J. Am. Ceram. Soc., 85 [2] 401– 407 (2002)

journal

Improving the Durability of a Biomedical-Grade Zirconia Ceramic by the Addition of Silica Laurent Gremillard, Jerome Chevalier, Thierry Epicier, and Gilbert Fantozzi Department of Metallurgy and Physical Properties of Materials, Associate Research Unit 5510, National Institute of Applied Sciences, 69621 Villeurbanne Cedex, France The lifetime of tetragonal zirconia ceramic implants is controlled by their resistance to slow crack growth and lowtemperature degradation (LTD). A decrease of the grain size is often proposed to decrease the effect of LTD on zirconia; however, this decrease also reduces the resistance to crack propagation. An alternative approach is reported here; such an approach consists of introducing a silica glass phase only at triple grain junctions, without changing the grain size. The addition of a small amount of silica (0.5 wt%) decreases the internal stresses, which improves the resistance to LTD without affecting the resistance to crack propagation. These results are discussed from a microstructural standpoint. I.

ceramic. In stage II, the speed is limited by the diffusion of corrosive species toward the crack tip (where the reaction occurs). Stage III (the highest KI values) represents fast fracture under vacuum conditions (essentially independent of the environment). A threshold below which no propagation occurs defines a “safety factor” for the use of bioceramics; such a threshold has been clearly observed in tetragonal zirconia ceramics.9 TZP has also been shown to exhibit good resistance to SCG,9 because the overall V–KI diagram of TZP is shifted toward higher KI, compared to that of alumina.10 Indeed, during propagation, stress concentrations at the crack tip provoke a local tetragonal to monoclinic (t–m) transformation. This transformation is accompanied by a volume increase of ⬃4%, which creates compressive stresses at the crack tip that act to reduce the crack driving force. The coarser the grain size, the greater the grain transformability and, consequently, the greater the transformation toughening. Thus, increasing the grain size improves the resistance to SCG.6 At the same time, increasing the grain size decreases the stability of tetragonal zirconia, thus favoring the isothermal transformation of tetragonal grains to the m-phase at the surface of zirconia. This mechanism is known as low-temperature degradation (LTD).11 The volume increase associated with the t–m transformation induces a microcracking and an increase of the surface roughness. Such an increase of roughness with time may increase the wear rate.12 The different experimental features of the LTD in TZP are as follows:13–16 (1) The transformation is slow at room temperature and shows a maximum rate at ⬃200°–300°C. (2) The transformation is accelerated in humid media (in presence of water or water vapor). (3) The transformation starts on the surface and extends into the bulk. Different mechanisms have been proposed to explain the aging phenomenon. In all models, the roles of the water molecule and internal stresses are emphasized.13–17 However, the real mechanism has not been identified yet. A recent investigation has shown that the transformation occurs via a nucleation and growth mechanism.16 Observations have shown that the LTD begins via the transformation of one or a few grains. Then, the transformation propagates from the initially transformed grains to their neighbors, because of the formation of stresses and microcracks around them (the growth step of the nucleation– growth aging kinetics). Increasing the grain size of zirconia ceramics decreases their stability; therefore, the LTD is even more important when the grain size is coarse.18 Some authors have emphasized the key role of internal stresses on LTD. Indeed, high internal stresses17 should increase the transformation rate. However, no experimental evidence for the real influence of residual stresses on LTD rates has been established. It is expected that a period of 25 years in vivo is necessary to induce consequent damage due to LTD of a sample surface.19 This time period is long enough for the case of modern orthopedic implants, which have an average lifetime of 10 years. However, considering an aging population and a growing demand for

Introduction

C

ERAMIC materials are now becoming a competitive alternative to the common metal–polyethylene or metal–metal joint prosthesis devices in orthopedic surgery, because they strongly reduce the amount of wear debris.1 Alumina became very popular in the 1980s, but fracture problems have been reported in the literature.2,3 Then, zirconia ceramics were introduced in orthopedics in the early 1990s, to solve the problem of fracture of ceramic femoral heads. Indeed, tetragonal zirconia ceramics that have been stabilized with yttria (so-called Y-TZP) exhibit high toughness (KIC) (up to 7 MPa䡠m1/2) and strength (up to 2 GPa). Such values are due to both small grain size and a phase-transformation toughening mechanism. Currently, more than 350 000 zirconia femoral heads have been implanted, with good clinical success.4 Zirconia ceramics can exist as two polymorphs at room temperature: a tetragonal, metastable phase (the t-phase, which can be obtained using stabilizers such as yttrium or cerium oxides), and a monoclinic, stable phase (the m-phase). A cubic phase exists in zirconia with higher concentrations of stabilizers. The t-phase is generally obtained after sintering yttria-stabilized tetragonal zirconia polycrystals (the so-called Y-TZP) ceramics that have 2–3 mol% Y2O3; however, this phase can transform to the m-phase, for example, if energy is added into the environment by elastic or chemical stimuli. When considering a given ceramic for orthopedic prostheses, one must focus on its resistance to slow crack growth (SCG). This phenomenon occurs at stress intensity factors (KI) less than the toughness KIC, because of stress-assisted corrosion by water molecules,5–7 which may be located at the origin of an implant fracture.8 It is illustrated by plotting the crack-propagation speed V versus the stress intensity factor KI (V–KI diagrams). These diagrams generally exhibit three stages:5–7 in stage I (lowest KI values), the crack speed is limited by the reaction rate between the water molecules and the chemical bonds of the

P. F. Becher—contributing editor

Manuscript No. 187847. Received March 19, 2001; approved October 15, 2001.

401