Recent advances in understanding the fatigue and

Author’s Accepted Manuscript Recent advances in understanding the fatigue and wear behavior of dental composites and ceramics Jamie J. Kruzic, Joseph A. Arsecularatne, Carina B. Tanaka, Mark J. Hoffman, Paulo F. Cesar www.elsevier.com/locate/jmbbm

PII: DOI: Reference:

S1751-6161(18)30756-2 https://doi.org/10.1016/j.jmbbm.2018.08.008 JMBBM2919

To appear in: Journal of the Mechanical Behavior of Biomedical Materials Received date: 24 May 2018 Revised date: 6 August 2018 Accepted date: 8 August 2018 Cite this article as: Jamie J. Kruzic, Joseph A. Arsecularatne, Carina B. Tanaka, Mark J. Hoffman and Paulo F. Cesar, Recent advances in understanding the fatigue and wear behavior of dental composites and ceramics, Journal of the Mechanical Behavior of Biomedical Materials, https://doi.org/10.1016/j.jmbbm.2018.08.008 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Recent advances in understanding the fatigue and wear behavior of dental composites and ceramics Jamie J. Kruzic1,*, Joseph A. Arsecularatne2, Carina B. Tanaka1 Mark J. Hoffman1,2, Paulo F. Cesar3 1

School of Mechanical and Manufacturing Engineering, UNSW Sydney, NSW 2052, Australia 2

3

School of Materials Science and Engineering, UNSW Sydney, NSW 2052, Australia

Department of Biomaterials and Oral Biology, School of Dentistry, University of São Paulo, São Paulo, Brazil

Abstract Dental composite and ceramic restorative materials are designed to closely mimic the aesthetics and function of natural tooth tissue, and their longevity in the oral environment depends to a large degree on their fatigue and wear properties. The purpose of this review is to highlight some recent advances in our understanding of fatigue and wear mechanisms, and how they contribute to restoration failures in the complex oral environment. Overall, fatigue and wear processes are found to be closely related, with wear of dental ceramic occlusal surfaces providing initiation sites for fatigue failures, and subsurface fatigue crack propagation driving key wear mechanisms for composites, ceramics, and enamel. Furthermore, both fatigue and wear loadings of composite restorations may be important in enabling secondary caries formation, which is the leading cause of composite restoration failures. Overall, developing a mechanistic description of fatigue, wear, and secondary caries formation, along with understanding the interconnectivity of all three processes, are together seen as essential keys to successfully using in vitro studies to predict in vivo outcomes and develop improved dental restorative materials. Graphical abstract:

Keywords: Fatigue, wear, dental restorations, ceramics, resin composites, enamel, secondary caries 1

Introduction

Mineralized tissues make up the load bearing structure of the human tooth. These tissues include enamel, dentin, and cementum, and are all composed of similar building blocks: nanocrystals of 1

carbonated hydroxyapatite, an organic protein phase, and water. In all cases, a high degree of mineralization gives these tissues the mechanical strength and hardness needed for teeth to perform their key function of mastication. For tooth tissues, the degree of mineralization is ~45% to 50% by volume for dentin and ~95% by volume for enamel (Kinney et al., 2003; Robinson et al., 1995). Long term degradation processes such as fatigue, wear, and caries often create a need to repair human teeth using dental restorative materials. Of all the possible restorative materials, dental composites and ceramics are excellent at mimicking the aesthetics of natural tooth tissue while also possessing a similar amount of hard ceramic phase. Resin based dental composites typically contain ~45 to 60% ceramic phase by volume (Randolph et al., 2016; Willems et al., 1992), closely mimicking dentin, while dental ceramics (100%) more closely mimic enamel, and ceramic/polymer hybrids (~75%) fall in-between (Coldea et al., 2014). Unfortunately, these restorative materials are also subject to degradation by fatigue and wear processes that can lead to the eventual failure, and replacement, of the restoration. Furthermore, recent research suggests that mechanical cyclic loading can accelerate the formation of secondary caries lesions that develop in the tooth tissue at the margins of restorations (Khvostenko et al., 2015), which is the most common reason for composite restoration replacement. There has been extensive research aimed at understanding the fatigue and wear processes that cause the degradation and failure of dental composites and ceramics in the oral environment due to masticatory and parafunctional forces. Some of the recent research suggests significant interconnectedness of fatigue and wear mechanisms in these materials, and the importance of these processes in secondary caries formation at composite restorations. Accordingly, the purpose of this article is to review those recent advances while putting them in context with our broader understanding of fatigue and wear mechanisms. 2

Resin Based Dental Composites

2.1 Overview Resin based dental composites are direct dental restorative materials used to restore cavities caused by caries, non-carious cervical lesions, and missing tooth tissue that has been lost by fracture or worn away due to bruxism (Ferracane, 2011; Perez et al., 2012; Zhou and Zheng, 2008). Their usage has increased significantly in recent years due to their good aesthetics, ability to bond to tooth structures, their ease of placement relative to indirect restorations, and the desire for an amalgam alternative to eliminate the use of mercury in dental practices. The microstructure of dental composites generally consists of an organic polymer matrix, typically dimethacrylate based resin composed of bisphenol A-glycidyl methacrylate (bis-GMA) mixed with triethylene glycol dimethacrylate (TEGDMA), urethane dimethacrylate (UDMA), or other polymers to reduce viscosity. More recently, quaternary ammonium methacrylate, silorane, and tricyclodecane (TCD)urethane based materials have been explored, and a recent review on dental composite resin development can be found in (Fugolin and Pfeifer, 2017). The polymer matrix is reinforced with hard inorganic particles such as glass, quartz, colloidal silica, or zirconia. The composition and size of the filler particles have a profound effect on the mechanical, tribological, thermal, optical, and polymerization shrinkage properties (Ferracane and Palin, 2013). Furthermore, there has been success in improving the mechanical and polymerization shrinkage properties of dental composites by the addition of short fibers (Garoushi et al., 2018), although clinical usage of short fiber reinforced composites (SFRCs) is still quite limited. 2.2 Classification of Microstructures While some classifications of dental composites focus on characteristics that define their viscosity and consistency for the clinician (e.g., flowable or packable), when considering the mechanical properties of dental composites, it is best to consider their microstructure classifications. The microstructures are often classified based on the size, and size distribution, of the filler 2

particles/fibers, as seen in Figure 1. An excellent review of the chronological development of the various microstructures has been published recently (Ferracane, 2011); accordingly, only the more recent and relevant composites are summarized here. The most commonly used composite types include (Ferracane, 2011; Ferracane and Palin, 2013; Velo et al., 2016): Microfilled composites. They contain silica particles (0.01 to 0.04 m) and are highly polishable due to small particle size. Because of their low strength they are used in anterior or cervical restorations. Hybrid composites. The filler loading in these composites is ~75 wt% and contains two types of particles: ground glass (1.0 to 10 m) and colloidal silica. Colloidal silica is ~ 20 wt% of the total filler loading. These composites are being replaced in the market by newer micro-/nano-hybrid and nanofilled composites. Micro-/nano-hybrid composites. These are universal (anterior and posterior) materials. They possess the strength and wear resistance required for posterior restorations, and also the polishability and aesthetics required for anterior restorations. These materials are a combination of microscale and nanoscale (~ 20 nm) particles. A typical filler loading is ~78 wt% with the microscale filler sizes in the range of 1 to 3.5 m for microhybrid and 0.4 to 1 m for nanohybrid composites. Nanofilled composites. These are also universal restoratives with nanofillers dispersed either individually or as agglomerates in the resin matrix. The nanofillers are in the range 5 to 20 nm and clusters vary from 0.6 to 1.4 m. Some of the recent composites have very high filler loading, up to 92 wt%. Short fiber reinforced composites (SFRCs). The inclusion of glass fibers along with traditional particle reinforcements can give improved polymerization shrinkage properties along with higher strength and fracture toughness, making them particularly attractive for high stress bearing areas (Garoushi et al., 2018). Composites with combined particulate and fiber fillers up to 84 wt% have been developed. An alternative classification system has been recently proposed by (Randolph et al., 2016) based simply on the inorganic filler volume content. In this classification, a resin composite containing less than 50% of filler is classified as ultra low-fill, between 50 and 74% is referred as low-fill, and above 74% is termed as compact. While the filler content appears reasonable for correlating some properties such as elastic modulus and water absorption (Randolph et al., 2016), it is not expected to correlate well to more complex properties such as fatigue and wear due to their dependence on filler size, size distribution, and morphology.

Figure 1. Schematic illustrating the microstructural differences between various classes of resin based dental composites: (a) Microfill; (b) Hybrid; (c) Micro/nanohybrid; (d) Nanofilled; (e) Short fiber reinforced. Panels (a) – (d) reprinted from (Ferracane, 2011) with permission from Elsevier.

To date, long term clinical performance of resin-based composites is not reported to be as good as amalgam in high load bearing posterior restorations (Afrashtehfar et al., 2017; Bernardo et al., 2007; Collins et al., 1998; Opdam et al., 2010; Soncini et al., 2007). Annual failure rates of composite restorations can be as high as 4.6% for high caries risk patients with 3

average lifetimes reported to be as short as six years (Beck et al., 2015; Downer et al., 1999; Hickel et al., 2005; Pallesen and van Dijken, 2015; van Dijken and Lindberg, 2015). Furthermore, the majority of all restorations are replacements of previously failed restorations (Deligeorgi et al., 2001). Secondary caries is thought to be the most common cause of composite restoration failures (Alvanforoush et al., 2017; Bernardo et al., 2007; Marks et al., 1999; Mjör, 1996; Mjör and Toffenetti, 2000; Nedeljkovic et al., 2015; Soncini et al., 2007; Wilson et al., 1997), with composite fracture generally considered the second leading cause (Al Kayed, 1999; Attin et al., 2001; Bernardo et al., 2007; Nedeljkovic et al., 2015). For posterior composite restorations, it has been reported that failures within the first five years are primarily due to composite fracture, with secondary caries dominating beyond five years (Brunthaler et al., 2003). However, this trend with time has not been observed in all clinical studies (Da Rosa Rodolpho et al., 2011) and a review of the numerous causes for restoration replacements based on multiple surveys may be found in (Deligeorgi et al., 2001). 2.3

Fatigue Behavior of Dental Composites

The traditional engineering definition of fatigue is the time dependent fracture of a material due to repetitive cyclic loading (Suresh, 1998). When considering the issue of composite restoration fractures, the importance of fatigue is clear. Teeth experience cyclic loading due to forces from mastication, bruxism, etc., and it is estimated that such cyclic stresses are repeated more than 3x105 times per year for the average person (Garoushi et al., 2007; Wiskott et al., 1995). Such repetitive cyclic loading can lead to subcritical crack propagation in both tooth tissue (Arola et al., 2010; Kruzic and Ritchie, 2008) and restorative materials (Drummond et al., 2009; Lohbauer et al., 2013; Shah et al., 2009b), and eventually cause fatigue fracture of teeth and/or restorations. However, an additional fatigue related concern for composite restorations is margin failure and gap propagation due to cyclic loading superimposed on the interfacial stresses caused by polymerization shrinkage of the resin during curing (Aggarwal et al., 2008; Arisu et al., 2008; Campos et al., 2008; Pongprueksa et al., 2007; Vandewalle et al., 2004). The presence of marginal gaps can enable secondary caries formation (Khvostenko et al., 2016; Khvostenko et al., 2015; Pinna et al., 2017), which, as mentioned above, is thought to be the leading cause of composite restoration replacement. Finally, fatigue behavior has also been related to the wear properties of resin-based composites. Microcrack propagation in a subsurface layer due to repeated occlusal loading has long been thought to be a precursor of clinical wear (Truong and Tyas, 1988) and higher wear resistance has been found for restoratives with higher fatigue crack growth resistance (Truong et al., 1990). Reduced clinical wear can enhance the lifetime of composite restorations as well as help them retain their favorable aesthetic appearance. While understanding the wear behavior of dental composites, and how the micromechanisms of wear are related to fatigue, will be the subject of the later section 2.4, the present section will focus on recent advances in our understanding of both traditional fatigue fractures and secondary caries formation promoted by fatigue loading. 2.3.1

In vitro Studies of Dental Composite Fatigue Fracture

Dental restorations are subjected to complex cyclic loadings and a varying biochemical environment in the oral cavity that makes laboratory simulation challenging. However, clinical trials are expensive, time-consuming, and have uncontrollable variables. Furthermore, it has been shown that other properties such as elastic modulus, flexural strength, or fracture toughness are not good predictors of the relative fatigue properties of dental composites (Belli et al., 2014). While there are challenges in relating measured fatigue properties to clinical outcomes (Garcia-Godoy et al., 2012), it is clear that poor fatigue properties can result in unacceptably high marginal and bulk fracture rates in vivo (Kramer et al., 2005). Thus, in vitro experiments to measure the fatigue resistance are essential to understanding and predicting the performance of dental composites. When studying the in vitro fatigue response of a material, one may adopt either a strength of materials approach, or a fracture mechanics approach. In the former case, cyclic stress is applied to 4

nominally flaw free specimens and the number of cycles to cause fracture, Nf, is measured. Plotting the applied cyclic stress amplitude, a, against the fatigue life gives the well-known stress-life (S/N) or Wöhler curve from which an average fatigue strength may be determined for a desired lifetime. A variant of this methodology commonly used in dental materials research is known as the staircase approach, whereby the goal is to determine the fatigue strength at one specific number of cycles. In this methodology, individual samples are cycled at progressively higher cyclic stress levels for a desired number of cycles until fracture occurs. Then the stress level can be stepped up and down for subsequent samples until sufficient confidence is gained regarding the cyclic stress to achieve that lifetime. In either case, the total fatigue life can be divided into three components: 1) fatigue crack initiation, 2) fatigue crack growth, and 3) catastrophic fast fracture. When determining the factors that affect the fatigue life of a material and how to make fatigue property improvements, a serious drawback of measuring the total lifetime is that it is difficult to 1) deconvolute how a single factor affects the various stages of the fatigue life and 2) deduce the micromechanisms responsible for controlling each stage. Moreover, a single factor may have offsetting effects on the various stages. For example, if a factor plasticizes a material to make crack initiation easier, it will also likely increase the fracture toughness and thus delay the onset of catastrophic fast fracture to larger numbers of cycles. In order to make informed decisions about how to balance the properties of a dental composite for improved in vivo lifetimes, it is important to understand the micromechanisms that govern each stage of the fatigue life, and how various factors affect those micromechanisms. Fracture mechanics methodologies allow one to separately study the second and third stages of the fatigue lifetime, crack propagation and catastrophic fast fracture. Moreover, cracks generally exist in the tooth enamel (Brown et al., 1972) and the fracture of resin based dental restorative composites is thought to initiate at preexisting surface or internal flaws (Rodrigues et al., 2008), suggesting that the first stage of crack initiation may be negligible. The mode I stress-intensity factor, KI, is the global parameter used in linear elastic fracture mechanics which fully characterizes the local stress and deformation fields in the immediate vicinity of a crack tip in a linear-elastic solid (Anderson, 2005). It is defined for a crack of length a under uniform applied stress app as: KI = Yapp(a)½

Equation 1

where Y is a geometric factor of the order of unity. Catastrophic failure (fast fracture) will occur when stresses concentrated around a pre-existing defect surpass the critical level of stress intensity, i.e., the fracture toughness: KIC = Yf (ac)½

Equation 2

where ac is now the critical crack size that causes fracture at the measured fracture stress, f. Furthermore, fatigue crack growth rates, da/dN, are generally characterized in terms of the stress intensity range, ∆K, where ∆K= Kmax –Kmin. Kmax and Kmin are defined as the maximum and minimum stress intensity experienced by the crack during each loading cycle, calculated according to Equation 1. It is common to discuss fatigue crack propagation curves (i.e., plots of da/dN vs. ∆K) in terms of three characteristic regimes of behavior (Figure 2). Regime I is the near threshold behavior, where the crack growth rate is highly sensitive to the applied ∆K and in many materials, the crack will arrest when the applied ∆K level is below a threshold value, ∆Kth. A power law equation, commonly called the Paris law, usually describes regime II (Paris and Erdogan, 1963): da  C(K ) m dN

Equation 3

where C and m are constants defined by the specific material and environment. While Kmax is embedded in Equation 3, for brittle materials its contribution to defining da/dN becomes relatively large and the Paris law is sometimes rewritten as (Chen and Liu, 1991; Dauskardt et al., 1992): 5

(

) (

a

) Equation 4

where C´ = C(1 − R)m, and R is the load ratio, R = Pmin/Pmax. For ductile materials, m >> p, and for highly brittle materials, p >> m (Ritchie, 1999). The duration of crack growth during fatigue also depends on the fracture toughness, which governs the onset of regime III where the crack growth rate accelerates to cause fast fracture as Kmax approaches KIC. It is often seen in fatigue crack growth studies of dental composites that only regime II of crack growth is measured, e.g., in (Kawakami et al., 2007; Loughran et al., 2005; Soappman et al., 2007; Takeshige et al., 2007; Truong et al., 1990). This regime is of high interest because the Paris law is commonly used to predict fatigue life based on the growth of a preexisting crack, and as will be described in section 2.4.3, it can also be used to describe the certain wear mechanisms.

Figure 2. Schematic plot of fatigue crack growth rates, da/dN, versus stress intensity range, ∆K.

During in vitro fatigue studies of resin based dental composites, it is important to maintain a cyclic frequency, υ, that represents the oral conditions. This is because the resin matrix is viscoelastic (Watts, 1994), and attempts to shorten the test duration by using higher frequencies can lead to internal heating during fatigue testing (Loughran et al., 2005). Internal heating can soften polymeric materials and degrade the apparent fatigue properties (Hertzberg et al., 1975). Accordingly, (Braem et al., 1994) have recommended that the testing frequency not exceed 2 Hz, i.e., the upper range of the typical human chewing frequency. Finally, the load ratio, R, is an important consideration in fatigue testing whereby higher load ratios are often associated with worse fatigue properties in materials. Because teeth are generally unloaded each cycle during mastication, load ratios close to zero are common for studies of dental materials, with slight positive or negative values (e.g., R = 0.1) used to maintain continuous loading to the sample and avoid vibration problems in the test setup. Most ductile metals have Paris exponents in the range of 2 to 4 (Ritchie, 1999), polymers range higher up to ~10 (Furmanski and Pruitt, 2007; Williams, 1977), while the m values for brittle materials and their composites are much greater, generally in the range of 10 to 100 (Ritchie, 1999). The m value for hydrated human dentin (υ = 5 Hz, R = 0.1) has been reported to be 13 to 28, with the higher values associated with older donors (Bajaj et al., 2006; Ivancik et al., 2012), and to a lesser degree with increasing depth from the dentin-enamel junction (Ivancik et al., 2011). Furthermore, the m value for hydrated human enamel (υ = 5 Hz, R = 0.1) has been reported to be ~8. Comprehensive reviews of the fatigue properties of tooth tissues can be found in (Arola et al., 2010; Kruzic and Ritchie, 2008; Yahyazadehfar et al., 2014). Paris law exponents for hydrated dental composites have been reported to range widely from 5 to 76 (Kawakami et al., 2007; Loughran et al., 2005; Shah et al., 2009b; Soappman et al., 2007; Takeshige et al., 2007; Truong et al., 1990). The wide range of m values found for dental 6

composites relative to traditional ductile and brittle materials suggest that both ductile and brittle fatigue mechanisms may operate to various degrees for specific composites and test conditions. When considering that the fatigue resistance of dentin varies significantly from donor to donor and by location in the tooth (Bajaj et al., 2006; Ivancik et al., 2012; Ivancik et al., 2011), the fatigue crack growth resistance of various dental composites overlaps this range considerably (Figure 3). Furthermore, composites with relatively lower fatigue resistance tend to overlap the behaviour of enamel (Figure 3). Finally, the authors are unaware of any fatigue crack growth studies on SFRCs; however, based on their relatively high fracture toughness they are expected to have relatively high fatigue crack growth resistance as well.

Figure 3 – A comparison of the fatigue crack growth resistance for various dental composites, human dentin and enamel based on literature data (Bajaj et al., 2008; Bajaj et al., 2006; Khvostenko et al., 2013; Shah et al., 2009b; Soappman et al., 2007; Takeshige et al., 2007).

Next, similar to the fracture toughness properties of human dentin and enamel (Bajaj and Arola, 2009; Ivancik and Arola, 2013; Nazari et al., 2009; Yahyazadehfar et al., 2016), as shown in Figure 4 the fracture resistance of resin based dental composites have been observed to rise with crack extension (De Souza et al., 2011; Shah et al., 2009a; Shah et al., 2009c; Wendler et al., 2018a). Fracture resistance, KR, can then be plotted as a function of crack extension on a rising fracture resistance curve, or R-curve. (Shah et al., 2009c) identified crack deflection and crack bridging as the extrinsic toughening micromechanisms responsible for this effect. Crack deflection (Figure 5) occurs in dental composites because fracture occurs with an interparticle crack path, generally through the resin matrix but in some cases along the particle interfaces (Chan et al., 2007; De Souza et al., 2011; Shah et al., 2009a; Shah et al., 2009c). Assuming the reinforcement particles are sufficiently large, the crack path may deviate significantly from the highest stress direction. The crack-tip stress intensity for a crack which forms an infinitesimal kink at an angle from the plane of the initial crack is given by (Cotterell and Rice, 1980): ( )

( )

Equation 5

Where kI is the local mode I stress intensity factor at the tip of the kink and KI is the stress intensity factor of the main crack, which is given by Equation 1. The coefficient C11 is a function of the angle  and is given by: ( )

( )

(

)

Equation 6

A simple crack kink can lower the crack tip stress intensity by up to ~2/3 as  increases up to 90°. When intact particles remain bridging the crack wake (Figure 5), their effect is to resist crack 7

opening and reduce the stresses intensity at the crack tip, Ktip, relative to the applied stress intensity, Kapp: Equation 7 where Kbr is the bridging stress intensity which has a negative (compressive) magnitude and is a function of crack e tension, Δa (Kruzic et al., 2008; Mai and Lawn, 1987). As additional bridges form with crack extension, the influence of Kbr becomes more significant, leading to the rising Rcurve behavior seen in Figure 4. However, once a significant number of bridges start to fracture in the wake of the crack, a steady state plateau toughness may be reached, also seen in Figure 4. Additionally, while it has not been a subject of study for dental composites, in general, toughening by crack bridging can affect the fatigue crack growth behavior as well (Kruzic, 2009; Kruzic et al., 2005a; Kruzic et al., 2004). While particle reinforced composites tend to dominate the industry due to their generally superior aesthetic properties, the bridging toughening effect can be greatly increased by the addition of short fibers, which can in turn give enhanced fracture resistance (KIC or R-curve behavior) and fatigue properties (Bijelic-Donova et al., 2016; Garoushi et al., 2013; Wendler et al., 2018a).

Figure 4. Example R-curves for a typical mycrohybrid composite, FiltekTMZ250, tested both as-cured and after 60 days of water hydration. Figure reprinted from (Shah et al., 2009a) with permission from Elsevier.

Figure 5. Micrograph showing how reinforcing particles can deflect and bridge a crack in a typical mycrohybrid composite, FiltekTM Z250. Figure reprinted from (Shah et al., 2009c) with permission from Elsevier.

2.3.2

Fatigue Micromechanisms for Dental Composites

As mentioned above, the wide range of m values found for dental composites suggest that 8

both ductile and brittle fatigue mechanisms may operate for specific composites and test conditions; accordingly, both classes of mechanisms will be discussed in this section. The classic mechanism for describing fatigue crack growth in ductile metals and polymers is known as alternating crack tip blunting and resharpening. The micromechanics of this mechanism were first described by (Rice, 1967), and a schematic is shown in Figure 6. In short, during loading, a volume conserving monotonic plastic deformation zone forms ahead of the crack tip while the crack shape blunts. Next, during unloading, a reversed flow, or cyclic, plastic zone forms as the surrounding elastic material compresses the monotonic plastic zone and resharpens the crack tip. During each loading cycle, the new crack area created during blunting is converted to crack length during unloading and resharpening. This mechanism has also been identified for biological tissues such as dentin and bone (Kruzic et al., 2005c; Nalla et al., 2005).

Figure 6. Schematic of the alternating crack blunting and resharpening mechanism. Figure reprinted from (Kruzic et al., 2005c) with permission from Elsevier.

