Pushing the Limits of an O-18 Water Target
J.A. Nye, D. W. Dick, and R.J. Nickles
Department of Medical Physics, University of Wisconsin, Madison, Wisconsin.
Abstract. A gridded-niobium target was constructed for the improvement of routine [18F]-fluorine production from 18Oenriched water on a CTI RDS 112 cyclotron. Niobium was chosen for its inertness and excellent thermal properties. The
target volume consists of a 400µL (active volume) niobium chamber mounted with a single entrance foil supported
against an array of 3mm hexagonal holes with 0.25mm aluminum septa, machined by EDM. The target operates at high
beam currents and elevated pressures and temperatures with significant reductions in maintenance intervals. Several
diagnostic tools such as autoradiography, activation, and neutron logging optimize the performance and yield of the
target. Entrance foils including Havar and Nb are used to assess the [18F] chemical compatibility, with FDG synthesis as
the test reaction. The gridded, single-foiled niobium target chamber appears to be an improvement compared to a
standard double-foiled helium cooled water target used with RDS cyclotrons.
Construction of a target using niobium relies on a
sufficient heat dissipation rate on the foil side to
overcome the inadequate thermal conductivity of the
water filled target chamber. To improve the burst
pressure of a niobium foil, the entrance foil can be
mounted on a grid made of a high thermally
conductive material such as aluminum or copper (3).
Subsequently, the grid will provide mechanical
strength allowing the pressure in the target volume to
be increased. An increase in pressure will raise the
boiling temperature and minimize the loss of useful
beam from the formation of voids in the target
chamber.
INTRODUCTION
Small volume water targets for the production of
aqueous [18F]-fluoride have yet to be firmly
established in a small clinical and research setting.
Several performance limitations prevent small volume
targets from operating reliably at beam currents of
greater than 30µA. Insufficient heat dissipation from
the target foil and target body limits beam currents,
thereby increasing the required irradiation period. In
addition, boiling and water voids formed within the
target volume reduce yields and place unwanted stress
on the inner surfaces of the target chamber. The most
significant rate-limiting step in a small volume design
is preventing the loss in reactivity of the [18F]-ion
caused by contaminates from both the target body and
knock-on ions created by surfaces in the beam strike.
Therefore, the challenge presented is how to increase
beam currents while minimizing losses in target
performance due to inadequate heat dissipation and
formation of reactive ions.
METHODS AND DISCUSSION
A small volume target is constructed by machining
a 400µL active chamber in a niobium cavity mounted
on a copper plate. Access to the target volume is
accomplished with two silver-induction brazed 0.5mm
I.D. stainless steel tubes. A circulating water cavity
machined in the copper plate actively cools the area
directly behind the niobium chamber. The copper plate
used to mount the niobium chamber is 20mm deep
with 3mm bored channels cut in the radial direction
allowing access to the stainless tubing leads from the
target’s exterior. A single entrance foil is mounted on
Niobium is the material of choice in reactive
environments because of its high melting point and
chemical resistiveness (1). Other favorable materials
are titanium and silver (2), however, silver targets are
known to require frequent cleaning due to ion
contamination. Titanium is a suitable counterpart to
niobium but it is more difficult machine and has lower
yield strength and heat conductivity.
CP680, Application of Accelerators in Research and Industry: 17th Int'l. Conference, edited by J. L. Duggan and I. L. Morgan
© 2003 American Institute of Physics 0-7354-0149-7/03/$20.00
1098
an aluminum grid consisting of a hex-hole array (3mm
across the flats, 0.25mm septa, 12mm in axial depth).
The grid is actively cooled by parallel-opposed water
lines.
FIGURE 2. Audioradiograph images of 65Zn
produced from the reaction 65Cu(p,n)65Zn by
bombarding natural 1mm thick copper foil in place of
the Havar window. The top images show activation of
the copper foil without the grid (top left) and with the
grid (top right). The graphs below each image show
the beam profile along a bisecting line without the grid
(bottom left) and with the grid (bottom right). The
beam profiles are normalized to their peak values for
presentation purposes.
FIGURE 1. Sectional view of the target assembly
showing the placement of the grid and niobium target
cell. The niobium chamber (bottom left) has a small
head spaced machined on the outside perimeter of the
beam strike area to act as a recombination space. The
grid is shown in grayscale and a frontal view is shown
under the sectional assembly (bottom right).
It is evident from the above figure that a significant
portion of the beam current enters through the center
holes. This would suggest that further optimization of
the grid would require an asymmetric hex-hole pattern
to improve beam transmittance reducing the amount of
beam lost on the septa. Furthermore, a diverging hexhole pattern, characteristic of the beam emittance,
would also increase the transmittance. However, these
modifications would only provide a small increase in
overall performance.
