eriments one needs i& beam-spll .l stretcher nng WhlCh ,s a d.c. storage ~ing having the same radius and installed in the same ~unnel of the synchrotron. Accelerated pulses of beam ~rom the synchrotron are injected and stored in the ; ~tretcher ring to be spilled out uniformly in t'me by a' resonant slow extraction system. The stretcher ring is : ~n ideal application for superconducting magnets. . I To use a CW accelerator, e.g. a cyclotron such as ~IUMF as injector one needs an accumulator ring which lis an injection-energy d.c. storage ring, again having ~he same radius and installed in the same tunnel of the ~ynchrotron. The Cw beam from the injector 1s accumu- ; ~ated 1n the accumulator during one complete cycle of i ~he synchrotron and is transferred to the synchrotron ! lin one turn to be accelerated as one pulse. The timing , tween the 3 rings -- accumulator, synchrotron and I tretcher -- is shown in Figure 1. Stacki ng a large number of turns of beam during one synchrotron cycle is pOSsible only with charge ex-change injection. Thus, the beam must be extracted from TRIUMF as H-. Even then, stacking in the accumu-lator is greatly facilitated by pre-stacking in TRIUMF on the extraction orbit. Approximately 100 turns of .beam can be stacked in the longitudinal phase space :(energy vs. rf phase) using either the rf stacking4 or the field-bump stacking5 schemes. The completed stack is then extracted vertically as one beam packet by a ,fast kicker. The parameters of the beam packet are: Kinetic energy 450 MeV Number of turns/packet 100 TRIUMF revolution frequency 4.6 MHz TRIUMF harmonic no. c beam bunches/packet 5 I Longitudinal emittance Phase spread Aq. Momentum spread t:.pjp Transverse emittance Horizontal Vertical 811 lIITl-mrad 2'11 lIITl-mrad Accumulator and Accumulat i on Scheme We consider an accumulator (hence also synchrotron and stretcher) with a circumference 10 times that of the extraction orbit of TRIUMF and an rf harmonic num-.ber 100. The revolution frequency of the accumulator is then 0.46 MHz and the rf frequency is 46 MHz, twice that of TRIUMF. The injected beam bunches are syn-chronously captured into every other rf buckets . The scheme proposed for beam accumulation and injection is as follows. i Some 150 packets of H- extracted from TRIUMF are stacked in the transverse phase space by charge ex-change stacking to form a "box-car" of protons. All during this time the proton beam travels on an orbit in t~e accumulator which passes through the stripper foil (the stacking orbit). The completed box-car is then switched onto and stored on another orbit in the same accumulator which does not pass through the foil (the storage orbi t). Ten box-cars are sequentially strung end-to-end to fill the entire storage orbit. The train of beam box-cars is then fast transferred to the synchrotron in one turn. For this scheme to work one must check first that the foil scattering during the stacking of a box-car is not excessive. Secondly, one has to show that two separate closed orbits can be Simultaneously contained in the same ring. In princi-~le. one can always use a second ring instead of the second orbit. But the addition of a fourth ring may be ~conomically pr emotionally untenable. I figure Timing between the accumulator ring, the synchrotron, and the spill stretcher ring. *Operated by the Universities Research Association. Inc. under contract with the U.S. Department of Energy. I Foil Scattering ! i I I ' To obtain a good stripping efficiency (>98:) we need a 200 Ug/cm2 thick carbon foil stripper6• The ~umber of beam packets to be stacked in one box-car is 4.6 K~z/(100xlOx30 Hz) i 153. The first protons will hav~ made -1~30 revolutions by the time the last protons enter the accumulator. On the average each proton will pass through the foil 1530/2 • 765 times or a total thickness of 0.153 g/cm2• This is actually an overestimate because after having been blown up by foil scattering the p beam will be larger than the foil I ,'WhiCh is just the size of the initial H- beam. Hence protons in the beam will not pass through foil every turn. The rms scattering angle due to multiple Cou-'10mb scattering in 0.153 g/cm2 of carbon (radiation length • 42.7 g/om2) at 450 MeV (pv • 754 MeV) is e 14.1 MeV ' to.T53 (1 11 0.153) - 0 8 d I • 754 MeV '142.'7 ' +9 og42:'7 • • mra. I If the ring magnet lattice amplitude function at the fo11 is 6foil a 10m this scattering angle will add !OPPosite sides by angles as a ±1.75 mrad. The two 'distorted orbits will both have betatron invariant :(relative to the undistorted central orbit) I I W • 's (2S:~" r · 68 ",-orad. I .The maximum excursions of the orbits are i I I I (t.x) • : rwr-. :41 m max ,,···max .! only 1I8fo11 ,2 a 71111111-mrad in quadrature to the trans 'verse emittance. Since the space charge limit for 2xl013 protons/pulse requires an emittance larger than -20 1I11111-mrad this increase in emittance is inconse-quential. ~nd occur diametrically opposite separator the excursions are the separator. At the I I 1 I (t.x)s • (t.x)max COS1l\l • :35 mm. Formation of Two Closed Orbits in the Accumulator Ring For the accumulator which has a circumference of 478.8 m we assume a simple magnet lattice made up of 32, 88 0 separated function FOOD cells. The choice of lattice is not crucial. We only need a specific ex-ample for ease of discussion. The only parameters we will use are ~his is enough to accommodate the half-widths of two 30'11 lI111-mrad emittance beams with ample c1earar.ce for Half cell length Length of dipole in half cell Phase advance per cell Number of cells Betatron tunes ~ximum amplitude function Minimum amplitude function 1 • 7.48 m 1B • 3 m ~ " 88 0 N • 32 \lh=\lvi\l • 7.82 Bmax i 25 m Smin i 4.5 m the separator septum. of the septum and the I The cross-sectional geometry beams is shown in Figure 2. I I . I I I , I I , STORAGE ~ ',...'