Applied Radiation and Isotopes 62 (2005) 69–72

ARTICLE IN PRESS

*Correspond

E-mail addr

0969-8043/$ - se

doi:10.1016/j.ap

A study of neutron spectra from medical linear accelerators

A. Facurea, R.C. Falcaoa, A.X. Silvab,*, V.R. Crispimb, J.C. Vitorellib

aComissao Nacional de Energia Nuclear, R. Gal. Severiano 90, sala 405, 22294-900, Rio de Janeiro, RJ, Brazilb[PEN/COPPE—DNC/EE]CT, Universidade Federal do Rio de Janeiro, Ilha do Fundao, Caixa Postal 68509, 21945-970,

Rio de Janeiro, RJ, Brazil

Received 1 January 2004; received in revised form 5 April 2004; accepted 10 May 2004

Abstract

Medical accelerators with photon energies over 10MeV generate an undesired fast neutron contamination in the

therapeutic beam. In this work, the Monte Carlo code MCNP was used to simulate the transport of these

photoneutrons across the head of various medical accelerators of high energy. The average and most probable neutron

energies were obtained from these spectra, before and after crossing the accelerator shielding. The degradation of these

spectra, when they cross concrete barriers with thickness which vary between 25 and 100 cm, was also studied.

r 2004 Elsevier Ltd. All rights reserved.

Keywords: Photoneutrons; Neutron spectra; Monte Carlo simulation; Linear accelerators

1. Introduction

Nowadays, teletherapy machines of cobalt and cesium

are being replaced by linear accelerators. The great

advantage of this equipment is not having a radioactive

source attached, which makes them safer in the

radiological point of view. The maximum photon energy

in these accelerators can vary between 4 and 25MeV.

When the photons have energies above 10MeV, they

can interact with the atomic nucleus of the high-Z

material, which constitute the target and the head of the

accelerator, and neutrons are ejected. Consequently, it is

important to know the neutron spectrum, which

contaminates the therapeutic beam, both to project the

room shielding as well as to evaluate the increase to the

patient dose.

This work presents a study based on Monte Carlo

simulations of the neutron spectra generated by accel-

erators with photon energies between 10 and 25MeV.

These spectra were obtained before and after crossing

the accelerator head. From this data, the average and

ing author. Fax: +55-21-2562-8444.

ess: ademir@con.ufrj.br (A.X. Silva).

e front matter r 2004 Elsevier Ltd. All rights reserve

radiso.2004.05.072

most probable neutron energies were found and it was

observed that these energies are under the values found

in literature (Swanson, 1980). This is a first step to

complete mock up a radiotherapy facility, with the real

neutron energy spectra emitted by medical accelerators

and actual room and maze configuration. It was also the

studied degradation of initial spectra after crossing

concrete walls of different thickness. Finally, the gamma

capture spectra generated in concrete were obtained.

2. Neutron spectra

Neutrons generated in accelerators can be classified in

two groups: the first has a Maxwellian energy distribu-

tion and are called evaporation neutrons. The second

are direct neutrons, which are produced through direct

interaction between the photon and a neutron in the

nucleus of the target atom. The so-called direct neutrons

are, approximately, 15% of the total produced by the

(photon, n) reactions and their energy are greater than

the evaporation neutrons. Nevertheless, it is the

evaporation neutrons that constitute the greatest part

of the photoneutrons and their spectra, as described by

d.

ARTICLE IN PRESS

0 2 4 6 8 10 12 14 16

10-1

10-2

10-3

25 MV 20 MV 18 MV 15 MV

dN/d

E

Neutron energy (MeV)

Fig. 1. Neutron spectra produced in 15,18, 20 and 25MV linear

accelerators, before crossing its head (source spectrum).

Table 1

Average photoneutron energies for high-energy (15, 18, 20 and

25MV) accelerators

Maximum photon energy

(MeV)

Average neutron energies

(MeV)

15 1.15

18 1.25

20 1.31

25 1.46

A. Facure et al. / Applied Radiation and Isotopes 62 (2005) 69–7270

Tosi et al. (1991), is the following:

dN

dEn¼

En

T2exp

�En

T

� �; ð1Þ

where En is the neutron energy in MeV and T is the

‘‘nuclear temperature’’ (in MeV) to a particular nucleus.

For instance, the corresponding temperature for the

production of neutrons in tungsten is 0.5MeV.