The aluminum grid is designed (4) for a bursting
pressure of 2000psi for a 0.025mm havar foil. The
open projected area is 85%. The incident beam profile
of the CTI RDS 112 is defined by 1cm diameter
entrance slits, with 6-8mm FWHM, resulting in a
beam transmittance of 80% with the grid in place.
Figure 2 shows a section profile of the incident beam
on the target produced by activation of copper foil.
Neutron logging during irradiation at pressure
intervals from 400psi to 1500psi using H2natO and 10%
enriched H218O water remain constant at beam currents
ranging from 10-45µA. A constant neutron production
rate suggests, even in the presence of boiling, the
target water is still ‘thick’. Gamma spectroscopy using
a heavily collimated BGO crystal post irradiation
showed no evidence of 93mMo (γ=685keV, t1/2=6.95hr),
from the reaction 93Nb(p,n)93mMo, produced in the
target cell.
Since the target’s inauguration in June 2002, the
grid and Havar combination has operated without
failure at currents of 45µA and pressure ranges from
400-1400psi. In order to increase the reactivity of
eluted [18F]-fluoride an ion exchange column is used to
decouple contaminant ions deposited in the target
1099
TABLE 1. Yield measurements for 94.6% enriched
H218O water using a 0.025mm Havar foil.
Interchanging the Havar with Nb foil results in similar
yields. The target is operated with a total volume of
800µL, which includes the beam strike volume and
associated fittings. The target over pressure in column
three is helium gas.
water during irradiation. FDG labeling yields are
consistent at 60-65% decay-corrected using the
Hamacher procedure (5).
To improve the FDG labeling phase, a 0.025mm
Nb-foil has recently replaced the Havar foil creating a
complete Nb-target. The target is limited to pressures
less than 450 psi as a result of niobium’s mechanical
properties. The target has operated with the niobium
foil at beam currents >30µA without foil failure and
FDG labeling yields have increased to 70-75% decay
corrected. Subsequent gamma spectroscopy on a
natural water sample irradiated for >2 hours shows
little to no measurable evidence of the primary
contaminant, 93mMo. Figure 3 is data taken on a
heavily shielded 10atm Xe well-detector capable of
measuring activity at levels of >10nCi.
Beam
Current
[µA]
Bmb.
Duration
[min]
Pressure
[psi]
Activity
at EOB
[mCi]
Saturation
Yield
[mCi/µA]
13
117
400
715
101.3
13
125
400
730
102.9
20
107
500
1003
102.0
20
74
1100
725
100.2
CONCLUSION
Development of a reliable low-maintenance water
target for routine production of aqueous [18F]-fluoride
is presented. The use of a niobium target insert and
entrance foil has increased FDG labeling yields
compared to a double-foiled (He cooled) silver target.
The current limit for the Havar foil combination has
yet to be reached. The target design allows the user to
exchange target chambers (niobium, silver, titanium)
and foil arrangements. Adaptation to flow through
targetry for the production of [13N]-ammonia further
demonstrates the versatility of the target. Total
production expenses for the gridded target are
approximately $2500.
FIGURE 3. A six decade decay of H2natO after
approximately two hours of irradiation with the Nbfoil.
Saturation yields for the niobium target are limited
by the grid transmittance. The thick target yield is
approximately 100mCi/µA, table 1, and agrees with an
80% transmittance for thick H218O yields (6).
Presently, the target has operated without a foil change
or cleaning for 93hrs at an average beam current of
20µA. Normal operating conditions are 20µA for
durations of approximately 120min with the Nb foil.
One accidental irradiation occurred with no water
cooling on the aluminum grid at a current of 20µA for
150min. Some heat damage occurred on the beam side
of the grid but no serious harm is evident.
ACKNOWLEDGEMENTS
The authors of this paper would like to
acknowledge Andrew Roberts and Todd Barnhart for
their assistance in FDG synthesis.
REFERENCES
1. Zeisler S. K., Becker R. A., Pavan R. A., Moschel
R., and Rühle H., Appl. Radiat. Isot. 53, 449-453
(2000).
1100
2. Mulholland G. K., Hichwa R. D., Kilbourn M.R., and
Moskwa J., J.Labelled Compd. Radiopharm. 26, 192-193
(1989).
3. Nickles R.J., Nucl. Instr. Meth. 177, 593-594 (1980).
4. Young W. G., Roark’s Formulas for Stress and Strain
6th ed., New York: McGraw-Hill, 1989, pp. 477-478.
5. Hamacher K., Coenen H. H., and Stocklin G., J. Nucl.
Med. 27, 235-238 (1986).
6. Ruth, T.J. and Wolf A.F., Radiochem. Acta 26, 21-24
(1979).
1101