Standard model factorization can account for the large rate for $B^\pm \to \eta' K^\pm$. The large rate for $B \to \eta' X_s$ is more problematic and requires a ``cocktail'' solution. Individual contributions from factorization, $b \to \eta' s g$, and ``intrinsic charm'' fall significantly short of the observed rate. In contrast, the observed $\eta'$ yields are easily explained in models with $B(b \to sg) \sim 15%$. Factorization can account for the entire semi-inclusive $\eta'$ yield without violating rare decay constraints. Implications for CP violation are discussed.
Categories: hep-ph