LIBO—a linac-booster for protontherapy: construction and tests of a prototype

https://doi.org/10.1016/j.nima.2003.07.062Get rights and content

Abstract

LIBO is a proton accelerator that operates at 3GHz, the same frequency as the one adopted in the about 7500 electron linacs used for radiotherapy all over the world. Such a high frequency was chosen to obtain a large gradient (on average more than 10MV/m), and thus a short linac (about 15m) to boost the energy of the protons, extracted at about 60MeV from a cyclotron, up to the 200MeV needed for the treatment of deep-seated tumours.

This paper describes the design study of the full 3GHz Side Coupled Linac (modular structure, nine modules) and the construction and tests of the LIBO prototype (first module), which was built to accelerate protons from 62 to 74MeV with an RF peak power of 4.4MW. The items discussed are the beam dynamics parameters of the module (longitudinal and transverse acceptances), the constructional elements and procedures, the accuracies of the various mechanical elements, the cooling system, the RF tuning, the RF measurement and the RF power tests. These tests showed that, after a short conditioning time, the gradient in each of the four tanks of the module could reach 28.5MV/m, much larger than the nominal project value (15.8MV/m). The last section of the paper describes the successful acceleration tests performed at the Laboratori Nazionali del Sud of INFN in Catania with a solid-state 3GHz modulator lent by IBA.

Section snippets

Global status of protontherapy

About 7500 linacs, accelerating electrons to energies in the range from 5 to 25MeV, are used in the radiotherapy departments of most large hospitals. This corresponds to about half of the accelerators operated around the world (data by Waldemar H. Scharf and collaborators quoted in Ref. [1]). For comparison, only 25 cyclotrons or synchrotrons are used for tumour treatment with protons and/or carbon ions, an irradiation modality now known as ‘hadrontherapy’.

This technique has greatly developed

The 3GHz proton linac

In 1993, one of us (U.A.) initiated the study of a novel proton linear accelerator based on the same high radio frequency (f=3GHz) as used by thousands of electron accelerators running in hospitals all over the world. This high frequency implies a linac more compact and shorter than the standard lower-frequency proton linacs, used as injector of most synchrotrons, since the permissible accelerating field is roughly proportional to f1/2 [10]. The small iris of a linac of such a high frequency is

The first design of LIBO

As mentioned above, a 60–200MeV SCL linac has a wide range of applications, since it opens the possibility to transform a cyclotron (which is useful for eye therapy, isotope production and low-energy physics research) into a 200MeV accelerator for the treatment of deep-seated tumours. An artist's view of LIBO is shown in Fig. 3.

A 3GHz LIBO would have a duty cycle of the order of 0.1–0.2%, corresponding to a repetition rate of 200–400Hz and a pulse duration of 3–5μs. The cyclotron frequencies

The LIBO prototype

At the end of 1998, it was decided to build and test a prototype of LIBO-62. For that, the first module, the most critical one, that has to accelerate protons from 62 to 74MeV, was chosen. It requires an RF peak power of about 4.4MW, at 2.998GHz. The construction of this ‘prototype module’ was completed in 2000, after which full RF power tests were performed at CERN. Fig. 10 represents this LIBO prototype. Details of construction and tests are also given in [18], [19].

Among the basic elements

RF aspects

The design of LIBO is based on a mean accelerating field value E0 on axis which is identical in all the tanks. To achieve this in the prototype, the accelerating cells increase in length from tank to tank (conforming to the increasing velocity of protons), while the dimensions of the coupling cells and of the coupling slots in all the half-cell plates of the module do not vary. As consequence, the corresponding coupling coefficient (≅4.3%) decreases ∝β−1/2, as required to have E0=const.

The

RF power tests

To adjust the LIBO accelerating field to the correct value during the power tests, the total shunt impedance rs of the module has to be known. This impedance was determined prior to the tests by means of an innovative combination of perturbation measurements and SUPERFISH computations.

As perturbative element a nylon wire was used, stretched along the axis of the whole accelerator. The dielectric constant of the wire had been determined beforehand in a parallelpiped test cavity with a known

Preliminary acceleration tests

In 2001, the LIBO prototype was transported to and installed at the Laboratorio Nazionale del Sud (LNS) in Catania to undergo tests with a proton beam. A view of LIBO placed in the beam line of the INFN superconducting cyclotron is shown in Fig. 18. The proton beam delivered by the cyclotron had an energy of (62±0.2)MeV and the operating value of the average current at the LIBO entrance was of the order of 1nA.

Conventional diagnostic systems like Faraday cups and alumina screens were placed

Conclusions and perspectives

A prototype module of LIBO has been built and successfully tested at CERN and LNS, where it has accelerated 62MeV protons to an energy of 73MeV, in agreement with the results of the simulations. The working principle of LIBO has thus been fully demonstrated.

The results obtained also strongly indicate that higher accelerating gradients can be applied. For a new design one can aim at an average accelerating field of 20MeV/m. Moreover, elaborating the construction technique applied for the present

Acknowledgements

For the construction at CERN of the LIBO prototype and the RF tests we acknowledge the help of D. Allard, R. Bossart, A. Catinaccio, Ch. Dutriat, J.C. Gervais, S. Haider, D. Leroy, M. Mezin, J. Mourier, M. O'Neil, B. Pincott, Ph. Potdevin, L. Rinolfi, G. Rossat, J. Stovall and G. Yvon. The care of Serge Mathot in the alignment and brazing of the components of the module has been instrumental to the success.

At the PS division of CERN, the RF, PO, PC and PP groups put several pieces of equipment

References (23)

  • A. Zurlo

    Int. J. Radiat. Oncol. Biol. Phys.

    (2000)
  • E. Pedroni

    Med. Phys.

    (1995)
  • U. Amaldi

    Nucl. Phys. Meth. Phys. Res. A

    (1999)
  • C. De Martinis, et al., Beam tests on a proton linac booster for hadrontherapy, Proceedings of the European Particle...
  • U. Amaldi

    Europhys. News

    (2000)
  • R.R. Wilson

    Radiology

    (1946)
  • J. Sisterson (Ed.), Particles No. 27, January...
  • G. Kraft

    Strahlenther. Onkol.

    (1990)
    G. Kraft

    Prog. Part. Nucl. Phys.

    (2000)
  • A. Brahme

    Nucl. Instr. and Meth. B

    (2001)
  • Th. Haberer

    Nucl. Instr. and Meth. A

    (1993)
  • W.D. Kilpatrick

    Rev. Sci. Instrum.

    (1957)
  • Cited by (0)

    1

    Also at University of Milano Bicocca.

    View full text