Spin measurements for 147Sm+n resonances:: Further evidence for nonstatistical effects

We have determined the spins J of resonances in the Sm-147(n,gamma) reaction by measuring multiplicities of gamma-ray cascades following neutron capture. Using this technique, we were able to determine J values for all but 14 of the 141 known resonances below E-n=1 keV, including 41 firm J assignments for resonances whose spins previously were either unknown or tentative. These new spin assignments, together with previously determined resonance parameters, allowed us to extract level spacings (D-0,D-3=11.76 +/- 0.93 and D-0,D-4=11.21 +/- 0.85 eV) and neutron strength functions (10(4)S(0,3)=4.70 +/- 0.91 and 10(4)S(0,4)=4.93 +/- 0.92) for J=3 and 4 resonances, respectively. Furthermore, cumulative numbers of resonances and cumulative reduced neutron widths as functions of resonance energy indicate that very few resonances of either spin have been missed below E-n=700 eV. This conclusion is strengthened by the facts that, over this energy range, Wigner distributions calculated using these D-0 values agree with the measured nearest-neighbor level spacings to within the experimental uncertainties, and that the Delta(3) values calculated from the data also agree with the expected values. Because a nonstatistical effect recently was reported near E-n=350 eV from an analysis of Sm-147(n,alpha) data, we divided the data into two regions; 0 < E-n < 350 eV and 350 < E-n < 700 eV. Using neutron widths from a previous measurement (corrected for new unresolved doublets identified in this work) and published techniques for correcting for missed resonances and for testing whether data are consistent with a Porter-Thomas distribution, we found that the Gamma(0)(n) distribution for resonances below 350 eV is consistent with the expected Porter-Thomas distribution. However, we found that Gamma(0)(n) data in the 350 < E-n < 700 eV region are inconsistent with a Porter-Thomas distribution, but in good agreement with a chi(2) distribution having nu >= 2 We discuss possible explanations for these observed nonstatistical effects and their possible relation to similar effects previously observed in other nuclides.