In contrast, fatigue of brittle ceramic materials occurs quite differently. For ceramic glasses and similar materials, an environmentally assisted cracking (EAC) process occurs under static loading. This type of subcritical crack growth is so eti es loosely called “static fatigue”, though it does not meet the traditional definition for fatigue (Suresh, 1998). That said, under cyclic loading EAC can be activated during each loading cycle when the stress intensity rises above a threshold level (KI0). Above the threshold, it is common to identify three characteristic regions of crack growth behavior (Figure 7). Region I is strongly dependent on both the stress intensity factor (KI) and the environmental conditions. Here the crack growth velocity is commonly described by a power law (Evans and Fuller, 1974):

da = AK In dt

Equation 8

where A and n are constants that depend on the material and environment. In region II, the crack propagation velocity is less dependent on, and may be completely independent of, KI because here the velocity is defined mainly by the diffusion rate of the environment to the crack tip. Finally, in region III, the crack propagation rate rapidly accelerates to cause final fast fracture as KI approaches KIC. It is important to note the EAC does not require cyclic loading; rather, it is activated as long as the stress intensity is sustained above the threshold KI0. In that regard, it is considered a time dependent, rather than cycle dependent, crack growth mechanism. 9

For materials with extrinsic toughening mechanisms, such as crack bridging in dental composites, cyclic loading will have the effect to degrade the toughening mechanism, thus promoting crack advance (Ritchie, 1999). For example, crack bridges in a dental composite (Figure 5) can wear, chip, and crack apart, allowing cracks to propagate more easily under cyclic loading. For materials susceptible to EAC, this cyclic degradation will work together with EAC to give both cycle and time dependent components to the crack growth mechanism (Kruzic et al., 2005b).

Figure 7. Schematic of an environmentally assisted subcritical crack growth rate curve showing regions I, II and III.

There have been relatively few studies examining the detailed micromechanisms of fatigue crack growth in resin based dental composites. One such study (Shah et al., 2009b) compared the behavior of a water saturated microhybrid composite (FiltekTMZ250) to that of a water saturated nanofill composite containing micrometer scale nanoparticle agglomerates (FiltekTM Supreme Plus) while continuously exposed to water (Figure 8). The microhybrid showed a classic fatigue crack growth curve like that seen in Figure 2, and a Paris law exponent of m = 5.2 in region II. An interparticle crack path was observed, with the crack propagating through the resin matrix. This observation, coupled with a Paris law exponent similar to ductile polymers, suggests that a blunting and resharpening mechanism was most likely active as the crack propagated though the resin matrix. In contrast, the nanofill co posite showed a large region between ∆K = 0.4 to 0.6 where the growth rate was independent of the stress intensity range (Figure 8). This behavior is similar to typical brittle EAC behavior shown in Figure 7. An examination of the fracture surfaces showed that the water promoted debonding of the particles and agglomerates from the matrix, giving rise to a more brittle fatigue crack propagation behavior than observed for the microhybrid composite. Water induced debonding was also observed in a study of the fracture and strength properties of the same nanofill composite (Shah et al., 2009a). It was suggested that the porous surfaces associated with the nanoparticle agglomerates (Curtis et al., 2008) may not have adequately accepted the silane treatment or may simply be more prone to hydrolytic degradation than the solid particles in the microhybrid composite. While toughening by crack bridging was also found in the same composites, based on the above observations, the fatigue degradation of toughening was likely a secondary or negligible effect in these composites. Overall, such work confirmed what could be inferred from the Paris exponents of various studies, i.e., that the fatigue crack growth mechanisms of resin based dental composites can range from ductile to brittle depending on the specific composite and environmental conditions.

10

Figure 8. Fatigue crack growth curves for some resin based dental composites. Figure reproduced from (Shah et al., 2009b) with permission from Elsevier.

2.3.3

Factors Affecting the Fatigue Properties of Dental Composites

Water hydration is well known to influence the mechanical properties of dental composites by affecting the matrix, filler, or matrix-filler interface (Calais and Söderholm, 1988; Söderholm et al., 1984; Söderholm and Roberts, 1990). The effect of water hydration on the fatigue properties can be negligible, or as described above in section 2.3.2, it can be significantly degrading. For example, (Lohbauer et al., 2003) observed no change in the flexural fatigue strength of a resin based composite after 90 days water storage. In contrast, (Braem et al., 1995) found a degradation in fatigue strength after only one month of water storage for four different resin based composites. (Mirmohammadi et al., 2011) found negligible differences on the fatigue strength of several composites soaked in distilled water, buffered saline, NaOH, or enzyme (esterase,  -amylase) containing solutions, suggesting little sensitivity to additives other than water. A similar story has been found in crack growth studies. (Takeshige et al., 2007) found a continuous decline in the fatigue crack propagation resistance for three commercial (hybrid and microhybrid) composites with increasing water storage times of 0, 1, 2, and 3 months. Similarly, (Khvostenko et al., 2013) found a lower fatigue threshold, lower m values, and higher growth rates at all stress intensity ranges for one commercial and several experimental composites hydrated for 60 days in a bacteria containing media. In contrast, (Truong et al., 1990) found a very weak effect of increasing Paris exponent for water saturated composite samples, which gives improved fatigue thresholds for the water saturated samples, but poorer fatigue resistance at high stress intensity ranges. Such apparently contradictory results should not be viewed as in conflict with one another. Rather, they likely reflect real differences in crack growth mechanisms, such as was found in (Shah et al., 2009b). Some potential effects of water that may accelerate fatigue crack growth include:  

Hydrolytic degradation of the silane coupling agent promoting interfacial debonding (Söderholm et al., 1984), as has been observed for a nanofill composite (Shah et al., 2009b). Plasticization and softening of the resin matrix (Ferracane, 2006a), which could give rise to increased crack blunting per cycle (Figure 6) and higher crack growth rates. Such an effect would tend to lower the fatigue threshold and m values, as observed in (Khvostenko et al., 2013). 11



Water induced EAC of the ceramic filler particles creating a transparticle cracking mechanism.

To the authors’ knowledge, the latter mechanism has not generally been observed to date and the fatigue crack path is mostly interparticle. Thus, the first two mechanisms should be of concern and composites that resist these effects should not show much sensitivity of the fatigue crack growth properties to water exposure. Overall, there should be a continued effort to understand the mechanisms of fatigue for resin based dental composites in order to further understand how critical factors, such as hydration, influence the observed fatigue response and how to develop composite formulations that resist degrading effects. 2.3.4

Fatigue of Restoration Margins and Secondary Caries

While understanding and improving the bulk fracture and fatigue properties of dental composites should lead to improved resistance to restoration fractures, secondary caries is considered the largest cause of restoration failures in vivo (Bernardo et al., 2007; Marks et al., 1999; Mjör, 1996; Mjör and Toffenetti, 2000; Soncini et al., 2007; Wilson et al., 1997). Bacteria in the biofilm (e.g., Streptococcus mutans) metabolize sucrose to lactic acid which can demineralize tooth tissues (Loesche, 1996; Mjör and Toffenetti, 2000). Increased roughness of the restoration caused by wear (see Section 2.4) increases the ability of bacteria to colonize a given area by affecting pellicle formation, potentially creating a favorable environment for secondary caries (Marsh, 1995; Reis et al., 2002). However, while bacterial biofilm formation is required to induce secondary caries, the presence of the biofilm alone does not guarantee a carious lesion will form (Fejerskov, 2004; Mjör and Toffenetti, 2000). Numerous studies have identified the presence of bacteria within marginal gaps between the restoration and the dentin (Preussker et al., 2003; Splieth et al., 2003; Zivkovic et al., 2001), and the presence of marginal gaps is also considered an important factor in secondary caries formation. Even when the resin based composites provide good sealing of the cavity with no marginal gaps as a result of perfect placement by the clinician, polymerization shrinkage stresses that arise during placement, combined with cyclic mechanical loading during function, may lead to local interface failure and marginal gaps (Peutzfeldt et al., 2018). Indeed, many studies have shown the degradation of restoration margins during cyclic mechanical loading (Aggarwal et al., 2008; Arisu et al., 2008; Campos et al., 2008; Pongprueksa et al., 2007; Vandewalle et al., 2004). Furthermore, in vitro studies of the fatigue behavior of bonded composite-dentin interfaces have shown them to have lower fatigue strength and fatigue crack growth resistance compared to dentin or the composite (Zhang et al., 2015). Moreover, the fatigue resistance depends significantly on the choice of adhesive, and in some cases can degrade after storage in artificial saliva (Zhang et al., 2016, 2017). Once marginal gaps form, they can provide sites for bacterial colonies to form (Choi et al., 2000; Pinna et al., 2017). Streptococcus mutans, Actinomyces naeslundii and Lactobacillus casei are the most abundant bacterial species associated with clinical caries (Kidd et al., 1993), with streptococcus mutans strains identified as the most abundant (Loesche et al., 1975; Loesche, 1986). A minimum gap size exceeding 0.4 mm has been suggested for significant bacterial colonization of dental amalgam (Kidd et al., 1995) while in vitro studies suggest the minimum gap size could be as low as 15 m for resin based dental composites (Khvostenko et al., 2016; Khvostenko et al., 2015). The difference between composites and amalgam is consistent with the finding of (Svanberg et al., 1990) that significantly larger Streptococcus mutans colony counts occur at the tooth interface with composite restorations compared to interfaces with amalgam. While in vitro studies tend to find that increasing gaps sizes lead to increased lesion severity at composite margins (Ferracane, 2017), clinical and in situ oral studies have found no correlation between marginal gap size and bacterial colonization for resin based composite restorations (Kuper et al., 2014; Rezwani-Kaminski et al., 2002). However, it is important to note that clinical data is limited and the clinical importance of the physical gap size is an open question. 12

It is important to note, however, that the presence of biofilm and a marginal gap does not guarantee the development of secondary caries (Heintze, 2007; Mjör and Toffenetti, 2000; Turkun et al., 2003). This finding suggested that there may be an additional role of fatigue loading in secondary caries formation beyond simply creating a marginal gap or growing it above a critical size. Several recent in vitro fatigue studies have produced results that support this hypothesis (Askar et al., 2017; Khvostenko et al., 2015; Kuper et al., 2013). In the first of these studies, the hypothesis of (Kuper et al., 2013) was that fluid motion in and out of a marginal gap caused by cyclic opening and closing forces would increase demineralization since the reaction products would be continuously carried away and replaced by fresh fluid. Those authors conducted a study using a lactic acidic (pH = 5) solution and compared both bonded (no gap) and unbonded (gap) bovine tooth-composite interface samples loaded at a frequency of 0.267 Hz using a toothbrushing simulation machine. They found significantly more demineralization occurred for the unbonded samples with the addition of cyclic loading, and demineralization increased with increasing applied cyclic forces. Those authors cited some limitations of their model were the use of bovine teeth and that the experiments were bacteria free. Subsequently, (Khvostenko et al., 2015) created a Streptococcus mutans bacteria based secondary caries model using human tooth samples. While a full description of the testing can be found in (Khvostenko et al., 2015), a brief description will be given here. Disk shaped slabs of human tooth with restorations containing pre-existing marginal gaps (~ 15 to 30 m wide) were placed in bioreactors with the restoration side down on top of ring shaped support (Figure 9a). A combined biaxial bending and shear loading was applied to the samples using a loading rod that contacted the center of the sample (Figure 9a), while bacteria were cultured on each sample at 37°C. Specimen loading was done by alternating 2-hour blocks of 1.5 Hz cyclic loading and 4-hour resting periods for a total of 56 times over ~2 weeks. Non-loaded samples only showed deep penetration of bacteria into the gaps and deep demineralization for two out of six samples while cyclic loading caused deep 100% bacterial penetration for all samples (Figure 10). On average, the bacteria penetration depth was found to be 67% vs. 100% for non-loaded versus load samples, respectively. Such results provide strong in vitro evidence of a synergetic effect of cyclic loading and bacteria exposure that aids bacterial biofilm penetration at the dentin-restoration marginal gaps. While the details of the mechanism are unclear, it was hypothesized that the pumping effect due to cyclic loading may allow nutrients and bacteria to flow more easily into the narrow gaps while refreshing the fluid and removing waste products. This pumping action would potentially make the gap more hospitable for bacteria colonization. This hypothesis was supported to a degree by a follow-up study that showed bacteria penetration into gaps could be slowed by adding mildly antimicrobial bioactive fillers into the composite formulation to make the gap chemistry less hospitable for colonization (Khvostenko et al., 2016).

13

Figure 9. Custom samples and loading configurations developed for fatigue studies of secondary caries. (a) Schematic of the simulated tooth filling sample and the loading configuration used by (Khvostenko et al., 2015). On the right-hand side of panel (a), the dentin is made transparent to observe the composite restoration and support ring. Figure reproduced from (Khvostenko et al., 2015) with permission from Elsevier. (b) Cross-section of specimen used by (Askar et al., 2017). Mineral loss was measured along the interfacial gap to evaluate the wall lesion (W1, W2) and the surface (S1, S2) demineralization. Figure reproduced from (Askar et al., 2017) with permission from Elsevier.

Figure 10. Panels a – d show cross sections of the marginal gaps of four different fatigue loaded samples. See Figure 9a for a schematic of the samples. Red staining shows the bacterial penetration, which was fully to the bottom of the restoration for all (6/6) loaded samples. Subsequent microhardness studies confirmed demineralization occurred to the extent to bacterial colonization. Figure reproduced from (Khvostenko et al., 2015) with permission from Elsevier.

Most recently (Askar et al., 2017) developed a novel Lactobacillus rhamnosus based bacterial model using bovine tooth-composite interface samples and somewhat larger 100 m wide gaps using the loading configuration shown in Figure 9b. Samples were loaded with a frequency of 0.16 Hz by alternating 4 s of loading and 2 s of unloading. Specimens were subjected to a total of 144,000 loading cycles. While demineralization at the dentin surface (locations S1, S2 in Figure 9b) showed no effect of loading, demineralization in the gap (locations W1, W2 in Figure 9b) showed increasing demineralization with increasing load when the load was above a critical threshold value. Furthermore, they found different amounts of demineralization for two different composites. 14

While the number of studies is limited, there is a growing consensus that fatigue loading plays an important role in secondary caries formation, not only with regard to mechanically promoting gap formation, but also by promoting the subsequent bacterial colonization and demineralization. Furthermore, both (Askar et al., 2017; Khvostenko et al., 2016) have found that this latter process can be affected by changing the composite formulation. Accordingly, since secondary caries is the number one cause of restoration failure, there is a clear need for further studies to understand the mechanisms of how fatigue loading promotes secondary caries, and how to formulate composites that better resist this effect to give improved clinical outcomes. 2.4

Wear Behavior of Dental Composites

While the importance of wear for small to medium size restorations may be limited to degrading aesthetics, failure rates are thought to be higher for large restorations, particularly, those involving the functional cusps (Ferracane, 2006b) and patients with bruxism and clenching habits (Pallesen and Qvist, 2003; van Dijken, 2000). Additionally, available evidence on the longevity of restorations often originates from studies in which severe tooth wear was an exclusion criterion and hence the results of these studies do not reveal the restoration longevity of severe wear cases (Hamburger, 2014; Mesko et al., 2016). Moreover, efforts to correlate the wear performance to other mechanical properties, such as strength, hardness, of toughness, have not found good correlations (Hahnel et al., 2011; Lu et al., 2006; Manhart et al., 2000; Nagarajan et al., 2004; Reich et al., 2004; Sagsoz et al., 2014; Schultz et al., 2010; Tamura et al., 2013; Turssi et al., 2003). Thus, there is an ongoing need for qualitative and quantitative assessment of the wear behaviour of dental composites. Although human teeth may suffer wear in the oral environment, the wear resistance of dental enamel is considered to be high despite the challenging and constantly changing oral conditions (Zhou and Zheng, 2008). An ideal restorative material is expected to possess similar tribological behavior to enamel which will minimize the potential to 1) diminish the vertical dimension of occlusion, 2) degrade the strength of the tooth/restoration, and 3) allow biofilm accumulation (Oh et al., 2002). The first macrofill dental composites introduced in the mid-1960s contained large filler particles and showed rapid wear when used on the biting surfaces of posterior teeth (Padipatvuthikul et al., 2010). With the introduction of composites containing smaller filler particles below 10 µm in the mid-1980s and more recent micro-/nano-hybrid and nano-filled composites (Figure 1), wear problems have been substantially reduced. For some recent composites, the wear rate at the occlusal contact area (OCA) is similar to that for human enamel, i.e., 110 – 149 m over three years for composites compared to ~122 m for enamel (Chan et al., 2010; Lambrechts et al., 2006b). Both in vivo and in vitro studies have been used to help understand dental composite wear (Ferracane, 2013; Lambrechts et al., 2006b). Wear is considered to be a particularly difficult parameter to measure clinically because quantitative assessments from casts are not considered accurate unless sophisticated techniques, such as 3D laser scanning with data matching software, are used (Ferracane, 2013; Lambrechts et al., 2006b). Because of the difficulties associated with in vivo methods, and also to aid in the development of new materials, in vitro wear studies are required (Finlay et al., 2013). In vitro wear testing devices can be categorized as tooth brushing machines, two-body wear machines and three-body wear/chewing simulators (Lambrechts et al., 2006a); however, no current machine can fully mimic the in vivo mechanisms (Benetti et al., 2016). The variations in the simulators, antagonist materials, testing parameters, environments, etc. seem to create significant variability in wear results (Heintze et al., 2005; Lambrechts et al., 2006a; Shortall et al., 2002) making comparisons between results of various studies problematic. Large variability between studies, e.g., up to 61% difference in measured wear depth for the same composite on 15

similar wear simulators (Benetti et al., 2016), and/or contradictory results (Iwasaki et al., 2014; Koottathape et al., 2012; Lu et al., 2006; Manhart et al., 2000) have been reported. Nonetheless, in vitro testing is considered an invaluable tool for developing an understanding of the underlying wear mechanisms (Lewis and Dwyer-Joyce, 2005). Such an understanding is essential to interpret in vivo observations and to predict the in vivo composite wear behaviour (Arsecularatne et al., 2016; Hu et al., 2003; Lee et al., 2012). Overall, if researchers focus on understanding the wear mechanisms while ensuring a reasonable correlation between in vitro and in vivo observations, then the results of in vitro wear studies should be useful for the development of improved, longer lasting dental composites. The focus of the following sections is the wear behaviour of dental composites in vitro, in particular the factors that influence wear and the underlying mechanisms. Additionally, some of the relevant in vivo results will be discussed. 2.4.1

Material Factors that Influence Wear Performance

A number of material factors influence the wear behaviour of dental composites including: 1) the properties of the matrix material and the degree of conversion, 2) the content/properties of the filler, 3) the filler-matrix interface bonding, and 4) the relative abrasiveness of filler against the matrix (Ferracane, 2013; Turssi et al., 2003). The resin matrix can significantly affect the wear rate of composites (Soderholm et al., 2001). For example, compared to bis-GMA/TEGDMA composites, UDMA/TEDMA composites have shown significantly less wear due to their higher degree of conversion. Increased resistance to wear is generally noted when:     

The filler particle volume is increased (Condon and Ferracane, 1997; Heintze et al., 2007; Zimmerli et al., 2010), Finer filler particles are used (Turssi et al., 2005), The degree and the depth of polymer matrix conversion is increased (Ferracane, 2013), Stronger interface bonding is achieved (Turssi et al., 2003; Zimmerli et al., 2010), or The difference between the hardness of the filler and that of the matrix is reduced.

To understand some of these trends it is important to consider the wear mechanisms involved, and how the mechanisms are influenced by the composite properties. 2.4.2

Wear Mechanisms for Dental Composites

The mechanisms associated with the wear of dental composites are: abrasive wear (two/three body), fatigue, and corrosion (Lambrechts et al., 2006b; Turssi et al., 2003). Although wear due to adhesion of the mating surfaces is possible, its influence is considered to be negligible under lubricated, oral conditions. Due to preferential wear of the resin matrix, the filler particles in the composite can become hard protuberances, on one or both mating surfaces, and can cause two body wear (Turssi et al., 2003). Moreover, dislodged hard particles from mating surfaces or from third body media can cause three-body abrasive wear. Fatigue wear occurs when repeated alternating loads during sliding contact form plastically deformed regions and initiate microcracks, which propagate under cyclic loading. Finally, corrosion of a composite surface can occur due to its exposure to the environment (e.g., acidic conditions). The surface may be softened or a reaction layer may be formed which can be removed during sliding contact, thereby exposing the underlying surface to continued chemical attack. For corrosion wear, material loss mainly occurs when acting simultaneously with a mechanical wear process such as abrasion. Fatigue wear is also considered to play a major role in the wear of occlusal contact areas (OCAs), while abrasion is considered to affect both OCAs and contact free areas (Schultz et al., 2010). While corrosion and/or corrosion-abrasion can cause substance loss from dental composite surfaces (Lambrechts et al., 2006b; Turssi et al., 2003; Wu et al., 1984), their detrimental influence is not as rapid as that observed for dental enamel (Sarkar, 2000). 16

Generally, the majority of published studies on the wear behaviour of dental composites can be categorized as two-/three-body wear, toothbrushing abrasion, corrosion and corrosion-abrasion. While wear in vivo is considered to be a three-body process (for non-bruxist individuals), wear at the occlusal contact areas that stabilize the vertical dimension has been found to correlate well with two body wear in vitro (Manhart et al., 2000). However, in two-/three-body configurations, due to absence/presence of the third body, the dominant wear mechanisms are likely to be different. In fact, a reversal of wear rankings of the tested composite materials was observed when testing was changed from two-body to three-body configuration (Hu et al., 2002). The following sections will outline the important findings and trends from recent studies in the above areas and connect them to our current understanding of the associated wear mechanisms. 2.4.3

Two-Body Wear

Two body wear studies are generally conducted either with, or without, liquid lubricants such as human/artificial saliva or water. Since the mating surfaces slide under contact load substance loss can occur due to abrasion of the resin matrix by hard asperities of the opposing surface and/or by fatigue wear (Arsecularatne et al., 2016; Iwasaki et al., 2014; Kon et al., 2006; Koottathape et al., 2012). The typical wear surface morphology of each case is depicted in Figure 11. A higher filler loading in a composite can reduce resin matrix abrasion (Heintze et al., 2007), while a higher fracture toughness retards crack propagation in the microstructure which can reduce fatigue wear (Arsecularatne et al., 2016; Schultz et al., 2010). Overall, two-body wear of a dental composite is directly related to its filler fraction and fracture toughness (Heintze et al., 2007). Important trends regarding two-body wear are summarized as follows. 

Adequate filler adhesion to the matrix is essential for wear resistance; indeed, composites with poor interface bonding wear at higher rate due to a fatigue wear mechanism caused by interface debonding assisted crack initiation/propagation (Arsecularatne et al., 2016; Shinkai et al., 2016). In contrast, with adequate interface bond strength, subsurface crack propagation is slow and shallow and fatigue wear mechanisms do not result in high wear loss (Arsecularatne et al., 2016).



Significantly higher two-body wear occurred when the antagonist contained hard large fillers (e.g., ~3.5 m zirconia) compared to antagonists that contained nano-sized and/or softer fillers (Osiewicz et al., 2015). Abrasion (ploughing) by hard large fillers is responsible for the observed high wear in those cases.



While filler loading is clearly an important factor affecting resin matrix abrasion (Heintze et al., 2007), studies on the optimal filler loading for minimum wear have not formed a consensus. While some studies suggest increasing wear resistance with increasing filler content (Hu et al., 2003; Tamura et al., 2013), others have found an optimal wear resistance at low to moderate weight percent filler, for example between 25-50% for nanofilled composites in (Lawson and Burgess, 2015) or closer to 75 wt% for uniform micrometer scale filler particles in (Johnsen et al., 2011). Studies on commercial composites have also shown filler content alone is not a good predictor of two-body wear resistance (Sagsoz et al., 2014; Topcu et al., 2010), although there are some concerns about the accuracy of quoted filler fraction by the composite manufacturers (Lazaridou et al., 2015; Suzuki et al., 2009).



Similarly, there is no clear trend for filler size. While some researchers reported a higher wear resistance of microfilled/nano-hybrid composites compared to nanofilled/hybrid/micro-hybrid materials (Iwasaki et al., 2014; Koottathape et al., 2012, 2014) others reported insignificant difference (Asefi et al., 2016; Hahnel et al., 2011; Suzuki and Wood, 2007; Turssi et al., 2007). Furthermore, some microfilled composites demonstrate relatively low wear from abrasion and/or fatigue wear (Arsecularatne et al., 2016; Iwasaki et al., 2014; Koottathape et al., 2014) because of their relatively low elastic modulus that gives low contact stresses and lower wear loss. 17



Short fiber reinforced composites (SFRCs), which are normally used as bulk base and not as the final filling surface, showed much higher wear than human enamel (Garoushi et al., 2017). Highest measured wear was associated with a SFRC containing larger filler particles.

