First devised over half a century ago, metal-on-polyethylene (MoP) hip prostheses have become the gold standard for total hip arthroplasty (THA), a surgical intervention for degenerative hip joint conditions. The accumulation of polyethylene wear debris after long-term, in vivo articulations, can induce adverse cellular reactions, osteolysis and aseptic loosening of the implant – ultimately resulting in the failure of the THA. Despite the distinct differences between the biotribology of MoP and MoM prostheses, there is a lack of congruent high resolution research investigating the biotribological interactions and surface structures of MoP hip prosthesis components. This study characterized the surface and subsurface microstructural changes in failed MoP hip prosthesis retrievals using advanced electron microscopy techniques. The samples were comprised of retrieved metallic cobalt-chromium-molybdenum (CoCrMo) alloy femoral head components, one ultra-high molecular weight polyethylene (UHMWPE) acetabular cup component, and unused CoCrMo reference samples. The surface of the reference samples contained linear, parallel, uniform scratches as a result of the manufacturing process;​ whereas the surface of the retrieval samples were covered in an abundance of scratches and a layer of residual deposits, attributable to in vivo articulation of the implant. Characteristic hard phases were observed and examined on the surface and from the cross-sectional preparation of the cast CoCrMo samples. The multiphasic hard phases on the cast samples can strengthen the material but also be sites of crack propagation and material detachment, contributing to the generation of wear particles. Lastly, a nanocrystalline layer, 20 to 400 nm in thickness was observed in the subsurface microstructure of all samples (including references). Previous MoM studies suggest that the nanocrystalline layer is a result of dynamic crystallization in response to multidirectional, chronic loading in vivo, however, the presence of the layer in the unimplanted references suggest that the nanocrystalline layer can be formed during the production of the prosthesis component and therefore, pre-exists implantation. The imperfections on new, unused implants can have protective effects (e.g. troughs from scratches can be a reservoir for wear debris) but may influence in vivo wear processes after implantation (e.g. scratches may be a source of wear debris). Higher resolution analyses on more retrieval and reference samples are required to pinpoint the exact mechanism of failure in MoP hip prostheses and extend the longevity and efficacy of THA.