See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/235698195

Prediction of (89)Zr production using the Monte Carlo code FLUKA

Data · February 2013

CITATIONS READS

11 365

8 authors, including:

Angelo Infantino Davide Pancaldi

CERN Policlinico S.Orsola-Malpighi
19 PUBLICATIONS   153 CITATIONS    35 PUBLICATIONS   241 CITATIONS   

SEE PROFILE SEE PROFILE

Ciarmatori A. Stefano Boschi

Azienda Ospedaliera Marche Nord Policlinico S.Orsola-Malpighi
62 PUBLICATIONS   204 CITATIONS    172 PUBLICATIONS   4,231 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

68Ga-PSMA PET/CT for restaging prostate cancer patients who experienced biochemical relapse after radical therapy View project

Tumor targeting diagnostic and therapy View project

All content following this page was uploaded by Davide Pancaldi on 01 June 2014.

The user has requested enhancement of the downloaded file.

 Applied Radiation and Isotopes ] (]]]]) ]]]–]]]

Contents lists available at ScienceDirect

Applied Radiation and Isotopes

journal homepage: www.elsevier.com/locate/apradiso

89
Prediction of Zr production using the Monte Carlo code FLUKA
A. Infantino a, G. Cicoria b, D. Pancaldi b, A. Ciarmatori b, S. Boschi c, S. Fanti b, M. Marengo b, D. Mostacci a,n
a
University of Bologna, Montecuccolino Laboratory, via dei Colli 16, I-40136 Bologna, Italy
b
Department of Medical Physics, University Hospital, S.Orsola—Malpighi, Bologna, Italy
c
PET Radiopharmacy Unit, Department of Nuclear Medicine, University Hospital, S.Orsola—Malpighi, Bologna, Italy

a r t i c l e in f o abstract

The widely used Monte Carlo simulation code FLUKA has been utilized to prototype a solid target for the
Keywords: production of 89Zr by irradiation of a metallic 89Y target foil in a 16.5 MeV proton biomedical cyclotron,
89
Zr through the reaction 89Y(p, n)89Zr.
FLUKA Simulations were performed with and without an Al energy degrader. In the setup of the geometry of
Monte Carlo the target, state of the art support tools, like SimpleGeo, were used for accurate, detailed modeling.
PET radionuclides The results permitted a quick assessment of all possible radionuclidic contaminants and confirmed
Targetry that the use of an energy degrader avoids production of the most important impurity, 88Zr. The estimated
value for the activity produced in one hour of irradiation at 20 mA is 384 7 42 MBq; this is encouraging,
indicating possible production of clinically significant amounts of activity with the relatively simple
target setup adopted. Initial experimental tests gave results in excellent agreement with simulations,
confirming the usefulness and accuracy of FLUKA as a tool for the design and optimization of targets for
the production of PET radionuclides.
& 2010 Elsevier Ltd. All rights reserved.

1. Introduction results were then compared with analytical calculations and

89
Zr has been proposed for use in nuclear medicine to label
monoclonal antibodies (Meijs et al., 1997; Verel et al., 2003), and in
association with 90Y to detect the bio-distribution of radiophar-
maceuticals labeled with this b  emitter (Perk et al., 2006). 89Zr has
a half life of adequate length (T1/2 ¼78.4 h); its b + decay branching
ratio (22.3%) and maximum energy (897 keV) make it an interest-
ing candidate for PET imaging, even though the presence of an
intense gamma emission (100% yield) with energy 908.9 keV
dictates careful calibration and setup of the imaging equipment
to preserve image quality.
89
Zr can be produced by proton irradiation of a target of Yttrium-
89, through the 89Y(p,n)89Zr reaction (Kandil et al., 2007; Hohn et al.,
2008; Holland et al., 2009). The fact that natural yttrium is
comprised of only Yttrium-89 affords an additional advantage to
the production of 89Zr with a proton biomedical cyclotron: the
target does not require a costly, enriched target material.
The aim of the work reported here was to obtain an a priori
assessment of the productivity of an irradiation line for the
production of 89Zr, and of the presence of possible radioactive
contaminants due to secondary nuclear reactions. The system was
modeled by means of a widely used Monte Carlo code: simulation
n
Corresponding author. Tel.: + 39 051 6441711; fax: + 39 051 6441747.
E-mail address: domiziano.mostacci@unibo.it (D. Mostacci).
experimental findings.
2. Materials and methods
FLUKA is a general purpose code for modeling particle transport
and interactions with matter; it covers an extended range of
applications spanning from proton and electron accelerator shielding
to calorimetry, dosimetry, detector design, radiotherapy and more
(Fasso et al., 2005; Battistoni et al., 2007). In particular, it can be used
to simulate irradiation of target materials with charged hadrons in
order to optimize target design and study activation of the materials.
The cyclotron used in irradiation tests and to which our
simulations setup makes reference is a PETtrace (GE Medical
Systems, Uppsala, Sweden), a compact cyclotron with vertical
acceleration plane, capable of accelerating H-ions in the first
harmonic mode to an energy of 16.5 MeV, and D-ions in the second
harmonic mode to 8.4 MeV (Gallerani et al., 2008; Marengo
et al., 2008). The unit installed at the University Hospital is
equipped with an experimental solid target station, developed
locally (Cicoria et al., 2006). This solid target station is fitted directly
on one of the beam ports of the cyclotron, thus not requiring a
beam line; the contraption accomodates round targets having a
diameter of 32 mm and thicknesses between 0.5 and 2.0 mm; it
includes an automated system for the unloading of the irradiated
material and its transportation to a shielded container outside the
cyclotron vault.