18

Figure 11. Different wear surface topographies (SEM) following 10,000 two-body wear cycles (Koottathape et al., 2012): (a) abrasion of a microhybrid composite; (b) fatigue wear of a microfilled composite. Figure reproduced from (Koottathape et al., 2012) with permission from Elsevier.

Some confounding factors that may have affected the above findings include (i) large variations in mechanical properties of composites within a given type (Ilie and Hickel, 2009) and (ii) differences in testing configurations/parameters including the antagonists used and the various methods for lubricating the wear surfaces. However, the above results indicate the complex relationships between wear performance, composite microstructure, and wear configuration that can only be fully understood by studying the related wear mechanisms. To understand the mechanisms of two-body wear, (Arsecularatne et al., 2016) recently investigated the wear scar surface/subsurface using scanning electron microscopy (SEM), focused ion beam (FIB) micromachining, and transmission electron microscopy (TEM) following in vitro reciprocating wear tests carried out using self-mating micro-hybrid (DC-1 and DC-2) and microfilled (DC-3) dental composite specimens using an artificial saliva lubricant. The FIB/SEM analyses revealed an abrasive wear mechanism for the microfilled composite (Arsecularatne et al., 2016). In contrast, for the micro-hybrid composites, a fatigue wear mechanism was found with clear subsurface cracks. TEM analysis revealed some filler particle debonding below the wear surface (Figure 12a) and that the subsurface cracks propagated through the resin matrix and around the particles (Figure 12b). For the microfilled composite, in addition to particle pull-out, small wear particles were generated due to lateral crack extension (Figure 12c).

19

Figure 12. TEM analysis showing (a) Wear surface and subsurface of DC-2 composite (arrows indicate crack propagation along particle-matrix interface); (b) Subsurface crack propagation in DC-1 during wear, a partially debonded filler particle (arrow) and crack tip (arrow head); (c) Wear surface and subsurface of DC-3 composite (arrows indicate surface depressions formed due to wear). Figure reproduced from (Arsecularatne et al., 2016) with permission under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

The measured wear volumes for the three dental composites with varying surface preparations (ground/polished) are shown in Figure 13. In addition, wear measurements for human enamel with lubricants of saline (Eisenburger and Addy, 2002), and citric and acetic acids (Wu et al., 2015) are also included in Figure 13.

20

Figure 13. Comparison of measured wear volume for mechanically ground (G) and polished (P) dental composites with artificial saliva (AS) lubricant and for human dental enamel with different lubricants: Saline; AA – acetic acid; CA – citric acid. Figure reproduced from (Arsecularatne et al., 2016) with permission under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

It can be seen that the measured wear volumes for the DC-2 composite are much higher than those for DC-1 and DC-3. Moreover, the wear volume for DC-2 depends on the contact pressure/load while those for DC-1 and DC-3 do not. The measured wear volumes for DC-1 and DC-3 with artificial saliva lubricant are similar to those for human enamel with citric and acetic acid (both at pH 5.5) lubricants. The measured wear rate for DC-2 is much higher than that reported for human enamel in acidic lubricants but lower than that for human enamel in saline lubricant. Despite the presence of a large number of partially detached flake shaped particles on the wear surfaces of the DC-1 composite, the measured wear rate was low (Figure 13). Considering the subsurface microcracks observed at the wear scar, wear of DC-1 seems to occur by fatigue. This is supported by the reported good correlations between both the fracture toughness and the fatigue resistance with clinically observed wear of dental composites (Ferracane, 2013; Truong et al., 1990).The high fracture toughness and/or fatigue resistance likely retarded the formation of large wear particles and hence reduced wear (Truong et al., 1990). In such composites, crack bridging (e.g., Figure 5), crack pinning and crack-deflection induced toughening mechanisms are thought to increase their fracture toughness (Ferracane and Palin, 2013; Kim et al., 1994; Shah et al., 2009a; Shah et al., 2009c). Although the measured wear rate for DC-2 increased at high load, the observed increase is lower than that for human enamel where the wear mechanism is delamination (Arsecularatne and Hoffman, 2012). This suggests that the wear mechanism associated with composite DC-2 was also fatigue. However, it possessed lower fracture toughness leading to higher wear rates. The lower fracture toughness was likely due to poor filler-matrix interface bonding and/or polymerization shrinkage, the latter of which amplified the stress intensity (Arsecularatne et al., 2016; Lohbauer et al., 2013). Such results reveal that even when the same wear mechanism (fatigue) dominates, two composites can show considerably different wear rates, which are influenced by filler-matrix interface bonding and/or polymerization shrinkage. The microfilled composite DC-3, possessed a much lower elastic modulus than DC-1 or DC-2 and the resulting lower Hertzian stresses decreased the onset of fatigue wear in DC-3 (Arsecularatne et al., 2016). It was revealed that the dominant mechanism of DC-3 was abrasion due to lateral crack extension and particle pull out and the measured wear rate of DC-3 was low and similar to that of DC-1. Due to the brittle nature of dental composites (Heintze et al., 2007), the tensile stress generated at the trailing edge of the cusp is considered responsible for the wear scar subsurface cracks (Kato and 21

Adachi, 2002) which causes formation of wear particles due to fatigue. The fatigue crack growth rate in dental composites can be related to the fracture mechanics parameters using the Paris law (Equation 3). By combining Equation 1 and Equation 3 and integrating, the number of cycles to form and remove a wear particle, Nf, may be expressed as: (

(

)

(

) (

)

)

Equation 9

In Equation 9, C, m are from Equation 3, Y is the geometrical factor from Equation 1, and ai and ac are initial and final crack lengths under cyclic applied stress range ∆. The relation between fracture toughness KIC, and ac can be seen in Equation 2. From Equation 9, it can be seen that a larger value for ai, e.g., due to debonding at the filler-matrix interface, will reduce the number of cycles required to generate a wear particle. Additionally, a composite material that displays greater filler-matrix interface debonding possesses a lower fracture toughness (Truong et al., 1990) and hence smaller ac (Equation 2) and lower Nf (Equation 9) than one that resists such debonding, e.g., composite DC-1. This explains why a composite with greater interface debonding gives a higher wear rate than one that resists such debonding (Arsecularatne et al., 2016; Shinkai et al., 2016). 2.4.4

Three-Body Wear

During a three-body wear test in vitro, the dental composite specimen and antagonist are in sliding contact under load with a third-body medium simulating the food bolus (Lambrechts et al., 2006a). Typical third-body media are aqueous suspensions containing: (i) ground rice/millet/poppy seeds; (ii) polymethyl methacrylate (PMMA) or glass beads; (iii) alumina particles (Benetti et al., 2016; Han et al., 2014; Lawson et al., 2012). It can be seen that the hardness of these media vary greatly (e.g., 0.23 GPa for poppy seeds and ~20 GPa for alumina) and hence will profoundly influence measured wear (Schultz et al., 2010). In general, some important trends regarding three-body wear include: 

Composites with greater filler loading show less three-body wear due to reduced abrasion of the softer matrix (Lambrechts et al., 2006a; Osiewicz et al., 2015). This is because large interspaces between filler particles expose the matrix to abrasion (ploughing), filler exfoliation and hence greater wear (Shinkai et al., 2016).



Compared to composites with relatively larger fillers (e.g., microfilled or hybrid materials), those with nanofillers (e.g., nanofilled) demonstrate superior three-body wear resistance (Schultz et al., 2010).



High hardness media promotes much higher wear rates than softer media. For example, glass beads (hardness H = 1.3 GPa) demonstrate more than 30 times higher wear rates compared to much softer ground millet, poppy seeds, or PMMA beads with H = 0.23 – 0.28 GPa (Lawson et al., 2012).



Three-body wear rates are fairly insensitive to whether traditional resin matrix formulations were used compared to newer silorane or TCD-urethane based matrices (Han et al., 2014). Similarly, flowable composite formulations have improved their three body wear rates to become comparable to their traditional counterparts (Shinkai et al., 2016; Sumino et al., 2013) suggesting that filler size, morphology, and loading are much more influential compared to resin formulation (Schultz et al., 2010; Shinkai et al., 2016; Sumino et al., 2013).



Unlike two-body wear, three-body wear rates appear to be insensitive to the size/hardness of composite antagonist fillers (Osiewicz et al., 2015).

22



Adequate filler adhesion to the matrix and fracture toughness are essential; indeed, poor three-body wear resistance occurs when microcracking and/or filler debonding occurs (Figure 14) (Sumino et al., 2013).

While the above trends appear to paint a somewhat simpler picture for understanding three-body wear relative to two-body wear, there are again complications related to the specific mechanisms and testing conditions. For example, wear rates are not necessarily constant and may increase with wear cycles (de Souza et al., 2013), or conversely there may be greater wear during the initial stage followed by a phase with constant wear rate (Schultz et al., 2010). In the latter case, the high wear during the initial stage (up to ~100,000 cycles) is mechanistically attributed to initial surface damage due to the grinding/polishing processes that were used during specimen preparation that later reaches a steady state condition. A comparison of in vitro data (from six wear simulators) with in vivo data revealed a significant correlation with only one simulator (Heintze et al., 2012). While the variability of clinical wear results was very high with mean coefficient of variation 53% due to the significant influence of individual patients (Heintze et al., 2012), there are well known challenges to creating representative wear conditions in vitro. These include choices of (i) testing configurations (linear, rotary sliding), (ii) antagonists (steel, human enamel, dental ceramic), and (iii) testing parameters (contact force, number of cycles, etc.) used in different simulators (Lambrechts et al., 2006a).

Figure 14. Three-body wear surface of a nanohybrid composite after 400,000 cycles revealing microcracking and filler debonding. Also note the large fillers (~20 m) in this nanohybrid material. Figure reproduced from (Sumino et al., 2013) with permission from Taylor & Francis.

Accordingly, there is still an important need for more mechanistic studies of three-body wear of dental composites in order to fully understand the trends and also design in vitro experiments that are representative of the in vivo conditions. Understanding both surface and subsurface wear scar processes, such as microcracking and filler debonding, is needed and may be accomplished through the use of electron microscopy techniques such as FIB/SEM/TEM. By relating such mechanisms to quantitative wear measurements and composite microstructures, a deeper insight into underlying three-body wear mechanisms and the influencing factors will be possible. 2.4.5

Toothbrushing Abrasion

Studies on the effects of toothbrushing abrasion focus not only on composite wear (substance loss), but also on changes in roughness and gloss since these parameters influence the aesthetics of restorations in vivo (Heintze et al., 2010). Additionally, roughness directly influences the deposition of biofilm and bacterial retention, which can promote secondary caries formation. During toothbrushing, abrasive particles trapped at brush filament tips (Lewis and Dwyer-Joyce, 2005) preferentially abrade the resin matrix of a composite restoration resulting in the exposure of the relatively larger filler particles which then dislodge, causing bumps and craters on a composite 23

surface (Al Khuraif, 2014; Braga et al., 2011; Jin et al., 2014; Malavasi et al., 2015; Suzuki et al., 2009). This mechanism gives more severe surface roughening for composites containing significant amounts of larger filler particles, e.g., for microhybrid composites (Figure 15). In contrast, for nanofilled composites the nanoscale particles and the matrix seem to be abraded off together giving a relatively smaller roughness increase (Braga et al., 2011). The increased surface roughness induced by toothbrushing abrasion is associated with reduced reflectivity and gloss (Heintze et al., 2010; Jin et al., 2014; Lefever et al., 2012); indeed, there is a strong correlation between loss of gloss and increased roughness (Heintze et al., 2010; Jin et al., 2014). It is interesting to note that, in contrast, enamel is much better at gloss retention following toothbrushing due to its higher hardness (Jin et al., 2014; Lefever et al., 2012). Some additional trends and observations regarding toothbrushing abrasion include:  

In general, a higher brushing load increases toothbrushing abrasion (Heintze et al., 2010; Kon et al., 2006) due to the larger contact area between the brush filaments and composite surface, increasing the number of abrasive particles involved. Similarly, roughening and loss of gloss by toothbrushing abrasion tends to increase with the number of brushing cycles (Al Khuraif, 2014; Braga et al., 2011; Heintze et al., 2010; Teixeira et al., 2005) although measurable weight loss is not always apparent (Al Khuraif, 2014; Braga et al., 2011).

Figure 15. Resin matrix abrasion and filler dislodgement due to toothbrushing abrasion: (a) SEM image of a microhybrid composite surface before toothbrushing; (b) composite surface after 100,000 brushing cycles. Figure reproduced from (Teixeira et al., 2005) with permission from John Wiley and Sons.



The measured wear or surface roughness following toothbrushing abrasion is not consistent within a composite microstructure type (Suzuki et al., 2009). Moreover, no correlation was found between: (i) surface roughness and wear (Suzuki et al., 2009); (ii) filler size and surface roughness (Heintze and Forjanic, 2005). This is due to the dependency of wear on a number of parameters such as filler shape, distance between fillers, composition of the resin matrix, and bonding between the fillers and the matrix (Heintze et al, 2010).

While some of the mechanisms responsible for composite wear during toothbrushing abrasion have been elucidated, not all observations are readily explained. For example, some studies have found a poor correlation between filler size and surface roughness (Heintze and Forjanic, 2005). Similarly, depth of wear and roughness are not necessarily correlated (Suzuki et al., 2009) since fast material removal can happen in nanofill composites where the surfaces stay quite smooth. Both of these 24

examples suggest that the above mechanistic description of the role of relatively large fillers is overly simplistic. While it is thought that parameters such as the shape of fillers, distance between fillers, composition of the resin matrix and bonding between fillers and matrix are important (Heintze et al., 2010), our current mechanistic understanding of these factors remains incomplete. Moreover, deterioration of composite surfaces (quantified by roughness/gloss) following toothbrushing abrasion observed in vitro does not necessarily show correlation to in vivo observations since patient-related factors such as food/drinks, saliva, and brushing frequency, and variable brushing force all seem to play roles (Heintze et al., 2010). Most mechanistic observations have been limited to composite surface analyses before/after toothbrushing abrasion using AFM and/or SEM (Heintze et al., 2010; Teixeira et al., 2005). Considering that microscopic mechanisms such as filler debonding or subsurface crack initiation/propagation can occur well below the wear surface, in the future subsurface analyses using FIB/SEM/TEM following toothbrushing abrasion is recommended to reveal valuable information relating to the wear mechanisms and the influencing factors involved. 2.4.6

Corrosion and corrosion-abrasion

Chemical degradation of dental composites can occur due to their exposure to water, saliva, enzymes, alcohol, and acids from bacteria, food, drinks, or medicines (Bagheri et al., 2007; Soares et al., 2012; Valinoti et al., 2008), making them more susceptible to wear processes. However, the corrosive damage occurs slowly and it may take a year or longer to develop following the placement of a restorative (Sarkar, 2000). Nonetheless, over time this degradation reduces the hardness and wear resistance of the composite surface. The processes that occur during corrosion of a dental composite are (i) absorption of water into the polymer, (ii) chemical polymer degradation resulting in the formation of oligomers and monomers plus the degradation of silane coupling and filler particles, and finally (iii) formation of pores in the microstructure and release of oligomers and monomers via these pores (Gopferich, 1996; Prakki et al., 2005; Sarkar, 2000). Thus, corrosion leads to interfacial debonding, filler particle dissolution, and matrix cracking resulting in the formation of a soft and porous layer (Sarkar, 2000) which, in turn, may negatively influence the wear behavior. In vitro corrosion testing in water, artificial saliva, tooth bleaching agents (Turker and Biskin, 2003), acidic solutions (Turssi et al., 2002; Valinoti et al., 2008), alcohol containing mouthwashes, or soft drinks (Soares et al., 2012) all produce various amounts of corrosion damage. However, for in vitro corrosion/tribocorrosion tests, a short-term alkaline treatment (e.g., 0.1N NaOH solution) seems to produce surface/subsurface damage similar to that observed in vivo (Bagheri et al., 2007; Sarkar, 2000). Corrosion followed by toothbrushing abrasion in vivo results in the mechanical removal of the softened layer formed on the dental composite surface. Simulations of this process in vitro by corrosive exposure to an alkaline solution followed by toothbrushing abrasion carried out with 11 composites have shown that the increase in wear was in the range 60 – 1400% (Sarkar, 2000). Such results illustrate the importance of considering the oral environment since it can substantially increase the material loss. When acidic soft drinks and artificial saliva are used for corrosion, toothbrushing abrasion caused a significant increase in surface roughness and a decrease in microhardness with acidic drinks being more aggressive than artificial saliva (de Paula et al., 2015). The lower hardness measured following corrosion and toothbrushing abrasion suggests that not all of the softened layer is removed by toothbrushing abrasion. Greater surface deterioration (increase in roughness) was also observed following toothbrushing when composites were treated with hydrochloric acid at pH 1.2 to simulate regurgitated gastric acid (Roque et al., 2015). Hydrochloric acid corroded the composite surfaces, increasing the negative effects of toothbrushing. In addition to the alkaline/acidic solutions/drinks, the influence of tooth bleaching agents on corrosion-abrasion behavior of composite surfaces is also of interest. Hydrogen peroxide based 25

bleaching agents have been found to corrode composite surfaces and form pores and cracks (Turker and Biskin, 2003). Such treated composite surfaces showed greater wear following toothbrushing abrasion (Hajizadeh et al., 2013). The corrosive damage and the associated decrease in microhardness (Atali and Topbaşi, 2011) appears to be responsible for the observed high wear. While hybrid and micro-hybrid composites suffered greater surface damage (increase in roughness/wear), nanocomposites have shown better resistance to the chemical and mechanical assault (Atali and Topbaşi, 2011; Hajizadeh et al., 2013) although the mechanism responsible for this difference is still unclear. 3

Dental Ceramics

3.1 Overview Ceramics are commonly used today as indirect restorative materials for single or multi-unit restorations for dental crowns because of their superior chemical resistance, wear properties, etchability (ability to be bonded), biocompatibility, and aesthetic features that mimic dental structures (Hickel, 2009; Lohbauer et al., 2017; Zhou and Zheng, 2008). Clinically, there are two major applications for this group of materials: 1) as a veneering layer that masks a less aesthetic framework of metal or opaque ceramic, or 2) as an all-ceramic restoration, which includes inlays, onlays, laminates (thin veneers), crowns and bridges. Additionally, orthodontic brackets, dental implants, implant abutments, and ceramic denture teeth are available from different types of ceramics (Powers and Sakaguchi, 2006). That said, their brittleness and tendency to fracture and/or chip over time due to fatigue mechanisms remains a challenge. Another issue is the observed high wear of either the opposing enamel or both the enamel and ceramic itself (Etman et al., 2008; Krämer et al., 2006; Mitov et al., 2012; Suputtamongkol et al., 2008). In particular, wear of opposing teeth is prevalent in patients with bruxism (Theocharopoulos et al., 2013a). 3.2 Classification of Dental Ceramics Since the first application of ceramics in dentistry in 1774, a wide range of dental ceramic materials have been developed with large differences in chemical composition and microstructure (Zhou and Zheng, 2008). Dental ceramics can be classified according to the clinical application, fabrication method or microstructure (Mainjot, 2016). Since the clinical applications of dental ceramics are directly related to their mechanical properties (Powers and Sakaguchi, 2006), classifying these materials according to their microstructure is the most relevant for this review (Table 1). Dental porcelains are mostly composed of feldspar (more than 90%) and therefore are named feldspathic porcelains. After sintering and glazing, the microstructure of dental porcelains consists of a predominant a glassy phase with up to 30% of leucite crystals heterogeneously distributed within the material. These microstructures result in relatively low strength and fracture toughness, on the order of 70 MPa and 1.0 MPa.m1/2, respectively (Cesar et al., 2005). To improve porcelain properties, a variety of glass-ceramics have been developed based on different crystalline phases (Mainjot, 2016). Typical hot-pressed, leucite-reinforced glass-ceramics contain up to 45% by volume of homogeneously dispersed tetragonal leucite crystals, resulting in somewhat higher improved strength and fracture toughness, on the order of 100 MPa and 1.2 MPa.m1/2, respectively (Guazzato et al., 2004). Glass-ceramics reinforced with ~70% by volume of lithium disilicate crystals show even better properties with strength and fracture toughness values ranging up to approximately 300 MPa and 3 MPa.m1/2, respectively (Guazzato et al., 2004). The enhanced mechanical properties are mainly attributed to a microstructure consisting of elongated highly interlocked crystals and reduction of the glass content. Slip casting can be used to create a ceramic framework with high crystalline content (up to 74 vol.% of Al2O3, MgAlO4, or zirconia-toughened alumina) infiltrated with glass (Swain et al., 2016). However, such infiltrated ceramics are being progressively discontinued due to increasing usage of polycrystalline ceramics such as alumina and zirconia. Zirconia has the highest mechanical 26

properties available to use as framework material in multi-unit and posterior prostheses and is generally preferred over alumina, but often has the disadvantage of being opaque. More recently developed translucent zirconia ceramics sacrifice those favorable mechanical properties. Translucent zirconia ceramics are sometimes described as fully stabilized due to their high content of cubic phase (53%) to reduce light scatter, but they have been reported to have lower strengths (Flinn et al., 2017; Pereira et al., 2018; Stawarczyk et al., 2017). Recent developments in dental ceramics include glass-ceramics with zirconium oxide additions dissolved in the glass matrix along with lithium disilicate crystals (e.g., Suprinity from Vita), hybrid ceramic-polymer materials, such as polymer infiltrated ceramics (e.g., Enamic from Vita), thirdgeneration translucent zirconia ceramics, and new processing methods that improve the mechanical behavior of veneering layers for multilayered restorations (e.g., CAD-on, Rapid Layer, or Press-on methods) (Basso et al., 2016; Li et al., 2014).

Table 1- microstructural classification of dental ceramics Material Type

Group

Feldspathic porcelain

crystalline phase

Leucite (from 1 to 30 vol%)

Manufacturer

Clinical indication

IPS d.Sign (Ivoclar Vivadent); VM7, VM9, VM13, Mark II (VITA Zahnfabrik); EX-3, Cerabien ZR (Noritake)

Mark II (VITA Zahnfabrik); Cerec blocks (Sirona)

IPS Ceramic 35 vol% of leucite

Glass-ceramics

70 vol% of lithium disilicate

Disadvantages

veneers, inlays, onlays and veneering of metallic and ceramic frameworks

very aesthetic high adhesion

hand-made feldspathic ceramics present high quantities of flaws and porosity

Veneers; inlays; onlays; anterior and posterior crowns; overlay veneers for multi-unit frameworks.

material is more homogeneous, containing fewer flaws, and then exhibits more strength than hand-made feldspathic ceramics

less aesthetic compared with materials made by stratification

CAD/CAM blocks

Although the strength is approximately twice of the conventional feldspathic porcelains, it remains limited application for posterior crowns and bridges.

Powder/liquid layering ceramic Hot pressed, injectionmolded ceramic and, CAD/CAM blocks

The optical properties of the blocks are completely different from the final restoration

CAD/CAM (blue blocks)

Veneers; inlays; onlays; anterior and posterior crowns.

More homogenous leucite distribution and slightly higher strength than feldspathic porcelain;

Veneers; inlays; onlays; anterior and posterior crowns; anterior and posterior implant abutments; three-unit bridges up to premolars; overlay veneers for multi-unit frameworks.

Nearly three times more resistant than leucite reinforced glassceramics

e.max

Ceram IPS Empress Esthetic IPS Empress® CAD (IvoclarVivadent)

IPS e.max CAD (IvoclarVivadent)

Processing technique

Advantages

Sintered Porcelains Powder/liquid

27

Glass Infiltrated ceramics (crystalline phase: 70 vol%)* These materials are no longer commercially available

lithium silicate with 10 wt% of zirconia dissolved in the glass matrix

Vita Suprinity (VITA Zahnfabrik); Celtra Duo (Dentsply DeTrey)

Veneers; inlays; onlays; anterior and posterior crowns.

According to the manufacturer it has high degree of translucency, fluorescence and opalescence.

Magnesium (MgAl2O4) and alumina (Al2O3)

In-Ceram Spinell (VITA Zahnfabrik)

framework of an all ceramic restoration

good translucency

Alumina

In-Ceram Alumina (VITA Zahnfabrik)

Laminate cores, crowns, 4-unit bridges.

