Intracell activity distributions were measured in three natural uranium, heavy water lattices of 1. 010 inch diameter, aluminum clad rods on triangular spacings of 4. 5 inches, 5. 0 inches, and 5. 75 inches, respectively, and in a uranium, heavy water lattice of 0. 25 inch diameter, 1. 03% U 2235, aluminum-clad rods on a triangular spacing of 1. 25 inches. The distributions were measured with bare and cadmium-covered foils of gold, lutetium, and europium. The gold was used as a 1/v absorber to measure the thermal neutron density distribution. Because the activation cross sections of lutetium and europium depart considerably from 1/v behavior, their activation depends strongly on the thermal neutron energy spectrum. Hence, they were used to make integral measurements of the change in the neutron energy spectrum with position in the lattice cell. A method was developed for treating the partial absorption, by cadmium covers, of neutrons at the 0. 46 ev europium resonance, and it was found possible to correct the europium activations to energy cutoffs just above and just below the resonance. The measured activity distributions were compared with those computed with the THERMOS code. In the natural uranium lattices, THERMOS gave excellent agreement with the measured gold activity distributions and very good agreement with the lutetium and europium distributions, indicating that THERMOS gives a very good estimate of the spatial and energy distribution of thermal neutrons in these lattices. In the enriched lattice, THERMOS gave a large overestimate of the activity dip in the fuel for all three detectors. The discrepancy was attributed to a breakdown in the Wigner-Seitz cylindrical cell approximation at small cell radii. However, the measured ratios of lutetium and europium activity to gold activity were in good agreement with the THERMOS values, indicating that THERMOS still gave a good estimate of the degree of spectral hardening. Neutron temperature calculations were made from the data by using Westcott effective cross sections. The temperature changes so calculated agreed well with those predicted by THERMOS. Disadvantage factors calculated by the Amouyal-Benoist-Horowitz (ABH) method were in excellent agreement with the measured values in the natural uranium lattices. The agreement was not as good in the enriched lattice because of an expected breakdown in the ABH method at small cell radii. Values of the thermal utilization obtained from experiment, from THERMOS, and with the ABH method were in excellent agreement for all the lattices studied. Radial and axial buckling measurements made with lutetium were in excellent agreement with similar measurements made with gold, indicating that the thermal neutron spectrum was uniform throughout the lattice tank. Measurements of intracell gold activity distributions made in off-center cells differed only slightly from those made in the central cell of the lattice, indicating that the radial flux distribution was almost completely separable into a macroscopic Jo and a microscopic cell distribution.

Measurements of the intracellular distribution of the activation of foils by neutrons were made in lattices of 1/4-inch diameter, 1.03% U-235, uranium rods moderated by heavy water, with bare and cadmium-covered foils of gold, depleted uranium, lutetium, europium and copper. The measurements were made in the M.I.T. Heavy Water Lattice Facility with source neutrons from the M.I.T. Reactor. Two lattices were studied in detail in this work. The more closely packed lattice had a triangular spacing of 1.25 inches, and the less closely packed lattice had a triangular spacing of 2.5 inches. The results of the experiments were compared to one-dimensional, 30-energy group, THERMOS calculations based on the available energy exchange kernels. The comparison indicated that the approximation that the hexagonal cell may be replaced by an equivalent circular cell (the Wigner-Seitz approximation) can lead to serious discrepancies in closely packed lattices moderated by heavy water. A modified one-dimensional, and a two-dimensional, calculation were shown to predict the intracellular activation distribution in the closely packed lattice. An analytical treatment of the problem of the flux perturbation in a foil was developed and compared to the experimental results obtained by using gold foils of four different thicknesses in the lattice cell; the method was shown to be adequate. An analytical method to treat the effect of leakage from an exponential assembly was formulated; the results indicated that only in small exponential assemblies would leakage be a significant problem in intracellular flux measurements. A method was developed to predict the cadmium ratio of the foils used in the lattice cell; comparison with available measurements with gold foils indicated good agreement between theory and experiment, except for a lattice having very large ratios of moderator volume, to fuel volume, e.g., 100:1. Calculations of the fuel disadvantage factor by the method of successive generations for gold, lutetium and europium detector foils were compared to the results of THERMOS calculations, because THERMOS was shown to predict the experimental distributions. The comparison indicated that the method of successive generations is a good alternative to the THERMOS calculation, if all that is required is 17 and the thermal utilization.