Neutron Producing Target for Accelerator Based Neutron Capture Therapy*

Pilot innovative accelerator based neutron source for neutron capture therapy of cancer [1] is now on the threshold of its operation at the BINP, Russia. One of the main elements of the facility is lithium target producing neutrons via threshold Li(p,n)Be reaction at 25 kW proton beam with energies 1.915 MeV or 2.5 MeV. In the present report, choice of target is substantiated, the main problems of lithium target are determined, the conception of optimal target is proposed. The results of investigation of radiation blistering and lithium layer are presented. Design of target for the neutron source constructed at BINP is shown. Reactions. Neutrons with energies of 0.5 eV – 10 keV are necessary for neutron capture therapy. Neutrons with higher or lower energies and γ-radiation are extremely undesirable. The following neutron-producing charged particles reactions are considered mainly for use in accelerator based neutron capture therapy: Li(p,n), Be(p,n), Be(d,n) and C(d,n) [3]. Table 1 shows the properties of these reactions. The Li(p,n) reaction is excellent neutronically: neutron producing is high and relatively soft spectrum requires less moderation than those generated in other reactions. Unfortunately, lithium melting point is low, its thermal conductivity is poor, and finally, lithium is a very reactive metal, forming compounds immediately upon exposure to air. Alternate targets from beryllium and carbon overcome these difficulties in manufacture and cooling. However, more extensive moderators are required, and higher accelerator current is needed. Li(p,n)Be reaction is a threshold one and it is characterized by unusually high increase in reaction cross section near threshold. This allows considering an addition opportunity for near threshold operation when proton energy exceeds 30 – 40 keV the reaction threshold (1.882 MeV). Neutron beam with mean energy of 40 keV kinematically collimated is generated in this case. To decrease the neutron energy to epithermal value, thin water moderator (about 2 cm thick) is sufficient, therefore patient may be placed near the target. This provides the same treatment time that the standard mode at proton energy of 2.5 MeV. Thus, for neutron generation, just Li(p,n)Be reaction is proposed to be used with proton beam energy of 1.915 or 2.5 MeV, in spite of poor lithium properties. This choice intends developing much more complicated lithium target than the Be or C ones. Purity. Pure lithium is more efficient in neutron generating than lithium hydride, oxide, nitride, or fluoride [4] (Table 2) and possesses higher thermal-conductivity, but incomparably lower melting temperature, therefore it requires efficient heat removal at as low lithium layer temperature as possible. Use of target with liquid lithium layer is also possible, but considerable lithium evaporation results in decrease in high voltage electric durability due to lithium vapour inflow and expansion of nascent radioactive beryllium over the whole facility. Thickness. Inelastic proton scattering on lithium nuclei leads to considerable γ-rays flux with energy of 478 keV that sometimes exceeds neutron flux [5]. In Table 3 the gamma yield is shown depending on proton energy for thick lithium target which stops a proton, and for thin one, braking a proton only to 1.882 MeV (energy of neutron generation reaction threshold). It is seen that thin target decreases considerably the gamma flow. In case of thin target protons should further brake in tungsten, molybdenum or any other substance whose inelastic scattering does not result in gamma radiation. This condition is met for almost all nuclei harder than aluminium one [6]. Target lifetime. Radiation blistering is main processes determining lifetime of a target. Appearance of blisters increases the lithium layer evaporation due to rise of temperature because of swelling and flaking, and generally spoils the target. Fluence which certainly causes blistering is about 2 10 cm for copper and a hundred times as much for metals that solve hydrogen well. At current of 10 mA and target diameter of 10 cm, blistering appears on copper substrate after several hours, and such a target will require daily replacement. But, copper substrate is not expensive and quite easy to manufacture, its heat conductivity is high, and all these features make this substrate the main solution. Substrate made of metals that solve hydrogen well may be replaced once a week. But heat conductivity of these metals is lower than copper, their manufacturing is harder; and what is the most important, effect of interaction for hydrogen diffusing to substrate surface with lithium is not yet clear. Induced activity is another problem preventing the target from long-term operation. Every act of neutron production in the Li(p,n)Be reaction is accompanied by radioactive nucleus of beryllium isotope. Beryllium isotope Be becomes stable lithium isotope Li due to capture of orbital electron with half-life of 53.6 days. The capture causes no radiation in 89,7 % cases, and in 10.3 % it radiates gamma-quantum with energy of 478 keV. Operation with open source of activity higher than 10 Bq is allowed in an isolated room only. As 10 Bq activity may be achieved quite quickly (12 min at ___________________________________________ * Work supported by ISTC (www.istc.ru) # [email protected] CHOICE OF TARGET 360 Proceedings of RuPAC 2006, Novosibirsk, Russia