Zirconia (onethird ZrO2) and Alumina (two-thirds Al2O3)

In-Ceram YZ (VITA Zahnfabrik)

3.3 3.3.1

Glassy network infiltrated with dimethacrylates (UDMA and TEGDMA)

CAD/CAM block

Slip casting or CAD/CAM block

slightly stronger than lithium disilicate reinforced glassceramics

poor translucency No adhesion

Slip casting or CAD/CAM block

High strength (around 50% higher than InCeram Alumina)

very poor translucency

Slip casting or CAD/CAM block

exhibit mechanical properties similar to alumina-based infiltrated ceramic

Slip casting or CAD/CAM block

High strength, biocompatibility, metal-free esthetics.

Alumina

Procera AllCeram, Nobel Biocare;

Zirconia TZP)

Procera Zirconia (Nobel Biocare); Lava/Lava Plus (3M ESPE); Zirkon (DCS); Katana Zirconia ML (Noritake); Cercon ht (Dentsply); Prettau Zirconia (Zirkonzahn); IPS e.max ZirCAD (Ivoclar); Zenostar (Wieland)

Crowns (monolithic or layered restoration), framework for singleand multi-unit (3– 12 elements) fixed-partial dentures.

High strength and fracture toughness.. The newest translucent zirconia systems enable good esthetic monolithic restorations.

Poor adhesion with veneer material; low temperature degradation (LTD), poor translucency of the traditional zirconia.

CAD/CAM block

Enamic (VITA Zahnfabrik)

Veneers; inlays; onlays; anterior and posterior crowns.

Good occlusal contact damage resistance

Low toughness

CAD/CAM block

Polycrystalline ceramic (100% crystalline)

Hybrid materials (ceramic/polymer composites)

Crown and bridge (up to 3-elements) framework in posterior region all-ceramic crowns and bridges, endodontic posts, orthodontic brackets, dental implants and abutments

The optical properties of the blocks are completely different from the final restoration moderate strength (in the lower range of lithium disilicate reinforced glassceramics)

(Y-

None (Ramos et al., 2016)

Fatigue of Dental Ceramics Toughening Mechanisms in Dental Ceramics

Unlike composites, ceramics can’t generally deform plastically to initiate fatigue cracks. Fracture of a ceramic piece can usually be traced to an initial flaw that forms during processing (i.e. pores, inclusions), handling (i.e. adjustment, polishing), or service. Additionally, the material 28

microstructure also has significant influence on the fracture and fatigue behavior by toughening mechanisms such as crack bridging, crack deflection, or transformation toughening. Originally, leucite crystals were added to feldspathic porcelains in a content varying from 1 to 30% with the primary goal to match the coefficient of thermal expansion (CTE) of the metal frameworks of porcelain-fused-to metal (PFM) restorations (Denry, 1996). However, a secondary effect is improved mechanical properties due to toughening by crack deflection (Apel et al., 2008; Cesar et al., 2005; Holand et al., 2009; Yoshimura et al., 2005). Cracks deflect around crystals and cluster boundaries due to the residual stress fields developed as the result of the CTE mismatch between the amorphous and crystalline phases (Cesar et al., 2005). The presence of crystals can further enhance the toughness by crack bridging mechanisms. Overall, crack deflection and bridging, toughening mechanisms that were described in detail in section 2.3.1, are important toughening mechanisms for most dental ceramics, including lithium disilicate glass-ceramics (Figure 16a) (Apel et al., 2008; Holand et al., 2009), glass-infiltrated ceramics (Guazzato et al., 2004), and polycrystalline alumina (Kruzic et al., 2004; Swanson et al., 1987). An example of crack bridging in a glass-infiltrated alumina can be seen in Figure 16b.

Figure 16 – (a) Microstructure and crack deflection in a lithium disilicate glass-ceramic. Figure reproduced from (Holand et al., 2009) with permission from Elsevier. (b) Crack deflection and bridging in an alumina glass-infiltrated ceramic. Figure reproduced from (Guazzato et al., 2004) with permission from Elsevier.

In contrast, yttria-stabilized tetragonal zirconia polycrystals are unique in that they obtain their toughening, and excellent mechanical properties, from a stress-induced martensitic transformation toughening mechanism. When these ceramics experience external stresses (e.g., occlusal load, bruxism), the stress intensification at any existing crack tips causes the metastable tetragonal grains to transform into the monoclinic (t m) phase, resulting in a 4% increase of volume (Evans and Cannon, 1986; Garvie et al., 1975; McMeeking and Evans, 1982). This volume expansion is constrained by the surrounding material and creates compressive stresses on crack tips, lowering the crack tip stress intensity, Ktip, acting much in the same way as crack bridging (Equation 7). Thus, it can be described as: Equation 10 where Ktr is the transformation stress intensity which, like Kbr, has a negative (compressive) agnitude and is a function of crack e tension, Δa. As additional tetragonal crystals transform with crack extension, the influence of Ktr becomes more significant, leading to rising R-curve behavior much like with the crack bridging mechanism (e.g., as in Figure 4). Enamic is a hybrid material composed of a porous feldspathic ceramic structure (Figure 17a) with relatively high elastic properties compared to the low modulus interpenetrating polymer phase. This material was developed to bridge the gap between the properties of the enamel and dentin and mimic the microstructure of natural biomaterials (Swain et al., 2016). The toughening mechanisms 29

identified in interpenetrating polymer network materials are crack bridging and a plastic deformation zone about the crack tip (Swain et al., 2016). The resin phase is able to distribute stresses more effectively in all directions, resulting in excellent resistance to fatigue contact and flexural damage (El Zhawi et al., 2016).

Figure 17 – SEM images of the microstructure of a) Enamic resin infiltrated glass-ceramic and b) Suprinity lithium silicate glass-ceramic with zirconium oxide additions dissolved in the glass matrix. Figures reproduced from (Belli et al., 2017) with permission from Elsevier.

Suprinity and Celtra Duo are glass-ceramics with zirconium oxide additions dissolved in the glass matrix (Figure 17b). However, strength and toughness values are comparable to those of the existing lithium disilicate glass-ceramics (Wendler et al., 2017). One study pointed out that the role of the zirconium oxide as a reinforcing phase may be annulled due to its dissolution in the glass matrix (Belli et al., 2017). The toughening mechanisms present in this glass-ceramic have not been completely understood; however, microstructural analyses suggested that microcracks may form due to the thermal incompatibility among the different phases present in the material (Wendler et al., 2017). Examination of the fracture surface of Suprinity showed that crack propagation occurred preferentially in the glass matrix in contrast with the intergranular path of their lithium disilicate counterpart (Ramos et al., 2016). Despite the efforts to improve the material microstructure, the mechanical properties of recently developed Enamic and Suprinity/Cetra Duo materials have not significantly improved over more traditional feldspathic porcelain and lithium disilicate (Della Bona et al., 2014; Ramos et al., 2016; Wendler et al., 2017). 3.3.2

Environmentally Assisted Cracking (EAC)

It is common that, before fast fracture occurs, the pre-existing defects in ceramic materials will propagate in a slow and stable manner as described in section 2.3.2. For oxide based ceramics and glasses, it is the combination of subcritical stress with the presence of water that progressively breaks the primary chemical bonds of the ceramic material at the crack tip (Michalske and Freiman, 1983), causing subcritical propagation until the crack reached the critical crack size needed for unstable, catastrophic failure. The EAC exponent (Equation 8) is an important parameter that indicates the mechanical strength degradation of the ceramic structure over time. The higher the value of n, the lower is the susceptibility to EAC. However, a recent study has suggested that having a high threshold KI0 for EAC (Figure 7) is a better predictor than n for the reliability of ceramic restorations in vivo (Wendler et al., 2018b). In the case of Y-TZP, simultaneously with EAC, another phenomenon called low temperature degradation (LTD) has been reported; however, since there no clinical evidence available indicating that long-term LTD has a significant effect on the restoration lifetime (Flinn et al., 2017; Sulaiman et al., 2016), it won’t be discussed in detail here.

30

3.3.3

Fatigue in the Oral Environment

Like for composite restorations, laboratory simulation of in vivo fatigue conditions is challenging. The three most important components for consider for dental ceramic fatigue are: a) immersion in artificial saliva, b) temperature and pH variations, and c) load application with movement, frequency and force similar to those found in human mastication (DeLong and Douglas, 1983). The aqueous oral environment induces dissolution of ceramic materials with a predominant glassy matrix, such as porcelains and glass-ceramic materials. While only minimal dissolution occurs for crystalline ceramics (alumina and zirconia) (Swain, 2014), the severity of this dissolution for glass containing ceramics depends on both the glass composition and the pH (Pinto et al., 2008). Further improvements to laboratory simulations are found when adding not only artificial saliva, but also temperature and pH variations (DeLong and Douglas, 1983). This enables the chewing simulation to more accurately take into consideration the chemical influence on both dissolution and EAC. Next, during mastication, the chewing process may be divided into 1) initial surface contact, 2) sliding contact of tooth surfaces across each other, and 3) separation (DeLong and Douglas, 1983). The crushing phase (contact and sliding) lasts approximately 30 s and the force magnitude lies in the range of 9-180 N (Rues et al., 2011). Accordingly, a methodology was developed to reproduce clinical fatigue failure modes fatigue loading such that the load is applied with a 30° inclination (Figure 18) and the cyclic movements include three stages: loading, sliding, and lifting-off (Coelho et al., 2009a). Controlled load application allows quantification of the critical loads necessary for crack initiation, and variations on the load frequency and magnitude result in distinct fatigue response of the material.

Figure 18 – Schematic of posterior teeth in eccentric occlusal position of right side first molar.

3.4

Fatigue Failure Modes for Dental Ceramics

Quantitative fractography has been applied in order to understand the clinical failure of ceramic crowns (Scherrer et al., 2017). Ceramic failure modes can be classified according to the fracture origin location: at the occlusal contact, at the internal cementation surface, at the cervical margin, or at the bridge connector (Zhang et al., 2013). Factors that determine the fracture mode include the material microstructure, prosthesis design, type of load applied, type of origin flaw, presence of residual stresses, quality of the bonding between layers, etc. Understanding the various failure modes is important to identify the source of a problem (manufacturer, dental laboratory, clinician, or patient) and enable corrective actions.

31

3.4.1

Occlusal Contact Damage

Bulk fractures that initiate due to occlusal contact are considered a major cause of failure for allceramic restorations (Lohbauer et al., 2008). Fatigue simulation with spherical indenters on flat surfaces was initially used to explore the fracture modes developed in occlusal contact (Figure 18), with the blunt spheres representing the dental cusp anatomy (Kim et al., 2007b; Lawn and Lee, 2009). Cyclic loading in this configuration gives cracks that initiate on the outside surface of the blunt indenter and propagate downwards and outwards in a stable manner driven by EAC (Lawn, 1998). As the applied load increases, multiple cone cracks expand around the contact circle until they merge with first inner crack on the top surface resulting in a truncated cone configuration. Such brittle damage modes are associated with materials that have a predominant glassy matrix (Figure 19A), such as porcelains and glass-ceramics. However, for coarse-grained and tougher ceramics such as alumina and zirconia, failure occurs due to the coalescence of microcracks in the subsurface zone in a quasiplastic damage mode (Figure 19C). Finally, for glass-infiltrated ceramics with a predominant crystalline phase, a mixed mode is observed, in which a brittle-plastic transition between fracture (cone crack) and quasiplastic damage modes is present (Figure 19B) (Peterson et al., 1998). Finally, a liquid environment has been found to be an essential requirement for inner cone cracks to initiate within the contact zone. Indeed, when cyclic testing was conducted in air, outer cone cracks initiate much more slowly and inner-cone cracks were not observed (Zhang et al., 2005). Moreover, water has been found to facilitate crack propagation downward by a hydraulic pumping mechanism.

Figure 19 – Micrographs of half-surface (upper) and section (lower) of damage modes in glass-ceramics with different grain size (A-fine, B-medium, C- coarse) embedded in a glass matrix. Cone cracking is evident in the fine-grained material (A) and quasi-plasticity damage on the coarse grained material (C). A mixed of both damage modes occurs in medium grained material (B). Figure reproduced from (Peterson et al., 1998) with permission from Elsevier.

While such experiments have helped elucidate the importance of the ceramic microstructure and environment in determining the dominant damage modes of dental ceramics; unfortunately, these observed in vitro cracking modes (outer, inner cone cracks and medial–radial cracks) have rarely been observed in qualitative fractography analysis of clinically fractured ceramic restorations. The fractographic features revealed that most clinical fractures instead originate at wear facets (Scherrer et al., 2017; Scherrer et al., 2007). These origins may also be associated with defects left by the occlusal dentition or by surface finishing procedures conducted during occlusal adjustments (Guess et al., 2010; Sailer et al., 2009). To better represent clinical observations, fatigue testing models evolved progressively to more complex systems involving off-axis cycling load more akin to the actual mastication conditions (Figure 18), first adding sliding contact to studies of flat samples (Kim et al., 2007b), and then using crown shaped specimens (Coelho et al., 2009a) to evaluate the damage mode under mouth-motion sliding contacts (Zhang et al., 2013). The sliding fatigue test (Figure 20) under water has been an 32

important advance in the understanding of the biomechanical behavior of all-ceramic systems under occlusal conditions. This modern in vitro approach provides an evaluation of the interaction between the mechanical fatigue (cyclic load) and wear (sliding friction). In this model, the friction component plays a significant role in strength degradation, as it intensifies the tensile stresses at the crack tip. Additionally, a more distributed damage effect occurs on the surface in which removal of fine particles in the friction zone results in multiple microshearing stresses over intersecting planes and the generation of a large number of microcracks. The propagation of these microcracks is assisted by hydraulic pumping (EAC) and a relatively lower number of cycles are needed to cause fracture relative to pure axial loading (Kim et al., 2007b; Lawn et al., 1984). Indeed, Kim et al. showed that a 30° inclination of the ceramic flat surface can reduce the load required for partial cone cracks to nucleate by 13% when compared with outer cone cracks induced by axial loading. Furthermore, the cracking pattern changes significantly, being described as partial cone cracks, which are a modification of the classical Hertzian cone crack (Figure 20). Partial cone cracks can propagate faster towards the interface, resulting in loss of the veneering material (chipping). On the other hand, off-axis (inclined) loading is less likely to cause fractures from the cementation surface induced by radial cracks due to flexural stresses, as is seen frequently in pure axial loading. The mouth-motion fatigue testing set-up employed in these studies represented an important advance for in vitro simulation and were able to reproduce the failure modes observed clinically, therefore giving a more reliable lifetime prediction for dental ceramics.

33

Figure 20 – Schematic fatigue testing setting (upper) for axial loading (a,b,c) and off-axis loading (d,e,f), in which the specimen is placed with  = 30° inclination. Schematic model demonstrating cone crack evolution assisted by water entrapment (bottom) for fatigue loading with axial (b,c) and off-axis (e,f) configurations. (b) In axial loading the hydraulic pumping process starts with water entering the inner cone crack prior to contact engulfment. (c) As the indenter contact expands, the water is trapped and is squeezed toward the crack tip, causing downward penetration (Kim et al., 2007b). Outer cone cracks, (O) always lies in the Hertzian tensile field outside the contact and water is never trapped in this crack. (e) In off-axis loading water enters the partial cone crack (P) at the trailing edge of the contact at the n cycle, as the indenter slides across the surface in the n + 1 cycle, compressive crack mouth pinching squeezes the water toward the crack tip. (f) Cyclic contact repeats the process, forcing more water into the crack in successive cycles. The schematic is based on (Kim et al., 2007b).

Fatigue studies on anatomically correct crowns have also been conducted by (Basso et al., 2016; Corazza et al., 2015; Guess et al., 2010) using a staircase testing methodology to accelerate the experiments. Hand veneered Y-TZP-based crowns have been found to fail at relatively low load levels (~200N), with 90% of specimens failing from cone crack propagation through the veneer layer resulting in chipping fracture by 100,000 cycles (Guess et al., 2010). Monolithic lithium disilicate crowns were less susceptible to fatigue damage with a threshold for fatigue damage as high as 1,100 to 1,200 N (Guess et al., 2010). For monolithic zirconia crowns, the fatigue damage was characterized only by a wear crater with some microcracks underneath the contact zone (Ren et al., 2011). While these studies apply much higher loads than those found in normal chewing, the in vitro findings of (Coelho et al., 2009a; Coelho et al., 2009b; Guess et al., 2010; Zhang and Kim, 2009) are compatible with observed failures reported in clinical trials (Oilo and Gjerdet, 2013; Pjetursson et al., 2015; Sailer et al., 2015), suggesting that the mouth-motion methodology with staircase loading is a good accelerated in vitro simulator of dental ceramic behavior under clinical scenarios. The authors suggest that this method is suitable for gathering data that can be extrapolated to better understand clinical fractures (Coelho et al., 2009a). A less aggressive approach has been used to understand the fatigue and wear damage in the resin infiltrated ceramic network structure materials (El Zhawi et al., 2016). The polymer infiltrated ceramic material was subjected to a long-term mouth-motion sliding-contact using a clinically relevant load of 200 N. Approximately 5 years in the oral cavity has been simulated using the fatigue/wear testing profile with an extended number of cycles (1.25 million cycles) at a frequency of 2 Hz (El Zhawi et al., 2016). The result of such low load fatigue testing was that none of the crowns tested fractured or presented significant fatigue damage other than minor wear (El Zhawi et al., 2016; Swain et al., 2016). 3.4.2 Chipping - Delamination A common clinical consequence of occlusal contact damage in veneering ceramics is chipping fracture (Pang et al., 2015). The clinical concern regarding chipping fractures grew significantly when Y-TZP was introduced into the market as a core material (Sailer et al., 2007). Indeed, a relatively high clinical incidence of this type of failure has been reported for veneered zirconia restorations (Pjetursson et al., 2015; Sailer et al., 2009; Sailer et al., 2015) compared to the porcelain-fused to metal (PFM) and other all-ceramic restorations (Sailer et al., 2007) Cracks leading to chipping will generally start from occlusal wear facets, as described in section 3.4.1, and when the crack encounters a core material that has higher fracture toughness compared to the porcelain veneer layer, the crack will deflect resulting in chipping fracture or delamination (Belli et al., 2013; Guess et al., 2010). Feldspathic porcelain is the most common material used for veneering purposes and is also the ceramic with the lowest toughness. Even in the case of large chipping fractures, the core material usually is not exposed and a thin veneer ceramic layer remains attached to the core surface, indicating that good adhesion exists between the ceramic materials (Belli et al., 2013; Guess et al., 2010). In addition to occlusal contact damage, residual thermal stress is also an important contributor to the chipping of veneered zirconia crowns in clinical trials (Sailer et al., 2015). Residual thermal 34

stresses are introduced in the veneering material during cooling after the last sintering cycle and are influenced by the cooling rate, coefficient of thermal expansion, and thickness ratio between the core and veneer (Swain, 2009). While compressive stresses in the veneer layer are commonly used for suppressing crack initiation and propagation, potentially damaging tensile stresses can occur for veneered zirconia crowns, especially when produced using fast cooling rates (Lohbauer et al., 2017). Non-uniform solidification from the external surface to the center of the veneer occurs when fast cooling is applied from temperature above Ts (liquid) to temperature below Tg (solid) in the last sintering firing cycle. The outer surface solidifies and contracts first, and then when the inner part starts to solidify, it experiences a rigid external surface that will not allow contraction at the same rate (Meira et al., 2013). Due to the low thermal conductivity of both core and veneering ceramics, applying a fast cooling protocol will develop high levels of thermal gradients and deleterious interfacial stresses that negatively affect the lifetime of veneered zirconia restorations (Belli et al., 2013). In order to reduce the thermal gradient effects, manufacturers have modified their cooling protocol, and recommend that veneered zirconia should be cooled down inside a closed furnace until it reached temperature below the Tg (~600°C). Studies have shown that slow cooling protocols significantly reduce the magnitude of tensile residual stresses within the veneering layer (Belli et al., 2012; Meira et al., 2013; Swain, 2009). However, to the authors’ knowledge, no long-term clinical study has been performed to evaluate the fatigue performance of slow cooled restorations. Thermal expansion coefficient (CTE) mismatch also plays an important role in the development of tensile stress within the veneer layer (Meira et al., 2013; Swain, 2009). The ideal situation to reduce the risk of chipping is to apply a veneering ceramic with a lower CTE than the framework to reduce the magnitude of residual stresses in the porcelain after solidification and consequently increase the chipping resistance (Isgro et al., 2005; Mainjot et al., 2015; Swain, 2009). Finally, there is an additional effect of the thickness ratio on the thermal gradients and residual stress profile of bilayered dental ceramics. In order to reduce the thermal gradient effects on the veneer layer its recommended that 1) the veneer thickness should not exceed two times that of the core (Swain, 2009) 2) the veneering ceramic is fully supported by the core structure and 3) the veneering ceramic has a constant thickness (Benetti et al., 2014). Achieving these requirements can be challenging when the veneering is done manually by a technician. Some recent proposed solutions to reduce veneer chipping failure involve new veneering techniques, such as Rapid Layer and CAD -on (Kanat et al., 2014; Kanat-Erturk et al., 2015). The Rapid Layer Technique uses pre-fabricated CAD/CAM blocks for both components whereby the core Y-TZP and the veneering ceramic are milled separately and then both parts are bonded together via a cementation procedure. Alternatively, the CAD-on technique involves milling a veneering lithium disilicate block that is sintered on top of a zirconia core with a fusion glassceramic between them. These new technologies aim to improve the veneer strength by reducing the shortcomings of hand-veneering technique (Guess et al., 2010), such as high material porosity, residual stresses due to CTE mismatch between core and veneering layers, and non-uniform layer thickness of the veneering material (Lameira et al., 2015). 3.4.3

Cervical Margin (gingival level) Fatigue

Recent fractography studies (Oilo and Gjerdet, 2013; Oilo and Quinn, 2016) have shown that various types of all-ceramic crowns have their fracture origin located on the cervical margin, in the approximal area (Figure 21). These fracture origin sites at the margins are usually related to machining of the ceramic piece. With the increasing usage of CAD/CAM technology, machining defects have frequently been found at the margins of prosthetic crowns. These defects become even more problematic when very thin crown margins are designed. From the crown margins, tensile hoop stresses cause propagation of the flaws along the axial wall of the crown. The crack then moves towards the occlusal surface, and finally propagates to the opposite side of the crown. 35

An in vitro model developed by (Oilo et al., 2014) aimed to explain how the tensile hoop stresses develop. Their hypothesis was that deformation occurs along the tooth axis due to occlusal loading (masticatory forces), resulting in an increase in the diameter of the tooth around the cervical region (Figure 21). The stiff ceramic does not easily accommodate this deformation, generating tensile hoop stresses at the crown margins (Oilo et al., 2014). The model corresponded well with clinically fractured all-ceramic crowns. Also, cervical defects are more prone to propagate trough the framework in the case of less stiff core materials (Oilo and Gjerdet, 2013).

Figure 21 – (a) Schematic of the hoop tensile stress (arrows) acting at the cervical margin of a cemented crown. (b) Image of a fracture veneered zirconia crown located at the approximal area. Panel (b) reproduced from (Oilo et al., 2014) with permission under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

3.4.4

Fatigue of Bridge Connectors

All previously mentioned fracture mechanisms focused on single unit restorations (single crowns, onlays, inlays, veneer laminates). However, the replacement of one or more teeth may be done with multi-unit restorations (fixed partial dentures (FPDs), i.e., bridges). An FPD will bridge an existing gap of missing teeth and on both sides of the prosthetic space, and abutments will give support to the restoration. Abutments can be prepared from natural teeth, tooth roots, and/or dental implants. Studies of all-ceramic dental bridges under cyclic fatigue have shown that fractures are more likely to occur first in the veneering layer, due to the non-toughened nature of the glassy ceramic, before the rupture of the framework material (Studart et al., 2007). In addition to cracks induced by occlusal loading, multi-unit bridges experience high tensile stress at the gingival side of the connector during mastication (Figure 22). The magnitudes of the stresses are determined by the load conditions and by the shape and size of the FPD. For bridges with same shape and size, an increase of the maximum stress on the connector is proportional to the decrease of the connector diameter. Thus, to achieve satisfactory long-term performance of dental bridges under occlusal loading, a correct design of the connector is essential in order to prevent failure of both the veneering layer and the framework. Furthermore, the material used for framework construction also effects the failure mode and stress distribution in fixed partial dentures. Fracture originating from occlusal contact damage will lead to chipping (see section 3.4.2) for tough framework materials such as with veneered zirconia bridges (Pjetursson et al., 2015; Taskonak et al., 2008). In contrast, a higher incidence of bulk fracture has been reported for FDP frameworks made of reinforced glass-ceramic and glass-infiltrated alumina. FDPs have a beam-like geometry, and therefore experience bending forces similar to a three-point bending test. The maximum tensile stresses are thus located on the gingival side of the connector, between the two abutments (Figure 22). Accordingly, radial cracks initiate at the flexural zone and propagate upwards to the occlusal surface through the framework material resulting in bulk fracture. The fracture origin can be either from critical flaw on the porcelain or framework surface, 36

located at the veneer/framework interface, or at the cervical margin. In some cases, the crack has been observed to propagate through both structures, as if it were a homogeneous material, without any crack deflection at the veneer–framework interface (Borba et al., 2015).