In order to evaluate the relative contribution of the

distinct components of the total spectra, an isotropic

neutron source with the energy spectra described by Tosi

et al. (1991) was considered:

dN

dEn¼0:8929En

ð0:5Þ2exp

�En

0:5

� �

þ0:1071 ln

Emax

En þ 7:34

� �

R Emax�7:340 ln

Emax

En þ 7:34

� �dEn

: ð2Þ

Considering X-ray energies of 25MeV (Emax) the

photoneutron emission spectrum results:

dN

dEn¼ 3:5716En expð�2EnÞ

þ 0:0123607 ln25

En þ 7:34

� �: ð3Þ

In this work the Monte Carlo radiation transport

code MCNP, version 4B (Briesmeister, 1997), and the

Evaluated Nuclear Data File B-VI (ENDF/B-VI)

continuous energy neutron cross section library were

employed to perform the calculations of photoneutrons

transport. In the simulations, the source term was

considered as an isotropic point-like spectrum given by

Eq. (2). The accelerator head was modelled as a 10 cm

tungsten sphere around the source where neutrons are

produced. Other authors have also used the same model

for the accelerator head in their simulation (Agosteo

et al., 1993; Carinou and Kamenopoulou, 1999).

3. Results

Fig. 1 shows the source spectra, obtained from

Eq. (2), for accelerators (15, 18, 20 and 25MV) and

used as input data for the MCNP simulation. The

photoneutron emission spectra have a predominance of

neutrons with energies up to 2MeV. The average

neutron energies of spectra are presented in Table 1,

and the most probable energy is about 0.5MeV (due to

evaporation term) to all spectrum. These values are in

accordance with the ones reported in the literature.

Swanson (Swanson, 1980) reported that the median

energy of the photoneutron spectra is around 1.5MeV

and, in a recent study, d’Errico (d’Errico et al., 2001)

reported effective energies in the 1.8–2.1MeV range for

direct photoneutrons from 10 to 18MeV X-ray beams.

The neutron spectrum produced by linear accelerators

is slowed down in energy as a result of particle

interaction with the accelerator structure. Few neutrons

are lost or have their energy reduced when crossing a

given lead structure, and a 15% decrease in the original

neutron fluence occurs when these particles cross a

tungsten head. In Fig. 2 this degradation, for a tungsten

shielding, is shown.

The weighted average photoneutron energy, %E; foreach spectrum, after filtration by 10 cm of tungsten was

calculated as follows:

%E ¼PN

i¼1 EiFðEiÞPNi¼1 FðEiÞ

; ð4Þ

where Ei is the transmitted neutron energy of the ith

energy interval; FðEiÞ corresponds to its fluence

obtained by MCNP calculations and the sum symbol

is extended to all N energy intervals of the each

spectrum.

According to Eq. (4), the average neutron energy was

found to be about 0.4MeV for the four studied spectra.

The most probable energy also in all cases is about

0.2MeV. These values are in good agreement with the

ones reported by Swanson (Swanson, 1980) and d’Errico

et al. (2001).

ARTICLE IN PRESS

0 1 2 3 6

10-1

10-2

10-3

10-4

10-5

10-6

100

10-1

10-2

10-3

10-4

10-5

10-6

100

10-1

10-2

10-3

10-4

10-5

10-6

100

15 MV

Source spectrum Leakage spectrum

Source spectrum Leakage spectrum

Source spectrum Leakage spectrum

Nor

mal

ized

neu

tron

flue

nce

(n/c

m2 )

Energy (MeV)

0 3 8 10 11 12 13

20 MV

Nor

mal

ized

neu

tron

flue

nce

(n/c

m2 )

Energy (MeV)

0 2 4 6 8 10 12 14 16 18

25 MV

Nor

mal

ized

neu

tron

flue

nce

(n/c

m2 )

Energy (MeV)

4 5 7 8

1 2 4 5 6 7 9

Fig. 2. Neutron spectra produced in 15, 20 and 25MV

accelerators, before (source) and after (leakage) crossing the

tungsten head.

10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100 101 102

10-1

10-2

10-3

10-4

10-5

10-6

10-7

10-8

10-9

100

Epithermal FastThermal

25 cm

50 cm

75 cm

100 cm

Nor

mal

ized

neu

tron

flue

nce

(n/c

m2 )

Energy (MeV)

Fig. 3. Neutron spectra attenuation from a 15MV linear

accelerator (leakage spectrum), after concrete barriers of 25, 50,

75 and 100 cm.

10-1

10-2

10-3

10-4

10-5

10-6

10-7

10-8

10-9

10-10

10-2 10-1 100 101

100

50 cm thick

25 cm thick

Nor

mal

ized

pho

ton

fluen

ce (

phot

ons/

cm2 )

Photon energy (MeV)

Fig. 4. Capture gamma rays spectra produced by neutrons

from a 15MV accelerator, after crossing concrete barriers of 25

and 50 cm, respectively.

