The MYRRHA accelerator is a linear accelerator delivering a Continuous Wave (CW) proton beam of up to 4 mA at an energy of 600 MeV. The beam penetrates the reactor vertically from the top.
The design of the Linac aims at simultaneously maximizing:
- Its efficiency, both in terms of its wall plug electricity consumption and of its geometrical length.
- The beam availability, and hence its fault tolerance capability.
A “superconducting Linac”
The combination of requirements leads to the option of a “superconducting Linac”, indicating a Linac which, above a certain beam energy, entirely consists of a sequence of individually controlled superconducting RF cavities. Below this energy the particle velocity changes rapidly, and then it is more useful to apply multi-cell accelerating cavities. These may be normal conducting (made of copper) or superconducting (made of niobium). The lowest energy part of the Linac, in which beam losses are unavoidable, is definitely normal conducting.
So, the superconducting Linac for MYRRHA has 2 sections, called "the front end" and "the independently phased superconducting section". The transition between the 2 sections is placed at 17 MeV. Below this energy, distributed redundancy is not applicable, and hence the high availability aimed at can only be obtained by a classical redundant doubling of this full section. Above 17 MeV the availability is obtained by the full implementation of the fault tolerant scheme.
Why superconducting cavities?
The choice of using superconducting RF cavities (made of Nb) is based on the following arguments:
The possibility of obtaining higher accelerating fields: Especially in a CW linac, the fields obtainable in normal conducting cavities are limited by their capability of dissipating the ohmic losses. Not so in superconducting cavities, in which the physical limits of the electric fields may be reached. This results in a shorter linac, but more importantly it allows to leave ample margin between the nominal operation points and the operational limits in an economically viable way. This is the situation that is strongly needed both for reliability and for the fault tolerance based on distributed redundancy.
Larger beam apertures: Superconductivity gives the freedom to design RF cavities with significantly larger apertures for the beam passage than in the case of copper cavities. This leads to well reduced beam losses along the linac, and therefore to a reduced activation. For a high intensity machine this is a very important point.
Better energy efficiency: In spite of the overhead due to the cryogenic plant, the absence of ohmic losses strongly reduces the linac's power consumption and the directly related exploitation costs.
The front end
The front end starts with an Electron Cyclotron Resonance (ECR) ion source. Its technology is well proven at significantly higher beam currents (100 mA) than our specification, thereby providing ample margin for reliability improvements.
After a short matching section the DC beam from the source is injected into a RadioFrequency Quadrupole (RFQ), which is used for bunching and initial acceleration of the beam.
The RFQ is strongly focusing and has a perfect beam buncher, but it is a relatively inefficient accelerator. It is therefore replaced by a more efficient accelerating structure as early as possible. A well-suited structure at low particle velocity is of the so-called Cross-bar H-type (CH). Normal conducting CH cavities accelerate the beam up to 5 MeV. At this point the unavoidable low energy beam losses have occurred, and from there on, superconducting structures can be envisaged.
The acceleration to 17 MeV can be obtained through 4 superconducting CH cavities, combined into one single cryomodule together with the necessary focusing elements. With this layout the total length of the front end section will be less than 16 m. The frequency will be 176 MHz or 352 MHz, to be decided. In the former case the RFQ will be of the 4-rod type, in the latter case it will be a 4-vane RFQ.
The independently phased superconducting section
The independently phased superconducting section is further subdivided according to the increasing beam velocity:
- Acceleration from 17 MeV up to 90 MeV is obtained by a series of spoke cavities. The frequency is 352 MHz.
- At this point the RF frequency changes to 704 MHz, and the main part of the Linac, up to the final energy, is made of 2 families of elliptical cavities.
All along this Linac section the focusing quadrupoles are normal conducting magnets, organized in doublets and installed in between the cryomodules that carry the cavities. According to the requirements of the transverse beam optics, each cryomodule houses 2 or 3 independent cavities. The total length of the independently phased superconducting section is estimated at 232 m.
The path to the MYRRHA spallation target is the High Energy Beam Transport (HEBT). This beam line has achromatic optics in order to guarantee the beam stability on target, and it houses the AC magnets which allow scanning the beam on target in the specified doughnut shape. A full power beam dump is foreseen in the alignment of the Linac, allowing for the commissioning of the MYRRHA accelerator fully independently from the reactor.
In the context of ADS, the beam interruption system of the driver accelerator not only has a machine protection function (like in any high power proton accelerator), but it also has a fundamental safety purpose in the ADS as a whole. Fast electronic systems will act at the level of the ion source and of the RF control, whereas slower mechanical systems will be able to physically intercept and block any potentially remaining low energy beam.
General layout of the superconducting linac