The naturally occurring radioisotope 32 Si represents a potentially limiting background in future dark matter directdetection experiments. We investigate sources of 32 Si and the vectors by which it comes to reside in silicon crystals used for fabrication of radiation detectors. We infer that the 32 Si concentration in commercial single-crystal silicon is likely variable, dependent upon the specific geologic and hydrologic history of the source (or sources) of silicon "ore" and the details of the silicon-refinement process. The silicon production industry is large, highly segmented by refining step, and multifaceted in terms of final product type, from which we conclude that production of 32 Si-mitigated crystals requires both targeted silicon material selection and a dedicated refinement-through-crystal-production process. We review options for source material selection, including quartz from an underground source and silicon isotopically reduced in 32 Si. To quantitatively evaluate the 32 Si content in silicon metal and precursor materials, we propose analytic methods employing chemical processing and radiometric measurements. Ultimately, it appears feasible to produce silicon detectors with low levels of 32 Si, though significant assay method development is required to validate this claim and thereby enable a quality assurance program during an actual controlled silicon-detector production cycle.