Figure 22 – Stress distribution in a 3-unit fixed partial denture framework of a glass infiltrated zirconia-reinforced. The magnitudes of stresses are represented by color gradients. The maximum positive value in the scale (red) represents the maximum tensile stress. The maximum tensile stresses were located underneath the bridge (gingival side) and near the connectors. Figure based on data from (Borba et al., 2015).

3.5

Wear of dental ceramics

As described above, wear facets are often the clinically observed initiation sites for fatigue failures originating at the occlusal surface of ceramic restorations (Kim et al., 2007a; Scherrer et al., 2007). Furthermore, wear of the opposing enamel has long been recognized as a shortcoming of ceramic restorations (Mahalick et al., 1971). Accordingly, there has been a growing interest in understanding both the wear mechanisms of dental ceramics and the opposing enamel. 3.5.1

Factors that influence ceramic wear performance

In general, wear of a brittle materials, such as dental ceramics, occurs primarily due to crack growth mechanisms rather than plastic deformation (Oh et al., 2002). Factors that influence the wear behaviour of dental ceramics include ceramic type, microstructure, fracture toughness, surface quality/roughness, contact load, environment, and the number of cycles (Amer et al., 2014; Etman et al., 2013; Hmaidouch and Weigl, 2013; Hopp and Land, 2013; Miyazaki et al., 2013; Mundhe et al., 2015; Oh et al., 2002; Peng et al., 2016; Stawarczyk et al., 2013; Suputtamongkol et al., 2010; Wang et al., 2012). However, perhaps surprisingly, there is no strong correlation between a cera ic’s hardness and wear of opposing enamel (Hmaidouch and Weigl, 2013; Miyazaki et al., 2013). Overall, lower ceramic and/or opposing enamel wear is generally noted when: 

The ceramic material possesses high flexural strength or fracture toughness (Albashaireh et al., 2010; Dupriez et al., 2015; Santos et al., 2016; Theocharopoulos et al., 2013a; Theocharopoulos et al., 2013b).



The ceramic material, particularly for feldspathic porcelains or leucite glass-ceramics, contains a high fraction of small, uniformly distributed crystals (Min et al., 2015; Theocharopoulos et al., 2013a; Theocharopoulos et al., 2013b).



When the ceramic contains fewer internal defects, e.g., cracks, pores (Min et al., 2015; Theocharopoulos et al., 2013a; Zhang et al., 2008). 37



The surface roughness of ceramic is reduced, e.g., by polishing (Chong et al., 2015; Janyavula et al., 2013; Mitov et al., 2012; Peng et al., 2016; Wang et al., 2012)

In the latter case, glazing a ceramic surface to improve aesthetics tends to increase its surface roughness (Heintze and Forjanic, 2005) and associated enamel wear (Heintze et al., 2008; Passos et al., 2014). Additionally, increases in surface roughness, and ceramic/enamel wear, can result from occlusal adjustment of ceramic crowns at the time of insertion that roughens the ceramic surface (Janyavula et al., 2013). In the literature, both in vivo and in vitro methods have been used to investigate dental ceramic/enamel wear, and as argued in the above sections, in vitro testing provides an invaluable tool. The focus of the following sections is the wear behaviour and mechanisms of dental ceramics and opposing human enamel in vitro, while some of the relevant in vivo results will be discussed to provide clinical context. 3.5.2

Wear mechanisms

As described above in section 3.3, dental ceramics show varying susceptibility to EAC and varying toughening mechanisms, and this may in turn influence the wear mechanisms. When human enamel is in sliding contact with a dental ceramic under high contact load, high tensile stresses are generated due to the high elastic modulus of the ceramic and/or high friction. This can result in subsurface crack propagation in enamel leading to delamination wear (Figure 23) (Arsecularatne et al., 2015; Lee et al., 2014; Wang et al., 2012). Delamination wear of enamel can be extremely damaging since it is associated with severe wear and an exponential increase in wear rate can occur (Arsecularatne and Hoffman, 2010; Mair et al., 1996). Under lower contact loads, however, hard crystals in ceramic can cause abrasive wear of enamel (Oh et al., 2002; Theocharopoulos et al., 2013a) which is normally associated with mild wear.

Figure 23. Topographies during delamination wear of human enamel: (a) wear surface revealing formation of large wear particles. Figure reproduced from (Lee et al., 2014) with permission from Elsevier; (b) subsurface cracks (arrowed). Figure reproduced from (Arsecularatne et al., 2015) with permission under the CC BY-NCND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

In the presence of hard asperities on the opposing surface, abrasive wear under low contact loads can also occur on the ceramic surface due to plastic flow and/or superficial surface cracks and is associated with low friction, relatively smooth wear surfaces and mild wear (Arsecularatne et al., 2015; Guo et al., 2014; Mair et al., 1996; Zhang et al., 2008). Under higher contact loads and high friction, however, subsurface crack propagation (brittle fracture) results in the formation of large wear particles due to delamination (Figure 24). Delamination wear of glass-ceramics is associated with the presence or formation of microcracks (Zhang et al., 2008). Because of this, ceramic materials that possess higher fracture toughness, e.g., zirconia, are able to resist delamination wear 38

better than those ceramics that possess lower fracture toughness, e.g., porcelain (Dupriez et al., 2015; Janyavula et al., 2013). In vitro wear studies have revealed partially detached wear particles and/or subsurface cracks running parallel to the sliding surface in both dental ceramics and enamel (Arsecularatne et al., 2015; Lee et al., 2014; Wang et al., 2012). These indicate that the wear particle formation can occur gradually with a lateral crack extending due to cyclic tensile loading by the sliding antagonist, i.e., delamination wear controlled by a fatigue process. However, under severe contact conditions (high load and friction), delamination wear can instead be due to fast fracture (Kato and Adachi, 2002). The delamination process can thus result in relatively low wear or high wear, for example when considering enamel opposing polished zirconia under low load or enamel opposing rough zirconia under high load, respectively (Wang et al., 2012).

Figure 24. Topographies during delamination wear of a leucite glass-ceramic (LGC): (a) wear surface revealing formation of large wear particles (circled). Figure reproduced from (Zhang et al., 2008) with permission from John Wiley and Sons; (b) subsurface cracks (arrowed). Figure reproduced from (Arsecularatne et al., 2015) with permission under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

3.5.3 Dental ceramic – human enamel system 3.5.3.1 Human enamel antagonist on dental ceramic flat-surface/inlay A review of 20 in vitro wear studies revealed large variations in how the studies were conducted, (Heintze et al., 2008):    

Contact forces ranged from 0.4 – 75 N. The number of wear cycles ranged from 10,000 to 1,200,000. Both flat surface ceramic specimens and inlays were used. The enamel antagonist cusps were either used as obtained or following grinding/polishing processes to form flat/spherical contacting surfaces of identical shape (standardized cusps).

Additionally, only a few studies considered the same ceramic materials. Accordingly, only limited quantitative comparisons were possible and they were inconsistent or contradictory (Heintze et al., 2008). The variability of wear results was very high: 5 – 64% for ceramic materials and 6 – 97% for opposing enamel. A review of five in vivo wear studies also noted contradictory results and hence a comparison of the wear rankings or wear rates of different ceramic materials was impossible (Hmaidouch and Weigl, 2013). This motivates us in the present work to focus on where direct fair comparisons can be made between ceramics, on studies where in vitro and in vivo studies find agreement, and on understanding the relevant mechanisms of the wear process. A comparison of in vitro wear results for three ceramic materials: a feldspathic porcelain, a leucite glass-ceramic, and a lithium disilicate glass-ceramic, revealed greater antagonist enamel wear with 39

lithium disilicate glass-ceramic which was in agreement with in vivo observations (Heintze et al., 2008). This is attributed to excessive abrasion of enamel caused by the hard lithium disilicate crystals at the surfaces and in the wear debris (Figueiredo-Pina et al., 2016). In contrast, for leucite glass-ceramics, the hard ceramic wear particles embedded in the enamel and protected it from abrasive wear (Arsecularatne et al., 2015). When compared to glass-ceramics containing micrometre scale leucite crystals, materials containing sub-micrometre/nanometre scale crystals contained 1) fewer internal defects, 2) displayed higher flexural strength, and 3) exhibited lower opposing enamel wear (Theocharopoulos et al., 2013a). The opposing enamel surfaces of the coarser reinforced ceramics were rougher than those of the latter materials, which indicate higher friction and associated delamination wear of the enamel (Arsecularatne and Hoffman, 2012; Douglas et al., 1985). In recent years, the use of zirconia has increased with full-contour crown restorations added to the market (Jung et al., 2010; Passos et al., 2014). Zirconia ceramic materials in general, cause significantly lower opposing enamel wear compared to feldspathic porcelain (Kim et al., 2012; Lawson et al., 2014; Nakashima et al., 2016; Preis et al., 2012; Preis et al., 2011) and lithium disilicate glass-ceramics (Kim et al., 2012; Nakashima et al., 2016). This is attributed to the much higher hardness and fracture toughness of zirconia (Dupriez et al., 2015) that allows it to maintain a smooth contacting surface with reduced friction (Janyavula et al., 2013). Low friction in turn leads to low abrasion/delamination wear of the opposing enamel (Arsecularatne and Hoffman, 2012; Janyavula et al., 2013). Moreover, feldspathic porcelain and lithium disilicate glass-ceramics suffer greater ceramic wear due to their lower fracture toughness. Furthermore, the rougher wear surfaces and high friction causes the higher abrasive and/or delamination wear of the enamel relative to when zirconia is used (Arsecularatne and Hoffman, 2012; Douglas et al., 1985; Lee et al., 2014). Also, excessive abrasion of enamel by hard lithium disilicate crystals in surfaces and wear debris has been noted (Figueiredo-Pina et al., 2016; Nakashima et al., 2016). Two-body wear can be further reduced by polishing the zirconia relative to glazed or stained zirconia surfaces (Amer et al., 2014; Chong et al., 2015; Janyavula et al., 2013; Jung et al., 2010; Mitov et al., 2012; Park et al., 2014; Passos et al., 2014; Stawarczyk et al., 2013; Wang et al., 2012). Again, this can be attributed to higher roughness, which results in high friction and hence high abrasive/delamination wear. In contrast, the difference becomes insignificant under three-body wear conditions (Amer et al., 2014) where the influence of surface roughness is diminished due to the presence of abrasive media. Although bruxism and other parafunction activities are considered a contraindication for all ceramic dental prostheses, zirconia is sometimes recommended for such patients (Soxman, 2015). While zirconia caused lower opposing enamel wear in vitro (Kim et al., 2012; Preis et al., 2012; Preis et al., 2011), the associated enamel wear mechanism was delamination (Wang et al., 2012). A characteristic of delamination wear is an exponential increase in the wear rate above a threshold contact load (Arsecularatne and Hoffman, 2012). Since the biting forces of bruxist patients can be large (Diracoglu et al., 2011), caution should be exercised when using zirconia for such patients. Additionally, rough ceramic surfaces due to chair-side adjustments or generated during service may increase friction during sliding contact. High friction can initiate delamination wear and cause rapid substance loss even in non-bruxist individuals (Arsecularatne and Hoffman, 2012). Since rough zirconia can cause greater opposing enamel wear (Kwon et al., 2015; Lawson et al., 2014; Mitov et al., 2012), the biting surface of zirconia prostheses should be well polished before service to minimise opposing enamel wear in vivo. Since these surfaces are likely to roughen during clinical service, periodic inspection and polishing may also be necessary. While one in vitro study that used as-extracted enamel cusps reported insignificant differences in enamel wear when the cusps opposed zirconia, lithium disilicate glass-ceramic, or porcelain (Zandparsa et al., 2016), images of worn cusps revealed multiple contact wear regions which are likely to vary between cusps. This is expected to cause large variations in contact stresses between 40

tests and hence measured wear. Such results reveal the profound influence of cusp geometry on measured wear and the need to select/prepare cusps with single contact for wear tests. Overall, various in vitro studies have used: (i) standardized cusps (Arsecularatne et al., 2015; Janyavula et al., 2013; Nakashima et al., 2016); (ii) as extracted cusps (Figueiredo-Pina et al., 2016; Jung et al., 2010; Zandparsa et al., 2016); (iii) both standardized and as extracted cusps (Heintze et al., 2008). While some of the contradictory results can be attributed to the use of non-standard cusps with multiple contacts, the variability of wear with standardized cusps can also be large due to: (i) changes in composition and mechanical properties of enamel; (ii) pores and other defects in ceramic/enamel (iii) differences in ceramic surface roughness/treatment (Heintze et al., 2008). Compared to zirconia, recently introduced hybrid ceramics (Ramos et al., 2016; Zhi et al., 2016) have shown significantly greater wear of both the ceramic and the opposing enamel (Mormann et al., 2013). However, the combined wear of novel hybrid ceramics and opposing human enamel appears to be similar to that of feldspathic porcelain or glass-ceramics (Mormann et al., 2013; Stawarczyk et al., 2015). The measured fracture toughness of the hybrid ceramics is similar to that of feldspathic porcelain (Ramos et al., 2016) and the wear behaviour appears to be similar to the major constituent of glass-ceramic. 3.5.3.2 Ceramic antagonist on human enamel surface With feldspathic porcelain or glass-ceramic antagonists, under relatively low contact loads, the enamel wear is due to abrasion (Wang et al., 2012). However, as the load is increased, both abrasion and delamination mechanisms occur (Arsecularatne et al., 2015; Lee et al., 2014; Nakashima et al., 2016). In contrast, with a zirconia antagonist, delamination is dominant even at low loads (Wang et al., 2012). This can be attributed to the high elastic modulus of zirconia and associated high contact stresses resulting in subsurface crack formation in the enamel causing delamination wear (Arsecularatne and Hoffman, 2012; Dupriez et al., 2015). Because of relatively lower elastic modulus of feldspathic porcelain and glass-ceramic, the contact stresses under low loads are not high enough to initiate subsurface cracks and delamination wear. Enamel wear when using feldspathic porcelain or glass-ceramic antagonists is similar to that with an enamel antagonist (Nakashima et al., 2016). However, similar to the results discussed above for enamel antagonists on zirconia, polished zirconia antagonists cause the lowest enamel wear while exhibiting the lowest wear rates. A parameter that was overlooked in most of these studies is the age of the tooth donor. A co parison of ena el’s brittleness inde (which is a function of hardness, elastic odulus and fracture toughness) reveals a higher value for old enamel (e.g., from individuals above 50 years of age) compared to young enamel, e.g., individuals 18 – 25 years of age (Park et al., 2008a). A material with a high brittleness index is more prone to fracture than one with low brittleness index. At the tooth surface, both hardness and elastic modulus of old enamel was significantly higher than those of young enamel (Park et al., 2008b) but latter shows higher wear resistance (Zheng and Zhou, 2006). Thus, the use of human teeth from a wide donor age range is likely to cause greater variability of measured wear. To reduce this variability, studies using human teeth from a narrow age range is likely needed. 3.5.4

Dental ceramic opposing non-enamel antagonist

Other reported antagonist materials include dental ceramics (glass-ceramics, alumina, zirconia), steatite and stainless steel. With hard zirconia or alumina antagonists, the wear loss of porcelain materials was influenced by the crystal size, content and distribution as well as the matrix (Min et al., 2015). Porcelain materials with a dense matrix and containing a high fraction of small, uniformly distributed crystals showed lower wear dominated by abrasion. However, materials containing large crystals and/or subsurface microcracks displayed higher wear dominated by delamination (Min et al., 2015; Zhang et al., 2008). Again, the wear loss of a dental ceramic against a hard antagonist can be related to the fracture toughness and/or flexural strength of the ceramic 41

material. Materials with high fracture toughness and/or flexural strength show lower wear and vice versa (Albashaireh et al., 2010; Dupriez et al., 2015; Santos et al., 2016). The observed wear loss of ceramic materials was: zirconia < lithium disilicate ceramic < leucite glass-ceramic < fluroapatite/nano-fluorapatite glass-ceramic (Albashaireh et al., 2010). While zirconia seems to wear by abrasion, delamination was dominant for the other ceramic materials. Wear generally increases with contact load whereby for a lithium disilicate glass-ceramics both abrasion and delamination mechanisms were active under low load (e.g., 5 N) while delamination was dominant under high load, e.g., 25 N (Peng et al., 2016). Additionally, the wear loss was greater for rougher surfaces (D’Arcangelo et al., 2016; Peng et al., 2016). Surprisingly, with relatively softer steatite antagonist, glass-ceramics with relatively high fracture toughness actually wear at higher rates (Dupriez et al., 2015). For example, a lithium disilicate glass-ceramic with fracture toughness KIC = 2.5 – 3 MPa.m1/2 showed greater wear than a leucite glass-ceramic with fracture toughness KIC = 1.3 MPa.m1/2. The mechanism appears to be that highly abrasive wear particles embed in the softer antagonist and/or act as third body media to cause greater wear of the higher toughness glass-ceramic. However, consistent with studies using other antagonists, zirconia shows the lowest wear of both ceramic and antagonist and the behaviour improved with polishing of the zirconia (Kontos et al., 2013; Preis et al., 2016). With stainless steel antagonists, the wear mechanism of porcelain transitioned with increasing contact pressure (Guo et al., 2017). Under low pressure, the porcelain wear was mild and dominated by abrasion while under high pressure wear was severe and dominated by delamination (Guo et al., 2017; Guo et al., 2014). 3.5.5

Wear scar subsurface observations

Despite many previous studies on the wear behaviour of dental ceramics, there is a clear need for further understanding of the underlying wear mechanisms (Arsecularatne et al., 2015; Mitov et al., 2012). To gain a more in-depth knowledge on these mechanisms, wear scar subsurface analyses were conducted using FIB/SEM/TEM following in vitro wear tests for a leucite glass-ceramic with ceramic/enamel antagonists. This analysis revealed a dominant abrasive wear mechanism caused by superficial surface cracks (Arsecularatne et al., 2015). With a ceramic antagonist, hard leucite crystals seem to directly act as asperities while with an enamel antagonist, hard ceramic wear particles embed in the enamel surface to then act as asperities. Moreover, localized delamination resulted in the formation of subsurface cracks (Figure 24b). Similar wear behaviour of glassceramics was also observed with both hard (alumina) and softer (steatite) antagonists (Dupriez et al., 2015). A TEM analysis of leucite glass-ceramic wear subsurfaces revealed (Figure 25a): (i) crack propagation along the ceramic-glass interface (at A); (ii) an intragranular crack continuing into surrounding glass matrix or alternatively a crack formed in the glass continuing right through a ceramic grain (at B); (iii) fracture of leucite crystals accompanying crack deflection (at C) (Arsecularatne et al., 2015). Similar crack propagation micromechanisms were also observed during macro indentation (Apel et al., 2008). A TEM analysis of enamel wear subsurfaces (when opposed by leucite glass-ceramic antagonists) revealed (Figure 25b) a rough wear surface and cracks underneath (arrowed) indicating a delamination wear mechanism (Arsecularatne et al., 2015). These wear surface/subsurface observations are consistent with the enamel wear surface topography reported by other researchers (Lee et al., 2014; Wang et al., 2012).

42

Figure 25. TEM analysis of wear surface/subsurface: (a) glass-ceramic; (b) human enamel opposed by glass-ceramic. Figure reproduced from (Arsecularatne et al., 2015) with permission with permission under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

3.5.6

Toothbrushing abrasion

Short duration in vitro toothbrushing tests (equivalent to 1 month of brushing) revealed a significant gloss reduction for feldspathic porcelain and glass-ceramic materials compared to zirconia (Mormann et al., 2013). However, none of these materials showed a significant change in their surface roughness due to short term toothbrushing simulation. In contrast, long duration tests (equivalent to 8.5 years of brushing) of extrinsically (surface layer) stained feldspathic porcelain revealed a significant substance loss and change in surface roughness and shade (Anil and Bolay, 2002). With leucite glass-ceramics, up to 12 years of simulated brushing did not reveal a significant change in shade/gloss or surface roughness (Garza et al., 2016; Heintze et al., 2010). However, a significant change in shade and surface roughness for a lithium disilicate glass-ceramic was noted (Garza et al., 2016). 3.5.7

Attrition-corrosion

As extracted (unpolished) human enamel cusp opposing a lithium disilicate glass-ceramic in vitro caused lower enamel wear in artificial saliva and citric acid solution (at pH 3) than in artificial saliva (at pH 7) (Figueiredo-Pina et al., 2016). However, a leucite glass-ceramic cusp opposing polished human enamel surface revealed greater enamel wear in citric acid solution (at pH 2.6) than in artificial saliva (at pH 7) (Wiegand et al., 2017). These contradictory results can be attributed to: 

The difference in using polished versus unpolished enamel as polished enamel is more susceptible to corrosion than unpolished enamel (Ganss et al., 2000).



Differences in the pH of acidic solutions since a lower pH can profoundly influence attrition-corrosion wear of human enamel (Kaidonis et al., 1998; Wu et al., 2015).



Differences in the wear testing configurations – an enamel cusp (opposing a ceramic flat surface) is in continuous contact and hence less exposed to acid attack compared to an enamel flat surface (opposed by ceramic cusp) which is frequently exposed to the acid solution.

43

3.5.8

Analytical Models of Wear

So far only a few attempts have been made in modelling of dental ceramic wear. The FIB/SEM/TEM analyses of (Arsecularatne et al., 2015) revealed a wear mechanism due to the formation of superficial cracks with ceramic crystals in the antagonist surface acting as asperities (Oh et al., 2002; Theocharopoulos et al., 2013a). Under these conditions, the contact force is assumed to exceed a fracture threshold force and the depth of lateral fracture can be used to estimate the wear volume (Evans and Marshall, 1981). Based on that work, a relation between wear volume V and contact pressure, p, for the leucite glass-ceramic IPS-Empress can be written (Arsecularatne et al., 2015): Equation 11 For the experimental conditions of (Arsecularatne et al., 2015), the relation between V and p as given by Equation 11 is represented by line 1 (Figure 26) in which the symbols represent the experimental wear results. The Archard wear equation (Archard, 1961) describes a linear relation between wear volume and, sliding distance and real area of contact. For the conditions used in (Arsecularatne et al., 2015), the relation between V and p is represented by line 2 (Figure 26). The curve of best fit for the experimental wear data (R2 = 0.934) is represented by curve 3 with corresponding equation: Equation 12 From Figure 26, it can be seen that both the Archard equation and Evans-Marshall equation (Equation 12) underestimate the wear volume considerably, particularly, at high contact pressures. Under these conditions localized delamination also contributed to the wear loss and its influence is not adequately accounted for in the above wear models. There remains a need for the further development of more predictive mechanistic models.