0969-8043/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.apradiso.2010.11.027

89
Please cite this article as: Infantino, A., et al., Prediction of Zr production using the Monte Carlo code FLUKA. Appl. Radiat. Isotopes
(2010), doi:10.1016/j.apradiso.2010.11.027
2 A. Infantino et al. / Applied Radiation and Isotopes ] (]]]]) ]]]–]]]

Reference cross section data can be calculated using the soft-

ware TALYS 1.0 (Koning et al., 2008); TALYS is a software for the
simulation of nuclear reactions that includes many state-of-the-art
nuclear models, making it possible to model all major reaction
mechanisms encountered in light particle-induced nuclear
reactions.
Besides the main reaction, 89Y(p,n)89Zr, also the 89Y(p,n)89m Zr
reaction has a significant cross section; 89mZr decays in 94% of the
cases to 89Zr ground state with a half life of 4.18 min, emitting a
588 keV photon, and in 6% of the cases directly to stable 89Y.
Planning to download the target after a waiting time of about 1 h, it
is then possible to combine the two cross sections for the ground
and the metastable levels into a single ‘‘effective’’ cross section
(Fig. 1). It is to be noted that in FLUKA, the tallying of the nuclei
produced does not distinguish between ground state and isomeric
Fig. 2. Rendering of the core of the target: the yttrium target foil is shown in its
state, and they are tallied as one and the same isotope. All residual
copper backing. In the simulations the aluminum degrader was or was not present.
nuclei are tallied when they have been fully de-excited down to This rendering was produced by SimpleGeo 4.1 (Theis, 2006).
their ground or isomeric state; in other words, FLUKA cannot
predict the correct branching for the production of different states
of the same radionuclide; instead it distributes the cross section
equally over the different states. Thus, we expect that simulations
will in any case correctly predict the production in the case of 89Zr
in the previously described situation, after an appropriate waiting
time following the end of the irradiation.
Fig. 2 shows the model of the target insert: it is composed
of a disc of metallic Yttrium-89 (+¼14 mm and thickness
0.15–0.30 mm), placed into a 0.3 mm deep cavity carved in a
copper holder (+¼ 32 mm and thickness¼1 mm). The choice of
the target foil thickness was made on the basis of a tradeoff
between ease of availability, expected productivity and likelihood
of thermal damage: a thicker target would entail greater produc-
tion, but also more energy absorption in the target foil and
increased demand on the cooling system. On the other hand,
foreseeable uses of 89Zr labeled radiopharmaceuticals in pre-
clinical research do not require the production of large activities,
such as would be needed in routine clinical use. In the present
preliminary evaluation it was decide to use a target thickness of Fig. 3. Rendering of the whole target assembly as designed with SimpleGeo 4.1
0.3 mm, for which ready to use metal foils are available at a limited (Theis, 2006).

