Shielding data for 100–250 MeV proton accelerators: Attenuation of secondary radiation in thick iron and concrete/iron shields

https://doi.org/10.1016/j.nimb.2008.05.002Get rights and content

Abstract

Double differential distributions of neutrons produced by 100, 150, 200 and 250 MeV protons stopped in a thick iron target were calculated with the FLUKA Monte Carlo code at four emission angles: forward, 45°, transverse and 135° backwards. The attenuation in thick iron shields of the dose equivalent due to neutrons, protons, photons and electrons was also calculated. The contribution to the total ambient dose equivalent from photons and protons is limited to a few percent at maximum. Source terms and attenuation lengths are given as a function of energy and emission angle, along with fits to the Monte Carlo data, for shallow depth and deep penetration in the shield. A brief discussion of simulations performed with composite iron/concrete shields is also given, showing the need for further investigations.

Introduction

A recent paper [1] has revised, updated and complemented source term and attenuation length data for concrete (the most common material used in accelerator shielding) for 100–250 MeV proton accelerators, as a function of proton energy and emission angle. Accelerators in this energy range, in the past employed for nuclear physics research, are now mainly used as injectors to high energy accelerators at large research centres but also find increasing application in cancer radiation therapy (hadron therapy) and are more and more frequently installed in large medical centres. Shielding of these accelerators is a critical issue, either because of the large beam intensity required as injectors or, in the case of radiation therapy, because hospitals are located in highly populated areas. The main radiation to be shielded is the neutron field produced by the interaction of the proton beam with the structures of the accelerator, of the beam transfer lines and – in the case of medical machines – of the beam delivery system used to irradiate the patient (such as collimators and field-shaping devices) and with the patient himself (where the remaining beam is ultimately lost).

In designing accelerator shielding a first assessment of the required shield thickness is often performed by using a simplified model, to be verified at a more advanced stage of the project with a Monte Carlo simulation in a more realistic geometry of the facility. Handy data for simplified calculations are given for instance in [2]. Often this simplified approach provides a conservative estimate of the required shielding [3].

This paper extends the previous work [1] providing shielding data for iron, as very few are available in the literature (either experimental or computational). It is well known (see for example [4], [5], [6]) that iron alone is not a very effective shielding material for neutrons. Its main property is to slow down neutrons with energy above 847 keV (the first excited nuclear energy level of 56Fe) via inelastic scattering reactions, whilst below this threshold neutrons can only lose energy via elastic scattering on the dominant 56Fe nucleus, a very inefficient mechanism. The dose equivalent past an iron shield is thus dominated by neutrons of energy less than about 1 MeV. However, because of this same property and of its high density, iron becomes more effective when backed by a comparatively small amount of concrete (or other hydrogenated material), which absorbs (most of) the surviving lower energy neutrons. This paper provides shielding data for iron (source terms and attenuation lengths) for secondary radiation from 100 to 250 MeV protons. What finally counts in a shielding design is its effectiveness in dose equivalent attenuation and the amount of space it requires. For this reason, in this paper comparisons between concrete and iron shields are always made in terms of linear dimensions rather than mass thickness.

The attenuation through an iron shield of the total dose equivalent produced by 100, 150, 200 and 250 MeV protons stopped in an iron target thicker than the proton range were calculated with the FLUKA Monte Carlo code [7], [8] at four emission angles: 0° (forward), 45°, 90° (transverse) and 135° (backwards). The present calculations simulate the dominant secondary radiation field created by a beam loss in a thick metallic target such as a magnet, a collimator or a vacuum chamber. Iron was chosen as target element as it is representative of other materials of similar density and atomic number (such as copper and stainless steel), which are the main constituents of accelerator components. Shielding data for backward angles may also be of interest to account for special conditions found in modern hadron therapy facilities, where the beam extracted from the accelerator can be rotated 360° around the patient by means of a large mechanical structure (isocentric gantry). The ambient dose equivalent, H(10), behind the shield includes contribution from neutrons, photons, charged particles (protons and electrons) and their secondaries produced in the shielding material itself. The results of the calculations were fitted by the classical two-parameter formula of a point-source line-of-sight model. As mentioned above iron is often used in combination with concrete, and the present paper also investigates the effect of composite iron/concrete barriers versus pure iron or concrete ones.

Section snippets

Monte carlo simulations

The simulations were performed with the version 2006.3 of the FLUKA [7], [8] Monte Carlo code, with the identical geometry set-up and the same settings used in the previous paper [1], except that concrete was replaced by iron (Section 3) as shielding material. Neutrons, photons, protons and electrons produced by a monoenergetic and monodirectional proton beam impinging on a thick iron target located at the centre of a large spherical shield made of iron were transported in four angular bins

Shielding material: iron

In FLUKA, the group structure used for neutron transport below 19.6 MeV is necessarily coarse with respect to the resonance structure in many materials. If an isotope is very pure and is present in large amounts (for instance, iron as shielding material, as in the present case), it can act as a “neutron sink”, causing sharp dips in the neutron spectrum corresponding to each resonance: this is called “self-shielding” effect.