Figure 26. Comparison of Archard and Evans-Marshall wear models with experimental results for leucite glassceramic. Figure reproduced from (Arsecularatne et al., 2015) with permission under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

4

Summary and Outlook

The longevity of dental restorations will inevitably be dependent on their fatigue and wear properties. This review has highlighted some recent advances in our understanding of fatigue and wear processes, and also the ways that their mechanisms are intimately related. For example, while most mechanical properties are poor predictors of wear response, good fracture and fatigue crack growth properties often predict good wear resistance because delamination wear mechanisms are 44

governed by subsurface fatigue crack propagation in both dental composites and ceramics. Additionally, wear facets are common in vivo occlusal origins of ceramic restoration fatigue failures, and the in vitro study of this in vivo failure mechanism requires adding a sliding (wear) motion to indentation fatigue tests. An open question remains: “How well can the in vivo fatigue/wear/secondary caries performance of a restoration be predicted from in vitro testing?” Here it is argued that in vitro testing can only be truly useful if careful attention is payed to matching the in vivo failure mechanisms and observations. Based on the above review, we can identify some areas where our understanding is becoming well developed and useful, and other that need significant further research. For both dental composites and ceramics, we have a good understanding of the mechanisms of toughening composites by crack deflection and bridging, and that that can correlate well to enhanced fatigue crack growth and wear resistance. However, there are only a handful of studies that measure quantitative in vitro fatigue or wear parameters associated with the underlying mechanisms (e.g., using the Paris law, EAC models, wear models, etc.), and there is very little understanding of how those parameters correlate with in vivo outcomes. One example of this is a recent study showing how the EAC threshold appears to be more important that the exponent in predicting in vivo outcomes for various ceramics. In this regard, there needs to be a continued effort to 1) identify the appropriate mechanisms responsible for failures, 2) formulate the appropriate quantitative models to describe the mechanisms, 3) develop in vitro testing methods representative of in vivo conditions and that to extract those quantitative parameters and 4) understand how those parameters correlate with in vivo outcomes. In some cases, we are moving well along in this process. For example, the evolution of in vitro indentation fatigue testing methods for dental ceramics to include sliding (wear) motion for dental ceramics demonstrates an excellent case study for using in vivo mechanism observations to guide the development of progressively more representative in vitro studies. In other cases, we are not very far at all. Quantitative models for wear of dental ceramics and composites are very limited. While we have a reasonably good understanding of fatigue mechanisms for dental composites, the details of in vivo composite fractures, and how to simulate them, are poorly understood. Also, the majority of direct composite restoration failures occur by secondary caries, which is thought to be promoted by wear and interface fatigue mechanisms. Mechanistically, it is thought that wear roughens composite restoration surfaces making biofilm formation easier, and fatigue loading helps create and propagate marginal gaps and aid in the bacterial colonization and demineralization of those gaps by pumping fluid in and out of them. However, mechanistic models are lacking, our current methods of in vitro testing in this area are at their beginning, and appropriate quantitative parameters have not yet been identified to correlate with in vivo outcomes. There is much research that needs to be done. Further mechanistic studies are needed and there are many open questions to be answered about the complex interactions between fatigue, wear, secondary caries, and the complex oral environment.

Acknowledgements JJK and CBT acknowledge support from the University of New South Wales/Sao Paulo Research Foundation (FAPESP) Joint Program Grant # 2017/50290-6. CFP acknowledges support from Sao Paulo Research Foundation (FAPESP) Grant # 2017/11913-8.

45

References Afrashtehfar, K.I., Emami, E., Ahmadi, M., Eilayyan, O., Abi-Nader, S., Tamimi, F., 2017. Failure rate of single-unit restorations on posterior vital teeth: A systematic review. The Journal of prosthetic dentistry 117, 345-353 e348. Aggarwal, V., Logani, A., Jain, V., Shah, N., 2008. Effect of cyclic loading on marginal adaptation and bond strength in direct vs. indirect class II MO composite restorations. Operative dentistry 33, 587-592. Al Kayed, M.A.S., 1999. Clinical Study of Placement and Replacement of Composite restoration in Jordan. The Saudi Dental Journal 11, 53-59. Al Khuraif, A.A., 2014. An in vitro evaluation of wear and surface roughness of particulate filler composite resin after tooth brushing. Acta Odontol Scand 72, 977-983. Albashaireh, Z.S., Ghazal, M., Kern, M., 2010. Two-body wear of different ceramic materials opposed to zirconia ceramic. The Journal of prosthetic dentistry, 104, 105–113. Alvanforoush, N., Palamara, J., Wong, R.H., Burrow, M.F., 2017. Comparison between published clinical success of direct resin composite restorations in vital posterior teeth in 1995-2005 and 2006-2016 periods. Australian dental journal 62, 132-145. Amer, R., Kürklü Kateeb, E., Seghi, R.R., 2014. Three-body wear potential of dental yttriumstabilized zirconia ceramic after grinding, polishing, and glazing treatments. J Prosthet Dent 112:1151-1155. The Journal of prosthetic dentistry 112, 1151-1155. Anderson, T.L., 2005. Fracture Mechanics: Fundamentals and Applications, third ed. ed. CRC, Taylor and Francis Group, Florida. Anil, N., Bolay, S., 2002. Effect of toothbrushing on the material loss, roughness, and colour of intrinsically and extrinsically stained porcelain used in metal-ceramic restorations: an in vitro study The International journal of prosthodontics 15 483-487. Apel, E., Deubener, J., Bernard, A., Holand, M., Muller, R., Kappert, H., Rheinberger, V., Holand, R., 2008. Phenomina and mechanisms of crack propagation in glass ceramics. . Journal of the Mechanical Behaviour of Biomedical Materials 1, 313-325. Archard, J.F., 1961. Single Contacts and Multiple Encounters. J Appl Phys 32, 1420-1425. Arisu, H.D., Uctasli, M.B., Eliguzeloglu, E., Ozcan, S., Omurlu, H., 2008. The effect of occlusal loading on the microleakage of class V restorations. Operative dentistry 33, 135-141. Arola, D., Bajaj, D., Ivancik, J., Majd, H., Zhang, D., 2010. Fatigue of biomaterials: Hard tissues. Int J Fatigue 32, 1400-1412. Arsecularatne, J.A., Chung, N.R., Hoffman, M., 2016. An in vitro study of the wear behaviour of dental composites. Biosurface and Biotribology 2, 102-113. Arsecularatne, J.A., Dingeldein, J.P., Hoffman, M., 2015. An in vitro study of the wear mechanism of a leucite glass dental ceramic. . Biosurface and Biotribology 1 50 – 61. Arsecularatne, J.A., Hoffman, M., 2010. On the Wear Mechanism of Human Dental Enamel Journal of the Mechanical Behaviour of Biomedical Materials 3:, 347-356. Arsecularatne, J.A., Hoffman, M., 2012. Ceramic-like Wear Behaviour of Human Dental Enamel. Journal of the Mechanical Behaviour of Biomedical Materials 8 47-57. Asefi, S., Eskandarion, S., Hamidiaval, S., 2016. Fissure sealant materials: Wear resistance of flowable composite resins. J Dent Res Dent Clin Dent Prospects 10, 194-199. Askar, H., Brouwer, F., Lehmensiek, M., Paris, S., Schwendicke, F., 2017. The association between loading of restorations and secondary caries lesions is moderated by the restoration material elasticity. J. Dent. 58, 74-79. Atali, P.Y., Topbaşi, F.B., 2011. The effect of different bleaching ethods on the surface roughness and hardness of resin composites. Journal of Dentistry and Oral Hygiene 3, 10-17. Attin, T., Opatowski, A., Meyer, C., Zingg-Meyer, B., Buchalla, W., Monting, J.S., 2001. Threeyear follow up assessment of Class II restorations In primary molars with a polyacid-modified composite resin and a hybrid composite. Am. J. Dent. 14, 148-152. 46

Bagheri, R., Tyas, M.J., Burrow, M.F., 2007. Subsurface degradation of resin-based composites. Dent. Mater. 23, 944-951. Bajaj, D., Arola, D.D., 2009. On the R-curve behavior of human tooth enamel. Biomaterials 30, 4037-4046. Bajaj, D., Nazari, A., Eidelman, N., Arola, D.D., 2008. A comparison of fatigue crack growth in human enamel and hydroxyapatite. Biomaterials 29, 4847-4854. Bajaj, D., Sundaram, N., Nazari, A., Arola, D., 2006. Age, dehydration and fatigue crack growth in dentin. Biomaterials 27, 2507-2517. Basso, G.R., Moraes, R.R., Borba, M., Duan, Y., Griggs, J.A., Della Bona, A., 2016. Reliability and failure behavior of CAD-on fixed partial dentures. Dental materials : official publication of the Academy of Dental Materials 32, 624-630. Beck, F., Lettner, S., Graf, A., Bitriol, B., Dumitrescu, N., Bauer, P., Moritz, A., Schedle, A., 2015. Survival of direct resin restorations in posterior teeth within a 19-year period (1996–2015): A metaanalysis of prospective studies. Dental Materials 31, 958-985. Belli, R., Frankenberger, R., Appelt, A., Schmitt, J., Baratieri, L.N., Greil, P., Lohbauer, U., 2013. Thermal-induced residual stresses affect the lifetime of zirconia-veneer crowns. Dental materials : official publication of the Academy of Dental Materials 29, 181-190. Belli, R., Monteiro, S., Jr., Baratieri, L.N., Katte, H., Petschelt, A., Lohbauer, U., 2012. A photoelastic assessment of residual stresses in zirconia-veneer crowns. Journal of dental research 91, 316-320. Belli, R., Petschelt, A., LohbauerLaboratory, U., 2014. Are linear elastic material properties relevant predictors of the cyclic fatigue resistance of dental resin composites? Dental Materials 30, 381-391. Belli, R., Wendler, M., de Ligny, D., Cicconi, M.R., Petschelt, A., Peterlik, H., Lohbauer, U., 2017. Chairside CAD/CAM materials. Part 1: Measurement of elastic constants and microstructural characterization. Dental materials : official publication of the Academy of Dental Materials 33, 8498. Benetti, A.R., Larsen, L., Dowling, A.H., Fleming, G.J.P., 2016. Assessment of wear facets produced by the ACTA wear machine. J. Dent. 45, 19-25. Benetti, P., Kelly, J.R., Sanchez, M., Della Bona, A., 2014. Influence of thermal gradients on stress state of veneered restorations. Dental materials : official publication of the Academy of Dental Materials 30, 554-563. Bernardo, M., Luis, H., Martin, M.D., Leroux, B.G., Rue, T., Leitao, J., DeRouen, T.A., 2007. Survival and reasons for failure of amalgam versus composite posterior restorations placed in a randomized clinical trial. J. Am. Dent. Assoc. 138, 775-783. Bijelic-Donova, J., Garoushi, S., Vallittu, P.K., Lassila, L.V.J., 2016. Mechanical properties, fracture resistance, and fatigue limits of short fiber reinforced dental composite resin. Journal of Prosthetic Dentistry 115, 95-102. Borba, M., Duan, Y., Griggs, J.A., Cesar, P.F., Della Bona, A., 2015. Effect of ceramic infrastructure on the failure behavior and stress distribution of fixed partial dentures. Dental materials : official publication of the Academy of Dental Materials 31, 413-422. Braem, M., Lambrechts, P., Vanherle, G., 1994. Clinical relevance of laboratory fatigue studies. J. Dent. 22, 97-102. Braem, M.J.A., Lambrechts, P., Gladys, S., Vanherle, G., 1995. In vitro fatigue behavior of restorative composites and glass ionomers. Dent. Mater. 11, 137-141. Braga, S.R.M., Tachibana, T.Y., Garone-Netto, N., Sobral, M.A.P., 2011. Abrasion resistance of different resin composites. Journal of the Health Sciences Institute 29, 85-88. Brown, W.S., Jacobs, H.R., Thompson, R.E., 1972. Thermal fatigue in teeth. J. Dent. Res. 51, 461467. Brunthaler, A., Konig, F., Lucas, T., Sperr, W., Schedle, A., 2003. Longevity of direct resin composite restorations in posterior teeth. Clin Oral Investig 7, 63-70. 47

Calais, J.G., Söderholm, K.J.M., 1988. Influence of filler type and water exposure on flexural strength of experimental composite resins J. Dent. Res. 67, 836-840. Campos, P.E., Barceleiro Mde, O., Sampaio-Filho, H.R., Martins, L.R., 2008. Evaluation of the cervical integrity during occlusal loading of Class II restorations. Operative dentistry 33, 59-64. Cesar, P.F., Yoshimura, H.N., Miranda, W.G., Okada, C.Y., 2005. Correlation between fracture toughness and leucite content in dental porcelains. Journal of dentistry 33, 721-729. Chan, K.H.S., Mai, Y., Kim, H., Tong, K.C.T., Ng, D., Hsiao, J.C.M., 2010. Review: Resin Composite Filling. Materials 3, 1228-1243. Chan, K.S., Lee, Y.D., Nicolella, D.P., Furman, B.R., Wellinghoff, S., Rawls, H.R., 2007. Improving fracture toughness of dental nanocomposites by interface engineering and micromechanics. Eng Fract Mech 74, 1857-1871. Chen, I.-W., Liu, S.-Y., 1991. Fatigue of yttria-stabilized zirconia. I. Fatigue damage, fracture origins, and lifetime prediction. J. Am. Ceram. Soc. 74, 1197-1205. Choi, K.K., Condon, J.R., Ferracane, J.L., 2000. The effects of adhesive thickness on polymerization contraction stress of composite. J. Dent. Res. 79, 812-817. Chong, B.J., Thangavel, A.K., Rolton, S.B., Guazzato, M., Klineberg, I.J., 2015. Clinical and laboratory surface finishing procedures for zirconia on opposing human enamel wear: a laboratory study J. Mech. Behav. Biomed. Mater. 50 93–103. J. Mech. Behav. Biomed. Mater. 50 93–103. Coelho, P.G., Bonfante, E.A., Silva, N.R., Rekow, E.D., Thompson, V.P., 2009a. Laboratory simulation of Y-TZP all-ceramic crown clinical failures. Journal of dental research 88, 382-386. Coelho, P.G., Silva, N.R., Bonfante, E.A., Guess, P.C., Rekow, E.D., Thompson, V.P., 2009b. Fatigue testing of two porcelain-zirconia all-ceramic crown systems. Dental materials : official publication of the Academy of Dental Materials 25, 1122-1127. Coldea, A., Swain, M.V., Thiel, N., 2014. Hertzian contact response and damage tolerance of dental ceramics. Journal of the mechanical behavior of biomedical materials 34, 124-133. Collins, C.J., Bryant, R.W., Hodge, K.-L.V., 1998. A clinical evaluation of posterior composite resin restorations: 8-year findings. Journal of dentistry 26, 311-317. Condon, J.R., Ferracane, J.L., 1997. Factors effecting dental composite wear in vitro. J. Biomed. Mater. Res. 38, 303-313. Corazza, P.H., Duan, Y., Kimpara, E.T., Griggs, J.A., Della Bona, A., 2015. Lifetime comparison of Y-TZP/porcelain crowns under different loading conditions. Journal of dentistry 43, 450-457. Cotterell, B., Rice, J.R., 1980. Slightly Curved or Kinked Cracks. Int J Fracture 16, 155-169. Curtis, A.R., Shortall, A.C., Marquis, P.M., Palin, W.M., 2008. Water uptake and strength characteristics of a nanofilled resin-based composite. Journal of dentistry 36, 186-193. D’Arcangelo, C., Vanini, L., Rondoni, G.D., De Angelis, F., 2016. Wear properties of dental ceramics and porcelains compared with human enamel. . J. Prosthet. Dent. 115, 350–355. Da Rosa Rodolpho, P.A., Donassollo, T.A., Cenci, M.S., Loguercio, A.D., Moraes, R.R., Bronkhorst, E.M., Opdam, N.J.M., Demarco, F.F., 2011. 22-Year clinical evaluation of the performance of two posterior composites with different filler characteristics. Dent. Mater. 27, 955963. Dauskardt, R.H., James, M.R., Porter, J.R., Ritchie, R.O., 1992. Cyclic fatigue-crack growth in a SiC-whisker-reinforced alumina ceramic composite: long- and small-crack behavior. J. Am. Ceram. Soc. 75, 759-771. de Paula, A.B., Alonso, R.C.B., de Araújo, G.A.S., Rontani, J.P., Correr-Sobrinho, L., PuppinRontani, R.M., 2015. Influence of chemical degradation and abrasion on surface properties of nanorestorative materials. Brazilian Journal of Oral Sciences 14, 100-105. de Souza, J.A., Dolavale, L.C., Camargo, S.A.D., 2013. Wear Mechanisms of Dental Composite Restorative Materials by Two Different in-vitro Methods. Mater Res-Ibero-Am J 16, 333-340. De Souza, J.A., Goutianos, S., Skovgaard, M., Sørensen, B.F., 2011. Fracture resistance curves and toughening mechanisms in polymer based dental composites. Journal of the mechanical behavior of biomedical materials 4, 558-571. 48

Deligeorgi, V., Mjör, I.A., Wilson, N.H., 2001. An overview of reasons for the placement and replacement of restorations. Prim Dent Care 8, 5-11. Della Bona, A., Corazza, P.H., Zhang, Y., 2014. Characterization of a polymer-infiltrated ceramicnetwork material. Dent. Mater. 30, 564-569. DeLong, R., Douglas, W.H., 1983. Development of an artificial oral environment for the testing of dental restoratives: bi-axial force and movement control. Journal of dental research 62, 32-36. Denry, I.L., 1996. Recent advances in ceramics for dentistry. Crit Rev Oral Biol M 7, 134-143. Diracoglu, D., Alptekin, K., Cifter, E.D., Guclu, B., Karan, A., Aksoy, C., 2011. Relationship between maximal bite force and tooth wear in bruxist and non-bruxist individuals. . Archives of Oral Biology 56, 1569–1575. Douglas, W.H., Sakaguchi, R.L., DeLong, R., 1985. Frictional effects between natural teeth in an artificial mouth. Dental Materials 1, 115-119. Downer, M., Azli, N., Bedi, R., Moles, D., Setchell, D., 1999. Dental Restorations: How long do routine dental restorations last? A systematic review. British dental journal 187, 432-439. Drummond, J.L., Lin, L.H., Al-Turki, L.A., Hurley, R.K., 2009. Fatigue behaviour of dental composite materials. J. Dent. 37, 321-330. Dupriez, N.D., Koeckritz, A.K., Kunzelmann, K.H., 2015. A comparative study of sliding wear of non-metallic dental restorative materials with emphasis on micromechanical wear mechanisms. . J Biomed Mater Res Part B 103, 925–934. Eisenburger, M., Addy, M., 2002. Erosion and attrition of human enamel in vitro Part II: Influence of time and loading. J. Dent. 30, 349-352. El Zhawi, H., Kaizer, M.R., Chughtai, A., Moraes, R.R., Zhang, Y., 2016. Polymer infiltrated ceramic network structures for resistance to fatigue fracture and wear. Dental materials : official publication of the Academy of Dental Materials 32, 1352-1361. Etman, M.K., Woolford, M., Dunne, S., 2008. Quantitative measurement of tooth and ceramic wear: in vivo study. . Int. J. Prosthodontics 21, 245-252. Etman, M.K., Woolford, M., Dunne, S., 2013. Wear properties of dental ceramics in:. Woodhead Publishing, Cambridge, UK. Evans, A.G., Cannon, R.M., 1986. Toughening of brittle solids by martensitic transformations. Acta Metall. 34, 761-800. Evans, A.G., Fuller, E.R., 1974. Crack propagation in ceramic materials under cyclic loading conditions. Metall. Trans. 5, 27-33. Evans, A.G., Marshall, D.B., 1981. Wear mechanisms in ceramics. In: , in: DA, R. (Ed.), Fundamental of friction and wear of materials, ASM, Metals Park, OH, USA, pp. 439-452. Fejerskov, O., 2004. Changing paradigms in concepts on dental caries: consequences for oral health care. Caries Res 38, 182-191. Ferracane, J.L., 2006a. Hygroscopic and hydrolytic effects in dental polymer networks. Dent. Mater. 22, 211-222. Ferracane, J.L., 2006b. Is the wear of dental composites still a clinical concern?: Is there still a need for in vitro wear simulating devices? Dent. Mater. 22, 689-692. Ferracane, J.L., 2011. Resin composite-State of the art. Dent. Mater. 27, 29-38. Ferracane, J.L., 2013. Resin-based co posite perfor ance: Are there so e things we can’t predict? Dent. Mater. 29, 51-58. Ferracane, J.L., 2017. Models of Caries Formation around Dental Composite Restorations. J. Dent. Res. 96, 364-371. Ferracane, J.L., Palin, W.M., 2013. Effects of particulate filler systems on the properties and performance of dental polymer composites. Woodh Publ Ser Biom 53, 294-335. Figueiredo-Pina, C.G., Patas, N., Canhoto, J., Claudio, R., Olhero, S.M., Serro, A.P., Ferro, A.C., Guedes, M., 2016. Tribological behaviour of unveneered and veneered lithium disilicate dental material. . J. Mech. Behav. Biomed. Mater. 53 226–238.

49

Finlay, N., Hahnel, S., Dowling, A.H., Fleming, G.J.P., 2013. The in vitro wear behavior of experimental resin-based composites derived from a commercial formulation. Dent. Mater. 29, 365374. Flinn, B.D., Raigrodski, A.J., Mancl, L.A., Toivola, R., Kuykendall, T., 2017. Influence of aging on flexural strength of translucent zirconia for monolithic restorations. The Journal of prosthetic dentistry 117, 303-309. Fugolin, A.P.P., Pfeifer, C.S., 2017. New Resins for Dental Composites. J. Dent. Res. 96, 10851091. Furmanski, J., Pruitt, L.A., 2007. Peak stress intensity dictates fatigue crack propagation in UHMWPE. Polymer 48, 3512-3519. Ganss, C., Klimek, J., Schwarz, N., 2000. A comparative profilometric in vitro study of the susceptibility of polished and natural human enamel and dentine surfaces to erosive demineralization. . Archives of Oral Biology 45 897-902. Garcia-Godoy, F., Frankenberger, R., Lohbauer, U., Feilzer, A.J., Kramer, N., 2012. Fatigue behavior of dental resin composites: Flexural fatigue in vitro versus 6 years in vivo. J. Biomed. Mater. Res. Part B 100b, 903-910. Garoushi, S., Gargoum, A., Vallittu, P.K., Lassila, L., 2018. Short fiber-reinforced composite restorations: A review of the current literature. J Investig Clin Dent. Garoushi, S., Lassila, L.V.J., Tezvergil, A., Vallittu, P.K., 2007. Static and fatigue compression test for particulate filler composite resin with fiber-reinforced composite substructure. Dental materials : official publication of the Academy of Dental Materials 23, 17-23. Garoushi, S., Sailynoja, E., Vallittu, P.K., Lassila, L., 2013. Physical properties and depth of cure of a new short fiber reinforced composite. Dent. Mater. 29, 835-841. Garvie, R.C., Hannink, R.H., Pascoe, R.T., 1975. Ceramic steel? Nature 258, 703. Garza, L.A., Thompson, G., Cho, S.H., Berzins, D.W., 2016. Effect of toothbrushing on shade and surface roughness of extrinsically stained pressable ceramics. . The Journal of prosthetic dentistry 115 489-494. Gopferich, A., 1996. Mechanisms of polymer degradation and erosion. Biomaterials 17, 103-114. Guazzato, M., Albakry, M., Ringer, S.P., Swain, M.V., 2004. Strength, fracture toughness and microstructure of a selection of all-ceramic materials. Part I. Pressable and alumina glass-infiltrated ceramics. Dental Materials 20, 441-448. Guess, P.C., Zavanelli, R.A., Silva, N.R., Bonfante, E.A., Coelho, P.G., Thompson, V.P., 2010. Monolithic CAD/CAM lithium disilicate versus veneered Y-TZP crowns: comparison of failure modes and reliability after fatigue. Int J Prosthodont 23, 434-442. Guo, J., Li, D., Wang, H., Yang, Y., Wang, L., Guan, D., Qiu, Y., He, L., Zhang, S., 2017. Effect of contact stress on the cycle-dependent wear behaviour of ceramic restoration. J. Mech. Behav. Biomed. Mater. 68, 16–25. Guo, J., Tian, B., Wei, R., Wang, W., Zhang, H., Wu, X., He, L., Zhang, S., 2014. Investigation of the time-dependent wear behaviour of veneering ceramic in porcelain fused to metal crowns during chewing simulations. . J. Mech. Behav. Biomed. Mater. 40, 23–32. Hahnel, S., Schultz, S., Trempler, C., Ach, B., Handel, G., Rosentritt, M., 2011. Two-body wear of dental restorative materials. Journal of the Mechanical Behavior of Biomedical Materials 4, 237244. Hajizadeh, H., Ameri, H., Eslami, S., Mirzaeepoor, B., 2013. The effect of bleaching on toothbrush abrasion of resin composites. J Conserv Dent 16, 17-20. Hamburger, J.T., 2014. Treatment of Severe Tooth Wear: A minimally invasive approach. Radboud Universiteit Nijmegen, Netherlands, p. 119. Han, J.M., Zhang, H.Y., Choe, H.S., Lin, H., Zheng, G., Hong, G., 2014. Abrasive wear and surface roughness of contemporary dental composite resin. Dental materials journal 33, 725-732. Heintze, S.D., 2007. Systematic reviews: I. The correlation between laboratory tests on marginal quality and bond strength. II. The correlation between marginal quality and clinical outcome. Journal of Adhesive Dentistry 9, 77-106. 50