--7 ew----I STACKING cs To form two orbits in a horizontal plane we place ! !a separator at a horizontally focusing Quadrupole ! iwhere e :: 6s • &max • 25 m. The separator is an elec- ' :trostatic septum 60 em long and producing fields of : Fioure 2. '~2 kV/om on opposite Sides. One can also use a cur-lrent septum of the same length producing magnetic fieldS_ of :100 G. The separator kinks the orbits on Cross-sectional geometry of the separated orbits and beams at the location of the electrostatic septum. The transverse emit-tances of the beam are taken to be 30 To rrrn-mrad. OIPOLE -" % :r u 00 t Fi INCOUING . 101- 0"81T S~IPP£R FOIL D, ~--+-----~~i~------------" III FfguM! 3. 1L o I Z ~ ... Geometry of the stacking and the'stJrage orbits and beams at a location diametrically opposite to the separator, showing the charge-exchange stripper foil, the orbit-switching kicker. and the extraction kicker. The qu~drupoles '--_______ -:-..IU.Jii...-lo..LoiI.>&l!.Il"""'-.A.c....b.L1h ilLAnd-thueam eDvel.cpes_Are...Appr.QX ima tecLby -s tra i ght . .l illes. _ -2 Dz Injection Geometry We denote the focusing quadrupole diametrically opposite to the separator by F} and the downstream , quadrupoles sequentially by 01, F2, 02 etc. The strip~ per foil could be placed between Do and FI, across the I "outside orbit (stacking orbit) as shown in Figure 3. I The 3 rn dipole should be cut up into 2 sections 1 m ,and 2 m in length respectively and the foil is placed lin a 25 cm spacing between the two sections. The in- I cident H- beam clears Do on the outside and is deflec- i ted onto the stacking orbit by the 1 m section of di- ! pole. At the stripper we have Bh :: 10 m and Bv :: 13 m.1 The cross-sectional dimensions of the H- beam, hence I also of the stripper foil are 18 rrrn(h) x 10 rrrn(v). But I in order to support the foil it may have to be larger. I As discussed earlier these S values at the foil yield acceptab 1 e emittance growths due to foi 1 . scattering. The fully stacked beam box-car is switched over to the storage orbit by a fast kicker magnet placed just upstream of Fo where the two orbits cross and where Sk-= Smax (Figure 3). The angle to be kicked is ek • 2;;; • 3.3 mrad. he kicker should be 1 m long, and have a peak field 0 ; 113 G and a rise time shorter than the separation be- I en rf bunches in TRIUMF, namely 2 accumulator rf periods or 43 nsec. It may be better to stack 153 (10/9) z 170 beam packets in one box-car and string ~n1y 9 box-cars together with gaps 3 accumulator rf I~riods long. The rise time of the kicker could then ! ~ 65 nsec. Clearly by reducing the number of box-cars i ,further one can accommodate even longer kicker rise i times. This kicker should flat-top for -180 nsec and i should operate at a pulse rate of -300 Hz. Extraction from the Storace Orbit The extraction kicker should be located at Fl (Sei! Figure 3). It can span both the stacking and , e storage orbits, but the separation between the &ms at Fl is large enough to accommodate the side rame of a kicker which spans only the storage orbit. e specifications of this kicker could be the same as Ith,se ,f the-"bit switchi,. kicke'. except the flat-~op should be -2.2 usee. The kick angle of ee • ek. I :3.3 mrad at Fl will give a horizontal displacement at iF2 of I i (bX)e • Smax ee sinu • 83 mm. ~h;s ;s ample for the beam to clear a current septum and enter an extraction channel which further deflects the beam out of the accumulator. It is convenient to ~eave out the ~ipole in this half-cell F2 to 02. To cancel the di spersion caused by this omission we should , omit another dipole half a wave length or 2 cells up-pr down-stream, i.e. in either half-cell Fo or Do or half-cell F4 to 04' To obtain the minimum 2-fold symmetry in t he ring geometry one should omit also the : ~iametricall y opposite dipoles. These empty half-cells : provide convenient locations for rf cavities. i ! The ring magnets of the accumulator will naturally ! 3 [ave to have a wide aperture. Good field aperture of ! 4 cm(h) x 6 cm(v) is adequate. This aperture will I ccommodate beams with 30 ~ mm-mrad transverse emit- I ~ances. Aside from the special problem created by in- I jection from a cyclotron, TRIUMF, the design of the ; fast cycling synchrotron and of the superconducting ~tretcher ring is standard and needs no elaboration. 2. 3. 4. 5. 6. References K.R. Symon, D.W. Kerst, L.W. Jones, L.J. Laslett and K.M. Terwilliger, Phys. Rev. 103, No.6. pp. 1837-1859 (1956) See e.g . A. Roberts, Ann. of Phys. 4, pp. 115-165 (1958) -See e~g . L.C. Teng, Proc. of the Workshop on Nuclear and Particle Physics at Energies up to 31 GeV (January 1981) p. 477, Los Alamos Scientific Laboratory Report LA-8775-C W. Joho , Particle Accelerators 6, 41 (1974) R. Laxdal, R. Lee, G. Mackenzi~ TRIUMF design note TRI-ON-82-l0; M.K. Craddock, TRIUMF design note TRI-DN-82-ll R. Baartman; R. Laxdal, G. Mackenzie, MKaon Factory,: Neutral Beam Injection". TRIUMF draft design note (22 October 1981) .. . . . 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items