cost and that is expected to yield sufficient productivity, as will be

discussed further in the following.
Fig. 1. Cross sections of the 89Y(p,n)89Zr and the 89Y(p,n)89mZr reactions and their
total combination. Data calculated using the software TALYS 1.0 (Koning, 2008).
Simulations were performed both with and without an alumi-
num degrader (+¼32 mm and thickness¼0.5 mm) covering this
assembly; without the degrader, the proton beam energy upon
entering the target is 16.1 MeV, while after exiting the yttrium foil
it is 13.0 MeV. Since cross section data for the 89Y(p, n)89Zr reaction
show that the effective cross section has a maximum slightly above
13 MeV, the presence of the degrader, that lowers the entering
energy to 12.6 MeV, with an energy at the exit from the target of
9.5 MeV, is expected to preserve the production of 89Zr, while
excluding the concurring reaction 89Y(p, 2n)88Zr, that has its
threshold at ca. 13 MeV. The whole target assembly model is
shown in Fig. 3, as designed with SimpleGeo 4.1 (Theis et al., 2006),
an interactive solid modeler which allows for flexible and easy
creation of the models for FLUKA. The front flange for fitting the
target to the cyclotron is made of aluminum; a HAVAR foil with a
thickness of 25 mm insulates the vacuum chamber of the cyclotron
from the frontal Helium cooling chamber, 16 mm deep.
The simulated proton beam had an elliptical shape with
a Gaussian profile with Full Width at Half Maximum (FWHM) of
8.5–6.3 mm in the x and y directions, respectively, and impinges on
the target in the z direction. In long term irradiation, however, the
beam oscillates an uncertainty of 10% in the beam area has thus
been adopted in the following calculations. Repeated simulations
with a statistics of 10,000,000 primary particles have been run.
Results of the simulations are reported by FLUKA in terms of
activation (nuclei/proton.cm3) in the yttrium target, the aluminum

89
Please cite this article as: Infantino, A., et al., Prediction of Zr production using the Monte Carlo code FLUKA. Appl. Radiat. Isotopes
(2010), doi:10.1016/j.apradiso.2010.11.027
 A. Infantino et al. / Applied Radiation and Isotopes ] (]]]]) ]]]–]]] 3

degrader and the copper backing. The uncertainty is given as

standard deviation of the results of 5 repeated runs.
Analytical calculation of the theoretical activity production
were made using the evaluated cross sections obtained from TALYS
and the stopping power calculated using SRIM 2008 (Ziegler et al.,
1985). Considering the uncertainties in the energy values entering
and exiting the target foil and the accuracy in the calibration of the
beam current, the overall accuracy of analytical calculation was
evaluated at a level of 10%.
The Havar foil insulating the cyclotron vacuum chamber from
the target material is clearly activated in the irradiation; however,
information on the Havar activation and its consequences on the
purity of radionuclide production is discussed at large in several
references (O’Donnell et al., 2004; Marengo et al., 2008; Manickam
et al., 2009), therefore studying this process was not considered
necessary for the scope of this work.
Experimental irradiation runs were conducted for 60 min at
20 mA; the irradiated material was downloaded automatically from Fig. 5. Activation products estimated by FLUKA in the irradiation of the target
the target assembly after a waiting time of 1 h, sufficient to allow adopting an Al degrader.

decay of 89mZr to the ground state; activity was then measured by

means of a radionuclide activity meter.