Heintze, S.D., Cavalleri, A., Forjanic, M., Zellweger, G., Rousson, V., 2008. Wear of ceramic and antagonist—A systematic evaluation of influencing factors in vitro Dental Materials 24, 433–449. Heintze, S.D., Faouzi, M., Rousson, V., Ozcan, M., 2012. Correlation of wear in vivo and six laboratory wear methods. Dent. Mater. 28, 961-973. Heintze, S.D., Forjanic, M., 2005. Surface roughness of different dental materials before and after simulated toothbrushing in vitro. Oper Dent 30, 617-626. Heintze, S.D., Forjanic, M., Ohmiti, K., Rousson, V., 2010. Surface deterioration of dental materials after simulated toothbrushing in relation to brushing time and load. Dent. Mater. 26, 306319. Heintze, S.D., Zappini, G., Rousson, V., 2005. Wear of ten dental restorative materials in five wear simulators - Results of a round robin test. Dent. Mater. 21, 304-317. Heintze, S.D., Zellweger, G., Zappini, G., 2007. The relationship between physical parameters and wear of dental composites. Wear 263, 1138-1146. Hertzberg, R.W., Manson, J.A., Skibo, M., 1975. Frequency Sensitivity of Fatigue Processes in Polymeric Solids. Polym Eng Sci 15, 252-260. Hickel, R., 2009. Trends in materials science from the point of view of a practicing dentist. J Eur Ceram Soc 29, 1283-1289. Hickel, R., Kaaden, C., Paschos, E., Buerkle, V., Garcia-Godoy, F., Manhart, J., 2005. Longevity of occlusally-stressed restorations in posterior primary teeth. Am. J. Dent. 18, 198-211. Hmaidouch, R., Weigl, P., 2013. Tooth wear against ceramic crowns in posterior region: a systematic literature review. . International journal of oral science 5, 183–190. Holand, W., Rheinberger, V., Apel, E., Ritzberger, C., Rothbrust, F., Kappert, H., Krumeich, F., Nesper, R., 2009. Future perspectives of biomaterials for dental restoration. J Eur Ceram Soc 29, 1291-1297. Hopp, C.D., Land, M.F., 2013. Considerations for ceramic inlays in posterior teeth: a review. Clinical, Cosmetic and Investigational Dentistry 5, 21–32. Hu, X., Marquis, P.M., Shortall, A.C., 2003. Influence of filler loading on the two-body wear of a dental composite. Journal of Oral Rehabilitation 30, 729-737. Hu, X., Shortall, A.C., Marquis, P.M., 2002. Wear of three dental composites under different testing conditions. Journal of Oral Rehabilitation 29, 756-764. Ilie, N., Hickel, R., 2009. Investigations on mechanical behaviour of dental composites. Clin Oral Invest 13, 427-438. Isgro, G., Wang, H., Kleverlaan, C.J., Feilzer, A.J., 2005. The effects of thermal mismatch and fabrication procedures on the deflection of layered all-ceramic discs. Dental Materials 21, 649-655. Ivancik, J., Arola, D.D., 2013. The importance of microstructural variations on the fracture toughness of human dentin. Biomaterials 34, 864-874. Ivancik, J., Majd, H., Bajaj, D., Romberg, E., Arola, D., 2012. Contributions of aging to the fatigue crack growth resistance of human dentin. Acta Biomaterialia 8, 2737-2746. Ivancik, J., Neerchal, N.K., Romberg, E., Arola, D., 2011. The Reduction in Fatigue Crack Growth Resistance of Dentin with Depth. J. Dent. Res. 90, 1031-1036. Iwasaki, N., Takahashi, H., Koottathape, N., Kanehira, M., Finger, W.J., Sasaki, K., 2014. Texture of composite resins exposed to two- and three-body wear in vitro. The journal of contemporary dental practice 15, 232-241. Janyavula, S., Lawson, N., Cakir, D., Beck, P., Ramp, L.C., Burgess, J.O., 2013. The wear of polished and glazed zirconia against enamel. J. Prosthet. Dent. 109, 22–29. Jin, J., Takahashi, R., Hickel, R., Kunzelmann, K.H., 2014. Surface properties of universal and flowable nanohybrid composites after simulated tooth brushing. Am. J. Dent. 27, 149-154. Johnsen, G.F., Taxt-Lamolle, S.F., Haugen, H.J., 2011. Wear model simulating clinical abrasion on composite filling materials. Dent. Mater. J. 30, 739-748. Jung, Y.-S., Lee, J.-W., Choi, Y.-J., Ahn, J.-S., Shin, S.-W., Huh, J.-B., 2010. A study on the invitro wear of the natural tooth structure by opposing zirconia or dental porcelain. . J Adv Prosthodont 2 1111–1115. 51

Kaidonis, J.A., Richards, L.C., Townsend, G.C., Tansley, G.D., 1998. Wear of human enamel: a quantitative in vitro assessment. . Journal of dental research 77 1983-1990. Kanat, B., Comlekoglu, E.M., Dundar-Comlekoglu, M., Hakan Sen, B., Ozcan, M., Ali Gungor, M., 2014. Effect of various veneering techniques on mechanical strength of computer-controlled zirconia framework designs. Journal of prosthodontics : official journal of the American College of Prosthodontists 23, 445-455. Kanat-Erturk, B., Comlekoglu, E.M., Dundar-Comlekoglu, M., Ozcan, M., Gungor, M.A., 2015. Effect of Veneering Methods on Zirconia Framework-Veneer Ceramic Adhesion and Fracture Resistance of Single Crowns. Journal of prosthodontics : official journal of the American College of Prosthodontists 24, 620-628. Kato, K., Adachi, K., 2002. Wear of advanced ceramics. Wear 253, 1097-1104. Kawakami, Y., Takeshige, F., Hayashi, M., Ebisu, S., 2007. Fatigue of tooth-colored restoratives in aqueous environment. Dent Mater 26, 1-6. Khvostenko, D., Hilton, T.J., Ferracane, J.L., Mitchell, J.C., Kruzic, J.J., 2016. Bioactive glass fillers reduce bacterial penetration into marginal gaps for composite restorations. Dent. Mater. 32, 73-81. Khvostenko, D., Mitchell, J.C., Hilton, T.J., Ferracane, J.L., Kruzic, J.J., 2013. Mechanical performance of novel bioactive glass containing dental restorative composites. Dent. Mater. 29, 1139-1148. Khvostenko, D., Salehi, S., Naleway, S.E., Hilton, T.J., Ferracane, J.L., Mitchell, J.C., Kruzic, J.J., 2015. Cyclic mechanical loading promotes bacterial penetration along composite restoration marginal gaps. Dent. Mater. 31, 702-710. Kidd, E.A.M., Joystonbechal, S., Beighton, D., 1993. Microbiological Validation of Assessments of Caries Activity during Cavity Preparation. Caries Research 27, 402-408. Kidd, E.A.M., Joystonbechal, S., Beighton, D., 1995. Marginal Ditching and Staining as a Predictor of Secondary Caries around Amalgam Restorations - a Clinical and Microbiological Study. J. Dent. Res. 74, 1206-1211. Kim, B., Zhang, Y., Pines, M., Thompson, V.P., 2007a. Fracture of porcelain-veneered structures in fatigue. Journal of dental research 86, 142-146. Kim, J.W., Kim, J.H., Thompson, V.P., Zhang, Y., 2007b. Sliding contact fatigue damage in layered ceramic structures. Journal of dental research 86, 1046-1050. Kim, K.H., Park, J.H., Imai, Y., Kishi, T., 1994. Microfracture mechanisms of dental resin composites containing spherically-shaped filler particles. J Dent Res 73, 499-504. Kim, M.-J., Oh, S.-H., Kim, J.-H., Ju, S.-W., Seo, D.-G., Jun, S.-H., Ahn, J.-S., Ryu, J.-J., 2012. Wear evaluation of the human enamel opposing different Y-TZP dental ceramics and other porcelains. J. Dent. 40, 979-988. Kinney, J.H., Marshall, S.J., Marshall, G.W., 2003. The mechanical properties of human dentin: A critical review and re-evaluation of the dental literature. Critical Reviews in Oral Biology and Medicine 14, 13-29. Kon, M., Kakuta, K., Ogura, H., 2006. Effects of occlusal and brushing forces on wear of composite resins. Dent. Mater. J. 25, 183-194. Kontos, L., Schille, C., Schweizer, E., Geis-Gerstorfer, J., 2013. Influence of surface treatment on the wear of solid zirconia. Acta odontologica Scandinavica 71 482–487. Koottathape, N., Takahashi, H., Iwasaki, N., Kanehira, M., Finger, W.J., 2012. Two- and threebody wear of composite resins. Dent. Mater. 28, 1261-1270. Koottathape, N., Takahashi, H., Iwasaki, N., Kanehira, M., Finger, W.J., 2014. Quantitative wear and wear damage analysis of composite resins in vitro. Journal of the Mechanical Behavior of Biomedical Materials 29, 508-516. Kramer, N., Garcia-Godoy, F., Frankenberger, R., 2005. Evaluation of resin composite materials. Part II: In vivo investigations. Am. J. Dent. 18, 75-81.

52

Krämer, N., Kunzelmann, K.H., Taschner, M., Mehl, A., Garcia-Godoy, F., Frankenberger, R., 2006. Antagonist Enamel Wears More Than Ceramic Inlays. Journal of dental research 85 10971100. Kruzic, J.J., 2009. Predicting Fatigue Failures. Science 325, 156-158. Kruzic, J.J., Cannon, R.M., Ager III, J.W., Ritchie, R.O., 2005a. Fatigue threshold R-curves for predicting reliability of ceramics under cyclic loading. Acta Mater. 53, 2595-2605. Kruzic, J.J., Cannon, R.M., Ritchie, R.O., 2004. Crack size effects on cyclic and monotonic crack growth in polycrystalline alumina: Quantification of the role of grain bridging. J. Am. Ceram. Soc. 87, 93-103. Kruzic, J.J., Cannon, R.M., Ritchie, R.O., 2005b. Effects of moisture on grain-boundary strength, fracture and fatigue properties of alumina. J. Am. Ceram. Soc. 88, 2236-2245. Kruzic, J.J., Nalla, R.K., Kinney, J.H., Ritchie, R.O., 2005c. Mechanistic aspects of in vitro fatiguecrack growth in dentin. Biomaterials 26, 1195-1204. Kruzic, J.J., Ritchie, R.O., 2008. Fatigue of mineralized tissues: Cortical bone and dentin. Journal of the Mechanical Behavior of Biomedical Materials 1, 3-17. Kruzic, J.J., Satet, R.L., Hoffmann, M.J., Cannon, R.M., Ritchie, R.O., 2008. The utility of Rcurves for understanding fracture toughness-strength relations in bridging ceramics. J Am Ceram Soc 91, 1986-1994. Kuper, N.K., Opdam, N.J.M., Bronkhorst, E.M., Ruben, J.L., Huysmans, M.C.D.N.J.M., 2013. Hydrodynamic Flow through Loading and in vitro Secondary Caries Development. J. Dent. Res. 92, 383-387. Kuper, N.K., Opdam, N.J.M., Ruben, J.L., de Soet, J.J., Cenci, M.S., Bronkhorst, E.M., Huysmans, M.C.D.N.J.M., 2014. Gap Size and Wall Lesion Development Next to Composite. J. Dent. Res. 93, 108-113. Kwon, M.S., Oh, S.Y., Cho, S.A., 2015. Two-body wear comparison of zirconia crown, gold crown and enamel against zirconia. J. Mech. Behav. Biomed. Mater. 47, 21–28. Lambrechts, P., Debels, E., Van Landuyt, K., Peumans, M., Van Meerbeek, B., 2006a. How to simulate wear?: Overview of existing methods. Dent. Mater. 22, 693-701. Lambrechts, P., Goovaerts, K., Bharadwaj, D., De Munck, J., Bergmans, L., Peumans, M., Van Meerbeek, B., 2006b. Degradation of tooth structure and restorative materials: A review. Wear 261, 980-986. Lameira, D.P., Buarque e Silva, W.A., Andrade e Silva, F., De Souza, G.M., 2015. Fracture Strength of Aged Monolithic and Bilayer Zirconia-Based Crowns. Biomed Res Int 2015, 418641. Lawn, B.R., 1998. Indentation of ceramics with spheres: A century after Hertz. J Am Ceram Soc 81, 1977-1994. Lawn, B.R., Lee, J.J., 2009. Analysis of fracture and deformation modes in teeth subjected to occlusal loading. Acta Biomater 5, 2213-2221. Lawn, B.R., Wiederhorn, S.M., Roberts, D.E., 1984. Effect of Sliding Friction Forces on the Strength of Brittle Materials. J Mater Sci 19, 2561-2569. Lawson, N., Janyavula, S., Syklawer, S., McLaren, E.A., Burgess, J.O., 2014. Wear of enamel opposing zirconia and lithium disilicate after adjustment, polishing and glazing. . Journal of dental research 42 1586-1591. Lawson, N.C., Burgess, J.O., 2015. Wear of nanofilled dental composites at varying filler concentrations. J. Biomed. Mater. Res. Part B 103, 424-429. Lawson, N.C., Cakir, D., Beck, P., Litaker, M.S., Burgess, J.O., 2012. Characterization of thirdbody media particles and their effect on in vitro composite wear. Dent. Mater. 28, E118-E126. Lazaridou, D., Belli, R., Petschelt, A., Lohbauer, U., 2015. Are resin composites suitable replacements for amalgam? A study of two-body wear. Clin Oral Invest 19, 1485-1492. Lee, A., He, L.H., Lyons, K., Swain, M.V., 2012. Tooth wear and wear investigations in dentistry. Journal of Oral Rehabilitation 39, 217-225. Lee, A., Swain, M.V., He, L., Lyons, K., 2014. Wear behaviour of human enamel against lithium disilicate glass ceramic and type III gold. The Journal of prosthetic dentistry. 53

Lefever, D., Perakis, N., Roig, M., Krejci, I., Ardu, S., 2012. The effect of toothbrushing on surface gloss of resin composites. Am J Dent 25, 54-58. Lewis, R., Dwyer-Joyce, R.S., 2005. Wear of human teeth: a tribological perspective. . Proceedings of the Institution of Mechanical Engineers Part J: J Engineering Tribology 219, 1-18. Li, R.W., Chow, T.W., Matinlinna, J.P., 2014. Ceramic dental biomaterials and CAD/CAM technology: state of the art. J Prosthodont Res 58, 208-216. Loesche, W., Rowan, J., Straffon, L., Loos, P., 1975. Association of Streptococcus mutans with human dental decay. Infection and immunity 11, 1252-1260. Loesche, W.J., 1986. Role of Streptococcus mutans in human dental decay. Microbiological reviews 50, 353-380. Loesche, W.J., 1996. Microbiology of Dental Decay and Periodontal Disease, in: Baron, S. (Ed.), Medical Microbiology. 4th edition. University of Texas Medical Branch at Galveston, Galveston, TX. Lohbauer, U., Belli, R., Ferracane, J.L., 2013. Factors Involved in Mechanical Fatigue Degradation of Dental Resin Composites. J. Dent. Res. 92, 584-591. Lohbauer, U., Frankenberger, R., Kramer, N., Petschelt, A., 2003. Time-dependent strength and fatigue resistance of dental direct restorative materials. J Mater Sci-Mater M 14, 1047-1053. Lohbauer, U., Kramer, N., Petschelt, A., Frankenberger, R., 2008. Correlation of in vitro fatigue data and in vivo clinical performance of a glassceramic material. Dental materials : official publication of the Academy of Dental Materials 24, 39-44. Lohbauer, U., Scherrer, S.S., Della Bona, A., Tholey, M., van Noort, R., Vichi, A., Kelly, J.R., Cesar, P.F., 2017. ADM guidance-Ceramics: all-ceramic multilayer interfaces in dentistry. Dental materials : official publication of the Academy of Dental Materials. Loughran, G.M., Versluis, A., Douglas, W.H., 2005. Evaluation of sub-critical fatigue crack propagation in a restorative composite. Dental materials : official publication of the Academy of Dental Materials 21, 252-261. Lu, H., Lee, Y.K., Oguri, M., Powers, J.M., 2006. Properties of a dental resin composite with a spherical inorganic filler. Oper Dent 31, 734-740. Mahalick, J.A., Knap, F.J., Weiter, E.J., 1971. Occlusal Wear in Prosthodontics. J. Am. Dent. Assoc. 82, 154-159. Mai, Y.-W., Lawn, B.R., 1987. Crack-interface grain bridging as a fracture resistance mechanism in ceramics: II, Theoretical fracture mechanics model. J. Am. Ceram. Soc. 70, 289-294. Mainjot, A., 2016. Dental ceramics: Classification and indications. Esthetics in Dentistry, 251-268. Mainjot, A.K., Najjar, A., Jakubowicz-Kohen, B.D., Sadoun, M.J., 2015. Influence of thermal expansion mismatch on residual stress profile in veneering ceramic layered on zirconia: Measurement by hole-drilling. Dental materials : official publication of the Academy of Dental Materials 31, 1142-1149. Mair, L.H., Stolarski, T.A., Vowles, R.W., Lloyd, C.H., 1996. Wear: mechanisms, manifestations and measurement. Report of a workshop. . Journal of dentistry 24 141-148. Malavasi, C.V., Macedo, E.M., Souza Kda, C., Rego, G.F., Schneider, L.F., Cavalcante, L.M., 2015. Surface Texture and Optical Properties of Self-Adhering Composite Materials after Toothbrush Abrasion. The journal of contemporary dental practice 16, 775-782. Manhart, J., Kunzelmann, K.H., Chen, H.Y., Hickel, R., 2000. Mechanical properties of new composite restorative materials. J. Biomed. Mater. Res. 53, 353-361. Marks, L.A.M., Weerheijm, K.L., van Amerongen, W.E., Groen, H.J., Martens, L.C., 1999. Dyract versus Tytin class II restorations in primary molars: 36 months evaluation. Caries Research 33, 387392. Marsh, P.D., 1995. The Role of Microbiology in Models of Dental Caries. Advances in Dental Research 9, 244-254. McMeeking, R.M., Evans, A.G., 1982. Mechanics of transformation-toughening in brittle materials. J. Am. Ceram. Soc. 65, 242-246. 54

Meira, J.B.C., Reis, B.R., Tanaka, C.B., Ballester, R.Y., Cesar, P.F., Versluis, A., Swain, M.V., 2013. Residual stresses in Y-TZP crowns due to changes in the thermal contraction coefficient of veneers. Dental Materials 29, 594-601. Mesko, M.E., Sarkis-Onofre, R., Cenci, M.S., Opdam, N.J., Loomans, B., Pereira-Cenci, T., 2016. Rehabilitation of severely worn teeth: A systematic review. J Dent 48, 9-15. Michalske, T.A., Freiman, S.W., 1983. A Molecular Mechanism for Stress-Corrosion in Vitreous Silica. J Am Ceram Soc 66, 284-288. Min, J., Zhang, Q., Qiu, X., Zhu, M., Yu, H.Y., Gao, S., 2015. Investigation on the Tribological Behavior and Wear Mechanism of Five Different Veneering Porcelains. . PloS one. Mirmohammadi, H., Kleverlaan, C.J., Aboushelib, M.N., Feilzer, A.J., 2011. Influence of salivary enzymes and alkaline pH environment on fatigue behavior of resin composites. Am. J. Dent. 24, 3136. Mitov, G., Heintze, S.D., Walz, S., Woll, K., Muecklich, F., Pospiech, P., 2012. Wear behaviour of dental Y-TZP ceramic against natural enamel after different finishing procedures. Dental materials 28 909-918. Miyazaki, T., Nakamura, T., Matsumura, H., Ban, S., Kobayashi, T., 2013. Current status of zirconia restoration. Journal of Prosthodontic Research 57 236–261. Mjör, I.A., 1996. Glass-ionomer cement restorations and secondary caries: a preliminary report. Quintessence Int 27, 171-174. Mjör, I.A., Toffenetti, O.F., 2000. Secondary caries: A literature review with case reports. Quintessence Int. 31, 165-179. Mormann, W.H., Stawarczyk, B., Ender, A., Sener, B., Attin, T., Mehl, A., 2013. Wear characteristics of current esthetic dental restorative CAD/CAM materials: two-body wear, gloss retention, roughness and Martens hardness. J. Mech. Behav. Biomed. Mater. 20, 113–125. Mundhe, K., Jain, V., Pruthi, G., Shah, N., 2015. Clinical study to evaluate the wear of natural enamel antagonist to zirconia and metal ceramic crowns. The Journal of prosthetic dentistry 114, 358-363. Nagarajan, V.S., Jahanmir, S., Thompson, V.P., 2004. In vitro contact wear of dental composites. Dent. Mater. 20, 63-71. Nakashima, J., Taira, Y., Sawase, T., 2016. In vitro wear of four ceramic materials and human enamel on enamel antagonist. European journal of oral sciences 124, 295–300. Nalla, R.K., Kruzic, J.J., Kinney, J.H., Ritchie, R.O., 2005. Aspects of in vitro fatigue in human cortical bone: time and cycle dependent crack growth. Biomaterials 26, 2183-2195. Nazari, A., Bajaj, D., Zhang, D., Romberg, E., Arola, D., 2009. Aging and the reduction in fracture toughness of human dentin. Journal of the Mechanical Behavior of Biomedical Materials 2, 550559. Nedeljkovic, I., Teughels, W., De Munck, J., Van Meerbeek, B., Van Landuyt, K.L., 2015. Is secondary caries with composites a material-based problem? Dental materials : official publication of the Academy of Dental Materials 31, e247-277. Oh, W.S., DeLong, R., Anusavice, K., 2002. Factors affecting enamel and ceramic wear: A literature review. The Journal of prosthetic dentistry 87 451-459. Oilo, M., Gjerdet, N.R., 2013. Fractographic analyses of all-ceramic crowns: a study of 27 clinically fractured crowns. Dent Mater 29, e78-84. Oilo, M., Kvam, K., Gjerdet, N.R., 2014. Simulation of clinical fractures for three different allceramic crowns. Eur J Oral Sci 122, 245-250. Oilo, M., Quinn, G.D., 2016. Fracture origins in twenty-two dental alumina crowns. Journal of the mechanical behavior of biomedical materials 53, 93-103. Opdam, N.J.M., Bronkhorst, E.M., Loomans, B.A.C., Huysmans, M.C.D.N.J.M., 2010. 12-year Survival of Composite vs. Amalgam Restorations. Journal of dental research 89, 1063-1067. Osiewicz, M.A., Werner, A., Pytko-Polonczyk, J., Roeters, F.J.M., Kleverlaan, C.J., 2015. Contactand contact-free wear between various resin composites. Dent. Mater. 31, 134-140. 55

