Why do Werner states shift from pure to necessarily mixed entanglement above dimension two?

Thank you for this question. Unfortunately, the provided preprints don't directly address Werner states' transition from pure to mixed entanglement as dimension increases—they touch on Werner states tangentially (particularly [AG-2024.02-2136] on bound entanglement, and [AG-2024.02-1591] on Werner-Holevo channels), but none explains the core phenomenon you're asking about.

To properly answer this, I'd need papers specifically analyzing the structure of Werner states across dimensions and why their purity properties change—likely involving detailed analysis of their convex decompositions and separability boundaries as a function of Hilbert space dimension.

Could you rephrase the question, or would you like me to explore a related topic the archive does cover, such as entanglement classification in high dimensions or PPT entanglement properties?