In order to make adequate corrections, two additional neutron cross

Energy and angular distributions of source neutrons

In a previous paper [1] the double differential distributions of neutrons produced by 100 to 250 MeV protons impinging on the same thick iron targets were calculated with the FLUKA code. For every spectrum the statistical uncertainty associated to each energy bin was analysed, giving particular attention to the highest energy bins for which the uncertainties were higher than 20%, due to the importance of these bins in deep penetration problems. The same analysis performed for the present

Shielding parameters

The attenuation of the total ambient dose equivalent through a thick shield can be fitted with the classical two-parameter formula (see, for example, [4]):H(Ep,θ,d/λ)=H0(Ep,θ)r2exp-dλ(θ)g(α)where H is the ambient dose equivalent beyond the shield, Ep the proton energy, r the distance between the radiation source (the target stopping the protons) and the scoring position, θ the angle between the direction r and the beam axis, H0 the source term, d the shield thickness, λ(θ) the attenuation

Shielding data for iron

The attenuation plots of the total dose equivalent in iron at the four emission angles are shown in Fig. 3, Fig. 4, Fig. 5, Fig. 6 for 100, 150, 200 and 250 MeV protons. The statistical uncertainties on the data points are less than 4% for all proton energies and angular bins, the maximum values always being at the largest depths in either the forward direction or in the 40°–50° bin. This is because most of the ambient dose equivalent is due to neutrons in the 1-MeV range. In general the small

Comparison with literature data

Shielding data for iron are scarce in the literature, and are always found for specific simulation geometries and experimental set-ups. The only source for a comparison is [11], where the attenuation of the dose equivalent through shields up to 6 m thick was simulated in a slab geometry, in order to compare results from different simulation codes. A comparison with the present calculations can only be made on the attenuation length in the transverse direction (80°–90° angular bin) for 200 MeV

Comparison between concrete and iron shielding barriers

In order to better understand the difference in neutron attenuation in iron and concrete, Fig. 18 shows the relative cumulative neutron ambient dose equivalent versus neutron energy past concrete and iron shields of the same thickness (800 g cm−2), for forward emission and 250 MeV protons. This case provides the hardest spectrum exiting the target. Both the iron and the concrete shields are thick enough to lead to an equilibrium spectrum. High-energy neutrons contribute very little to the dose

Composite shielding barriers

As iron is often used in combination with concrete as shielding material, the effect of composite iron/concrete shields versus pure iron or concrete barriers was investigated. Three cases were studied with the identical geometry set-up and scoring method, for 250 MeV protons in the forward, 40°–50° and transverse directions. Each shield was made up of an inner spherical layer of iron and an outer spherical layer of concrete: 100 cm iron + 100 cm concrete (A), 200 cm iron + 50 cm concrete (B), 50 cm iron + 

Conclusions

This paper has complemented previous calculations for intermediate energy proton accelerators for concrete shielding [1] to provide counterpart data for iron barriers. Although concrete is the most common material used in accelerator shielding, iron is also used, for which few data are available. The two papers provide a set of source terms and attenuation lengths to be used in the shielding design of intermediate energy proton accelerators, with a view to reduce to a minimum the need for

Acknowledgements

The authors wish to thank Alfredo Ferrari (CERN) for useful discussions.

References (11)

  • S. Agosteo et al.

    Nucl. Instr. and Meth. B

    (2007)
  • A.H. Sullivan

    A Guide to Radiation and Radioactivity Levels Near High Energy Particle Accelerators

    (1992)
  • M. Magistris et al.

    Radiat. Protect. Dosim.

    (2005)
  • R.H. Thomas, G.R. Stevenson, Radiological Safety Aspects of the Operation of Proton Accelerators, Technical report...
  • H.W. Patterson et al.

    Aceelerator Health Physics

    (1973)
There are more references available in the full text version of this article.

Cited by (28)

  • Heavy concrete shielding properties for carbon therapy

    2023, Nuclear Engineering and Technology
  • Calculation of shielding performance of CRT concrete for proton therapy and optimal shielding design of treatment delivery room

    2022, Applied Radiation and Isotopes
    Citation Excerpt :

    When designing treatment delivery room shielding, the required shielding thickness is usually evaluated for the first time using a simplified model, which can significantly reduce the calculation time. This simplified method provides a conservative estimate of the required shielding (Agosteo et al., 2008). Based on our previous study and literature, more than 95% of the reflected neutrons and secondary neutrons are concentrated in the proton incident direction, and the reflected neutrons in the remaining direction can be ignored.

  • Shielding data for hadron-therapy ion accelerators: Attenuation of secondary radiation in concrete

    2014, Nuclear Instruments and Methods in Physics Research, Section B: Beam Interactions with Materials and Atoms
    Citation Excerpt :

    In real situations the beam is usually lost in a rather thick target (e.g. a magnet or a collimator) and thick-target data should normally be used [5]. As done in various previous studies [4,7,8,10,11,20], the target was located at the centre of a large spherical shield made of ordinary concrete (type TFS 5.5 [21], Table 3), the inner radius of which was sufficiently large (90 m) to ensure that curvature-related effects are negligible (Fig. 1). The volume of the cavity delimited by the inner surface of the shield was “filled” with vacuum.

  • Attenuation in iron of neutrons produced by 120 GeV/c positive hadrons on a thick copper target

    2013, Nuclear Instruments and Methods in Physics Research, Section B: Beam Interactions with Materials and Atoms
  • Deep-penetration calculations in concrete and iron for shielding of proton therapy accelerators

    2012, Nuclear Instruments and Methods in Physics Research, Section B: Beam Interactions with Materials and Atoms
View all citing articles on Scopus
View full text