RFD Linac Structure
|
|
|
|
RFD Linac Structure |
|
|
The particles, traveling along the axis of the linac, traverse two distinct regions, namely gaps between drift tubes where the acceleration takes place, and regions inside the drift tubes where the rf quadrupole focusing takes place. This structure uses both phases of the rf fields to affect the beam; one for accelerating the beam and the other for focusing the beam. In this case, the "reverse phase" does not decelerate the beam because the fields inside the drift tubes are distorted into transverse focusing fields with little longitudinal component. The orientation of the fingers in the focusing regions alternate so as to create an alternating focusing and defocusing action on the beam in each transverse plane.
As originally conceived and shown in Figure 1, the rf quadrupole lens electrodes were supported through thin rings of ceramic from a water cooled drift tube body, supported from the tank wall on a single stem. Subsequently, an inductive stem concept was developed where the lens electrodes were isolated from each other by an inductive feature of the support stem. The current design calls for the two lens electrodes to be supported on two water-cooled support blades, which emanate from a single stem, which mounts in a single hole through the linac tank wall. Two views of an RFD drift tube, supported on a "two-bladed" drift tube stem, are shown in Figure 2.
The drift tube lens electrodes are capacitively coupled to the longitudinal electric field of the linac, while the two-bladed drift tube support stems are inductively coupled to the transverse magnetic field of the linac. The lens excitation is controlled by controlling the magnitude of these couplings.
As originally conceived, this design still offers a complete drift tube package supported from the tank wall at a single point. The relative alignment of the four fingers can be assembled and inspected "on the bench" to insure adequate alignment. This preserves the simplicity of installation and alignment of our original configuration.
The stem/blade assembly will be a 4-layer stainless steel sandwich requiring two hydrogen-furnace braze cycles for completion. In the first braze cycle, two halves of the blades will be joined together to form a support blade complete with a machined cooling channel. In the second braze cycle, two blade assemblies will be brazed together to form a complete stem/blade assembly with connections to the parallel blade cooling channels. The stem portion of this assembly will be machined to a circular cross section and precision ground to a cylindrical surface for installation into a reamed hole in the tank wall. The stem and hole will be keyed to control the angular orientation of the drift tube. At this point, portions of the stem/blade assembly will be copper plated.
To achieve the required precision in the location of the body halves relative to the stem base, the stem/blade base will be held in a precision jig, coolant will be circulated through the stem/blade assembly, and the inner surface of the circular annuli at the end of the blades will be cut by the wire EDM process to form a precision seat for the body halves, accurately located relative to the precision ground cylindrical surface of the stem base.
The body electrodes are made from oxygen-free copper. The body electrodes for each drift tube will be different to accommodate the changing velocity of the accelerated particles, whereas the two halves of each drift tube will be identical. The drift tube half-bodies are attached to the stem/blade assembly by a third and final hydrogen-furnace braze cycle.
The beam bunches arrive at the centers of the gaps at the times when the electric fields are optimum for acceleration, and the beam bunches arrive at the centers of the drift tubes one half cycle later when the electric fields have reversed their directions and are suitable for focusing the beam. The upper part of Figure 3 shows the RFD linac structure, with greatly exaggerated finger spacing, showing the distribution of electric fields (arrows) and electric charges (+ and - signs) within the structure at the "acceleration phase". The lower part of Figure 3 shows the same structure with the field directions and particle bunch locations as they would be at the "focusing phase". The directions, shown for the fields inside the drift tubes, pertain only to the component of the fields in the plane of the Figure. The components of the fields normal to the Figure are in the opposite direction relative to the axis. Particles traveling along the axis experience no focusing force, as the transverse fields vanish on the axis. Off-axis particles A and B will experience a "focusing" action on their motion, while particles C and D will experience a "defocusing" action on their motion.
A PARMILA-like beam dynamics code, PARMIR (Phase And Radial Motion In RFDs), was written to facilitate the study of the beam dynamics in this new linac structure. PARMIR simulates multiparticle beam dynamics in drift tube linacs that employ rf focusing inside the drift tubes. The formulation includes hard- and soft-edged quadrupole fringe fields and duodecapole effects. This code has been used extensively in our studies of the RFD structure.
Useful information of the performance of these structures can be obtained from the well known linear beam dynamics code, TRACE-3D. The RFD structure can be described to TRACE by using three types of elements: an RFQ element with no acceleration, a drift element, and an rf gap element. Even though it cannot simulate nonlinear fields or space charge forces, it can yield valuable information on the properties of the matched beam in the structures and a measure of their relative performance.
The RFD linac structure provides a graceful way to accelerate the small diameter, tightly bunched beams that come from RFQ linacs to higher energies. Because of the rf electric focusing, the RFD linac structure operates well at much lower energies than the conventional magnetically focused DTL. Consequently, the transition energy between the RFQ linac, required to capture the unbunched beam from the injector, and the RFD linac can be significantly lower than for conventional RFQ/DTL combinations, perhaps in the 0.5 to 1 MeV range. We believe that this new structure will become the structure of choice to follow RFQ linacs in many applications.
At these lower energies, the acceleration efficiency of the RFQ is relatively high and similar to that of the RFD . Likewise, the transverse focusing (and beam size) in the RFD is similar to that in the RFQ. Consequently, no complex matching section is required between the two linacs.
RFD linacs will have the same small diameter beams found in RFQ linacs. These small diameter beams allow higher frequency operation, which in turn, implies smaller, lighter-weight, and more efficient linac structures, which in turn, means less rf power to generate, less thermal load to cool, and less surface area to evacuate. The higher frequencies also offer the possibility of higher gradient operation (shorter structures). These compact linac structures will be more transportable than their predecessors, easier to shield, and less expensive to procure, operate, and maintain.
A very important advantage that the RFD structure has over the RFQ structure is acceleration efficiency. A comparison of the rf efficiencies for the RFD and RFQ linac structures, expressed in terms of shunt impedance in megohms/meter, is shown in Figure 4. In the range of 1-to-5 MeV, the RFD structure has approximately 4 times the shunt impedance of the RFQ structure. Note that, in the range where we propose to use the RFQ (up to 0.8 MeV), the RFQ still has a respectable shunt impedance.
This new linac structure will lead to the development of a new class
of ion linac systems for a whole host of ion beam applications. These new
linac systems will enhance the reliability and reduce the cost, and complexity
of these sources of radiation. Applications, which formerly were out of
range for linacs, may now find that they are within the range of practicality
for linac-based systems. This new linac structure could be as revolutionary
to accelerator technology as the conventional RFQ was on its introduction
to the world in 1980.
|