A. Facure et al. / Applied Radiation and Isotopes 62 (2005) 69–72 71

In order to study the energy attenuation of neutrons

after crossing the treatment room walls, some simula-

tions were made with these particles going through

concrete barriers of 25, 50, 75 and 100 cm. The incident

energy spectrum in these simulations was the one

generated by 15MV accelerator (leakage spectrum

shown in Fig. 2). The concrete barrier, with density of

2.26 g/cm3, was constituted (in percentage by weight

composition) by 0.92% of hydrogen; 49.83% of oxygen;

1.71% of sodium; 4.56% of aluminum; 31.58% of

silicon; 1.92% of potassium; 8.26% of calcium and

1.22% of iron (National Council on Radiation Protec-

tion Measurements, 1976).

The resulting spectra are shown in Fig. 3. It can be

observed that fast neutrons (100 keVoEo10MeV) areattenuated in intensity and energy by scattering with the

concrete atoms constituents. The slowing down process

generates epithermal (1 eVoEo100 keV) and thermal(1 meVoEo1 eV) neutron fluxes. With the increase ofconcrete thickness, a considerable reduction of the

normalized flux of fast and epithermal neutrons is

observed. This hardening effect of the fast part of the

photoneutron spectra was previously observed inside

water phantoms (d’Errico et al., 1998). On the other

hand, the thermal neutron flux remains nearly constant;

which can be explained by the transformation of fast

and epithermal into thermal neutrons.

Neutrons, after crossing the accelerator head and the

concrete barriers, are moderated and thermalized and (n,

ARTICLE IN PRESSA. Facure et al. / Applied Radiation and Isotopes 62 (2005) 69–7272

photon) reactions start. Fig. 4 shows the rate of gamma

production as a function of their energies. When primary

neutron spectrum from a 15MV linear accelerator (see

Fig. 2) cross a concrete barrier, it is observed that gamma

rays with energies up to 10MeV, which are highly

penetrating, are produced. Nevertheless, this production

is drastically reduced when the barrier thickness is

increased, for example from 25 to 50 cm. It is important

to note that those capture gamma generation does not

occur at the same concrete deepness.

4. Conclusion

The undesired neutrons produced in medical linear

accelerators are high-energy neutrons, and the energy

distributions can be described by Eq. (2). The photoneu-

tron average energy values obtained in this work, before

crossing the accelerator head, range from 1.15 to

1.46MeV for 15 and 25MeV beam, respectively. On the

other hand, after crossing this shielding the average

energy is the same, i.e., 0.4MeV.

According to Fig. 3 it can be seen that the rapid

component of the spectrum is highly attenuated in a

concrete barrier, increasing the thermal and epithermal

parts of it. The capture gamma rays, which are produced

due to neutron interaction with concrete, have maximum

energy of 10MeV, with a peak in the 0.1 and 1MeV

region.

References

Agosteo, S., Foglio Para, A., Maggioni, B., 1993. Neutron

fluxes in radiotherapy rooms. Med. Phys. 20 (2),

407–414.

Briesmeister, J.F. (Ed.), 1997. MCNP—AGeneral Monte Carlo

N-particle Transport Code, Version 4B. Los Alamos

National Laboratory, Los Alamos, NM; LA-12625-M.

Carinou, E., Kamenopoulou, V., 1999. Evaluation of neutron

dose in the maze of medical accelerators. Med. Phys. 26

(12), 2520–2525.

d’Errico, F., Nath, R., Tana, L., Curzio, G., Alberts, W., 1998.

In-phantom dosimetry and spectrometry of photoneutrons

from an 18 MV linear accelerator. Med. Phys. 25 (9),

1717–1724.

d’Errico, F., Luszik-Bhadra, M., Nath, R., Siebert, B.R.L.,

Wolf, U., 2001. Depth dose-equivalent and effective

energies of photoneutrons generated by 6–18MV X-ray

beams for radiotherapy. Health Phys. 80 (1), 4–11.

National Council on Radiation Protection Measure-

ments, 1976. Structural shielding design and evaluation

for medical use of X-rays gamma rays of energies up

to10MeV. NCRP Report No. 49, Bethesda, MD,

USA.

Swanson, W.P., 1980. Estimate of the risk in radiation therapy

due to unwanted neutrons. Med. Phys. 7 (2), 141–144.

Tosi, G., Torresin, A., Agosteo, S., Folgio Para, A., Sangiust,

V., Zeni, L., Silari, M., 1991. Neutron measur-

ements around medical electron accelerators by active

and passive detection techniques. Med. Phys. 18 (1),

54–60.