Alpha particles emitted from the decay of uranium in a UF<
sub>
6<
/sub>
matrix can interact with fluorine and generate neutrons via the <
sup>
19<
/sup>
F(?,n)<
sup>
22<
/sup>
Na reaction. These neutrons can be used to determine the uranium content in a UF<
sub>
6<
/sub>
storage cylinder. The accuracy of this self-interrogating, non-destructive assay (NDA) technique is, however, limited by the uncertainty of the <
sup>
19<
/sup>
F(?,n)<
sup>
22<
/sup>
Na cross section. We have performed complementary measurements of the <
sup>
19<
/sup>
F(?,n)<
sup>
22<
/sup>
Na reaction with both <
sup>
4<
/sup>
He and <
sup>
19<
/sup>
F beams to improve the precision of the <
sup>
19<
/sup>
F(?,n)<
sup>
22<
/sup>
Na cross section over the alpha energy range that encompasses common actinide alpha decay needed for NDA studies. We have determined an absolute cross section for the <
sup>
19<
/sup>
F(?,n)<
sup>
22<
/sup>
Na reaction to an average precision of 7.6% over the alpha energy range of 3.9 ? 6.7 MeV. We utilized this cross section in a simulation of a 100 g spherical UF<
sub>
6<
/sub>
assembly and obtained a change in neutron emission rate values of approximately 10-12%, and a significant (factor of 3.6) decrease in the neutron emission rate uncertainty (from 50-51% to 13-14%), compared to simulations using the old cross section. Our new absolute cross section enables improved interpretations of NDAs of containers of arbitrary size and configuration.