Progress on light-element standards at Los Alamos using EDA

Full text: Over the past year, good progress was made on the R-matrix analyses of the A=2,4, and 7 systems in order to provide the standard cross sections for n+p scattering, and for the 3He(n,p) and 6Li(n,t) reactions. Lately, the N-N analysis, which also includes data for n+p capture and γ+d photodisintegration, has concentrated on including all the reasonable measurements of the n+p total cross section at energies up to 30 or 40 MeV. A final fit was obtained to more than 5000 data points having a chi-square per degree of freedom of 0.83. The new total cross sections deviate by no more than 0.5 % from the previous evaluation, but the differential cross section for small-angle proton scattering is about 2.5% lower than before at 10 MeV. These differences will have some impact on the actinide fission cross section standards measured relative to hydrogen. New data were added to the 4He analysis that affected the n+3He cross sections, including that for the 3He(n,p) standard reaction. These additions led to highly constrained predictions of the spin-dependent neutron scattering lengths that were consistent with the existing experimental values. Quite recently, however, high-precision measurements of the scattering lengths have been reported that are inconsistent (within uncertainties) with the R-matrix predictions (but in very good agreement, incidentally, with the microscopic RRGM value for the coherent scattering length). We are reviewing the situation with respect to adding these new data to the analysis, and the possible conflicts with the other measurements recently included. The analysis of reactions in the 7Li system continues to represent all the data included reasonably well, especially the high-precision measurements of cross sections and analyzing powers for t+α elastic scattering. Recently a solution was obtained with the original uncertainties of the t+α data doubled, as well as those of a shape measurement by Macklin of the 6Li(n,t) cross section relative to 235U(n,f) over a wide energy range. Outlying data points (χ2>10) were also eliminated in this analysis, giving an overall χ2 per degree of freedom of 1.16, comparable to the value obtained by RAC. Differences persist with the 6Li(n,t) cross section calculated by RAC, especially at energies above 1 MeV, reflecting differences in the 7Li level structure obtained by the two analyses at energies above the 5/2- resonance. In the standards, region, however, the RAC calculation was adjusted to agree closely enough with the EDA result that the CRP members felt the two could be averaged to obtain the standard cross section for this reaction. Another persistent difference between the two codes is in the output covariances (correlations). These are being explored further in a test of the performance of the three R-matrix codes EDA, RAC, and SAMMY, using a simplified (but realistic) data base from the 7Li system. The covariances from EDA were tested extensively against pseudo-data generated with known statistical properties, and they matched exactly with expectations. EDA's chi-square expression, when 'marginalized' over normalizations, was shown to be essentially the same near a solution as that used by RAC, SAMMY, and the generalized least-square codes. (author)