Each year, roughly 750,000 patients in the US receive a total joint replacement (TJR), or a synthetic medical device that serves to replace the natural joint to restore function and relieve pain. TJRs have had a long history of use in the hip, knee and shoulder, yet still retain the same standard design of a hard-on-soft bearing coupling. Today, those bearings are primarily composed of hard cobalt chrome (CoCr) surface articulating against ultrahigh molecular weight polyethylene (UHMWPE), a polymer with notable mechanical toughness and biocompatibility that have driven its 50-plus years of use in vivo. TJRs have seen tremendous success for older patient cohorts for whom these devices were designed. However, increasing demand from younger patients has motivated the need for more durable materials that can sustain higher, more variable loading for a longer time. UHMWPE has become central to this mission, with limitations in its long-term wear resistance, oxidative stability and fracture/fatigue properties coming under the limelight in recent years.

Explanted devices that have been retrieved from patients following failure have provided significant insight into the vulnerabilities of different UHMWPE formulations currently in the market, especially with respect to component fracture. Three cases examined in this work describe fracture failures of UHMWPE components seen in the knee, hip and shoulder to demonstrate existing tradeoffs in material sterilization and processing. Irradiation crosslinked blends, initially introduced to the orthopedic market to improve wear resistance, exhibits reduced resistance to both oxidation and fatigue crack propagation. Such tradeoffs are shown to contribute to the failure of two fractured knee tibial inserts, a single fractured hip acetabular cup, and a series of severely fractured glenoid components used in the shoulder. The former two cases exhibit an additional precursor to fracture that has been recognized but largely ignored in the fatigue characterization of crosslinked UHMWPE: the existence of stress concentrations (notches).

Incidences of component fracture like the three reviewed in this work have motivated significant study of UHMWPE’s mechanical deformation as a function of microstructural changes. In addition, computational studies have sought to establish how changes in design may reduce local stresses in UHMWPE components to mitigate failure. Material and design influences have thus predominantly been studied in isolation. The latter half of this work demonstrates how previous methodologies using linear elastic fracture mechanics (LEFM) and theoretical approaches to notch fatigue can been merged to elucidate the influence that notch geometry (notch-root radius) has on crack behavior in UHMWPE formulations. A robust computational analysis of existing theories regarding LEFM notch stress intensity and notch plasticity is presented, lending credibility to the use of the Dowling (1979) approach in characterizing the crack driving force in the vicinity of the notch. Standard compact tension (CT) specimens were modified to include one of five notch geometries: a sharp (0.13 mm) radius (control specimen), two crack-like radii (0.75 and 1 mm) and two blunt keyhole-type radii (2 and 3 mm). All radii were chosen to reflect geometries seen in modern TJR features, such as a tibial post in the knee or a liner locking mechanism in a hip. Cracks from 0.1 to 1 mm in length were created at each notch root using a sharp razor blade. Cyclic tensile loading (stress ratio, R = 0.1) was applied to impose an increasing cyclic stress intensity (ΔK), and crack growth was monitored optically.

The fatigue crack propagation (FCP) behavior was mapped using the Paris law to compare the relative crack speeds at a given applied ΔK for each notch geometry and material formulation. Crack growth ahead of each notch was found to overlap with sharp crack data, further supporting the use of the Dowling approach in characterizing near-notch crack growth. This overlap in data implied that mechanisms of crack growth near the notch were similar to those further away from the notch (outside the “notch-affected zone”, as calculated for each notch radius using the Dowling approach). The congruency of all notch-emanating crack data also revealed microstructure-driven trends between each material cohort that have been noted in previous sharp crack studies. Highly crosslinked and remelted UHMWPE (RXLPE) was found to display the least resistance to FCP, while highly crosslinked and Vitamin E blended (VXLPE) formulations demonstrated a notable improvement. Virgin UHMWPE consistently demonstrated the best resistance to FCP. The reduction in FCP resistance seen in highly crosslinked materials was associated with reduced local plasticity in amorphous regions that otherwise serves to mitigate crack advance. Above-melt annealing (“remelting”) in RXLPE has been shown to minimize oxidation and subsequent embrittlement in vivo through free radical elimination, but can result in reduced percent crystallinity and lamellae quality that can further diminish FCP resistance. Blending Vitamin E similarly serves to reduce oxidation for VXLPE, but retains the crystalline quality and quantity of virgin UHMWPE. However, blended Vitamin E also diminishes crosslinking efficiency, thereby improving VXLPE’s resistance to FCP relative to RXLPE but with a tradeoff in optimal wear resistance.

This study demonstrated that fatigue crack growth in UHMWPE primarily defers to microstructural influences, even when considering varying notch geometries within the vicinity of a crack. This work demonstrates that this methodology of investigating notch effects on crack behavior can be leveraged for polymeric materials, despite its primary origin from crystalline metals. Furthermore, by mimicking previous specimen types, sample dimensions and loading conditions, the methodology used in this work is easily translatable to orthopedic manufactures or research groups seeking to evaluate notch effects in novel UHMWPE formulations. Results shown here reveal that blunter notches do serve to mitigate catastrophic failure by reducing local driving forces (lower ΔK) within a larger notch-affected zone than sharp notches. However, this reduction may not offset the tradeoff in fatigue properties exhibited by highly crosslinked UHMWPE and a focus on optimizing the microstructure of this polymer may be more prudent for increasing its durability in vivo.

There are currently over 2 million total joint replacement procedures completed every

year worldwide. The number of total joint replacement procedures is predicted to grow. At

the same time, the percentage of younger, more active patients undergoing joint replacement

procedures is also predicted to increase. Such predictions motivate a need for joint replacements to last longer and perform better in more active patients. The greatest challenge for current joint replacement device designs is wear-related failures. One potential solution to

improve wear performance is to use alternative materials with superior wear performance.

Polycarbonate polyurethane (PCU) has been proposed as an alternative material to improve

the performance of joint replacements. It is softer and more elastomeric than the current

standard polymer, ultra-high molecular weight polyethylene. It has been hypothesized that,

due to its lower elastic modulus, PCU will have improved lubrication performance, reducing

wear. Historically, improvement to ultra-high molecular weight polyethylene wear performance

has come at the expense of fracture resistance. Therefore, in the highly cyclic loading

environment of a total joint replacement, the fatigue and fracture properties are also important

to consider.

This thesis evaluated the long-term mechanical performance of PCU in orthopedic implant

applications. First, we characterized the fatigue crack growth mechanisms in PCU

as a function of loading, frequency, hydration, and thermal annealing treatment. We found

highly time-dependent behavior as mechanisms of crack growth and failure changed with

loading regime. Second, we looked at changes in the structural organization of PCU as a

function of thermal treatments and strain. We found trends of increasing ductility with

increasing annealing temperature and increasing hydrogen bonding in the ordered hard domain. The amount of hydrogen bonding decreased with increasing strain. Finally, we used an 3D-transient elastohydrodynamic lubrication model to characterize the lubrication regimes in PCU during a simulated gait cycle. Our model contradicts the optimistic predictions

of 1D-steady state modeling and may explain the disconnect between early modeling and

the experimental wear results for PCU. Moving forward, this work serves as a foundation

for many questions that still need to be answered to understand the fatigue, fracture, and

tribological performance of this complex material in long-term clinical implant applications.