Padipatvuthikul, P., Jarad, F.D., Mair, L., 2010. Determination of surface and subsurface fatigue damage in dental composites. Wear 268, 1483-1489. Pallesen, U., Qvist, V., 2003. Composite resin fillings and inlays. An 11-year evaluation. Clin Oral Investig 7, 71-79. Pallesen, U., van Dijken, J.W., 2015. A randomized controlled 30 years follow up of three conventional resin composites in Class II restorations. Dental materials : official publication of the Academy of Dental Materials 31, 1232-1244. Pang, Z., Chughtai, A., Sailer, I., Zhang, Y., 2015. A fractographic study of clinically retrieved zirconia-ceramic and metal-ceramic fixed dental prostheses. Dental materials : official publication of the Academy of Dental Materials 31, 1198-1206. Paris, P.C., Erdogan, F., 1963. A critical analysis of crack propagation laws. J. Basic Eng. 85, 528534. Park, J.-H., Park, S., Lee, K., Yun, K.-D., Lim, H.-P., 2014. Antagonist wear of three CAD/CAM anatomic contour zirconia ceramics. . J. Prosthet. Dent. 111, 20–29. Park, S., Quinn, J.B., Romberg, E., Arola, D., 2008a. On the brittleness of enamel and selected dental materials. Dent Mater. 24 1477–1485. Park, S., Wang, D.H., Zhang, D.S., Romberg, E., Arola, D., 2008b. Mechanical properties of human enamel as a function of age and location in the tooth. . J Mater Sci: Mater Med. Passos, S.P., Torrealba, Y., Major, P., Linke, B., Flores-Mir, C., Nychka, J.A., 2014. In vitro wear behaviour of zirconia opposing enamel: a systematic review. Journal of prosthodontics : official journal of the American College of Prosthodontists 23, 593-601. Peng, Z.X., Rahman, M.I.A., Zhang, Y., Yin, L., 2016. Wear Behaviour of Pressible Lithium Disilicate Glass Ceramic. J Biomed Mater Res B Appl Biomater. 104, 968–978. Pereira, G.K.R., Guilardi, L.F., Dapieve, K.S., Kleverlaan, C.J., Rippe, M.P., Valandro, L.F., 2018. Mechanical reliability, fatigue strength and survival analysis of new polycrystalline translucent zirconia ceramics for monolithic restorations. Journal of the mechanical behavior of biomedical materials 85, 57-65. Perez, C.D., Gonzalez, M.R., Prado, N.A.S., de Miranda, M.S.F., Macedo, M.D., Fernandes, B.M.P., 2012. Restoration of Noncarious Cervical Lesions: When, Why, and How. Int J Dent. Peterson, I.M., Pajares, A., Lawn, B.R., Thompson, V.P., Rekow, E.D., 1998. Mechanical characterization of dental ceramics by hertzian contacts. Journal of dental research 77, 589-602. Peutzfeldt, A., Muhlebach, S., Lussi, A., Flury, S., 2018. Marginal Gap Formation in Approximal "Bulk Fill" Resin Composite Restorations After Artificial Ageing. Oper Dent 43, 180-189. Pinna, R., Usai, P., Filigheddu, E., Garcia-Godoy, F., Milia, E., 2017. The role of adhesive materials and oral biofilm in the failure of adhesive resin restorations. Am. J. Dent. 30, 285-292. Pinto, M.M., Cesar, P.F., Rosa, V., Yoshimura, H.N., 2008. Influence of pH on slow crack growth of dental porcelains. Dental materials : official publication of the Academy of Dental Materials 24, 814-823. Pjetursson, B.E., Sailer, I., Makarov, N.A., Zwahlen, M., Thoma, D.S., 2015. All-ceramic or metalceramic tooth-supported fixed dental prostheses (FDPs)? A systematic review of the survival and complication rates. Part II: Multiple-unit FDPs. Dental materials : official publication of the Academy of Dental Materials 31, 624-639. Pongprueksa, P., Kuphasuk, W., Senawongse, P., 2007. Effect of elastic cavity wall and occlusal loading on microleakage and dentin bond strength. Operative dentistry 32, 466-475. Powers, J.M., Sakaguchi, R.L., 2006. Craig's Restorative Dental Materials. Mosby Elsevier. Prakki, A., Cilli, R., Mondelli, R.F.L., Kalachandra, S., Pereira, J.C., 2005. Influence of pH environment on polymer based dental material properties. J. Dent. 33, 91-98. Preis, V., Behr, M., Handel, G., Schneider-Feyrer, S., Hahnel, S., Rosentritt, M., 2012. Wear performance of dental ceramics after grinding and polishing treatments. . J. Mech. Behav. Biomed. Mater. 10, 13–22. Preis, V., Behr, M., Kolbeck, C., Hahnel, S., Handel, G., Rosentritt, M., 2011. Wear performance of substructure ceramics and veneering porcelains. Dent. Mater. 27, 796–804. 56

Preis, V., Grumser, K., Schneider-Feyrer, S., Behr, M., Rosentritt, M., 2016. Cycle-dependent in vitro wear performance of dental ceramics after clinical surface treatments. . J. Mech. Behav. Biomed. Mater. 53, 49–58. Preussker, S., Klimm, W., Poschmann, M., Koch, R., 2003. Microbial ingrowth around single- and multi-component adhesives studied in vitro. Caries Res 37, 345-351. Ramos, N.d.C., Campos, T.M.B., Paz, I.S.d.L., Machado, J.P.B., Bottino, M.A., Cesar, P.F., Melo, R.M.d., 2016. Microstructure characterization and SCG of newly engineered dental ceramics. Dental Materials 32, 870-878. Randolph, L.D., Palin, W.M., Leloup, G., Leprince, J.G., 2016. Filler characteristics of modern dental resin composites and their influence on physico-mechanical properties. Dental Materials 32, 1586-1599. Reich, S.M., Petschelt, A., Wichmann, M., Frankenberger, R., 2004. Mechanical properties and three-body wear of veneering composites and their matrices. J. Biomed. Mater. Res. Part A 69a, 6569. Reis, A.F., Giannini, M., Lovadino, J.R., Dias, C.T.D., 2002. The effect of six polishing systems on the surface roughness of two packable resin-based composites. Am. J. Dent. 15, 193-197. Ren, L., Janal, M.N., Zhang, Y., 2011. Sliding Contact Fatigue of Graded Zirconia with External Esthetic Glass. Journal of dental research 90, 1116-1121. Rezwani-Kaminski, T., Kamann, W., Gaengler, P., 2002. Secondary caries susceptibility of teeth with long-term performing composite restorations. Journal of oral rehabilitation 29, 1131-1138. Rice, J.R., 1967. Mechanics of crack tip deformation and extension by fatigue, Fatigue Crack Propagation, ASTM STP 415. ASTM, Philadelphia, PA, pp. 247-309. Ritchie, R.O., 1999. Mechanisms of fatigue-crack propagation in ductile and brittle solids. Int. J. Fract. 100, 55-83. Robinson, C., Kirkham, J., Brookes, S.J., Shore, R.C., 1995. Chemistry of Mature Enamel, in: Robinson, C., Kirkham, J., Shore, R.C. (Eds.), Dental Enamel Formation to Destruction. CRC Press, pp. 167-191. Rodrigues, S.A., Ferracane, J.L., Della Bona, A., 2008. Flexural strength and Weibull analysis of a microhybrid and a nanofill composite evaluated by 3- and 4-point bending tests. Dental materials : official publication of the Academy of Dental Materials 24, 426-431. Roque, A.C., Bohner, L.O., de Godoi, A.P., Colucci, V., Corona, S.A., Catirse, A.B., 2015. Surface roughness of composite resins subjected to hydrochloric acid. Brazilian dental journal 26, 268-271. Rues, S., Lenz, J., Turp, J.C., Schweizerhof, K., Schindler, H.J., 2011. Muscle and joint forces under variable equilibrium states of the mandible. Clinical oral investigations 15, 737-747. Sagsoz, O., Ilday, N.O., Sagsoz, N.P., Bayindir, Y.Z., Alsaran, A., 2014. Investigation of Hardness and Wear Behavior of Dental Composite Resins. International Journal of Composite Materials 4, 179-184. Sailer, I., Feher, A., Filser, F., Gauckler, L.J., Luthy, H., Hammerle, C.H., 2007. Five-year clinical results of zirconia frameworks for posterior fixed partial dentures. Int J Prosthodont 20, 383-388. Sailer, I., Gottnerb, J., Kanelb, S., Hammerle, C.H., 2009. Randomized controlled clinical trial of zirconia-ceramic and metal-ceramic posterior fixed dental prostheses: a 3-year follow-up. The International journal of prosthodontics 22, 553-560. Sailer, I., Makarov, N.A., Thoma, D.S., Zwahlen, M., Pjetursson, B.E., 2015. All-ceramic or metalceramic tooth-supported fixed dental prostheses (FDPs)? A systematic review of the survival and complication rates. Part I: Single crowns (SCs). Dental materials : official publication of the Academy of Dental Materials 31, 603-623. Santos, R.L.P., Buciumeanu, M., Silva, F.S., Souza, J.C.M., Nascimento, R.M., Motta, F.V., Henriques, B., 2016. Tribological behaviour of zirconia-reinforced glass–ceramic composites in artificial saliva. . Tribology International ,. Sarkar, N.K., 2000. Internal corrosion in dental composite wear: Its significance and simulation. J. Biomed. Mater. Res. 53, 371-380. 57

Scherrer, S.S., Lohbauer, U., Della Bona, A., Vichi, A., Tholey, M.J., Kelly, J.R., van Noort, R., Cesar, P.F., 2017. ADM guidance-Ceramics: guidance to the use of fractography in failure analysis of brittle materials. Dental materials : official publication of the Academy of Dental Materials. Scherrer, S.S., Quinn, J.B., Quinn, G.D., Wiskott, H.W., 2007. Fractographic ceramic failure analysis using the replica technique. Dental materials : official publication of the Academy of Dental Materials 23, 1397-1404. Schultz, S., Rosentritt, M., Behr, M., Handel, G., 2010. Mechanical properties and three-body wear of dental restoratives and their comparative flowable materials. Quintessence Int 41, e1-10. Shah, M., Ferracane, J.L., Kruzic, J.J., 2009a. R-curve behavior and toughening mechanisms of resin based dental composites: Effects of hydration and post-cure heat treatment. Dent. Mater. 25, 760-770. Shah, M.B., Ferracane, J.L., Kruzic, J.J., 2009b. Mechanistic aspects of fatigue crack growth behavior in resin based dental restorative composites. Dent. Mater. 25, 909-916. Shah, M.B., Ferracane, J.L., Kruzic, J.J., 2009c. R-curve behavior and micromechanisms of fracture in resin based dental restorative composites. Journal of the Mechanical Behavior of Biomedical Materials 2, 502-511. Shinkai, K., Taira, Y., Suzuki, S., Suzuki, M., 2016. In vitro wear of flowable resin composite for posterior restorations. Dent. Mater. J. 35, 37-44. Shortall, A.C., Hu, X.Q., Marquis, P.M., 2002. Potential countersample materials for in vitro simulation wear testing. Dent. Mater. 18, 246-254. Soappman, M., Nazari, A., Porter, J., Arola, D., 2007. A comparison of fatigue crack growth in resin composite, dentin and the interface. Dental materials : official publication of the Academy of Dental Materials 23, 608 - 614. Soares, L.E.S., Cortez, L.R., Zarur, R.D., Martin, A.A., 2012. Scanning Electron Microscopy and Roughness Study of Dental Composite Degradation. Microsc Microanal 18, 289-294. Söderholm, K.J., Zigan, M., Ragan, M., Fischlschweiger, W., Bergman, M., 1984. Hydrolytic degradation of dental composites. J. Dent. Res. 63, 1248-1254. Soderholm, K.J.M., Lambrechts, P., Sarrett, D., Abe, Y., Yang, M.C.K., Labella, R., Yildiz, E., Willems, G., 2001. Clinical wear performance of eight experimental dental composites over three years determined by two measuring methods. Eur J Oral Sci 109, 273-281. Söderholm, K.J.M., Roberts, M.J., 1990. Influence of water exposure on the tensile strength of composites. J. Dent. Res. 69, 1812-1816. Soncini, J.A., Maserejian, N.N., Trachtenberg, F., Tavares, M., Hayes, C., 2007. The. longevity of amalgam versus compomer/composite restorations in posterior primary and permanent teeth Findings from the new England children's amalgam trial. J. Am. Dent. Assoc. 138, 763-772. Soxman, J.A., 2015. Non-nutritive sucking and parafunctional habits, Handbook of Clinical Techniques in Pediatric Dentistry. John Wiley & Sons, Inc, pp. 163-171. Splieth, C., Bernhardt, O., Heinrich, A., Bernhardt, H., Meyer, G., 2003. Anaerobic microflora under Class I and Class II composite and amalgam restorations. Quintessence international 34, 497503. Stawarczyk, B., Keul, C., Eichberger, M., Figge, D., Edelhoff, D., Lumkemann, N., 2017. Three generations of zirconia: From veneered to monolithic. Part I. Quintessence international 48, 369380. Stawarczyk, B., Liebermann, A., Eichberger, M., Guth, J.F., 2015. Evaluation of mechanical and optical behaviour of current esthetic dental restorative CAD/CAM composites. . J. Mech. Behav. Biomed. Mater. 55, 1–11. Stawarczyk, B., Özcan, M., Schmutz, F., Trottmann, A., Roos, M., Hämmerle, C.H.F., 2013. Twobody wear of monolithic, veneered and glazed zirconia and their corresponding enamel antagonists. . Acta odontologica Scandinavica 71, 102-112. Studart, A.R., Filser, F., Kocher, P., Luthy, H., Gauckler, L.J., 2007. Cyclic fatigue in water of veneer-framework composites for all-ceramic dental bridges. Dental Materials 23, 177-185. 58

Sulaiman, T.A., Abdulmajeed, A.A., Donovan, T.E., Cooper, L.F., Walter, R., 2016. Fracture rate of monolithic zirconia restorations up to 5 years: A dental laboratory survey. The Journal of prosthetic dentistry 116, 436-439. Sumino, N., Tsubota, K., Takamizawa, T., Shiratsuchi, K., Miyazaki, M., Latta, M.A., 2013. Comparison of the wear and flexural characteristics of flowable resin composites for posterior lesions. Acta Odontologica Scandinavica 71, 820-827. Suputtamongkol, K., Anusavice, K.J., Suchatlampong, C., Sithiamnuai, P., Tulapornchai, C., 2008. Clinical performance and wear characteristics of veneered lithiadisilicate-based ceramic crowns. . Dent. Mater. 24 667–673. Suputtamongkol, K., Wonglamsam, A., Eiampongpaiboon, T., Malla, S., Anusavice, K.J., 2010. Surface roughness resulting from wear of lithia-disilicate-based posterior crowns. Wear 269 317– 322. Suresh, S., 1998. Fatigue of Materials, 2nd ed. Cambridge University Press, Cambridge. Suzuki, S., Wood, M.B., 2007. Mutual wear evaluation of resin composites when opposing enamel and dentin. J. Biomed. Mater. Res. Part B 81b, 76-81. Suzuki, T., Kyoizumi, H., Finger, W.J., Kanehira, M., Endo, T., Utterodt, A., Hisamitsu, H., Komatsu, M., 2009. Resistance of nanofill and nanohybrid resin composites to toothbrush abrasion with calcium carbonate slurry. Dent. Mater. J. 28, 708-716. Svanberg, M., Mjör, I.A., Orstavik, D., 1990. Mutans Streptococci in Plaque from Margins of Amalgam, Composite, and Glass-Ionomer Restorations. J. Dent. Res. 69, 861-864. Swain, M.V., 2009. Unstable cracking (chipping) of veneering porcelain on all-ceramic dental crowns and fixed partial dentures. Acta Biomater 5, 1668-1677. Swain, M.V., 2014. Impact of oral fluids on dental ceramics: what is the clinical relevance? Dental materials : official publication of the Academy of Dental Materials 30, 33-42. Swain, M.V., Coldea, A., Bilkhair, A., Guess, P.C., 2016. Interpenetrating network ceramic-resin composite dental restorative materials. Dental materials : official publication of the Academy of Dental Materials 32, 34-42. Swanson, P.L., Fairbanks, C.J., Lawn, B.R., Mai, Y.-W., Hockey, B.J., 1987. Crack-interface grain bridging as a fracture resistance mechanism in ceramics: I, Experimental study on alumina. J. Am. Ceram. Soc. 70, 279-289. Takeshige, F., Kawakami, Y., Hayashi, M., Ebisu, S., 2007. Fatigue behavior of resin composites in aqueous environments. Dental materials : official publication of the Academy of Dental Materials 23, 893 - 899. Tamura, Y., Kakuta, K., Ogura, H., 2013. Wear and mechanical properties of composite resins consisting of different filler particles. Odontology 101, 156-169. Taskonak, B., Yan, J., Mecholsky, J.J., Jr., Sertgoz, A., Kocak, A., 2008. Fractographic analyses of zirconia-based fixed partial dentures. Dent Mater 24, 1077-1082. Teixeira, E.C., Thompson, J.L., Piascik, J.R., Thompson, J.Y., 2005. In vitro toothbrush-dentifrice abrasion of two restorative composites. J Esthet Restor Dent 17, 172-180; discussion 181-172. Theocharopoulos, A., Chen, X., Hill, R., Cattell, M.J., 2013a. Reduced wear of enamel with novel fine and nano-scale leucite glass-ceramics. . Journal of dentistry 41, 561-568. Theocharopoulos, A., Chen, X., Wilson, R.M., Hill, R., Cattell, M.J., 2013b. Crystallization of high-strength nano-scale leucite glass-ceramics. Dental Materials 29, 1149-1157. Topcu, F.T., Erdemir, U., Sahinkesen, G., Yildiz, E., Uslan, I., Acikel, C., 2010. Evaluation of Microhardness, Surface Roughness, and Wear Behavior of Different Types of Resin Composites Polymerized With Two Different Light Sources. J. Biomed. Mater. Res. Part B 92b, 470-478. Truong, V.T., Cock, D.J., Padmanathan, N., 1990. Fatigue Crack-Propagation in Posterior Dental Composites and Prediction of Clinical Wear. J Appl Biomater 1, 21-30. Truong, V.T., Tyas, M.J., 1988. Prediction of in vivo wear in posterior composite resins: a fracture mechanics approach. Dental materials : official publication of the Academy of Dental Materials 4, 318-327. 59

Turker, S.B., Biskin, T., 2003. Effect of three bleaching agents on the surface properties of three different esthetic restorative materials. Journal of Prosthetic Dentistry 89, 466-473. Turkun, L.S., Aktener, O., Ates, M., 2003. Clinical evaluation of different posterior resin composite materials: A 7-year report. Quintessence Int. 34, 418-426. Turssi, C.P., Faraoni-Romano, J.J., de Menezes, M., Serra, M.C., 2007. Comparative study of the wear behavior of composites for posterior restorations. J. Mater. Sci.-Mater. Med. 18, 143-147. Turssi, C.P., Ferracane, J.L., Vogel, K., 2005. Filler features and their effects on wear and degree of conversion of particulate dental resin composites. Biomaterials 26, 4932-4937. Turssi, C.P., Hara, A.T., Serra, M.C., Rodrigues, A.L., 2002. Effect of storage media upon the surface micromorphology of resin-based restorative materials. Journal of Oral Rehabilitation 29, 864-871. Turssi, C.P., Purquerio, B.D.M., Serra, M.C., 2003. Wear of dental resin composites: Insights into underlying processes and assessment methods - A review. J. Biomed. Mater. Res. Part B 65b, 280285. Valinoti, A.C., Neves, B.G., da Silva, E.M., Maia, L.C., 2008. Surface degradation of composite resins by acidic medicines and pH-cycling. J Appl Oral Sci 16, 257-265. van Dijken, J.W.V., 2000. Direct resin composite inlays/onlays: an 11 year follow-up. J. Dent. 28, 299-306. van Dijken, J.W.V., Lindberg, A., 2015. A 15-year randomized controlled study of a reduced shrinkage stress resin composite. Dental Materials 31, 1150-1158. Vandewalle, K.S., Ferracane, J.L., Hilton, T.J., Erickson, R.L., Sakaguchi, R.L., 2004. Effect of energy density on properties and marginal integrity of posterior resin composite restorations. Dental materials : official publication of the Academy of Dental Materials 20, 96-106. Velo, M.M.d.A.C., Coelho, L.V.B.F., Basting, R.T., Amaral, F.L.B.d., França, F.M.G., 2016. Longevity of restorations in direct composite resin: literature review. RGO - Revista Gaúcha de Odontologia 64, 320-326. Wang, L., Liu, Y., Si, W., Feng, H., Tao, Y., Ma, Z., 2012. Friction and wear behaviors of dental ceramics against natural tooth enamel. . J Eur Ceram Soc 32 2599–2606. Watts, D.C., 1994. Elastic moduli and viscoelastic relaxation. J. Dent. 22, 154-158. Wendler, M., Belli, R., Petschelt, A., Mevec, D., Harrer, W., Lube, T., Danzer, R., Lohbauer, U., 2017. Chairside CAD/CAM materials. Part 2: Flexural strength testing. Dental materials : official publication of the Academy of Dental Materials 33, 99-109. Wendler, M., Belli, R., Schachtner, M., Amberger, G., Petschelt, A., Fey, T., Lohbauer, U., 2018a. Resistance curves of short-fiber reinforced methacrylate-based biomedical composites. Eng. Fract. Mech. 190, 146-158. Wendler, M., Belli, R., Valladares, D., Petschelt, A., Lohbauer, U., 2018b. Chairside CAD/CAM materials. Part 3: Cyclic fatigue parameters and lifetime predictions. Dental Materials 34, 910-921. Wiegand, A., Credé, A., Tschammler, C., Attin, T., Tauböck, T.T., 2017. Enamel wear by antagonistic restorative materials under erosive conditions. Clinical oral investigations 21, 26892693. Willems, G., Lambrechts, P., Braem, M., Celis, J.P., Vanherle, G., 1992. A Classification of Dental Composites According to Their Morphological and Mechanical Characteristics. Dent. Mater. 8, 310-319. Williams, J.G., 1977. Model of Fatigue Crack Growth in Polymers. J. Mater. Sci. 12, 2525-2533. Wilson, N.H., Burke, F.J., Mjör, I.A., 1997. Reasons for placement and replacement of restorations of direct restorative materials by a selected group of practitioners in the United Kingdom. Quintessence Int 28, 245-248. Wiskott, H.W., Nicholls, J.I., Belser, U.C., 1995. Stress fatigue: basic principles and prosthodontic implications. The International journal of prosthodontics 8, 105-116. Wu, W., Toth, E.E., Moffa, J.F., Ellison, J.A., 1984. Subsurface damage layer of in vivo worn dental composite restorations. J Dent Res 63, 675-680. 60

Wu, Y.Q., Arsecularatne, J.A., Hoffman, M., 2015. Effect of acidity upon attrition-corrosion of human dental enamel. Journal of the Mechanical Behavior of Biomedical Materials 44, 23-34. Yahyazadehfar, M., Ivancik, J., Majd, H., An, B.B., Zhang, D.S., Arola, D., 2014. On the Mechanics of Fatigue and Fracture in Teeth. Appl Mech Rev 66. Yahyazadehfar, M., Zhang, D.S., Arola, D., 2016. On the importance of aging to the crack growth resistance of human enamel. Acta Biomaterialia 32, 264-274. Yoshimura, H.N., Cesar, P.F., Miranda, W.G., Gonzaga, C.C., Okada, C.Y., Goldenstein, H., 2005. Fracture toughness of dental porcelains evaluated by IF, SCF, and SEPB methods. J Am Ceram Soc 88, 1680-1683. Zandparsa, R., El Huni, R.M., Hirayama, H., Johnson, M.I., 2016. Effect of different dental ceramic systems on the wear of human enamel: An in vitro study. . The Journal of prosthetic dentistry 115 230-237. Zhang, Y., Kim, J.W., 2009. Graded structures for damage resistant and aesthetic all-ceramic restorations. Dental materials : official publication of the Academy of Dental Materials 25, 781790. Zhang, Y., Sailer, I., Lawn, B.R., 2013. Fatigue of dental ceramics. Journal of dentistry 41, 11351147. Zhang, Y., Song, J.K., Lawn, B.R., 2005. Deep-penetrating conical cracks in brittle layers from hydraulic cyclic contact. J Biomed Mater Res B 73b, 186-193. Zhang, Y., Xu, D.X., Rao, P.G., Lu, M., Wu, J.Q., 2008. Friction behavior of dental porcelain with different leucite particle sizes. J Am Ceram Soc 91, 1678-1681. Zhang, Z.H., Beitzel, D., Majd, H., Mutluay, M., Tezvergil-Mutluay, A., Tay, F.R., Pashley, D.H., Arola, D., 2016. Effect of carbodiimide on the fatigue crack growth resistance of resin-dentin bonds. Dent. Mater. 32, 211-222. Zhang, Z.H., Beitzel, D., Majd, H., Mutluay, M., Tezvergil-Mutluay, A., Tay, F.R., Pashley, D.H., Arola, D., 2017. Fatigue resistance of dentin bonds prepared with two- vs. three-step adhesives: Effect of carbodiimide. Dent. Mater. 33, 1340-1350. Zhang, Z.H., Beitzel, D., Mutluay, M., Tay, F.R., Pashley, D.H., Arola, D., 2015. On the durability of resin-dentin bonds: Identifying the weakest links. Dent. Mater. 31, 1109-1118. Zheng, J., Zhou, Z.R., 2006. Effect of age on the friction and wear behaviours of human teeth. Tribology International 39 278-284. Zhi, L., Bortolotto, T., Krejci, I., 2016. Comparative in vitro wear resistance of CAD/CAM composite resin and ceramic materials. . The Journal of prosthetic dentistry 115, 199-202. Zhou, Z.R., Zheng, J., 2008. Tribology of dental materials: A review. . Journal of Physics D: Applied Physics 41 1-22. Zimmerli, B., Strub, M., Jeger, F., Stadler, O., Lussi, A., 2010. Composite materials: composition, properties and clinical applications. A literature review. Schweiz Monatsschr Zahnmed 120, 972986. Zivkovic, S., Bojovic, S., Pavlica, D., 2001. Bacterial penetration of restored cavities. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 91, 353-358.

61

Highlights:     

Dental restorations fail due to fatigue and wear in the complex oral environment Material microstructure plays key role in restoration failure by fatigue and wear The mechanistic processes of fatigue and wear are interrelated Restoration failure by secondary caries is influenced by fatigue and wear mechanisms Understanding the mechanistic processes of failure links in vitro and in vivo studies

62