3. Results

Results of the simulated irradiations without the aluminum

energy degrader are shown in Fig. 4; the graphical representation
produced by FLUKA illustrates with an intuitive, semi-quantitative
scale the relative importance of the different activation products.
It includes all nuclear species, both stable and radioactive. Insofar
as the yttrium metallic target, contamination from 88Zr (T1/2 ¼83.4
days, parent of 88Y) is predicted to be significant, while limited, yet
not negligible production of 86Y and 90Y is expected. The production
of 89Zr, in the simulated condition of 10,000,000 proton histories, is
evaluated as (2.1670.02)  10  4 nuclei/proton.cm3. Taking into
account the half life of the nuclide and volume of irradiated
material, production is estimated at 189721 MBq in one hour of
irradiation at 20 mA.
According to the results of the simulations, the use of an
aluminum degrader to reduce the beam energy entering in the
target to below 13 MeV limits radionuclidic impurities signifi-
cantly, as shown in Fig. 5: only production of 89Zr, that decays back
to stable 89Y, is expected for an evaluated activation of
(4.3970.02)  10  4 nuclei/proton.cm3; this choice proves better
in that it limits impurities, at the same time improving activation
Fig. 4. Activation products estimated by FLUKA in the irradiation of the target
without Al degrader.
Fig. 6. Activation products estimated by FLUKA in the irradiation of the Al degrader.
by a factor of 2. In one hour at 20 mA the expected production is
384742 MBq.
Figs. 6 and 7 show the results as regards activation expected in
the aluminum degrader and in the copper backing: the most
significant radionuclide are reported in Table 1. As can be seen,
the higher activity levels are expected for very short half lived 27Si
(T1/2 ¼4.13 s); concerning longer lived radionuclides, again for an
irradiation of one hour at 20 mA, the production of 26Al in the
degrader can be estimated as o1 kBq , while the production of 65Zn
in the copper backing is estimated in 190740 MBq, with a
significant uncertainty due to the scattering of the beam in the
preceding layers.
In summary, FLUKA simulations with a target thickness of
300 mm of metallic 89Y, predict a satisfactory productivity; with a
saturation yield of 2200 7200 MBq/mA.
First experimental tests point to an average yield of
417736 MBq in a 60 min production at 20 mA of beam current
(Ciarmatori et al., accepted for publication), in excellent agreement
with the result of 384 742 MBq expected on the basis of the
FLUKA simulations. The analytical calculation yields a limit value of
412741 MBq for the activity produced in the same conditions; we
consider that our simulation model reproduces with sufficient
accuracy the experimental setup and that our target assembly
seems capable of approaching closely the theoretical limit of
production in the irradiation geometry adopted in this study.

89
Please cite this article as: Infantino, A., et al., Prediction of Zr production using the Monte Carlo code FLUKA. Appl. Radiat. Isotopes
(2010), doi:10.1016/j.apradiso.2010.11.027
4 A. Infantino et al. / Applied Radiation and Isotopes ] (]]]]) ]]]–]]]

FLUKA proved to be a useful and satisfactorily accurate tool for

the prediction of productivity of PET radionuclides and for the
optimization of the design of the related targetry, encouraging its
systematic use in the modeling of other irradiation lines that our
group is developing, such as 124Te(p,n)124I and natCu(p,x)65Zn.

