The following was presented at the 34th Annual Meeting of the Society of Nuclear Medicine. The abstract is printed in the Journal of Nuclear Medicine, Vol. 28, No. 4, April, 1987. A full report is printed in Health Physics of Radiation-Generating Machines Proceedings of the 20th Midyear Topical meeting of the Health Physics Society, Reno NV, Feb. 1987.

return to list of publications

return to CRA home page.

Radiation Measurements Related to the Design of a Self-Shielded Accelerator System for Routine Use in PET

L.R. Carroll, E. Pekrul, G.O. Hendry, R.J. Nickles, J. Votaw, CTI Inc., Berkeley, CA, and the University of Wisconsin, Madison, WI


An 11 MeV Negative Hydrogen Ion Cyclotron, including gas and liquid targets for automated production of positron-emitting isotopes, has been developed by CTI, Inc., Knoxville, TN. Tests were conducted to: 1) Characterize the target reactions with respect to neutron and gamma dose as measured through various thicknesses of shield material. 2) Characterize the accelerator itself regarding its contribution to the overall dose; evaluate materials used internally such as copper, carbon, aluminum, tantalum, etc. with particular attention to components in the path of direct or stripped beam. 3) Evaluate a prototype neutron and gamma shield.

A liquid-scintillator detector (NE213) with pulse-shape discrimination gave separate and distinct spectra for gamma (Compton) interactions and proton recoils from fast neutron collisions. First-moment integral computations plus a simple unfolding algorithm were applied to the data to yield dose rate and incident neutron spectra, which were then used for fluence and transport calculations.

The shield measurements were in accord with prediction: The target zone utilizes 20 cm lead (gamma shielding), plus 80 cm hydrogenous material (neutron shielding) to attenuate radiation to safe levels. The cyclotron magnet yoke provides substantial shielding for the remainder of the accelerator, with at least an additional 30 cm hydrogenous material required overall to absorb neutrons which penetrate the steel magnet.

Target reactions and Neutron Spectra

The spectra shown below were obtained using an NE213 liquid scintillator spectrometer. Pulse shape discrimination permitted the relatively slow-decay pulses due to neutron interactions (proton recoils) to be distinguished from faster pulses associated with gamma ray (Compton scatter) interactions.(1)

The intensity of scintillation light output from ionization by a proton recoil is a non-linear, though well-known and predictable function of energy. (2, 3) After calibrating against standard gamma ray spectral components and correcting for the intrinsic nonlinearity, the raw NE213 spectra were "unfolded" by differentiating with respect to energy and weighting the result by an energy-dependent inverse efficiency factor derived from the macroscopic scattering cross-section for hydrogen, which is a major constituent of the liquid-scintillator detector.

The spectral unfolding algorithm was validated (to the degree appropriate for the present work) by comparing a measured spectrum of neutrons from an Americium-Berylium (AmBe) source against an ensemble of measured and calculated AmBe spectra found in the literature.(4-5)

Unfolded NE213 spectra are shown for some PET isotope target reactions of major interest: 18O(p,n)18F, and 15N(p,n)15O. The spectra are all normalized to the same charge on target. Each graph has two traces; a bare-target spectrum and a spectrum measured through 20 cm lead.

The benchmark, or reference bare-target neutron dose-equivalent for 10.4 MeV protons incident on an H218O target, measured 1.8 meters from the target (0 deg. forward angle), is 1 rem / µampere-hour.

20 cm lead is required to attenuate the intense, high-energy gamma radiation which accompanies the nuclear reactions occuring in the target. The lead also serves to degrade the energy of neutrons through inelastic and non-elastic scatter processes but, except for some resonance capture in antimony (which is usually alloyed with lead to give it better strength and hardness) few neutrons are actually absorbed by lead.

Low-energy neutrons are not displayed in these plots because the detector is essentially "blind" to neutrons below approximately 1 MeV. However, the large number of energy-degraded neutrons which penetrate the lead are easily absorbed by the outer shield-layer, which is made of boron-loaded hydrogenous material.

Prompt gamma radiation from target reactions

Proton bombardment of the target causes target and product nuclei to be elevated to high-energy nuclear-excited states. Subsequent transitions to the ground state are accompanied by intense Gamma radiation. Gamma dose rates were measured using the NE213 Compton spectrometer (set to reject neutrons) and using a G-M chamber specially constructed to minimize response to neutrons (E.G.G. Santa Barbara model GM-1).

Detailed spectral unfolding was not attempted. Principle components could be inferred from prominent edges in the NE213 Compton spectra and from nuclear data on excited states. Examples are tabulated here for several isotopes of interest.(6)

The benchmark, or reference bare-target gamma doserate for 10.4 MeV protons incident on an H218O target is 0.1 rad / µampere-hour measured 1.8 meters from the target. The first 5 cm lead shielding provides a nominal tenth-value gamma-attenuation; as more lead is added the apparent tenth-value thickness increases to 10 cm due to neutron-induced secondary gamma radiation produced in the lead itself.


  1. Knoll, Glenn F., chapter 15 in Radiation Detection and Measurement John Wiley and Sons, New York, 1979
  2. Cecil, R.A. Anderson, B.D., Madey, R, Improved Predictions of Neutron Detection Efficiency for Hydrocarbon Scintillators from 1 MeV to About 300 MeV. Nuclear Instruments and Methods 161 (1979) 439-447.
  3. Maier, K.H., Nutschke, J., Die Lichtausbeute Eines NE213-Szintillators fur Protonen, Nuclear Instruments and Methods 59 (1968) 227-228.
  4. Geiger, K.W., Van der Zwan, L., Radioactive Neutron Source Spectra from 9Be(a,n) Cross Section Data. Nuclear Instruments and Methods 131 (1975) 315-321).
  5. Notarrigo, Parto, et. al., Experimental and Calculated Energy Spectra of AmBe and PuBe Neutron Sources. Nuclear Physics A125 (1969) 28-32.
  6. Table of Isotopes, Seventh Edition Lederer, M., Shirley, V., Ed's. John Wiley and Sons, Inc. New York, 1978.

return to list of publications

return to CRA home page.