References

 A., Ranft, J.,

Battistoni, G., Muraro, S., Sala, P.R., Cerutti, F., Ferrari, A., Roesler, S., Fasso,
2007. The FLUKA code: Description and benchmarking. In: Proceedings of the
Hadronic Shower Simulation Workshop, 2006, AIP Conference Proceeding 896,
pp. 31–49.
Cicoria, G., Pancaldi, D., Piancastelli, L., Giovaniello, G., Bianconi, D., Bollini, D.,
Menapace, E., Civollani, S., Pettinato, C., Spinelli, A., Boschi, S., Bergamini, C.,
Marengo, M., 2006. Development and operational tests of a solid target for the
PET-trace cyclotron. Q. J. Nucl. Med. 50 (suppl. 1), 3–4.
Ciarmatori, A., Cicoria, G. ,Pancaldi, D., Infantino, A., Boschi, S., Fanti, S., Marengo, M.,
accepted for publication. Experimental Studies on 89Zr production. Radio-
chimica Acta.
Fig. 7. Activation products estimated by FLUKA in the copper backing hosting the  A., Ferrari, A., Ranft, J., Sala, P.R., 2005. FLUKA: a multi-particle transport code.
Fasso,
aluminum target foil. CERN-2005-10, INFN/TC_05/11, SLAC-R-773.
Gallerani, R., Cicoria, G., Fantuzzi, E., Marengo, M., Mostacci, D., 2008. Neutron
production in the operation of a 16.5 MeV PET trace cyclotron. Prog. Nucl.
Energy 50, 939–943.
Table 1
Hohn, A., Zimmermann, K., Schaub, E., Hirzel, W., Schubiger, P.A., Schibli, R., 2008.
Summary of the most significant activation products in the Aluminum degrader and
Production and separation of ‘‘non-standard’’ PET nuclides at a large cyclotron
in the Copper backing.
facility: the experiences at the Paul Scherrer Institute in Switzerland. Q. J. Nucl.
Med. Mol. Imaging 52, 145–150.
Product T½ Decay mode Nuclei/proton.(cm3) Holland, J.P., Sheh, Y., Lewis, J.S., 2009. Standardized methods for the production of
high specific-activity zirconium-89. Nucl. Med. Biol. 36, 729–739.
27
Si 4.13 s b + (100%) (2.597 0.09)  10  4 Kandil, S., Scholten, B., Saleh, Z., Youssef, A., Qaim, S., Coenen, H., 2007. A comparative
26 study on the separation of radiozirconium via ion-exchange and solvent
Al 72,0000 year EC (18%)
(2.487 0.16)  10  5 extraction techniques, with particular reference to the production of 88Zr and
b + (82%) 89
Zr in proton induced reactions on ytrrium. J. Radioanal. Nucl. Chem. 274,
65 45–52.
Zn 244.1 day EC (98.5%)
(7.707 0.41)  10  5 Koning, A.J., Hilaire, S., Duijvestijn, M.C., 2008. TALYS-1.0. In: Proceedings of the
b + (1.5%)
International Conference on Nuclear Data for Science and Technology—
63
Zn 38.1 min EC (7%) ND2007, Nice, France. pp. 211–214.
(8.527 0.24)  10  5
b + (93%) Meijs, W.E., Haisma, H.J., Klok, R.P., Vangog, F.B., Kievit, E., Pinedo, H.M., Herscheid,
J.D., 1997. Zirconium-labeled monoclonal antibodies and their distribution in
tumor-bearing nude mice. J. Nucl. Med. 38, 112–118.
Marengo, M., Lodi, F., Magi, S., Cicoria, G., Pancaldi, D., Boschi, S., 2008. Assessment of
4. Conclusions radionuclidic impurities in 2-[18 F]fluoro-2-deoxy-Dglucose ([18 F]-FDG) rou-
tine production. Appl. Radiat. Isot. 66, 295–302.
The results of the simulations confirmed that an energy Manickam, V., Brey, R.R., Jenkins, P.A., Christian, P.E., 2009. Measurements of
activation products associated with Havar foils from a GE PETtrace medical
degrader, by shifting the beam energy upon entering the target,
cyclotron using high resolution gamma spectroscopy. Health Phys. 96, S37–S42.
is effective in controlling the production of undesired radionuclide O’Donnell, R.G., León Vintró, L., Duffy, G.J., Mitchell, P.I., 2004. Measurement of the
impurities, yet preserving simplicity in the assembly and produc- residual radioactivity induced in the front foil of a target assembly in a modern
tion yield of the desired radionuclide. medical cyclotron. Appl. Radiat. Isot. 60, 539–542.
Perk, L.R., Visser, O.J., Stigter-van Walsum, M., Vosjan, M.J., Visser, G.W., Zijlstra, J.M.,
Estimates of the saturation yield for the target assembly devel- Huijgens, P.C., van Dongen, G.A., 2006. Preparation and evaluation of (89)Zr-
oped are in excellent agreement with the results of analytical Zevalin for monitoring of (90)Y-Zevalin biodistribution with positron emission
calculations. On the basis of simulations, the production of 89Zr tomography. Eur. J. Nucl. Med. Mol. Imaging 33, 1337–1345.
Theis, C., Buchegger, K.H., Brugger, M., Forkel-Wirth, D., Roesler, S., Vincke, H.,
by proton irradiation seems feasible with this easy-to-assemble 2006Zr. Interactive three dimensional visualization and creation of geometries
metallic 89Y target, and appears satisfactory as far as yield and for Monte Carlo calculations. Nucl. Instrum. Methods .Phys. Res. A 562, 827–829.
radionuclidic purity. Analytical calculations and preliminary experi- Verel, I., Visser, G.W., Boellaard, R., Stigter-van Walsum, M., Snow, G.B., van Dongen,
G.A., 2003. 89Zr immuno-PET: comprehensive procedures for the production of
mental tests conducted on the basis of the encouraging results 89
Zr-labeled monoclonal antibodies. J. Nucl. Med. 44, 1271–1281.
described (Ciarmatori et al., 2010) confirm the overall validity of the Ziegler, J.F., Biersack, J.P., Littmark, U., 1985. The Stopping and Range of Ions in Solids.
simulations. Pergamon Press, New York.

89
Please cite this article as: Infantino, A., et al., Prediction of Zr production using the Monte Carlo code FLUKA. Appl. Radiat. Isotopes
(2010), doi:10.1016/j.apradiso.2010.11.027
View publication stats